Re: [lwip-users] http server and pbuf overflow

[Bernhard 'Gustl' Bauer]

Simon Goldschmidt schrieb:
> p == NULL is perfectly normal and tells your application that the
> remote side has sent a FIN. However, this should not lead to memory-
> or pbuf leaks... I'll see if I can reproduce that.

In debug_21_01.pcap is an example with p==NULL and FIN. But this is not 
the situation when pbuf leaks! The leak is shown in debug_20_01_c.pcap!

There the [SYN, ACK] (106) from LWIP is not ACKed from remote (107, 108: 
packet miss). My remote terminates after 2 sec. TCP_SYN_RCVD_TIMEOUT is 
20 sec. So [FIN, ACK] (128) is received before TCP_SYN_RCVD_TIMEOUT runs 
out. pcb->state should be still SYN_RCVD. Can you explain to me what 
will happen if [FIN, ACK] (128) is received? It looks pretty similar 
than the missing 107.

Gustl