/*
- * Copyright (c) 2000-2007 Apple Inc. All rights reserved.
+ * Copyright (c) 2000-2010 Apple Inc. All rights reserved.
*
* @APPLE_OSREFERENCE_LICENSE_HEADER_START@
*
SYSCTL_INT(_net_inet_tcp, OID_AUTO, newreno, CTLFLAG_RW, &tcp_do_newreno,
0, "Enable NewReno Algorithms");
+int tcp_do_tso = 1;
+SYSCTL_INT(_net_inet_tcp, OID_AUTO, tso, CTLFLAG_RW,
+ &tcp_do_tso, 0, "Enable TCP Segmentation Offload");
+
+
int tcp_ecn_outbound = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, ecn_initiate_out, CTLFLAG_RW, &tcp_ecn_outbound,
0, "Initiate ECN for outbound connections");
SYSCTL_INT(_net_inet_tcp, OID_AUTO, socket_unlocked_on_output, CTLFLAG_RW, &tcp_output_unlocked,
0, "Unlock TCP when sending packets down to IP");
-static long packchain_newlist = 0;
-static long packchain_looped = 0;
-static long packchain_sent = 0;
-
+static int32_t packchain_newlist = 0;
+static int32_t packchain_looped = 0;
+static int32_t packchain_sent = 0;
/* temporary: for testing */
#if IPSEC
#endif
extern int slowlink_wsize; /* window correction for slow links */
-extern u_long route_generation;
#if IPFIREWALL
extern int fw_enable; /* firewall check for packet chaining */
extern int fw_bypass; /* firewall check: disable packet chaining if there is rules */
#endif /* IPFIREWALL */
extern vm_size_t so_cache_zone_element_size;
+#if RANDOM_IP_ID
+extern int ip_use_randomid;
+#endif /* RANDOM_IP_ID */
+extern u_int32_t dlil_filter_count;
+extern u_int32_t kipf_count;
static int tcp_ip_output(struct socket *, struct tcpcb *, struct mbuf *, int,
- struct mbuf *, int);
+ struct mbuf *, int, int, int32_t);
static __inline__ u_int16_t
get_socket_id(struct socket * s)
if (so_cache_zone_element_size == 0) {
return (0);
}
- val = (u_int16_t)(((u_int32_t)s) / so_cache_zone_element_size);
+ val = (u_int16_t)(((uintptr_t)s) / so_cache_zone_element_size);
if (val == 0) {
val = 0xffff;
}
tcp_output(struct tcpcb *tp)
{
struct socket *so = tp->t_inpcb->inp_socket;
- long len, recwin, sendwin;
- int off, flags, error;
+ int32_t len, recwin, sendwin, off;
+ int flags, error;
register struct mbuf *m;
struct ip *ip = NULL;
register struct ipovly *ipov = NULL;
unsigned ipoptlen, optlen, hdrlen;
int idle, sendalot, lost = 0;
int i, sack_rxmit;
+ int tso = 0;
int sack_bytes_rxmt;
struct sackhole *p;
-
+#ifdef IPSEC
+ unsigned ipsec_optlen = 0;
+#endif
int maxburst = TCP_MAXBURST;
int last_off = 0;
int m_off;
* come back before the TCP connection times out).
*/
rt = tp->t_inpcb->inp_route.ro_rt;
- if (rt != NULL && rt->generation_id != route_generation) {
+ if (rt != NULL && (!(rt->rt_flags & RTF_UP) ||
+ rt->generation_id != route_generation)) {
struct ifnet *ifp;
+ struct in_ifaddr *ia;
/* disable multipages at the socket */
somultipages(so, FALSE);
+ /* Disable TSO for the socket until we know more */
+ tp->t_flags &= ~TF_TSO;
+
/* check that the source address is still valid */
- if (ifa_foraddr(tp->t_inpcb->inp_laddr.s_addr) == 0) {
+ if ((ia = ifa_foraddr(tp->t_inpcb->inp_laddr.s_addr)) == NULL) {
if (tp->t_state >= TCPS_CLOSE_WAIT) {
tcp_drop(tp, EADDRNOTAVAIL);
else
return(0); /* silently ignore, keep data in socket: address may be back */
}
+ ifafree(&ia->ia_ifa);
/*
* Address is still valid; check for multipages capability
* again in case the outgoing interface has changed.
*/
- lck_mtx_lock(rt_mtx);
- rt = tp->t_inpcb->inp_route.ro_rt;
- if (rt != NULL && (ifp = rt->rt_ifp) != NULL)
+ RT_LOCK(rt);
+ if ((ifp = rt->rt_ifp) != NULL) {
somultipages(so, (ifp->if_hwassist & IFNET_MULTIPAGES));
- if (rt != NULL && rt->generation_id != route_generation)
+ tcp_set_tso(tp, ifp);
+ }
+ if (rt->rt_flags & RTF_UP)
rt->generation_id = route_generation;
/*
* See if we should do MTU discovery. Don't do it if:
else
tp->t_flags |= TF_PMTUD;
- lck_mtx_unlock(rt_mtx);
+ RT_UNLOCK(rt);
}
}
p = NULL;
if (tp->sack_enable && IN_FASTRECOVERY(tp) &&
(p = tcp_sack_output(tp, &sack_bytes_rxmt))) {
- long cwin;
+ int32_t cwin;
cwin = min(tp->snd_wnd, tp->snd_cwnd) - sack_bytes_rxmt;
if (cwin < 0)
goto after_sack_rexmit;
} else
/* Can rexmit part of the current hole */
- len = ((long)ulmin(cwin,
+ len = ((int32_t)min(cwin,
tp->snd_recover - p->rxmit));
} else
- len = ((long)ulmin(cwin, p->end - p->rxmit));
- off = p->rxmit - tp->snd_una;
+ len = ((int32_t)min(cwin, p->end - p->rxmit));
if (len > 0) {
+ off = p->rxmit - tp->snd_una; /* update off only if we really transmit SACK data */
sack_rxmit = 1;
sendalot = 1;
tcpstat.tcps_sack_rexmits++;
tcpstat.tcps_sack_rexmit_bytes +=
min(len, tp->t_maxseg);
}
+ else
+ len = 0;
}
after_sack_rexmit:
/*
*/
if (sack_rxmit == 0) {
if (sack_bytes_rxmt == 0)
- len = ((long)ulmin(so->so_snd.sb_cc, sendwin) - off);
+ len = min(so->so_snd.sb_cc, sendwin) - off;
else {
- long cwin;
+ int32_t cwin;
/*
* We are inside of a SACK recovery episode and are
* sending new data, having retransmitted all the
* data possible in the scoreboard.
*/
- len = ((long)ulmin(so->so_snd.sb_cc, tp->snd_wnd)
- - off);
+ len = min(so->so_snd.sb_cc, tp->snd_wnd)
+ - off;
/*
* Don't remove this (len > 0) check !
* We explicitly check for len > 0 here (although it
sack_bytes_rxmt;
if (cwin < 0)
cwin = 0;
- len = lmin(len, cwin);
+ len = imin(len, cwin);
}
+ else
+ len = 0;
}
}
error = tcp_ip_output(so, tp, packetlist,
packchain_listadd, tp_inp_options,
- (so_options & SO_DONTROUTE));
+ (so_options & SO_DONTROUTE), (sack_rxmit | (sack_bytes_rxmt != 0)), 0);
tp->t_flags &= ~TF_SENDINPROG;
}
}
/*
- * len will be >= 0 after this point. Truncate to the maximum
- * segment length and ensure that FIN is removed if the length
- * no longer contains the last data byte.
+ * Truncate to the maximum segment length or enable TCP Segmentation
+ * Offloading (if supported by hardware) and ensure that FIN is removed
+ * if the length no longer contains the last data byte.
+ *
+ * TSO may only be used if we are in a pure bulk sending state. The
+ * presence of TCP-MD5, SACK retransmits, SACK advertizements, ipfw rules
+ * and IP options prevent using TSO. With TSO the TCP header is the same
+ * (except for the sequence number) for all generated packets. This
+ * makes it impossible to transmit any options which vary per generated
+ * segment or packet.
+ *
+ * The length of TSO bursts is limited to TCP_MAXWIN. That limit and
+ * removal of FIN (if not already catched here) are handled later after
+ * the exact length of the TCP options are known.
*/
+#if IPSEC
+ /*
+ * Pre-calculate here as we save another lookup into the darknesses
+ * of IPsec that way and can actually decide if TSO is ok.
+ */
+ if (ipsec_bypass == 0)
+ ipsec_optlen = ipsec_hdrsiz_tcp(tp);
+#endif
+
if (len > tp->t_maxseg) {
- len = tp->t_maxseg;
- sendalot = 1;
+ if ((tp->t_flags & TF_TSO) && tcp_do_tso &&
+#if RANDOM_IP_ID
+ ip_use_randomid &&
+#endif /* RANDOM_IP_ID */
+ kipf_count == 0 && dlil_filter_count == 0 &&
+ tp->rcv_numsacks == 0 && sack_rxmit == 0 && sack_bytes_rxmt == 0 &&
+ tp->t_inpcb->inp_options == NULL &&
+ tp->t_inpcb->in6p_options == NULL
+#if IPSEC
+ && ipsec_optlen == 0
+#endif
+#if IPFIREWALL
+ && (fw_enable == 0 || fw_bypass)
+#endif
+ ) {
+ tso = 1;
+ sendalot = 0;
+ } else {
+ len = tp->t_maxseg;
+ sendalot = 1;
+ tso = 0;
+ }
}
if (sack_rxmit) {
if (SEQ_LT(p->rxmit + len, tp->snd_una + so->so_snd.sb_cc))
* Sender silly window avoidance. We transmit under the following
* conditions when len is non-zero:
*
- * - We have a full segment
+ * - We have a full segment (or more with TSO)
* - This is the last buffer in a write()/send() and we are
* either idle or running NODELAY
* - we've timed out (e.g. persist timer)
* - we need to retransmit
*/
if (len) {
- if (len == tp->t_maxseg) {
+ if (len >= tp->t_maxseg) {
tp->t_flags |= TF_MAXSEGSNT;
goto send;
}
* taking into account that we are limited by
* TCP_MAXWIN << tp->rcv_scale.
*/
- long adv = lmin(recwin, (long)TCP_MAXWIN << tp->rcv_scale) -
+ int32_t adv = imin(recwin, (int)TCP_MAXWIN << tp->rcv_scale) -
(tp->rcv_adv - tp->rcv_nxt);
- if (adv >= (long) (2 * tp->t_maxseg)) {
+ if (adv >= (int32_t) (2 * tp->t_maxseg)) {
/*
* Update only if the resulting scaled value of the window changed, or
if ((tp->last_ack_sent != tp->rcv_nxt) || (((recwin + adv) >> tp->rcv_scale) > recwin))
goto send;
}
- if (2 * adv >= (long) so->so_rcv.sb_hiwat)
+ if (2 * adv >= (int32_t) so->so_rcv.sb_hiwat)
goto send;
}
* after the retransmission timer has been turned off. Make sure
* that the retransmission timer is set.
*/
- if (tp->sack_enable && SEQ_GT(tp->snd_max, tp->snd_una) &&
+ if (tp->sack_enable && (tp->t_state >= TCPS_ESTABLISHED) && SEQ_GT(tp->snd_max, tp->snd_una) &&
tp->t_timer[TCPT_REXMT] == 0 &&
tp->t_timer[TCPT_PERSIST] == 0) {
tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
tp->t_flags |= TF_SENDINPROG;
error = tcp_ip_output(so, tp, packetlist, packchain_listadd,
- tp_inp_options, (so_options & SO_DONTROUTE));
+ tp_inp_options, (so_options & SO_DONTROUTE), (sack_rxmit | (sack_bytes_rxmt != 0)), recwin);
tp->t_flags &= ~TF_SENDINPROG;
}
ipoptlen = 0;
}
#if IPSEC
- if (ipsec_bypass == 0)
- ipoptlen += ipsec_hdrsiz_tcp(tp);
+ ipoptlen += ipsec_optlen;
#endif
/*
* bump the packet length beyond the t_maxopd length.
* Clear the FIN bit because we cut off the tail of
* the segment.
+ *
+ * When doing TSO limit a burst to TCP_MAXWIN minus the
+ * IP, TCP and Options length to keep ip->ip_len from
+ * overflowing. Prevent the last segment from being
+ * fractional thus making them all equal sized and set
+ * the flag to continue sending. TSO is disabled when
+ * IP options or IPSEC are present.
*/
if (len + optlen + ipoptlen > tp->t_maxopd) {
/*
* If there is still more to send, don't close the connection.
*/
flags &= ~TH_FIN;
- len = tp->t_maxopd - optlen - ipoptlen;
- sendalot = 1;
+ if (tso) {
+ int32_t tso_maxlen;
+
+ tso_maxlen = tp->tso_max_segment_size ? tp->tso_max_segment_size : TCP_MAXWIN;
+
+ if (len > tso_maxlen - hdrlen - optlen) {
+ len = tso_maxlen - hdrlen - optlen;
+ len = len - (len % (tp->t_maxopd - optlen));
+ sendalot = 1;
+ } else if (tp->t_flags & TF_NEEDFIN)
+ sendalot = 1;
+ } else {
+ len = tp->t_maxopd - optlen - ipoptlen;
+ sendalot = 1;
+ }
}
/*#ifdef DIAGNOSTIC*/
if (len) {
if (tp->t_force && len == 1)
tcpstat.tcps_sndprobe++;
- else if (SEQ_LT(tp->snd_nxt, tp->snd_max)) {
+ else if (SEQ_LT(tp->snd_nxt, tp->snd_max) || sack_rxmit) {
tcpstat.tcps_sndrexmitpack++;
tcpstat.tcps_sndrexmitbyte += len;
} else {
m->m_len = hdrlen;
}
/* makes sure we still have data left to be sent at this point */
- if (so->so_snd.sb_mb == NULL || off == -1) {
+ if (so->so_snd.sb_mb == NULL || off < 0) {
if (m != NULL) m_freem(m);
error = 0; /* should we return an error? */
goto out;
* m_copym_with_hdrs to avoid rescanning from the beginning of the socket buffer mbuf list.
* setting the mbuf pointer to NULL is sufficient to disable the hint mechanism.
*/
- if (m_head != so->so_snd.sb_mb || last_off != off)
+ if (m_head != so->so_snd.sb_mb || sack_rxmit || last_off != off)
m_last = NULL;
last_off = off + len;
m_head = so->so_snd.sb_mb;
* m_copym_with_hdrs will always return the last mbuf pointer and the offset into it that
* it acted on to fullfill the current request, whether a valid 'hint' was passed in or not
*/
- if ((m = m_copym_with_hdrs(so->so_snd.sb_mb, off, (int) len, M_DONTWAIT, &m_last, &m_off)) == NULL) {
+ if ((m = m_copym_with_hdrs(so->so_snd.sb_mb, off, len, M_DONTWAIT, &m_last, &m_off)) == NULL) {
error = ENOBUFS;
goto out;
}
* Calculate receive window. Don't shrink window,
* but avoid silly window syndrome.
*/
- if (recwin < (long)(so->so_rcv.sb_hiwat / 4) && recwin < (long)tp->t_maxseg)
+ if (recwin < (int32_t)(so->so_rcv.sb_hiwat / 4) && recwin < (int)tp->t_maxseg)
recwin = 0;
- if (recwin < (long)(tp->rcv_adv - tp->rcv_nxt))
- recwin = (long)(tp->rcv_adv - tp->rcv_nxt);
+ if (recwin < (int32_t)(tp->rcv_adv - tp->rcv_nxt))
+ recwin = (int32_t)(tp->rcv_adv - tp->rcv_nxt);
if (tp->t_flags & TF_SLOWLINK && slowlink_wsize > 0) {
- if (recwin > (long)slowlink_wsize)
+ if (recwin > (int32_t)slowlink_wsize)
recwin = slowlink_wsize;
- th->th_win = htons((u_short) (recwin>>tp->rcv_scale));
+ th->th_win = htons((u_short) (recwin>>tp->rcv_scale));
}
else {
- if (recwin > (long)(TCP_MAXWIN << tp->rcv_scale))
- recwin = (long)(TCP_MAXWIN << tp->rcv_scale);
+ if (recwin > (int32_t)(TCP_MAXWIN << tp->rcv_scale))
+ recwin = (int32_t)(TCP_MAXWIN << tp->rcv_scale);
th->th_win = htons((u_short) (recwin>>tp->rcv_scale));
}
htons((u_short)(optlen + len)));
}
+ /*
+ * Enable TSO and specify the size of the segments.
+ * The TCP pseudo header checksum is always provided.
+ * XXX: Fixme: This is currently not the case for IPv6.
+ */
+ if (tso) {
+#if INET6
+ if (isipv6)
+ m->m_pkthdr.csum_flags = CSUM_TSO_IPV6;
+ else
+#endif /* INET6 */
+ m->m_pkthdr.csum_flags = CSUM_TSO_IPV4;
+
+ m->m_pkthdr.tso_segsz = tp->t_maxopd - optlen;
+ }
+ else
+ m->m_pkthdr.tso_segsz = 0;
+
/*
* In transmit state, time the transmission and arrange for
* the retransmit. In persist state, just set snd_max.
}
#endif /*IPSEC*/
m->m_pkthdr.socket_id = socket_id;
+
+#if PKT_PRIORITY
+ set_traffic_class(m, so, MBUF_TC_NONE);
+#endif /* PKT_PRIORITY */
error = ip6_output(m,
inp6_pktopts,
&tp->t_inpcb->in6p_route,
lost = 0;
m->m_pkthdr.socket_id = socket_id;
m->m_nextpkt = NULL;
+#if PKT_PRIORITY
+ set_traffic_class(m, so, MBUF_TC_NONE);
+#endif /* PKT_PRIORITY */
tp->t_pktlist_sentlen += len;
tp->t_lastchain++;
if (tp->t_pktlist_head != NULL) {
error = tcp_ip_output(so, tp, packetlist,
packchain_listadd, tp_inp_options,
- (so_options & SO_DONTROUTE));
+ (so_options & SO_DONTROUTE), (sack_rxmit | (sack_bytes_rxmt != 0)), recwin);
tp->t_flags &= ~TF_SENDINPROG;
if (error) {
packchain_looped++;
tcpstat.tcps_sndtotal++;
- if (recwin > 0 && SEQ_GT(tp->rcv_nxt+recwin, tp->rcv_adv))
- tp->rcv_adv = tp->rcv_nxt + recwin;
- tp->last_ack_sent = tp->rcv_nxt;
- tp->t_flags &= ~(TF_ACKNOW|TF_DELACK);
goto again;
}
}
TCP_PKTLIST_CLEAR(tp);
if (error == ENOBUFS) {
- if (!tp->t_timer[TCPT_REXMT] &&
- !tp->t_timer[TCPT_PERSIST])
- tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
- tcp_quench(tp->t_inpcb, 0);
+ if (!tp->t_timer[TCPT_REXMT] &&
+ !tp->t_timer[TCPT_PERSIST])
+ tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
+
+ tp->snd_cwnd = tp->t_maxseg;
+ tp->t_bytes_acked = 0;
+
KERNEL_DEBUG(DBG_FNC_TCP_OUTPUT | DBG_FUNC_END, 0,0,0,0,0);
return (0);
}
* for us. tcp_mtudisc() will, as its last action,
* initiate retransmission, so it is important to
* not do so here.
+ *
+ * If TSO was active we either got an interface
+ * without TSO capabilits or TSO was turned off.
+ * Disable it for this connection as too and
+ * immediatly retry with MSS sized segments generated
+ * by this function.
*/
+ if (tso)
+ tp->t_flags &= ~TF_TSO;
+
tcp_mtudisc(tp->t_inpcb, 0);
KERNEL_DEBUG(DBG_FNC_TCP_OUTPUT | DBG_FUNC_END, 0,0,0,0,0);
return 0;
* Data sent (as far as we can tell).
* If this advertises a larger window than any other segment,
* then remember the size of the advertised window.
- * Any pending ACK has now been sent.
+ * Make sure ACK/DELACK conditions are cleared before
+ * we unlock the socket.
+ * NOTE: for now, this is done in tcp_ip_output for IPv4
*/
- if (recwin > 0 && SEQ_GT(tp->rcv_nxt+recwin, tp->rcv_adv))
- tp->rcv_adv = tp->rcv_nxt + recwin;
- tp->last_ack_sent = tp->rcv_nxt;
- tp->t_flags &= ~(TF_ACKNOW|TF_DELACK);
+#if INET6
+ if (isipv6) {
+ if (recwin > 0 && SEQ_GT(tp->rcv_nxt + recwin, tp->rcv_adv))
+ tp->rcv_adv = tp->rcv_nxt + recwin;
+ tp->last_ack_sent = tp->rcv_nxt;
+ tp->t_flags &= ~(TF_ACKNOW | TF_DELACK);
+ }
+#endif
KERNEL_DEBUG(DBG_FNC_TCP_OUTPUT | DBG_FUNC_END,0,0,0,0,0);
if (sendalot && (!tcp_do_newreno || --maxburst))
static int
tcp_ip_output(struct socket *so, struct tcpcb *tp, struct mbuf *pkt,
- int cnt, struct mbuf *opt, int flags)
+ int cnt, struct mbuf *opt, int flags, int sack_in_progress, int recwin)
{
int error = 0;
boolean_t chain;
boolean_t unlocked = FALSE;
+ struct inpcb *inp = tp->t_inpcb;
+ struct ip_out_args ipoa;
+ struct route ro;
+#if CONFIG_OUT_IF
+ unsigned int outif;
+#endif /* CONFIG_OUT_IF */
- /* Make sure ACK/DELACK conditions are cleared before
+ /* If socket was bound to an ifindex, tell ip_output about it */
+ ipoa.ipoa_ifscope = (inp->inp_flags & INP_BOUND_IF) ?
+ inp->inp_boundif : IFSCOPE_NONE;
+ flags |= IP_OUTARGS;
+
+ /* Copy the cached route and take an extra reference */
+ inp_route_copyout(inp, &ro);
+
+ /*
+ * Data sent (as far as we can tell).
+ * If this advertises a larger window than any other segment,
+ * then remember the size of the advertised window.
+ * Make sure ACK/DELACK conditions are cleared before
* we unlock the socket.
*/
-
+ if (recwin > 0 && SEQ_GT(tp->rcv_nxt + recwin, tp->rcv_adv))
+ tp->rcv_adv = tp->rcv_nxt + recwin;
+ tp->last_ack_sent = tp->rcv_nxt;
tp->t_flags &= ~(TF_ACKNOW | TF_DELACK);
+
/*
- * If allowed, unlock TCP socket while in IP
+ * If allowed, unlock TCP socket while in IP
* but only if the connection is established and
* if we're not sending from an upcall.
- */
-
+ */
if (tcp_output_unlocked && ((so->so_flags & SOF_UPCALLINUSE) == 0) &&
- (tp->t_state == TCPS_ESTABLISHED)) {
- unlocked = TRUE;
- socket_unlock(so, 0);
+ (tp->t_state == TCPS_ESTABLISHED) && (sack_in_progress == 0)) {
+ unlocked = TRUE;
+ socket_unlock(so, 0);
}
-
+
/*
* Don't send down a chain of packets when:
* - TCP chaining is disabled
*/
cnt = 0;
}
-#if CONFIG_FORCE_OUT_IFP
- error = ip_output_list(pkt, cnt, opt, &tp->t_inpcb->inp_route,
- flags, 0, tp->t_inpcb->pdp_ifp);
-#else
- error = ip_output_list(pkt, cnt, opt, &tp->t_inpcb->inp_route,
- flags, 0, NULL);
-#endif
+
+ error = ip_output_list(pkt, cnt, opt, &ro, flags, 0, &ipoa);
if (chain || error) {
/*
* If we sent down a chain then we are done since
if (unlocked)
socket_lock(so, 0);
+ /* Synchronize cached PCB route */
+ inp_route_copyin(inp, &ro);
+
return (error);
}
projects / apple / xnu.git / blobdiff