From e4de5fab806f74622600ab7fd6ed22b7f911a8c5 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:31 -0500 Subject: tipc: use negative error return values in functions In some places, TIPC functions returns positive integers as return codes. This goes against standard Linux coding practice, and may even cause problems in some cases. We now change the return values of the functions filter_rcv() and filter_connect() to become signed integers, and return negative error codes when needed. The codes we use in these particular cases are still TIPC specific, since they are both part of the TIPC API and have no correspondence in errno.h Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 46 +++++++++++++++++++++++----------------------- 1 file changed, 23 insertions(+), 23 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index ef0475568f9e..9a95dab13b7e 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -1269,17 +1269,16 @@ static void tipc_data_ready(struct sock *sk) * @tsk: TIPC socket * @msg: message * - * Returns TIPC error status code and socket error status code - * once it encounters some errors + * Returns 0 (TIPC_OK) if everyting ok, -TIPC_ERR_NO_PORT otherwise */ -static u32 filter_connect(struct tipc_sock *tsk, struct sk_buff **buf) +static int filter_connect(struct tipc_sock *tsk, struct sk_buff **buf) { struct sock *sk = &tsk->sk; struct tipc_port *port = &tsk->port; struct socket *sock = sk->sk_socket; struct tipc_msg *msg = buf_msg(*buf); - u32 retval = TIPC_ERR_NO_PORT; + int retval = -TIPC_ERR_NO_PORT; int res; if (msg_mcast(msg)) @@ -1382,32 +1381,33 @@ static unsigned int rcvbuf_limit(struct sock *sk, struct sk_buff *buf) * * Called with socket lock already taken; port lock may also be taken. * - * Returns TIPC error status code (TIPC_OK if message is not to be rejected) + * Returns 0 (TIPC_OK) if message was consumed, -TIPC error code if message + * to be rejected. */ -static u32 filter_rcv(struct sock *sk, struct sk_buff *buf) +static int filter_rcv(struct sock *sk, struct sk_buff *buf) { struct socket *sock = sk->sk_socket; struct tipc_sock *tsk = tipc_sk(sk); struct tipc_msg *msg = buf_msg(buf); unsigned int limit = rcvbuf_limit(sk, buf); - u32 res = TIPC_OK; + int rc = TIPC_OK; /* Reject message if it is wrong sort of message for socket */ if (msg_type(msg) > TIPC_DIRECT_MSG) - return TIPC_ERR_NO_PORT; + return -TIPC_ERR_NO_PORT; if (sock->state == SS_READY) { if (msg_connected(msg)) - return TIPC_ERR_NO_PORT; + return -TIPC_ERR_NO_PORT; } else { - res = filter_connect(tsk, &buf); - if (res != TIPC_OK || buf == NULL) - return res; + rc = filter_connect(tsk, &buf); + if (rc != TIPC_OK || buf == NULL) + return rc; } /* Reject message if there isn't room to queue it */ if (sk_rmem_alloc_get(sk) + buf->truesize >= limit) - return TIPC_ERR_OVERLOAD; + return -TIPC_ERR_OVERLOAD; /* Enqueue message */ TIPC_SKB_CB(buf)->handle = NULL; @@ -1429,13 +1429,13 @@ static u32 filter_rcv(struct sock *sk, struct sk_buff *buf) */ static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *buf) { - u32 res; + int rc; struct tipc_sock *tsk = tipc_sk(sk); uint truesize = buf->truesize; - res = filter_rcv(sk, buf); - if (unlikely(res)) - tipc_reject_msg(buf, res); + rc = filter_rcv(sk, buf); + if (unlikely(rc)) + tipc_reject_msg(buf, -rc); if (atomic_read(&tsk->dupl_rcvcnt) < TIPC_CONN_OVERLOAD_LIMIT) atomic_add(truesize, &tsk->dupl_rcvcnt); @@ -1455,7 +1455,7 @@ int tipc_sk_rcv(struct sk_buff *buf) struct tipc_port *port; struct sock *sk; u32 dport = msg_destport(buf_msg(buf)); - int err = TIPC_OK; + int rc = TIPC_OK; uint limit; /* Forward unresolved named message */ @@ -1467,7 +1467,7 @@ int tipc_sk_rcv(struct sk_buff *buf) /* Validate destination */ port = tipc_port_lock(dport); if (unlikely(!port)) { - err = TIPC_ERR_NO_PORT; + rc = -TIPC_ERR_NO_PORT; goto exit; } @@ -1478,22 +1478,22 @@ int tipc_sk_rcv(struct sk_buff *buf) bh_lock_sock(sk); if (!sock_owned_by_user(sk)) { - err = filter_rcv(sk, buf); + rc = filter_rcv(sk, buf); } else { if (sk->sk_backlog.len == 0) atomic_set(&tsk->dupl_rcvcnt, 0); limit = rcvbuf_limit(sk, buf) + atomic_read(&tsk->dupl_rcvcnt); if (sk_add_backlog(sk, buf, limit)) - err = TIPC_ERR_OVERLOAD; + rc = -TIPC_ERR_OVERLOAD; } bh_unlock_sock(sk); tipc_port_unlock(port); - if (likely(!err)) + if (likely(!rc)) return 0; exit: - tipc_reject_msg(buf, err); + tipc_reject_msg(buf, -rc); return -EHOSTUNREACH; } -- cgit v1.2.3 From 8db1bae30b7cd3c3abc05f467d0f7c69b33b80e9 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:35 -0500 Subject: tipc: separate building and sending of rejected messages The way we build and send rejected message is currenty perceived as hard to follow, partly because we let the transmission go via deep call chains through functions such as tipc_reject_msg() and net_route_msg(). We want to remove those functions, and make the call sequences shallower and simpler. For this purpose, we separate building and sending of rejected messages. We build the reject message using the new function tipc_msg_reverse(), and let the transmission go via the newly introduced tipc_link_xmit2() function, as all transmission eventually will do. We also ensure that all calls to tipc_link_xmit2() are made outside port_lock/bh_lock_sock. Finally, we replace all calls to tipc_reject_msg() with the two new calls at all locations in the code that we want to keep. The remaining calls are made from code that we are planning to remove, along with tipc_reject_msg() itself. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 28 ++++++++++++++++++++-------- 1 file changed, 20 insertions(+), 8 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 9a95dab13b7e..5961a6335f49 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -39,6 +39,7 @@ #include "node.h" #include +#include "link.h" #define SS_LISTENING -1 /* socket is listening */ #define SS_READY -2 /* socket is connectionless */ @@ -123,9 +124,12 @@ static void advance_rx_queue(struct sock *sk) static void reject_rx_queue(struct sock *sk) { struct sk_buff *buf; + u32 dnode; - while ((buf = __skb_dequeue(&sk->sk_receive_queue))) - tipc_reject_msg(buf, TIPC_ERR_NO_PORT); + while ((buf = __skb_dequeue(&sk->sk_receive_queue))) { + if (tipc_msg_reverse(buf, &dnode, TIPC_ERR_NO_PORT)) + tipc_link_xmit2(buf, dnode, 0); + } } /** @@ -303,6 +307,7 @@ static int tipc_release(struct socket *sock) struct tipc_sock *tsk; struct tipc_port *port; struct sk_buff *buf; + u32 dnode; /* * Exit if socket isn't fully initialized (occurs when a failed accept() @@ -331,7 +336,8 @@ static int tipc_release(struct socket *sock) sock->state = SS_DISCONNECTING; tipc_port_disconnect(port->ref); } - tipc_reject_msg(buf, TIPC_ERR_NO_PORT); + if (tipc_msg_reverse(buf, &dnode, TIPC_ERR_NO_PORT)) + tipc_link_xmit2(buf, dnode, 0); } } @@ -1430,14 +1436,15 @@ static int filter_rcv(struct sock *sk, struct sk_buff *buf) static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *buf) { int rc; + u32 onode; struct tipc_sock *tsk = tipc_sk(sk); uint truesize = buf->truesize; rc = filter_rcv(sk, buf); - if (unlikely(rc)) - tipc_reject_msg(buf, -rc); - if (atomic_read(&tsk->dupl_rcvcnt) < TIPC_CONN_OVERLOAD_LIMIT) + if (unlikely(rc && tipc_msg_reverse(buf, &onode, -rc))) + tipc_link_xmit2(buf, onode, 0); + else if (atomic_read(&tsk->dupl_rcvcnt) < TIPC_CONN_OVERLOAD_LIMIT) atomic_add(truesize, &tsk->dupl_rcvcnt); return 0; @@ -1457,6 +1464,7 @@ int tipc_sk_rcv(struct sk_buff *buf) u32 dport = msg_destport(buf_msg(buf)); int rc = TIPC_OK; uint limit; + u32 dnode; /* Forward unresolved named message */ if (unlikely(!dport)) { @@ -1493,7 +1501,9 @@ int tipc_sk_rcv(struct sk_buff *buf) if (likely(!rc)) return 0; exit: - tipc_reject_msg(buf, -rc); + if (!tipc_msg_reverse(buf, &dnode, -rc)) + return -EHOSTUNREACH; + tipc_link_xmit2(buf, dnode, 0); return -EHOSTUNREACH; } @@ -1758,6 +1768,7 @@ static int tipc_shutdown(struct socket *sock, int how) struct tipc_sock *tsk = tipc_sk(sk); struct tipc_port *port = &tsk->port; struct sk_buff *buf; + u32 peer; int res; if (how != SHUT_RDWR) @@ -1778,7 +1789,8 @@ restart: goto restart; } tipc_port_disconnect(port->ref); - tipc_reject_msg(buf, TIPC_CONN_SHUTDOWN); + if (tipc_msg_reverse(buf, &peer, TIPC_CONN_SHUTDOWN)) + tipc_link_xmit2(buf, peer, 0); } else { tipc_port_shutdown(port->ref); } -- cgit v1.2.3 From 5a379074a7dd6d288ec9e6472769ba0e0c54dd85 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:36 -0500 Subject: tipc: introduce message evaluation function When a message arrives in a node and finds no destination socket, we may need to drop it, reject it, or forward it after a secondary destination lookup. The latter two cases currently results in a code path that is perceived as complex, because it follows a deep call chain via obscure functions such as net_route_named_msg() and net_route_msg(). We now introduce a function, tipc_msg_eval(), that takes the decision about whether such a message should be rejected or forwarded, but leaves it to the caller to actually perform the indicated action. If the decision is 'reject', it is still the task of the recently introduced function tipc_msg_reverse() to take the final decision about whether the message is rejectable or not. In the latter case it drops the message. As a result of this change, we can finally eliminate the function net_route_named_msg(), and hence become independent of net_route_msg(). Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 16 +++++----------- 1 file changed, 5 insertions(+), 11 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 5961a6335f49..e642ed5b3602 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -1466,16 +1466,10 @@ int tipc_sk_rcv(struct sk_buff *buf) uint limit; u32 dnode; - /* Forward unresolved named message */ - if (unlikely(!dport)) { - tipc_net_route_msg(buf); - return 0; - } - - /* Validate destination */ + /* Validate destination and message */ port = tipc_port_lock(dport); if (unlikely(!port)) { - rc = -TIPC_ERR_NO_PORT; + rc = tipc_msg_eval(buf, &dnode); goto exit; } @@ -1494,17 +1488,17 @@ int tipc_sk_rcv(struct sk_buff *buf) if (sk_add_backlog(sk, buf, limit)) rc = -TIPC_ERR_OVERLOAD; } - bh_unlock_sock(sk); tipc_port_unlock(port); if (likely(!rc)) return 0; exit: - if (!tipc_msg_reverse(buf, &dnode, -rc)) + if ((rc < 0) && !tipc_msg_reverse(buf, &dnode, -rc)) return -EHOSTUNREACH; + tipc_link_xmit2(buf, dnode, 0); - return -EHOSTUNREACH; + return (rc < 0) ? -EHOSTUNREACH : 0; } static int tipc_wait_for_connect(struct socket *sock, long *timeo_p) -- cgit v1.2.3 From e2dafe87d328774a94fdd77718422b9cbd97ed47 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:37 -0500 Subject: tipc: RDM/DGRAM transport uses new fragmenting and sending functions We merge the code for sending port name and port identity addressed messages into the corresponding send functions in socket.c, and start using the new fragmenting and transmit functions we just have introduced. This saves a call level and quite a few code lines, as well as making this part of the code easier to follow. As a consequence, the functions tipc_send2name() and tipc_send2port() in port.c can be removed. For practical reasons, we break out the code for sending multicast messages from tipc_sendmsg() and move it into a separate function, tipc_sendmcast(), but we do not yet convert it into using the new build/send functions. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 158 +++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 110 insertions(+), 48 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index e642ed5b3602..5690751cbfa5 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -36,8 +36,9 @@ #include "core.h" #include "port.h" +#include "name_table.h" #include "node.h" - +#include "link.h" #include #include "link.h" @@ -545,6 +546,8 @@ static int dest_name_check(struct sockaddr_tipc *dest, struct msghdr *m) { struct tipc_cfg_msg_hdr hdr; + if (unlikely(dest->addrtype == TIPC_ADDR_ID)) + return 0; if (likely(dest->addr.name.name.type >= TIPC_RESERVED_TYPES)) return 0; if (likely(dest->addr.name.name.type == TIPC_TOP_SRV)) @@ -587,13 +590,49 @@ static int tipc_wait_for_sndmsg(struct socket *sock, long *timeo_p) return 0; } +/** + * tipc_sendmcast - send multicast message + * @sock: socket structure + * @seq: destination address + * @iov: message data to send + * @dsz: total length of message data + * @timeo: timeout to wait for wakeup + * + * Called from function tipc_sendmsg(), which has done all sanity checks + * Returns the number of bytes sent on success, or errno + */ +static int tipc_sendmcast(struct socket *sock, struct tipc_name_seq *seq, + struct iovec *iov, size_t dsz, long timeo) +{ + struct sock *sk = sock->sk; + struct tipc_sock *tsk = tipc_sk(sk); + int rc; + + do { + if (sock->state != SS_READY) { + rc = -EOPNOTSUPP; + break; + } + rc = tipc_port_mcast_xmit(&tsk->port, seq, iov, dsz); + if (likely(rc >= 0)) { + if (sock->state != SS_READY) + sock->state = SS_CONNECTING; + break; + } + if (rc != -ELINKCONG) + break; + rc = tipc_wait_for_sndmsg(sock, &timeo); + } while (!rc); + + return rc; +} /** * tipc_sendmsg - send message in connectionless manner * @iocb: if NULL, indicates that socket lock is already held * @sock: socket structure * @m: message to send - * @total_len: length of message + * @dsz: amount of user data to be sent * * Message must have an destination specified explicitly. * Used for SOCK_RDM and SOCK_DGRAM messages, @@ -603,93 +642,116 @@ static int tipc_wait_for_sndmsg(struct socket *sock, long *timeo_p) * Returns the number of bytes sent on success, or errno otherwise */ static int tipc_sendmsg(struct kiocb *iocb, struct socket *sock, - struct msghdr *m, size_t total_len) + struct msghdr *m, size_t dsz) { + DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name); struct sock *sk = sock->sk; struct tipc_sock *tsk = tipc_sk(sk); struct tipc_port *port = &tsk->port; - DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name); - int needs_conn; + struct tipc_msg *mhdr = &port->phdr; + struct iovec *iov = m->msg_iov; + u32 dnode, dport; + struct sk_buff *buf; + struct tipc_name_seq *seq = &dest->addr.nameseq; + u32 mtu; long timeo; - int res = -EINVAL; + int rc = -EINVAL; if (unlikely(!dest)) return -EDESTADDRREQ; + if (unlikely((m->msg_namelen < sizeof(*dest)) || (dest->family != AF_TIPC))) return -EINVAL; - if (total_len > TIPC_MAX_USER_MSG_SIZE) + + if (dsz > TIPC_MAX_USER_MSG_SIZE) return -EMSGSIZE; if (iocb) lock_sock(sk); - needs_conn = (sock->state != SS_READY); - if (unlikely(needs_conn)) { + if (unlikely(sock->state != SS_READY)) { if (sock->state == SS_LISTENING) { - res = -EPIPE; + rc = -EPIPE; goto exit; } if (sock->state != SS_UNCONNECTED) { - res = -EISCONN; + rc = -EISCONN; goto exit; } if (tsk->port.published) { - res = -EOPNOTSUPP; + rc = -EOPNOTSUPP; goto exit; } if (dest->addrtype == TIPC_ADDR_NAME) { tsk->port.conn_type = dest->addr.name.name.type; tsk->port.conn_instance = dest->addr.name.name.instance; } - - /* Abort any pending connection attempts (very unlikely) */ - reject_rx_queue(sk); } + rc = dest_name_check(dest, m); + if (rc) + goto exit; timeo = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT); - do { - if (dest->addrtype == TIPC_ADDR_NAME) { - res = dest_name_check(dest, m); - if (res) - break; - res = tipc_send2name(port, - &dest->addr.name.name, - dest->addr.name.domain, - m->msg_iov, - total_len); - } else if (dest->addrtype == TIPC_ADDR_ID) { - res = tipc_send2port(port, - &dest->addr.id, - m->msg_iov, - total_len); - } else if (dest->addrtype == TIPC_ADDR_MCAST) { - if (needs_conn) { - res = -EOPNOTSUPP; - break; - } - res = dest_name_check(dest, m); - if (res) - break; - res = tipc_port_mcast_xmit(port, - &dest->addr.nameseq, - m->msg_iov, - total_len); + + if (dest->addrtype == TIPC_ADDR_MCAST) { + rc = tipc_sendmcast(sock, seq, iov, dsz, timeo); + goto exit; + } else if (dest->addrtype == TIPC_ADDR_NAME) { + u32 type = dest->addr.name.name.type; + u32 inst = dest->addr.name.name.instance; + u32 domain = dest->addr.name.domain; + + dnode = domain; + msg_set_type(mhdr, TIPC_NAMED_MSG); + msg_set_hdr_sz(mhdr, NAMED_H_SIZE); + msg_set_nametype(mhdr, type); + msg_set_nameinst(mhdr, inst); + msg_set_lookup_scope(mhdr, tipc_addr_scope(domain)); + dport = tipc_nametbl_translate(type, inst, &dnode); + msg_set_destnode(mhdr, dnode); + msg_set_destport(mhdr, dport); + if (unlikely(!dport && !dnode)) { + rc = -EHOSTUNREACH; + goto exit; } - if (likely(res != -ELINKCONG)) { - if (needs_conn && (res >= 0)) + } else if (dest->addrtype == TIPC_ADDR_ID) { + dnode = dest->addr.id.node; + msg_set_type(mhdr, TIPC_DIRECT_MSG); + msg_set_lookup_scope(mhdr, 0); + msg_set_destnode(mhdr, dnode); + msg_set_destport(mhdr, dest->addr.id.ref); + msg_set_hdr_sz(mhdr, BASIC_H_SIZE); + } + +new_mtu: + mtu = tipc_node_get_mtu(dnode, tsk->port.ref); + rc = tipc_msg_build2(mhdr, iov, 0, dsz, mtu, &buf); + if (rc < 0) + goto exit; + + do { + rc = tipc_link_xmit2(buf, dnode, tsk->port.ref); + if (likely(rc >= 0)) { + if (sock->state != SS_READY) sock->state = SS_CONNECTING; + rc = dsz; break; } - res = tipc_wait_for_sndmsg(sock, &timeo); - if (res) + if (rc == -EMSGSIZE) + goto new_mtu; + + if (rc != -ELINKCONG) break; - } while (1); + + rc = tipc_wait_for_sndmsg(sock, &timeo); + } while (!rc); exit: if (iocb) release_sock(sk); - return res; + + return rc; } static int tipc_wait_for_sndpkt(struct socket *sock, long *timeo_p) -- cgit v1.2.3 From 4ccfe5e0419eefcab3010ff6a87ffb03aef86c5d Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:38 -0500 Subject: tipc: connection oriented transport uses new send functions We move the message sending across established connections to use the message preparation and send functions introduced earlier in this series. We now do the message preparation and call to the link send function directly from the socket, instead of going via the port layer. As a consequence of this change, the functions tipc_send(), tipc_port_iovec_rcv(), tipc_port_iovec_reject() and tipc_reject_msg() become unreferenced and can be eliminated from port.c. For the same reason, the functions tipc_link_xmit_fast(), tipc_link_iovec_xmit_long() and tipc_link_iovec_fast() can be eliminated from link.c. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 180 +++++++++++++++++++++--------------------------------- 1 file changed, 68 insertions(+), 112 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 5690751cbfa5..bfe79bbd83a1 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -206,6 +206,7 @@ static int tipc_sk_create(struct net *net, struct socket *sock, sk->sk_data_ready = tipc_data_ready; sk->sk_write_space = tipc_write_space; tsk->conn_timeout = CONN_TIMEOUT_DEFAULT; + tsk->port.sent = 0; atomic_set(&tsk->dupl_rcvcnt, 0); tipc_port_unlock(port); @@ -784,30 +785,40 @@ static int tipc_wait_for_sndpkt(struct socket *sock, long *timeo_p) } /** - * tipc_send_packet - send a connection-oriented message - * @iocb: if NULL, indicates that socket lock is already held + * tipc_send_stream - send stream-oriented data + * @iocb: (unused) * @sock: socket structure - * @m: message to send - * @total_len: length of message + * @m: data to send + * @dsz: total length of data to be transmitted * - * Used for SOCK_SEQPACKET messages and SOCK_STREAM data. + * Used for SOCK_STREAM data. * - * Returns the number of bytes sent on success, or errno otherwise + * Returns the number of bytes sent on success (or partial success), + * or errno if no data sent */ -static int tipc_send_packet(struct kiocb *iocb, struct socket *sock, - struct msghdr *m, size_t total_len) +static int tipc_send_stream(struct kiocb *iocb, struct socket *sock, + struct msghdr *m, size_t dsz) { struct sock *sk = sock->sk; struct tipc_sock *tsk = tipc_sk(sk); + struct tipc_port *port = &tsk->port; + struct tipc_msg *mhdr = &port->phdr; + struct sk_buff *buf; DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name); - int res = -EINVAL; + u32 ref = port->ref; + int rc = -EINVAL; long timeo; + u32 dnode; + uint mtu, send, sent = 0; /* Handle implied connection establishment */ - if (unlikely(dest)) - return tipc_sendmsg(iocb, sock, m, total_len); - - if (total_len > TIPC_MAX_USER_MSG_SIZE) + if (unlikely(dest)) { + rc = tipc_sendmsg(iocb, sock, m, dsz); + if (dsz && (dsz == rc)) + tsk->port.sent = 1; + return rc; + } + if (dsz > (uint)INT_MAX) return -EMSGSIZE; if (iocb) @@ -815,123 +826,68 @@ static int tipc_send_packet(struct kiocb *iocb, struct socket *sock, if (unlikely(sock->state != SS_CONNECTED)) { if (sock->state == SS_DISCONNECTING) - res = -EPIPE; + rc = -EPIPE; else - res = -ENOTCONN; + rc = -ENOTCONN; goto exit; } timeo = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT); + dnode = tipc_port_peernode(port); + port->congested = 1; + +next: + mtu = port->max_pkt; + send = min_t(uint, dsz - sent, TIPC_MAX_USER_MSG_SIZE); + rc = tipc_msg_build2(mhdr, m->msg_iov, sent, send, mtu, &buf); + if (unlikely(rc < 0)) + goto exit; do { - res = tipc_send(&tsk->port, m->msg_iov, total_len); - if (likely(res != -ELINKCONG)) - break; - res = tipc_wait_for_sndpkt(sock, &timeo); - if (res) - break; - } while (1); + port->congested = 1; + if (likely(!tipc_port_congested(port))) { + rc = tipc_link_xmit2(buf, dnode, ref); + if (likely(!rc)) { + port->sent++; + sent += send; + if (sent == dsz) + break; + goto next; + } + if (rc == -EMSGSIZE) { + port->max_pkt = tipc_node_get_mtu(dnode, ref); + goto next; + } + if (rc != -ELINKCONG) + break; + } + rc = tipc_wait_for_sndpkt(sock, &timeo); + } while (!rc); + + port->congested = 0; exit: if (iocb) release_sock(sk); - return res; + return sent ? sent : rc; } /** - * tipc_send_stream - send stream-oriented data - * @iocb: (unused) + * tipc_send_packet - send a connection-oriented message + * @iocb: if NULL, indicates that socket lock is already held * @sock: socket structure - * @m: data to send - * @total_len: total length of data to be sent + * @m: message to send + * @dsz: length of data to be transmitted * - * Used for SOCK_STREAM data. + * Used for SOCK_SEQPACKET messages. * - * Returns the number of bytes sent on success (or partial success), - * or errno if no data sent + * Returns the number of bytes sent on success, or errno otherwise */ -static int tipc_send_stream(struct kiocb *iocb, struct socket *sock, - struct msghdr *m, size_t total_len) +static int tipc_send_packet(struct kiocb *iocb, struct socket *sock, + struct msghdr *m, size_t dsz) { - struct sock *sk = sock->sk; - struct tipc_sock *tsk = tipc_sk(sk); - struct msghdr my_msg; - struct iovec my_iov; - struct iovec *curr_iov; - int curr_iovlen; - char __user *curr_start; - u32 hdr_size; - int curr_left; - int bytes_to_send; - int bytes_sent; - int res; - - lock_sock(sk); - - /* Handle special cases where there is no connection */ - if (unlikely(sock->state != SS_CONNECTED)) { - if (sock->state == SS_UNCONNECTED) - res = tipc_send_packet(NULL, sock, m, total_len); - else - res = sock->state == SS_DISCONNECTING ? -EPIPE : -ENOTCONN; - goto exit; - } - - if (unlikely(m->msg_name)) { - res = -EISCONN; - goto exit; - } - - if (total_len > (unsigned int)INT_MAX) { - res = -EMSGSIZE; - goto exit; - } - - /* - * Send each iovec entry using one or more messages - * - * Note: This algorithm is good for the most likely case - * (i.e. one large iovec entry), but could be improved to pass sets - * of small iovec entries into send_packet(). - */ - curr_iov = m->msg_iov; - curr_iovlen = m->msg_iovlen; - my_msg.msg_iov = &my_iov; - my_msg.msg_iovlen = 1; - my_msg.msg_flags = m->msg_flags; - my_msg.msg_name = NULL; - bytes_sent = 0; - - hdr_size = msg_hdr_sz(&tsk->port.phdr); - - while (curr_iovlen--) { - curr_start = curr_iov->iov_base; - curr_left = curr_iov->iov_len; - - while (curr_left) { - bytes_to_send = tsk->port.max_pkt - hdr_size; - if (bytes_to_send > TIPC_MAX_USER_MSG_SIZE) - bytes_to_send = TIPC_MAX_USER_MSG_SIZE; - if (curr_left < bytes_to_send) - bytes_to_send = curr_left; - my_iov.iov_base = curr_start; - my_iov.iov_len = bytes_to_send; - res = tipc_send_packet(NULL, sock, &my_msg, - bytes_to_send); - if (res < 0) { - if (bytes_sent) - res = bytes_sent; - goto exit; - } - curr_left -= bytes_to_send; - curr_start += bytes_to_send; - bytes_sent += bytes_to_send; - } + if (dsz > TIPC_MAX_USER_MSG_SIZE) + return -EMSGSIZE; - curr_iov++; - } - res = bytes_sent; -exit: - release_sock(sk); - return res; + return tipc_send_stream(iocb, sock, m, dsz); } /** -- cgit v1.2.3 From ec8a2e5621db2da24badb3969eda7fd359e1869f Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:40 -0500 Subject: tipc: same receive code path for connection protocol and data messages As a preparation to eliminate port_lock we need to bring reception of connection protocol messages under proper protection of bh_lock_sock or socket owner. We fix this by letting those messages follow the same code path as incoming data messages. As a side effect of this change, the last reference to the function net_route_msg() disappears, and we can eliminate that function. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index bfe79bbd83a1..d838a4b66d3a 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -1416,6 +1416,11 @@ static int filter_rcv(struct sock *sk, struct sk_buff *buf) unsigned int limit = rcvbuf_limit(sk, buf); int rc = TIPC_OK; + if (unlikely(msg_user(msg) == CONN_MANAGER)) { + tipc_port_proto_rcv(&tsk->port, buf); + return TIPC_OK; + } + /* Reject message if it is wrong sort of message for socket */ if (msg_type(msg) > TIPC_DIRECT_MSG) return -TIPC_ERR_NO_PORT; -- cgit v1.2.3 From ac0074ee70ddb32f62d918b31cb20e3c947c75a1 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:41 -0500 Subject: tipc: clean up connection protocol reception function We simplify the code for receiving connection probes, leveraging the recently introduced tipc_msg_reverse() function. We also stick to the principle of sending a possible response message directly from the calling (tipc_sk_rcv or backlog_rcv) functions, hence making the call chain shallower and easier to follow. We make one small protocol change here, allowed according to the spec. If a protocol message arrives from a remote socket that is not the one we are connected to, we are currently generating a connection abort message and send it to the source. This behavior is unnecessary, and might even be a security risk, so instead we now choose to only ignore the message. The consequnce for the sender is that he will need longer time to discover his mistake (until the next timeout), but this is an extreme corner case, and may happen anyway under other circumstances, so we deem this change acceptable. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 64 +++++++++++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 55 insertions(+), 9 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index d838a4b66d3a..1762323156af 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -46,6 +46,7 @@ #define SS_READY -2 /* socket is connectionless */ #define CONN_TIMEOUT_DEFAULT 8000 /* default connect timeout = 8s */ +#define TIPC_FWD_MSG 1 static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *skb); static void tipc_data_ready(struct sock *sk); @@ -533,6 +534,46 @@ static unsigned int tipc_poll(struct file *file, struct socket *sock, return mask; } +/** + * tipc_sk_proto_rcv - receive a connection mng protocol message + * @tsk: receiving socket + * @dnode: node to send response message to, if any + * @buf: buffer containing protocol message + * Returns 0 (TIPC_OK) if message was consumed, 1 (TIPC_FWD_MSG) if + * (CONN_PROBE_REPLY) message should be forwarded. + */ +int tipc_sk_proto_rcv(struct tipc_sock *tsk, u32 *dnode, struct sk_buff *buf) +{ + struct tipc_msg *msg = buf_msg(buf); + struct tipc_port *port = &tsk->port; + int wakeable; + + /* Ignore if connection cannot be validated: */ + if (!port->connected || !tipc_port_peer_msg(port, msg)) + goto exit; + + port->probing_state = TIPC_CONN_OK; + + if (msg_type(msg) == CONN_ACK) { + wakeable = tipc_port_congested(port) && port->congested; + port->acked += msg_msgcnt(msg); + if (!tipc_port_congested(port)) { + port->congested = 0; + if (wakeable) + tipc_port_wakeup(port); + } + } else if (msg_type(msg) == CONN_PROBE) { + if (!tipc_msg_reverse(buf, dnode, TIPC_OK)) + return TIPC_OK; + msg_set_type(msg, CONN_PROBE_REPLY); + return TIPC_FWD_MSG; + } + /* Do nothing if msg_type() == CONN_PROBE_REPLY */ +exit: + kfree_skb(buf); + return TIPC_OK; +} + /** * dest_name_check - verify user is permitted to send to specified port name * @dest: destination address @@ -1406,7 +1447,7 @@ static unsigned int rcvbuf_limit(struct sock *sk, struct sk_buff *buf) * Called with socket lock already taken; port lock may also be taken. * * Returns 0 (TIPC_OK) if message was consumed, -TIPC error code if message - * to be rejected. + * to be rejected, 1 (TIPC_FWD_MSG) if (CONN_MANAGER) message to be forwarded */ static int filter_rcv(struct sock *sk, struct sk_buff *buf) { @@ -1414,12 +1455,11 @@ static int filter_rcv(struct sock *sk, struct sk_buff *buf) struct tipc_sock *tsk = tipc_sk(sk); struct tipc_msg *msg = buf_msg(buf); unsigned int limit = rcvbuf_limit(sk, buf); + u32 onode; int rc = TIPC_OK; - if (unlikely(msg_user(msg) == CONN_MANAGER)) { - tipc_port_proto_rcv(&tsk->port, buf); - return TIPC_OK; - } + if (unlikely(msg_user(msg) == CONN_MANAGER)) + return tipc_sk_proto_rcv(tsk, &onode, buf); /* Reject message if it is wrong sort of message for socket */ if (msg_type(msg) > TIPC_DIRECT_MSG) @@ -1465,10 +1505,16 @@ static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *buf) rc = filter_rcv(sk, buf); - if (unlikely(rc && tipc_msg_reverse(buf, &onode, -rc))) - tipc_link_xmit2(buf, onode, 0); - else if (atomic_read(&tsk->dupl_rcvcnt) < TIPC_CONN_OVERLOAD_LIMIT) - atomic_add(truesize, &tsk->dupl_rcvcnt); + if (likely(!rc)) { + if (atomic_read(&tsk->dupl_rcvcnt) < TIPC_CONN_OVERLOAD_LIMIT) + atomic_add(truesize, &tsk->dupl_rcvcnt); + return 0; + } + + if ((rc < 0) && !tipc_msg_reverse(buf, &onode, -rc)) + return 0; + + tipc_link_xmit2(buf, onode, 0); return 0; } -- cgit v1.2.3 From 60120526c26f42fd658e32bf4a6d548483d09da8 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 25 Jun 2014 20:41:42 -0500 Subject: tipc: simplify connection congestion handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As a consequence of the recently introduced serialized access to the socket in commit 8d94168a761819d10252bab1f8de6d7b202c3baa ("tipc: same receive code path for connection protocol and data messages") we can make a number of simplifications in the detection and handling of connection congestion situations. - We don't need to keep two counters, one for sent messages and one for acked messages. There is no longer any risk for races between acknowledge messages arriving in BH and data message sending running in user context. So we merge this into one counter, 'sent_unacked', which is incremented at sending and subtracted from at acknowledge reception. - We don't need to set the 'congested' field in tipc_port to true before we sent the message, and clear it when sending is successful. (As a matter of fact, it was never necessary; the field was set in link_schedule_port() before any wakeup could arrive anyway.) - We keep the conditions for link congestion and connection connection congestion separated. There would otherwise be a risk that an arriving acknowledge message may wake up a user sleeping because of link congestion. - We can simplify reception of acknowledge messages. We also make some cosmetic/structural changes: - We rename the 'congested' field to the more correct 'link_cong´. - We rename 'conn_unacked' to 'rcv_unacked' - We move the above mentioned fields from struct tipc_port to struct tipc_sock. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 48 +++++++++++++++++++++++------------------------- 1 file changed, 23 insertions(+), 25 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 1762323156af..ede78b144dcf 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -207,7 +207,7 @@ static int tipc_sk_create(struct net *net, struct socket *sock, sk->sk_data_ready = tipc_data_ready; sk->sk_write_space = tipc_write_space; tsk->conn_timeout = CONN_TIMEOUT_DEFAULT; - tsk->port.sent = 0; + tsk->sent_unacked = 0; atomic_set(&tsk->dupl_rcvcnt, 0); tipc_port_unlock(port); @@ -513,12 +513,12 @@ static unsigned int tipc_poll(struct file *file, struct socket *sock, switch ((int)sock->state) { case SS_UNCONNECTED: - if (!tsk->port.congested) + if (!tsk->link_cong) mask |= POLLOUT; break; case SS_READY: case SS_CONNECTED: - if (!tsk->port.congested) + if (!tsk->link_cong && !tipc_sk_conn_cong(tsk)) mask |= POLLOUT; /* fall thru' */ case SS_CONNECTING: @@ -546,7 +546,7 @@ int tipc_sk_proto_rcv(struct tipc_sock *tsk, u32 *dnode, struct sk_buff *buf) { struct tipc_msg *msg = buf_msg(buf); struct tipc_port *port = &tsk->port; - int wakeable; + int conn_cong; /* Ignore if connection cannot be validated: */ if (!port->connected || !tipc_port_peer_msg(port, msg)) @@ -555,13 +555,10 @@ int tipc_sk_proto_rcv(struct tipc_sock *tsk, u32 *dnode, struct sk_buff *buf) port->probing_state = TIPC_CONN_OK; if (msg_type(msg) == CONN_ACK) { - wakeable = tipc_port_congested(port) && port->congested; - port->acked += msg_msgcnt(msg); - if (!tipc_port_congested(port)) { - port->congested = 0; - if (wakeable) - tipc_port_wakeup(port); - } + conn_cong = tipc_sk_conn_cong(tsk); + tsk->sent_unacked -= msg_msgcnt(msg); + if (conn_cong) + tipc_sock_wakeup(tsk); } else if (msg_type(msg) == CONN_PROBE) { if (!tipc_msg_reverse(buf, dnode, TIPC_OK)) return TIPC_OK; @@ -626,7 +623,7 @@ static int tipc_wait_for_sndmsg(struct socket *sock, long *timeo_p) return sock_intr_errno(*timeo_p); prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE); - done = sk_wait_event(sk, timeo_p, !tsk->port.congested); + done = sk_wait_event(sk, timeo_p, !tsk->link_cong); finish_wait(sk_sleep(sk), &wait); } while (!done); return 0; @@ -800,7 +797,6 @@ static int tipc_wait_for_sndpkt(struct socket *sock, long *timeo_p) { struct sock *sk = sock->sk; struct tipc_sock *tsk = tipc_sk(sk); - struct tipc_port *port = &tsk->port; DEFINE_WAIT(wait); int done; @@ -819,7 +815,9 @@ static int tipc_wait_for_sndpkt(struct socket *sock, long *timeo_p) prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE); done = sk_wait_event(sk, timeo_p, - (!port->congested || !port->connected)); + (!tsk->link_cong && + !tipc_sk_conn_cong(tsk)) || + !tsk->port.connected); finish_wait(sk_sleep(sk), &wait); } while (!done); return 0; @@ -856,7 +854,7 @@ static int tipc_send_stream(struct kiocb *iocb, struct socket *sock, if (unlikely(dest)) { rc = tipc_sendmsg(iocb, sock, m, dsz); if (dsz && (dsz == rc)) - tsk->port.sent = 1; + tsk->sent_unacked = 1; return rc; } if (dsz > (uint)INT_MAX) @@ -875,7 +873,6 @@ static int tipc_send_stream(struct kiocb *iocb, struct socket *sock, timeo = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT); dnode = tipc_port_peernode(port); - port->congested = 1; next: mtu = port->max_pkt; @@ -884,11 +881,10 @@ next: if (unlikely(rc < 0)) goto exit; do { - port->congested = 1; - if (likely(!tipc_port_congested(port))) { + if (likely(!tipc_sk_conn_cong(tsk))) { rc = tipc_link_xmit2(buf, dnode, ref); if (likely(!rc)) { - port->sent++; + tsk->sent_unacked++; sent += send; if (sent == dsz) break; @@ -903,8 +899,6 @@ next: } rc = tipc_wait_for_sndpkt(sock, &timeo); } while (!rc); - - port->congested = 0; exit: if (iocb) release_sock(sk); @@ -1169,8 +1163,10 @@ restart: /* Consume received message (optional) */ if (likely(!(flags & MSG_PEEK))) { if ((sock->state != SS_READY) && - (++port->conn_unacked >= TIPC_CONNACK_INTV)) - tipc_acknowledge(port->ref, port->conn_unacked); + (++tsk->rcv_unacked >= TIPC_CONNACK_INTV)) { + tipc_acknowledge(port->ref, tsk->rcv_unacked); + tsk->rcv_unacked = 0; + } advance_rx_queue(sk); } exit: @@ -1278,8 +1274,10 @@ restart: /* Consume received message (optional) */ if (likely(!(flags & MSG_PEEK))) { - if (unlikely(++port->conn_unacked >= TIPC_CONNACK_INTV)) - tipc_acknowledge(port->ref, port->conn_unacked); + if (unlikely(++tsk->rcv_unacked >= TIPC_CONNACK_INTV)) { + tipc_acknowledge(port->ref, tsk->rcv_unacked); + tsk->rcv_unacked = 0; + } advance_rx_queue(sk); } -- cgit v1.2.3 From 70452dcb6d401349fbd6db55dfab112fb44639be Mon Sep 17 00:00:00 2001 From: Erik Hugne Date: Sun, 6 Jul 2014 20:38:50 -0400 Subject: tipc: fix a memleak when sending data This fixes a regression bug caused by: 067608e9d019d6477fd45dd948e81af0e5bf599f ("tipc: introduce direct iovec to buffer chain fragmentation function") If data is sent on a nonblocking socket and the destination link is congested, the buffer chain is leaked. We fix this by freeing the chain in this case. Signed-off-by: Erik Hugne Signed-off-by: Jon Maloy Acked-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index ede78b144dcf..8c5600cfcc3e 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -784,8 +784,9 @@ new_mtu: break; rc = tipc_wait_for_sndmsg(sock, &timeo); + if (rc) + kfree_skb_list(buf); } while (!rc); - exit: if (iocb) release_sock(sk); @@ -898,6 +899,8 @@ next: break; } rc = tipc_wait_for_sndpkt(sock, &timeo); + if (rc) + kfree_skb_list(buf); } while (!rc); exit: if (iocb) -- cgit v1.2.3 From 7ceaa583becacb985ad5803bac0970b8f8cc0d74 Mon Sep 17 00:00:00 2001 From: Fabian Frederick Date: Mon, 14 Jul 2014 18:30:52 +0200 Subject: tipc: remove unnecessary break after return Signed-off-by: Fabian Frederick Signed-off-by: David S. Miller --- net/tipc/socket.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 8c5600cfcc3e..de01622672b2 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -2028,7 +2028,6 @@ int tipc_ioctl(struct socket *sk, unsigned int cmd, unsigned long arg) return 0; } return -EADDRNOTAVAIL; - break; default: return -ENOIOCTLCMD; } -- cgit v1.2.3 From 078bec826f7b73cf2a2397680537bcb7e075b492 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 16 Jul 2014 20:41:00 -0400 Subject: tipc: add new functions for multicast and broadcast distribution We add a new broadcast link transmit function in bclink.c and a new receive function in socket.c. The purpose is to move the branching between external and internal destination down to the link layer, just as we have done with unicast in earlier commits. We also make use of the new link-independent fragmentation support that was introduced in an earlier commit series. This gives a shorter and simpler code path, and makes it possible to obtain copy-free buffer delivery to all node local destination sockets. The new transmission code is added in parallel with the existing one, and will be used by the socket multicast send function in the next commit in this series. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 40 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 40 insertions(+) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index de01622672b2..8d30995682b1 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -534,6 +534,46 @@ static unsigned int tipc_poll(struct file *file, struct socket *sock, return mask; } +/* tipc_sk_mcast_rcv - Deliver multicast message to all destination sockets + */ +void tipc_sk_mcast_rcv(struct sk_buff *buf) +{ + struct tipc_msg *msg = buf_msg(buf); + struct tipc_port_list dports = {0, NULL, }; + struct tipc_port_list *item; + struct sk_buff *b; + uint i, last, dst = 0; + u32 scope = TIPC_CLUSTER_SCOPE; + + if (in_own_node(msg_orignode(msg))) + scope = TIPC_NODE_SCOPE; + + /* Create destination port list: */ + tipc_nametbl_mc_translate(msg_nametype(msg), + msg_namelower(msg), + msg_nameupper(msg), + scope, + &dports); + last = dports.count; + if (!last) { + kfree_skb(buf); + return; + } + + for (item = &dports; item; item = item->next) { + for (i = 0; i < PLSIZE && ++dst <= last; i++) { + b = (dst != last) ? skb_clone(buf, GFP_ATOMIC) : buf; + if (!b) { + pr_warn("Failed do clone mcast rcv buffer\n"); + continue; + } + msg_set_destport(msg, item->ports[i]); + tipc_sk_rcv(b); + } + } + tipc_port_list_free(&dports); +} + /** * tipc_sk_proto_rcv - receive a connection mng protocol message * @tsk: receiving socket -- cgit v1.2.3 From 0abd8ff21f19adddc465538354e9baaca63df073 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 16 Jul 2014 20:41:01 -0400 Subject: tipc: start using the new multicast functions In this commit, we convert the socket multicast send function to directly call the new multicast/broadcast function (tipc_bclink_xmit2()) introduced in the previous commit. We do this instead of letting the call go via the now obsolete tipc_port_mcast_xmit(), hence saving a call level and some code complexity. We also remove the initial destination lookup at the message sending side, and replace that with an unconditional lookup at the receiving side, including on the sending node itself. This makes the destination lookup and message transfer more uniform than before. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 90 ++++++++++++++++++++++++++++++++----------------------- 1 file changed, 53 insertions(+), 37 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 8d30995682b1..b28eea50c7fc 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -53,6 +53,7 @@ static void tipc_data_ready(struct sock *sk); static void tipc_write_space(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); static const struct proto_ops packet_ops; static const struct proto_ops stream_ops; @@ -534,6 +535,58 @@ static unsigned int tipc_poll(struct file *file, struct socket *sock, return mask; } +/** + * tipc_sendmcast - send multicast message + * @sock: socket structure + * @seq: destination address + * @iov: message data to send + * @dsz: total length of message data + * @timeo: timeout to wait for wakeup + * + * Called from function tipc_sendmsg(), which has done all sanity checks + * Returns the number of bytes sent on success, or errno + */ +static int tipc_sendmcast(struct socket *sock, struct tipc_name_seq *seq, + struct iovec *iov, size_t dsz, long timeo) +{ + struct sock *sk = sock->sk; + struct tipc_msg *mhdr = &tipc_sk(sk)->port.phdr; + struct sk_buff *buf; + uint mtu; + int rc; + + msg_set_type(mhdr, TIPC_MCAST_MSG); + msg_set_lookup_scope(mhdr, TIPC_CLUSTER_SCOPE); + msg_set_destport(mhdr, 0); + msg_set_destnode(mhdr, 0); + msg_set_nametype(mhdr, seq->type); + msg_set_namelower(mhdr, seq->lower); + msg_set_nameupper(mhdr, seq->upper); + msg_set_hdr_sz(mhdr, MCAST_H_SIZE); + +new_mtu: + mtu = tipc_bclink_get_mtu(); + rc = tipc_msg_build2(mhdr, iov, 0, dsz, mtu, &buf); + if (unlikely(rc < 0)) + return rc; + + do { + rc = tipc_bclink_xmit(buf); + if (likely(rc >= 0)) { + rc = dsz; + break; + } + if (rc == -EMSGSIZE) + goto new_mtu; + if (rc != -ELINKCONG) + break; + rc = tipc_wait_for_sndmsg(sock, &timeo); + if (rc) + kfree_skb_list(buf); + } while (!rc); + return rc; +} + /* tipc_sk_mcast_rcv - Deliver multicast message to all destination sockets */ void tipc_sk_mcast_rcv(struct sk_buff *buf) @@ -669,43 +722,6 @@ static int tipc_wait_for_sndmsg(struct socket *sock, long *timeo_p) return 0; } -/** - * tipc_sendmcast - send multicast message - * @sock: socket structure - * @seq: destination address - * @iov: message data to send - * @dsz: total length of message data - * @timeo: timeout to wait for wakeup - * - * Called from function tipc_sendmsg(), which has done all sanity checks - * Returns the number of bytes sent on success, or errno - */ -static int tipc_sendmcast(struct socket *sock, struct tipc_name_seq *seq, - struct iovec *iov, size_t dsz, long timeo) -{ - struct sock *sk = sock->sk; - struct tipc_sock *tsk = tipc_sk(sk); - int rc; - - do { - if (sock->state != SS_READY) { - rc = -EOPNOTSUPP; - break; - } - rc = tipc_port_mcast_xmit(&tsk->port, seq, iov, dsz); - if (likely(rc >= 0)) { - if (sock->state != SS_READY) - sock->state = SS_CONNECTING; - break; - } - if (rc != -ELINKCONG) - break; - rc = tipc_wait_for_sndmsg(sock, &timeo); - } while (!rc); - - return rc; -} - /** * tipc_sendmsg - send message in connectionless manner * @iocb: if NULL, indicates that socket lock is already held -- cgit v1.2.3 From 9fbfb8b120bd4fe89cd70d6c8841e6e1cfab2609 Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Wed, 16 Jul 2014 20:41:03 -0400 Subject: tipc: rename temporarily named functions After the previous commit, we can now give the functions with temporary names, such as tipc_link_xmit2(), tipc_msg_build2() etc., their proper names. There are no functional changes in this commit. Signed-off-by: Jon Maloy Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index b28eea50c7fc..5ad42f6137d8 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -131,7 +131,7 @@ static void reject_rx_queue(struct sock *sk) while ((buf = __skb_dequeue(&sk->sk_receive_queue))) { if (tipc_msg_reverse(buf, &dnode, TIPC_ERR_NO_PORT)) - tipc_link_xmit2(buf, dnode, 0); + tipc_link_xmit(buf, dnode, 0); } } @@ -341,7 +341,7 @@ static int tipc_release(struct socket *sock) tipc_port_disconnect(port->ref); } if (tipc_msg_reverse(buf, &dnode, TIPC_ERR_NO_PORT)) - tipc_link_xmit2(buf, dnode, 0); + tipc_link_xmit(buf, dnode, 0); } } @@ -566,7 +566,7 @@ static int tipc_sendmcast(struct socket *sock, struct tipc_name_seq *seq, new_mtu: mtu = tipc_bclink_get_mtu(); - rc = tipc_msg_build2(mhdr, iov, 0, dsz, mtu, &buf); + rc = tipc_msg_build(mhdr, iov, 0, dsz, mtu, &buf); if (unlikely(rc < 0)) return rc; @@ -821,12 +821,12 @@ static int tipc_sendmsg(struct kiocb *iocb, struct socket *sock, new_mtu: mtu = tipc_node_get_mtu(dnode, tsk->port.ref); - rc = tipc_msg_build2(mhdr, iov, 0, dsz, mtu, &buf); + rc = tipc_msg_build(mhdr, iov, 0, dsz, mtu, &buf); if (rc < 0) goto exit; do { - rc = tipc_link_xmit2(buf, dnode, tsk->port.ref); + rc = tipc_link_xmit(buf, dnode, tsk->port.ref); if (likely(rc >= 0)) { if (sock->state != SS_READY) sock->state = SS_CONNECTING; @@ -934,12 +934,12 @@ static int tipc_send_stream(struct kiocb *iocb, struct socket *sock, next: mtu = port->max_pkt; send = min_t(uint, dsz - sent, TIPC_MAX_USER_MSG_SIZE); - rc = tipc_msg_build2(mhdr, m->msg_iov, sent, send, mtu, &buf); + rc = tipc_msg_build(mhdr, m->msg_iov, sent, send, mtu, &buf); if (unlikely(rc < 0)) goto exit; do { if (likely(!tipc_sk_conn_cong(tsk))) { - rc = tipc_link_xmit2(buf, dnode, ref); + rc = tipc_link_xmit(buf, dnode, ref); if (likely(!rc)) { tsk->sent_unacked++; sent += send; @@ -1571,7 +1571,7 @@ static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *buf) if ((rc < 0) && !tipc_msg_reverse(buf, &onode, -rc)) return 0; - tipc_link_xmit2(buf, onode, 0); + tipc_link_xmit(buf, onode, 0); return 0; } @@ -1623,7 +1623,7 @@ exit: if ((rc < 0) && !tipc_msg_reverse(buf, &dnode, -rc)) return -EHOSTUNREACH; - tipc_link_xmit2(buf, dnode, 0); + tipc_link_xmit(buf, dnode, 0); return (rc < 0) ? -EHOSTUNREACH : 0; } @@ -1910,7 +1910,7 @@ restart: } tipc_port_disconnect(port->ref); if (tipc_msg_reverse(buf, &peer, TIPC_CONN_SHUTDOWN)) - tipc_link_xmit2(buf, peer, 0); + tipc_link_xmit(buf, peer, 0); } else { tipc_port_shutdown(port->ref); } -- cgit v1.2.3 From 52f50ce556af7008e1c2088d8b2b778ef5e79ff1 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Sun, 20 Jul 2014 13:14:28 +0800 Subject: tipc: fix sparse non static symbol warnings Fixes the following sparse warnings: net/tipc/socket.c:545:5: warning: symbol 'tipc_sk_proto_rcv' was not declared. Should it be static? net/tipc/socket.c:2015:5: warning: symbol 'tipc_ioctl' was not declared. Should it be static? Signed-off-by: Wei Yongjun Acked-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/socket.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 5ad42f6137d8..8477d08a6aa0 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -635,7 +635,8 @@ void tipc_sk_mcast_rcv(struct sk_buff *buf) * Returns 0 (TIPC_OK) if message was consumed, 1 (TIPC_FWD_MSG) if * (CONN_PROBE_REPLY) message should be forwarded. */ -int tipc_sk_proto_rcv(struct tipc_sock *tsk, u32 *dnode, struct sk_buff *buf) +static int tipc_sk_proto_rcv(struct tipc_sock *tsk, u32 *dnode, + struct sk_buff *buf) { struct tipc_msg *msg = buf_msg(buf); struct tipc_port *port = &tsk->port; @@ -2068,7 +2069,7 @@ static int tipc_getsockopt(struct socket *sock, int lvl, int opt, return put_user(sizeof(value), ol); } -int tipc_ioctl(struct socket *sk, unsigned int cmd, unsigned long arg) +static int tipc_ioctl(struct socket *sk, unsigned int cmd, unsigned long arg) { struct tipc_sioc_ln_req lnr; void __user *argp = (void __user *)arg; -- cgit v1.2.3 From ad025a56a5487d5d443ee790c04517581d39b21b Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Mon, 28 Jul 2014 21:30:14 +0800 Subject: tipc: remove duplicated include from socket.c Remove duplicated include. Signed-off-by: Wei Yongjun Signed-off-by: David S. Miller --- net/tipc/socket.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net/tipc/socket.c') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 8477d08a6aa0..7d423ee10897 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -40,7 +40,6 @@ #include "node.h" #include "link.h" #include -#include "link.h" #define SS_LISTENING -1 /* socket is listening */ #define SS_READY -2 /* socket is connectionless */ -- cgit v1.2.3