d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Cosmetic-only changes for readability.

Browse Source 
Reviewed by:	(early form passed by) bde
Approved by:	itojun (from core@kame.net)
This commit is contained in:
Jeffrey Hsu 2002-08-17 02:05:25 +00:00
parent 74495bcdab
commit c068736a61
Notes: svn2git 2020-12-20 02:59:44 +00:00 
svn path=/head/; revision=102002
2 changed files with 556 additions and 652 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

604
sys/netinet/tcp_input.c
View File

@ -68,32 +68,25 @@
#include <netinet/ip_icmp.h> /* for ICMP_BANDLIM */
#include <netinet/icmp_var.h> /* for ICMP_BANDLIM */
#include <netinet/ip_var.h>
#ifdef INET6
#include <netinet/ip6.h>
#include <netinet/icmp6.h>
#include <netinet6/in6_pcb.h>
#include <netinet6/ip6_var.h>
#include <netinet6/nd6.h>
#endif
#include <netinet/tcp.h>
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#ifdef INET6
#include <netinet6/tcp6_var.h>
#endif
#include <netinet/tcpip.h>
#ifdef TCPDEBUG
#include <netinet/tcp_debug.h>
#endif /* TCPDEBUG */
#ifdef IPSEC
#include <netinet6/ipsec.h>
#ifdef INET6
#include <netinet6/ipsec6.h>
#endif
#include <netkey/key.h>
#endif /*IPSEC*/
@ -101,7 +94,7 @@
MALLOC_DEFINE(M_TSEGQ, "tseg_qent", "TCP segment queue entry");
static int tcprexmtthresh = 3;
static const int tcprexmtthresh = 3;
tcp_cc tcp_ccgen;
struct tcpstat tcpstat;
@ -138,7 +131,7 @@ static void tcp_pulloutofband(struct socket *,
static int tcp_reass(struct tcpcb *, struct tcphdr *, int *,
struct mbuf *);
static void tcp_xmit_timer(struct tcpcb *, int);
static int tcp_newreno(struct tcpcb *, struct tcphdr *);
static void tcp_newreno_partial_ack(struct tcpcb *, struct tcphdr *);
/* Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. */
#ifdef INET6
@ -350,29 +343,35 @@ tcp_input(m, off0)
struct rmxp_tao *taop; /* pointer to our TAO cache entry */
struct rmxp_tao tao_noncached; /* in case there's no cached entry */
int headlocked = 0;
struct sockaddr_in *next_hop = NULL;
int rstreason; /* For badport_bandlim accounting purposes */
struct ip6_hdr *ip6 = NULL;
#ifdef INET6
int isipv6;
#else
const int isipv6 = 0;
#endif
#ifdef TCPDEBUG
/*
* The size of tcp_saveipgen must be the size of the max ip header,
* now IPv6.
*/
u_char tcp_saveipgen[40];
/* the size of the above must be of max ip header, now IPv6 */
struct tcphdr tcp_savetcp;
short ostate = 0;
#endif
#ifdef INET6
struct ip6_hdr *ip6 = NULL;
int isipv6;
#endif /* INET6 */
struct sockaddr_in *next_hop = NULL;
#ifdef MAC
int error;
#endif
int rstreason; /* For badport_bandlim accounting purposes */
/* Grab info from MT_TAG mbufs prepended to the chain. */
for (;m && m->m_type == MT_TAG; m = m->m_next) {
if (m->m_tag_id == PACKET_TAG_IPFORWARD)
next_hop = (struct sockaddr_in *)m->m_hdr.mh_data;
}
#ifdef INET6
isipv6 = (mtod(m, struct ip *)->ip_v == 6) ? 1 : 0;
#endif
@ -380,7 +379,6 @@ tcp_input(m, off0)
tcpstat.tcps_rcvtotal++;
#ifdef INET6
if (isipv6) {
/* IP6_EXTHDR_CHECK() is already done at tcp6_input() */
ip6 = mtod(m, struct ip6_hdr *);
@ -403,55 +401,55 @@ tcp_input(m, off0)
/* XXX stat */
goto drop;
}
} else
#endif /* INET6 */
{
/*
* Get IP and TCP header together in first mbuf.
* Note: IP leaves IP header in first mbuf.
*/
if (off0 > sizeof (struct ip)) {
ip_stripoptions(m, (struct mbuf *)0);
off0 = sizeof(struct ip);
}
if (m->m_len < sizeof (struct tcpiphdr)) {
if ((m = m_pullup(m, sizeof (struct tcpiphdr))) == 0) {
tcpstat.tcps_rcvshort++;
return;
}
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
tlen = ip->ip_len;
if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) {
if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
th->th_sum = m->m_pkthdr.csum_data;
else
th->th_sum = in_pseudo(ip->ip_src.s_addr,
ip->ip_dst.s_addr, htonl(m->m_pkthdr.csum_data +
ip->ip_len + IPPROTO_TCP));
th->th_sum ^= 0xffff;
} else {
/*
* Checksum extended TCP header and data.
* Get IP and TCP header together in first mbuf.
* Note: IP leaves IP header in first mbuf.
*/
len = sizeof (struct ip) + tlen;
bzero(ipov->ih_x1, sizeof(ipov->ih_x1));
ipov->ih_len = (u_short)tlen;
ipov->ih_len = htons(ipov->ih_len);
th->th_sum = in_cksum(m, len);
}
if (th->th_sum) {
tcpstat.tcps_rcvbadsum++;
goto drop;
}
if (off0 > sizeof (struct ip)) {
ip_stripoptions(m, (struct mbuf *)0);
off0 = sizeof(struct ip);
}
if (m->m_len < sizeof (struct tcpiphdr)) {
if ((m = m_pullup(m, sizeof (struct tcpiphdr))) == 0) {
tcpstat.tcps_rcvshort++;
return;
}
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
tlen = ip->ip_len;
if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) {
if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
th->th_sum = m->m_pkthdr.csum_data;
else
th->th_sum = in_pseudo(ip->ip_src.s_addr,
ip->ip_dst.s_addr,
htonl(m->m_pkthdr.csum_data +
ip->ip_len +
IPPROTO_TCP));
th->th_sum ^= 0xffff;
} else {
/*
* Checksum extended TCP header and data.
*/
len = sizeof (struct ip) + tlen;
bzero(ipov->ih_x1, sizeof(ipov->ih_x1));
ipov->ih_len = (u_short)tlen;
ipov->ih_len = htons(ipov->ih_len);
th->th_sum = in_cksum(m, len);
}
if (th->th_sum) {
tcpstat.tcps_rcvbadsum++;
goto drop;
}
#ifdef INET6
/* Re-initialization for later version check */
ip->ip_v = IPVERSION;
/* Re-initialization for later version check */
ip->ip_v = IPVERSION;
#endif
}
}
/*
* Check that TCP offset makes sense,
@ -464,24 +462,22 @@ tcp_input(m, off0)
}
tlen -= off; /* tlen is used instead of ti->ti_len */
if (off > sizeof (struct tcphdr)) {
#ifdef INET6
if (isipv6) {
IP6_EXTHDR_CHECK(m, off0, off, );
ip6 = mtod(m, struct ip6_hdr *);
th = (struct tcphdr *)((caddr_t)ip6 + off0);
} else
#endif /* INET6 */
{
if (m->m_len < sizeof(struct ip) + off) {
if ((m = m_pullup(m, sizeof (struct ip) + off)) == 0) {
tcpstat.tcps_rcvshort++;
return;
} else {
if (m->m_len < sizeof(struct ip) + off) {
if ((m = m_pullup(m, sizeof (struct ip) + off))
== 0) {
tcpstat.tcps_rcvshort++;
return;
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
}
}
optlen = off - sizeof (struct tcphdr);
optp = (u_char *)(th + 1);
}
@ -526,60 +522,50 @@ tcp_input(m, off0)
headlocked = 1;
findpcb:
/* IPFIREWALL_FORWARD section */
if (next_hop != NULL
#ifdef INET6
&& isipv6 == NULL /* IPv6 support is not yet */
#endif /* INET6 */
) {
if (next_hop != NULL && isipv6 == 0) { /* IPv6 support is not yet */
/*
* Transparently forwarded. Pretend to be the destination.
* already got one like this?
*/
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src, th->th_sport,
ip->ip_dst, th->th_dport, 0, m->m_pkthdr.rcvif);
ip->ip_dst, th->th_dport,
0, m->m_pkthdr.rcvif);
if (!inp) {
/*
* No, then it's new. Try find the ambushing socket
*/
if (next_hop->sin_port == 0) {
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src,
th->th_sport, next_hop->sin_addr,
th->th_dport, 1, m->m_pkthdr.rcvif);
} else {
inp = in_pcblookup_hash(&tcbinfo,
ip->ip_src, th->th_sport,
next_hop->sin_addr,
ntohs(next_hop->sin_port), 1,
m->m_pkthdr.rcvif);
}
/* It's new. Try find the ambushing socket. */
inp = in_pcblookup_hash(&tcbinfo,
ip->ip_src, th->th_sport,
next_hop->sin_addr,
next_hop->sin_port ?
ntohs(next_hop->sin_port) :
th->th_dport,
1, m->m_pkthdr.rcvif);
}
} else
{
#ifdef INET6
if (isipv6)
inp = in6_pcblookup_hash(&tcbinfo, &ip6->ip6_src, th->th_sport,
&ip6->ip6_dst, th->th_dport, 1,
m->m_pkthdr.rcvif);
else
#endif /* INET6 */
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src, th->th_sport,
ip->ip_dst, th->th_dport, 1, m->m_pkthdr.rcvif);
} else {
if (isipv6)
inp = in6_pcblookup_hash(&tcbinfo,
&ip6->ip6_src, th->th_sport,
&ip6->ip6_dst, th->th_dport,
1, m->m_pkthdr.rcvif);
else
inp = in_pcblookup_hash(&tcbinfo,
ip->ip_src, th->th_sport,
ip->ip_dst, th->th_dport,
1, m->m_pkthdr.rcvif);
}
#ifdef IPSEC
#ifdef INET6
if (isipv6) {
if (inp != NULL && ipsec6_in_reject_so(m, inp->inp_socket)) {
ipsec6stat.in_polvio++;
goto drop;
}
} else
#endif /* INET6 */
if (inp != NULL && ipsec4_in_reject_so(m, inp->inp_socket)) {
ipsecstat.in_polvio++;
goto drop;
} else {
if (inp != NULL && ipsec4_in_reject_so(m, inp->inp_socket)) {
ipsecstat.in_polvio++;
goto drop;
}
}
#endif /*IPSEC*/
#endif
/*
* If the state is CLOSED (i.e., TCB does not exist) then
@ -591,34 +577,32 @@ tcp_input(m, off0)
if (log_in_vain) {
#ifdef INET6
char dbuf[INET6_ADDRSTRLEN], sbuf[INET6_ADDRSTRLEN];
#else /* INET6 */
#else
char dbuf[4*sizeof "123"], sbuf[4*sizeof "123"];
#endif /* INET6 */
#endif
#ifdef INET6
if (isipv6) {
strcpy(dbuf, ip6_sprintf(&ip6->ip6_dst));
strcpy(sbuf, ip6_sprintf(&ip6->ip6_src));
} else
#endif
{
strcpy(dbuf, inet_ntoa(ip->ip_dst));
strcpy(sbuf, inet_ntoa(ip->ip_src));
}
} else {
strcpy(dbuf, inet_ntoa(ip->ip_dst));
strcpy(sbuf, inet_ntoa(ip->ip_src));
}
switch (log_in_vain) {
case 1:
if(thflags & TH_SYN)
if (thflags & TH_SYN)
log(LOG_INFO,
"Connection attempt to TCP %s:%d from %s:%d\n",
dbuf, ntohs(th->th_dport),
sbuf,
ntohs(th->th_sport));
"Connection attempt to TCP %s:%d "
"from %s:%d\n",
dbuf, ntohs(th->th_dport), sbuf,
ntohs(th->th_sport));
break;
case 2:
log(LOG_INFO,
"Connection attempt to TCP %s:%d from %s:%d flags:0x%x\n",
dbuf, ntohs(th->th_dport), sbuf,
ntohs(th->th_sport), thflags);
"Connection attempt to TCP %s:%d "
"from %s:%d flags:0x%x\n",
dbuf, ntohs(th->th_dport), sbuf,
ntohs(th->th_sport), thflags);
break;
default:
break;
@ -666,13 +650,12 @@ tcp_input(m, off0)
#ifdef TCPDEBUG
if (so->so_options & SO_DEBUG) {
ostate = tp->t_state;
#ifdef INET6
if (isipv6)
bcopy((char *)ip6, (char *)tcp_saveipgen,
sizeof(*ip6));
sizeof(*ip6));
else
#endif /* INET6 */
bcopy((char *)ip, (char *)tcp_saveipgen, sizeof(*ip));
bcopy((char *)ip, (char *)tcp_saveipgen,
sizeof(*ip));
tcp_savetcp = *th;
}
#endif
@ -681,14 +664,12 @@ tcp_input(m, off0)
goto after_listen;
#ifdef INET6
inc.inc_isipv6 = isipv6;
#endif
if (isipv6) {
inc.inc6_faddr = ip6->ip6_src;
inc.inc6_laddr = ip6->ip6_dst;
inc.inc6_route.ro_rt = NULL; /* XXX */
} else
#endif /* INET6 */
{
} else {
inc.inc_faddr = ip->ip_src;
inc.inc_laddr = ip->ip_dst;
inc.inc_route.ro_rt = NULL; /* XXX */
@ -817,15 +798,14 @@ tcp_input(m, off0)
* Don't bother responding if the destination was a broadcast.
*/
if (th->th_dport == th->th_sport) {
#ifdef INET6
if (isipv6) {
if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
&ip6->ip6_src))
goto drop;
} else
#endif /* INET6 */
if (ip->ip_dst.s_addr == ip->ip_src.s_addr)
goto drop;
} else {
if (ip->ip_dst.s_addr == ip->ip_src.s_addr)
goto drop;
}
}
/*
* RFC1122 4.2.3.10, p. 104: discard bcast/mcast SYN
@ -836,18 +816,17 @@ tcp_input(m, off0)
*/
if (m->m_flags & (M_BCAST|M_MCAST))
goto drop;
#ifdef INET6
if (isipv6) {
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src))
goto drop;
} else
#endif
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
} else {
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
}
/*
* SYN appears to be valid; create compressed TCP state
* for syncache, or perform t/tcp connection.
@ -884,20 +863,13 @@ tcp_input(m, off0)
* segment. Otherwise must send ACK now in case
* the other side is slow starting.
*/
if (DELAY_ACK(tp) && ((thflags & TH_FIN) ||
(tlen != 0 &&
#ifdef INET6
((isipv6 && in6_localaddr(&inp->in6p_faddr))
||
(!isipv6 &&
#endif
in_localaddr(inp->inp_faddr)
#ifdef INET6
))
#endif
))) {
callout_reset(tp->tt_delack, tcp_delacktime,
tcp_timer_delack, tp);
if (DELAY_ACK(tp) &&
((thflags & TH_FIN) ||
(tlen != 0 &&
((isipv6 && in6_localaddr(&inp->in6p_faddr)) ||
(!isipv6 && in_localaddr(inp->inp_faddr)))))) {
callout_reset(tp->tt_delack, tcp_delacktime,
tcp_timer_delack, tp);
tp->t_flags |= TF_NEEDSYN;
} else
tp->t_flags |= (TF_ACKNOW | TF_NEEDSYN);
@ -984,7 +956,7 @@ tcp_input(m, off0)
* proposal of the tcplw@cray.com list (Braden 1993/04/26).
*/
if ((to.to_flags & TOF_TS) != 0 &&
SEQ_LEQ(th->th_seq, tp->last_ack_sent)) {
SEQ_LEQ(th->th_seq, tp->last_ack_sent)) {
tp->ts_recent_age = ticks;
tp->ts_recent = to.to_tsval;
}
@ -1017,7 +989,8 @@ tcp_input(m, off0)
ticks - to.to_tsecr + 1);
else if (tp->t_rtttime &&
SEQ_GT(th->th_ack, tp->t_rtseq))
tcp_xmit_timer(tp, ticks - tp->t_rtttime);
tcp_xmit_timer(tp,
ticks - tp->t_rtttime);
acked = th->th_ack - tp->snd_una;
tcpstat.tcps_rcvackpack++;
tcpstat.tcps_rcvackbyte += acked;
@ -1223,15 +1196,16 @@ tcp_input(m, off0)
tcp_timer_keep, tp);
}
} else {
/*
* Received initial SYN in SYN-SENT[*] state => simul-
* taneous open. If segment contains CC option and there is
* a cached CC, apply TAO test; if it succeeds, connection is
* half-synchronized. Otherwise, do 3-way handshake:
* SYN-SENT -> SYN-RECEIVED
* SYN-SENT* -> SYN-RECEIVED*
* If there was no CC option, clear cached CC value.
*/
/*
* Received initial SYN in SYN-SENT[*] state =>
* simultaneous open. If segment contains CC option
* and there is a cached CC, apply TAO test.
* If it succeeds, connection is * half-synchronized.
* Otherwise, do 3-way handshake:
* SYN-SENT -> SYN-RECEIVED
* SYN-SENT* -> SYN-RECEIVED*
* If there was no CC option, clear cached CC value.
*/
tp->t_flags |= TF_ACKNOW;
callout_stop(tp->tt_rexmt);
if (to.to_flags & TOF_CC) {
@ -1282,15 +1256,17 @@ tcp_input(m, off0)
tp->snd_wl1 = th->th_seq - 1;
tp->rcv_up = th->th_seq;
/*
* Client side of transaction: already sent SYN and data.
* If the remote host used T/TCP to validate the SYN,
* our data will be ACK'd; if so, enter normal data segment
* processing in the middle of step 5, ack processing.
* Otherwise, goto step 6.
* Client side of transaction: already sent SYN and data.
* If the remote host used T/TCP to validate the SYN,
* our data will be ACK'd; if so, enter normal data segment
* processing in the middle of step 5, ack processing.
* Otherwise, goto step 6.
*/
if (thflags & TH_ACK)
goto process_ACK;
goto step6;
/*
* If the state is LAST_ACK or CLOSING or TIME_WAIT:
* if segment contains a SYN and CC [not CC.NEW] option:
@ -1701,8 +1677,9 @@ tcp_input(m, off0)
u_int win =
min(tp->snd_wnd, tp->snd_cwnd) / 2 /
tp->t_maxseg;
if (tcp_do_newreno && SEQ_LT(th->th_ack,
tp->snd_recover)) {
if (tcp_do_newreno &&
SEQ_LT(th->th_ack,
tp->snd_recover)) {
/* False retransmit, should not
* cut window
*/
@ -1721,7 +1698,7 @@ tcp_input(m, off0)
tp->snd_cwnd = tp->t_maxseg;
(void) tcp_output(tp);
tp->snd_cwnd = tp->snd_ssthresh +
tp->t_maxseg * tp->t_dupacks;
tp->t_maxseg * tp->t_dupacks;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
goto drop;
@ -1738,34 +1715,48 @@ tcp_input(m, off0)
* If the congestion window was inflated to account
* for the other side's cached packets, retract it.
*/
if (tcp_do_newreno == 0) {
if (tcp_do_newreno) {
int is_partialack = SEQ_LT(th->th_ack, tp->snd_recover);
if (tp->t_dupacks >= tcprexmtthresh) {
if (is_partialack) {
tcp_newreno_partial_ack(tp, th);
} else {
/*
* Window inflation should have left us
* with approximately snd_ssthresh
* outstanding data.
* But in case we would be inclined to
* send a burst, better to do it via
* the slow start mechanism.
*/
if (SEQ_GT(th->th_ack +
tp->snd_ssthresh,
tp->snd_max))
tp->snd_cwnd = tp->snd_max -
th->th_ack +
tp->t_maxseg;
else
tp->snd_cwnd = tp->snd_ssthresh;
}
}
/*
* Reset dupacks, except on partial acks in
* fast recovery.
*/
if (!(tp->t_dupacks >= tcprexmtthresh && is_partialack))
tp->t_dupacks = 0;
} else {
if (tp->t_dupacks >= tcprexmtthresh &&
tp->snd_cwnd > tp->snd_ssthresh)
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0;
} else if (tp->t_dupacks >= tcprexmtthresh &&
!tcp_newreno(tp, th)) {
/*
* Window inflation should have left us with approx.
* snd_ssthresh outstanding data. But in case we
* would be inclined to send a burst, better to do
* it via the slow start mechanism.
*/
if (SEQ_GT(th->th_ack + tp->snd_ssthresh, tp->snd_max))
tp->snd_cwnd =
tp->snd_max - th->th_ack + tp->t_maxseg;
else
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0;
}
if (tp->t_dupacks < tcprexmtthresh)
tp->snd_cwnd > tp->snd_ssthresh)
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0;
}
if (SEQ_GT(th->th_ack, tp->snd_max)) {
tcpstat.tcps_rcvacktoomuch++;
goto dropafterack;
}
/*
* If we reach this point, ACK is not a duplicate,
* If we reach this point, ACK is not a duplicate,
* i.e., it ACKs something we sent.
*/
if (tp->t_flags & TF_NEEDSYN) {
@ -1858,7 +1849,7 @@ tcp_input(m, off0)
* in NewReno fast recovery mode, so we leave the congestion
* window alone.
*/
if (tcp_do_newreno == 0 || tp->t_dupacks == 0)
if (!tcp_do_newreno || tp->t_dupacks == 0)
tp->snd_cwnd = min(cw + incr,TCP_MAXWIN<<tp->snd_scale);
}
if (acked > so->so_snd.sb_cc) {
@ -2022,7 +2013,7 @@ tcp_input(m, off0)
)
tcp_pulloutofband(so, th, m,
drop_hdrlen); /* hdr drop is delayed */
} else
} else {
/*
* If no out of band data is expected,
* pull receive urgent pointer along
@ -2030,6 +2021,7 @@ tcp_input(m, off0)
*/
if (SEQ_GT(tp->rcv_nxt, tp->rcv_up))
tp->rcv_up = tp->rcv_nxt;
}
dodata: /* XXX */
KASSERT(headlocked, ("headlocked"));
INP_INFO_WUNLOCK(&tcbinfo);
@ -2042,25 +2034,27 @@ tcp_input(m, off0)
* case PRU_RCVD). If a FIN has already been received on this
* connection then we just ignore the text.
*/
if ((tlen || (thflags&TH_FIN)) &&
if ((tlen || (thflags & TH_FIN)) &&
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
m_adj(m, drop_hdrlen); /* delayed header drop */
/*
* Insert segment which inludes th into reassembly queue of tcp with
* control block tp. Return TH_FIN if reassembly now includes
* a segment with FIN. This handle the common case inline (segment
* is the next to be received on an established connection, and the
* queue is empty), avoiding linkage into and removal from the queue
* and repetition of various conversions.
* Set DELACK for segments received in order, but ack immediately
* when segments are out of order (so fast retransmit can work).
* Insert segment which includes th into TCP reassembly queue
* with control block tp. Set thflags to whether reassembly now
* includes a segment with FIN. This handles the common case
* inline (segment is the next to be received on an established
* connection, and the queue is empty), avoiding linkage into
* and removal from the queue and repetition of various
* conversions.
* Set DELACK for segments received in order, but ack
* immediately when segments are out of order (so
* fast retransmit can work).
*/
if (th->th_seq == tp->rcv_nxt &&
LIST_EMPTY(&tp->t_segq) &&
TCPS_HAVEESTABLISHED(tp->t_state)) {
if (DELAY_ACK(tp))
callout_reset(tp->tt_delack, tcp_delacktime,
tcp_timer_delack, tp);
tcp_timer_delack, tp);
else
tp->t_flags |= TF_ACKNOW;
tp->rcv_nxt += tlen;
@ -2094,11 +2088,11 @@ tcp_input(m, off0)
if (TCPS_HAVERCVDFIN(tp->t_state) == 0) {
socantrcvmore(so);
/*
* If connection is half-synchronized
* (ie NEEDSYN flag on) then delay ACK,
* so it may be piggybacked when SYN is sent.
* Otherwise, since we received a FIN then no
* more input can be expected, send ACK now.
* If connection is half-synchronized
* (ie NEEDSYN flag on) then delay ACK,
* so it may be piggybacked when SYN is sent.
* Otherwise, since we received a FIN then no
* more input can be expected, send ACK now.
*/
if (DELAY_ACK(tp) && (tp->t_flags & TF_NEEDSYN))
callout_reset(tp->tt_delack, tcp_delacktime,
@ -2217,18 +2211,17 @@ tcp_input(m, off0)
*/
if ((thflags & TH_RST) || m->m_flags & (M_BCAST|M_MCAST))
goto drop;
#ifdef INET6
if (isipv6) {
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src))
goto drop;
} else
#endif /* INET6 */
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
} else {
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
}
/* IPv6 anycast check is done at tcp6_input() */
/*
@ -2518,35 +2511,27 @@ tcp_mss(tp, offer)
struct ifnet *ifp;
register int rtt, mss;
u_long bufsize;
struct inpcb *inp;
struct inpcb *inp = tp->t_inpcb;
struct socket *so;
struct rmxp_tao *taop;
int origoffer = offer;
#ifdef INET6
int isipv6;
int min_protoh;
int isipv6 = ((inp->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
size_t min_protoh = isipv6 ?
sizeof (struct ip6_hdr) + sizeof (struct tcphdr) :
sizeof (struct tcpiphdr);
#else
const int isipv6 = 0;
const size_t min_protoh = sizeof (struct tcpiphdr);
#endif
inp = tp->t_inpcb;
#ifdef INET6
isipv6 = ((inp->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
min_protoh = isipv6 ? sizeof (struct ip6_hdr) + sizeof (struct tcphdr)
: sizeof (struct tcpiphdr);
#else
#define min_protoh (sizeof (struct tcpiphdr))
#endif
#ifdef INET6
if (isipv6)
rt = tcp_rtlookup6(&inp->inp_inc);
else
#endif
rt = tcp_rtlookup(&inp->inp_inc);
rt = tcp_rtlookup(&inp->inp_inc);
if (rt == NULL) {
tp->t_maxopd = tp->t_maxseg =
#ifdef INET6
isipv6 ? tcp_v6mssdflt :
#endif /* INET6 */
tcp_mssdflt;
isipv6 ? tcp_v6mssdflt : tcp_mssdflt;
return;
}
ifp = rt->rt_ifp;
@ -2564,11 +2549,7 @@ tcp_mss(tp, offer)
* in this case we use tcp_mssdflt.
*/
if (offer == 0)
offer =
#ifdef INET6
isipv6 ? tcp_v6mssdflt :
#endif /* INET6 */
tcp_mssdflt;
offer = isipv6 ? tcp_v6mssdflt : tcp_mssdflt;
else
/*
* Sanity check: make sure that maxopd will be large
@ -2612,25 +2593,17 @@ tcp_mss(tp, offer)
*/
if (rt->rt_rmx.rmx_mtu)
mss = rt->rt_rmx.rmx_mtu - min_protoh;
else
{
mss =
#ifdef INET6
(isipv6 ? nd_ifinfo[rt->rt_ifp->if_index].linkmtu :
#endif
ifp->if_mtu
#ifdef INET6
)
#endif
- min_protoh;
#ifdef INET6
else {
if (isipv6) {
mss = nd_ifinfo[rt->rt_ifp->if_index].linkmtu -
min_protoh;
if (!in6_localaddr(&inp->in6p_faddr))
mss = min(mss, tcp_v6mssdflt);
} else
#endif
if (!in_localaddr(inp->inp_faddr))
mss = min(mss, tcp_mssdflt);
} else {
mss = ifp->if_mtu - min_protoh;
if (!in_localaddr(inp->inp_faddr))
mss = min(mss, tcp_mssdflt);
}
}
mss = min(mss, offer);
/*
@ -2700,16 +2673,8 @@ tcp_mss(tp, offer)
* Set the slow-start flight size depending on whether this
* is a local network or not.
*/
if (
#ifdef INET6
(isipv6 && in6_localaddr(&inp->in6p_faddr)) ||
(!isipv6 &&
#endif
in_localaddr(inp->inp_faddr)
#ifdef INET6
)
#endif
)
if ((isipv6 && in6_localaddr(&inp->in6p_faddr)) ||
(!isipv6 && in_localaddr(inp->inp_faddr)))
tp->snd_cwnd = mss * ss_fltsz_local;
else
tp->snd_cwnd = mss * ss_fltsz;
@ -2735,68 +2700,55 @@ tcp_mssopt(tp)
{
struct rtentry *rt;
#ifdef INET6
int isipv6;
int min_protoh;
int isipv6 = ((tp->t_inpcb->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
size_t min_protoh = isipv6 ?
sizeof (struct ip6_hdr) + sizeof (struct tcphdr) :
sizeof (struct tcpiphdr);
#else
const int isipv6 = 0;
const size_t min_protoh = sizeof (struct tcpiphdr);
#endif
#ifdef INET6
isipv6 = ((tp->t_inpcb->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
min_protoh = isipv6 ? sizeof (struct ip6_hdr) + sizeof (struct tcphdr)
: sizeof (struct tcpiphdr);
#else
#define min_protoh (sizeof (struct tcpiphdr))
#endif
#ifdef INET6
if (isipv6)
rt = tcp_rtlookup6(&tp->t_inpcb->inp_inc);
else
#endif /* INET6 */
rt = tcp_rtlookup(&tp->t_inpcb->inp_inc);
rt = tcp_rtlookup(&tp->t_inpcb->inp_inc);
if (rt == NULL)
return
#ifdef INET6
isipv6 ? tcp_v6mssdflt :
#endif /* INET6 */
tcp_mssdflt;
return (isipv6 ? tcp_v6mssdflt : tcp_mssdflt);
return rt->rt_ifp->if_mtu - min_protoh;
return (rt->rt_ifp->if_mtu - min_protoh);
}
/*
* Checks for partial ack. If partial ack arrives, force the retransmission
* of the next unacknowledged segment, do not clear tp->t_dupacks, and return
* 1. By setting snd_nxt to ti_ack, this forces retransmission timer to
* be started again. If the ack advances at least to tp->snd_recover, return 0.
* On a partial ack arrives, force the retransmission of the
* next unacknowledged segment. Do not clear tp->t_dupacks.
* By setting snd_nxt to ti_ack, this forces retransmission timer to
* be started again.
*/
static int
tcp_newreno(tp, th)
static void
tcp_newreno_partial_ack(tp, th)
struct tcpcb *tp;
struct tcphdr *th;
{
if (SEQ_LT(th->th_ack, tp->snd_recover)) {
tcp_seq onxt = tp->snd_nxt;
u_long ocwnd = tp->snd_cwnd;
tcp_seq onxt = tp->snd_nxt;
u_long ocwnd = tp->snd_cwnd;
callout_stop(tp->tt_rexmt);
tp->t_rtttime = 0;
tp->snd_nxt = th->th_ack;
/*
* Set snd_cwnd to one segment beyond acknowledged offset
* (tp->snd_una has not yet been updated when this function
* is called)
*/
tp->snd_cwnd = tp->t_maxseg + (th->th_ack - tp->snd_una);
(void) tcp_output(tp);
tp->snd_cwnd = ocwnd;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
/*
* Partial window deflation. Relies on fact that tp->snd_una
* not updated yet.
*/
tp->snd_cwnd -= (th->th_ack - tp->snd_una - tp->t_maxseg);
return (1);
}
return (0);
callout_stop(tp->tt_rexmt);
tp->t_rtttime = 0;
tp->snd_nxt = th->th_ack;
/*
* Set snd_cwnd to one segment beyond acknowledged offset.
* (tp->snd_una has not yet been updated when this function is called.)
*/
tp->snd_cwnd = tp->t_maxseg + (th->th_ack - tp->snd_una);
(void) tcp_output(tp);
tp->snd_cwnd = ocwnd;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
/*
* Partial window deflation. Relies on fact that tp->snd_una
* not updated yet.
*/
tp->snd_cwnd -= (th->th_ack - tp->snd_una - tp->t_maxseg);
}

604
sys/netinet/tcp_reass.c
View File

@ -68,32 +68,25 @@
#include <netinet/ip_icmp.h> /* for ICMP_BANDLIM */
#include <netinet/icmp_var.h> /* for ICMP_BANDLIM */
#include <netinet/ip_var.h>
#ifdef INET6
#include <netinet/ip6.h>
#include <netinet/icmp6.h>
#include <netinet6/in6_pcb.h>
#include <netinet6/ip6_var.h>
#include <netinet6/nd6.h>
#endif
#include <netinet/tcp.h>
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#ifdef INET6
#include <netinet6/tcp6_var.h>
#endif
#include <netinet/tcpip.h>
#ifdef TCPDEBUG
#include <netinet/tcp_debug.h>
#endif /* TCPDEBUG */
#ifdef IPSEC
#include <netinet6/ipsec.h>
#ifdef INET6
#include <netinet6/ipsec6.h>
#endif
#include <netkey/key.h>
#endif /*IPSEC*/
@ -101,7 +94,7 @@
MALLOC_DEFINE(M_TSEGQ, "tseg_qent", "TCP segment queue entry");
static int tcprexmtthresh = 3;
static const int tcprexmtthresh = 3;
tcp_cc tcp_ccgen;
struct tcpstat tcpstat;
@ -138,7 +131,7 @@ static void tcp_pulloutofband(struct socket *,
static int tcp_reass(struct tcpcb *, struct tcphdr *, int *,
struct mbuf *);
static void tcp_xmit_timer(struct tcpcb *, int);
static int tcp_newreno(struct tcpcb *, struct tcphdr *);
static void tcp_newreno_partial_ack(struct tcpcb *, struct tcphdr *);
/* Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. */
#ifdef INET6
@ -350,29 +343,35 @@ tcp_input(m, off0)
struct rmxp_tao *taop; /* pointer to our TAO cache entry */
struct rmxp_tao tao_noncached; /* in case there's no cached entry */
int headlocked = 0;
struct sockaddr_in *next_hop = NULL;
int rstreason; /* For badport_bandlim accounting purposes */
struct ip6_hdr *ip6 = NULL;
#ifdef INET6
int isipv6;
#else
const int isipv6 = 0;
#endif
#ifdef TCPDEBUG
/*
* The size of tcp_saveipgen must be the size of the max ip header,
* now IPv6.
*/
u_char tcp_saveipgen[40];
/* the size of the above must be of max ip header, now IPv6 */
struct tcphdr tcp_savetcp;
short ostate = 0;
#endif
#ifdef INET6
struct ip6_hdr *ip6 = NULL;
int isipv6;
#endif /* INET6 */
struct sockaddr_in *next_hop = NULL;
#ifdef MAC
int error;
#endif
int rstreason; /* For badport_bandlim accounting purposes */
/* Grab info from MT_TAG mbufs prepended to the chain. */
for (;m && m->m_type == MT_TAG; m = m->m_next) {
if (m->m_tag_id == PACKET_TAG_IPFORWARD)
next_hop = (struct sockaddr_in *)m->m_hdr.mh_data;
}
#ifdef INET6
isipv6 = (mtod(m, struct ip *)->ip_v == 6) ? 1 : 0;
#endif
@ -380,7 +379,6 @@ tcp_input(m, off0)
tcpstat.tcps_rcvtotal++;
#ifdef INET6
if (isipv6) {
/* IP6_EXTHDR_CHECK() is already done at tcp6_input() */
ip6 = mtod(m, struct ip6_hdr *);
@ -403,55 +401,55 @@ tcp_input(m, off0)
/* XXX stat */
goto drop;
}
} else
#endif /* INET6 */
{
/*
* Get IP and TCP header together in first mbuf.
* Note: IP leaves IP header in first mbuf.
*/
if (off0 > sizeof (struct ip)) {
ip_stripoptions(m, (struct mbuf *)0);
off0 = sizeof(struct ip);
}
if (m->m_len < sizeof (struct tcpiphdr)) {
if ((m = m_pullup(m, sizeof (struct tcpiphdr))) == 0) {
tcpstat.tcps_rcvshort++;
return;
}
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
tlen = ip->ip_len;
if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) {
if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
th->th_sum = m->m_pkthdr.csum_data;
else
th->th_sum = in_pseudo(ip->ip_src.s_addr,
ip->ip_dst.s_addr, htonl(m->m_pkthdr.csum_data +
ip->ip_len + IPPROTO_TCP));
th->th_sum ^= 0xffff;
} else {
/*
* Checksum extended TCP header and data.
* Get IP and TCP header together in first mbuf.
* Note: IP leaves IP header in first mbuf.
*/
len = sizeof (struct ip) + tlen;
bzero(ipov->ih_x1, sizeof(ipov->ih_x1));
ipov->ih_len = (u_short)tlen;
ipov->ih_len = htons(ipov->ih_len);
th->th_sum = in_cksum(m, len);
}
if (th->th_sum) {
tcpstat.tcps_rcvbadsum++;
goto drop;
}
if (off0 > sizeof (struct ip)) {
ip_stripoptions(m, (struct mbuf *)0);
off0 = sizeof(struct ip);
}
if (m->m_len < sizeof (struct tcpiphdr)) {
if ((m = m_pullup(m, sizeof (struct tcpiphdr))) == 0) {
tcpstat.tcps_rcvshort++;
return;
}
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
tlen = ip->ip_len;
if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) {
if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
th->th_sum = m->m_pkthdr.csum_data;
else
th->th_sum = in_pseudo(ip->ip_src.s_addr,
ip->ip_dst.s_addr,
htonl(m->m_pkthdr.csum_data +
ip->ip_len +
IPPROTO_TCP));
th->th_sum ^= 0xffff;
} else {
/*
* Checksum extended TCP header and data.
*/
len = sizeof (struct ip) + tlen;
bzero(ipov->ih_x1, sizeof(ipov->ih_x1));
ipov->ih_len = (u_short)tlen;
ipov->ih_len = htons(ipov->ih_len);
th->th_sum = in_cksum(m, len);
}
if (th->th_sum) {
tcpstat.tcps_rcvbadsum++;
goto drop;
}
#ifdef INET6
/* Re-initialization for later version check */
ip->ip_v = IPVERSION;
/* Re-initialization for later version check */
ip->ip_v = IPVERSION;
#endif
}
}
/*
* Check that TCP offset makes sense,
@ -464,24 +462,22 @@ tcp_input(m, off0)
}
tlen -= off; /* tlen is used instead of ti->ti_len */
if (off > sizeof (struct tcphdr)) {
#ifdef INET6
if (isipv6) {
IP6_EXTHDR_CHECK(m, off0, off, );
ip6 = mtod(m, struct ip6_hdr *);
th = (struct tcphdr *)((caddr_t)ip6 + off0);
} else
#endif /* INET6 */
{
if (m->m_len < sizeof(struct ip) + off) {
if ((m = m_pullup(m, sizeof (struct ip) + off)) == 0) {
tcpstat.tcps_rcvshort++;
return;
} else {
if (m->m_len < sizeof(struct ip) + off) {
if ((m = m_pullup(m, sizeof (struct ip) + off))
== 0) {
tcpstat.tcps_rcvshort++;
return;
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
}
ip = mtod(m, struct ip *);
ipov = (struct ipovly *)ip;
th = (struct tcphdr *)((caddr_t)ip + off0);
}
}
optlen = off - sizeof (struct tcphdr);
optp = (u_char *)(th + 1);
}
@ -526,60 +522,50 @@ tcp_input(m, off0)
headlocked = 1;
findpcb:
/* IPFIREWALL_FORWARD section */
if (next_hop != NULL
#ifdef INET6
&& isipv6 == NULL /* IPv6 support is not yet */
#endif /* INET6 */
) {
if (next_hop != NULL && isipv6 == 0) { /* IPv6 support is not yet */
/*
* Transparently forwarded. Pretend to be the destination.
* already got one like this?
*/
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src, th->th_sport,
ip->ip_dst, th->th_dport, 0, m->m_pkthdr.rcvif);
ip->ip_dst, th->th_dport,
0, m->m_pkthdr.rcvif);
if (!inp) {
/*
* No, then it's new. Try find the ambushing socket
*/
if (next_hop->sin_port == 0) {
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src,
th->th_sport, next_hop->sin_addr,
th->th_dport, 1, m->m_pkthdr.rcvif);
} else {
inp = in_pcblookup_hash(&tcbinfo,
ip->ip_src, th->th_sport,
next_hop->sin_addr,
ntohs(next_hop->sin_port), 1,
m->m_pkthdr.rcvif);
}
/* It's new. Try find the ambushing socket. */
inp = in_pcblookup_hash(&tcbinfo,
ip->ip_src, th->th_sport,
next_hop->sin_addr,
next_hop->sin_port ?
ntohs(next_hop->sin_port) :
th->th_dport,
1, m->m_pkthdr.rcvif);
}
} else
{
#ifdef INET6
if (isipv6)
inp = in6_pcblookup_hash(&tcbinfo, &ip6->ip6_src, th->th_sport,
&ip6->ip6_dst, th->th_dport, 1,
m->m_pkthdr.rcvif);
else
#endif /* INET6 */
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src, th->th_sport,
ip->ip_dst, th->th_dport, 1, m->m_pkthdr.rcvif);
} else {
if (isipv6)
inp = in6_pcblookup_hash(&tcbinfo,
&ip6->ip6_src, th->th_sport,
&ip6->ip6_dst, th->th_dport,
1, m->m_pkthdr.rcvif);
else
inp = in_pcblookup_hash(&tcbinfo,
ip->ip_src, th->th_sport,
ip->ip_dst, th->th_dport,
1, m->m_pkthdr.rcvif);
}
#ifdef IPSEC
#ifdef INET6
if (isipv6) {
if (inp != NULL && ipsec6_in_reject_so(m, inp->inp_socket)) {
ipsec6stat.in_polvio++;
goto drop;
}
} else
#endif /* INET6 */
if (inp != NULL && ipsec4_in_reject_so(m, inp->inp_socket)) {
ipsecstat.in_polvio++;
goto drop;
} else {
if (inp != NULL && ipsec4_in_reject_so(m, inp->inp_socket)) {
ipsecstat.in_polvio++;
goto drop;
}
}
#endif /*IPSEC*/
#endif
/*
* If the state is CLOSED (i.e., TCB does not exist) then
@ -591,34 +577,32 @@ tcp_input(m, off0)
if (log_in_vain) {
#ifdef INET6
char dbuf[INET6_ADDRSTRLEN], sbuf[INET6_ADDRSTRLEN];
#else /* INET6 */
#else
char dbuf[4*sizeof "123"], sbuf[4*sizeof "123"];
#endif /* INET6 */
#endif
#ifdef INET6
if (isipv6) {
strcpy(dbuf, ip6_sprintf(&ip6->ip6_dst));
strcpy(sbuf, ip6_sprintf(&ip6->ip6_src));
} else
#endif
{
strcpy(dbuf, inet_ntoa(ip->ip_dst));
strcpy(sbuf, inet_ntoa(ip->ip_src));
}
} else {
strcpy(dbuf, inet_ntoa(ip->ip_dst));
strcpy(sbuf, inet_ntoa(ip->ip_src));
}
switch (log_in_vain) {
case 1:
if(thflags & TH_SYN)
if (thflags & TH_SYN)
log(LOG_INFO,
"Connection attempt to TCP %s:%d from %s:%d\n",
dbuf, ntohs(th->th_dport),
sbuf,
ntohs(th->th_sport));
"Connection attempt to TCP %s:%d "
"from %s:%d\n",
dbuf, ntohs(th->th_dport), sbuf,
ntohs(th->th_sport));
break;
case 2:
log(LOG_INFO,
"Connection attempt to TCP %s:%d from %s:%d flags:0x%x\n",
dbuf, ntohs(th->th_dport), sbuf,
ntohs(th->th_sport), thflags);
"Connection attempt to TCP %s:%d "
"from %s:%d flags:0x%x\n",
dbuf, ntohs(th->th_dport), sbuf,
ntohs(th->th_sport), thflags);
break;
default:
break;
@ -666,13 +650,12 @@ tcp_input(m, off0)
#ifdef TCPDEBUG
if (so->so_options & SO_DEBUG) {
ostate = tp->t_state;
#ifdef INET6
if (isipv6)
bcopy((char *)ip6, (char *)tcp_saveipgen,
sizeof(*ip6));
sizeof(*ip6));
else
#endif /* INET6 */
bcopy((char *)ip, (char *)tcp_saveipgen, sizeof(*ip));
bcopy((char *)ip, (char *)tcp_saveipgen,
sizeof(*ip));
tcp_savetcp = *th;
}
#endif
@ -681,14 +664,12 @@ tcp_input(m, off0)
goto after_listen;
#ifdef INET6
inc.inc_isipv6 = isipv6;
#endif
if (isipv6) {
inc.inc6_faddr = ip6->ip6_src;
inc.inc6_laddr = ip6->ip6_dst;
inc.inc6_route.ro_rt = NULL; /* XXX */
} else
#endif /* INET6 */
{
} else {
inc.inc_faddr = ip->ip_src;
inc.inc_laddr = ip->ip_dst;
inc.inc_route.ro_rt = NULL; /* XXX */
@ -817,15 +798,14 @@ tcp_input(m, off0)
* Don't bother responding if the destination was a broadcast.
*/
if (th->th_dport == th->th_sport) {
#ifdef INET6
if (isipv6) {
if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
&ip6->ip6_src))
goto drop;
} else
#endif /* INET6 */
if (ip->ip_dst.s_addr == ip->ip_src.s_addr)
goto drop;
} else {
if (ip->ip_dst.s_addr == ip->ip_src.s_addr)
goto drop;
}
}
/*
* RFC1122 4.2.3.10, p. 104: discard bcast/mcast SYN
@ -836,18 +816,17 @@ tcp_input(m, off0)
*/
if (m->m_flags & (M_BCAST|M_MCAST))
goto drop;
#ifdef INET6
if (isipv6) {
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src))
goto drop;
} else
#endif
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
} else {
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
}
/*
* SYN appears to be valid; create compressed TCP state
* for syncache, or perform t/tcp connection.
@ -884,20 +863,13 @@ tcp_input(m, off0)
* segment. Otherwise must send ACK now in case
* the other side is slow starting.
*/
if (DELAY_ACK(tp) && ((thflags & TH_FIN) ||
(tlen != 0 &&
#ifdef INET6
((isipv6 && in6_localaddr(&inp->in6p_faddr))
||
(!isipv6 &&
#endif
in_localaddr(inp->inp_faddr)
#ifdef INET6
))
#endif
))) {
callout_reset(tp->tt_delack, tcp_delacktime,
tcp_timer_delack, tp);
if (DELAY_ACK(tp) &&
((thflags & TH_FIN) ||
(tlen != 0 &&
((isipv6 && in6_localaddr(&inp->in6p_faddr)) ||
(!isipv6 && in_localaddr(inp->inp_faddr)))))) {
callout_reset(tp->tt_delack, tcp_delacktime,
tcp_timer_delack, tp);
tp->t_flags |= TF_NEEDSYN;
} else
tp->t_flags |= (TF_ACKNOW | TF_NEEDSYN);
@ -984,7 +956,7 @@ tcp_input(m, off0)
* proposal of the tcplw@cray.com list (Braden 1993/04/26).
*/
if ((to.to_flags & TOF_TS) != 0 &&
SEQ_LEQ(th->th_seq, tp->last_ack_sent)) {
SEQ_LEQ(th->th_seq, tp->last_ack_sent)) {
tp->ts_recent_age = ticks;
tp->ts_recent = to.to_tsval;
}
@ -1017,7 +989,8 @@ tcp_input(m, off0)
ticks - to.to_tsecr + 1);
else if (tp->t_rtttime &&
SEQ_GT(th->th_ack, tp->t_rtseq))
tcp_xmit_timer(tp, ticks - tp->t_rtttime);
tcp_xmit_timer(tp,
ticks - tp->t_rtttime);
acked = th->th_ack - tp->snd_una;
tcpstat.tcps_rcvackpack++;
tcpstat.tcps_rcvackbyte += acked;
@ -1223,15 +1196,16 @@ tcp_input(m, off0)
tcp_timer_keep, tp);
}
} else {
/*
* Received initial SYN in SYN-SENT[*] state => simul-
* taneous open. If segment contains CC option and there is
* a cached CC, apply TAO test; if it succeeds, connection is
* half-synchronized. Otherwise, do 3-way handshake:
* SYN-SENT -> SYN-RECEIVED
* SYN-SENT* -> SYN-RECEIVED*
* If there was no CC option, clear cached CC value.
*/
/*
* Received initial SYN in SYN-SENT[*] state =>
* simultaneous open. If segment contains CC option
* and there is a cached CC, apply TAO test.
* If it succeeds, connection is * half-synchronized.
* Otherwise, do 3-way handshake:
* SYN-SENT -> SYN-RECEIVED
* SYN-SENT* -> SYN-RECEIVED*
* If there was no CC option, clear cached CC value.
*/
tp->t_flags |= TF_ACKNOW;
callout_stop(tp->tt_rexmt);
if (to.to_flags & TOF_CC) {
@ -1282,15 +1256,17 @@ tcp_input(m, off0)
tp->snd_wl1 = th->th_seq - 1;
tp->rcv_up = th->th_seq;
/*
* Client side of transaction: already sent SYN and data.
* If the remote host used T/TCP to validate the SYN,
* our data will be ACK'd; if so, enter normal data segment
* processing in the middle of step 5, ack processing.
* Otherwise, goto step 6.
* Client side of transaction: already sent SYN and data.
* If the remote host used T/TCP to validate the SYN,
* our data will be ACK'd; if so, enter normal data segment
* processing in the middle of step 5, ack processing.
* Otherwise, goto step 6.
*/
if (thflags & TH_ACK)
goto process_ACK;
goto step6;
/*
* If the state is LAST_ACK or CLOSING or TIME_WAIT:
* if segment contains a SYN and CC [not CC.NEW] option:
@ -1701,8 +1677,9 @@ tcp_input(m, off0)
u_int win =
min(tp->snd_wnd, tp->snd_cwnd) / 2 /
tp->t_maxseg;
if (tcp_do_newreno && SEQ_LT(th->th_ack,
tp->snd_recover)) {
if (tcp_do_newreno &&
SEQ_LT(th->th_ack,
tp->snd_recover)) {
/* False retransmit, should not
* cut window
*/
@ -1721,7 +1698,7 @@ tcp_input(m, off0)
tp->snd_cwnd = tp->t_maxseg;
(void) tcp_output(tp);
tp->snd_cwnd = tp->snd_ssthresh +
tp->t_maxseg * tp->t_dupacks;
tp->t_maxseg * tp->t_dupacks;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
goto drop;
@ -1738,34 +1715,48 @@ tcp_input(m, off0)
* If the congestion window was inflated to account
* for the other side's cached packets, retract it.
*/
if (tcp_do_newreno == 0) {
if (tcp_do_newreno) {
int is_partialack = SEQ_LT(th->th_ack, tp->snd_recover);
if (tp->t_dupacks >= tcprexmtthresh) {
if (is_partialack) {
tcp_newreno_partial_ack(tp, th);
} else {
/*
* Window inflation should have left us
* with approximately snd_ssthresh
* outstanding data.
* But in case we would be inclined to
* send a burst, better to do it via
* the slow start mechanism.
*/
if (SEQ_GT(th->th_ack +
tp->snd_ssthresh,
tp->snd_max))
tp->snd_cwnd = tp->snd_max -
th->th_ack +
tp->t_maxseg;
else
tp->snd_cwnd = tp->snd_ssthresh;
}
}
/*
* Reset dupacks, except on partial acks in
* fast recovery.
*/
if (!(tp->t_dupacks >= tcprexmtthresh && is_partialack))
tp->t_dupacks = 0;
} else {
if (tp->t_dupacks >= tcprexmtthresh &&
tp->snd_cwnd > tp->snd_ssthresh)
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0;
} else if (tp->t_dupacks >= tcprexmtthresh &&
!tcp_newreno(tp, th)) {
/*
* Window inflation should have left us with approx.
* snd_ssthresh outstanding data. But in case we
* would be inclined to send a burst, better to do
* it via the slow start mechanism.
*/
if (SEQ_GT(th->th_ack + tp->snd_ssthresh, tp->snd_max))
tp->snd_cwnd =
tp->snd_max - th->th_ack + tp->t_maxseg;
else
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0;
}
if (tp->t_dupacks < tcprexmtthresh)
tp->snd_cwnd > tp->snd_ssthresh)
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0;
}
if (SEQ_GT(th->th_ack, tp->snd_max)) {
tcpstat.tcps_rcvacktoomuch++;
goto dropafterack;
}
/*
* If we reach this point, ACK is not a duplicate,
* If we reach this point, ACK is not a duplicate,
* i.e., it ACKs something we sent.
*/
if (tp->t_flags & TF_NEEDSYN) {
@ -1858,7 +1849,7 @@ tcp_input(m, off0)
* in NewReno fast recovery mode, so we leave the congestion
* window alone.
*/
if (tcp_do_newreno == 0 || tp->t_dupacks == 0)
if (!tcp_do_newreno || tp->t_dupacks == 0)
tp->snd_cwnd = min(cw + incr,TCP_MAXWIN<<tp->snd_scale);
}
if (acked > so->so_snd.sb_cc) {
@ -2022,7 +2013,7 @@ tcp_input(m, off0)
)
tcp_pulloutofband(so, th, m,
drop_hdrlen); /* hdr drop is delayed */
} else
} else {
/*
* If no out of band data is expected,
* pull receive urgent pointer along
@ -2030,6 +2021,7 @@ tcp_input(m, off0)
*/
if (SEQ_GT(tp->rcv_nxt, tp->rcv_up))
tp->rcv_up = tp->rcv_nxt;
}
dodata: /* XXX */
KASSERT(headlocked, ("headlocked"));
INP_INFO_WUNLOCK(&tcbinfo);
@ -2042,25 +2034,27 @@ tcp_input(m, off0)
* case PRU_RCVD). If a FIN has already been received on this
* connection then we just ignore the text.
*/
if ((tlen || (thflags&TH_FIN)) &&
if ((tlen || (thflags & TH_FIN)) &&
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
m_adj(m, drop_hdrlen); /* delayed header drop */
/*
* Insert segment which inludes th into reassembly queue of tcp with
* control block tp. Return TH_FIN if reassembly now includes
* a segment with FIN. This handle the common case inline (segment
* is the next to be received on an established connection, and the
* queue is empty), avoiding linkage into and removal from the queue
* and repetition of various conversions.
* Set DELACK for segments received in order, but ack immediately
* when segments are out of order (so fast retransmit can work).
* Insert segment which includes th into TCP reassembly queue
* with control block tp. Set thflags to whether reassembly now
* includes a segment with FIN. This handles the common case
* inline (segment is the next to be received on an established
* connection, and the queue is empty), avoiding linkage into
* and removal from the queue and repetition of various
* conversions.
* Set DELACK for segments received in order, but ack
* immediately when segments are out of order (so
* fast retransmit can work).
*/
if (th->th_seq == tp->rcv_nxt &&
LIST_EMPTY(&tp->t_segq) &&
TCPS_HAVEESTABLISHED(tp->t_state)) {
if (DELAY_ACK(tp))
callout_reset(tp->tt_delack, tcp_delacktime,
tcp_timer_delack, tp);
tcp_timer_delack, tp);
else
tp->t_flags |= TF_ACKNOW;
tp->rcv_nxt += tlen;
@ -2094,11 +2088,11 @@ tcp_input(m, off0)
if (TCPS_HAVERCVDFIN(tp->t_state) == 0) {
socantrcvmore(so);
/*
* If connection is half-synchronized
* (ie NEEDSYN flag on) then delay ACK,
* so it may be piggybacked when SYN is sent.
* Otherwise, since we received a FIN then no
* more input can be expected, send ACK now.
* If connection is half-synchronized
* (ie NEEDSYN flag on) then delay ACK,
* so it may be piggybacked when SYN is sent.
* Otherwise, since we received a FIN then no
* more input can be expected, send ACK now.
*/
if (DELAY_ACK(tp) && (tp->t_flags & TF_NEEDSYN))
callout_reset(tp->tt_delack, tcp_delacktime,
@ -2217,18 +2211,17 @@ tcp_input(m, off0)
*/
if ((thflags & TH_RST) || m->m_flags & (M_BCAST|M_MCAST))
goto drop;
#ifdef INET6
if (isipv6) {
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src))
goto drop;
} else
#endif /* INET6 */
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
} else {
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
goto drop;
}
/* IPv6 anycast check is done at tcp6_input() */
/*
@ -2518,35 +2511,27 @@ tcp_mss(tp, offer)
struct ifnet *ifp;
register int rtt, mss;
u_long bufsize;
struct inpcb *inp;
struct inpcb *inp = tp->t_inpcb;
struct socket *so;
struct rmxp_tao *taop;
int origoffer = offer;
#ifdef INET6
int isipv6;
int min_protoh;
int isipv6 = ((inp->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
size_t min_protoh = isipv6 ?
sizeof (struct ip6_hdr) + sizeof (struct tcphdr) :
sizeof (struct tcpiphdr);
#else
const int isipv6 = 0;
const size_t min_protoh = sizeof (struct tcpiphdr);
#endif
inp = tp->t_inpcb;
#ifdef INET6
isipv6 = ((inp->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
min_protoh = isipv6 ? sizeof (struct ip6_hdr) + sizeof (struct tcphdr)
: sizeof (struct tcpiphdr);
#else
#define min_protoh (sizeof (struct tcpiphdr))
#endif
#ifdef INET6
if (isipv6)
rt = tcp_rtlookup6(&inp->inp_inc);
else
#endif
rt = tcp_rtlookup(&inp->inp_inc);
rt = tcp_rtlookup(&inp->inp_inc);
if (rt == NULL) {
tp->t_maxopd = tp->t_maxseg =
#ifdef INET6
isipv6 ? tcp_v6mssdflt :
#endif /* INET6 */
tcp_mssdflt;
isipv6 ? tcp_v6mssdflt : tcp_mssdflt;
return;
}
ifp = rt->rt_ifp;
@ -2564,11 +2549,7 @@ tcp_mss(tp, offer)
* in this case we use tcp_mssdflt.
*/
if (offer == 0)
offer =
#ifdef INET6
isipv6 ? tcp_v6mssdflt :
#endif /* INET6 */
tcp_mssdflt;
offer = isipv6 ? tcp_v6mssdflt : tcp_mssdflt;
else
/*
* Sanity check: make sure that maxopd will be large
@ -2612,25 +2593,17 @@ tcp_mss(tp, offer)
*/
if (rt->rt_rmx.rmx_mtu)
mss = rt->rt_rmx.rmx_mtu - min_protoh;
else
{
mss =
#ifdef INET6
(isipv6 ? nd_ifinfo[rt->rt_ifp->if_index].linkmtu :
#endif
ifp->if_mtu
#ifdef INET6
)
#endif
- min_protoh;
#ifdef INET6
else {
if (isipv6) {
mss = nd_ifinfo[rt->rt_ifp->if_index].linkmtu -
min_protoh;
if (!in6_localaddr(&inp->in6p_faddr))
mss = min(mss, tcp_v6mssdflt);
} else
#endif
if (!in_localaddr(inp->inp_faddr))
mss = min(mss, tcp_mssdflt);
} else {
mss = ifp->if_mtu - min_protoh;
if (!in_localaddr(inp->inp_faddr))
mss = min(mss, tcp_mssdflt);
}
}
mss = min(mss, offer);
/*
@ -2700,16 +2673,8 @@ tcp_mss(tp, offer)
* Set the slow-start flight size depending on whether this
* is a local network or not.
*/
if (
#ifdef INET6
(isipv6 && in6_localaddr(&inp->in6p_faddr)) ||
(!isipv6 &&
#endif
in_localaddr(inp->inp_faddr)
#ifdef INET6
)
#endif
)
if ((isipv6 && in6_localaddr(&inp->in6p_faddr)) ||
(!isipv6 && in_localaddr(inp->inp_faddr)))
tp->snd_cwnd = mss * ss_fltsz_local;
else
tp->snd_cwnd = mss * ss_fltsz;
@ -2735,68 +2700,55 @@ tcp_mssopt(tp)
{
struct rtentry *rt;
#ifdef INET6
int isipv6;
int min_protoh;
int isipv6 = ((tp->t_inpcb->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
size_t min_protoh = isipv6 ?
sizeof (struct ip6_hdr) + sizeof (struct tcphdr) :
sizeof (struct tcpiphdr);
#else
const int isipv6 = 0;
const size_t min_protoh = sizeof (struct tcpiphdr);
#endif
#ifdef INET6
isipv6 = ((tp->t_inpcb->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
min_protoh = isipv6 ? sizeof (struct ip6_hdr) + sizeof (struct tcphdr)
: sizeof (struct tcpiphdr);
#else
#define min_protoh (sizeof (struct tcpiphdr))
#endif
#ifdef INET6
if (isipv6)
rt = tcp_rtlookup6(&tp->t_inpcb->inp_inc);
else
#endif /* INET6 */
rt = tcp_rtlookup(&tp->t_inpcb->inp_inc);
rt = tcp_rtlookup(&tp->t_inpcb->inp_inc);
if (rt == NULL)
return
#ifdef INET6
isipv6 ? tcp_v6mssdflt :
#endif /* INET6 */
tcp_mssdflt;
return (isipv6 ? tcp_v6mssdflt : tcp_mssdflt);
return rt->rt_ifp->if_mtu - min_protoh;
return (rt->rt_ifp->if_mtu - min_protoh);
}
/*
* Checks for partial ack. If partial ack arrives, force the retransmission
* of the next unacknowledged segment, do not clear tp->t_dupacks, and return
* 1. By setting snd_nxt to ti_ack, this forces retransmission timer to
* be started again. If the ack advances at least to tp->snd_recover, return 0.
* On a partial ack arrives, force the retransmission of the
* next unacknowledged segment. Do not clear tp->t_dupacks.
* By setting snd_nxt to ti_ack, this forces retransmission timer to
* be started again.
*/
static int
tcp_newreno(tp, th)
static void
tcp_newreno_partial_ack(tp, th)
struct tcpcb *tp;
struct tcphdr *th;
{
if (SEQ_LT(th->th_ack, tp->snd_recover)) {
tcp_seq onxt = tp->snd_nxt;
u_long ocwnd = tp->snd_cwnd;
tcp_seq onxt = tp->snd_nxt;
u_long ocwnd = tp->snd_cwnd;
callout_stop(tp->tt_rexmt);
tp->t_rtttime = 0;
tp->snd_nxt = th->th_ack;
/*
* Set snd_cwnd to one segment beyond acknowledged offset
* (tp->snd_una has not yet been updated when this function
* is called)
*/
tp->snd_cwnd = tp->t_maxseg + (th->th_ack - tp->snd_una);
(void) tcp_output(tp);
tp->snd_cwnd = ocwnd;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
/*
* Partial window deflation. Relies on fact that tp->snd_una
* not updated yet.
*/
tp->snd_cwnd -= (th->th_ack - tp->snd_una - tp->t_maxseg);
return (1);
}
return (0);
callout_stop(tp->tt_rexmt);
tp->t_rtttime = 0;
tp->snd_nxt = th->th_ack;
/*
* Set snd_cwnd to one segment beyond acknowledged offset.
* (tp->snd_una has not yet been updated when this function is called.)
*/
tp->snd_cwnd = tp->t_maxseg + (th->th_ack - tp->snd_una);
(void) tcp_output(tp);
tp->snd_cwnd = ocwnd;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
/*
* Partial window deflation. Relies on fact that tp->snd_una
* not updated yet.
*/
tp->snd_cwnd -= (th->th_ack - tp->snd_una - tp->t_maxseg);
}