[NSE] rpcinfo - handling of an initial receive_bytes() causes not very graceful exit

[jah <jah () zadkiel plus com>]

Hi folks,

If rpcinfo.nse sends its rpc dump packet to a portmapper which then FIN
ACKs the connection then the initial receive_bytes(), which is wrapped
in a try(), will result in an error being thrown:

NSE: rpcinfo against X.X.194.130:111 threw an error!
EOF
stack traceback:
    [C]: in function 'try'
    C:\Program Files\Nmap\scripts\rpcinfo.nse:55: in function
<C:\Program Files\Nmap\scripts\rpcinfo.nse:41>
    (tail call): ?

Similarly, for any other error, e.g. TIMEOUT.  This is OK, but do we
really need the traceback?  I suggest a more graceful exit of the script:

NSE: rpcinfo failed to receive a response from X.X.194.130:111 with
error: EOF

and have attached a patch.

Regards,

jah

--- rpcinfo.nse.orig    2009-08-06 11:48:54.531250000 +0100
+++ rpcinfo.nse 2009-08-06 11:51:29.328125000 +0100
@@ -52,7 +52,15 @@
   local request = bin.pack('>IIIIIIILL',0x80000028,transaction_id,0,2,100000,2,4,0,0)
   try(socket:send(request))
 
-  local answer = try(socket:receive_bytes(1))
+  local status, answer = socket:receive_bytes(1)
+  if not status then
+    stdnse.print_debug(1, "%s failed to receive a response from %s:%d with error: %s",
+      filename:match( "[\\/]([^\\/]+)\.nse$" ) or filename,
+      host.ip, port.number,
+      answer or "unknown")
+    socket:close()
+    return nil
+  end
 
   local _,offset,header,length,tx_id,msg_type,reply_state,accept_state,value,payload,last_fragment
   last_fragment = false; offset = 1; payload = ''


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