Re: 3.90 Segmentation Fault ;-(

[Paul Tarjan <ptarjan () gmail com>]

Well, this is the same thing I was working on before. If you take out
the -O2, it should work.. also, the -fno-aliasing directive should
also stop the problem.

Libpcap has some errors on GCC4.0. It builds a linked list with a null
node that causes a segfault when it is being dereferenced.

I'm looking into it. Try these directives and it should help. (I
thought Fyodor put them in the config file.. I'll check on it when I'm
back from Camping)

Paul

On 9/8/05, Craig Humphrey <Craig.Humphrey () chapmantripp com> wrote:
> OK, two issues.  Fedora Core 4, just did a yum upgrade (though was lazy
> and didn't reboot).
> Gcc (GCC) 4.0.1 20050727 (Red Hat 4.0.1-5)
> GNU gdb Red Hat Linux (6.3.0.0-1.21rh)
>
> > From configure:
> configure: WARNING: linux/netfilter_ipv4/ipchains_core.h: present but
> cannot be compiled
> configure: WARNING: linux/netfilter_ipv4/ipchains_core.h:     check for
> missing prerequisite headers?
> configure: WARNING: linux/netfilter_ipv4/ipchains_core.h: see the
> Autoconf documentation
> configure: WARNING: linux/netfilter_ipv4/ipchains_core.h:     section
> "Present But Cannot Be Compiled"
> configure: WARNING: linux/netfilter_ipv4/ipchains_core.h: proceeding
> with the preprocessor's result
> configure: WARNING: linux/netfilter_ipv4/ipchains_core.h: in the future,
> the compiler will take recedence
> configure: WARNING:     ## ------------------------------------------ ##
> configure: WARNING:     ## Report this to the AC_PACKAGE_NAME lists.  ##
> configure: WARNING:     ## ------------------------------------------ ##
>
> But that's probably not a biggie.
>
> The main problem is the segfault:
>
> Fired up nmap in emacs/gdb:
> (gdb) run -sP 10.9.9.1
> warning: cannot close "shared object read from target memory": File in
> wrong format
> Starting program: /home/craigh/nmap-3.90/nmap -sP 10.9.9.1
> Reading symbols from shared object read from target memory...done.
> Loaded system supplied DSO at 0xac5000
>
> Starting nmap 3.90 ( http://www.insecure.org/nmap/ ) at 2005-09-09 10:36
> NZST
> Warning: File ./nmap-services exists, but Nmap is using
> /usr/local/share/nmap/nmap-services for security and consistency
> reasons.  set NMAPDIR=. to gi\
> ve priority to files in your local directory (may affect the other data
> files too).
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x08090cda in convert_code_r (p=0x9072f20) at ./optimize.c:2064
> Current language:  auto; currently c
>
> Line 2064 in optimize.c:  off = JF(p)->offset - (p->offset + slen) - 1;
> (gdb) p offset
> $1 = (struct slist **) 0x90730f0
> (gdb) p slen
> $2 = 1
> (gdb) p p
> $3 = (struct block *) 0x9072f20
> (gdb) p p->offset + slen
> $4 = 3
> (gdb) p JF
> No symbol "JF" in current context.
>
> Umm to my uneducated (well, it's been a while since I had to debug C
> code) mind, that looks to be the problem.  Is JF (and JT) some global or
> macro?  They both report "No symbol ... In current context".
>
> Hope that helps someone.
>
> Later'ish
> Craig
>
>
> _______________________________________________
> Sent through the nmap-dev mailing list
> http://cgi.insecure.org/mailman/listinfo/nmap-dev


_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev