diff options
author | José Pekkarinen <jose.pekkarinen@nokia.com> | 2016-04-11 10:41:07 +0300 |
---|---|---|
committer | José Pekkarinen <jose.pekkarinen@nokia.com> | 2016-04-13 08:17:18 +0300 |
commit | e09b41010ba33a20a87472ee821fa407a5b8da36 (patch) | |
tree | d10dc367189862e7ca5c592f033dc3726e1df4e3 /kernel/net/tipc | |
parent | f93b97fd65072de626c074dbe099a1fff05ce060 (diff) |
These changes are the raw update to linux-4.4.6-rt14. Kernel sources
are taken from kernel.org, and rt patch from the rt wiki download page.
During the rebasing, the following patch collided:
Force tick interrupt and get rid of softirq magic(I70131fb85).
Collisions have been removed because its logic was found on the
source already.
Change-Id: I7f57a4081d9deaa0d9ccfc41a6c8daccdee3b769
Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>
Diffstat (limited to 'kernel/net/tipc')
-rw-r--r-- | kernel/net/tipc/addr.c | 7 | ||||
-rw-r--r-- | kernel/net/tipc/addr.h | 8 | ||||
-rw-r--r-- | kernel/net/tipc/bcast.c | 963 | ||||
-rw-r--r-- | kernel/net/tipc/bcast.h | 120 | ||||
-rw-r--r-- | kernel/net/tipc/bearer.c | 142 | ||||
-rw-r--r-- | kernel/net/tipc/bearer.h | 14 | ||||
-rw-r--r-- | kernel/net/tipc/core.c | 13 | ||||
-rw-r--r-- | kernel/net/tipc/core.h | 55 | ||||
-rw-r--r-- | kernel/net/tipc/discover.c | 150 | ||||
-rw-r--r-- | kernel/net/tipc/link.c | 2575 | ||||
-rw-r--r-- | kernel/net/tipc/link.h | 234 | ||||
-rw-r--r-- | kernel/net/tipc/msg.c | 201 | ||||
-rw-r--r-- | kernel/net/tipc/msg.h | 133 | ||||
-rw-r--r-- | kernel/net/tipc/name_distr.c | 6 | ||||
-rw-r--r-- | kernel/net/tipc/name_table.c | 34 | ||||
-rw-r--r-- | kernel/net/tipc/net.c | 7 | ||||
-rw-r--r-- | kernel/net/tipc/netlink_compat.c | 139 | ||||
-rw-r--r-- | kernel/net/tipc/node.c | 1131 | ||||
-rw-r--r-- | kernel/net/tipc/node.h | 127 | ||||
-rw-r--r-- | kernel/net/tipc/server.c | 6 | ||||
-rw-r--r-- | kernel/net/tipc/socket.c | 409 | ||||
-rw-r--r-- | kernel/net/tipc/socket.h | 2 | ||||
-rw-r--r-- | kernel/net/tipc/subscr.c | 251 | ||||
-rw-r--r-- | kernel/net/tipc/subscr.h | 18 | ||||
-rw-r--r-- | kernel/net/tipc/udp_media.c | 29 |
25 files changed, 3491 insertions, 3283 deletions
diff --git a/kernel/net/tipc/addr.c b/kernel/net/tipc/addr.c index ba7daa864..48fd3b5a7 100644 --- a/kernel/net/tipc/addr.c +++ b/kernel/net/tipc/addr.c @@ -38,13 +38,6 @@ #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/kernel/net/tipc/addr.h b/kernel/net/tipc/addr.h index 7ba6d5c8a..93f7c983b 100644 --- a/kernel/net/tipc/addr.h +++ b/kernel/net/tipc/addr.h @@ -41,10 +41,18 @@ #include <linux/tipc.h> #include <net/net_namespace.h> #include <net/netns/generic.h> +#include "core.h" #define TIPC_ZONE_MASK 0xff000000u #define TIPC_CLUSTER_MASK 0xfffff000u +static inline u32 tipc_own_addr(struct net *net) +{ + struct tipc_net *tn = net_generic(net, tipc_net_id); + + return tn->own_addr; +} + static inline u32 tipc_zone_mask(u32 addr) { return addr & TIPC_ZONE_MASK; diff --git a/kernel/net/tipc/bcast.c b/kernel/net/tipc/bcast.c index c5cbdcb1f..92e367a0a 100644 --- a/kernel/net/tipc/bcast.c +++ b/kernel/net/tipc/bcast.c @@ -35,690 +35,301 @@ * POSSIBILITY OF SUCH DAMAGE. */ +#include <linux/tipc_config.h> #include "socket.h" #include "msg.h" #include "bcast.h" #include "name_distr.h" -#include "core.h" +#include "link.h" +#include "node.h" -#define MAX_PKT_DEFAULT_MCAST 1500 /* bcast link max packet size (fixed) */ -#define BCLINK_WIN_DEFAULT 20 /* bcast link window size (default) */ +#define BCLINK_WIN_DEFAULT 50 /* bcast link window size (default) */ +#define BCLINK_WIN_MIN 32 /* bcast minimum link window size */ const char tipc_bclink_name[] = "broadcast-link"; -static void tipc_nmap_diff(struct tipc_node_map *nm_a, - struct tipc_node_map *nm_b, - struct tipc_node_map *nm_diff); -static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node); -static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node); - -static void tipc_bclink_lock(struct net *net) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); - - spin_lock_bh(&tn->bclink->lock); -} - -static void tipc_bclink_unlock(struct net *net) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); - - spin_unlock_bh(&tn->bclink->lock); -} - -void tipc_bclink_input(struct net *net) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); - - tipc_sk_mcast_rcv(net, &tn->bclink->arrvq, &tn->bclink->inputq); -} - -uint tipc_bclink_get_mtu(void) -{ - return MAX_PKT_DEFAULT_MCAST; -} - -static u32 bcbuf_acks(struct sk_buff *buf) -{ - return (u32)(unsigned long)TIPC_SKB_CB(buf)->handle; -} - -static void bcbuf_set_acks(struct sk_buff *buf, u32 acks) -{ - TIPC_SKB_CB(buf)->handle = (void *)(unsigned long)acks; -} - -static void bcbuf_decr_acks(struct sk_buff *buf) -{ - bcbuf_set_acks(buf, bcbuf_acks(buf) - 1); -} +/** + * struct tipc_bc_base - base structure for keeping broadcast send state + * @link: broadcast send link structure + * @inputq: data input queue; will only carry SOCK_WAKEUP messages + * @dest: array keeping number of reachable destinations per bearer + * @primary_bearer: a bearer having links to all broadcast destinations, if any + */ +struct tipc_bc_base { + struct tipc_link *link; + struct sk_buff_head inputq; + int dests[MAX_BEARERS]; + int primary_bearer; +}; -void tipc_bclink_add_node(struct net *net, u32 addr) +static struct tipc_bc_base *tipc_bc_base(struct net *net) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - - tipc_bclink_lock(net); - tipc_nmap_add(&tn->bclink->bcast_nodes, addr); - tipc_bclink_unlock(net); + return tipc_net(net)->bcbase; } -void tipc_bclink_remove_node(struct net *net, u32 addr) +int tipc_bcast_get_mtu(struct net *net) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - - tipc_bclink_lock(net); - tipc_nmap_remove(&tn->bclink->bcast_nodes, addr); - tipc_bclink_unlock(net); + return tipc_link_mtu(tipc_bc_sndlink(net)); } -static void bclink_set_last_sent(struct net *net) +/* tipc_bcbase_select_primary(): find a bearer with links to all destinations, + * if any, and make it primary bearer + */ +static void tipc_bcbase_select_primary(struct net *net) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_link *bcl = tn->bcl; - struct sk_buff *skb = skb_peek(&bcl->backlogq); + struct tipc_bc_base *bb = tipc_bc_base(net); + int all_dests = tipc_link_bc_peers(bb->link); + int i, mtu; - if (skb) - bcl->fsm_msg_cnt = mod(buf_seqno(skb) - 1); - else - bcl->fsm_msg_cnt = mod(bcl->next_out_no - 1); -} - -u32 tipc_bclink_get_last_sent(struct net *net) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); + bb->primary_bearer = INVALID_BEARER_ID; - return tn->bcl->fsm_msg_cnt; -} + if (!all_dests) + return; -static void bclink_update_last_sent(struct tipc_node *node, u32 seqno) -{ - node->bclink.last_sent = less_eq(node->bclink.last_sent, seqno) ? - seqno : node->bclink.last_sent; -} + for (i = 0; i < MAX_BEARERS; i++) { + if (!bb->dests[i]) + continue; -/** - * tipc_bclink_retransmit_to - get most recent node to request retransmission - * - * Called with bclink_lock locked - */ -struct tipc_node *tipc_bclink_retransmit_to(struct net *net) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); + mtu = tipc_bearer_mtu(net, i); + if (mtu < tipc_link_mtu(bb->link)) + tipc_link_set_mtu(bb->link, mtu); - return tn->bclink->retransmit_to; -} + if (bb->dests[i] < all_dests) + continue; -/** - * bclink_retransmit_pkt - retransmit broadcast packets - * @after: sequence number of last packet to *not* retransmit - * @to: sequence number of last packet to retransmit - * - * Called with bclink_lock locked - */ -static void bclink_retransmit_pkt(struct tipc_net *tn, u32 after, u32 to) -{ - struct sk_buff *skb; - struct tipc_link *bcl = tn->bcl; + bb->primary_bearer = i; - skb_queue_walk(&bcl->transmq, skb) { - if (more(buf_seqno(skb), after)) { - tipc_link_retransmit(bcl, skb, mod(to - after)); + /* Reduce risk that all nodes select same primary */ + if ((i ^ tipc_own_addr(net)) & 1) break; - } } } -/** - * tipc_bclink_wakeup_users - wake up pending users - * - * Called with no locks taken - */ -void tipc_bclink_wakeup_users(struct net *net) +void tipc_bcast_inc_bearer_dst_cnt(struct net *net, int bearer_id) { - struct tipc_net *tn = net_generic(net, tipc_net_id); + struct tipc_bc_base *bb = tipc_bc_base(net); - tipc_sk_rcv(net, &tn->bclink->link.wakeupq); + tipc_bcast_lock(net); + bb->dests[bearer_id]++; + tipc_bcbase_select_primary(net); + tipc_bcast_unlock(net); } -/** - * tipc_bclink_acknowledge - handle acknowledgement of broadcast packets - * @n_ptr: node that sent acknowledgement info - * @acked: broadcast sequence # that has been acknowledged - * - * Node is locked, bclink_lock unlocked. - */ -void tipc_bclink_acknowledge(struct tipc_node *n_ptr, u32 acked) +void tipc_bcast_dec_bearer_dst_cnt(struct net *net, int bearer_id) { - struct sk_buff *skb, *tmp; - unsigned int released = 0; - struct net *net = n_ptr->net; - struct tipc_net *tn = net_generic(net, tipc_net_id); - - if (unlikely(!n_ptr->bclink.recv_permitted)) - return; - - tipc_bclink_lock(net); - - /* Bail out if tx queue is empty (no clean up is required) */ - skb = skb_peek(&tn->bcl->transmq); - if (!skb) - goto exit; - - /* Determine which messages need to be acknowledged */ - if (acked == INVALID_LINK_SEQ) { - /* - * Contact with specified node has been lost, so need to - * acknowledge sent messages only (if other nodes still exist) - * or both sent and unsent messages (otherwise) - */ - if (tn->bclink->bcast_nodes.count) - acked = tn->bcl->fsm_msg_cnt; - else - acked = tn->bcl->next_out_no; - } else { - /* - * Bail out if specified sequence number does not correspond - * to a message that has been sent and not yet acknowledged - */ - if (less(acked, buf_seqno(skb)) || - less(tn->bcl->fsm_msg_cnt, acked) || - less_eq(acked, n_ptr->bclink.acked)) - goto exit; - } - - /* Skip over packets that node has previously acknowledged */ - skb_queue_walk(&tn->bcl->transmq, skb) { - if (more(buf_seqno(skb), n_ptr->bclink.acked)) - break; - } + struct tipc_bc_base *bb = tipc_bc_base(net); - /* Update packets that node is now acknowledging */ - skb_queue_walk_from_safe(&tn->bcl->transmq, skb, tmp) { - if (more(buf_seqno(skb), acked)) - break; - bcbuf_decr_acks(skb); - bclink_set_last_sent(net); - if (bcbuf_acks(skb) == 0) { - __skb_unlink(skb, &tn->bcl->transmq); - kfree_skb(skb); - released = 1; - } - } - n_ptr->bclink.acked = acked; - - /* Try resolving broadcast link congestion, if necessary */ - if (unlikely(skb_peek(&tn->bcl->backlogq))) { - tipc_link_push_packets(tn->bcl); - bclink_set_last_sent(net); - } - if (unlikely(released && !skb_queue_empty(&tn->bcl->wakeupq))) - n_ptr->action_flags |= TIPC_WAKEUP_BCAST_USERS; -exit: - tipc_bclink_unlock(net); + tipc_bcast_lock(net); + bb->dests[bearer_id]--; + tipc_bcbase_select_primary(net); + tipc_bcast_unlock(net); } -/** - * tipc_bclink_update_link_state - update broadcast link state +/* tipc_bcbase_xmit - broadcast a packet queue across one or more bearers * - * RCU and node lock set + * Note that number of reachable destinations, as indicated in the dests[] + * array, may transitionally differ from the number of destinations indicated + * in each sent buffer. We can sustain this. Excess destination nodes will + * drop and never acknowledge the unexpected packets, and missing destinations + * will either require retransmission (if they are just about to be added to + * the bearer), or be removed from the buffer's 'ackers' counter (if they + * just went down) */ -void tipc_bclink_update_link_state(struct tipc_node *n_ptr, - u32 last_sent) +static void tipc_bcbase_xmit(struct net *net, struct sk_buff_head *xmitq) { - struct sk_buff *buf; - struct net *net = n_ptr->net; - struct tipc_net *tn = net_generic(net, tipc_net_id); + int bearer_id; + struct tipc_bc_base *bb = tipc_bc_base(net); + struct sk_buff *skb, *_skb; + struct sk_buff_head _xmitq; - /* Ignore "stale" link state info */ - if (less_eq(last_sent, n_ptr->bclink.last_in)) + if (skb_queue_empty(xmitq)) return; - /* Update link synchronization state; quit if in sync */ - bclink_update_last_sent(n_ptr, last_sent); - - if (n_ptr->bclink.last_sent == n_ptr->bclink.last_in) + /* The typical case: at least one bearer has links to all nodes */ + bearer_id = bb->primary_bearer; + if (bearer_id >= 0) { + tipc_bearer_bc_xmit(net, bearer_id, xmitq); return; - - /* Update out-of-sync state; quit if loss is still unconfirmed */ - if ((++n_ptr->bclink.oos_state) == 1) { - if (n_ptr->bclink.deferred_size < (TIPC_MIN_LINK_WIN / 2)) - return; - n_ptr->bclink.oos_state++; } - /* Don't NACK if one has been recently sent (or seen) */ - if (n_ptr->bclink.oos_state & 0x1) - return; + /* We have to transmit across all bearers */ + skb_queue_head_init(&_xmitq); + for (bearer_id = 0; bearer_id < MAX_BEARERS; bearer_id++) { + if (!bb->dests[bearer_id]) + continue; - /* Send NACK */ - buf = tipc_buf_acquire(INT_H_SIZE); - if (buf) { - struct tipc_msg *msg = buf_msg(buf); - struct sk_buff *skb = skb_peek(&n_ptr->bclink.deferdq); - u32 to = skb ? buf_seqno(skb) - 1 : n_ptr->bclink.last_sent; - - tipc_msg_init(tn->own_addr, msg, BCAST_PROTOCOL, STATE_MSG, - INT_H_SIZE, n_ptr->addr); - msg_set_non_seq(msg, 1); - msg_set_mc_netid(msg, tn->net_id); - msg_set_bcast_ack(msg, n_ptr->bclink.last_in); - msg_set_bcgap_after(msg, n_ptr->bclink.last_in); - msg_set_bcgap_to(msg, to); - - tipc_bclink_lock(net); - tipc_bearer_send(net, MAX_BEARERS, buf, NULL); - tn->bcl->stats.sent_nacks++; - tipc_bclink_unlock(net); - kfree_skb(buf); - - n_ptr->bclink.oos_state++; + skb_queue_walk(xmitq, skb) { + _skb = pskb_copy_for_clone(skb, GFP_ATOMIC); + if (!_skb) + break; + __skb_queue_tail(&_xmitq, _skb); + } + tipc_bearer_bc_xmit(net, bearer_id, &_xmitq); } + __skb_queue_purge(xmitq); + __skb_queue_purge(&_xmitq); } -/** - * bclink_peek_nack - monitor retransmission requests sent by other nodes - * - * Delay any upcoming NACK by this node if another node has already - * requested the first message this node is going to ask for. - */ -static void bclink_peek_nack(struct net *net, struct tipc_msg *msg) -{ - struct tipc_node *n_ptr = tipc_node_find(net, msg_destnode(msg)); - - if (unlikely(!n_ptr)) - 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 +/* tipc_bcast_xmit - deliver buffer chain to all nodes in cluster * and to identified node local sockets * @net: the applicable net namespace * @list: chain of buffers containing message * Consumes the buffer chain, except when returning -ELINKCONG * Returns 0 if success, otherwise errno: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE */ -int tipc_bclink_xmit(struct net *net, struct sk_buff_head *list) +int tipc_bcast_xmit(struct net *net, struct sk_buff_head *list) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_link *bcl = tn->bcl; - struct tipc_bclink *bclink = tn->bclink; + struct tipc_link *l = tipc_bc_sndlink(net); + struct sk_buff_head xmitq, inputq, rcvq; int rc = 0; - int bc = 0; - struct sk_buff *skb; - struct sk_buff_head arrvq; - struct sk_buff_head inputq; - /* Prepare clone of message for local node */ - skb = tipc_msg_reassemble(list); - if (unlikely(!skb)) { - __skb_queue_purge(list); + __skb_queue_head_init(&rcvq); + __skb_queue_head_init(&xmitq); + skb_queue_head_init(&inputq); + + /* Prepare message clone for local node */ + if (unlikely(!tipc_msg_reassemble(list, &rcvq))) return -EHOSTUNREACH; - } - /* Broadcast to all nodes */ - if (likely(bclink)) { - tipc_bclink_lock(net); - if (likely(bclink->bcast_nodes.count)) { - rc = __tipc_link_xmit(net, bcl, list); - if (likely(!rc)) { - u32 len = skb_queue_len(&bcl->transmq); - - bclink_set_last_sent(net); - bcl->stats.queue_sz_counts++; - bcl->stats.accu_queue_sz += len; - } - bc = 1; - } - tipc_bclink_unlock(net); - } - if (unlikely(!bc)) - __skb_queue_purge(list); + tipc_bcast_lock(net); + if (tipc_link_bc_peers(l)) + rc = tipc_link_xmit(l, list, &xmitq); + tipc_bcast_unlock(net); + /* Don't send to local node if adding to link failed */ if (unlikely(rc)) { - kfree_skb(skb); + __skb_queue_purge(&rcvq); return rc; } - /* Deliver message clone */ - __skb_queue_head_init(&arrvq); - skb_queue_head_init(&inputq); - __skb_queue_tail(&arrvq, skb); - tipc_sk_mcast_rcv(net, &arrvq, &inputq); - return rc; -} -/** - * bclink_accept_pkt - accept an incoming, in-sequence broadcast packet - * - * Called with both sending node's lock and bclink_lock taken. - */ -static void bclink_accept_pkt(struct tipc_node *node, u32 seqno) -{ - struct tipc_net *tn = net_generic(node->net, tipc_net_id); - - bclink_update_last_sent(node, seqno); - node->bclink.last_in = seqno; - node->bclink.oos_state = 0; - tn->bcl->stats.recv_info++; - - /* - * Unicast an ACK periodically, ensuring that - * all nodes in the cluster don't ACK at the same time - */ - 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); - tn->bcl->stats.sent_acks++; - } + /* Broadcast to all nodes, inluding local node */ + tipc_bcbase_xmit(net, &xmitq); + tipc_sk_mcast_rcv(net, &rcvq, &inputq); + __skb_queue_purge(list); + return 0; } -/** - * tipc_bclink_rcv - receive a broadcast packet, and deliver upwards +/* tipc_bcast_rcv - receive a broadcast packet, and deliver to rcv link * * RCU is locked, no other locks set */ -void tipc_bclink_rcv(struct net *net, struct sk_buff *buf) +int tipc_bcast_rcv(struct net *net, struct tipc_link *l, struct sk_buff *skb) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_link *bcl = tn->bcl; - struct tipc_msg *msg = buf_msg(buf); - struct tipc_node *node; - u32 next_in; - u32 seqno; - int deferred = 0; - int pos = 0; - struct sk_buff *iskb; - struct sk_buff_head *arrvq, *inputq; - - /* Screen out unwanted broadcast messages */ - if (msg_mc_netid(msg) != tn->net_id) - goto exit; - - node = tipc_node_find(net, msg_prevnode(msg)); - if (unlikely(!node)) - goto exit; - - tipc_node_lock(node); - if (unlikely(!node->bclink.recv_permitted)) - goto unlock; - - /* Handle broadcast protocol message */ - if (unlikely(msg_user(msg) == BCAST_PROTOCOL)) { - if (msg_type(msg) != STATE_MSG) - goto unlock; - if (msg_destnode(msg) == tn->own_addr) { - tipc_bclink_acknowledge(node, msg_bcast_ack(msg)); - 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; - } - - /* Handle in-sequence broadcast message */ - seqno = msg_seqno(msg); - next_in = mod(node->bclink.last_in + 1); - arrvq = &tn->bclink->arrvq; - inputq = &tn->bclink->inputq; - - if (likely(seqno == next_in)) { -receive: - /* Deliver message to destination */ - if (likely(msg_isdata(msg))) { - tipc_bclink_lock(net); - bclink_accept_pkt(node, seqno); - spin_lock_bh(&inputq->lock); - __skb_queue_tail(arrvq, buf); - spin_unlock_bh(&inputq->lock); - node->action_flags |= TIPC_BCAST_MSG_EVT; - tipc_bclink_unlock(net); - tipc_node_unlock(node); - } else if (msg_user(msg) == MSG_BUNDLER) { - tipc_bclink_lock(net); - bclink_accept_pkt(node, seqno); - bcl->stats.recv_bundles++; - bcl->stats.recv_bundled += msg_msgcnt(msg); - pos = 0; - while (tipc_msg_extract(buf, &iskb, &pos)) { - spin_lock_bh(&inputq->lock); - __skb_queue_tail(arrvq, iskb); - spin_unlock_bh(&inputq->lock); - } - node->action_flags |= TIPC_BCAST_MSG_EVT; - tipc_bclink_unlock(net); - tipc_node_unlock(node); - } else if (msg_user(msg) == MSG_FRAGMENTER) { - 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++; - msg = buf_msg(buf); - tipc_bclink_unlock(net); - goto receive; - } - tipc_bclink_unlock(net); - tipc_node_unlock(node); - } else { - tipc_bclink_lock(net); - bclink_accept_pkt(node, seqno); - tipc_bclink_unlock(net); - tipc_node_unlock(node); - kfree_skb(buf); - } - buf = NULL; - - /* Determine new synchronization state */ - tipc_node_lock(node); - if (unlikely(!tipc_node_is_up(node))) - goto unlock; - - if (node->bclink.last_in == node->bclink.last_sent) - goto unlock; + struct tipc_msg *hdr = buf_msg(skb); + struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq; + struct sk_buff_head xmitq; + int rc; - if (skb_queue_empty(&node->bclink.deferdq)) { - node->bclink.oos_state = 1; - goto unlock; - } - - msg = buf_msg(skb_peek(&node->bclink.deferdq)); - seqno = msg_seqno(msg); - next_in = mod(next_in + 1); - if (seqno != next_in) - goto unlock; - - /* Take in-sequence message from deferred queue & deliver it */ - buf = __skb_dequeue(&node->bclink.deferdq); - goto receive; - } + __skb_queue_head_init(&xmitq); - /* Handle out-of-sequence broadcast message */ - if (less(next_in, seqno)) { - deferred = tipc_link_defer_pkt(&node->bclink.deferdq, - buf); - bclink_update_last_sent(node, seqno); - buf = NULL; + if (msg_mc_netid(hdr) != tipc_netid(net) || !tipc_link_is_up(l)) { + kfree_skb(skb); + return 0; } - tipc_bclink_lock(net); - - if (deferred) - bcl->stats.deferred_recv++; + tipc_bcast_lock(net); + if (msg_user(hdr) == BCAST_PROTOCOL) + rc = tipc_link_bc_nack_rcv(l, skb, &xmitq); else - bcl->stats.duplicates++; + rc = tipc_link_rcv(l, skb, NULL); + tipc_bcast_unlock(net); - tipc_bclink_unlock(net); + tipc_bcbase_xmit(net, &xmitq); -unlock: - tipc_node_unlock(node); - tipc_node_put(node); -exit: - kfree_skb(buf); -} + /* Any socket wakeup messages ? */ + if (!skb_queue_empty(inputq)) + tipc_sk_rcv(net, inputq); -u32 tipc_bclink_acks_missing(struct tipc_node *n_ptr) -{ - return (n_ptr->bclink.recv_permitted && - (tipc_bclink_get_last_sent(n_ptr->net) != n_ptr->bclink.acked)); + return rc; } - -/** - * tipc_bcbearer_send - send a packet through the broadcast pseudo-bearer - * - * Send packet over as many bearers as necessary to reach all nodes - * that have joined the broadcast link. +/* tipc_bcast_ack_rcv - receive and handle a broadcast acknowledge * - * Returns 0 (packet sent successfully) under all circumstances, - * since the broadcast link's pseudo-bearer never blocks + * RCU is locked, no other locks set */ -static int tipc_bcbearer_send(struct net *net, struct sk_buff *buf, - struct tipc_bearer *unused1, - struct tipc_media_addr *unused2) +void tipc_bcast_ack_rcv(struct net *net, struct tipc_link *l, u32 acked) { - int bp_index; - struct tipc_msg *msg = buf_msg(buf); - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_bcbearer *bcbearer = tn->bcbearer; - struct tipc_bclink *bclink = tn->bclink; - - /* Prepare broadcast link message for reliable transmission, - * if first time trying to send it; - * preparation is skipped for broadcast link protocol messages - * since they are sent in an unreliable manner and don't need it - */ - if (likely(!msg_non_seq(buf_msg(buf)))) { - bcbuf_set_acks(buf, bclink->bcast_nodes.count); - msg_set_non_seq(msg, 1); - msg_set_mc_netid(msg, tn->net_id); - tn->bcl->stats.sent_info++; - if (WARN_ON(!bclink->bcast_nodes.count)) { - dump_stack(); - return 0; - } - } + struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq; + struct sk_buff_head xmitq; - /* Send buffer over bearers until all targets reached */ - bcbearer->remains = bclink->bcast_nodes; - - for (bp_index = 0; bp_index < MAX_BEARERS; bp_index++) { - struct tipc_bearer *p = bcbearer->bpairs[bp_index].primary; - struct tipc_bearer *s = bcbearer->bpairs[bp_index].secondary; - struct tipc_bearer *bp[2] = {p, s}; - struct tipc_bearer *b = bp[msg_link_selector(msg)]; - struct sk_buff *tbuf; - - if (!p) - break; /* No more bearers to try */ - if (!b) - b = p; - tipc_nmap_diff(&bcbearer->remains, &b->nodes, - &bcbearer->remains_new); - if (bcbearer->remains_new.count == bcbearer->remains.count) - continue; /* Nothing added by bearer pair */ - - if (bp_index == 0) { - /* Use original buffer for first bearer */ - tipc_bearer_send(net, b->identity, buf, &b->bcast_addr); - } else { - /* Avoid concurrent buffer access */ - tbuf = pskb_copy_for_clone(buf, GFP_ATOMIC); - if (!tbuf) - break; - tipc_bearer_send(net, b->identity, tbuf, - &b->bcast_addr); - kfree_skb(tbuf); /* Bearer keeps a clone */ - } - if (bcbearer->remains_new.count == 0) - break; /* All targets reached */ + __skb_queue_head_init(&xmitq); - bcbearer->remains = bcbearer->remains_new; - } + tipc_bcast_lock(net); + tipc_link_bc_ack_rcv(l, acked, &xmitq); + tipc_bcast_unlock(net); - return 0; + tipc_bcbase_xmit(net, &xmitq); + + /* Any socket wakeup messages ? */ + if (!skb_queue_empty(inputq)) + tipc_sk_rcv(net, inputq); } -/** - * tipc_bcbearer_sort - create sets of bearer pairs used by broadcast bearer +/* tipc_bcast_synch_rcv - check and update rcv link with peer's send state + * + * RCU is locked, no other locks set */ -void tipc_bcbearer_sort(struct net *net, struct tipc_node_map *nm_ptr, - u32 node, bool action) +void tipc_bcast_sync_rcv(struct net *net, struct tipc_link *l, + struct tipc_msg *hdr) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_bcbearer *bcbearer = tn->bcbearer; - struct tipc_bcbearer_pair *bp_temp = bcbearer->bpairs_temp; - struct tipc_bcbearer_pair *bp_curr; - struct tipc_bearer *b; - int b_index; - int pri; - - tipc_bclink_lock(net); + struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq; + struct sk_buff_head xmitq; - if (action) - tipc_nmap_add(nm_ptr, node); - else - tipc_nmap_remove(nm_ptr, node); + __skb_queue_head_init(&xmitq); - /* Group bearers by priority (can assume max of two per priority) */ - memset(bp_temp, 0, sizeof(bcbearer->bpairs_temp)); + tipc_bcast_lock(net); + if (msg_type(hdr) == STATE_MSG) { + tipc_link_bc_ack_rcv(l, msg_bcast_ack(hdr), &xmitq); + tipc_link_bc_sync_rcv(l, hdr, &xmitq); + } else { + tipc_link_bc_init_rcv(l, hdr); + } + tipc_bcast_unlock(net); - rcu_read_lock(); - for (b_index = 0; b_index < MAX_BEARERS; b_index++) { - b = rcu_dereference_rtnl(tn->bearer_list[b_index]); - if (!b || !b->nodes.count) - continue; + tipc_bcbase_xmit(net, &xmitq); - if (!bp_temp[b->priority].primary) - bp_temp[b->priority].primary = b; - else - bp_temp[b->priority].secondary = b; - } - rcu_read_unlock(); + /* Any socket wakeup messages ? */ + if (!skb_queue_empty(inputq)) + tipc_sk_rcv(net, inputq); +} - /* Create array of bearer pairs for broadcasting */ - bp_curr = bcbearer->bpairs; - memset(bcbearer->bpairs, 0, sizeof(bcbearer->bpairs)); +/* tipc_bcast_add_peer - add a peer node to broadcast link and bearer + * + * RCU is locked, node lock is set + */ +void tipc_bcast_add_peer(struct net *net, struct tipc_link *uc_l, + struct sk_buff_head *xmitq) +{ + struct tipc_link *snd_l = tipc_bc_sndlink(net); - for (pri = TIPC_MAX_LINK_PRI; pri >= 0; pri--) { + tipc_bcast_lock(net); + tipc_link_add_bc_peer(snd_l, uc_l, xmitq); + tipc_bcbase_select_primary(net); + tipc_bcast_unlock(net); +} - if (!bp_temp[pri].primary) - continue; +/* tipc_bcast_remove_peer - remove a peer node from broadcast link and bearer + * + * RCU is locked, node lock is set + */ +void tipc_bcast_remove_peer(struct net *net, struct tipc_link *rcv_l) +{ + struct tipc_link *snd_l = tipc_bc_sndlink(net); + struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq; + struct sk_buff_head xmitq; - bp_curr->primary = bp_temp[pri].primary; + __skb_queue_head_init(&xmitq); - if (bp_temp[pri].secondary) { - if (tipc_nmap_equal(&bp_temp[pri].primary->nodes, - &bp_temp[pri].secondary->nodes)) { - bp_curr->secondary = bp_temp[pri].secondary; - } else { - bp_curr++; - bp_curr->primary = bp_temp[pri].secondary; - } - } + tipc_bcast_lock(net); + tipc_link_remove_bc_peer(snd_l, rcv_l, &xmitq); + tipc_bcbase_select_primary(net); + tipc_bcast_unlock(net); - bp_curr++; - } + tipc_bcbase_xmit(net, &xmitq); - tipc_bclink_unlock(net); + /* Any socket wakeup messages ? */ + if (!skb_queue_empty(inputq)) + tipc_sk_rcv(net, inputq); } static int __tipc_nl_add_bc_link_stat(struct sk_buff *skb, @@ -784,12 +395,14 @@ int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg) if (!bcl) return 0; - tipc_bclink_lock(net); + tipc_bcast_lock(net); hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family, NLM_F_MULTI, TIPC_NL_LINK_GET); - if (!hdr) + if (!hdr) { + tipc_bcast_unlock(net); return -EMSGSIZE; + } attrs = nla_nest_start(msg->skb, TIPC_NLA_LINK); if (!attrs) @@ -803,9 +416,9 @@ int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg) goto attr_msg_full; if (nla_put_string(msg->skb, TIPC_NLA_LINK_NAME, bcl->name)) goto attr_msg_full; - if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, bcl->next_in_no)) + if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, bcl->rcv_nxt)) goto attr_msg_full; - if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, bcl->next_out_no)) + if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, bcl->snd_nxt)) goto attr_msg_full; prop = nla_nest_start(msg->skb, TIPC_NLA_LINK_PROP); @@ -819,7 +432,7 @@ int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg) if (err) goto attr_msg_full; - tipc_bclink_unlock(net); + tipc_bcast_unlock(net); nla_nest_end(msg->skb, attrs); genlmsg_end(msg->skb, hdr); @@ -830,7 +443,7 @@ prop_msg_full: attr_msg_full: nla_nest_cancel(msg->skb, attrs); msg_full: - tipc_bclink_unlock(net); + tipc_bcast_unlock(net); genlmsg_cancel(msg->skb, hdr); return -EMSGSIZE; @@ -844,143 +457,91 @@ int tipc_bclink_reset_stats(struct net *net) if (!bcl) return -ENOPROTOOPT; - tipc_bclink_lock(net); + tipc_bcast_lock(net); memset(&bcl->stats, 0, sizeof(bcl->stats)); - tipc_bclink_unlock(net); + tipc_bcast_unlock(net); return 0; } -int tipc_bclink_set_queue_limits(struct net *net, u32 limit) +static int tipc_bc_link_set_queue_limits(struct net *net, u32 limit) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_link *bcl = tn->bcl; + struct tipc_link *l = tipc_bc_sndlink(net); - if (!bcl) + if (!l) return -ENOPROTOOPT; - if ((limit < TIPC_MIN_LINK_WIN) || (limit > TIPC_MAX_LINK_WIN)) + if (limit < BCLINK_WIN_MIN) + limit = BCLINK_WIN_MIN; + if (limit > TIPC_MAX_LINK_WIN) return -EINVAL; - - tipc_bclink_lock(net); - tipc_link_set_queue_limits(bcl, limit); - tipc_bclink_unlock(net); + tipc_bcast_lock(net); + tipc_link_set_queue_limits(l, limit); + tipc_bcast_unlock(net); return 0; } -int tipc_bclink_init(struct net *net) +int tipc_nl_bc_link_set(struct net *net, struct nlattr *attrs[]) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_bcbearer *bcbearer; - struct tipc_bclink *bclink; - struct tipc_link *bcl; - - bcbearer = kzalloc(sizeof(*bcbearer), GFP_ATOMIC); - if (!bcbearer) - return -ENOMEM; - - bclink = kzalloc(sizeof(*bclink), GFP_ATOMIC); - if (!bclink) { - kfree(bcbearer); - return -ENOMEM; - } + int err; + u32 win; + struct nlattr *props[TIPC_NLA_PROP_MAX + 1]; - bcl = &bclink->link; - bcbearer->bearer.media = &bcbearer->media; - bcbearer->media.send_msg = tipc_bcbearer_send; - sprintf(bcbearer->media.name, "tipc-broadcast"); - - spin_lock_init(&bclink->lock); - __skb_queue_head_init(&bcl->transmq); - __skb_queue_head_init(&bcl->backlogq); - __skb_queue_head_init(&bcl->deferdq); - skb_queue_head_init(&bcl->wakeupq); - bcl->next_out_no = 1; - spin_lock_init(&bclink->node.lock); - __skb_queue_head_init(&bclink->arrvq); - skb_queue_head_init(&bclink->inputq); - bcl->owner = &bclink->node; - bcl->owner->net = net; - 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); - bcl->state = WORKING_WORKING; - bcl->pmsg = (struct tipc_msg *)&bcl->proto_msg; - msg_set_prevnode(bcl->pmsg, tn->own_addr); - strlcpy(bcl->name, tipc_bclink_name, TIPC_MAX_LINK_NAME); - tn->bcbearer = bcbearer; - tn->bclink = bclink; - tn->bcl = bcl; - return 0; -} + if (!attrs[TIPC_NLA_LINK_PROP]) + return -EINVAL; -void tipc_bclink_stop(struct net *net) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); + err = tipc_nl_parse_link_prop(attrs[TIPC_NLA_LINK_PROP], props); + if (err) + return err; - tipc_bclink_lock(net); - tipc_link_purge_queues(tn->bcl); - tipc_bclink_unlock(net); + if (!props[TIPC_NLA_PROP_WIN]) + return -EOPNOTSUPP; - RCU_INIT_POINTER(tn->bearer_list[BCBEARER], NULL); - synchronize_net(); - kfree(tn->bcbearer); - kfree(tn->bclink); + win = nla_get_u32(props[TIPC_NLA_PROP_WIN]); + + return tipc_bc_link_set_queue_limits(net, win); } -/** - * tipc_nmap_add - add a node to a node map - */ -static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node) +int tipc_bcast_init(struct net *net) { - int n = tipc_node(node); - int w = n / WSIZE; - u32 mask = (1 << (n % WSIZE)); + struct tipc_net *tn = tipc_net(net); + struct tipc_bc_base *bb = NULL; + struct tipc_link *l = NULL; - if ((nm_ptr->map[w] & mask) == 0) { - nm_ptr->count++; - nm_ptr->map[w] |= mask; - } + bb = kzalloc(sizeof(*bb), GFP_ATOMIC); + if (!bb) + goto enomem; + tn->bcbase = bb; + spin_lock_init(&tipc_net(net)->bclock); + + if (!tipc_link_bc_create(net, 0, 0, + U16_MAX, + BCLINK_WIN_DEFAULT, + 0, + &bb->inputq, + NULL, + NULL, + &l)) + goto enomem; + bb->link = l; + tn->bcl = l; + return 0; +enomem: + kfree(bb); + kfree(l); + return -ENOMEM; } -/** - * tipc_nmap_remove - remove a node from a node map - */ -static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node) +void tipc_bcast_reinit(struct net *net) { - int n = tipc_node(node); - int w = n / WSIZE; - u32 mask = (1 << (n % WSIZE)); + struct tipc_bc_base *b = tipc_bc_base(net); - if ((nm_ptr->map[w] & mask) != 0) { - nm_ptr->map[w] &= ~mask; - nm_ptr->count--; - } + msg_set_prevnode(b->link->pmsg, tipc_own_addr(net)); } -/** - * tipc_nmap_diff - find differences between node maps - * @nm_a: input node map A - * @nm_b: input node map B - * @nm_diff: output node map A-B (i.e. nodes of A that are not in B) - */ -static void tipc_nmap_diff(struct tipc_node_map *nm_a, - struct tipc_node_map *nm_b, - struct tipc_node_map *nm_diff) +void tipc_bcast_stop(struct net *net) { - int stop = ARRAY_SIZE(nm_a->map); - int w; - int b; - u32 map; - - memset(nm_diff, 0, sizeof(*nm_diff)); - for (w = 0; w < stop; w++) { - map = nm_a->map[w] ^ (nm_a->map[w] & nm_b->map[w]); - nm_diff->map[w] = map; - if (map != 0) { - for (b = 0 ; b < WSIZE; b++) { - if (map & (1 << b)) - nm_diff->count++; - } - } - } + struct tipc_net *tn = net_generic(net, tipc_net_id); + + synchronize_net(); + kfree(tn->bcbase); + kfree(tn->bcl); } diff --git a/kernel/net/tipc/bcast.h b/kernel/net/tipc/bcast.h index 4bdc12277..2855b9356 100644 --- a/kernel/net/tipc/bcast.h +++ b/kernel/net/tipc/bcast.h @@ -37,100 +37,44 @@ #ifndef _TIPC_BCAST_H #define _TIPC_BCAST_H -#include <linux/tipc_config.h> -#include "link.h" -#include "node.h" +#include "core.h" -/** - * struct tipc_bcbearer_pair - a pair of bearers used by broadcast link - * @primary: pointer to primary bearer - * @secondary: pointer to secondary bearer - * - * Bearers must have same priority and same set of reachable destinations - * to be paired. - */ - -struct tipc_bcbearer_pair { - struct tipc_bearer *primary; - struct tipc_bearer *secondary; -}; - -#define BCBEARER MAX_BEARERS - -/** - * struct tipc_bcbearer - bearer used by broadcast link - * @bearer: (non-standard) broadcast bearer structure - * @media: (non-standard) broadcast media structure - * @bpairs: array of bearer pairs - * @bpairs_temp: temporary array of bearer pairs used by tipc_bcbearer_sort() - * @remains: temporary node map used by tipc_bcbearer_send() - * @remains_new: temporary node map used tipc_bcbearer_send() - * - * Note: The fields labelled "temporary" are incorporated into the bearer - * to avoid consuming potentially limited stack space through the use of - * large local variables within multicast routines. Concurrent access is - * prevented through use of the spinlock "bclink_lock". - */ -struct tipc_bcbearer { - struct tipc_bearer bearer; - struct tipc_media media; - struct tipc_bcbearer_pair bpairs[MAX_BEARERS]; - struct tipc_bcbearer_pair bpairs_temp[TIPC_MAX_LINK_PRI + 1]; - struct tipc_node_map remains; - struct tipc_node_map remains_new; -}; +struct tipc_node; +struct tipc_msg; +struct tipc_nl_msg; +struct tipc_node_map; -/** - * struct tipc_bclink - link used for broadcast messages - * @lock: spinlock governing access to structure - * @link: (non-standard) broadcast link structure - * @node: (non-standard) node structure representing b'cast link's peer node - * @bcast_nodes: map of broadcast-capable nodes - * @retransmit_to: node that most recently requested a retransmit - * - * Handles sequence numbering, fragmentation, bundling, etc. - */ -struct tipc_bclink { - spinlock_t lock; - struct tipc_link link; - struct tipc_node node; - struct sk_buff_head arrvq; - struct sk_buff_head inputq; - struct tipc_node_map bcast_nodes; - struct tipc_node *retransmit_to; -}; +int tipc_bcast_init(struct net *net); +void tipc_bcast_reinit(struct net *net); +void tipc_bcast_stop(struct net *net); +void tipc_bcast_add_peer(struct net *net, struct tipc_link *l, + struct sk_buff_head *xmitq); +void tipc_bcast_remove_peer(struct net *net, struct tipc_link *rcv_bcl); +void tipc_bcast_inc_bearer_dst_cnt(struct net *net, int bearer_id); +void tipc_bcast_dec_bearer_dst_cnt(struct net *net, int bearer_id); +int tipc_bcast_get_mtu(struct net *net); +int tipc_bcast_xmit(struct net *net, struct sk_buff_head *list); +int tipc_bcast_rcv(struct net *net, struct tipc_link *l, struct sk_buff *skb); +void tipc_bcast_ack_rcv(struct net *net, struct tipc_link *l, u32 acked); +void tipc_bcast_sync_rcv(struct net *net, struct tipc_link *l, + struct tipc_msg *hdr); +int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg); +int tipc_nl_bc_link_set(struct net *net, struct nlattr *attrs[]); +int tipc_bclink_reset_stats(struct net *net); -struct tipc_node; -extern const char tipc_bclink_name[]; +static inline void tipc_bcast_lock(struct net *net) +{ + spin_lock_bh(&tipc_net(net)->bclock); +} -/** - * tipc_nmap_equal - test for equality of node maps - */ -static inline int tipc_nmap_equal(struct tipc_node_map *nm_a, - struct tipc_node_map *nm_b) +static inline void tipc_bcast_unlock(struct net *net) { - return !memcmp(nm_a, nm_b, sizeof(*nm_a)); + spin_unlock_bh(&tipc_net(net)->bclock); } -int tipc_bclink_init(struct net *net); -void tipc_bclink_stop(struct net *net); -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); -void tipc_bclink_acknowledge(struct tipc_node *n_ptr, u32 acked); -void tipc_bclink_rcv(struct net *net, struct sk_buff *buf); -u32 tipc_bclink_get_last_sent(struct net *net); -u32 tipc_bclink_acks_missing(struct tipc_node *n_ptr); -void tipc_bclink_update_link_state(struct tipc_node *node, - u32 last_sent); -int tipc_bclink_reset_stats(struct net *net); -int tipc_bclink_set_queue_limits(struct net *net, u32 limit); -void tipc_bcbearer_sort(struct net *net, struct tipc_node_map *nm_ptr, - u32 node, bool action); -uint tipc_bclink_get_mtu(void); -int tipc_bclink_xmit(struct net *net, struct sk_buff_head *list); -void tipc_bclink_wakeup_users(struct net *net); -int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg); -void tipc_bclink_input(struct net *net); +static inline struct tipc_link *tipc_bc_sndlink(struct net *net) +{ + return tipc_net(net)->bcl; +} #endif diff --git a/kernel/net/tipc/bearer.c b/kernel/net/tipc/bearer.c index 70e3dacbf..648f2a67f 100644 --- a/kernel/net/tipc/bearer.c +++ b/kernel/net/tipc/bearer.c @@ -71,8 +71,7 @@ static const struct nla_policy tipc_nl_media_policy[TIPC_NLA_MEDIA_MAX + 1] = { [TIPC_NLA_MEDIA_PROP] = { .type = NLA_NESTED } }; -static void bearer_disable(struct net *net, struct tipc_bearer *b_ptr, - bool shutting_down); +static void bearer_disable(struct net *net, struct tipc_bearer *b_ptr); /** * tipc_media_find - locates specified media object by name @@ -194,10 +193,8 @@ void tipc_bearer_add_dest(struct net *net, u32 bearer_id, u32 dest) rcu_read_lock(); b_ptr = rcu_dereference_rtnl(tn->bearer_list[bearer_id]); - if (b_ptr) { - tipc_bcbearer_sort(net, &b_ptr->nodes, dest, true); + if (b_ptr) tipc_disc_add_dest(b_ptr->link_req); - } rcu_read_unlock(); } @@ -208,10 +205,8 @@ void tipc_bearer_remove_dest(struct net *net, u32 bearer_id, u32 dest) rcu_read_lock(); b_ptr = rcu_dereference_rtnl(tn->bearer_list[bearer_id]); - if (b_ptr) { - tipc_bcbearer_sort(net, &b_ptr->nodes, dest, false); + if (b_ptr) tipc_disc_remove_dest(b_ptr->link_req); - } rcu_read_unlock(); } @@ -324,7 +319,7 @@ restart: res = tipc_disc_create(net, b_ptr, &b_ptr->bcast_addr); if (res) { - bearer_disable(net, b_ptr, false); + bearer_disable(net, b_ptr); pr_warn("Bearer <%s> rejected, discovery object creation failed\n", name); return -EINVAL; @@ -344,7 +339,7 @@ restart: static int tipc_reset_bearer(struct net *net, struct tipc_bearer *b_ptr) { pr_info("Resetting bearer <%s>\n", b_ptr->name); - tipc_link_reset_list(net, b_ptr->identity); + tipc_node_delete_links(net, b_ptr->identity); tipc_disc_reset(net, b_ptr); return 0; } @@ -354,8 +349,7 @@ static int tipc_reset_bearer(struct net *net, struct tipc_bearer *b_ptr) * * Note: This routine assumes caller holds RTNL lock. */ -static void bearer_disable(struct net *net, struct tipc_bearer *b_ptr, - bool shutting_down) +static void bearer_disable(struct net *net, struct tipc_bearer *b_ptr) { struct tipc_net *tn = net_generic(net, tipc_net_id); u32 i; @@ -363,7 +357,8 @@ static void bearer_disable(struct net *net, struct tipc_bearer *b_ptr, pr_info("Disabling bearer <%s>\n", b_ptr->name); b_ptr->media->disable_media(b_ptr); - tipc_link_delete_list(net, b_ptr->identity, shutting_down); + tipc_node_delete_links(net, b_ptr->identity); + RCU_INIT_POINTER(b_ptr->media_ptr, NULL); if (b_ptr->link_req) tipc_disc_delete(b_ptr->link_req); @@ -401,16 +396,13 @@ int tipc_enable_l2_media(struct net *net, struct tipc_bearer *b, /* tipc_disable_l2_media - detach TIPC bearer from an L2 interface * - * Mark L2 bearer as inactive so that incoming buffers are thrown away, - * then get worker thread to complete bearer cleanup. (Can't do cleanup - * here because cleanup code needs to sleep and caller holds spinlocks.) + * Mark L2 bearer as inactive so that incoming buffers are thrown away */ void tipc_disable_l2_media(struct tipc_bearer *b) { struct net_device *dev; dev = (struct net_device *)rtnl_dereference(b->media_ptr); - RCU_INIT_POINTER(b->media_ptr, NULL); RCU_INIT_POINTER(dev->tipc_ptr, NULL); synchronize_net(); dev_put(dev); @@ -422,10 +414,9 @@ void tipc_disable_l2_media(struct tipc_bearer *b) * @b_ptr: the bearer through which the packet is to be sent * @dest: peer destination address */ -int tipc_l2_send_msg(struct net *net, struct sk_buff *buf, +int tipc_l2_send_msg(struct net *net, struct sk_buff *skb, struct tipc_bearer *b, struct tipc_media_addr *dest) { - struct sk_buff *clone; struct net_device *dev; int delta; @@ -433,42 +424,97 @@ int tipc_l2_send_msg(struct net *net, struct sk_buff *buf, if (!dev) return 0; - clone = skb_clone(buf, GFP_ATOMIC); - if (!clone) - return 0; - - delta = dev->hard_header_len - skb_headroom(buf); + delta = dev->hard_header_len - skb_headroom(skb); if ((delta > 0) && - pskb_expand_head(clone, SKB_DATA_ALIGN(delta), 0, GFP_ATOMIC)) { - kfree_skb(clone); + pskb_expand_head(skb, SKB_DATA_ALIGN(delta), 0, GFP_ATOMIC)) { + kfree_skb(skb); return 0; } - skb_reset_network_header(clone); - clone->dev = dev; - clone->protocol = htons(ETH_P_TIPC); - dev_hard_header(clone, dev, ETH_P_TIPC, dest->value, - dev->dev_addr, clone->len); - dev_queue_xmit(clone); + skb_reset_network_header(skb); + skb->dev = dev; + skb->protocol = htons(ETH_P_TIPC); + dev_hard_header(skb, dev, ETH_P_TIPC, dest->value, + dev->dev_addr, skb->len); + dev_queue_xmit(skb); return 0; } -/* tipc_bearer_send- sends buffer to destination over bearer - * - * IMPORTANT: - * The media send routine must not alter the buffer being passed in - * as it may be needed for later retransmission! +int tipc_bearer_mtu(struct net *net, u32 bearer_id) +{ + int mtu = 0; + struct tipc_bearer *b; + + rcu_read_lock(); + b = rcu_dereference_rtnl(tipc_net(net)->bearer_list[bearer_id]); + if (b) + mtu = b->mtu; + rcu_read_unlock(); + return mtu; +} + +/* tipc_bearer_xmit_skb - sends buffer to destination over bearer + */ +void tipc_bearer_xmit_skb(struct net *net, u32 bearer_id, + struct sk_buff *skb, + struct tipc_media_addr *dest) +{ + struct tipc_net *tn = tipc_net(net); + struct tipc_bearer *b; + + rcu_read_lock(); + b = rcu_dereference_rtnl(tn->bearer_list[bearer_id]); + if (likely(b)) + b->media->send_msg(net, skb, b, dest); + rcu_read_unlock(); +} + +/* tipc_bearer_xmit() -send buffer to destination over bearer */ -void tipc_bearer_send(struct net *net, u32 bearer_id, struct sk_buff *buf, - struct tipc_media_addr *dest) +void tipc_bearer_xmit(struct net *net, u32 bearer_id, + struct sk_buff_head *xmitq, + struct tipc_media_addr *dst) { struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_bearer *b_ptr; + struct tipc_bearer *b; + struct sk_buff *skb, *tmp; + + if (skb_queue_empty(xmitq)) + return; rcu_read_lock(); - b_ptr = rcu_dereference_rtnl(tn->bearer_list[bearer_id]); - if (likely(b_ptr)) - b_ptr->media->send_msg(net, buf, b_ptr, dest); + b = rcu_dereference_rtnl(tn->bearer_list[bearer_id]); + if (likely(b)) { + skb_queue_walk_safe(xmitq, skb, tmp) { + __skb_dequeue(xmitq); + b->media->send_msg(net, skb, b, dst); + } + } + rcu_read_unlock(); +} + +/* tipc_bearer_bc_xmit() - broadcast buffers to all destinations + */ +void tipc_bearer_bc_xmit(struct net *net, u32 bearer_id, + struct sk_buff_head *xmitq) +{ + struct tipc_net *tn = tipc_net(net); + int net_id = tn->net_id; + struct tipc_bearer *b; + struct sk_buff *skb, *tmp; + struct tipc_msg *hdr; + + rcu_read_lock(); + b = rcu_dereference_rtnl(tn->bearer_list[bearer_id]); + if (likely(b)) { + skb_queue_walk_safe(xmitq, skb, tmp) { + hdr = buf_msg(skb); + msg_set_non_seq(hdr, 1); + msg_set_mc_netid(hdr, net_id); + __skb_dequeue(xmitq); + b->media->send_msg(net, skb, b, &b->bcast_addr); + } + } rcu_read_unlock(); } @@ -530,7 +576,7 @@ static int tipc_l2_device_event(struct notifier_block *nb, unsigned long evt, case NETDEV_CHANGE: if (netif_carrier_ok(dev)) break; - case NETDEV_DOWN: + case NETDEV_GOING_DOWN: case NETDEV_CHANGEMTU: tipc_reset_bearer(net, b_ptr); break; @@ -541,7 +587,7 @@ static int tipc_l2_device_event(struct notifier_block *nb, unsigned long evt, break; case NETDEV_UNREGISTER: case NETDEV_CHANGENAME: - bearer_disable(dev_net(dev), b_ptr, false); + bearer_disable(dev_net(dev), b_ptr); break; } return NOTIFY_OK; @@ -583,7 +629,7 @@ void tipc_bearer_stop(struct net *net) for (i = 0; i < MAX_BEARERS; i++) { b_ptr = rtnl_dereference(tn->bearer_list[i]); if (b_ptr) { - bearer_disable(net, b_ptr, true); + bearer_disable(net, b_ptr); tn->bearer_list[i] = NULL; } } @@ -747,7 +793,7 @@ int tipc_nl_bearer_disable(struct sk_buff *skb, struct genl_info *info) return -EINVAL; } - bearer_disable(net, bearer, false); + bearer_disable(net, bearer); rtnl_unlock(); return 0; @@ -812,7 +858,7 @@ int tipc_nl_bearer_set(struct sk_buff *skb, struct genl_info *info) char *name; struct tipc_bearer *b; struct nlattr *attrs[TIPC_NLA_BEARER_MAX + 1]; - struct net *net = genl_info_net(info); + struct net *net = sock_net(skb->sk); if (!info->attrs[TIPC_NLA_BEARER]) return -EINVAL; diff --git a/kernel/net/tipc/bearer.h b/kernel/net/tipc/bearer.h index 5cad243ee..552185bc4 100644 --- a/kernel/net/tipc/bearer.h +++ b/kernel/net/tipc/bearer.h @@ -38,9 +38,9 @@ #define _TIPC_BEARER_H #include "netlink.h" +#include "core.h" #include <net/genetlink.h> -#define MAX_BEARERS 2 #define MAX_MEDIA 3 #define MAX_NODES 4096 #define WSIZE 32 @@ -163,6 +163,7 @@ struct tipc_bearer { u32 identity; struct tipc_link_req *link_req; char net_plane; + int node_cnt; struct tipc_node_map nodes; }; @@ -215,7 +216,14 @@ struct tipc_media *tipc_media_find(const char *name); int tipc_bearer_setup(void); void tipc_bearer_cleanup(void); void tipc_bearer_stop(struct net *net); -void tipc_bearer_send(struct net *net, u32 bearer_id, struct sk_buff *buf, - struct tipc_media_addr *dest); +int tipc_bearer_mtu(struct net *net, u32 bearer_id); +void tipc_bearer_xmit_skb(struct net *net, u32 bearer_id, + struct sk_buff *skb, + struct tipc_media_addr *dest); +void tipc_bearer_xmit(struct net *net, u32 bearer_id, + struct sk_buff_head *xmitq, + struct tipc_media_addr *dst); +void tipc_bearer_bc_xmit(struct net *net, u32 bearer_id, + struct sk_buff_head *xmitq); #endif /* _TIPC_BEARER_H */ diff --git a/kernel/net/tipc/core.c b/kernel/net/tipc/core.c index be1c9fa60..03a842870 100644 --- a/kernel/net/tipc/core.c +++ b/kernel/net/tipc/core.c @@ -42,6 +42,7 @@ #include "bearer.h" #include "net.h" #include "socket.h" +#include "bcast.h" #include <linux/module.h> @@ -68,11 +69,18 @@ static int __net_init tipc_init_net(struct net *net) if (err) goto out_nametbl; - err = tipc_subscr_start(net); + err = tipc_topsrv_start(net); if (err) goto out_subscr; + + err = tipc_bcast_init(net); + if (err) + goto out_bclink; + return 0; +out_bclink: + tipc_bcast_stop(net); out_subscr: tipc_nametbl_stop(net); out_nametbl: @@ -83,8 +91,9 @@ out_sk_rht: static void __net_exit tipc_exit_net(struct net *net) { - tipc_subscr_stop(net); + tipc_topsrv_stop(net); tipc_net_stop(net); + tipc_bcast_stop(net); tipc_nametbl_stop(net); tipc_sk_rht_destroy(net); } diff --git a/kernel/net/tipc/core.h b/kernel/net/tipc/core.h index 3dc68c7a9..18e95a802 100644 --- a/kernel/net/tipc/core.h +++ b/kernel/net/tipc/core.h @@ -60,16 +60,18 @@ #include <net/netns/generic.h> #include <linux/rhashtable.h> -#include "node.h" -#include "bearer.h" -#include "bcast.h" -#include "netlink.h" -#include "link.h" -#include "node.h" -#include "msg.h" +struct tipc_node; +struct tipc_bearer; +struct tipc_bc_base; +struct tipc_link; +struct tipc_name_table; +struct tipc_server; #define TIPC_MOD_VER "2.0.0" +#define NODE_HTABLE_SIZE 512 +#define MAX_BEARERS 3 + extern int tipc_net_id __read_mostly; extern int sysctl_tipc_rmem[3] __read_mostly; extern int sysctl_tipc_named_timeout __read_mostly; @@ -90,8 +92,8 @@ struct tipc_net { struct tipc_bearer __rcu *bearer_list[MAX_BEARERS + 1]; /* Broadcast link */ - struct tipc_bcbearer *bcbearer; - struct tipc_bclink *bclink; + spinlock_t bclock; + struct tipc_bc_base *bcbase; struct tipc_link *bcl; /* Socket hash table */ @@ -106,6 +108,41 @@ struct tipc_net { atomic_t subscription_count; }; +static inline struct tipc_net *tipc_net(struct net *net) +{ + return net_generic(net, tipc_net_id); +} + +static inline int tipc_netid(struct net *net) +{ + return tipc_net(net)->net_id; +} + +static inline u16 mod(u16 x) +{ + return x & 0xffffu; +} + +static inline int less_eq(u16 left, u16 right) +{ + return mod(right - left) < 32768u; +} + +static inline int more(u16 left, u16 right) +{ + return !less_eq(left, right); +} + +static inline int less(u16 left, u16 right) +{ + return less_eq(left, right) && (mod(right) != mod(left)); +} + +static inline int in_range(u16 val, u16 min, u16 max) +{ + return !less(val, min) && !more(val, max); +} + #ifdef CONFIG_SYSCTL int tipc_register_sysctl(void); void tipc_unregister_sysctl(void); diff --git a/kernel/net/tipc/discover.c b/kernel/net/tipc/discover.c index 967e292f5..afe8c47c4 100644 --- a/kernel/net/tipc/discover.c +++ b/kernel/net/tipc/discover.c @@ -35,7 +35,7 @@ */ #include "core.h" -#include "link.h" +#include "node.h" #include "discover.h" /* min delay during bearer start up */ @@ -89,7 +89,7 @@ static void tipc_disc_init_msg(struct net *net, struct sk_buff *buf, u32 type, MAX_H_SIZE, dest_domain); msg_set_non_seq(msg, 1); msg_set_node_sig(msg, tn->random); - msg_set_node_capabilities(msg, 0); + msg_set_node_capabilities(msg, TIPC_NODE_CAPABILITIES); msg_set_dest_domain(msg, dest_domain); msg_set_bc_netid(msg, tn->net_id); b_ptr->media->addr2msg(msg_media_addr(msg), &b_ptr->addr); @@ -120,30 +120,24 @@ static void disc_dupl_alert(struct tipc_bearer *b_ptr, u32 node_addr, * @buf: buffer containing message * @bearer: bearer that message arrived on */ -void tipc_disc_rcv(struct net *net, struct sk_buff *buf, +void tipc_disc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *bearer) { struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_node *node; - struct tipc_link *link; struct tipc_media_addr maddr; - struct sk_buff *rbuf; - struct tipc_msg *msg = buf_msg(buf); - u32 ddom = msg_dest_domain(msg); - u32 onode = msg_prevnode(msg); - u32 net_id = msg_bc_netid(msg); - u32 mtyp = msg_type(msg); - u32 signature = msg_node_sig(msg); - u16 caps = msg_node_capabilities(msg); - bool addr_match = false; - bool sign_match = false; - bool link_up = false; - bool accept_addr = false; - bool accept_sign = false; + struct sk_buff *rskb; + struct tipc_msg *hdr = buf_msg(skb); + u32 ddom = msg_dest_domain(hdr); + u32 onode = msg_prevnode(hdr); + u32 net_id = msg_bc_netid(hdr); + u32 mtyp = msg_type(hdr); + u32 signature = msg_node_sig(hdr); + u16 caps = msg_node_capabilities(hdr); bool respond = false; + bool dupl_addr = false; - bearer->media->msg2addr(bearer, &maddr, msg_media_addr(msg)); - kfree_skb(buf); + bearer->media->msg2addr(bearer, &maddr, msg_media_addr(hdr)); + kfree_skb(skb); /* Ensure message from node is valid and communication is permitted */ if (net_id != tn->net_id) @@ -165,102 +159,19 @@ void tipc_disc_rcv(struct net *net, struct sk_buff *buf, if (!tipc_in_scope(bearer->domain, onode)) return; - node = tipc_node_create(net, onode); - if (!node) - return; - tipc_node_lock(node); - node->capabilities = caps; - link = node->links[bearer->identity]; - - /* Prepare to validate requesting node's signature and media address */ - sign_match = (signature == node->signature); - addr_match = link && !memcmp(&link->media_addr, &maddr, sizeof(maddr)); - link_up = link && tipc_link_is_up(link); - - - /* These three flags give us eight permutations: */ - - if (sign_match && addr_match && link_up) { - /* All is fine. Do nothing. */ - } else if (sign_match && addr_match && !link_up) { - /* Respond. The link will come up in due time */ - respond = true; - } else if (sign_match && !addr_match && link_up) { - /* Peer has changed i/f address without rebooting. - * If so, the link will reset soon, and the next - * discovery will be accepted. So we can ignore it. - * It may also be an cloned or malicious peer having - * chosen the same node address and signature as an - * existing one. - * Ignore requests until the link goes down, if ever. - */ - disc_dupl_alert(bearer, onode, &maddr); - } else if (sign_match && !addr_match && !link_up) { - /* Peer link has changed i/f address without rebooting. - * It may also be a cloned or malicious peer; we can't - * distinguish between the two. - * The signature is correct, so we must accept. - */ - accept_addr = true; - respond = true; - } else if (!sign_match && addr_match && link_up) { - /* Peer node rebooted. Two possibilities: - * - Delayed re-discovery; this link endpoint has already - * reset and re-established contact with the peer, before - * receiving a discovery message from that node. - * (The peer happened to receive one from this node first). - * - The peer came back so fast that our side has not - * discovered it yet. Probing from this side will soon - * reset the link, since there can be no working link - * endpoint at the peer end, and the link will re-establish. - * Accept the signature, since it comes from a known peer. - */ - accept_sign = true; - } else if (!sign_match && addr_match && !link_up) { - /* The peer node has rebooted. - * Accept signature, since it is a known peer. - */ - accept_sign = true; - respond = true; - } else if (!sign_match && !addr_match && link_up) { - /* Peer rebooted with new address, or a new/duplicate peer. - * Ignore until the link goes down, if ever. - */ + tipc_node_check_dest(net, onode, bearer, caps, signature, + &maddr, &respond, &dupl_addr); + if (dupl_addr) disc_dupl_alert(bearer, onode, &maddr); - } else if (!sign_match && !addr_match && !link_up) { - /* Peer rebooted with new address, or it is a new peer. - * Accept signature and address. - */ - accept_sign = true; - accept_addr = true; - respond = true; - } - - if (accept_sign) - node->signature = signature; - - if (accept_addr) { - if (!link) - link = tipc_link_create(node, bearer, &maddr); - if (link) { - memcpy(&link->media_addr, &maddr, sizeof(maddr)); - tipc_link_reset(link); - } else { - respond = false; - } - } /* Send response, if necessary */ if (respond && (mtyp == DSC_REQ_MSG)) { - rbuf = tipc_buf_acquire(MAX_H_SIZE); - if (rbuf) { - tipc_disc_init_msg(net, rbuf, DSC_RESP_MSG, bearer); - tipc_bearer_send(net, bearer->identity, rbuf, &maddr); - kfree_skb(rbuf); - } + rskb = tipc_buf_acquire(MAX_H_SIZE); + if (!rskb) + return; + tipc_disc_init_msg(net, rskb, DSC_RESP_MSG, bearer); + tipc_bearer_xmit_skb(net, bearer->identity, rskb, &maddr); } - tipc_node_unlock(node); - tipc_node_put(node); } /** @@ -313,6 +224,7 @@ void tipc_disc_remove_dest(struct tipc_link_req *req) static void disc_timeout(unsigned long data) { struct tipc_link_req *req = (struct tipc_link_req *)data; + struct sk_buff *skb; int max_delay; spin_lock_bh(&req->lock); @@ -330,9 +242,9 @@ static void disc_timeout(unsigned long data) * hold at fast polling rate if don't have any associated nodes, * otherwise hold at slow polling rate */ - tipc_bearer_send(req->net, req->bearer_id, req->buf, &req->dest); - - + skb = skb_clone(req->buf, GFP_ATOMIC); + if (skb) + tipc_bearer_xmit_skb(req->net, req->bearer_id, skb, &req->dest); req->timer_intv *= 2; if (req->num_nodes) max_delay = TIPC_LINK_REQ_SLOW; @@ -359,6 +271,7 @@ int tipc_disc_create(struct net *net, struct tipc_bearer *b_ptr, struct tipc_media_addr *dest) { struct tipc_link_req *req; + struct sk_buff *skb; req = kmalloc(sizeof(*req), GFP_ATOMIC); if (!req) @@ -380,7 +293,9 @@ int tipc_disc_create(struct net *net, struct tipc_bearer *b_ptr, setup_timer(&req->timer, disc_timeout, (unsigned long)req); mod_timer(&req->timer, jiffies + req->timer_intv); b_ptr->link_req = req; - tipc_bearer_send(net, req->bearer_id, req->buf, &req->dest); + skb = skb_clone(req->buf, GFP_ATOMIC); + if (skb) + tipc_bearer_xmit_skb(net, req->bearer_id, skb, &req->dest); return 0; } @@ -404,6 +319,7 @@ void tipc_disc_delete(struct tipc_link_req *req) void tipc_disc_reset(struct net *net, struct tipc_bearer *b_ptr) { struct tipc_link_req *req = b_ptr->link_req; + struct sk_buff *skb; spin_lock_bh(&req->lock); tipc_disc_init_msg(net, req->buf, DSC_REQ_MSG, b_ptr); @@ -413,6 +329,8 @@ void tipc_disc_reset(struct net *net, struct tipc_bearer *b_ptr) req->num_nodes = 0; req->timer_intv = TIPC_LINK_REQ_INIT; mod_timer(&req->timer, jiffies + req->timer_intv); - tipc_bearer_send(net, req->bearer_id, req->buf, &req->dest); + skb = skb_clone(req->buf, GFP_ATOMIC); + if (skb) + tipc_bearer_xmit_skb(net, req->bearer_id, skb, &req->dest); spin_unlock_bh(&req->lock); } diff --git a/kernel/net/tipc/link.c b/kernel/net/tipc/link.c index 43a515dc9..91aea071a 100644 --- a/kernel/net/tipc/link.c +++ b/kernel/net/tipc/link.c @@ -48,9 +48,9 @@ /* * Error message prefixes */ -static const char *link_co_err = "Link changeover error, "; +static const char *link_co_err = "Link tunneling error, "; static const char *link_rst_msg = "Resetting link "; -static const char *link_unk_evt = "Unknown link event "; +static const char tipc_bclink_name[] = "broadcast-link"; static const struct nla_policy tipc_nl_link_policy[TIPC_NLA_LINK_MAX + 1] = { [TIPC_NLA_LINK_UNSPEC] = { .type = NLA_UNSPEC }, @@ -76,257 +76,530 @@ static const struct nla_policy tipc_nl_prop_policy[TIPC_NLA_PROP_MAX + 1] = { [TIPC_NLA_PROP_WIN] = { .type = NLA_U32 } }; +/* Send states for broadcast NACKs + */ +enum { + BC_NACK_SND_CONDITIONAL, + BC_NACK_SND_UNCONDITIONAL, + BC_NACK_SND_SUPPRESS, +}; + +/* + * Interval between NACKs when packets arrive out of order + */ +#define TIPC_NACK_INTV (TIPC_MIN_LINK_WIN * 2) /* * Out-of-range value for link session numbers */ -#define INVALID_SESSION 0x10000 +#define WILDCARD_SESSION 0x10000 -/* - * Link state events: +/* Link FSM states: */ -#define STARTING_EVT 856384768 /* link processing trigger */ -#define TRAFFIC_MSG_EVT 560815u /* rx'd ??? */ -#define TIMEOUT_EVT 560817u /* link timer expired */ +enum { + LINK_ESTABLISHED = 0xe, + LINK_ESTABLISHING = 0xe << 4, + LINK_RESET = 0x1 << 8, + LINK_RESETTING = 0x2 << 12, + LINK_PEER_RESET = 0xd << 16, + LINK_FAILINGOVER = 0xf << 20, + LINK_SYNCHING = 0xc << 24 +}; -/* - * State value stored in 'failover_pkts' +/* Link FSM state checking routines */ -#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 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 int link_is_up(struct tipc_link *l) +{ + return l->state & (LINK_ESTABLISHED | LINK_SYNCHING); +} + +static int tipc_link_proto_rcv(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *xmitq); +static void tipc_link_build_proto_msg(struct tipc_link *l, int mtyp, bool probe, + u16 rcvgap, int tolerance, int priority, + struct sk_buff_head *xmitq); static void link_reset_statistics(struct tipc_link *l_ptr); static void link_print(struct tipc_link *l_ptr, const char *str); -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); +static void tipc_link_build_nack_msg(struct tipc_link *l, + struct sk_buff_head *xmitq); +static void tipc_link_build_bc_init_msg(struct tipc_link *l, + struct sk_buff_head *xmitq); +static bool tipc_link_release_pkts(struct tipc_link *l, u16 to); + /* - * Simple link routines + * Simple non-static link routines (i.e. referenced outside this file) */ -static unsigned int align(unsigned int i) +bool tipc_link_is_up(struct tipc_link *l) { - return (i + 3) & ~3u; + return link_is_up(l); } -static void tipc_link_release(struct kref *kref) +bool tipc_link_peer_is_down(struct tipc_link *l) { - kfree(container_of(kref, struct tipc_link, ref)); + return l->state == LINK_PEER_RESET; } -static void tipc_link_get(struct tipc_link *l_ptr) +bool tipc_link_is_reset(struct tipc_link *l) { - kref_get(&l_ptr->ref); + return l->state & (LINK_RESET | LINK_FAILINGOVER | LINK_ESTABLISHING); } -static void tipc_link_put(struct tipc_link *l_ptr) +bool tipc_link_is_establishing(struct tipc_link *l) { - kref_put(&l_ptr->ref, tipc_link_release); + return l->state == LINK_ESTABLISHING; } -static struct tipc_link *tipc_parallel_link(struct tipc_link *l) +bool tipc_link_is_synching(struct tipc_link *l) { - if (l->owner->active_links[0] != l) - return l->owner->active_links[0]; - return l->owner->active_links[1]; + return l->state == LINK_SYNCHING; } -/* - * Simple non-static link routines (i.e. referenced outside this file) - */ -int tipc_link_is_up(struct tipc_link *l_ptr) +bool tipc_link_is_failingover(struct tipc_link *l) { - if (!l_ptr) - return 0; - return link_working_working(l_ptr) || link_working_unknown(l_ptr); + return l->state == LINK_FAILINGOVER; } -int tipc_link_is_active(struct tipc_link *l_ptr) +bool tipc_link_is_blocked(struct tipc_link *l) { - return (l_ptr->owner->active_links[0] == l_ptr) || - (l_ptr->owner->active_links[1] == l_ptr); + return l->state & (LINK_RESETTING | LINK_PEER_RESET | LINK_FAILINGOVER); } -/** - * link_timeout - handle expiration of link timer - * @l_ptr: pointer to link - */ -static void link_timeout(unsigned long data) +static bool link_is_bc_sndlink(struct tipc_link *l) { - struct tipc_link *l_ptr = (struct tipc_link *)data; - struct sk_buff *skb; + return !l->bc_sndlink; +} - tipc_node_lock(l_ptr->owner); +static bool link_is_bc_rcvlink(struct tipc_link *l) +{ + return ((l->bc_rcvlink == l) && !link_is_bc_sndlink(l)); +} - /* update counters used in statistical profiling of send traffic */ - l_ptr->stats.accu_queue_sz += skb_queue_len(&l_ptr->transmq); - l_ptr->stats.queue_sz_counts++; +int tipc_link_is_active(struct tipc_link *l) +{ + return l->active; +} - skb = skb_peek(&l_ptr->transmq); - if (skb) { - struct tipc_msg *msg = buf_msg(skb); - u32 length = msg_size(msg); +void tipc_link_set_active(struct tipc_link *l, bool active) +{ + l->active = active; +} - if ((msg_user(msg) == MSG_FRAGMENTER) && - (msg_type(msg) == FIRST_FRAGMENT)) { - length = msg_size(msg_get_wrapped(msg)); - } - if (length) { - l_ptr->stats.msg_lengths_total += length; - l_ptr->stats.msg_length_counts++; - if (length <= 64) - l_ptr->stats.msg_length_profile[0]++; - else if (length <= 256) - l_ptr->stats.msg_length_profile[1]++; - else if (length <= 1024) - l_ptr->stats.msg_length_profile[2]++; - else if (length <= 4096) - l_ptr->stats.msg_length_profile[3]++; - else if (length <= 16384) - l_ptr->stats.msg_length_profile[4]++; - else if (length <= 32768) - l_ptr->stats.msg_length_profile[5]++; - else - l_ptr->stats.msg_length_profile[6]++; - } +void tipc_link_add_bc_peer(struct tipc_link *snd_l, + struct tipc_link *uc_l, + struct sk_buff_head *xmitq) +{ + struct tipc_link *rcv_l = uc_l->bc_rcvlink; + + snd_l->ackers++; + rcv_l->acked = snd_l->snd_nxt - 1; + snd_l->state = LINK_ESTABLISHED; + tipc_link_build_bc_init_msg(uc_l, xmitq); +} + +void tipc_link_remove_bc_peer(struct tipc_link *snd_l, + struct tipc_link *rcv_l, + struct sk_buff_head *xmitq) +{ + u16 ack = snd_l->snd_nxt - 1; + + snd_l->ackers--; + tipc_link_bc_ack_rcv(rcv_l, ack, xmitq); + tipc_link_reset(rcv_l); + rcv_l->state = LINK_RESET; + if (!snd_l->ackers) { + tipc_link_reset(snd_l); + snd_l->state = LINK_RESET; + __skb_queue_purge(xmitq); } +} - /* do all other link processing performed on a periodic basis */ - link_state_event(l_ptr, TIMEOUT_EVT); +int tipc_link_bc_peers(struct tipc_link *l) +{ + return l->ackers; +} - if (skb_queue_len(&l_ptr->backlogq)) - tipc_link_push_packets(l_ptr); +void tipc_link_set_mtu(struct tipc_link *l, int mtu) +{ + l->mtu = mtu; +} - tipc_node_unlock(l_ptr->owner); - tipc_link_put(l_ptr); +int tipc_link_mtu(struct tipc_link *l) +{ + return l->mtu; } -static void link_set_timer(struct tipc_link *link, unsigned long time) +static u32 link_own_addr(struct tipc_link *l) { - if (!mod_timer(&link->timer, jiffies + time)) - tipc_link_get(link); + return msg_prevnode(l->pmsg); } /** * tipc_link_create - create a new link - * @n_ptr: pointer to associated node - * @b_ptr: pointer to associated bearer - * @media_addr: media address to use when sending messages over link + * @n: pointer to associated node + * @if_name: associated interface name + * @bearer_id: id (index) of associated bearer + * @tolerance: link tolerance to be used by link + * @net_plane: network plane (A,B,c..) this link belongs to + * @mtu: mtu to be advertised by link + * @priority: priority to be used by link + * @window: send window to be used by link + * @session: session to be used by link + * @ownnode: identity of own node + * @peer: node id of peer node + * @peer_caps: bitmap describing peer node capabilities + * @bc_sndlink: the namespace global link used for broadcast sending + * @bc_rcvlink: the peer specific link used for broadcast reception + * @inputq: queue to put messages ready for delivery + * @namedq: queue to put binding table update messages ready for delivery + * @link: return value, pointer to put the created link * - * Returns pointer to link. + * Returns true if link was created, otherwise false */ -struct tipc_link *tipc_link_create(struct tipc_node *n_ptr, - struct tipc_bearer *b_ptr, - const struct tipc_media_addr *media_addr) +bool tipc_link_create(struct net *net, char *if_name, int bearer_id, + int tolerance, char net_plane, u32 mtu, int priority, + int window, u32 session, u32 ownnode, u32 peer, + u16 peer_caps, + struct tipc_link *bc_sndlink, + struct tipc_link *bc_rcvlink, + struct sk_buff_head *inputq, + struct sk_buff_head *namedq, + struct tipc_link **link) { - struct tipc_net *tn = net_generic(n_ptr->net, tipc_net_id); - struct tipc_link *l_ptr; - struct tipc_msg *msg; - char *if_name; - char addr_string[16]; - u32 peer = n_ptr->addr; - - if (n_ptr->link_cnt >= MAX_BEARERS) { - tipc_addr_string_fill(addr_string, n_ptr->addr); - pr_err("Attempt to establish %uth link to %s. Max %u allowed.\n", - n_ptr->link_cnt, addr_string, MAX_BEARERS); - return NULL; - } - - if (n_ptr->links[b_ptr->identity]) { - tipc_addr_string_fill(addr_string, n_ptr->addr); - pr_err("Attempt to establish second link on <%s> to %s\n", - b_ptr->name, addr_string); - return NULL; - } + struct tipc_link *l; + struct tipc_msg *hdr; - l_ptr = kzalloc(sizeof(*l_ptr), GFP_ATOMIC); - if (!l_ptr) { - pr_warn("Link creation failed, no memory\n"); - return NULL; - } - kref_init(&l_ptr->ref); - l_ptr->addr = peer; - if_name = strchr(b_ptr->name, ':') + 1; - sprintf(l_ptr->name, "%u.%u.%u:%s-%u.%u.%u:unknown", - tipc_zone(tn->own_addr), tipc_cluster(tn->own_addr), - tipc_node(tn->own_addr), - if_name, - tipc_zone(peer), tipc_cluster(peer), tipc_node(peer)); - /* note: peer i/f name is updated by reset/activate message */ - memcpy(&l_ptr->media_addr, media_addr, sizeof(*media_addr)); - l_ptr->owner = n_ptr; - l_ptr->checkpoint = 1; - l_ptr->peer_session = INVALID_SESSION; - l_ptr->bearer_id = b_ptr->identity; - link_set_supervision_props(l_ptr, b_ptr->tolerance); - l_ptr->state = RESET_UNKNOWN; - - l_ptr->pmsg = (struct tipc_msg *)&l_ptr->proto_msg; - msg = l_ptr->pmsg; - tipc_msg_init(tn->own_addr, msg, LINK_PROTOCOL, RESET_MSG, INT_H_SIZE, - l_ptr->addr); - msg_set_size(msg, sizeof(l_ptr->proto_msg)); - msg_set_session(msg, (tn->random & 0xffff)); - msg_set_bearer_id(msg, b_ptr->identity); - strcpy((char *)msg_data(msg), if_name); - l_ptr->net_plane = b_ptr->net_plane; - 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); - __skb_queue_head_init(&l_ptr->deferdq); - skb_queue_head_init(&l_ptr->wakeupq); - skb_queue_head_init(&l_ptr->inputq); - skb_queue_head_init(&l_ptr->namedq); - link_reset_statistics(l_ptr); - tipc_node_attach_link(n_ptr, l_ptr); - setup_timer(&l_ptr->timer, link_timeout, (unsigned long)l_ptr); - link_state_event(l_ptr, STARTING_EVT); - - return l_ptr; + l = kzalloc(sizeof(*l), GFP_ATOMIC); + if (!l) + return false; + *link = l; + l->pmsg = (struct tipc_msg *)&l->proto_msg; + hdr = l->pmsg; + tipc_msg_init(ownnode, hdr, LINK_PROTOCOL, RESET_MSG, INT_H_SIZE, peer); + msg_set_size(hdr, sizeof(l->proto_msg)); + msg_set_session(hdr, session); + msg_set_bearer_id(hdr, l->bearer_id); + + /* Note: peer i/f name is completed by reset/activate message */ + sprintf(l->name, "%u.%u.%u:%s-%u.%u.%u:unknown", + tipc_zone(ownnode), tipc_cluster(ownnode), tipc_node(ownnode), + if_name, tipc_zone(peer), tipc_cluster(peer), tipc_node(peer)); + strcpy((char *)msg_data(hdr), if_name); + + l->addr = peer; + l->peer_caps = peer_caps; + l->net = net; + l->peer_session = WILDCARD_SESSION; + l->bearer_id = bearer_id; + l->tolerance = tolerance; + l->net_plane = net_plane; + l->advertised_mtu = mtu; + l->mtu = mtu; + l->priority = priority; + tipc_link_set_queue_limits(l, window); + l->ackers = 1; + l->bc_sndlink = bc_sndlink; + l->bc_rcvlink = bc_rcvlink; + l->inputq = inputq; + l->namedq = namedq; + l->state = LINK_RESETTING; + __skb_queue_head_init(&l->transmq); + __skb_queue_head_init(&l->backlogq); + __skb_queue_head_init(&l->deferdq); + skb_queue_head_init(&l->wakeupq); + skb_queue_head_init(l->inputq); + return true; } /** - * tipc_link_delete - Delete a link - * @l: link to be deleted + * tipc_link_bc_create - create new link to be used for broadcast + * @n: pointer to associated node + * @mtu: mtu to be used + * @window: send window to be used + * @inputq: queue to put messages ready for delivery + * @namedq: queue to put binding table update messages ready for delivery + * @link: return value, pointer to put the created link + * + * Returns true if link was created, otherwise false */ -void tipc_link_delete(struct tipc_link *l) +bool tipc_link_bc_create(struct net *net, u32 ownnode, u32 peer, + int mtu, int window, u16 peer_caps, + struct sk_buff_head *inputq, + struct sk_buff_head *namedq, + struct tipc_link *bc_sndlink, + struct tipc_link **link) { + struct tipc_link *l; + + if (!tipc_link_create(net, "", MAX_BEARERS, 0, 'Z', mtu, 0, window, + 0, ownnode, peer, peer_caps, bc_sndlink, + NULL, inputq, namedq, link)) + return false; + + l = *link; + strcpy(l->name, tipc_bclink_name); 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); + l->state = LINK_RESET; + l->ackers = 0; + l->bc_rcvlink = l; + + /* Broadcast send link is always up */ + if (link_is_bc_sndlink(l)) + l->state = LINK_ESTABLISHED; + + return true; } -void tipc_link_delete_list(struct net *net, unsigned int bearer_id, - bool shutting_down) +/** + * tipc_link_fsm_evt - link finite state machine + * @l: pointer to link + * @evt: state machine event to be processed + */ +int tipc_link_fsm_evt(struct tipc_link *l, int evt) { - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_link *link; - struct tipc_node *node; + int rc = 0; - 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_link_delete(link); - tipc_node_unlock(node); + switch (l->state) { + case LINK_RESETTING: + switch (evt) { + case LINK_PEER_RESET_EVT: + l->state = LINK_PEER_RESET; + break; + case LINK_RESET_EVT: + l->state = LINK_RESET; + break; + case LINK_FAILURE_EVT: + case LINK_FAILOVER_BEGIN_EVT: + case LINK_ESTABLISH_EVT: + case LINK_FAILOVER_END_EVT: + case LINK_SYNCH_BEGIN_EVT: + case LINK_SYNCH_END_EVT: + default: + goto illegal_evt; + } + break; + case LINK_RESET: + switch (evt) { + case LINK_PEER_RESET_EVT: + l->state = LINK_ESTABLISHING; + break; + case LINK_FAILOVER_BEGIN_EVT: + l->state = LINK_FAILINGOVER; + case LINK_FAILURE_EVT: + case LINK_RESET_EVT: + case LINK_ESTABLISH_EVT: + case LINK_FAILOVER_END_EVT: + break; + case LINK_SYNCH_BEGIN_EVT: + case LINK_SYNCH_END_EVT: + default: + goto illegal_evt; + } + break; + case LINK_PEER_RESET: + switch (evt) { + case LINK_RESET_EVT: + l->state = LINK_ESTABLISHING; + break; + case LINK_PEER_RESET_EVT: + case LINK_ESTABLISH_EVT: + case LINK_FAILURE_EVT: + break; + case LINK_SYNCH_BEGIN_EVT: + case LINK_SYNCH_END_EVT: + case LINK_FAILOVER_BEGIN_EVT: + case LINK_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + case LINK_FAILINGOVER: + switch (evt) { + case LINK_FAILOVER_END_EVT: + l->state = LINK_RESET; + break; + case LINK_PEER_RESET_EVT: + case LINK_RESET_EVT: + case LINK_ESTABLISH_EVT: + case LINK_FAILURE_EVT: + break; + case LINK_FAILOVER_BEGIN_EVT: + case LINK_SYNCH_BEGIN_EVT: + case LINK_SYNCH_END_EVT: + default: + goto illegal_evt; + } + break; + case LINK_ESTABLISHING: + switch (evt) { + case LINK_ESTABLISH_EVT: + l->state = LINK_ESTABLISHED; + break; + case LINK_FAILOVER_BEGIN_EVT: + l->state = LINK_FAILINGOVER; + break; + case LINK_RESET_EVT: + l->state = LINK_RESET; + break; + case LINK_FAILURE_EVT: + case LINK_PEER_RESET_EVT: + case LINK_SYNCH_BEGIN_EVT: + case LINK_FAILOVER_END_EVT: + break; + case LINK_SYNCH_END_EVT: + default: + goto illegal_evt; + } + break; + case LINK_ESTABLISHED: + switch (evt) { + case LINK_PEER_RESET_EVT: + l->state = LINK_PEER_RESET; + rc |= TIPC_LINK_DOWN_EVT; + break; + case LINK_FAILURE_EVT: + l->state = LINK_RESETTING; + rc |= TIPC_LINK_DOWN_EVT; + break; + case LINK_RESET_EVT: + l->state = LINK_RESET; + break; + case LINK_ESTABLISH_EVT: + case LINK_SYNCH_END_EVT: + break; + case LINK_SYNCH_BEGIN_EVT: + l->state = LINK_SYNCHING; + break; + case LINK_FAILOVER_BEGIN_EVT: + case LINK_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + case LINK_SYNCHING: + switch (evt) { + case LINK_PEER_RESET_EVT: + l->state = LINK_PEER_RESET; + rc |= TIPC_LINK_DOWN_EVT; + break; + case LINK_FAILURE_EVT: + l->state = LINK_RESETTING; + rc |= TIPC_LINK_DOWN_EVT; + break; + case LINK_RESET_EVT: + l->state = LINK_RESET; + break; + case LINK_ESTABLISH_EVT: + case LINK_SYNCH_BEGIN_EVT: + break; + case LINK_SYNCH_END_EVT: + l->state = LINK_ESTABLISHED; + break; + case LINK_FAILOVER_BEGIN_EVT: + case LINK_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + default: + pr_err("Unknown FSM state %x in %s\n", l->state, l->name); } - rcu_read_unlock(); + return rc; +illegal_evt: + pr_err("Illegal FSM event %x in state %x on link %s\n", + evt, l->state, l->name); + return rc; +} + +/* link_profile_stats - update statistical profiling of traffic + */ +static void link_profile_stats(struct tipc_link *l) +{ + struct sk_buff *skb; + struct tipc_msg *msg; + int length; + + /* Update counters used in statistical profiling of send traffic */ + l->stats.accu_queue_sz += skb_queue_len(&l->transmq); + l->stats.queue_sz_counts++; + + skb = skb_peek(&l->transmq); + if (!skb) + return; + msg = buf_msg(skb); + length = msg_size(msg); + + if (msg_user(msg) == MSG_FRAGMENTER) { + if (msg_type(msg) != FIRST_FRAGMENT) + return; + length = msg_size(msg_get_wrapped(msg)); + } + l->stats.msg_lengths_total += length; + l->stats.msg_length_counts++; + if (length <= 64) + l->stats.msg_length_profile[0]++; + else if (length <= 256) + l->stats.msg_length_profile[1]++; + else if (length <= 1024) + l->stats.msg_length_profile[2]++; + else if (length <= 4096) + l->stats.msg_length_profile[3]++; + else if (length <= 16384) + l->stats.msg_length_profile[4]++; + else if (length <= 32768) + l->stats.msg_length_profile[5]++; + else + l->stats.msg_length_profile[6]++; +} + +/* tipc_link_timeout - perform periodic task as instructed from node timeout + */ +/* tipc_link_timeout - perform periodic task as instructed from node timeout + */ +int tipc_link_timeout(struct tipc_link *l, struct sk_buff_head *xmitq) +{ + int rc = 0; + int mtyp = STATE_MSG; + bool xmit = false; + bool prb = false; + u16 bc_snt = l->bc_sndlink->snd_nxt - 1; + u16 bc_acked = l->bc_rcvlink->acked; + bool bc_up = link_is_up(l->bc_rcvlink); + + link_profile_stats(l); + + switch (l->state) { + case LINK_ESTABLISHED: + case LINK_SYNCHING: + if (!l->silent_intv_cnt) { + if (bc_up && (bc_acked != bc_snt)) + xmit = true; + } else if (l->silent_intv_cnt <= l->abort_limit) { + xmit = true; + prb = true; + } else { + rc |= tipc_link_fsm_evt(l, LINK_FAILURE_EVT); + } + l->silent_intv_cnt++; + break; + case LINK_RESET: + xmit = true; + mtyp = RESET_MSG; + break; + case LINK_ESTABLISHING: + xmit = true; + mtyp = ACTIVATE_MSG; + break; + case LINK_PEER_RESET: + case LINK_RESETTING: + case LINK_FAILINGOVER: + break; + default: + break; + } + + if (xmit) + tipc_link_build_proto_msg(l, mtyp, prb, 0, 0, 0, xmitq); + + return rc; } /** @@ -334,7 +607,7 @@ void tipc_link_delete_list(struct net *net, unsigned int bearer_id, * @link: congested link * @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 + * Does not consume buffer list */ static int link_schedule_user(struct tipc_link *link, struct sk_buff_head *list) { @@ -347,8 +620,7 @@ static int link_schedule_user(struct tipc_link *link, struct sk_buff_head *list) /* 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; + return -ENOBUFS; } /* Non-blocking sender: */ if (TIPC_SKB_CB(skb_peek(list))->wakeup_pending) @@ -358,15 +630,12 @@ static int link_schedule_user(struct tipc_link *link, struct sk_buff_head *list) skb = tipc_msg_create(SOCK_WAKEUP, 0, INT_H_SIZE, 0, addr, addr, oport, 0, 0); if (!skb) - goto err; + return -ENOBUFS; 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 -ELINKCONG; -err: - __skb_queue_purge(list); - return -ENOBUFS; } /** @@ -388,782 +657,202 @@ void link_prepare_wakeup(struct tipc_link *l) if ((pnd[imp] + l->backlog[imp].len) >= lim) break; skb_unlink(skb, &l->wakeupq); - skb_queue_tail(&l->inputq, skb); - l->owner->inputq = &l->inputq; - l->owner->action_flags |= TIPC_MSG_EVT; + skb_queue_tail(l->inputq, skb); } } -/** - * tipc_link_reset_fragments - purge link's inbound message fragments queue - * @l_ptr: pointer to link - */ -void tipc_link_reset_fragments(struct tipc_link *l_ptr) +void tipc_link_reset(struct tipc_link *l) { - kfree_skb(l_ptr->reasm_buf); - l_ptr->reasm_buf = NULL; -} + /* Link is down, accept any session */ + l->peer_session = WILDCARD_SESSION; -static void tipc_link_purge_backlog(struct tipc_link *l) -{ + /* If peer is up, it only accepts an incremented session number */ + msg_set_session(l->pmsg, msg_session(l->pmsg) + 1); + + /* Prepare for renewed mtu size negotiation */ + l->mtu = l->advertised_mtu; + + /* Clean up all queues and counters: */ + __skb_queue_purge(&l->transmq); + __skb_queue_purge(&l->deferdq); + skb_queue_splice_init(&l->wakeupq, l->inputq); __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; + kfree_skb(l->reasm_buf); + kfree_skb(l->failover_reasm_skb); + l->reasm_buf = NULL; + l->failover_reasm_skb = NULL; + l->rcv_unacked = 0; + l->snd_nxt = 1; + l->rcv_nxt = 1; + l->acked = 0; + l->silent_intv_cnt = 0; + l->stats.recv_info = 0; + l->stale_count = 0; + l->bc_peer_is_up = false; + link_reset_statistics(l); } /** - * tipc_link_purge_queues - purge all pkt queues associated with link - * @l_ptr: pointer to link - */ -void tipc_link_purge_queues(struct tipc_link *l_ptr) -{ - __skb_queue_purge(&l_ptr->deferdq); - __skb_queue_purge(&l_ptr->transmq); - 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; - 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 renewed mtu size negotiation */ - l_ptr->mtu = l_ptr->advertised_mtu; - - l_ptr->state = RESET_UNKNOWN; - - if ((prev_state == RESET_UNKNOWN) || (prev_state == RESET_RESET)) - return; - - 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_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->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; - l_ptr->fsm_msg_cnt = 0; - l_ptr->stale_count = 0; - link_reset_statistics(l_ptr); -} - -void tipc_link_reset_list(struct net *net, unsigned int bearer_id) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct tipc_link *l_ptr; - struct tipc_node *n_ptr; - - rcu_read_lock(); - list_for_each_entry_rcu(n_ptr, &tn->node_list, list) { - tipc_node_lock(n_ptr); - l_ptr = n_ptr->links[bearer_id]; - if (l_ptr) - tipc_link_reset(l_ptr); - tipc_node_unlock(n_ptr); - } - rcu_read_unlock(); -} - -static void link_activate(struct tipc_link *link) -{ - struct tipc_node *node = link->owner; - - link->next_in_no = 1; - link->stats.recv_info = 1; - tipc_node_link_up(node, link); - tipc_bearer_add_dest(node->net, link->bearer_id, link->addr); -} - -/** - * link_state_event - link finite state machine - * @l_ptr: pointer to link - * @event: state machine event to process - */ -static void link_state_event(struct tipc_link *l_ptr, unsigned int event) -{ - struct tipc_link *other; - unsigned long cont_intv = l_ptr->cont_intv; - - if (l_ptr->flags & LINK_STOPPED) - return; - - if (!(l_ptr->flags & LINK_STARTED) && (event != STARTING_EVT)) - return; /* Not yet. */ - - if (l_ptr->flags & LINK_FAILINGOVER) { - if (event == TIMEOUT_EVT) - link_set_timer(l_ptr, cont_intv); - return; - } - - switch (l_ptr->state) { - case WORKING_WORKING: - switch (event) { - case TRAFFIC_MSG_EVT: - case ACTIVATE_MSG: - break; - case TIMEOUT_EVT: - if (l_ptr->next_in_no != l_ptr->checkpoint) { - 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); - l_ptr->fsm_msg_cnt++; - } - link_set_timer(l_ptr, cont_intv); - break; - } - l_ptr->state = WORKING_UNKNOWN; - l_ptr->fsm_msg_cnt = 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; - case RESET_MSG: - pr_debug("%s<%s>, requested by peer\n", - link_rst_msg, l_ptr->name); - tipc_link_reset(l_ptr); - l_ptr->state = RESET_RESET; - l_ptr->fsm_msg_cnt = 0; - tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG, - 0, 0, 0, 0); - l_ptr->fsm_msg_cnt++; - link_set_timer(l_ptr, cont_intv); - break; - default: - pr_debug("%s%u in WW state\n", link_unk_evt, event); - } - break; - case WORKING_UNKNOWN: - switch (event) { - case TRAFFIC_MSG_EVT: - case ACTIVATE_MSG: - l_ptr->state = WORKING_WORKING; - l_ptr->fsm_msg_cnt = 0; - link_set_timer(l_ptr, cont_intv); - break; - case RESET_MSG: - pr_debug("%s<%s>, requested by peer while probing\n", - link_rst_msg, l_ptr->name); - tipc_link_reset(l_ptr); - l_ptr->state = RESET_RESET; - l_ptr->fsm_msg_cnt = 0; - tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG, - 0, 0, 0, 0); - l_ptr->fsm_msg_cnt++; - link_set_timer(l_ptr, cont_intv); - break; - case TIMEOUT_EVT: - if (l_ptr->next_in_no != l_ptr->checkpoint) { - l_ptr->state = WORKING_WORKING; - l_ptr->fsm_msg_cnt = 0; - 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); - 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); - l_ptr->fsm_msg_cnt++; - link_set_timer(l_ptr, cont_intv / 4); - } else { /* Link has failed */ - pr_debug("%s<%s>, peer not responding\n", - link_rst_msg, l_ptr->name); - tipc_link_reset(l_ptr); - l_ptr->state = RESET_UNKNOWN; - l_ptr->fsm_msg_cnt = 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; - default: - pr_err("%s%u in WU state\n", link_unk_evt, event); - } - break; - case RESET_UNKNOWN: - switch (event) { - case TRAFFIC_MSG_EVT: - break; - case ACTIVATE_MSG: - other = l_ptr->owner->active_links[0]; - if (other && link_working_unknown(other)) - break; - 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); - l_ptr->fsm_msg_cnt++; - if (l_ptr->owner->working_links == 1) - tipc_link_sync_xmit(l_ptr); - link_set_timer(l_ptr, cont_intv); - break; - case RESET_MSG: - l_ptr->state = RESET_RESET; - l_ptr->fsm_msg_cnt = 0; - tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG, - 1, 0, 0, 0); - l_ptr->fsm_msg_cnt++; - link_set_timer(l_ptr, cont_intv); - break; - case STARTING_EVT: - l_ptr->flags |= LINK_STARTED; - l_ptr->fsm_msg_cnt++; - link_set_timer(l_ptr, cont_intv); - break; - case TIMEOUT_EVT: - 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; - default: - pr_err("%s%u in RU state\n", link_unk_evt, event); - } - break; - case RESET_RESET: - switch (event) { - case TRAFFIC_MSG_EVT: - case ACTIVATE_MSG: - other = l_ptr->owner->active_links[0]; - if (other && link_working_unknown(other)) - break; - 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); - l_ptr->fsm_msg_cnt++; - if (l_ptr->owner->working_links == 1) - tipc_link_sync_xmit(l_ptr); - link_set_timer(l_ptr, cont_intv); - break; - case RESET_MSG: - break; - case TIMEOUT_EVT: - tipc_link_proto_xmit(l_ptr, ACTIVATE_MSG, - 0, 0, 0, 0); - l_ptr->fsm_msg_cnt++; - link_set_timer(l_ptr, cont_intv); - break; - default: - pr_err("%s%u in RR state\n", link_unk_evt, event); - } - break; - default: - pr_err("Unknown link state %u/%u\n", l_ptr->state, event); - } -} - -/** - * __tipc_link_xmit(): same as tipc_link_xmit, but destlink is known & locked + * tipc_link_xmit(): enqueue buffer list according to queue situation * @link: link to use * @list: chain of buffers containing message + * @xmitq: returned list of packets to be sent by caller * * 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) +int tipc_link_xmit(struct tipc_link *l, struct sk_buff_head *list, + struct sk_buff_head *xmitq) { - struct tipc_msg *msg = buf_msg(skb_peek(list)); - unsigned int maxwin = link->window; - unsigned int imp = msg_importance(msg); - 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; - struct tipc_media_addr *addr = &link->media_addr; - struct sk_buff_head *transmq = &link->transmq; - struct sk_buff_head *backlogq = &link->backlogq; - struct sk_buff *skb, *tmp; - - /* Match backlog limit against msg importance: */ - if (unlikely(link->backlog[imp].len >= link->backlog[imp].limit)) - return link_schedule_user(link, list); - - if (unlikely(msg_size(msg) > mtu)) { - __skb_queue_purge(list); - return -EMSGSIZE; + struct tipc_msg *hdr = buf_msg(skb_peek(list)); + unsigned int maxwin = l->window; + unsigned int i, imp = msg_importance(hdr); + unsigned int mtu = l->mtu; + u16 ack = l->rcv_nxt - 1; + u16 seqno = l->snd_nxt; + u16 bc_ack = l->bc_rcvlink->rcv_nxt - 1; + struct sk_buff_head *transmq = &l->transmq; + struct sk_buff_head *backlogq = &l->backlogq; + struct sk_buff *skb, *_skb, *bskb; + + /* Match msg importance against this and all higher backlog limits: */ + for (i = imp; i <= TIPC_SYSTEM_IMPORTANCE; i++) { + if (unlikely(l->backlog[i].len >= l->backlog[i].limit)) + return link_schedule_user(l, list); } + if (unlikely(msg_size(hdr) > mtu)) + return -EMSGSIZE; + /* Prepare each packet for sending, and add to relevant queue: */ - skb_queue_walk_safe(list, skb, tmp) { - __skb_unlink(skb, list); - msg = buf_msg(skb); - msg_set_seqno(msg, seqno); - msg_set_ack(msg, ack); - msg_set_bcast_ack(msg, bc_last_in); + while (skb_queue_len(list)) { + skb = skb_peek(list); + hdr = buf_msg(skb); + msg_set_seqno(hdr, seqno); + msg_set_ack(hdr, ack); + msg_set_bcast_ack(hdr, bc_ack); if (likely(skb_queue_len(transmq) < maxwin)) { + _skb = skb_clone(skb, GFP_ATOMIC); + if (!_skb) + return -ENOBUFS; + __skb_dequeue(list); __skb_queue_tail(transmq, skb); - tipc_bearer_send(net, link->bearer_id, skb, addr); - link->rcv_unacked = 0; + __skb_queue_tail(xmitq, _skb); + TIPC_SKB_CB(skb)->ackers = l->ackers; + l->rcv_unacked = 0; seqno++; continue; } - if (tipc_msg_bundle(skb_peek_tail(backlogq), skb, mtu)) { - link->stats.sent_bundled++; + if (tipc_msg_bundle(skb_peek_tail(backlogq), hdr, mtu)) { + kfree_skb(__skb_dequeue(list)); + l->stats.sent_bundled++; continue; } - if (tipc_msg_make_bundle(&skb, mtu, link->addr)) { - link->stats.sent_bundled++; - link->stats.sent_bundles++; - imp = msg_importance(buf_msg(skb)); + if (tipc_msg_make_bundle(&bskb, hdr, mtu, l->addr)) { + kfree_skb(__skb_dequeue(list)); + __skb_queue_tail(backlogq, bskb); + l->backlog[msg_importance(buf_msg(bskb))].len++; + l->stats.sent_bundled++; + l->stats.sent_bundles++; + continue; } - __skb_queue_tail(backlogq, skb); - link->backlog[imp].len++; - seqno++; + l->backlog[imp].len += skb_queue_len(list); + skb_queue_splice_tail_init(list, backlogq); } - link->next_out_no = seqno; + l->snd_nxt = seqno; return 0; } -static void skb2list(struct sk_buff *skb, struct sk_buff_head *list) -{ - skb_queue_head_init(list); - __skb_queue_tail(list, skb); -} - -static int __tipc_link_xmit_skb(struct tipc_link *link, struct sk_buff *skb) +void tipc_link_advance_backlog(struct tipc_link *l, struct sk_buff_head *xmitq) { - struct sk_buff_head head; - - skb2list(skb, &head); - 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); - rc = tipc_link_xmit(net, &head, dnode, selector); - if (rc == -ELINKCONG) - kfree_skb(skb); - return 0; -} - -/** - * tipc_link_xmit() is the general link level function for message sending - * @net: the applicable net namespace - * @list: chain of buffers containing message - * @dsz: amount of user data to be sent - * @dnode: address of destination node - * @selector: a number used for deterministic link selection - * Consumes the buffer chain, except when returning -ELINKCONG - * Returns 0 if success, otherwise errno: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE - */ -int tipc_link_xmit(struct net *net, struct sk_buff_head *list, u32 dnode, - u32 selector) -{ - struct tipc_link *link = NULL; - struct tipc_node *node; - int rc = -EHOSTUNREACH; - - node = tipc_node_find(net, dnode); - if (node) { - tipc_node_lock(node); - link = node->active_links[selector & 1]; - 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))) { - tipc_sk_rcv(net, list); - return 0; - } - - __skb_queue_purge(list); - return rc; -} - -/* - * tipc_link_sync_xmit - synchronize broadcast link endpoints. - * - * Give a newly added peer node the sequence number where it should - * start receiving and acking broadcast packets. - * - * Called with node locked - */ -static void tipc_link_sync_xmit(struct tipc_link *link) -{ - struct sk_buff *skb; - struct tipc_msg *msg; - - skb = tipc_buf_acquire(INT_H_SIZE); - if (!skb) - return; - - msg = buf_msg(skb); - tipc_msg_init(link_own_addr(link), msg, BCAST_PROTOCOL, STATE_MSG, - INT_H_SIZE, link->addr); - msg_set_last_bcast(msg, link->owner->bclink.acked); - __tipc_link_xmit_skb(link, skb); -} - -/* - * tipc_link_sync_rcv - synchronize broadcast link endpoints. - * Receive the sequence number where we should start receiving and - * acking broadcast packets from a newly added peer node, and open - * up for reception of such packets. - * - * Called with node locked - */ -static void tipc_link_sync_rcv(struct tipc_node *n, struct sk_buff *buf) -{ - struct tipc_msg *msg = buf_msg(buf); - - n->bclink.last_sent = n->bclink.last_in = msg_last_bcast(msg); - n->bclink.recv_permitted = true; - kfree_skb(buf); -} - -/* - * tipc_link_push_packets - push unsent packets to bearer - * - * Push out the unsent messages of a link where congestion - * has abated. Node is locked. - * - * Called with node locked - */ -void tipc_link_push_packets(struct tipc_link *link) -{ - struct sk_buff *skb; - struct tipc_msg *msg; - unsigned int ack = mod(link->next_in_no - 1); - - while (skb_queue_len(&link->transmq) < link->window) { - skb = __skb_dequeue(&link->backlogq); + struct sk_buff *skb, *_skb; + struct tipc_msg *hdr; + u16 seqno = l->snd_nxt; + u16 ack = l->rcv_nxt - 1; + u16 bc_ack = l->bc_rcvlink->rcv_nxt - 1; + + while (skb_queue_len(&l->transmq) < l->window) { + skb = skb_peek(&l->backlogq); 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; - __skb_queue_tail(&link->transmq, skb); - tipc_bearer_send(link->owner->net, link->bearer_id, - skb, &link->media_addr); - } -} - -void tipc_link_reset_all(struct tipc_node *node) -{ - char addr_string[16]; - u32 i; - - tipc_node_lock(node); - - pr_warn("Resetting all links to %s\n", - tipc_addr_string_fill(addr_string, node->addr)); - - for (i = 0; i < MAX_BEARERS; i++) { - if (node->links[i]) { - link_print(node->links[i], "Resetting link\n"); - tipc_link_reset(node->links[i]); - } + _skb = skb_clone(skb, GFP_ATOMIC); + if (!_skb) + break; + __skb_dequeue(&l->backlogq); + hdr = buf_msg(skb); + l->backlog[msg_importance(hdr)].len--; + __skb_queue_tail(&l->transmq, skb); + __skb_queue_tail(xmitq, _skb); + TIPC_SKB_CB(skb)->ackers = l->ackers; + msg_set_seqno(hdr, seqno); + msg_set_ack(hdr, ack); + msg_set_bcast_ack(hdr, bc_ack); + l->rcv_unacked = 0; + seqno++; } - - tipc_node_unlock(node); + l->snd_nxt = seqno; } -static void link_retransmit_failure(struct tipc_link *l_ptr, - struct sk_buff *buf) +static void link_retransmit_failure(struct tipc_link *l, struct sk_buff *skb) { - struct tipc_msg *msg = buf_msg(buf); - struct net *net = l_ptr->owner->net; - - pr_warn("Retransmission failure on link <%s>\n", l_ptr->name); - - if (l_ptr->addr) { - /* Handle failure on standard link */ - link_print(l_ptr, "Resetting link\n"); - tipc_link_reset(l_ptr); - - } else { - /* Handle failure on broadcast link */ - struct tipc_node *n_ptr; - char addr_string[16]; - - pr_info("Msg seq number: %u, ", msg_seqno(msg)); - pr_cont("Outstanding acks: %lu\n", - (unsigned long) TIPC_SKB_CB(buf)->handle); - - n_ptr = tipc_bclink_retransmit_to(net); - - tipc_addr_string_fill(addr_string, n_ptr->addr); - pr_info("Broadcast link info for %s\n", addr_string); - pr_info("Reception permitted: %d, Acked: %u\n", - n_ptr->bclink.recv_permitted, - n_ptr->bclink.acked); - pr_info("Last in: %u, Oos state: %u, Last sent: %u\n", - n_ptr->bclink.last_in, - n_ptr->bclink.oos_state, - n_ptr->bclink.last_sent); - - n_ptr->action_flags |= TIPC_BCAST_RESET; - l_ptr->stale_count = 0; - } + struct tipc_msg *hdr = buf_msg(skb); + + pr_warn("Retransmission failure on link <%s>\n", l->name); + link_print(l, "Resetting link "); + pr_info("Failed msg: usr %u, typ %u, len %u, err %u\n", + msg_user(hdr), msg_type(hdr), msg_size(hdr), msg_errcode(hdr)); + pr_info("sqno %u, prev: %x, src: %x\n", + msg_seqno(hdr), msg_prevnode(hdr), msg_orignode(hdr)); } -void tipc_link_retransmit(struct tipc_link *l_ptr, struct sk_buff *skb, - u32 retransmits) +int tipc_link_retrans(struct tipc_link *l, u16 from, u16 to, + struct sk_buff_head *xmitq) { - struct tipc_msg *msg; + struct sk_buff *_skb, *skb = skb_peek(&l->transmq); + struct tipc_msg *hdr; + u16 ack = l->rcv_nxt - 1; + u16 bc_ack = l->bc_rcvlink->rcv_nxt - 1; if (!skb) - return; - - msg = buf_msg(skb); + return 0; - /* Detect repeated retransmit failures */ - if (l_ptr->last_retransmitted == msg_seqno(msg)) { - if (++l_ptr->stale_count > 100) { - link_retransmit_failure(l_ptr, skb); - return; - } - } else { - l_ptr->last_retransmitted = msg_seqno(msg); - l_ptr->stale_count = 1; + /* Detect repeated retransmit failures on same packet */ + if (likely(l->last_retransm != buf_seqno(skb))) { + l->last_retransm = buf_seqno(skb); + l->stale_count = 1; + } else if (++l->stale_count > 100) { + link_retransmit_failure(l, skb); + return tipc_link_fsm_evt(l, LINK_FAILURE_EVT); } - skb_queue_walk_from(&l_ptr->transmq, skb) { - if (!retransmits) + /* Move forward to where retransmission should start */ + skb_queue_walk(&l->transmq, skb) { + if (!less(buf_seqno(skb), from)) break; - msg = buf_msg(skb); - msg_set_ack(msg, mod(l_ptr->next_in_no - 1)); - msg_set_bcast_ack(msg, l_ptr->owner->bclink.last_in); - tipc_bearer_send(l_ptr->owner->net, l_ptr->bearer_id, skb, - &l_ptr->media_addr); - retransmits--; - l_ptr->stats.retransmitted++; } -} - -/* 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) -{ - u32 seq_no; - if (skb_queue_empty(&link->deferdq)) - return; - - seq_no = buf_seqno(skb_peek(&link->deferdq)); - if (seq_no == mod(link->next_in_no)) - skb_queue_splice_tail_init(&link->deferdq, list); -} - -/** - * tipc_rcv - process TIPC packets/messages arriving from off-node - * @net: the applicable net namespace - * @skb: TIPC packet - * @b_ptr: pointer to bearer message arrived on - * - * Invoked with no locks held. Bearer pointer must point to a valid bearer - * structure (i.e. cannot be NULL), but bearer can be inactive. - */ -void tipc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *b_ptr) -{ - struct tipc_net *tn = net_generic(net, tipc_net_id); - struct sk_buff_head head; - struct tipc_node *n_ptr; - struct tipc_link *l_ptr; - struct sk_buff *skb1, *tmp; - struct tipc_msg *msg; - u32 seq_no; - u32 ackd; - u32 released; - - skb2list(skb, &head); - - while ((skb = __skb_dequeue(&head))) { - /* Ensure message is well-formed */ - if (unlikely(!tipc_msg_validate(skb))) - goto discard; - - /* Handle arrival of a non-unicast link message */ - msg = buf_msg(skb); - if (unlikely(msg_non_seq(msg))) { - if (msg_user(msg) == LINK_CONFIG) - tipc_disc_rcv(net, skb, b_ptr); - else - tipc_bclink_rcv(net, skb); - continue; - } - - /* Discard unicast link messages destined for another node */ - if (unlikely(!msg_short(msg) && - (msg_destnode(msg) != tn->own_addr))) - goto discard; - - /* Locate neighboring node that sent message */ - n_ptr = tipc_node_find(net, msg_prevnode(msg)); - if (unlikely(!n_ptr)) - goto discard; - - 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)) - goto unlock; - - /* Verify that communication with node is currently allowed */ - if ((n_ptr->action_flags & TIPC_WAIT_PEER_LINKS_DOWN) && - msg_user(msg) == LINK_PROTOCOL && - (msg_type(msg) == RESET_MSG || - msg_type(msg) == ACTIVATE_MSG) && - !msg_redundant_link(msg)) - n_ptr->action_flags &= ~TIPC_WAIT_PEER_LINKS_DOWN; - - if (tipc_node_blocked(n_ptr)) - goto unlock; - - /* Validate message sequence number info */ - seq_no = msg_seqno(msg); - ackd = msg_ack(msg); - - /* Release acked messages */ - if (unlikely(n_ptr->bclink.acked != msg_bcast_ack(msg))) - tipc_bclink_acknowledge(n_ptr, msg_bcast_ack(msg)); - - released = 0; - skb_queue_walk_safe(&l_ptr->transmq, skb1, tmp) { - if (more(buf_seqno(skb1), ackd)) - break; - __skb_unlink(skb1, &l_ptr->transmq); - kfree_skb(skb1); - released = 1; - } - - /* Try sending any messages link endpoint has pending */ - if (unlikely(skb_queue_len(&l_ptr->backlogq))) - tipc_link_push_packets(l_ptr); - - if (released && !skb_queue_empty(&l_ptr->wakeupq)) - link_prepare_wakeup(l_ptr); - - /* Process the incoming packet */ - if (unlikely(!link_working_working(l_ptr))) { - if (msg_user(msg) == LINK_PROTOCOL) { - tipc_link_proto_rcv(l_ptr, skb); - link_retrieve_defq(l_ptr, &head); - skb = NULL; - goto unlock; - } - - /* Traffic message. Conditionally activate link */ - link_state_event(l_ptr, TRAFFIC_MSG_EVT); - - if (link_working_working(l_ptr)) { - /* Re-insert buffer in front of queue */ - __skb_queue_head(&head, skb); - skb = NULL; - goto unlock; - } - goto unlock; - } - - /* Link is now in state WORKING_WORKING */ - if (unlikely(seq_no != mod(l_ptr->next_in_no))) { - link_handle_out_of_seq_msg(l_ptr, skb); - link_retrieve_defq(l_ptr, &head); - skb = NULL; - goto unlock; - } - /* Synchronize with parallel link if applicable */ - if (unlikely((l_ptr->flags & LINK_SYNCHING) && !msg_dup(msg))) { - if (!link_synch(l_ptr)) - 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); - } - 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); + skb_queue_walk_from(&l->transmq, skb) { + if (more(buf_seqno(skb), to)) + break; + hdr = buf_msg(skb); + _skb = __pskb_copy(skb, MIN_H_SIZE, GFP_ATOMIC); + if (!_skb) + return 0; + hdr = buf_msg(_skb); + msg_set_ack(hdr, ack); + msg_set_bcast_ack(hdr, bc_ack); + _skb->priority = TC_PRIO_CONTROL; + __skb_queue_tail(xmitq, _skb); + l->stats.retransmitted++; } + return 0; } /* tipc_data_input - deliver data and name distr msgs to upper layer @@ -1171,29 +860,20 @@ discard: * Consumes buffer if message is of right type * Node lock must be held */ -static bool tipc_data_input(struct tipc_link *link, struct sk_buff *skb) +static bool tipc_data_input(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *inputq) { - struct tipc_node *node = link->owner; - struct tipc_msg *msg = buf_msg(skb); - u32 dport = msg_destport(msg); - - switch (msg_user(msg)) { + switch (msg_user(buf_msg(skb))) { case TIPC_LOW_IMPORTANCE: case TIPC_MEDIUM_IMPORTANCE: case TIPC_HIGH_IMPORTANCE: case TIPC_CRITICAL_IMPORTANCE: case CONN_MANAGER: - if (tipc_skb_queue_tail(&link->inputq, skb, dport)) { - node->inputq = &link->inputq; - node->action_flags |= TIPC_MSG_EVT; - } + skb_queue_tail(inputq, skb); return true; case NAME_DISTRIBUTOR: - node->bclink.recv_permitted = true; - node->namedq = &link->namedq; - skb_queue_tail(&link->namedq, skb); - if (skb_queue_len(&link->namedq) == 1) - node->action_flags |= TIPC_NAMED_MSG_EVT; + l->bc_rcvlink->state = LINK_ESTABLISHED; + skb_queue_tail(l->namedq, skb); return true; case MSG_BUNDLER: case TUNNEL_PROTOCOL: @@ -1210,540 +890,629 @@ static bool tipc_data_input(struct tipc_link *link, struct sk_buff *skb) /* tipc_link_input - process packet that has passed link protocol check * * Consumes buffer - * Node lock must be held */ -static void tipc_link_input(struct tipc_link *link, struct sk_buff *skb) +static int tipc_link_input(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *inputq) { - struct tipc_node *node = link->owner; - struct tipc_msg *msg = buf_msg(skb); + struct tipc_msg *hdr = buf_msg(skb); + struct sk_buff **reasm_skb = &l->reasm_buf; struct sk_buff *iskb; + struct sk_buff_head tmpq; + int usr = msg_user(hdr); + int rc = 0; int pos = 0; + int ipos = 0; - if (likely(tipc_data_input(link, skb))) - return; - - switch (msg_user(msg)) { - 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); - break; + if (unlikely(usr == TUNNEL_PROTOCOL)) { + if (msg_type(hdr) == SYNCH_MSG) { + __skb_queue_purge(&l->deferdq); + goto drop; } - case MSG_BUNDLER: - link->stats.recv_bundles++; - link->stats.recv_bundled += msg_msgcnt(msg); + if (!tipc_msg_extract(skb, &iskb, &ipos)) + return rc; + kfree_skb(skb); + skb = iskb; + hdr = buf_msg(skb); + if (less(msg_seqno(hdr), l->drop_point)) + goto drop; + if (tipc_data_input(l, skb, inputq)) + return rc; + usr = msg_user(hdr); + reasm_skb = &l->failover_reasm_skb; + } + if (usr == MSG_BUNDLER) { + skb_queue_head_init(&tmpq); + l->stats.recv_bundles++; + l->stats.recv_bundled += msg_msgcnt(hdr); while (tipc_msg_extract(skb, &iskb, &pos)) - tipc_data_input(link, iskb); - break; - case MSG_FRAGMENTER: - link->stats.recv_fragments++; - if (tipc_buf_append(&link->reasm_buf, &skb)) { - link->stats.recv_fragmented++; - tipc_data_input(link, skb); - } else if (!link->reasm_buf) { - tipc_link_reset(link); + tipc_data_input(l, iskb, &tmpq); + tipc_skb_queue_splice_tail(&tmpq, inputq); + return 0; + } else if (usr == MSG_FRAGMENTER) { + l->stats.recv_fragments++; + if (tipc_buf_append(reasm_skb, &skb)) { + l->stats.recv_fragmented++; + tipc_data_input(l, skb, inputq); + } else if (!*reasm_skb && !link_is_bc_rcvlink(l)) { + pr_warn_ratelimited("Unable to build fragment list\n"); + return tipc_link_fsm_evt(l, LINK_FAILURE_EVT); } - break; - case BCAST_PROTOCOL: - tipc_link_sync_rcv(node, skb); - break; - default: - break; - }; + return 0; + } else if (usr == BCAST_PROTOCOL) { + tipc_bcast_lock(l->net); + tipc_link_bc_init_rcv(l->bc_rcvlink, hdr); + tipc_bcast_unlock(l->net); + } +drop: + kfree_skb(skb); + return 0; } -/** - * tipc_link_defer_pkt - Add out-of-sequence message to deferred reception queue - * - * Returns increase in queue length (i.e. 0 or 1) - */ -u32 tipc_link_defer_pkt(struct sk_buff_head *list, struct sk_buff *skb) +static bool tipc_link_release_pkts(struct tipc_link *l, u16 acked) { - struct sk_buff *skb1; - u32 seq_no = buf_seqno(skb); + bool released = false; + struct sk_buff *skb, *tmp; - /* Empty queue ? */ - if (skb_queue_empty(list)) { - __skb_queue_tail(list, skb); - return 1; + skb_queue_walk_safe(&l->transmq, skb, tmp) { + if (more(buf_seqno(skb), acked)) + break; + __skb_unlink(skb, &l->transmq); + kfree_skb(skb); + released = true; } + return released; +} - /* Last ? */ - if (less(buf_seqno(skb_peek_tail(list)), seq_no)) { - __skb_queue_tail(list, skb); - return 1; +/* tipc_link_build_ack_msg: prepare link acknowledge message for transmission + * + * Note that sending of broadcast ack is coordinated among nodes, to reduce + * risk of ack storms towards the sender + */ +int tipc_link_build_ack_msg(struct tipc_link *l, struct sk_buff_head *xmitq) +{ + if (!l) + return 0; + + /* Broadcast ACK must be sent via a unicast link => defer to caller */ + if (link_is_bc_rcvlink(l)) { + if (((l->rcv_nxt ^ link_own_addr(l)) & 0xf) != 0xf) + return 0; + l->rcv_unacked = 0; + return TIPC_LINK_SND_BC_ACK; } - /* Locate insertion point in queue, then insert; discard if duplicate */ - skb_queue_walk(list, skb1) { - u32 curr_seqno = buf_seqno(skb1); + /* Unicast ACK */ + l->rcv_unacked = 0; + l->stats.sent_acks++; + tipc_link_build_proto_msg(l, STATE_MSG, 0, 0, 0, 0, xmitq); + return 0; +} - if (seq_no == curr_seqno) { - kfree_skb(skb); - return 0; - } +/* tipc_link_build_reset_msg: prepare link RESET or ACTIVATE message + */ +void tipc_link_build_reset_msg(struct tipc_link *l, struct sk_buff_head *xmitq) +{ + int mtyp = RESET_MSG; - if (less(seq_no, curr_seqno)) - break; - } + if (l->state == LINK_ESTABLISHING) + mtyp = ACTIVATE_MSG; - __skb_queue_before(list, skb1, skb); - return 1; + tipc_link_build_proto_msg(l, mtyp, 0, 0, 0, 0, xmitq); } -/* - * link_handle_out_of_seq_msg - handle arrival of out-of-sequence packet +/* tipc_link_build_nack_msg: prepare link nack message for transmission */ -static void link_handle_out_of_seq_msg(struct tipc_link *l_ptr, - struct sk_buff *buf) +static void tipc_link_build_nack_msg(struct tipc_link *l, + struct sk_buff_head *xmitq) { - u32 seq_no = buf_seqno(buf); + u32 def_cnt = ++l->stats.deferred_recv; - if (likely(msg_user(buf_msg(buf)) == LINK_PROTOCOL)) { - tipc_link_proto_rcv(l_ptr, buf); + if (link_is_bc_rcvlink(l)) return; - } - /* Record OOS packet arrival (force mismatch on next timeout) */ - l_ptr->checkpoint--; + if ((skb_queue_len(&l->deferdq) == 1) || !(def_cnt % TIPC_NACK_INTV)) + tipc_link_build_proto_msg(l, STATE_MSG, 0, 0, 0, 0, xmitq); +} - /* - * Discard packet if a duplicate; otherwise add it to deferred queue - * and notify peer of gap as per protocol specification - */ - if (less(seq_no, mod(l_ptr->next_in_no))) { - l_ptr->stats.duplicates++; - kfree_skb(buf); - return; - } +/* tipc_link_rcv - process TIPC packets/messages arriving from off-node + * @l: the link that should handle the message + * @skb: TIPC packet + * @xmitq: queue to place packets to be sent after this call + */ +int tipc_link_rcv(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *xmitq) +{ + struct sk_buff_head *defq = &l->deferdq; + struct tipc_msg *hdr; + u16 seqno, rcv_nxt, win_lim; + int rc = 0; + + do { + hdr = buf_msg(skb); + seqno = msg_seqno(hdr); + rcv_nxt = l->rcv_nxt; + win_lim = rcv_nxt + TIPC_MAX_LINK_WIN; + + /* Verify and update link state */ + if (unlikely(msg_user(hdr) == LINK_PROTOCOL)) + return tipc_link_proto_rcv(l, skb, xmitq); + + if (unlikely(!link_is_up(l))) { + if (l->state == LINK_ESTABLISHING) + rc = TIPC_LINK_UP_EVT; + goto drop; + } - 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); - } else { - l_ptr->stats.duplicates++; - } + /* Don't send probe at next timeout expiration */ + l->silent_intv_cnt = 0; + + /* Drop if outside receive window */ + if (unlikely(less(seqno, rcv_nxt) || more(seqno, win_lim))) { + l->stats.duplicates++; + goto drop; + } + + /* Forward queues and wake up waiting users */ + if (likely(tipc_link_release_pkts(l, msg_ack(hdr)))) { + tipc_link_advance_backlog(l, xmitq); + if (unlikely(!skb_queue_empty(&l->wakeupq))) + link_prepare_wakeup(l); + } + + /* Defer delivery if sequence gap */ + if (unlikely(seqno != rcv_nxt)) { + __tipc_skb_queue_sorted(defq, seqno, skb); + tipc_link_build_nack_msg(l, xmitq); + break; + } + + /* Deliver packet */ + l->rcv_nxt++; + l->stats.recv_info++; + if (!tipc_data_input(l, skb, l->inputq)) + rc |= tipc_link_input(l, skb, l->inputq); + if (unlikely(++l->rcv_unacked >= TIPC_MIN_LINK_WIN)) + rc |= tipc_link_build_ack_msg(l, xmitq); + if (unlikely(rc & ~TIPC_LINK_SND_BC_ACK)) + break; + } while ((skb = __skb_dequeue(defq))); + + return rc; +drop: + kfree_skb(skb); + return rc; } /* * Send protocol message to the other endpoint. */ -void tipc_link_proto_xmit(struct tipc_link *l_ptr, u32 msg_typ, int probe_msg, +void tipc_link_proto_xmit(struct tipc_link *l, u32 msg_typ, int probe_msg, 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 failover */ - if (l_ptr->flags & LINK_FAILINGOVER) - return; + struct sk_buff *skb = NULL; + struct sk_buff_head xmitq; - /* Abort non-RESET send if communication with node is prohibited */ - if ((tipc_node_blocked(l_ptr->owner)) && (msg_typ != RESET_MSG)) + __skb_queue_head_init(&xmitq); + tipc_link_build_proto_msg(l, msg_typ, probe_msg, gap, + tolerance, priority, &xmitq); + skb = __skb_dequeue(&xmitq); + if (!skb) return; + tipc_bearer_xmit_skb(l->net, l->bearer_id, skb, l->media_addr); + l->rcv_unacked = 0; +} - /* Create protocol message with "out-of-sequence" sequence number */ - msg_set_type(msg, msg_typ); - msg_set_net_plane(msg, l_ptr->net_plane); - msg_set_bcast_ack(msg, l_ptr->owner->bclink.last_in); - msg_set_last_bcast(msg, tipc_bclink_get_last_sent(l_ptr->owner->net)); +static void tipc_link_build_proto_msg(struct tipc_link *l, int mtyp, bool probe, + u16 rcvgap, int tolerance, int priority, + struct sk_buff_head *xmitq) +{ + struct sk_buff *skb = NULL; + struct tipc_msg *hdr = l->pmsg; + bool node_up = link_is_up(l->bc_rcvlink); - if (msg_typ == STATE_MSG) { - u32 next_sent = mod(l_ptr->next_out_no); + /* Don't send protocol message during reset or link failover */ + if (tipc_link_is_blocked(l)) + return; - if (!tipc_link_is_up(l_ptr)) + msg_set_type(hdr, mtyp); + msg_set_net_plane(hdr, l->net_plane); + msg_set_next_sent(hdr, l->snd_nxt); + msg_set_ack(hdr, l->rcv_nxt - 1); + msg_set_bcast_ack(hdr, l->bc_rcvlink->rcv_nxt - 1); + msg_set_last_bcast(hdr, l->bc_sndlink->snd_nxt - 1); + msg_set_link_tolerance(hdr, tolerance); + msg_set_linkprio(hdr, priority); + msg_set_redundant_link(hdr, node_up); + msg_set_seq_gap(hdr, 0); + + /* Compatibility: created msg must not be in sequence with pkt flow */ + msg_set_seqno(hdr, l->snd_nxt + U16_MAX / 2); + + if (mtyp == STATE_MSG) { + if (!tipc_link_is_up(l)) return; - if (skb_queue_len(&l_ptr->backlogq)) - next_sent = buf_seqno(skb_peek(&l_ptr->backlogq)); - msg_set_next_sent(msg, next_sent); - if (!skb_queue_empty(&l_ptr->deferdq)) { - u32 rec = buf_seqno(skb_peek(&l_ptr->deferdq)); - gap = mod(rec - mod(l_ptr->next_in_no)); + + /* Override rcvgap if there are packets in deferred queue */ + if (!skb_queue_empty(&l->deferdq)) + rcvgap = buf_seqno(skb_peek(&l->deferdq)) - l->rcv_nxt; + if (rcvgap) { + msg_set_seq_gap(hdr, rcvgap); + l->stats.sent_nacks++; } - msg_set_seq_gap(msg, gap); - if (gap) - l_ptr->stats.sent_nacks++; - msg_set_link_tolerance(msg, tolerance); - msg_set_linkprio(msg, priority); - 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) - l_ptr->stats.sent_probes++; - l_ptr->stats.sent_states++; - } else { /* RESET_MSG or ACTIVATE_MSG */ - 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->advertised_mtu); + msg_set_probe(hdr, probe); + if (probe) + l->stats.sent_probes++; + l->stats.sent_states++; + l->rcv_unacked = 0; + } else { + /* RESET_MSG or ACTIVATE_MSG */ + msg_set_max_pkt(hdr, l->advertised_mtu); + msg_set_ack(hdr, l->rcv_nxt - 1); + msg_set_next_sent(hdr, 1); } + skb = tipc_buf_acquire(msg_size(hdr)); + if (!skb) + return; + skb_copy_to_linear_data(skb, hdr, msg_size(hdr)); + skb->priority = TC_PRIO_CONTROL; + __skb_queue_tail(xmitq, skb); +} - r_flag = (l_ptr->owner->working_links > tipc_link_is_up(l_ptr)); - msg_set_redundant_link(msg, r_flag); - msg_set_linkprio(msg, l_ptr->priority); - msg_set_size(msg, msg_size); +/* tipc_link_tnl_prepare(): prepare and return a list of tunnel packets + * with contents of the link's transmit and backlog queues. + */ +void tipc_link_tnl_prepare(struct tipc_link *l, struct tipc_link *tnl, + int mtyp, struct sk_buff_head *xmitq) +{ + struct sk_buff *skb, *tnlskb; + struct tipc_msg *hdr, tnlhdr; + struct sk_buff_head *queue = &l->transmq; + struct sk_buff_head tmpxq, tnlq; + u16 pktlen, pktcnt, seqno = l->snd_nxt; - msg_set_seqno(msg, mod(l_ptr->next_out_no + (0xffff/2))); + if (!tnl) + return; - buf = tipc_buf_acquire(msg_size); - if (!buf) + skb_queue_head_init(&tnlq); + skb_queue_head_init(&tmpxq); + + /* At least one packet required for safe algorithm => add dummy */ + skb = tipc_msg_create(TIPC_LOW_IMPORTANCE, TIPC_DIRECT_MSG, + BASIC_H_SIZE, 0, l->addr, link_own_addr(l), + 0, 0, TIPC_ERR_NO_PORT); + if (!skb) { + pr_warn("%sunable to create tunnel packet\n", link_co_err); return; + } + skb_queue_tail(&tnlq, skb); + tipc_link_xmit(l, &tnlq, &tmpxq); + __skb_queue_purge(&tmpxq); + + /* Initialize reusable tunnel packet header */ + tipc_msg_init(link_own_addr(l), &tnlhdr, TUNNEL_PROTOCOL, + mtyp, INT_H_SIZE, l->addr); + pktcnt = skb_queue_len(&l->transmq) + skb_queue_len(&l->backlogq); + msg_set_msgcnt(&tnlhdr, pktcnt); + msg_set_bearer_id(&tnlhdr, l->peer_bearer_id); +tnl: + /* Wrap each packet into a tunnel packet */ + skb_queue_walk(queue, skb) { + hdr = buf_msg(skb); + if (queue == &l->backlogq) + msg_set_seqno(hdr, seqno++); + pktlen = msg_size(hdr); + msg_set_size(&tnlhdr, pktlen + INT_H_SIZE); + tnlskb = tipc_buf_acquire(pktlen + INT_H_SIZE); + if (!tnlskb) { + pr_warn("%sunable to send packet\n", link_co_err); + return; + } + skb_copy_to_linear_data(tnlskb, &tnlhdr, INT_H_SIZE); + skb_copy_to_linear_data_offset(tnlskb, INT_H_SIZE, hdr, pktlen); + __skb_queue_tail(&tnlq, tnlskb); + } + if (queue != &l->backlogq) { + queue = &l->backlogq; + goto tnl; + } + + tipc_link_xmit(tnl, &tnlq, xmitq); - skb_copy_to_linear_data(buf, msg, sizeof(l_ptr->proto_msg)); - buf->priority = TC_PRIO_CONTROL; - tipc_bearer_send(l_ptr->owner->net, l_ptr->bearer_id, buf, - &l_ptr->media_addr); - l_ptr->rcv_unacked = 0; - kfree_skb(buf); + if (mtyp == FAILOVER_MSG) { + tnl->drop_point = l->rcv_nxt; + tnl->failover_reasm_skb = l->reasm_buf; + l->reasm_buf = NULL; + } } -/* - * Receive protocol message : +/* tipc_link_proto_rcv(): receive link level protocol message : * Note that network plane id propagates through the network, and may - * change at any time. The node with lowest address rules + * change at any time. The node with lowest numerical id determines + * network plane */ -static void tipc_link_proto_rcv(struct tipc_link *l_ptr, - struct sk_buff *buf) +static int tipc_link_proto_rcv(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *xmitq) { - u32 rec_gap = 0; - u32 msg_tol; - struct tipc_msg *msg = buf_msg(buf); + struct tipc_msg *hdr = buf_msg(skb); + u16 rcvgap = 0; + u16 ack = msg_ack(hdr); + u16 gap = msg_seq_gap(hdr); + u16 peers_snd_nxt = msg_next_sent(hdr); + u16 peers_tol = msg_link_tolerance(hdr); + u16 peers_prio = msg_linkprio(hdr); + u16 rcv_nxt = l->rcv_nxt; + int mtyp = msg_type(hdr); + char *if_name; + int rc = 0; - if (l_ptr->flags & LINK_FAILINGOVER) + if (tipc_link_is_blocked(l) || !xmitq) goto exit; - if (l_ptr->net_plane != msg_net_plane(msg)) - if (link_own_addr(l_ptr) > msg_prevnode(msg)) - l_ptr->net_plane = msg_net_plane(msg); - - switch (msg_type(msg)) { + if (link_own_addr(l) > msg_prevnode(hdr)) + l->net_plane = msg_net_plane(hdr); + switch (mtyp) { case RESET_MSG: - if (!link_working_unknown(l_ptr) && - (l_ptr->peer_session != INVALID_SESSION)) { - if (less_eq(msg_session(msg), l_ptr->peer_session)) - break; /* duplicate or old reset: ignore */ - } - - if (!msg_redundant_link(msg) && (link_working_working(l_ptr) || - link_working_unknown(l_ptr))) { - /* - * peer has lost contact -- don't allow peer's links - * to reactivate before we recognize loss & clean up - */ - l_ptr->owner->action_flags |= TIPC_WAIT_OWN_LINKS_DOWN; - } - - link_state_event(l_ptr, RESET_MSG); + /* Ignore duplicate RESET with old session number */ + if ((less_eq(msg_session(hdr), l->peer_session)) && + (l->peer_session != WILDCARD_SESSION)) + break; /* fall thru' */ + case ACTIVATE_MSG: - /* Update link settings according other endpoint's values */ - strcpy((strrchr(l_ptr->name, ':') + 1), (char *)msg_data(msg)); - msg_tol = msg_link_tolerance(msg); - if (msg_tol > l_ptr->tolerance) - link_set_supervision_props(l_ptr, msg_tol); + /* Complete own link name with peer's interface name */ + if_name = strrchr(l->name, ':') + 1; + if (sizeof(l->name) - (if_name - l->name) <= TIPC_MAX_IF_NAME) + break; + if (msg_data_sz(hdr) < TIPC_MAX_IF_NAME) + break; + strncpy(if_name, msg_data(hdr), TIPC_MAX_IF_NAME); - if (msg_linkprio(msg) > l_ptr->priority) - l_ptr->priority = msg_linkprio(msg); + /* Update own tolerance if peer indicates a non-zero value */ + if (in_range(peers_tol, TIPC_MIN_LINK_TOL, TIPC_MAX_LINK_TOL)) + l->tolerance = peers_tol; - if (l_ptr->mtu > msg_max_pkt(msg)) - l_ptr->mtu = msg_max_pkt(msg); + /* Update own priority if peer's priority is higher */ + if (in_range(peers_prio, l->priority + 1, TIPC_MAX_LINK_PRI)) + l->priority = peers_prio; - /* Synchronize broadcast link info, if not done previously */ - if (!tipc_node_is_up(l_ptr->owner)) { - l_ptr->owner->bclink.last_sent = - l_ptr->owner->bclink.last_in = - msg_last_bcast(msg); - l_ptr->owner->bclink.oos_state = 0; - } + /* ACTIVATE_MSG serves as PEER_RESET if link is already down */ + if ((mtyp == RESET_MSG) || !link_is_up(l)) + rc = tipc_link_fsm_evt(l, LINK_PEER_RESET_EVT); - l_ptr->peer_session = msg_session(msg); - l_ptr->peer_bearer_id = msg_bearer_id(msg); + /* ACTIVATE_MSG takes up link if it was already locally reset */ + if ((mtyp == ACTIVATE_MSG) && (l->state == LINK_ESTABLISHING)) + rc = TIPC_LINK_UP_EVT; - if (msg_type(msg) == ACTIVATE_MSG) - link_state_event(l_ptr, ACTIVATE_MSG); + l->peer_session = msg_session(hdr); + l->peer_bearer_id = msg_bearer_id(hdr); + if (l->mtu > msg_max_pkt(hdr)) + l->mtu = msg_max_pkt(hdr); break; + case STATE_MSG: - msg_tol = msg_link_tolerance(msg); - if (msg_tol) - link_set_supervision_props(l_ptr, msg_tol); - - if (msg_linkprio(msg) && - (msg_linkprio(msg) != l_ptr->priority)) { - pr_debug("%s<%s>, priority change %u->%u\n", - link_rst_msg, l_ptr->name, - l_ptr->priority, msg_linkprio(msg)); - l_ptr->priority = msg_linkprio(msg); - tipc_link_reset(l_ptr); /* Enforce change to take effect */ - break; - } + /* Update own tolerance if peer indicates a non-zero value */ + if (in_range(peers_tol, TIPC_MIN_LINK_TOL, TIPC_MAX_LINK_TOL)) + l->tolerance = peers_tol; - /* Record reception; force mismatch at next timeout: */ - l_ptr->checkpoint--; + l->silent_intv_cnt = 0; + l->stats.recv_states++; + if (msg_probe(hdr)) + l->stats.recv_probes++; - link_state_event(l_ptr, TRAFFIC_MSG_EVT); - l_ptr->stats.recv_states++; - if (link_reset_unknown(l_ptr)) + if (!link_is_up(l)) { + if (l->state == LINK_ESTABLISHING) + rc = TIPC_LINK_UP_EVT; break; - - if (less_eq(mod(l_ptr->next_in_no), msg_next_sent(msg))) { - rec_gap = mod(msg_next_sent(msg) - - mod(l_ptr->next_in_no)); } - if (msg_probe(msg)) - l_ptr->stats.recv_probes++; - - /* Protocol message before retransmits, reduce loss risk */ - if (l_ptr->owner->bclink.recv_permitted) - tipc_bclink_update_link_state(l_ptr->owner, - msg_last_bcast(msg)); - - if (rec_gap || (msg_probe(msg))) { - tipc_link_proto_xmit(l_ptr, STATE_MSG, 0, - rec_gap, 0, 0); + /* Send NACK if peer has sent pkts we haven't received yet */ + if (more(peers_snd_nxt, rcv_nxt) && !tipc_link_is_synching(l)) + rcvgap = peers_snd_nxt - l->rcv_nxt; + if (rcvgap || (msg_probe(hdr))) + tipc_link_build_proto_msg(l, STATE_MSG, 0, rcvgap, + 0, 0, xmitq); + tipc_link_release_pkts(l, ack); + + /* If NACK, retransmit will now start at right position */ + if (gap) { + rc = tipc_link_retrans(l, ack + 1, ack + gap, xmitq); + l->stats.recv_nacks++; } - if (msg_seq_gap(msg)) { - l_ptr->stats.recv_nacks++; - tipc_link_retransmit(l_ptr, skb_peek(&l_ptr->transmq), - msg_seq_gap(msg)); - } - break; + + tipc_link_advance_backlog(l, xmitq); + if (unlikely(!skb_queue_empty(&l->wakeupq))) + link_prepare_wakeup(l); } exit: - kfree_skb(buf); + kfree_skb(skb); + return rc; } - -/* tipc_link_tunnel_xmit(): Tunnel one packet via a link belonging to - * a different bearer. Owner node is locked. +/* tipc_link_build_bc_proto_msg() - create broadcast protocol message */ -static void tipc_link_tunnel_xmit(struct tipc_link *l_ptr, - struct tipc_msg *tunnel_hdr, - struct tipc_msg *msg, - u32 selector) +static bool tipc_link_build_bc_proto_msg(struct tipc_link *l, bool bcast, + u16 peers_snd_nxt, + struct sk_buff_head *xmitq) { - struct tipc_link *tunnel; struct sk_buff *skb; - u32 length = msg_size(msg); + struct tipc_msg *hdr; + struct sk_buff *dfrd_skb = skb_peek(&l->deferdq); + u16 ack = l->rcv_nxt - 1; + u16 gap_to = peers_snd_nxt - 1; - tunnel = l_ptr->owner->active_links[selector & 1]; - if (!tipc_link_is_up(tunnel)) { - pr_warn("%stunnel link no longer available\n", link_co_err); - return; - } - msg_set_size(tunnel_hdr, length + INT_H_SIZE); - skb = tipc_buf_acquire(length + INT_H_SIZE); - if (!skb) { - pr_warn("%sunable to send tunnel msg\n", link_co_err); - return; - } - skb_copy_to_linear_data(skb, tunnel_hdr, INT_H_SIZE); - skb_copy_to_linear_data_offset(skb, INT_H_SIZE, msg, length); - __tipc_link_xmit_skb(tunnel, skb); + skb = tipc_msg_create(BCAST_PROTOCOL, STATE_MSG, INT_H_SIZE, + 0, l->addr, link_own_addr(l), 0, 0, 0); + if (!skb) + return false; + hdr = buf_msg(skb); + msg_set_last_bcast(hdr, l->bc_sndlink->snd_nxt - 1); + msg_set_bcast_ack(hdr, ack); + msg_set_bcgap_after(hdr, ack); + if (dfrd_skb) + gap_to = buf_seqno(dfrd_skb) - 1; + msg_set_bcgap_to(hdr, gap_to); + msg_set_non_seq(hdr, bcast); + __skb_queue_tail(xmitq, skb); + return true; } +/* tipc_link_build_bc_init_msg() - synchronize broadcast link endpoints. + * + * Give a newly added peer node the sequence number where it should + * start receiving and acking broadcast packets. + */ +static void tipc_link_build_bc_init_msg(struct tipc_link *l, + struct sk_buff_head *xmitq) +{ + struct sk_buff_head list; + + __skb_queue_head_init(&list); + if (!tipc_link_build_bc_proto_msg(l->bc_rcvlink, false, 0, &list)) + return; + tipc_link_xmit(l, &list, xmitq); +} -/* tipc_link_failover_send_queue(): A link has gone down, but a second - * link is still active. We can do failover. Tunnel the failing link's - * whole send queue via the remaining link. This way, we don't lose - * any packets, and sequence order is preserved for subsequent traffic - * sent over the remaining link. Owner node is locked. +/* tipc_link_bc_init_rcv - receive initial broadcast synch data from peer */ -void tipc_link_failover_send_queue(struct tipc_link *l_ptr) +void tipc_link_bc_init_rcv(struct tipc_link *l, struct tipc_msg *hdr) { - int msgcount; - struct tipc_link *tunnel = l_ptr->owner->active_links[0]; - struct tipc_msg tunnel_hdr; - struct sk_buff *skb; - int split_bundles; + int mtyp = msg_type(hdr); + u16 peers_snd_nxt = msg_bc_snd_nxt(hdr); - if (!tunnel) + if (link_is_up(l)) return; - 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); - - if (skb_queue_empty(&l_ptr->transmq)) { - skb = tipc_buf_acquire(INT_H_SIZE); - if (skb) { - skb_copy_to_linear_data(skb, &tunnel_hdr, INT_H_SIZE); - msg_set_size(&tunnel_hdr, INT_H_SIZE); - __tipc_link_xmit_skb(tunnel, skb); - } else { - pr_warn("%sunable to send changeover msg\n", - link_co_err); - } + if (msg_user(hdr) == BCAST_PROTOCOL) { + l->rcv_nxt = peers_snd_nxt; + l->state = LINK_ESTABLISHED; return; } - split_bundles = (l_ptr->owner->active_links[0] != - l_ptr->owner->active_links[1]); - - skb_queue_walk(&l_ptr->transmq, skb) { - struct tipc_msg *msg = buf_msg(skb); + if (l->peer_caps & TIPC_BCAST_SYNCH) + return; - if ((msg_user(msg) == MSG_BUNDLER) && split_bundles) { - struct tipc_msg *m = msg_get_wrapped(msg); - unchar *pos = (unchar *)m; + if (msg_peer_node_is_up(hdr)) + return; - msgcount = msg_msgcnt(msg); - while (msgcount--) { - msg_set_seqno(m, msg_seqno(msg)); - tipc_link_tunnel_xmit(l_ptr, &tunnel_hdr, m, - msg_link_selector(m)); - pos += align(msg_size(m)); - m = (struct tipc_msg *)pos; - } - } else { - tipc_link_tunnel_xmit(l_ptr, &tunnel_hdr, msg, - msg_link_selector(msg)); - } - } + /* Compatibility: accept older, less safe initial synch data */ + if ((mtyp == RESET_MSG) || (mtyp == ACTIVATE_MSG)) + l->rcv_nxt = peers_snd_nxt; } -/* tipc_link_dup_queue_xmit(): A second link has become active. Tunnel a - * duplicate of the first link's send queue via the new link. This way, we - * are guaranteed that currently queued packets from a socket are delivered - * before future traffic from the same socket, even if this is using the - * new link. The last arriving copy of each duplicate packet is dropped at - * the receiving end by the regular protocol check, so packet cardinality - * and sequence order is preserved per sender/receiver socket pair. - * Owner node is locked. +/* tipc_link_bc_sync_rcv - update rcv link according to peer's send state */ -void tipc_link_dup_queue_xmit(struct tipc_link *link, - struct tipc_link *tnl) +void tipc_link_bc_sync_rcv(struct tipc_link *l, struct tipc_msg *hdr, + struct sk_buff_head *xmitq) { - struct sk_buff *skb; - struct tipc_msg tnl_hdr; - struct sk_buff_head *queue = &link->transmq; - int mcnt; + u16 peers_snd_nxt = msg_bc_snd_nxt(hdr); - 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); + if (!link_is_up(l)) + return; -tunnel_queue: - skb_queue_walk(queue, skb) { - struct sk_buff *outskb; - struct tipc_msg *msg = buf_msg(skb); - u32 len = msg_size(msg); - - msg_set_ack(msg, mod(link->next_in_no - 1)); - msg_set_bcast_ack(msg, link->owner->bclink.last_in); - msg_set_size(&tnl_hdr, len + INT_H_SIZE); - outskb = tipc_buf_acquire(len + INT_H_SIZE); - if (outskb == NULL) { - pr_warn("%sunable to send duplicate msg\n", - link_co_err); - return; - } - skb_copy_to_linear_data(outskb, &tnl_hdr, INT_H_SIZE); - skb_copy_to_linear_data_offset(outskb, INT_H_SIZE, - skb->data, len); - __tipc_link_xmit_skb(tnl, outskb); - if (!tipc_link_is_up(link)) - return; - } - if (queue == &link->backlogq) + if (!msg_peer_node_is_up(hdr)) return; - queue = &link->backlogq; - goto tunnel_queue; -} -/* tipc_link_failover_rcv(): Receive a tunnelled FAILOVER_MSG packet - * Owner node is locked. - */ -static bool tipc_link_failover_rcv(struct tipc_link *link, - struct sk_buff **skb) -{ - 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; + l->bc_peer_is_up = true; - if (msg_type(msg) != FAILOVER_MSG) { - pr_warn("%sunknown tunnel pkt received\n", link_co_err); - goto exit; + /* Ignore if peers_snd_nxt goes beyond receive window */ + if (more(peers_snd_nxt, l->rcv_nxt + l->window)) + return; + + if (!more(peers_snd_nxt, l->rcv_nxt)) { + l->nack_state = BC_NACK_SND_CONDITIONAL; + return; } - if (bearer_id >= MAX_BEARERS) - goto exit; - if (bearer_id == link->bearer_id) - goto exit; + /* Don't NACK if one was recently sent or peeked */ + if (l->nack_state == BC_NACK_SND_SUPPRESS) { + l->nack_state = BC_NACK_SND_UNCONDITIONAL; + return; + } - pl = link->owner->links[bearer_id]; - if (pl && tipc_link_is_up(pl)) - tipc_link_reset(pl); + /* Conditionally delay NACK sending until next synch rcv */ + if (l->nack_state == BC_NACK_SND_CONDITIONAL) { + l->nack_state = BC_NACK_SND_UNCONDITIONAL; + if ((peers_snd_nxt - l->rcv_nxt) < TIPC_MIN_LINK_WIN) + return; + } - if (link->failover_pkts == FIRST_FAILOVER) - link->failover_pkts = msg_msgcnt(msg); + /* Send NACK now but suppress next one */ + tipc_link_build_bc_proto_msg(l, true, peers_snd_nxt, xmitq); + l->nack_state = BC_NACK_SND_SUPPRESS; +} - /* Should we expect an inner packet? */ - if (!link->failover_pkts) - goto exit; +void tipc_link_bc_ack_rcv(struct tipc_link *l, u16 acked, + struct sk_buff_head *xmitq) +{ + struct sk_buff *skb, *tmp; + struct tipc_link *snd_l = l->bc_sndlink; - 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 (!link_is_up(l) || !l->bc_peer_is_up) + return; - /* Was this packet already delivered? */ - if (less(buf_seqno(iskb), link->failover_checkpt)) { - kfree_skb(iskb); - iskb = NULL; - goto exit; + if (!more(acked, l->acked)) + return; + + /* Skip over packets peer has already acked */ + skb_queue_walk(&snd_l->transmq, skb) { + if (more(buf_seqno(skb), l->acked)) + break; } - if (msg_user(buf_msg(iskb)) == MSG_FRAGMENTER) { - link->stats.recv_fragments++; - tipc_buf_append(&link->failover_skb, &iskb); + + /* Update/release the packets peer is acking now */ + skb_queue_walk_from_safe(&snd_l->transmq, skb, tmp) { + if (more(buf_seqno(skb), acked)) + break; + if (!--TIPC_SKB_CB(skb)->ackers) { + __skb_unlink(skb, &snd_l->transmq); + kfree_skb(skb); + } } -exit: - if (!link->failover_pkts && pl) - pl->flags &= ~LINK_FAILINGOVER; - kfree_skb(*skb); - *skb = iskb; - return *skb; + l->acked = acked; + tipc_link_advance_backlog(snd_l, xmitq); + if (unlikely(!skb_queue_empty(&snd_l->wakeupq))) + link_prepare_wakeup(snd_l); } -static void link_set_supervision_props(struct tipc_link *l_ptr, u32 tol) +/* tipc_link_bc_nack_rcv(): receive broadcast nack message + */ +int tipc_link_bc_nack_rcv(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *xmitq) { - unsigned long intv = ((tol / 4) > 500) ? 500 : tol / 4; + struct tipc_msg *hdr = buf_msg(skb); + u32 dnode = msg_destnode(hdr); + int mtyp = msg_type(hdr); + u16 acked = msg_bcast_ack(hdr); + u16 from = acked + 1; + u16 to = msg_bcgap_to(hdr); + u16 peers_snd_nxt = to + 1; + int rc = 0; + + kfree_skb(skb); + + if (!tipc_link_is_up(l) || !l->bc_peer_is_up) + return 0; - if ((tol < TIPC_MIN_LINK_TOL) || (tol > TIPC_MAX_LINK_TOL)) - return; + if (mtyp != STATE_MSG) + return 0; + + if (dnode == link_own_addr(l)) { + tipc_link_bc_ack_rcv(l, acked, xmitq); + rc = tipc_link_retrans(l->bc_sndlink, from, to, xmitq); + l->stats.recv_nacks++; + return rc; + } + + /* Msg for other node => suppress own NACK at next sync if applicable */ + if (more(peers_snd_nxt, l->rcv_nxt) && !less(l->rcv_nxt, from)) + l->nack_state = BC_NACK_SND_SUPPRESS; - l_ptr->tolerance = tol; - l_ptr->cont_intv = msecs_to_jiffies(intv); - l_ptr->abort_limit = tol / (jiffies_to_msecs(l_ptr->cont_intv) / 4); + return 0; } void tipc_link_set_queue_limits(struct tipc_link *l, u32 win) @@ -1780,7 +1549,7 @@ static struct tipc_node *tipc_link_find_owner(struct net *net, list_for_each_entry_rcu(n_ptr, &tn->node_list, list) { tipc_node_lock(n_ptr); for (i = 0; i < MAX_BEARERS; i++) { - l_ptr = n_ptr->links[i]; + l_ptr = n_ptr->links[i].link; if (l_ptr && !strcmp(l_ptr->name, link_name)) { *bearer_id = i; found_node = n_ptr; @@ -1803,31 +1572,20 @@ static struct tipc_node *tipc_link_find_owner(struct net *net, static void link_reset_statistics(struct tipc_link *l_ptr) { memset(&l_ptr->stats, 0, sizeof(l_ptr->stats)); - l_ptr->stats.sent_info = l_ptr->next_out_no; - l_ptr->stats.recv_info = l_ptr->next_in_no; + l_ptr->stats.sent_info = l_ptr->snd_nxt; + l_ptr->stats.recv_info = l_ptr->rcv_nxt; } -static void link_print(struct tipc_link *l_ptr, const char *str) +static void link_print(struct tipc_link *l, const char *str) { - struct tipc_net *tn = net_generic(l_ptr->owner->net, tipc_net_id); - struct tipc_bearer *b_ptr; - - rcu_read_lock(); - b_ptr = rcu_dereference_rtnl(tn->bearer_list[l_ptr->bearer_id]); - if (b_ptr) - pr_info("%s Link %x<%s>:", str, l_ptr->addr, b_ptr->name); - rcu_read_unlock(); - - if (link_working_unknown(l_ptr)) - pr_cont(":WU\n"); - else if (link_reset_reset(l_ptr)) - pr_cont(":RR\n"); - else if (link_reset_unknown(l_ptr)) - pr_cont(":RU\n"); - else if (link_working_working(l_ptr)) - pr_cont(":WW\n"); - else - pr_cont("\n"); + struct sk_buff *hskb = skb_peek(&l->transmq); + u16 head = hskb ? msg_seqno(buf_msg(hskb)) : l->snd_nxt - 1; + u16 tail = l->snd_nxt - 1; + + pr_info("%s Link <%s> state %x\n", str, l->name, l->state); + pr_info("XMTQ: %u [%u-%u], BKLGQ: %u, SNDNX: %u, RCVNX: %u\n", + skb_queue_len(&l->transmq), head, tail, + skb_queue_len(&l->backlogq), l->snd_nxt, l->rcv_nxt); } /* Parse and validate nested (link) properties valid for media, bearer and link @@ -1893,13 +1651,16 @@ int tipc_nl_link_set(struct sk_buff *skb, struct genl_info *info) name = nla_data(attrs[TIPC_NLA_LINK_NAME]); + if (strcmp(name, tipc_bclink_name) == 0) + return tipc_nl_bc_link_set(net, attrs); + node = tipc_link_find_owner(net, name, &bearer_id); if (!node) return -EINVAL; tipc_node_lock(node); - link = node->links[bearer_id]; + link = node->links[bearer_id].link; if (!link) { res = -EINVAL; goto out; @@ -1919,7 +1680,7 @@ int tipc_nl_link_set(struct sk_buff *skb, struct genl_info *info) u32 tol; tol = nla_get_u32(props[TIPC_NLA_PROP_TOL]); - link_set_supervision_props(link, tol); + link->tolerance = tol; tipc_link_proto_xmit(link, STATE_MSG, 0, 0, tol, 0); } if (props[TIPC_NLA_PROP_PRIO]) { @@ -2034,15 +1795,15 @@ static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg, goto attr_msg_full; 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)) + if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, link->rcv_nxt)) goto attr_msg_full; - if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, link->next_out_no)) + if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, link->snd_nxt)) goto attr_msg_full; if (tipc_link_is_up(link)) if (nla_put_flag(msg->skb, TIPC_NLA_LINK_UP)) goto attr_msg_full; - if (tipc_link_is_active(link)) + if (link->active) if (nla_put_flag(msg->skb, TIPC_NLA_LINK_ACTIVE)) goto attr_msg_full; @@ -2089,10 +1850,11 @@ static int __tipc_nl_add_node_links(struct net *net, struct tipc_nl_msg *msg, for (i = *prev_link; i < MAX_BEARERS; i++) { *prev_link = i; - if (!node->links[i]) + if (!node->links[i].link) continue; - err = __tipc_nl_add_link(net, msg, node->links[i], NLM_F_MULTI); + err = __tipc_nl_add_link(net, msg, + node->links[i].link, NLM_F_MULTI); if (err) return err; } @@ -2175,50 +1937,53 @@ out: int tipc_nl_link_get(struct sk_buff *skb, struct genl_info *info) { struct net *net = genl_info_net(info); - struct sk_buff *ans_skb; struct tipc_nl_msg msg; - struct tipc_link *link; - struct tipc_node *node; char *name; - int bearer_id; int err; + msg.portid = info->snd_portid; + msg.seq = info->snd_seq; + if (!info->attrs[TIPC_NLA_LINK_NAME]) return -EINVAL; - name = nla_data(info->attrs[TIPC_NLA_LINK_NAME]); - node = tipc_link_find_owner(net, name, &bearer_id); - if (!node) - return -EINVAL; - ans_skb = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL); - if (!ans_skb) + msg.skb = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL); + if (!msg.skb) return -ENOMEM; - msg.skb = ans_skb; - msg.portid = info->snd_portid; - msg.seq = info->snd_seq; - - tipc_node_lock(node); - link = node->links[bearer_id]; - if (!link) { - err = -EINVAL; - goto err_out; - } - - err = __tipc_nl_add_link(net, &msg, link, 0); - if (err) - goto err_out; + if (strcmp(name, tipc_bclink_name) == 0) { + err = tipc_nl_add_bc_link(net, &msg); + if (err) { + nlmsg_free(msg.skb); + return err; + } + } else { + int bearer_id; + struct tipc_node *node; + struct tipc_link *link; - tipc_node_unlock(node); + node = tipc_link_find_owner(net, name, &bearer_id); + if (!node) + return -EINVAL; - return genlmsg_reply(ans_skb, info); + tipc_node_lock(node); + link = node->links[bearer_id].link; + if (!link) { + tipc_node_unlock(node); + nlmsg_free(msg.skb); + return -EINVAL; + } -err_out: - tipc_node_unlock(node); - nlmsg_free(ans_skb); + err = __tipc_nl_add_link(net, &msg, link, 0); + tipc_node_unlock(node); + if (err) { + nlmsg_free(msg.skb); + return err; + } + } - return err; + return genlmsg_reply(msg.skb, info); } int tipc_nl_link_reset_stats(struct sk_buff *skb, struct genl_info *info) @@ -2258,7 +2023,7 @@ int tipc_nl_link_reset_stats(struct sk_buff *skb, struct genl_info *info) tipc_node_lock(node); - link = node->links[bearer_id]; + link = node->links[bearer_id].link; if (!link) { tipc_node_unlock(node); return -EINVAL; diff --git a/kernel/net/tipc/link.h b/kernel/net/tipc/link.h index b5b4e3554..66d859b66 100644 --- a/kernel/net/tipc/link.h +++ b/kernel/net/tipc/link.h @@ -49,19 +49,26 @@ */ #define INVALID_LINK_SEQ 0x10000 -/* Link working states +/* Link FSM events: */ -#define WORKING_WORKING 560810u -#define WORKING_UNKNOWN 560811u -#define RESET_UNKNOWN 560812u -#define RESET_RESET 560813u +enum { + LINK_ESTABLISH_EVT = 0xec1ab1e, + LINK_PEER_RESET_EVT = 0x9eed0e, + LINK_FAILURE_EVT = 0xfa110e, + LINK_RESET_EVT = 0x10ca1d0e, + LINK_FAILOVER_BEGIN_EVT = 0xfa110bee, + LINK_FAILOVER_END_EVT = 0xfa110ede, + LINK_SYNCH_BEGIN_EVT = 0xc1ccbee, + LINK_SYNCH_END_EVT = 0xc1ccede +}; -/* Link endpoint execution states +/* Events returned from link at packet reception or at timeout */ -#define LINK_STARTED 0x0001 -#define LINK_STOPPED 0x0002 -#define LINK_SYNCHING 0x0004 -#define LINK_FAILINGOVER 0x0008 +enum { + TIPC_LINK_UP_EVT = 1, + TIPC_LINK_DOWN_EVT = (1 << 1), + TIPC_LINK_SND_BC_ACK = (1 << 2) +}; /* Starting value for maximum packet size negotiation on unicast links * (unless bearer MTU is less) @@ -104,33 +111,34 @@ struct tipc_stats { * @name: link name character string * @media_addr: media address to use when sending messages over link * @timer: link timer - * @owner: pointer to peer node + * @net: pointer to namespace struct * @refcnt: reference counter for permanent references (owner node & timer) - * @flags: execution state flags for link endpoint instance - * @checkpoint: reference point for triggering link continuity checking * @peer_session: link session # being used by peer end of link * @peer_bearer_id: bearer id used by link's peer endpoint * @bearer_id: local bearer id used by link * @tolerance: minimum link continuity loss needed to reset link [in ms] - * @cont_intv: link continuity testing interval + * @keepalive_intv: link keepalive timer interval * @abort_limit: # of unacknowledged continuity probes needed to reset link * @state: current state of link FSM - * @fsm_msg_cnt: # of protocol messages link FSM has sent in current state + * @peer_caps: bitmap describing capabilities of peer node + * @silent_intv_cnt: # of timer intervals without any reception from peer * @proto_msg: template for control messages generated by link * @pmsg: convenience pointer to "proto_msg" field * @priority: current link priority * @net_plane: current link network plane ('A' through 'H') * @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 + * @reset_rcv_checkpt: seq # of last acknowledged message at time of link reset * @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 + * @snt_nxt: next sequence number to use for outbound messages * @last_retransmitted: sequence number of most recently retransmitted message * @stale_count: # of identical retransmit requests made by peer - * @next_in_no: next sequence number to expect for inbound messages + * @ackers: # of peers that needs to ack each packet before it can be released + * @acked: # last packet acked by a certain peer. Used for broadcast. + * @rcv_nxt: next sequence number to expect for inbound messages * @deferred_queue: deferred queue saved OOS b'cast message received from node * @unacked_window: # of inbound messages rx'd without ack'ing back to peer * @inputq: buffer queue for messages to be delivered upwards @@ -139,27 +147,26 @@ struct tipc_stats { * @wakeupq: linked list of wakeup msgs waiting for link congestion to abate * @long_msg_seq_no: next identifier to use for outbound fragmented messages * @reasm_buf: head of partially reassembled inbound message fragments + * @bc_rcvr: marks that this is a broadcast receiver link * @stats: collects statistics regarding link activity */ struct tipc_link { u32 addr; char name[TIPC_MAX_LINK_NAME]; - struct tipc_media_addr media_addr; - struct timer_list timer; - struct tipc_node *owner; - struct kref ref; + struct tipc_media_addr *media_addr; + struct net *net; /* Management and link supervision data */ - unsigned int flags; - u32 checkpoint; u32 peer_session; u32 peer_bearer_id; u32 bearer_id; u32 tolerance; - unsigned long cont_intv; + unsigned long keepalive_intv; u32 abort_limit; - int state; - u32 fsm_msg_cnt; + u32 state; + u16 peer_caps; + bool active; + u32 silent_intv_cnt; struct { unchar hdr[INT_H_SIZE]; unchar body[TIPC_MAX_IF_NAME]; @@ -167,12 +174,10 @@ struct tipc_link { struct tipc_msg *pmsg; u32 priority; char net_plane; - u16 synch_point; - /* Failover */ - u16 failover_pkts; - u16 failover_checkpt; - struct sk_buff *failover_skb; + /* Failover/synch */ + u16 drop_point; + struct sk_buff *failover_reasm_skb; /* Max packet negotiation */ u16 mtu; @@ -185,17 +190,17 @@ struct tipc_link { u16 len; u16 limit; } backlog[5]; - u32 next_out_no; - u32 window; - u32 last_retransmitted; + u16 snd_nxt; + u16 last_retransm; + u16 window; u32 stale_count; /* Reception */ - u32 next_in_no; + u16 rcv_nxt; u32 rcv_unacked; struct sk_buff_head deferdq; - struct sk_buff_head inputq; - struct sk_buff_head namedq; + struct sk_buff_head *inputq; + struct sk_buff_head *namedq; /* Congestion handling */ struct sk_buff_head wakeupq; @@ -203,109 +208,76 @@ struct tipc_link { /* Fragmentation/reassembly */ struct sk_buff *reasm_buf; + /* Broadcast */ + u16 ackers; + u16 acked; + struct tipc_link *bc_rcvlink; + struct tipc_link *bc_sndlink; + int nack_state; + bool bc_peer_is_up; + /* Statistics */ struct tipc_stats stats; }; -struct tipc_port; - -struct tipc_link *tipc_link_create(struct tipc_node *n_ptr, - struct tipc_bearer *b_ptr, - const struct tipc_media_addr *media_addr); -void tipc_link_delete(struct tipc_link *link); -void tipc_link_delete_list(struct net *net, unsigned int bearer_id, - bool shutting_down); -void tipc_link_failover_send_queue(struct tipc_link *l_ptr); -void tipc_link_dup_queue_xmit(struct tipc_link *l_ptr, struct tipc_link *dest); +bool tipc_link_create(struct net *net, char *if_name, int bearer_id, + int tolerance, char net_plane, u32 mtu, int priority, + int window, u32 session, u32 ownnode, u32 peer, + u16 peer_caps, + struct tipc_link *bc_sndlink, + struct tipc_link *bc_rcvlink, + struct sk_buff_head *inputq, + struct sk_buff_head *namedq, + struct tipc_link **link); +bool tipc_link_bc_create(struct net *net, u32 ownnode, u32 peer, + int mtu, int window, u16 peer_caps, + struct sk_buff_head *inputq, + struct sk_buff_head *namedq, + struct tipc_link *bc_sndlink, + struct tipc_link **link); +void tipc_link_tnl_prepare(struct tipc_link *l, struct tipc_link *tnl, + int mtyp, struct sk_buff_head *xmitq); +void tipc_link_build_reset_msg(struct tipc_link *l, struct sk_buff_head *xmitq); +int tipc_link_fsm_evt(struct tipc_link *l, int evt); void tipc_link_reset_fragments(struct tipc_link *l_ptr); -int tipc_link_is_up(struct tipc_link *l_ptr); -int tipc_link_is_active(struct tipc_link *l_ptr); -void tipc_link_purge_queues(struct tipc_link *l_ptr); -void tipc_link_reset_all(struct tipc_node *node); +bool tipc_link_is_up(struct tipc_link *l); +bool tipc_link_peer_is_down(struct tipc_link *l); +bool tipc_link_is_reset(struct tipc_link *l); +bool tipc_link_is_establishing(struct tipc_link *l); +bool tipc_link_is_synching(struct tipc_link *l); +bool tipc_link_is_failingover(struct tipc_link *l); +bool tipc_link_is_blocked(struct tipc_link *l); +void tipc_link_set_active(struct tipc_link *l, bool active); void tipc_link_reset(struct tipc_link *l_ptr); -void tipc_link_reset_list(struct net *net, unsigned int bearer_id); -int tipc_link_xmit_skb(struct net *net, struct sk_buff *skb, u32 dest, - u32 selector); -int tipc_link_xmit(struct net *net, struct sk_buff_head *list, u32 dest, - u32 selector); -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); -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); -void tipc_link_retransmit(struct tipc_link *l_ptr, - struct sk_buff *start, u32 retransmits); -struct sk_buff *tipc_skb_queue_next(const struct sk_buff_head *list, - const struct sk_buff *skb); +int tipc_link_xmit(struct tipc_link *link, struct sk_buff_head *list, + struct sk_buff_head *xmitq); +void tipc_link_set_queue_limits(struct tipc_link *l, u32 window); int tipc_nl_link_dump(struct sk_buff *skb, struct netlink_callback *cb); int tipc_nl_link_get(struct sk_buff *skb, struct genl_info *info); int tipc_nl_link_set(struct sk_buff *skb, struct genl_info *info); int tipc_nl_link_reset_stats(struct sk_buff *skb, struct genl_info *info); int tipc_nl_parse_link_prop(struct nlattr *prop, struct nlattr *props[]); -void link_prepare_wakeup(struct tipc_link *l); - -/* - * Link sequence number manipulation routines (uses modulo 2**16 arithmetic) - */ -static inline u32 buf_seqno(struct sk_buff *buf) -{ - return msg_seqno(buf_msg(buf)); -} - -static inline u32 mod(u32 x) -{ - return x & 0xffffu; -} - -static inline int less_eq(u32 left, u32 right) -{ - return mod(right - left) < 32768u; -} - -static inline int more(u32 left, u32 right) -{ - return !less_eq(left, right); -} - -static inline int less(u32 left, u32 right) -{ - return less_eq(left, right) && (mod(right) != mod(left)); -} - -static inline u32 lesser(u32 left, u32 right) -{ - return less_eq(left, right) ? left : right; -} - -static inline u32 link_own_addr(struct tipc_link *l) -{ - return msg_prevnode(l->pmsg); -} - -/* - * Link status checking routines - */ -static inline int link_working_working(struct tipc_link *l_ptr) -{ - return l_ptr->state == WORKING_WORKING; -} - -static inline int link_working_unknown(struct tipc_link *l_ptr) -{ - return l_ptr->state == WORKING_UNKNOWN; -} - -static inline int link_reset_unknown(struct tipc_link *l_ptr) -{ - return l_ptr->state == RESET_UNKNOWN; -} - -static inline int link_reset_reset(struct tipc_link *l_ptr) -{ - return l_ptr->state == RESET_RESET; -} - +int tipc_link_timeout(struct tipc_link *l, struct sk_buff_head *xmitq); +int tipc_link_rcv(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *xmitq); +int tipc_link_build_ack_msg(struct tipc_link *l, struct sk_buff_head *xmitq); +void tipc_link_add_bc_peer(struct tipc_link *snd_l, + struct tipc_link *uc_l, + struct sk_buff_head *xmitq); +void tipc_link_remove_bc_peer(struct tipc_link *snd_l, + struct tipc_link *rcv_l, + struct sk_buff_head *xmitq); +int tipc_link_bc_peers(struct tipc_link *l); +void tipc_link_set_mtu(struct tipc_link *l, int mtu); +int tipc_link_mtu(struct tipc_link *l); +void tipc_link_bc_ack_rcv(struct tipc_link *l, u16 acked, + struct sk_buff_head *xmitq); +void tipc_link_build_bc_sync_msg(struct tipc_link *l, + struct sk_buff_head *xmitq); +void tipc_link_bc_init_rcv(struct tipc_link *l, struct tipc_msg *hdr); +void tipc_link_bc_sync_rcv(struct tipc_link *l, struct tipc_msg *hdr, + struct sk_buff_head *xmitq); +int tipc_link_bc_nack_rcv(struct tipc_link *l, struct sk_buff *skb, + struct sk_buff_head *xmitq); #endif diff --git a/kernel/net/tipc/msg.c b/kernel/net/tipc/msg.c index c3e96e815..8740930f0 100644 --- a/kernel/net/tipc/msg.c +++ b/kernel/net/tipc/msg.c @@ -121,7 +121,7 @@ int tipc_buf_append(struct sk_buff **headbuf, struct sk_buff **buf) { struct sk_buff *head = *headbuf; struct sk_buff *frag = *buf; - struct sk_buff *tail; + struct sk_buff *tail = NULL; struct tipc_msg *msg; u32 fragid; int delta; @@ -141,9 +141,15 @@ int tipc_buf_append(struct sk_buff **headbuf, struct sk_buff **buf) if (unlikely(skb_unclone(frag, GFP_ATOMIC))) goto err; head = *headbuf = frag; - skb_frag_list_init(head); - TIPC_SKB_CB(head)->tail = NULL; *buf = NULL; + TIPC_SKB_CB(head)->tail = NULL; + if (skb_is_nonlinear(head)) { + skb_walk_frags(head, tail) { + TIPC_SKB_CB(head)->tail = tail; + } + } else { + skb_frag_list_init(head); + } return 0; } @@ -176,7 +182,6 @@ int tipc_buf_append(struct sk_buff **headbuf, struct sk_buff **buf) *buf = NULL; return 0; err: - pr_warn_ratelimited("Unable to build fragment list\n"); kfree_skb(*buf); kfree_skb(*headbuf); *buf = *headbuf = NULL; @@ -331,16 +336,15 @@ error: /** * tipc_msg_bundle(): Append contents of a buffer to tail of an existing one - * @bskb: the buffer to append to ("bundle") - * @skb: buffer to be appended + * @skb: the buffer to append to ("bundle") + * @msg: message to be appended * @mtu: max allowable size for the bundle buffer * Consumes buffer if successful * Returns true if bundling could be performed, otherwise false */ -bool tipc_msg_bundle(struct sk_buff *bskb, struct sk_buff *skb, u32 mtu) +bool tipc_msg_bundle(struct sk_buff *skb, struct tipc_msg *msg, u32 mtu) { struct tipc_msg *bmsg; - struct tipc_msg *msg = buf_msg(skb); unsigned int bsz; unsigned int msz = msg_size(msg); u32 start, pad; @@ -348,9 +352,9 @@ bool tipc_msg_bundle(struct sk_buff *bskb, struct sk_buff *skb, u32 mtu) if (likely(msg_user(msg) == MSG_FRAGMENTER)) return false; - if (!bskb) + if (!skb) return false; - bmsg = buf_msg(bskb); + bmsg = buf_msg(skb); bsz = msg_size(bmsg); start = align(bsz); pad = start - bsz; @@ -359,18 +363,20 @@ bool tipc_msg_bundle(struct sk_buff *bskb, struct sk_buff *skb, u32 mtu) return false; if (unlikely(msg_user(msg) == BCAST_PROTOCOL)) return false; - if (likely(msg_user(bmsg) != MSG_BUNDLER)) + if (unlikely(msg_user(bmsg) != MSG_BUNDLER)) return false; - if (unlikely(skb_tailroom(bskb) < (pad + msz))) + if (unlikely(skb_tailroom(skb) < (pad + msz))) return false; if (unlikely(max < (start + msz))) return false; + if ((msg_importance(msg) < TIPC_SYSTEM_IMPORTANCE) && + (msg_importance(bmsg) == TIPC_SYSTEM_IMPORTANCE)) + return false; - skb_put(bskb, pad + msz); - skb_copy_to_linear_data_offset(bskb, start, skb->data, msz); + skb_put(skb, pad + msz); + skb_copy_to_linear_data_offset(skb, start, msg, msz); msg_set_size(bmsg, start + msz); msg_set_msgcnt(bmsg, msg_msgcnt(bmsg) + 1); - kfree_skb(skb); return true; } @@ -416,18 +422,18 @@ none: /** * tipc_msg_make_bundle(): Create bundle buf and append message to its tail - * @list: the buffer chain - * @skb: buffer to be appended and replaced + * @list: the buffer chain, where head is the buffer to replace/append + * @skb: buffer to be created, appended to and returned in case of success + * @msg: message to be appended * @mtu: max allowable size for the bundle buffer, inclusive header * @dnode: destination node for message. (Not always present in header) - * Replaces buffer if successful * Returns true if success, otherwise false */ -bool tipc_msg_make_bundle(struct sk_buff **skb, u32 mtu, u32 dnode) +bool tipc_msg_make_bundle(struct sk_buff **skb, struct tipc_msg *msg, + u32 mtu, u32 dnode) { - struct sk_buff *bskb; + struct sk_buff *_skb; struct tipc_msg *bmsg; - struct tipc_msg *msg = buf_msg(*skb); u32 msz = msg_size(msg); u32 max = mtu - INT_H_SIZE; @@ -440,78 +446,94 @@ bool tipc_msg_make_bundle(struct sk_buff **skb, u32 mtu, u32 dnode) if (msz > (max / 2)) return false; - bskb = tipc_buf_acquire(max); - if (!bskb) + _skb = tipc_buf_acquire(max); + if (!_skb) return false; - skb_trim(bskb, INT_H_SIZE); - bmsg = buf_msg(bskb); + skb_trim(_skb, INT_H_SIZE); + bmsg = buf_msg(_skb); tipc_msg_init(msg_prevnode(msg), bmsg, MSG_BUNDLER, 0, INT_H_SIZE, dnode); + if (msg_isdata(msg)) + msg_set_importance(bmsg, TIPC_CRITICAL_IMPORTANCE); + else + msg_set_importance(bmsg, TIPC_SYSTEM_IMPORTANCE); msg_set_seqno(bmsg, msg_seqno(msg)); msg_set_ack(bmsg, msg_ack(msg)); msg_set_bcast_ack(bmsg, msg_bcast_ack(msg)); - tipc_msg_bundle(bskb, *skb, mtu); - *skb = bskb; + tipc_msg_bundle(_skb, msg, mtu); + *skb = _skb; return true; } /** * tipc_msg_reverse(): swap source and destination addresses and add error code - * @buf: buffer containing message to be reversed - * @dnode: return value: node where to send message after reversal - * @err: error code to be set in message - * Consumes buffer if failure + * @own_node: originating node id for reversed message + * @skb: buffer containing message to be reversed; may be replaced. + * @err: error code to be set in message, if any + * Consumes buffer at failure * Returns true if success, otherwise false */ -bool tipc_msg_reverse(u32 own_addr, struct sk_buff *buf, u32 *dnode, - int err) +bool tipc_msg_reverse(u32 own_node, struct sk_buff **skb, int err) { - struct tipc_msg *msg = buf_msg(buf); + struct sk_buff *_skb = *skb; + struct tipc_msg *hdr = buf_msg(_skb); struct tipc_msg ohdr; - uint rdsz = min_t(uint, msg_data_sz(msg), MAX_FORWARD_SIZE); + int dlen = min_t(uint, msg_data_sz(hdr), MAX_FORWARD_SIZE); - if (skb_linearize(buf)) + if (skb_linearize(_skb)) goto exit; - msg = buf_msg(buf); - if (msg_dest_droppable(msg)) + hdr = buf_msg(_skb); + if (msg_dest_droppable(hdr)) goto exit; - if (msg_errcode(msg)) + if (msg_errcode(hdr)) goto exit; - memcpy(&ohdr, msg, msg_hdr_sz(msg)); - msg_set_errcode(msg, err); - msg_set_origport(msg, msg_destport(&ohdr)); - msg_set_destport(msg, msg_origport(&ohdr)); - msg_set_prevnode(msg, own_addr); - if (!msg_short(msg)) { - msg_set_orignode(msg, msg_destnode(&ohdr)); - msg_set_destnode(msg, msg_orignode(&ohdr)); + + /* Take a copy of original header before altering message */ + memcpy(&ohdr, hdr, msg_hdr_sz(hdr)); + + /* Never return SHORT header; expand by replacing buffer if necessary */ + if (msg_short(hdr)) { + *skb = tipc_buf_acquire(BASIC_H_SIZE + dlen); + if (!*skb) + goto exit; + memcpy((*skb)->data + BASIC_H_SIZE, msg_data(hdr), dlen); + kfree_skb(_skb); + _skb = *skb; + hdr = buf_msg(_skb); + memcpy(hdr, &ohdr, BASIC_H_SIZE); + msg_set_hdr_sz(hdr, BASIC_H_SIZE); } - msg_set_size(msg, msg_hdr_sz(msg) + rdsz); - skb_trim(buf, msg_size(msg)); - skb_orphan(buf); - *dnode = msg_orignode(&ohdr); + + /* Now reverse the concerned fields */ + msg_set_errcode(hdr, err); + msg_set_origport(hdr, msg_destport(&ohdr)); + msg_set_destport(hdr, msg_origport(&ohdr)); + msg_set_destnode(hdr, msg_prevnode(&ohdr)); + msg_set_prevnode(hdr, own_node); + msg_set_orignode(hdr, own_node); + msg_set_size(hdr, msg_hdr_sz(hdr) + dlen); + skb_trim(_skb, msg_size(hdr)); + skb_orphan(_skb); return true; exit: - kfree_skb(buf); - *dnode = 0; + kfree_skb(_skb); + *skb = NULL; return false; } /** * tipc_msg_lookup_dest(): try to find new destination for named message * @skb: the buffer containing the message. - * @dnode: return value: next-hop node, if destination found - * @err: return value: error code to use, if message to be rejected + * @err: error code to be used by caller if lookup fails * Does not consume buffer * Returns true if a destination is found, false otherwise */ -bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb, - u32 *dnode, int *err) +bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb, int *err) { struct tipc_msg *msg = buf_msg(skb); - u32 dport; - u32 own_addr = tipc_own_addr(net); + u32 dport, dnode; + u32 onode = tipc_own_addr(net); if (!msg_isdata(msg)) return false; @@ -522,17 +544,18 @@ bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb, *err = -TIPC_ERR_NO_NAME; if (skb_linearize(skb)) return false; + msg = buf_msg(skb); if (msg_reroute_cnt(msg)) return false; - *dnode = addr_domain(net, msg_lookup_scope(msg)); + dnode = addr_domain(net, msg_lookup_scope(msg)); dport = tipc_nametbl_translate(net, msg_nametype(msg), - msg_nameinst(msg), dnode); + msg_nameinst(msg), &dnode); if (!dport) return false; msg_incr_reroute_cnt(msg); - if (*dnode != own_addr) - msg_set_prevnode(msg, own_addr); - msg_set_destnode(msg, *dnode); + if (dnode != onode) + msg_set_prevnode(msg, onode); + msg_set_destnode(msg, dnode); msg_set_destport(msg, dport); *err = TIPC_OK; return true; @@ -541,18 +564,22 @@ bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb, /* tipc_msg_reassemble() - clone a buffer chain of fragments and * reassemble the clones into one message */ -struct sk_buff *tipc_msg_reassemble(struct sk_buff_head *list) +bool tipc_msg_reassemble(struct sk_buff_head *list, struct sk_buff_head *rcvq) { - struct sk_buff *skb; + struct sk_buff *skb, *_skb; struct sk_buff *frag = NULL; struct sk_buff *head = NULL; - int hdr_sz; + int hdr_len; /* Copy header if single buffer */ if (skb_queue_len(list) == 1) { skb = skb_peek(list); - hdr_sz = skb_headroom(skb) + msg_hdr_sz(buf_msg(skb)); - return __pskb_copy(skb, hdr_sz, GFP_ATOMIC); + hdr_len = skb_headroom(skb) + msg_hdr_sz(buf_msg(skb)); + _skb = __pskb_copy(skb, hdr_len, GFP_ATOMIC); + if (!_skb) + return false; + __skb_queue_tail(rcvq, _skb); + return true; } /* Clone all fragments and reassemble */ @@ -566,9 +593,41 @@ struct sk_buff *tipc_msg_reassemble(struct sk_buff_head *list) if (!head) goto error; } - return frag; + __skb_queue_tail(rcvq, frag); + return true; error: pr_warn("Failed do clone local mcast rcv buffer\n"); kfree_skb(head); - return NULL; + return false; +} + +/* tipc_skb_queue_sorted(); sort pkt into list according to sequence number + * @list: list to be appended to + * @seqno: sequence number of buffer to add + * @skb: buffer to add + */ +void __tipc_skb_queue_sorted(struct sk_buff_head *list, u16 seqno, + struct sk_buff *skb) +{ + struct sk_buff *_skb, *tmp; + + if (skb_queue_empty(list) || less(seqno, buf_seqno(skb_peek(list)))) { + __skb_queue_head(list, skb); + return; + } + + if (more(seqno, buf_seqno(skb_peek_tail(list)))) { + __skb_queue_tail(list, skb); + return; + } + + skb_queue_walk_safe(list, _skb, tmp) { + if (more(seqno, buf_seqno(_skb))) + continue; + if (seqno == buf_seqno(_skb)) + break; + __skb_queue_before(list, _skb, skb); + return; + } + kfree_skb(skb); } diff --git a/kernel/net/tipc/msg.h b/kernel/net/tipc/msg.h index e1d3595e2..55778a0ae 100644 --- a/kernel/net/tipc/msg.h +++ b/kernel/net/tipc/msg.h @@ -38,6 +38,7 @@ #define _TIPC_MSG_H #include <linux/tipc.h> +#include "core.h" /* * Constants and routines used to read and write TIPC payload message headers @@ -109,9 +110,9 @@ struct tipc_skb_cb { struct sk_buff *tail; bool validated; bool wakeup_pending; - bool bundling; u16 chain_sz; u16 chain_imp; + u16 ackers; }; #define TIPC_SKB_CB(__skb) ((struct tipc_skb_cb *)&((__skb)->cb[0])) @@ -313,12 +314,12 @@ static inline void msg_set_lookup_scope(struct tipc_msg *m, u32 n) msg_set_bits(m, 1, 19, 0x3, n); } -static inline u32 msg_bcast_ack(struct tipc_msg *m) +static inline u16 msg_bcast_ack(struct tipc_msg *m) { return msg_bits(m, 1, 0, 0xffff); } -static inline void msg_set_bcast_ack(struct tipc_msg *m, u32 n) +static inline void msg_set_bcast_ack(struct tipc_msg *m, u16 n) { msg_set_bits(m, 1, 0, 0xffff, n); } @@ -327,22 +328,22 @@ static inline void msg_set_bcast_ack(struct tipc_msg *m, u32 n) /* * Word 2 */ -static inline u32 msg_ack(struct tipc_msg *m) +static inline u16 msg_ack(struct tipc_msg *m) { return msg_bits(m, 2, 16, 0xffff); } -static inline void msg_set_ack(struct tipc_msg *m, u32 n) +static inline void msg_set_ack(struct tipc_msg *m, u16 n) { msg_set_bits(m, 2, 16, 0xffff, n); } -static inline u32 msg_seqno(struct tipc_msg *m) +static inline u16 msg_seqno(struct tipc_msg *m) { return msg_bits(m, 2, 0, 0xffff); } -static inline void msg_set_seqno(struct tipc_msg *m, u32 n) +static inline void msg_set_seqno(struct tipc_msg *m, u16 n) { msg_set_bits(m, 2, 0, 0xffff, n); } @@ -352,18 +353,22 @@ static inline void msg_set_seqno(struct tipc_msg *m, u32 n) */ static inline u32 msg_importance(struct tipc_msg *m) { - if (unlikely(msg_user(m) == MSG_FRAGMENTER)) - return msg_bits(m, 5, 13, 0x7); - if (likely(msg_isdata(m) && !msg_errcode(m))) - return msg_user(m); + int usr = msg_user(m); + + if (likely((usr <= TIPC_CRITICAL_IMPORTANCE) && !msg_errcode(m))) + return usr; + if ((usr == MSG_FRAGMENTER) || (usr == MSG_BUNDLER)) + return msg_bits(m, 9, 0, 0x7); return TIPC_SYSTEM_IMPORTANCE; } static inline void msg_set_importance(struct tipc_msg *m, u32 i) { - if (unlikely(msg_user(m) == MSG_FRAGMENTER)) - msg_set_bits(m, 5, 13, 0x7, i); - else if (likely(i < TIPC_SYSTEM_IMPORTANCE)) + int usr = msg_user(m); + + if (likely((usr == MSG_FRAGMENTER) || (usr == MSG_BUNDLER))) + msg_set_bits(m, 9, 0, 0x7, i); + else if (i < TIPC_SYSTEM_IMPORTANCE) msg_set_user(m, i); else pr_warn("Trying to set illegal importance in message\n"); @@ -554,15 +559,6 @@ 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 */ @@ -605,6 +601,11 @@ static inline u32 msg_last_bcast(struct tipc_msg *m) return msg_bits(m, 4, 16, 0xffff); } +static inline u32 msg_bc_snd_nxt(struct tipc_msg *m) +{ + return msg_last_bcast(m) + 1; +} + static inline void msg_set_last_bcast(struct tipc_msg *m, u32 n) { msg_set_bits(m, 4, 16, 0xffff, n); @@ -616,12 +617,12 @@ static inline void msg_set_fragm_no(struct tipc_msg *m, u32 n) } -static inline u32 msg_next_sent(struct tipc_msg *m) +static inline u16 msg_next_sent(struct tipc_msg *m) { return msg_bits(m, 4, 0, 0xffff); } -static inline void msg_set_next_sent(struct tipc_msg *m, u32 n) +static inline void msg_set_next_sent(struct tipc_msg *m, u16 n) { msg_set_bits(m, 4, 0, 0xffff, n); } @@ -654,12 +655,12 @@ static inline void msg_set_link_selector(struct tipc_msg *m, u32 n) /* * Word 5 */ -static inline u32 msg_session(struct tipc_msg *m) +static inline u16 msg_session(struct tipc_msg *m) { return msg_bits(m, 5, 16, 0xffff); } -static inline void msg_set_session(struct tipc_msg *m, u32 n) +static inline void msg_set_session(struct tipc_msg *m, u16 n) { msg_set_bits(m, 5, 16, 0xffff, n); } @@ -722,12 +723,12 @@ static inline char *msg_media_addr(struct tipc_msg *m) /* * Word 9 */ -static inline u32 msg_msgcnt(struct tipc_msg *m) +static inline u16 msg_msgcnt(struct tipc_msg *m) { return msg_bits(m, 9, 16, 0xffff); } -static inline void msg_set_msgcnt(struct tipc_msg *m, u32 n) +static inline void msg_set_msgcnt(struct tipc_msg *m, u16 n) { msg_set_bits(m, 9, 16, 0xffff, n); } @@ -762,25 +763,46 @@ static inline void msg_set_link_tolerance(struct tipc_msg *m, u32 n) msg_set_bits(m, 9, 0, 0xffff, n); } +static inline bool msg_peer_link_is_up(struct tipc_msg *m) +{ + if (likely(msg_user(m) != LINK_PROTOCOL)) + return true; + if (msg_type(m) == STATE_MSG) + return true; + return false; +} + +static inline bool msg_peer_node_is_up(struct tipc_msg *m) +{ + if (msg_peer_link_is_up(m)) + return true; + return msg_redundant_link(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, - int err); +bool tipc_msg_reverse(u32 own_addr, struct sk_buff **skb, int err); void tipc_msg_init(u32 own_addr, struct tipc_msg *m, u32 user, u32 type, u32 hsize, u32 destnode); struct sk_buff *tipc_msg_create(uint user, uint type, uint hdr_sz, uint data_sz, u32 dnode, u32 onode, u32 dport, u32 oport, int errcode); int tipc_buf_append(struct sk_buff **headbuf, struct sk_buff **buf); -bool tipc_msg_bundle(struct sk_buff *bskb, struct sk_buff *skb, u32 mtu); - -bool tipc_msg_make_bundle(struct sk_buff **skb, u32 mtu, u32 dnode); +bool tipc_msg_bundle(struct sk_buff *skb, struct tipc_msg *msg, u32 mtu); +bool tipc_msg_make_bundle(struct sk_buff **skb, struct tipc_msg *msg, + u32 mtu, u32 dnode); bool tipc_msg_extract(struct sk_buff *skb, struct sk_buff **iskb, int *pos); int tipc_msg_build(struct tipc_msg *mhdr, struct msghdr *m, int offset, int dsz, int mtu, struct sk_buff_head *list); -bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb, u32 *dnode, - int *err); -struct sk_buff *tipc_msg_reassemble(struct sk_buff_head *list); +bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb, int *err); +bool tipc_msg_reassemble(struct sk_buff_head *list, struct sk_buff_head *rcvq); +void __tipc_skb_queue_sorted(struct sk_buff_head *list, u16 seqno, + struct sk_buff *skb); + +static inline u16 buf_seqno(struct sk_buff *skb) +{ + return msg_seqno(buf_msg(skb)); +} /* tipc_skb_peek(): peek and reserve first buffer in list * @list: list to be peeked in @@ -848,26 +870,33 @@ static inline struct sk_buff *tipc_skb_dequeue(struct sk_buff_head *list, return skb; } -/* tipc_skb_queue_tail(): add buffer to tail of list; - * @list: list to be appended to - * @skb: buffer to append. Always appended - * @dport: the destination port of the buffer - * returns true if dport differs from previous destination +/* tipc_skb_queue_splice_tail - append an skb list to lock protected list + * @list: the new list to append. Not lock protected + * @head: target list. Lock protected. */ -static inline bool tipc_skb_queue_tail(struct sk_buff_head *list, - struct sk_buff *skb, u32 dport) +static inline void tipc_skb_queue_splice_tail(struct sk_buff_head *list, + struct sk_buff_head *head) { - struct sk_buff *_skb = NULL; - bool rv = false; + spin_lock_bh(&head->lock); + skb_queue_splice_tail(list, head); + spin_unlock_bh(&head->lock); +} + +/* tipc_skb_queue_splice_tail_init - merge two lock protected skb lists + * @list: the new list to add. Lock protected. Will be reinitialized + * @head: target list. Lock protected. + */ +static inline void tipc_skb_queue_splice_tail_init(struct sk_buff_head *list, + struct sk_buff_head *head) +{ + struct sk_buff_head tmp; + + __skb_queue_head_init(&tmp); spin_lock_bh(&list->lock); - _skb = skb_peek_tail(list); - if (!_skb || (msg_destport(buf_msg(_skb)) != dport) || - (skb_queue_len(list) > 32)) - rv = true; - __skb_queue_tail(list, skb); + skb_queue_splice_tail_init(list, &tmp); spin_unlock_bh(&list->lock); - return rv; + tipc_skb_queue_splice_tail(&tmp, head); } #endif diff --git a/kernel/net/tipc/name_distr.c b/kernel/net/tipc/name_distr.c index 41e7b7e4d..c07612bab 100644 --- a/kernel/net/tipc/name_distr.c +++ b/kernel/net/tipc/name_distr.c @@ -96,13 +96,13 @@ void named_cluster_distribute(struct net *net, struct sk_buff *skb) dnode = node->addr; if (in_own_node(net, dnode)) continue; - if (!tipc_node_active_links(node)) + if (!tipc_node_is_up(node)) continue; oskb = pskb_copy(skb, GFP_ATOMIC); if (!oskb) break; msg_set_destnode(buf_msg(oskb), dnode); - tipc_link_xmit_skb(net, oskb, dnode, dnode); + tipc_node_xmit_skb(net, oskb, dnode, 0); } rcu_read_unlock(); @@ -223,7 +223,7 @@ void tipc_named_node_up(struct net *net, u32 dnode) &tn->nametbl->publ_list[TIPC_ZONE_SCOPE]); rcu_read_unlock(); - tipc_link_xmit(net, &head, dnode, dnode); + tipc_node_xmit(net, &head, dnode, 0); } static void tipc_publ_subscribe(struct net *net, struct publication *publ, diff --git a/kernel/net/tipc/name_table.c b/kernel/net/tipc/name_table.c index ab0ac62a1..0f47f08bf 100644 --- a/kernel/net/tipc/name_table.c +++ b/kernel/net/tipc/name_table.c @@ -330,13 +330,9 @@ static struct publication *tipc_nameseq_insert_publ(struct net *net, /* Any subscriptions waiting for notification? */ list_for_each_entry_safe(s, st, &nseq->subscriptions, nameseq_list) { - tipc_subscr_report_overlap(s, - publ->lower, - publ->upper, - TIPC_PUBLISHED, - publ->ref, - publ->node, - created_subseq); + tipc_subscrp_report_overlap(s, publ->lower, publ->upper, + TIPC_PUBLISHED, publ->ref, + publ->node, created_subseq); } return publ; } @@ -404,13 +400,9 @@ found: /* Notify any waiting subscriptions */ list_for_each_entry_safe(s, st, &nseq->subscriptions, nameseq_list) { - tipc_subscr_report_overlap(s, - publ->lower, - publ->upper, - TIPC_WITHDRAWN, - publ->ref, - publ->node, - removed_subseq); + tipc_subscrp_report_overlap(s, publ->lower, publ->upper, + TIPC_WITHDRAWN, publ->ref, + publ->node, removed_subseq); } return publ; @@ -432,19 +424,17 @@ static void tipc_nameseq_subscribe(struct name_seq *nseq, return; while (sseq != &nseq->sseqs[nseq->first_free]) { - if (tipc_subscr_overlap(s, sseq->lower, sseq->upper)) { + if (tipc_subscrp_check_overlap(s, sseq->lower, sseq->upper)) { struct publication *crs; struct name_info *info = sseq->info; int must_report = 1; list_for_each_entry(crs, &info->zone_list, zone_list) { - tipc_subscr_report_overlap(s, - sseq->lower, - sseq->upper, - TIPC_PUBLISHED, - crs->ref, - crs->node, - must_report); + tipc_subscrp_report_overlap(s, sseq->lower, + sseq->upper, + TIPC_PUBLISHED, + crs->ref, crs->node, + must_report); must_report = 0; } } diff --git a/kernel/net/tipc/net.c b/kernel/net/tipc/net.c index a54f3cbe2..77bf9113c 100644 --- a/kernel/net/tipc/net.c +++ b/kernel/net/tipc/net.c @@ -40,6 +40,7 @@ #include "subscr.h" #include "socket.h" #include "node.h" +#include "bcast.h" static const struct nla_policy tipc_nl_net_policy[TIPC_NLA_NET_MAX + 1] = { [TIPC_NLA_NET_UNSPEC] = { .type = NLA_UNSPEC }, @@ -111,14 +112,11 @@ int tipc_net_start(struct net *net, u32 addr) { struct tipc_net *tn = net_generic(net, tipc_net_id); char addr_string[16]; - int res; tn->own_addr = addr; tipc_named_reinit(net); tipc_sk_reinit(net); - res = tipc_bclink_init(net); - if (res) - return res; + tipc_bcast_reinit(net); tipc_nametbl_publish(net, TIPC_CFG_SRV, tn->own_addr, tn->own_addr, TIPC_ZONE_SCOPE, 0, tn->own_addr); @@ -141,7 +139,6 @@ void tipc_net_stop(struct net *net) tn->own_addr); rtnl_lock(); tipc_bearer_stop(net); - tipc_bclink_stop(net); tipc_node_stop(net); rtnl_unlock(); diff --git a/kernel/net/tipc/netlink_compat.c b/kernel/net/tipc/netlink_compat.c index ce9121e8e..1eadc95e1 100644 --- a/kernel/net/tipc/netlink_compat.c +++ b/kernel/net/tipc/netlink_compat.c @@ -55,6 +55,7 @@ struct tipc_nl_compat_msg { int rep_type; int rep_size; int req_type; + struct net *net; struct sk_buff *rep; struct tlv_desc *req; struct sock *dst_sk; @@ -68,7 +69,8 @@ struct tipc_nl_compat_cmd_dump { struct tipc_nl_compat_cmd_doit { int (*doit)(struct sk_buff *skb, struct genl_info *info); - int (*transcode)(struct sk_buff *skb, struct tipc_nl_compat_msg *msg); + int (*transcode)(struct tipc_nl_compat_cmd_doit *cmd, + struct sk_buff *skb, struct tipc_nl_compat_msg *msg); }; static int tipc_skb_tailroom(struct sk_buff *skb) @@ -281,7 +283,7 @@ static int __tipc_nl_compat_doit(struct tipc_nl_compat_cmd_doit *cmd, if (!trans_buf) return -ENOMEM; - err = (*cmd->transcode)(trans_buf, msg); + err = (*cmd->transcode)(cmd, trans_buf, msg); if (err) goto trans_out; @@ -353,7 +355,8 @@ static int tipc_nl_compat_bearer_dump(struct tipc_nl_compat_msg *msg, nla_len(bearer[TIPC_NLA_BEARER_NAME])); } -static int tipc_nl_compat_bearer_enable(struct sk_buff *skb, +static int tipc_nl_compat_bearer_enable(struct tipc_nl_compat_cmd_doit *cmd, + struct sk_buff *skb, struct tipc_nl_compat_msg *msg) { struct nlattr *prop; @@ -385,7 +388,8 @@ static int tipc_nl_compat_bearer_enable(struct sk_buff *skb, return 0; } -static int tipc_nl_compat_bearer_disable(struct sk_buff *skb, +static int tipc_nl_compat_bearer_disable(struct tipc_nl_compat_cmd_doit *cmd, + struct sk_buff *skb, struct tipc_nl_compat_msg *msg) { char *name; @@ -576,11 +580,81 @@ static int tipc_nl_compat_link_dump(struct tipc_nl_compat_msg *msg, &link_info, sizeof(link_info)); } -static int tipc_nl_compat_link_set(struct sk_buff *skb, - struct tipc_nl_compat_msg *msg) +static int __tipc_add_link_prop(struct sk_buff *skb, + struct tipc_nl_compat_msg *msg, + struct tipc_link_config *lc) +{ + switch (msg->cmd) { + case TIPC_CMD_SET_LINK_PRI: + return nla_put_u32(skb, TIPC_NLA_PROP_PRIO, ntohl(lc->value)); + case TIPC_CMD_SET_LINK_TOL: + return nla_put_u32(skb, TIPC_NLA_PROP_TOL, ntohl(lc->value)); + case TIPC_CMD_SET_LINK_WINDOW: + return nla_put_u32(skb, TIPC_NLA_PROP_WIN, ntohl(lc->value)); + } + + return -EINVAL; +} + +static int tipc_nl_compat_media_set(struct sk_buff *skb, + struct tipc_nl_compat_msg *msg) { - struct nlattr *link; struct nlattr *prop; + struct nlattr *media; + struct tipc_link_config *lc; + + lc = (struct tipc_link_config *)TLV_DATA(msg->req); + + media = nla_nest_start(skb, TIPC_NLA_MEDIA); + if (!media) + return -EMSGSIZE; + + if (nla_put_string(skb, TIPC_NLA_MEDIA_NAME, lc->name)) + return -EMSGSIZE; + + prop = nla_nest_start(skb, TIPC_NLA_MEDIA_PROP); + if (!prop) + return -EMSGSIZE; + + __tipc_add_link_prop(skb, msg, lc); + nla_nest_end(skb, prop); + nla_nest_end(skb, media); + + return 0; +} + +static int tipc_nl_compat_bearer_set(struct sk_buff *skb, + struct tipc_nl_compat_msg *msg) +{ + struct nlattr *prop; + struct nlattr *bearer; + struct tipc_link_config *lc; + + lc = (struct tipc_link_config *)TLV_DATA(msg->req); + + bearer = nla_nest_start(skb, TIPC_NLA_BEARER); + if (!bearer) + return -EMSGSIZE; + + if (nla_put_string(skb, TIPC_NLA_BEARER_NAME, lc->name)) + return -EMSGSIZE; + + prop = nla_nest_start(skb, TIPC_NLA_BEARER_PROP); + if (!prop) + return -EMSGSIZE; + + __tipc_add_link_prop(skb, msg, lc); + nla_nest_end(skb, prop); + nla_nest_end(skb, bearer); + + return 0; +} + +static int __tipc_nl_compat_link_set(struct sk_buff *skb, + struct tipc_nl_compat_msg *msg) +{ + struct nlattr *prop; + struct nlattr *link; struct tipc_link_config *lc; lc = (struct tipc_link_config *)TLV_DATA(msg->req); @@ -596,24 +670,40 @@ static int tipc_nl_compat_link_set(struct sk_buff *skb, if (!prop) return -EMSGSIZE; - if (msg->cmd == TIPC_CMD_SET_LINK_PRI) { - if (nla_put_u32(skb, TIPC_NLA_PROP_PRIO, ntohl(lc->value))) - return -EMSGSIZE; - } else if (msg->cmd == TIPC_CMD_SET_LINK_TOL) { - if (nla_put_u32(skb, TIPC_NLA_PROP_TOL, ntohl(lc->value))) - return -EMSGSIZE; - } else if (msg->cmd == TIPC_CMD_SET_LINK_WINDOW) { - if (nla_put_u32(skb, TIPC_NLA_PROP_WIN, ntohl(lc->value))) - return -EMSGSIZE; - } - + __tipc_add_link_prop(skb, msg, lc); nla_nest_end(skb, prop); nla_nest_end(skb, link); return 0; } -static int tipc_nl_compat_link_reset_stats(struct sk_buff *skb, +static int tipc_nl_compat_link_set(struct tipc_nl_compat_cmd_doit *cmd, + struct sk_buff *skb, + struct tipc_nl_compat_msg *msg) +{ + struct tipc_link_config *lc; + struct tipc_bearer *bearer; + struct tipc_media *media; + + lc = (struct tipc_link_config *)TLV_DATA(msg->req); + + media = tipc_media_find(lc->name); + if (media) { + cmd->doit = &tipc_nl_media_set; + return tipc_nl_compat_media_set(skb, msg); + } + + bearer = tipc_bearer_find(msg->net, lc->name); + if (bearer) { + cmd->doit = &tipc_nl_bearer_set; + return tipc_nl_compat_bearer_set(skb, msg); + } + + return __tipc_nl_compat_link_set(skb, msg); +} + +static int tipc_nl_compat_link_reset_stats(struct tipc_nl_compat_cmd_doit *cmd, + struct sk_buff *skb, struct tipc_nl_compat_msg *msg) { char *name; @@ -851,7 +941,8 @@ static int tipc_nl_compat_node_dump(struct tipc_nl_compat_msg *msg, sizeof(node_info)); } -static int tipc_nl_compat_net_set(struct sk_buff *skb, +static int tipc_nl_compat_net_set(struct tipc_nl_compat_cmd_doit *cmd, + struct sk_buff *skb, struct tipc_nl_compat_msg *msg) { u32 val; @@ -1007,7 +1098,6 @@ static int tipc_nl_compat_recv(struct sk_buff *skb, struct genl_info *info) struct nlmsghdr *req_nlh; struct nlmsghdr *rep_nlh; struct tipc_genlmsghdr *req_userhdr = info->userhdr; - struct net *net = genl_info_net(info); memset(&msg, 0, sizeof(msg)); @@ -1015,6 +1105,7 @@ static int tipc_nl_compat_recv(struct sk_buff *skb, struct genl_info *info) msg.req = nlmsg_data(req_nlh) + GENL_HDRLEN + TIPC_GENL_HDRLEN; msg.cmd = req_userhdr->cmd; msg.dst_sk = info->dst_sk; + msg.net = genl_info_net(info); if ((msg.cmd & 0xC000) && (!netlink_net_capable(skb, CAP_NET_ADMIN))) { msg.rep = tipc_get_err_tlv(TIPC_CFG_NOT_NET_ADMIN); @@ -1023,14 +1114,14 @@ static int tipc_nl_compat_recv(struct sk_buff *skb, struct genl_info *info) } len = nlmsg_attrlen(req_nlh, GENL_HDRLEN + TIPC_GENL_HDRLEN); - if (TLV_GET_LEN(msg.req) && !TLV_OK(msg.req, len)) { + if (len && !TLV_OK(msg.req, len)) { msg.rep = tipc_get_err_tlv(TIPC_CFG_NOT_SUPPORTED); err = -EOPNOTSUPP; goto send; } err = tipc_nl_compat_handle(&msg); - if (err == -EOPNOTSUPP) + if ((err == -EOPNOTSUPP) || (err == -EPERM)) msg.rep = tipc_get_err_tlv(TIPC_CFG_NOT_SUPPORTED); else if (err == -EINVAL) msg.rep = tipc_get_err_tlv(TIPC_CFG_TLV_ERROR); @@ -1043,7 +1134,7 @@ send: rep_nlh = nlmsg_hdr(msg.rep); memcpy(rep_nlh, info->nlhdr, len); rep_nlh->nlmsg_len = msg.rep->len; - genlmsg_unicast(net, msg.rep, NETLINK_CB(skb).portid); + genlmsg_unicast(msg.net, msg.rep, NETLINK_CB(skb).portid); return err; } diff --git a/kernel/net/tipc/node.c b/kernel/net/tipc/node.c index 22c059ad2..3926b561f 100644 --- a/kernel/net/tipc/node.c +++ b/kernel/net/tipc/node.c @@ -1,7 +1,7 @@ /* * net/tipc/node.c: TIPC node management routines * - * Copyright (c) 2000-2006, 2012-2014, Ericsson AB + * Copyright (c) 2000-2006, 2012-2015, Ericsson AB * Copyright (c) 2005-2006, 2010-2014, Wind River Systems * All rights reserved. * @@ -39,10 +39,42 @@ #include "node.h" #include "name_distr.h" #include "socket.h" +#include "bcast.h" +#include "discover.h" -static void node_lost_contact(struct tipc_node *n_ptr); -static void node_established_contact(struct tipc_node *n_ptr); +/* Node FSM states and events: + */ +enum { + SELF_DOWN_PEER_DOWN = 0xdd, + SELF_UP_PEER_UP = 0xaa, + SELF_DOWN_PEER_LEAVING = 0xd1, + SELF_UP_PEER_COMING = 0xac, + SELF_COMING_PEER_UP = 0xca, + SELF_LEAVING_PEER_DOWN = 0x1d, + NODE_FAILINGOVER = 0xf0, + NODE_SYNCHING = 0xcc +}; + +enum { + SELF_ESTABL_CONTACT_EVT = 0xece, + SELF_LOST_CONTACT_EVT = 0x1ce, + PEER_ESTABL_CONTACT_EVT = 0x9ece, + PEER_LOST_CONTACT_EVT = 0x91ce, + NODE_FAILOVER_BEGIN_EVT = 0xfbe, + NODE_FAILOVER_END_EVT = 0xfee, + NODE_SYNCH_BEGIN_EVT = 0xcbe, + NODE_SYNCH_END_EVT = 0xcee +}; + +static void __tipc_node_link_down(struct tipc_node *n, int *bearer_id, + struct sk_buff_head *xmitq, + struct tipc_media_addr **maddr); +static void tipc_node_link_down(struct tipc_node *n, int bearer_id, + bool delete); +static void node_lost_contact(struct tipc_node *n, struct sk_buff_head *inputq); static void tipc_node_delete(struct tipc_node *node); +static void tipc_node_timeout(unsigned long data); +static void tipc_node_fsm_evt(struct tipc_node *n, int evt); struct tipc_sock_conn { u32 port; @@ -109,7 +141,7 @@ struct tipc_node *tipc_node_find(struct net *net, u32 addr) return NULL; } -struct tipc_node *tipc_node_create(struct net *net, u32 addr) +struct tipc_node *tipc_node_create(struct net *net, u32 addr, u16 capabilities) { struct tipc_net *tn = net_generic(net, tipc_net_id); struct tipc_node *n_ptr, *temp_node; @@ -125,31 +157,66 @@ struct tipc_node *tipc_node_create(struct net *net, u32 addr) } n_ptr->addr = addr; n_ptr->net = net; + n_ptr->capabilities = capabilities; kref_init(&n_ptr->kref); spin_lock_init(&n_ptr->lock); INIT_HLIST_NODE(&n_ptr->hash); INIT_LIST_HEAD(&n_ptr->list); INIT_LIST_HEAD(&n_ptr->publ_list); INIT_LIST_HEAD(&n_ptr->conn_sks); - __skb_queue_head_init(&n_ptr->bclink.deferdq); + skb_queue_head_init(&n_ptr->bc_entry.namedq); + skb_queue_head_init(&n_ptr->bc_entry.inputq1); + __skb_queue_head_init(&n_ptr->bc_entry.arrvq); + skb_queue_head_init(&n_ptr->bc_entry.inputq2); + n_ptr->state = SELF_DOWN_PEER_LEAVING; + n_ptr->signature = INVALID_NODE_SIG; + n_ptr->active_links[0] = INVALID_BEARER_ID; + n_ptr->active_links[1] = INVALID_BEARER_ID; + if (!tipc_link_bc_create(net, tipc_own_addr(net), n_ptr->addr, + U16_MAX, tipc_bc_sndlink(net)->window, + n_ptr->capabilities, + &n_ptr->bc_entry.inputq1, + &n_ptr->bc_entry.namedq, + tipc_bc_sndlink(net), + &n_ptr->bc_entry.link)) { + pr_warn("Broadcast rcv link creation failed, no memory\n"); + kfree(n_ptr); + n_ptr = NULL; + goto exit; + } + tipc_node_get(n_ptr); + setup_timer(&n_ptr->timer, tipc_node_timeout, (unsigned long)n_ptr); + n_ptr->keepalive_intv = U32_MAX; hlist_add_head_rcu(&n_ptr->hash, &tn->node_htable[tipc_hashfn(addr)]); list_for_each_entry_rcu(temp_node, &tn->node_list, list) { if (n_ptr->addr < temp_node->addr) break; } 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_calculate_timer(struct tipc_node *n, struct tipc_link *l) +{ + unsigned long tol = l->tolerance; + unsigned long intv = ((tol / 4) > 500) ? 500 : tol / 4; + unsigned long keepalive_intv = msecs_to_jiffies(intv); + + /* Link with lowest tolerance determines timer interval */ + if (keepalive_intv < n->keepalive_intv) + n->keepalive_intv = keepalive_intv; + + /* Ensure link's abort limit corresponds to current interval */ + l->abort_limit = l->tolerance / jiffies_to_msecs(n->keepalive_intv); +} + static void tipc_node_delete(struct tipc_node *node) { list_del_rcu(&node->list); hlist_del_rcu(&node->hash); + kfree(node->bc_entry.link); kfree_rcu(node, rcu); } @@ -159,8 +226,11 @@ void tipc_node_stop(struct net *net) struct tipc_node *node, *t_node; spin_lock_bh(&tn->node_list_lock); - list_for_each_entry_safe(node, t_node, &tn->node_list, list) + list_for_each_entry_safe(node, t_node, &tn->node_list, list) { + if (del_timer(&node->timer)) + tipc_node_put(node); tipc_node_put(node); + } spin_unlock_bh(&tn->node_list_lock); } @@ -218,212 +288,604 @@ void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port) tipc_node_put(node); } +/* tipc_node_timeout - handle expiration of node timer + */ +static void tipc_node_timeout(unsigned long data) +{ + struct tipc_node *n = (struct tipc_node *)data; + struct tipc_link_entry *le; + struct sk_buff_head xmitq; + int bearer_id; + int rc = 0; + + __skb_queue_head_init(&xmitq); + + for (bearer_id = 0; bearer_id < MAX_BEARERS; bearer_id++) { + tipc_node_lock(n); + le = &n->links[bearer_id]; + if (le->link) { + /* Link tolerance may change asynchronously: */ + tipc_node_calculate_timer(n, le->link); + rc = tipc_link_timeout(le->link, &xmitq); + } + tipc_node_unlock(n); + tipc_bearer_xmit(n->net, bearer_id, &xmitq, &le->maddr); + if (rc & TIPC_LINK_DOWN_EVT) + tipc_node_link_down(n, bearer_id, false); + } + if (!mod_timer(&n->timer, jiffies + n->keepalive_intv)) + tipc_node_get(n); + tipc_node_put(n); +} + /** - * tipc_node_link_up - handle addition of link - * + * __tipc_node_link_up - handle addition of link + * Node lock must be held by caller * Link becomes active (alone or shared) or standby, depending on its priority. */ -void tipc_node_link_up(struct tipc_node *n_ptr, struct tipc_link *l_ptr) +static void __tipc_node_link_up(struct tipc_node *n, int bearer_id, + struct sk_buff_head *xmitq) { - struct tipc_link **active = &n_ptr->active_links[0]; + int *slot0 = &n->active_links[0]; + int *slot1 = &n->active_links[1]; + struct tipc_link *ol = node_active_link(n, 0); + struct tipc_link *nl = n->links[bearer_id].link; - n_ptr->working_links++; - n_ptr->action_flags |= TIPC_NOTIFY_LINK_UP; - n_ptr->link_id = l_ptr->peer_bearer_id << 16 | l_ptr->bearer_id; + if (!nl) + return; - pr_debug("Established link <%s> on network plane %c\n", - l_ptr->name, l_ptr->net_plane); + tipc_link_fsm_evt(nl, LINK_ESTABLISH_EVT); + if (!tipc_link_is_up(nl)) + return; - if (!active[0]) { - active[0] = active[1] = l_ptr; - node_established_contact(n_ptr); - goto exit; - } - if (l_ptr->priority < active[0]->priority) { - pr_debug("New link <%s> becomes standby\n", l_ptr->name); - goto exit; + n->working_links++; + n->action_flags |= TIPC_NOTIFY_LINK_UP; + n->link_id = nl->peer_bearer_id << 16 | bearer_id; + + /* Leave room for tunnel header when returning 'mtu' to users: */ + n->links[bearer_id].mtu = nl->mtu - INT_H_SIZE; + + tipc_bearer_add_dest(n->net, bearer_id, n->addr); + tipc_bcast_inc_bearer_dst_cnt(n->net, bearer_id); + + pr_debug("Established link <%s> on network plane %c\n", + nl->name, nl->net_plane); + + /* First link? => give it both slots */ + if (!ol) { + *slot0 = bearer_id; + *slot1 = bearer_id; + tipc_node_fsm_evt(n, SELF_ESTABL_CONTACT_EVT); + n->action_flags |= TIPC_NOTIFY_NODE_UP; + tipc_bcast_add_peer(n->net, nl, xmitq); + return; } - tipc_link_dup_queue_xmit(active[0], l_ptr); - if (l_ptr->priority == active[0]->priority) { - active[0] = l_ptr; - goto exit; + + /* Second link => redistribute slots */ + if (nl->priority > ol->priority) { + pr_debug("Old link <%s> becomes standby\n", ol->name); + *slot0 = bearer_id; + *slot1 = bearer_id; + tipc_link_set_active(nl, true); + tipc_link_set_active(ol, false); + } else if (nl->priority == ol->priority) { + tipc_link_set_active(nl, true); + *slot1 = bearer_id; + } else { + pr_debug("New link <%s> is standby\n", nl->name); } - pr_debug("Old link <%s> becomes standby\n", active[0]->name); - if (active[1] != active[0]) - pr_debug("Old link <%s> becomes standby\n", active[1]->name); - active[0] = active[1] = l_ptr; -exit: - /* Leave room for changeover header when returning 'mtu' to users: */ - n_ptr->act_mtus[0] = active[0]->mtu - INT_H_SIZE; - n_ptr->act_mtus[1] = active[1]->mtu - INT_H_SIZE; + + /* Prepare synchronization with first link */ + tipc_link_tnl_prepare(ol, nl, SYNCH_MSG, xmitq); } /** - * node_select_active_links - select active link + * tipc_node_link_up - handle addition of link + * + * Link becomes active (alone or shared) or standby, depending on its priority. */ -static void node_select_active_links(struct tipc_node *n_ptr) +static void tipc_node_link_up(struct tipc_node *n, int bearer_id, + struct sk_buff_head *xmitq) { - struct tipc_link **active = &n_ptr->active_links[0]; - u32 i; - u32 highest_prio = 0; + tipc_node_lock(n); + __tipc_node_link_up(n, bearer_id, xmitq); + tipc_node_unlock(n); +} - active[0] = active[1] = NULL; +/** + * __tipc_node_link_down - handle loss of link + */ +static void __tipc_node_link_down(struct tipc_node *n, int *bearer_id, + struct sk_buff_head *xmitq, + struct tipc_media_addr **maddr) +{ + struct tipc_link_entry *le = &n->links[*bearer_id]; + int *slot0 = &n->active_links[0]; + int *slot1 = &n->active_links[1]; + int i, highest = 0; + struct tipc_link *l, *_l, *tnl; + + l = n->links[*bearer_id].link; + if (!l || tipc_link_is_reset(l)) + return; - for (i = 0; i < MAX_BEARERS; i++) { - struct tipc_link *l_ptr = n_ptr->links[i]; + n->working_links--; + n->action_flags |= TIPC_NOTIFY_LINK_DOWN; + n->link_id = l->peer_bearer_id << 16 | *bearer_id; - if (!l_ptr || !tipc_link_is_up(l_ptr) || - (l_ptr->priority < highest_prio)) - continue; + tipc_bearer_remove_dest(n->net, *bearer_id, n->addr); - if (l_ptr->priority > highest_prio) { - highest_prio = l_ptr->priority; - active[0] = active[1] = l_ptr; - } else { - active[1] = l_ptr; + pr_debug("Lost link <%s> on network plane %c\n", + l->name, l->net_plane); + + /* Select new active link if any available */ + *slot0 = INVALID_BEARER_ID; + *slot1 = INVALID_BEARER_ID; + for (i = 0; i < MAX_BEARERS; i++) { + _l = n->links[i].link; + if (!_l || !tipc_link_is_up(_l)) + continue; + if (_l == l) + continue; + if (_l->priority < highest) + continue; + if (_l->priority > highest) { + highest = _l->priority; + *slot0 = i; + *slot1 = i; + continue; } + *slot1 = i; } + + if (!tipc_node_is_up(n)) { + if (tipc_link_peer_is_down(l)) + tipc_node_fsm_evt(n, PEER_LOST_CONTACT_EVT); + tipc_node_fsm_evt(n, SELF_LOST_CONTACT_EVT); + tipc_link_fsm_evt(l, LINK_RESET_EVT); + tipc_link_reset(l); + tipc_link_build_reset_msg(l, xmitq); + *maddr = &n->links[*bearer_id].maddr; + node_lost_contact(n, &le->inputq); + tipc_bcast_dec_bearer_dst_cnt(n->net, *bearer_id); + return; + } + tipc_bcast_dec_bearer_dst_cnt(n->net, *bearer_id); + + /* There is still a working link => initiate failover */ + tnl = node_active_link(n, 0); + tipc_link_fsm_evt(tnl, LINK_SYNCH_END_EVT); + tipc_node_fsm_evt(n, NODE_SYNCH_END_EVT); + n->sync_point = tnl->rcv_nxt + (U16_MAX / 2 - 1); + tipc_link_tnl_prepare(l, tnl, FAILOVER_MSG, xmitq); + tipc_link_reset(l); + tipc_link_fsm_evt(l, LINK_RESET_EVT); + tipc_link_fsm_evt(l, LINK_FAILOVER_BEGIN_EVT); + tipc_node_fsm_evt(n, NODE_FAILOVER_BEGIN_EVT); + *maddr = &n->links[tnl->bearer_id].maddr; + *bearer_id = tnl->bearer_id; } -/** - * tipc_node_link_down - handle loss of link - */ -void tipc_node_link_down(struct tipc_node *n_ptr, struct tipc_link *l_ptr) +static void tipc_node_link_down(struct tipc_node *n, int bearer_id, bool delete) { - struct tipc_net *tn = net_generic(n_ptr->net, tipc_net_id); - struct tipc_link **active; - - n_ptr->working_links--; - n_ptr->action_flags |= TIPC_NOTIFY_LINK_DOWN; - n_ptr->link_id = l_ptr->peer_bearer_id << 16 | l_ptr->bearer_id; + struct tipc_link_entry *le = &n->links[bearer_id]; + struct tipc_link *l = le->link; + struct tipc_media_addr *maddr; + struct sk_buff_head xmitq; - if (!tipc_link_is_active(l_ptr)) { - pr_debug("Lost standby link <%s> on network plane %c\n", - l_ptr->name, l_ptr->net_plane); + if (!l) return; - } - pr_debug("Lost link <%s> on network plane %c\n", - l_ptr->name, l_ptr->net_plane); - - active = &n_ptr->active_links[0]; - if (active[0] == l_ptr) - active[0] = active[1]; - if (active[1] == l_ptr) - active[1] = active[0]; - if (active[0] == l_ptr) - node_select_active_links(n_ptr); - if (tipc_node_is_up(n_ptr)) - tipc_link_failover_send_queue(l_ptr); - else - node_lost_contact(n_ptr); - /* Leave room for changeover header when returning 'mtu' to users: */ - if (active[0]) { - 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; - n_ptr->act_mtus[1] = MAX_MSG_SIZE; + __skb_queue_head_init(&xmitq); + + tipc_node_lock(n); + if (!tipc_link_is_establishing(l)) { + __tipc_node_link_down(n, &bearer_id, &xmitq, &maddr); + if (delete) { + kfree(l); + le->link = NULL; + n->link_cnt--; + } + } else { + /* Defuse pending tipc_node_link_up() */ + tipc_link_fsm_evt(l, LINK_RESET_EVT); } + tipc_node_unlock(n); + tipc_bearer_xmit(n->net, bearer_id, &xmitq, maddr); + tipc_sk_rcv(n->net, &le->inputq); } -int tipc_node_active_links(struct tipc_node *n_ptr) +bool tipc_node_is_up(struct tipc_node *n) { - return n_ptr->active_links[0] != NULL; + return n->active_links[0] != INVALID_BEARER_ID; } -int tipc_node_is_up(struct tipc_node *n_ptr) +void tipc_node_check_dest(struct net *net, u32 onode, + struct tipc_bearer *b, + u16 capabilities, u32 signature, + struct tipc_media_addr *maddr, + bool *respond, bool *dupl_addr) { - return tipc_node_active_links(n_ptr); + struct tipc_node *n; + struct tipc_link *l; + struct tipc_link_entry *le; + bool addr_match = false; + bool sign_match = false; + bool link_up = false; + bool accept_addr = false; + bool reset = true; + char *if_name; + + *dupl_addr = false; + *respond = false; + + n = tipc_node_create(net, onode, capabilities); + if (!n) + return; + + tipc_node_lock(n); + + le = &n->links[b->identity]; + + /* Prepare to validate requesting node's signature and media address */ + l = le->link; + link_up = l && tipc_link_is_up(l); + addr_match = l && !memcmp(&le->maddr, maddr, sizeof(*maddr)); + sign_match = (signature == n->signature); + + /* These three flags give us eight permutations: */ + + if (sign_match && addr_match && link_up) { + /* All is fine. Do nothing. */ + reset = false; + } else if (sign_match && addr_match && !link_up) { + /* Respond. The link will come up in due time */ + *respond = true; + } else if (sign_match && !addr_match && link_up) { + /* Peer has changed i/f address without rebooting. + * If so, the link will reset soon, and the next + * discovery will be accepted. So we can ignore it. + * It may also be an cloned or malicious peer having + * chosen the same node address and signature as an + * existing one. + * Ignore requests until the link goes down, if ever. + */ + *dupl_addr = true; + } else if (sign_match && !addr_match && !link_up) { + /* Peer link has changed i/f address without rebooting. + * It may also be a cloned or malicious peer; we can't + * distinguish between the two. + * The signature is correct, so we must accept. + */ + accept_addr = true; + *respond = true; + } else if (!sign_match && addr_match && link_up) { + /* Peer node rebooted. Two possibilities: + * - Delayed re-discovery; this link endpoint has already + * reset and re-established contact with the peer, before + * receiving a discovery message from that node. + * (The peer happened to receive one from this node first). + * - The peer came back so fast that our side has not + * discovered it yet. Probing from this side will soon + * reset the link, since there can be no working link + * endpoint at the peer end, and the link will re-establish. + * Accept the signature, since it comes from a known peer. + */ + n->signature = signature; + } else if (!sign_match && addr_match && !link_up) { + /* The peer node has rebooted. + * Accept signature, since it is a known peer. + */ + n->signature = signature; + *respond = true; + } else if (!sign_match && !addr_match && link_up) { + /* Peer rebooted with new address, or a new/duplicate peer. + * Ignore until the link goes down, if ever. + */ + *dupl_addr = true; + } else if (!sign_match && !addr_match && !link_up) { + /* Peer rebooted with new address, or it is a new peer. + * Accept signature and address. + */ + n->signature = signature; + accept_addr = true; + *respond = true; + } + + if (!accept_addr) + goto exit; + + /* Now create new link if not already existing */ + if (!l) { + if (n->link_cnt == 2) { + pr_warn("Cannot establish 3rd link to %x\n", n->addr); + goto exit; + } + if_name = strchr(b->name, ':') + 1; + if (!tipc_link_create(net, if_name, b->identity, b->tolerance, + b->net_plane, b->mtu, b->priority, + b->window, mod(tipc_net(net)->random), + tipc_own_addr(net), onode, + n->capabilities, + tipc_bc_sndlink(n->net), n->bc_entry.link, + &le->inputq, + &n->bc_entry.namedq, &l)) { + *respond = false; + goto exit; + } + tipc_link_reset(l); + tipc_link_fsm_evt(l, LINK_RESET_EVT); + if (n->state == NODE_FAILINGOVER) + tipc_link_fsm_evt(l, LINK_FAILOVER_BEGIN_EVT); + le->link = l; + n->link_cnt++; + tipc_node_calculate_timer(n, l); + if (n->link_cnt == 1) + if (!mod_timer(&n->timer, jiffies + n->keepalive_intv)) + tipc_node_get(n); + } + memcpy(&le->maddr, maddr, sizeof(*maddr)); +exit: + tipc_node_unlock(n); + if (reset && !tipc_link_is_reset(l)) + tipc_node_link_down(n, b->identity, false); + tipc_node_put(n); } -void tipc_node_attach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr) +void tipc_node_delete_links(struct net *net, int bearer_id) { - n_ptr->links[l_ptr->bearer_id] = l_ptr; - n_ptr->link_cnt++; + struct tipc_net *tn = net_generic(net, tipc_net_id); + struct tipc_node *n; + + rcu_read_lock(); + list_for_each_entry_rcu(n, &tn->node_list, list) { + tipc_node_link_down(n, bearer_id, true); + } + rcu_read_unlock(); } -void tipc_node_detach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr) +static void tipc_node_reset_links(struct tipc_node *n) { + char addr_string[16]; int i; + pr_warn("Resetting all links to %s\n", + tipc_addr_string_fill(addr_string, n->addr)); + for (i = 0; i < MAX_BEARERS; i++) { - if (l_ptr != n_ptr->links[i]) - continue; - n_ptr->links[i] = NULL; - n_ptr->link_cnt--; + tipc_node_link_down(n, i, false); + } +} + +/* tipc_node_fsm_evt - node finite state machine + * Determines when contact is allowed with peer node + */ +static void tipc_node_fsm_evt(struct tipc_node *n, int evt) +{ + int state = n->state; + + switch (state) { + case SELF_DOWN_PEER_DOWN: + switch (evt) { + case SELF_ESTABL_CONTACT_EVT: + state = SELF_UP_PEER_COMING; + break; + case PEER_ESTABL_CONTACT_EVT: + state = SELF_COMING_PEER_UP; + break; + case SELF_LOST_CONTACT_EVT: + case PEER_LOST_CONTACT_EVT: + break; + case NODE_SYNCH_END_EVT: + case NODE_SYNCH_BEGIN_EVT: + case NODE_FAILOVER_BEGIN_EVT: + case NODE_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + case SELF_UP_PEER_UP: + switch (evt) { + case SELF_LOST_CONTACT_EVT: + state = SELF_DOWN_PEER_LEAVING; + break; + case PEER_LOST_CONTACT_EVT: + state = SELF_LEAVING_PEER_DOWN; + break; + case NODE_SYNCH_BEGIN_EVT: + state = NODE_SYNCHING; + break; + case NODE_FAILOVER_BEGIN_EVT: + state = NODE_FAILINGOVER; + break; + case SELF_ESTABL_CONTACT_EVT: + case PEER_ESTABL_CONTACT_EVT: + case NODE_SYNCH_END_EVT: + case NODE_FAILOVER_END_EVT: + break; + default: + goto illegal_evt; + } + break; + case SELF_DOWN_PEER_LEAVING: + switch (evt) { + case PEER_LOST_CONTACT_EVT: + state = SELF_DOWN_PEER_DOWN; + break; + case SELF_ESTABL_CONTACT_EVT: + case PEER_ESTABL_CONTACT_EVT: + case SELF_LOST_CONTACT_EVT: + break; + case NODE_SYNCH_END_EVT: + case NODE_SYNCH_BEGIN_EVT: + case NODE_FAILOVER_BEGIN_EVT: + case NODE_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + case SELF_UP_PEER_COMING: + switch (evt) { + case PEER_ESTABL_CONTACT_EVT: + state = SELF_UP_PEER_UP; + break; + case SELF_LOST_CONTACT_EVT: + state = SELF_DOWN_PEER_LEAVING; + break; + case SELF_ESTABL_CONTACT_EVT: + case PEER_LOST_CONTACT_EVT: + case NODE_SYNCH_END_EVT: + case NODE_FAILOVER_BEGIN_EVT: + break; + case NODE_SYNCH_BEGIN_EVT: + case NODE_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + case SELF_COMING_PEER_UP: + switch (evt) { + case SELF_ESTABL_CONTACT_EVT: + state = SELF_UP_PEER_UP; + break; + case PEER_LOST_CONTACT_EVT: + state = SELF_LEAVING_PEER_DOWN; + break; + case SELF_LOST_CONTACT_EVT: + case PEER_ESTABL_CONTACT_EVT: + break; + case NODE_SYNCH_END_EVT: + case NODE_SYNCH_BEGIN_EVT: + case NODE_FAILOVER_BEGIN_EVT: + case NODE_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + case SELF_LEAVING_PEER_DOWN: + switch (evt) { + case SELF_LOST_CONTACT_EVT: + state = SELF_DOWN_PEER_DOWN; + break; + case SELF_ESTABL_CONTACT_EVT: + case PEER_ESTABL_CONTACT_EVT: + case PEER_LOST_CONTACT_EVT: + break; + case NODE_SYNCH_END_EVT: + case NODE_SYNCH_BEGIN_EVT: + case NODE_FAILOVER_BEGIN_EVT: + case NODE_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + case NODE_FAILINGOVER: + switch (evt) { + case SELF_LOST_CONTACT_EVT: + state = SELF_DOWN_PEER_LEAVING; + break; + case PEER_LOST_CONTACT_EVT: + state = SELF_LEAVING_PEER_DOWN; + break; + case NODE_FAILOVER_END_EVT: + state = SELF_UP_PEER_UP; + break; + case NODE_FAILOVER_BEGIN_EVT: + case SELF_ESTABL_CONTACT_EVT: + case PEER_ESTABL_CONTACT_EVT: + break; + case NODE_SYNCH_BEGIN_EVT: + case NODE_SYNCH_END_EVT: + default: + goto illegal_evt; + } + break; + case NODE_SYNCHING: + switch (evt) { + case SELF_LOST_CONTACT_EVT: + state = SELF_DOWN_PEER_LEAVING; + break; + case PEER_LOST_CONTACT_EVT: + state = SELF_LEAVING_PEER_DOWN; + break; + case NODE_SYNCH_END_EVT: + state = SELF_UP_PEER_UP; + break; + case NODE_FAILOVER_BEGIN_EVT: + state = NODE_FAILINGOVER; + break; + case NODE_SYNCH_BEGIN_EVT: + case SELF_ESTABL_CONTACT_EVT: + case PEER_ESTABL_CONTACT_EVT: + break; + case NODE_FAILOVER_END_EVT: + default: + goto illegal_evt; + } + break; + default: + pr_err("Unknown node fsm state %x\n", state); + break; } + n->state = state; + return; + +illegal_evt: + pr_err("Illegal node fsm evt %x in state %x\n", evt, state); } -static void node_established_contact(struct tipc_node *n_ptr) +bool tipc_node_filter_pkt(struct tipc_node *n, struct tipc_msg *hdr) { - n_ptr->action_flags |= TIPC_NOTIFY_NODE_UP; - n_ptr->bclink.oos_state = 0; - n_ptr->bclink.acked = tipc_bclink_get_last_sent(n_ptr->net); - tipc_bclink_add_node(n_ptr->net, n_ptr->addr); + int state = n->state; + + if (likely(state == SELF_UP_PEER_UP)) + return true; + + if (state == SELF_LEAVING_PEER_DOWN) + return false; + + if (state == SELF_DOWN_PEER_LEAVING) { + if (msg_peer_node_is_up(hdr)) + return false; + } + + return true; } -static void node_lost_contact(struct tipc_node *n_ptr) +static void node_lost_contact(struct tipc_node *n, + struct sk_buff_head *inputq) { char addr_string[16]; struct tipc_sock_conn *conn, *safe; - struct list_head *conns = &n_ptr->conn_sks; + struct tipc_link *l; + struct list_head *conns = &n->conn_sks; struct sk_buff *skb; - struct tipc_net *tn = net_generic(n_ptr->net, tipc_net_id); uint i; pr_debug("Lost contact with %s\n", - tipc_addr_string_fill(addr_string, n_ptr->addr)); - - /* Flush broadcast link info associated with lost node */ - if (n_ptr->bclink.recv_permitted) { - __skb_queue_purge(&n_ptr->bclink.deferdq); - - if (n_ptr->bclink.reasm_buf) { - kfree_skb(n_ptr->bclink.reasm_buf); - n_ptr->bclink.reasm_buf = NULL; - } - - tipc_bclink_remove_node(n_ptr->net, n_ptr->addr); - tipc_bclink_acknowledge(n_ptr, INVALID_LINK_SEQ); + tipc_addr_string_fill(addr_string, n->addr)); - n_ptr->bclink.recv_permitted = false; - } + /* Clean up broadcast state */ + tipc_bcast_remove_peer(n->net, n->bc_entry.link); /* 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->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); + l = n->links[i].link; + if (l) + tipc_link_fsm_evt(l, LINK_FAILOVER_END_EVT); } - n_ptr->action_flags &= ~TIPC_WAIT_OWN_LINKS_DOWN; - - /* Prevent re-contact with node until cleanup is done */ - n_ptr->action_flags |= TIPC_WAIT_PEER_LINKS_DOWN; - /* Notify publications from this node */ - n_ptr->action_flags |= TIPC_NOTIFY_NODE_DOWN; + n->action_flags |= TIPC_NOTIFY_NODE_DOWN; /* Notify sockets connected to node */ list_for_each_entry_safe(conn, safe, conns, list) { skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE, TIPC_CONN_MSG, - SHORT_H_SIZE, 0, tn->own_addr, + SHORT_H_SIZE, 0, tipc_own_addr(n->net), conn->peer_node, conn->port, conn->peer_port, TIPC_ERR_NO_NODE); - if (likely(skb)) { - skb_queue_tail(n_ptr->inputq, skb); - n_ptr->action_flags |= TIPC_MSG_EVT; - } + if (likely(skb)) + skb_queue_tail(inputq, skb); list_del(&conn->list); kfree(conn); } @@ -452,7 +914,7 @@ int tipc_node_get_linkname(struct net *net, u32 bearer_id, u32 addr, goto exit; tipc_node_lock(node); - link = node->links[bearer_id]; + link = node->links[bearer_id].link; if (link) { strncpy(linkname, link->name, len); err = 0; @@ -470,36 +932,24 @@ void tipc_node_unlock(struct tipc_node *node) u32 flags = node->action_flags; u32 link_id = 0; struct list_head *publ_list; - struct sk_buff_head *inputq = node->inputq; - struct sk_buff_head *namedq; - if (likely(!flags || (flags == TIPC_MSG_EVT))) { - node->action_flags = 0; + if (likely(!flags)) { spin_unlock_bh(&node->lock); - if (flags == TIPC_MSG_EVT) - tipc_sk_rcv(net, inputq); return; } addr = node->addr; link_id = node->link_id; - namedq = node->namedq; publ_list = &node->publ_list; - node->action_flags &= ~(TIPC_MSG_EVT | - 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_BCAST_RESET); + node->action_flags &= ~(TIPC_NOTIFY_NODE_DOWN | TIPC_NOTIFY_NODE_UP | + TIPC_NOTIFY_LINK_DOWN | TIPC_NOTIFY_LINK_UP); spin_unlock_bh(&node->lock); if (flags & TIPC_NOTIFY_NODE_DOWN) tipc_publ_notify(net, publ_list, addr); - if (flags & TIPC_WAKEUP_BCAST_USERS) - tipc_bclink_wakeup_users(net); - if (flags & TIPC_NOTIFY_NODE_UP) tipc_named_node_up(net, addr); @@ -511,17 +961,6 @@ void tipc_node_unlock(struct tipc_node *node) tipc_nametbl_withdraw(net, TIPC_LINK_STATE, addr, link_id, addr); - if (flags & TIPC_MSG_EVT) - tipc_sk_rcv(net, inputq); - - if (flags & TIPC_NAMED_MSG_EVT) - tipc_named_rcv(net, namedq); - - 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 */ @@ -558,6 +997,350 @@ msg_full: return -EMSGSIZE; } +static struct tipc_link *tipc_node_select_link(struct tipc_node *n, int sel, + int *bearer_id, + struct tipc_media_addr **maddr) +{ + int id = n->active_links[sel & 1]; + + if (unlikely(id < 0)) + return NULL; + + *bearer_id = id; + *maddr = &n->links[id].maddr; + return n->links[id].link; +} + +/** + * tipc_node_xmit() is the general link level function for message sending + * @net: the applicable net namespace + * @list: chain of buffers containing message + * @dnode: address of destination node + * @selector: a number used for deterministic link selection + * Consumes the buffer chain, except when returning -ELINKCONG + * Returns 0 if success, otherwise errno: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE + */ +int tipc_node_xmit(struct net *net, struct sk_buff_head *list, + u32 dnode, int selector) +{ + struct tipc_link *l = NULL; + struct tipc_node *n; + struct sk_buff_head xmitq; + struct tipc_media_addr *maddr; + int bearer_id; + int rc = -EHOSTUNREACH; + + __skb_queue_head_init(&xmitq); + n = tipc_node_find(net, dnode); + if (likely(n)) { + tipc_node_lock(n); + l = tipc_node_select_link(n, selector, &bearer_id, &maddr); + if (likely(l)) + rc = tipc_link_xmit(l, list, &xmitq); + tipc_node_unlock(n); + if (unlikely(rc == -ENOBUFS)) + tipc_node_link_down(n, bearer_id, false); + tipc_node_put(n); + } + if (likely(!rc)) { + tipc_bearer_xmit(net, bearer_id, &xmitq, maddr); + return 0; + } + if (likely(in_own_node(net, dnode))) { + tipc_sk_rcv(net, list); + return 0; + } + return rc; +} + +/* tipc_node_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_node_xmit_skb(struct net *net, struct sk_buff *skb, u32 dnode, + u32 selector) +{ + struct sk_buff_head head; + int rc; + + skb_queue_head_init(&head); + __skb_queue_tail(&head, skb); + rc = tipc_node_xmit(net, &head, dnode, selector); + if (rc == -ELINKCONG) + kfree_skb(skb); + return 0; +} + +/** + * tipc_node_bc_rcv - process TIPC broadcast packet arriving from off-node + * @net: the applicable net namespace + * @skb: TIPC packet + * @bearer_id: id of bearer message arrived on + * + * Invoked with no locks held. + */ +static void tipc_node_bc_rcv(struct net *net, struct sk_buff *skb, int bearer_id) +{ + int rc; + struct sk_buff_head xmitq; + struct tipc_bclink_entry *be; + struct tipc_link_entry *le; + struct tipc_msg *hdr = buf_msg(skb); + int usr = msg_user(hdr); + u32 dnode = msg_destnode(hdr); + struct tipc_node *n; + + __skb_queue_head_init(&xmitq); + + /* If NACK for other node, let rcv link for that node peek into it */ + if ((usr == BCAST_PROTOCOL) && (dnode != tipc_own_addr(net))) + n = tipc_node_find(net, dnode); + else + n = tipc_node_find(net, msg_prevnode(hdr)); + if (!n) { + kfree_skb(skb); + return; + } + be = &n->bc_entry; + le = &n->links[bearer_id]; + + rc = tipc_bcast_rcv(net, be->link, skb); + + /* Broadcast link reset may happen at reassembly failure */ + if (rc & TIPC_LINK_DOWN_EVT) + tipc_node_reset_links(n); + + /* Broadcast ACKs are sent on a unicast link */ + if (rc & TIPC_LINK_SND_BC_ACK) { + tipc_node_lock(n); + tipc_link_build_ack_msg(le->link, &xmitq); + tipc_node_unlock(n); + } + + if (!skb_queue_empty(&xmitq)) + tipc_bearer_xmit(net, bearer_id, &xmitq, &le->maddr); + + /* Deliver. 'arrvq' is under inputq2's lock protection */ + if (!skb_queue_empty(&be->inputq1)) { + spin_lock_bh(&be->inputq2.lock); + spin_lock_bh(&be->inputq1.lock); + skb_queue_splice_tail_init(&be->inputq1, &be->arrvq); + spin_unlock_bh(&be->inputq1.lock); + spin_unlock_bh(&be->inputq2.lock); + tipc_sk_mcast_rcv(net, &be->arrvq, &be->inputq2); + } + tipc_node_put(n); +} + +/** + * tipc_node_check_state - check and if necessary update node state + * @skb: TIPC packet + * @bearer_id: identity of bearer delivering the packet + * Returns true if state is ok, otherwise consumes buffer and returns false + */ +static bool tipc_node_check_state(struct tipc_node *n, struct sk_buff *skb, + int bearer_id, struct sk_buff_head *xmitq) +{ + struct tipc_msg *hdr = buf_msg(skb); + int usr = msg_user(hdr); + int mtyp = msg_type(hdr); + u16 oseqno = msg_seqno(hdr); + u16 iseqno = msg_seqno(msg_get_wrapped(hdr)); + u16 exp_pkts = msg_msgcnt(hdr); + u16 rcv_nxt, syncpt, dlv_nxt; + int state = n->state; + struct tipc_link *l, *tnl, *pl = NULL; + struct tipc_media_addr *maddr; + int i, pb_id; + + l = n->links[bearer_id].link; + if (!l) + return false; + rcv_nxt = l->rcv_nxt; + + + if (likely((state == SELF_UP_PEER_UP) && (usr != TUNNEL_PROTOCOL))) + return true; + + /* Find parallel link, if any */ + for (i = 0; i < MAX_BEARERS; i++) { + if ((i != bearer_id) && n->links[i].link) { + pl = n->links[i].link; + break; + } + } + + /* Update node accesibility if applicable */ + if (state == SELF_UP_PEER_COMING) { + if (!tipc_link_is_up(l)) + return true; + if (!msg_peer_link_is_up(hdr)) + return true; + tipc_node_fsm_evt(n, PEER_ESTABL_CONTACT_EVT); + } + + if (state == SELF_DOWN_PEER_LEAVING) { + if (msg_peer_node_is_up(hdr)) + return false; + tipc_node_fsm_evt(n, PEER_LOST_CONTACT_EVT); + } + + /* Ignore duplicate packets */ + if ((usr != LINK_PROTOCOL) && less(oseqno, rcv_nxt)) + return true; + + /* Initiate or update failover mode if applicable */ + if ((usr == TUNNEL_PROTOCOL) && (mtyp == FAILOVER_MSG)) { + syncpt = oseqno + exp_pkts - 1; + if (pl && tipc_link_is_up(pl)) { + pb_id = pl->bearer_id; + __tipc_node_link_down(n, &pb_id, xmitq, &maddr); + tipc_skb_queue_splice_tail_init(pl->inputq, l->inputq); + } + /* If pkts arrive out of order, use lowest calculated syncpt */ + if (less(syncpt, n->sync_point)) + n->sync_point = syncpt; + } + + /* Open parallel link when tunnel link reaches synch point */ + if ((n->state == NODE_FAILINGOVER) && tipc_link_is_up(l)) { + if (!more(rcv_nxt, n->sync_point)) + return true; + tipc_node_fsm_evt(n, NODE_FAILOVER_END_EVT); + if (pl) + tipc_link_fsm_evt(pl, LINK_FAILOVER_END_EVT); + return true; + } + + /* No synching needed if only one link */ + if (!pl || !tipc_link_is_up(pl)) + return true; + + /* Initiate synch mode if applicable */ + if ((usr == TUNNEL_PROTOCOL) && (mtyp == SYNCH_MSG) && (oseqno == 1)) { + syncpt = iseqno + exp_pkts - 1; + if (!tipc_link_is_up(l)) { + tipc_link_fsm_evt(l, LINK_ESTABLISH_EVT); + __tipc_node_link_up(n, bearer_id, xmitq); + } + if (n->state == SELF_UP_PEER_UP) { + n->sync_point = syncpt; + tipc_link_fsm_evt(l, LINK_SYNCH_BEGIN_EVT); + tipc_node_fsm_evt(n, NODE_SYNCH_BEGIN_EVT); + } + if (less(syncpt, n->sync_point)) + n->sync_point = syncpt; + } + + /* Open tunnel link when parallel link reaches synch point */ + if ((n->state == NODE_SYNCHING) && tipc_link_is_synching(l)) { + if (tipc_link_is_synching(l)) { + tnl = l; + } else { + tnl = pl; + pl = l; + } + dlv_nxt = pl->rcv_nxt - mod(skb_queue_len(pl->inputq)); + if (more(dlv_nxt, n->sync_point)) { + tipc_link_fsm_evt(tnl, LINK_SYNCH_END_EVT); + tipc_node_fsm_evt(n, NODE_SYNCH_END_EVT); + return true; + } + if (l == pl) + return true; + if ((usr == TUNNEL_PROTOCOL) && (mtyp == SYNCH_MSG)) + return true; + if (usr == LINK_PROTOCOL) + return true; + return false; + } + return true; +} + +/** + * tipc_rcv - process TIPC packets/messages arriving from off-node + * @net: the applicable net namespace + * @skb: TIPC packet + * @bearer: pointer to bearer message arrived on + * + * Invoked with no locks held. Bearer pointer must point to a valid bearer + * structure (i.e. cannot be NULL), but bearer can be inactive. + */ +void tipc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *b) +{ + struct sk_buff_head xmitq; + struct tipc_node *n; + struct tipc_msg *hdr = buf_msg(skb); + int usr = msg_user(hdr); + int bearer_id = b->identity; + struct tipc_link_entry *le; + u16 bc_ack = msg_bcast_ack(hdr); + int rc = 0; + + __skb_queue_head_init(&xmitq); + + /* Ensure message is well-formed */ + if (unlikely(!tipc_msg_validate(skb))) + goto discard; + + /* Handle arrival of discovery or broadcast packet */ + if (unlikely(msg_non_seq(hdr))) { + if (unlikely(usr == LINK_CONFIG)) + return tipc_disc_rcv(net, skb, b); + else + return tipc_node_bc_rcv(net, skb, bearer_id); + } + + /* Locate neighboring node that sent packet */ + n = tipc_node_find(net, msg_prevnode(hdr)); + if (unlikely(!n)) + goto discard; + le = &n->links[bearer_id]; + + /* Ensure broadcast reception is in synch with peer's send state */ + if (unlikely(usr == LINK_PROTOCOL)) + tipc_bcast_sync_rcv(net, n->bc_entry.link, hdr); + else if (unlikely(n->bc_entry.link->acked != bc_ack)) + tipc_bcast_ack_rcv(net, n->bc_entry.link, bc_ack); + + tipc_node_lock(n); + + /* Is reception permitted at the moment ? */ + if (!tipc_node_filter_pkt(n, hdr)) + goto unlock; + + /* Check and if necessary update node state */ + if (likely(tipc_node_check_state(n, skb, bearer_id, &xmitq))) { + rc = tipc_link_rcv(le->link, skb, &xmitq); + skb = NULL; + } +unlock: + tipc_node_unlock(n); + + if (unlikely(rc & TIPC_LINK_UP_EVT)) + tipc_node_link_up(n, bearer_id, &xmitq); + + if (unlikely(rc & TIPC_LINK_DOWN_EVT)) + tipc_node_link_down(n, bearer_id, false); + + if (unlikely(!skb_queue_empty(&n->bc_entry.namedq))) + tipc_named_rcv(net, &n->bc_entry.namedq); + + if (!skb_queue_empty(&le->inputq)) + tipc_sk_rcv(net, &le->inputq); + + if (!skb_queue_empty(&xmitq)) + tipc_bearer_xmit(net, bearer_id, &xmitq, &le->maddr); + + tipc_node_put(n); +discard: + kfree_skb(skb); +} + int tipc_nl_node_dump(struct sk_buff *skb, struct netlink_callback *cb) { int err; diff --git a/kernel/net/tipc/node.h b/kernel/net/tipc/node.h index 02d5c20dc..6734562d3 100644 --- a/kernel/net/tipc/node.h +++ b/kernel/net/tipc/node.h @@ -45,50 +45,41 @@ /* Out-of-range value for node signature */ #define INVALID_NODE_SIG 0x10000 -#define NODE_HTABLE_SIZE 512 +#define INVALID_BEARER_ID -1 /* Flags used to take different actions according to flag type - * TIPC_WAIT_PEER_LINKS_DOWN: wait to see that peer's links are down - * TIPC_WAIT_OWN_LINKS_DOWN: wait until peer node is declared down * TIPC_NOTIFY_NODE_DOWN: notify node is down * TIPC_NOTIFY_NODE_UP: notify node is up * TIPC_DISTRIBUTE_NAME: publish or withdraw link state name type */ enum { - TIPC_MSG_EVT = 1, - TIPC_WAIT_PEER_LINKS_DOWN = (1 << 1), - TIPC_WAIT_OWN_LINKS_DOWN = (1 << 2), TIPC_NOTIFY_NODE_DOWN = (1 << 3), TIPC_NOTIFY_NODE_UP = (1 << 4), - TIPC_WAKEUP_BCAST_USERS = (1 << 5), 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_RESET = (1 << 10) + TIPC_NOTIFY_LINK_DOWN = (1 << 7) }; -/** - * struct tipc_node_bclink - TIPC node bclink structure - * @acked: sequence # of last outbound b'cast message acknowledged by node - * @last_in: sequence # of last in-sequence b'cast message received from node - * @last_sent: sequence # of last b'cast message sent by node - * @oos_state: state tracker for handling OOS b'cast messages - * @deferred_queue: deferred queue saved OOS b'cast message received from node - * @reasm_buf: broadcast reassembly queue head from node - * @inputq_map: bitmap indicating which inqueues should be kicked - * @recv_permitted: true if node is allowed to receive b'cast messages +/* Optional capabilities supported by this code version */ -struct tipc_node_bclink { - u32 acked; - u32 last_in; - u32 last_sent; - u32 oos_state; - u32 deferred_size; - struct sk_buff_head deferdq; - struct sk_buff *reasm_buf; - int inputq_map; - bool recv_permitted; +enum { + TIPC_BCAST_SYNCH = (1 << 1) +}; + +#define TIPC_NODE_CAPABILITIES TIPC_BCAST_SYNCH + +struct tipc_link_entry { + struct tipc_link *link; + u32 mtu; + struct sk_buff_head inputq; + struct tipc_media_addr maddr; +}; + +struct tipc_bclink_entry { + struct tipc_link *link; + struct sk_buff_head inputq1; + struct sk_buff_head arrvq; + struct sk_buff_head inputq2; + struct sk_buff_head namedq; }; /** @@ -100,11 +91,11 @@ struct tipc_node_bclink { * @hash: links to adjacent nodes in unsorted hash chain * @inputq: pointer to input queue containing messages for msg event * @namedq: pointer to name table input queue with name table messages - * @curr_link: the link holding the node lock, if any - * @active_links: pointers to active links to node - * @links: pointers to all links to node + * @active_links: bearer ids of active links, used as index into links[] array + * @links: array containing references to all links to node * @action_flags: bit mask of different types of node actions - * @bclink: broadcast-related info + * @state: connectivity state vs peer node + * @sync_point: sequence number where synch/failover is finished * @list: links to adjacent nodes in sorted list of cluster's nodes * @working_links: number of working links to node (both active and standby) * @link_cnt: number of links to node @@ -120,14 +111,13 @@ struct tipc_node { spinlock_t lock; struct net *net; struct hlist_node hash; - struct sk_buff_head *inputq; - struct sk_buff_head *namedq; - struct tipc_link *active_links[2]; - u32 act_mtus[2]; - struct tipc_link *links[MAX_BEARERS]; + int active_links[2]; + struct tipc_link_entry links[MAX_BEARERS]; + struct tipc_bclink_entry bc_entry; int action_flags; - struct tipc_node_bclink bclink; struct list_head list; + int state; + u16 sync_point; int link_cnt; u16 working_links; u16 capabilities; @@ -135,25 +125,32 @@ struct tipc_node { u32 link_id; struct list_head publ_list; struct list_head conn_sks; + unsigned long keepalive_intv; + struct timer_list timer; struct rcu_head rcu; }; 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_check_dest(struct net *net, u32 onode, + struct tipc_bearer *bearer, + u16 capabilities, u32 signature, + struct tipc_media_addr *maddr, + bool *respond, bool *dupl_addr); +void tipc_node_delete_links(struct net *net, int bearer_id); void tipc_node_attach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr); void tipc_node_detach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr); -void tipc_node_link_down(struct tipc_node *n_ptr, struct tipc_link *l_ptr); -void tipc_node_link_up(struct tipc_node *n_ptr, struct tipc_link *l_ptr); -int tipc_node_active_links(struct tipc_node *n_ptr); -int tipc_node_is_up(struct tipc_node *n_ptr); +bool tipc_node_is_up(struct tipc_node *n); int tipc_node_get_linkname(struct net *net, u32 bearer_id, u32 node, char *linkname, size_t len); void tipc_node_unlock(struct tipc_node *node); +int tipc_node_xmit(struct net *net, struct sk_buff_head *list, u32 dnode, + int selector); +int tipc_node_xmit_skb(struct net *net, struct sk_buff *skb, u32 dest, + u32 selector); int tipc_node_add_conn(struct net *net, u32 dnode, u32 port, u32 peer_port); void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port); - int tipc_nl_node_dump(struct sk_buff *skb, struct netlink_callback *cb); static inline void tipc_node_lock(struct tipc_node *node) @@ -161,26 +158,30 @@ static inline void tipc_node_lock(struct tipc_node *node) spin_lock_bh(&node->lock); } -static inline bool tipc_node_blocked(struct tipc_node *node) +static inline struct tipc_link *node_active_link(struct tipc_node *n, int sel) { - return (node->action_flags & (TIPC_WAIT_PEER_LINKS_DOWN | - TIPC_NOTIFY_NODE_DOWN | TIPC_WAIT_OWN_LINKS_DOWN)); -} + int bearer_id = n->active_links[sel & 1]; -static inline uint tipc_node_get_mtu(struct net *net, u32 addr, u32 selector) -{ - struct tipc_node *node; - u32 mtu; + if (unlikely(bearer_id == INVALID_BEARER_ID)) + return NULL; - node = tipc_node_find(net, addr); - - if (likely(node)) { - mtu = node->act_mtus[selector & 1]; - tipc_node_put(node); - } else { - mtu = MAX_MSG_SIZE; - } + return n->links[bearer_id].link; +} +static inline unsigned int tipc_node_get_mtu(struct net *net, u32 addr, u32 sel) +{ + struct tipc_node *n; + int bearer_id; + unsigned int mtu = MAX_MSG_SIZE; + + n = tipc_node_find(net, addr); + if (unlikely(!n)) + return mtu; + + bearer_id = n->active_links[sel & 1]; + if (likely(bearer_id != INVALID_BEARER_ID)) + mtu = n->links[bearer_id].mtu; + tipc_node_put(n); return mtu; } diff --git a/kernel/net/tipc/server.c b/kernel/net/tipc/server.c index 77ff03ed1..922e04a43 100644 --- a/kernel/net/tipc/server.c +++ b/kernel/net/tipc/server.c @@ -309,6 +309,10 @@ static int tipc_accept_from_sock(struct tipc_conn *con) /* Notify that new connection is incoming */ newcon->usr_data = s->tipc_conn_new(newcon->conid); + if (!newcon->usr_data) { + sock_release(newsock); + return -ENOMEM; + } /* Wake up receive process in case of 'SYN+' message */ newsock->sk->sk_data_ready(newsock->sk); @@ -321,7 +325,7 @@ static struct socket *tipc_create_listen_sock(struct tipc_conn *con) struct socket *sock = NULL; int ret; - ret = __sock_create(s->net, AF_TIPC, SOCK_SEQPACKET, 0, &sock, 1); + ret = sock_create_kern(s->net, AF_TIPC, SOCK_SEQPACKET, 0, &sock); if (ret < 0) return NULL; ret = kernel_setsockopt(sock, SOL_TIPC, TIPC_IMPORTANCE, diff --git a/kernel/net/tipc/socket.c b/kernel/net/tipc/socket.c index 20cc6df07..b53246fb0 100644 --- a/kernel/net/tipc/socket.c +++ b/kernel/net/tipc/socket.c @@ -41,6 +41,7 @@ #include "link.h" #include "name_distr.h" #include "socket.h" +#include "bcast.h" #define SS_LISTENING -1 /* socket is listening */ #define SS_READY -2 /* socket is connectionless */ @@ -104,6 +105,7 @@ struct tipc_sock { static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *skb); static void tipc_data_ready(struct sock *sk); static void tipc_write_space(struct sock *sk); +static void tipc_sock_destruct(struct sock *sk); static int tipc_release(struct socket *sock); static int tipc_accept(struct socket *sock, struct socket *new_sock, int flags); static int tipc_wait_for_sndmsg(struct socket *sock, long *timeo_p); @@ -247,6 +249,22 @@ static void tsk_advance_rx_queue(struct sock *sk) kfree_skb(__skb_dequeue(&sk->sk_receive_queue)); } +/* tipc_sk_respond() : send response message back to sender + */ +static void tipc_sk_respond(struct sock *sk, struct sk_buff *skb, int err) +{ + u32 selector; + u32 dnode; + u32 onode = tipc_own_addr(sock_net(sk)); + + if (!tipc_msg_reverse(onode, &skb, err)) + return; + + dnode = msg_destnode(buf_msg(skb)); + selector = msg_origport(buf_msg(skb)); + tipc_node_xmit_skb(sock_net(sk), skb, dnode, selector); +} + /** * tsk_rej_rx_queue - reject all buffers in socket receive queue * @@ -255,13 +273,9 @@ static void tsk_advance_rx_queue(struct sock *sk) static void tsk_rej_rx_queue(struct sock *sk) { struct sk_buff *skb; - u32 dnode; - u32 own_node = tsk_own_node(tipc_sk(sk)); - while ((skb = __skb_dequeue(&sk->sk_receive_queue))) { - if (tipc_msg_reverse(own_node, skb, &dnode, TIPC_ERR_NO_PORT)) - tipc_link_xmit_skb(sock_net(sk), skb, dnode, 0); - } + while ((skb = __skb_dequeue(&sk->sk_receive_queue))) + tipc_sk_respond(sk, skb, TIPC_ERR_NO_PORT); } /* tsk_peer_msg - verify if message was sent by connected port's peer @@ -342,7 +356,7 @@ static int tipc_sk_create(struct net *net, struct socket *sock, } /* Allocate socket's protocol area */ - sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto); + sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto, kern); if (sk == NULL) return -ENOMEM; @@ -368,6 +382,7 @@ static int tipc_sk_create(struct net *net, struct socket *sock, sk->sk_rcvbuf = sysctl_tipc_rmem[1]; sk->sk_data_ready = tipc_data_ready; sk->sk_write_space = tipc_write_space; + sk->sk_destruct = tipc_sock_destruct; tsk->conn_timeout = CONN_TIMEOUT_DEFAULT; tsk->sent_unacked = 0; atomic_set(&tsk->dupl_rcvcnt, 0); @@ -409,7 +424,7 @@ static int tipc_release(struct socket *sock) struct net *net; struct tipc_sock *tsk; struct sk_buff *skb; - u32 dnode, probing_state; + u32 dnode; /* * Exit if socket isn't fully initialized (occurs when a failed accept() @@ -440,17 +455,12 @@ static int tipc_release(struct socket *sock) tsk->connected = 0; tipc_node_remove_conn(net, dnode, tsk->portid); } - if (tipc_msg_reverse(tsk_own_node(tsk), skb, &dnode, - TIPC_ERR_NO_PORT)) - tipc_link_xmit_skb(net, skb, dnode, 0); + tipc_sk_respond(sk, skb, TIPC_ERR_NO_PORT); } } tipc_sk_withdraw(tsk, 0, NULL); - probing_state = tsk->probing_state; - if (del_timer_sync(&sk->sk_timer) && - probing_state != TIPC_CONN_PROBING) - sock_put(sk); + sk_stop_timer(sk, &sk->sk_timer); tipc_sk_remove(tsk); if (tsk->connected) { skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE, @@ -458,13 +468,10 @@ static int tipc_release(struct socket *sock) tsk_own_node(tsk), tsk_peer_port(tsk), tsk->portid, TIPC_ERR_NO_PORT); if (skb) - tipc_link_xmit_skb(net, skb, dnode, tsk->portid); + tipc_node_xmit_skb(net, skb, dnode, tsk->portid); tipc_node_remove_conn(net, dnode, tsk->portid); } - /* Discard any remaining (connection-based) messages in receive queue */ - __skb_queue_purge(&sk->sk_receive_queue); - /* Reject any messages that accumulated in backlog queue */ sock->state = SS_DISCONNECTING; release_sock(sk); @@ -681,28 +688,29 @@ static int tipc_sendmcast(struct socket *sock, struct tipc_name_seq *seq, msg_set_hdr_sz(mhdr, MCAST_H_SIZE); new_mtu: - mtu = tipc_bclink_get_mtu(); + mtu = tipc_bcast_get_mtu(net); rc = tipc_msg_build(mhdr, msg, 0, dsz, mtu, pktchain); if (unlikely(rc < 0)) return rc; do { - rc = tipc_bclink_xmit(net, pktchain); - if (likely(rc >= 0)) { - rc = dsz; - break; + rc = tipc_bcast_xmit(net, pktchain); + if (likely(!rc)) + return dsz; + + if (rc == -ELINKCONG) { + tsk->link_cong = 1; + rc = tipc_wait_for_sndmsg(sock, &timeo); + if (!rc) + continue; } + __skb_queue_purge(pktchain); if (rc == -EMSGSIZE) { msg->msg_iter = save; goto new_mtu; } - if (rc != -ELINKCONG) - break; - tipc_sk(sk)->link_cong = 1; - rc = tipc_wait_for_sndmsg(sock, &timeo); - if (rc) - __skb_queue_purge(pktchain); - } while (!rc); + break; + } while (1); return rc; } @@ -765,35 +773,35 @@ void tipc_sk_mcast_rcv(struct net *net, struct sk_buff_head *arrvq, /** * tipc_sk_proto_rcv - receive a connection mng protocol message * @tsk: receiving socket - * @skb: pointer to message buffer. Set to NULL if buffer is consumed. + * @skb: pointer to message buffer. */ -static void tipc_sk_proto_rcv(struct tipc_sock *tsk, struct sk_buff **skb) +static void tipc_sk_proto_rcv(struct tipc_sock *tsk, struct sk_buff *skb) { - struct tipc_msg *msg = buf_msg(*skb); + struct sock *sk = &tsk->sk; + struct tipc_msg *hdr = buf_msg(skb); + int mtyp = msg_type(hdr); int conn_cong; - u32 dnode; - u32 own_node = tsk_own_node(tsk); + /* Ignore if connection cannot be validated: */ - if (!tsk_peer_msg(tsk, msg)) + if (!tsk_peer_msg(tsk, hdr)) goto exit; tsk->probing_state = TIPC_CONN_OK; - if (msg_type(msg) == CONN_ACK) { + if (mtyp == CONN_PROBE) { + msg_set_type(hdr, CONN_PROBE_REPLY); + tipc_sk_respond(sk, skb, TIPC_OK); + return; + } else if (mtyp == CONN_ACK) { conn_cong = tsk_conn_cong(tsk); - tsk->sent_unacked -= msg_msgcnt(msg); + tsk->sent_unacked -= msg_msgcnt(hdr); if (conn_cong) - tsk->sk.sk_write_space(&tsk->sk); - } else if (msg_type(msg) == CONN_PROBE) { - if (tipc_msg_reverse(own_node, *skb, &dnode, TIPC_OK)) { - msg_set_type(msg, CONN_PROBE_REPLY); - return; - } + sk->sk_write_space(sk); + } else if (mtyp != CONN_PROBE_REPLY) { + pr_warn("Received unknown CONN_PROTO msg\n"); } - /* Do nothing if msg_type() == CONN_PROBE_REPLY */ exit: - kfree_skb(*skb); - *skb = NULL; + kfree_skb(skb); } static int tipc_wait_for_sndmsg(struct socket *sock, long *timeo_p) @@ -926,24 +934,25 @@ new_mtu: do { skb = skb_peek(pktchain); TIPC_SKB_CB(skb)->wakeup_pending = tsk->link_cong; - rc = tipc_link_xmit(net, pktchain, dnode, tsk->portid); - if (likely(rc >= 0)) { + rc = tipc_node_xmit(net, pktchain, dnode, tsk->portid); + if (likely(!rc)) { if (sock->state != SS_READY) sock->state = SS_CONNECTING; - rc = dsz; - break; + return dsz; } + if (rc == -ELINKCONG) { + tsk->link_cong = 1; + rc = tipc_wait_for_sndmsg(sock, &timeo); + if (!rc) + continue; + } + __skb_queue_purge(pktchain); if (rc == -EMSGSIZE) { m->msg_iter = save; goto new_mtu; } - if (rc != -ELINKCONG) - break; - tsk->link_cong = 1; - rc = tipc_wait_for_sndmsg(sock, &timeo); - if (rc) - __skb_queue_purge(pktchain); - } while (!rc); + break; + } while (1); return rc; } @@ -1045,15 +1054,16 @@ next: return rc; do { if (likely(!tsk_conn_cong(tsk))) { - rc = tipc_link_xmit(net, pktchain, dnode, portid); + rc = tipc_node_xmit(net, pktchain, dnode, portid); if (likely(!rc)) { tsk->sent_unacked++; sent += send; if (sent == dsz) - break; + return dsz; goto next; } if (rc == -EMSGSIZE) { + __skb_queue_purge(pktchain); tsk->max_pkt = tipc_node_get_mtu(net, dnode, portid); m->msg_iter = save; @@ -1061,13 +1071,13 @@ next: } if (rc != -ELINKCONG) break; + tsk->link_cong = 1; } rc = tipc_wait_for_sndpkt(sock, &timeo); - if (rc) - __skb_queue_purge(pktchain); } while (!rc); + __skb_queue_purge(pktchain); return sent ? sent : rc; } @@ -1223,7 +1233,7 @@ static void tipc_sk_send_ack(struct tipc_sock *tsk, uint ack) return; msg = buf_msg(skb); msg_set_msgcnt(msg, ack); - tipc_link_xmit_skb(net, skb, dnode, msg_link_selector(msg)); + tipc_node_xmit_skb(net, skb, dnode, msg_link_selector(msg)); } static int tipc_wait_for_rcvmsg(struct socket *sock, long *timeop) @@ -1504,87 +1514,91 @@ static void tipc_data_ready(struct sock *sk) rcu_read_unlock(); } +static void tipc_sock_destruct(struct sock *sk) +{ + __skb_queue_purge(&sk->sk_receive_queue); +} + /** * filter_connect - Handle all incoming messages for a connection-based socket * @tsk: TIPC socket * @skb: pointer to message buffer. Set to NULL if buffer is consumed * - * Returns 0 (TIPC_OK) if everything ok, -TIPC_ERR_NO_PORT otherwise + * Returns true if everything ok, false otherwise */ -static int filter_connect(struct tipc_sock *tsk, struct sk_buff **skb) +static bool filter_connect(struct tipc_sock *tsk, struct sk_buff *skb) { struct sock *sk = &tsk->sk; struct net *net = sock_net(sk); struct socket *sock = sk->sk_socket; - struct tipc_msg *msg = buf_msg(*skb); - int retval = -TIPC_ERR_NO_PORT; + struct tipc_msg *hdr = buf_msg(skb); - if (msg_mcast(msg)) - return retval; + if (unlikely(msg_mcast(hdr))) + return false; switch ((int)sock->state) { case SS_CONNECTED: + /* Accept only connection-based messages sent by peer */ - if (tsk_peer_msg(tsk, msg)) { - if (unlikely(msg_errcode(msg))) { - sock->state = SS_DISCONNECTING; - tsk->connected = 0; - /* let timer expire on it's own */ - tipc_node_remove_conn(net, tsk_peer_node(tsk), - tsk->portid); - } - retval = TIPC_OK; + if (unlikely(!tsk_peer_msg(tsk, hdr))) + return false; + + if (unlikely(msg_errcode(hdr))) { + sock->state = SS_DISCONNECTING; + tsk->connected = 0; + /* Let timer expire on it's own */ + tipc_node_remove_conn(net, tsk_peer_node(tsk), + tsk->portid); } - break; + return true; + case SS_CONNECTING: - /* Accept only ACK or NACK message */ - if (unlikely(!msg_connected(msg))) - break; + /* Accept only ACK or NACK message */ + if (unlikely(!msg_connected(hdr))) + return false; - if (unlikely(msg_errcode(msg))) { + if (unlikely(msg_errcode(hdr))) { sock->state = SS_DISCONNECTING; sk->sk_err = ECONNREFUSED; - retval = TIPC_OK; - break; + return true; } - if (unlikely(msg_importance(msg) > TIPC_CRITICAL_IMPORTANCE)) { + if (unlikely(!msg_isdata(hdr))) { sock->state = SS_DISCONNECTING; sk->sk_err = EINVAL; - retval = TIPC_OK; - break; + return true; } - tipc_sk_finish_conn(tsk, msg_origport(msg), msg_orignode(msg)); - msg_set_importance(&tsk->phdr, msg_importance(msg)); + tipc_sk_finish_conn(tsk, msg_origport(hdr), msg_orignode(hdr)); + msg_set_importance(&tsk->phdr, msg_importance(hdr)); sock->state = SS_CONNECTED; - /* If an incoming message is an 'ACK-', it should be - * discarded here because it doesn't contain useful - * data. In addition, we should try to wake up - * connect() routine if sleeping. - */ - if (msg_data_sz(msg) == 0) { - kfree_skb(*skb); - *skb = NULL; - if (waitqueue_active(sk_sleep(sk))) - wake_up_interruptible(sk_sleep(sk)); - } - retval = TIPC_OK; - break; + /* If 'ACK+' message, add to socket receive queue */ + if (msg_data_sz(hdr)) + return true; + + /* If empty 'ACK-' message, wake up sleeping connect() */ + if (waitqueue_active(sk_sleep(sk))) + wake_up_interruptible(sk_sleep(sk)); + + /* 'ACK-' message is neither accepted nor rejected: */ + msg_set_dest_droppable(hdr, 1); + return false; + case SS_LISTENING: case SS_UNCONNECTED: + /* Accept only SYN message */ - if (!msg_connected(msg) && !(msg_errcode(msg))) - retval = TIPC_OK; + if (!msg_connected(hdr) && !(msg_errcode(hdr))) + return true; break; case SS_DISCONNECTING: break; default: pr_err("Unknown socket state %u\n", sock->state); } - return retval; + return false; } /** @@ -1619,61 +1633,70 @@ static unsigned int rcvbuf_limit(struct sock *sk, struct sk_buff *buf) /** * filter_rcv - validate incoming message * @sk: socket - * @skb: pointer to message. Set to NULL if buffer is consumed. + * @skb: pointer to message. * * Enqueues message on receive queue if acceptable; optionally handles * disconnect indication for a connected socket. * * Called with socket lock already taken * - * Returns 0 (TIPC_OK) if message was ok, -TIPC error code if rejected + * Returns true if message was added to socket receive queue, otherwise false */ -static int filter_rcv(struct sock *sk, struct sk_buff **skb) +static bool filter_rcv(struct sock *sk, struct sk_buff *skb) { struct socket *sock = sk->sk_socket; struct tipc_sock *tsk = tipc_sk(sk); - struct tipc_msg *msg = buf_msg(*skb); - unsigned int limit = rcvbuf_limit(sk, *skb); - int rc = TIPC_OK; + struct tipc_msg *hdr = buf_msg(skb); + unsigned int limit = rcvbuf_limit(sk, skb); + int err = TIPC_OK; + int usr = msg_user(hdr); - if (unlikely(msg_user(msg) == CONN_MANAGER)) { + if (unlikely(msg_user(hdr) == CONN_MANAGER)) { tipc_sk_proto_rcv(tsk, skb); - return TIPC_OK; + return false; } - if (unlikely(msg_user(msg) == SOCK_WAKEUP)) { - kfree_skb(*skb); + if (unlikely(usr == SOCK_WAKEUP)) { + kfree_skb(skb); tsk->link_cong = 0; sk->sk_write_space(sk); - *skb = NULL; - return TIPC_OK; + return false; } - /* Reject message if it is wrong sort of message for socket */ - if (msg_type(msg) > TIPC_DIRECT_MSG) - return -TIPC_ERR_NO_PORT; + /* Drop if illegal message type */ + if (unlikely(msg_type(hdr) > TIPC_DIRECT_MSG)) { + kfree_skb(skb); + return false; + } - if (sock->state == SS_READY) { - if (msg_connected(msg)) - return -TIPC_ERR_NO_PORT; - } else { - rc = filter_connect(tsk, skb); - if (rc != TIPC_OK || !*skb) - return rc; + /* Reject if wrong message type for current socket state */ + if (unlikely(sock->state == SS_READY)) { + if (msg_connected(hdr)) { + err = TIPC_ERR_NO_PORT; + goto reject; + } + } else if (unlikely(!filter_connect(tsk, skb))) { + err = TIPC_ERR_NO_PORT; + goto reject; } /* Reject message if there isn't room to queue it */ - if (sk_rmem_alloc_get(sk) + (*skb)->truesize >= limit) - return -TIPC_ERR_OVERLOAD; + if (unlikely(sk_rmem_alloc_get(sk) + skb->truesize >= limit)) { + err = TIPC_ERR_OVERLOAD; + goto reject; + } /* Enqueue message */ - TIPC_SKB_CB(*skb)->handle = NULL; - __skb_queue_tail(&sk->sk_receive_queue, *skb); - skb_set_owner_r(*skb, sk); + TIPC_SKB_CB(skb)->handle = NULL; + __skb_queue_tail(&sk->sk_receive_queue, skb); + skb_set_owner_r(skb, sk); sk->sk_data_ready(sk); - *skb = NULL; - return TIPC_OK; + return true; + +reject: + tipc_sk_respond(sk, skb, err); + return false; } /** @@ -1687,22 +1710,10 @@ static int filter_rcv(struct sock *sk, struct sk_buff **skb) */ static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *skb) { - int err; - atomic_t *dcnt; - u32 dnode; - struct tipc_sock *tsk = tipc_sk(sk); - struct net *net = sock_net(sk); - uint truesize = skb->truesize; + unsigned int truesize = skb->truesize; - err = filter_rcv(sk, &skb); - if (likely(!skb)) { - dcnt = &tsk->dupl_rcvcnt; - if (atomic_read(dcnt) < TIPC_CONN_OVERLOAD_LIMIT) - atomic_add(truesize, dcnt); - return 0; - } - if (!err || tipc_msg_reverse(tsk_own_node(tsk), skb, &dnode, -err)) - tipc_link_xmit_skb(net, skb, dnode, tsk->portid); + if (likely(filter_rcv(sk, skb))) + atomic_add(truesize, &tipc_sk(sk)->dupl_rcvcnt); return 0; } @@ -1712,45 +1723,43 @@ static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *skb) * @inputq: list of incoming buffers with potentially different destinations * @sk: socket where the buffers should be enqueued * @dport: port number for the socket - * @_skb: returned buffer to be forwarded or rejected, if applicable * * Caller must hold socket lock - * - * Returns TIPC_OK if all buffers enqueued, otherwise -TIPC_ERR_OVERLOAD - * or -TIPC_ERR_NO_PORT */ -static int tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk, - u32 dport, struct sk_buff **_skb) +static void tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk, + u32 dport) { unsigned int lim; atomic_t *dcnt; - int err; struct sk_buff *skb; unsigned long time_limit = jiffies + 2; while (skb_queue_len(inputq)) { if (unlikely(time_after_eq(jiffies, time_limit))) - return TIPC_OK; + return; + skb = tipc_skb_dequeue(inputq, dport); if (unlikely(!skb)) - return TIPC_OK; + return; + + /* Add message directly to receive queue if possible */ if (!sock_owned_by_user(sk)) { - err = filter_rcv(sk, &skb); - if (likely(!skb)) - continue; - *_skb = skb; - return err; + filter_rcv(sk, skb); + continue; } + + /* Try backlog, compensating for double-counted bytes */ dcnt = &tipc_sk(sk)->dupl_rcvcnt; if (sk->sk_backlog.len) atomic_set(dcnt, 0); lim = rcvbuf_limit(sk, skb) + atomic_read(dcnt); if (likely(!sk_add_backlog(sk, skb, lim))) continue; - *_skb = skb; - return -TIPC_ERR_OVERLOAD; + + /* Overload => reject message back to sender */ + tipc_sk_respond(sk, skb, TIPC_ERR_OVERLOAD); + break; } - return TIPC_OK; } /** @@ -1758,49 +1767,46 @@ static int tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk, * @inputq: buffer list containing the buffers * Consumes all buffers in list until inputq is empty * Note: may be called in multiple threads referring to the same queue - * Returns 0 if last buffer was accepted, otherwise -EHOSTUNREACH - * Only node local calls check the return value, sending single-buffer queues */ -int tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq) +void tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq) { u32 dnode, dport = 0; int err; - struct sk_buff *skb; struct tipc_sock *tsk; - struct tipc_net *tn; struct sock *sk; + struct sk_buff *skb; while (skb_queue_len(inputq)) { - err = -TIPC_ERR_NO_PORT; - skb = NULL; dport = tipc_skb_peek_port(inputq, dport); tsk = tipc_sk_lookup(net, dport); + if (likely(tsk)) { sk = &tsk->sk; if (likely(spin_trylock_bh(&sk->sk_lock.slock))) { - err = tipc_sk_enqueue(inputq, sk, dport, &skb); + tipc_sk_enqueue(inputq, sk, dport); spin_unlock_bh(&sk->sk_lock.slock); - dport = 0; } sock_put(sk); - } else { - skb = tipc_skb_dequeue(inputq, dport); - } - if (likely(!skb)) continue; - if (tipc_msg_lookup_dest(net, skb, &dnode, &err)) - goto xmit; - if (!err) { - dnode = msg_destnode(buf_msg(skb)); - goto xmit; } - tn = net_generic(net, tipc_net_id); - if (!tipc_msg_reverse(tn->own_addr, skb, &dnode, -err)) + + /* No destination socket => dequeue skb if still there */ + skb = tipc_skb_dequeue(inputq, dport); + if (!skb) + return; + + /* Try secondary lookup if unresolved named message */ + err = TIPC_ERR_NO_PORT; + if (tipc_msg_lookup_dest(net, skb, &err)) + goto xmit; + + /* Prepare for message rejection */ + if (!tipc_msg_reverse(tipc_own_addr(net), &skb, err)) continue; xmit: - tipc_link_xmit_skb(net, skb, dnode, dport); + dnode = msg_destnode(buf_msg(skb)); + tipc_node_xmit_skb(net, skb, dnode, dport); } - return err ? -EHOSTUNREACH : 0; } static int tipc_wait_for_connect(struct socket *sock, long *timeo_p) @@ -2069,7 +2075,10 @@ static int tipc_shutdown(struct socket *sock, int how) struct net *net = sock_net(sk); struct tipc_sock *tsk = tipc_sk(sk); struct sk_buff *skb; - u32 dnode; + u32 dnode = tsk_peer_node(tsk); + u32 dport = tsk_peer_port(tsk); + u32 onode = tipc_own_addr(net); + u32 oport = tsk->portid; int res; if (how != SHUT_RDWR) @@ -2082,6 +2091,8 @@ static int tipc_shutdown(struct socket *sock, int how) case SS_CONNECTED: restart: + dnode = tsk_peer_node(tsk); + /* Disconnect and send a 'FIN+' or 'FIN-' message to peer */ skb = __skb_dequeue(&sk->sk_receive_queue); if (skb) { @@ -2089,19 +2100,13 @@ restart: kfree_skb(skb); goto restart; } - if (tipc_msg_reverse(tsk_own_node(tsk), skb, &dnode, - TIPC_CONN_SHUTDOWN)) - tipc_link_xmit_skb(net, skb, dnode, - tsk->portid); + tipc_sk_respond(sk, skb, TIPC_CONN_SHUTDOWN); } else { - dnode = tsk_peer_node(tsk); - skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE, TIPC_CONN_MSG, SHORT_H_SIZE, - 0, dnode, tsk_own_node(tsk), - tsk_peer_port(tsk), - tsk->portid, TIPC_CONN_SHUTDOWN); - tipc_link_xmit_skb(net, skb, dnode, tsk->portid); + 0, dnode, onode, dport, oport, + TIPC_CONN_SHUTDOWN); + tipc_node_xmit_skb(net, skb, dnode, tsk->portid); } tsk->connected = 0; sock->state = SS_DISCONNECTING; @@ -2163,7 +2168,7 @@ static void tipc_sk_timeout(unsigned long data) } bh_unlock_sock(sk); if (skb) - tipc_link_xmit_skb(sock_net(sk), skb, peer_node, tsk->portid); + tipc_node_xmit_skb(sock_net(sk), skb, peer_node, tsk->portid); exit: sock_put(sk); } diff --git a/kernel/net/tipc/socket.h b/kernel/net/tipc/socket.h index bf6551389..4241f2206 100644 --- a/kernel/net/tipc/socket.h +++ b/kernel/net/tipc/socket.h @@ -44,7 +44,7 @@ SKB_TRUESIZE(TIPC_MAX_USER_MSG_SIZE)) int tipc_socket_init(void); void tipc_socket_stop(void); -int tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq); +void 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); void tipc_sk_reinit(struct net *net); diff --git a/kernel/net/tipc/subscr.c b/kernel/net/tipc/subscr.c index 1c147c869..69ee2eeef 100644 --- a/kernel/net/tipc/subscr.c +++ b/kernel/net/tipc/subscr.c @@ -40,16 +40,21 @@ /** * struct tipc_subscriber - TIPC network topology subscriber + * @kref: reference counter to tipc_subscription object * @conid: connection identifier to server connecting to subscriber * @lock: control access to subscriber - * @subscription_list: list of subscription objects for this subscriber + * @subscrp_list: list of subscription objects for this subscriber */ struct tipc_subscriber { + struct kref kref; int conid; spinlock_t lock; - struct list_head subscription_list; + struct list_head subscrp_list; }; +static void tipc_subscrp_delete(struct tipc_subscription *sub); +static void tipc_subscrb_put(struct tipc_subscriber *subscriber); + /** * htohl - convert value to endianness used by destination * @in: value to convert @@ -62,9 +67,9 @@ static u32 htohl(u32 in, int swap) return swap ? swab32(in) : in; } -static void subscr_send_event(struct tipc_subscription *sub, u32 found_lower, - u32 found_upper, u32 event, u32 port_ref, - u32 node) +static void tipc_subscrp_send_event(struct tipc_subscription *sub, + u32 found_lower, u32 found_upper, + u32 event, u32 port_ref, u32 node) { struct tipc_net *tn = net_generic(sub->net, tipc_net_id); struct tipc_subscriber *subscriber = sub->subscriber; @@ -82,12 +87,13 @@ static void subscr_send_event(struct tipc_subscription *sub, u32 found_lower, } /** - * tipc_subscr_overlap - test for subscription overlap with the given values + * tipc_subscrp_check_overlap - test for subscription overlap with the + * given values * * Returns 1 if there is overlap, otherwise 0. */ -int tipc_subscr_overlap(struct tipc_subscription *sub, u32 found_lower, - u32 found_upper) +int tipc_subscrp_check_overlap(struct tipc_subscription *sub, u32 found_lower, + u32 found_upper) { if (found_lower < sub->seq.lower) found_lower = sub->seq.lower; @@ -98,138 +104,121 @@ int tipc_subscr_overlap(struct tipc_subscription *sub, u32 found_lower, return 1; } -/** - * tipc_subscr_report_overlap - issue event if there is subscription overlap - * - * Protected by nameseq.lock in name_table.c - */ -void tipc_subscr_report_overlap(struct tipc_subscription *sub, u32 found_lower, - u32 found_upper, u32 event, u32 port_ref, - u32 node, int must) +void tipc_subscrp_report_overlap(struct tipc_subscription *sub, u32 found_lower, + u32 found_upper, u32 event, u32 port_ref, + u32 node, int must) { - if (!tipc_subscr_overlap(sub, found_lower, found_upper)) + if (!tipc_subscrp_check_overlap(sub, found_lower, found_upper)) return; if (!must && !(sub->filter & TIPC_SUB_PORTS)) return; - subscr_send_event(sub, found_lower, found_upper, event, port_ref, node); + tipc_subscrp_send_event(sub, found_lower, found_upper, event, port_ref, + node); } -static void subscr_timeout(unsigned long data) +static void tipc_subscrp_timeout(unsigned long data) { struct tipc_subscription *sub = (struct tipc_subscription *)data; struct tipc_subscriber *subscriber = sub->subscriber; - struct tipc_net *tn = net_generic(sub->net, tipc_net_id); - /* The spin lock per subscriber is used to protect its members */ - spin_lock_bh(&subscriber->lock); + /* Notify subscriber of timeout */ + tipc_subscrp_send_event(sub, sub->evt.s.seq.lower, sub->evt.s.seq.upper, + TIPC_SUBSCR_TIMEOUT, 0, 0); - /* Validate timeout (in case subscription is being cancelled) */ - if (sub->timeout == TIPC_WAIT_FOREVER) { - spin_unlock_bh(&subscriber->lock); - return; - } + spin_lock_bh(&subscriber->lock); + tipc_subscrp_delete(sub); + spin_unlock_bh(&subscriber->lock); - /* Unlink subscription from name table */ - tipc_nametbl_unsubscribe(sub); + tipc_subscrb_put(subscriber); +} - /* Unlink subscription from subscriber */ - list_del(&sub->subscription_list); +static void tipc_subscrb_kref_release(struct kref *kref) +{ + struct tipc_subscriber *subcriber = container_of(kref, + struct tipc_subscriber, kref); - spin_unlock_bh(&subscriber->lock); + kfree(subcriber); +} - /* Notify subscriber of timeout */ - subscr_send_event(sub, sub->evt.s.seq.lower, sub->evt.s.seq.upper, - TIPC_SUBSCR_TIMEOUT, 0, 0); +static void tipc_subscrb_put(struct tipc_subscriber *subscriber) +{ + kref_put(&subscriber->kref, tipc_subscrb_kref_release); +} - /* Now destroy subscription */ - kfree(sub); - atomic_dec(&tn->subscription_count); +static void tipc_subscrb_get(struct tipc_subscriber *subscriber) +{ + kref_get(&subscriber->kref); } -/** - * subscr_del - delete a subscription within a subscription list - * - * Called with subscriber lock held. - */ -static void subscr_del(struct tipc_subscription *sub) +static struct tipc_subscriber *tipc_subscrb_create(int conid) { - struct tipc_net *tn = net_generic(sub->net, tipc_net_id); + struct tipc_subscriber *subscriber; - tipc_nametbl_unsubscribe(sub); - list_del(&sub->subscription_list); - kfree(sub); - atomic_dec(&tn->subscription_count); + subscriber = kzalloc(sizeof(*subscriber), GFP_ATOMIC); + if (!subscriber) { + pr_warn("Subscriber rejected, no memory\n"); + return NULL; + } + kref_init(&subscriber->kref); + INIT_LIST_HEAD(&subscriber->subscrp_list); + subscriber->conid = conid; + spin_lock_init(&subscriber->lock); + + return subscriber; } -static void subscr_release(struct tipc_subscriber *subscriber) +static void tipc_subscrb_delete(struct tipc_subscriber *subscriber) { - struct tipc_subscription *sub; - struct tipc_subscription *sub_temp; + struct tipc_subscription *sub, *temp; spin_lock_bh(&subscriber->lock); - /* Destroy any existing subscriptions for subscriber */ - list_for_each_entry_safe(sub, sub_temp, &subscriber->subscription_list, - subscription_list) { - if (sub->timeout != TIPC_WAIT_FOREVER) { - spin_unlock_bh(&subscriber->lock); - del_timer_sync(&sub->timer); - spin_lock_bh(&subscriber->lock); + list_for_each_entry_safe(sub, temp, &subscriber->subscrp_list, + subscrp_list) { + if (del_timer(&sub->timer)) { + tipc_subscrp_delete(sub); + tipc_subscrb_put(subscriber); } - subscr_del(sub); } spin_unlock_bh(&subscriber->lock); - /* Now destroy subscriber */ - kfree(subscriber); + tipc_subscrb_put(subscriber); } -/** - * subscr_cancel - handle subscription cancellation request - * - * Called with subscriber lock held. Routine must temporarily release lock - * to enable the subscription timeout routine to finish without deadlocking; - * the lock is then reclaimed to allow caller to release it upon return. - * - * Note that fields of 's' use subscriber's endianness! - */ -static void subscr_cancel(struct tipc_subscr *s, - struct tipc_subscriber *subscriber) +static void tipc_subscrp_delete(struct tipc_subscription *sub) { - struct tipc_subscription *sub; - struct tipc_subscription *sub_temp; - int found = 0; + struct tipc_net *tn = net_generic(sub->net, tipc_net_id); + + tipc_nametbl_unsubscribe(sub); + list_del(&sub->subscrp_list); + kfree(sub); + atomic_dec(&tn->subscription_count); +} + +static void tipc_subscrp_cancel(struct tipc_subscr *s, + struct tipc_subscriber *subscriber) +{ + struct tipc_subscription *sub, *temp; + spin_lock_bh(&subscriber->lock); /* Find first matching subscription, exit if not found */ - list_for_each_entry_safe(sub, sub_temp, &subscriber->subscription_list, - subscription_list) { + list_for_each_entry_safe(sub, temp, &subscriber->subscrp_list, + subscrp_list) { if (!memcmp(s, &sub->evt.s, sizeof(struct tipc_subscr))) { - found = 1; + if (del_timer(&sub->timer)) { + tipc_subscrp_delete(sub); + tipc_subscrb_put(subscriber); + } break; } } - if (!found) - return; - - /* Cancel subscription timer (if used), then delete subscription */ - if (sub->timeout != TIPC_WAIT_FOREVER) { - sub->timeout = TIPC_WAIT_FOREVER; - spin_unlock_bh(&subscriber->lock); - del_timer_sync(&sub->timer); - spin_lock_bh(&subscriber->lock); - } - subscr_del(sub); + spin_unlock_bh(&subscriber->lock); } -/** - * subscr_subscribe - create subscription for subscriber - * - * Called with subscriber lock held. - */ -static int subscr_subscribe(struct net *net, struct tipc_subscr *s, - struct tipc_subscriber *subscriber, - struct tipc_subscription **sub_p) +static int tipc_subscrp_create(struct net *net, struct tipc_subscr *s, + struct tipc_subscriber *subscriber, + struct tipc_subscription **sub_p) { struct tipc_net *tn = net_generic(net, tipc_net_id); struct tipc_subscription *sub; @@ -241,7 +230,7 @@ static int subscr_subscribe(struct net *net, struct tipc_subscr *s, /* Detect & process a subscription cancellation request */ if (s->filter & htohl(TIPC_SUB_CANCEL, swap)) { s->filter &= ~htohl(TIPC_SUB_CANCEL, swap); - subscr_cancel(s, subscriber); + tipc_subscrp_cancel(s, subscriber); return 0; } @@ -273,62 +262,50 @@ static int subscr_subscribe(struct net *net, struct tipc_subscr *s, kfree(sub); return -EINVAL; } - list_add(&sub->subscription_list, &subscriber->subscription_list); + spin_lock_bh(&subscriber->lock); + list_add(&sub->subscrp_list, &subscriber->subscrp_list); + spin_unlock_bh(&subscriber->lock); sub->subscriber = subscriber; sub->swap = swap; - memcpy(&sub->evt.s, s, sizeof(struct tipc_subscr)); + memcpy(&sub->evt.s, s, sizeof(*s)); atomic_inc(&tn->subscription_count); - if (sub->timeout != TIPC_WAIT_FOREVER) { - setup_timer(&sub->timer, subscr_timeout, (unsigned long)sub); - mod_timer(&sub->timer, jiffies + sub->timeout); - } + setup_timer(&sub->timer, tipc_subscrp_timeout, (unsigned long)sub); + if (sub->timeout != TIPC_WAIT_FOREVER) + sub->timeout += jiffies; + if (!mod_timer(&sub->timer, sub->timeout)) + tipc_subscrb_get(subscriber); *sub_p = sub; return 0; } /* Handle one termination request for the subscriber */ -static void subscr_conn_shutdown_event(int conid, void *usr_data) +static void tipc_subscrb_shutdown_cb(int conid, void *usr_data) { - subscr_release((struct tipc_subscriber *)usr_data); + tipc_subscrb_delete((struct tipc_subscriber *)usr_data); } /* Handle one request to create a new subscription for the subscriber */ -static void subscr_conn_msg_event(struct net *net, int conid, - struct sockaddr_tipc *addr, void *usr_data, - void *buf, size_t len) +static void tipc_subscrb_rcv_cb(struct net *net, int conid, + struct sockaddr_tipc *addr, void *usr_data, + void *buf, size_t len) { - struct tipc_subscriber *subscriber = usr_data; + struct tipc_subscriber *subscrb = usr_data; struct tipc_subscription *sub = NULL; struct tipc_net *tn = net_generic(net, tipc_net_id); - spin_lock_bh(&subscriber->lock); - subscr_subscribe(net, (struct tipc_subscr *)buf, subscriber, &sub); - if (sub) - tipc_nametbl_subscribe(sub); - else - tipc_conn_terminate(tn->topsrv, subscriber->conid); - spin_unlock_bh(&subscriber->lock); + if (tipc_subscrp_create(net, (struct tipc_subscr *)buf, subscrb, &sub)) + return tipc_conn_terminate(tn->topsrv, subscrb->conid); + + tipc_nametbl_subscribe(sub); } /* Handle one request to establish a new subscriber */ -static void *subscr_named_msg_event(int conid) +static void *tipc_subscrb_connect_cb(int conid) { - struct tipc_subscriber *subscriber; - - /* Create subscriber object */ - subscriber = kzalloc(sizeof(struct tipc_subscriber), GFP_ATOMIC); - if (subscriber == NULL) { - pr_warn("Subscriber rejected, no memory\n"); - return NULL; - } - INIT_LIST_HEAD(&subscriber->subscription_list); - subscriber->conid = conid; - spin_lock_init(&subscriber->lock); - - return (void *)subscriber; + return (void *)tipc_subscrb_create(conid); } -int tipc_subscr_start(struct net *net) +int tipc_topsrv_start(struct net *net) { struct tipc_net *tn = net_generic(net, tipc_net_id); const char name[] = "topology_server"; @@ -355,9 +332,9 @@ int tipc_subscr_start(struct net *net) topsrv->imp = TIPC_CRITICAL_IMPORTANCE; topsrv->type = SOCK_SEQPACKET; topsrv->max_rcvbuf_size = sizeof(struct tipc_subscr); - topsrv->tipc_conn_recvmsg = subscr_conn_msg_event; - topsrv->tipc_conn_new = subscr_named_msg_event; - topsrv->tipc_conn_shutdown = subscr_conn_shutdown_event; + topsrv->tipc_conn_recvmsg = tipc_subscrb_rcv_cb; + topsrv->tipc_conn_new = tipc_subscrb_connect_cb; + topsrv->tipc_conn_shutdown = tipc_subscrb_shutdown_cb; strncpy(topsrv->name, name, strlen(name) + 1); tn->topsrv = topsrv; @@ -366,7 +343,7 @@ int tipc_subscr_start(struct net *net) return tipc_server_start(topsrv); } -void tipc_subscr_stop(struct net *net) +void tipc_topsrv_stop(struct net *net) { struct tipc_net *tn = net_generic(net, tipc_net_id); struct tipc_server *topsrv = tn->topsrv; diff --git a/kernel/net/tipc/subscr.h b/kernel/net/tipc/subscr.h index 33488bd9f..92ee18cc5 100644 --- a/kernel/net/tipc/subscr.h +++ b/kernel/net/tipc/subscr.h @@ -54,7 +54,7 @@ struct tipc_subscriber; * @filter: event filtering to be done for subscription * @timer: timer governing subscription duration (optional) * @nameseq_list: adjacent subscriptions in name sequence's subscription list - * @subscription_list: adjacent subscriptions in subscriber's subscription list + * @subscrp_list: adjacent subscriptions in subscriber's subscription list * @server_ref: object reference of server port associated with subscription * @swap: indicates if subscriber uses opposite endianness in its messages * @evt: template for events generated by subscription @@ -67,17 +67,17 @@ struct tipc_subscription { u32 filter; struct timer_list timer; struct list_head nameseq_list; - struct list_head subscription_list; + struct list_head subscrp_list; int swap; struct tipc_event evt; }; -int tipc_subscr_overlap(struct tipc_subscription *sub, u32 found_lower, - u32 found_upper); -void tipc_subscr_report_overlap(struct tipc_subscription *sub, u32 found_lower, - u32 found_upper, u32 event, u32 port_ref, - u32 node, int must); -int tipc_subscr_start(struct net *net); -void tipc_subscr_stop(struct net *net); +int tipc_subscrp_check_overlap(struct tipc_subscription *sub, u32 found_lower, + u32 found_upper); +void tipc_subscrp_report_overlap(struct tipc_subscription *sub, + u32 found_lower, u32 found_upper, u32 event, + u32 port_ref, u32 node, int must); +int tipc_topsrv_start(struct net *net); +void tipc_topsrv_stop(struct net *net); #endif diff --git a/kernel/net/tipc/udp_media.c b/kernel/net/tipc/udp_media.c index 66deebc66..70c03271b 100644 --- a/kernel/net/tipc/udp_media.c +++ b/kernel/net/tipc/udp_media.c @@ -48,10 +48,13 @@ #include <linux/tipc_netlink.h> #include "core.h" #include "bearer.h" +#include "msg.h" /* IANA assigned UDP port */ #define UDP_PORT_DEFAULT 6118 +#define UDP_MIN_HEADROOM 28 + static const struct nla_policy tipc_nl_udp_policy[TIPC_NLA_UDP_MAX + 1] = { [TIPC_NLA_UDP_UNSPEC] = {.type = NLA_UNSPEC}, [TIPC_NLA_UDP_LOCAL] = {.type = NLA_BINARY, @@ -153,11 +156,15 @@ static int tipc_udp_send_msg(struct net *net, struct sk_buff *skb, struct udp_bearer *ub; struct udp_media_addr *dst = (struct udp_media_addr *)&dest->value; struct udp_media_addr *src = (struct udp_media_addr *)&b->addr.value; - struct sk_buff *clone; struct rtable *rt; - clone = skb_clone(skb, GFP_ATOMIC); - skb_set_inner_protocol(clone, htons(ETH_P_TIPC)); + if (skb_headroom(skb) < UDP_MIN_HEADROOM) { + err = pskb_expand_head(skb, UDP_MIN_HEADROOM, 0, GFP_ATOMIC); + if (err) + goto tx_error; + } + + skb_set_inner_protocol(skb, htons(ETH_P_TIPC)); ub = rcu_dereference_rtnl(b->media_ptr); if (!ub) { err = -ENODEV; @@ -167,7 +174,7 @@ static int tipc_udp_send_msg(struct net *net, struct sk_buff *skb, struct flowi4 fl = { .daddr = dst->ipv4.s_addr, .saddr = src->ipv4.s_addr, - .flowi4_mark = clone->mark, + .flowi4_mark = skb->mark, .flowi4_proto = IPPROTO_UDP }; rt = ip_route_output_key(net, &fl); @@ -176,7 +183,7 @@ static int tipc_udp_send_msg(struct net *net, struct sk_buff *skb, goto tx_error; } ttl = ip4_dst_hoplimit(&rt->dst); - err = udp_tunnel_xmit_skb(rt, ub->ubsock->sk, clone, + err = udp_tunnel_xmit_skb(rt, ub->ubsock->sk, skb, src->ipv4.s_addr, dst->ipv4.s_addr, 0, ttl, 0, src->udp_port, dst->udp_port, @@ -194,11 +201,12 @@ static int tipc_udp_send_msg(struct net *net, struct sk_buff *skb, .saddr = src->ipv6, .flowi6_proto = IPPROTO_UDP }; - err = ipv6_stub->ipv6_dst_lookup(ub->ubsock->sk, &ndst, &fl6); + err = ipv6_stub->ipv6_dst_lookup(net, ub->ubsock->sk, &ndst, + &fl6); if (err) goto tx_error; ttl = ip6_dst_hoplimit(ndst); - err = udp_tunnel6_xmit_skb(ndst, ub->ubsock->sk, clone, + err = udp_tunnel6_xmit_skb(ndst, ub->ubsock->sk, skb, ndst->dev, &src->ipv6, &dst->ipv6, 0, ttl, src->udp_port, dst->udp_port, false); @@ -207,7 +215,7 @@ static int tipc_udp_send_msg(struct net *net, struct sk_buff *skb, return err; tx_error: - kfree_skb(clone); + kfree_skb(skb); return err; } @@ -216,6 +224,10 @@ static int tipc_udp_recv(struct sock *sk, struct sk_buff *skb) { struct udp_bearer *ub; struct tipc_bearer *b; + int usr = msg_user(buf_msg(skb)); + + if ((usr == LINK_PROTOCOL) || (usr == NAME_DISTRIBUTOR)) + skb_linearize(skb); ub = rcu_dereference_sk_user_data(sk); if (!ub) { @@ -424,7 +436,6 @@ static void tipc_udp_disable(struct tipc_bearer *b) } if (ub->ubsock) sock_set_flag(ub->ubsock->sk, SOCK_DEAD); - RCU_INIT_POINTER(b->media_ptr, NULL); RCU_INIT_POINTER(ub->bearer, NULL); /* sock_release need to be done outside of rtnl lock */ |