Re: UCD-4.2.2 and UCD-4.2.3 snmptrapd verification

[Wes Hardaker <wes () hardakers net>]

> > > > > On Mon, 18 Feb 2002 07:58:27 -0500, KF <dotslash () snosoft com> said:

dotslash> I am using -o with snmptrapd to send outout to file... no
dotslash> other options.

Yep, and I've tried that.  I can only reproduce it on a ppc machine
with a really old glibc and it's dying within a call to vsnprintf that
otherwise looks perfectly correct (hence my theory).

Can you send me the packet you're using to cause the crash (though
I've tried the entire test suite without problems on the other various
machines).

dotslash> Program terminated with signal 11, Segmentation fault.
dotslash> #0  0x0fe19090 in strlen () from /lib/libc.so.6
dotslash> 721
dotslash> 722     #ifdef notused
dotslash> 723         in_addr_t myaddr;
dotslash> 724         oid src[MAX_OID_LEN], dst[MAX_OID_LEN], context[MAX_OID_LEN];
dotslash> 725         int srclen, dstlen, contextlen;
dotslash> 726         char ctmp[300];
dotslash> 727     #endif
dotslash> 728
dotslash> 729         /* register our configuration handlers now so -H properly
dotslash> displays them */
dotslash> 730         register_config_handler("snmptrapd", "traphandle",
dotslash> #0  0x0fe19090 in strlen () from /lib/libc.so.6
dotslash> #1  0x0fde8bfc in vfprintf () from /lib/libc.so.6

That's not a proper listing dump, by the way.  It doesn't have the
source for the place it broke (strlen) so what it's showing you is
somewhat random.

dotslash> (gdb) bt
dotslash> #0  0x0fe19090 in strlen () from /lib/libc.so.6
dotslash> #1  0x0fde8bfc in vfprintf () from /lib/libc.so.6
dotslash> #2  0x0fe09220 in vsnprintf () from /lib/libc.so.6

Right.  It's dying in a valid (most likely) call to vsnprintf.

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."