Re: [PATCH]in big-endian platform, nmap failed to resolve the service name

[David Fifield <david () bamsoftware com>]

On Tue, Sep 25, 2012 at 12:01:58PM +0800, hejianet wrote:
> in bigendian platform, nmap failed to get the mapping relationship because 
> all protocols are stored in protocol_table[0].
>
> [root@host nmap-6.01]# ./nmap -sO localhost
>
> Starting Nmap 6.01 ( http://nmap.org ) at 2012-09-24 20:41 CDT
> Nmap scan report for localhost (127.0.0.1)
> Host is up (0.000016s latency).
> Other addresses for localhost (not scanned): 127.0.0.1
> Not shown: 249 closed protocols
> PROTOCOL STATE         SERVICE
> 1        open          unknown
> 2        open|filtered unknown
> 6        open          unknown
> 17       open          unknown
> 103      open|filtered unknown
> 136      open|filtered unknown
> 255      open|filtered unknown
>
> Nmap done: 1 IP address (1 host up) scanned in 1.60 seconds
> [root@host nmap-6.01]# uname -m
> ppc64
>
> Signed-off-by: Jia<hejianet () linux vnet ibm com>
> Tested-by: Jia<hejianet () linux vnet ibm com> 
> --- nmap-6.01/protocols.cc.ori        2012-09-24 20:42:45.019985471 -0500
> +++ nmap-6.01/protocols.cc    2012-09-24 20:44:01.130938317 -0500
> @@ -145,7 +145,7 @@ static int nmap_protocols_init() {
>      protno = htons(protno);
>  
>      /* Now we make sure our protocols don't have duplicates */
> -    for(current = protocol_table[0], previous = NULL;
> +    for(current = protocol_table[protno%PROTOCOL_TABLE_SIZE], previous = NULL;
>       current; current = current->next) {
>        if (protno == current->protoent->p_proto) {
>       if (o.debugging) {
> @@ -164,7 +164,7 @@ static int nmap_protocols_init() {
>      current->protoent = (struct protoent *) cp_alloc(sizeof(struct protoent));
>      current->next = NULL;
>      if (previous == NULL) {
> -      protocol_table[protno] = current;
> +      protocol_table[protno%PROTOCOL_TABLE_SIZE] = current;
>      } else {
>        previous->next = current;
>      }
>
> test results:
> PROTOCOL STATE         SERVICE
> 1        open          icmp
> 2        open|filtered igmp
> 6        open          tcp
> 17       open          udp
> 103      open|filtered pim
> 136      open|filtered udplite
> 255      open|filtered unknown

Thank you for finding this. I looked into the problem and decided that
there was more wrong than just the hard-coded insertion into bucket 0.
The bigger problem was the call
        protno = htons(protno);
which mapped byte values into integers like 0x0100, 0x0200, 0x0300, ....
This caused all protocols to end up in bucket 0, even with your fix. I
committed a more extensive revision in r29859.

David Fifield
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/