void
ieee80211_notify_country(struct ieee80211vap *vap,
const uint8_t bssid[IEEE80211_ADDR_LEN], const uint8_t cc[2])
{
struct ifnet *ifp = vap->iv_ifp;
struct ieee80211_country_event iev;
memset(&iev, 0, sizeof(iev));
IEEE80211_ADDR_COPY(iev.iev_addr, bssid);
iev.iev_cc[0] = cc[0];
iev.iev_cc[1] = cc[1];
CURVNET_SET(ifp->if_vnet);
rt_ieee80211msg(ifp, RTM_IEEE80211_COUNTRY, &iev, sizeof(iev));
CURVNET_RESTORE();
}
void
ieee80211_notify_radar(struct ieee80211com *ic,
const struct ieee80211_channel *c)
{
struct ifnet *ifp = ic->ic_ifp;
struct ieee80211_radar_event iev;
memset(&iev, 0, sizeof(iev));
iev.iev_flags = c->ic_flags;
iev.iev_freq = c->ic_freq;
iev.iev_ieee = c->ic_ieee;
CURVNET_SET(ifp->if_vnet);
rt_ieee80211msg(ifp, RTM_IEEE80211_RADAR, &iev, sizeof(iev));
CURVNET_RESTORE();
}
/*
* Tcp protocol timeout routine called every 500 ms.
* Updates timestamps used for TCP
* causes finite state machine actions if timers expire.
*/
void
tcp_slowtimo(void)
{
VNET_ITERATOR_DECL(vnet_iter);
VNET_LIST_RLOCK_NOSLEEP();
VNET_FOREACH(vnet_iter) {
CURVNET_SET(vnet_iter);
INP_INFO_WLOCK(&V_tcbinfo);
(void) tcp_tw_2msl_scan(0);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
}
VNET_LIST_RUNLOCK_NOSLEEP();
}
void
ieee80211_notify_cac(struct ieee80211com *ic,
const struct ieee80211_channel *c, enum ieee80211_notify_cac_event type)
{
struct ifnet *ifp = ic->ic_ifp;
struct ieee80211_cac_event iev;
memset(&iev, 0, sizeof(iev));
iev.iev_flags = c->ic_flags;
iev.iev_freq = c->ic_freq;
iev.iev_ieee = c->ic_ieee;
iev.iev_type = type;
CURVNET_SET(ifp->if_vnet);
rt_ieee80211msg(ifp, RTM_IEEE80211_CAC, &iev, sizeof(iev));
CURVNET_RESTORE();
}
void
ieee80211_notify_csa(struct ieee80211com *ic,
const struct ieee80211_channel *c, int mode, int count)
{
struct ifnet *ifp = ic->ic_ifp;
struct ieee80211_csa_event iev;
memset(&iev, 0, sizeof(iev));
iev.iev_flags = c->ic_flags;
iev.iev_freq = c->ic_freq;
iev.iev_ieee = c->ic_ieee;
iev.iev_mode = mode;
iev.iev_count = count;
CURVNET_SET(ifp->if_vnet);
rt_ieee80211msg(ifp, RTM_IEEE80211_CSA, &iev, sizeof(iev));
CURVNET_RESTORE();
}
/* ARGSUSED */
int
soo_read(struct file *fp, struct uio *uio, struct ucred *active_cred,
int flags, struct thread *td)
{
struct socket *so = fp->f_data;
int error;
#ifdef MAC
error = mac_socket_check_receive(active_cred, so);
if (error)
return (error);
#endif
CURVNET_SET(so->so_vnet);
error = soreceive(so, 0, uio, 0, 0, 0);
CURVNET_RESTORE();
return (error);
}
static void
in_rtqtimo(void *rock)
{
CURVNET_SET((struct vnet *) rock);
int fibnum;
void *newrock;
struct timeval atv;
for (fibnum = 0; fibnum < rt_numfibs; fibnum++) {
newrock = rt_tables_get_rnh(fibnum, AF_INET);
if (newrock != NULL)
in_rtqtimo_one(newrock);
}
atv.tv_usec = 0;
atv.tv_sec = V_rtq_timeout;
callout_reset(&V_rtq_timer, tvtohz(&atv), in_rtqtimo, rock);
CURVNET_RESTORE();
}
ipfpoll(dev_t dev, int events, struct proc *td)
#endif
{
int unit = GET_MINOR(dev);
int revents;
if (unit < 0 || unit > IPL_LOGMAX)
return 0;
revents = 0;
CURVNET_SET(TD_TO_VNET(td));
switch (unit)
{
case IPL_LOGIPF :
case IPL_LOGNAT :
case IPL_LOGSTATE :
#ifdef IPFILTER_LOG
if ((events & (POLLIN | POLLRDNORM)) && ipf_log_canread(&V_ipfmain, unit))
revents |= events & (POLLIN | POLLRDNORM);
#endif
break;
case IPL_LOGAUTH :
if ((events & (POLLIN | POLLRDNORM)) && ipf_auth_waiting(&V_ipfmain))
revents |= events & (POLLIN | POLLRDNORM);
break;
case IPL_LOGSYNC :
if ((events & (POLLIN | POLLRDNORM)) && ipf_sync_canread(&V_ipfmain))
revents |= events & (POLLIN | POLLRDNORM);
if ((events & (POLLOUT | POLLWRNORM)) && ipf_sync_canwrite(&V_ipfmain))
revents |= events & (POLLOUT | POLLWRNORM);
break;
case IPL_LOGSCAN :
case IPL_LOGLOOKUP :
default :
break;
}
if ((revents == 0) && ((events & (POLLIN|POLLRDNORM)) != 0))
selrecord(td, &V_ipfmain.ipf_selwait[unit]);
CURVNET_RESTORE();
return revents;
}
/*
* Detach net80211 state on device detach. Tear down
* all vap's and reclaim all common state prior to the
* device state going away. Note we may call back into
* driver; it must be prepared for this.
*/
void
ieee80211_ifdetach(struct ieee80211com *ic)
{
struct ifnet *ifp = ic->ic_ifp;
struct ieee80211vap *vap;
/*
* This detaches the main interface, but not the vaps.
* Each VAP may be in a separate VIMAGE.
*/
CURVNET_SET(ifp->if_vnet);
if_detach(ifp);
CURVNET_RESTORE();
/*
* The VAP is responsible for setting and clearing
* the VIMAGE context.
*/
while ((vap = TAILQ_FIRST(&ic->ic_vaps)) != NULL)
ieee80211_vap_destroy(vap);
ieee80211_waitfor_parent(ic);
ieee80211_sysctl_detach(ic);
ieee80211_dfs_detach(ic);
ieee80211_regdomain_detach(ic);
ieee80211_scan_detach(ic);
#ifdef IEEE80211_SUPPORT_SUPERG
ieee80211_superg_detach(ic);
#endif
ieee80211_ht_detach(ic);
/* NB: must be called before ieee80211_node_detach */
ieee80211_proto_detach(ic);
ieee80211_crypto_detach(ic);
ieee80211_power_detach(ic);
ieee80211_node_detach(ic);
/* XXX VNET needed? */
ifmedia_removeall(&ic->ic_media);
taskqueue_free(ic->ic_tq);
IEEE80211_TX_LOCK_DESTROY(ic);
IEEE80211_LOCK_DESTROY(ic);
}
void
act_open_failure_cleanup(struct adapter *sc, u_int atid, u_int status)
{
struct toepcb *toep = lookup_atid(sc, atid);
struct inpcb *inp = toep->inp;
struct toedev *tod = &toep->td->tod;
free_atid(sc, atid);
toep->tid = -1;
CURVNET_SET(toep->vnet);
if (status != EAGAIN)
INP_INFO_RLOCK(&V_tcbinfo);
INP_WLOCK(inp);
toe_connect_failed(tod, inp, status);
final_cpl_received(toep); /* unlocks inp */
if (status != EAGAIN)
INP_INFO_RUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
}
/*
* Active open succeeded.
*/
static int
do_act_establish(struct sge_iq *iq, const struct rss_header *rss,
struct mbuf *m)
{
struct adapter *sc = iq->adapter;
const struct cpl_act_establish *cpl = (const void *)(rss + 1);
u_int tid = GET_TID(cpl);
u_int atid = G_TID_TID(ntohl(cpl->tos_atid));
struct toepcb *toep = lookup_atid(sc, atid);
struct inpcb *inp = toep->inp;
KASSERT(m == NULL, ("%s: wasn't expecting payload", __func__));
KASSERT(toep->tid == atid, ("%s: toep tid/atid mismatch", __func__));
CTR3(KTR_CXGBE, "%s: atid %u, tid %u", __func__, atid, tid);
free_atid(sc, atid);
CURVNET_SET(toep->vnet);
INP_WLOCK(inp);
toep->tid = tid;
insert_tid(sc, tid, toep, inp->inp_vflag & INP_IPV6 ? 2 : 1);
if (inp->inp_flags & INP_DROPPED) {
/* socket closed by the kernel before hw told us it connected */
send_flowc_wr(toep, NULL);
send_reset(sc, toep, be32toh(cpl->snd_isn));
goto done;
}
make_established(toep, be32toh(cpl->snd_isn) - 1,
be32toh(cpl->rcv_isn) - 1, cpl->tcp_opt);
if (toep->ulp_mode == ULP_MODE_TLS)
tls_establish(toep);
done:
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return (0);
}
void
tcp_timer_delack(void *xtp)
{
struct tcpcb *tp = xtp;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
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_delack: inp == NULL"));
*/
if (inp == NULL) {
tcp_timer_race++;
CURVNET_RESTORE();
return;
}
INP_WLOCK(inp);
if (callout_pending(&tp->t_timers->tt_delack) ||
!callout_active(&tp->t_timers->tt_delack)) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_delack);
if ((inp->inp_flags & INP_DROPPED) != 0) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
tp->t_flags |= TF_ACKNOW;
TCPSTAT_INC(tcps_delack);
(void) tcp_output(tp);
INP_WUNLOCK(inp);
CURVNET_RESTORE();
}
void
ieee80211_notify_michael_failure(struct ieee80211vap *vap,
const struct ieee80211_frame *wh, u_int keyix)
{
struct ifnet *ifp = vap->iv_ifp;
IEEE80211_NOTE_MAC(vap, IEEE80211_MSG_CRYPTO, wh->i_addr2,
"michael MIC verification failed <keyix %u>", keyix);
vap->iv_stats.is_rx_tkipmic++;
if (ifp != NULL) { /* NB: for cipher test modules */
struct ieee80211_michael_event iev;
IEEE80211_ADDR_COPY(iev.iev_dst, wh->i_addr1);
IEEE80211_ADDR_COPY(iev.iev_src, wh->i_addr2);
iev.iev_cipher = IEEE80211_CIPHER_TKIP;
iev.iev_keyix = keyix;
CURVNET_SET(ifp->if_vnet);
rt_ieee80211msg(ifp, RTM_IEEE80211_MICHAEL, &iev, sizeof(iev));
CURVNET_RESTORE();
}
}
static int
soo_ioctl(struct file *fp, u_long cmd, void *data, struct ucred *active_cred,
struct thread *td)
{
struct socket *so = fp->f_data;
int error = 0;
switch (cmd) {
case FIONBIO:
SOCK_LOCK(so);
if (*(int *)data)
so->so_state |= SS_NBIO;
else
so->so_state &= ~SS_NBIO;
SOCK_UNLOCK(so);
break;
case FIOASYNC:
/*
* XXXRW: This code separately acquires SOCK_LOCK(so) and
* SOCKBUF_LOCK(&so->so_rcv) even though they are the same
* mutex to avoid introducing the assumption that they are
* the same.
*/
if (*(int *)data) {
SOCK_LOCK(so);
so->so_state |= SS_ASYNC;
SOCK_UNLOCK(so);
SOCKBUF_LOCK(&so->so_rcv);
so->so_rcv.sb_flags |= SB_ASYNC;
SOCKBUF_UNLOCK(&so->so_rcv);
SOCKBUF_LOCK(&so->so_snd);
so->so_snd.sb_flags |= SB_ASYNC;
SOCKBUF_UNLOCK(&so->so_snd);
} else {
SOCK_LOCK(so);
so->so_state &= ~SS_ASYNC;
SOCK_UNLOCK(so);
SOCKBUF_LOCK(&so->so_rcv);
so->so_rcv.sb_flags &= ~SB_ASYNC;
SOCKBUF_UNLOCK(&so->so_rcv);
SOCKBUF_LOCK(&so->so_snd);
so->so_snd.sb_flags &= ~SB_ASYNC;
SOCKBUF_UNLOCK(&so->so_snd);
}
break;
case FIONREAD:
/* Unlocked read. */
*(int *)data = sbavail(&so->so_rcv);
break;
case FIONWRITE:
/* Unlocked read. */
*(int *)data = sbavail(&so->so_snd);
break;
case FIONSPACE:
/* Unlocked read. */
if ((so->so_snd.sb_hiwat < sbused(&so->so_snd)) ||
(so->so_snd.sb_mbmax < so->so_snd.sb_mbcnt))
*(int *)data = 0;
else
*(int *)data = sbspace(&so->so_snd);
break;
case FIOSETOWN:
error = fsetown(*(int *)data, &so->so_sigio);
break;
case FIOGETOWN:
*(int *)data = fgetown(&so->so_sigio);
break;
case SIOCSPGRP:
error = fsetown(-(*(int *)data), &so->so_sigio);
break;
case SIOCGPGRP:
*(int *)data = -fgetown(&so->so_sigio);
break;
case SIOCATMARK:
/* Unlocked read. */
*(int *)data = (so->so_rcv.sb_state & SBS_RCVATMARK) != 0;
break;
default:
/*
* Interface/routing/protocol specific ioctls: interface and
* routing ioctls should have a different entry since a
* socket is unnecessary.
*/
if (IOCGROUP(cmd) == 'i')
error = ifioctl(so, cmd, data, td);
else if (IOCGROUP(cmd) == 'r') {
CURVNET_SET(so->so_vnet);
error = rtioctl_fib(cmd, data, so->so_fibnum);
CURVNET_RESTORE();
} else {
CURVNET_SET(so->so_vnet);
error = ((*so->so_proto->pr_usrreqs->pru_control)
(so, cmd, data, 0, td));
CURVNET_RESTORE();
}
break;
}
return (error);
}
void
tcp_timer_2msl(void *xtp)
{
struct tcpcb *tp = xtp;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
/*
* XXXRW: Does this actually happen?
*/
INP_INFO_WLOCK(&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_2msl: inp == NULL"));
*/
if (inp == NULL) {
tcp_timer_race++;
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
INP_WLOCK(inp);
tcp_free_sackholes(tp);
if ((inp->inp_flags & INP_DROPPED) || callout_pending(&tp->t_timers->tt_2msl) ||
!callout_active(&tp->t_timers->tt_2msl)) {
INP_WUNLOCK(tp->t_inpcb);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_2msl);
/*
* 2 MSL timeout in shutdown went off. If we're closed but
* still waiting for peer to close and connection has been idle
* too long, or if 2MSL time is up from TIME_WAIT, delete connection
* control block. Otherwise, check again in a bit.
*
* If fastrecycle of FIN_WAIT_2, in FIN_WAIT_2 and receiver has closed,
* there's no point in hanging onto FIN_WAIT_2 socket. Just close it.
* Ignore fact that there were recent incoming segments.
*/
if (tcp_fast_finwait2_recycle && tp->t_state == TCPS_FIN_WAIT_2 &&
tp->t_inpcb && tp->t_inpcb->inp_socket &&
(tp->t_inpcb->inp_socket->so_rcv.sb_state & SBS_CANTRCVMORE)) {
TCPSTAT_INC(tcps_finwait2_drops);
tp = tcp_close(tp);
} else {
if (tp->t_state != TCPS_TIME_WAIT &&
ticks - tp->t_rcvtime <= TP_MAXIDLE(tp))
callout_reset_on(&tp->t_timers->tt_2msl,
TP_KEEPINTVL(tp), tcp_timer_2msl, tp, INP_CPU(inp));
else
tp = tcp_close(tp);
}
#ifdef TCPDEBUG
if (tp != NULL && (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
if (tp != NULL)
INP_WUNLOCK(inp);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
}
/*
* Mount a remote root fs via. nfs. This depends on the info in the
* nfs_diskless structure that has been filled in properly by some primary
* bootstrap.
* It goes something like this:
* - do enough of "ifconfig" by calling ifioctl() so that the system
* can talk to the server
* - If nfs_diskless.mygateway is filled in, use that address as
* a default gateway.
* - build the rootfs mount point and call mountnfs() to do the rest.
*
* It is assumed to be safe to read, modify, and write the nfsv3_diskless
* structure, as well as other global NFS client variables here, as
* nfs_mountroot() will be called once in the boot before any other NFS
* client activity occurs.
*/
int
nfs_mountroot(struct mount *mp)
{
struct thread *td = curthread;
struct nfsv3_diskless *nd = &nfsv3_diskless;
struct socket *so;
struct vnode *vp;
struct ifreq ir;
int error;
u_long l;
char buf[128];
char *cp;
#if defined(BOOTP_NFSROOT) && defined(BOOTP)
bootpc_init(); /* use bootp to get nfs_diskless filled in */
#elif defined(NFS_ROOT)
nfs_setup_diskless();
#endif
if (nfs_diskless_valid == 0) {
return (-1);
}
if (nfs_diskless_valid == 1)
nfs_convert_diskless();
/*
* XXX splnet, so networks will receive...
*/
splnet();
/*
* Do enough of ifconfig(8) so that the critical net interface can
* talk to the server.
*/
error = socreate(nd->myif.ifra_addr.sa_family, &so, nd->root_args.sotype, 0,
td->td_ucred, td);
if (error)
panic("nfs_mountroot: socreate(%04x): %d",
nd->myif.ifra_addr.sa_family, error);
#if 0 /* XXX Bad idea */
/*
* We might not have been told the right interface, so we pass
* over the first ten interfaces of the same kind, until we get
* one of them configured.
*/
for (i = strlen(nd->myif.ifra_name) - 1;
nd->myif.ifra_name[i] >= '0' &&
nd->myif.ifra_name[i] <= '9';
nd->myif.ifra_name[i] ++) {
error = ifioctl(so, SIOCAIFADDR, (caddr_t)&nd->myif, td);
if(!error)
break;
}
#endif
error = ifioctl(so, SIOCAIFADDR, (caddr_t)&nd->myif, td);
if (error)
panic("nfs_mountroot: SIOCAIFADDR: %d", error);
if ((cp = getenv("boot.netif.mtu")) != NULL) {
ir.ifr_mtu = strtol(cp, NULL, 10);
bcopy(nd->myif.ifra_name, ir.ifr_name, IFNAMSIZ);
freeenv(cp);
error = ifioctl(so, SIOCSIFMTU, (caddr_t)&ir, td);
if (error)
printf("nfs_mountroot: SIOCSIFMTU: %d", error);
}
soclose(so);
/*
* If the gateway field is filled in, set it as the default route.
* Note that pxeboot will set a default route of 0 if the route
* is not set by the DHCP server. Check also for a value of 0
* to avoid panicking inappropriately in that situation.
*/
if (nd->mygateway.sin_len != 0 &&
nd->mygateway.sin_addr.s_addr != 0) {
struct sockaddr_in mask, sin;
bzero((caddr_t)&mask, sizeof(mask));
sin = mask;
sin.sin_family = AF_INET;
sin.sin_len = sizeof(sin);
/* XXX MRT use table 0 for this sort of thing */
CURVNET_SET(TD_TO_VNET(td));
error = rtrequest_fib(RTM_ADD, (struct sockaddr *)&sin,
(struct sockaddr *)&nd->mygateway,
(struct sockaddr *)&mask,
RTF_UP | RTF_GATEWAY, NULL, RT_DEFAULT_FIB);
CURVNET_RESTORE();
if (error)
panic("nfs_mountroot: RTM_ADD: %d", error);
}
/*
* Create the rootfs mount point.
*/
nd->root_args.fh = nd->root_fh;
nd->root_args.fhsize = nd->root_fhsize;
l = ntohl(nd->root_saddr.sin_addr.s_addr);
snprintf(buf, sizeof(buf), "%ld.%ld.%ld.%ld:%s",
(l >> 24) & 0xff, (l >> 16) & 0xff,
(l >> 8) & 0xff, (l >> 0) & 0xff, nd->root_hostnam);
printf("NFS ROOT: %s\n", buf);
nd->root_args.hostname = buf;
if ((error = nfs_mountdiskless(buf,
&nd->root_saddr, &nd->root_args, td, &vp, mp)) != 0) {
return (error);
}
/*
* This is not really an nfs issue, but it is much easier to
* set hostname here and then let the "/etc/rc.xxx" files
* mount the right /var based upon its preset value.
*/
mtx_lock(&prison0.pr_mtx);
strlcpy(prison0.pr_hostname, nd->my_hostnam,
sizeof (prison0.pr_hostname));
mtx_unlock(&prison0.pr_mtx);
inittodr(ntohl(nd->root_time));
return (0);
}
/*
* Tear down vap state and reclaim the ifnet.
* The driver is assumed to have prepared for
* this; e.g. by turning off interrupts for the
* underlying device.
*/
void
ieee80211_vap_detach(struct ieee80211vap *vap)
{
struct ieee80211com *ic = vap->iv_ic;
struct ifnet *ifp = vap->iv_ifp;
CURVNET_SET(ifp->if_vnet);
IEEE80211_DPRINTF(vap, IEEE80211_MSG_STATE, "%s: %s parent %s\n",
__func__, ieee80211_opmode_name[vap->iv_opmode],
ic->ic_ifp->if_xname);
/* NB: bpfdetach is called by ether_ifdetach and claims all taps */
ether_ifdetach(ifp);
ieee80211_stop(vap);
/*
* Flush any deferred vap tasks.
*/
ieee80211_draintask(ic, &vap->iv_nstate_task);
ieee80211_draintask(ic, &vap->iv_swbmiss_task);
/* XXX band-aid until ifnet handles this for us */
taskqueue_drain(taskqueue_swi, &ifp->if_linktask);
IEEE80211_LOCK(ic);
KASSERT(vap->iv_state == IEEE80211_S_INIT , ("vap still running"));
TAILQ_REMOVE(&ic->ic_vaps, vap, iv_next);
ieee80211_syncflag_locked(ic, IEEE80211_F_WME);
#ifdef IEEE80211_SUPPORT_SUPERG
ieee80211_syncflag_locked(ic, IEEE80211_F_TURBOP);
#endif
ieee80211_syncflag_locked(ic, IEEE80211_F_PCF);
ieee80211_syncflag_locked(ic, IEEE80211_F_BURST);
ieee80211_syncflag_ht_locked(ic, IEEE80211_FHT_HT);
ieee80211_syncflag_ht_locked(ic, IEEE80211_FHT_USEHT40);
/* NB: this handles the bpfdetach done below */
ieee80211_syncflag_ext_locked(ic, IEEE80211_FEXT_BPF);
ieee80211_syncifflag_locked(ic, IFF_PROMISC);
ieee80211_syncifflag_locked(ic, IFF_ALLMULTI);
IEEE80211_UNLOCK(ic);
ifmedia_removeall(&vap->iv_media);
ieee80211_radiotap_vdetach(vap);
ieee80211_regdomain_vdetach(vap);
ieee80211_scan_vdetach(vap);
#ifdef IEEE80211_SUPPORT_SUPERG
ieee80211_superg_vdetach(vap);
#endif
ieee80211_ht_vdetach(vap);
/* NB: must be before ieee80211_node_vdetach */
ieee80211_proto_vdetach(vap);
ieee80211_crypto_vdetach(vap);
ieee80211_power_vdetach(vap);
ieee80211_node_vdetach(vap);
ieee80211_sysctl_vdetach(vap);
if_free(ifp);
CURVNET_RESTORE();
}
/*
* The timer handler for dummynet. Time is computed in ticks, but
* but the code is tolerant to the actual rate at which this is called.
* Once complete, the function reschedules itself for the next tick.
*/
void
dummynet_task(void *context, int pending)
{
struct timeval t;
struct mq q = { NULL, NULL }; /* queue to accumulate results */
CURVNET_SET((struct vnet *)context);
DN_BH_WLOCK();
/* Update number of lost(coalesced) ticks. */
tick_lost += pending - 1;
getmicrouptime(&t);
/* Last tick duration (usec). */
tick_last = (t.tv_sec - dn_cfg.prev_t.tv_sec) * 1000000 +
(t.tv_usec - dn_cfg.prev_t.tv_usec);
/* Last tick vs standard tick difference (usec). */
tick_delta = (tick_last * hz - 1000000) / hz;
/* Accumulated tick difference (usec). */
tick_delta_sum += tick_delta;
dn_cfg.prev_t = t;
/*
* Adjust curr_time if the accumulated tick difference is
* greater than the 'standard' tick. Since curr_time should
* be monotonically increasing, we do positive adjustments
* as required, and throttle curr_time in case of negative
* adjustment.
*/
dn_cfg.curr_time++;
if (tick_delta_sum - tick >= 0) {
int diff = tick_delta_sum / tick;
dn_cfg.curr_time += diff;
tick_diff += diff;
tick_delta_sum %= tick;
tick_adjustment++;
} else if (tick_delta_sum + tick <= 0) {
dn_cfg.curr_time--;
tick_diff--;
tick_delta_sum += tick;
tick_adjustment++;
}
/* serve pending events, accumulate in q */
for (;;) {
struct dn_id *p; /* generic parameter to handler */
if (dn_cfg.evheap.elements == 0 ||
DN_KEY_LT(dn_cfg.curr_time, HEAP_TOP(&dn_cfg.evheap)->key))
break;
p = HEAP_TOP(&dn_cfg.evheap)->object;
heap_extract(&dn_cfg.evheap, NULL);
if (p->type == DN_SCH_I) {
serve_sched(&q, (struct dn_sch_inst *)p, dn_cfg.curr_time);
} else { /* extracted a delay line */
transmit_event(&q, (struct delay_line *)p, dn_cfg.curr_time);
}
}
if (dn_cfg.expire && ++dn_cfg.expire_cycle >= dn_cfg.expire) {
dn_cfg.expire_cycle = 0;
dn_drain_scheduler();
dn_drain_queue();
}
DN_BH_WUNLOCK();
dn_reschedule();
if (q.head != NULL)
dummynet_send(q.head);
CURVNET_RESTORE();
}
/*
* Attach/setup the common net80211 state. Called by
* the driver on attach to prior to creating any vap's.
*/
void
ieee80211_ifattach(struct ieee80211com *ic,
const uint8_t macaddr[IEEE80211_ADDR_LEN])
{
struct ifnet *ifp = ic->ic_ifp;
struct sockaddr_dl *sdl;
struct ifaddr *ifa;
KASSERT(ifp->if_type == IFT_IEEE80211, ("if_type %d", ifp->if_type));
IEEE80211_LOCK_INIT(ic, ifp->if_xname);
IEEE80211_TX_LOCK_INIT(ic, ifp->if_xname);
TAILQ_INIT(&ic->ic_vaps);
/* Create a taskqueue for all state changes */
ic->ic_tq = taskqueue_create("ic_taskq", M_WAITOK | M_ZERO,
taskqueue_thread_enqueue, &ic->ic_tq);
taskqueue_start_threads(&ic->ic_tq, 1, PI_NET, "%s net80211 taskq",
ifp->if_xname);
/*
* Fill in 802.11 available channel set, mark all
* available channels as active, and pick a default
* channel if not already specified.
*/
ieee80211_media_init(ic);
ic->ic_update_mcast = null_update_mcast;
ic->ic_update_promisc = null_update_promisc;
ic->ic_update_chw = null_update_chw;
ic->ic_hash_key = arc4random();
ic->ic_bintval = IEEE80211_BINTVAL_DEFAULT;
ic->ic_lintval = ic->ic_bintval;
ic->ic_txpowlimit = IEEE80211_TXPOWER_MAX;
ieee80211_crypto_attach(ic);
ieee80211_node_attach(ic);
ieee80211_power_attach(ic);
ieee80211_proto_attach(ic);
#ifdef IEEE80211_SUPPORT_SUPERG
ieee80211_superg_attach(ic);
#endif
ieee80211_ht_attach(ic);
ieee80211_scan_attach(ic);
ieee80211_regdomain_attach(ic);
ieee80211_dfs_attach(ic);
ieee80211_sysctl_attach(ic);
ifp->if_addrlen = IEEE80211_ADDR_LEN;
ifp->if_hdrlen = 0;
CURVNET_SET(vnet0);
if_attach(ifp);
ifp->if_mtu = IEEE80211_MTU_MAX;
ifp->if_broadcastaddr = ieee80211broadcastaddr;
ifp->if_output = null_output;
ifp->if_input = null_input; /* just in case */
ifp->if_resolvemulti = NULL; /* NB: callers check */
ifa = ifaddr_byindex(ifp->if_index);
KASSERT(ifa != NULL, ("%s: no lladdr!\n", __func__));
sdl = (struct sockaddr_dl *)ifa->ifa_addr;
sdl->sdl_type = IFT_ETHER; /* XXX IFT_IEEE80211? */
sdl->sdl_alen = IEEE80211_ADDR_LEN;
IEEE80211_ADDR_COPY(LLADDR(sdl), macaddr);
ifa_free(ifa);
CURVNET_RESTORE();
}
void
tcp_timer_persist(void *xtp)
{
struct tcpcb *tp = xtp;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
inp = tp->t_inpcb;
KASSERT(inp != NULL, ("%s: tp %p tp->t_inpcb == NULL", __func__, tp));
INP_WLOCK(inp);
if (callout_pending(&tp->t_timers->tt_persist) ||
!callout_active(&tp->t_timers->tt_persist)) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_persist);
if ((inp->inp_flags & INP_DROPPED) != 0) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
KASSERT((tp->t_timers->tt_flags & TT_STOPPED) == 0,
("%s: tp %p tcpcb can't be stopped here", __func__, tp));
/*
* Persistence timer into zero window.
* Force a byte to be output, if possible.
*/
TCPSTAT_INC(tcps_persisttimeo);
/*
* Hack: if the peer is dead/unreachable, we do not
* time out if the window is closed. After a full
* backoff, drop the connection if the idle time
* (no responses to probes) reaches the maximum
* backoff that we would use if retransmitting.
*/
if (tp->t_rxtshift == TCP_MAXRXTSHIFT &&
(ticks - tp->t_rcvtime >= tcp_maxpersistidle ||
ticks - tp->t_rcvtime >= TCP_REXMTVAL(tp) * tcp_totbackoff)) {
TCPSTAT_INC(tcps_persistdrop);
if (tcp_inpinfo_lock_add(inp)) {
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
tp = tcp_drop(tp, ETIMEDOUT);
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
/*
* If the user has closed the socket then drop a persisting
* connection after a much reduced timeout.
*/
if (tp->t_state > TCPS_CLOSE_WAIT &&
(ticks - tp->t_rcvtime) >= TCPTV_PERSMAX) {
TCPSTAT_INC(tcps_persistdrop);
if (tcp_inpinfo_lock_add(inp)) {
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
tp = tcp_drop(tp, ETIMEDOUT);
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
tcp_setpersist(tp);
tp->t_flags |= TF_FORCEDATA;
(void) tp->t_fb->tfb_tcp_output(tp);
tp->t_flags &= ~TF_FORCEDATA;
#ifdef TCPDEBUG
if (tp != NULL && tp->t_inpcb->inp_socket->so_options & SO_DEBUG)
tcp_trace(TA_USER, ostate, tp, NULL, NULL, PRU_SLOWTIMO);
#endif
TCP_PROBE2(debug__user, tp, PRU_SLOWTIMO);
INP_WUNLOCK(inp);
out:
CURVNET_RESTORE();
}
void
tcp_timer_rexmt(void * xtp)
{
struct tcpcb *tp = xtp;
CURVNET_SET(tp->t_vnet);
int rexmt;
struct inpcb *inp;
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
inp = tp->t_inpcb;
KASSERT(inp != NULL, ("%s: tp %p tp->t_inpcb == NULL", __func__, tp));
INP_WLOCK(inp);
if (callout_pending(&tp->t_timers->tt_rexmt) ||
!callout_active(&tp->t_timers->tt_rexmt)) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_rexmt);
if ((inp->inp_flags & INP_DROPPED) != 0) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
KASSERT((tp->t_timers->tt_flags & TT_STOPPED) == 0,
("%s: tp %p tcpcb can't be stopped here", __func__, tp));
tcp_free_sackholes(tp);
if (tp->t_fb->tfb_tcp_rexmit_tmr) {
/* The stack has a timer action too. */
(*tp->t_fb->tfb_tcp_rexmit_tmr)(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);
if (tcp_inpinfo_lock_add(inp)) {
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
tp = tcp_drop(tp, tp->t_softerror ?
tp->t_softerror : ETIMEDOUT);
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
if (tp->t_state == TCPS_SYN_SENT) {
/*
* If the SYN was retransmitted, indicate CWND to be
* limited to 1 segment in cc_conn_init().
*/
tp->snd_cwnd = 1;
} else 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
void
tcp_timer_2msl(void *xtp)
{
struct tcpcb *tp = xtp;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
inp = tp->t_inpcb;
KASSERT(inp != NULL, ("%s: tp %p tp->t_inpcb == NULL", __func__, tp));
INP_WLOCK(inp);
tcp_free_sackholes(tp);
if (callout_pending(&tp->t_timers->tt_2msl) ||
!callout_active(&tp->t_timers->tt_2msl)) {
INP_WUNLOCK(tp->t_inpcb);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_2msl);
if ((inp->inp_flags & INP_DROPPED) != 0) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
KASSERT((tp->t_timers->tt_flags & TT_STOPPED) == 0,
("%s: tp %p tcpcb can't be stopped here", __func__, tp));
/*
* 2 MSL timeout in shutdown went off. If we're closed but
* still waiting for peer to close and connection has been idle
* too long delete connection control block. Otherwise, check
* again in a bit.
*
* If in TIME_WAIT state just ignore as this timeout is handled in
* tcp_tw_2msl_scan().
*
* If fastrecycle of FIN_WAIT_2, in FIN_WAIT_2 and receiver has closed,
* there's no point in hanging onto FIN_WAIT_2 socket. Just close it.
* Ignore fact that there were recent incoming segments.
*/
if ((inp->inp_flags & INP_TIMEWAIT) != 0) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
if (tcp_fast_finwait2_recycle && tp->t_state == TCPS_FIN_WAIT_2 &&
tp->t_inpcb && tp->t_inpcb->inp_socket &&
(tp->t_inpcb->inp_socket->so_rcv.sb_state & SBS_CANTRCVMORE)) {
TCPSTAT_INC(tcps_finwait2_drops);
if (tcp_inpinfo_lock_add(inp)) {
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
tp = tcp_close(tp);
tcp_inpinfo_lock_del(inp, tp);
goto out;
} else {
if (ticks - tp->t_rcvtime <= TP_MAXIDLE(tp)) {
callout_reset(&tp->t_timers->tt_2msl,
TP_KEEPINTVL(tp), tcp_timer_2msl, tp);
} else {
if (tcp_inpinfo_lock_add(inp)) {
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
tp = tcp_close(tp);
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
}
#ifdef TCPDEBUG
if (tp != NULL && (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
TCP_PROBE2(debug__user, tp, PRU_SLOWTIMO);
if (tp != NULL)
INP_WUNLOCK(inp);
out:
CURVNET_RESTORE();
}
void
tcp_timer_keep(void *xtp)
{
struct tcpcb *tp = xtp;
struct tcptemp *t_template;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
inp = tp->t_inpcb;
KASSERT(inp != NULL, ("%s: tp %p tp->t_inpcb == NULL", __func__, tp));
INP_WLOCK(inp);
if (callout_pending(&tp->t_timers->tt_keep) ||
!callout_active(&tp->t_timers->tt_keep)) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_keep);
if ((inp->inp_flags & INP_DROPPED) != 0) {
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
KASSERT((tp->t_timers->tt_flags & TT_STOPPED) == 0,
("%s: tp %p tcpcb can't be stopped here", __func__, tp));
/*
* Because we don't regularly reset the keepalive callout in
* the ESTABLISHED state, it may be that we don't actually need
* to send a keepalive yet. If that occurs, schedule another
* call for the next time the keepalive timer might expire.
*/
if (TCPS_HAVEESTABLISHED(tp->t_state)) {
u_int idletime;
idletime = ticks - tp->t_rcvtime;
if (idletime < TP_KEEPIDLE(tp)) {
callout_reset(&tp->t_timers->tt_keep,
TP_KEEPIDLE(tp) - idletime, tcp_timer_keep, tp);
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
}
}
/*
* Keep-alive timer went off; send something
* or drop connection if idle for too long.
*/
TCPSTAT_INC(tcps_keeptimeo);
if (tp->t_state < TCPS_ESTABLISHED)
goto dropit;
if ((always_keepalive || inp->inp_socket->so_options & SO_KEEPALIVE) &&
tp->t_state <= TCPS_CLOSING) {
if (ticks - tp->t_rcvtime >= TP_KEEPIDLE(tp) + TP_MAXIDLE(tp))
goto dropit;
/*
* Send a packet designed to force a response
* if the peer is up and reachable:
* either an ACK if the connection is still alive,
* or an RST if the peer has closed the connection
* due to timeout or reboot.
* Using sequence number tp->snd_una-1
* causes the transmitted zero-length segment
* to lie outside the receive window;
* by the protocol spec, this requires the
* correspondent TCP to respond.
*/
TCPSTAT_INC(tcps_keepprobe);
t_template = tcpip_maketemplate(inp);
if (t_template) {
tcp_respond(tp, t_template->tt_ipgen,
&t_template->tt_t, (struct mbuf *)NULL,
tp->rcv_nxt, tp->snd_una - 1, 0);
free(t_template, M_TEMP);
}
callout_reset(&tp->t_timers->tt_keep, TP_KEEPINTVL(tp),
tcp_timer_keep, tp);
} else
callout_reset(&tp->t_timers->tt_keep, TP_KEEPIDLE(tp),
tcp_timer_keep, tp);
#ifdef TCPDEBUG
if (inp->inp_socket->so_options & SO_DEBUG)
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
TCP_PROBE2(debug__user, tp, PRU_SLOWTIMO);
INP_WUNLOCK(inp);
CURVNET_RESTORE();
return;
dropit:
TCPSTAT_INC(tcps_keepdrops);
if (tcp_inpinfo_lock_add(inp)) {
tcp_inpinfo_lock_del(inp, tp);
goto out;
}
tp = tcp_drop(tp, ETIMEDOUT);
#ifdef TCPDEBUG
if (tp != NULL && (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
TCP_PROBE2(debug__user, tp, PRU_SLOWTIMO);
tcp_inpinfo_lock_del(inp, tp);
out:
CURVNET_RESTORE();
}
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
void
tcp_timer_persist(void *xtp)
{
struct tcpcb *tp = xtp;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
INP_INFO_WLOCK(&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_persist: inp == NULL"));
*/
if (inp == NULL) {
tcp_timer_race++;
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
INP_WLOCK(inp);
if ((inp->inp_flags & INP_DROPPED) || callout_pending(&tp->t_timers->tt_persist)
|| !callout_active(&tp->t_timers->tt_persist)) {
INP_WUNLOCK(inp);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_persist);
/*
* Persistance timer into zero window.
* Force a byte to be output, if possible.
*/
TCPSTAT_INC(tcps_persisttimeo);
/*
* Hack: if the peer is dead/unreachable, we do not
* time out if the window is closed. After a full
* backoff, drop the connection if the idle time
* (no responses to probes) reaches the maximum
* backoff that we would use if retransmitting.
*/
if (tp->t_rxtshift == TCP_MAXRXTSHIFT &&
(ticks - tp->t_rcvtime >= tcp_maxpersistidle ||
ticks - tp->t_rcvtime >= TCP_REXMTVAL(tp) * tcp_totbackoff)) {
TCPSTAT_INC(tcps_persistdrop);
tp = tcp_drop(tp, ETIMEDOUT);
goto out;
}
tcp_setpersist(tp);
tp->t_flags |= TF_FORCEDATA;
(void) tcp_output(tp);
tp->t_flags &= ~TF_FORCEDATA;
out:
#ifdef TCPDEBUG
if (tp != NULL && tp->t_inpcb->inp_socket->so_options & SO_DEBUG)
tcp_trace(TA_USER, ostate, tp, NULL, NULL, PRU_SLOWTIMO);
#endif
if (tp != NULL)
INP_WUNLOCK(inp);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
}
void
bootpc_init(void)
{
struct bootpc_ifcontext *ifctx, *nctx; /* Interface BOOTP contexts */
struct bootpc_globalcontext *gctx; /* Global BOOTP context */
struct ifnet *ifp;
int error;
#ifndef BOOTP_WIRED_TO
int ifcnt;
#endif
struct nfsv3_diskless *nd;
struct thread *td;
nd = &nfsv3_diskless;
td = curthread;
/*
* If already filled in, don't touch it here
*/
if (nfs_diskless_valid != 0)
return;
gctx = malloc(sizeof(*gctx), M_TEMP, M_WAITOK | M_ZERO);
if (gctx == NULL)
panic("Failed to allocate bootp global context structure");
gctx->xid = ~0xFFFF;
gctx->starttime = time_second;
/*
* Find a network interface.
*/
CURVNET_SET(TD_TO_VNET(td));
#ifdef BOOTP_WIRED_TO
printf("bootpc_init: wired to interface '%s'\n",
__XSTRING(BOOTP_WIRED_TO));
allocifctx(gctx);
#else
/*
* Preallocate interface context storage, if another interface
* attaches and wins the race, it won't be eligible for bootp.
*/
IFNET_RLOCK();
for (ifp = TAILQ_FIRST(&V_ifnet), ifcnt = 0;
ifp != NULL;
ifp = TAILQ_NEXT(ifp, if_link)) {
if ((ifp->if_flags &
(IFF_LOOPBACK | IFF_POINTOPOINT | IFF_BROADCAST)) !=
IFF_BROADCAST)
continue;
ifcnt++;
}
IFNET_RUNLOCK();
if (ifcnt == 0)
panic("bootpc_init: no eligible interfaces");
for (; ifcnt > 0; ifcnt--)
allocifctx(gctx);
#endif
IFNET_RLOCK();
for (ifp = TAILQ_FIRST(&V_ifnet), ifctx = gctx->interfaces;
ifp != NULL && ifctx != NULL;
ifp = TAILQ_NEXT(ifp, if_link)) {
strlcpy(ifctx->ireq.ifr_name, ifp->if_xname,
sizeof(ifctx->ireq.ifr_name));
#ifdef BOOTP_WIRED_TO
if (strcmp(ifctx->ireq.ifr_name,
__XSTRING(BOOTP_WIRED_TO)) != 0)
continue;
#else
if ((ifp->if_flags &
(IFF_LOOPBACK | IFF_POINTOPOINT | IFF_BROADCAST)) !=
IFF_BROADCAST)
continue;
#endif
ifctx->ifp = ifp;
ifctx = ifctx->next;
}
IFNET_RUNLOCK();
CURVNET_RESTORE();
if (gctx->interfaces == NULL || gctx->interfaces->ifp == NULL) {
#ifdef BOOTP_WIRED_TO
panic("bootpc_init: Could not find interface specified "
"by BOOTP_WIRED_TO: "
__XSTRING(BOOTP_WIRED_TO));
#else
panic("bootpc_init: no suitable interface");
#endif
}
for (ifctx = gctx->interfaces; ifctx != NULL; ifctx = ifctx->next)
bootpc_fakeup_interface(ifctx, gctx, td);
for (ifctx = gctx->interfaces; ifctx != NULL; ifctx = ifctx->next)
bootpc_compose_query(ifctx, gctx, td);
error = bootpc_call(gctx, td);
if (error != 0) {
#ifdef BOOTP_NFSROOT
panic("BOOTP call failed");
#else
printf("BOOTP call failed\n");
#endif
}
rootdevnames[0] = "nfs:";
#ifdef NFSCLIENT
rootdevnames[1] = "oldnfs:";
#endif
mountopts(&nd->root_args, NULL);
for (ifctx = gctx->interfaces; ifctx != NULL; ifctx = ifctx->next)
if (bootpc_ifctx_isresolved(ifctx) != 0)
bootpc_decode_reply(nd, ifctx, gctx);
#ifdef BOOTP_NFSROOT
if (gctx->gotrootpath == 0)
panic("bootpc: No root path offered");
#endif
for (ifctx = gctx->interfaces; ifctx != NULL; ifctx = ifctx->next) {
bootpc_adjust_interface(ifctx, gctx, td);
soclose(ifctx->so);
}
for (ifctx = gctx->interfaces; ifctx != NULL; ifctx = ifctx->next)
if (ifctx->gotrootpath != 0)
break;
if (ifctx == NULL) {
for (ifctx = gctx->interfaces;
ifctx != NULL;
ifctx = ifctx->next)
if (bootpc_ifctx_isresolved(ifctx) != 0)
break;
}
if (ifctx == NULL)
goto out;
if (gctx->gotrootpath != 0) {
setenv("boot.netif.name", ifctx->ifp->if_xname);
error = md_mount(&nd->root_saddr, nd->root_hostnam,
nd->root_fh, &nd->root_fhsize,
&nd->root_args, td);
if (error != 0)
panic("nfs_boot: mountd root, error=%d", error);
nfs_diskless_valid = 3;
}
strcpy(nd->myif.ifra_name, ifctx->ireq.ifr_name);
bcopy(&ifctx->myaddr, &nd->myif.ifra_addr, sizeof(ifctx->myaddr));
bcopy(&ifctx->myaddr, &nd->myif.ifra_broadaddr, sizeof(ifctx->myaddr));
((struct sockaddr_in *) &nd->myif.ifra_broadaddr)->sin_addr.s_addr =
ifctx->myaddr.sin_addr.s_addr |
~ ifctx->netmask.sin_addr.s_addr;
bcopy(&ifctx->netmask, &nd->myif.ifra_mask, sizeof(ifctx->netmask));
out:
for (ifctx = gctx->interfaces; ifctx != NULL; ifctx = nctx) {
nctx = ifctx->next;
free(ifctx, M_TEMP);
}
free(gctx, M_TEMP);
}
/*
* Create a new transport for a socket optained via soaccept().
*/
SVCXPRT *
svc_vc_create_conn(SVCPOOL *pool, struct socket *so, struct sockaddr *raddr)
{
SVCXPRT *xprt = NULL;
struct cf_conn *cd = NULL;
struct sockaddr* sa = NULL;
struct sockopt opt;
int one = 1;
int error;
bzero(&opt, sizeof(struct sockopt));
opt.sopt_dir = SOPT_SET;
opt.sopt_level = SOL_SOCKET;
opt.sopt_name = SO_KEEPALIVE;
opt.sopt_val = &one;
opt.sopt_valsize = sizeof(one);
CURVNET_SET(so->so_vnet);
error = sosetopt(so, &opt);
if (error) {
CURVNET_RESTORE();
return (NULL);
}
if (so->so_proto->pr_protocol == IPPROTO_TCP) {
bzero(&opt, sizeof(struct sockopt));
opt.sopt_dir = SOPT_SET;
opt.sopt_level = IPPROTO_TCP;
opt.sopt_name = TCP_NODELAY;
opt.sopt_val = &one;
opt.sopt_valsize = sizeof(one);
error = sosetopt(so, &opt);
if (error) {
CURVNET_RESTORE();
return (NULL);
}
}
CURVNET_RESTORE();
cd = mem_alloc(sizeof(*cd));
cd->strm_stat = XPRT_IDLE;
xprt = svc_xprt_alloc();
sx_init(&xprt->xp_lock, "xprt->xp_lock");
xprt->xp_pool = pool;
xprt->xp_socket = so;
xprt->xp_p1 = cd;
xprt->xp_p2 = NULL;
xprt->xp_ops = &svc_vc_ops;
/*
* See http://www.connectathon.org/talks96/nfstcp.pdf - client
* has a 5 minute timer, server has a 6 minute timer.
*/
xprt->xp_idletimeout = 6 * 60;
memcpy(&xprt->xp_rtaddr, raddr, raddr->sa_len);
error = so->so_proto->pr_usrreqs->pru_sockaddr(so, &sa);
if (error)
goto cleanup_svc_vc_create;
memcpy(&xprt->xp_ltaddr, sa, sa->sa_len);
free(sa, M_SONAME);
xprt_register(xprt);
SOCKBUF_LOCK(&so->so_rcv);
xprt->xp_upcallset = 1;
soupcall_set(so, SO_RCV, svc_vc_soupcall, xprt);
SOCKBUF_UNLOCK(&so->so_rcv);
/*
* Throw the transport into the active list in case it already
* has some data buffered.
*/
sx_xlock(&xprt->xp_lock);
xprt_active(xprt);
sx_xunlock(&xprt->xp_lock);
return (xprt);
cleanup_svc_vc_create:
if (xprt) {
mem_free(xprt, sizeof(*xprt));
}
if (cd)
mem_free(cd, sizeof(*cd));
return (NULL);
}
void
tcp_timer_keep(void *xtp)
{
struct tcpcb *tp = xtp;
struct tcptemp *t_template;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
INP_INFO_RLOCK(&V_tcbinfo);
inp = tp->t_inpcb;
KASSERT(inp != NULL, ("%s: tp %p tp->t_inpcb == NULL", __func__, tp));
INP_WLOCK(inp);
if (callout_pending(&tp->t_timers->tt_keep) ||
!callout_active(&tp->t_timers->tt_keep)) {
INP_WUNLOCK(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_keep);
if ((inp->inp_flags & INP_DROPPED) != 0) {
INP_WUNLOCK(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
KASSERT((tp->t_timers->tt_flags & TT_STOPPED) == 0,
("%s: tp %p tcpcb can't be stopped here", __func__, tp));
KASSERT((tp->t_timers->tt_flags & TT_KEEP) != 0,
("%s: tp %p keep callout should be running", __func__, tp));
/*
* Keep-alive timer went off; send something
* or drop connection if idle for too long.
*/
TCPSTAT_INC(tcps_keeptimeo);
if (tp->t_state < TCPS_ESTABLISHED)
goto dropit;
if ((always_keepalive || inp->inp_socket->so_options & SO_KEEPALIVE) &&
tp->t_state <= TCPS_CLOSING) {
if (ticks - tp->t_rcvtime >= TP_KEEPIDLE(tp) + TP_MAXIDLE(tp))
goto dropit;
/*
* Send a packet designed to force a response
* if the peer is up and reachable:
* either an ACK if the connection is still alive,
* or an RST if the peer has closed the connection
* due to timeout or reboot.
* Using sequence number tp->snd_una-1
* causes the transmitted zero-length segment
* to lie outside the receive window;
* by the protocol spec, this requires the
* correspondent TCP to respond.
*/
TCPSTAT_INC(tcps_keepprobe);
t_template = tcpip_maketemplate(inp);
if (t_template) {
tcp_respond(tp, t_template->tt_ipgen,
&t_template->tt_t, (struct mbuf *)NULL,
tp->rcv_nxt, tp->snd_una - 1, 0);
free(t_template, M_TEMP);
}
if (!callout_reset(&tp->t_timers->tt_keep, TP_KEEPINTVL(tp),
tcp_timer_keep, tp)) {
tp->t_timers->tt_flags &= ~TT_KEEP_RST;
}
} else if (!callout_reset(&tp->t_timers->tt_keep, TP_KEEPIDLE(tp),
tcp_timer_keep, tp)) {
tp->t_timers->tt_flags &= ~TT_KEEP_RST;
}
#ifdef TCPDEBUG
if (inp->inp_socket->so_options & SO_DEBUG)
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
TCP_PROBE2(debug__user, tp, PRU_SLOWTIMO);
INP_WUNLOCK(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
dropit:
TCPSTAT_INC(tcps_keepdrops);
tp = tcp_drop(tp, ETIMEDOUT);
#ifdef TCPDEBUG
if (tp != NULL && (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
TCP_PROBE2(debug__user, tp, PRU_SLOWTIMO);
if (tp != NULL)
INP_WUNLOCK(tp->t_inpcb);
INP_INFO_RUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
}
static bool_t
svc_vc_recv(SVCXPRT *xprt, struct rpc_msg *msg,
struct sockaddr **addrp, struct mbuf **mp)
{
struct cf_conn *cd = (struct cf_conn *) xprt->xp_p1;
struct uio uio;
struct mbuf *m;
XDR xdrs;
int error, rcvflag;
/*
* Serialise access to the socket and our own record parsing
* state.
*/
sx_xlock(&xprt->xp_lock);
for (;;) {
/*
* If we have an mbuf chain in cd->mpending, try to parse a
* record from it, leaving the result in cd->mreq. If we don't
* have a complete record, leave the partial result in
* cd->mreq and try to read more from the socket.
*/
if (cd->mpending) {
/*
* If cd->resid is non-zero, we have part of the
* record already, otherwise we are expecting a record
* marker.
*/
if (!cd->resid) {
/*
* See if there is enough data buffered to
* make up a record marker. Make sure we can
* handle the case where the record marker is
* split across more than one mbuf.
*/
size_t n = 0;
uint32_t header;
m = cd->mpending;
while (n < sizeof(uint32_t) && m) {
n += m->m_len;
m = m->m_next;
}
if (n < sizeof(uint32_t))
goto readmore;
if (cd->mpending->m_len < sizeof(uint32_t))
cd->mpending = m_pullup(cd->mpending,
sizeof(uint32_t));
memcpy(&header, mtod(cd->mpending, uint32_t *),
sizeof(header));
header = ntohl(header);
cd->eor = (header & 0x80000000) != 0;
cd->resid = header & 0x7fffffff;
m_adj(cd->mpending, sizeof(uint32_t));
}
/*
* Start pulling off mbufs from cd->mpending
* until we either have a complete record or
* we run out of data. We use m_split to pull
* data - it will pull as much as possible and
* split the last mbuf if necessary.
*/
while (cd->mpending && cd->resid) {
m = cd->mpending;
if (cd->mpending->m_next
|| cd->mpending->m_len > cd->resid)
cd->mpending = m_split(cd->mpending,
cd->resid, M_WAIT);
else
cd->mpending = NULL;
if (cd->mreq)
m_last(cd->mreq)->m_next = m;
else
cd->mreq = m;
while (m) {
cd->resid -= m->m_len;
m = m->m_next;
}
}
/*
* If cd->resid is zero now, we have managed to
* receive a record fragment from the stream. Check
* for the end-of-record mark to see if we need more.
*/
if (cd->resid == 0) {
if (!cd->eor)
continue;
/*
* Success - we have a complete record in
* cd->mreq.
*/
xdrmbuf_create(&xdrs, cd->mreq, XDR_DECODE);
cd->mreq = NULL;
sx_xunlock(&xprt->xp_lock);
if (! xdr_callmsg(&xdrs, msg)) {
XDR_DESTROY(&xdrs);
return (FALSE);
}
*addrp = NULL;
*mp = xdrmbuf_getall(&xdrs);
XDR_DESTROY(&xdrs);
return (TRUE);
}
}
readmore:
/*
* The socket upcall calls xprt_active() which will eventually
* cause the server to call us here. We attempt to
* read as much as possible from the socket and put
* the result in cd->mpending. If the read fails,
* we have drained both cd->mpending and the socket so
* we can call xprt_inactive().
*/
uio.uio_resid = 1000000000;
uio.uio_td = curthread;
m = NULL;
rcvflag = MSG_DONTWAIT;
CURVNET_SET(xprt->xp_socket->so_vnet);
error = soreceive(xprt->xp_socket, NULL, &uio, &m, NULL,
&rcvflag);
CURVNET_RESTORE();
if (error == EWOULDBLOCK) {
/*
* We must re-test for readability after
* taking the lock to protect us in the case
* where a new packet arrives on the socket
* after our call to soreceive fails with
* EWOULDBLOCK. The pool lock protects us from
* racing the upcall after our soreadable()
* call returns false.
*/
mtx_lock(&xprt->xp_pool->sp_lock);
if (!soreadable(xprt->xp_socket))
xprt_inactive_locked(xprt);
mtx_unlock(&xprt->xp_pool->sp_lock);
sx_xunlock(&xprt->xp_lock);
return (FALSE);
}
if (error) {
SOCKBUF_LOCK(&xprt->xp_socket->so_rcv);
if (xprt->xp_upcallset) {
xprt->xp_upcallset = 0;
soupcall_clear(xprt->xp_socket, SO_RCV);
}
SOCKBUF_UNLOCK(&xprt->xp_socket->so_rcv);
xprt_inactive(xprt);
cd->strm_stat = XPRT_DIED;
sx_xunlock(&xprt->xp_lock);
return (FALSE);
}
if (!m) {
/*
* EOF - the other end has closed the socket.
*/
xprt_inactive(xprt);
cd->strm_stat = XPRT_DIED;
sx_xunlock(&xprt->xp_lock);
return (FALSE);
}
if (cd->mpending)
m_last(cd->mpending)->m_next = m;
else
cd->mpending = m;
}
void
tcp_timer_keep(void *xtp)
{
struct tcpcb *tp = xtp;
struct tcptemp *t_template;
struct inpcb *inp;
CURVNET_SET(tp->t_vnet);
#ifdef TCPDEBUG
int ostate;
ostate = tp->t_state;
#endif
INP_INFO_WLOCK(&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_keep: inp == NULL"));
*/
if (inp == NULL) {
tcp_timer_race++;
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
INP_WLOCK(inp);
if ((inp->inp_flags & INP_DROPPED) || callout_pending(&tp->t_timers->tt_keep)
|| !callout_active(&tp->t_timers->tt_keep)) {
INP_WUNLOCK(inp);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
}
callout_deactivate(&tp->t_timers->tt_keep);
/*
* Keep-alive timer went off; send something
* or drop connection if idle for too long.
*/
TCPSTAT_INC(tcps_keeptimeo);
if (tp->t_state < TCPS_ESTABLISHED)
goto dropit;
if ((always_keepalive || inp->inp_socket->so_options & SO_KEEPALIVE) &&
tp->t_state <= TCPS_CLOSING) {
if (ticks - tp->t_rcvtime >= TP_KEEPIDLE(tp) + TP_MAXIDLE(tp))
goto dropit;
/*
* Send a packet designed to force a response
* if the peer is up and reachable:
* either an ACK if the connection is still alive,
* or an RST if the peer has closed the connection
* due to timeout or reboot.
* Using sequence number tp->snd_una-1
* causes the transmitted zero-length segment
* to lie outside the receive window;
* by the protocol spec, this requires the
* correspondent TCP to respond.
*/
TCPSTAT_INC(tcps_keepprobe);
t_template = tcpip_maketemplate(inp);
if (t_template) {
tcp_respond(tp, t_template->tt_ipgen,
&t_template->tt_t, (struct mbuf *)NULL,
tp->rcv_nxt, tp->snd_una - 1, 0);
free(t_template, M_TEMP);
}
callout_reset_on(&tp->t_timers->tt_keep, TP_KEEPINTVL(tp),
tcp_timer_keep, tp, INP_CPU(inp));
} else
callout_reset_on(&tp->t_timers->tt_keep, TP_KEEPIDLE(tp),
tcp_timer_keep, tp, INP_CPU(inp));
#ifdef TCPDEBUG
if (inp->inp_socket->so_options & SO_DEBUG)
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
INP_WUNLOCK(inp);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
return;
dropit:
TCPSTAT_INC(tcps_keepdrops);
tp = tcp_drop(tp, ETIMEDOUT);
#ifdef TCPDEBUG
if (tp != NULL && (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
tcp_trace(TA_USER, ostate, tp, (void *)0, (struct tcphdr *)0,
PRU_SLOWTIMO);
#endif
if (tp != NULL)
INP_WUNLOCK(tp->t_inpcb);
INP_INFO_WUNLOCK(&V_tcbinfo);
CURVNET_RESTORE();
}