RE: [Ntop-dev] BUG: build on FC4

Burton Strauss Wed, 03 Aug 2005 08:13:33 -0700

Nope, that's pretty clear WHAT is wrong ... myGlobals.runningPrefs.enableSessionHandling isn't readable.

What makes no sense is that none of this is pointer based ...

typedef struct ntopGlobals {

...

/* User-configurable parameters via the command line and the web page. */
UserPref savedPref; /* this is what is saved */
UserPref runningPref; /* this is what is currently used */
...

} NtopGlobals;

with:

typedef struct _userPref {
...

bool enableSessionHandling; /* -z | --disable-sessions */

...
} UserPref;

With myGlobals static in globals-core.c:

NtopGlobals myGlobals;

I wonder if - in the interest of 'performance' or some such, RH has chosen not to initialize memory? But that wouldn't explain why myGlobals wasn't zeroed before the failing line, even so. All that actually happened before the memset() in 3.1 was the command line pre-processing.

All I can say is that there's a ton of changes in that area in the cvs, related to the single state variable enhancement and to try that version.

For example, the main() code in the cvs now begins thusly:

#ifdef WIN32
int ntop_main(int argc, char *argv[]) {
#else
int main(int argc, char *argv[]) {
#endif
int i, rc, userSpecified;

...
char main_buf[LEN_GENERAL_WORK_BUFFER],
       lib[LEN_GENERAL_WORK_BUFFER],
       env[LEN_GENERAL_WORK_BUFFER],
       buf[LEN_GENERAL_WORK_BUFFER];

/* printf("Wait please: ntop is coming up...\n"); */

#ifdef MTRACE
mtrace();
#endif

#ifdef MEMORY_DEBUG
initLeaks(); /* Don't move this below nor above */
#endif

/* VERY FIRST THING is to clear myGlobals, so myGlobals.ntopRunState can be used */
memset(&myGlobals, 0, sizeof(myGlobals));
setRunState(FLAG_NTOPSTATE_PREINIT);

...

Anyway, back to my earlier comments:

(1) We don't support other people's RPMs. Just no way to know what version it is, what patches, etc.

(3) Use the source and see if it still fails.

-----Burton

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Randy Gordey
Sent: Wednesday, August 03, 2005 9:32 AM
To: [email protected]
Subject: RE: [Ntop-dev] BUG: build on FC4

#wget http://easynews.dl.sourceforge.net/sourceforge/ntop/ntop-3.1.tgz

#tar -xvzf ntop-3.1.tgz

#cd ntop

#./configure (I can attach config.log if you think it will help it didn’t complain about anything other than xmldump plugin)

#make

#make install

[EMAIL PROTECTED] ~]# gdb ntop

GNU gdb Red Hat Linux (6.3.0.0-1.21rh)

GDB is free software, covered by the GNU General Public License, and you are

welcome to change it and/or distribute copies of it under certain conditions.

Type "show copying" to see the conditions.

There is absolutely no warranty for GDB. Type "show warranty" for details.

This GDB was configured as "i386-redhat-linux-gnu"...Using host libthread_db library "/lib/libthread_db.so.1".

(gdb) set args -u root -K

(gdb) r

Starting program: /usr/local/bin/ntop -u root -K

Reading symbols from shared object read from target memory...done.

Loaded system supplied DSO at 0xc08000

[Thread debugging using libthread_db enabled]

[New Thread -1208338752 (LWP 32287)]

Wed Aug 3 10:06:18 2005 Initializing gdbm databases

Wed Aug 3 10:06:18 2005 ntop v.3.1 MT (SSL)

Wed Aug 3 10:06:18 2005 Configured on Aug 3 2005 10:02:06, built on Aug 3 2005 10:03:25.

Wed Aug 3 10:06:18 2005 Get the freshest ntop from http://www.ntop.org/

Wed Aug 3 10:06:18 2005 Initializing ntop

Wed Aug 3 10:06:18 2005 Checking eth0 for additional devices

Wed Aug 3 10:06:18 2005 Added virtual interface: 'eth0:1'

Wed Aug 3 10:06:18 2005 Resetting traffic statistics for device eth0

Wed Aug 3 10:06:18 2005 DLT: Device 0 [eth0] is 1, mtu 1514, header 14

Program received signal SIGSEGV, Segmentation fault.

[Switching to Thread -1208338752 (LWP 32287)]

initNtop (devices=Cannot access memory at address 0x8

) at globals-core.c:602

602 if(myGlobals.runningPref.enableSessionHandling)

(gdb) list

597 */

598 if (myGlobals.capturePackets == FLAG_NTOPSTATE_RUN) {

599 initDevices(devices);

600 }

601

602 if(myGlobals.runningPref.enableSessionHandling)

603 initPassiveSessions();

604

605 /* ********************************** */

606

(gdb) frame 5

Cannot access memory at address 0x4

(gdb) bt full

#0 initNtop (devices=Cannot access memory at address 0x8

) at globals-core.c:602

value = Cannot access memory at address 0xffffffd0

(gdb)

Burton, what else would you like to see and I would be happy to try that too.

Kind Regards

_______________________________________________
Ntop-dev mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

RE: [Ntop-dev] BUG: build on FC4

Reply via email to