diff options
Diffstat (limited to 'drivers/staging/lustre/lnet/klnds')
10 files changed, 2053 insertions, 1701 deletions
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c index cb74ae731..0d32e6541 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c @@ -42,15 +42,7 @@ #include <asm/page.h> #include "o2iblnd.h" -static lnd_t the_o2iblnd = { - .lnd_type = O2IBLND, - .lnd_startup = kiblnd_startup, - .lnd_shutdown = kiblnd_shutdown, - .lnd_ctl = kiblnd_ctl, - .lnd_query = kiblnd_query, - .lnd_send = kiblnd_send, - .lnd_recv = kiblnd_recv, -}; +static lnd_t the_o2iblnd; kib_data_t kiblnd_data; @@ -63,7 +55,7 @@ static __u32 kiblnd_cksum(void *ptr, int nob) sum = ((sum << 1) | (sum >> 31)) + *c++; /* ensure I don't return 0 (== no checksum) */ - return (sum == 0) ? 1 : sum; + return !sum ? 1 : sum; } static char *kiblnd_msgtype2str(int type) @@ -145,7 +137,7 @@ static int kiblnd_unpack_rd(kib_msg_t *msg, int flip) int i; LASSERT(msg->ibm_type == IBLND_MSG_GET_REQ || - msg->ibm_type == IBLND_MSG_PUT_ACK); + msg->ibm_type == IBLND_MSG_PUT_ACK); rd = msg->ibm_type == IBLND_MSG_GET_REQ ? &msg->ibm_u.get.ibgm_rd : @@ -189,8 +181,10 @@ void kiblnd_pack_msg(lnet_ni_t *ni, kib_msg_t *msg, int version, { kib_net_t *net = ni->ni_data; - /* CAVEAT EMPTOR! all message fields not set here should have been - * initialised previously. */ + /* + * CAVEAT EMPTOR! all message fields not set here should have been + * initialised previously. + */ msg->ibm_magic = IBLND_MSG_MAGIC; msg->ibm_version = version; /* ibm_type */ @@ -249,11 +243,13 @@ int kiblnd_unpack_msg(kib_msg_t *msg, int nob) return -EPROTO; } - /* checksum must be computed with ibm_cksum zero and BEFORE anything - * gets flipped */ + /* + * checksum must be computed with ibm_cksum zero and BEFORE anything + * gets flipped + */ msg_cksum = flip ? __swab32(msg->ibm_cksum) : msg->ibm_cksum; msg->ibm_cksum = 0; - if (msg_cksum != 0 && + if (msg_cksum && msg_cksum != kiblnd_cksum(msg, msg_nob)) { CERROR("Bad checksum\n"); return -EPROTO; @@ -326,21 +322,21 @@ int kiblnd_create_peer(lnet_ni_t *ni, kib_peer_t **peerp, lnet_nid_t nid) int cpt = lnet_cpt_of_nid(nid); unsigned long flags; - LASSERT(net != NULL); + LASSERT(net); LASSERT(nid != LNET_NID_ANY); LIBCFS_CPT_ALLOC(peer, lnet_cpt_table(), cpt, sizeof(*peer)); - if (peer == NULL) { + if (!peer) { CERROR("Cannot allocate peer\n"); return -ENOMEM; } - memset(peer, 0, sizeof(*peer)); /* zero flags etc */ - peer->ibp_ni = ni; peer->ibp_nid = nid; peer->ibp_error = 0; peer->ibp_last_alive = 0; + peer->ibp_max_frags = IBLND_CFG_RDMA_FRAGS; + peer->ibp_queue_depth = *kiblnd_tunables.kib_peertxcredits; atomic_set(&peer->ibp_refcount, 1); /* 1 ref for caller */ INIT_LIST_HEAD(&peer->ibp_list); /* not in the peer table yet */ @@ -350,7 +346,7 @@ int kiblnd_create_peer(lnet_ni_t *ni, kib_peer_t **peerp, lnet_nid_t nid) write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); /* always called with a ref on ni, which prevents ni being shutdown */ - LASSERT(net->ibn_shutdown == 0); + LASSERT(!net->ibn_shutdown); /* npeers only grows with the global lock held */ atomic_inc(&net->ibn_npeers); @@ -365,38 +361,36 @@ void kiblnd_destroy_peer(kib_peer_t *peer) { kib_net_t *net = peer->ibp_ni->ni_data; - LASSERT(net != NULL); - LASSERT(atomic_read(&peer->ibp_refcount) == 0); + LASSERT(net); + LASSERT(!atomic_read(&peer->ibp_refcount)); LASSERT(!kiblnd_peer_active(peer)); - LASSERT(peer->ibp_connecting == 0); - LASSERT(peer->ibp_accepting == 0); - LASSERT(list_empty(&peer->ibp_conns)); + LASSERT(kiblnd_peer_idle(peer)); LASSERT(list_empty(&peer->ibp_tx_queue)); LIBCFS_FREE(peer, sizeof(*peer)); - /* NB a peer's connections keep a reference on their peer until + /* + * NB a peer's connections keep a reference on their peer until * they are destroyed, so we can be assured that _all_ state to do * with this peer has been cleaned up when its refcount drops to - * zero. */ + * zero. + */ atomic_dec(&net->ibn_npeers); } kib_peer_t *kiblnd_find_peer_locked(lnet_nid_t nid) { - /* the caller is responsible for accounting the additional reference - * that this creates */ + /* + * the caller is responsible for accounting the additional reference + * that this creates + */ struct list_head *peer_list = kiblnd_nid2peerlist(nid); struct list_head *tmp; kib_peer_t *peer; list_for_each(tmp, peer_list) { - peer = list_entry(tmp, kib_peer_t, ibp_list); - - LASSERT(peer->ibp_connecting > 0 || /* creating conns */ - peer->ibp_accepting > 0 || - !list_empty(&peer->ibp_conns)); /* active conn */ + LASSERT(!kiblnd_peer_idle(peer)); if (peer->ibp_nid != nid) continue; @@ -431,13 +425,9 @@ static int kiblnd_get_peer_info(lnet_ni_t *ni, int index, read_lock_irqsave(&kiblnd_data.kib_global_lock, flags); for (i = 0; i < kiblnd_data.kib_peer_hash_size; i++) { - list_for_each(ptmp, &kiblnd_data.kib_peers[i]) { - peer = list_entry(ptmp, kib_peer_t, ibp_list); - LASSERT(peer->ibp_connecting > 0 || - peer->ibp_accepting > 0 || - !list_empty(&peer->ibp_conns)); + LASSERT(!kiblnd_peer_idle(peer)); if (peer->ibp_ni != ni) continue; @@ -474,8 +464,10 @@ static void kiblnd_del_peer_locked(kib_peer_t *peer) } /* NB closing peer's last conn unlinked it. */ } - /* NB peer now unlinked; might even be freed if the peer table had the - * last ref on it. */ + /* + * NB peer now unlinked; might even be freed if the peer table had the + * last ref on it. + */ } static int kiblnd_del_peer(lnet_ni_t *ni, lnet_nid_t nid) @@ -493,7 +485,8 @@ static int kiblnd_del_peer(lnet_ni_t *ni, lnet_nid_t nid) write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); if (nid != LNET_NID_ANY) { - lo = hi = kiblnd_nid2peerlist(nid) - kiblnd_data.kib_peers; + lo = kiblnd_nid2peerlist(nid) - kiblnd_data.kib_peers; + hi = kiblnd_nid2peerlist(nid) - kiblnd_data.kib_peers; } else { lo = 0; hi = kiblnd_data.kib_peer_hash_size - 1; @@ -502,9 +495,7 @@ static int kiblnd_del_peer(lnet_ni_t *ni, lnet_nid_t nid) for (i = lo; i <= hi; i++) { list_for_each_safe(ptmp, pnxt, &kiblnd_data.kib_peers[i]) { peer = list_entry(ptmp, kib_peer_t, ibp_list); - LASSERT(peer->ibp_connecting > 0 || - peer->ibp_accepting > 0 || - !list_empty(&peer->ibp_conns)); + LASSERT(!kiblnd_peer_idle(peer)); if (peer->ibp_ni != ni) continue; @@ -516,7 +507,7 @@ static int kiblnd_del_peer(lnet_ni_t *ni, lnet_nid_t nid) LASSERT(list_empty(&peer->ibp_conns)); list_splice_init(&peer->ibp_tx_queue, - &zombies); + &zombies); } kiblnd_del_peer_locked(peer); @@ -544,11 +535,8 @@ static kib_conn_t *kiblnd_get_conn_by_idx(lnet_ni_t *ni, int index) for (i = 0; i < kiblnd_data.kib_peer_hash_size; i++) { list_for_each(ptmp, &kiblnd_data.kib_peers[i]) { - peer = list_entry(ptmp, kib_peer_t, ibp_list); - LASSERT(peer->ibp_connecting > 0 || - peer->ibp_accepting > 0 || - !list_empty(&peer->ibp_conns)); + LASSERT(!kiblnd_peer_idle(peer)); if (peer->ibp_ni != ni) continue; @@ -558,7 +546,7 @@ static kib_conn_t *kiblnd_get_conn_by_idx(lnet_ni_t *ni, int index) continue; conn = list_entry(ctmp, kib_conn_t, - ibc_list); + ibc_list); kiblnd_conn_addref(conn); read_unlock_irqrestore( &kiblnd_data.kib_global_lock, @@ -597,12 +585,12 @@ static void kiblnd_setup_mtu_locked(struct rdma_cm_id *cmid) int mtu; /* XXX There is no path record for iWARP, set by netdev->change_mtu? */ - if (cmid->route.path_rec == NULL) + if (!cmid->route.path_rec) return; mtu = kiblnd_translate_mtu(*kiblnd_tunables.kib_ib_mtu); LASSERT(mtu >= 0); - if (mtu != 0) + if (mtu) cmid->route.path_rec->mtu = mtu; } @@ -619,13 +607,13 @@ static int kiblnd_get_completion_vector(kib_conn_t *conn, int cpt) return 0; mask = cfs_cpt_cpumask(lnet_cpt_table(), cpt); - if (mask == NULL) + if (!mask) return 0; /* hash NID to CPU id in this partition... */ off = do_div(nid, cpumask_weight(mask)); for_each_cpu(i, mask) { - if (off-- == 0) + if (!off--) return i % vectors; } @@ -634,15 +622,17 @@ static int kiblnd_get_completion_vector(kib_conn_t *conn, int cpt) } kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, - int state, int version) + int state, int version) { - /* CAVEAT EMPTOR: + /* + * CAVEAT EMPTOR: * If the new conn is created successfully it takes over the caller's * ref on 'peer'. It also "owns" 'cmid' and destroys it when it itself * is destroyed. On failure, the caller's ref on 'peer' remains and * she must dispose of 'cmid'. (Actually I'd block forever if I tried * to destroy 'cmid' here since I'm called from the CM which still has - * its ref on 'cmid'). */ + * its ref on 'cmid'). + */ rwlock_t *glock = &kiblnd_data.kib_global_lock; kib_net_t *net = peer->ibp_ni->ni_data; kib_dev_t *dev; @@ -656,7 +646,7 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, int rc; int i; - LASSERT(net != NULL); + LASSERT(net); LASSERT(!in_interrupt()); dev = net->ibn_dev; @@ -668,14 +658,14 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, LIBCFS_CPT_ALLOC(init_qp_attr, lnet_cpt_table(), cpt, sizeof(*init_qp_attr)); - if (init_qp_attr == NULL) { + if (!init_qp_attr) { CERROR("Can't allocate qp_attr for %s\n", libcfs_nid2str(peer->ibp_nid)); goto failed_0; } LIBCFS_CPT_ALLOC(conn, lnet_cpt_table(), cpt, sizeof(*conn)); - if (conn == NULL) { + if (!conn) { CERROR("Can't allocate connection for %s\n", libcfs_nid2str(peer->ibp_nid)); goto failed_1; @@ -686,6 +676,8 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, conn->ibc_peer = peer; /* I take the caller's ref */ cmid->context = conn; /* for future CM callbacks */ conn->ibc_cmid = cmid; + conn->ibc_max_frags = peer->ibp_max_frags; + conn->ibc_queue_depth = peer->ibp_queue_depth; INIT_LIST_HEAD(&conn->ibc_early_rxs); INIT_LIST_HEAD(&conn->ibc_tx_noops); @@ -697,7 +689,7 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, LIBCFS_CPT_ALLOC(conn->ibc_connvars, lnet_cpt_table(), cpt, sizeof(*conn->ibc_connvars)); - if (conn->ibc_connvars == NULL) { + if (!conn->ibc_connvars) { CERROR("Can't allocate in-progress connection state\n"); goto failed_2; } @@ -731,42 +723,42 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, write_unlock_irqrestore(glock, flags); LIBCFS_CPT_ALLOC(conn->ibc_rxs, lnet_cpt_table(), cpt, - IBLND_RX_MSGS(version) * sizeof(kib_rx_t)); - if (conn->ibc_rxs == NULL) { + IBLND_RX_MSGS(conn) * sizeof(kib_rx_t)); + if (!conn->ibc_rxs) { CERROR("Cannot allocate RX buffers\n"); goto failed_2; } rc = kiblnd_alloc_pages(&conn->ibc_rx_pages, cpt, - IBLND_RX_MSG_PAGES(version)); - if (rc != 0) + IBLND_RX_MSG_PAGES(conn)); + if (rc) goto failed_2; kiblnd_map_rx_descs(conn); - cq_attr.cqe = IBLND_CQ_ENTRIES(version); + cq_attr.cqe = IBLND_CQ_ENTRIES(conn); cq_attr.comp_vector = kiblnd_get_completion_vector(conn, cpt); cq = ib_create_cq(cmid->device, kiblnd_cq_completion, kiblnd_cq_event, conn, &cq_attr); if (IS_ERR(cq)) { - CERROR("Can't create CQ: %ld, cqe: %d\n", - PTR_ERR(cq), IBLND_CQ_ENTRIES(version)); + CERROR("Failed to create CQ with %d CQEs: %ld\n", + IBLND_CQ_ENTRIES(conn), PTR_ERR(cq)); goto failed_2; } conn->ibc_cq = cq; rc = ib_req_notify_cq(cq, IB_CQ_NEXT_COMP); - if (rc != 0) { - CERROR("Can't request completion notificiation: %d\n", rc); + if (rc) { + CERROR("Can't request completion notification: %d\n", rc); goto failed_2; } init_qp_attr->event_handler = kiblnd_qp_event; init_qp_attr->qp_context = conn; - init_qp_attr->cap.max_send_wr = IBLND_SEND_WRS(version); - init_qp_attr->cap.max_recv_wr = IBLND_RECV_WRS(version); + init_qp_attr->cap.max_send_wr = IBLND_SEND_WRS(conn); + init_qp_attr->cap.max_recv_wr = IBLND_RECV_WRS(conn); init_qp_attr->cap.max_send_sge = 1; init_qp_attr->cap.max_recv_sge = 1; init_qp_attr->sq_sig_type = IB_SIGNAL_REQ_WR; @@ -777,7 +769,7 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, conn->ibc_sched = sched; rc = rdma_create_qp(cmid, conn->ibc_hdev->ibh_pd, init_qp_attr); - if (rc != 0) { + if (rc) { CERROR("Can't create QP: %d, send_wr: %d, recv_wr: %d\n", rc, init_qp_attr->cap.max_send_wr, init_qp_attr->cap.max_recv_wr); @@ -787,33 +779,37 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, LIBCFS_FREE(init_qp_attr, sizeof(*init_qp_attr)); /* 1 ref for caller and each rxmsg */ - atomic_set(&conn->ibc_refcount, 1 + IBLND_RX_MSGS(version)); - conn->ibc_nrx = IBLND_RX_MSGS(version); + atomic_set(&conn->ibc_refcount, 1 + IBLND_RX_MSGS(conn)); + conn->ibc_nrx = IBLND_RX_MSGS(conn); /* post receives */ - for (i = 0; i < IBLND_RX_MSGS(version); i++) { + for (i = 0; i < IBLND_RX_MSGS(conn); i++) { rc = kiblnd_post_rx(&conn->ibc_rxs[i], IBLND_POSTRX_NO_CREDIT); - if (rc != 0) { + if (rc) { CERROR("Can't post rxmsg: %d\n", rc); /* Make posted receives complete */ kiblnd_abort_receives(conn); - /* correct # of posted buffers - * NB locking needed now I'm racing with completion */ + /* + * correct # of posted buffers + * NB locking needed now I'm racing with completion + */ spin_lock_irqsave(&sched->ibs_lock, flags); - conn->ibc_nrx -= IBLND_RX_MSGS(version) - i; + conn->ibc_nrx -= IBLND_RX_MSGS(conn) - i; spin_unlock_irqrestore(&sched->ibs_lock, flags); - /* cmid will be destroyed by CM(ofed) after cm_callback + /* + * cmid will be destroyed by CM(ofed) after cm_callback * returned, so we can't refer it anymore - * (by kiblnd_connd()->kiblnd_destroy_conn) */ + * (by kiblnd_connd()->kiblnd_destroy_conn) + */ rdma_destroy_qp(conn->ibc_cmid); conn->ibc_cmid = NULL; /* Drop my own and unused rxbuffer refcounts */ - while (i++ <= IBLND_RX_MSGS(version)) + while (i++ <= IBLND_RX_MSGS(conn)) kiblnd_conn_decref(conn); return NULL; @@ -822,7 +818,7 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, /* Init successful! */ LASSERT(state == IBLND_CONN_ACTIVE_CONNECT || - state == IBLND_CONN_PASSIVE_WAIT); + state == IBLND_CONN_PASSIVE_WAIT); conn->ibc_state = state; /* 1 more conn */ @@ -830,29 +826,29 @@ kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, return conn; failed_2: - kiblnd_destroy_conn(conn); + kiblnd_destroy_conn(conn, true); failed_1: LIBCFS_FREE(init_qp_attr, sizeof(*init_qp_attr)); failed_0: return NULL; } -void kiblnd_destroy_conn(kib_conn_t *conn) +void kiblnd_destroy_conn(kib_conn_t *conn, bool free_conn) { struct rdma_cm_id *cmid = conn->ibc_cmid; kib_peer_t *peer = conn->ibc_peer; int rc; LASSERT(!in_interrupt()); - LASSERT(atomic_read(&conn->ibc_refcount) == 0); + LASSERT(!atomic_read(&conn->ibc_refcount)); LASSERT(list_empty(&conn->ibc_early_rxs)); LASSERT(list_empty(&conn->ibc_tx_noops)); LASSERT(list_empty(&conn->ibc_tx_queue)); LASSERT(list_empty(&conn->ibc_tx_queue_rsrvd)); LASSERT(list_empty(&conn->ibc_tx_queue_nocred)); LASSERT(list_empty(&conn->ibc_active_txs)); - LASSERT(conn->ibc_noops_posted == 0); - LASSERT(conn->ibc_nsends_posted == 0); + LASSERT(!conn->ibc_noops_posted); + LASSERT(!conn->ibc_nsends_posted); switch (conn->ibc_state) { default: @@ -861,7 +857,7 @@ void kiblnd_destroy_conn(kib_conn_t *conn) case IBLND_CONN_DISCONNECTED: /* connvars should have been freed already */ - LASSERT(conn->ibc_connvars == NULL); + LASSERT(!conn->ibc_connvars); break; case IBLND_CONN_INIT: @@ -869,28 +865,27 @@ void kiblnd_destroy_conn(kib_conn_t *conn) } /* conn->ibc_cmid might be destroyed by CM already */ - if (cmid != NULL && cmid->qp != NULL) + if (cmid && cmid->qp) rdma_destroy_qp(cmid); - if (conn->ibc_cq != NULL) { + if (conn->ibc_cq) { rc = ib_destroy_cq(conn->ibc_cq); - if (rc != 0) + if (rc) CWARN("Error destroying CQ: %d\n", rc); } - if (conn->ibc_rx_pages != NULL) + if (conn->ibc_rx_pages) kiblnd_unmap_rx_descs(conn); - if (conn->ibc_rxs != NULL) { + if (conn->ibc_rxs) { LIBCFS_FREE(conn->ibc_rxs, - IBLND_RX_MSGS(conn->ibc_version) - * sizeof(kib_rx_t)); + IBLND_RX_MSGS(conn) * sizeof(kib_rx_t)); } - if (conn->ibc_connvars != NULL) + if (conn->ibc_connvars) LIBCFS_FREE(conn->ibc_connvars, sizeof(*conn->ibc_connvars)); - if (conn->ibc_hdev != NULL) + if (conn->ibc_hdev) kiblnd_hdev_decref(conn->ibc_hdev); /* See CAVEAT EMPTOR above in kiblnd_create_conn */ @@ -927,7 +922,7 @@ int kiblnd_close_peer_conns_locked(kib_peer_t *peer, int why) } int kiblnd_close_stale_conns_locked(kib_peer_t *peer, - int version, __u64 incarnation) + int version, __u64 incarnation) { kib_conn_t *conn; struct list_head *ctmp; @@ -967,20 +962,18 @@ static int kiblnd_close_matching_conns(lnet_ni_t *ni, lnet_nid_t nid) write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); - if (nid != LNET_NID_ANY) - lo = hi = kiblnd_nid2peerlist(nid) - kiblnd_data.kib_peers; - else { + if (nid != LNET_NID_ANY) { + lo = kiblnd_nid2peerlist(nid) - kiblnd_data.kib_peers; + hi = kiblnd_nid2peerlist(nid) - kiblnd_data.kib_peers; + } else { lo = 0; hi = kiblnd_data.kib_peer_hash_size - 1; } for (i = lo; i <= hi; i++) { list_for_each_safe(ptmp, pnxt, &kiblnd_data.kib_peers[i]) { - peer = list_entry(ptmp, kib_peer_t, ibp_list); - LASSERT(peer->ibp_connecting > 0 || - peer->ibp_accepting > 0 || - !list_empty(&peer->ibp_conns)); + LASSERT(!kiblnd_peer_idle(peer)); if (peer->ibp_ni != ni) continue; @@ -998,10 +991,10 @@ static int kiblnd_close_matching_conns(lnet_ni_t *ni, lnet_nid_t nid) if (nid == LNET_NID_ANY) return 0; - return (count == 0) ? -ENOENT : 0; + return !count ? -ENOENT : 0; } -int kiblnd_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) +static int kiblnd_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) { struct libcfs_ioctl_data *data = arg; int rc = -EINVAL; @@ -1027,14 +1020,14 @@ int kiblnd_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) rc = 0; conn = kiblnd_get_conn_by_idx(ni, data->ioc_count); - if (conn == NULL) { + if (!conn) { rc = -ENOENT; break; } - LASSERT(conn->ibc_cmid != NULL); + LASSERT(conn->ibc_cmid); data->ioc_nid = conn->ibc_peer->ibp_nid; - if (conn->ibc_cmid->route.path_rec == NULL) + if (!conn->ibc_cmid->route.path_rec) data->ioc_u32[0] = 0; /* iWarp has no path MTU */ else data->ioc_u32[0] = @@ -1054,7 +1047,7 @@ int kiblnd_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) return rc; } -void kiblnd_query(lnet_ni_t *ni, lnet_nid_t nid, unsigned long *when) +static void kiblnd_query(lnet_ni_t *ni, lnet_nid_t nid, unsigned long *when) { unsigned long last_alive = 0; unsigned long now = cfs_time_current(); @@ -1065,21 +1058,19 @@ void kiblnd_query(lnet_ni_t *ni, lnet_nid_t nid, unsigned long *when) read_lock_irqsave(glock, flags); peer = kiblnd_find_peer_locked(nid); - if (peer != NULL) { - LASSERT(peer->ibp_connecting > 0 || /* creating conns */ - peer->ibp_accepting > 0 || - !list_empty(&peer->ibp_conns)); /* active conn */ + if (peer) last_alive = peer->ibp_last_alive; - } read_unlock_irqrestore(glock, flags); - if (last_alive != 0) + if (last_alive) *when = last_alive; - /* peer is not persistent in hash, trigger peer creation - * and connection establishment with a NULL tx */ - if (peer == NULL) + /* + * peer is not persistent in hash, trigger peer creation + * and connection establishment with a NULL tx + */ + if (!peer) kiblnd_launch_tx(ni, NULL, nid); CDEBUG(D_NET, "Peer %s %p, alive %ld secs ago\n", @@ -1087,13 +1078,13 @@ void kiblnd_query(lnet_ni_t *ni, lnet_nid_t nid, unsigned long *when) last_alive ? cfs_duration_sec(now - last_alive) : -1); } -void kiblnd_free_pages(kib_pages_t *p) +static void kiblnd_free_pages(kib_pages_t *p) { int npages = p->ibp_npages; int i; for (i = 0; i < npages; i++) { - if (p->ibp_pages[i] != NULL) + if (p->ibp_pages[i]) __free_page(p->ibp_pages[i]); } @@ -1107,7 +1098,7 @@ int kiblnd_alloc_pages(kib_pages_t **pp, int cpt, int npages) LIBCFS_CPT_ALLOC(p, lnet_cpt_table(), cpt, offsetof(kib_pages_t, ibp_pages[npages])); - if (p == NULL) { + if (!p) { CERROR("Can't allocate descriptor for %d pages\n", npages); return -ENOMEM; } @@ -1119,7 +1110,7 @@ int kiblnd_alloc_pages(kib_pages_t **pp, int cpt, int npages) p->ibp_pages[i] = alloc_pages_node( cfs_cpt_spread_node(lnet_cpt_table(), cpt), GFP_NOFS, 0); - if (p->ibp_pages[i] == NULL) { + if (!p->ibp_pages[i]) { CERROR("Can't allocate page %d of %d\n", i, npages); kiblnd_free_pages(p); return -ENOMEM; @@ -1135,10 +1126,10 @@ void kiblnd_unmap_rx_descs(kib_conn_t *conn) kib_rx_t *rx; int i; - LASSERT(conn->ibc_rxs != NULL); - LASSERT(conn->ibc_hdev != NULL); + LASSERT(conn->ibc_rxs); + LASSERT(conn->ibc_hdev); - for (i = 0; i < IBLND_RX_MSGS(conn->ibc_version); i++) { + for (i = 0; i < IBLND_RX_MSGS(conn); i++) { rx = &conn->ibc_rxs[i]; LASSERT(rx->rx_nob >= 0); /* not posted */ @@ -1162,7 +1153,7 @@ void kiblnd_map_rx_descs(kib_conn_t *conn) int ipg; int i; - for (pg_off = ipg = i = 0; i < IBLND_RX_MSGS(conn->ibc_version); i++) { + for (pg_off = ipg = i = 0; i < IBLND_RX_MSGS(conn); i++) { pg = conn->ibc_rx_pages->ibp_pages[ipg]; rx = &conn->ibc_rxs[i]; @@ -1174,7 +1165,7 @@ void kiblnd_map_rx_descs(kib_conn_t *conn) IBLND_MSG_SIZE, DMA_FROM_DEVICE); LASSERT(!kiblnd_dma_mapping_error(conn->ibc_hdev->ibh_ibdev, - rx->rx_msgaddr)); + rx->rx_msgaddr)); KIBLND_UNMAP_ADDR_SET(rx, rx_msgunmap, rx->rx_msgaddr); CDEBUG(D_NET, "rx %d: %p %#llx(%#llx)\n", @@ -1187,7 +1178,7 @@ void kiblnd_map_rx_descs(kib_conn_t *conn) if (pg_off == PAGE_SIZE) { pg_off = 0; ipg++; - LASSERT(ipg <= IBLND_RX_MSG_PAGES(conn->ibc_version)); + LASSERT(ipg <= IBLND_RX_MSG_PAGES(conn)); } } } @@ -1198,9 +1189,9 @@ static void kiblnd_unmap_tx_pool(kib_tx_pool_t *tpo) kib_tx_t *tx; int i; - LASSERT(tpo->tpo_pool.po_allocated == 0); + LASSERT(!tpo->tpo_pool.po_allocated); - if (hdev == NULL) + if (!hdev) return; for (i = 0; i < tpo->tpo_pool.po_size; i++) { @@ -1224,9 +1215,10 @@ static kib_hca_dev_t *kiblnd_current_hdev(kib_dev_t *dev) read_lock_irqsave(&kiblnd_data.kib_global_lock, flags); while (dev->ibd_failover) { read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); - if (i++ % 50 == 0) + if (!(i++ % 50)) CDEBUG(D_NET, "%s: Wait for failover\n", dev->ibd_ifname); + set_current_state(TASK_INTERRUPTIBLE); schedule_timeout(cfs_time_seconds(1) / 100); read_lock_irqsave(&kiblnd_data.kib_global_lock, flags); @@ -1252,7 +1244,7 @@ static void kiblnd_map_tx_pool(kib_tx_pool_t *tpo) int ipage; int i; - LASSERT(net != NULL); + LASSERT(net); dev = net->ibn_dev; @@ -1260,7 +1252,7 @@ static void kiblnd_map_tx_pool(kib_tx_pool_t *tpo) CLASSERT(IBLND_MSG_SIZE <= PAGE_SIZE); /* No fancy arithmetic when we do the buffer calculations */ - CLASSERT(PAGE_SIZE % IBLND_MSG_SIZE == 0); + CLASSERT(!(PAGE_SIZE % IBLND_MSG_SIZE)); tpo->tpo_hdev = kiblnd_current_hdev(dev); @@ -1275,7 +1267,7 @@ static void kiblnd_map_tx_pool(kib_tx_pool_t *tpo) tpo->tpo_hdev->ibh_ibdev, tx->tx_msg, IBLND_MSG_SIZE, DMA_TO_DEVICE); LASSERT(!kiblnd_dma_mapping_error(tpo->tpo_hdev->ibh_ibdev, - tx->tx_msgaddr)); + tx->tx_msgaddr)); KIBLND_UNMAP_ADDR_SET(tx, tx_msgunmap, tx->tx_msgaddr); list_add(&tx->tx_list, &pool->po_free_list); @@ -1291,68 +1283,32 @@ static void kiblnd_map_tx_pool(kib_tx_pool_t *tpo) } } -struct ib_mr *kiblnd_find_dma_mr(kib_hca_dev_t *hdev, __u64 addr, __u64 size) +struct ib_mr *kiblnd_find_rd_dma_mr(kib_hca_dev_t *hdev, kib_rdma_desc_t *rd, + int negotiated_nfrags) { - __u64 index; - - LASSERT(hdev->ibh_mrs[0] != NULL); - - if (hdev->ibh_nmrs == 1) - return hdev->ibh_mrs[0]; - - index = addr >> hdev->ibh_mr_shift; + __u16 nfrags = (negotiated_nfrags != -1) ? + negotiated_nfrags : *kiblnd_tunables.kib_map_on_demand; - if (index < hdev->ibh_nmrs && - index == ((addr + size - 1) >> hdev->ibh_mr_shift)) - return hdev->ibh_mrs[index]; - - return NULL; -} - -struct ib_mr *kiblnd_find_rd_dma_mr(kib_hca_dev_t *hdev, kib_rdma_desc_t *rd) -{ - struct ib_mr *prev_mr; - struct ib_mr *mr; - int i; - - LASSERT(hdev->ibh_mrs[0] != NULL); + LASSERT(hdev->ibh_mrs); if (*kiblnd_tunables.kib_map_on_demand > 0 && - *kiblnd_tunables.kib_map_on_demand <= rd->rd_nfrags) + nfrags <= rd->rd_nfrags) return NULL; - if (hdev->ibh_nmrs == 1) - return hdev->ibh_mrs[0]; - - for (i = 0, mr = prev_mr = NULL; - i < rd->rd_nfrags; i++) { - mr = kiblnd_find_dma_mr(hdev, - rd->rd_frags[i].rf_addr, - rd->rd_frags[i].rf_nob); - if (prev_mr == NULL) - prev_mr = mr; - - if (mr == NULL || prev_mr != mr) { - /* Can't covered by one single MR */ - mr = NULL; - break; - } - } - - return mr; + return hdev->ibh_mrs; } static void kiblnd_destroy_fmr_pool(kib_fmr_pool_t *pool) { - LASSERT(pool->fpo_map_count == 0); + LASSERT(!pool->fpo_map_count); - if (pool->fpo_fmr_pool != NULL) + if (pool->fpo_fmr_pool) ib_destroy_fmr_pool(pool->fpo_fmr_pool); - if (pool->fpo_hdev != NULL) + if (pool->fpo_hdev) kiblnd_hdev_decref(pool->fpo_hdev); - LIBCFS_FREE(pool, sizeof(kib_fmr_pool_t)); + LIBCFS_FREE(pool, sizeof(*pool)); } static void kiblnd_destroy_fmr_pool_list(struct list_head *head) @@ -1387,7 +1343,7 @@ static int kiblnd_create_fmr_pool(kib_fmr_poolset_t *fps, kib_dev_t *dev = fps->fps_net->ibn_dev; kib_fmr_pool_t *fpo; struct ib_fmr_pool_param param = { - .max_pages_per_fmr = LNET_MAX_PAYLOAD/PAGE_SIZE, + .max_pages_per_fmr = LNET_MAX_PAYLOAD / PAGE_SIZE, .page_shift = PAGE_SHIFT, .access = (IB_ACCESS_LOCAL_WRITE | IB_ACCESS_REMOTE_WRITE), @@ -1399,7 +1355,7 @@ static int kiblnd_create_fmr_pool(kib_fmr_poolset_t *fps, int rc; LIBCFS_CPT_ALLOC(fpo, lnet_cpt_table(), fps->fps_cpt, sizeof(*fpo)); - if (fpo == NULL) + if (!fpo) return -ENOMEM; fpo->fpo_hdev = kiblnd_current_hdev(dev); @@ -1410,7 +1366,7 @@ static int kiblnd_create_fmr_pool(kib_fmr_poolset_t *fps, CERROR("Failed to create FMR pool: %d\n", rc); kiblnd_hdev_decref(fpo->fpo_hdev); - LIBCFS_FREE(fpo, sizeof(kib_fmr_pool_t)); + LIBCFS_FREE(fpo, sizeof(*fpo)); return rc; } @@ -1424,7 +1380,7 @@ static int kiblnd_create_fmr_pool(kib_fmr_poolset_t *fps, static void kiblnd_fail_fmr_poolset(kib_fmr_poolset_t *fps, struct list_head *zombies) { - if (fps->fps_net == NULL) /* intialized? */ + if (!fps->fps_net) /* intialized? */ return; spin_lock(&fps->fps_lock); @@ -1434,7 +1390,7 @@ static void kiblnd_fail_fmr_poolset(kib_fmr_poolset_t *fps, kib_fmr_pool_t, fpo_list); fpo->fpo_failed = 1; list_del(&fpo->fpo_list); - if (fpo->fpo_map_count == 0) + if (!fpo->fpo_map_count) list_add(&fpo->fpo_list, zombies); else list_add(&fpo->fpo_list, &fps->fps_failed_pool_list); @@ -1445,7 +1401,7 @@ static void kiblnd_fail_fmr_poolset(kib_fmr_poolset_t *fps, static void kiblnd_fini_fmr_poolset(kib_fmr_poolset_t *fps) { - if (fps->fps_net != NULL) { /* initialized? */ + if (fps->fps_net) { /* initialized? */ kiblnd_destroy_fmr_pool_list(&fps->fps_failed_pool_list); kiblnd_destroy_fmr_pool_list(&fps->fps_pool_list); } @@ -1458,7 +1414,7 @@ static int kiblnd_init_fmr_poolset(kib_fmr_poolset_t *fps, int cpt, kib_fmr_pool_t *fpo; int rc; - memset(fps, 0, sizeof(kib_fmr_poolset_t)); + memset(fps, 0, sizeof(*fps)); fps->fps_net = net; fps->fps_cpt = cpt; @@ -1469,7 +1425,7 @@ static int kiblnd_init_fmr_poolset(kib_fmr_poolset_t *fps, int cpt, INIT_LIST_HEAD(&fps->fps_failed_pool_list); rc = kiblnd_create_fmr_pool(fps, &fpo); - if (rc == 0) + if (!rc) list_add_tail(&fpo->fpo_list, &fps->fps_pool_list); return rc; @@ -1477,7 +1433,7 @@ static int kiblnd_init_fmr_poolset(kib_fmr_poolset_t *fps, int cpt, static int kiblnd_fmr_pool_is_idle(kib_fmr_pool_t *fpo, unsigned long now) { - if (fpo->fpo_map_count != 0) /* still in use */ + if (fpo->fpo_map_count) /* still in use */ return 0; if (fpo->fpo_failed) return 1; @@ -1494,11 +1450,11 @@ void kiblnd_fmr_pool_unmap(kib_fmr_t *fmr, int status) int rc; rc = ib_fmr_pool_unmap(fmr->fmr_pfmr); - LASSERT(rc == 0); + LASSERT(!rc); - if (status != 0) { + if (status) { rc = ib_flush_fmr_pool(fpo->fpo_fmr_pool); - LASSERT(rc == 0); + LASSERT(!rc); } fmr->fmr_pool = NULL; @@ -1563,11 +1519,9 @@ int kiblnd_fmr_pool_map(kib_fmr_poolset_t *fps, __u64 *pages, int npages, if (fps->fps_increasing) { spin_unlock(&fps->fps_lock); - CDEBUG(D_NET, - "Another thread is allocating new FMR pool, waiting for her to complete\n"); + CDEBUG(D_NET, "Another thread is allocating new FMR pool, waiting for her to complete\n"); schedule(); goto again; - } if (time_before(cfs_time_current(), fps->fps_next_retry)) { @@ -1583,7 +1537,7 @@ int kiblnd_fmr_pool_map(kib_fmr_poolset_t *fps, __u64 *pages, int npages, rc = kiblnd_create_fmr_pool(fps, &fpo); spin_lock(&fps->fps_lock); fps->fps_increasing = 0; - if (rc == 0) { + if (!rc) { fps->fps_version++; list_add_tail(&fpo->fpo_list, &fps->fps_pool_list); } else { @@ -1597,7 +1551,7 @@ int kiblnd_fmr_pool_map(kib_fmr_poolset_t *fps, __u64 *pages, int npages, static void kiblnd_fini_pool(kib_pool_t *pool) { LASSERT(list_empty(&pool->po_free_list)); - LASSERT(pool->po_allocated == 0); + LASSERT(!pool->po_allocated); CDEBUG(D_NET, "Finalize %s pool\n", pool->po_owner->ps_name); } @@ -1606,7 +1560,7 @@ static void kiblnd_init_pool(kib_poolset_t *ps, kib_pool_t *pool, int size) { CDEBUG(D_NET, "Initialize %s pool\n", ps->ps_name); - memset(pool, 0, sizeof(kib_pool_t)); + memset(pool, 0, sizeof(*pool)); INIT_LIST_HEAD(&pool->po_free_list); pool->po_deadline = cfs_time_shift(IBLND_POOL_DEADLINE); pool->po_owner = ps; @@ -1621,14 +1575,14 @@ static void kiblnd_destroy_pool_list(struct list_head *head) pool = list_entry(head->next, kib_pool_t, po_list); list_del(&pool->po_list); - LASSERT(pool->po_owner != NULL); + LASSERT(pool->po_owner); pool->po_owner->ps_pool_destroy(pool); } } static void kiblnd_fail_poolset(kib_poolset_t *ps, struct list_head *zombies) { - if (ps->ps_net == NULL) /* intialized? */ + if (!ps->ps_net) /* intialized? */ return; spin_lock(&ps->ps_lock); @@ -1637,7 +1591,7 @@ static void kiblnd_fail_poolset(kib_poolset_t *ps, struct list_head *zombies) kib_pool_t, po_list); po->po_failed = 1; list_del(&po->po_list); - if (po->po_allocated == 0) + if (!po->po_allocated) list_add(&po->po_list, zombies); else list_add(&po->po_list, &ps->ps_failed_pool_list); @@ -1647,7 +1601,7 @@ static void kiblnd_fail_poolset(kib_poolset_t *ps, struct list_head *zombies) static void kiblnd_fini_poolset(kib_poolset_t *ps) { - if (ps->ps_net != NULL) { /* initialized? */ + if (ps->ps_net) { /* initialized? */ kiblnd_destroy_pool_list(&ps->ps_failed_pool_list); kiblnd_destroy_pool_list(&ps->ps_pool_list); } @@ -1663,7 +1617,7 @@ static int kiblnd_init_poolset(kib_poolset_t *ps, int cpt, kib_pool_t *pool; int rc; - memset(ps, 0, sizeof(kib_poolset_t)); + memset(ps, 0, sizeof(*ps)); ps->ps_cpt = cpt; ps->ps_net = net; @@ -1680,7 +1634,7 @@ static int kiblnd_init_poolset(kib_poolset_t *ps, int cpt, INIT_LIST_HEAD(&ps->ps_failed_pool_list); rc = ps->ps_pool_create(ps, size, &pool); - if (rc == 0) + if (!rc) list_add(&pool->po_list, &ps->ps_pool_list); else CERROR("Failed to create the first pool for %s\n", ps->ps_name); @@ -1690,7 +1644,7 @@ static int kiblnd_init_poolset(kib_poolset_t *ps, int cpt, static int kiblnd_pool_is_idle(kib_pool_t *pool, unsigned long now) { - if (pool->po_allocated != 0) /* still in use */ + if (pool->po_allocated) /* still in use */ return 0; if (pool->po_failed) return 1; @@ -1706,7 +1660,7 @@ void kiblnd_pool_free_node(kib_pool_t *pool, struct list_head *node) spin_lock(&ps->ps_lock); - if (ps->ps_node_fini != NULL) + if (ps->ps_node_fini) ps->ps_node_fini(pool, node); LASSERT(pool->po_allocated > 0); @@ -1731,6 +1685,9 @@ struct list_head *kiblnd_pool_alloc_node(kib_poolset_t *ps) { struct list_head *node; kib_pool_t *pool; + unsigned int interval = 1; + unsigned long time_before; + unsigned int trips = 0; int rc; again: @@ -1744,7 +1701,7 @@ struct list_head *kiblnd_pool_alloc_node(kib_poolset_t *ps) node = pool->po_free_list.next; list_del(node); - if (ps->ps_node_init != NULL) { + if (ps->ps_node_init) { /* still hold the lock */ ps->ps_node_init(pool, node); } @@ -1756,9 +1713,15 @@ struct list_head *kiblnd_pool_alloc_node(kib_poolset_t *ps) if (ps->ps_increasing) { /* another thread is allocating a new pool */ spin_unlock(&ps->ps_lock); - CDEBUG(D_NET, "Another thread is allocating new %s pool, waiting for her to complete\n", - ps->ps_name); - schedule(); + trips++; + CDEBUG(D_NET, "Another thread is allocating new %s pool, waiting %d HZs for her to complete. trips = %d\n", + ps->ps_name, interval, trips); + + set_current_state(TASK_INTERRUPTIBLE); + schedule_timeout(interval); + if (interval < cfs_time_seconds(1)) + interval *= 2; + goto again; } @@ -1772,12 +1735,14 @@ struct list_head *kiblnd_pool_alloc_node(kib_poolset_t *ps) spin_unlock(&ps->ps_lock); CDEBUG(D_NET, "%s pool exhausted, allocate new pool\n", ps->ps_name); - + time_before = cfs_time_current(); rc = ps->ps_pool_create(ps, ps->ps_pool_size, &pool); + CDEBUG(D_NET, "ps_pool_create took %lu HZ to complete", + cfs_time_current() - time_before); spin_lock(&ps->ps_lock); ps->ps_increasing = 0; - if (rc == 0) { + if (!rc) { list_add_tail(&pool->po_list, &ps->ps_pool_list); } else { ps->ps_next_retry = cfs_time_shift(IBLND_POOL_RETRY); @@ -1794,37 +1759,37 @@ static void kiblnd_destroy_tx_pool(kib_pool_t *pool) kib_tx_pool_t *tpo = container_of(pool, kib_tx_pool_t, tpo_pool); int i; - LASSERT(pool->po_allocated == 0); + LASSERT(!pool->po_allocated); - if (tpo->tpo_tx_pages != NULL) { + if (tpo->tpo_tx_pages) { kiblnd_unmap_tx_pool(tpo); kiblnd_free_pages(tpo->tpo_tx_pages); } - if (tpo->tpo_tx_descs == NULL) + if (!tpo->tpo_tx_descs) goto out; for (i = 0; i < pool->po_size; i++) { kib_tx_t *tx = &tpo->tpo_tx_descs[i]; list_del(&tx->tx_list); - if (tx->tx_pages != NULL) + if (tx->tx_pages) LIBCFS_FREE(tx->tx_pages, LNET_MAX_IOV * sizeof(*tx->tx_pages)); - if (tx->tx_frags != NULL) + if (tx->tx_frags) LIBCFS_FREE(tx->tx_frags, IBLND_MAX_RDMA_FRAGS * sizeof(*tx->tx_frags)); - if (tx->tx_wrq != NULL) + if (tx->tx_wrq) LIBCFS_FREE(tx->tx_wrq, (1 + IBLND_MAX_RDMA_FRAGS) * sizeof(*tx->tx_wrq)); - if (tx->tx_sge != NULL) + if (tx->tx_sge) LIBCFS_FREE(tx->tx_sge, (1 + IBLND_MAX_RDMA_FRAGS) * sizeof(*tx->tx_sge)); - if (tx->tx_rd != NULL) + if (tx->tx_rd) LIBCFS_FREE(tx->tx_rd, offsetof(kib_rdma_desc_t, rd_frags[IBLND_MAX_RDMA_FRAGS])); @@ -1834,7 +1799,7 @@ static void kiblnd_destroy_tx_pool(kib_pool_t *pool) pool->po_size * sizeof(kib_tx_t)); out: kiblnd_fini_pool(pool); - LIBCFS_FREE(tpo, sizeof(kib_tx_pool_t)); + LIBCFS_FREE(tpo, sizeof(*tpo)); } static int kiblnd_tx_pool_size(int ncpts) @@ -1853,7 +1818,7 @@ static int kiblnd_create_tx_pool(kib_poolset_t *ps, int size, kib_tx_pool_t *tpo; LIBCFS_CPT_ALLOC(tpo, lnet_cpt_table(), ps->ps_cpt, sizeof(*tpo)); - if (tpo == NULL) { + if (!tpo) { CERROR("Failed to allocate TX pool\n"); return -ENOMEM; } @@ -1864,15 +1829,15 @@ static int kiblnd_create_tx_pool(kib_poolset_t *ps, int size, tpo->tpo_tx_pages = NULL; npg = (size * IBLND_MSG_SIZE + PAGE_SIZE - 1) / PAGE_SIZE; - if (kiblnd_alloc_pages(&tpo->tpo_tx_pages, ps->ps_cpt, npg) != 0) { + if (kiblnd_alloc_pages(&tpo->tpo_tx_pages, ps->ps_cpt, npg)) { CERROR("Can't allocate tx pages: %d\n", npg); - LIBCFS_FREE(tpo, sizeof(kib_tx_pool_t)); + LIBCFS_FREE(tpo, sizeof(*tpo)); return -ENOMEM; } LIBCFS_CPT_ALLOC(tpo->tpo_tx_descs, lnet_cpt_table(), ps->ps_cpt, size * sizeof(kib_tx_t)); - if (tpo->tpo_tx_descs == NULL) { + if (!tpo->tpo_tx_descs) { CERROR("Can't allocate %d tx descriptors\n", size); ps->ps_pool_destroy(pool); return -ENOMEM; @@ -1884,17 +1849,17 @@ static int kiblnd_create_tx_pool(kib_poolset_t *ps, int size, kib_tx_t *tx = &tpo->tpo_tx_descs[i]; tx->tx_pool = tpo; - if (ps->ps_net->ibn_fmr_ps != NULL) { + if (ps->ps_net->ibn_fmr_ps) { LIBCFS_CPT_ALLOC(tx->tx_pages, lnet_cpt_table(), ps->ps_cpt, LNET_MAX_IOV * sizeof(*tx->tx_pages)); - if (tx->tx_pages == NULL) + if (!tx->tx_pages) break; } LIBCFS_CPT_ALLOC(tx->tx_frags, lnet_cpt_table(), ps->ps_cpt, IBLND_MAX_RDMA_FRAGS * sizeof(*tx->tx_frags)); - if (tx->tx_frags == NULL) + if (!tx->tx_frags) break; sg_init_table(tx->tx_frags, IBLND_MAX_RDMA_FRAGS); @@ -1902,19 +1867,19 @@ static int kiblnd_create_tx_pool(kib_poolset_t *ps, int size, LIBCFS_CPT_ALLOC(tx->tx_wrq, lnet_cpt_table(), ps->ps_cpt, (1 + IBLND_MAX_RDMA_FRAGS) * sizeof(*tx->tx_wrq)); - if (tx->tx_wrq == NULL) + if (!tx->tx_wrq) break; LIBCFS_CPT_ALLOC(tx->tx_sge, lnet_cpt_table(), ps->ps_cpt, (1 + IBLND_MAX_RDMA_FRAGS) * sizeof(*tx->tx_sge)); - if (tx->tx_sge == NULL) + if (!tx->tx_sge) break; LIBCFS_CPT_ALLOC(tx->tx_rd, lnet_cpt_table(), ps->ps_cpt, offsetof(kib_rdma_desc_t, rd_frags[IBLND_MAX_RDMA_FRAGS])); - if (tx->tx_rd == NULL) + if (!tx->tx_rd) break; } @@ -1945,23 +1910,23 @@ static void kiblnd_net_fini_pools(kib_net_t *net) kib_tx_poolset_t *tps; kib_fmr_poolset_t *fps; - if (net->ibn_tx_ps != NULL) { + if (net->ibn_tx_ps) { tps = net->ibn_tx_ps[i]; kiblnd_fini_poolset(&tps->tps_poolset); } - if (net->ibn_fmr_ps != NULL) { + if (net->ibn_fmr_ps) { fps = net->ibn_fmr_ps[i]; kiblnd_fini_fmr_poolset(fps); } } - if (net->ibn_tx_ps != NULL) { + if (net->ibn_tx_ps) { cfs_percpt_free(net->ibn_tx_ps); net->ibn_tx_ps = NULL; } - if (net->ibn_fmr_ps != NULL) { + if (net->ibn_fmr_ps) { cfs_percpt_free(net->ibn_fmr_ps); net->ibn_fmr_ps = NULL; } @@ -1975,8 +1940,7 @@ static int kiblnd_net_init_pools(kib_net_t *net, __u32 *cpts, int ncpts) int i; read_lock_irqsave(&kiblnd_data.kib_global_lock, flags); - if (*kiblnd_tunables.kib_map_on_demand == 0 && - net->ibn_dev->ibd_hdev->ibh_nmrs == 1) { + if (!*kiblnd_tunables.kib_map_on_demand) { read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); goto create_tx_pool; } @@ -1996,7 +1960,7 @@ static int kiblnd_net_init_pools(kib_net_t *net, __u32 *cpts, int ncpts) * TX pool must be created later than FMR, see LU-2268 * for details */ - LASSERT(net->ibn_tx_ps == NULL); + LASSERT(!net->ibn_tx_ps); /* * premapping can fail if ibd_nmr > 1, so we always create @@ -2005,56 +1969,45 @@ static int kiblnd_net_init_pools(kib_net_t *net, __u32 *cpts, int ncpts) net->ibn_fmr_ps = cfs_percpt_alloc(lnet_cpt_table(), sizeof(kib_fmr_poolset_t)); - if (net->ibn_fmr_ps == NULL) { + if (!net->ibn_fmr_ps) { CERROR("Failed to allocate FMR pool array\n"); rc = -ENOMEM; goto failed; } for (i = 0; i < ncpts; i++) { - cpt = (cpts == NULL) ? i : cpts[i]; + cpt = !cpts ? i : cpts[i]; rc = kiblnd_init_fmr_poolset(net->ibn_fmr_ps[cpt], cpt, net, kiblnd_fmr_pool_size(ncpts), kiblnd_fmr_flush_trigger(ncpts)); - if (rc == -ENOSYS && i == 0) /* no FMR */ - break; - - if (rc != 0) { /* a real error */ + if (rc) { CERROR("Can't initialize FMR pool for CPT %d: %d\n", cpt, rc); goto failed; } } - if (i > 0) { + if (i > 0) LASSERT(i == ncpts); - goto create_tx_pool; - } - - cfs_percpt_free(net->ibn_fmr_ps); - net->ibn_fmr_ps = NULL; - - CWARN("Device does not support FMR\n"); - goto failed; create_tx_pool: net->ibn_tx_ps = cfs_percpt_alloc(lnet_cpt_table(), sizeof(kib_tx_poolset_t)); - if (net->ibn_tx_ps == NULL) { + if (!net->ibn_tx_ps) { CERROR("Failed to allocate tx pool array\n"); rc = -ENOMEM; goto failed; } for (i = 0; i < ncpts; i++) { - cpt = (cpts == NULL) ? i : cpts[i]; + cpt = !cpts ? i : cpts[i]; rc = kiblnd_init_poolset(&net->ibn_tx_ps[cpt]->tps_poolset, cpt, net, "TX", kiblnd_tx_pool_size(ncpts), kiblnd_create_tx_pool, kiblnd_destroy_tx_pool, kiblnd_tx_init, NULL); - if (rc != 0) { + if (rc) { CERROR("Can't initialize TX pool for CPT %d: %d\n", cpt, rc); goto failed; @@ -2064,14 +2017,16 @@ static int kiblnd_net_init_pools(kib_net_t *net, __u32 *cpts, int ncpts) return 0; failed: kiblnd_net_fini_pools(net); - LASSERT(rc != 0); + LASSERT(rc); return rc; } static int kiblnd_hdev_get_attr(kib_hca_dev_t *hdev) { - /* It's safe to assume a HCA can handle a page size - * matching that of the native system */ + /* + * It's safe to assume a HCA can handle a page size + * matching that of the native system + */ hdev->ibh_page_shift = PAGE_SHIFT; hdev->ibh_page_size = 1 << PAGE_SHIFT; hdev->ibh_page_mask = ~((__u64)hdev->ibh_page_size - 1); @@ -2082,44 +2037,28 @@ static int kiblnd_hdev_get_attr(kib_hca_dev_t *hdev) return 0; } - for (hdev->ibh_mr_shift = 0; - hdev->ibh_mr_shift < 64; hdev->ibh_mr_shift++) { - if (hdev->ibh_mr_size == (1ULL << hdev->ibh_mr_shift) || - hdev->ibh_mr_size == (1ULL << hdev->ibh_mr_shift) - 1) - return 0; - } - CERROR("Invalid mr size: %#llx\n", hdev->ibh_mr_size); return -EINVAL; } static void kiblnd_hdev_cleanup_mrs(kib_hca_dev_t *hdev) { - int i; - - if (hdev->ibh_nmrs == 0 || hdev->ibh_mrs == NULL) + if (!hdev->ibh_mrs) return; - for (i = 0; i < hdev->ibh_nmrs; i++) { - if (hdev->ibh_mrs[i] == NULL) - break; + ib_dereg_mr(hdev->ibh_mrs); - ib_dereg_mr(hdev->ibh_mrs[i]); - } - - LIBCFS_FREE(hdev->ibh_mrs, sizeof(*hdev->ibh_mrs) * hdev->ibh_nmrs); - hdev->ibh_mrs = NULL; - hdev->ibh_nmrs = 0; + hdev->ibh_mrs = NULL; } void kiblnd_hdev_destroy(kib_hca_dev_t *hdev) { kiblnd_hdev_cleanup_mrs(hdev); - if (hdev->ibh_pd != NULL) + if (hdev->ibh_pd) ib_dealloc_pd(hdev->ibh_pd); - if (hdev->ibh_cmid != NULL) + if (hdev->ibh_cmid) rdma_destroy_id(hdev->ibh_cmid); LIBCFS_FREE(hdev, sizeof(*hdev)); @@ -2132,18 +2071,9 @@ static int kiblnd_hdev_setup_mrs(kib_hca_dev_t *hdev) int acflags = IB_ACCESS_LOCAL_WRITE | IB_ACCESS_REMOTE_WRITE; rc = kiblnd_hdev_get_attr(hdev); - if (rc != 0) + if (rc) return rc; - LIBCFS_ALLOC(hdev->ibh_mrs, 1 * sizeof(*hdev->ibh_mrs)); - if (hdev->ibh_mrs == NULL) { - CERROR("Failed to allocate MRs table\n"); - return -ENOMEM; - } - - hdev->ibh_mrs[0] = NULL; - hdev->ibh_nmrs = 1; - mr = ib_get_dma_mr(hdev->ibh_pd, acflags); if (IS_ERR(mr)) { CERROR("Failed ib_get_dma_mr : %ld\n", PTR_ERR(mr)); @@ -2151,7 +2081,7 @@ static int kiblnd_hdev_setup_mrs(kib_hca_dev_t *hdev) return PTR_ERR(mr); } - hdev->ibh_mrs[0] = mr; + hdev->ibh_mrs = mr; return 0; } @@ -2170,12 +2100,13 @@ static int kiblnd_dev_need_failover(kib_dev_t *dev) struct sockaddr_in dstaddr; int rc; - if (dev->ibd_hdev == NULL || /* initializing */ - dev->ibd_hdev->ibh_cmid == NULL || /* listener is dead */ + if (!dev->ibd_hdev || /* initializing */ + !dev->ibd_hdev->ibh_cmid || /* listener is dead */ *kiblnd_tunables.kib_dev_failover > 1) /* debugging */ return 1; - /* XXX: it's UGLY, but I don't have better way to find + /* + * XXX: it's UGLY, but I don't have better way to find * ib-bonding HCA failover because: * * a. no reliable CM event for HCA failover... @@ -2184,7 +2115,8 @@ static int kiblnd_dev_need_failover(kib_dev_t *dev) * We have only two choices at this point: * * a. rdma_bind_addr(), it will conflict with listener cmid - * b. rdma_resolve_addr() to zero addr */ + * b. rdma_resolve_addr() to zero addr + */ cmid = kiblnd_rdma_create_id(kiblnd_dummy_callback, dev, RDMA_PS_TCP, IB_QPT_RC); if (IS_ERR(cmid)) { @@ -2201,7 +2133,7 @@ static int kiblnd_dev_need_failover(kib_dev_t *dev) dstaddr.sin_family = AF_INET; rc = rdma_resolve_addr(cmid, (struct sockaddr *)&srcaddr, (struct sockaddr *)&dstaddr, 1); - if (rc != 0 || cmid->device == NULL) { + if (rc || !cmid->device) { CERROR("Failed to bind %s:%pI4h to device(%p): %d\n", dev->ibd_ifname, &dev->ibd_ifip, cmid->device, rc); @@ -2230,24 +2162,27 @@ int kiblnd_dev_failover(kib_dev_t *dev) int i; LASSERT(*kiblnd_tunables.kib_dev_failover > 1 || - dev->ibd_can_failover || - dev->ibd_hdev == NULL); + dev->ibd_can_failover || !dev->ibd_hdev); rc = kiblnd_dev_need_failover(dev); if (rc <= 0) goto out; - if (dev->ibd_hdev != NULL && - dev->ibd_hdev->ibh_cmid != NULL) { - /* XXX it's not good to close old listener at here, + if (dev->ibd_hdev && + dev->ibd_hdev->ibh_cmid) { + /* + * XXX it's not good to close old listener at here, * because we can fail to create new listener. * But we have to close it now, otherwise rdma_bind_addr - * will return EADDRINUSE... How crap! */ + * will return EADDRINUSE... How crap! + */ write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); cmid = dev->ibd_hdev->ibh_cmid; - /* make next schedule of kiblnd_dev_need_failover() - * return 1 for me */ + /* + * make next schedule of kiblnd_dev_need_failover() + * return 1 for me + */ dev->ibd_hdev->ibh_cmid = NULL; write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); @@ -2269,7 +2204,7 @@ int kiblnd_dev_failover(kib_dev_t *dev) /* Bind to failover device or port */ rc = rdma_bind_addr(cmid, (struct sockaddr *)&addr); - if (rc != 0 || cmid->device == NULL) { + if (rc || !cmid->device) { CERROR("Failed to bind %s:%pI4h to device(%p): %d\n", dev->ibd_ifname, &dev->ibd_ifip, cmid->device, rc); @@ -2278,7 +2213,7 @@ int kiblnd_dev_failover(kib_dev_t *dev) } LIBCFS_ALLOC(hdev, sizeof(*hdev)); - if (hdev == NULL) { + if (!hdev) { CERROR("Failed to allocate kib_hca_dev\n"); rdma_destroy_id(cmid); rc = -ENOMEM; @@ -2300,13 +2235,13 @@ int kiblnd_dev_failover(kib_dev_t *dev) hdev->ibh_pd = pd; rc = rdma_listen(cmid, 0); - if (rc != 0) { + if (rc) { CERROR("Can't start new listener: %d\n", rc); goto out; } rc = kiblnd_hdev_setup_mrs(hdev); - if (rc != 0) { + if (rc) { CERROR("Can't setup device: %d\n", rc); goto out; } @@ -2334,10 +2269,10 @@ int kiblnd_dev_failover(kib_dev_t *dev) kiblnd_destroy_pool_list(&zombie_ppo); if (!list_empty(&zombie_fpo)) kiblnd_destroy_fmr_pool_list(&zombie_fpo); - if (hdev != NULL) + if (hdev) kiblnd_hdev_decref(hdev); - if (rc != 0) + if (rc) dev->ibd_failed_failover++; else dev->ibd_failed_failover = 0; @@ -2347,13 +2282,13 @@ int kiblnd_dev_failover(kib_dev_t *dev) void kiblnd_destroy_dev(kib_dev_t *dev) { - LASSERT(dev->ibd_nnets == 0); + LASSERT(!dev->ibd_nnets); LASSERT(list_empty(&dev->ibd_nets)); list_del(&dev->ibd_fail_list); list_del(&dev->ibd_list); - if (dev->ibd_hdev != NULL) + if (dev->ibd_hdev) kiblnd_hdev_decref(dev->ibd_hdev); LIBCFS_FREE(dev, sizeof(*dev)); @@ -2369,7 +2304,7 @@ static kib_dev_t *kiblnd_create_dev(char *ifname) int rc; rc = lnet_ipif_query(ifname, &up, &ip, &netmask); - if (rc != 0) { + if (rc) { CERROR("Can't query IPoIB interface %s: %d\n", ifname, rc); return NULL; @@ -2381,11 +2316,11 @@ static kib_dev_t *kiblnd_create_dev(char *ifname) } LIBCFS_ALLOC(dev, sizeof(*dev)); - if (dev == NULL) + if (!dev) return NULL; netdev = dev_get_by_name(&init_net, ifname); - if (netdev == NULL) { + if (!netdev) { dev->ibd_can_failover = 0; } else { dev->ibd_can_failover = !!(netdev->flags & IFF_MASTER); @@ -2400,14 +2335,13 @@ static kib_dev_t *kiblnd_create_dev(char *ifname) /* initialize the device */ rc = kiblnd_dev_failover(dev); - if (rc != 0) { + if (rc) { CERROR("Can't initialize device: %d\n", rc); LIBCFS_FREE(dev, sizeof(*dev)); return NULL; } - list_add_tail(&dev->ibd_list, - &kiblnd_data.kib_devs); + list_add_tail(&dev->ibd_list, &kiblnd_data.kib_devs); return dev; } @@ -2424,18 +2358,22 @@ static void kiblnd_base_shutdown(void) case IBLND_INIT_ALL: case IBLND_INIT_DATA: - LASSERT(kiblnd_data.kib_peers != NULL); + LASSERT(kiblnd_data.kib_peers); for (i = 0; i < kiblnd_data.kib_peer_hash_size; i++) LASSERT(list_empty(&kiblnd_data.kib_peers[i])); LASSERT(list_empty(&kiblnd_data.kib_connd_zombies)); LASSERT(list_empty(&kiblnd_data.kib_connd_conns)); + LASSERT(list_empty(&kiblnd_data.kib_reconn_list)); + LASSERT(list_empty(&kiblnd_data.kib_reconn_wait)); /* flag threads to terminate; wake and wait for them to die */ kiblnd_data.kib_shutdown = 1; - /* NB: we really want to stop scheduler threads net by net + /* + * NB: we really want to stop scheduler threads net by net * instead of the whole module, this should be improved - * with dynamic configuration LNet */ + * with dynamic configuration LNet + */ cfs_percpt_for_each(sched, i, kiblnd_data.kib_scheds) wake_up_all(&sched->ibs_waitq); @@ -2443,7 +2381,7 @@ static void kiblnd_base_shutdown(void) wake_up_all(&kiblnd_data.kib_failover_waitq); i = 2; - while (atomic_read(&kiblnd_data.kib_nthreads) != 0) { + while (atomic_read(&kiblnd_data.kib_nthreads)) { i++; /* power of 2 ? */ CDEBUG(((i & (-i)) == i) ? D_WARNING : D_NET, @@ -2459,20 +2397,20 @@ static void kiblnd_base_shutdown(void) break; } - if (kiblnd_data.kib_peers != NULL) { + if (kiblnd_data.kib_peers) { LIBCFS_FREE(kiblnd_data.kib_peers, sizeof(struct list_head) * kiblnd_data.kib_peer_hash_size); } - if (kiblnd_data.kib_scheds != NULL) + if (kiblnd_data.kib_scheds) cfs_percpt_free(kiblnd_data.kib_scheds); kiblnd_data.kib_init = IBLND_INIT_NOTHING; module_put(THIS_MODULE); } -void kiblnd_shutdown(lnet_ni_t *ni) +static void kiblnd_shutdown(lnet_ni_t *ni) { kib_net_t *net = ni->ni_data; rwlock_t *g_lock = &kiblnd_data.kib_global_lock; @@ -2481,7 +2419,7 @@ void kiblnd_shutdown(lnet_ni_t *ni) LASSERT(kiblnd_data.kib_init == IBLND_INIT_ALL); - if (net == NULL) + if (!net) goto out; write_lock_irqsave(g_lock, flags); @@ -2498,7 +2436,7 @@ void kiblnd_shutdown(lnet_ni_t *ni) /* Wait for all peer state to clean up */ i = 2; - while (atomic_read(&net->ibn_npeers) != 0) { + while (atomic_read(&net->ibn_npeers)) { i++; CDEBUG(((i & (-i)) == i) ? D_WARNING : D_NET, /* 2**n? */ "%s: waiting for %d peers to disconnect\n", @@ -2519,10 +2457,9 @@ void kiblnd_shutdown(lnet_ni_t *ni) /* fall through */ case IBLND_INIT_NOTHING: - LASSERT(atomic_read(&net->ibn_nconns) == 0); + LASSERT(!atomic_read(&net->ibn_nconns)); - if (net->ibn_dev != NULL && - net->ibn_dev->ibd_nnets == 0) + if (net->ibn_dev && !net->ibn_dev->ibd_nnets) kiblnd_destroy_dev(net->ibn_dev); break; @@ -2558,7 +2495,7 @@ static int kiblnd_base_startup(void) kiblnd_data.kib_peer_hash_size = IBLND_PEER_HASH_SIZE; LIBCFS_ALLOC(kiblnd_data.kib_peers, sizeof(struct list_head) * kiblnd_data.kib_peer_hash_size); - if (kiblnd_data.kib_peers == NULL) + if (!kiblnd_data.kib_peers) goto failed; for (i = 0; i < kiblnd_data.kib_peer_hash_size; i++) INIT_LIST_HEAD(&kiblnd_data.kib_peers[i]); @@ -2566,12 +2503,15 @@ static int kiblnd_base_startup(void) spin_lock_init(&kiblnd_data.kib_connd_lock); INIT_LIST_HEAD(&kiblnd_data.kib_connd_conns); INIT_LIST_HEAD(&kiblnd_data.kib_connd_zombies); + INIT_LIST_HEAD(&kiblnd_data.kib_reconn_list); + INIT_LIST_HEAD(&kiblnd_data.kib_reconn_wait); + init_waitqueue_head(&kiblnd_data.kib_connd_waitq); init_waitqueue_head(&kiblnd_data.kib_failover_waitq); kiblnd_data.kib_scheds = cfs_percpt_alloc(lnet_cpt_table(), sizeof(*sched)); - if (kiblnd_data.kib_scheds == NULL) + if (!kiblnd_data.kib_scheds) goto failed; cfs_percpt_for_each(sched, i, kiblnd_data.kib_scheds) { @@ -2585,8 +2525,10 @@ static int kiblnd_base_startup(void) if (*kiblnd_tunables.kib_nscheds > 0) { nthrs = min(nthrs, *kiblnd_tunables.kib_nscheds); } else { - /* max to half of CPUs, another half is reserved for - * upper layer modules */ + /* + * max to half of CPUs, another half is reserved for + * upper layer modules + */ nthrs = min(max(IBLND_N_SCHED, nthrs >> 1), nthrs); } @@ -2601,16 +2543,16 @@ static int kiblnd_base_startup(void) /*****************************************************/ rc = kiblnd_thread_start(kiblnd_connd, NULL, "kiblnd_connd"); - if (rc != 0) { + if (rc) { CERROR("Can't spawn o2iblnd connd: %d\n", rc); goto failed; } - if (*kiblnd_tunables.kib_dev_failover != 0) + if (*kiblnd_tunables.kib_dev_failover) rc = kiblnd_thread_start(kiblnd_failover_thread, NULL, "kiblnd_failover"); - if (rc != 0) { + if (rc) { CERROR("Can't spawn o2iblnd failover thread: %d\n", rc); goto failed; } @@ -2632,7 +2574,7 @@ static int kiblnd_start_schedulers(struct kib_sched_info *sched) int nthrs; int i; - if (sched->ibs_nthreads == 0) { + if (!sched->ibs_nthreads) { if (*kiblnd_tunables.kib_nscheds > 0) { nthrs = sched->ibs_nthreads_max; } else { @@ -2655,7 +2597,7 @@ static int kiblnd_start_schedulers(struct kib_sched_info *sched) snprintf(name, sizeof(name), "kiblnd_sd_%02ld_%02ld", KIB_THREAD_CPT(id), KIB_THREAD_TID(id)); rc = kiblnd_thread_start(kiblnd_scheduler, (void *)id, name); - if (rc == 0) + if (!rc) continue; CERROR("Can't spawn thread %d for scheduler[%d]: %d\n", @@ -2677,14 +2619,14 @@ static int kiblnd_dev_start_threads(kib_dev_t *dev, int newdev, __u32 *cpts, for (i = 0; i < ncpts; i++) { struct kib_sched_info *sched; - cpt = (cpts == NULL) ? i : cpts[i]; + cpt = !cpts ? i : cpts[i]; sched = kiblnd_data.kib_scheds[cpt]; if (!newdev && sched->ibs_nthreads > 0) continue; rc = kiblnd_start_schedulers(kiblnd_data.kib_scheds[cpt]); - if (rc != 0) { + if (rc) { CERROR("Failed to start scheduler threads for %s\n", dev->ibd_ifname); return rc; @@ -2702,30 +2644,30 @@ static kib_dev_t *kiblnd_dev_search(char *ifname) colon = strchr(ifname, ':'); list_for_each_entry(dev, &kiblnd_data.kib_devs, ibd_list) { - if (strcmp(&dev->ibd_ifname[0], ifname) == 0) + if (!strcmp(&dev->ibd_ifname[0], ifname)) return dev; - if (alias != NULL) + if (alias) continue; colon2 = strchr(dev->ibd_ifname, ':'); - if (colon != NULL) + if (colon) *colon = 0; - if (colon2 != NULL) + if (colon2) *colon2 = 0; - if (strcmp(&dev->ibd_ifname[0], ifname) == 0) + if (!strcmp(&dev->ibd_ifname[0], ifname)) alias = dev; - if (colon != NULL) + if (colon) *colon = ':'; - if (colon2 != NULL) + if (colon2) *colon2 = ':'; } return alias; } -int kiblnd_startup(lnet_ni_t *ni) +static int kiblnd_startup(lnet_ni_t *ni) { char *ifname; kib_dev_t *ibdev = NULL; @@ -2739,13 +2681,13 @@ int kiblnd_startup(lnet_ni_t *ni) if (kiblnd_data.kib_init == IBLND_INIT_NOTHING) { rc = kiblnd_base_startup(); - if (rc != 0) + if (rc) return rc; } LIBCFS_ALLOC(net, sizeof(*net)); ni->ni_data = net; - if (net == NULL) + if (!net) goto net_failed; ktime_get_real_ts64(&tv); @@ -2757,11 +2699,11 @@ int kiblnd_startup(lnet_ni_t *ni) ni->ni_peertxcredits = *kiblnd_tunables.kib_peertxcredits; ni->ni_peerrtrcredits = *kiblnd_tunables.kib_peerrtrcredits; - if (ni->ni_interfaces[0] != NULL) { + if (ni->ni_interfaces[0]) { /* Use the IPoIB interface specified in 'networks=' */ CLASSERT(LNET_MAX_INTERFACES > 1); - if (ni->ni_interfaces[1] != NULL) { + if (ni->ni_interfaces[1]) { CERROR("Multiple interfaces not supported\n"); goto failed; } @@ -2778,12 +2720,12 @@ int kiblnd_startup(lnet_ni_t *ni) ibdev = kiblnd_dev_search(ifname); - newdev = ibdev == NULL; + newdev = !ibdev; /* hmm...create kib_dev even for alias */ - if (ibdev == NULL || strcmp(&ibdev->ibd_ifname[0], ifname) != 0) + if (!ibdev || strcmp(&ibdev->ibd_ifname[0], ifname)) ibdev = kiblnd_create_dev(ifname); - if (ibdev == NULL) + if (!ibdev) goto failed; net->ibn_dev = ibdev; @@ -2791,11 +2733,11 @@ int kiblnd_startup(lnet_ni_t *ni) rc = kiblnd_dev_start_threads(ibdev, newdev, ni->ni_cpts, ni->ni_ncpts); - if (rc != 0) + if (rc) goto failed; rc = kiblnd_net_init_pools(net, ni->ni_cpts, ni->ni_ncpts); - if (rc != 0) { + if (rc) { CERROR("Failed to initialize NI pools: %d\n", rc); goto failed; } @@ -2810,7 +2752,7 @@ int kiblnd_startup(lnet_ni_t *ni) return 0; failed: - if (net->ibn_dev == NULL && ibdev != NULL) + if (!net->ibn_dev && ibdev) kiblnd_destroy_dev(ibdev); net_failed: @@ -2820,25 +2762,35 @@ net_failed: return -ENETDOWN; } -static void __exit kiblnd_module_fini(void) +static lnd_t the_o2iblnd = { + .lnd_type = O2IBLND, + .lnd_startup = kiblnd_startup, + .lnd_shutdown = kiblnd_shutdown, + .lnd_ctl = kiblnd_ctl, + .lnd_query = kiblnd_query, + .lnd_send = kiblnd_send, + .lnd_recv = kiblnd_recv, +}; + +static void __exit ko2iblnd_exit(void) { lnet_unregister_lnd(&the_o2iblnd); } -static int __init kiblnd_module_init(void) +static int __init ko2iblnd_init(void) { int rc; CLASSERT(sizeof(kib_msg_t) <= IBLND_MSG_SIZE); CLASSERT(offsetof(kib_msg_t, - ibm_u.get.ibgm_rd.rd_frags[IBLND_MAX_RDMA_FRAGS]) - <= IBLND_MSG_SIZE); + ibm_u.get.ibgm_rd.rd_frags[IBLND_MAX_RDMA_FRAGS]) + <= IBLND_MSG_SIZE); CLASSERT(offsetof(kib_msg_t, - ibm_u.putack.ibpam_rd.rd_frags[IBLND_MAX_RDMA_FRAGS]) - <= IBLND_MSG_SIZE); + ibm_u.putack.ibpam_rd.rd_frags[IBLND_MAX_RDMA_FRAGS]) + <= IBLND_MSG_SIZE); rc = kiblnd_tunables_init(); - if (rc != 0) + if (rc) return rc; lnet_register_lnd(&the_o2iblnd); @@ -2847,8 +2799,9 @@ static int __init kiblnd_module_init(void) } MODULE_AUTHOR("OpenSFS, Inc. <http://www.lustre.org/>"); -MODULE_DESCRIPTION("Kernel OpenIB gen2 LND v2.00"); +MODULE_DESCRIPTION("OpenIB gen2 LNet Network Driver"); +MODULE_VERSION("2.7.0"); MODULE_LICENSE("GPL"); -module_init(kiblnd_module_init); -module_exit(kiblnd_module_fini); +module_init(ko2iblnd_init); +module_exit(ko2iblnd_exit); diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h index 025faa9f8..bfcbdd167 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h @@ -60,17 +60,17 @@ #include <net/sock.h> #include <linux/in.h> +#include <rdma/rdma_cm.h> +#include <rdma/ib_cm.h> +#include <rdma/ib_verbs.h> +#include <rdma/ib_fmr_pool.h> + #define DEBUG_SUBSYSTEM S_LND #include "../../../include/linux/libcfs/libcfs.h" #include "../../../include/linux/lnet/lnet.h" #include "../../../include/linux/lnet/lib-lnet.h" -#include <rdma/rdma_cm.h> -#include <rdma/ib_cm.h> -#include <rdma/ib_verbs.h> -#include <rdma/ib_fmr_pool.h> - #define IBLND_PEER_HASH_SIZE 101 /* # peer lists */ /* # scheduler loops before reschedule */ #define IBLND_RESCHED 100 @@ -146,9 +146,9 @@ kiblnd_concurrent_sends_v1(void) #define IBLND_OOB_CAPABLE(v) ((v) != IBLND_MSG_VERSION_1) #define IBLND_OOB_MSGS(v) (IBLND_OOB_CAPABLE(v) ? 2 : 0) -#define IBLND_MSG_SIZE (4<<10) /* max size of queued messages (inc hdr) */ +#define IBLND_MSG_SIZE (4 << 10) /* max size of queued messages (inc hdr) */ #define IBLND_MAX_RDMA_FRAGS LNET_MAX_IOV /* max # of fragments supported */ -#define IBLND_CFG_RDMA_FRAGS (*kiblnd_tunables.kib_map_on_demand != 0 ? \ +#define IBLND_CFG_RDMA_FRAGS (*kiblnd_tunables.kib_map_on_demand ? \ *kiblnd_tunables.kib_map_on_demand : \ IBLND_MAX_RDMA_FRAGS) /* max # of fragments configured by user */ #define IBLND_RDMA_FRAGS(v) ((v) == IBLND_MSG_VERSION_1 ? \ @@ -162,18 +162,17 @@ kiblnd_concurrent_sends_v1(void) #define IBLND_FMR_POOL 256 #define IBLND_FMR_POOL_FLUSH 192 -/* TX messages (shared by all connections) */ -#define IBLND_TX_MSGS() (*kiblnd_tunables.kib_ntx) - -/* RX messages (per connection) */ -#define IBLND_RX_MSGS(v) (IBLND_MSG_QUEUE_SIZE(v) * 2 + IBLND_OOB_MSGS(v)) -#define IBLND_RX_MSG_BYTES(v) (IBLND_RX_MSGS(v) * IBLND_MSG_SIZE) -#define IBLND_RX_MSG_PAGES(v) ((IBLND_RX_MSG_BYTES(v) + PAGE_SIZE - 1) / PAGE_SIZE) +#define IBLND_RX_MSGS(c) \ + ((c->ibc_queue_depth) * 2 + IBLND_OOB_MSGS(c->ibc_version)) +#define IBLND_RX_MSG_BYTES(c) (IBLND_RX_MSGS(c) * IBLND_MSG_SIZE) +#define IBLND_RX_MSG_PAGES(c) \ + ((IBLND_RX_MSG_BYTES(c) + PAGE_SIZE - 1) / PAGE_SIZE) /* WRs and CQEs (per connection) */ -#define IBLND_RECV_WRS(v) IBLND_RX_MSGS(v) -#define IBLND_SEND_WRS(v) ((IBLND_RDMA_FRAGS(v) + 1) * IBLND_CONCURRENT_SENDS(v)) -#define IBLND_CQ_ENTRIES(v) (IBLND_RECV_WRS(v) + IBLND_SEND_WRS(v)) +#define IBLND_RECV_WRS(c) IBLND_RX_MSGS(c) +#define IBLND_SEND_WRS(c) \ + ((c->ibc_max_frags + 1) * IBLND_CONCURRENT_SENDS(c->ibc_version)) +#define IBLND_CQ_ENTRIES(c) (IBLND_RECV_WRS(c) + IBLND_SEND_WRS(c)) struct kib_hca_dev; @@ -209,8 +208,7 @@ typedef struct kib_hca_dev { __u64 ibh_page_mask; /* page mask of current HCA */ int ibh_mr_shift; /* bits shift of max MR size */ __u64 ibh_mr_size; /* size of MR */ - int ibh_nmrs; /* # of global MRs */ - struct ib_mr **ibh_mrs; /* global MR */ + struct ib_mr *ibh_mrs; /* global MR */ struct ib_pd *ibh_pd; /* PD */ kib_dev_t *ibh_dev; /* owner */ atomic_t ibh_ref; /* refcount */ @@ -350,6 +348,16 @@ typedef struct { void *kib_connd; /* the connd task (serialisation assertions) */ struct list_head kib_connd_conns; /* connections to setup/teardown */ struct list_head kib_connd_zombies; /* connections with zero refcount */ + /* connections to reconnect */ + struct list_head kib_reconn_list; + /* peers wait for reconnection */ + struct list_head kib_reconn_wait; + /** + * The second that peers are pulled out from \a kib_reconn_wait + * for reconnection. + */ + time64_t kib_reconn_sec; + wait_queue_head_t kib_connd_waitq; /* connection daemon sleeps here */ spinlock_t kib_connd_lock; /* serialise */ struct ib_qp_attr kib_error_qpa; /* QP->ERROR */ @@ -465,10 +473,10 @@ typedef struct { #define IBLND_REJECT_FATAL 3 /* Anything else */ #define IBLND_REJECT_CONN_UNCOMPAT 4 /* incompatible version peer */ #define IBLND_REJECT_CONN_STALE 5 /* stale peer */ -#define IBLND_REJECT_RDMA_FRAGS 6 /* Fatal: peer's rdma frags can't match */ - /* mine */ -#define IBLND_REJECT_MSG_QUEUE_SIZE 7 /* Fatal: peer's msg queue size can't */ - /* match mine */ +/* peer's rdma frags doesn't match mine */ +#define IBLND_REJECT_RDMA_FRAGS 6 +/* peer's msg queue size doesn't match mine */ +#define IBLND_REJECT_MSG_QUEUE_SIZE 7 /***********************************************************************/ @@ -527,6 +535,8 @@ typedef struct kib_conn { struct list_head ibc_list; /* stash on peer's conn list */ struct list_head ibc_sched_list; /* schedule for attention */ __u16 ibc_version; /* version of connection */ + /* reconnect later */ + __u16 ibc_reconnect:1; __u64 ibc_incarnation; /* which instance of the peer */ atomic_t ibc_refcount; /* # users */ int ibc_state; /* what's happening */ @@ -536,6 +546,10 @@ typedef struct kib_conn { int ibc_outstanding_credits; /* # credits to return */ int ibc_reserved_credits; /* # ACK/DONE msg credits */ int ibc_comms_error; /* set on comms error */ + /* connections queue depth */ + __u16 ibc_queue_depth; + /* connections max frags */ + __u16 ibc_max_frags; unsigned int ibc_nrx:16; /* receive buffers owned */ unsigned int ibc_scheduled:1; /* scheduled for attention */ unsigned int ibc_ready:1; /* CQ callback fired */ @@ -572,18 +586,29 @@ typedef struct kib_peer { struct list_head ibp_list; /* stash on global peer list */ lnet_nid_t ibp_nid; /* who's on the other end(s) */ lnet_ni_t *ibp_ni; /* LNet interface */ - atomic_t ibp_refcount; /* # users */ struct list_head ibp_conns; /* all active connections */ struct list_head ibp_tx_queue; /* msgs waiting for a conn */ - __u16 ibp_version; /* version of peer */ __u64 ibp_incarnation; /* incarnation of peer */ - int ibp_connecting; /* current active connection attempts - */ - int ibp_accepting; /* current passive connection attempts - */ - int ibp_error; /* errno on closing this peer */ - unsigned long ibp_last_alive; /* when (in jiffies) I was last alive - */ + /* when (in jiffies) I was last alive */ + unsigned long ibp_last_alive; + /* # users */ + atomic_t ibp_refcount; + /* version of peer */ + __u16 ibp_version; + /* current passive connection attempts */ + unsigned short ibp_accepting; + /* current active connection attempts */ + unsigned short ibp_connecting; + /* reconnect this peer later */ + unsigned short ibp_reconnecting:1; + /* # consecutive reconnection attempts to this peer */ + unsigned int ibp_reconnected; + /* errno on closing this peer */ + int ibp_error; + /* max map_on_demand */ + __u16 ibp_max_frags; + /* max_peer_credits */ + __u16 ibp_queue_depth; } kib_peer_t; extern kib_data_t kiblnd_data; @@ -611,7 +636,7 @@ kiblnd_dev_can_failover(kib_dev_t *dev) if (!list_empty(&dev->ibd_fail_list)) /* already scheduled */ return 0; - if (*kiblnd_tunables.kib_dev_failover == 0) /* disabled */ + if (!*kiblnd_tunables.kib_dev_failover) /* disabled */ return 0; if (*kiblnd_tunables.kib_dev_failover > 1) /* force failover */ @@ -661,6 +686,20 @@ do { \ kiblnd_destroy_peer(peer); \ } while (0) +static inline bool +kiblnd_peer_connecting(kib_peer_t *peer) +{ + return peer->ibp_connecting || + peer->ibp_reconnecting || + peer->ibp_accepting; +} + +static inline bool +kiblnd_peer_idle(kib_peer_t *peer) +{ + return !kiblnd_peer_connecting(peer) && list_empty(&peer->ibp_conns); +} + static inline struct list_head * kiblnd_nid2peerlist(lnet_nid_t nid) { @@ -691,7 +730,8 @@ kiblnd_send_keepalive(kib_conn_t *conn) { return (*kiblnd_tunables.kib_keepalive > 0) && cfs_time_after(jiffies, conn->ibc_last_send + - *kiblnd_tunables.kib_keepalive*HZ); + msecs_to_jiffies(*kiblnd_tunables.kib_keepalive * + MSEC_PER_SEC)); } static inline int @@ -710,16 +750,16 @@ kiblnd_need_noop(kib_conn_t *conn) /* No tx to piggyback NOOP onto or no credit to send a tx */ return (list_empty(&conn->ibc_tx_queue) || - conn->ibc_credits == 0); + !conn->ibc_credits); } if (!list_empty(&conn->ibc_tx_noops) || /* NOOP already queued */ !list_empty(&conn->ibc_tx_queue_nocred) || /* piggyback NOOP */ - conn->ibc_credits == 0) /* no credit */ + !conn->ibc_credits) /* no credit */ return 0; if (conn->ibc_credits == 1 && /* last credit reserved for */ - conn->ibc_outstanding_credits == 0) /* giving back credits */ + !conn->ibc_outstanding_credits) /* giving back credits */ return 0; /* No tx to piggyback NOOP onto or no credit to send a tx */ @@ -755,18 +795,19 @@ kiblnd_queue2str(kib_conn_t *conn, struct list_head *q) /* CAVEAT EMPTOR: We rely on descriptor alignment to allow us to use the */ /* lowest bits of the work request id to stash the work item type. */ -#define IBLND_WID_TX 0 -#define IBLND_WID_RDMA 1 -#define IBLND_WID_RX 2 -#define IBLND_WID_MASK 3UL +#define IBLND_WID_INVAL 0 +#define IBLND_WID_TX 1 +#define IBLND_WID_RX 2 +#define IBLND_WID_RDMA 3 +#define IBLND_WID_MASK 3UL static inline __u64 kiblnd_ptr2wreqid(void *ptr, int type) { unsigned long lptr = (unsigned long)ptr; - LASSERT((lptr & IBLND_WID_MASK) == 0); - LASSERT((type & ~IBLND_WID_MASK) == 0); + LASSERT(!(lptr & IBLND_WID_MASK)); + LASSERT(!(type & ~IBLND_WID_MASK)); return (__u64)(lptr | type); } @@ -907,9 +948,8 @@ static inline unsigned int kiblnd_sg_dma_len(struct ib_device *dev, #define KIBLND_CONN_PARAM_LEN(e) ((e)->param.conn.private_data_len) struct ib_mr *kiblnd_find_rd_dma_mr(kib_hca_dev_t *hdev, - kib_rdma_desc_t *rd); -struct ib_mr *kiblnd_find_dma_mr(kib_hca_dev_t *hdev, - __u64 addr, __u64 size); + kib_rdma_desc_t *rd, + int negotiated_nfrags); void kiblnd_map_rx_descs(kib_conn_t *conn); void kiblnd_unmap_rx_descs(kib_conn_t *conn); void kiblnd_pool_free_node(kib_pool_t *pool, struct list_head *node); @@ -919,11 +959,6 @@ int kiblnd_fmr_pool_map(kib_fmr_poolset_t *fps, __u64 *pages, int npages, __u64 iov, kib_fmr_t *fmr); void kiblnd_fmr_pool_unmap(kib_fmr_t *fmr, int status); -int kiblnd_startup(lnet_ni_t *ni); -void kiblnd_shutdown(lnet_ni_t *ni); -int kiblnd_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg); -void kiblnd_query(struct lnet_ni *ni, lnet_nid_t nid, unsigned long *when); - int kiblnd_tunables_init(void); void kiblnd_tunables_fini(void); @@ -933,7 +968,6 @@ int kiblnd_thread_start(int (*fn)(void *arg), void *arg, char *name); int kiblnd_failover_thread(void *arg); int kiblnd_alloc_pages(kib_pages_t **pp, int cpt, int npages); -void kiblnd_free_pages(kib_pages_t *p); int kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event); @@ -942,39 +976,30 @@ int kiblnd_translate_mtu(int value); int kiblnd_dev_failover(kib_dev_t *dev); int kiblnd_create_peer(lnet_ni_t *ni, kib_peer_t **peerp, lnet_nid_t nid); void kiblnd_destroy_peer(kib_peer_t *peer); +bool kiblnd_reconnect_peer(kib_peer_t *peer); void kiblnd_destroy_dev(kib_dev_t *dev); void kiblnd_unlink_peer_locked(kib_peer_t *peer); -void kiblnd_peer_alive(kib_peer_t *peer); kib_peer_t *kiblnd_find_peer_locked(lnet_nid_t nid); -void kiblnd_peer_connect_failed(kib_peer_t *peer, int active, int error); int kiblnd_close_stale_conns_locked(kib_peer_t *peer, - int version, __u64 incarnation); + int version, __u64 incarnation); int kiblnd_close_peer_conns_locked(kib_peer_t *peer, int why); -void kiblnd_connreq_done(kib_conn_t *conn, int status); kib_conn_t *kiblnd_create_conn(kib_peer_t *peer, struct rdma_cm_id *cmid, - int state, int version); -void kiblnd_destroy_conn(kib_conn_t *conn); + int state, int version); +void kiblnd_destroy_conn(kib_conn_t *conn, bool free_conn); void kiblnd_close_conn(kib_conn_t *conn, int error); void kiblnd_close_conn_locked(kib_conn_t *conn, int error); -int kiblnd_init_rdma(kib_conn_t *conn, kib_tx_t *tx, int type, - int nob, kib_rdma_desc_t *dstrd, __u64 dstcookie); - void kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid); -void kiblnd_queue_tx_locked(kib_tx_t *tx, kib_conn_t *conn); -void kiblnd_queue_tx(kib_tx_t *tx, kib_conn_t *conn); -void kiblnd_init_tx_msg(lnet_ni_t *ni, kib_tx_t *tx, int type, int body_nob); void kiblnd_txlist_done(lnet_ni_t *ni, struct list_head *txlist, - int status); -void kiblnd_check_sends (kib_conn_t *conn); + int status); void kiblnd_qp_event(struct ib_event *event, void *arg); void kiblnd_cq_event(struct ib_event *event, void *arg); void kiblnd_cq_completion(struct ib_cq *cq, void *arg); void kiblnd_pack_msg(lnet_ni_t *ni, kib_msg_t *msg, int version, - int credits, lnet_nid_t dstnid, __u64 dststamp); + int credits, lnet_nid_t dstnid, __u64 dststamp); int kiblnd_unpack_msg(kib_msg_t *msg, int nob); int kiblnd_post_rx(kib_rx_t *rx, int credit); diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c index c7b9ccb13..2323e8d3a 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c @@ -40,6 +40,15 @@ #include "o2iblnd.h" +static void kiblnd_peer_alive(kib_peer_t *peer); +static void kiblnd_peer_connect_failed(kib_peer_t *peer, int active, int error); +static void kiblnd_check_sends(kib_conn_t *conn); +static void kiblnd_init_tx_msg(lnet_ni_t *ni, kib_tx_t *tx, + int type, int body_nob); +static int kiblnd_init_rdma(kib_conn_t *conn, kib_tx_t *tx, int type, + int resid, kib_rdma_desc_t *dstrd, __u64 dstcookie); +static void kiblnd_queue_tx_locked(kib_tx_t *tx, kib_conn_t *conn); +static void kiblnd_queue_tx(kib_tx_t *tx, kib_conn_t *conn); static void kiblnd_unmap_tx(lnet_ni_t *ni, kib_tx_t *tx); static void @@ -50,12 +59,12 @@ kiblnd_tx_done(lnet_ni_t *ni, kib_tx_t *tx) int rc; int i; - LASSERT(net != NULL); + LASSERT(net); LASSERT(!in_interrupt()); LASSERT(!tx->tx_queued); /* mustn't be queued for sending */ - LASSERT(tx->tx_sending == 0); /* mustn't be awaiting sent callback */ + LASSERT(!tx->tx_sending); /* mustn't be awaiting sent callback */ LASSERT(!tx->tx_waiting); /* mustn't be awaiting peer response */ - LASSERT(tx->tx_pool != NULL); + LASSERT(tx->tx_pool); kiblnd_unmap_tx(ni, tx); @@ -64,7 +73,7 @@ kiblnd_tx_done(lnet_ni_t *ni, kib_tx_t *tx) lntmsg[1] = tx->tx_lntmsg[1]; tx->tx_lntmsg[1] = NULL; rc = tx->tx_status; - if (tx->tx_conn != NULL) { + if (tx->tx_conn) { LASSERT(ni == tx->tx_conn->ibc_peer->ibp_ni); kiblnd_conn_decref(tx->tx_conn); @@ -78,7 +87,7 @@ kiblnd_tx_done(lnet_ni_t *ni, kib_tx_t *tx) /* delay finalize until my descs have been freed */ for (i = 0; i < 2; i++) { - if (lntmsg[i] == NULL) + if (!lntmsg[i]) continue; lnet_finalize(ni, lntmsg[i], rc); @@ -111,19 +120,19 @@ kiblnd_get_idle_tx(lnet_ni_t *ni, lnet_nid_t target) tps = net->ibn_tx_ps[lnet_cpt_of_nid(target)]; node = kiblnd_pool_alloc_node(&tps->tps_poolset); - if (node == NULL) + if (!node) return NULL; - tx = container_of(node, kib_tx_t, tx_list); + tx = list_entry(node, kib_tx_t, tx_list); - LASSERT(tx->tx_nwrq == 0); + LASSERT(!tx->tx_nwrq); LASSERT(!tx->tx_queued); - LASSERT(tx->tx_sending == 0); + LASSERT(!tx->tx_sending); LASSERT(!tx->tx_waiting); - LASSERT(tx->tx_status == 0); - LASSERT(tx->tx_conn == NULL); - LASSERT(tx->tx_lntmsg[0] == NULL); - LASSERT(tx->tx_lntmsg[1] == NULL); - LASSERT(tx->tx_nfrags == 0); + LASSERT(!tx->tx_status); + LASSERT(!tx->tx_conn); + LASSERT(!tx->tx_lntmsg[0]); + LASSERT(!tx->tx_lntmsg[1]); + LASSERT(!tx->tx_nfrags); return tx; } @@ -149,17 +158,15 @@ kiblnd_post_rx(kib_rx_t *rx, int credit) kib_conn_t *conn = rx->rx_conn; kib_net_t *net = conn->ibc_peer->ibp_ni->ni_data; struct ib_recv_wr *bad_wrq = NULL; - struct ib_mr *mr; + struct ib_mr *mr = conn->ibc_hdev->ibh_mrs; int rc; - LASSERT(net != NULL); + LASSERT(net); LASSERT(!in_interrupt()); LASSERT(credit == IBLND_POSTRX_NO_CREDIT || credit == IBLND_POSTRX_PEER_CREDIT || credit == IBLND_POSTRX_RSRVD_CREDIT); - - mr = kiblnd_find_dma_mr(conn->ibc_hdev, rx->rx_msgaddr, IBLND_MSG_SIZE); - LASSERT(mr != NULL); + LASSERT(mr); rx->rx_sge.lkey = mr->lkey; rx->rx_sge.addr = rx->rx_msgaddr; @@ -185,7 +192,7 @@ kiblnd_post_rx(kib_rx_t *rx, int credit) */ kiblnd_conn_addref(conn); rc = ib_post_recv(conn->ibc_cmid->qp, &rx->rx_wrq, &bad_wrq); - if (unlikely(rc != 0)) { + if (unlikely(rc)) { CERROR("Can't post rx for %s: %d, bad_wrq: %p\n", libcfs_nid2str(conn->ibc_peer->ibp_nid), rc, bad_wrq); rx->rx_nob = 0; @@ -194,7 +201,7 @@ kiblnd_post_rx(kib_rx_t *rx, int credit) if (conn->ibc_state < IBLND_CONN_ESTABLISHED) /* Initial post */ goto out; - if (unlikely(rc != 0)) { + if (unlikely(rc)) { kiblnd_close_conn(conn, rc); kiblnd_drop_rx(rx); /* No more posts for this rx */ goto out; @@ -225,7 +232,7 @@ kiblnd_find_waiting_tx_locked(kib_conn_t *conn, int txtype, __u64 cookie) kib_tx_t *tx = list_entry(tmp, kib_tx_t, tx_list); LASSERT(!tx->tx_queued); - LASSERT(tx->tx_sending != 0 || tx->tx_waiting); + LASSERT(tx->tx_sending || tx->tx_waiting); if (tx->tx_cookie != cookie) continue; @@ -251,7 +258,7 @@ kiblnd_handle_completion(kib_conn_t *conn, int txtype, int status, __u64 cookie) spin_lock(&conn->ibc_lock); tx = kiblnd_find_waiting_tx_locked(conn, txtype, cookie); - if (tx == NULL) { + if (!tx) { spin_unlock(&conn->ibc_lock); CWARN("Unmatched completion type %x cookie %#llx from %s\n", @@ -260,7 +267,7 @@ kiblnd_handle_completion(kib_conn_t *conn, int txtype, int status, __u64 cookie) return; } - if (tx->tx_status == 0) { /* success so far */ + if (!tx->tx_status) { /* success so far */ if (status < 0) /* failed? */ tx->tx_status = status; else if (txtype == IBLND_MSG_GET_REQ) @@ -269,7 +276,7 @@ kiblnd_handle_completion(kib_conn_t *conn, int txtype, int status, __u64 cookie) tx->tx_waiting = 0; - idle = !tx->tx_queued && (tx->tx_sending == 0); + idle = !tx->tx_queued && !tx->tx_sending; if (idle) list_del(&tx->tx_list); @@ -285,7 +292,7 @@ kiblnd_send_completion(kib_conn_t *conn, int type, int status, __u64 cookie) lnet_ni_t *ni = conn->ibc_peer->ibp_ni; kib_tx_t *tx = kiblnd_get_idle_tx(ni, conn->ibc_peer->ibp_nid); - if (tx == NULL) { + if (!tx) { CERROR("Can't get tx for completion %x for %s\n", type, libcfs_nid2str(conn->ibc_peer->ibp_nid)); return; @@ -316,19 +323,18 @@ kiblnd_handle_rx(kib_rx_t *rx) msg->ibm_type, credits, libcfs_nid2str(conn->ibc_peer->ibp_nid)); - if (credits != 0) { + if (credits) { /* Have I received credits that will let me send? */ spin_lock(&conn->ibc_lock); if (conn->ibc_credits + credits > - IBLND_MSG_QUEUE_SIZE(conn->ibc_version)) { + conn->ibc_queue_depth) { rc2 = conn->ibc_credits; spin_unlock(&conn->ibc_lock); CERROR("Bad credits from %s: %d + %d > %d\n", libcfs_nid2str(conn->ibc_peer->ibp_nid), - rc2, credits, - IBLND_MSG_QUEUE_SIZE(conn->ibc_version)); + rc2, credits, conn->ibc_queue_depth); kiblnd_close_conn(conn, -EPROTO); kiblnd_post_rx(rx, IBLND_POSTRX_NO_CREDIT); @@ -360,7 +366,7 @@ kiblnd_handle_rx(kib_rx_t *rx) break; } - if (credits != 0) /* credit already posted */ + if (credits) /* credit already posted */ post_credit = IBLND_POSTRX_NO_CREDIT; else /* a keepalive NOOP */ post_credit = IBLND_POSTRX_PEER_CREDIT; @@ -396,12 +402,12 @@ kiblnd_handle_rx(kib_rx_t *rx) spin_lock(&conn->ibc_lock); tx = kiblnd_find_waiting_tx_locked(conn, IBLND_MSG_PUT_REQ, - msg->ibm_u.putack.ibpam_src_cookie); - if (tx != NULL) + msg->ibm_u.putack.ibpam_src_cookie); + if (tx) list_del(&tx->tx_list); spin_unlock(&conn->ibc_lock); - if (tx == NULL) { + if (!tx) { CERROR("Unmatched PUT_ACK from %s\n", libcfs_nid2str(conn->ibc_peer->ibp_nid)); rc = -EPROTO; @@ -409,10 +415,11 @@ kiblnd_handle_rx(kib_rx_t *rx) } LASSERT(tx->tx_waiting); - /* CAVEAT EMPTOR: I could be racing with tx_complete, but... + /* + * CAVEAT EMPTOR: I could be racing with tx_complete, but... * (a) I can overwrite tx_msg since my peer has received it! - * (b) tx_waiting set tells tx_complete() it's not done. */ - + * (b) tx_waiting set tells tx_complete() it's not done. + */ tx->tx_nwrq = 0; /* overwrite PUT_REQ */ rc2 = kiblnd_init_rdma(conn, tx, IBLND_MSG_PUT_DONE, @@ -469,7 +476,7 @@ kiblnd_rx_complete(kib_rx_t *rx, int status, int nob) int rc; int err = -EIO; - LASSERT(net != NULL); + LASSERT(net); LASSERT(rx->rx_nob < 0); /* was posted */ rx->rx_nob = 0; /* isn't now */ @@ -486,9 +493,9 @@ kiblnd_rx_complete(kib_rx_t *rx, int status, int nob) rx->rx_nob = nob; rc = kiblnd_unpack_msg(msg, rx->rx_nob); - if (rc != 0) { + if (rc) { CERROR("Error %d unpacking rx from %s\n", - rc, libcfs_nid2str(conn->ibc_peer->ibp_nid)); + rc, libcfs_nid2str(conn->ibc_peer->ibp_nid)); goto failed; } @@ -497,7 +504,7 @@ kiblnd_rx_complete(kib_rx_t *rx, int status, int nob) msg->ibm_srcstamp != conn->ibc_incarnation || msg->ibm_dststamp != net->ibn_incarnation) { CERROR("Stale rx from %s\n", - libcfs_nid2str(conn->ibc_peer->ibp_nid)); + libcfs_nid2str(conn->ibc_peer->ibp_nid)); err = -ESTALE; goto failed; } @@ -537,7 +544,7 @@ kiblnd_kvaddr_to_page(unsigned long vaddr) if (is_vmalloc_addr((void *)vaddr)) { page = vmalloc_to_page((void *)vaddr); - LASSERT(page != NULL); + LASSERT(page); return page; } #ifdef CONFIG_HIGHMEM @@ -549,7 +556,7 @@ kiblnd_kvaddr_to_page(unsigned long vaddr) } #endif page = virt_to_page(vaddr); - LASSERT(page != NULL); + LASSERT(page); return page; } @@ -565,8 +572,8 @@ kiblnd_fmr_map_tx(kib_net_t *net, kib_tx_t *tx, kib_rdma_desc_t *rd, int nob) int rc; int i; - LASSERT(tx->tx_pool != NULL); - LASSERT(tx->tx_pool->tpo_pool.po_owner != NULL); + LASSERT(tx->tx_pool); + LASSERT(tx->tx_pool->tpo_pool.po_owner); hdev = tx->tx_pool->tpo_hdev; @@ -582,13 +589,15 @@ kiblnd_fmr_map_tx(kib_net_t *net, kib_tx_t *tx, kib_rdma_desc_t *rd, int nob) fps = net->ibn_fmr_ps[cpt]; rc = kiblnd_fmr_pool_map(fps, pages, npages, 0, &tx->fmr); - if (rc != 0) { + if (rc) { CERROR("Can't map %d pages: %d\n", npages, rc); return rc; } - /* If rd is not tx_rd, it's going to get sent to a peer, who will need - * the rkey */ + /* + * If rd is not tx_rd, it's going to get sent to a peer, who will need + * the rkey + */ rd->rd_key = (rd != tx->tx_rd) ? tx->fmr.fmr_pfmr->fmr->rkey : tx->fmr.fmr_pfmr->fmr->lkey; rd->rd_frags[0].rf_addr &= ~hdev->ibh_page_mask; @@ -602,14 +611,14 @@ static void kiblnd_unmap_tx(lnet_ni_t *ni, kib_tx_t *tx) { kib_net_t *net = ni->ni_data; - LASSERT(net != NULL); + LASSERT(net); if (net->ibn_fmr_ps && tx->fmr.fmr_pfmr) { kiblnd_fmr_pool_unmap(&tx->fmr, tx->tx_status); tx->fmr.fmr_pfmr = NULL; } - if (tx->tx_nfrags != 0) { + if (tx->tx_nfrags) { kiblnd_dma_unmap_sg(tx->tx_pool->tpo_hdev->ibh_ibdev, tx->tx_frags, tx->tx_nfrags, tx->tx_dmadir); tx->tx_nfrags = 0; @@ -625,8 +634,10 @@ static int kiblnd_map_tx(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, __u32 nob; int i; - /* If rd is not tx_rd, it's going to get sent to a peer and I'm the - * RDMA sink */ + /* + * If rd is not tx_rd, it's going to get sent to a peer and I'm the + * RDMA sink + */ tx->tx_dmadir = (rd != tx->tx_rd) ? DMA_FROM_DEVICE : DMA_TO_DEVICE; tx->tx_nfrags = nfrags; @@ -641,15 +652,15 @@ static int kiblnd_map_tx(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, nob += rd->rd_frags[i].rf_nob; } - /* looking for pre-mapping MR */ - mr = kiblnd_find_rd_dma_mr(hdev, rd); - if (mr != NULL) { + mr = kiblnd_find_rd_dma_mr(hdev, rd, tx->tx_conn ? + tx->tx_conn->ibc_max_frags : -1); + if (mr) { /* found pre-mapping MR */ rd->rd_key = (rd != tx->tx_rd) ? mr->rkey : mr->lkey; return 0; } - if (net->ibn_fmr_ps != NULL) + if (net->ibn_fmr_ps) return kiblnd_fmr_map_tx(net, tx, rd, nob); return -EINVAL; @@ -668,7 +679,7 @@ kiblnd_setup_rd_iov(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, LASSERT(nob > 0); LASSERT(niov > 0); - LASSERT(net != NULL); + LASSERT(net); while (offset >= iov->iov_len) { offset -= iov->iov_len; @@ -684,7 +695,7 @@ kiblnd_setup_rd_iov(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, vaddr = ((unsigned long)iov->iov_base) + offset; page_offset = vaddr & (PAGE_SIZE - 1); page = kiblnd_kvaddr_to_page(vaddr); - if (page == NULL) { + if (!page) { CERROR("Can't find page\n"); return -EFAULT; } @@ -710,7 +721,7 @@ kiblnd_setup_rd_iov(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, static int kiblnd_setup_rd_kiov(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, - int nkiov, lnet_kiov_t *kiov, int offset, int nob) + int nkiov, lnet_kiov_t *kiov, int offset, int nob) { kib_net_t *net = ni->ni_data; struct scatterlist *sg; @@ -720,7 +731,7 @@ kiblnd_setup_rd_kiov(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, LASSERT(nob > 0); LASSERT(nkiov > 0); - LASSERT(net != NULL); + LASSERT(net); while (offset >= kiov->kiov_len) { offset -= kiov->kiov_len; @@ -750,26 +761,24 @@ kiblnd_setup_rd_kiov(lnet_ni_t *ni, kib_tx_t *tx, kib_rdma_desc_t *rd, static int kiblnd_post_tx_locked(kib_conn_t *conn, kib_tx_t *tx, int credit) - __releases(conn->ibc_lock) - __acquires(conn->ibc_lock) + __must_hold(&conn->ibc_lock) { kib_msg_t *msg = tx->tx_msg; kib_peer_t *peer = conn->ibc_peer; int ver = conn->ibc_version; int rc; int done; - struct ib_send_wr *bad_wrq; LASSERT(tx->tx_queued); /* We rely on this for QP sizing */ LASSERT(tx->tx_nwrq > 0); - LASSERT(tx->tx_nwrq <= 1 + IBLND_RDMA_FRAGS(ver)); + LASSERT(tx->tx_nwrq <= 1 + conn->ibc_max_frags); - LASSERT(credit == 0 || credit == 1); + LASSERT(!credit || credit == 1); LASSERT(conn->ibc_outstanding_credits >= 0); - LASSERT(conn->ibc_outstanding_credits <= IBLND_MSG_QUEUE_SIZE(ver)); + LASSERT(conn->ibc_outstanding_credits <= conn->ibc_queue_depth); LASSERT(conn->ibc_credits >= 0); - LASSERT(conn->ibc_credits <= IBLND_MSG_QUEUE_SIZE(ver)); + LASSERT(conn->ibc_credits <= conn->ibc_queue_depth); if (conn->ibc_nsends_posted == IBLND_CONCURRENT_SENDS(ver)) { /* tx completions outstanding... */ @@ -778,13 +787,13 @@ kiblnd_post_tx_locked(kib_conn_t *conn, kib_tx_t *tx, int credit) return -EAGAIN; } - if (credit != 0 && conn->ibc_credits == 0) { /* no credits */ + if (credit && !conn->ibc_credits) { /* no credits */ CDEBUG(D_NET, "%s: no credits\n", libcfs_nid2str(peer->ibp_nid)); return -EAGAIN; } - if (credit != 0 && !IBLND_OOB_CAPABLE(ver) && + if (credit && !IBLND_OOB_CAPABLE(ver) && conn->ibc_credits == 1 && /* last credit reserved */ msg->ibm_type != IBLND_MSG_NOOP) { /* for NOOP */ CDEBUG(D_NET, "%s: not using last credit\n", @@ -800,9 +809,11 @@ kiblnd_post_tx_locked(kib_conn_t *conn, kib_tx_t *tx, int credit) (!kiblnd_need_noop(conn) || /* redundant NOOP */ (IBLND_OOB_CAPABLE(ver) && /* posted enough NOOP */ conn->ibc_noops_posted == IBLND_OOB_MSGS(ver)))) { - /* OK to drop when posted enough NOOPs, since + /* + * OK to drop when posted enough NOOPs, since * kiblnd_check_sends will queue NOOP again when - * posted NOOPs complete */ + * posted NOOPs complete + */ spin_unlock(&conn->ibc_lock); kiblnd_tx_done(peer->ibp_ni, tx); spin_lock(&conn->ibc_lock); @@ -821,12 +832,14 @@ kiblnd_post_tx_locked(kib_conn_t *conn, kib_tx_t *tx, int credit) if (msg->ibm_type == IBLND_MSG_NOOP) conn->ibc_noops_posted++; - /* CAVEAT EMPTOR! This tx could be the PUT_DONE of an RDMA + /* + * CAVEAT EMPTOR! This tx could be the PUT_DONE of an RDMA * PUT. If so, it was first queued here as a PUT_REQ, sent and * stashed on ibc_active_txs, matched by an incoming PUT_ACK, * and then re-queued here. It's (just) possible that * tx_sending is non-zero if we've not done the tx_complete() - * from the first send; hence the ++ rather than = below. */ + * from the first send; hence the ++ rather than = below. + */ tx->tx_sending++; list_add(&tx->tx_list, &conn->ibc_active_txs); @@ -838,16 +851,25 @@ kiblnd_post_tx_locked(kib_conn_t *conn, kib_tx_t *tx, int credit) /* close_conn will launch failover */ rc = -ENETDOWN; } else { - rc = ib_post_send(conn->ibc_cmid->qp, &tx->tx_wrq->wr, &bad_wrq); + struct ib_send_wr *wrq = &tx->tx_wrq[tx->tx_nwrq - 1].wr; + + LASSERTF(wrq->wr_id == kiblnd_ptr2wreqid(tx, IBLND_WID_TX), + "bad wr_id %llx, opc %d, flags %d, peer: %s\n", + wrq->wr_id, wrq->opcode, wrq->send_flags, + libcfs_nid2str(conn->ibc_peer->ibp_nid)); + wrq = NULL; + rc = ib_post_send(conn->ibc_cmid->qp, &tx->tx_wrq->wr, &wrq); } conn->ibc_last_send = jiffies; - if (rc == 0) + if (!rc) return 0; - /* NB credits are transferred in the actual - * message, which can only be the last work item */ + /* + * NB credits are transferred in the actual + * message, which can only be the last work item + */ conn->ibc_credits += credit; conn->ibc_outstanding_credits += msg->ibm_credits; conn->ibc_nsends_posted--; @@ -858,7 +880,7 @@ kiblnd_post_tx_locked(kib_conn_t *conn, kib_tx_t *tx, int credit) tx->tx_waiting = 0; tx->tx_sending--; - done = (tx->tx_sending == 0); + done = !tx->tx_sending; if (done) list_del(&tx->tx_list); @@ -881,7 +903,7 @@ kiblnd_post_tx_locked(kib_conn_t *conn, kib_tx_t *tx, int credit) return -EIO; } -void +static void kiblnd_check_sends(kib_conn_t *conn) { int ver = conn->ibc_version; @@ -899,13 +921,13 @@ kiblnd_check_sends(kib_conn_t *conn) LASSERT(conn->ibc_nsends_posted <= IBLND_CONCURRENT_SENDS(ver)); LASSERT(!IBLND_OOB_CAPABLE(ver) || - conn->ibc_noops_posted <= IBLND_OOB_MSGS(ver)); + conn->ibc_noops_posted <= IBLND_OOB_MSGS(ver)); LASSERT(conn->ibc_reserved_credits >= 0); while (conn->ibc_reserved_credits > 0 && !list_empty(&conn->ibc_tx_queue_rsrvd)) { tx = list_entry(conn->ibc_tx_queue_rsrvd.next, - kib_tx_t, tx_list); + kib_tx_t, tx_list); list_del(&tx->tx_list); list_add_tail(&tx->tx_list, &conn->ibc_tx_queue); conn->ibc_reserved_credits--; @@ -915,23 +937,21 @@ kiblnd_check_sends(kib_conn_t *conn) spin_unlock(&conn->ibc_lock); tx = kiblnd_get_idle_tx(ni, conn->ibc_peer->ibp_nid); - if (tx != NULL) + if (tx) kiblnd_init_tx_msg(ni, tx, IBLND_MSG_NOOP, 0); spin_lock(&conn->ibc_lock); - if (tx != NULL) + if (tx) kiblnd_queue_tx_locked(tx, conn); } - kiblnd_conn_addref(conn); /* 1 ref for me.... (see b21911) */ - for (;;) { int credit; if (!list_empty(&conn->ibc_tx_queue_nocred)) { credit = 0; tx = list_entry(conn->ibc_tx_queue_nocred.next, - kib_tx_t, tx_list); + kib_tx_t, tx_list); } else if (!list_empty(&conn->ibc_tx_noops)) { LASSERT(!IBLND_OOB_CAPABLE(ver)); credit = 1; @@ -940,17 +960,16 @@ kiblnd_check_sends(kib_conn_t *conn) } else if (!list_empty(&conn->ibc_tx_queue)) { credit = 1; tx = list_entry(conn->ibc_tx_queue.next, - kib_tx_t, tx_list); - } else + kib_tx_t, tx_list); + } else { break; + } - if (kiblnd_post_tx_locked(conn, tx, credit) != 0) + if (kiblnd_post_tx_locked(conn, tx, credit)) break; } spin_unlock(&conn->ibc_lock); - - kiblnd_conn_decref(conn); /* ...until here */ } static void @@ -976,9 +995,10 @@ kiblnd_tx_complete(kib_tx_t *tx, int status) spin_lock(&conn->ibc_lock); - /* I could be racing with rdma completion. Whoever makes 'tx' idle - * gets to free it, which also drops its ref on 'conn'. */ - + /* + * I could be racing with rdma completion. Whoever makes 'tx' idle + * gets to free it, which also drops its ref on 'conn'. + */ tx->tx_sending--; conn->ibc_nsends_posted--; if (tx->tx_msg->ibm_type == IBLND_MSG_NOOP) @@ -989,7 +1009,7 @@ kiblnd_tx_complete(kib_tx_t *tx, int status) tx->tx_status = -EIO; } - idle = (tx->tx_sending == 0) && /* This is the final callback */ + idle = !tx->tx_sending && /* This is the final callback */ !tx->tx_waiting && /* Not waiting for peer */ !tx->tx_queued; /* Not re-queued (PUT_DONE) */ if (idle) @@ -1007,24 +1027,22 @@ kiblnd_tx_complete(kib_tx_t *tx, int status) kiblnd_conn_decref(conn); /* ...until here */ } -void +static void kiblnd_init_tx_msg(lnet_ni_t *ni, kib_tx_t *tx, int type, int body_nob) { kib_hca_dev_t *hdev = tx->tx_pool->tpo_hdev; struct ib_sge *sge = &tx->tx_sge[tx->tx_nwrq]; struct ib_rdma_wr *wrq = &tx->tx_wrq[tx->tx_nwrq]; int nob = offsetof(kib_msg_t, ibm_u) + body_nob; - struct ib_mr *mr; + struct ib_mr *mr = hdev->ibh_mrs; LASSERT(tx->tx_nwrq >= 0); LASSERT(tx->tx_nwrq < IBLND_MAX_RDMA_FRAGS + 1); LASSERT(nob <= IBLND_MSG_SIZE); + LASSERT(mr); kiblnd_init_msg(tx->tx_msg, type, body_nob); - mr = kiblnd_find_dma_mr(hdev, tx->tx_msgaddr, nob); - LASSERT(mr != NULL); - sge->lkey = mr->lkey; sge->addr = tx->tx_msgaddr; sge->length = nob; @@ -1041,25 +1059,23 @@ kiblnd_init_tx_msg(lnet_ni_t *ni, kib_tx_t *tx, int type, int body_nob) tx->tx_nwrq++; } -int +static int kiblnd_init_rdma(kib_conn_t *conn, kib_tx_t *tx, int type, - int resid, kib_rdma_desc_t *dstrd, __u64 dstcookie) + int resid, kib_rdma_desc_t *dstrd, __u64 dstcookie) { kib_msg_t *ibmsg = tx->tx_msg; kib_rdma_desc_t *srcrd = tx->tx_rd; struct ib_sge *sge = &tx->tx_sge[0]; struct ib_rdma_wr *wrq = &tx->tx_wrq[0], *next; int rc = resid; - int srcidx; - int dstidx; + int srcidx = 0; + int dstidx = 0; int wrknob; LASSERT(!in_interrupt()); - LASSERT(tx->tx_nwrq == 0); + LASSERT(!tx->tx_nwrq); LASSERT(type == IBLND_MSG_GET_DONE || - type == IBLND_MSG_PUT_DONE); - - srcidx = dstidx = 0; + type == IBLND_MSG_PUT_DONE); while (resid > 0) { if (srcidx >= srcrd->rd_nfrags) { @@ -1074,10 +1090,10 @@ kiblnd_init_rdma(kib_conn_t *conn, kib_tx_t *tx, int type, break; } - if (tx->tx_nwrq == IBLND_RDMA_FRAGS(conn->ibc_version)) { - CERROR("RDMA too fragmented for %s (%d): %d/%d src %d/%d dst frags\n", + if (tx->tx_nwrq >= conn->ibc_max_frags) { + CERROR("RDMA has too many fragments for peer %s (%d), src idx/frags: %d/%d dst idx/frags: %d/%d\n", libcfs_nid2str(conn->ibc_peer->ibp_nid), - IBLND_RDMA_FRAGS(conn->ibc_version), + conn->ibc_max_frags, srcidx, srcrd->rd_nfrags, dstidx, dstrd->rd_nfrags); rc = -EMSGSIZE; @@ -1127,7 +1143,7 @@ kiblnd_init_rdma(kib_conn_t *conn, kib_tx_t *tx, int type, return rc; } -void +static void kiblnd_queue_tx_locked(kib_tx_t *tx, kib_conn_t *conn) { struct list_head *q; @@ -1137,9 +1153,11 @@ kiblnd_queue_tx_locked(kib_tx_t *tx, kib_conn_t *conn) LASSERT(conn->ibc_state >= IBLND_CONN_ESTABLISHED); tx->tx_queued = 1; - tx->tx_deadline = jiffies + (*kiblnd_tunables.kib_timeout * HZ); + tx->tx_deadline = jiffies + + msecs_to_jiffies(*kiblnd_tunables.kib_timeout * + MSEC_PER_SEC); - if (tx->tx_conn == NULL) { + if (!tx->tx_conn) { kiblnd_conn_addref(conn); tx->tx_conn = conn; LASSERT(tx->tx_msg->ibm_type != IBLND_MSG_PUT_DONE); @@ -1180,7 +1198,7 @@ kiblnd_queue_tx_locked(kib_tx_t *tx, kib_conn_t *conn) list_add_tail(&tx->tx_list, q); } -void +static void kiblnd_queue_tx(kib_tx_t *tx, kib_conn_t *conn) { spin_lock(&conn->ibc_lock); @@ -1200,19 +1218,19 @@ static int kiblnd_resolve_addr(struct rdma_cm_id *cmid, /* allow the port to be reused */ rc = rdma_set_reuseaddr(cmid, 1); - if (rc != 0) { + if (rc) { CERROR("Unable to set reuse on cmid: %d\n", rc); return rc; } /* look for a free privileged port */ - for (port = PROT_SOCK-1; port > 0; port--) { + for (port = PROT_SOCK - 1; port > 0; port--) { srcaddr->sin_port = htons(port); rc = rdma_resolve_addr(cmid, (struct sockaddr *)srcaddr, (struct sockaddr *)dstaddr, timeout_ms); - if (rc == 0) { + if (!rc) { CDEBUG(D_NET, "bound to port %hu\n", port); return 0; } else if (rc == -EADDRINUSE || rc == -EADDRNOTAVAIL) { @@ -1237,8 +1255,9 @@ kiblnd_connect_peer(kib_peer_t *peer) struct sockaddr_in dstaddr; int rc; - LASSERT(net != NULL); + LASSERT(net); LASSERT(peer->ibp_connecting > 0); + LASSERT(!peer->ibp_reconnecting); cmid = kiblnd_rdma_create_id(kiblnd_cm_callback, peer, RDMA_PS_TCP, IB_QPT_RC); @@ -1271,14 +1290,14 @@ kiblnd_connect_peer(kib_peer_t *peer) (struct sockaddr *)&dstaddr, *kiblnd_tunables.kib_timeout * 1000); } - if (rc != 0) { + if (rc) { /* Can't initiate address resolution: */ CERROR("Can't resolve addr for %s: %d\n", libcfs_nid2str(peer->ibp_nid), rc); goto failed2; } - LASSERT(cmid->device != NULL); + LASSERT(cmid->device); CDEBUG(D_NET, "%s: connection bound to %s:%pI4h:%s\n", libcfs_nid2str(peer->ibp_nid), dev->ibd_ifname, &dev->ibd_ifip, cmid->device->name); @@ -1286,12 +1305,64 @@ kiblnd_connect_peer(kib_peer_t *peer) return; failed2: + kiblnd_peer_connect_failed(peer, 1, rc); kiblnd_peer_decref(peer); /* cmid's ref */ rdma_destroy_id(cmid); + return; failed: kiblnd_peer_connect_failed(peer, 1, rc); } +bool +kiblnd_reconnect_peer(kib_peer_t *peer) +{ + rwlock_t *glock = &kiblnd_data.kib_global_lock; + char *reason = NULL; + struct list_head txs; + unsigned long flags; + + INIT_LIST_HEAD(&txs); + + write_lock_irqsave(glock, flags); + if (!peer->ibp_reconnecting) { + if (peer->ibp_accepting) + reason = "accepting"; + else if (peer->ibp_connecting) + reason = "connecting"; + else if (!list_empty(&peer->ibp_conns)) + reason = "connected"; + else /* connected then closed */ + reason = "closed"; + + goto no_reconnect; + } + + LASSERT(!peer->ibp_accepting && !peer->ibp_connecting && + list_empty(&peer->ibp_conns)); + peer->ibp_reconnecting = 0; + + if (!kiblnd_peer_active(peer)) { + list_splice_init(&peer->ibp_tx_queue, &txs); + reason = "unlinked"; + goto no_reconnect; + } + + peer->ibp_connecting++; + peer->ibp_reconnected++; + write_unlock_irqrestore(glock, flags); + + kiblnd_connect_peer(peer); + return true; + +no_reconnect: + write_unlock_irqrestore(glock, flags); + + CWARN("Abort reconnection of %s: %s\n", + libcfs_nid2str(peer->ibp_nid), reason); + kiblnd_txlist_done(peer->ibp_ni, &txs, -ECONNABORTED); + return false; +} + void kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) { @@ -1302,25 +1373,28 @@ kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) unsigned long flags; int rc; - /* If I get here, I've committed to send, so I complete the tx with - * failure on any problems */ - - LASSERT(tx == NULL || tx->tx_conn == NULL); /* only set when assigned a conn */ - LASSERT(tx == NULL || tx->tx_nwrq > 0); /* work items have been set up */ + /* + * If I get here, I've committed to send, so I complete the tx with + * failure on any problems + */ + LASSERT(!tx || !tx->tx_conn); /* only set when assigned a conn */ + LASSERT(!tx || tx->tx_nwrq > 0); /* work items have been set up */ - /* First time, just use a read lock since I expect to find my peer - * connected */ + /* + * First time, just use a read lock since I expect to find my peer + * connected + */ read_lock_irqsave(g_lock, flags); peer = kiblnd_find_peer_locked(nid); - if (peer != NULL && !list_empty(&peer->ibp_conns)) { + if (peer && !list_empty(&peer->ibp_conns)) { /* Found a peer with an established connection */ conn = kiblnd_get_conn_locked(peer); kiblnd_conn_addref(conn); /* 1 ref for me... */ read_unlock_irqrestore(g_lock, flags); - if (tx != NULL) + if (tx) kiblnd_queue_tx(tx, conn); kiblnd_conn_decref(conn); /* ...to here */ return; @@ -1331,14 +1405,13 @@ kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) write_lock(g_lock); peer = kiblnd_find_peer_locked(nid); - if (peer != NULL) { + if (peer) { if (list_empty(&peer->ibp_conns)) { /* found a peer, but it's still connecting... */ - LASSERT(peer->ibp_connecting != 0 || - peer->ibp_accepting != 0); - if (tx != NULL) + LASSERT(kiblnd_peer_connecting(peer)); + if (tx) list_add_tail(&tx->tx_list, - &peer->ibp_tx_queue); + &peer->ibp_tx_queue); write_unlock_irqrestore(g_lock, flags); } else { conn = kiblnd_get_conn_locked(peer); @@ -1346,7 +1419,7 @@ kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) write_unlock_irqrestore(g_lock, flags); - if (tx != NULL) + if (tx) kiblnd_queue_tx(tx, conn); kiblnd_conn_decref(conn); /* ...to here */ } @@ -1357,9 +1430,9 @@ kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) /* Allocate a peer ready to add to the peer table and retry */ rc = kiblnd_create_peer(ni, &peer, nid); - if (rc != 0) { + if (rc) { CERROR("Can't create peer %s\n", libcfs_nid2str(nid)); - if (tx != NULL) { + if (tx) { tx->tx_status = -EHOSTUNREACH; tx->tx_waiting = 0; kiblnd_tx_done(ni, tx); @@ -1370,14 +1443,13 @@ kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) write_lock_irqsave(g_lock, flags); peer2 = kiblnd_find_peer_locked(nid); - if (peer2 != NULL) { + if (peer2) { if (list_empty(&peer2->ibp_conns)) { /* found a peer, but it's still connecting... */ - LASSERT(peer2->ibp_connecting != 0 || - peer2->ibp_accepting != 0); - if (tx != NULL) + LASSERT(kiblnd_peer_connecting(peer2)); + if (tx) list_add_tail(&tx->tx_list, - &peer2->ibp_tx_queue); + &peer2->ibp_tx_queue); write_unlock_irqrestore(g_lock, flags); } else { conn = kiblnd_get_conn_locked(peer2); @@ -1385,7 +1457,7 @@ kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) write_unlock_irqrestore(g_lock, flags); - if (tx != NULL) + if (tx) kiblnd_queue_tx(tx, conn); kiblnd_conn_decref(conn); /* ...to here */ } @@ -1395,13 +1467,13 @@ kiblnd_launch_tx(lnet_ni_t *ni, kib_tx_t *tx, lnet_nid_t nid) } /* Brand new peer */ - LASSERT(peer->ibp_connecting == 0); + LASSERT(!peer->ibp_connecting); peer->ibp_connecting = 1; /* always called with a ref on ni, which prevents ni being shutdown */ - LASSERT(((kib_net_t *)ni->ni_data)->ibn_shutdown == 0); + LASSERT(!((kib_net_t *)ni->ni_data)->ibn_shutdown); - if (tx != NULL) + if (tx) list_add_tail(&tx->tx_list, &peer->ibp_tx_queue); kiblnd_peer_addref(peer); @@ -1437,13 +1509,13 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) CDEBUG(D_NET, "sending %d bytes in %d frags to %s\n", payload_nob, payload_niov, libcfs_id2str(target)); - LASSERT(payload_nob == 0 || payload_niov > 0); + LASSERT(!payload_nob || payload_niov > 0); LASSERT(payload_niov <= LNET_MAX_IOV); /* Thread context */ LASSERT(!in_interrupt()); /* payload is either all vaddrs or all pages */ - LASSERT(!(payload_kiov != NULL && payload_iov != NULL)); + LASSERT(!(payload_kiov && payload_iov)); switch (type) { default: @@ -1451,7 +1523,7 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) return -EIO; case LNET_MSG_ACK: - LASSERT(payload_nob == 0); + LASSERT(!payload_nob); break; case LNET_MSG_GET: @@ -1464,7 +1536,7 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) break; /* send IMMEDIATE */ tx = kiblnd_get_idle_tx(ni, target.nid); - if (tx == NULL) { + if (!tx) { CERROR("Can't allocate txd for GET to %s\n", libcfs_nid2str(target.nid)); return -ENOMEM; @@ -1472,7 +1544,7 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) ibmsg = tx->tx_msg; rd = &ibmsg->ibm_u.get.ibgm_rd; - if ((lntmsg->msg_md->md_options & LNET_MD_KIOV) == 0) + if (!(lntmsg->msg_md->md_options & LNET_MD_KIOV)) rc = kiblnd_setup_rd_iov(ni, tx, rd, lntmsg->msg_md->md_niov, lntmsg->msg_md->md_iov.iov, @@ -1482,7 +1554,7 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) lntmsg->msg_md->md_niov, lntmsg->msg_md->md_iov.kiov, 0, lntmsg->msg_md->md_length); - if (rc != 0) { + if (rc) { CERROR("Can't setup GET sink for %s: %d\n", libcfs_nid2str(target.nid), rc); kiblnd_tx_done(ni, tx); @@ -1496,7 +1568,7 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) kiblnd_init_tx_msg(ni, tx, IBLND_MSG_GET_REQ, nob); tx->tx_lntmsg[1] = lnet_create_reply_msg(ni, lntmsg); - if (tx->tx_lntmsg[1] == NULL) { + if (!tx->tx_lntmsg[1]) { CERROR("Can't create reply for GET -> %s\n", libcfs_nid2str(target.nid)); kiblnd_tx_done(ni, tx); @@ -1516,14 +1588,14 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) break; /* send IMMEDIATE */ tx = kiblnd_get_idle_tx(ni, target.nid); - if (tx == NULL) { + if (!tx) { CERROR("Can't allocate %s txd for %s\n", type == LNET_MSG_PUT ? "PUT" : "REPLY", libcfs_nid2str(target.nid)); return -ENOMEM; } - if (payload_kiov == NULL) + if (!payload_kiov) rc = kiblnd_setup_rd_iov(ni, tx, tx->tx_rd, payload_niov, payload_iov, payload_offset, payload_nob); @@ -1531,7 +1603,7 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) rc = kiblnd_setup_rd_kiov(ni, tx, tx->tx_rd, payload_niov, payload_kiov, payload_offset, payload_nob); - if (rc != 0) { + if (rc) { CERROR("Can't setup PUT src for %s: %d\n", libcfs_nid2str(target.nid), rc); kiblnd_tx_done(ni, tx); @@ -1555,16 +1627,16 @@ kiblnd_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) <= IBLND_MSG_SIZE); tx = kiblnd_get_idle_tx(ni, target.nid); - if (tx == NULL) { + if (!tx) { CERROR("Can't send %d to %s: tx descs exhausted\n", - type, libcfs_nid2str(target.nid)); + type, libcfs_nid2str(target.nid)); return -ENOMEM; } ibmsg = tx->tx_msg; ibmsg->ibm_u.immediate.ibim_hdr = *hdr; - if (payload_kiov != NULL) + if (payload_kiov) lnet_copy_kiov2flat(IBLND_MSG_SIZE, ibmsg, offsetof(kib_msg_t, ibm_u.immediate.ibim_payload), payload_niov, payload_kiov, @@ -1596,22 +1668,22 @@ kiblnd_reply(lnet_ni_t *ni, kib_rx_t *rx, lnet_msg_t *lntmsg) int rc; tx = kiblnd_get_idle_tx(ni, rx->rx_conn->ibc_peer->ibp_nid); - if (tx == NULL) { + if (!tx) { CERROR("Can't get tx for REPLY to %s\n", libcfs_nid2str(target.nid)); goto failed_0; } - if (nob == 0) + if (!nob) rc = 0; - else if (kiov == NULL) + else if (!kiov) rc = kiblnd_setup_rd_iov(ni, tx, tx->tx_rd, niov, iov, offset, nob); else rc = kiblnd_setup_rd_kiov(ni, tx, tx->tx_rd, niov, kiov, offset, nob); - if (rc != 0) { + if (rc) { CERROR("Can't setup GET src for %s: %d\n", libcfs_nid2str(target.nid), rc); goto failed_1; @@ -1627,12 +1699,11 @@ kiblnd_reply(lnet_ni_t *ni, kib_rx_t *rx, lnet_msg_t *lntmsg) goto failed_1; } - if (nob == 0) { + if (!nob) { /* No RDMA: local completion may happen now! */ lnet_finalize(ni, lntmsg, 0); } else { - /* RDMA: lnet_finalize(lntmsg) when it - * completes */ + /* RDMA: lnet_finalize(lntmsg) when it completes */ tx->tx_lntmsg[0] = lntmsg; } @@ -1647,8 +1718,8 @@ kiblnd_reply(lnet_ni_t *ni, kib_rx_t *rx, lnet_msg_t *lntmsg) int kiblnd_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, - unsigned int niov, struct kvec *iov, lnet_kiov_t *kiov, - unsigned int offset, unsigned int mlen, unsigned int rlen) + unsigned int niov, struct kvec *iov, lnet_kiov_t *kiov, + unsigned int offset, unsigned int mlen, unsigned int rlen) { kib_rx_t *rx = private; kib_msg_t *rxmsg = rx->rx_msg; @@ -1661,7 +1732,7 @@ kiblnd_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, LASSERT(mlen <= rlen); LASSERT(!in_interrupt()); /* Either all pages or all vaddrs */ - LASSERT(!(kiov != NULL && iov != NULL)); + LASSERT(!(kiov && iov)); switch (rxmsg->ibm_type) { default: @@ -1671,13 +1742,13 @@ kiblnd_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, nob = offsetof(kib_msg_t, ibm_u.immediate.ibim_payload[rlen]); if (nob > rx->rx_nob) { CERROR("Immediate message from %s too big: %d(%d)\n", - libcfs_nid2str(rxmsg->ibm_u.immediate.ibim_hdr.src_nid), - nob, rx->rx_nob); + libcfs_nid2str(rxmsg->ibm_u.immediate.ibim_hdr.src_nid), + nob, rx->rx_nob); rc = -EPROTO; break; } - if (kiov != NULL) + if (kiov) lnet_copy_flat2kiov(niov, kiov, offset, IBLND_MSG_SIZE, rxmsg, offsetof(kib_msg_t, ibm_u.immediate.ibim_payload), @@ -1694,7 +1765,7 @@ kiblnd_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, kib_msg_t *txmsg; kib_rdma_desc_t *rd; - if (mlen == 0) { + if (!mlen) { lnet_finalize(ni, lntmsg, 0); kiblnd_send_completion(rx->rx_conn, IBLND_MSG_PUT_NAK, 0, rxmsg->ibm_u.putreq.ibprm_cookie); @@ -1702,7 +1773,7 @@ kiblnd_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, } tx = kiblnd_get_idle_tx(ni, conn->ibc_peer->ibp_nid); - if (tx == NULL) { + if (!tx) { CERROR("Can't allocate tx for %s\n", libcfs_nid2str(conn->ibc_peer->ibp_nid)); /* Not replying will break the connection */ @@ -1712,13 +1783,13 @@ kiblnd_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, txmsg = tx->tx_msg; rd = &txmsg->ibm_u.putack.ibpam_rd; - if (kiov == NULL) + if (!kiov) rc = kiblnd_setup_rd_iov(ni, tx, rd, niov, iov, offset, mlen); else rc = kiblnd_setup_rd_kiov(ni, tx, rd, niov, kiov, offset, mlen); - if (rc != 0) { + if (rc) { CERROR("Can't setup PUT sink for %s: %d\n", libcfs_nid2str(conn->ibc_peer->ibp_nid), rc); kiblnd_tx_done(ni, tx); @@ -1744,7 +1815,7 @@ kiblnd_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, } case IBLND_MSG_GET_REQ: - if (lntmsg != NULL) { + if (lntmsg) { /* Optimized GET; RDMA lntmsg's payload */ kiblnd_reply(ni, rx, lntmsg); } else { @@ -1778,7 +1849,7 @@ kiblnd_thread_fini(void) atomic_dec(&kiblnd_data.kib_nthreads); } -void +static void kiblnd_peer_alive(kib_peer_t *peer) { /* This is racy, but everyone's only writing cfs_time_current() */ @@ -1795,10 +1866,7 @@ kiblnd_peer_notify(kib_peer_t *peer) read_lock_irqsave(&kiblnd_data.kib_global_lock, flags); - if (list_empty(&peer->ibp_conns) && - peer->ibp_accepting == 0 && - peer->ibp_connecting == 0 && - peer->ibp_error != 0) { + if (kiblnd_peer_idle(peer) && peer->ibp_error) { error = peer->ibp_error; peer->ibp_error = 0; @@ -1807,7 +1875,7 @@ kiblnd_peer_notify(kib_peer_t *peer) read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); - if (error != 0) + if (error) lnet_notify(peer->ibp_ni, peer->ibp_nid, 0, last_alive); } @@ -1815,25 +1883,27 @@ kiblnd_peer_notify(kib_peer_t *peer) void kiblnd_close_conn_locked(kib_conn_t *conn, int error) { - /* This just does the immediate housekeeping. 'error' is zero for a + /* + * This just does the immediate housekeeping. 'error' is zero for a * normal shutdown which can happen only after the connection has been * established. If the connection is established, schedule the - * connection to be finished off by the connd. Otherwise the connd is + * connection to be finished off by the connd. Otherwise the connd is * already dealing with it (either to set it up or tear it down). - * Caller holds kib_global_lock exclusively in irq context */ + * Caller holds kib_global_lock exclusively in irq context + */ kib_peer_t *peer = conn->ibc_peer; kib_dev_t *dev; unsigned long flags; - LASSERT(error != 0 || conn->ibc_state >= IBLND_CONN_ESTABLISHED); + LASSERT(error || conn->ibc_state >= IBLND_CONN_ESTABLISHED); - if (error != 0 && conn->ibc_comms_error == 0) + if (error && !conn->ibc_comms_error) conn->ibc_comms_error = error; if (conn->ibc_state != IBLND_CONN_ESTABLISHED) return; /* already being handled */ - if (error == 0 && + if (!error && list_empty(&conn->ibc_tx_noops) && list_empty(&conn->ibc_tx_queue) && list_empty(&conn->ibc_tx_queue_rsrvd) && @@ -1843,12 +1913,12 @@ kiblnd_close_conn_locked(kib_conn_t *conn, int error) libcfs_nid2str(peer->ibp_nid)); } else { CNETERR("Closing conn to %s: error %d%s%s%s%s%s\n", - libcfs_nid2str(peer->ibp_nid), error, - list_empty(&conn->ibc_tx_queue) ? "" : "(sending)", - list_empty(&conn->ibc_tx_noops) ? "" : "(sending_noops)", - list_empty(&conn->ibc_tx_queue_rsrvd) ? "" : "(sending_rsrvd)", - list_empty(&conn->ibc_tx_queue_nocred) ? "" : "(sending_nocred)", - list_empty(&conn->ibc_active_txs) ? "" : "(waiting)"); + libcfs_nid2str(peer->ibp_nid), error, + list_empty(&conn->ibc_tx_queue) ? "" : "(sending)", + list_empty(&conn->ibc_tx_noops) ? "" : "(sending_noops)", + list_empty(&conn->ibc_tx_queue_rsrvd) ? "" : "(sending_rsrvd)", + list_empty(&conn->ibc_tx_queue_nocred) ? "" : "(sending_nocred)", + list_empty(&conn->ibc_active_txs) ? "" : "(waiting)"); } dev = ((kib_net_t *)peer->ibp_ni->ni_data)->ibn_dev; @@ -1865,7 +1935,7 @@ kiblnd_close_conn_locked(kib_conn_t *conn, int error) kiblnd_set_conn_state(conn, IBLND_CONN_CLOSING); - if (error != 0 && + if (error && kiblnd_dev_can_failover(dev)) { list_add_tail(&dev->ibd_fail_list, &kiblnd_data.kib_failed_devs); @@ -1929,8 +1999,7 @@ kiblnd_abort_txs(kib_conn_t *conn, struct list_head *txs) if (txs == &conn->ibc_active_txs) { LASSERT(!tx->tx_queued); - LASSERT(tx->tx_waiting || - tx->tx_sending != 0); + LASSERT(tx->tx_waiting || tx->tx_sending); } else { LASSERT(tx->tx_queued); } @@ -1938,7 +2007,7 @@ kiblnd_abort_txs(kib_conn_t *conn, struct list_head *txs) tx->tx_status = -ECONNABORTED; tx->tx_waiting = 0; - if (tx->tx_sending == 0) { + if (!tx->tx_sending) { tx->tx_queued = 0; list_del(&tx->tx_list); list_add(&tx->tx_list, &zombies); @@ -1958,14 +2027,17 @@ kiblnd_finalise_conn(kib_conn_t *conn) kiblnd_set_conn_state(conn, IBLND_CONN_DISCONNECTED); - /* abort_receives moves QP state to IB_QPS_ERR. This is only required + /* + * abort_receives moves QP state to IB_QPS_ERR. This is only required * for connections that didn't get as far as being connected, because - * rdma_disconnect() does this for free. */ + * rdma_disconnect() does this for free. + */ kiblnd_abort_receives(conn); - /* Complete all tx descs not waiting for sends to complete. - * NB we should be safe from RDMA now that the QP has changed state */ - + /* + * Complete all tx descs not waiting for sends to complete. + * NB we should be safe from RDMA now that the QP has changed state + */ kiblnd_abort_txs(conn, &conn->ibc_tx_noops); kiblnd_abort_txs(conn, &conn->ibc_tx_queue); kiblnd_abort_txs(conn, &conn->ibc_tx_queue_rsrvd); @@ -1975,13 +2047,13 @@ kiblnd_finalise_conn(kib_conn_t *conn) kiblnd_handle_early_rxs(conn); } -void +static void kiblnd_peer_connect_failed(kib_peer_t *peer, int active, int error) { LIST_HEAD(zombies); unsigned long flags; - LASSERT(error != 0); + LASSERT(error); LASSERT(!in_interrupt()); write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); @@ -1994,14 +2066,14 @@ kiblnd_peer_connect_failed(kib_peer_t *peer, int active, int error) peer->ibp_accepting--; } - if (peer->ibp_connecting != 0 || - peer->ibp_accepting != 0) { + if (kiblnd_peer_connecting(peer)) { /* another connection attempt under way... */ write_unlock_irqrestore(&kiblnd_data.kib_global_lock, - flags); + flags); return; } + peer->ibp_reconnected = 0; if (list_empty(&peer->ibp_conns)) { /* Take peer's blocked transmits to complete with error */ list_add(&zombies, &peer->ibp_tx_queue); @@ -2029,7 +2101,7 @@ kiblnd_peer_connect_failed(kib_peer_t *peer, int active, int error) kiblnd_txlist_done(peer->ibp_ni, &zombies, -EHOSTUNREACH); } -void +static void kiblnd_connreq_done(kib_conn_t *conn, int status) { kib_peer_t *peer = conn->ibc_peer; @@ -2047,14 +2119,14 @@ kiblnd_connreq_done(kib_conn_t *conn, int status) LASSERT(!in_interrupt()); LASSERT((conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT && - peer->ibp_connecting > 0) || + peer->ibp_connecting > 0) || (conn->ibc_state == IBLND_CONN_PASSIVE_WAIT && - peer->ibp_accepting > 0)); + peer->ibp_accepting > 0)); LIBCFS_FREE(conn->ibc_connvars, sizeof(*conn->ibc_connvars)); conn->ibc_connvars = NULL; - if (status != 0) { + if (status) { /* failed to establish connection */ kiblnd_peer_connect_failed(peer, active, status); kiblnd_finalise_conn(conn); @@ -2068,16 +2140,19 @@ kiblnd_connreq_done(kib_conn_t *conn, int status) kiblnd_set_conn_state(conn, IBLND_CONN_ESTABLISHED); kiblnd_peer_alive(peer); - /* Add conn to peer's list and nuke any dangling conns from a different - * peer instance... */ + /* + * Add conn to peer's list and nuke any dangling conns from a different + * peer instance... + */ kiblnd_conn_addref(conn); /* +1 ref for ibc_list */ list_add(&conn->ibc_list, &peer->ibp_conns); + peer->ibp_reconnected = 0; if (active) peer->ibp_connecting--; else peer->ibp_accepting--; - if (peer->ibp_version == 0) { + if (!peer->ibp_version) { peer->ibp_version = conn->ibc_version; peer->ibp_incarnation = conn->ibc_incarnation; } @@ -2095,7 +2170,7 @@ kiblnd_connreq_done(kib_conn_t *conn, int status) list_del_init(&peer->ibp_tx_queue); if (!kiblnd_peer_active(peer) || /* peer has been deleted */ - conn->ibc_comms_error != 0) { /* error has happened already */ + conn->ibc_comms_error) { /* error has happened already */ lnet_ni_t *ni = peer->ibp_ni; /* start to shut down connection */ @@ -2107,6 +2182,16 @@ kiblnd_connreq_done(kib_conn_t *conn, int status) return; } + /** + * refcount taken by cmid is not reliable after I released the glock + * because this connection is visible to other threads now, another + * thread can find and close this connection right after I released + * the glock, if kiblnd_cm_callback for RDMA_CM_EVENT_DISCONNECTED is + * called, it can release the connection refcount taken by cmid. + * It means the connection could be destroyed before I finish my + * operations on it. + */ + kiblnd_conn_addref(conn); write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); /* Schedule blocked txs */ @@ -2122,6 +2207,8 @@ kiblnd_connreq_done(kib_conn_t *conn, int status) /* schedule blocked rxs */ kiblnd_handle_early_rxs(conn); + + kiblnd_conn_decref(conn); } static void @@ -2131,7 +2218,7 @@ kiblnd_reject(struct rdma_cm_id *cmid, kib_rej_t *rej) rc = rdma_reject(cmid, rej, sizeof(*rej)); - if (rc != 0) + if (rc) CWARN("Error %d sending reject\n", rc); } @@ -2159,14 +2246,14 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) /* cmid inherits 'context' from the corresponding listener id */ ibdev = (kib_dev_t *)cmid->context; - LASSERT(ibdev != NULL); + LASSERT(ibdev); memset(&rej, 0, sizeof(rej)); rej.ibr_magic = IBLND_MSG_MAGIC; rej.ibr_why = IBLND_REJECT_FATAL; rej.ibr_cp.ibcp_max_msg_size = IBLND_MSG_SIZE; - peer_addr = (struct sockaddr_in *)&(cmid->route.addr.dst_addr); + peer_addr = (struct sockaddr_in *)&cmid->route.addr.dst_addr; if (*kiblnd_tunables.kib_require_priv_port && ntohs(peer_addr->sin_port) >= PROT_SOCK) { __u32 ip = ntohl(peer_addr->sin_addr.s_addr); @@ -2181,12 +2268,14 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) goto failed; } - /* Future protocol version compatibility support! If the + /* + * Future protocol version compatibility support! If the * o2iblnd-specific protocol changes, or when LNET unifies * protocols over all LNDs, the initial connection will * negotiate a protocol version. I trap this here to avoid * console errors; the reject tells the peer which protocol I - * speak. */ + * speak. + */ if (reqmsg->ibm_magic == LNET_PROTO_MAGIC || reqmsg->ibm_magic == __swab32(LNET_PROTO_MAGIC)) goto failed; @@ -2200,7 +2289,7 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) goto failed; rc = kiblnd_unpack_msg(reqmsg, priv_nob); - if (rc != 0) { + if (rc) { CERROR("Can't parse connection request: %d\n", rc); goto failed; } @@ -2208,17 +2297,17 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) nid = reqmsg->ibm_srcnid; ni = lnet_net2ni(LNET_NIDNET(reqmsg->ibm_dstnid)); - if (ni != NULL) { + if (ni) { net = (kib_net_t *)ni->ni_data; rej.ibr_incarnation = net->ibn_incarnation; } - if (ni == NULL || /* no matching net */ + if (!ni || /* no matching net */ ni->ni_nid != reqmsg->ibm_dstnid || /* right NET, wrong NID! */ net->ibn_dev != ibdev) { /* wrong device */ - CERROR("Can't accept %s on %s (%s:%d:%pI4h): bad dst nid %s\n", + CERROR("Can't accept conn from %s on %s (%s:%d:%pI4h): bad dst nid %s\n", libcfs_nid2str(nid), - ni == NULL ? "NA" : libcfs_nid2str(ni->ni_nid), + !ni ? "NA" : libcfs_nid2str(ni->ni_nid), ibdev->ibd_ifname, ibdev->ibd_nnets, &ibdev->ibd_ifip, libcfs_nid2str(reqmsg->ibm_dstnid)); @@ -2227,7 +2316,7 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) } /* check time stamp as soon as possible */ - if (reqmsg->ibm_dststamp != 0 && + if (reqmsg->ibm_dststamp && reqmsg->ibm_dststamp != net->ibn_incarnation) { CWARN("Stale connection request\n"); rej.ibr_why = IBLND_REJECT_CONN_STALE; @@ -2243,10 +2332,11 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) goto failed; } - if (reqmsg->ibm_u.connparams.ibcp_queue_depth != + if (reqmsg->ibm_u.connparams.ibcp_queue_depth > IBLND_MSG_QUEUE_SIZE(version)) { - CERROR("Can't accept %s: incompatible queue depth %d (%d wanted)\n", - libcfs_nid2str(nid), reqmsg->ibm_u.connparams.ibcp_queue_depth, + CERROR("Can't accept conn from %s, queue depth too large: %d (<=%d wanted)\n", + libcfs_nid2str(nid), + reqmsg->ibm_u.connparams.ibcp_queue_depth, IBLND_MSG_QUEUE_SIZE(version)); if (version == IBLND_MSG_VERSION) @@ -2255,18 +2345,28 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) goto failed; } - if (reqmsg->ibm_u.connparams.ibcp_max_frags != + if (reqmsg->ibm_u.connparams.ibcp_max_frags > IBLND_RDMA_FRAGS(version)) { - CERROR("Can't accept %s(version %x): incompatible max_frags %d (%d wanted)\n", - libcfs_nid2str(nid), version, - reqmsg->ibm_u.connparams.ibcp_max_frags, - IBLND_RDMA_FRAGS(version)); + CWARN("Can't accept conn from %s (version %x): max_frags %d too large (%d wanted)\n", + libcfs_nid2str(nid), version, + reqmsg->ibm_u.connparams.ibcp_max_frags, + IBLND_RDMA_FRAGS(version)); - if (version == IBLND_MSG_VERSION) + if (version >= IBLND_MSG_VERSION) rej.ibr_why = IBLND_REJECT_RDMA_FRAGS; goto failed; + } else if (reqmsg->ibm_u.connparams.ibcp_max_frags < + IBLND_RDMA_FRAGS(version) && !net->ibn_fmr_ps) { + CWARN("Can't accept conn from %s (version %x): max_frags %d incompatible without FMR pool (%d wanted)\n", + libcfs_nid2str(nid), version, + reqmsg->ibm_u.connparams.ibcp_max_frags, + IBLND_RDMA_FRAGS(version)); + + if (version >= IBLND_MSG_VERSION) + rej.ibr_why = IBLND_REJECT_RDMA_FRAGS; + goto failed; } if (reqmsg->ibm_u.connparams.ibcp_max_msg_size > IBLND_MSG_SIZE) { @@ -2279,17 +2379,21 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) /* assume 'nid' is a new peer; create */ rc = kiblnd_create_peer(ni, &peer, nid); - if (rc != 0) { + if (rc) { CERROR("Can't create peer for %s\n", libcfs_nid2str(nid)); rej.ibr_why = IBLND_REJECT_NO_RESOURCES; goto failed; } + /* We have validated the peer's parameters so use those */ + peer->ibp_max_frags = reqmsg->ibm_u.connparams.ibcp_max_frags; + peer->ibp_queue_depth = reqmsg->ibm_u.connparams.ibcp_queue_depth; + write_lock_irqsave(g_lock, flags); peer2 = kiblnd_find_peer_locked(nid); - if (peer2 != NULL) { - if (peer2->ibp_version == 0) { + if (peer2) { + if (!peer2->ibp_version) { peer2->ibp_version = version; peer2->ibp_incarnation = reqmsg->ibm_srcstamp; } @@ -2298,10 +2402,16 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) if (peer2->ibp_incarnation != reqmsg->ibm_srcstamp || peer2->ibp_version != version) { kiblnd_close_peer_conns_locked(peer2, -ESTALE); + + if (kiblnd_peer_active(peer2)) { + peer2->ibp_incarnation = reqmsg->ibm_srcstamp; + peer2->ibp_version = version; + } write_unlock_irqrestore(g_lock, flags); - CWARN("Conn stale %s [old ver: %x, new ver: %x]\n", - libcfs_nid2str(nid), peer2->ibp_version, version); + CWARN("Conn stale %s version %x/%x incarnation %llu/%llu\n", + libcfs_nid2str(nid), peer2->ibp_version, version, + peer2->ibp_incarnation, reqmsg->ibm_srcstamp); kiblnd_peer_decref(peer); rej.ibr_why = IBLND_REJECT_CONN_STALE; @@ -2309,7 +2419,7 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) } /* tie-break connection race in favour of the higher NID */ - if (peer2->ibp_connecting != 0 && + if (peer2->ibp_connecting && nid < ni->ni_nid) { write_unlock_irqrestore(g_lock, flags); @@ -2320,24 +2430,37 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) goto failed; } + /** + * passive connection is allowed even this peer is waiting for + * reconnection. + */ + peer2->ibp_reconnecting = 0; peer2->ibp_accepting++; kiblnd_peer_addref(peer2); + /** + * Race with kiblnd_launch_tx (active connect) to create peer + * so copy validated parameters since we now know what the + * peer's limits are + */ + peer2->ibp_max_frags = peer->ibp_max_frags; + peer2->ibp_queue_depth = peer->ibp_queue_depth; + write_unlock_irqrestore(g_lock, flags); kiblnd_peer_decref(peer); peer = peer2; } else { /* Brand new peer */ - LASSERT(peer->ibp_accepting == 0); - LASSERT(peer->ibp_version == 0 && - peer->ibp_incarnation == 0); + LASSERT(!peer->ibp_accepting); + LASSERT(!peer->ibp_version && + !peer->ibp_incarnation); peer->ibp_accepting = 1; peer->ibp_version = version; peer->ibp_incarnation = reqmsg->ibm_srcstamp; /* I have a ref on ni that prevents it being shutdown */ - LASSERT(net->ibn_shutdown == 0); + LASSERT(!net->ibn_shutdown); kiblnd_peer_addref(peer); list_add_tail(&peer->ibp_list, kiblnd_nid2peerlist(nid)); @@ -2345,31 +2468,33 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) write_unlock_irqrestore(g_lock, flags); } - conn = kiblnd_create_conn(peer, cmid, IBLND_CONN_PASSIVE_WAIT, version); - if (conn == NULL) { + conn = kiblnd_create_conn(peer, cmid, IBLND_CONN_PASSIVE_WAIT, + version); + if (!conn) { kiblnd_peer_connect_failed(peer, 0, -ENOMEM); kiblnd_peer_decref(peer); rej.ibr_why = IBLND_REJECT_NO_RESOURCES; goto failed; } - /* conn now "owns" cmid, so I return success from here on to ensure the - * CM callback doesn't destroy cmid. */ - + /* + * conn now "owns" cmid, so I return success from here on to ensure the + * CM callback doesn't destroy cmid. + */ conn->ibc_incarnation = reqmsg->ibm_srcstamp; - conn->ibc_credits = IBLND_MSG_QUEUE_SIZE(version); - conn->ibc_reserved_credits = IBLND_MSG_QUEUE_SIZE(version); - LASSERT(conn->ibc_credits + conn->ibc_reserved_credits + IBLND_OOB_MSGS(version) - <= IBLND_RX_MSGS(version)); + conn->ibc_credits = conn->ibc_queue_depth; + conn->ibc_reserved_credits = conn->ibc_queue_depth; + LASSERT(conn->ibc_credits + conn->ibc_reserved_credits + + IBLND_OOB_MSGS(version) <= IBLND_RX_MSGS(conn)); ackmsg = &conn->ibc_connvars->cv_msg; memset(ackmsg, 0, sizeof(*ackmsg)); kiblnd_init_msg(ackmsg, IBLND_MSG_CONNACK, sizeof(ackmsg->ibm_u.connparams)); - ackmsg->ibm_u.connparams.ibcp_queue_depth = IBLND_MSG_QUEUE_SIZE(version); + ackmsg->ibm_u.connparams.ibcp_queue_depth = conn->ibc_queue_depth; + ackmsg->ibm_u.connparams.ibcp_max_frags = conn->ibc_max_frags; ackmsg->ibm_u.connparams.ibcp_max_msg_size = IBLND_MSG_SIZE; - ackmsg->ibm_u.connparams.ibcp_max_frags = IBLND_RDMA_FRAGS(version); kiblnd_pack_msg(ni, ackmsg, version, 0, nid, reqmsg->ibm_srcstamp); @@ -2385,7 +2510,7 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) CDEBUG(D_NET, "Accept %s\n", libcfs_nid2str(nid)); rc = rdma_accept(cmid, &cp); - if (rc != 0) { + if (rc) { CERROR("Can't accept %s: %d\n", libcfs_nid2str(nid), rc); rej.ibr_version = version; rej.ibr_why = IBLND_REJECT_FATAL; @@ -2399,7 +2524,7 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) return 0; failed: - if (ni != NULL) + if (ni) lnet_ni_decref(ni); rej.ibr_version = version; @@ -2411,45 +2536,82 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) } static void -kiblnd_reconnect(kib_conn_t *conn, int version, - __u64 incarnation, int why, kib_connparams_t *cp) +kiblnd_check_reconnect(kib_conn_t *conn, int version, + __u64 incarnation, int why, kib_connparams_t *cp) { + rwlock_t *glock = &kiblnd_data.kib_global_lock; kib_peer_t *peer = conn->ibc_peer; char *reason; - int retry = 0; + int msg_size = IBLND_MSG_SIZE; + int frag_num = -1; + int queue_dep = -1; + bool reconnect; unsigned long flags; LASSERT(conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT); LASSERT(peer->ibp_connecting > 0); /* 'conn' at least */ + LASSERT(!peer->ibp_reconnecting); - write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); + if (cp) { + msg_size = cp->ibcp_max_msg_size; + frag_num = cp->ibcp_max_frags; + queue_dep = cp->ibcp_queue_depth; + } - /* retry connection if it's still needed and no other connection + write_lock_irqsave(glock, flags); + /** + * retry connection if it's still needed and no other connection * attempts (active or passive) are in progress * NB: reconnect is still needed even when ibp_tx_queue is * empty if ibp_version != version because reconnect may be - * initiated by kiblnd_query() */ - if ((!list_empty(&peer->ibp_tx_queue) || - peer->ibp_version != version) && - peer->ibp_connecting == 1 && - peer->ibp_accepting == 0) { - retry = 1; - peer->ibp_connecting++; - - peer->ibp_version = version; - peer->ibp_incarnation = incarnation; + * initiated by kiblnd_query() + */ + reconnect = (!list_empty(&peer->ibp_tx_queue) || + peer->ibp_version != version) && + peer->ibp_connecting == 1 && + !peer->ibp_accepting; + if (!reconnect) { + reason = "no need"; + goto out; } - write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); - - if (!retry) - return; - switch (why) { default: reason = "Unknown"; break; + case IBLND_REJECT_RDMA_FRAGS: + if (!cp) { + reason = "can't negotiate max frags"; + goto out; + } + if (!*kiblnd_tunables.kib_map_on_demand) { + reason = "map_on_demand must be enabled"; + goto out; + } + if (conn->ibc_max_frags <= frag_num) { + reason = "unsupported max frags"; + goto out; + } + + peer->ibp_max_frags = frag_num; + reason = "rdma fragments"; + break; + + case IBLND_REJECT_MSG_QUEUE_SIZE: + if (!cp) { + reason = "can't negotiate queue depth"; + goto out; + } + if (conn->ibc_queue_depth <= queue_dep) { + reason = "unsupported queue depth"; + goto out; + } + + peer->ibp_queue_depth = queue_dep; + reason = "queue depth"; + break; + case IBLND_REJECT_CONN_STALE: reason = "stale"; break; @@ -2463,14 +2625,24 @@ kiblnd_reconnect(kib_conn_t *conn, int version, break; } - CNETERR("%s: retrying (%s), %x, %x, queue_dep: %d, max_frag: %d, msg_size: %d\n", - libcfs_nid2str(peer->ibp_nid), - reason, IBLND_MSG_VERSION, version, - cp != NULL ? cp->ibcp_queue_depth : IBLND_MSG_QUEUE_SIZE(version), - cp != NULL ? cp->ibcp_max_frags : IBLND_RDMA_FRAGS(version), - cp != NULL ? cp->ibcp_max_msg_size : IBLND_MSG_SIZE); + conn->ibc_reconnect = 1; + peer->ibp_reconnecting = 1; + peer->ibp_version = version; + if (incarnation) + peer->ibp_incarnation = incarnation; +out: + write_unlock_irqrestore(glock, flags); - kiblnd_connect_peer(peer); + CNETERR("%s: %s (%s), %x, %x, msg_size: %d, queue_depth: %d/%d, max_frags: %d/%d\n", + libcfs_nid2str(peer->ibp_nid), + reconnect ? "reconnect" : "don't reconnect", + reason, IBLND_MSG_VERSION, version, msg_size, + conn->ibc_queue_depth, queue_dep, + conn->ibc_max_frags, frag_num); + /** + * if conn::ibc_reconnect is TRUE, connd will reconnect to the peer + * while destroying the zombie + */ } static void @@ -2483,8 +2655,8 @@ kiblnd_rejected(kib_conn_t *conn, int reason, void *priv, int priv_nob) switch (reason) { case IB_CM_REJ_STALE_CONN: - kiblnd_reconnect(conn, IBLND_MSG_VERSION, 0, - IBLND_REJECT_CONN_STALE, NULL); + kiblnd_check_reconnect(conn, IBLND_MSG_VERSION, 0, + IBLND_REJECT_CONN_STALE, NULL); break; case IB_CM_REJ_INVALID_SERVICE_ID: @@ -2521,9 +2693,11 @@ kiblnd_rejected(kib_conn_t *conn, int reason, void *priv, int priv_nob) if (priv_nob >= sizeof(kib_rej_t) && rej->ibr_version > IBLND_MSG_VERSION_1) { - /* priv_nob is always 148 in current version + /* + * priv_nob is always 148 in current version * of OFED, so we still need to check version. - * (define of IB_CM_REJ_PRIVATE_DATA_SIZE) */ + * (define of IB_CM_REJ_PRIVATE_DATA_SIZE) + */ cp = &rej->ibr_cp; if (flip) { @@ -2564,24 +2738,11 @@ kiblnd_rejected(kib_conn_t *conn, int reason, void *priv, int priv_nob) case IBLND_REJECT_CONN_RACE: case IBLND_REJECT_CONN_STALE: case IBLND_REJECT_CONN_UNCOMPAT: - kiblnd_reconnect(conn, rej->ibr_version, - incarnation, rej->ibr_why, cp); - break; - case IBLND_REJECT_MSG_QUEUE_SIZE: - CERROR("%s rejected: incompatible message queue depth %d, %d\n", - libcfs_nid2str(peer->ibp_nid), - cp != NULL ? cp->ibcp_queue_depth : - IBLND_MSG_QUEUE_SIZE(rej->ibr_version), - IBLND_MSG_QUEUE_SIZE(conn->ibc_version)); - break; - case IBLND_REJECT_RDMA_FRAGS: - CERROR("%s rejected: incompatible # of RDMA fragments %d, %d\n", - libcfs_nid2str(peer->ibp_nid), - cp != NULL ? cp->ibcp_max_frags : - IBLND_RDMA_FRAGS(rej->ibr_version), - IBLND_RDMA_FRAGS(conn->ibc_version)); + kiblnd_check_reconnect(conn, rej->ibr_version, + incarnation, + rej->ibr_why, cp); break; case IBLND_REJECT_NO_RESOURCES: @@ -2623,9 +2784,9 @@ kiblnd_check_connreply(kib_conn_t *conn, void *priv, int priv_nob) int rc = kiblnd_unpack_msg(msg, priv_nob); unsigned long flags; - LASSERT(net != NULL); + LASSERT(net); - if (rc != 0) { + if (rc) { CERROR("Can't unpack connack from %s: %d\n", libcfs_nid2str(peer->ibp_nid), rc); goto failed; @@ -2645,22 +2806,22 @@ kiblnd_check_connreply(kib_conn_t *conn, void *priv, int priv_nob) goto failed; } - if (msg->ibm_u.connparams.ibcp_queue_depth != - IBLND_MSG_QUEUE_SIZE(ver)) { - CERROR("%s has incompatible queue depth %d(%d wanted)\n", + if (msg->ibm_u.connparams.ibcp_queue_depth > + conn->ibc_queue_depth) { + CERROR("%s has incompatible queue depth %d (<=%d wanted)\n", libcfs_nid2str(peer->ibp_nid), msg->ibm_u.connparams.ibcp_queue_depth, - IBLND_MSG_QUEUE_SIZE(ver)); + conn->ibc_queue_depth); rc = -EPROTO; goto failed; } - if (msg->ibm_u.connparams.ibcp_max_frags != - IBLND_RDMA_FRAGS(ver)) { - CERROR("%s has incompatible max_frags %d (%d wanted)\n", + if (msg->ibm_u.connparams.ibcp_max_frags > + conn->ibc_max_frags) { + CERROR("%s has incompatible max_frags %d (<=%d wanted)\n", libcfs_nid2str(peer->ibp_nid), msg->ibm_u.connparams.ibcp_max_frags, - IBLND_RDMA_FRAGS(ver)); + conn->ibc_max_frags); rc = -EPROTO; goto failed; } @@ -2682,7 +2843,7 @@ kiblnd_check_connreply(kib_conn_t *conn, void *priv, int priv_nob) rc = -ESTALE; read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); - if (rc != 0) { + if (rc) { CERROR("Bad connection reply from %s, rc = %d, version: %x max_frags: %d\n", libcfs_nid2str(peer->ibp_nid), rc, msg->ibm_version, msg->ibm_u.connparams.ibcp_max_frags); @@ -2690,21 +2851,24 @@ kiblnd_check_connreply(kib_conn_t *conn, void *priv, int priv_nob) } conn->ibc_incarnation = msg->ibm_srcstamp; - conn->ibc_credits = - conn->ibc_reserved_credits = IBLND_MSG_QUEUE_SIZE(ver); - LASSERT(conn->ibc_credits + conn->ibc_reserved_credits + IBLND_OOB_MSGS(ver) - <= IBLND_RX_MSGS(ver)); + conn->ibc_credits = msg->ibm_u.connparams.ibcp_queue_depth; + conn->ibc_reserved_credits = msg->ibm_u.connparams.ibcp_queue_depth; + conn->ibc_queue_depth = msg->ibm_u.connparams.ibcp_queue_depth; + conn->ibc_max_frags = msg->ibm_u.connparams.ibcp_max_frags; + LASSERT(conn->ibc_credits + conn->ibc_reserved_credits + + IBLND_OOB_MSGS(ver) <= IBLND_RX_MSGS(conn)); kiblnd_connreq_done(conn, 0); return; failed: - /* NB My QP has already established itself, so I handle anything going + /* + * NB My QP has already established itself, so I handle anything going * wrong here by setting ibc_comms_error. * kiblnd_connreq_done(0) moves the conn state to ESTABLISHED, but then - * immediately tears it down. */ - - LASSERT(rc != 0); + * immediately tears it down. + */ + LASSERT(rc); conn->ibc_comms_error = rc; kiblnd_connreq_done(conn, 0); } @@ -2724,28 +2888,30 @@ kiblnd_active_connect(struct rdma_cm_id *cmid) read_lock_irqsave(&kiblnd_data.kib_global_lock, flags); incarnation = peer->ibp_incarnation; - version = (peer->ibp_version == 0) ? IBLND_MSG_VERSION : - peer->ibp_version; + version = !peer->ibp_version ? IBLND_MSG_VERSION : + peer->ibp_version; read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); - conn = kiblnd_create_conn(peer, cmid, IBLND_CONN_ACTIVE_CONNECT, version); - if (conn == NULL) { + conn = kiblnd_create_conn(peer, cmid, IBLND_CONN_ACTIVE_CONNECT, + version); + if (!conn) { kiblnd_peer_connect_failed(peer, 1, -ENOMEM); kiblnd_peer_decref(peer); /* lose cmid's ref */ return -ENOMEM; } - /* conn "owns" cmid now, so I return success from here on to ensure the + /* + * conn "owns" cmid now, so I return success from here on to ensure the * CM callback doesn't destroy cmid. conn also takes over cmid's ref - * on peer */ - + * on peer + */ msg = &conn->ibc_connvars->cv_msg; memset(msg, 0, sizeof(*msg)); kiblnd_init_msg(msg, IBLND_MSG_CONNREQ, sizeof(msg->ibm_u.connparams)); - msg->ibm_u.connparams.ibcp_queue_depth = IBLND_MSG_QUEUE_SIZE(version); - msg->ibm_u.connparams.ibcp_max_frags = IBLND_RDMA_FRAGS(version); + msg->ibm_u.connparams.ibcp_queue_depth = conn->ibc_queue_depth; + msg->ibm_u.connparams.ibcp_max_frags = conn->ibc_max_frags; msg->ibm_u.connparams.ibcp_max_msg_size = IBLND_MSG_SIZE; kiblnd_pack_msg(peer->ibp_ni, msg, version, @@ -2764,7 +2930,7 @@ kiblnd_active_connect(struct rdma_cm_id *cmid) LASSERT(conn->ibc_cmid == cmid); rc = rdma_connect(cmid, &cp); - if (rc != 0) { + if (rc) { CERROR("Can't connect to %s: %d\n", libcfs_nid2str(peer->ibp_nid), rc); kiblnd_connreq_done(conn, rc); @@ -2798,10 +2964,10 @@ kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event) case RDMA_CM_EVENT_ADDR_ERROR: peer = (kib_peer_t *)cmid->context; CNETERR("%s: ADDR ERROR %d\n", - libcfs_nid2str(peer->ibp_nid), event->status); + libcfs_nid2str(peer->ibp_nid), event->status); kiblnd_peer_connect_failed(peer, 1, -EHOSTUNREACH); kiblnd_peer_decref(peer); - return -EHOSTUNREACH; /* rc != 0 destroys cmid */ + return -EHOSTUNREACH; /* rc destroys cmid */ case RDMA_CM_EVENT_ADDR_RESOLVED: peer = (kib_peer_t *)cmid->context; @@ -2809,14 +2975,14 @@ kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event) CDEBUG(D_NET, "%s Addr resolved: %d\n", libcfs_nid2str(peer->ibp_nid), event->status); - if (event->status != 0) { + if (event->status) { CNETERR("Can't resolve address for %s: %d\n", libcfs_nid2str(peer->ibp_nid), event->status); rc = event->status; } else { rc = rdma_resolve_route( cmid, *kiblnd_tunables.kib_timeout * 1000); - if (rc == 0) + if (!rc) return 0; /* Can't initiate route resolution */ CERROR("Can't resolve route for %s: %d\n", @@ -2824,7 +2990,7 @@ kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event) } kiblnd_peer_connect_failed(peer, 1, rc); kiblnd_peer_decref(peer); - return rc; /* rc != 0 destroys cmid */ + return rc; /* rc destroys cmid */ case RDMA_CM_EVENT_ROUTE_ERROR: peer = (kib_peer_t *)cmid->context; @@ -2832,28 +2998,28 @@ kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event) libcfs_nid2str(peer->ibp_nid), event->status); kiblnd_peer_connect_failed(peer, 1, -EHOSTUNREACH); kiblnd_peer_decref(peer); - return -EHOSTUNREACH; /* rc != 0 destroys cmid */ + return -EHOSTUNREACH; /* rc destroys cmid */ case RDMA_CM_EVENT_ROUTE_RESOLVED: peer = (kib_peer_t *)cmid->context; CDEBUG(D_NET, "%s Route resolved: %d\n", libcfs_nid2str(peer->ibp_nid), event->status); - if (event->status == 0) + if (!event->status) return kiblnd_active_connect(cmid); CNETERR("Can't resolve route for %s: %d\n", - libcfs_nid2str(peer->ibp_nid), event->status); + libcfs_nid2str(peer->ibp_nid), event->status); kiblnd_peer_connect_failed(peer, 1, event->status); kiblnd_peer_decref(peer); - return event->status; /* rc != 0 destroys cmid */ + return event->status; /* rc destroys cmid */ case RDMA_CM_EVENT_UNREACHABLE: conn = (kib_conn_t *)cmid->context; LASSERT(conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT || conn->ibc_state == IBLND_CONN_PASSIVE_WAIT); CNETERR("%s: UNREACHABLE %d\n", - libcfs_nid2str(conn->ibc_peer->ibp_nid), event->status); + libcfs_nid2str(conn->ibc_peer->ibp_nid), event->status); kiblnd_connreq_done(conn, -ENETDOWN); kiblnd_conn_decref(conn); return 0; @@ -2876,8 +3042,8 @@ kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event) case IBLND_CONN_PASSIVE_WAIT: CERROR("%s: REJECTED %d\n", - libcfs_nid2str(conn->ibc_peer->ibp_nid), - event->status); + libcfs_nid2str(conn->ibc_peer->ibp_nid), + event->status); kiblnd_connreq_done(conn, -ECONNRESET); break; @@ -2933,8 +3099,10 @@ kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event) LCONSOLE_ERROR_MSG(0x131, "Received notification of device removal\n" "Please shutdown LNET to allow this to proceed\n"); - /* Can't remove network from underneath LNET for now, so I have - * to ignore this */ + /* + * Can't remove network from underneath LNET for now, so I have + * to ignore this + */ return 0; case RDMA_CM_EVENT_ADDR_CHANGE: @@ -2956,7 +3124,7 @@ kiblnd_check_txs_locked(kib_conn_t *conn, struct list_head *txs) LASSERT(tx->tx_queued); } else { LASSERT(!tx->tx_queued); - LASSERT(tx->tx_waiting || tx->tx_sending != 0); + LASSERT(tx->tx_waiting || tx->tx_sending); } if (cfs_time_aftereq(jiffies, tx->tx_deadline)) { @@ -2989,13 +3157,16 @@ kiblnd_check_conns(int idx) struct list_head *ptmp; kib_peer_t *peer; kib_conn_t *conn; + kib_conn_t *temp; kib_conn_t *tmp; struct list_head *ctmp; unsigned long flags; - /* NB. We expect to have a look at all the peers and not find any + /* + * NB. We expect to have a look at all the peers and not find any * RDMAs to time out, so we just use a shared lock while we - * take a look... */ + * take a look... + */ read_lock_irqsave(&kiblnd_data.kib_global_lock, flags); list_for_each(ptmp, peers) { @@ -3028,8 +3199,7 @@ kiblnd_check_conns(int idx) conn->ibc_reserved_credits); list_add(&conn->ibc_connd_list, &closes); } else { - list_add(&conn->ibc_connd_list, - &checksends); + list_add(&conn->ibc_connd_list, &checksends); } /* +ref for 'closes' or 'checksends' */ kiblnd_conn_addref(conn); @@ -3040,21 +3210,23 @@ kiblnd_check_conns(int idx) read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); - /* Handle timeout by closing the whole + /* + * Handle timeout by closing the whole * connection. We can only be sure RDMA activity - * has ceased once the QP has been modified. */ + * has ceased once the QP has been modified. + */ list_for_each_entry_safe(conn, tmp, &closes, ibc_connd_list) { list_del(&conn->ibc_connd_list); kiblnd_close_conn(conn, -ETIMEDOUT); kiblnd_conn_decref(conn); } - /* In case we have enough credits to return via a + /* + * In case we have enough credits to return via a * NOOP, but there were no non-blocking tx descs - * free to do it last time... */ - while (!list_empty(&checksends)) { - conn = list_entry(checksends.next, - kib_conn_t, ibc_connd_list); + * free to do it last time... + */ + list_for_each_entry_safe(conn, temp, &checksends, ibc_connd_list) { list_del(&conn->ibc_connd_list); kiblnd_check_sends(conn); kiblnd_conn_decref(conn); @@ -3074,9 +3246,21 @@ kiblnd_disconnect_conn(kib_conn_t *conn) kiblnd_peer_notify(conn->ibc_peer); } +/** + * High-water for reconnection to the same peer, reconnection attempt should + * be delayed after trying more than KIB_RECONN_HIGH_RACE. + */ +#define KIB_RECONN_HIGH_RACE 10 +/** + * Allow connd to take a break and handle other things after consecutive + * reconnection attemps. + */ +#define KIB_RECONN_BREAK 100 + int kiblnd_connd(void *arg) { + spinlock_t *lock= &kiblnd_data.kib_connd_lock; wait_queue_t wait; unsigned long flags; kib_conn_t *conn; @@ -3091,39 +3275,79 @@ kiblnd_connd(void *arg) init_waitqueue_entry(&wait, current); kiblnd_data.kib_connd = current; - spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); + spin_lock_irqsave(lock, flags); while (!kiblnd_data.kib_shutdown) { + int reconn = 0; dropped_lock = 0; if (!list_empty(&kiblnd_data.kib_connd_zombies)) { + kib_peer_t *peer = NULL; + conn = list_entry(kiblnd_data.kib_connd_zombies.next, - kib_conn_t, ibc_list); + kib_conn_t, ibc_list); list_del(&conn->ibc_list); + if (conn->ibc_reconnect) { + peer = conn->ibc_peer; + kiblnd_peer_addref(peer); + } - spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, - flags); + spin_unlock_irqrestore(lock, flags); dropped_lock = 1; - kiblnd_destroy_conn(conn); + kiblnd_destroy_conn(conn, !peer); + + spin_lock_irqsave(lock, flags); + if (!peer) + continue; - spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); + conn->ibc_peer = peer; + if (peer->ibp_reconnected < KIB_RECONN_HIGH_RACE) + list_add_tail(&conn->ibc_list, + &kiblnd_data.kib_reconn_list); + else + list_add_tail(&conn->ibc_list, + &kiblnd_data.kib_reconn_wait); } if (!list_empty(&kiblnd_data.kib_connd_conns)) { conn = list_entry(kiblnd_data.kib_connd_conns.next, - kib_conn_t, ibc_list); + kib_conn_t, ibc_list); list_del(&conn->ibc_list); - spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, - flags); + spin_unlock_irqrestore(lock, flags); dropped_lock = 1; kiblnd_disconnect_conn(conn); kiblnd_conn_decref(conn); - spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); + spin_lock_irqsave(lock, flags); + } + + while (reconn < KIB_RECONN_BREAK) { + if (kiblnd_data.kib_reconn_sec != + ktime_get_real_seconds()) { + kiblnd_data.kib_reconn_sec = ktime_get_real_seconds(); + list_splice_init(&kiblnd_data.kib_reconn_wait, + &kiblnd_data.kib_reconn_list); + } + + if (list_empty(&kiblnd_data.kib_reconn_list)) + break; + + conn = list_entry(kiblnd_data.kib_reconn_list.next, + kib_conn_t, ibc_list); + list_del(&conn->ibc_list); + + spin_unlock_irqrestore(lock, flags); + dropped_lock = 1; + + reconn += kiblnd_reconnect_peer(conn->ibc_peer); + kiblnd_peer_decref(conn->ibc_peer); + LIBCFS_FREE(conn, sizeof(*conn)); + + spin_lock_irqsave(lock, flags); } /* careful with the jiffy wrap... */ @@ -3133,21 +3357,22 @@ kiblnd_connd(void *arg) const int p = 1; int chunk = kiblnd_data.kib_peer_hash_size; - spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, flags); + spin_unlock_irqrestore(lock, flags); dropped_lock = 1; - /* Time to check for RDMA timeouts on a few more + /* + * Time to check for RDMA timeouts on a few more * peers: I do checks every 'p' seconds on a * proportion of the peer table and I need to check * every connection 'n' times within a timeout * interval, to ensure I detect a timeout on any * connection within (n+1)/n times the timeout - * interval. */ - + * interval. + */ if (*kiblnd_tunables.kib_timeout > n * p) chunk = (chunk * n * p) / *kiblnd_tunables.kib_timeout; - if (chunk == 0) + if (!chunk) chunk = 1; for (i = 0; i < chunk; i++) { @@ -3156,8 +3381,8 @@ kiblnd_connd(void *arg) kiblnd_data.kib_peer_hash_size; } - deadline += p * HZ; - spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); + deadline += msecs_to_jiffies(p * MSEC_PER_SEC); + spin_lock_irqsave(lock, flags); } if (dropped_lock) @@ -3166,15 +3391,15 @@ kiblnd_connd(void *arg) /* Nothing to do for 'timeout' */ set_current_state(TASK_INTERRUPTIBLE); add_wait_queue(&kiblnd_data.kib_connd_waitq, &wait); - spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, flags); + spin_unlock_irqrestore(lock, flags); schedule_timeout(timeout); remove_wait_queue(&kiblnd_data.kib_connd_waitq, &wait); - spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); + spin_lock_irqsave(lock, flags); } - spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, flags); + spin_unlock_irqrestore(lock, flags); kiblnd_thread_fini(); return 0; @@ -3206,12 +3431,14 @@ kiblnd_complete(struct ib_wc *wc) LBUG(); case IBLND_WID_RDMA: - /* We only get RDMA completion notification if it fails. All + /* + * We only get RDMA completion notification if it fails. All * subsequent work items, including the final SEND will fail * too. However we can't print out any more info about the * failing RDMA because 'tx' might be back on the idle list or * even reused already if we didn't manage to post all our work - * items */ + * items + */ CNETERR("RDMA (tx: %p) failed: %d\n", kiblnd_wreqid2ptr(wc->wr_id), wc->status); return; @@ -3230,11 +3457,13 @@ kiblnd_complete(struct ib_wc *wc) void kiblnd_cq_completion(struct ib_cq *cq, void *arg) { - /* NB I'm not allowed to schedule this conn once its refcount has + /* + * NB I'm not allowed to schedule this conn once its refcount has * reached 0. Since fundamentally I'm racing with scheduler threads * consuming my CQ I could be called after all completions have - * occurred. But in this case, ibc_nrx == 0 && ibc_nsends_posted == 0 - * and this CQ is about to be destroyed so I NOOP. */ + * occurred. But in this case, !ibc_nrx && !ibc_nsends_posted + * and this CQ is about to be destroyed so I NOOP. + */ kib_conn_t *conn = arg; struct kib_sched_info *sched = conn->ibc_sched; unsigned long flags; @@ -3288,7 +3517,7 @@ kiblnd_scheduler(void *arg) sched = kiblnd_data.kib_scheds[KIB_THREAD_CPT(id)]; rc = cfs_cpt_bind(lnet_cpt_table(), sched->ibs_cpt); - if (rc != 0) { + if (rc) { CWARN("Failed to bind on CPT %d, please verify whether all CPUs are healthy and reload modules if necessary, otherwise your system might under risk of low performance\n", sched->ibs_cpt); } @@ -3308,8 +3537,8 @@ kiblnd_scheduler(void *arg) did_something = 0; if (!list_empty(&sched->ibs_conns)) { - conn = list_entry(sched->ibs_conns.next, - kib_conn_t, ibc_sched_list); + conn = list_entry(sched->ibs_conns.next, kib_conn_t, + ibc_sched_list); /* take over kib_sched_conns' ref on conn... */ LASSERT(conn->ibc_scheduled); list_del(&conn->ibc_sched_list); @@ -3317,8 +3546,10 @@ kiblnd_scheduler(void *arg) spin_unlock_irqrestore(&sched->ibs_lock, flags); + wc.wr_id = IBLND_WID_INVAL; + rc = ib_poll_cq(conn->ibc_cq, 1, &wc); - if (rc == 0) { + if (!rc) { rc = ib_req_notify_cq(conn->ibc_cq, IB_CQ_NEXT_COMP); if (rc < 0) { @@ -3327,13 +3558,22 @@ kiblnd_scheduler(void *arg) kiblnd_close_conn(conn, -EIO); kiblnd_conn_decref(conn); spin_lock_irqsave(&sched->ibs_lock, - flags); + flags); continue; } rc = ib_poll_cq(conn->ibc_cq, 1, &wc); } + if (unlikely(rc > 0 && wc.wr_id == IBLND_WID_INVAL)) { + LCONSOLE_ERROR("ib_poll_cq (rc: %d) returned invalid wr_id, opcode %d, status: %d, vendor_err: %d, conn: %s status: %d\nplease upgrade firmware and OFED or contact vendor.\n", + rc, wc.opcode, wc.status, + wc.vendor_err, + libcfs_nid2str(conn->ibc_peer->ibp_nid), + conn->ibc_state); + rc = -EINVAL; + } + if (rc < 0) { CWARN("%s: ib_poll_cq failed: %d, closing connection\n", libcfs_nid2str(conn->ibc_peer->ibp_nid), @@ -3346,21 +3586,23 @@ kiblnd_scheduler(void *arg) spin_lock_irqsave(&sched->ibs_lock, flags); - if (rc != 0 || conn->ibc_ready) { - /* There may be another completion waiting; get + if (rc || conn->ibc_ready) { + /* + * There may be another completion waiting; get * another scheduler to check while I handle - * this one... */ + * this one... + */ /* +1 ref for sched_conns */ kiblnd_conn_addref(conn); list_add_tail(&conn->ibc_sched_list, - &sched->ibs_conns); + &sched->ibs_conns); if (waitqueue_active(&sched->ibs_waitq)) wake_up(&sched->ibs_waitq); } else { conn->ibc_scheduled = 0; } - if (rc != 0) { + if (rc) { spin_unlock_irqrestore(&sched->ibs_lock, flags); kiblnd_complete(&wc); @@ -3400,7 +3642,7 @@ kiblnd_failover_thread(void *arg) unsigned long flags; int rc; - LASSERT(*kiblnd_tunables.kib_dev_failover != 0); + LASSERT(*kiblnd_tunables.kib_dev_failover); cfs_block_allsigs(); @@ -3459,13 +3701,15 @@ kiblnd_failover_thread(void *arg) remove_wait_queue(&kiblnd_data.kib_failover_waitq, &wait); write_lock_irqsave(glock, flags); - if (!long_sleep || rc != 0) + if (!long_sleep || rc) continue; - /* have a long sleep, routine check all active devices, + /* + * have a long sleep, routine check all active devices, * we need checking like this because if there is not active * connection on the dev and no SEND from local, we may listen - * on wrong HCA for ever while there is a bonding failover */ + * on wrong HCA for ever while there is a bonding failover + */ list_for_each_entry(dev, &kiblnd_data.kib_devs, ibd_list) { if (kiblnd_dev_can_failover(dev)) { list_add_tail(&dev->ibd_fail_list, diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_modparams.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_modparams.c index 1d4e7efb5..b4607dad3 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_modparams.c +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_modparams.c @@ -52,8 +52,10 @@ static int timeout = 50; module_param(timeout, int, 0644); MODULE_PARM_DESC(timeout, "timeout (seconds)"); -/* Number of threads in each scheduler pool which is percpt, - * we will estimate reasonable value based on CPUs if it's set to zero. */ +/* + * Number of threads in each scheduler pool which is percpt, + * we will estimate reasonable value based on CPUs if it's set to zero. + */ static int nscheds; module_param(nscheds, int, 0444); MODULE_PARM_DESC(nscheds, "number of threads in each scheduler pool"); @@ -200,7 +202,7 @@ kiblnd_tunables_init(void) if (*kiblnd_tunables.kib_map_on_demand == 1) *kiblnd_tunables.kib_map_on_demand = 2; /* don't make sense to create map if only one fragment */ - if (*kiblnd_tunables.kib_concurrent_sends == 0) { + if (!*kiblnd_tunables.kib_concurrent_sends) { if (*kiblnd_tunables.kib_map_on_demand > 0 && *kiblnd_tunables.kib_map_on_demand <= IBLND_MAX_RDMA_FRAGS / 8) *kiblnd_tunables.kib_concurrent_sends = (*kiblnd_tunables.kib_peertxcredits) * 2; diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c index 05aa90ea5..cca7b2f7f 100644 --- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c +++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c @@ -70,7 +70,7 @@ ksocknal_create_route(__u32 ipaddr, int port) ksock_route_t *route; LIBCFS_ALLOC(route, sizeof(*route)); - if (route == NULL) + if (!route) return NULL; atomic_set(&route->ksnr_refcount, 1); @@ -91,9 +91,9 @@ ksocknal_create_route(__u32 ipaddr, int port) void ksocknal_destroy_route(ksock_route_t *route) { - LASSERT(atomic_read(&route->ksnr_refcount) == 0); + LASSERT(!atomic_read(&route->ksnr_refcount)); - if (route->ksnr_peer != NULL) + if (route->ksnr_peer) ksocknal_peer_decref(route->ksnr_peer); LIBCFS_FREE(route, sizeof(*route)); @@ -102,6 +102,7 @@ ksocknal_destroy_route(ksock_route_t *route) static int ksocknal_create_peer(ksock_peer_t **peerp, lnet_ni_t *ni, lnet_process_id_t id) { + int cpt = lnet_cpt_of_nid(id.nid); ksock_net_t *net = ni->ni_data; ksock_peer_t *peer; @@ -109,8 +110,8 @@ ksocknal_create_peer(ksock_peer_t **peerp, lnet_ni_t *ni, lnet_process_id_t id) LASSERT(id.pid != LNET_PID_ANY); LASSERT(!in_interrupt()); - LIBCFS_ALLOC(peer, sizeof(*peer)); - if (peer == NULL) + LIBCFS_CPT_ALLOC(peer, lnet_cpt_table(), cpt, sizeof(*peer)); + if (!peer) return -ENOMEM; peer->ksnp_ni = ni; @@ -152,10 +153,10 @@ ksocknal_destroy_peer(ksock_peer_t *peer) ksock_net_t *net = peer->ksnp_ni->ni_data; CDEBUG(D_NET, "peer %s %p deleted\n", - libcfs_id2str(peer->ksnp_id), peer); + libcfs_id2str(peer->ksnp_id), peer); - LASSERT(atomic_read(&peer->ksnp_refcount) == 0); - LASSERT(peer->ksnp_accepting == 0); + LASSERT(!atomic_read(&peer->ksnp_refcount)); + LASSERT(!peer->ksnp_accepting); LASSERT(list_empty(&peer->ksnp_conns)); LASSERT(list_empty(&peer->ksnp_routes)); LASSERT(list_empty(&peer->ksnp_tx_queue)); @@ -163,10 +164,12 @@ ksocknal_destroy_peer(ksock_peer_t *peer) LIBCFS_FREE(peer, sizeof(*peer)); - /* NB a peer's connections and routes keep a reference on their peer + /* + * NB a peer's connections and routes keep a reference on their peer * until they are destroyed, so we can be assured that _all_ state to * do with this peer has been cleaned up when its refcount drops to - * zero. */ + * zero. + */ spin_lock_bh(&net->ksnn_lock); net->ksnn_npeers--; spin_unlock_bh(&net->ksnn_lock); @@ -180,7 +183,6 @@ ksocknal_find_peer_locked(lnet_ni_t *ni, lnet_process_id_t id) ksock_peer_t *peer; list_for_each(tmp, peer_list) { - peer = list_entry(tmp, ksock_peer_t, ksnp_list); LASSERT(!peer->ksnp_closing); @@ -207,7 +209,7 @@ ksocknal_find_peer(lnet_ni_t *ni, lnet_process_id_t id) read_lock(&ksocknal_data.ksnd_global_lock); peer = ksocknal_find_peer_locked(ni, id); - if (peer != NULL) /* +1 ref for caller? */ + if (peer) /* +1 ref for caller? */ ksocknal_peer_addref(peer); read_unlock(&ksocknal_data.ksnd_global_lock); @@ -226,9 +228,11 @@ ksocknal_unlink_peer_locked(ksock_peer_t *peer) ip = peer->ksnp_passive_ips[i]; iface = ksocknal_ip2iface(peer->ksnp_ni, ip); - /* All IPs in peer->ksnp_passive_ips[] come from the - * interface list, therefore the call must succeed. */ - LASSERT(iface != NULL); + /* + * All IPs in peer->ksnp_passive_ips[] come from the + * interface list, therefore the call must succeed. + */ + LASSERT(iface); CDEBUG(D_NET, "peer=%p iface=%p ksni_nroutes=%d\n", peer, iface, iface->ksni_nroutes); @@ -246,8 +250,8 @@ ksocknal_unlink_peer_locked(ksock_peer_t *peer) static int ksocknal_get_peer_info(lnet_ni_t *ni, int index, - lnet_process_id_t *id, __u32 *myip, __u32 *peer_ip, - int *port, int *conn_count, int *share_count) + lnet_process_id_t *id, __u32 *myip, __u32 *peer_ip, + int *port, int *conn_count, int *share_count) { ksock_peer_t *peer; struct list_head *ptmp; @@ -260,14 +264,13 @@ ksocknal_get_peer_info(lnet_ni_t *ni, int index, read_lock(&ksocknal_data.ksnd_global_lock); for (i = 0; i < ksocknal_data.ksnd_peer_hash_size; i++) { - list_for_each(ptmp, &ksocknal_data.ksnd_peers[i]) { peer = list_entry(ptmp, ksock_peer_t, ksnp_list); if (peer->ksnp_ni != ni) continue; - if (peer->ksnp_n_passive_ips == 0 && + if (!peer->ksnp_n_passive_ips && list_empty(&peer->ksnp_routes)) { if (index-- > 0) continue; @@ -301,7 +304,7 @@ ksocknal_get_peer_info(lnet_ni_t *ni, int index, continue; route = list_entry(rtmp, ksock_route_t, - ksnr_list); + ksnr_list); *id = peer->ksnp_id; *myip = route->ksnr_myipaddr; @@ -330,7 +333,7 @@ ksocknal_associate_route_conn_locked(ksock_route_t *route, ksock_conn_t *conn) ksocknal_route_addref(route); if (route->ksnr_myipaddr != conn->ksnc_myipaddr) { - if (route->ksnr_myipaddr == 0) { + if (!route->ksnr_myipaddr) { /* route wasn't bound locally yet (the initial route) */ CDEBUG(D_NET, "Binding %s %pI4h to %pI4h\n", libcfs_id2str(peer->ksnp_id), @@ -345,21 +348,23 @@ ksocknal_associate_route_conn_locked(ksock_route_t *route, ksock_conn_t *conn) iface = ksocknal_ip2iface(route->ksnr_peer->ksnp_ni, route->ksnr_myipaddr); - if (iface != NULL) + if (iface) iface->ksni_nroutes--; } route->ksnr_myipaddr = conn->ksnc_myipaddr; iface = ksocknal_ip2iface(route->ksnr_peer->ksnp_ni, route->ksnr_myipaddr); - if (iface != NULL) + if (iface) iface->ksni_nroutes++; } - route->ksnr_connected |= (1<<type); + route->ksnr_connected |= (1 << type); route->ksnr_conn_count++; - /* Successful connection => further attempts can - * proceed immediately */ + /* + * Successful connection => further attempts can + * proceed immediately + */ route->ksnr_retry_interval = 0; } @@ -371,10 +376,10 @@ ksocknal_add_route_locked(ksock_peer_t *peer, ksock_route_t *route) ksock_route_t *route2; LASSERT(!peer->ksnp_closing); - LASSERT(route->ksnr_peer == NULL); + LASSERT(!route->ksnr_peer); LASSERT(!route->ksnr_scheduled); LASSERT(!route->ksnr_connecting); - LASSERT(route->ksnr_connected == 0); + LASSERT(!route->ksnr_connected); /* LASSERT(unique) */ list_for_each(tmp, &peer->ksnp_routes) { @@ -382,8 +387,8 @@ ksocknal_add_route_locked(ksock_peer_t *peer, ksock_route_t *route) if (route2->ksnr_ipaddr == route->ksnr_ipaddr) { CERROR("Duplicate route %s %pI4h\n", - libcfs_id2str(peer->ksnp_id), - &route->ksnr_ipaddr); + libcfs_id2str(peer->ksnp_id), + &route->ksnr_ipaddr); LBUG(); } } @@ -425,10 +430,10 @@ ksocknal_del_route_locked(ksock_route_t *route) ksocknal_close_conn_locked(conn, 0); } - if (route->ksnr_myipaddr != 0) { + if (route->ksnr_myipaddr) { iface = ksocknal_ip2iface(route->ksnr_peer->ksnp_ni, route->ksnr_myipaddr); - if (iface != NULL) + if (iface) iface->ksni_nroutes--; } @@ -438,8 +443,10 @@ ksocknal_del_route_locked(ksock_route_t *route) if (list_empty(&peer->ksnp_routes) && list_empty(&peer->ksnp_conns)) { - /* I've just removed the last route to a peer with no active - * connections */ + /* + * I've just removed the last route to a peer with no active + * connections + */ ksocknal_unlink_peer_locked(peer); } } @@ -460,11 +467,11 @@ ksocknal_add_peer(lnet_ni_t *ni, lnet_process_id_t id, __u32 ipaddr, int port) /* Have a brand new peer ready... */ rc = ksocknal_create_peer(&peer, ni, id); - if (rc != 0) + if (rc) return rc; route = ksocknal_create_route(ipaddr, port); - if (route == NULL) { + if (!route) { ksocknal_peer_decref(peer); return -ENOMEM; } @@ -472,16 +479,16 @@ ksocknal_add_peer(lnet_ni_t *ni, lnet_process_id_t id, __u32 ipaddr, int port) write_lock_bh(&ksocknal_data.ksnd_global_lock); /* always called with a ref on ni, so shutdown can't have started */ - LASSERT(((ksock_net_t *) ni->ni_data)->ksnn_shutdown == 0); + LASSERT(!((ksock_net_t *) ni->ni_data)->ksnn_shutdown); peer2 = ksocknal_find_peer_locked(ni, id); - if (peer2 != NULL) { + if (peer2) { ksocknal_peer_decref(peer); peer = peer2; } else { /* peer table takes my ref on peer */ list_add_tail(&peer->ksnp_list, - ksocknal_nid2peerlist(id.nid)); + ksocknal_nid2peerlist(id.nid)); } route2 = NULL; @@ -493,7 +500,7 @@ ksocknal_add_peer(lnet_ni_t *ni, lnet_process_id_t id, __u32 ipaddr, int port) route2 = NULL; } - if (route2 == NULL) { + if (!route2) { ksocknal_add_route_locked(peer, route); route->ksnr_share_count++; } else { @@ -524,7 +531,7 @@ ksocknal_del_peer_locked(ksock_peer_t *peer, __u32 ip) route = list_entry(tmp, ksock_route_t, ksnr_list); /* no match */ - if (!(ip == 0 || route->ksnr_ipaddr == ip)) + if (!(!ip || route->ksnr_ipaddr == ip)) continue; route->ksnr_share_count = 0; @@ -538,15 +545,16 @@ ksocknal_del_peer_locked(ksock_peer_t *peer, __u32 ip) nshared += route->ksnr_share_count; } - if (nshared == 0) { - /* remove everything else if there are no explicit entries - * left */ - + if (!nshared) { + /* + * remove everything else if there are no explicit entries + * left + */ list_for_each_safe(tmp, nxt, &peer->ksnp_routes) { route = list_entry(tmp, ksock_route_t, ksnr_list); /* we should only be removing auto-entries */ - LASSERT(route->ksnr_share_count == 0); + LASSERT(!route->ksnr_share_count); ksocknal_del_route_locked(route); } @@ -575,16 +583,16 @@ ksocknal_del_peer(lnet_ni_t *ni, lnet_process_id_t id, __u32 ip) write_lock_bh(&ksocknal_data.ksnd_global_lock); - if (id.nid != LNET_NID_ANY) - lo = hi = (int)(ksocknal_nid2peerlist(id.nid) - ksocknal_data.ksnd_peers); - else { + if (id.nid != LNET_NID_ANY) { + lo = (int)(ksocknal_nid2peerlist(id.nid) - ksocknal_data.ksnd_peers); + hi = (int)(ksocknal_nid2peerlist(id.nid) - ksocknal_data.ksnd_peers); + } else { lo = 0; hi = ksocknal_data.ksnd_peer_hash_size - 1; } for (i = lo; i <= hi; i++) { - list_for_each_safe(ptmp, pnxt, - &ksocknal_data.ksnd_peers[i]) { + list_for_each_safe(ptmp, pnxt, &ksocknal_data.ksnd_peers[i]) { peer = list_entry(ptmp, ksock_peer_t, ksnp_list); if (peer->ksnp_ni != ni) @@ -604,7 +612,7 @@ ksocknal_del_peer(lnet_ni_t *ni, lnet_process_id_t id, __u32 ip) LASSERT(list_empty(&peer->ksnp_routes)); list_splice_init(&peer->ksnp_tx_queue, - &zombies); + &zombies); } ksocknal_peer_decref(peer); /* ...till here */ @@ -645,7 +653,7 @@ ksocknal_get_conn_by_idx(lnet_ni_t *ni, int index) continue; conn = list_entry(ctmp, ksock_conn_t, - ksnc_list); + ksnc_list); ksocknal_conn_addref(conn); read_unlock(&ksocknal_data.ksnd_global_lock); return conn; @@ -692,8 +700,10 @@ ksocknal_local_ipvec(lnet_ni_t *ni, __u32 *ipaddrs) nip = net->ksnn_ninterfaces; LASSERT(nip <= LNET_MAX_INTERFACES); - /* Only offer interfaces for additional connections if I have - * more than one. */ + /* + * Only offer interfaces for additional connections if I have + * more than one. + */ if (nip < 2) { read_unlock(&ksocknal_data.ksnd_global_lock); return 0; @@ -701,7 +711,7 @@ ksocknal_local_ipvec(lnet_ni_t *ni, __u32 *ipaddrs) for (i = 0; i < nip; i++) { ipaddrs[i] = net->ksnn_interfaces[i].ksni_ipaddr; - LASSERT(ipaddrs[i] != 0); + LASSERT(ipaddrs[i]); } read_unlock(&ksocknal_data.ksnd_global_lock); @@ -719,11 +729,11 @@ ksocknal_match_peerip(ksock_interface_t *iface, __u32 *ips, int nips) int i; for (i = 0; i < nips; i++) { - if (ips[i] == 0) + if (!ips[i]) continue; this_xor = ips[i] ^ iface->ksni_ipaddr; - this_netmatch = ((this_xor & iface->ksni_netmask) == 0) ? 1 : 0; + this_netmatch = !(this_xor & iface->ksni_netmask) ? 1 : 0; if (!(best < 0 || best_netmatch < this_netmatch || @@ -757,38 +767,45 @@ ksocknal_select_ips(ksock_peer_t *peer, __u32 *peerips, int n_peerips) int best_netmatch; int best_npeers; - /* CAVEAT EMPTOR: We do all our interface matching with an + /* + * CAVEAT EMPTOR: We do all our interface matching with an * exclusive hold of global lock at IRQ priority. We're only * expecting to be dealing with small numbers of interfaces, so the - * O(n**3)-ness shouldn't matter */ - - /* Also note that I'm not going to return more than n_peerips - * interfaces, even if I have more myself */ - + * O(n**3)-ness shouldn't matter + */ + /* + * Also note that I'm not going to return more than n_peerips + * interfaces, even if I have more myself + */ write_lock_bh(global_lock); LASSERT(n_peerips <= LNET_MAX_INTERFACES); LASSERT(net->ksnn_ninterfaces <= LNET_MAX_INTERFACES); - /* Only match interfaces for additional connections - * if I have > 1 interface */ + /* + * Only match interfaces for additional connections + * if I have > 1 interface + */ n_ips = (net->ksnn_ninterfaces < 2) ? 0 : min(n_peerips, net->ksnn_ninterfaces); for (i = 0; peer->ksnp_n_passive_ips < n_ips; i++) { /* ^ yes really... */ - /* If we have any new interfaces, first tick off all the + /* + * If we have any new interfaces, first tick off all the * peer IPs that match old interfaces, then choose new * interfaces to match the remaining peer IPS. * We don't forget interfaces we've stopped using; we might - * start using them again... */ - + * start using them again... + */ if (i < peer->ksnp_n_passive_ips) { /* Old interface. */ ip = peer->ksnp_passive_ips[i]; best_iface = ksocknal_ip2iface(peer->ksnp_ni, ip); + /* peer passive ips are kept up to date */ + LASSERT(best_iface); } else { /* choose a new interface */ LASSERT(i == peer->ksnp_n_passive_ips); @@ -810,9 +827,9 @@ ksocknal_select_ips(ksock_peer_t *peer, __u32 *peerips, int n_peerips) k = ksocknal_match_peerip(iface, peerips, n_peerips); xor = ip ^ peerips[k]; - this_netmatch = ((xor & iface->ksni_netmask) == 0) ? 1 : 0; + this_netmatch = !(xor & iface->ksni_netmask) ? 1 : 0; - if (!(best_iface == NULL || + if (!(!best_iface || best_netmatch < this_netmatch || (best_netmatch == this_netmatch && best_npeers > iface->ksni_npeers))) @@ -823,10 +840,12 @@ ksocknal_select_ips(ksock_peer_t *peer, __u32 *peerips, int n_peerips) best_npeers = iface->ksni_npeers; } + LASSERT(best_iface); + best_iface->ksni_npeers++; ip = best_iface->ksni_ipaddr; peer->ksnp_passive_ips[i] = ip; - peer->ksnp_n_passive_ips = i+1; + peer->ksnp_n_passive_ips = i + 1; } /* mark the best matching peer IP used */ @@ -860,16 +879,19 @@ ksocknal_create_routes(ksock_peer_t *peer, int port, int i; int j; - /* CAVEAT EMPTOR: We do all our interface matching with an + /* + * CAVEAT EMPTOR: We do all our interface matching with an * exclusive hold of global lock at IRQ priority. We're only * expecting to be dealing with small numbers of interfaces, so the - * O(n**3)-ness here shouldn't matter */ - + * O(n**3)-ness here shouldn't matter + */ write_lock_bh(global_lock); if (net->ksnn_ninterfaces < 2) { - /* Only create additional connections - * if I have > 1 interface */ + /* + * Only create additional connections + * if I have > 1 interface + */ write_unlock_bh(global_lock); return; } @@ -877,13 +899,13 @@ ksocknal_create_routes(ksock_peer_t *peer, int port, LASSERT(npeer_ipaddrs <= LNET_MAX_INTERFACES); for (i = 0; i < npeer_ipaddrs; i++) { - if (newroute != NULL) { + if (newroute) { newroute->ksnr_ipaddr = peer_ipaddrs[i]; } else { write_unlock_bh(global_lock); newroute = ksocknal_create_route(peer_ipaddrs[i], port); - if (newroute == NULL) + if (!newroute) return; write_lock_bh(global_lock); @@ -904,7 +926,7 @@ ksocknal_create_routes(ksock_peer_t *peer, int port, route = NULL; } - if (route != NULL) + if (route) continue; best_iface = NULL; @@ -920,21 +942,21 @@ ksocknal_create_routes(ksock_peer_t *peer, int port, /* Using this interface already? */ list_for_each(rtmp, &peer->ksnp_routes) { route = list_entry(rtmp, ksock_route_t, - ksnr_list); + ksnr_list); if (route->ksnr_myipaddr == iface->ksni_ipaddr) break; route = NULL; } - if (route != NULL) + if (route) continue; - this_netmatch = (((iface->ksni_ipaddr ^ + this_netmatch = (!((iface->ksni_ipaddr ^ newroute->ksnr_ipaddr) & - iface->ksni_netmask) == 0) ? 1 : 0; + iface->ksni_netmask)) ? 1 : 0; - if (!(best_iface == NULL || + if (!(!best_iface || best_netmatch < this_netmatch || (best_netmatch == this_netmatch && best_nroutes > iface->ksni_nroutes))) @@ -945,7 +967,7 @@ ksocknal_create_routes(ksock_peer_t *peer, int port, best_nroutes = iface->ksni_nroutes; } - if (best_iface == NULL) + if (!best_iface) continue; newroute->ksnr_myipaddr = best_iface->ksni_ipaddr; @@ -956,7 +978,7 @@ ksocknal_create_routes(ksock_peer_t *peer, int port, } write_unlock_bh(global_lock); - if (newroute != NULL) + if (newroute) ksocknal_route_decref(newroute); } @@ -969,10 +991,10 @@ ksocknal_accept(lnet_ni_t *ni, struct socket *sock) int peer_port; rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port); - LASSERT(rc == 0); /* we succeeded before */ + LASSERT(!rc); /* we succeeded before */ LIBCFS_ALLOC(cr, sizeof(*cr)); - if (cr == NULL) { + if (!cr) { LCONSOLE_ERROR_MSG(0x12f, "Dropping connection request from %pI4h: memory exhausted\n", &peer_ip); return -ENOMEM; @@ -997,7 +1019,6 @@ ksocknal_connecting(ksock_peer_t *peer, __u32 ipaddr) ksock_route_t *route; list_for_each_entry(route, &peer->ksnp_routes, ksnr_list) { - if (route->ksnr_ipaddr == ipaddr) return route->ksnr_connecting; } @@ -1006,7 +1027,7 @@ ksocknal_connecting(ksock_peer_t *peer, __u32 ipaddr) int ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, - struct socket *sock, int type) + struct socket *sock, int type) { rwlock_t *global_lock = &ksocknal_data.ksnd_global_lock; LIST_HEAD(zombies); @@ -1026,12 +1047,12 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, int active; char *warn = NULL; - active = (route != NULL); + active = !!route; LASSERT(active == (type != SOCKLND_CONN_NONE)); LIBCFS_ALLOC(conn, sizeof(*conn)); - if (conn == NULL) { + if (!conn) { rc = -ENOMEM; goto failed_0; } @@ -1039,8 +1060,10 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, conn->ksnc_peer = NULL; conn->ksnc_route = NULL; conn->ksnc_sock = sock; - /* 2 ref, 1 for conn, another extra ref prevents socket - * being closed before establishment of connection */ + /* + * 2 ref, 1 for conn, another extra ref prevents socket + * being closed before establishment of connection + */ atomic_set(&conn->ksnc_sock_refcount, 2); conn->ksnc_type = type; ksocknal_lib_save_callback(sock, conn); @@ -1057,21 +1080,22 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, LIBCFS_ALLOC(hello, offsetof(ksock_hello_msg_t, kshm_ips[LNET_MAX_INTERFACES])); - if (hello == NULL) { + if (!hello) { rc = -ENOMEM; goto failed_1; } /* stash conn's local and remote addrs */ rc = ksocknal_lib_get_conn_addrs(conn); - if (rc != 0) + if (rc) goto failed_1; - /* Find out/confirm peer's NID and connection type and get the + /* + * Find out/confirm peer's NID and connection type and get the * vector of interfaces she's willing to let me connect to. * Passive connections use the listener timeout since the peer sends - * eagerly */ - + * eagerly + */ if (active) { peer = route->ksnr_peer; LASSERT(ni == peer->ksnp_ni); @@ -1084,7 +1108,7 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, conn->ksnc_proto = peer->ksnp_proto; write_unlock_bh(global_lock); - if (conn->ksnc_proto == NULL) { + if (!conn->ksnc_proto) { conn->ksnc_proto = &ksocknal_protocol_v3x; #if SOCKNAL_VERSION_DEBUG if (*ksocknal_tunables.ksnd_protocol == 2) @@ -1095,7 +1119,7 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, } rc = ksocknal_send_hello(ni, conn, peerid.nid, hello); - if (rc != 0) + if (rc) goto failed_1; } else { peerid.nid = LNET_NID_ANY; @@ -1109,8 +1133,8 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, if (rc < 0) goto failed_1; - LASSERT(rc == 0 || active); - LASSERT(conn->ksnc_proto != NULL); + LASSERT(!rc || active); + LASSERT(conn->ksnc_proto); LASSERT(peerid.nid != LNET_NID_ANY); cpt = lnet_cpt_of_nid(peerid.nid); @@ -1120,20 +1144,22 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, write_lock_bh(global_lock); } else { rc = ksocknal_create_peer(&peer, ni, peerid); - if (rc != 0) + if (rc) goto failed_1; write_lock_bh(global_lock); /* called with a ref on ni, so shutdown can't have started */ - LASSERT(((ksock_net_t *) ni->ni_data)->ksnn_shutdown == 0); + LASSERT(!((ksock_net_t *) ni->ni_data)->ksnn_shutdown); peer2 = ksocknal_find_peer_locked(ni, peerid); - if (peer2 == NULL) { - /* NB this puts an "empty" peer in the peer - * table (which takes my ref) */ + if (!peer2) { + /* + * NB this puts an "empty" peer in the peer + * table (which takes my ref) + */ list_add_tail(&peer->ksnp_list, - ksocknal_nid2peerlist(peerid.nid)); + ksocknal_nid2peerlist(peerid.nid)); } else { ksocknal_peer_decref(peer); peer = peer2; @@ -1143,8 +1169,10 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, ksocknal_peer_addref(peer); peer->ksnp_accepting++; - /* Am I already connecting to this guy? Resolve in - * favour of higher NID... */ + /* + * Am I already connecting to this guy? Resolve in + * favour of higher NID... + */ if (peerid.nid < ni->ni_nid && ksocknal_connecting(peer, conn->ksnc_ipaddr)) { rc = EALREADY; @@ -1161,8 +1189,9 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, goto failed_2; } - if (peer->ksnp_proto == NULL) { - /* Never connected before. + if (!peer->ksnp_proto) { + /* + * Never connected before. * NB recv_hello may have returned EPROTO to signal my peer * wants a different protocol than the one I asked for. */ @@ -1198,8 +1227,10 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, goto failed_2; } - /* Refuse to duplicate an existing connection, unless this is a - * loopback connection */ + /* + * Refuse to duplicate an existing connection, unless this is a + * loopback connection + */ if (conn->ksnc_ipaddr != conn->ksnc_myipaddr) { list_for_each(tmp, &peer->ksnp_conns) { conn2 = list_entry(tmp, ksock_conn_t, ksnc_list); @@ -1209,9 +1240,11 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, conn2->ksnc_type != conn->ksnc_type) continue; - /* Reply on a passive connection attempt so the peer - * realises we're connected. */ - LASSERT(rc == 0); + /* + * Reply on a passive connection attempt so the peer + * realises we're connected. + */ + LASSERT(!rc); if (!active) rc = EALREADY; @@ -1220,9 +1253,11 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, } } - /* If the connection created by this route didn't bind to the IP + /* + * If the connection created by this route didn't bind to the IP * address the route connected to, the connection/route matching - * code below probably isn't going to work. */ + * code below probably isn't going to work. + */ if (active && route->ksnr_ipaddr != conn->ksnc_ipaddr) { CERROR("Route %s %pI4h connected to %pI4h\n", @@ -1231,10 +1266,12 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, &conn->ksnc_ipaddr); } - /* Search for a route corresponding to the new connection and + /* + * Search for a route corresponding to the new connection and * create an association. This allows incoming connections created * by routes in my peer to match my own route entries so I don't - * continually create duplicate routes. */ + * continually create duplicate routes. + */ list_for_each(tmp, &peer->ksnp_routes) { route = list_entry(tmp, ksock_route_t, ksnr_list); @@ -1278,14 +1315,14 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, write_unlock_bh(global_lock); - /* We've now got a new connection. Any errors from here on are just + /* + * We've now got a new connection. Any errors from here on are just * like "normal" comms errors and we close the connection normally. * NB (a) we still have to send the reply HELLO for passive * connections, * (b) normal I/O on the conn is blocked until I setup and call the * socket callbacks. */ - CDEBUG(D_NET, "New conn %s p %d.x %pI4h -> %pI4h/%d incarnation:%lld sched[%d:%d]\n", libcfs_id2str(peerid), conn->ksnc_proto->pro_version, &conn->ksnc_myipaddr, &conn->ksnc_ipaddr, @@ -1305,12 +1342,14 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t, kshm_ips[LNET_MAX_INTERFACES])); - /* setup the socket AFTER I've received hello (it disables + /* + * setup the socket AFTER I've received hello (it disables * SO_LINGER). I might call back to the acceptor who may want * to send a protocol version response and then close the * socket; this ensures the socket only tears down after the - * response has been sent. */ - if (rc == 0) + * response has been sent. + */ + if (!rc) rc = ksocknal_lib_setup_sock(sock); write_lock_bh(global_lock); @@ -1323,14 +1362,14 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, write_unlock_bh(global_lock); - if (rc != 0) { + if (rc) { write_lock_bh(global_lock); if (!conn->ksnc_closing) { /* could be closed by another thread */ ksocknal_close_conn_locked(conn, rc); } write_unlock_bh(global_lock); - } else if (ksocknal_connsock_addref(conn) == 0) { + } else if (!ksocknal_connsock_addref(conn)) { /* Allow I/O to proceed. */ ksocknal_read_callback(conn); ksocknal_write_callback(conn); @@ -1352,19 +1391,21 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, write_unlock_bh(global_lock); - if (warn != NULL) { + if (warn) { if (rc < 0) CERROR("Not creating conn %s type %d: %s\n", libcfs_id2str(peerid), conn->ksnc_type, warn); else CDEBUG(D_NET, "Not creating conn %s type %d: %s\n", - libcfs_id2str(peerid), conn->ksnc_type, warn); + libcfs_id2str(peerid), conn->ksnc_type, warn); } if (!active) { if (rc > 0) { - /* Request retry by replying with CONN_NONE - * ksnc_proto has been set already */ + /* + * Request retry by replying with CONN_NONE + * ksnc_proto has been set already + */ conn->ksnc_type = SOCKLND_CONN_NONE; hello->kshm_nips = 0; ksocknal_send_hello(ni, conn, peerid.nid, hello); @@ -1379,7 +1420,7 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route, ksocknal_peer_decref(peer); failed_1: - if (hello != NULL) + if (hello) LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t, kshm_ips[LNET_MAX_INTERFACES])); @@ -1393,15 +1434,17 @@ failed_0: void ksocknal_close_conn_locked(ksock_conn_t *conn, int error) { - /* This just does the immmediate housekeeping, and queues the + /* + * This just does the immmediate housekeeping, and queues the * connection for the reaper to terminate. - * Caller holds ksnd_global_lock exclusively in irq context */ + * Caller holds ksnd_global_lock exclusively in irq context + */ ksock_peer_t *peer = conn->ksnc_peer; ksock_route_t *route; ksock_conn_t *conn2; struct list_head *tmp; - LASSERT(peer->ksnp_error == 0); + LASSERT(!peer->ksnp_error); LASSERT(!conn->ksnc_closing); conn->ksnc_closing = 1; @@ -1409,10 +1452,10 @@ ksocknal_close_conn_locked(ksock_conn_t *conn, int error) list_del(&conn->ksnc_list); route = conn->ksnc_route; - if (route != NULL) { + if (route) { /* dissociate conn from route... */ LASSERT(!route->ksnr_deleted); - LASSERT((route->ksnr_connected & (1 << conn->ksnc_type)) != 0); + LASSERT(route->ksnr_connected & (1 << conn->ksnc_type)); conn2 = NULL; list_for_each(tmp, &peer->ksnp_conns) { @@ -1424,7 +1467,7 @@ ksocknal_close_conn_locked(ksock_conn_t *conn, int error) conn2 = NULL; } - if (conn2 == NULL) + if (!conn2) route->ksnr_connected &= ~(1 << conn->ksnc_type); conn->ksnc_route = NULL; @@ -1445,15 +1488,17 @@ ksocknal_close_conn_locked(ksock_conn_t *conn, int error) LASSERT(conn->ksnc_proto == &ksocknal_protocol_v3x); - /* throw them to the last connection..., - * these TXs will be send to /dev/null by scheduler */ + /* + * throw them to the last connection..., + * these TXs will be send to /dev/null by scheduler + */ list_for_each_entry(tx, &peer->ksnp_tx_queue, - tx_list) + tx_list) ksocknal_tx_prep(conn, tx); spin_lock_bh(&conn->ksnc_scheduler->kss_lock); list_splice_init(&peer->ksnp_tx_queue, - &conn->ksnc_tx_queue); + &conn->ksnc_tx_queue); spin_unlock_bh(&conn->ksnc_scheduler->kss_lock); } @@ -1461,8 +1506,10 @@ ksocknal_close_conn_locked(ksock_conn_t *conn, int error) peer->ksnp_error = error; /* stash last conn close reason */ if (list_empty(&peer->ksnp_routes)) { - /* I've just closed last conn belonging to a - * peer with no routes to it */ + /* + * I've just closed last conn belonging to a + * peer with no routes to it + */ ksocknal_unlink_peer_locked(peer); } } @@ -1470,7 +1517,7 @@ ksocknal_close_conn_locked(ksock_conn_t *conn, int error) spin_lock_bh(&ksocknal_data.ksnd_reaper_lock); list_add_tail(&conn->ksnc_list, - &ksocknal_data.ksnd_deathrow_conns); + &ksocknal_data.ksnd_deathrow_conns); wake_up(&ksocknal_data.ksnd_reaper_waitq); spin_unlock_bh(&ksocknal_data.ksnd_reaper_lock); @@ -1482,16 +1529,17 @@ ksocknal_peer_failed(ksock_peer_t *peer) int notify = 0; unsigned long last_alive = 0; - /* There has been a connection failure or comms error; but I'll only + /* + * There has been a connection failure or comms error; but I'll only * tell LNET I think the peer is dead if it's to another kernel and - * there are no connections or connection attempts in existence. */ - + * there are no connections or connection attempts in existence. + */ read_lock(&ksocknal_data.ksnd_global_lock); - if ((peer->ksnp_id.pid & LNET_PID_USERFLAG) == 0 && + if (!(peer->ksnp_id.pid & LNET_PID_USERFLAG) && list_empty(&peer->ksnp_conns) && - peer->ksnp_accepting == 0 && - ksocknal_find_connecting_route_locked(peer) == NULL) { + !peer->ksnp_accepting && + !ksocknal_find_connecting_route_locked(peer)) { notify = 1; last_alive = peer->ksnp_last_alive; } @@ -1500,7 +1548,7 @@ ksocknal_peer_failed(ksock_peer_t *peer) if (notify) lnet_notify(peer->ksnp_ni, peer->ksnp_id.nid, 0, - last_alive); + last_alive); } void @@ -1508,12 +1556,15 @@ ksocknal_finalize_zcreq(ksock_conn_t *conn) { ksock_peer_t *peer = conn->ksnc_peer; ksock_tx_t *tx; + ksock_tx_t *temp; ksock_tx_t *tmp; LIST_HEAD(zlist); - /* NB safe to finalize TXs because closing of socket will - * abort all buffered data */ - LASSERT(conn->ksnc_sock == NULL); + /* + * NB safe to finalize TXs because closing of socket will + * abort all buffered data + */ + LASSERT(!conn->ksnc_sock); spin_lock(&peer->ksnp_lock); @@ -1521,7 +1572,7 @@ ksocknal_finalize_zcreq(ksock_conn_t *conn) if (tx->tx_conn != conn) continue; - LASSERT(tx->tx_msg.ksm_zc_cookies[0] != 0); + LASSERT(tx->tx_msg.ksm_zc_cookies[0]); tx->tx_msg.ksm_zc_cookies[0] = 0; tx->tx_zc_aborted = 1; /* mark it as not-acked */ @@ -1531,9 +1582,7 @@ ksocknal_finalize_zcreq(ksock_conn_t *conn) spin_unlock(&peer->ksnp_lock); - while (!list_empty(&zlist)) { - tx = list_entry(zlist.next, ksock_tx_t, tx_zc_list); - + list_for_each_entry_safe(tx, temp, &zlist, tx_zc_list) { list_del(&tx->tx_zc_list); ksocknal_tx_decref(tx); } @@ -1542,10 +1591,12 @@ ksocknal_finalize_zcreq(ksock_conn_t *conn) void ksocknal_terminate_conn(ksock_conn_t *conn) { - /* This gets called by the reaper (guaranteed thread context) to + /* + * This gets called by the reaper (guaranteed thread context) to * disengage the socket from its callbacks and close it. * ksnc_refcount will eventually hit zero, and then the reaper will - * destroy it. */ + * destroy it. + */ ksock_peer_t *peer = conn->ksnc_peer; ksock_sched_t *sched = conn->ksnc_scheduler; int failed = 0; @@ -1561,7 +1612,7 @@ ksocknal_terminate_conn(ksock_conn_t *conn) if (!conn->ksnc_tx_scheduled && !list_empty(&conn->ksnc_tx_queue)) { list_add_tail(&conn->ksnc_tx_list, - &sched->kss_tx_conns); + &sched->kss_tx_conns); conn->ksnc_tx_scheduled = 1; /* extra ref for scheduler */ ksocknal_conn_addref(conn); @@ -1576,11 +1627,13 @@ ksocknal_terminate_conn(ksock_conn_t *conn) ksocknal_lib_reset_callback(conn->ksnc_sock, conn); - /* OK, so this conn may not be completely disengaged from its - * scheduler yet, but it _has_ committed to terminate... */ + /* + * OK, so this conn may not be completely disengaged from its + * scheduler yet, but it _has_ committed to terminate... + */ conn->ksnc_scheduler->kss_nconns--; - if (peer->ksnp_error != 0) { + if (peer->ksnp_error) { /* peer's last conn closed in error */ LASSERT(list_empty(&peer->ksnp_conns)); failed = 1; @@ -1592,11 +1645,13 @@ ksocknal_terminate_conn(ksock_conn_t *conn) if (failed) ksocknal_peer_failed(peer); - /* The socket is closed on the final put; either here, or in + /* + * The socket is closed on the final put; either here, or in * ksocknal_{send,recv}msg(). Since we set up the linger2 option * when the connection was established, this will close the socket * immediately, aborting anything buffered in it. Any hung - * zero-copy transmits will therefore complete in finite time. */ + * zero-copy transmits will therefore complete in finite time. + */ ksocknal_connsock_decref(conn); } @@ -1605,7 +1660,7 @@ ksocknal_queue_zombie_conn(ksock_conn_t *conn) { /* Queue the conn for the reaper to destroy */ - LASSERT(atomic_read(&conn->ksnc_conn_refcount) == 0); + LASSERT(!atomic_read(&conn->ksnc_conn_refcount)); spin_lock_bh(&ksocknal_data.ksnd_reaper_lock); list_add_tail(&conn->ksnc_list, &ksocknal_data.ksnd_zombie_conns); @@ -1622,10 +1677,10 @@ ksocknal_destroy_conn(ksock_conn_t *conn) /* Final coup-de-grace of the reaper */ CDEBUG(D_NET, "connection %p\n", conn); - LASSERT(atomic_read(&conn->ksnc_conn_refcount) == 0); - LASSERT(atomic_read(&conn->ksnc_sock_refcount) == 0); - LASSERT(conn->ksnc_sock == NULL); - LASSERT(conn->ksnc_route == NULL); + LASSERT(!atomic_read(&conn->ksnc_conn_refcount)); + LASSERT(!atomic_read(&conn->ksnc_sock_refcount)); + LASSERT(!conn->ksnc_sock); + LASSERT(!conn->ksnc_route); LASSERT(!conn->ksnc_tx_scheduled); LASSERT(!conn->ksnc_rx_scheduled); LASSERT(list_empty(&conn->ksnc_tx_queue)); @@ -1642,7 +1697,7 @@ ksocknal_destroy_conn(ksock_conn_t *conn) cfs_duration_sec(cfs_time_sub(cfs_time_current(), last_rcv))); lnet_finalize(conn->ksnc_peer->ksnp_ni, - conn->ksnc_cookie, -EIO); + conn->ksnc_cookie, -EIO); break; case SOCKNAL_RX_LNET_HEADER: if (conn->ksnc_rx_started) @@ -1685,8 +1740,7 @@ ksocknal_close_peer_conns_locked(ksock_peer_t *peer, __u32 ipaddr, int why) list_for_each_safe(ctmp, cnxt, &peer->ksnp_conns) { conn = list_entry(ctmp, ksock_conn_t, ksnc_list); - if (ipaddr == 0 || - conn->ksnc_ipaddr == ipaddr) { + if (!ipaddr || conn->ksnc_ipaddr == ipaddr) { count++; ksocknal_close_conn_locked(conn, why); } @@ -1724,17 +1778,17 @@ ksocknal_close_matching_conns(lnet_process_id_t id, __u32 ipaddr) write_lock_bh(&ksocknal_data.ksnd_global_lock); - if (id.nid != LNET_NID_ANY) - lo = hi = (int)(ksocknal_nid2peerlist(id.nid) - ksocknal_data.ksnd_peers); - else { + if (id.nid != LNET_NID_ANY) { + lo = (int)(ksocknal_nid2peerlist(id.nid) - ksocknal_data.ksnd_peers); + hi = (int)(ksocknal_nid2peerlist(id.nid) - ksocknal_data.ksnd_peers); + } else { lo = 0; hi = ksocknal_data.ksnd_peer_hash_size - 1; } for (i = lo; i <= hi; i++) { list_for_each_safe(ptmp, pnxt, - &ksocknal_data.ksnd_peers[i]) { - + &ksocknal_data.ksnd_peers[i]) { peer = list_entry(ptmp, ksock_peer_t, ksnp_list); if (!((id.nid == LNET_NID_ANY || id.nid == peer->ksnp_id.nid) && @@ -1748,10 +1802,10 @@ ksocknal_close_matching_conns(lnet_process_id_t id, __u32 ipaddr) write_unlock_bh(&ksocknal_data.ksnd_global_lock); /* wildcards always succeed */ - if (id.nid == LNET_NID_ANY || id.pid == LNET_PID_ANY || ipaddr == 0) + if (id.nid == LNET_NID_ANY || id.pid == LNET_PID_ANY || !ipaddr) return 0; - if (count == 0) + if (!count) return -ENOENT; else return 0; @@ -1760,15 +1814,17 @@ ksocknal_close_matching_conns(lnet_process_id_t id, __u32 ipaddr) void ksocknal_notify(lnet_ni_t *ni, lnet_nid_t gw_nid, int alive) { - /* The router is telling me she's been notified of a change in - * gateway state.... */ + /* + * The router is telling me she's been notified of a change in + * gateway state.... + */ lnet_process_id_t id = {0}; id.nid = gw_nid; id.pid = LNET_PID_ANY; CDEBUG(D_NET, "gw %s %s\n", libcfs_nid2str(gw_nid), - alive ? "up" : "down"); + alive ? "up" : "down"); if (!alive) { /* If the gateway crashed, close all open connections... */ @@ -1776,8 +1832,10 @@ ksocknal_notify(lnet_ni_t *ni, lnet_nid_t gw_nid, int alive) return; } - /* ...otherwise do nothing. We can only establish new connections - * if we have autroutes, and these connect on demand. */ + /* + * ...otherwise do nothing. We can only establish new connections + * if we have autroutes, and these connect on demand. + */ } void @@ -1788,12 +1846,15 @@ ksocknal_query(lnet_ni_t *ni, lnet_nid_t nid, unsigned long *when) unsigned long now = cfs_time_current(); ksock_peer_t *peer = NULL; rwlock_t *glock = &ksocknal_data.ksnd_global_lock; - lnet_process_id_t id = {.nid = nid, .pid = LUSTRE_SRV_LNET_PID}; + lnet_process_id_t id = { + .nid = nid, + .pid = LNET_PID_LUSTRE, + }; read_lock(glock); peer = ksocknal_find_peer_locked(ni, id); - if (peer != NULL) { + if (peer) { struct list_head *tmp; ksock_conn_t *conn; int bufnob; @@ -1812,13 +1873,13 @@ ksocknal_query(lnet_ni_t *ni, lnet_nid_t nid, unsigned long *when) } last_alive = peer->ksnp_last_alive; - if (ksocknal_find_connectable_route_locked(peer) == NULL) + if (!ksocknal_find_connectable_route_locked(peer)) connect = 0; } read_unlock(glock); - if (last_alive != 0) + if (last_alive) *when = last_alive; CDEBUG(D_NET, "Peer %s %p, alive %ld secs ago, connect %d\n", @@ -1834,7 +1895,7 @@ ksocknal_query(lnet_ni_t *ni, lnet_nid_t nid, unsigned long *when) write_lock_bh(glock); peer = ksocknal_find_peer_locked(ni, id); - if (peer != NULL) + if (peer) ksocknal_launch_all_connections_locked(peer); write_unlock_bh(glock); @@ -1857,7 +1918,7 @@ ksocknal_push_peer(ksock_peer_t *peer) list_for_each(tmp, &peer->ksnp_conns) { if (i++ == index) { conn = list_entry(tmp, ksock_conn_t, - ksnc_list); + ksnc_list); ksocknal_conn_addref(conn); break; } @@ -1865,7 +1926,7 @@ ksocknal_push_peer(ksock_peer_t *peer) read_unlock(&ksocknal_data.ksnd_global_lock); - if (conn == NULL) + if (!conn) break; ksocknal_lib_push_conn(conn); @@ -1885,7 +1946,8 @@ static int ksocknal_push(lnet_ni_t *ni, lnet_process_id_t id) start = &ksocknal_data.ksnd_peers[0]; end = &ksocknal_data.ksnd_peers[hsize - 1]; } else { - start = end = ksocknal_nid2peerlist(id.nid); + start = ksocknal_nid2peerlist(id.nid); + end = ksocknal_nid2peerlist(id.nid); } for (tmp = start; tmp <= end; tmp++) { @@ -1910,7 +1972,7 @@ static int ksocknal_push(lnet_ni_t *ni, lnet_process_id_t id) } read_unlock(&ksocknal_data.ksnd_global_lock); - if (i == 0) /* no match */ + if (!i) /* no match */ break; rc = 0; @@ -1934,14 +1996,13 @@ ksocknal_add_interface(lnet_ni_t *ni, __u32 ipaddress, __u32 netmask) struct list_head *rtmp; ksock_route_t *route; - if (ipaddress == 0 || - netmask == 0) + if (!ipaddress || !netmask) return -EINVAL; write_lock_bh(&ksocknal_data.ksnd_global_lock); iface = ksocknal_ip2iface(ni, ipaddress); - if (iface != NULL) { + if (iface) { /* silently ignore dups */ rc = 0; } else if (net->ksnn_ninterfaces == LNET_MAX_INTERFACES) { @@ -1957,16 +2018,15 @@ ksocknal_add_interface(lnet_ni_t *ni, __u32 ipaddress, __u32 netmask) for (i = 0; i < ksocknal_data.ksnd_peer_hash_size; i++) { list_for_each(ptmp, &ksocknal_data.ksnd_peers[i]) { peer = list_entry(ptmp, ksock_peer_t, - ksnp_list); + ksnp_list); for (j = 0; j < peer->ksnp_n_passive_ips; j++) if (peer->ksnp_passive_ips[j] == ipaddress) iface->ksni_npeers++; list_for_each(rtmp, &peer->ksnp_routes) { - route = list_entry(rtmp, - ksock_route_t, - ksnr_list); + route = list_entry(rtmp, ksock_route_t, + ksnr_list); if (route->ksnr_myipaddr == ipaddress) iface->ksni_nroutes++; @@ -1995,8 +2055,8 @@ ksocknal_peer_del_interface_locked(ksock_peer_t *peer, __u32 ipaddr) for (i = 0; i < peer->ksnp_n_passive_ips; i++) if (peer->ksnp_passive_ips[i] == ipaddr) { - for (j = i+1; j < peer->ksnp_n_passive_ips; j++) - peer->ksnp_passive_ips[j-1] = + for (j = i + 1; j < peer->ksnp_n_passive_ips; j++) + peer->ksnp_passive_ips[j - 1] = peer->ksnp_passive_ips[j]; peer->ksnp_n_passive_ips--; break; @@ -2008,7 +2068,7 @@ ksocknal_peer_del_interface_locked(ksock_peer_t *peer, __u32 ipaddr) if (route->ksnr_myipaddr != ipaddr) continue; - if (route->ksnr_share_count != 0) { + if (route->ksnr_share_count) { /* Manually created; keep, but unbind */ route->ksnr_myipaddr = 0; } else { @@ -2041,23 +2101,21 @@ ksocknal_del_interface(lnet_ni_t *ni, __u32 ipaddress) for (i = 0; i < net->ksnn_ninterfaces; i++) { this_ip = net->ksnn_interfaces[i].ksni_ipaddr; - if (!(ipaddress == 0 || - ipaddress == this_ip)) + if (!(!ipaddress || ipaddress == this_ip)) continue; rc = 0; - for (j = i+1; j < net->ksnn_ninterfaces; j++) - net->ksnn_interfaces[j-1] = + for (j = i + 1; j < net->ksnn_ninterfaces; j++) + net->ksnn_interfaces[j - 1] = net->ksnn_interfaces[j]; net->ksnn_ninterfaces--; for (j = 0; j < ksocknal_data.ksnd_peer_hash_size; j++) { list_for_each_safe(tmp, nxt, - &ksocknal_data.ksnd_peers[j]) { - peer = list_entry(tmp, ksock_peer_t, - ksnp_list); + &ksocknal_data.ksnd_peers[j]) { + peer = list_entry(tmp, ksock_peer_t, ksnp_list); if (peer->ksnp_ni != ni) continue; @@ -2121,7 +2179,7 @@ ksocknal_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) rc = ksocknal_get_peer_info(ni, data->ioc_count, &id, &myip, &ip, &port, &conn_count, &share_count); - if (rc != 0) + if (rc) return rc; data->ioc_nid = id.nid; @@ -2136,7 +2194,7 @@ ksocknal_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) case IOC_LIBCFS_ADD_PEER: id.nid = data->ioc_nid; - id.pid = LUSTRE_SRV_LNET_PID; + id.pid = LNET_PID_LUSTRE; return ksocknal_add_peer(ni, id, data->ioc_u32[0], /* IP */ data->ioc_u32[1]); /* port */ @@ -2153,7 +2211,7 @@ ksocknal_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) int nagle; ksock_conn_t *conn = ksocknal_get_conn_by_idx(ni, data->ioc_count); - if (conn == NULL) + if (!conn) return -ENOENT; ksocknal_lib_get_conn_tunables(conn, &txmem, &rxmem, &nagle); @@ -2202,14 +2260,14 @@ ksocknal_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg) static void ksocknal_free_buffers(void) { - LASSERT(atomic_read(&ksocknal_data.ksnd_nactive_txs) == 0); + LASSERT(!atomic_read(&ksocknal_data.ksnd_nactive_txs)); - if (ksocknal_data.ksnd_sched_info != NULL) { + if (ksocknal_data.ksnd_sched_info) { struct ksock_sched_info *info; int i; cfs_percpt_for_each(info, i, ksocknal_data.ksnd_sched_info) { - if (info->ksi_scheds != NULL) { + if (info->ksi_scheds) { LIBCFS_FREE(info->ksi_scheds, info->ksi_nthreads_max * sizeof(info->ksi_scheds[0])); @@ -2219,21 +2277,21 @@ ksocknal_free_buffers(void) } LIBCFS_FREE(ksocknal_data.ksnd_peers, - sizeof(struct list_head) * - ksocknal_data.ksnd_peer_hash_size); + sizeof(struct list_head) * + ksocknal_data.ksnd_peer_hash_size); spin_lock(&ksocknal_data.ksnd_tx_lock); if (!list_empty(&ksocknal_data.ksnd_idle_noop_txs)) { struct list_head zlist; ksock_tx_t *tx; + ksock_tx_t *temp; list_add(&zlist, &ksocknal_data.ksnd_idle_noop_txs); list_del_init(&ksocknal_data.ksnd_idle_noop_txs); spin_unlock(&ksocknal_data.ksnd_tx_lock); - while (!list_empty(&zlist)) { - tx = list_entry(zlist.next, ksock_tx_t, tx_list); + list_for_each_entry_safe(tx, temp, &zlist, tx_list) { list_del(&tx->tx_list); LIBCFS_FREE(tx, tx->tx_desc_size); } @@ -2250,7 +2308,7 @@ ksocknal_base_shutdown(void) int i; int j; - LASSERT(ksocknal_data.ksnd_nnets == 0); + LASSERT(!ksocknal_data.ksnd_nnets); switch (ksocknal_data.ksnd_init) { default: @@ -2258,7 +2316,7 @@ ksocknal_base_shutdown(void) case SOCKNAL_INIT_ALL: case SOCKNAL_INIT_DATA: - LASSERT(ksocknal_data.ksnd_peers != NULL); + LASSERT(ksocknal_data.ksnd_peers); for (i = 0; i < ksocknal_data.ksnd_peer_hash_size; i++) LASSERT(list_empty(&ksocknal_data.ksnd_peers[i])); @@ -2268,14 +2326,13 @@ ksocknal_base_shutdown(void) LASSERT(list_empty(&ksocknal_data.ksnd_connd_connreqs)); LASSERT(list_empty(&ksocknal_data.ksnd_connd_routes)); - if (ksocknal_data.ksnd_sched_info != NULL) { + if (ksocknal_data.ksnd_sched_info) { cfs_percpt_for_each(info, i, ksocknal_data.ksnd_sched_info) { - if (info->ksi_scheds == NULL) + if (!info->ksi_scheds) continue; for (j = 0; j < info->ksi_nthreads_max; j++) { - sched = &info->ksi_scheds[j]; LASSERT(list_empty( &sched->kss_tx_conns)); @@ -2283,7 +2340,7 @@ ksocknal_base_shutdown(void) &sched->kss_rx_conns)); LASSERT(list_empty( &sched->kss_zombie_noop_txs)); - LASSERT(sched->kss_nconns == 0); + LASSERT(!sched->kss_nconns); } } } @@ -2293,10 +2350,10 @@ ksocknal_base_shutdown(void) wake_up_all(&ksocknal_data.ksnd_connd_waitq); wake_up_all(&ksocknal_data.ksnd_reaper_waitq); - if (ksocknal_data.ksnd_sched_info != NULL) { + if (ksocknal_data.ksnd_sched_info) { cfs_percpt_for_each(info, i, ksocknal_data.ksnd_sched_info) { - if (info->ksi_scheds == NULL) + if (!info->ksi_scheds) continue; for (j = 0; j < info->ksi_nthreads_max; j++) { @@ -2308,7 +2365,7 @@ ksocknal_base_shutdown(void) i = 4; read_lock(&ksocknal_data.ksnd_global_lock); - while (ksocknal_data.ksnd_nthreads != 0) { + while (ksocknal_data.ksnd_nthreads) { i++; CDEBUG(((i & (-i)) == i) ? D_WARNING : D_NET, /* power of 2? */ "waiting for %d threads to terminate\n", @@ -2332,7 +2389,6 @@ ksocknal_base_shutdown(void) static __u64 ksocknal_new_incarnation(void) { - /* The incarnation number is the time this module loaded and it * identifies this particular instance of the socknal. */ @@ -2347,15 +2403,15 @@ ksocknal_base_startup(void) int i; LASSERT(ksocknal_data.ksnd_init == SOCKNAL_INIT_NOTHING); - LASSERT(ksocknal_data.ksnd_nnets == 0); + LASSERT(!ksocknal_data.ksnd_nnets); memset(&ksocknal_data, 0, sizeof(ksocknal_data)); /* zero pointers */ ksocknal_data.ksnd_peer_hash_size = SOCKNAL_PEER_HASH_SIZE; LIBCFS_ALLOC(ksocknal_data.ksnd_peers, - sizeof(struct list_head) * - ksocknal_data.ksnd_peer_hash_size); - if (ksocknal_data.ksnd_peers == NULL) + sizeof(struct list_head) * + ksocknal_data.ksnd_peer_hash_size); + if (!ksocknal_data.ksnd_peers) return -ENOMEM; for (i = 0; i < ksocknal_data.ksnd_peer_hash_size; i++) @@ -2386,7 +2442,7 @@ ksocknal_base_startup(void) ksocknal_data.ksnd_sched_info = cfs_percpt_alloc(lnet_cpt_table(), sizeof(*info)); - if (ksocknal_data.ksnd_sched_info == NULL) + if (!ksocknal_data.ksnd_sched_info) goto failed; cfs_percpt_for_each(info, i, ksocknal_data.ksnd_sched_info) { @@ -2397,8 +2453,10 @@ ksocknal_base_startup(void) if (*ksocknal_tunables.ksnd_nscheds > 0) { nthrs = min(nthrs, *ksocknal_tunables.ksnd_nscheds); } else { - /* max to half of CPUs, assume another half should be - * reserved for upper layer modules */ + /* + * max to half of CPUs, assume another half should be + * reserved for upper layer modules + */ nthrs = min(max(SOCKNAL_NSCHEDS, nthrs >> 1), nthrs); } @@ -2407,7 +2465,7 @@ ksocknal_base_startup(void) LIBCFS_CPT_ALLOC(info->ksi_scheds, lnet_cpt_table(), i, info->ksi_nthreads_max * sizeof(*sched)); - if (info->ksi_scheds == NULL) + if (!info->ksi_scheds) goto failed; for (; nthrs > 0; nthrs--) { @@ -2425,8 +2483,10 @@ ksocknal_base_startup(void) ksocknal_data.ksnd_connd_starting = 0; ksocknal_data.ksnd_connd_failed_stamp = 0; ksocknal_data.ksnd_connd_starting_stamp = ktime_get_real_seconds(); - /* must have at least 2 connds to remain responsive to accepts while - * connecting */ + /* + * must have at least 2 connds to remain responsive to accepts while + * connecting + */ if (*ksocknal_tunables.ksnd_nconnds < SOCKNAL_CONND_RESV + 1) *ksocknal_tunables.ksnd_nconnds = SOCKNAL_CONND_RESV + 1; @@ -2446,7 +2506,7 @@ ksocknal_base_startup(void) snprintf(name, sizeof(name), "socknal_cd%02d", i); rc = ksocknal_thread_start(ksocknal_connd, (void *)((ulong_ptr_t)i), name); - if (rc != 0) { + if (rc) { spin_lock_bh(&ksocknal_data.ksnd_connd_lock); ksocknal_data.ksnd_connd_starting--; spin_unlock_bh(&ksocknal_data.ksnd_connd_lock); @@ -2456,7 +2516,7 @@ ksocknal_base_startup(void) } rc = ksocknal_thread_start(ksocknal_reaper, NULL, "socknal_reaper"); - if (rc != 0) { + if (rc) { CERROR("Can't spawn socknal reaper: %d\n", rc); goto failed; } @@ -2491,7 +2551,7 @@ ksocknal_debug_peerhash(lnet_ni_t *ni) } } - if (peer != NULL) { + if (peer) { ksock_route_t *route; ksock_conn_t *conn; @@ -2515,9 +2575,9 @@ ksocknal_debug_peerhash(lnet_ni_t *ni) list_for_each(tmp, &peer->ksnp_conns) { conn = list_entry(tmp, ksock_conn_t, ksnc_list); CWARN("Conn: ref %d, sref %d, t %d, c %d\n", - atomic_read(&conn->ksnc_conn_refcount), - atomic_read(&conn->ksnc_sock_refcount), - conn->ksnc_type, conn->ksnc_closing); + atomic_read(&conn->ksnc_conn_refcount), + atomic_read(&conn->ksnc_sock_refcount), + conn->ksnc_type, conn->ksnc_closing); } } @@ -2548,7 +2608,7 @@ ksocknal_shutdown(lnet_ni_t *ni) /* Wait for all peer state to clean up */ i = 2; spin_lock_bh(&net->ksnn_lock); - while (net->ksnn_npeers != 0) { + while (net->ksnn_npeers) { spin_unlock_bh(&net->ksnn_lock); i++; @@ -2565,15 +2625,15 @@ ksocknal_shutdown(lnet_ni_t *ni) spin_unlock_bh(&net->ksnn_lock); for (i = 0; i < net->ksnn_ninterfaces; i++) { - LASSERT(net->ksnn_interfaces[i].ksni_npeers == 0); - LASSERT(net->ksnn_interfaces[i].ksni_nroutes == 0); + LASSERT(!net->ksnn_interfaces[i].ksni_npeers); + LASSERT(!net->ksnn_interfaces[i].ksni_nroutes); } list_del(&net->ksnn_list); LIBCFS_FREE(net, sizeof(*net)); ksocknal_data.ksnd_nnets--; - if (ksocknal_data.ksnd_nnets == 0) + if (!ksocknal_data.ksnd_nnets) ksocknal_base_shutdown(); } @@ -2601,7 +2661,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net) continue; rc = lnet_ipif_query(names[i], &up, &ip, &mask); - if (rc != 0) { + if (rc) { CWARN("Can't get interface %s info: %d\n", names[i], rc); continue; @@ -2628,7 +2688,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net) lnet_ipif_free_enumeration(names, n); - if (j == 0) + if (!j) CERROR("Can't find any usable interfaces\n"); return j; @@ -2647,21 +2707,20 @@ ksocknal_search_new_ipif(ksock_net_t *net) ksock_net_t *tmp; int j; - if (colon != NULL) /* ignore alias device */ + if (colon) /* ignore alias device */ *colon = 0; - list_for_each_entry(tmp, &ksocknal_data.ksnd_nets, - ksnn_list) { + list_for_each_entry(tmp, &ksocknal_data.ksnd_nets, ksnn_list) { for (j = 0; !found && j < tmp->ksnn_ninterfaces; j++) { char *ifnam2 = &tmp->ksnn_interfaces[j].ksni_name[0]; char *colon2 = strchr(ifnam2, ':'); - if (colon2 != NULL) + if (colon2) *colon2 = 0; - found = strcmp(ifnam, ifnam2) == 0; - if (colon2 != NULL) + found = !strcmp(ifnam, ifnam2); + if (colon2) *colon2 = ':'; } if (found) @@ -2669,7 +2728,7 @@ ksocknal_search_new_ipif(ksock_net_t *net) } new_ipif += !found; - if (colon != NULL) + if (colon) *colon = ':'; } @@ -2683,7 +2742,7 @@ ksocknal_start_schedulers(struct ksock_sched_info *info) int rc = 0; int i; - if (info->ksi_nthreads == 0) { + if (!info->ksi_nthreads) { if (*ksocknal_tunables.ksnd_nscheds > 0) { nthrs = info->ksi_nthreads_max; } else { @@ -2711,7 +2770,7 @@ ksocknal_start_schedulers(struct ksock_sched_info *info) rc = ksocknal_thread_start(ksocknal_scheduler, (void *)id, name); - if (rc == 0) + if (!rc) continue; CERROR("Can't spawn thread %d for scheduler[%d]: %d\n", @@ -2734,7 +2793,7 @@ ksocknal_net_start_threads(ksock_net_t *net, __u32 *cpts, int ncpts) for (i = 0; i < ncpts; i++) { struct ksock_sched_info *info; - int cpt = (cpts == NULL) ? i : cpts[i]; + int cpt = !cpts ? i : cpts[i]; LASSERT(cpt < cfs_cpt_number(lnet_cpt_table())); info = ksocknal_data.ksnd_sched_info[cpt]; @@ -2743,7 +2802,7 @@ ksocknal_net_start_threads(ksock_net_t *net, __u32 *cpts, int ncpts) continue; rc = ksocknal_start_schedulers(info); - if (rc != 0) + if (rc) return rc; } return 0; @@ -2760,12 +2819,12 @@ ksocknal_startup(lnet_ni_t *ni) if (ksocknal_data.ksnd_init == SOCKNAL_INIT_NOTHING) { rc = ksocknal_base_startup(); - if (rc != 0) + if (rc) return rc; } LIBCFS_ALLOC(net, sizeof(*net)); - if (net == NULL) + if (!net) goto fail_0; spin_lock_init(&net->ksnn_lock); @@ -2776,7 +2835,7 @@ ksocknal_startup(lnet_ni_t *ni) ni->ni_peertxcredits = *ksocknal_tunables.ksnd_peertxcredits; ni->ni_peerrtrcredits = *ksocknal_tunables.ksnd_peerrtrcredits; - if (ni->ni_interfaces[0] == NULL) { + if (!ni->ni_interfaces[0]) { rc = ksocknal_enumerate_interfaces(net); if (rc <= 0) goto fail_1; @@ -2786,14 +2845,14 @@ ksocknal_startup(lnet_ni_t *ni) for (i = 0; i < LNET_MAX_INTERFACES; i++) { int up; - if (ni->ni_interfaces[i] == NULL) + if (!ni->ni_interfaces[i]) break; rc = lnet_ipif_query(ni->ni_interfaces[i], &up, - &net->ksnn_interfaces[i].ksni_ipaddr, - &net->ksnn_interfaces[i].ksni_netmask); + &net->ksnn_interfaces[i].ksni_ipaddr, + &net->ksnn_interfaces[i].ksni_netmask); - if (rc != 0) { + if (rc) { CERROR("Can't get interface %s info: %d\n", ni->ni_interfaces[i], rc); goto fail_1; @@ -2814,7 +2873,7 @@ ksocknal_startup(lnet_ni_t *ni) /* call it before add it to ksocknal_data.ksnd_nets */ rc = ksocknal_net_start_threads(net, ni->ni_cpts, ni->ni_ncpts); - if (rc != 0) + if (rc) goto fail_1; ni->ni_nid = LNET_MKNID(LNET_NIDNET(ni->ni_nid), @@ -2828,20 +2887,18 @@ ksocknal_startup(lnet_ni_t *ni) fail_1: LIBCFS_FREE(net, sizeof(*net)); fail_0: - if (ksocknal_data.ksnd_nnets == 0) + if (!ksocknal_data.ksnd_nnets) ksocknal_base_shutdown(); return -ENETDOWN; } -static void __exit -ksocknal_module_fini(void) +static void __exit ksocklnd_exit(void) { lnet_unregister_lnd(&the_ksocklnd); } -static int __init -ksocknal_module_init(void) +static int __init ksocklnd_init(void) { int rc; @@ -2861,7 +2918,7 @@ ksocknal_module_init(void) the_ksocklnd.lnd_accept = ksocknal_accept; rc = ksocknal_tunables_init(); - if (rc != 0) + if (rc) return rc; lnet_register_lnd(&the_ksocklnd); @@ -2870,9 +2927,9 @@ ksocknal_module_init(void) } MODULE_AUTHOR("OpenSFS, Inc. <http://www.lustre.org/>"); -MODULE_DESCRIPTION("Kernel TCP Socket LND v3.0.0"); +MODULE_DESCRIPTION("TCP Socket LNet Network Driver"); +MODULE_VERSION("2.7.0"); MODULE_LICENSE("GPL"); -MODULE_VERSION("3.0.0"); -module_init(ksocknal_module_init); -module_exit(ksocknal_module_fini); +module_init(ksocklnd_init); +module_exit(ksocklnd_exit); diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h index f4fa72550..a60d72f94 100644 --- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h +++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h @@ -19,10 +19,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with Portals; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - * */ #ifndef _SOCKLND_SOCKLND_H_ @@ -69,8 +65,10 @@ #define SOCKNAL_VERSION_DEBUG 0 /* enable protocol version debugging */ -/* risk kmap deadlock on multi-frag I/O (backs off to single-frag if disabled). - * no risk if we're not running on a CONFIG_HIGHMEM platform. */ +/* + * risk kmap deadlock on multi-frag I/O (backs off to single-frag if disabled). + * no risk if we're not running on a CONFIG_HIGHMEM platform. + */ #ifdef CONFIG_HIGHMEM # define SOCKNAL_RISK_KMAP_DEADLOCK 0 #else @@ -237,15 +235,16 @@ typedef struct { #define SOCKNAL_INIT_DATA 1 #define SOCKNAL_INIT_ALL 2 -/* A packet just assembled for transmission is represented by 1 or more +/* + * A packet just assembled for transmission is represented by 1 or more * struct iovec fragments (the first frag contains the portals header), * followed by 0 or more lnet_kiov_t fragments. * * On the receive side, initially 1 struct iovec fragment is posted for * receive (the header). Once the header has been received, the payload is * received into either struct iovec or lnet_kiov_t fragments, depending on - * what the header matched or whether the message needs forwarding. */ - + * what the header matched or whether the message needs forwarding. + */ struct ksock_conn; /* forward ref */ struct ksock_peer; /* forward ref */ struct ksock_route; /* forward ref */ @@ -284,12 +283,14 @@ typedef struct /* transmit packet */ } tx_frags; } ksock_tx_t; -#define KSOCK_NOOP_TX_SIZE ((int)offsetof(ksock_tx_t, tx_frags.paged.kiov[0])) +#define KSOCK_NOOP_TX_SIZE (offsetof(ksock_tx_t, tx_frags.paged.kiov[0])) /* network zero copy callback descriptor embedded in ksock_tx_t */ -/* space for the rx frag descriptors; we either read a single contiguous - * header, or up to LNET_MAX_IOV frags of payload of either type. */ +/* + * space for the rx frag descriptors; we either read a single contiguous + * header, or up to LNET_MAX_IOV frags of payload of either type. + */ typedef union { struct kvec iov[LNET_MAX_IOV]; lnet_kiov_t kiov[LNET_MAX_IOV]; @@ -463,11 +464,13 @@ typedef struct ksock_proto { /* handle ZC ACK */ int (*pro_handle_zcack)(ksock_conn_t *, __u64, __u64); - /* msg type matches the connection type: + /* + * msg type matches the connection type: * return value: * return MATCH_NO : no * return MATCH_YES : matching type - * return MATCH_MAY : can be backup */ + * return MATCH_MAY : can be backup + */ int (*pro_match_tx)(ksock_conn_t *, ksock_tx_t *, int); } ksock_proto_t; diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c index 477b385f1..976fd7892 100644 --- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c +++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c @@ -19,9 +19,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with Portals; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ #include "socklnd.h" @@ -47,10 +44,10 @@ ksocknal_alloc_tx(int type, int size) spin_unlock(&ksocknal_data.ksnd_tx_lock); } - if (tx == NULL) + if (!tx) LIBCFS_ALLOC(tx, size); - if (tx == NULL) + if (!tx) return NULL; atomic_set(&tx->tx_refcount, 1); @@ -70,7 +67,7 @@ ksocknal_alloc_tx_noop(__u64 cookie, int nonblk) ksock_tx_t *tx; tx = ksocknal_alloc_tx(KSOCK_MSG_NOOP, KSOCK_NOOP_TX_SIZE); - if (tx == NULL) { + if (!tx) { CERROR("Can't allocate noop tx desc\n"); return NULL; } @@ -90,11 +87,11 @@ ksocknal_alloc_tx_noop(__u64 cookie, int nonblk) } void -ksocknal_free_tx (ksock_tx_t *tx) +ksocknal_free_tx(ksock_tx_t *tx) { atomic_dec(&ksocknal_data.ksnd_nactive_txs); - if (tx->tx_lnetmsg == NULL && tx->tx_desc_size == KSOCK_NOOP_TX_SIZE) { + if (!tx->tx_lnetmsg && tx->tx_desc_size == KSOCK_NOOP_TX_SIZE) { /* it's a noop tx */ spin_lock(&ksocknal_data.ksnd_tx_lock); @@ -107,7 +104,7 @@ ksocknal_free_tx (ksock_tx_t *tx) } static int -ksocknal_send_iov (ksock_conn_t *conn, ksock_tx_t *tx) +ksocknal_send_iov(ksock_conn_t *conn, ksock_tx_t *tx) { struct kvec *iov = tx->tx_iov; int nob; @@ -122,7 +119,7 @@ ksocknal_send_iov (ksock_conn_t *conn, ksock_tx_t *tx) return rc; nob = rc; - LASSERT (nob <= tx->tx_resid); + LASSERT(nob <= tx->tx_resid); tx->tx_resid -= nob; /* "consume" iov */ @@ -138,19 +135,19 @@ ksocknal_send_iov (ksock_conn_t *conn, ksock_tx_t *tx) nob -= iov->iov_len; tx->tx_iov = ++iov; tx->tx_niov--; - } while (nob != 0); + } while (nob); return rc; } static int -ksocknal_send_kiov (ksock_conn_t *conn, ksock_tx_t *tx) +ksocknal_send_kiov(ksock_conn_t *conn, ksock_tx_t *tx) { lnet_kiov_t *kiov = tx->tx_kiov; int nob; int rc; - LASSERT(tx->tx_niov == 0); + LASSERT(!tx->tx_niov); LASSERT(tx->tx_nkiov > 0); /* Never touch tx->tx_kiov inside ksocknal_lib_send_kiov() */ @@ -160,7 +157,7 @@ ksocknal_send_kiov (ksock_conn_t *conn, ksock_tx_t *tx) return rc; nob = rc; - LASSERT (nob <= tx->tx_resid); + LASSERT(nob <= tx->tx_resid); tx->tx_resid -= nob; /* "consume" kiov */ @@ -176,27 +173,27 @@ ksocknal_send_kiov (ksock_conn_t *conn, ksock_tx_t *tx) nob -= (int)kiov->kiov_len; tx->tx_kiov = ++kiov; tx->tx_nkiov--; - } while (nob != 0); + } while (nob); return rc; } static int -ksocknal_transmit (ksock_conn_t *conn, ksock_tx_t *tx) +ksocknal_transmit(ksock_conn_t *conn, ksock_tx_t *tx) { int rc; int bufnob; - if (ksocknal_data.ksnd_stall_tx != 0) { + if (ksocknal_data.ksnd_stall_tx) { set_current_state(TASK_UNINTERRUPTIBLE); schedule_timeout(cfs_time_seconds(ksocknal_data.ksnd_stall_tx)); } - LASSERT(tx->tx_resid != 0); + LASSERT(tx->tx_resid); rc = ksocknal_connsock_addref(conn); - if (rc != 0) { - LASSERT (conn->ksnc_closing); + if (rc) { + LASSERT(conn->ksnc_closing); return -ESHUTDOWN; } @@ -205,10 +202,10 @@ ksocknal_transmit (ksock_conn_t *conn, ksock_tx_t *tx) /* testing... */ ksocknal_data.ksnd_enomem_tx--; rc = -EAGAIN; - } else if (tx->tx_niov != 0) { - rc = ksocknal_send_iov (conn, tx); + } else if (tx->tx_niov) { + rc = ksocknal_send_iov(conn, tx); } else { - rc = ksocknal_send_kiov (conn, tx); + rc = ksocknal_send_kiov(conn, tx); } bufnob = conn->ksnc_sock->sk->sk_wmem_queued; @@ -216,8 +213,10 @@ ksocknal_transmit (ksock_conn_t *conn, ksock_tx_t *tx) conn->ksnc_tx_bufnob += rc; /* account it */ if (bufnob < conn->ksnc_tx_bufnob) { - /* allocated send buffer bytes < computed; infer - * something got ACKed */ + /* + * allocated send buffer bytes < computed; infer + * something got ACKed + */ conn->ksnc_tx_deadline = cfs_time_shift(*ksocknal_tunables.ksnd_timeout); conn->ksnc_peer->ksnp_last_alive = cfs_time_current(); @@ -227,7 +226,7 @@ ksocknal_transmit (ksock_conn_t *conn, ksock_tx_t *tx) if (rc <= 0) { /* Didn't write anything? */ - if (rc == 0) /* some stacks return 0 instead of -EAGAIN */ + if (!rc) /* some stacks return 0 instead of -EAGAIN */ rc = -EAGAIN; /* Check if EAGAIN is due to memory pressure */ @@ -238,17 +237,17 @@ ksocknal_transmit (ksock_conn_t *conn, ksock_tx_t *tx) } /* socket's wmem_queued now includes 'rc' bytes */ - atomic_sub (rc, &conn->ksnc_tx_nob); + atomic_sub(rc, &conn->ksnc_tx_nob); rc = 0; - } while (tx->tx_resid != 0); + } while (tx->tx_resid); ksocknal_connsock_decref(conn); return rc; } static int -ksocknal_recv_iov (ksock_conn_t *conn) +ksocknal_recv_iov(ksock_conn_t *conn) { struct kvec *iov = conn->ksnc_rx_iov; int nob; @@ -256,8 +255,10 @@ ksocknal_recv_iov (ksock_conn_t *conn) LASSERT(conn->ksnc_rx_niov > 0); - /* Never touch conn->ksnc_rx_iov or change connection - * status inside ksocknal_lib_recv_iov */ + /* + * Never touch conn->ksnc_rx_iov or change connection + * status inside ksocknal_lib_recv_iov + */ rc = ksocknal_lib_recv_iov(conn); if (rc <= 0) @@ -287,13 +288,13 @@ ksocknal_recv_iov (ksock_conn_t *conn) nob -= iov->iov_len; conn->ksnc_rx_iov = ++iov; conn->ksnc_rx_niov--; - } while (nob != 0); + } while (nob); return rc; } static int -ksocknal_recv_kiov (ksock_conn_t *conn) +ksocknal_recv_kiov(ksock_conn_t *conn) { lnet_kiov_t *kiov = conn->ksnc_rx_kiov; int nob; @@ -301,8 +302,10 @@ ksocknal_recv_kiov (ksock_conn_t *conn) LASSERT(conn->ksnc_rx_nkiov > 0); - /* Never touch conn->ksnc_rx_kiov or change connection - * status inside ksocknal_lib_recv_iov */ + /* + * Never touch conn->ksnc_rx_kiov or change connection + * status inside ksocknal_lib_recv_iov + */ rc = ksocknal_lib_recv_kiov(conn); if (rc <= 0) @@ -332,41 +335,43 @@ ksocknal_recv_kiov (ksock_conn_t *conn) nob -= kiov->kiov_len; conn->ksnc_rx_kiov = ++kiov; conn->ksnc_rx_nkiov--; - } while (nob != 0); + } while (nob); return 1; } static int -ksocknal_receive (ksock_conn_t *conn) +ksocknal_receive(ksock_conn_t *conn) { - /* Return 1 on success, 0 on EOF, < 0 on error. + /* + * Return 1 on success, 0 on EOF, < 0 on error. * Caller checks ksnc_rx_nob_wanted to determine - * progress/completion. */ + * progress/completion. + */ int rc; - if (ksocknal_data.ksnd_stall_rx != 0) { + if (ksocknal_data.ksnd_stall_rx) { set_current_state(TASK_UNINTERRUPTIBLE); schedule_timeout(cfs_time_seconds(ksocknal_data.ksnd_stall_rx)); } rc = ksocknal_connsock_addref(conn); - if (rc != 0) { - LASSERT (conn->ksnc_closing); + if (rc) { + LASSERT(conn->ksnc_closing); return -ESHUTDOWN; } for (;;) { - if (conn->ksnc_rx_niov != 0) - rc = ksocknal_recv_iov (conn); + if (conn->ksnc_rx_niov) + rc = ksocknal_recv_iov(conn); else - rc = ksocknal_recv_kiov (conn); + rc = ksocknal_recv_kiov(conn); if (rc <= 0) { /* error/EOF or partial receive */ if (rc == -EAGAIN) { rc = 1; - } else if (rc == 0 && conn->ksnc_rx_started) { + } else if (!rc && conn->ksnc_rx_started) { /* EOF in the middle of a message */ rc = -EPROTO; } @@ -375,7 +380,7 @@ ksocknal_receive (ksock_conn_t *conn) /* Completed a fragment */ - if (conn->ksnc_rx_nob_wanted == 0) { + if (!conn->ksnc_rx_nob_wanted) { rc = 1; break; } @@ -386,36 +391,36 @@ ksocknal_receive (ksock_conn_t *conn) } void -ksocknal_tx_done (lnet_ni_t *ni, ksock_tx_t *tx) +ksocknal_tx_done(lnet_ni_t *ni, ksock_tx_t *tx) { lnet_msg_t *lnetmsg = tx->tx_lnetmsg; - int rc = (tx->tx_resid == 0 && !tx->tx_zc_aborted) ? 0 : -EIO; + int rc = (!tx->tx_resid && !tx->tx_zc_aborted) ? 0 : -EIO; - LASSERT(ni != NULL || tx->tx_conn != NULL); + LASSERT(ni || tx->tx_conn); - if (tx->tx_conn != NULL) + if (tx->tx_conn) ksocknal_conn_decref(tx->tx_conn); - if (ni == NULL && tx->tx_conn != NULL) + if (!ni && tx->tx_conn) ni = tx->tx_conn->ksnc_peer->ksnp_ni; - ksocknal_free_tx (tx); - if (lnetmsg != NULL) /* KSOCK_MSG_NOOP go without lnetmsg */ - lnet_finalize (ni, lnetmsg, rc); + ksocknal_free_tx(tx); + if (lnetmsg) /* KSOCK_MSG_NOOP go without lnetmsg */ + lnet_finalize(ni, lnetmsg, rc); } void -ksocknal_txlist_done (lnet_ni_t *ni, struct list_head *txlist, int error) +ksocknal_txlist_done(lnet_ni_t *ni, struct list_head *txlist, int error) { ksock_tx_t *tx; - while (!list_empty (txlist)) { + while (!list_empty(txlist)) { tx = list_entry(txlist->next, ksock_tx_t, tx_list); - if (error && tx->tx_lnetmsg != NULL) { + if (error && tx->tx_lnetmsg) { CNETERR("Deleting packet type %d len %d %s->%s\n", - le32_to_cpu (tx->tx_lnetmsg->msg_hdr.type), - le32_to_cpu (tx->tx_lnetmsg->msg_hdr.payload_length), + le32_to_cpu(tx->tx_lnetmsg->msg_hdr.type), + le32_to_cpu(tx->tx_lnetmsg->msg_hdr.payload_length), libcfs_nid2str(le64_to_cpu(tx->tx_lnetmsg->msg_hdr.src_nid)), libcfs_nid2str(le64_to_cpu(tx->tx_lnetmsg->msg_hdr.dest_nid))); } else if (error) { @@ -435,12 +440,14 @@ ksocknal_check_zc_req(ksock_tx_t *tx) ksock_conn_t *conn = tx->tx_conn; ksock_peer_t *peer = conn->ksnc_peer; - /* Set tx_msg.ksm_zc_cookies[0] to a unique non-zero cookie and add tx + /* + * Set tx_msg.ksm_zc_cookies[0] to a unique non-zero cookie and add tx * to ksnp_zc_req_list if some fragment of this message should be sent * zero-copy. Our peer will send an ACK containing this cookie when * she has received this message to tell us we can signal completion. * tx_msg.ksm_zc_cookies[0] remains non-zero while tx is on - * ksnp_zc_req_list. */ + * ksnp_zc_req_list. + */ LASSERT(tx->tx_msg.ksm_type != KSOCK_MSG_NOOP); LASSERT(tx->tx_zc_capable); @@ -450,9 +457,10 @@ ksocknal_check_zc_req(ksock_tx_t *tx) !conn->ksnc_zc_capable) return; - /* assign cookie and queue tx to pending list, it will be released when - * a matching ack is received. See ksocknal_handle_zcack() */ - + /* + * assign cookie and queue tx to pending list, it will be released when + * a matching ack is received. See ksocknal_handle_zcack() + */ ksocknal_tx_addref(tx); spin_lock(&peer->ksnp_lock); @@ -461,11 +469,11 @@ ksocknal_check_zc_req(ksock_tx_t *tx) tx->tx_deadline = cfs_time_shift(*ksocknal_tunables.ksnd_timeout); - LASSERT(tx->tx_msg.ksm_zc_cookies[0] == 0); + LASSERT(!tx->tx_msg.ksm_zc_cookies[0]); tx->tx_msg.ksm_zc_cookies[0] = peer->ksnp_zc_next_cookie++; - if (peer->ksnp_zc_next_cookie == 0) + if (!peer->ksnp_zc_next_cookie) peer->ksnp_zc_next_cookie = SOCKNAL_KEEPALIVE_PING + 1; list_add_tail(&tx->tx_zc_list, &peer->ksnp_zc_req_list); @@ -485,7 +493,7 @@ ksocknal_uncheck_zc_req(ksock_tx_t *tx) spin_lock(&peer->ksnp_lock); - if (tx->tx_msg.ksm_zc_cookies[0] == 0) { + if (!tx->tx_msg.ksm_zc_cookies[0]) { /* Not waiting for an ACK */ spin_unlock(&peer->ksnp_lock); return; @@ -500,20 +508,20 @@ ksocknal_uncheck_zc_req(ksock_tx_t *tx) } static int -ksocknal_process_transmit (ksock_conn_t *conn, ksock_tx_t *tx) +ksocknal_process_transmit(ksock_conn_t *conn, ksock_tx_t *tx) { int rc; if (tx->tx_zc_capable && !tx->tx_zc_checked) ksocknal_check_zc_req(tx); - rc = ksocknal_transmit (conn, tx); + rc = ksocknal_transmit(conn, tx); CDEBUG(D_NET, "send(%d) %d\n", tx->tx_resid, rc); - if (tx->tx_resid == 0) { + if (!tx->tx_resid) { /* Sent everything OK */ - LASSERT (rc == 0); + LASSERT(!rc); return 0; } @@ -532,13 +540,13 @@ ksocknal_process_transmit (ksock_conn_t *conn, ksock_tx_t *tx) spin_lock_bh(&ksocknal_data.ksnd_reaper_lock); /* enomem list takes over scheduler's ref... */ - LASSERT (conn->ksnc_tx_scheduled); + LASSERT(conn->ksnc_tx_scheduled); list_add_tail(&conn->ksnc_tx_list, - &ksocknal_data.ksnd_enomem_conns); + &ksocknal_data.ksnd_enomem_conns); if (!cfs_time_aftereq(cfs_time_add(cfs_time_current(), SOCKNAL_ENOMEM_RETRY), ksocknal_data.ksnd_reaper_waketime)) - wake_up (&ksocknal_data.ksnd_reaper_waitq); + wake_up(&ksocknal_data.ksnd_reaper_waitq); spin_unlock_bh(&ksocknal_data.ksnd_reaper_lock); return rc; @@ -569,21 +577,19 @@ ksocknal_process_transmit (ksock_conn_t *conn, ksock_tx_t *tx) ksocknal_uncheck_zc_req(tx); /* it's not an error if conn is being closed */ - ksocknal_close_conn_and_siblings (conn, - (conn->ksnc_closing) ? 0 : rc); + ksocknal_close_conn_and_siblings(conn, (conn->ksnc_closing) ? 0 : rc); return rc; } static void -ksocknal_launch_connection_locked (ksock_route_t *route) +ksocknal_launch_connection_locked(ksock_route_t *route) { - /* called holding write lock on ksnd_global_lock */ LASSERT(!route->ksnr_scheduled); LASSERT(!route->ksnr_connecting); - LASSERT((ksocknal_route_mask() & ~route->ksnr_connected) != 0); + LASSERT(ksocknal_route_mask() & ~route->ksnr_connected); route->ksnr_scheduled = 1; /* scheduling conn for connd */ ksocknal_route_addref(route); /* extra ref for connd */ @@ -591,14 +597,14 @@ ksocknal_launch_connection_locked (ksock_route_t *route) spin_lock_bh(&ksocknal_data.ksnd_connd_lock); list_add_tail(&route->ksnr_connd_list, - &ksocknal_data.ksnd_connd_routes); + &ksocknal_data.ksnd_connd_routes); wake_up(&ksocknal_data.ksnd_connd_waitq); spin_unlock_bh(&ksocknal_data.ksnd_connd_lock); } void -ksocknal_launch_all_connections_locked (ksock_peer_t *peer) +ksocknal_launch_all_connections_locked(ksock_peer_t *peer) { ksock_route_t *route; @@ -606,7 +612,7 @@ ksocknal_launch_all_connections_locked (ksock_peer_t *peer) for (;;) { /* launch any/all connections that need it */ route = ksocknal_find_connectable_route_locked(peer); - if (route == NULL) + if (!route) return; ksocknal_launch_connection_locked(route); @@ -623,15 +629,15 @@ ksocknal_find_conn_locked(ksock_peer_t *peer, ksock_tx_t *tx, int nonblk) int tnob = 0; int fnob = 0; - list_for_each (tmp, &peer->ksnp_conns) { + list_for_each(tmp, &peer->ksnp_conns) { ksock_conn_t *c = list_entry(tmp, ksock_conn_t, ksnc_list); int nob = atomic_read(&c->ksnc_tx_nob) + c->ksnc_sock->sk->sk_wmem_queued; int rc; LASSERT(!c->ksnc_closing); - LASSERT(c->ksnc_proto != NULL && - c->ksnc_proto->pro_match_tx != NULL); + LASSERT(c->ksnc_proto && + c->ksnc_proto->pro_match_tx); rc = c->ksnc_proto->pro_match_tx(c, tx, nonblk); @@ -642,7 +648,7 @@ ksocknal_find_conn_locked(ksock_peer_t *peer, ksock_tx_t *tx, int nonblk) continue; case SOCKNAL_MATCH_YES: /* typed connection */ - if (typed == NULL || tnob > nob || + if (!typed || tnob > nob || (tnob == nob && *ksocknal_tunables.ksnd_round_robin && cfs_time_after(typed->ksnc_tx_last_post, c->ksnc_tx_last_post))) { typed = c; @@ -651,7 +657,7 @@ ksocknal_find_conn_locked(ksock_peer_t *peer, ksock_tx_t *tx, int nonblk) break; case SOCKNAL_MATCH_MAY: /* fallback connection */ - if (fallback == NULL || fnob > nob || + if (!fallback || fnob > nob || (fnob == nob && *ksocknal_tunables.ksnd_round_robin && cfs_time_after(fallback->ksnc_tx_last_post, c->ksnc_tx_last_post))) { fallback = c; @@ -662,9 +668,9 @@ ksocknal_find_conn_locked(ksock_peer_t *peer, ksock_tx_t *tx, int nonblk) } /* prefer the typed selection */ - conn = (typed != NULL) ? typed : fallback; + conn = (typed) ? typed : fallback; - if (conn != NULL) + if (conn) conn->ksnc_tx_last_post = cfs_time_current(); return conn; @@ -675,48 +681,51 @@ ksocknal_tx_prep(ksock_conn_t *conn, ksock_tx_t *tx) { conn->ksnc_proto->pro_pack(tx); - atomic_add (tx->tx_nob, &conn->ksnc_tx_nob); + atomic_add(tx->tx_nob, &conn->ksnc_tx_nob); ksocknal_conn_addref(conn); /* +1 ref for tx */ tx->tx_conn = conn; } void -ksocknal_queue_tx_locked (ksock_tx_t *tx, ksock_conn_t *conn) +ksocknal_queue_tx_locked(ksock_tx_t *tx, ksock_conn_t *conn) { ksock_sched_t *sched = conn->ksnc_scheduler; ksock_msg_t *msg = &tx->tx_msg; ksock_tx_t *ztx = NULL; int bufnob = 0; - /* called holding global lock (read or irq-write) and caller may + /* + * called holding global lock (read or irq-write) and caller may * not have dropped this lock between finding conn and calling me, * so we don't need the {get,put}connsock dance to deref - * ksnc_sock... */ + * ksnc_sock... + */ LASSERT(!conn->ksnc_closing); CDEBUG(D_NET, "Sending to %s ip %pI4h:%d\n", - libcfs_id2str(conn->ksnc_peer->ksnp_id), - &conn->ksnc_ipaddr, - conn->ksnc_port); + libcfs_id2str(conn->ksnc_peer->ksnp_id), + &conn->ksnc_ipaddr, conn->ksnc_port); ksocknal_tx_prep(conn, tx); - /* Ensure the frags we've been given EXACTLY match the number of + /* + * Ensure the frags we've been given EXACTLY match the number of * bytes we want to send. Many TCP/IP stacks disregard any total * size parameters passed to them and just look at the frags. * * We always expect at least 1 mapped fragment containing the - * complete ksocknal message header. */ - LASSERT(lnet_iov_nob (tx->tx_niov, tx->tx_iov) + + * complete ksocknal message header. + */ + LASSERT(lnet_iov_nob(tx->tx_niov, tx->tx_iov) + lnet_kiov_nob(tx->tx_nkiov, tx->tx_kiov) == (unsigned int)tx->tx_nob); LASSERT(tx->tx_niov >= 1); LASSERT(tx->tx_resid == tx->tx_nob); - CDEBUG (D_NET, "Packet %p type %d, nob %d niov %d nkiov %d\n", - tx, (tx->tx_lnetmsg != NULL) ? tx->tx_lnetmsg->msg_hdr.type : - KSOCK_MSG_NOOP, - tx->tx_nob, tx->tx_niov, tx->tx_nkiov); + CDEBUG(D_NET, "Packet %p type %d, nob %d niov %d nkiov %d\n", + tx, (tx->tx_lnetmsg) ? tx->tx_lnetmsg->msg_hdr.type : + KSOCK_MSG_NOOP, + tx->tx_nob, tx->tx_niov, tx->tx_nkiov); /* * FIXME: SOCK_WMEM_QUEUED and SOCK_ERROR could block in __DARWIN8__ @@ -725,7 +734,7 @@ ksocknal_queue_tx_locked (ksock_tx_t *tx, ksock_conn_t *conn) bufnob = conn->ksnc_sock->sk->sk_wmem_queued; spin_lock_bh(&sched->kss_lock); - if (list_empty(&conn->ksnc_tx_queue) && bufnob == 0) { + if (list_empty(&conn->ksnc_tx_queue) && !bufnob) { /* First packet starts the timeout */ conn->ksnc_tx_deadline = cfs_time_shift(*ksocknal_tunables.ksnd_timeout); @@ -736,26 +745,30 @@ ksocknal_queue_tx_locked (ksock_tx_t *tx, ksock_conn_t *conn) } if (msg->ksm_type == KSOCK_MSG_NOOP) { - /* The packet is noop ZC ACK, try to piggyback the ack_cookie - * on a normal packet so I don't need to send it */ - LASSERT(msg->ksm_zc_cookies[1] != 0); - LASSERT(conn->ksnc_proto->pro_queue_tx_zcack != NULL); + /* + * The packet is noop ZC ACK, try to piggyback the ack_cookie + * on a normal packet so I don't need to send it + */ + LASSERT(msg->ksm_zc_cookies[1]); + LASSERT(conn->ksnc_proto->pro_queue_tx_zcack); if (conn->ksnc_proto->pro_queue_tx_zcack(conn, tx, 0)) ztx = tx; /* ZC ACK piggybacked on ztx release tx later */ } else { - /* It's a normal packet - can it piggback a noop zc-ack that - * has been queued already? */ - LASSERT(msg->ksm_zc_cookies[1] == 0); - LASSERT(conn->ksnc_proto->pro_queue_tx_msg != NULL); + /* + * It's a normal packet - can it piggback a noop zc-ack that + * has been queued already? + */ + LASSERT(!msg->ksm_zc_cookies[1]); + LASSERT(conn->ksnc_proto->pro_queue_tx_msg); ztx = conn->ksnc_proto->pro_queue_tx_msg(conn, tx); /* ztx will be released later */ } - if (ztx != NULL) { - atomic_sub (ztx->tx_nob, &conn->ksnc_tx_nob); + if (ztx) { + atomic_sub(ztx->tx_nob, &conn->ksnc_tx_nob); list_add_tail(&ztx->tx_list, &sched->kss_zombie_noop_txs); } @@ -763,24 +776,23 @@ ksocknal_queue_tx_locked (ksock_tx_t *tx, ksock_conn_t *conn) !conn->ksnc_tx_scheduled) { /* not scheduled to send */ /* +1 ref for scheduler */ ksocknal_conn_addref(conn); - list_add_tail (&conn->ksnc_tx_list, - &sched->kss_tx_conns); + list_add_tail(&conn->ksnc_tx_list, &sched->kss_tx_conns); conn->ksnc_tx_scheduled = 1; - wake_up (&sched->kss_waitq); + wake_up(&sched->kss_waitq); } spin_unlock_bh(&sched->kss_lock); } ksock_route_t * -ksocknal_find_connectable_route_locked (ksock_peer_t *peer) +ksocknal_find_connectable_route_locked(ksock_peer_t *peer) { unsigned long now = cfs_time_current(); struct list_head *tmp; ksock_route_t *route; - list_for_each (tmp, &peer->ksnp_routes) { - route = list_entry (tmp, ksock_route_t, ksnr_list); + list_for_each(tmp, &peer->ksnp_routes) { + route = list_entry(tmp, ksock_route_t, ksnr_list); LASSERT(!route->ksnr_connecting || route->ksnr_scheduled); @@ -788,10 +800,10 @@ ksocknal_find_connectable_route_locked (ksock_peer_t *peer) continue; /* all route types connected ? */ - if ((ksocknal_route_mask() & ~route->ksnr_connected) == 0) + if (!(ksocknal_route_mask() & ~route->ksnr_connected)) continue; - if (!(route->ksnr_retry_interval == 0 || /* first attempt */ + if (!(!route->ksnr_retry_interval || /* first attempt */ cfs_time_aftereq(now, route->ksnr_timeout))) { CDEBUG(D_NET, "Too soon to retry route %pI4h (cnted %d, interval %ld, %ld secs later)\n", @@ -809,13 +821,13 @@ ksocknal_find_connectable_route_locked (ksock_peer_t *peer) } ksock_route_t * -ksocknal_find_connecting_route_locked (ksock_peer_t *peer) +ksocknal_find_connecting_route_locked(ksock_peer_t *peer) { struct list_head *tmp; ksock_route_t *route; - list_for_each (tmp, &peer->ksnp_routes) { - route = list_entry (tmp, ksock_route_t, ksnr_list); + list_for_each(tmp, &peer->ksnp_routes) { + route = list_entry(tmp, ksock_route_t, ksnr_list); LASSERT(!route->ksnr_connecting || route->ksnr_scheduled); @@ -827,7 +839,7 @@ ksocknal_find_connecting_route_locked (ksock_peer_t *peer) } int -ksocknal_launch_packet (lnet_ni_t *ni, ksock_tx_t *tx, lnet_process_id_t id) +ksocknal_launch_packet(lnet_ni_t *ni, ksock_tx_t *tx, lnet_process_id_t id) { ksock_peer_t *peer; ksock_conn_t *conn; @@ -835,21 +847,23 @@ ksocknal_launch_packet (lnet_ni_t *ni, ksock_tx_t *tx, lnet_process_id_t id) int retry; int rc; - LASSERT(tx->tx_conn == NULL); + LASSERT(!tx->tx_conn); g_lock = &ksocknal_data.ksnd_global_lock; for (retry = 0;; retry = 1) { read_lock(g_lock); peer = ksocknal_find_peer_locked(ni, id); - if (peer != NULL) { - if (ksocknal_find_connectable_route_locked(peer) == NULL) { + if (peer) { + if (!ksocknal_find_connectable_route_locked(peer)) { conn = ksocknal_find_conn_locked(peer, tx, tx->tx_nonblk); - if (conn != NULL) { - /* I've got no routes that need to be + if (conn) { + /* + * I've got no routes that need to be * connecting and I do have an actual - * connection... */ - ksocknal_queue_tx_locked (tx, conn); + * connection... + */ + ksocknal_queue_tx_locked(tx, conn); read_unlock(g_lock); return 0; } @@ -862,12 +876,12 @@ ksocknal_launch_packet (lnet_ni_t *ni, ksock_tx_t *tx, lnet_process_id_t id) write_lock_bh(g_lock); peer = ksocknal_find_peer_locked(ni, id); - if (peer != NULL) + if (peer) break; write_unlock_bh(g_lock); - if ((id.pid & LNET_PID_USERFLAG) != 0) { + if (id.pid & LNET_PID_USERFLAG) { CERROR("Refusing to create a connection to userspace process %s\n", libcfs_id2str(id)); return -EHOSTUNREACH; @@ -881,7 +895,7 @@ ksocknal_launch_packet (lnet_ni_t *ni, ksock_tx_t *tx, lnet_process_id_t id) rc = ksocknal_add_peer(ni, id, LNET_NIDADDR(id.nid), lnet_acceptor_port()); - if (rc != 0) { + if (rc) { CERROR("Can't add peer %s: %d\n", libcfs_id2str(id), rc); return rc; @@ -891,21 +905,21 @@ ksocknal_launch_packet (lnet_ni_t *ni, ksock_tx_t *tx, lnet_process_id_t id) ksocknal_launch_all_connections_locked(peer); conn = ksocknal_find_conn_locked(peer, tx, tx->tx_nonblk); - if (conn != NULL) { + if (conn) { /* Connection exists; queue message on it */ - ksocknal_queue_tx_locked (tx, conn); + ksocknal_queue_tx_locked(tx, conn); write_unlock_bh(g_lock); return 0; } if (peer->ksnp_accepting > 0 || - ksocknal_find_connecting_route_locked (peer) != NULL) { + ksocknal_find_connecting_route_locked(peer)) { /* the message is going to be pinned to the peer */ tx->tx_deadline = cfs_time_shift(*ksocknal_tunables.ksnd_timeout); /* Queue the message until a connection is established */ - list_add_tail (&tx->tx_list, &peer->ksnp_tx_queue); + list_add_tail(&tx->tx_list, &peer->ksnp_tx_queue); write_unlock_bh(g_lock); return 0; } @@ -932,19 +946,20 @@ ksocknal_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) int desc_size; int rc; - /* NB 'private' is different depending on what we're sending. - * Just ignore it... */ - + /* + * NB 'private' is different depending on what we're sending. + * Just ignore it... + */ CDEBUG(D_NET, "sending %u bytes in %d frags to %s\n", payload_nob, payload_niov, libcfs_id2str(target)); - LASSERT(payload_nob == 0 || payload_niov > 0); + LASSERT(!payload_nob || payload_niov > 0); LASSERT(payload_niov <= LNET_MAX_IOV); /* payload is either all vaddrs or all pages */ - LASSERT (!(payload_kiov != NULL && payload_iov != NULL)); - LASSERT (!in_interrupt ()); + LASSERT(!(payload_kiov && payload_iov)); + LASSERT(!in_interrupt()); - if (payload_iov != NULL) + if (payload_iov) desc_size = offsetof(ksock_tx_t, tx_frags.virt.iov[1 + payload_niov]); else @@ -954,7 +969,7 @@ ksocknal_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) if (lntmsg->msg_vmflush) mpflag = cfs_memory_pressure_get_and_set(); tx = ksocknal_alloc_tx(KSOCK_MSG_LNET, desc_size); - if (tx == NULL) { + if (!tx) { CERROR("Can't allocate tx desc type %d size %d\n", type, desc_size); if (lntmsg->msg_vmflush) @@ -965,7 +980,7 @@ ksocknal_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) tx->tx_conn = NULL; /* set when assigned a conn */ tx->tx_lnetmsg = lntmsg; - if (payload_iov != NULL) { + if (payload_iov) { tx->tx_kiov = NULL; tx->tx_nkiov = 0; tx->tx_iov = tx->tx_frags.virt.iov; @@ -992,7 +1007,7 @@ ksocknal_send(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg) if (!mpflag) cfs_memory_pressure_restore(mpflag); - if (rc == 0) + if (!rc) return 0; ksocknal_free_tx(tx); @@ -1014,7 +1029,7 @@ ksocknal_thread_start(int (*fn)(void *arg), void *arg, char *name) } void -ksocknal_thread_fini (void) +ksocknal_thread_fini(void) { write_lock_bh(&ksocknal_data.ksnd_global_lock); ksocknal_data.ksnd_nthreads--; @@ -1022,7 +1037,7 @@ ksocknal_thread_fini (void) } int -ksocknal_new_packet (ksock_conn_t *conn, int nob_to_skip) +ksocknal_new_packet(ksock_conn_t *conn, int nob_to_skip) { static char ksocknal_slop_buffer[4096]; @@ -1030,14 +1045,14 @@ ksocknal_new_packet (ksock_conn_t *conn, int nob_to_skip) unsigned int niov; int skipped; - LASSERT(conn->ksnc_proto != NULL); + LASSERT(conn->ksnc_proto); - if ((*ksocknal_tunables.ksnd_eager_ack & conn->ksnc_type) != 0) { + if (*ksocknal_tunables.ksnd_eager_ack & conn->ksnc_type) { /* Remind the socket to ack eagerly... */ ksocknal_lib_eager_ack(conn); } - if (nob_to_skip == 0) { /* right at next packet boundary now */ + if (!nob_to_skip) { /* right at next packet boundary now */ conn->ksnc_rx_started = 0; mb(); /* racing with timeout thread */ @@ -1061,11 +1076,11 @@ ksocknal_new_packet (ksock_conn_t *conn, int nob_to_skip) conn->ksnc_rx_iov = (struct kvec *)&conn->ksnc_rx_iov_space; conn->ksnc_rx_iov[0].iov_base = &conn->ksnc_msg.ksm_u.lnetmsg; - conn->ksnc_rx_iov[0].iov_len = sizeof (lnet_hdr_t); + conn->ksnc_rx_iov[0].iov_len = sizeof(lnet_hdr_t); break; default: - LBUG (); + LBUG(); } conn->ksnc_rx_niov = 1; @@ -1075,9 +1090,10 @@ ksocknal_new_packet (ksock_conn_t *conn, int nob_to_skip) return 1; } - /* Set up to skip as much as possible now. If there's more left - * (ran out of iov entries) we'll get called again */ - + /* + * Set up to skip as much as possible now. If there's more left + * (ran out of iov entries) we'll get called again + */ conn->ksnc_rx_state = SOCKNAL_RX_SLOP; conn->ksnc_rx_nob_left = nob_to_skip; conn->ksnc_rx_iov = (struct kvec *)&conn->ksnc_rx_iov_space; @@ -1093,8 +1109,8 @@ ksocknal_new_packet (ksock_conn_t *conn, int nob_to_skip) skipped += nob; nob_to_skip -= nob; - } while (nob_to_skip != 0 && /* mustn't overflow conn's rx iov */ - niov < sizeof(conn->ksnc_rx_iov_space) / sizeof (struct iovec)); + } while (nob_to_skip && /* mustn't overflow conn's rx iov */ + niov < sizeof(conn->ksnc_rx_iov_space) / sizeof(struct iovec)); conn->ksnc_rx_niov = niov; conn->ksnc_rx_kiov = NULL; @@ -1104,13 +1120,13 @@ ksocknal_new_packet (ksock_conn_t *conn, int nob_to_skip) } static int -ksocknal_process_receive (ksock_conn_t *conn) +ksocknal_process_receive(ksock_conn_t *conn) { lnet_hdr_t *lhdr; lnet_process_id_t *id; int rc; - LASSERT (atomic_read(&conn->ksnc_conn_refcount) > 0); + LASSERT(atomic_read(&conn->ksnc_conn_refcount) > 0); /* NB: sched lock NOT held */ /* SOCKNAL_RX_LNET_HEADER is here for backward compatibility */ @@ -1119,13 +1135,13 @@ ksocknal_process_receive (ksock_conn_t *conn) conn->ksnc_rx_state == SOCKNAL_RX_LNET_HEADER || conn->ksnc_rx_state == SOCKNAL_RX_SLOP); again: - if (conn->ksnc_rx_nob_wanted != 0) { + if (conn->ksnc_rx_nob_wanted) { rc = ksocknal_receive(conn); if (rc <= 0) { - LASSERT (rc != -EAGAIN); + LASSERT(rc != -EAGAIN); - if (rc == 0) + if (!rc) CDEBUG(D_NET, "[%p] EOF from %s ip %pI4h:%d\n", conn, libcfs_id2str(conn->ksnc_peer->ksnp_id), @@ -1139,12 +1155,12 @@ ksocknal_process_receive (ksock_conn_t *conn) conn->ksnc_port); /* it's not an error if conn is being closed */ - ksocknal_close_conn_and_siblings (conn, - (conn->ksnc_closing) ? 0 : rc); - return (rc == 0 ? -ESHUTDOWN : rc); + ksocknal_close_conn_and_siblings(conn, + (conn->ksnc_closing) ? 0 : rc); + return (!rc ? -ESHUTDOWN : rc); } - if (conn->ksnc_rx_nob_wanted != 0) { + if (conn->ksnc_rx_nob_wanted) { /* short read */ return -EAGAIN; } @@ -1169,7 +1185,7 @@ ksocknal_process_receive (ksock_conn_t *conn) } if (conn->ksnc_msg.ksm_type == KSOCK_MSG_NOOP && - conn->ksnc_msg.ksm_csum != 0 && /* has checksum */ + conn->ksnc_msg.ksm_csum && /* has checksum */ conn->ksnc_msg.ksm_csum != conn->ksnc_rx_csum) { /* NOOP Checksum error */ CERROR("%s: Checksum error, wire:0x%08X data:0x%08X\n", @@ -1180,10 +1196,10 @@ ksocknal_process_receive (ksock_conn_t *conn) return -EIO; } - if (conn->ksnc_msg.ksm_zc_cookies[1] != 0) { + if (conn->ksnc_msg.ksm_zc_cookies[1]) { __u64 cookie = 0; - LASSERT (conn->ksnc_proto != &ksocknal_protocol_v1x); + LASSERT(conn->ksnc_proto != &ksocknal_protocol_v1x); if (conn->ksnc_msg.ksm_type == KSOCK_MSG_NOOP) cookie = conn->ksnc_msg.ksm_zc_cookies[0]; @@ -1191,7 +1207,7 @@ ksocknal_process_receive (ksock_conn_t *conn) rc = conn->ksnc_proto->pro_handle_zcack(conn, cookie, conn->ksnc_msg.ksm_zc_cookies[1]); - if (rc != 0) { + if (rc) { CERROR("%s: Unknown ZC-ACK cookie: %llu, %llu\n", libcfs_id2str(conn->ksnc_peer->ksnp_id), cookie, conn->ksnc_msg.ksm_zc_cookies[1]); @@ -1202,7 +1218,7 @@ ksocknal_process_receive (ksock_conn_t *conn) } if (conn->ksnc_msg.ksm_type == KSOCK_MSG_NOOP) { - ksocknal_new_packet (conn, 0); + ksocknal_new_packet(conn, 0); return 0; /* NOOP is done and just return */ } @@ -1224,7 +1240,7 @@ ksocknal_process_receive (ksock_conn_t *conn) /* unpack message header */ conn->ksnc_proto->pro_unpack(&conn->ksnc_msg); - if ((conn->ksnc_peer->ksnp_id.pid & LNET_PID_USERFLAG) != 0) { + if (conn->ksnc_peer->ksnp_id.pid & LNET_PID_USERFLAG) { /* Userspace peer */ lhdr = &conn->ksnc_msg.ksm_u.lnetmsg.ksnm_hdr; id = &conn->ksnc_peer->ksnp_id; @@ -1243,14 +1259,14 @@ ksocknal_process_receive (ksock_conn_t *conn) if (rc < 0) { /* I just received garbage: give up on this conn */ ksocknal_new_packet(conn, 0); - ksocknal_close_conn_and_siblings (conn, rc); + ksocknal_close_conn_and_siblings(conn, rc); ksocknal_conn_decref(conn); return -EPROTO; } /* I'm racing with ksocknal_recv() */ - LASSERT (conn->ksnc_rx_state == SOCKNAL_RX_PARSE || - conn->ksnc_rx_state == SOCKNAL_RX_LNET_PAYLOAD); + LASSERT(conn->ksnc_rx_state == SOCKNAL_RX_PARSE || + conn->ksnc_rx_state == SOCKNAL_RX_LNET_PAYLOAD); if (conn->ksnc_rx_state != SOCKNAL_RX_LNET_PAYLOAD) return 0; @@ -1262,8 +1278,8 @@ ksocknal_process_receive (ksock_conn_t *conn) /* payload all received */ rc = 0; - if (conn->ksnc_rx_nob_left == 0 && /* not truncating */ - conn->ksnc_msg.ksm_csum != 0 && /* has checksum */ + if (!conn->ksnc_rx_nob_left && /* not truncating */ + conn->ksnc_msg.ksm_csum && /* has checksum */ conn->ksnc_msg.ksm_csum != conn->ksnc_rx_csum) { CERROR("%s: Checksum error, wire:0x%08X data:0x%08X\n", libcfs_id2str(conn->ksnc_peer->ksnp_id), @@ -1271,7 +1287,7 @@ ksocknal_process_receive (ksock_conn_t *conn) rc = -EIO; } - if (rc == 0 && conn->ksnc_msg.ksm_zc_cookies[0] != 0) { + if (!rc && conn->ksnc_msg.ksm_zc_cookies[0]) { LASSERT(conn->ksnc_proto != &ksocknal_protocol_v1x); lhdr = &conn->ksnc_msg.ksm_u.lnetmsg.ksnm_hdr; @@ -1285,16 +1301,16 @@ ksocknal_process_receive (ksock_conn_t *conn) lnet_finalize(conn->ksnc_peer->ksnp_ni, conn->ksnc_cookie, rc); - if (rc != 0) { + if (rc) { ksocknal_new_packet(conn, 0); - ksocknal_close_conn_and_siblings (conn, rc); + ksocknal_close_conn_and_siblings(conn, rc); return -EPROTO; } /* Fall through */ case SOCKNAL_RX_SLOP: /* starting new packet? */ - if (ksocknal_new_packet (conn, conn->ksnc_rx_nob_left)) + if (ksocknal_new_packet(conn, conn->ksnc_rx_nob_left)) return 0; /* come back later */ goto again; /* try to finish reading slop now */ @@ -1308,9 +1324,9 @@ ksocknal_process_receive (ksock_conn_t *conn) } int -ksocknal_recv (lnet_ni_t *ni, void *private, lnet_msg_t *msg, int delayed, - unsigned int niov, struct kvec *iov, lnet_kiov_t *kiov, - unsigned int offset, unsigned int mlen, unsigned int rlen) +ksocknal_recv(lnet_ni_t *ni, void *private, lnet_msg_t *msg, int delayed, + unsigned int niov, struct kvec *iov, lnet_kiov_t *kiov, + unsigned int offset, unsigned int mlen, unsigned int rlen) { ksock_conn_t *conn = private; ksock_sched_t *sched = conn->ksnc_scheduler; @@ -1322,7 +1338,7 @@ ksocknal_recv (lnet_ni_t *ni, void *private, lnet_msg_t *msg, int delayed, conn->ksnc_rx_nob_wanted = mlen; conn->ksnc_rx_nob_left = rlen; - if (mlen == 0 || iov != NULL) { + if (!mlen || iov) { conn->ksnc_rx_nkiov = 0; conn->ksnc_rx_kiov = NULL; conn->ksnc_rx_iov = conn->ksnc_rx_iov_space.iov; @@ -1349,8 +1365,8 @@ ksocknal_recv (lnet_ni_t *ni, void *private, lnet_msg_t *msg, int delayed, switch (conn->ksnc_rx_state) { case SOCKNAL_RX_PARSE_WAIT: list_add_tail(&conn->ksnc_rx_list, &sched->kss_rx_conns); - wake_up (&sched->kss_waitq); - LASSERT (conn->ksnc_rx_ready); + wake_up(&sched->kss_waitq); + LASSERT(conn->ksnc_rx_ready); break; case SOCKNAL_RX_PARSE: @@ -1396,7 +1412,7 @@ int ksocknal_scheduler(void *arg) cfs_block_allsigs(); rc = cfs_cpt_bind(lnet_cpt_table(), info->ksi_cpt); - if (rc != 0) { + if (rc) { CERROR("Can't set CPT affinity to %d: %d\n", info->ksi_cpt, rc); } @@ -1408,18 +1424,20 @@ int ksocknal_scheduler(void *arg) /* Ensure I progress everything semi-fairly */ - if (!list_empty (&sched->kss_rx_conns)) { + if (!list_empty(&sched->kss_rx_conns)) { conn = list_entry(sched->kss_rx_conns.next, - ksock_conn_t, ksnc_rx_list); + ksock_conn_t, ksnc_rx_list); list_del(&conn->ksnc_rx_list); LASSERT(conn->ksnc_rx_scheduled); LASSERT(conn->ksnc_rx_ready); - /* clear rx_ready in case receive isn't complete. + /* + * clear rx_ready in case receive isn't complete. * Do it BEFORE we call process_recv, since * data_ready can set it any time after we release - * kss_lock. */ + * kss_lock. + */ conn->ksnc_rx_ready = 0; spin_unlock_bh(&sched->kss_lock); @@ -1431,18 +1449,20 @@ int ksocknal_scheduler(void *arg) LASSERT(conn->ksnc_rx_scheduled); /* Did process_receive get everything it wanted? */ - if (rc == 0) + if (!rc) conn->ksnc_rx_ready = 1; if (conn->ksnc_rx_state == SOCKNAL_RX_PARSE) { - /* Conn blocked waiting for ksocknal_recv() + /* + * Conn blocked waiting for ksocknal_recv() * I change its state (under lock) to signal - * it can be rescheduled */ + * it can be rescheduled + */ conn->ksnc_rx_state = SOCKNAL_RX_PARSE_WAIT; } else if (conn->ksnc_rx_ready) { /* reschedule for rx */ - list_add_tail (&conn->ksnc_rx_list, - &sched->kss_rx_conns); + list_add_tail(&conn->ksnc_rx_list, + &sched->kss_rx_conns); } else { conn->ksnc_rx_scheduled = 0; /* drop my ref */ @@ -1452,25 +1472,24 @@ int ksocknal_scheduler(void *arg) did_something = 1; } - if (!list_empty (&sched->kss_tx_conns)) { + if (!list_empty(&sched->kss_tx_conns)) { LIST_HEAD(zlist); if (!list_empty(&sched->kss_zombie_noop_txs)) { - list_add(&zlist, - &sched->kss_zombie_noop_txs); + list_add(&zlist, &sched->kss_zombie_noop_txs); list_del_init(&sched->kss_zombie_noop_txs); } conn = list_entry(sched->kss_tx_conns.next, - ksock_conn_t, ksnc_tx_list); - list_del (&conn->ksnc_tx_list); + ksock_conn_t, ksnc_tx_list); + list_del(&conn->ksnc_tx_list); LASSERT(conn->ksnc_tx_scheduled); LASSERT(conn->ksnc_tx_ready); LASSERT(!list_empty(&conn->ksnc_tx_queue)); tx = list_entry(conn->ksnc_tx_queue.next, - ksock_tx_t, tx_list); + ksock_tx_t, tx_list); if (conn->ksnc_tx_carrier == tx) ksocknal_next_tx_carrier(conn); @@ -1478,16 +1497,20 @@ int ksocknal_scheduler(void *arg) /* dequeue now so empty list => more to send */ list_del(&tx->tx_list); - /* Clear tx_ready in case send isn't complete. Do + /* + * Clear tx_ready in case send isn't complete. Do * it BEFORE we call process_transmit, since * write_space can set it any time after we release - * kss_lock. */ + * kss_lock. + */ conn->ksnc_tx_ready = 0; spin_unlock_bh(&sched->kss_lock); if (!list_empty(&zlist)) { - /* free zombie noop txs, it's fast because - * noop txs are just put in freelist */ + /* + * free zombie noop txs, it's fast because + * noop txs are just put in freelist + */ ksocknal_txlist_done(NULL, &zlist, 0); } @@ -1496,8 +1519,7 @@ int ksocknal_scheduler(void *arg) if (rc == -ENOMEM || rc == -EAGAIN) { /* Incomplete send: replace tx on HEAD of tx_queue */ spin_lock_bh(&sched->kss_lock); - list_add(&tx->tx_list, - &conn->ksnc_tx_queue); + list_add(&tx->tx_list, &conn->ksnc_tx_queue); } else { /* Complete send; tx -ref */ ksocknal_tx_decref(tx); @@ -1508,13 +1530,15 @@ int ksocknal_scheduler(void *arg) } if (rc == -ENOMEM) { - /* Do nothing; after a short timeout, this - * conn will be reposted on kss_tx_conns. */ + /* + * Do nothing; after a short timeout, this + * conn will be reposted on kss_tx_conns. + */ } else if (conn->ksnc_tx_ready && !list_empty(&conn->ksnc_tx_queue)) { /* reschedule for tx */ list_add_tail(&conn->ksnc_tx_list, - &sched->kss_tx_conns); + &sched->kss_tx_conns); } else { conn->ksnc_tx_scheduled = 0; /* drop my ref */ @@ -1533,7 +1557,7 @@ int ksocknal_scheduler(void *arg) rc = wait_event_interruptible_exclusive( sched->kss_waitq, !ksocknal_sched_cansleep(sched)); - LASSERT (rc == 0); + LASSERT(!rc); } else { cond_resched(); } @@ -1551,7 +1575,7 @@ int ksocknal_scheduler(void *arg) * Add connection to kss_rx_conns of scheduler * and wakeup the scheduler. */ -void ksocknal_read_callback (ksock_conn_t *conn) +void ksocknal_read_callback(ksock_conn_t *conn) { ksock_sched_t *sched; @@ -1562,13 +1586,12 @@ void ksocknal_read_callback (ksock_conn_t *conn) conn->ksnc_rx_ready = 1; if (!conn->ksnc_rx_scheduled) { /* not being progressed */ - list_add_tail(&conn->ksnc_rx_list, - &sched->kss_rx_conns); + list_add_tail(&conn->ksnc_rx_list, &sched->kss_rx_conns); conn->ksnc_rx_scheduled = 1; /* extra ref for scheduler */ ksocknal_conn_addref(conn); - wake_up (&sched->kss_waitq); + wake_up(&sched->kss_waitq); } spin_unlock_bh(&sched->kss_lock); } @@ -1577,7 +1600,7 @@ void ksocknal_read_callback (ksock_conn_t *conn) * Add connection to kss_tx_conns of scheduler * and wakeup the scheduler. */ -void ksocknal_write_callback (ksock_conn_t *conn) +void ksocknal_write_callback(ksock_conn_t *conn) { ksock_sched_t *sched; @@ -1589,20 +1612,19 @@ void ksocknal_write_callback (ksock_conn_t *conn) if (!conn->ksnc_tx_scheduled && /* not being progressed */ !list_empty(&conn->ksnc_tx_queue)) { /* packets to send */ - list_add_tail (&conn->ksnc_tx_list, - &sched->kss_tx_conns); + list_add_tail(&conn->ksnc_tx_list, &sched->kss_tx_conns); conn->ksnc_tx_scheduled = 1; /* extra ref for scheduler */ ksocknal_conn_addref(conn); - wake_up (&sched->kss_waitq); + wake_up(&sched->kss_waitq); } spin_unlock_bh(&sched->kss_lock); } static ksock_proto_t * -ksocknal_parse_proto_version (ksock_hello_msg_t *hello) +ksocknal_parse_proto_version(ksock_hello_msg_t *hello) { __u32 version = 0; @@ -1611,7 +1633,7 @@ ksocknal_parse_proto_version (ksock_hello_msg_t *hello) else if (hello->kshm_magic == __swab32(LNET_PROTO_MAGIC)) version = __swab32(hello->kshm_version); - if (version != 0) { + if (version) { #if SOCKNAL_VERSION_DEBUG if (*ksocknal_tunables.ksnd_protocol == 1) return NULL; @@ -1632,11 +1654,11 @@ ksocknal_parse_proto_version (ksock_hello_msg_t *hello) if (hello->kshm_magic == le32_to_cpu(LNET_PROTO_TCP_MAGIC)) { lnet_magicversion_t *hmv = (lnet_magicversion_t *)hello; - CLASSERT(sizeof (lnet_magicversion_t) == - offsetof (ksock_hello_msg_t, kshm_src_nid)); + CLASSERT(sizeof(lnet_magicversion_t) == + offsetof(ksock_hello_msg_t, kshm_src_nid)); - if (hmv->version_major == cpu_to_le16 (KSOCK_PROTO_V1_MAJOR) && - hmv->version_minor == cpu_to_le16 (KSOCK_PROTO_V1_MINOR)) + if (hmv->version_major == cpu_to_le16(KSOCK_PROTO_V1_MAJOR) && + hmv->version_minor == cpu_to_le16(KSOCK_PROTO_V1_MINOR)) return &ksocknal_protocol_v1x; } @@ -1644,8 +1666,8 @@ ksocknal_parse_proto_version (ksock_hello_msg_t *hello) } int -ksocknal_send_hello (lnet_ni_t *ni, ksock_conn_t *conn, - lnet_nid_t peer_nid, ksock_hello_msg_t *hello) +ksocknal_send_hello(lnet_ni_t *ni, ksock_conn_t *conn, + lnet_nid_t peer_nid, ksock_hello_msg_t *hello) { /* CAVEAT EMPTOR: this byte flips 'ipaddrs' */ ksock_net_t *net = (ksock_net_t *)ni->ni_data; @@ -1653,7 +1675,7 @@ ksocknal_send_hello (lnet_ni_t *ni, ksock_conn_t *conn, LASSERT(hello->kshm_nips <= LNET_MAX_INTERFACES); /* rely on caller to hold a ref on socket so it wouldn't disappear */ - LASSERT(conn->ksnc_proto != NULL); + LASSERT(conn->ksnc_proto); hello->kshm_src_nid = ni->ni_nid; hello->kshm_dst_nid = peer_nid; @@ -1682,9 +1704,9 @@ ksocknal_invert_type(int type) } int -ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, - ksock_hello_msg_t *hello, lnet_process_id_t *peerid, - __u64 *incarnation) +ksocknal_recv_hello(lnet_ni_t *ni, ksock_conn_t *conn, + ksock_hello_msg_t *hello, lnet_process_id_t *peerid, + __u64 *incarnation) { /* Return < 0 fatal error * 0 success @@ -1692,7 +1714,7 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, * EPROTO protocol version mismatch */ struct socket *sock = conn->ksnc_sock; - int active = (conn->ksnc_proto != NULL); + int active = !!conn->ksnc_proto; int timeout; int proto_match; int rc; @@ -1705,20 +1727,20 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, timeout = active ? *ksocknal_tunables.ksnd_timeout : lnet_acceptor_timeout(); - rc = lnet_sock_read(sock, &hello->kshm_magic, sizeof (hello->kshm_magic), timeout); - if (rc != 0) { + rc = lnet_sock_read(sock, &hello->kshm_magic, sizeof(hello->kshm_magic), timeout); + if (rc) { CERROR("Error %d reading HELLO from %pI4h\n", - rc, &conn->ksnc_ipaddr); - LASSERT (rc < 0); + rc, &conn->ksnc_ipaddr); + LASSERT(rc < 0); return rc; } if (hello->kshm_magic != LNET_PROTO_MAGIC && hello->kshm_magic != __swab32(LNET_PROTO_MAGIC) && - hello->kshm_magic != le32_to_cpu (LNET_PROTO_TCP_MAGIC)) { + hello->kshm_magic != le32_to_cpu(LNET_PROTO_TCP_MAGIC)) { /* Unexpected magic! */ CERROR("Bad magic(1) %#08x (%#08x expected) from %pI4h\n", - __cpu_to_le32 (hello->kshm_magic), + __cpu_to_le32(hello->kshm_magic), LNET_PROTO_TCP_MAGIC, &conn->ksnc_ipaddr); return -EPROTO; @@ -1726,15 +1748,15 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, rc = lnet_sock_read(sock, &hello->kshm_version, sizeof(hello->kshm_version), timeout); - if (rc != 0) { + if (rc) { CERROR("Error %d reading HELLO from %pI4h\n", - rc, &conn->ksnc_ipaddr); + rc, &conn->ksnc_ipaddr); LASSERT(rc < 0); return rc; } proto = ksocknal_parse_proto_version(hello); - if (proto == NULL) { + if (!proto) { if (!active) { /* unknown protocol from peer, tell peer my protocol */ conn->ksnc_proto = &ksocknal_protocol_v3x; @@ -1760,7 +1782,7 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, /* receive the rest of hello message anyway */ rc = conn->ksnc_proto->pro_recv_hello(conn, hello, timeout); - if (rc != 0) { + if (rc) { CERROR("Error %d reading or checking hello from from %pI4h\n", rc, &conn->ksnc_ipaddr); LASSERT(rc < 0); @@ -1792,8 +1814,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, conn->ksnc_type = ksocknal_invert_type(hello->kshm_ctype); if (conn->ksnc_type == SOCKLND_CONN_NONE) { CERROR("Unexpected type %d from %s ip %pI4h\n", - hello->kshm_ctype, libcfs_id2str(*peerid), - &conn->ksnc_ipaddr); + hello->kshm_ctype, libcfs_id2str(*peerid), + &conn->ksnc_ipaddr); return -EPROTO; } @@ -1816,9 +1838,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, if (ksocknal_invert_type(hello->kshm_ctype) != conn->ksnc_type) { CERROR("Mismatched types: me %d, %s ip %pI4h %d\n", - conn->ksnc_type, libcfs_id2str(*peerid), - &conn->ksnc_ipaddr, - hello->kshm_ctype); + conn->ksnc_type, libcfs_id2str(*peerid), + &conn->ksnc_ipaddr, hello->kshm_ctype); return -EPROTO; } @@ -1826,7 +1847,7 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, } static int -ksocknal_connect (ksock_route_t *route) +ksocknal_connect(ksock_route_t *route) { LIST_HEAD(zombies); ksock_peer_t *peer = route->ksnr_peer; @@ -1850,10 +1871,12 @@ ksocknal_connect (ksock_route_t *route) for (;;) { wanted = ksocknal_route_mask() & ~route->ksnr_connected; - /* stop connecting if peer/route got closed under me, or - * route got connected while queued */ + /* + * stop connecting if peer/route got closed under me, or + * route got connected while queued + */ if (peer->ksnp_closing || route->ksnr_deleted || - wanted == 0) { + !wanted) { retry_later = 0; break; } @@ -1869,14 +1892,14 @@ ksocknal_connect (ksock_route_t *route) if (retry_later) /* needs reschedule */ break; - if ((wanted & (1 << SOCKLND_CONN_ANY)) != 0) { + if (wanted & (1 << SOCKLND_CONN_ANY)) { type = SOCKLND_CONN_ANY; - } else if ((wanted & (1 << SOCKLND_CONN_CONTROL)) != 0) { + } else if (wanted & (1 << SOCKLND_CONN_CONTROL)) { type = SOCKLND_CONN_CONTROL; - } else if ((wanted & (1 << SOCKLND_CONN_BULK_IN)) != 0) { + } else if (wanted & (1 << SOCKLND_CONN_BULK_IN)) { type = SOCKLND_CONN_BULK_IN; } else { - LASSERT ((wanted & (1 << SOCKLND_CONN_BULK_OUT)) != 0); + LASSERT(wanted & (1 << SOCKLND_CONN_BULK_OUT)); type = SOCKLND_CONN_BULK_OUT; } @@ -1893,7 +1916,7 @@ ksocknal_connect (ksock_route_t *route) rc = lnet_connect(&sock, peer->ksnp_id.nid, route->ksnr_myipaddr, route->ksnr_ipaddr, route->ksnr_port); - if (rc != 0) + if (rc) goto failed; rc = ksocknal_create_conn(peer->ksnp_ni, route, sock, type); @@ -1904,9 +1927,11 @@ ksocknal_connect (ksock_route_t *route) goto failed; } - /* A +ve RC means I have to retry because I lost the connection - * race or I have to renegotiate protocol version */ - retry_later = (rc != 0); + /* + * A +ve RC means I have to retry because I lost the connection + * race or I have to renegotiate protocol version + */ + retry_later = (rc); if (retry_later) CDEBUG(D_NET, "peer %s: conn race, retry later.\n", libcfs_nid2str(peer->ksnp_id.nid)); @@ -1918,17 +1943,20 @@ ksocknal_connect (ksock_route_t *route) route->ksnr_connecting = 0; if (retry_later) { - /* re-queue for attention; this frees me up to handle - * the peer's incoming connection request */ - + /* + * re-queue for attention; this frees me up to handle + * the peer's incoming connection request + */ if (rc == EALREADY || - (rc == 0 && peer->ksnp_accepting > 0)) { - /* We want to introduce a delay before next + (!rc && peer->ksnp_accepting > 0)) { + /* + * We want to introduce a delay before next * attempt to connect if we lost conn race, * but the race is resolved quickly usually, - * so min_reconnectms should be good heuristic */ + * so min_reconnectms should be good heuristic + */ route->ksnr_retry_interval = - cfs_time_seconds(*ksocknal_tunables.ksnd_min_reconnectms)/1000; + cfs_time_seconds(*ksocknal_tunables.ksnd_min_reconnectms) / 1000; route->ksnr_timeout = cfs_time_add(cfs_time_current(), route->ksnr_retry_interval); } @@ -1949,30 +1977,34 @@ ksocknal_connect (ksock_route_t *route) route->ksnr_retry_interval *= 2; route->ksnr_retry_interval = max(route->ksnr_retry_interval, - cfs_time_seconds(*ksocknal_tunables.ksnd_min_reconnectms)/1000); + cfs_time_seconds(*ksocknal_tunables.ksnd_min_reconnectms) / 1000); route->ksnr_retry_interval = min(route->ksnr_retry_interval, - cfs_time_seconds(*ksocknal_tunables.ksnd_max_reconnectms)/1000); + cfs_time_seconds(*ksocknal_tunables.ksnd_max_reconnectms) / 1000); - LASSERT (route->ksnr_retry_interval != 0); + LASSERT(route->ksnr_retry_interval); route->ksnr_timeout = cfs_time_add(cfs_time_current(), route->ksnr_retry_interval); if (!list_empty(&peer->ksnp_tx_queue) && - peer->ksnp_accepting == 0 && - ksocknal_find_connecting_route_locked(peer) == NULL) { + !peer->ksnp_accepting && + !ksocknal_find_connecting_route_locked(peer)) { ksock_conn_t *conn; - /* ksnp_tx_queue is queued on a conn on successful - * connection for V1.x and V2.x */ - if (!list_empty (&peer->ksnp_conns)) { + /* + * ksnp_tx_queue is queued on a conn on successful + * connection for V1.x and V2.x + */ + if (!list_empty(&peer->ksnp_conns)) { conn = list_entry(peer->ksnp_conns.next, - ksock_conn_t, ksnc_list); - LASSERT (conn->ksnc_proto == &ksocknal_protocol_v3x); + ksock_conn_t, ksnc_list); + LASSERT(conn->ksnc_proto == &ksocknal_protocol_v3x); } - /* take all the blocked packets while I've got the lock and - * complete below... */ + /* + * take all the blocked packets while I've got the lock and + * complete below... + */ list_splice_init(&peer->ksnp_tx_queue, &zombies); } @@ -2011,8 +2043,10 @@ ksocknal_connd_check_start(time64_t sec, long *timeout) if (total >= *ksocknal_tunables.ksnd_nconnds_max || total > ksocknal_data.ksnd_connd_connecting + SOCKNAL_CONND_RESV) { - /* can't create more connd, or still have enough - * threads to handle more connecting */ + /* + * can't create more connd, or still have enough + * threads to handle more connecting + */ return 0; } @@ -2041,7 +2075,7 @@ ksocknal_connd_check_start(time64_t sec, long *timeout) rc = ksocknal_thread_start(ksocknal_connd, NULL, name); spin_lock_bh(&ksocknal_data.ksnd_connd_lock); - if (rc == 0) + if (!rc) return 1; /* we tried ... */ @@ -2093,8 +2127,10 @@ ksocknal_connd_check_stop(time64_t sec, long *timeout) ksocknal_data.ksnd_connd_connecting + SOCKNAL_CONND_RESV; } -/* Go through connd_routes queue looking for a route that we can process - * right now, @timeout_p can be updated if we need to come back later */ +/* + * Go through connd_routes queue looking for a route that we can process + * right now, @timeout_p can be updated if we need to come back later + */ static ksock_route_t * ksocknal_connd_get_route_locked(signed long *timeout_p) { @@ -2104,10 +2140,9 @@ ksocknal_connd_get_route_locked(signed long *timeout_p) now = cfs_time_current(); /* connd_routes can contain both pending and ordinary routes */ - list_for_each_entry (route, &ksocknal_data.ksnd_connd_routes, - ksnr_connd_list) { - - if (route->ksnr_retry_interval == 0 || + list_for_each_entry(route, &ksocknal_data.ksnd_connd_routes, + ksnr_connd_list) { + if (!route->ksnr_retry_interval || cfs_time_aftereq(now, route->ksnr_timeout)) return route; @@ -2120,7 +2155,7 @@ ksocknal_connd_get_route_locked(signed long *timeout_p) } int -ksocknal_connd (void *arg) +ksocknal_connd(void *arg) { spinlock_t *connd_lock = &ksocknal_data.ksnd_connd_lock; ksock_connreq_t *cr; @@ -2172,15 +2207,17 @@ ksocknal_connd (void *arg) spin_lock_bh(connd_lock); } - /* Only handle an outgoing connection request if there + /* + * Only handle an outgoing connection request if there * is a thread left to handle incoming connections and - * create new connd */ + * create new connd + */ if (ksocknal_data.ksnd_connd_connecting + SOCKNAL_CONND_RESV < ksocknal_data.ksnd_connd_running) { route = ksocknal_connd_get_route_locked(&timeout); } - if (route != NULL) { - list_del (&route->ksnr_connd_list); + if (route) { + list_del(&route->ksnr_connd_list); ksocknal_data.ksnd_connd_connecting++; spin_unlock_bh(connd_lock); dropped_lock = 1; @@ -2231,24 +2268,26 @@ ksocknal_connd (void *arg) } static ksock_conn_t * -ksocknal_find_timed_out_conn (ksock_peer_t *peer) +ksocknal_find_timed_out_conn(ksock_peer_t *peer) { /* We're called with a shared lock on ksnd_global_lock */ ksock_conn_t *conn; struct list_head *ctmp; - list_for_each (ctmp, &peer->ksnp_conns) { + list_for_each(ctmp, &peer->ksnp_conns) { int error; - conn = list_entry (ctmp, ksock_conn_t, ksnc_list); + conn = list_entry(ctmp, ksock_conn_t, ksnc_list); /* Don't need the {get,put}connsock dance to deref ksnc_sock */ LASSERT(!conn->ksnc_closing); - /* SOCK_ERROR will reset error code of socket in - * some platform (like Darwin8.x) */ + /* + * SOCK_ERROR will reset error code of socket in + * some platform (like Darwin8.x) + */ error = conn->ksnc_sock->sk->sk_err; - if (error != 0) { + if (error) { ksocknal_conn_addref(conn); switch (error) { @@ -2292,11 +2331,13 @@ ksocknal_find_timed_out_conn (ksock_peer_t *peer) } if ((!list_empty(&conn->ksnc_tx_queue) || - conn->ksnc_sock->sk->sk_wmem_queued != 0) && + conn->ksnc_sock->sk->sk_wmem_queued) && cfs_time_aftereq(cfs_time_current(), conn->ksnc_tx_deadline)) { - /* Timed out messages queued for sending or - * buffered in the socket's send buffer */ + /* + * Timed out messages queued for sending or + * buffered in the socket's send buffer + */ ksocknal_conn_addref(conn); CNETERR("Timeout sending data to %s (%pI4h:%d) the network or that node may be down.\n", libcfs_id2str(peer->ksnp_id), @@ -2313,20 +2354,18 @@ static inline void ksocknal_flush_stale_txs(ksock_peer_t *peer) { ksock_tx_t *tx; + ksock_tx_t *tmp; LIST_HEAD(stale_txs); write_lock_bh(&ksocknal_data.ksnd_global_lock); - while (!list_empty (&peer->ksnp_tx_queue)) { - tx = list_entry (peer->ksnp_tx_queue.next, - ksock_tx_t, tx_list); - + list_for_each_entry_safe(tx, tmp, &peer->ksnp_tx_queue, tx_list) { if (!cfs_time_aftereq(cfs_time_current(), tx->tx_deadline)) break; - list_del (&tx->tx_list); - list_add_tail (&tx->tx_list, &stale_txs); + list_del(&tx->tx_list); + list_add_tail(&tx->tx_list, &stale_txs); } write_unlock_bh(&ksocknal_data.ksnd_global_lock); @@ -2336,6 +2375,7 @@ ksocknal_flush_stale_txs(ksock_peer_t *peer) static int ksocknal_send_keepalive_locked(ksock_peer_t *peer) + __must_hold(&ksocknal_data.ksnd_global_lock) { ksock_sched_t *sched; ksock_conn_t *conn; @@ -2356,12 +2396,14 @@ ksocknal_send_keepalive_locked(ksock_peer_t *peer) if (time_before(cfs_time_current(), peer->ksnp_send_keepalive)) return 0; - /* retry 10 secs later, so we wouldn't put pressure - * on this peer if we failed to send keepalive this time */ + /* + * retry 10 secs later, so we wouldn't put pressure + * on this peer if we failed to send keepalive this time + */ peer->ksnp_send_keepalive = cfs_time_shift(10); conn = ksocknal_find_conn_locked(peer, NULL, 1); - if (conn != NULL) { + if (conn) { sched = conn->ksnc_scheduler; spin_lock_bh(&sched->kss_lock); @@ -2378,12 +2420,12 @@ ksocknal_send_keepalive_locked(ksock_peer_t *peer) /* cookie = 1 is reserved for keepalive PING */ tx = ksocknal_alloc_tx_noop(1, 1); - if (tx == NULL) { + if (!tx) { read_lock(&ksocknal_data.ksnd_global_lock); return -ENOMEM; } - if (ksocknal_launch_packet(peer->ksnp_ni, tx, peer->ksnp_id) == 0) { + if (!ksocknal_launch_packet(peer->ksnp_ni, tx, peer->ksnp_id)) { read_lock(&ksocknal_data.ksnd_global_lock); return 1; } @@ -2395,7 +2437,7 @@ ksocknal_send_keepalive_locked(ksock_peer_t *peer) } static void -ksocknal_check_peer_timeouts (int idx) +ksocknal_check_peer_timeouts(int idx) { struct list_head *peers = &ksocknal_data.ksnd_peers[idx]; ksock_peer_t *peer; @@ -2403,9 +2445,11 @@ ksocknal_check_peer_timeouts (int idx) ksock_tx_t *tx; again: - /* NB. We expect to have a look at all the peers and not find any + /* + * NB. We expect to have a look at all the peers and not find any * connections to time out, so we just use a shared lock while we - * take a look... */ + * take a look... + */ read_lock(&ksocknal_data.ksnd_global_lock); list_for_each_entry(peer, peers, ksnp_list) { @@ -2413,35 +2457,37 @@ ksocknal_check_peer_timeouts (int idx) int resid = 0; int n = 0; - if (ksocknal_send_keepalive_locked(peer) != 0) { + if (ksocknal_send_keepalive_locked(peer)) { read_unlock(&ksocknal_data.ksnd_global_lock); goto again; } - conn = ksocknal_find_timed_out_conn (peer); + conn = ksocknal_find_timed_out_conn(peer); - if (conn != NULL) { + if (conn) { read_unlock(&ksocknal_data.ksnd_global_lock); - ksocknal_close_conn_and_siblings (conn, -ETIMEDOUT); + ksocknal_close_conn_and_siblings(conn, -ETIMEDOUT); - /* NB we won't find this one again, but we can't + /* + * NB we won't find this one again, but we can't * just proceed with the next peer, since we dropped - * ksnd_global_lock and it might be dead already! */ + * ksnd_global_lock and it might be dead already! + */ ksocknal_conn_decref(conn); goto again; } - /* we can't process stale txs right here because we're - * holding only shared lock */ - if (!list_empty (&peer->ksnp_tx_queue)) { - ksock_tx_t *tx = - list_entry (peer->ksnp_tx_queue.next, - ksock_tx_t, tx_list); + /* + * we can't process stale txs right here because we're + * holding only shared lock + */ + if (!list_empty(&peer->ksnp_tx_queue)) { + ksock_tx_t *tx = list_entry(peer->ksnp_tx_queue.next, + ksock_tx_t, tx_list); if (cfs_time_aftereq(cfs_time_current(), tx->tx_deadline)) { - ksocknal_peer_addref(peer); read_unlock(&ksocknal_data.ksnd_global_lock); @@ -2466,13 +2512,13 @@ ksocknal_check_peer_timeouts (int idx) n++; } - if (n == 0) { + if (!n) { spin_unlock(&peer->ksnp_lock); continue; } tx = list_entry(peer->ksnp_zc_req_list.next, - ksock_tx_t, tx_zc_list); + ksock_tx_t, tx_zc_list); deadline = tx->tx_deadline; resid = tx->tx_resid; conn = tx->tx_conn; @@ -2486,7 +2532,7 @@ ksocknal_check_peer_timeouts (int idx) cfs_duration_sec(cfs_time_current() - deadline), resid, conn->ksnc_sock->sk->sk_wmem_queued); - ksocknal_close_conn_and_siblings (conn, -ETIMEDOUT); + ksocknal_close_conn_and_siblings(conn, -ETIMEDOUT); ksocknal_conn_decref(conn); goto again; } @@ -2495,7 +2541,7 @@ ksocknal_check_peer_timeouts (int idx) } int -ksocknal_reaper (void *arg) +ksocknal_reaper(void *arg) { wait_queue_t wait; ksock_conn_t *conn; @@ -2515,12 +2561,10 @@ ksocknal_reaper (void *arg) spin_lock_bh(&ksocknal_data.ksnd_reaper_lock); while (!ksocknal_data.ksnd_shuttingdown) { - - if (!list_empty (&ksocknal_data.ksnd_deathrow_conns)) { - conn = list_entry (ksocknal_data. \ - ksnd_deathrow_conns.next, - ksock_conn_t, ksnc_list); - list_del (&conn->ksnc_list); + if (!list_empty(&ksocknal_data.ksnd_deathrow_conns)) { + conn = list_entry(ksocknal_data.ksnd_deathrow_conns.next, + ksock_conn_t, ksnc_list); + list_del(&conn->ksnc_list); spin_unlock_bh(&ksocknal_data.ksnd_reaper_lock); @@ -2531,10 +2575,10 @@ ksocknal_reaper (void *arg) continue; } - if (!list_empty (&ksocknal_data.ksnd_zombie_conns)) { - conn = list_entry (ksocknal_data.ksnd_zombie_conns.\ - next, ksock_conn_t, ksnc_list); - list_del (&conn->ksnc_list); + if (!list_empty(&ksocknal_data.ksnd_zombie_conns)) { + conn = list_entry(ksocknal_data.ksnd_zombie_conns.next, + ksock_conn_t, ksnc_list); + list_del(&conn->ksnc_list); spin_unlock_bh(&ksocknal_data.ksnd_reaper_lock); @@ -2544,9 +2588,9 @@ ksocknal_reaper (void *arg) continue; } - if (!list_empty (&ksocknal_data.ksnd_enomem_conns)) { + if (!list_empty(&ksocknal_data.ksnd_enomem_conns)) { list_add(&enomem_conns, - &ksocknal_data.ksnd_enomem_conns); + &ksocknal_data.ksnd_enomem_conns); list_del_init(&ksocknal_data.ksnd_enomem_conns); } @@ -2554,10 +2598,10 @@ ksocknal_reaper (void *arg) /* reschedule all the connections that stalled with ENOMEM... */ nenomem_conns = 0; - while (!list_empty (&enomem_conns)) { - conn = list_entry (enomem_conns.next, - ksock_conn_t, ksnc_tx_list); - list_del (&conn->ksnc_tx_list); + while (!list_empty(&enomem_conns)) { + conn = list_entry(enomem_conns.next, ksock_conn_t, + ksnc_tx_list); + list_del(&conn->ksnc_tx_list); sched = conn->ksnc_scheduler; @@ -2566,7 +2610,7 @@ ksocknal_reaper (void *arg) LASSERT(conn->ksnc_tx_scheduled); conn->ksnc_tx_ready = 1; list_add_tail(&conn->ksnc_tx_list, - &sched->kss_tx_conns); + &sched->kss_tx_conns); wake_up(&sched->kss_waitq); spin_unlock_bh(&sched->kss_lock); @@ -2580,21 +2624,22 @@ ksocknal_reaper (void *arg) const int p = 1; int chunk = ksocknal_data.ksnd_peer_hash_size; - /* Time to check for timeouts on a few more peers: I do + /* + * Time to check for timeouts on a few more peers: I do * checks every 'p' seconds on a proportion of the peer * table and I need to check every connection 'n' times * within a timeout interval, to ensure I detect a * timeout on any connection within (n+1)/n times the - * timeout interval. */ - + * timeout interval. + */ if (*ksocknal_tunables.ksnd_timeout > n * p) chunk = (chunk * n * p) / *ksocknal_tunables.ksnd_timeout; - if (chunk == 0) + if (!chunk) chunk = 1; for (i = 0; i < chunk; i++) { - ksocknal_check_peer_timeouts (peer_index); + ksocknal_check_peer_timeouts(peer_index); peer_index = (peer_index + 1) % ksocknal_data.ksnd_peer_hash_size; } @@ -2602,25 +2647,27 @@ ksocknal_reaper (void *arg) deadline = cfs_time_add(deadline, cfs_time_seconds(p)); } - if (nenomem_conns != 0) { - /* Reduce my timeout if I rescheduled ENOMEM conns. + if (nenomem_conns) { + /* + * Reduce my timeout if I rescheduled ENOMEM conns. * This also prevents me getting woken immediately - * if any go back on my enomem list. */ + * if any go back on my enomem list. + */ timeout = SOCKNAL_ENOMEM_RETRY; } ksocknal_data.ksnd_reaper_waketime = cfs_time_add(cfs_time_current(), timeout); - set_current_state (TASK_INTERRUPTIBLE); - add_wait_queue (&ksocknal_data.ksnd_reaper_waitq, &wait); + set_current_state(TASK_INTERRUPTIBLE); + add_wait_queue(&ksocknal_data.ksnd_reaper_waitq, &wait); if (!ksocknal_data.ksnd_shuttingdown && - list_empty (&ksocknal_data.ksnd_deathrow_conns) && - list_empty (&ksocknal_data.ksnd_zombie_conns)) + list_empty(&ksocknal_data.ksnd_deathrow_conns) && + list_empty(&ksocknal_data.ksnd_zombie_conns)) schedule_timeout(timeout); - set_current_state (TASK_RUNNING); - remove_wait_queue (&ksocknal_data.ksnd_reaper_waitq, &wait); + set_current_state(TASK_RUNNING); + remove_wait_queue(&ksocknal_data.ksnd_reaper_waitq, &wait); spin_lock_bh(&ksocknal_data.ksnd_reaper_lock); } diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib.c index cf8e43bd3..d4ce06d0a 100644 --- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib.c +++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib.c @@ -45,13 +45,13 @@ ksocknal_lib_get_conn_addrs(ksock_conn_t *conn) /* Didn't need the {get,put}connsock dance to deref ksnc_sock... */ LASSERT(!conn->ksnc_closing); - if (rc != 0) { + if (rc) { CERROR("Error %d getting sock peer IP\n", rc); return rc; } rc = lnet_sock_getaddr(conn->ksnc_sock, 0, &conn->ksnc_myipaddr, NULL); - if (rc != 0) { + if (rc) { CERROR("Error %d getting sock local IP\n", rc); return rc; } @@ -67,9 +67,11 @@ ksocknal_lib_zc_capable(ksock_conn_t *conn) if (conn->ksnc_proto == &ksocknal_protocol_v1x) return 0; - /* ZC if the socket supports scatter/gather and doesn't need software - * checksums */ - return ((caps & NETIF_F_SG) != 0 && (caps & NETIF_F_CSUM_MASK) != 0); + /* + * ZC if the socket supports scatter/gather and doesn't need software + * checksums + */ + return ((caps & NETIF_F_SG) && (caps & NETIF_F_CSUM_MASK)); } int @@ -82,12 +84,13 @@ ksocknal_lib_send_iov(ksock_conn_t *conn, ksock_tx_t *tx) if (*ksocknal_tunables.ksnd_enable_csum && /* checksum enabled */ conn->ksnc_proto == &ksocknal_protocol_v2x && /* V2.x connection */ tx->tx_nob == tx->tx_resid && /* frist sending */ - tx->tx_msg.ksm_csum == 0) /* not checksummed */ + !tx->tx_msg.ksm_csum) /* not checksummed */ ksocknal_lib_csum_tx(tx); - /* NB we can't trust socket ops to either consume our iovs - * or leave them alone. */ - + /* + * NB we can't trust socket ops to either consume our iovs + * or leave them alone. + */ { #if SOCKNAL_SINGLE_FRAG_TX struct kvec scratch; @@ -123,11 +126,13 @@ ksocknal_lib_send_kiov(ksock_conn_t *conn, ksock_tx_t *tx) int nob; /* Not NOOP message */ - LASSERT(tx->tx_lnetmsg != NULL); + LASSERT(tx->tx_lnetmsg); - /* NB we can't trust socket ops to either consume our iovs - * or leave them alone. */ - if (tx->tx_msg.ksm_zc_cookies[0] != 0) { + /* + * NB we can't trust socket ops to either consume our iovs + * or leave them alone. + */ + if (tx->tx_msg.ksm_zc_cookies[0]) { /* Zero copy is enabled */ struct sock *sk = sock->sk; struct page *page = kiov->kiov_page; @@ -136,13 +141,13 @@ ksocknal_lib_send_kiov(ksock_conn_t *conn, ksock_tx_t *tx) int msgflg = MSG_DONTWAIT; CDEBUG(D_NET, "page %p + offset %x for %d\n", - page, offset, kiov->kiov_len); + page, offset, kiov->kiov_len); if (!list_empty(&conn->ksnc_tx_queue) || fragsize < tx->tx_resid) msgflg |= MSG_MORE; - if (sk->sk_prot->sendpage != NULL) { + if (sk->sk_prot->sendpage) { rc = sk->sk_prot->sendpage(sk, page, offset, fragsize, msgflg); } else { @@ -187,13 +192,14 @@ ksocknal_lib_eager_ack(ksock_conn_t *conn) int opt = 1; struct socket *sock = conn->ksnc_sock; - /* Remind the socket to ACK eagerly. If I don't, the socket might + /* + * Remind the socket to ACK eagerly. If I don't, the socket might * think I'm about to send something it could piggy-back the ACK * on, introducing delay in completing zero-copy sends in my - * peer. */ - - kernel_setsockopt(sock, SOL_TCP, TCP_QUICKACK, - (char *)&opt, sizeof(opt)); + * peer. + */ + kernel_setsockopt(sock, SOL_TCP, TCP_QUICKACK, (char *)&opt, + sizeof(opt)); } int @@ -218,8 +224,10 @@ ksocknal_lib_recv_iov(ksock_conn_t *conn) int sum; __u32 saved_csum; - /* NB we can't trust socket ops to either consume our iovs - * or leave them alone. */ + /* + * NB we can't trust socket ops to either consume our iovs + * or leave them alone. + */ LASSERT(niov > 0); for (nob = i = 0; i < niov; i++) { @@ -228,8 +236,8 @@ ksocknal_lib_recv_iov(ksock_conn_t *conn) } LASSERT(nob <= conn->ksnc_rx_nob_wanted); - rc = kernel_recvmsg(conn->ksnc_sock, &msg, - scratchiov, niov, nob, MSG_DONTWAIT); + rc = kernel_recvmsg(conn->ksnc_sock, &msg, scratchiov, niov, nob, + MSG_DONTWAIT); saved_csum = 0; if (conn->ksnc_proto == &ksocknal_protocol_v2x) { @@ -237,7 +245,7 @@ ksocknal_lib_recv_iov(ksock_conn_t *conn) conn->ksnc_msg.ksm_csum = 0; } - if (saved_csum != 0) { + if (saved_csum) { /* accumulate checksum */ for (i = 0, sum = rc; sum > 0; i++, sum -= fragnob) { LASSERT(i < niov); @@ -258,7 +266,7 @@ ksocknal_lib_recv_iov(ksock_conn_t *conn) static void ksocknal_lib_kiov_vunmap(void *addr) { - if (addr == NULL) + if (!addr) return; vunmap(addr); @@ -272,7 +280,7 @@ ksocknal_lib_kiov_vmap(lnet_kiov_t *kiov, int niov, int nob; int i; - if (!*ksocknal_tunables.ksnd_zc_recv || pages == NULL) + if (!*ksocknal_tunables.ksnd_zc_recv || !pages) return NULL; LASSERT(niov <= LNET_MAX_IOV); @@ -282,8 +290,8 @@ ksocknal_lib_kiov_vmap(lnet_kiov_t *kiov, int niov, return NULL; for (nob = i = 0; i < niov; i++) { - if ((kiov[i].kiov_offset != 0 && i > 0) || - (kiov[i].kiov_offset + kiov[i].kiov_len != PAGE_CACHE_SIZE && i < niov - 1)) + if ((kiov[i].kiov_offset && i > 0) || + (kiov[i].kiov_offset + kiov[i].kiov_len != PAGE_SIZE && i < niov - 1)) return NULL; pages[i] = kiov[i].kiov_page; @@ -291,7 +299,7 @@ ksocknal_lib_kiov_vmap(lnet_kiov_t *kiov, int niov, } addr = vmap(pages, niov, VM_MAP, PAGE_KERNEL); - if (addr == NULL) + if (!addr) return NULL; iov->iov_base = addr + kiov[0].kiov_offset; @@ -329,10 +337,12 @@ ksocknal_lib_recv_kiov(ksock_conn_t *conn) int fragnob; int n; - /* NB we can't trust socket ops to either consume our iovs - * or leave them alone. */ + /* + * NB we can't trust socket ops to either consume our iovs + * or leave them alone. + */ addr = ksocknal_lib_kiov_vmap(kiov, niov, scratchiov, pages); - if (addr != NULL) { + if (addr) { nob = scratchiov[0].iov_len; n = 1; @@ -347,17 +357,19 @@ ksocknal_lib_recv_kiov(ksock_conn_t *conn) LASSERT(nob <= conn->ksnc_rx_nob_wanted); - rc = kernel_recvmsg(conn->ksnc_sock, &msg, - (struct kvec *)scratchiov, n, nob, MSG_DONTWAIT); + rc = kernel_recvmsg(conn->ksnc_sock, &msg, (struct kvec *)scratchiov, + n, nob, MSG_DONTWAIT); - if (conn->ksnc_msg.ksm_csum != 0) { + if (conn->ksnc_msg.ksm_csum) { for (i = 0, sum = rc; sum > 0; i++, sum -= fragnob) { LASSERT(i < niov); - /* Dang! have to kmap again because I have nowhere to + /* + * Dang! have to kmap again because I have nowhere to * stash the mapped address. But by doing it while the * page is still mapped, the kernel just bumps the map - * count and returns me the address it stashed. */ + * count and returns me the address it stashed. + */ base = kmap(kiov[i].kiov_page) + kiov[i].kiov_offset; fragnob = kiov[i].kiov_len; if (fragnob > sum) @@ -370,7 +382,7 @@ ksocknal_lib_recv_kiov(ksock_conn_t *conn) } } - if (addr != NULL) { + if (addr) { ksocknal_lib_kiov_vunmap(addr); } else { for (i = 0; i < niov; i++) @@ -388,7 +400,7 @@ ksocknal_lib_csum_tx(ksock_tx_t *tx) void *base; LASSERT(tx->tx_iov[0].iov_base == &tx->tx_msg); - LASSERT(tx->tx_conn != NULL); + LASSERT(tx->tx_conn); LASSERT(tx->tx_conn->ksnc_proto == &ksocknal_protocol_v2x); tx->tx_msg.ksm_csum = 0; @@ -396,7 +408,7 @@ ksocknal_lib_csum_tx(ksock_tx_t *tx) csum = ksocknal_csum(~0, tx->tx_iov[0].iov_base, tx->tx_iov[0].iov_len); - if (tx->tx_kiov != NULL) { + if (tx->tx_kiov) { for (i = 0; i < tx->tx_nkiov; i++) { base = kmap(tx->tx_kiov[i].kiov_page) + tx->tx_kiov[i].kiov_offset; @@ -427,22 +439,22 @@ ksocknal_lib_get_conn_tunables(ksock_conn_t *conn, int *txmem, int *rxmem, int * int rc; rc = ksocknal_connsock_addref(conn); - if (rc != 0) { + if (rc) { LASSERT(conn->ksnc_closing); *txmem = *rxmem = *nagle = 0; return -ESHUTDOWN; } rc = lnet_sock_getbuf(sock, txmem, rxmem); - if (rc == 0) { + if (!rc) { len = sizeof(*nagle); rc = kernel_getsockopt(sock, SOL_TCP, TCP_NODELAY, - (char *)nagle, &len); + (char *)nagle, &len); } ksocknal_connsock_decref(conn); - if (rc == 0) + if (!rc) *nagle = !*nagle; else *txmem = *rxmem = *nagle = 0; @@ -463,23 +475,24 @@ ksocknal_lib_setup_sock(struct socket *sock) sock->sk->sk_allocation = GFP_NOFS; - /* Ensure this socket aborts active sends immediately when we close - * it. */ - + /* + * Ensure this socket aborts active sends immediately when we close + * it. + */ linger.l_onoff = 0; linger.l_linger = 0; - rc = kernel_setsockopt(sock, SOL_SOCKET, SO_LINGER, - (char *)&linger, sizeof(linger)); - if (rc != 0) { + rc = kernel_setsockopt(sock, SOL_SOCKET, SO_LINGER, (char *)&linger, + sizeof(linger)); + if (rc) { CERROR("Can't set SO_LINGER: %d\n", rc); return rc; } option = -1; - rc = kernel_setsockopt(sock, SOL_TCP, TCP_LINGER2, - (char *)&option, sizeof(option)); - if (rc != 0) { + rc = kernel_setsockopt(sock, SOL_TCP, TCP_LINGER2, (char *)&option, + sizeof(option)); + if (rc) { CERROR("Can't set SO_LINGER2: %d\n", rc); return rc; } @@ -488,8 +501,8 @@ ksocknal_lib_setup_sock(struct socket *sock) option = 1; rc = kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, - (char *)&option, sizeof(option)); - if (rc != 0) { + (char *)&option, sizeof(option)); + if (rc) { CERROR("Can't disable nagle: %d\n", rc); return rc; } @@ -497,10 +510,10 @@ ksocknal_lib_setup_sock(struct socket *sock) rc = lnet_sock_setbuf(sock, *ksocknal_tunables.ksnd_tx_buffer_size, *ksocknal_tunables.ksnd_rx_buffer_size); - if (rc != 0) { + if (rc) { CERROR("Can't set buffer tx %d, rx %d buffers: %d\n", - *ksocknal_tunables.ksnd_tx_buffer_size, - *ksocknal_tunables.ksnd_rx_buffer_size, rc); + *ksocknal_tunables.ksnd_tx_buffer_size, + *ksocknal_tunables.ksnd_rx_buffer_size, rc); return rc; } @@ -514,9 +527,9 @@ ksocknal_lib_setup_sock(struct socket *sock) do_keepalive = (keep_idle > 0 && keep_count > 0 && keep_intvl > 0); option = (do_keepalive ? 1 : 0); - rc = kernel_setsockopt(sock, SOL_SOCKET, SO_KEEPALIVE, - (char *)&option, sizeof(option)); - if (rc != 0) { + rc = kernel_setsockopt(sock, SOL_SOCKET, SO_KEEPALIVE, (char *)&option, + sizeof(option)); + if (rc) { CERROR("Can't set SO_KEEPALIVE: %d\n", rc); return rc; } @@ -524,23 +537,23 @@ ksocknal_lib_setup_sock(struct socket *sock) if (!do_keepalive) return 0; - rc = kernel_setsockopt(sock, SOL_TCP, TCP_KEEPIDLE, - (char *)&keep_idle, sizeof(keep_idle)); - if (rc != 0) { + rc = kernel_setsockopt(sock, SOL_TCP, TCP_KEEPIDLE, (char *)&keep_idle, + sizeof(keep_idle)); + if (rc) { CERROR("Can't set TCP_KEEPIDLE: %d\n", rc); return rc; } rc = kernel_setsockopt(sock, SOL_TCP, TCP_KEEPINTVL, - (char *)&keep_intvl, sizeof(keep_intvl)); - if (rc != 0) { + (char *)&keep_intvl, sizeof(keep_intvl)); + if (rc) { CERROR("Can't set TCP_KEEPINTVL: %d\n", rc); return rc; } - rc = kernel_setsockopt(sock, SOL_TCP, TCP_KEEPCNT, - (char *)&keep_count, sizeof(keep_count)); - if (rc != 0) { + rc = kernel_setsockopt(sock, SOL_TCP, TCP_KEEPCNT, (char *)&keep_count, + sizeof(keep_count)); + if (rc) { CERROR("Can't set TCP_KEEPCNT: %d\n", rc); return rc; } @@ -558,7 +571,7 @@ ksocknal_lib_push_conn(ksock_conn_t *conn) int rc; rc = ksocknal_connsock_addref(conn); - if (rc != 0) /* being shut down */ + if (rc) /* being shut down */ return; sk = conn->ksnc_sock->sk; @@ -570,8 +583,8 @@ ksocknal_lib_push_conn(ksock_conn_t *conn) release_sock(sk); rc = kernel_setsockopt(conn->ksnc_sock, SOL_TCP, TCP_NODELAY, - (char *)&val, sizeof(val)); - LASSERT(rc == 0); + (char *)&val, sizeof(val)); + LASSERT(!rc); lock_sock(sk); tp->nonagle = nonagle; @@ -593,11 +606,12 @@ ksocknal_data_ready(struct sock *sk) read_lock(&ksocknal_data.ksnd_global_lock); conn = sk->sk_user_data; - if (conn == NULL) { /* raced with ksocknal_terminate_conn */ + if (!conn) { /* raced with ksocknal_terminate_conn */ LASSERT(sk->sk_data_ready != &ksocknal_data_ready); sk->sk_data_ready(sk); - } else + } else { ksocknal_read_callback(conn); + } read_unlock(&ksocknal_data.ksnd_global_lock); } @@ -619,14 +633,14 @@ ksocknal_write_space(struct sock *sk) CDEBUG(D_NET, "sk %p wspace %d low water %d conn %p%s%s%s\n", sk, wspace, min_wpace, conn, - (conn == NULL) ? "" : (conn->ksnc_tx_ready ? + !conn ? "" : (conn->ksnc_tx_ready ? " ready" : " blocked"), - (conn == NULL) ? "" : (conn->ksnc_tx_scheduled ? + !conn ? "" : (conn->ksnc_tx_scheduled ? " scheduled" : " idle"), - (conn == NULL) ? "" : (list_empty(&conn->ksnc_tx_queue) ? + !conn ? "" : (list_empty(&conn->ksnc_tx_queue) ? " empty" : " queued")); - if (conn == NULL) { /* raced with ksocknal_terminate_conn */ + if (!conn) { /* raced with ksocknal_terminate_conn */ LASSERT(sk->sk_write_space != &ksocknal_write_space); sk->sk_write_space(sk); @@ -637,10 +651,11 @@ ksocknal_write_space(struct sock *sk) if (wspace >= min_wpace) { /* got enough space */ ksocknal_write_callback(conn); - /* Clear SOCK_NOSPACE _after_ ksocknal_write_callback so the + /* + * Clear SOCK_NOSPACE _after_ ksocknal_write_callback so the * ENOMEM check in ksocknal_transmit is race-free (think about - * it). */ - + * it). + */ clear_bit(SOCK_NOSPACE, &sk->sk_socket->flags); } @@ -666,15 +681,19 @@ ksocknal_lib_set_callback(struct socket *sock, ksock_conn_t *conn) void ksocknal_lib_reset_callback(struct socket *sock, ksock_conn_t *conn) { - /* Remove conn's network callbacks. + /* + * Remove conn's network callbacks. * NB I _have_ to restore the callback, rather than storing a noop, - * since the socket could survive past this module being unloaded!! */ + * since the socket could survive past this module being unloaded!! + */ sock->sk->sk_data_ready = conn->ksnc_saved_data_ready; sock->sk->sk_write_space = conn->ksnc_saved_write_space; - /* A callback could be in progress already; they hold a read lock + /* + * A callback could be in progress already; they hold a read lock * on ksnd_global_lock (to serialise with me) and NOOP if - * sk_user_data is NULL. */ + * sk_user_data is NULL. + */ sock->sk->sk_user_data = NULL; return ; @@ -691,14 +710,16 @@ ksocknal_lib_memory_pressure(ksock_conn_t *conn) if (!test_bit(SOCK_NOSPACE, &conn->ksnc_sock->flags) && !conn->ksnc_tx_ready) { - /* SOCK_NOSPACE is set when the socket fills + /* + * SOCK_NOSPACE is set when the socket fills * and cleared in the write_space callback * (which also sets ksnc_tx_ready). If * SOCK_NOSPACE and ksnc_tx_ready are BOTH * zero, I didn't fill the socket and * write_space won't reschedule me, so I * return -ENOMEM to get my caller to retry - * after a timeout */ + * after a timeout + */ rc = -ENOMEM; } diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_modparams.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_modparams.c index fdb2b23e2..6329cbe66 100644 --- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_modparams.c +++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_modparams.c @@ -14,9 +14,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with Portals; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ #include "socklnd.h" @@ -41,8 +38,10 @@ static int peer_timeout = 180; module_param(peer_timeout, int, 0444); MODULE_PARM_DESC(peer_timeout, "Seconds without aliveness news to declare peer dead (<=0 to disable)"); -/* Number of daemons in each thread pool which is percpt, - * we will estimate reasonable value based on CPUs if it's not set. */ +/* + * Number of daemons in each thread pool which is percpt, + * we will estimate reasonable value based on CPUs if it's not set. + */ static unsigned int nscheds; module_param(nscheds, int, 0444); MODULE_PARM_DESC(nscheds, "# scheduler daemons in each pool while starting"); @@ -72,7 +71,7 @@ static int typed_conns = 1; module_param(typed_conns, int, 0444); MODULE_PARM_DESC(typed_conns, "use different sockets for bulk"); -static int min_bulk = 1<<10; +static int min_bulk = 1 << 10; module_param(min_bulk, int, 0644); MODULE_PARM_DESC(min_bulk, "smallest 'large' message"); diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c index 986bce4c9..32cc31e4c 100644 --- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c +++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c @@ -19,9 +19,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with Portals; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ #include "socklnd.h" @@ -56,15 +53,14 @@ ksocknal_next_tx_carrier(ksock_conn_t *conn) /* Called holding BH lock: conn->ksnc_scheduler->kss_lock */ LASSERT(!list_empty(&conn->ksnc_tx_queue)); - LASSERT(tx != NULL); + LASSERT(tx); /* Next TX that can carry ZC-ACK or LNet message */ if (tx->tx_list.next == &conn->ksnc_tx_queue) { /* no more packets queued */ conn->ksnc_tx_carrier = NULL; } else { - conn->ksnc_tx_carrier = list_entry(tx->tx_list.next, - ksock_tx_t, tx_list); + conn->ksnc_tx_carrier = list_next_entry(tx, tx_list); LASSERT(conn->ksnc_tx_carrier->tx_msg.ksm_type == tx->tx_msg.ksm_type); } } @@ -75,8 +71,8 @@ ksocknal_queue_tx_zcack_v2(ksock_conn_t *conn, { ksock_tx_t *tx = conn->ksnc_tx_carrier; - LASSERT(tx_ack == NULL || - tx_ack->tx_msg.ksm_type == KSOCK_MSG_NOOP); + LASSERT(!tx_ack || + tx_ack->tx_msg.ksm_type == KSOCK_MSG_NOOP); /* * Enqueue or piggyback tx_ack / cookie @@ -85,10 +81,10 @@ ksocknal_queue_tx_zcack_v2(ksock_conn_t *conn, * . There is tx can piggyback cookie of tx_ack (or cookie), * piggyback the cookie and return the tx. */ - if (tx == NULL) { - if (tx_ack != NULL) { + if (!tx) { + if (tx_ack) { list_add_tail(&tx_ack->tx_list, - &conn->ksnc_tx_queue); + &conn->ksnc_tx_queue); conn->ksnc_tx_carrier = tx_ack; } return 0; @@ -96,16 +92,16 @@ ksocknal_queue_tx_zcack_v2(ksock_conn_t *conn, if (tx->tx_msg.ksm_type == KSOCK_MSG_NOOP) { /* tx is noop zc-ack, can't piggyback zc-ack cookie */ - if (tx_ack != NULL) + if (tx_ack) list_add_tail(&tx_ack->tx_list, - &conn->ksnc_tx_queue); + &conn->ksnc_tx_queue); return 0; } LASSERT(tx->tx_msg.ksm_type == KSOCK_MSG_LNET); - LASSERT(tx->tx_msg.ksm_zc_cookies[1] == 0); + LASSERT(!tx->tx_msg.ksm_zc_cookies[1]); - if (tx_ack != NULL) + if (tx_ack) cookie = tx_ack->tx_msg.ksm_zc_cookies[1]; /* piggyback the zc-ack cookie */ @@ -128,7 +124,7 @@ ksocknal_queue_tx_msg_v2(ksock_conn_t *conn, ksock_tx_t *tx_msg) * . If there is NOOP on the connection, piggyback the cookie * and replace the NOOP tx, and return the NOOP tx. */ - if (tx == NULL) { /* nothing on queue */ + if (!tx) { /* nothing on queue */ list_add_tail(&tx_msg->tx_list, &conn->ksnc_tx_queue); conn->ksnc_tx_carrier = tx_msg; return NULL; @@ -162,22 +158,22 @@ ksocknal_queue_tx_zcack_v3(ksock_conn_t *conn, return ksocknal_queue_tx_zcack_v2(conn, tx_ack, cookie); /* non-blocking ZC-ACK (to router) */ - LASSERT(tx_ack == NULL || - tx_ack->tx_msg.ksm_type == KSOCK_MSG_NOOP); + LASSERT(!tx_ack || + tx_ack->tx_msg.ksm_type == KSOCK_MSG_NOOP); tx = conn->ksnc_tx_carrier; - if (tx == NULL) { - if (tx_ack != NULL) { + if (!tx) { + if (tx_ack) { list_add_tail(&tx_ack->tx_list, - &conn->ksnc_tx_queue); + &conn->ksnc_tx_queue); conn->ksnc_tx_carrier = tx_ack; } return 0; } - /* conn->ksnc_tx_carrier != NULL */ + /* conn->ksnc_tx_carrier */ - if (tx_ack != NULL) + if (tx_ack) cookie = tx_ack->tx_msg.ksm_zc_cookies[1]; if (cookie == SOCKNAL_KEEPALIVE_PING) /* ignore keepalive PING */ @@ -185,7 +181,7 @@ ksocknal_queue_tx_zcack_v3(ksock_conn_t *conn, if (tx->tx_msg.ksm_zc_cookies[1] == SOCKNAL_KEEPALIVE_PING) { /* replace the keepalive PING with a real ACK */ - LASSERT(tx->tx_msg.ksm_zc_cookies[0] == 0); + LASSERT(!tx->tx_msg.ksm_zc_cookies[0]); tx->tx_msg.ksm_zc_cookies[1] = cookie; return 1; } @@ -197,7 +193,7 @@ ksocknal_queue_tx_zcack_v3(ksock_conn_t *conn, return 1; /* XXX return error in the future */ } - if (tx->tx_msg.ksm_zc_cookies[0] == 0) { + if (!tx->tx_msg.ksm_zc_cookies[0]) { /* NOOP tx has only one ZC-ACK cookie, can carry at least one more */ if (tx->tx_msg.ksm_zc_cookies[1] > cookie) { tx->tx_msg.ksm_zc_cookies[0] = tx->tx_msg.ksm_zc_cookies[1]; @@ -233,7 +229,7 @@ ksocknal_queue_tx_zcack_v3(ksock_conn_t *conn, tmp = tx->tx_msg.ksm_zc_cookies[0]; } - if (tmp != 0) { + if (tmp) { /* range of cookies */ tx->tx_msg.ksm_zc_cookies[0] = tmp - 1; tx->tx_msg.ksm_zc_cookies[1] = tmp + 1; @@ -261,7 +257,7 @@ ksocknal_queue_tx_zcack_v3(ksock_conn_t *conn, } /* failed to piggyback ZC-ACK */ - if (tx_ack != NULL) { + if (tx_ack) { list_add_tail(&tx_ack->tx_list, &conn->ksnc_tx_queue); /* the next tx can piggyback at least 1 ACK */ ksocknal_next_tx_carrier(conn); @@ -280,7 +276,7 @@ ksocknal_match_tx(ksock_conn_t *conn, ksock_tx_t *tx, int nonblk) return SOCKNAL_MATCH_YES; #endif - if (tx == NULL || tx->tx_lnetmsg == NULL) { + if (!tx || !tx->tx_lnetmsg) { /* noop packet */ nob = offsetof(ksock_msg_t, ksm_u); } else { @@ -319,7 +315,7 @@ ksocknal_match_tx_v3(ksock_conn_t *conn, ksock_tx_t *tx, int nonblk) { int nob; - if (tx == NULL || tx->tx_lnetmsg == NULL) + if (!tx || !tx->tx_lnetmsg) nob = offsetof(ksock_msg_t, ksm_u); else nob = tx->tx_lnetmsg->msg_len + sizeof(ksock_msg_t); @@ -334,7 +330,7 @@ ksocknal_match_tx_v3(ksock_conn_t *conn, ksock_tx_t *tx, int nonblk) case SOCKLND_CONN_ACK: if (nonblk) return SOCKNAL_MATCH_YES; - else if (tx == NULL || tx->tx_lnetmsg == NULL) + else if (!tx || !tx->tx_lnetmsg) return SOCKNAL_MATCH_MAY; else return SOCKNAL_MATCH_NO; @@ -369,10 +365,10 @@ ksocknal_handle_zcreq(ksock_conn_t *c, __u64 cookie, int remote) read_lock(&ksocknal_data.ksnd_global_lock); conn = ksocknal_find_conn_locked(peer, NULL, !!remote); - if (conn != NULL) { + if (conn) { ksock_sched_t *sched = conn->ksnc_scheduler; - LASSERT(conn->ksnc_proto->pro_queue_tx_zcack != NULL); + LASSERT(conn->ksnc_proto->pro_queue_tx_zcack); spin_lock_bh(&sched->kss_lock); @@ -390,11 +386,11 @@ ksocknal_handle_zcreq(ksock_conn_t *c, __u64 cookie, int remote) /* ACK connection is not ready, or can't piggyback the ACK */ tx = ksocknal_alloc_tx_noop(cookie, !!remote); - if (tx == NULL) + if (!tx) return -ENOMEM; rc = ksocknal_launch_packet(peer->ksnp_ni, tx, peer->ksnp_id); - if (rc == 0) + if (!rc) return 0; ksocknal_free_tx(tx); @@ -407,11 +403,12 @@ ksocknal_handle_zcack(ksock_conn_t *conn, __u64 cookie1, __u64 cookie2) { ksock_peer_t *peer = conn->ksnc_peer; ksock_tx_t *tx; + ksock_tx_t *temp; ksock_tx_t *tmp; LIST_HEAD(zlist); int count; - if (cookie1 == 0) + if (!cookie1) cookie1 = cookie2; count = (cookie1 > cookie2) ? 2 : (cookie2 - cookie1 + 1); @@ -424,8 +421,8 @@ ksocknal_handle_zcack(ksock_conn_t *conn, __u64 cookie1, __u64 cookie2) spin_lock(&peer->ksnp_lock); - list_for_each_entry_safe(tx, tmp, - &peer->ksnp_zc_req_list, tx_zc_list) { + list_for_each_entry_safe(tx, tmp, &peer->ksnp_zc_req_list, + tx_zc_list) { __u64 c = tx->tx_msg.ksm_zc_cookies[0]; if (c == cookie1 || c == cookie2 || (cookie1 < c && c < cookie2)) { @@ -433,20 +430,19 @@ ksocknal_handle_zcack(ksock_conn_t *conn, __u64 cookie1, __u64 cookie2) list_del(&tx->tx_zc_list); list_add(&tx->tx_zc_list, &zlist); - if (--count == 0) + if (!--count) break; } } spin_unlock(&peer->ksnp_lock); - while (!list_empty(&zlist)) { - tx = list_entry(zlist.next, ksock_tx_t, tx_zc_list); + list_for_each_entry_safe(tx, temp, &zlist, tx_zc_list) { list_del(&tx->tx_zc_list); ksocknal_tx_decref(tx); } - return count == 0 ? 0 : -EPROTO; + return !count ? 0 : -EPROTO; } static int @@ -461,58 +457,59 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello) CLASSERT(sizeof(lnet_magicversion_t) == offsetof(lnet_hdr_t, src_nid)); LIBCFS_ALLOC(hdr, sizeof(*hdr)); - if (hdr == NULL) { + if (!hdr) { CERROR("Can't allocate lnet_hdr_t\n"); return -ENOMEM; } hmv = (lnet_magicversion_t *)&hdr->dest_nid; - /* Re-organize V2.x message header to V1.x (lnet_hdr_t) - * header and send out */ - hmv->magic = cpu_to_le32 (LNET_PROTO_TCP_MAGIC); - hmv->version_major = cpu_to_le16 (KSOCK_PROTO_V1_MAJOR); - hmv->version_minor = cpu_to_le16 (KSOCK_PROTO_V1_MINOR); + /* + * Re-organize V2.x message header to V1.x (lnet_hdr_t) + * header and send out + */ + hmv->magic = cpu_to_le32(LNET_PROTO_TCP_MAGIC); + hmv->version_major = cpu_to_le16(KSOCK_PROTO_V1_MAJOR); + hmv->version_minor = cpu_to_le16(KSOCK_PROTO_V1_MINOR); - if (the_lnet.ln_testprotocompat != 0) { + if (the_lnet.ln_testprotocompat) { /* single-shot proto check */ LNET_LOCK(); - if ((the_lnet.ln_testprotocompat & 1) != 0) { + if (the_lnet.ln_testprotocompat & 1) { hmv->version_major++; /* just different! */ the_lnet.ln_testprotocompat &= ~1; } - if ((the_lnet.ln_testprotocompat & 2) != 0) { + if (the_lnet.ln_testprotocompat & 2) { hmv->magic = LNET_PROTO_MAGIC; the_lnet.ln_testprotocompat &= ~2; } LNET_UNLOCK(); } - hdr->src_nid = cpu_to_le64 (hello->kshm_src_nid); - hdr->src_pid = cpu_to_le32 (hello->kshm_src_pid); - hdr->type = cpu_to_le32 (LNET_MSG_HELLO); - hdr->payload_length = cpu_to_le32 (hello->kshm_nips * sizeof(__u32)); - hdr->msg.hello.type = cpu_to_le32 (hello->kshm_ctype); - hdr->msg.hello.incarnation = cpu_to_le64 (hello->kshm_src_incarnation); + hdr->src_nid = cpu_to_le64(hello->kshm_src_nid); + hdr->src_pid = cpu_to_le32(hello->kshm_src_pid); + hdr->type = cpu_to_le32(LNET_MSG_HELLO); + hdr->payload_length = cpu_to_le32(hello->kshm_nips * sizeof(__u32)); + hdr->msg.hello.type = cpu_to_le32(hello->kshm_ctype); + hdr->msg.hello.incarnation = cpu_to_le64(hello->kshm_src_incarnation); rc = lnet_sock_write(sock, hdr, sizeof(*hdr), lnet_acceptor_timeout()); - if (rc != 0) { + if (rc) { CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n", rc, &conn->ksnc_ipaddr, conn->ksnc_port); goto out; } - if (hello->kshm_nips == 0) + if (!hello->kshm_nips) goto out; - for (i = 0; i < (int) hello->kshm_nips; i++) { - hello->kshm_ips[i] = __cpu_to_le32 (hello->kshm_ips[i]); - } + for (i = 0; i < (int) hello->kshm_nips; i++) + hello->kshm_ips[i] = __cpu_to_le32(hello->kshm_ips[i]); rc = lnet_sock_write(sock, hello->kshm_ips, hello->kshm_nips * sizeof(__u32), lnet_acceptor_timeout()); - if (rc != 0) { + if (rc) { CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n", rc, hello->kshm_nips, &conn->ksnc_ipaddr, conn->ksnc_port); @@ -532,10 +529,10 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello) hello->kshm_magic = LNET_PROTO_MAGIC; hello->kshm_version = conn->ksnc_proto->pro_version; - if (the_lnet.ln_testprotocompat != 0) { + if (the_lnet.ln_testprotocompat) { /* single-shot proto check */ LNET_LOCK(); - if ((the_lnet.ln_testprotocompat & 1) != 0) { + if (the_lnet.ln_testprotocompat & 1) { hello->kshm_version++; /* just different! */ the_lnet.ln_testprotocompat &= ~1; } @@ -544,19 +541,19 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello) rc = lnet_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips), lnet_acceptor_timeout()); - if (rc != 0) { + if (rc) { CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n", rc, &conn->ksnc_ipaddr, conn->ksnc_port); return rc; } - if (hello->kshm_nips == 0) + if (!hello->kshm_nips) return 0; rc = lnet_sock_write(sock, hello->kshm_ips, hello->kshm_nips * sizeof(__u32), lnet_acceptor_timeout()); - if (rc != 0) { + if (rc) { CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n", rc, hello->kshm_nips, &conn->ksnc_ipaddr, conn->ksnc_port); @@ -575,7 +572,7 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello, int i; LIBCFS_ALLOC(hdr, sizeof(*hdr)); - if (hdr == NULL) { + if (!hdr) { CERROR("Can't allocate lnet_hdr_t\n"); return -ENOMEM; } @@ -583,15 +580,15 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello, rc = lnet_sock_read(sock, &hdr->src_nid, sizeof(*hdr) - offsetof(lnet_hdr_t, src_nid), timeout); - if (rc != 0) { + if (rc) { CERROR("Error %d reading rest of HELLO hdr from %pI4h\n", - rc, &conn->ksnc_ipaddr); + rc, &conn->ksnc_ipaddr); LASSERT(rc < 0 && rc != -EALREADY); goto out; } /* ...and check we got what we expected */ - if (hdr->type != cpu_to_le32 (LNET_MSG_HELLO)) { + if (hdr->type != cpu_to_le32(LNET_MSG_HELLO)) { CERROR("Expecting a HELLO hdr, but got type %d from %pI4h\n", le32_to_cpu(hdr->type), &conn->ksnc_ipaddr); @@ -613,14 +610,14 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello, goto out; } - if (hello->kshm_nips == 0) + if (!hello->kshm_nips) goto out; rc = lnet_sock_read(sock, hello->kshm_ips, hello->kshm_nips * sizeof(__u32), timeout); - if (rc != 0) { + if (rc) { CERROR("Error %d reading IPs from ip %pI4h\n", - rc, &conn->ksnc_ipaddr); + rc, &conn->ksnc_ipaddr); LASSERT(rc < 0 && rc != -EALREADY); goto out; } @@ -628,7 +625,7 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello, for (i = 0; i < (int) hello->kshm_nips; i++) { hello->kshm_ips[i] = __le32_to_cpu(hello->kshm_ips[i]); - if (hello->kshm_ips[i] == 0) { + if (!hello->kshm_ips[i]) { CERROR("Zero IP[%d] from ip %pI4h\n", i, &conn->ksnc_ipaddr); rc = -EPROTO; @@ -657,9 +654,9 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout offsetof(ksock_hello_msg_t, kshm_ips) - offsetof(ksock_hello_msg_t, kshm_src_nid), timeout); - if (rc != 0) { + if (rc) { CERROR("Error %d reading HELLO from %pI4h\n", - rc, &conn->ksnc_ipaddr); + rc, &conn->ksnc_ipaddr); LASSERT(rc < 0 && rc != -EALREADY); return rc; } @@ -681,14 +678,14 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout return -EPROTO; } - if (hello->kshm_nips == 0) + if (!hello->kshm_nips) return 0; rc = lnet_sock_read(sock, hello->kshm_ips, hello->kshm_nips * sizeof(__u32), timeout); - if (rc != 0) { + if (rc) { CERROR("Error %d reading IPs from ip %pI4h\n", - rc, &conn->ksnc_ipaddr); + rc, &conn->ksnc_ipaddr); LASSERT(rc < 0 && rc != -EALREADY); return rc; } @@ -697,7 +694,7 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout if (conn->ksnc_flip) __swab32s(&hello->kshm_ips[i]); - if (hello->kshm_ips[i] == 0) { + if (!hello->kshm_ips[i]) { CERROR("Zero IP[%d] from ip %pI4h\n", i, &conn->ksnc_ipaddr); return -EPROTO; @@ -712,12 +709,13 @@ ksocknal_pack_msg_v1(ksock_tx_t *tx) { /* V1.x has no KSOCK_MSG_NOOP */ LASSERT(tx->tx_msg.ksm_type != KSOCK_MSG_NOOP); - LASSERT(tx->tx_lnetmsg != NULL); + LASSERT(tx->tx_lnetmsg); tx->tx_iov[0].iov_base = &tx->tx_lnetmsg->msg_hdr; tx->tx_iov[0].iov_len = sizeof(lnet_hdr_t); - tx->tx_resid = tx->tx_nob = tx->tx_lnetmsg->msg_len + sizeof(lnet_hdr_t); + tx->tx_nob = tx->tx_lnetmsg->msg_len + sizeof(lnet_hdr_t); + tx->tx_resid = tx->tx_lnetmsg->msg_len + sizeof(lnet_hdr_t); } static void @@ -725,17 +723,19 @@ ksocknal_pack_msg_v2(ksock_tx_t *tx) { tx->tx_iov[0].iov_base = &tx->tx_msg; - if (tx->tx_lnetmsg != NULL) { + if (tx->tx_lnetmsg) { LASSERT(tx->tx_msg.ksm_type != KSOCK_MSG_NOOP); tx->tx_msg.ksm_u.lnetmsg.ksnm_hdr = tx->tx_lnetmsg->msg_hdr; tx->tx_iov[0].iov_len = sizeof(ksock_msg_t); - tx->tx_resid = tx->tx_nob = sizeof(ksock_msg_t) + tx->tx_lnetmsg->msg_len; + tx->tx_nob = sizeof(ksock_msg_t) + tx->tx_lnetmsg->msg_len; + tx->tx_resid = sizeof(ksock_msg_t) + tx->tx_lnetmsg->msg_len; } else { LASSERT(tx->tx_msg.ksm_type == KSOCK_MSG_NOOP); tx->tx_iov[0].iov_len = offsetof(ksock_msg_t, ksm_u.lnetmsg.ksnm_hdr); - tx->tx_resid = tx->tx_nob = offsetof(ksock_msg_t, ksm_u.lnetmsg.ksnm_hdr); + tx->tx_nob = offsetof(ksock_msg_t, ksm_u.lnetmsg.ksnm_hdr); + tx->tx_resid = offsetof(ksock_msg_t, ksm_u.lnetmsg.ksnm_hdr); } /* Don't checksum before start sending, because packet can be piggybacked with ACK */ } @@ -745,7 +745,8 @@ ksocknal_unpack_msg_v1(ksock_msg_t *msg) { msg->ksm_csum = 0; msg->ksm_type = KSOCK_MSG_LNET; - msg->ksm_zc_cookies[0] = msg->ksm_zc_cookies[1] = 0; + msg->ksm_zc_cookies[0] = 0; + msg->ksm_zc_cookies[1] = 0; } static void |