On 21 Mar 2002, Michel Arboi wrote:
> Vincent Renardias <[EMAIL PROTECTED]> writes:
>
> > - The find_services.nes plugin timeouts on almost every host I tried it on
> > (even when raising the timeout for this plugin to 10 minutes). Is this a
> > known bug?
>
> Maybe you could trigger the logic bomb in read_stream_connection()
> /*
> * This piece of code is not perfect: we may still hang on some
> * vicious server
> * We could use the alarm+longjmp trick (see the open function) but
> * I really dislike it
> */
>
> IIRC, this only happens on a buggy (or vicious) _SSL_ server. A
> discard service is not a candidate for this.
After compiling find_service.nes with DEBUG activated, it seems to hang
after (or while) probing the last port:
Extract from the nessusd.dump:
Port 21 is open. "Transport" is 1
Port 22 is open. "Transport" is 1
[...18 more open ports...]
Could not read anything from port 33111
check_for_socks_proxy: port=33111
Port 54964 is open. "Transport" is 1
[...find_service.nes killed here after a 500 seconds timeout...]
Any clue or point I should look at ?
Another thing:
I set 'plugin_timeout=500' in the client, and I added the line
'plug_set_timeout(desc, 1000);' in the plugin_init() function of
find_service.c, but the timeout of the plugin remain 500 seconds, not
1000. Is this the expected behavior ? I expected the plugin specific
timeout to override the global timeout, not the contrary...
Cordialement,
--
Vincent RENARDIAS
Directeur Technique
StrongHoldNET / http://www.strongholdnet.com
