/*
* Set up the socket for TCP offload.
*/
void
offload_socket(struct socket *so, struct toepcb *toep)
{
struct tom_data *td = toep->td;
struct inpcb *inp = sotoinpcb(so);
struct tcpcb *tp = intotcpcb(inp);
struct sockbuf *sb;
INP_WLOCK_ASSERT(inp);
/* Update socket */
sb = &so->so_snd;
SOCKBUF_LOCK(sb);
sb->sb_flags |= SB_NOCOALESCE;
SOCKBUF_UNLOCK(sb);
sb = &so->so_rcv;
SOCKBUF_LOCK(sb);
sb->sb_flags |= SB_NOCOALESCE;
SOCKBUF_UNLOCK(sb);
/* Update TCP PCB */
tp->tod = &td->tod;
tp->t_toe = toep;
tp->t_flags |= TF_TOE;
/* Install an extra hold on inp */
toep->inp = inp;
toepcb_set_flag(toep, TPF_ATTACHED);
in_pcbref(inp);
/* Add the TOE PCB to the active list */
mtx_lock(&td->toep_list_lock);
TAILQ_INSERT_HEAD(&td->toep_list, toep, link);
mtx_unlock(&td->toep_list_lock);
}
static int
udp6_append(struct inpcb *inp, struct mbuf *n, int off,
struct sockaddr_in6 *fromsa)
{
struct socket *so;
struct mbuf *opts;
struct udpcb *up;
INP_LOCK_ASSERT(inp);
/*
* Engage the tunneling protocol.
*/
up = intoudpcb(inp);
if (up->u_tun_func != NULL) {
in_pcbref(inp);
INP_RUNLOCK(inp);
(*up->u_tun_func)(n, off, inp, (struct sockaddr *)fromsa,
up->u_tun_ctx);
INP_RLOCK(inp);
return (in_pcbrele_rlocked(inp));
}
#ifdef IPSEC
/* Check AH/ESP integrity. */
if (ipsec6_in_reject(n, inp)) {
m_freem(n);
return (0);
}
#endif /* IPSEC */
#ifdef MAC
if (mac_inpcb_check_deliver(inp, n) != 0) {
m_freem(n);
return (0);
}
#endif
opts = NULL;
if (inp->inp_flags & INP_CONTROLOPTS ||
inp->inp_socket->so_options & SO_TIMESTAMP)
ip6_savecontrol(inp, n, &opts);
m_adj(n, off + sizeof(struct udphdr));
so = inp->inp_socket;
SOCKBUF_LOCK(&so->so_rcv);
if (sbappendaddr_locked(&so->so_rcv, (struct sockaddr *)fromsa, n,
opts) == 0) {
SOCKBUF_UNLOCK(&so->so_rcv);
m_freem(n);
if (opts)
m_freem(opts);
UDPSTAT_INC(udps_fullsock);
} else
sorwakeup_locked(so);
return (0);
}
int
tcp_inpinfo_lock_add(struct inpcb *inp)
{
in_pcbref(inp);
INP_WUNLOCK(inp);
INP_INFO_RLOCK(&V_tcbinfo);
INP_WLOCK(inp);
if (inp->inp_flags & (INP_TIMEWAIT | INP_DROPPED)) {
return(1);
}
return(0);
}
static int
udp_pcblist(SYSCTL_HANDLER_ARGS)
{
int error, i, n;
struct inpcb *inp, **inp_list;
inp_gen_t gencnt;
struct xinpgen xig;
/*
* The process of preparing the PCB list is too time-consuming and
* resource-intensive to repeat twice on every request.
*/
if (req->oldptr == 0) {
n = V_udbinfo.ipi_count;
n += imax(n / 8, 10);
req->oldidx = 2 * (sizeof xig) + n * sizeof(struct xinpcb);
return (0);
}
if (req->newptr != 0)
return (EPERM);
/*
* OK, now we're committed to doing something.
*/
INP_INFO_RLOCK(&V_udbinfo);
gencnt = V_udbinfo.ipi_gencnt;
n = V_udbinfo.ipi_count;
INP_INFO_RUNLOCK(&V_udbinfo);
error = sysctl_wire_old_buffer(req, 2 * (sizeof xig)
+ n * sizeof(struct xinpcb));
if (error != 0)
return (error);
xig.xig_len = sizeof xig;
xig.xig_count = n;
xig.xig_gen = gencnt;
xig.xig_sogen = so_gencnt;
error = SYSCTL_OUT(req, &xig, sizeof xig);
if (error)
return (error);
inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK);
if (inp_list == 0)
return (ENOMEM);
INP_INFO_RLOCK(&V_udbinfo);
for (inp = LIST_FIRST(V_udbinfo.ipi_listhead), i = 0; inp && i < n;
inp = LIST_NEXT(inp, inp_list)) {
INP_WLOCK(inp);
if (inp->inp_gencnt <= gencnt &&
cr_canseeinpcb(req->td->td_ucred, inp) == 0) {
in_pcbref(inp);
inp_list[i++] = inp;
}
INP_WUNLOCK(inp);
}
INP_INFO_RUNLOCK(&V_udbinfo);
n = i;
error = 0;
for (i = 0; i < n; i++) {
inp = inp_list[i];
INP_RLOCK(inp);
if (inp->inp_gencnt <= gencnt) {
struct xinpcb xi;
bzero(&xi, sizeof(xi));
xi.xi_len = sizeof xi;
/* XXX should avoid extra copy */
bcopy(inp, &xi.xi_inp, sizeof *inp);
if (inp->inp_socket)
sotoxsocket(inp->inp_socket, &xi.xi_socket);
xi.xi_inp.inp_gencnt = inp->inp_gencnt;
INP_RUNLOCK(inp);
error = SYSCTL_OUT(req, &xi, sizeof xi);
} else
INP_RUNLOCK(inp);
}
INP_INFO_WLOCK(&V_udbinfo);
for (i = 0; i < n; i++) {
inp = inp_list[i];
INP_RLOCK(inp);
if (!in_pcbrele_rlocked(inp))
INP_RUNLOCK(inp);
}
INP_INFO_WUNLOCK(&V_udbinfo);
if (!error) {
/*
* Give the user an updated idea of our state. If the
* generation differs from what we told her before, she knows
* that something happened while we were processing this
* request, and it might be necessary to retry.
*/
INP_INFO_RLOCK(&V_udbinfo);
xig.xig_gen = V_udbinfo.ipi_gencnt;
xig.xig_sogen = so_gencnt;
xig.xig_count = V_udbinfo.ipi_count;
INP_INFO_RUNLOCK(&V_udbinfo);
error = SYSCTL_OUT(req, &xig, sizeof xig);
}
free(inp_list, M_TEMP);
return (error);
}
void
tcp_timer_rexmt(void * xtp)
{
struct tcpcb *tp = xtp;
CURVNET_SET(tp->t_vnet);
int rexmt;
int headlocked;
struct inpcb *inp;
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
INP_INFO_RLOCK(&V_tcbinfo);
inp = tp->t_inpcb;
/*
* XXXRW: While this assert is in fact correct, bugs in the tcpcb
* tear-down mean we need it as a work-around for races between
* timers and tcp_discardcb().
*
* KASSERT(inp != NULL, ("tcp_timer_rexmt: inp == NULL"));
*/
if (inp == NULL) {
tcp_timer_race++;
INP_INFO_RUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
INP_WLOCK(inp);
if ((inp->inp_flags & INP_DROPPED) || callout_pending(&tp->t_timers->tt_rexmt)
|| !callout_active(&tp->t_timers->tt_rexmt)) {
INP_WUNLOCK(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_rexmt);
tcp_free_sackholes(tp);
/*
* Retransmission timer went off. Message has not
* been acked within retransmit interval. Back off
* to a longer retransmit interval and retransmit one segment.
*/
if (++tp->t_rxtshift > TCP_MAXRXTSHIFT) {
tp->t_rxtshift = TCP_MAXRXTSHIFT;
TCPSTAT_INC(tcps_timeoutdrop);
in_pcbref(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
INP_WUNLOCK(inp);
INP_INFO_WLOCK(&V_tcbinfo);
INP_WLOCK(inp);
if (in_pcbrele_wlocked(inp)) {
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
if (inp->inp_flags & INP_DROPPED) {
INP_WUNLOCK(inp);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
tp = tcp_drop(tp, tp->t_softerror ?
tp->t_softerror : ETIMEDOUT);
headlocked = 1;
goto out;
}
INP_INFO_RUNLOCK(&V_tcbinfo);
headlocked = 0;
if (tp->t_rxtshift == 1) {
/*
* first retransmit; record ssthresh and cwnd so they can
* be recovered if this turns out to be a "bad" retransmit.
* A retransmit is considered "bad" if an ACK for this
* segment is received within RTT/2 interval; the assumption
* here is that the ACK was already in flight. See
* "On Estimating End-to-End Network Path Properties" by
* Allman and Paxson for more details.
*/
tp->snd_cwnd_prev = tp->snd_cwnd;
tp->snd_ssthresh_prev = tp->snd_ssthresh;
tp->snd_recover_prev = tp->snd_recover;
if (IN_FASTRECOVERY(tp->t_flags))
tp->t_flags |= TF_WASFRECOVERY;
else
tp->t_flags &= ~TF_WASFRECOVERY;
if (IN_CONGRECOVERY(tp->t_flags))
tp->t_flags |= TF_WASCRECOVERY;
else
tp->t_flags &= ~TF_WASCRECOVERY;
tp->t_badrxtwin = ticks + (tp->t_srtt >> (TCP_RTT_SHIFT + 1));
tp->t_flags |= TF_PREVVALID;
} else