X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lnet%2Fklnds%2Fo2iblnd%2Fo2iblnd_cb.c;h=0905627b10beafaa7fa1c87354a52d340459d259;hp=355260f003f5241c374606f5128a30e1b79c4480;hb=ca08e2c30897b345baedd4cf87e4b17714a092b1;hpb=08aa217ce49aba1ded52e0f7adb8a607035123fd diff --git a/lnet/klnds/o2iblnd/o2iblnd_cb.c b/lnet/klnds/o2iblnd/o2iblnd_cb.c index 355260f..0905627 100644 --- a/lnet/klnds/o2iblnd/o2iblnd_cb.c +++ b/lnet/klnds/o2iblnd/o2iblnd_cb.c @@ -43,44 +43,44 @@ void kiblnd_tx_done (lnet_ni_t *ni, kib_tx_t *tx) { - lnet_msg_t *lntmsg[2]; - kib_net_t *net = ni->ni_data; - int rc; - int i; + lnet_msg_t *lntmsg[2]; + kib_net_t *net = ni->ni_data; + int rc; + int i; - LASSERT (net != NULL); - LASSERT (!cfs_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_waiting); /* mustn't be awaiting peer response */ - LASSERT (tx->tx_pool != NULL); + LASSERT (net != NULL); + 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_waiting); /* mustn't be awaiting peer response */ + LASSERT (tx->tx_pool != NULL); - kiblnd_unmap_tx(ni, tx); + kiblnd_unmap_tx(ni, tx); - /* tx may have up to 2 lnet msgs to finalise */ - lntmsg[0] = tx->tx_lntmsg[0]; tx->tx_lntmsg[0] = NULL; - lntmsg[1] = tx->tx_lntmsg[1]; tx->tx_lntmsg[1] = NULL; - rc = tx->tx_status; + /* tx may have up to 2 lnet msgs to finalise */ + lntmsg[0] = tx->tx_lntmsg[0]; tx->tx_lntmsg[0] = NULL; + lntmsg[1] = tx->tx_lntmsg[1]; tx->tx_lntmsg[1] = NULL; + rc = tx->tx_status; - if (tx->tx_conn != NULL) { - LASSERT (ni == tx->tx_conn->ibc_peer->ibp_ni); + if (tx->tx_conn != NULL) { + LASSERT (ni == tx->tx_conn->ibc_peer->ibp_ni); - kiblnd_conn_decref(tx->tx_conn); - tx->tx_conn = NULL; - } + kiblnd_conn_decref(tx->tx_conn); + tx->tx_conn = NULL; + } - tx->tx_nwrq = 0; - tx->tx_status = 0; + tx->tx_nwrq = 0; + tx->tx_status = 0; - kiblnd_pool_free_node(&tx->tx_pool->tpo_pool, &tx->tx_list); + kiblnd_pool_free_node(&tx->tx_pool->tpo_pool, &tx->tx_list); - /* delay finalize until my descs have been freed */ - for (i = 0; i < 2; i++) { - if (lntmsg[i] == NULL) - continue; + /* delay finalize until my descs have been freed */ + for (i = 0; i < 2; i++) { + if (lntmsg[i] == NULL) + continue; - lnet_finalize(ni, lntmsg[i], rc); - } + lnet_finalize(ni, lntmsg[i], rc); + } } void @@ -145,20 +145,20 @@ kiblnd_drop_rx(kib_rx_t *rx) int 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; - int rc; + 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; + int rc; - LASSERT (net != NULL); - LASSERT (!cfs_in_interrupt()); - LASSERT (credit == IBLND_POSTRX_NO_CREDIT || - credit == IBLND_POSTRX_PEER_CREDIT || - credit == IBLND_POSTRX_RSRVD_CREDIT); + LASSERT (net != NULL); + 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); + mr = kiblnd_find_dma_mr(conn->ibc_hdev, rx->rx_msgaddr, IBLND_MSG_SIZE); + LASSERT (mr != NULL); rx->rx_sge.lkey = mr->lkey; rx->rx_sge.addr = rx->rx_msgaddr; @@ -529,8 +529,7 @@ kiblnd_kvaddr_to_page (unsigned long vaddr) { struct page *page; - if (vaddr >= VMALLOC_START && - vaddr < VMALLOC_END) { + if (is_vmalloc_addr((void *)vaddr)) { page = vmalloc_to_page ((void *)vaddr); LASSERT (page != NULL); return page; @@ -551,7 +550,7 @@ kiblnd_kvaddr_to_page (unsigned long vaddr) static int kiblnd_fmr_map_tx(kib_net_t *net, kib_tx_t *tx, kib_rdma_desc_t *rd, int nob) { - kib_hca_dev_t *hdev = tx->tx_pool->tpo_hdev; + kib_hca_dev_t *hdev; __u64 *pages = tx->tx_pages; kib_fmr_poolset_t *fps; int npages; @@ -560,6 +559,11 @@ 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); + + hdev = tx->tx_pool->tpo_hdev; + for (i = 0, npages = 0; i < rd->rd_nfrags; i++) { for (size = 0; size < rd->rd_frags[i].rf_nob; size += hdev->ibh_page_size) { @@ -568,9 +572,6 @@ kiblnd_fmr_map_tx(kib_net_t *net, kib_tx_t *tx, kib_rdma_desc_t *rd, int nob) } } - LASSERT(tx->tx_pool != NULL); - LASSERT(tx->tx_pool->tpo_pool.po_owner != NULL); - cpt = tx->tx_pool->tpo_pool.po_owner->ps_cpt; fps = net->ibn_fmr_ps[cpt]; @@ -594,17 +595,19 @@ kiblnd_fmr_map_tx(kib_net_t *net, kib_tx_t *tx, kib_rdma_desc_t *rd, int nob) static int kiblnd_pmr_map_tx(kib_net_t *net, kib_tx_t *tx, kib_rdma_desc_t *rd, int nob) { - kib_hca_dev_t *hdev = tx->tx_pool->tpo_hdev; + kib_hca_dev_t *hdev; kib_pmr_poolset_t *pps; __u64 iova; int cpt; int rc; - iova = rd->rd_frags[0].rf_addr & ~hdev->ibh_page_mask; - LASSERT(tx->tx_pool != NULL); LASSERT(tx->tx_pool->tpo_pool.po_owner != NULL); + hdev = tx->tx_pool->tpo_hdev; + + iova = rd->rd_frags[0].rf_addr & ~hdev->ibh_page_mask; + cpt = tx->tx_pool->tpo_pool.po_owner->ps_cpt; pps = net->ibn_pmr_ps[cpt]; @@ -1080,23 +1083,23 @@ kiblnd_init_tx_msg (lnet_ni_t *ni, kib_tx_t *tx, int type, int body_nob) 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_send_wr *wrq = &tx->tx_wrq[0]; - int rc = resid; - int srcidx; - int dstidx; - int wrknob; - - LASSERT (!cfs_in_interrupt()); - LASSERT (tx->tx_nwrq == 0); - LASSERT (type == IBLND_MSG_GET_DONE || - type == IBLND_MSG_PUT_DONE); - - srcidx = dstidx = 0; + 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_send_wr *wrq = &tx->tx_wrq[0]; + int rc = resid; + int srcidx; + int dstidx; + int wrknob; + + LASSERT (!in_interrupt()); + LASSERT (tx->tx_nwrq == 0); + LASSERT (type == IBLND_MSG_GET_DONE || + type == IBLND_MSG_PUT_DONE); + + srcidx = dstidx = 0; while (resid > 0) { if (srcidx >= srcrd->rd_nfrags) { @@ -1172,8 +1175,8 @@ kiblnd_queue_tx_locked (kib_tx_t *tx, kib_conn_t *conn) LASSERT (!tx->tx_queued); /* not queued for sending already */ LASSERT (conn->ibc_state >= IBLND_CONN_ESTABLISHED); - tx->tx_queued = 1; - tx->tx_deadline = jiffies + (*kiblnd_tunables.kib_timeout * CFS_HZ); + tx->tx_queued = 1; + tx->tx_deadline = jiffies + (*kiblnd_tunables.kib_timeout * HZ); if (tx->tx_conn == NULL) { kiblnd_conn_addref(conn); @@ -1477,18 +1480,18 @@ 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_niov <= LNET_MAX_IOV); + LASSERT (payload_nob == 0 || payload_niov > 0); + LASSERT (payload_niov <= LNET_MAX_IOV); - /* Thread context */ - LASSERT (!cfs_in_interrupt()); - /* payload is either all vaddrs or all pages */ - LASSERT (!(payload_kiov != NULL && payload_iov != NULL)); + /* Thread context */ + LASSERT (!in_interrupt()); + /* payload is either all vaddrs or all pages */ + LASSERT (!(payload_kiov != NULL && payload_iov != NULL)); - switch (type) { - default: - LBUG(); - return (-EIO); + switch (type) { + default: + LBUG(); + return (-EIO); case LNET_MSG_ACK: LASSERT (payload_nob == 0); @@ -1697,18 +1700,18 @@ kiblnd_recv (lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, kib_conn_t *conn = rx->rx_conn; kib_tx_t *tx; kib_msg_t *txmsg; - int nob; - int post_credit = IBLND_POSTRX_PEER_CREDIT; - int rc = 0; + int nob; + int post_credit = IBLND_POSTRX_PEER_CREDIT; + int rc = 0; - LASSERT (mlen <= rlen); - LASSERT (!cfs_in_interrupt()); - /* Either all pages or all vaddrs */ - LASSERT (!(kiov != NULL && iov != NULL)); + LASSERT (mlen <= rlen); + LASSERT (!in_interrupt()); + /* Either all pages or all vaddrs */ + LASSERT (!(kiov != NULL && iov != NULL)); - switch (rxmsg->ibm_type) { - default: - LBUG(); + switch (rxmsg->ibm_type) { + default: + LBUG(); case IBLND_MSG_IMMEDIATE: nob = offsetof(kib_msg_t, ibm_u.immediate.ibim_payload[rlen]); @@ -1801,29 +1804,29 @@ kiblnd_recv (lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, int delayed, } int -kiblnd_thread_start (int (*fn)(void *arg), void *arg) +kiblnd_thread_start(int (*fn)(void *arg), void *arg, char *name) { - long pid = cfs_create_thread (fn, arg, 0); + struct task_struct *task = kthread_run(fn, arg, name); - if (pid < 0) - return ((int)pid); + if (IS_ERR(task)) + return PTR_ERR(task); - cfs_atomic_inc (&kiblnd_data.kib_nthreads); - return (0); + atomic_inc(&kiblnd_data.kib_nthreads); + return 0; } void kiblnd_thread_fini (void) { - cfs_atomic_dec (&kiblnd_data.kib_nthreads); + atomic_dec (&kiblnd_data.kib_nthreads); } void kiblnd_peer_alive (kib_peer_t *peer) { - /* This is racy, but everyone's only writing cfs_time_current() */ - peer->ibp_last_alive = cfs_time_current(); - cfs_mb(); + /* This is racy, but everyone's only writing cfs_time_current() */ + peer->ibp_last_alive = cfs_time_current(); + smp_mb(); } void @@ -1905,17 +1908,17 @@ kiblnd_close_conn_locked (kib_conn_t *conn, int error) kiblnd_set_conn_state(conn, IBLND_CONN_CLOSING); - if (error != 0 && - kiblnd_dev_can_failover(dev)) { - cfs_list_add_tail(&dev->ibd_fail_list, - &kiblnd_data.kib_failed_devs); - cfs_waitq_signal(&kiblnd_data.kib_failover_waitq); - } + if (error != 0 && + kiblnd_dev_can_failover(dev)) { + cfs_list_add_tail(&dev->ibd_fail_list, + &kiblnd_data.kib_failed_devs); + wake_up(&kiblnd_data.kib_failover_waitq); + } spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); cfs_list_add_tail(&conn->ibc_list, &kiblnd_data.kib_connd_conns); - cfs_waitq_signal(&kiblnd_data.kib_connd_waitq); + wake_up(&kiblnd_data.kib_connd_waitq); spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, flags); } @@ -1938,7 +1941,7 @@ kiblnd_handle_early_rxs(kib_conn_t *conn) unsigned long flags; kib_rx_t *rx; - LASSERT(!cfs_in_interrupt()); + LASSERT(!in_interrupt()); LASSERT(conn->ibc_state >= IBLND_CONN_ESTABLISHED); write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); @@ -1951,7 +1954,7 @@ kiblnd_handle_early_rxs(kib_conn_t *conn) kiblnd_handle_rx(rx); write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); - } + } write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags); } @@ -1994,48 +1997,48 @@ kiblnd_abort_txs(kib_conn_t *conn, cfs_list_t *txs) void kiblnd_finalise_conn (kib_conn_t *conn) { - LASSERT (!cfs_in_interrupt()); - LASSERT (conn->ibc_state > IBLND_CONN_INIT); + LASSERT (!in_interrupt()); + LASSERT (conn->ibc_state > IBLND_CONN_INIT); - kiblnd_set_conn_state(conn, IBLND_CONN_DISCONNECTED); + kiblnd_set_conn_state(conn, IBLND_CONN_DISCONNECTED); - /* 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. */ - kiblnd_abort_receives(conn); + /* 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. */ + 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); - kiblnd_abort_txs(conn, &conn->ibc_tx_queue_nocred); - kiblnd_abort_txs(conn, &conn->ibc_active_txs); + 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); + kiblnd_abort_txs(conn, &conn->ibc_tx_queue_nocred); + kiblnd_abort_txs(conn, &conn->ibc_active_txs); - kiblnd_handle_early_rxs(conn); + kiblnd_handle_early_rxs(conn); } void kiblnd_peer_connect_failed (kib_peer_t *peer, int active, int error) { - CFS_LIST_HEAD (zombies); - unsigned long flags; + CFS_LIST_HEAD (zombies); + unsigned long flags; - LASSERT (error != 0); - LASSERT (!cfs_in_interrupt()); + LASSERT (error != 0); + LASSERT (!in_interrupt()); write_lock_irqsave(&kiblnd_data.kib_global_lock, flags); - if (active) { - LASSERT (peer->ibp_connecting > 0); - peer->ibp_connecting--; - } else { - LASSERT (peer->ibp_accepting > 0); - peer->ibp_accepting--; - } + if (active) { + LASSERT (peer->ibp_connecting > 0); + peer->ibp_connecting--; + } else { + LASSERT (peer->ibp_accepting > 0); + peer->ibp_accepting--; + } - if (peer->ibp_connecting != 0 || + if (peer->ibp_connecting != 0 || peer->ibp_accepting != 0) { /* another connection attempt under way... */ write_unlock_irqrestore(&kiblnd_data.kib_global_lock, @@ -2081,15 +2084,15 @@ kiblnd_connreq_done(kib_conn_t *conn, int status) active = (conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT); - CDEBUG(D_NET,"%s: active(%d), version(%x), status(%d)\n", - libcfs_nid2str(peer->ibp_nid), active, - conn->ibc_version, status); + CDEBUG(D_NET,"%s: active(%d), version(%x), status(%d)\n", + libcfs_nid2str(peer->ibp_nid), active, + conn->ibc_version, status); - LASSERT (!cfs_in_interrupt()); - LASSERT ((conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT && - peer->ibp_connecting > 0) || - (conn->ibc_state == IBLND_CONN_PASSIVE_WAIT && - peer->ibp_accepting > 0)); + LASSERT (!in_interrupt()); + LASSERT ((conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT && + peer->ibp_connecting > 0) || + (conn->ibc_state == IBLND_CONN_PASSIVE_WAIT && + peer->ibp_accepting > 0)); LIBCFS_FREE(conn->ibc_connvars, sizeof(*conn->ibc_connvars)); conn->ibc_connvars = NULL; @@ -2191,15 +2194,15 @@ kiblnd_passive_connect (struct rdma_cm_id *cmid, void *priv, int priv_nob) lnet_nid_t nid; struct rdma_conn_param cp; kib_rej_t rej; - int version = IBLND_MSG_VERSION; - unsigned long flags; - int rc; - struct sockaddr_in *peer_addr; - LASSERT (!cfs_in_interrupt()); + int version = IBLND_MSG_VERSION; + unsigned long flags; + int rc; + struct sockaddr_in *peer_addr; + LASSERT (!in_interrupt()); - /* cmid inherits 'context' from the corresponding listener id */ - ibdev = (kib_dev_t *)cmid->context; - LASSERT (ibdev != NULL); + /* cmid inherits 'context' from the corresponding listener id */ + ibdev = (kib_dev_t *)cmid->context; + LASSERT (ibdev != NULL); memset(&rej, 0, sizeof(rej)); rej.ibr_magic = IBLND_MSG_MAGIC; @@ -2517,16 +2520,16 @@ kiblnd_reconnect (kib_conn_t *conn, int version, void kiblnd_rejected (kib_conn_t *conn, int reason, void *priv, int priv_nob) { - kib_peer_t *peer = conn->ibc_peer; + kib_peer_t *peer = conn->ibc_peer; - LASSERT (!cfs_in_interrupt()); - LASSERT (conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT); + LASSERT (!in_interrupt()); + LASSERT (conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT); - switch (reason) { - case IB_CM_REJ_STALE_CONN: - kiblnd_reconnect(conn, IBLND_MSG_VERSION, 0, - IBLND_REJECT_CONN_STALE, NULL); - break; + switch (reason) { + case IB_CM_REJ_STALE_CONN: + kiblnd_reconnect(conn, IBLND_MSG_VERSION, 0, + IBLND_REJECT_CONN_STALE, NULL); + break; case IB_CM_REJ_INVALID_SERVICE_ID: CNETERR("%s rejected: no listener at %d\n", @@ -2611,14 +2614,18 @@ kiblnd_rejected (kib_conn_t *conn, int reason, void *priv, int priv_nob) case IBLND_REJECT_MSG_QUEUE_SIZE: CERROR("%s rejected: incompatible message queue depth %d, %d\n", - libcfs_nid2str(peer->ibp_nid), cp->ibcp_queue_depth, - IBLND_MSG_QUEUE_SIZE(conn->ibc_version)); + 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->ibcp_max_frags, - IBLND_RDMA_FRAGS(conn->ibc_version)); + libcfs_nid2str(peer->ibp_nid), + cp != NULL ? cp->ibcp_max_frags : + IBLND_RDMA_FRAGS(rej->ibr_version), + IBLND_RDMA_FRAGS(conn->ibc_version)); break; case IBLND_REJECT_NO_RESOURCES: @@ -3109,33 +3116,32 @@ kiblnd_check_conns (int idx) void kiblnd_disconnect_conn (kib_conn_t *conn) { - LASSERT (!cfs_in_interrupt()); - LASSERT (current == kiblnd_data.kib_connd); - LASSERT (conn->ibc_state == IBLND_CONN_CLOSING); + LASSERT (!in_interrupt()); + LASSERT (current == kiblnd_data.kib_connd); + LASSERT (conn->ibc_state == IBLND_CONN_CLOSING); - rdma_disconnect(conn->ibc_cmid); - kiblnd_finalise_conn(conn); + rdma_disconnect(conn->ibc_cmid); + kiblnd_finalise_conn(conn); - kiblnd_peer_notify(conn->ibc_peer); + kiblnd_peer_notify(conn->ibc_peer); } int kiblnd_connd (void *arg) { - cfs_waitlink_t wait; - unsigned long flags; - kib_conn_t *conn; - int timeout; - int i; - int dropped_lock; - int peer_index = 0; - unsigned long deadline = jiffies; + wait_queue_t wait; + unsigned long flags; + kib_conn_t *conn; + int timeout; + int i; + int dropped_lock; + int peer_index = 0; + unsigned long deadline = jiffies; - cfs_daemonize ("kiblnd_connd"); - cfs_block_allsigs (); + cfs_block_allsigs (); - cfs_waitlink_init (&wait); - kiblnd_data.kib_connd = current; + init_waitqueue_entry_current (&wait); + kiblnd_data.kib_connd = current; spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); @@ -3197,13 +3203,13 @@ kiblnd_connd (void *arg) if (chunk == 0) chunk = 1; - for (i = 0; i < chunk; i++) { - kiblnd_check_conns(peer_index); - peer_index = (peer_index + 1) % - kiblnd_data.kib_peer_hash_size; - } + for (i = 0; i < chunk; i++) { + kiblnd_check_conns(peer_index); + peer_index = (peer_index + 1) % + kiblnd_data.kib_peer_hash_size; + } - deadline += p * CFS_HZ; + deadline += p * HZ; spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); } @@ -3211,14 +3217,14 @@ kiblnd_connd (void *arg) continue; /* Nothing to do for 'timeout' */ - cfs_set_current_state(CFS_TASK_INTERRUPTIBLE); - cfs_waitq_add(&kiblnd_data.kib_connd_waitq, &wait); + set_current_state(TASK_INTERRUPTIBLE); + add_wait_queue(&kiblnd_data.kib_connd_waitq, &wait); spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, flags); - cfs_waitq_timedwait(&wait, CFS_TASK_INTERRUPTIBLE, timeout); + waitq_timedwait(&wait, TASK_INTERRUPTIBLE, timeout); - cfs_set_current_state(CFS_TASK_RUNNING); - cfs_waitq_del(&kiblnd_data.kib_connd_waitq, &wait); + set_current_state(TASK_RUNNING); + remove_wait_queue(&kiblnd_data.kib_connd_waitq, &wait); spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags); } @@ -3300,8 +3306,8 @@ kiblnd_cq_completion(struct ib_cq *cq, void *arg) conn->ibc_scheduled = 1; cfs_list_add_tail(&conn->ibc_sched_list, &sched->ibs_conns); - if (cfs_waitq_active(&sched->ibs_waitq)) - cfs_waitq_signal(&sched->ibs_waitq); + if (waitqueue_active(&sched->ibs_waitq)) + wake_up(&sched->ibs_waitq); } spin_unlock_irqrestore(&sched->ibs_lock, flags); @@ -3322,30 +3328,25 @@ kiblnd_scheduler(void *arg) long id = (long)arg; struct kib_sched_info *sched; kib_conn_t *conn; - cfs_waitlink_t wait; + wait_queue_t wait; unsigned long flags; struct ib_wc wc; - char name[20]; int did_something; int busy_loops = 0; int rc; - snprintf(name, sizeof(name), "kiblnd_sd_%02ld_%02ld", - KIB_THREAD_CPT(id), KIB_THREAD_TID(id)); - - cfs_daemonize(name); cfs_block_allsigs(); - cfs_waitlink_init(&wait); + init_waitqueue_entry_current(&wait); sched = kiblnd_data.kib_scheds[KIB_THREAD_CPT(id)]; rc = cfs_cpt_bind(lnet_cpt_table(), sched->ibs_cpt); if (rc != 0) { - CWARN("Failed to bind %s on CPT %d, please verify whether " + 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", name, sched->ibs_cpt); + "performance\n", sched->ibs_cpt); } spin_lock_irqsave(&sched->ibs_lock, flags); @@ -3354,7 +3355,7 @@ kiblnd_scheduler(void *arg) if (busy_loops++ >= IBLND_RESCHED) { spin_unlock_irqrestore(&sched->ibs_lock, flags); - cfs_cond_resched(); + cond_resched(); busy_loops = 0; spin_lock_irqsave(&sched->ibs_lock, flags); @@ -3411,8 +3412,8 @@ kiblnd_scheduler(void *arg) kiblnd_conn_addref(conn); cfs_list_add_tail(&conn->ibc_sched_list, &sched->ibs_conns); - if (cfs_waitq_active(&sched->ibs_waitq)) - cfs_waitq_signal(&sched->ibs_waitq); + if (waitqueue_active(&sched->ibs_waitq)) + wake_up(&sched->ibs_waitq); } else { conn->ibc_scheduled = 0; } @@ -3431,15 +3432,15 @@ kiblnd_scheduler(void *arg) if (did_something) continue; - cfs_set_current_state(CFS_TASK_INTERRUPTIBLE); - cfs_waitq_add_exclusive(&sched->ibs_waitq, &wait); + set_current_state(TASK_INTERRUPTIBLE); + add_wait_queue_exclusive(&sched->ibs_waitq, &wait); spin_unlock_irqrestore(&sched->ibs_lock, flags); - cfs_waitq_wait(&wait, CFS_TASK_INTERRUPTIBLE); + waitq_wait(&wait, TASK_INTERRUPTIBLE); busy_loops = 0; - cfs_waitq_del(&sched->ibs_waitq, &wait); - cfs_set_current_state(CFS_TASK_RUNNING); + remove_wait_queue(&sched->ibs_waitq, &wait); + set_current_state(TASK_RUNNING); spin_lock_irqsave(&sched->ibs_lock, flags); } @@ -3453,17 +3454,16 @@ int kiblnd_failover_thread(void *arg) { rwlock_t *glock = &kiblnd_data.kib_global_lock; - kib_dev_t *dev; - cfs_waitlink_t wait; - unsigned long flags; - int rc; + kib_dev_t *dev; + wait_queue_t wait; + unsigned long flags; + int rc; - LASSERT (*kiblnd_tunables.kib_dev_failover != 0); + LASSERT (*kiblnd_tunables.kib_dev_failover != 0); - cfs_daemonize ("kiblnd_failover"); - cfs_block_allsigs (); + cfs_block_allsigs (); - cfs_waitlink_init(&wait); + init_waitqueue_entry_current(&wait); write_lock_irqsave(glock, flags); while (!kiblnd_data.kib_shutdown) { @@ -3509,14 +3509,14 @@ kiblnd_failover_thread(void *arg) /* long sleep if no more pending failover */ long_sleep = cfs_list_empty(&kiblnd_data.kib_failed_devs); - cfs_set_current_state(CFS_TASK_INTERRUPTIBLE); - cfs_waitq_add(&kiblnd_data.kib_failover_waitq, &wait); + set_current_state(TASK_INTERRUPTIBLE); + add_wait_queue(&kiblnd_data.kib_failover_waitq, &wait); write_unlock_irqrestore(glock, flags); rc = schedule_timeout(long_sleep ? cfs_time_seconds(10) : cfs_time_seconds(1)); - cfs_set_current_state(CFS_TASK_RUNNING); - cfs_waitq_del(&kiblnd_data.kib_failover_waitq, &wait); + set_current_state(TASK_RUNNING); + remove_wait_queue(&kiblnd_data.kib_failover_waitq, &wait); write_lock_irqsave(glock, flags); if (!long_sleep || rc != 0)