diff options
Diffstat (limited to 'net')
-rw-r--r-- | net/sunrpc/svc_xprt.c | 233 | ||||
-rw-r--r-- | net/sunrpc/svcsock.c | 157 | ||||
-rw-r--r-- | net/sunrpc/xprtrdma/svc_rdma_transport.c | 4 |
3 files changed, 190 insertions, 204 deletions
diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index bac973a..194d865 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -208,6 +208,35 @@ static struct svc_xprt *__svc_xpo_create(struct svc_xprt_class *xcl, return xcl->xcl_ops->xpo_create(serv, net, sap, len, flags); } +/* + * svc_xprt_received conditionally queues the transport for processing + * by another thread. The caller must hold the XPT_BUSY bit and must + * not thereafter touch transport data. + * + * Note: XPT_DATA only gets cleared when a read-attempt finds no (or + * insufficient) data. + */ +static void svc_xprt_received(struct svc_xprt *xprt) +{ + BUG_ON(!test_bit(XPT_BUSY, &xprt->xpt_flags)); + /* As soon as we clear busy, the xprt could be closed and + * 'put', so we need a reference to call svc_xprt_enqueue with: + */ + svc_xprt_get(xprt); + clear_bit(XPT_BUSY, &xprt->xpt_flags); + svc_xprt_enqueue(xprt); + svc_xprt_put(xprt); +} + +void svc_add_new_perm_xprt(struct svc_serv *serv, struct svc_xprt *new) +{ + clear_bit(XPT_TEMP, &new->xpt_flags); + spin_lock_bh(&serv->sv_lock); + list_add(&new->xpt_list, &serv->sv_permsocks); + spin_unlock_bh(&serv->sv_lock); + svc_xprt_received(new); +} + int svc_create_xprt(struct svc_serv *serv, const char *xprt_name, struct net *net, const int family, const unsigned short port, int flags) @@ -232,13 +261,8 @@ int svc_create_xprt(struct svc_serv *serv, const char *xprt_name, module_put(xcl->xcl_owner); return PTR_ERR(newxprt); } - - clear_bit(XPT_TEMP, &newxprt->xpt_flags); - spin_lock_bh(&serv->sv_lock); - list_add(&newxprt->xpt_list, &serv->sv_permsocks); - spin_unlock_bh(&serv->sv_lock); + svc_add_new_perm_xprt(serv, newxprt); newport = svc_xprt_local_port(newxprt); - clear_bit(XPT_BUSY, &newxprt->xpt_flags); return newport; } err: @@ -394,27 +418,6 @@ static struct svc_xprt *svc_xprt_dequeue(struct svc_pool *pool) return xprt; } -/* - * svc_xprt_received conditionally queues the transport for processing - * by another thread. The caller must hold the XPT_BUSY bit and must - * not thereafter touch transport data. - * - * Note: XPT_DATA only gets cleared when a read-attempt finds no (or - * insufficient) data. - */ -void svc_xprt_received(struct svc_xprt *xprt) -{ - BUG_ON(!test_bit(XPT_BUSY, &xprt->xpt_flags)); - /* As soon as we clear busy, the xprt could be closed and - * 'put', so we need a reference to call svc_xprt_enqueue with: - */ - svc_xprt_get(xprt); - clear_bit(XPT_BUSY, &xprt->xpt_flags); - svc_xprt_enqueue(xprt); - svc_xprt_put(xprt); -} -EXPORT_SYMBOL_GPL(svc_xprt_received); - /** * svc_reserve - change the space reserved for the reply to a request. * @rqstp: The request in question @@ -565,33 +568,12 @@ static void svc_check_conn_limits(struct svc_serv *serv) } } -/* - * Receive the next request on any transport. This code is carefully - * organised not to touch any cachelines in the shared svc_serv - * structure, only cachelines in the local svc_pool. - */ -int svc_recv(struct svc_rqst *rqstp, long timeout) +int svc_alloc_arg(struct svc_rqst *rqstp) { - struct svc_xprt *xprt = NULL; - struct svc_serv *serv = rqstp->rq_server; - struct svc_pool *pool = rqstp->rq_pool; - int len, i; - int pages; - struct xdr_buf *arg; - DECLARE_WAITQUEUE(wait, current); - long time_left; - - dprintk("svc: server %p waiting for data (to = %ld)\n", - rqstp, timeout); - - if (rqstp->rq_xprt) - printk(KERN_ERR - "svc_recv: service %p, transport not NULL!\n", - rqstp); - if (waitqueue_active(&rqstp->rq_wait)) - printk(KERN_ERR - "svc_recv: service %p, wait queue active!\n", - rqstp); + struct svc_serv *serv = rqstp->rq_server; + struct xdr_buf *arg; + int pages; + int i; /* now allocate needed pages. If we get a failure, sleep briefly */ pages = (serv->sv_max_mesg + PAGE_SIZE) / PAGE_SIZE; @@ -621,11 +603,15 @@ int svc_recv(struct svc_rqst *rqstp, long timeout) arg->page_len = (pages-2)*PAGE_SIZE; arg->len = (pages-1)*PAGE_SIZE; arg->tail[0].iov_len = 0; + return 0; +} - try_to_freeze(); - cond_resched(); - if (signalled() || kthread_should_stop()) - return -EINTR; +struct svc_xprt *svc_get_next_xprt(struct svc_rqst *rqstp, long timeout) +{ + struct svc_xprt *xprt; + struct svc_pool *pool = rqstp->rq_pool; + DECLARE_WAITQUEUE(wait, current); + long time_left; /* Normally we will wait up to 5 seconds for any required * cache information to be provided. @@ -663,7 +649,7 @@ int svc_recv(struct svc_rqst *rqstp, long timeout) if (kthread_should_stop()) { set_current_state(TASK_RUNNING); spin_unlock_bh(&pool->sp_lock); - return -EINTR; + return ERR_PTR(-EINTR); } add_wait_queue(&rqstp->rq_wait, &wait); @@ -684,48 +670,58 @@ int svc_recv(struct svc_rqst *rqstp, long timeout) spin_unlock_bh(&pool->sp_lock); dprintk("svc: server %p, no data yet\n", rqstp); if (signalled() || kthread_should_stop()) - return -EINTR; + return ERR_PTR(-EINTR); else - return -EAGAIN; + return ERR_PTR(-EAGAIN); } } spin_unlock_bh(&pool->sp_lock); + return xprt; +} + +void svc_add_new_temp_xprt(struct svc_serv *serv, struct svc_xprt *newxpt) +{ + spin_lock_bh(&serv->sv_lock); + set_bit(XPT_TEMP, &newxpt->xpt_flags); + list_add(&newxpt->xpt_list, &serv->sv_tempsocks); + serv->sv_tmpcnt++; + if (serv->sv_temptimer.function == NULL) { + /* setup timer to age temp transports */ + setup_timer(&serv->sv_temptimer, svc_age_temp_xprts, + (unsigned long)serv); + mod_timer(&serv->sv_temptimer, + jiffies + svc_conn_age_period * HZ); + } + spin_unlock_bh(&serv->sv_lock); + svc_xprt_received(newxpt); +} + +static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt) +{ + struct svc_serv *serv = rqstp->rq_server; + int len = 0; - len = 0; if (test_bit(XPT_CLOSE, &xprt->xpt_flags)) { dprintk("svc_recv: found XPT_CLOSE\n"); svc_delete_xprt(xprt); /* Leave XPT_BUSY set on the dead xprt: */ - goto out; + return 0; } if (test_bit(XPT_LISTENER, &xprt->xpt_flags)) { struct svc_xprt *newxpt; + /* + * We know this module_get will succeed because the + * listener holds a reference too + */ + __module_get(xprt->xpt_class->xcl_owner); + svc_check_conn_limits(xprt->xpt_server); newxpt = xprt->xpt_ops->xpo_accept(xprt); - if (newxpt) { - /* - * We know this module_get will succeed because the - * listener holds a reference too - */ - __module_get(newxpt->xpt_class->xcl_owner); - svc_check_conn_limits(xprt->xpt_server); - spin_lock_bh(&serv->sv_lock); - set_bit(XPT_TEMP, &newxpt->xpt_flags); - list_add(&newxpt->xpt_list, &serv->sv_tempsocks); - serv->sv_tmpcnt++; - if (serv->sv_temptimer.function == NULL) { - /* setup timer to age temp transports */ - setup_timer(&serv->sv_temptimer, - svc_age_temp_xprts, - (unsigned long)serv); - mod_timer(&serv->sv_temptimer, - jiffies + svc_conn_age_period * HZ); - } - spin_unlock_bh(&serv->sv_lock); - svc_xprt_received(newxpt); - } + if (newxpt) + svc_add_new_temp_xprt(serv, newxpt); } else if (xprt->xpt_ops->xpo_has_wspace(xprt)) { + /* XPT_DATA|XPT_DEFERRED case: */ dprintk("svc: server %p, pool %u, transport %p, inuse=%d\n", - rqstp, pool->sp_id, xprt, + rqstp, rqstp->rq_pool->sp_id, xprt, atomic_read(&xprt->xpt_ref.refcount)); rqstp->rq_deferred = svc_deferred_dequeue(xprt); if (rqstp->rq_deferred) @@ -736,10 +732,51 @@ int svc_recv(struct svc_rqst *rqstp, long timeout) rqstp->rq_reserved = serv->sv_max_mesg; atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved); } + /* clear XPT_BUSY: */ svc_xprt_received(xprt); + return len; +} + +/* + * Receive the next request on any transport. This code is carefully + * organised not to touch any cachelines in the shared svc_serv + * structure, only cachelines in the local svc_pool. + */ +int svc_recv(struct svc_rqst *rqstp, long timeout) +{ + struct svc_xprt *xprt = NULL; + struct svc_serv *serv = rqstp->rq_server; + int len, err; + + dprintk("svc: server %p waiting for data (to = %ld)\n", + rqstp, timeout); + + if (rqstp->rq_xprt) + printk(KERN_ERR + "svc_recv: service %p, transport not NULL!\n", + rqstp); + if (waitqueue_active(&rqstp->rq_wait)) + printk(KERN_ERR + "svc_recv: service %p, wait queue active!\n", + rqstp); + + err = svc_alloc_arg(rqstp); + if (err) + return err; + + try_to_freeze(); + cond_resched(); + if (signalled() || kthread_should_stop()) + return -EINTR; + + xprt = svc_get_next_xprt(rqstp, timeout); + if (IS_ERR(xprt)) + return PTR_ERR(xprt); + + len = svc_handle_xprt(rqstp, xprt); /* No data, incomplete (TCP) read, or accept() */ - if (len == 0 || len == -EAGAIN) + if (len <= 0) goto out; clear_bit(XPT_OLD, &xprt->xpt_flags); @@ -917,16 +954,18 @@ void svc_close_xprt(struct svc_xprt *xprt) } EXPORT_SYMBOL_GPL(svc_close_xprt); -static void svc_close_list(struct list_head *xprt_list, struct net *net) +static void svc_close_list(struct svc_serv *serv, struct list_head *xprt_list, struct net *net) { struct svc_xprt *xprt; + spin_lock(&serv->sv_lock); list_for_each_entry(xprt, xprt_list, xpt_list) { if (xprt->xpt_net != net) continue; set_bit(XPT_CLOSE, &xprt->xpt_flags); set_bit(XPT_BUSY, &xprt->xpt_flags); } + spin_unlock(&serv->sv_lock); } static void svc_clear_pools(struct svc_serv *serv, struct net *net) @@ -949,24 +988,28 @@ static void svc_clear_pools(struct svc_serv *serv, struct net *net) } } -static void svc_clear_list(struct list_head *xprt_list, struct net *net) +static void svc_clear_list(struct svc_serv *serv, struct list_head *xprt_list, struct net *net) { struct svc_xprt *xprt; struct svc_xprt *tmp; + LIST_HEAD(victims); + spin_lock(&serv->sv_lock); list_for_each_entry_safe(xprt, tmp, xprt_list, xpt_list) { if (xprt->xpt_net != net) continue; - svc_delete_xprt(xprt); + list_move(&xprt->xpt_list, &victims); } - list_for_each_entry(xprt, xprt_list, xpt_list) - BUG_ON(xprt->xpt_net == net); + spin_unlock(&serv->sv_lock); + + list_for_each_entry_safe(xprt, tmp, &victims, xpt_list) + svc_delete_xprt(xprt); } void svc_close_net(struct svc_serv *serv, struct net *net) { - svc_close_list(&serv->sv_tempsocks, net); - svc_close_list(&serv->sv_permsocks, net); + svc_close_list(serv, &serv->sv_tempsocks, net); + svc_close_list(serv, &serv->sv_permsocks, net); svc_clear_pools(serv, net); /* @@ -974,8 +1017,8 @@ void svc_close_net(struct svc_serv *serv, struct net *net) * svc_xprt_enqueue will not add new entries without taking the * sp_lock and checking XPT_BUSY. */ - svc_clear_list(&serv->sv_tempsocks, net); - svc_clear_list(&serv->sv_permsocks, net); + svc_clear_list(serv, &serv->sv_tempsocks, net); + svc_clear_list(serv, &serv->sv_permsocks, net); } /* diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 998aa8c..03827ce 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -59,7 +59,7 @@ static struct svc_sock *svc_setup_socket(struct svc_serv *, struct socket *, - int *errp, int flags); + int flags); static void svc_udp_data_ready(struct sock *, int); static int svc_udp_recvfrom(struct svc_rqst *); static int svc_udp_sendto(struct svc_rqst *); @@ -305,57 +305,6 @@ static int svc_one_sock_name(struct svc_sock *svsk, char *buf, int remaining) return len; } -/** - * svc_sock_names - construct a list of listener names in a string - * @serv: pointer to RPC service - * @buf: pointer to a buffer to fill in with socket names - * @buflen: size of the buffer to be filled - * @toclose: pointer to '\0'-terminated C string containing the name - * of a listener to be closed - * - * Fills in @buf with a '\n'-separated list of names of listener - * sockets. If @toclose is not NULL, the socket named by @toclose - * is closed, and is not included in the output list. - * - * Returns positive length of the socket name string, or a negative - * errno value on error. - */ -int svc_sock_names(struct svc_serv *serv, char *buf, const size_t buflen, - const char *toclose) -{ - struct svc_sock *svsk, *closesk = NULL; - int len = 0; - - if (!serv) - return 0; - - spin_lock_bh(&serv->sv_lock); - list_for_each_entry(svsk, &serv->sv_permsocks, sk_xprt.xpt_list) { - int onelen = svc_one_sock_name(svsk, buf + len, buflen - len); - if (onelen < 0) { - len = onelen; - break; - } - if (toclose && strcmp(toclose, buf + len) == 0) { - closesk = svsk; - svc_xprt_get(&closesk->sk_xprt); - } else - len += onelen; - } - spin_unlock_bh(&serv->sv_lock); - - if (closesk) { - /* Should unregister with portmap, but you cannot - * unregister just one protocol... - */ - svc_close_xprt(&closesk->sk_xprt); - svc_xprt_put(&closesk->sk_xprt); - } else if (toclose) - return -ENOENT; - return len; -} -EXPORT_SYMBOL_GPL(svc_sock_names); - /* * Check input queue length */ @@ -598,11 +547,9 @@ static int svc_udp_recvfrom(struct svc_rqst *rqstp) dprintk("svc: recvfrom returned error %d\n", -err); set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags); } - return -EAGAIN; + return 0; } len = svc_addr_len(svc_addr(rqstp)); - if (len == 0) - return -EAFNOSUPPORT; rqstp->rq_addrlen = len; if (skb->tstamp.tv64 == 0) { skb->tstamp = ktime_get_real(); @@ -620,10 +567,7 @@ static int svc_udp_recvfrom(struct svc_rqst *rqstp) if (!svc_udp_get_dest_address(rqstp, cmh)) { net_warn_ratelimited("svc: received unknown control message %d/%d; dropping RPC reply datagram\n", cmh->cmsg_level, cmh->cmsg_type); -out_free: - trace_kfree_skb(skb, svc_udp_recvfrom); - skb_free_datagram_locked(svsk->sk_sk, skb); - return 0; + goto out_free; } rqstp->rq_daddrlen = svc_addr_len(svc_daddr(rqstp)); @@ -662,6 +606,10 @@ out_free: serv->sv_stats->netudpcnt++; return len; +out_free: + trace_kfree_skb(skb, svc_udp_recvfrom); + skb_free_datagram_locked(svsk->sk_sk, skb); + return 0; } static int @@ -900,8 +848,9 @@ static struct svc_xprt *svc_tcp_accept(struct svc_xprt *xprt) */ newsock->sk->sk_sndtimeo = HZ*30; - if (!(newsvsk = svc_setup_socket(serv, newsock, &err, - (SVC_SOCK_ANONYMOUS | SVC_SOCK_TEMPORARY)))) + newsvsk = svc_setup_socket(serv, newsock, + (SVC_SOCK_ANONYMOUS | SVC_SOCK_TEMPORARY)); + if (IS_ERR(newsvsk)) goto failed; svc_xprt_set_remote(&newsvsk->sk_xprt, sin, slen); err = kernel_getsockname(newsock, sin, &slen); @@ -1174,13 +1123,13 @@ error: if (len != -EAGAIN) goto err_other; dprintk("RPC: TCP recvfrom got EAGAIN\n"); - return -EAGAIN; + return 0; err_other: printk(KERN_NOTICE "%s: recvfrom returned errno %d\n", svsk->sk_xprt.xpt_server->sv_name, -len); set_bit(XPT_CLOSE, &svsk->sk_xprt.xpt_flags); err_noclose: - return -EAGAIN; /* record not complete */ + return 0; /* record not complete */ } /* @@ -1383,29 +1332,29 @@ EXPORT_SYMBOL_GPL(svc_sock_update_bufs); */ static struct svc_sock *svc_setup_socket(struct svc_serv *serv, struct socket *sock, - int *errp, int flags) + int flags) { struct svc_sock *svsk; struct sock *inet; int pmap_register = !(flags & SVC_SOCK_ANONYMOUS); + int err = 0; dprintk("svc: svc_setup_socket %p\n", sock); - if (!(svsk = kzalloc(sizeof(*svsk), GFP_KERNEL))) { - *errp = -ENOMEM; - return NULL; - } + svsk = kzalloc(sizeof(*svsk), GFP_KERNEL); + if (!svsk) + return ERR_PTR(-ENOMEM); inet = sock->sk; /* Register socket with portmapper */ - if (*errp >= 0 && pmap_register) - *errp = svc_register(serv, sock_net(sock->sk), inet->sk_family, + if (pmap_register) + err = svc_register(serv, sock_net(sock->sk), inet->sk_family, inet->sk_protocol, ntohs(inet_sk(inet)->inet_sport)); - if (*errp < 0) { + if (err < 0) { kfree(svsk); - return NULL; + return ERR_PTR(err); } inet->sk_user_data = svsk; @@ -1450,42 +1399,38 @@ int svc_addsock(struct svc_serv *serv, const int fd, char *name_return, int err = 0; struct socket *so = sockfd_lookup(fd, &err); struct svc_sock *svsk = NULL; + struct sockaddr_storage addr; + struct sockaddr *sin = (struct sockaddr *)&addr; + int salen; if (!so) return err; + err = -EAFNOSUPPORT; if ((so->sk->sk_family != PF_INET) && (so->sk->sk_family != PF_INET6)) - err = -EAFNOSUPPORT; - else if (so->sk->sk_protocol != IPPROTO_TCP && + goto out; + err = -EPROTONOSUPPORT; + if (so->sk->sk_protocol != IPPROTO_TCP && so->sk->sk_protocol != IPPROTO_UDP) - err = -EPROTONOSUPPORT; - else if (so->state > SS_UNCONNECTED) - err = -EISCONN; - else { - if (!try_module_get(THIS_MODULE)) - err = -ENOENT; - else - svsk = svc_setup_socket(serv, so, &err, - SVC_SOCK_DEFAULTS); - if (svsk) { - struct sockaddr_storage addr; - struct sockaddr *sin = (struct sockaddr *)&addr; - int salen; - if (kernel_getsockname(svsk->sk_sock, sin, &salen) == 0) - svc_xprt_set_local(&svsk->sk_xprt, sin, salen); - clear_bit(XPT_TEMP, &svsk->sk_xprt.xpt_flags); - spin_lock_bh(&serv->sv_lock); - list_add(&svsk->sk_xprt.xpt_list, &serv->sv_permsocks); - spin_unlock_bh(&serv->sv_lock); - svc_xprt_received(&svsk->sk_xprt); - err = 0; - } else - module_put(THIS_MODULE); - } - if (err) { - sockfd_put(so); - return err; + goto out; + err = -EISCONN; + if (so->state > SS_UNCONNECTED) + goto out; + err = -ENOENT; + if (!try_module_get(THIS_MODULE)) + goto out; + svsk = svc_setup_socket(serv, so, SVC_SOCK_DEFAULTS); + if (IS_ERR(svsk)) { + module_put(THIS_MODULE); + err = PTR_ERR(svsk); + goto out; } + if (kernel_getsockname(svsk->sk_sock, sin, &salen) == 0) + svc_xprt_set_local(&svsk->sk_xprt, sin, salen); + svc_add_new_perm_xprt(serv, &svsk->sk_xprt); return svc_one_sock_name(svsk, name_return, len); +out: + sockfd_put(so); + return err; } EXPORT_SYMBOL_GPL(svc_addsock); @@ -1563,11 +1508,13 @@ static struct svc_xprt *svc_create_socket(struct svc_serv *serv, goto bummer; } - if ((svsk = svc_setup_socket(serv, sock, &error, flags)) != NULL) { - svc_xprt_set_local(&svsk->sk_xprt, newsin, newlen); - return (struct svc_xprt *)svsk; + svsk = svc_setup_socket(serv, sock, flags); + if (IS_ERR(svsk)) { + error = PTR_ERR(svsk); + goto bummer; } - + svc_xprt_set_local(&svsk->sk_xprt, newsin, newlen); + return (struct svc_xprt *)svsk; bummer: dprintk("svc: svc_create_socket error = %d\n", -error); sock_release(sock); diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c b/net/sunrpc/xprtrdma/svc_rdma_transport.c index 73b428b..62e4f9b 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_transport.c +++ b/net/sunrpc/xprtrdma/svc_rdma_transport.c @@ -578,10 +578,6 @@ static void handle_connect_req(struct rdma_cm_id *new_cma_id, size_t client_ird) list_add_tail(&newxprt->sc_accept_q, &listen_xprt->sc_accept_q); spin_unlock_bh(&listen_xprt->sc_lock); - /* - * Can't use svc_xprt_received here because we are not on a - * rqstp thread - */ set_bit(XPT_CONN, &listen_xprt->sc_xprt.xpt_flags); svc_xprt_enqueue(&listen_xprt->sc_xprt); } |