aboutsummaryrefslogtreecommitdiff
path: root/net/tipc
diff options
context:
space:
mode:
Diffstat (limited to 'net/tipc')
-rw-r--r--net/tipc/addr.c7
-rw-r--r--net/tipc/addr.h1
-rw-r--r--net/tipc/bcast.c42
-rw-r--r--net/tipc/bcast.h4
-rw-r--r--net/tipc/core.c2
-rw-r--r--net/tipc/discover.c1
-rw-r--r--net/tipc/link.c562
-rw-r--r--net/tipc/link.h34
-rw-r--r--net/tipc/msg.c11
-rw-r--r--net/tipc/msg.h42
-rw-r--r--net/tipc/name_distr.c2
-rw-r--r--net/tipc/name_table.c4
-rw-r--r--net/tipc/node.c112
-rw-r--r--net/tipc/node.h12
-rw-r--r--net/tipc/server.c47
-rw-r--r--net/tipc/socket.c155
-rw-r--r--net/tipc/socket.h4
-rw-r--r--net/tipc/udp_media.c10
18 files changed, 502 insertions, 550 deletions
diff --git a/net/tipc/addr.c b/net/tipc/addr.c
index 48fd3b5a73fb..ba7daa864d44 100644
--- a/net/tipc/addr.c
+++ b/net/tipc/addr.c
@@ -38,6 +38,13 @@
#include "addr.h"
#include "core.h"
+u32 tipc_own_addr(struct net *net)
+{
+ struct tipc_net *tn = net_generic(net, tipc_net_id);
+
+ return tn->own_addr;
+}
+
/**
* in_own_cluster - test for cluster inclusion; <0.0.0> always matches
*/
diff --git a/net/tipc/addr.h b/net/tipc/addr.h
index c700c2d28e09..7ba6d5c8ae40 100644
--- a/net/tipc/addr.h
+++ b/net/tipc/addr.h
@@ -55,6 +55,7 @@ static inline u32 tipc_cluster_mask(u32 addr)
return addr & TIPC_CLUSTER_MASK;
}
+u32 tipc_own_addr(struct net *net);
int in_own_cluster(struct net *net, u32 addr);
int in_own_cluster_exact(struct net *net, u32 addr);
int in_own_node(struct net *net, u32 addr);
diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c
index 5aff0844d4d3..c5cbdcb1f0b5 100644
--- a/net/tipc/bcast.c
+++ b/net/tipc/bcast.c
@@ -62,21 +62,8 @@ static void tipc_bclink_lock(struct net *net)
static void tipc_bclink_unlock(struct net *net)
{
struct tipc_net *tn = net_generic(net, tipc_net_id);
- struct tipc_node *node = NULL;
- if (likely(!tn->bclink->flags)) {
- spin_unlock_bh(&tn->bclink->lock);
- return;
- }
-
- if (tn->bclink->flags & TIPC_BCLINK_RESET) {
- tn->bclink->flags &= ~TIPC_BCLINK_RESET;
- node = tipc_bclink_retransmit_to(net);
- }
spin_unlock_bh(&tn->bclink->lock);
-
- if (node)
- tipc_link_reset_all(node);
}
void tipc_bclink_input(struct net *net)
@@ -91,13 +78,6 @@ uint tipc_bclink_get_mtu(void)
return MAX_PKT_DEFAULT_MCAST;
}
-void tipc_bclink_set_flags(struct net *net, unsigned int flags)
-{
- struct tipc_net *tn = net_generic(net, tipc_net_id);
-
- tn->bclink->flags |= flags;
-}
-
static u32 bcbuf_acks(struct sk_buff *buf)
{
return (u32)(unsigned long)TIPC_SKB_CB(buf)->handle;
@@ -156,7 +136,6 @@ static void bclink_update_last_sent(struct tipc_node *node, u32 seqno)
seqno : node->bclink.last_sent;
}
-
/**
* tipc_bclink_retransmit_to - get most recent node to request retransmission
*
@@ -350,13 +329,12 @@ static void bclink_peek_nack(struct net *net, struct tipc_msg *msg)
return;
tipc_node_lock(n_ptr);
-
if (n_ptr->bclink.recv_permitted &&
(n_ptr->bclink.last_in != n_ptr->bclink.last_sent) &&
(n_ptr->bclink.last_in == msg_bcgap_after(msg)))
n_ptr->bclink.oos_state = 2;
-
tipc_node_unlock(n_ptr);
+ tipc_node_put(n_ptr);
}
/* tipc_bclink_xmit - deliver buffer chain to all nodes in cluster
@@ -435,7 +413,7 @@ static void bclink_accept_pkt(struct tipc_node *node, u32 seqno)
*/
if (((seqno - tn->own_addr) % TIPC_MIN_LINK_WIN) == 0) {
tipc_link_proto_xmit(node->active_links[node->addr & 1],
- STATE_MSG, 0, 0, 0, 0, 0);
+ STATE_MSG, 0, 0, 0, 0);
tn->bcl->stats.sent_acks++;
}
}
@@ -476,17 +454,18 @@ void tipc_bclink_rcv(struct net *net, struct sk_buff *buf)
goto unlock;
if (msg_destnode(msg) == tn->own_addr) {
tipc_bclink_acknowledge(node, msg_bcast_ack(msg));
- tipc_node_unlock(node);
tipc_bclink_lock(net);
bcl->stats.recv_nacks++;
tn->bclink->retransmit_to = node;
bclink_retransmit_pkt(tn, msg_bcgap_after(msg),
msg_bcgap_to(msg));
tipc_bclink_unlock(net);
+ tipc_node_unlock(node);
} else {
tipc_node_unlock(node);
bclink_peek_nack(net, msg);
}
+ tipc_node_put(node);
goto exit;
}
@@ -523,11 +502,13 @@ receive:
tipc_bclink_unlock(net);
tipc_node_unlock(node);
} else if (msg_user(msg) == MSG_FRAGMENTER) {
- tipc_buf_append(&node->bclink.reasm_buf, &buf);
- if (unlikely(!buf && !node->bclink.reasm_buf))
- goto unlock;
tipc_bclink_lock(net);
bclink_accept_pkt(node, seqno);
+ tipc_buf_append(&node->bclink.reasm_buf, &buf);
+ if (unlikely(!buf && !node->bclink.reasm_buf)) {
+ tipc_bclink_unlock(net);
+ goto unlock;
+ }
bcl->stats.recv_fragments++;
if (buf) {
bcl->stats.recv_fragmented++;
@@ -589,6 +570,7 @@ receive:
unlock:
tipc_node_unlock(node);
+ tipc_node_put(node);
exit:
kfree_skb(buf);
}
@@ -829,7 +811,7 @@ int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg)
prop = nla_nest_start(msg->skb, TIPC_NLA_LINK_PROP);
if (!prop)
goto attr_msg_full;
- if (nla_put_u32(msg->skb, TIPC_NLA_PROP_WIN, bcl->queue_limit[0]))
+ if (nla_put_u32(msg->skb, TIPC_NLA_PROP_WIN, bcl->window))
goto prop_msg_full;
nla_nest_end(msg->skb, prop);
@@ -917,7 +899,7 @@ int tipc_bclink_init(struct net *net)
skb_queue_head_init(&bclink->inputq);
bcl->owner = &bclink->node;
bcl->owner->net = net;
- bcl->max_pkt = MAX_PKT_DEFAULT_MCAST;
+ bcl->mtu = MAX_PKT_DEFAULT_MCAST;
tipc_link_set_queue_limits(bcl, BCLINK_WIN_DEFAULT);
bcl->bearer_id = MAX_BEARERS;
rcu_assign_pointer(tn->bearer_list[MAX_BEARERS], &bcbearer->bearer);
diff --git a/net/tipc/bcast.h b/net/tipc/bcast.h
index 43f397fbac55..4bdc12277d33 100644
--- a/net/tipc/bcast.h
+++ b/net/tipc/bcast.h
@@ -55,7 +55,6 @@ struct tipc_bcbearer_pair {
struct tipc_bearer *secondary;
};
-#define TIPC_BCLINK_RESET 1
#define BCBEARER MAX_BEARERS
/**
@@ -86,7 +85,6 @@ struct tipc_bcbearer {
* @lock: spinlock governing access to structure
* @link: (non-standard) broadcast link structure
* @node: (non-standard) node structure representing b'cast link's peer node
- * @flags: represent bclink states
* @bcast_nodes: map of broadcast-capable nodes
* @retransmit_to: node that most recently requested a retransmit
*
@@ -96,7 +94,6 @@ struct tipc_bclink {
spinlock_t lock;
struct tipc_link link;
struct tipc_node node;
- unsigned int flags;
struct sk_buff_head arrvq;
struct sk_buff_head inputq;
struct tipc_node_map bcast_nodes;
@@ -117,7 +114,6 @@ static inline int tipc_nmap_equal(struct tipc_node_map *nm_a,
int tipc_bclink_init(struct net *net);
void tipc_bclink_stop(struct net *net);
-void tipc_bclink_set_flags(struct net *tn, unsigned int flags);
void tipc_bclink_add_node(struct net *net, u32 addr);
void tipc_bclink_remove_node(struct net *net, u32 addr);
struct tipc_node *tipc_bclink_retransmit_to(struct net *tn);
diff --git a/net/tipc/core.c b/net/tipc/core.c
index 935205e6bcfe..be1c9fa60b09 100644
--- a/net/tipc/core.c
+++ b/net/tipc/core.c
@@ -152,11 +152,11 @@ out_netlink:
static void __exit tipc_exit(void)
{
tipc_bearer_cleanup();
+ unregister_pernet_subsys(&tipc_net_ops);
tipc_netlink_stop();
tipc_netlink_compat_stop();
tipc_socket_stop();
tipc_unregister_sysctl();
- unregister_pernet_subsys(&tipc_net_ops);
pr_info("Deactivated\n");
}
diff --git a/net/tipc/discover.c b/net/tipc/discover.c
index 169f3dd038b9..967e292f53c8 100644
--- a/net/tipc/discover.c
+++ b/net/tipc/discover.c
@@ -260,6 +260,7 @@ void tipc_disc_rcv(struct net *net, struct sk_buff *buf,
}
}
tipc_node_unlock(node);
+ tipc_node_put(node);
}
/**
diff --git a/net/tipc/link.c b/net/tipc/link.c
index bc49120bfb44..a6b30df6ec02 100644
--- a/net/tipc/link.c
+++ b/net/tipc/link.c
@@ -89,24 +89,14 @@ static const struct nla_policy tipc_nl_prop_policy[TIPC_NLA_PROP_MAX + 1] = {
#define TIMEOUT_EVT 560817u /* link timer expired */
/*
- * The following two 'message types' is really just implementation
- * data conveniently stored in the message header.
- * They must not be considered part of the protocol
+ * State value stored in 'failover_pkts'
*/
-#define OPEN_MSG 0
-#define CLOSED_MSG 1
-
-/*
- * State value stored in 'exp_msg_count'
- */
-#define START_CHANGEOVER 100000u
+#define FIRST_FAILOVER 0xffffu
static void link_handle_out_of_seq_msg(struct tipc_link *link,
struct sk_buff *skb);
static void tipc_link_proto_rcv(struct tipc_link *link,
struct sk_buff *skb);
-static int tipc_link_tunnel_rcv(struct tipc_node *node,
- struct sk_buff **skb);
static void link_set_supervision_props(struct tipc_link *l_ptr, u32 tol);
static void link_state_event(struct tipc_link *l_ptr, u32 event);
static void link_reset_statistics(struct tipc_link *l_ptr);
@@ -115,7 +105,7 @@ static void tipc_link_sync_xmit(struct tipc_link *l);
static void tipc_link_sync_rcv(struct tipc_node *n, struct sk_buff *buf);
static void tipc_link_input(struct tipc_link *l, struct sk_buff *skb);
static bool tipc_data_input(struct tipc_link *l, struct sk_buff *skb);
-
+static bool tipc_link_failover_rcv(struct tipc_link *l, struct sk_buff **skb);
/*
* Simple link routines
*/
@@ -139,32 +129,11 @@ static void tipc_link_put(struct tipc_link *l_ptr)
kref_put(&l_ptr->ref, tipc_link_release);
}
-static void link_init_max_pkt(struct tipc_link *l_ptr)
+static struct tipc_link *tipc_parallel_link(struct tipc_link *l)
{
- struct tipc_node *node = l_ptr->owner;
- struct tipc_net *tn = net_generic(node->net, tipc_net_id);
- struct tipc_bearer *b_ptr;
- u32 max_pkt;
-
- rcu_read_lock();
- b_ptr = rcu_dereference_rtnl(tn->bearer_list[l_ptr->bearer_id]);
- if (!b_ptr) {
- rcu_read_unlock();
- return;
- }
- max_pkt = (b_ptr->mtu & ~3);
- rcu_read_unlock();
-
- if (max_pkt > MAX_MSG_SIZE)
- max_pkt = MAX_MSG_SIZE;
-
- l_ptr->max_pkt_target = max_pkt;
- if (l_ptr->max_pkt_target < MAX_PKT_DEFAULT)
- l_ptr->max_pkt = l_ptr->max_pkt_target;
- else
- l_ptr->max_pkt = MAX_PKT_DEFAULT;
-
- l_ptr->max_pkt_probes = 0;
+ if (l->owner->active_links[0] != l)
+ return l->owner->active_links[0];
+ return l->owner->active_links[1];
}
/*
@@ -307,10 +276,10 @@ struct tipc_link *tipc_link_create(struct tipc_node *n_ptr,
msg_set_bearer_id(msg, b_ptr->identity);
strcpy((char *)msg_data(msg), if_name);
l_ptr->net_plane = b_ptr->net_plane;
- link_init_max_pkt(l_ptr);
+ l_ptr->advertised_mtu = b_ptr->mtu;
+ l_ptr->mtu = l_ptr->advertised_mtu;
l_ptr->priority = b_ptr->priority;
tipc_link_set_queue_limits(l_ptr, b_ptr->window);
-
l_ptr->next_out_no = 1;
__skb_queue_head_init(&l_ptr->transmq);
__skb_queue_head_init(&l_ptr->backlogq);
@@ -327,15 +296,19 @@ struct tipc_link *tipc_link_create(struct tipc_node *n_ptr,
}
/**
- * link_delete - Conditional deletion of link.
- * If timer still running, real delete is done when it expires
- * @link: link to be deleted
+ * tipc_link_delete - Delete a link
+ * @l: link to be deleted
*/
-void tipc_link_delete(struct tipc_link *link)
+void tipc_link_delete(struct tipc_link *l)
{
- tipc_link_reset_fragments(link);
- tipc_node_detach_link(link->owner, link);
- tipc_link_put(link);
+ tipc_link_reset(l);
+ if (del_timer(&l->timer))
+ tipc_link_put(l);
+ l->flags |= LINK_STOPPED;
+ /* Delete link now, or when timer is finished: */
+ tipc_link_reset_fragments(l);
+ tipc_node_detach_link(l->owner, l);
+ tipc_link_put(l);
}
void tipc_link_delete_list(struct net *net, unsigned int bearer_id,
@@ -344,23 +317,12 @@ void tipc_link_delete_list(struct net *net, unsigned int bearer_id,
struct tipc_net *tn = net_generic(net, tipc_net_id);
struct tipc_link *link;
struct tipc_node *node;
- bool del_link;
rcu_read_lock();
list_for_each_entry_rcu(node, &tn->node_list, list) {
tipc_node_lock(node);
link = node->links[bearer_id];
- if (!link) {
- tipc_node_unlock(node);
- continue;
- }
- del_link = !tipc_link_is_up(link) && !link->exp_msg_count;
- tipc_link_reset(link);
- if (del_timer(&link->timer))
- tipc_link_put(link);
- link->flags |= LINK_STOPPED;
- /* Delete link now, or when failover is finished: */
- if (shutting_down || !tipc_node_is_up(node) || del_link)
+ if (link)
tipc_link_delete(link);
tipc_node_unlock(node);
}
@@ -368,28 +330,43 @@ void tipc_link_delete_list(struct net *net, unsigned int bearer_id,
}
/**
- * link_schedule_user - schedule user for wakeup after congestion
+ * link_schedule_user - schedule a message sender for wakeup after congestion
* @link: congested link
- * @oport: sending port
- * @chain_sz: size of buffer chain that was attempted sent
- * @imp: importance of message attempted sent
+ * @list: message that was attempted sent
* Create pseudo msg to send back to user when congestion abates
+ * Only consumes message if there is an error
*/
-static bool link_schedule_user(struct tipc_link *link, u32 oport,
- uint chain_sz, uint imp)
+static int link_schedule_user(struct tipc_link *link, struct sk_buff_head *list)
{
- struct sk_buff *buf;
+ struct tipc_msg *msg = buf_msg(skb_peek(list));
+ int imp = msg_importance(msg);
+ u32 oport = msg_origport(msg);
+ u32 addr = link_own_addr(link);
+ struct sk_buff *skb;
- buf = tipc_msg_create(SOCK_WAKEUP, 0, INT_H_SIZE, 0,
- link_own_addr(link), link_own_addr(link),
- oport, 0, 0);
- if (!buf)
- return false;
- TIPC_SKB_CB(buf)->chain_sz = chain_sz;
- TIPC_SKB_CB(buf)->chain_imp = imp;
- skb_queue_tail(&link->wakeupq, buf);
+ /* This really cannot happen... */
+ if (unlikely(imp > TIPC_CRITICAL_IMPORTANCE)) {
+ pr_warn("%s<%s>, send queue full", link_rst_msg, link->name);
+ tipc_link_reset(link);
+ goto err;
+ }
+ /* Non-blocking sender: */
+ if (TIPC_SKB_CB(skb_peek(list))->wakeup_pending)
+ return -ELINKCONG;
+
+ /* Create and schedule wakeup pseudo message */
+ skb = tipc_msg_create(SOCK_WAKEUP, 0, INT_H_SIZE, 0,
+ addr, addr, oport, 0, 0);
+ if (!skb)
+ goto err;
+ TIPC_SKB_CB(skb)->chain_sz = skb_queue_len(list);
+ TIPC_SKB_CB(skb)->chain_imp = imp;
+ skb_queue_tail(&link->wakeupq, skb);
link->stats.link_congs++;
- return true;
+ return -ELINKCONG;
+err:
+ __skb_queue_purge(list);
+ return -ENOBUFS;
}
/**
@@ -398,19 +375,22 @@ static bool link_schedule_user(struct tipc_link *link, u32 oport,
* Move a number of waiting users, as permitted by available space in
* the send queue, from link wait queue to node wait queue for wakeup
*/
-void link_prepare_wakeup(struct tipc_link *link)
+void link_prepare_wakeup(struct tipc_link *l)
{
- uint pend_qsz = skb_queue_len(&link->backlogq);
+ int pnd[TIPC_SYSTEM_IMPORTANCE + 1] = {0,};
+ int imp, lim;
struct sk_buff *skb, *tmp;
- skb_queue_walk_safe(&link->wakeupq, skb, tmp) {
- if (pend_qsz >= link->queue_limit[TIPC_SKB_CB(skb)->chain_imp])
+ skb_queue_walk_safe(&l->wakeupq, skb, tmp) {
+ imp = TIPC_SKB_CB(skb)->chain_imp;
+ lim = l->window + l->backlog[imp].limit;
+ pnd[imp] += TIPC_SKB_CB(skb)->chain_sz;
+ if ((pnd[imp] + l->backlog[imp].len) >= lim)
break;
- pend_qsz += TIPC_SKB_CB(skb)->chain_sz;
- skb_unlink(skb, &link->wakeupq);
- skb_queue_tail(&link->inputq, skb);
- link->owner->inputq = &link->inputq;
- link->owner->action_flags |= TIPC_MSG_EVT;
+ skb_unlink(skb, &l->wakeupq);
+ skb_queue_tail(&l->inputq, skb);
+ l->owner->inputq = &l->inputq;
+ l->owner->action_flags |= TIPC_MSG_EVT;
}
}
@@ -424,6 +404,16 @@ void tipc_link_reset_fragments(struct tipc_link *l_ptr)
l_ptr->reasm_buf = NULL;
}
+static void tipc_link_purge_backlog(struct tipc_link *l)
+{
+ __skb_queue_purge(&l->backlogq);
+ l->backlog[TIPC_LOW_IMPORTANCE].len = 0;
+ l->backlog[TIPC_MEDIUM_IMPORTANCE].len = 0;
+ l->backlog[TIPC_HIGH_IMPORTANCE].len = 0;
+ l->backlog[TIPC_CRITICAL_IMPORTANCE].len = 0;
+ l->backlog[TIPC_SYSTEM_IMPORTANCE].len = 0;
+}
+
/**
* tipc_link_purge_queues - purge all pkt queues associated with link
* @l_ptr: pointer to link
@@ -432,24 +422,24 @@ void tipc_link_purge_queues(struct tipc_link *l_ptr)
{
__skb_queue_purge(&l_ptr->deferdq);
__skb_queue_purge(&l_ptr->transmq);
- __skb_queue_purge(&l_ptr->backlogq);
+ tipc_link_purge_backlog(l_ptr);
tipc_link_reset_fragments(l_ptr);
}
void tipc_link_reset(struct tipc_link *l_ptr)
{
u32 prev_state = l_ptr->state;
- u32 checkpoint = l_ptr->next_in_no;
int was_active_link = tipc_link_is_active(l_ptr);
struct tipc_node *owner = l_ptr->owner;
+ struct tipc_link *pl = tipc_parallel_link(l_ptr);
msg_set_session(l_ptr->pmsg, ((msg_session(l_ptr->pmsg) + 1) & 0xffff));
/* Link is down, accept any session */
l_ptr->peer_session = INVALID_SESSION;
- /* Prepare for max packet size negotiation */
- link_init_max_pkt(l_ptr);
+ /* Prepare for renewed mtu size negotiation */
+ l_ptr->mtu = l_ptr->advertised_mtu;
l_ptr->state = RESET_UNKNOWN;
@@ -459,20 +449,25 @@ void tipc_link_reset(struct tipc_link *l_ptr)
tipc_node_link_down(l_ptr->owner, l_ptr);
tipc_bearer_remove_dest(owner->net, l_ptr->bearer_id, l_ptr->addr);
- if (was_active_link && tipc_node_active_links(l_ptr->owner)) {
- l_ptr->reset_checkpoint = checkpoint;
- l_ptr->exp_msg_count = START_CHANGEOVER;
+ if (was_active_link && tipc_node_is_up(l_ptr->owner) && (pl != l_ptr)) {
+ l_ptr->flags |= LINK_FAILINGOVER;
+ l_ptr->failover_checkpt = l_ptr->next_in_no;
+ pl->failover_pkts = FIRST_FAILOVER;
+ pl->failover_checkpt = l_ptr->next_in_no;
+ pl->failover_skb = l_ptr->reasm_buf;
+ } else {
+ kfree_skb(l_ptr->reasm_buf);
}
-
/* Clean up all queues, except inputq: */
__skb_queue_purge(&l_ptr->transmq);
- __skb_queue_purge(&l_ptr->backlogq);
__skb_queue_purge(&l_ptr->deferdq);
if (!owner->inputq)
owner->inputq = &l_ptr->inputq;
skb_queue_splice_init(&l_ptr->wakeupq, owner->inputq);
if (!skb_queue_empty(owner->inputq))
owner->action_flags |= TIPC_MSG_EVT;
+ tipc_link_purge_backlog(l_ptr);
+ l_ptr->reasm_buf = NULL;
l_ptr->rcv_unacked = 0;
l_ptr->checkpoint = 1;
l_ptr->next_out_no = 1;
@@ -524,8 +519,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
if (!(l_ptr->flags & LINK_STARTED) && (event != STARTING_EVT))
return; /* Not yet. */
- /* Check whether changeover is going on */
- if (l_ptr->exp_msg_count) {
+ if (l_ptr->flags & LINK_FAILINGOVER) {
if (event == TIMEOUT_EVT)
link_set_timer(l_ptr, cont_intv);
return;
@@ -542,11 +536,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->checkpoint = l_ptr->next_in_no;
if (tipc_bclink_acks_missing(l_ptr->owner)) {
tipc_link_proto_xmit(l_ptr, STATE_MSG,
- 0, 0, 0, 0, 0);
- l_ptr->fsm_msg_cnt++;
- } else if (l_ptr->max_pkt < l_ptr->max_pkt_target) {
- tipc_link_proto_xmit(l_ptr, STATE_MSG,
- 1, 0, 0, 0, 0);
+ 0, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
}
link_set_timer(l_ptr, cont_intv);
@@ -554,7 +544,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
}
l_ptr->state = WORKING_UNKNOWN;
l_ptr->fsm_msg_cnt = 0;
- tipc_link_proto_xmit(l_ptr, STATE_MSG, 1, 0, 0, 0, 0);
+ tipc_link_proto_xmit(l_ptr, STATE_MSG, 1, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv / 4);
break;
@@ -565,7 +555,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->state = RESET_RESET;
l_ptr->fsm_msg_cnt = 0;
tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG,
- 0, 0, 0, 0, 0);
+ 0, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv);
break;
@@ -588,7 +578,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->state = RESET_RESET;
l_ptr->fsm_msg_cnt = 0;
tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG,
- 0, 0, 0, 0, 0);
+ 0, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv);
break;
@@ -599,13 +589,13 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->checkpoint = l_ptr->next_in_no;
if (tipc_bclink_acks_missing(l_ptr->owner)) {
tipc_link_proto_xmit(l_ptr, STATE_MSG,
- 0, 0, 0, 0, 0);
+ 0, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
}
link_set_timer(l_ptr, cont_intv);
} else if (l_ptr->fsm_msg_cnt < l_ptr->abort_limit) {
tipc_link_proto_xmit(l_ptr, STATE_MSG,
- 1, 0, 0, 0, 0);
+ 1, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv / 4);
} else { /* Link has failed */
@@ -615,7 +605,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->state = RESET_UNKNOWN;
l_ptr->fsm_msg_cnt = 0;
tipc_link_proto_xmit(l_ptr, RESET_MSG,
- 0, 0, 0, 0, 0);
+ 0, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv);
}
@@ -635,7 +625,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->state = WORKING_WORKING;
l_ptr->fsm_msg_cnt = 0;
link_activate(l_ptr);
- tipc_link_proto_xmit(l_ptr, STATE_MSG, 1, 0, 0, 0, 0);
+ tipc_link_proto_xmit(l_ptr, STATE_MSG, 1, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
if (l_ptr->owner->working_links == 1)
tipc_link_sync_xmit(l_ptr);
@@ -645,7 +635,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->state = RESET_RESET;
l_ptr->fsm_msg_cnt = 0;
tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG,
- 1, 0, 0, 0, 0);
+ 1, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv);
break;
@@ -655,7 +645,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
link_set_timer(l_ptr, cont_intv);
break;
case TIMEOUT_EVT:
- tipc_link_proto_xmit(l_ptr, RESET_MSG, 0, 0, 0, 0, 0);
+ tipc_link_proto_xmit(l_ptr, RESET_MSG, 0, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv);
break;
@@ -673,7 +663,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
l_ptr->state = WORKING_WORKING;
l_ptr->fsm_msg_cnt = 0;
link_activate(l_ptr);
- tipc_link_proto_xmit(l_ptr, STATE_MSG, 1, 0, 0, 0, 0);
+ tipc_link_proto_xmit(l_ptr, STATE_MSG, 1, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
if (l_ptr->owner->working_links == 1)
tipc_link_sync_xmit(l_ptr);
@@ -683,7 +673,7 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
break;
case TIMEOUT_EVT:
tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG,
- 0, 0, 0, 0, 0);
+ 0, 0, 0, 0);
l_ptr->fsm_msg_cnt++;
link_set_timer(l_ptr, cont_intv);
break;
@@ -696,48 +686,15 @@ static void link_state_event(struct tipc_link *l_ptr, unsigned int event)
}
}
-/* tipc_link_cong: determine return value and how to treat the
- * sent buffer during link congestion.
- * - For plain, errorless user data messages we keep the buffer and
- * return -ELINKONG.
- * - For all other messages we discard the buffer and return -EHOSTUNREACH
- * - For TIPC internal messages we also reset the link
- */
-static int tipc_link_cong(struct tipc_link *link, struct sk_buff_head *list)
-{
- struct sk_buff *skb = skb_peek(list);
- struct tipc_msg *msg = buf_msg(skb);
- int imp = msg_importance(msg);
- u32 oport = msg_tot_origport(msg);
-
- if (unlikely(imp > TIPC_CRITICAL_IMPORTANCE)) {
- pr_warn("%s<%s>, send queue full", link_rst_msg, link->name);
- tipc_link_reset(link);
- goto drop;
- }
- if (unlikely(msg_errcode(msg)))
- goto drop;
- if (unlikely(msg_reroute_cnt(msg)))
- goto drop;
- if (TIPC_SKB_CB(skb)->wakeup_pending)
- return -ELINKCONG;
- if (link_schedule_user(link, oport, skb_queue_len(list), imp))
- return -ELINKCONG;
-drop:
- __skb_queue_purge(list);
- return -EHOSTUNREACH;
-}
-
/**
* __tipc_link_xmit(): same as tipc_link_xmit, but destlink is known & locked
* @link: link to use
* @list: chain of buffers containing message
*
- * Consumes the buffer chain, except when returning -ELINKCONG
- * Returns 0 if success, otherwise errno: -ELINKCONG, -EMSGSIZE (plain socket
- * user data messages) or -EHOSTUNREACH (all other messages/senders)
- * Only the socket functions tipc_send_stream() and tipc_send_packet() need
- * to act on the return value, since they may need to do more send attempts.
+ * Consumes the buffer chain, except when returning -ELINKCONG,
+ * since the caller then may want to make more send attempts.
+ * Returns 0 if success, or errno: -ELINKCONG, -EMSGSIZE or -ENOBUFS
+ * Messages at TIPC_SYSTEM_IMPORTANCE are always accepted
*/
int __tipc_link_xmit(struct net *net, struct tipc_link *link,
struct sk_buff_head *list)
@@ -745,7 +702,7 @@ int __tipc_link_xmit(struct net *net, struct tipc_link *link,
struct tipc_msg *msg = buf_msg(skb_peek(list));
unsigned int maxwin = link->window;
unsigned int imp = msg_importance(msg);
- uint mtu = link->max_pkt;
+ uint mtu = link->mtu;
uint ack = mod(link->next_in_no - 1);
uint seqno = link->next_out_no;
uint bc_last_in = link->owner->bclink.last_in;
@@ -754,16 +711,14 @@ int __tipc_link_xmit(struct net *net, struct tipc_link *link,
struct sk_buff_head *backlogq = &link->backlogq;
struct sk_buff *skb, *tmp;
- /* Match queue limit against msg importance: */
- if (unlikely(skb_queue_len(backlogq) >= link->queue_limit[imp]))
- return tipc_link_cong(link, list);
+ /* Match backlog limit against msg importance: */
+ if (unlikely(link->backlog[imp].len >= link->backlog[imp].limit))
+ return link_schedule_user(link, list);
- /* Has valid packet limit been used ? */
if (unlikely(msg_size(msg) > mtu)) {
__skb_queue_purge(list);
return -EMSGSIZE;
}
-
/* Prepare each packet for sending, and add to relevant queue: */
skb_queue_walk_safe(list, skb, tmp) {
__skb_unlink(skb, list);
@@ -786,8 +741,10 @@ int __tipc_link_xmit(struct net *net, struct tipc_link *link,
if (tipc_msg_make_bundle(&skb, mtu, link->addr)) {
link->stats.sent_bundled++;
link->stats.sent_bundles++;
+ imp = msg_importance(buf_msg(skb));
}
__skb_queue_tail(backlogq, skb);
+ link->backlog[imp].len++;
seqno++;
}
link->next_out_no = seqno;
@@ -808,13 +765,25 @@ static int __tipc_link_xmit_skb(struct tipc_link *link, struct sk_buff *skb)
return __tipc_link_xmit(link->owner->net, link, &head);
}
+/* tipc_link_xmit_skb(): send single buffer to destination
+ * Buffers sent via this functon are generally TIPC_SYSTEM_IMPORTANCE
+ * messages, which will not be rejected
+ * The only exception is datagram messages rerouted after secondary
+ * lookup, which are rare and safe to dispose of anyway.
+ * TODO: Return real return value, and let callers use
+ * tipc_wait_for_sendpkt() where applicable
+ */
int tipc_link_xmit_skb(struct net *net, struct sk_buff *skb, u32 dnode,
u32 selector)
{
struct sk_buff_head head;
+ int rc;
skb2list(skb, &head);
- return tipc_link_xmit(net, &head, dnode, selector);
+ rc = tipc_link_xmit(net, &head, dnode, selector);
+ if (rc == -ELINKCONG)
+ kfree_skb(skb);
+ return 0;
}
/**
@@ -841,12 +810,15 @@ int tipc_link_xmit(struct net *net, struct sk_buff_head *list, u32 dnode,
if (link)
rc = __tipc_link_xmit(net, link, list);
tipc_node_unlock(node);
+ tipc_node_put(node);
}
if (link)
return rc;
- if (likely(in_own_node(net, dnode)))
- return tipc_sk_rcv(net, list);
+ if (likely(in_own_node(net, dnode))) {
+ tipc_sk_rcv(net, list);
+ return 0;
+ }
__skb_queue_purge(list);
return rc;
@@ -912,6 +884,7 @@ void tipc_link_push_packets(struct tipc_link *link)
if (!skb)
break;
msg = buf_msg(skb);
+ link->backlog[msg_importance(msg)].len--;
msg_set_ack(msg, ack);
msg_set_bcast_ack(msg, link->owner->bclink.last_in);
link->rcv_unacked = 0;
@@ -964,7 +937,6 @@ static void link_retransmit_failure(struct tipc_link *l_ptr,
(unsigned long) TIPC_SKB_CB(buf)->handle);
n_ptr = tipc_bclink_retransmit_to(net);
- tipc_node_lock(n_ptr);
tipc_addr_string_fill(addr_string, n_ptr->addr);
pr_info("Broadcast link info for %s\n", addr_string);
@@ -976,9 +948,7 @@ static void link_retransmit_failure(struct tipc_link *l_ptr,
n_ptr->bclink.oos_state,
n_ptr->bclink.last_sent);
- tipc_node_unlock(n_ptr);
-
- tipc_bclink_set_flags(net, TIPC_BCLINK_RESET);
+ n_ptr->action_flags |= TIPC_BCAST_RESET;
l_ptr->stale_count = 0;
}
}
@@ -1017,6 +987,32 @@ void tipc_link_retransmit(struct tipc_link *l_ptr, struct sk_buff *skb,
}
}
+/* link_synch(): check if all packets arrived before the synch
+ * point have been consumed
+ * Returns true if the parallel links are synched, otherwise false
+ */
+static bool link_synch(struct tipc_link *l)
+{
+ unsigned int post_synch;
+ struct tipc_link *pl;
+
+ pl = tipc_parallel_link(l);
+ if (pl == l)
+ goto synched;
+
+ /* Was last pre-synch packet added to input queue ? */
+ if (less_eq(pl->next_in_no, l->synch_point))
+ return false;
+
+ /* Is it still in the input queue ? */
+ post_synch = mod(pl->next_in_no - l->synch_point) - 1;
+ if (skb_queue_len(&pl->inputq) > post_synch)
+ return false;
+synched:
+ l->flags &= ~LINK_SYNCHING;
+ return true;
+}
+
static void link_retrieve_defq(struct tipc_link *link,
struct sk_buff_head *list)
{
@@ -1077,8 +1073,8 @@ void tipc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *b_ptr)
n_ptr = tipc_node_find(net, msg_prevnode(msg));
if (unlikely(!n_ptr))
goto discard;
- tipc_node_lock(n_ptr);
+ tipc_node_lock(n_ptr);
/* Locate unicast link endpoint that should handle message */
l_ptr = n_ptr->links[b_ptr->identity];
if (unlikely(!l_ptr))
@@ -1147,17 +1143,26 @@ void tipc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *b_ptr)
skb = NULL;
goto unlock;
}
+ /* Synchronize with parallel link if applicable */
+ if (unlikely((l_ptr->flags & LINK_SYNCHING) && !msg_dup(msg))) {
+ link_handle_out_of_seq_msg(l_ptr, skb);
+ if (link_synch(l_ptr))
+ link_retrieve_defq(l_ptr, &head);
+ skb = NULL;
+ goto unlock;
+ }
l_ptr->next_in_no++;
if (unlikely(!skb_queue_empty(&l_ptr->deferdq)))
link_retrieve_defq(l_ptr, &head);
if (unlikely(++l_ptr->rcv_unacked >= TIPC_MIN_LINK_WIN)) {
l_ptr->stats.sent_acks++;
- tipc_link_proto_xmit(l_ptr, STATE_MSG, 0, 0, 0, 0, 0);
+ tipc_link_proto_xmit(l_ptr, STATE_MSG, 0, 0, 0, 0);
}
tipc_link_input(l_ptr, skb);
skb = NULL;
unlock:
tipc_node_unlock(n_ptr);
+ tipc_node_put(n_ptr);
discard:
if (unlikely(skb))
kfree_skb(skb);
@@ -1194,7 +1199,7 @@ static bool tipc_data_input(struct tipc_link *link, struct sk_buff *skb)
node->action_flags |= TIPC_NAMED_MSG_EVT;
return true;
case MSG_BUNDLER:
- case CHANGEOVER_PROTOCOL:
+ case TUNNEL_PROTOCOL:
case MSG_FRAGMENTER:
case BCAST_PROTOCOL:
return false;
@@ -1221,8 +1226,14 @@ static void tipc_link_input(struct tipc_link *link, struct sk_buff *skb)
return;
switch (msg_user(msg)) {
- case CHANGEOVER_PROTOCOL:
- if (!tipc_link_tunnel_rcv(node, &skb))
+ case TUNNEL_PROTOCOL:
+ if (msg_dup(msg)) {
+ link->flags |= LINK_SYNCHING;
+ link->synch_point = msg_seqno(msg_get_wrapped(msg));
+ kfree_skb(skb);
+ break;
+ }
+ if (!tipc_link_failover_rcv(link, &skb))
break;
if (msg_user(buf_msg(skb)) != MSG_BUNDLER) {
tipc_data_input(link, skb);
@@ -1320,7 +1331,7 @@ static void link_handle_out_of_seq_msg(struct tipc_link *l_ptr,
if (tipc_link_defer_pkt(&l_ptr->deferdq, buf)) {
l_ptr->stats.deferred_recv++;
if ((skb_queue_len(&l_ptr->deferdq) % TIPC_MIN_LINK_WIN) == 1)
- tipc_link_proto_xmit(l_ptr, STATE_MSG, 0, 0, 0, 0, 0);
+ tipc_link_proto_xmit(l_ptr, STATE_MSG, 0, 0, 0, 0);
} else {
l_ptr->stats.duplicates++;
}
@@ -1330,15 +1341,15 @@ static void link_handle_out_of_seq_msg(struct tipc_link *l_ptr,
* Send protocol message to the other endpoint.
*/
void tipc_link_proto_xmit(struct tipc_link *l_ptr, u32 msg_typ, int probe_msg,
- u32 gap, u32 tolerance, u32 priority, u32 ack_mtu)
+ u32 gap, u32 tolerance, u32 priority)
{
struct sk_buff *buf = NULL;
struct tipc_msg *msg = l_ptr->pmsg;
u32 msg_size = sizeof(l_ptr->proto_msg);
int r_flag;
- /* Don't send protocol message during link changeover */
- if (l_ptr->exp_msg_count)
+ /* Don't send protocol message during link failover */
+ if (l_ptr->flags & LINK_FAILINGOVER)
return;
/* Abort non-RESET send if communication with node is prohibited */
@@ -1368,35 +1379,20 @@ void tipc_link_proto_xmit(struct tipc_link *l_ptr, u32 msg_typ, int probe_msg,
l_ptr->stats.sent_nacks++;
msg_set_link_tolerance(msg, tolerance);
msg_set_linkprio(msg, priority);
- msg_set_max_pkt(msg, ack_mtu);
+ msg_set_max_pkt(msg, l_ptr->mtu);
msg_set_ack(msg, mod(l_ptr->next_in_no - 1));
msg_set_probe(msg, probe_msg != 0);
- if (probe_msg) {
- u32 mtu = l_ptr->max_pkt;
-
- if ((mtu < l_ptr->max_pkt_target) &&
- link_working_working(l_ptr) &&
- l_ptr->fsm_msg_cnt) {
- msg_size = (mtu + (l_ptr->max_pkt_target - mtu)/2 + 2) & ~3;
- if (l_ptr->max_pkt_probes == 10) {
- l_ptr->max_pkt_target = (msg_size - 4);
- l_ptr->max_pkt_probes = 0;
- msg_size = (mtu + (l_ptr->max_pkt_target - mtu)/2 + 2) & ~3;
- }
- l_ptr->max_pkt_probes++;
- }
-
+ if (probe_msg)
l_ptr->stats.sent_probes++;
- }
l_ptr->stats.sent_states++;
} else { /* RESET_MSG or ACTIVATE_MSG */
- msg_set_ack(msg, mod(l_ptr->reset_checkpoint - 1));
+ msg_set_ack(msg, mod(l_ptr->failover_checkpt - 1));
msg_set_seq_gap(msg, 0);
msg_set_next_sent(msg, 1);
msg_set_probe(msg, 0);
msg_set_link_tolerance(msg, l_ptr->tolerance);
msg_set_linkprio(msg, l_ptr->priority);
- msg_set_max_pkt(msg, l_ptr->max_pkt_target);
+ msg_set_max_pkt(msg, l_ptr->advertised_mtu);
}
r_flag = (l_ptr->owner->working_links > tipc_link_is_up(l_ptr));
@@ -1427,13 +1423,10 @@ static void tipc_link_proto_rcv(struct tipc_link *l_ptr,
struct sk_buff *buf)
{
u32 rec_gap = 0;
- u32 max_pkt_info;
- u32 max_pkt_ack;
u32 msg_tol;
struct tipc_msg *msg = buf_msg(buf);
- /* Discard protocol message during link changeover */
- if (l_ptr->exp_msg_count)
+ if (l_ptr->flags & LINK_FAILINGOVER)
goto exit;
if (l_ptr->net_plane != msg_net_plane(msg))
@@ -1472,15 +1465,8 @@ static void tipc_link_proto_rcv(struct tipc_link *l_ptr,
if (msg_linkprio(msg) > l_ptr->priority)
l_ptr->priority = msg_linkprio(msg);
- max_pkt_info = msg_max_pkt(msg);
- if (max_pkt_info) {
- if (max_pkt_info < l_ptr->max_pkt_target)
- l_ptr->max_pkt_target = max_pkt_info;
- if (l_ptr->max_pkt > l_ptr->max_pkt_target)
- l_ptr->max_pkt = l_ptr->max_pkt_target;
- } else {
- l_ptr->max_pkt = l_ptr->max_pkt_target;
- }
+ if (l_ptr->mtu > msg_max_pkt(msg))
+ l_ptr->mtu = msg_max_pkt(msg);
/* Synchronize broadcast link info, if not done previously */
if (!tipc_node_is_up(l_ptr->owner)) {
@@ -1525,18 +1511,8 @@ static void tipc_link_proto_rcv(struct tipc_link *l_ptr,
mod(l_ptr->next_in_no));
}
- max_pkt_ack = msg_max_pkt(msg);
- if (max_pkt_ack > l_ptr->max_pkt) {
- l_ptr->max_pkt = max_pkt_ack;
- l_ptr->max_pkt_probes = 0;
- }
-
- max_pkt_ack = 0;
- if (msg_probe(msg)) {
+ if (msg_probe(msg))
l_ptr->stats.recv_probes++;
- if (msg_size(msg) > sizeof(l_ptr->proto_msg))
- max_pkt_ack = msg_size(msg);
- }
/* Protocol message before retransmits, reduce loss risk */
if (l_ptr->owner->bclink.recv_permitted)
@@ -1544,8 +1520,8 @@ static void tipc_link_proto_rcv(struct tipc_link *l_ptr,
msg_last_bcast(msg));
if (rec_gap || (msg_probe(msg))) {
- tipc_link_proto_xmit(l_ptr, STATE_MSG, 0, rec_gap, 0,
- 0, max_pkt_ack);
+ tipc_link_proto_xmit(l_ptr, STATE_MSG, 0,
+ rec_gap, 0, 0);
}
if (msg_seq_gap(msg)) {
l_ptr->stats.recv_nacks++;
@@ -1605,9 +1581,10 @@ void tipc_link_failover_send_queue(struct tipc_link *l_ptr)
if (!tunnel)
return;
- tipc_msg_init(link_own_addr(l_ptr), &tunnel_hdr, CHANGEOVER_PROTOCOL,
- ORIGINAL_MSG, INT_H_SIZE, l_ptr->addr);
+ tipc_msg_init(link_own_addr(l_ptr), &tunnel_hdr, TUNNEL_PROTOCOL,
+ FAILOVER_MSG, INT_H_SIZE, l_ptr->addr);
skb_queue_splice_tail_init(&l_ptr->backlogq, &l_ptr->transmq);
+ tipc_link_purge_backlog(l_ptr);
msgcount = skb_queue_len(&l_ptr->transmq);
msg_set_bearer_id(&tunnel_hdr, l_ptr->peer_bearer_id);
msg_set_msgcnt(&tunnel_hdr, msgcount);
@@ -1667,8 +1644,8 @@ void tipc_link_dup_queue_xmit(struct tipc_link *link,
struct sk_buff_head *queue = &link->transmq;
int mcnt;
- tipc_msg_init(link_own_addr(link), &tnl_hdr, CHANGEOVER_PROTOCOL,
- DUPLICATE_MSG, INT_H_SIZE, link->addr);
+ tipc_msg_init(link_own_addr(link), &tnl_hdr, TUNNEL_PROTOCOL,
+ SYNCH_MSG, INT_H_SIZE, link->addr);
mcnt = skb_queue_len(&link->transmq) + skb_queue_len(&link->backlogq);
msg_set_msgcnt(&tnl_hdr, mcnt);
msg_set_bearer_id(&tnl_hdr, link->peer_bearer_id);
@@ -1701,101 +1678,63 @@ tunnel_queue:
goto tunnel_queue;
}
-/* tipc_link_dup_rcv(): Receive a tunnelled DUPLICATE_MSG packet.
- * Owner node is locked.
- */
-static void tipc_link_dup_rcv(struct tipc_link *link,
- struct sk_buff *skb)
-{
- struct sk_buff *iskb;
- int pos = 0;
-
- if (!tipc_link_is_up(link))
- return;
-
- if (!tipc_msg_extract(skb, &iskb, &pos)) {
- pr_warn("%sfailed to extract inner dup pkt\n", link_co_err);
- return;
- }
- /* Append buffer to deferred queue, if applicable: */
- link_handle_out_of_seq_msg(link, iskb);
-}
-
-/* tipc_link_failover_rcv(): Receive a tunnelled ORIGINAL_MSG packet
+/* tipc_link_failover_rcv(): Receive a tunnelled FAILOVER_MSG packet
* Owner node is locked.
*/
-static struct sk_buff *tipc_link_failover_rcv(struct tipc_link *l_ptr,
- struct sk_buff *t_buf)
+static bool tipc_link_failover_rcv(struct tipc_link *link,
+ struct sk_buff **skb)
{
- struct tipc_msg *t_msg = buf_msg(t_buf);
- struct sk_buff *buf = NULL;
- struct tipc_msg *msg;
+ struct tipc_msg *msg = buf_msg(*skb);
+ struct sk_buff *iskb = NULL;
+ struct tipc_link *pl = NULL;
+ int bearer_id = msg_bearer_id(msg);
int pos = 0;
- if (tipc_link_is_up(l_ptr))
- tipc_link_reset(l_ptr);
-
- /* First failover packet? */
- if (l_ptr->exp_msg_count == START_CHANGEOVER)
- l_ptr->exp_msg_count = msg_msgcnt(t_msg);
-
- /* Should there be an inner packet? */
- if (l_ptr->exp_msg_count) {
- l_ptr->exp_msg_count--;
- if (!tipc_msg_extract(t_buf, &buf, &pos)) {
- pr_warn("%sno inner failover pkt\n", link_co_err);
- goto exit;
- }
- msg = buf_msg(buf);
-
- if (less(msg_seqno(msg), l_ptr->reset_checkpoint)) {
- kfree_skb(buf);
- buf = NULL;
- goto exit;
- }
- if (msg_user(msg) == MSG_FRAGMENTER) {
- l_ptr->stats.recv_fragments++;
- tipc_buf_append(&l_ptr->reasm_buf, &buf);
- }
+ if (msg_type(msg) != FAILOVER_MSG) {
+ pr_warn("%sunknown tunnel pkt received\n", link_co_err);
+ goto exit;
}
-exit:
- if ((!l_ptr->exp_msg_count) && (l_ptr->flags & LINK_STOPPED))
- tipc_link_delete(l_ptr);
- return buf;
-}
+ if (bearer_id >= MAX_BEARERS)
+ goto exit;
-/* tipc_link_tunnel_rcv(): Receive a tunnelled packet, sent
- * via other link as result of a failover (ORIGINAL_MSG) or
- * a new active link (DUPLICATE_MSG). Failover packets are
- * returned to the active link for delivery upwards.
- * Owner node is locked.
- */
-static int tipc_link_tunnel_rcv(struct tipc_node *n_ptr,
- struct sk_buff **buf)
-{
- struct sk_buff *t_buf = *buf;
- struct tipc_link *l_ptr;
- struct tipc_msg *t_msg = buf_msg(t_buf);
- u32 bearer_id = msg_bearer_id(t_msg);
+ if (bearer_id == link->bearer_id)
+ goto exit;
- *buf = NULL;
+ pl = link->owner->links[bearer_id];
+ if (pl && tipc_link_is_up(pl))
+ tipc_link_reset(pl);
- if (bearer_id >= MAX_BEARERS)
+ if (link->failover_pkts == FIRST_FAILOVER)
+ link->failover_pkts = msg_msgcnt(msg);
+
+ /* Should we expect an inner packet? */
+ if (!link->failover_pkts)
goto exit;
- l_ptr = n_ptr->links[bearer_id];
- if (!l_ptr)
+ if (!tipc_msg_extract(*skb, &iskb, &pos)) {
+ pr_warn("%sno inner failover pkt\n", link_co_err);
+ *skb = NULL;
goto exit;
+ }
+ link->failover_pkts--;
+ *skb = NULL;
- if (msg_type(t_msg) == DUPLICATE_MSG)
- tipc_link_dup_rcv(l_ptr, t_buf);
- else if (msg_type(t_msg) == ORIGINAL_MSG)
- *buf = tipc_link_failover_rcv(l_ptr, t_buf);
- else
- pr_warn("%sunknown tunnel pkt received\n", link_co_err);
+ /* Was this packet already delivered? */
+ if (less(buf_seqno(iskb), link->failover_checkpt)) {
+ kfree_skb(iskb);
+ iskb = NULL;
+ goto exit;
+ }
+ if (msg_user(buf_msg(iskb)) == MSG_FRAGMENTER) {
+ link->stats.recv_fragments++;
+ tipc_buf_append(&link->failover_skb, &iskb);
+ }
exit:
- kfree_skb(t_buf);
- return *buf != NULL;
+ if (!link->failover_pkts && pl)
+ pl->flags &= ~LINK_FAILINGOVER;
+ kfree_skb(*skb);
+ *skb = iskb;
+ return *skb;
}
static void link_set_supervision_props(struct tipc_link *l_ptr, u32 tol)
@@ -1812,14 +1751,14 @@ static void link_set_supervision_props(struct tipc_link *l_ptr, u32 tol)
void tipc_link_set_queue_limits(struct tipc_link *l, u32 win)
{
- int max_bulk = TIPC_MAX_PUBLICATIONS / (l->max_pkt / ITEM_SIZE);
+ int max_bulk = TIPC_MAX_PUBLICATIONS / (l->mtu / ITEM_SIZE);
l->window = win;
- l->queue_limit[TIPC_LOW_IMPORTANCE] = win / 2;
- l->queue_limit[TIPC_MEDIUM_IMPORTANCE] = win;
- l->queue_limit[TIPC_HIGH_IMPORTANCE] = win / 2 * 3;
- l->queue_limit[TIPC_CRITICAL_IMPORTANCE] = win * 2;
- l->queue_limit[TIPC_SYSTEM_IMPORTANCE] = max_bulk;
+ l->backlog[TIPC_LOW_IMPORTANCE].limit = win / 2;
+ l->backlog[TIPC_MEDIUM_IMPORTANCE].limit = win;
+ l->backlog[TIPC_HIGH_IMPORTANCE].limit = win / 2 * 3;
+ l->backlog[TIPC_CRITICAL_IMPORTANCE].limit = win * 2;
+ l->backlog[TIPC_SYSTEM_IMPORTANCE].limit = max_bulk;
}
/* tipc_link_find_owner - locate owner node of link by link's name
@@ -1984,14 +1923,14 @@ int tipc_nl_link_set(struct sk_buff *skb, struct genl_info *info)
tol = nla_get_u32(props[TIPC_NLA_PROP_TOL]);
link_set_supervision_props(link, tol);
- tipc_link_proto_xmit(link, STATE_MSG, 0, 0, tol, 0, 0);
+ tipc_link_proto_xmit(link, STATE_MSG, 0, 0, tol, 0);
}
if (props[TIPC_NLA_PROP_PRIO]) {
u32 prio;
prio = nla_get_u32(props[TIPC_NLA_PROP_PRIO]);
link->priority = prio;
- tipc_link_proto_xmit(link, STATE_MSG, 0, 0, 0, prio, 0);
+ tipc_link_proto_xmit(link, STATE_MSG, 0, 0, 0, prio);
}
if (props[TIPC_NLA_PROP_WIN]) {
u32 win;
@@ -2096,7 +2035,7 @@ static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg,
if (nla_put_u32(msg->skb, TIPC_NLA_LINK_DEST,
tipc_cluster_mask(tn->own_addr)))
goto attr_msg_full;
- if (nla_put_u32(msg->skb, TIPC_NLA_LINK_MTU, link->max_pkt))
+ if (nla_put_u32(msg->skb, TIPC_NLA_LINK_MTU, link->mtu))
goto attr_msg_full;
if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, link->next_in_no))
goto attr_msg_full;
@@ -2118,7 +2057,7 @@ static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg,
if (nla_put_u32(msg->skb, TIPC_NLA_PROP_TOL, link->tolerance))
goto prop_msg_full;
if (nla_put_u32(msg->skb, TIPC_NLA_PROP_WIN,
- link->queue_limit[TIPC_LOW_IMPORTANCE]))
+ link->window))
goto prop_msg_full;
if (nla_put_u32(msg->skb, TIPC_NLA_PROP_PRIO, link->priority))
goto prop_msg_full;
@@ -2184,7 +2123,6 @@ int tipc_nl_link_dump(struct sk_buff *skb, struct netlink_callback *cb)
msg.seq = cb->nlh->nlmsg_seq;
rcu_read_lock();
-
if (prev_node) {
node = tipc_node_find(net, prev_node);
if (!node) {
@@ -2197,6 +2135,7 @@ int tipc_nl_link_dump(struct sk_buff *skb, struct netlink_callback *cb)
cb->prev_seq = 1;
goto out;
}
+ tipc_node_put(node);
list_for_each_entry_continue_rcu(node, &tn->node_list,
list) {
@@ -2204,6 +2143,7 @@ int tipc_nl_link_dump(struct sk_buff *skb, struct netlink_callback *cb)
err = __tipc_nl_add_node_links(net, &msg, node,
&prev_link);
tipc_node_unlock(node);
+ tipc_node_put(node);
if (err)
goto out;
diff --git a/net/tipc/link.h b/net/tipc/link.h
index eec3ecf2d450..b5b4e3554d4e 100644
--- a/net/tipc/link.h
+++ b/net/tipc/link.h
@@ -58,8 +58,10 @@
/* Link endpoint execution states
*/
-#define LINK_STARTED 0x0001
-#define LINK_STOPPED 0x0002
+#define LINK_STARTED 0x0001
+#define LINK_STOPPED 0x0002
+#define LINK_SYNCHING 0x0004
+#define LINK_FAILINGOVER 0x0008
/* Starting value for maximum packet size negotiation on unicast links
* (unless bearer MTU is less)
@@ -118,12 +120,11 @@ struct tipc_stats {
* @pmsg: convenience pointer to "proto_msg" field
* @priority: current link priority
* @net_plane: current link network plane ('A' through 'H')
- * @queue_limit: outbound message queue congestion thresholds (indexed by user)
+ * @backlog_limit: backlog queue congestion thresholds (indexed by importance)
* @exp_msg_count: # of tunnelled messages expected during link changeover
* @reset_checkpoint: seq # of last acknowledged message at time of link reset
- * @max_pkt: current maximum packet size for this link
- * @max_pkt_target: desired maximum packet size for this link
- * @max_pkt_probes: # of probes based on current (max_pkt, max_pkt_target)
+ * @mtu: current maximum packet size for this link
+ * @advertised_mtu: advertised own mtu when link is being established
* @transmitq: queue for sent, non-acked messages
* @backlogq: queue for messages waiting to be sent
* @next_out_no: next sequence number to use for outbound messages
@@ -166,20 +167,24 @@ struct tipc_link {
struct tipc_msg *pmsg;
u32 priority;
char net_plane;
- u32 queue_limit[15]; /* queue_limit[0]==window limit */
+ u16 synch_point;
- /* Changeover */
- u32 exp_msg_count;
- u32 reset_checkpoint;
+ /* Failover */
+ u16 failover_pkts;
+ u16 failover_checkpt;
+ struct sk_buff *failover_skb;
/* Max packet negotiation */
- u32 max_pkt;
- u32 max_pkt_target;
- u32 max_pkt_probes;
+ u16 mtu;
+ u16 advertised_mtu;
/* Sending */
struct sk_buff_head transmq;
struct sk_buff_head backlogq;
+ struct {
+ u16 len;
+ u16 limit;
+ } backlog[5];
u32 next_out_no;
u32 window;
u32 last_retransmitted;
@@ -196,7 +201,6 @@ struct tipc_link {
struct sk_buff_head wakeupq;
/* Fragmentation/reassembly */
- u32 long_msg_seq_no;
struct sk_buff *reasm_buf;
/* Statistics */
@@ -227,7 +231,7 @@ int tipc_link_xmit(struct net *net, struct sk_buff_head *list, u32 dest,
int __tipc_link_xmit(struct net *net, struct tipc_link *link,
struct sk_buff_head *list);
void tipc_link_proto_xmit(struct tipc_link *l_ptr, u32 msg_typ, int prob,
- u32 gap, u32 tolerance, u32 priority, u32 acked_mtu);
+ u32 gap, u32 tolerance, u32 priority);
void tipc_link_push_packets(struct tipc_link *l_ptr);
u32 tipc_link_defer_pkt(struct sk_buff_head *list, struct sk_buff *buf);
void tipc_link_set_queue_limits(struct tipc_link *l_ptr, u32 window);
diff --git a/net/tipc/msg.c b/net/tipc/msg.c
index 0c6dad8180a0..c3e96e815418 100644
--- a/net/tipc/msg.c
+++ b/net/tipc/msg.c
@@ -355,7 +355,7 @@ bool tipc_msg_bundle(struct sk_buff *bskb, struct sk_buff *skb, u32 mtu)
start = align(bsz);
pad = start - bsz;
- if (unlikely(msg_user(msg) == CHANGEOVER_PROTOCOL))
+ if (unlikely(msg_user(msg) == TUNNEL_PROTOCOL))
return false;
if (unlikely(msg_user(msg) == BCAST_PROTOCOL))
return false;
@@ -433,7 +433,7 @@ bool tipc_msg_make_bundle(struct sk_buff **skb, u32 mtu, u32 dnode)
if (msg_user(msg) == MSG_FRAGMENTER)
return false;
- if (msg_user(msg) == CHANGEOVER_PROTOCOL)
+ if (msg_user(msg) == TUNNEL_PROTOCOL)
return false;
if (msg_user(msg) == BCAST_PROTOCOL)
return false;
@@ -511,15 +511,18 @@ bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb,
{
struct tipc_msg *msg = buf_msg(skb);
u32 dport;
+ u32 own_addr = tipc_own_addr(net);
if (!msg_isdata(msg))
return false;
if (!msg_named(msg))
return false;
+ if (msg_errcode(msg))
+ return false;
*err = -TIPC_ERR_NO_NAME;
if (skb_linearize(skb))
return false;
- if (msg_reroute_cnt(msg) > 0)
+ if (msg_reroute_cnt(msg))
return false;
*dnode = addr_domain(net, msg_lookup_scope(msg));
dport = tipc_nametbl_translate(net, msg_nametype(msg),
@@ -527,6 +530,8 @@ bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb,
if (!dport)
return false;
msg_incr_reroute_cnt(msg);
+ if (*dnode != own_addr)
+ msg_set_prevnode(msg, own_addr);
msg_set_destnode(msg, *dnode);
msg_set_destport(msg, dport);
*err = TIPC_OK;
diff --git a/net/tipc/msg.h b/net/tipc/msg.h
index bd3969a80dd4..e1d3595e2ee9 100644
--- a/net/tipc/msg.h
+++ b/net/tipc/msg.h
@@ -72,7 +72,7 @@ struct plist;
#define MSG_BUNDLER 6
#define LINK_PROTOCOL 7
#define CONN_MANAGER 8
-#define CHANGEOVER_PROTOCOL 10
+#define TUNNEL_PROTOCOL 10
#define NAME_DISTRIBUTOR 11
#define MSG_FRAGMENTER 12
#define LINK_CONFIG 13
@@ -240,6 +240,15 @@ static inline void msg_set_size(struct tipc_msg *m, u32 sz)
m->hdr[0] = htonl((msg_word(m, 0) & ~0x1ffff) | sz);
}
+static inline unchar *msg_data(struct tipc_msg *m)
+{
+ return ((unchar *)m) + msg_hdr_sz(m);
+}
+
+static inline struct tipc_msg *msg_get_wrapped(struct tipc_msg *m)
+{
+ return (struct tipc_msg *)msg_data(m);
+}
/*
* Word 1
@@ -372,6 +381,8 @@ static inline void msg_set_prevnode(struct tipc_msg *m, u32 a)
static inline u32 msg_origport(struct tipc_msg *m)
{
+ if (msg_user(m) == MSG_FRAGMENTER)
+ m = msg_get_wrapped(m);
return msg_word(m, 4);
}
@@ -467,16 +478,6 @@ static inline void msg_set_nameupper(struct tipc_msg *m, u32 n)
msg_set_word(m, 10, n);
}
-static inline unchar *msg_data(struct tipc_msg *m)
-{
- return ((unchar *)m) + msg_hdr_sz(m);
-}
-
-static inline struct tipc_msg *msg_get_wrapped(struct tipc_msg *m)
-{
- return (struct tipc_msg *)msg_data(m);
-}
-
/*
* Constants and routines used to read and write TIPC internal message headers
*/
@@ -511,8 +512,8 @@ static inline struct tipc_msg *msg_get_wrapped(struct tipc_msg *m)
/*
* Changeover tunnel message types
*/
-#define DUPLICATE_MSG 0
-#define ORIGINAL_MSG 1
+#define SYNCH_MSG 0
+#define FAILOVER_MSG 1
/*
* Config protocol message types
@@ -553,6 +554,14 @@ static inline void msg_set_node_capabilities(struct tipc_msg *m, u32 n)
msg_set_bits(m, 1, 15, 0x1fff, n);
}
+static inline bool msg_dup(struct tipc_msg *m)
+{
+ if (likely(msg_user(m) != TUNNEL_PROTOCOL))
+ return false;
+ if (msg_type(m) != SYNCH_MSG)
+ return false;
+ return true;
+}
/*
* Word 2
@@ -753,13 +762,6 @@ static inline void msg_set_link_tolerance(struct tipc_msg *m, u32 n)
msg_set_bits(m, 9, 0, 0xffff, n);
}
-static inline u32 msg_tot_origport(struct tipc_msg *m)
-{
- if ((msg_user(m) == MSG_FRAGMENTER) && (msg_type(m) == FIRST_FRAGMENT))
- return msg_origport(msg_get_wrapped(m));
- return msg_origport(m);
-}
-
struct sk_buff *tipc_buf_acquire(u32 size);
bool tipc_msg_validate(struct sk_buff *skb);
bool tipc_msg_reverse(u32 own_addr, struct sk_buff *buf, u32 *dnode,
diff --git a/net/tipc/name_distr.c b/net/tipc/name_distr.c
index 506aaa565da7..41e7b7e4dda0 100644
--- a/net/tipc/name_distr.c
+++ b/net/tipc/name_distr.c
@@ -244,6 +244,7 @@ static void tipc_publ_subscribe(struct net *net, struct publication *publ,
tipc_node_lock(node);
list_add_tail(&publ->nodesub_list, &node->publ_list);
tipc_node_unlock(node);
+ tipc_node_put(node);
}
static void tipc_publ_unsubscribe(struct net *net, struct publication *publ,
@@ -258,6 +259,7 @@ static void tipc_publ_unsubscribe(struct net *net, struct publication *publ,
tipc_node_lock(node);
list_del_init(&publ->nodesub_list);
tipc_node_unlock(node);
+ tipc_node_put(node);
}
/**
diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c
index 105ba7adf06f..ab0ac62a1287 100644
--- a/net/tipc/name_table.c
+++ b/net/tipc/name_table.c
@@ -811,8 +811,8 @@ static void tipc_purge_publications(struct net *net, struct name_seq *seq)
sseq = seq->sseqs;
info = sseq->info;
list_for_each_entry_safe(publ, safe, &info->zone_list, zone_list) {
- tipc_nametbl_remove_publ(net, publ->type, publ->lower,
- publ->node, publ->ref, publ->key);
+ tipc_nameseq_remove_publ(net, seq, publ->lower, publ->node,
+ publ->ref, publ->key);
kfree_rcu(publ, rcu);
}
hlist_del_init_rcu(&seq->ns_list);
diff --git a/net/tipc/node.c b/net/tipc/node.c
index 26d1de1bf34d..22c059ad2999 100644
--- a/net/tipc/node.c
+++ b/net/tipc/node.c
@@ -42,6 +42,7 @@
static void node_lost_contact(struct tipc_node *n_ptr);
static void node_established_contact(struct tipc_node *n_ptr);
+static void tipc_node_delete(struct tipc_node *node);
struct tipc_sock_conn {
u32 port;
@@ -67,6 +68,23 @@ static unsigned int tipc_hashfn(u32 addr)
return addr & (NODE_HTABLE_SIZE - 1);
}
+static void tipc_node_kref_release(struct kref *kref)
+{
+ struct tipc_node *node = container_of(kref, struct tipc_node, kref);
+
+ tipc_node_delete(node);
+}
+
+void tipc_node_put(struct tipc_node *node)
+{
+ kref_put(&node->kref, tipc_node_kref_release);
+}
+
+static void tipc_node_get(struct tipc_node *node)
+{
+ kref_get(&node->kref);
+}
+
/*
* tipc_node_find - locate specified node object, if it exists
*/
@@ -82,6 +100,7 @@ struct tipc_node *tipc_node_find(struct net *net, u32 addr)
hlist_for_each_entry_rcu(node, &tn->node_htable[tipc_hashfn(addr)],
hash) {
if (node->addr == addr) {
+ tipc_node_get(node);
rcu_read_unlock();
return node;
}
@@ -106,6 +125,7 @@ struct tipc_node *tipc_node_create(struct net *net, u32 addr)
}
n_ptr->addr = addr;
n_ptr->net = net;
+ kref_init(&n_ptr->kref);
spin_lock_init(&n_ptr->lock);
INIT_HLIST_NODE(&n_ptr->hash);
INIT_LIST_HEAD(&n_ptr->list);
@@ -120,16 +140,17 @@ struct tipc_node *tipc_node_create(struct net *net, u32 addr)
list_add_tail_rcu(&n_ptr->list, &temp_node->list);
n_ptr->action_flags = TIPC_WAIT_PEER_LINKS_DOWN;
n_ptr->signature = INVALID_NODE_SIG;
+ tipc_node_get(n_ptr);
exit:
spin_unlock_bh(&tn->node_list_lock);
return n_ptr;
}
-static void tipc_node_delete(struct tipc_net *tn, struct tipc_node *n_ptr)
+static void tipc_node_delete(struct tipc_node *node)
{
- list_del_rcu(&n_ptr->list);
- hlist_del_rcu(&n_ptr->hash);
- kfree_rcu(n_ptr, rcu);
+ list_del_rcu(&node->list);
+ hlist_del_rcu(&node->hash);
+ kfree_rcu(node, rcu);
}
void tipc_node_stop(struct net *net)
@@ -139,7 +160,7 @@ void tipc_node_stop(struct net *net)
spin_lock_bh(&tn->node_list_lock);
list_for_each_entry_safe(node, t_node, &tn->node_list, list)
- tipc_node_delete(tn, node);
+ tipc_node_put(node);
spin_unlock_bh(&tn->node_list_lock);
}
@@ -147,6 +168,7 @@ int tipc_node_add_conn(struct net *net, u32 dnode, u32 port, u32 peer_port)
{
struct tipc_node *node;
struct tipc_sock_conn *conn;
+ int err = 0;
if (in_own_node(net, dnode))
return 0;
@@ -157,8 +179,10 @@ int tipc_node_add_conn(struct net *net, u32 dnode, u32 port, u32 peer_port)
return -EHOSTUNREACH;
}
conn = kmalloc(sizeof(*conn), GFP_ATOMIC);
- if (!conn)
- return -EHOSTUNREACH;
+ if (!conn) {
+ err = -EHOSTUNREACH;
+ goto exit;
+ }
conn->peer_node = dnode;
conn->port = port;
conn->peer_port = peer_port;
@@ -166,7 +190,9 @@ int tipc_node_add_conn(struct net *net, u32 dnode, u32 port, u32 peer_port)
tipc_node_lock(node);
list_add_tail(&conn->list, &node->conn_sks);
tipc_node_unlock(node);
- return 0;
+exit:
+ tipc_node_put(node);
+ return err;
}
void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port)
@@ -189,6 +215,7 @@ void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port)
kfree(conn);
}
tipc_node_unlock(node);
+ tipc_node_put(node);
}
/**
@@ -227,8 +254,8 @@ void tipc_node_link_up(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
active[0] = active[1] = l_ptr;
exit:
/* Leave room for changeover header when returning 'mtu' to users: */
- n_ptr->act_mtus[0] = active[0]->max_pkt - INT_H_SIZE;
- n_ptr->act_mtus[1] = active[1]->max_pkt - INT_H_SIZE;
+ n_ptr->act_mtus[0] = active[0]->mtu - INT_H_SIZE;
+ n_ptr->act_mtus[1] = active[1]->mtu - INT_H_SIZE;
}
/**
@@ -292,11 +319,10 @@ void tipc_node_link_down(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
/* Leave room for changeover header when returning 'mtu' to users: */
if (active[0]) {
- n_ptr->act_mtus[0] = active[0]->max_pkt - INT_H_SIZE;
- n_ptr->act_mtus[1] = active[1]->max_pkt - INT_H_SIZE;
+ n_ptr->act_mtus[0] = active[0]->mtu - INT_H_SIZE;
+ n_ptr->act_mtus[1] = active[1]->mtu - INT_H_SIZE;
return;
}
-
/* Loopback link went down? No fragmentation needed from now on. */
if (n_ptr->addr == tn->own_addr) {
n_ptr->act_mtus[0] = MAX_MSG_SIZE;
@@ -367,18 +393,17 @@ static void node_lost_contact(struct tipc_node *n_ptr)
n_ptr->bclink.recv_permitted = false;
}
- /* Abort link changeover */
+ /* Abort any ongoing link failover */
for (i = 0; i < MAX_BEARERS; i++) {
struct tipc_link *l_ptr = n_ptr->links[i];
if (!l_ptr)
continue;
- l_ptr->reset_checkpoint = l_ptr->next_in_no;
- l_ptr->exp_msg_count = 0;
+ l_ptr->flags &= ~LINK_FAILINGOVER;
+ l_ptr->failover_checkpt = 0;
+ l_ptr->failover_pkts = 0;
+ kfree_skb(l_ptr->failover_skb);
+ l_ptr->failover_skb = NULL;
tipc_link_reset_fragments(l_ptr);
-
- /* Link marked for deletion after failover? => do it now */
- if (l_ptr->flags & LINK_STOPPED)
- tipc_link_delete(l_ptr);
}
n_ptr->action_flags &= ~TIPC_WAIT_OWN_LINKS_DOWN;
@@ -417,19 +442,25 @@ int tipc_node_get_linkname(struct net *net, u32 bearer_id, u32 addr,
char *linkname, size_t len)
{
struct tipc_link *link;
+ int err = -EINVAL;
struct tipc_node *node = tipc_node_find(net, addr);
- if ((bearer_id >= MAX_BEARERS) || !node)
- return -EINVAL;
+ if (!node)
+ return err;
+
+ if (bearer_id >= MAX_BEARERS)
+ goto exit;
+
tipc_node_lock(node);
link = node->links[bearer_id];
if (link) {
strncpy(linkname, link->name, len);
- tipc_node_unlock(node);
- return 0;
+ err = 0;
}
+exit:
tipc_node_unlock(node);
- return -EINVAL;
+ tipc_node_put(node);
+ return err;
}
void tipc_node_unlock(struct tipc_node *node)
@@ -459,7 +490,7 @@ void tipc_node_unlock(struct tipc_node *node)
TIPC_NOTIFY_NODE_DOWN | TIPC_NOTIFY_NODE_UP |
TIPC_NOTIFY_LINK_DOWN | TIPC_NOTIFY_LINK_UP |
TIPC_WAKEUP_BCAST_USERS | TIPC_BCAST_MSG_EVT |
- TIPC_NAMED_MSG_EVT);
+ TIPC_NAMED_MSG_EVT | TIPC_BCAST_RESET);
spin_unlock_bh(&node->lock);
@@ -488,6 +519,9 @@ void tipc_node_unlock(struct tipc_node *node)
if (flags & TIPC_BCAST_MSG_EVT)
tipc_bclink_input(net);
+
+ if (flags & TIPC_BCAST_RESET)
+ tipc_link_reset_all(node);
}
/* Caller should hold node lock for the passed node */
@@ -542,17 +576,21 @@ int tipc_nl_node_dump(struct sk_buff *skb, struct netlink_callback *cb)
msg.seq = cb->nlh->nlmsg_seq;
rcu_read_lock();
-
- if (last_addr && !tipc_node_find(net, last_addr)) {
- rcu_read_unlock();
- /* We never set seq or call nl_dump_check_consistent() this
- * means that setting prev_seq here will cause the consistence
- * check to fail in the netlink callback handler. Resulting in
- * the NLMSG_DONE message having the NLM_F_DUMP_INTR flag set if
- * the node state changed while we released the lock.
- */
- cb->prev_seq = 1;
- return -EPIPE;
+ if (last_addr) {
+ node = tipc_node_find(net, last_addr);
+ if (!node) {
+ rcu_read_unlock();
+ /* We never set seq or call nl_dump_check_consistent()
+ * this means that setting prev_seq here will cause the
+ * consistence check to fail in the netlink callback
+ * handler. Resulting in the NLMSG_DONE message having
+ * the NLM_F_DUMP_INTR flag set if the node state
+ * changed while we released the lock.
+ */
+ cb->prev_seq = 1;
+ return -EPIPE;
+ }
+ tipc_node_put(node);
}
list_for_each_entry_rcu(node, &tn->node_list, list) {
diff --git a/net/tipc/node.h b/net/tipc/node.h
index e89ac04ec2c3..02d5c20dc551 100644
--- a/net/tipc/node.h
+++ b/net/tipc/node.h
@@ -64,7 +64,8 @@ enum {
TIPC_NOTIFY_LINK_UP = (1 << 6),
TIPC_NOTIFY_LINK_DOWN = (1 << 7),
TIPC_NAMED_MSG_EVT = (1 << 8),
- TIPC_BCAST_MSG_EVT = (1 << 9)
+ TIPC_BCAST_MSG_EVT = (1 << 9),
+ TIPC_BCAST_RESET = (1 << 10)
};
/**
@@ -93,6 +94,7 @@ struct tipc_node_bclink {
/**
* struct tipc_node - TIPC node structure
* @addr: network address of node
+ * @ref: reference counter to node object
* @lock: spinlock governing access to structure
* @net: the applicable net namespace
* @hash: links to adjacent nodes in unsorted hash chain
@@ -114,6 +116,7 @@ struct tipc_node_bclink {
*/
struct tipc_node {
u32 addr;
+ struct kref kref;
spinlock_t lock;
struct net *net;
struct hlist_node hash;
@@ -136,6 +139,7 @@ struct tipc_node {
};
struct tipc_node *tipc_node_find(struct net *net, u32 addr);
+void tipc_node_put(struct tipc_node *node);
struct tipc_node *tipc_node_create(struct net *net, u32 addr);
void tipc_node_stop(struct net *net);
void tipc_node_attach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr);
@@ -170,10 +174,12 @@ static inline uint tipc_node_get_mtu(struct net *net, u32 addr, u32 selector)
node = tipc_node_find(net, addr);
- if (likely(node))
+ if (likely(node)) {
mtu = node->act_mtus[selector & 1];
- else
+ tipc_node_put(node);
+ } else {
mtu = MAX_MSG_SIZE;
+ }
return mtu;
}
diff --git a/net/tipc/server.c b/net/tipc/server.c
index eadd4ed45905..ab6183cdb121 100644
--- a/net/tipc/server.c
+++ b/net/tipc/server.c
@@ -37,11 +37,13 @@
#include "core.h"
#include "socket.h"
#include <net/sock.h>
+#include <linux/module.h>
/* Number of messages to send before rescheduling */
#define MAX_SEND_MSG_COUNT 25
#define MAX_RECV_MSG_COUNT 25
#define CF_CONNECTED 1
+#define CF_SERVER 2
#define sock2con(x) ((struct tipc_conn *)(x)->sk_user_data)
@@ -88,9 +90,19 @@ static void tipc_clean_outqueues(struct tipc_conn *con);
static void tipc_conn_kref_release(struct kref *kref)
{
struct tipc_conn *con = container_of(kref, struct tipc_conn, kref);
+ struct sockaddr_tipc *saddr = con->server->saddr;
+ struct socket *sock = con->sock;
+ struct sock *sk;
- if (con->sock) {
- tipc_sock_release_local(con->sock);
+ if (sock) {
+ sk = sock->sk;
+ if (test_bit(CF_SERVER, &con->flags)) {
+ __module_get(sock->ops->owner);
+ __module_get(sk->sk_prot_creator->owner);
+ }
+ saddr->scope = -TIPC_NODE_SCOPE;
+ kernel_bind(sock, (struct sockaddr *)saddr, sizeof(*saddr));
+ sk_release_kernel(sk);
con->sock = NULL;
}
@@ -281,7 +293,7 @@ static int tipc_accept_from_sock(struct tipc_conn *con)
struct tipc_conn *newcon;
int ret;
- ret = tipc_sock_accept_local(sock, &newsock, O_NONBLOCK);
+ ret = kernel_accept(sock, &newsock, O_NONBLOCK);
if (ret < 0)
return ret;
@@ -309,9 +321,12 @@ static struct socket *tipc_create_listen_sock(struct tipc_conn *con)
struct socket *sock = NULL;
int ret;
- ret = tipc_sock_create_local(s->net, s->type, &sock);
+ ret = sock_create_kern(AF_TIPC, SOCK_SEQPACKET, 0, &sock);
if (ret < 0)
return NULL;
+
+ sk_change_net(sock->sk, s->net);
+
ret = kernel_setsockopt(sock, SOL_TIPC, TIPC_IMPORTANCE,
(char *)&s->imp, sizeof(s->imp));
if (ret < 0)
@@ -337,11 +352,31 @@ static struct socket *tipc_create_listen_sock(struct tipc_conn *con)
pr_err("Unknown socket type %d\n", s->type);
goto create_err;
}
+
+ /* As server's listening socket owner and creator is the same module,
+ * we have to decrease TIPC module reference count to guarantee that
+ * it remains zero after the server socket is created, otherwise,
+ * executing "rmmod" command is unable to make TIPC module deleted
+ * after TIPC module is inserted successfully.
+ *
+ * However, the reference count is ever increased twice in
+ * sock_create_kern(): one is to increase the reference count of owner
+ * of TIPC socket's proto_ops struct; another is to increment the
+ * reference count of owner of TIPC proto struct. Therefore, we must
+ * decrement the module reference count twice to ensure that it keeps
+ * zero after server's listening socket is created. Of course, we
+ * must bump the module reference count twice as well before the socket
+ * is closed.
+ */
+ module_put(sock->ops->owner);
+ module_put(sock->sk->sk_prot_creator->owner);
+ set_bit(CF_SERVER, &con->flags);
+
return sock;
create_err:
- sock_release(sock);
- con->sock = NULL;
+ kernel_sock_shutdown(sock, SHUT_RDWR);
+ sk_release_kernel(sock->sk);
return NULL;
}
diff --git a/net/tipc/socket.c b/net/tipc/socket.c
index 934947f038b6..ee90d74d7516 100644
--- a/net/tipc/socket.c
+++ b/net/tipc/socket.c
@@ -35,7 +35,6 @@
*/
#include <linux/rhashtable.h>
-#include <linux/jhash.h>
#include "core.h"
#include "name_table.h"
#include "node.h"
@@ -74,6 +73,7 @@
* @link_cong: non-zero if owner must sleep because of link congestion
* @sent_unacked: # messages sent by socket, and not yet acked by peer
* @rcv_unacked: # messages read by user, but not yet acked back to peer
+ * @remote: 'connected' peer for dgram/rdm
* @node: hash table node
* @rcu: rcu struct for tipc_sock
*/
@@ -96,6 +96,7 @@ struct tipc_sock {
bool link_cong;
uint sent_unacked;
uint rcv_unacked;
+ struct sockaddr_tipc remote;
struct rhash_head node;
struct rcu_head rcu;
};
@@ -121,9 +122,7 @@ static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dsz);
static const struct proto_ops packet_ops;
static const struct proto_ops stream_ops;
static const struct proto_ops msg_ops;
-
static struct proto tipc_proto;
-static struct proto tipc_proto_kern;
static const struct nla_policy tipc_nl_sock_policy[TIPC_NLA_SOCK_MAX + 1] = {
[TIPC_NLA_SOCK_UNSPEC] = { .type = NLA_UNSPEC },
@@ -133,6 +132,8 @@ static const struct nla_policy tipc_nl_sock_policy[TIPC_NLA_SOCK_MAX + 1] = {
[TIPC_NLA_SOCK_HAS_PUBL] = { .type = NLA_FLAG }
};
+static const struct rhashtable_params tsk_rht_params;
+
/*
* Revised TIPC socket locking policy:
*
@@ -341,11 +342,7 @@ static int tipc_sk_create(struct net *net, struct socket *sock,
}
/* Allocate socket's protocol area */
- if (!kern)
- sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto);
- else
- sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto_kern);
-
+ sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto);
if (sk == NULL)
return -ENOMEM;
@@ -383,75 +380,6 @@ static int tipc_sk_create(struct net *net, struct socket *sock,
return 0;
}
-/**
- * tipc_sock_create_local - create TIPC socket from inside TIPC module
- * @type: socket type - SOCK_RDM or SOCK_SEQPACKET
- *
- * We cannot use sock_creat_kern here because it bumps module user count.
- * Since socket owner and creator is the same module we must make sure
- * that module count remains zero for module local sockets, otherwise
- * we cannot do rmmod.
- *
- * Returns 0 on success, errno otherwise
- */
-int tipc_sock_create_local(struct net *net, int type, struct socket **res)
-{
- int rc;
-
- rc = sock_create_lite(AF_TIPC, type, 0, res);
- if (rc < 0) {
- pr_err("Failed to create kernel socket\n");
- return rc;
- }
- tipc_sk_create(net, *res, 0, 1);
-
- return 0;
-}
-
-/**
- * tipc_sock_release_local - release socket created by tipc_sock_create_local
- * @sock: the socket to be released.
- *
- * Module reference count is not incremented when such sockets are created,
- * so we must keep it from being decremented when they are released.
- */
-void tipc_sock_release_local(struct socket *sock)
-{
- tipc_release(sock);
- sock->ops = NULL;
- sock_release(sock);
-}
-
-/**
- * tipc_sock_accept_local - accept a connection on a socket created
- * with tipc_sock_create_local. Use this function to avoid that
- * module reference count is inadvertently incremented.
- *
- * @sock: the accepting socket
- * @newsock: reference to the new socket to be created
- * @flags: socket flags
- */
-
-int tipc_sock_accept_local(struct socket *sock, struct socket **newsock,
- int flags)
-{
- struct sock *sk = sock->sk;
- int ret;
-
- ret = sock_create_lite(sk->sk_family, sk->sk_type,
- sk->sk_protocol, newsock);
- if (ret < 0)
- return ret;
-
- ret = tipc_accept(sock, *newsock, flags);
- if (ret < 0) {
- sock_release(*newsock);
- return ret;
- }
- (*newsock)->ops = sock->ops;
- return ret;
-}
-
static void tipc_sk_callback(struct rcu_head *head)
{
struct tipc_sock *tsk = container_of(head, struct tipc_sock, rcu);
@@ -929,22 +857,23 @@ static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dsz)
u32 dnode, dport;
struct sk_buff_head *pktchain = &sk->sk_write_queue;
struct sk_buff *skb;
- struct tipc_name_seq *seq = &dest->addr.nameseq;
+ struct tipc_name_seq *seq;
struct iov_iter save;
u32 mtu;
long timeo;
int rc;
- if (unlikely(!dest))
- return -EDESTADDRREQ;
-
- if (unlikely((m->msg_namelen < sizeof(*dest)) ||
- (dest->family != AF_TIPC)))
- return -EINVAL;
-
if (dsz > TIPC_MAX_USER_MSG_SIZE)
return -EMSGSIZE;
-
+ if (unlikely(!dest)) {
+ if (tsk->connected && sock->state == SS_READY)
+ dest = &tsk->remote;
+ else
+ return -EDESTADDRREQ;
+ } else if (unlikely(m->msg_namelen < sizeof(*dest)) ||
+ dest->family != AF_TIPC) {
+ return -EINVAL;
+ }
if (unlikely(sock->state != SS_READY)) {
if (sock->state == SS_LISTENING)
return -EPIPE;
@@ -957,7 +886,7 @@ static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dsz)
tsk->conn_instance = dest->addr.name.name.instance;
}
}
-
+ seq = &dest->addr.nameseq;
timeo = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
if (dest->addrtype == TIPC_ADDR_MCAST) {
@@ -1908,17 +1837,26 @@ static int tipc_connect(struct socket *sock, struct sockaddr *dest,
int destlen, int flags)
{
struct sock *sk = sock->sk;
+ struct tipc_sock *tsk = tipc_sk(sk);
struct sockaddr_tipc *dst = (struct sockaddr_tipc *)dest;
struct msghdr m = {NULL,};
- long timeout = (flags & O_NONBLOCK) ? 0 : tipc_sk(sk)->conn_timeout;
+ long timeout = (flags & O_NONBLOCK) ? 0 : tsk->conn_timeout;
socket_state previous;
- int res;
+ int res = 0;
lock_sock(sk);
- /* For now, TIPC does not allow use of connect() with DGRAM/RDM types */
+ /* DGRAM/RDM connect(), just save the destaddr */
if (sock->state == SS_READY) {
- res = -EOPNOTSUPP;
+ if (dst->family == AF_UNSPEC) {
+ memset(&tsk->remote, 0, sizeof(struct sockaddr_tipc));
+ tsk->connected = 0;
+ } else if (destlen != sizeof(struct sockaddr_tipc)) {
+ res = -EINVAL;
+ } else {
+ memcpy(&tsk->remote, dest, destlen);
+ tsk->connected = 1;
+ }
goto exit;
}
@@ -2153,7 +2091,6 @@ restart:
TIPC_CONN_SHUTDOWN))
tipc_link_xmit_skb(net, skb, dnode,
tsk->portid);
- tipc_node_remove_conn(net, dnode, tsk->portid);
} else {
dnode = tsk_peer_node(tsk);
@@ -2311,7 +2248,7 @@ static struct tipc_sock *tipc_sk_lookup(struct net *net, u32 portid)
struct tipc_sock *tsk;
rcu_read_lock();
- tsk = rhashtable_lookup(&tn->sk_rht, &portid);
+ tsk = rhashtable_lookup_fast(&tn->sk_rht, &portid, tsk_rht_params);
if (tsk)
sock_hold(&tsk->sk);
rcu_read_unlock();
@@ -2333,7 +2270,8 @@ static int tipc_sk_insert(struct tipc_sock *tsk)
portid = TIPC_MIN_PORT;
tsk->portid = portid;
sock_hold(&tsk->sk);
- if (rhashtable_lookup_insert(&tn->sk_rht, &tsk->node))
+ if (!rhashtable_lookup_insert_fast(&tn->sk_rht, &tsk->node,
+ tsk_rht_params))
return 0;
sock_put(&tsk->sk);
}
@@ -2346,26 +2284,27 @@ static void tipc_sk_remove(struct tipc_sock *tsk)
struct sock *sk = &tsk->sk;
struct tipc_net *tn = net_generic(sock_net(sk), tipc_net_id);
- if (rhashtable_remove(&tn->sk_rht, &tsk->node)) {
+ if (!rhashtable_remove_fast(&tn->sk_rht, &tsk->node, tsk_rht_params)) {
WARN_ON(atomic_read(&sk->sk_refcnt) == 1);
__sock_put(sk);
}
}
+static const struct rhashtable_params tsk_rht_params = {
+ .nelem_hint = 192,
+ .head_offset = offsetof(struct tipc_sock, node),
+ .key_offset = offsetof(struct tipc_sock, portid),
+ .key_len = sizeof(u32), /* portid */
+ .max_size = 1048576,
+ .min_size = 256,
+ .automatic_shrinking = true,
+};
+
int tipc_sk_rht_init(struct net *net)
{
struct tipc_net *tn = net_generic(net, tipc_net_id);
- struct rhashtable_params rht_params = {
- .nelem_hint = 192,
- .head_offset = offsetof(struct tipc_sock, node),
- .key_offset = offsetof(struct tipc_sock, portid),
- .key_len = sizeof(u32), /* portid */
- .hashfn = jhash,
- .max_shift = 20, /* 1M */
- .min_shift = 8, /* 256 */
- };
- return rhashtable_init(&tn->sk_rht, &rht_params);
+ return rhashtable_init(&tn->sk_rht, &tsk_rht_params);
}
void tipc_sk_rht_destroy(struct net *net)
@@ -2608,12 +2547,6 @@ static struct proto tipc_proto = {
.sysctl_rmem = sysctl_tipc_rmem
};
-static struct proto tipc_proto_kern = {
- .name = "TIPC",
- .obj_size = sizeof(struct tipc_sock),
- .sysctl_rmem = sysctl_tipc_rmem
-};
-
/**
* tipc_socket_init - initialize TIPC socket interface
*
diff --git a/net/tipc/socket.h b/net/tipc/socket.h
index 238f1b7bd9bd..bf6551389522 100644
--- a/net/tipc/socket.h
+++ b/net/tipc/socket.h
@@ -44,10 +44,6 @@
SKB_TRUESIZE(TIPC_MAX_USER_MSG_SIZE))
int tipc_socket_init(void);
void tipc_socket_stop(void);
-int tipc_sock_create_local(struct net *net, int type, struct socket **res);
-void tipc_sock_release_local(struct socket *sock);
-int tipc_sock_accept_local(struct socket *sock, struct socket **newsock,
- int flags);
int tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq);
void tipc_sk_mcast_rcv(struct net *net, struct sk_buff_head *arrvq,
struct sk_buff_head *inputq);
diff --git a/net/tipc/udp_media.c b/net/tipc/udp_media.c
index fc2fb11a354d..ef3d7aa2854a 100644
--- a/net/tipc/udp_media.c
+++ b/net/tipc/udp_media.c
@@ -44,6 +44,7 @@
#include <net/sock.h>
#include <net/ip.h>
#include <net/udp_tunnel.h>
+#include <net/addrconf.h>
#include <linux/tipc_netlink.h>
#include "core.h"
#include "bearer.h"
@@ -192,7 +193,7 @@ static int tipc_udp_send_msg(struct net *net, struct sk_buff *skb,
.saddr = src->ipv6,
.flowi6_proto = IPPROTO_UDP
};
- err = ip6_dst_lookup(ub->ubsock->sk, &ndst, &fl6);
+ err = ipv6_stub->ipv6_dst_lookup(ub->ubsock->sk, &ndst, &fl6);
if (err)
goto tx_error;
ttl = ip6_dst_hoplimit(ndst);
@@ -246,11 +247,14 @@ static int enable_mcast(struct udp_bearer *ub, struct udp_media_addr *remote)
return 0;
mreqn.imr_multiaddr = remote->ipv4;
mreqn.imr_ifindex = ub->ifindex;
- err = __ip_mc_join_group(sk, &mreqn);
+ err = ip_mc_join_group(sk, &mreqn);
+#if IS_ENABLED(CONFIG_IPV6)
} else {
if (!ipv6_addr_is_multicast(&remote->ipv6))
return 0;
- err = __ipv6_sock_mc_join(sk, ub->ifindex, &remote->ipv6);
+ err = ipv6_stub->ipv6_sock_mc_join(sk, ub->ifindex,
+ &remote->ipv6);
+#endif
}
return err;
}
Powered by cgit, Gruvboxed by Blaster4385.