Back out the change from TCP/IP vol 2, in revision 1.7, which removed TH_FIN
from the output flags for CLOSING state. There is no harm in retransmitting the FIN, and this change has unexpected side effects that break simultaneous close behaviour.
This commit is contained in:
parent
401accfd7a
commit
3a64270ca6
|
@ -1,4 +1,4 @@
|
|||
/* $NetBSD: tcp_fsm.h,v 1.9 1998/07/03 05:39:56 thorpej Exp $ */
|
||||
/* $NetBSD: tcp_fsm.h,v 1.10 1998/07/09 05:49:56 mycroft Exp $ */
|
||||
|
||||
/*
|
||||
* Copyright (c) 1982, 1986, 1993
|
||||
|
@ -75,7 +75,7 @@
|
|||
u_char tcp_outflags[TCP_NSTATES] = {
|
||||
TH_RST|TH_ACK, 0, TH_SYN, TH_SYN|TH_ACK,
|
||||
TH_ACK, TH_ACK,
|
||||
TH_FIN|TH_ACK, TH_ACK, TH_FIN|TH_ACK, TH_ACK, TH_ACK,
|
||||
TH_FIN|TH_ACK, TH_FIN|TH_ACK, TH_FIN|TH_ACK, TH_ACK, TH_ACK,
|
||||
};
|
||||
#endif
|
||||
|
||||
|
|
Loading…
Reference in New Issue