[lwip-devel] [bug #25467] Listen backlog is not reset on timeout in SYN_RCVD

[piero]

Follow-up Comment #2, bug #25467 (project lwip):

Hi tested the patch. There was some errors in compilation.

I changed the code (i think in a correct way):

...
    LWIP_DEBUGF(TCP_DEBUG, ("tcp_pcb_purgen"));
    
// piero
// patch from Simon in lwip forum
#if TCP_LISTEN_BACKLOG
    if (pcb->state == SYN_RCVD) {
      /* Need to find the corresponding listen_pcb and decrease its
accepts_pending */
      struct tcp_pcb_listen *lpcb;
      LWIP_ASSERT("tcp_pcb_purge: pcb->state == SYN_RCVD but tcp_listen_pcbs
is NULL", 
        (tcp_listen_pcbs.listen_pcbs != NULL) );
      for (lpcb = tcp_listen_pcbs.listen_pcbs; lpcb != NULL; lpcb =
lpcb->next) {
        if ( (lpcb->local_port == pcb->local_port) &&
            (ip_addr_isany(&(lpcb->local_ip)) ||
             ip_addr_cmp( &(pcb->local_ip), &(lpcb->local_ip) ) )) {
            /* port and address of the listen pcb match the timed-out pcb */
            LWIP_ASSERT("tcp_pcb_purge: listen pcb does not have accepts
pending",
              lpcb->accepts_pending > 0);
            lpcb->accepts_pending--;
            break;
          }
      }
    }
#endif /* TCP_LISTEN_BACKLOG */
//   

    if (pcb->refused_data != NULL) {
....


I did some tests using NMAP... it seems to work!
If i understood how Simon patch the lwip code, i suppose there is no impact
in other functionality... anyway, i did some test on my application, to check
normal behaviour.

Now, when SYN/RST is received (NMAP does this for scanning ports) OR only a
SYN packet is received, the accepts_pending is purged
(in the first case when RST is received... i checked with debugger, in the
second case, when a timeout for SYN_RCVD is elapsed (20000ms) in slow_timer...
i don't have a tool to test this specific situation, i'm finding it...)

I will post here future news about thsi bug and Simon's patch.

Piero

    _______________________________________________________

Reply to this item at:

  <http://savannah.nongnu.org/bugs/?25467>

_______________________________________________
  Messaggio inviato con/da Savannah
  http://savannah.nongnu.org/