Re: Ncat -w -i options bugfixing
Daniel Miller <[email protected]>
| Newsgroups | gmane.comp.security.nmap.devel |
|---|---|
| Message-ID | <CABmvJnOKQ+WZ8_Hv3823oNk1COsBSqKQnHLF5L0oPmQU35gswQ@mail.gmail.com> |
Carlos, Thanks for this patch. I have added it in r36697. Here are some things that would have made it simpler to review: First, I added a comment pointing out the connection between get_conn_count and the connection state. Something like "If the connection count is 0, then don't time out waiting for a new connection" or "The idle timer should only be running when there are active connections." Second, I will need to look at ncat_proxy.c, where there is a similar call to fselect with input from idletimeout. In this case, though, the loop is *only* for selecting listening sockets, not connected sockets. So simply removing the idletimeout check is probably the appropriate fix. This means, then, that we need more extensive changes to handle_connect() and in ncat/http.c to add idletimeout checks. I have added an issue to the tracker for this: http://issues.nmap.org/824 Thanks! Dan On Tue, Mar 28, 2017 at 2:35 PM, Carlos Manso <[email protected]> wrote: > Hello again, > > I've taken the liberty of making a simple patch for the behaviour I > commented on my last mail, -i option causing ncat in listen mode to quit > with no connections, I hope it is ok. > > Regards, > > Carlos Manso. > > _______________________________________________ > Sent through the dev mailing list > https://nmap.org/mailman/listinfo/dev > Archived at http://seclists.org/nmap-dev/ > _______________________________________________ Sent through the dev mailing list https://nmap.org/mailman/listinfo/dev Archived at http://seclists.org/nmap-dev/