Thank you for your help, but unfortunately it didnt solve the problem. I put
HANDLER_CAN_RONLY (HANDLER_CAN_READ causes compile-time error) but there is
still segmentation fault.
There are two things that are maybe interesting to mention:
1) Segmentation fault arises even if I avoid handlers, for instance this
code snippet also suffers from the same problem:
void sm_sessions_register()
{
int retval;
retval = netsnmp_register_int_instance("Example1", // --> THIS IS
LINE 479!
session_oid,
OID_LENGTH(session_oid),
&demo_int, NULL);
if(retval == MIB_REGISTERED_OK)
snmp_log(LOG_INFO,"Variable registered\n");
retval = netsnmp_register_int_instance("Example2",
session_oid2,
OID_LENGTH(session_oid2),
&demo_int2, NULL);
if(retval == MIB_REGISTERED_OK)
snmp_log(LOG_INFO,"Variable2 registered\n");
}
Demo_int and demo_int2 are globals. Backtrace is now more detailed because I
have included CFLAGS=-ggdb before running ./configure of net-snmp:
(gdb) bt
#0 0x00a75b49 in snmp_resend_request (slp=0xb08e1ff0, rp=0xab63dfcc,
incr_retries=1) at snmp_api.c:6137
#1 0x00a75e62 in snmp_sess_timeout (sessp=0xb08e1ff0) at snmp_api.c:6224
#2 0x00a75568 in snmp_timeout () at snmp_api.c:6018
#3 0x00a465ee in snmp_synch_response_cb (ss=0xb089df08, pdu=0xb0855f70,
response=0xacb20198, pcb=0x35eaac <agentx_synch_input>)
at snmp_client.c:1020
#4 0x0035ed55 in agentx_synch_response (ss=0xb089df08, pdu=0xb0855f70,
response=0xacb20198) at mibgroup/agentx/client.c:112
#5 0x0035f410 in agentx_register (ss=0xb089df08, start=0xb0837fdc,
startlen=9, priority=127, range_subid=0, range_ubound=0, timeout=0,
flags=1 '\001', contextName=0x0) at mibgroup/agentx/client.c:220
#6 0x0034c94a in agentx_registration_callback (majorID=1, minorID=1,
serverarg=0xacb20288, clientarg=0xb089df08)
at mibgroup/agentx/subagent.c:624
#7 0x00a97ff7 in snmp_call_callbacks (major=1, minor=1,
caller_arg=0xacb20288)
at callback.c:329
#8 0x0033ff7b in netsnmp_register_mib (moduleName=0xb0835ff0 "Example1",
var=0x0, varsize=0, numvars=0, mibloc=0xb0837fdc, mibloclen=9,
priority=127, range_subid=0, range_ubound=0, ss=0x0, context=0x0,
timeout=0, flags=1, reginfo=0xb0833fd0, perform_callback=1)
at agent_registry.c:731
#9 0x003464b0 in netsnmp_register_handler (reginfo=0xb0833fd0)
---Type <return> to continue, or q <return> to quit---
at agent_handler.c:255
#10 0x002f8191 in netsnmp_register_serialize (reginfo=0xb0833fd0)
at serialize.c:43
#11 0x002f3000 in netsnmp_register_instance (reginfo=0xb0833fd0)
at instance.c:78
#12 0x002f39f2 in netsnmp_register_int_instance (
name=0x8094a4c "Session state", reg_oid=0x80aa7a0, reg_oid_len=9,
it=0x80aa7c4, subhandler=0) at instance.c:469
#13 0x0806897f in sm_sessions_register () at sm.c:479
#14 0x0806e00f in sm_main_thread (data=0xbfdae4a8) at sm.c:7778
#15 0x00c1a8ef in g_thread_create_full () from /usr/lib/libglib-2.0.so.0
#16 0x001c43b6 in start_thread () from /lib/libpthread.so.0
#17 0x008af33e in clone () from /lib/libc.so.6
I am also confused why is function snmp_resend_request() even been called..
2) I have mentioned that I get segmentation fault occasionally, not always.
The circumstances are specific:
- my program registers several variables and then I do few snmpwalks
(program is still active, it acts as a agentx subagent)
- after that, my program tries to register few more variables (see the code
above) and THEN I get segmentation fault
Again, I would highly appreciate any help.
Miljen
2008/7/11 Syed Mujtaba Ahmed Razvi <[EMAIL PROTECTED]>:
>
>
> Miljen, I think your code snippet is trying to access a memory location
> that it is not allowed to access. For e.g. you might be writing some illegal
> location using some pointer (may be one of them my_test,my_test2). Try
> changing the HANDLER_CAN_RWRITE in the declaration portion to
> HANDLER_CAN_READ (in both). Your pointers might be trying to access
> eachothers location.
>
>
>
> Hope it helps!
>
> Regards,
>
> Syed Mujtaba
>
> ------------------------------
> Date: Fri, 11 Jul 2008 15:52:31 +0200
> From: [EMAIL PROTECTED]
> To: net-snmp-users@lists.sourceforge.net
> Subject: Segmentation fault problem
>
>
> Hi everyone,
>
> I am trying to add SNMP support in the IKEv2 project (
> http://sourceforge.net/projects/ikev2) and I am using net-snmp, 5.4.1.
>
> Firstly, there is a great number of variables which I plan to register. I
> am wondering whether is possible (or harmful) to use one oid variable as:
> oid session_oid[] = {.., 13, 1};
> and then change last element of this array (to 2,3,..etc.) before
> registering different variables? The alternative is to declare and use
> different oid array for each variable that I want to register.
>
> More importantly, there is a problem which gives me a headache - a
> segmentation fault occasionally happens. Here is a backtrace:
> (gdb) bt
> #0 0x080fc3e1 in snmp_sess_timeout ()
> #1 0x080fc579 in snmp_timeout ()
> #2 0x080d8cb5 in snmp_synch_response_cb ()
> #3 0x080c9f57 in agentx_synch_response ()
> #4 0x080ca702 in agentx_register ()
> #5 0x080ba735 in agentx_registration_callback ()
> #6 0x0811c0c9 in snmp_call_callbacks ()
> #7 0x080ab5b2 in netsnmp_register_mib ()
> #8 0x080b04cd in netsnmp_register_handler ()
> #9 0x080793aa in sm_sessions_register () at sm.c:441
> #10 0x0807e93f in sm_main_thread (data=0xbfd33c28) at sm.c:7778
> #11 0x00c1a8ef in g_thread_create_full () from /usr/lib/libglib-2.0.so.0
> #12 0x001b53b6 in start_thread () from /lib/libpthread.so.0
> #13 0x00b2933e in clone () from /lib/libc.so.6
>
> Function sm_sessions_register() is defined as:
>
> void sm_sessions_register()
> {
> netsnmp_handler_registration *my_test;
> netsnmp_handler_registration *my_test2;
>
> my_test =
> netsnmp_create_handler_registration("example",
> sm_sess_handle,
> session_oid,
> OID_LENGTH(session_oid),
> HANDLER_CAN_RWRITE);
>
> netsnmp_register_instance(my_test); // --> THIS IS LINE 441!
>
> my_test2 =
> netsnmp_create_handler_registration("example2",
> sm_sess_handle,
> session_oid2,
> OID_LENGTH(session_oid2),
> HANDLER_CAN_RWRITE);
>
> netsnmp_register_instance(my_test2);
> }
>
> session_oid and session_oid2 are defined as:
>
> oid session_oid[] = { 1, 3, 6, 1, 3, 411249, 13, 2, 1 };
> oid session_oid2[] = { 1, 3, 6, 1, 3, 411249, 13, 2, 2 };
>
> and they are globals. Sm_sess_handle is written by example in your tutorial
> (delayed_instance.c) and I can include this code also if necessary.
>
> I would be very grateful for any help or hint!
>
> Miljen
>
> ------------------------------
> Connect to the next generation of MSN Messenger Get it now!
> <http://imagine-msn.com/messenger/launch80/default.aspx?locale=en-us&source=wlmailtagline>
>
-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
_______________________________________________
Net-snmp-users mailing list
Net-snmp-users@lists.sourceforge.net
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users
