diff options
Diffstat (limited to 'drivers/infiniband')
53 files changed, 648 insertions, 454 deletions
diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c index 60d3fbdd216..7767a11b689 100644 --- a/drivers/infiniband/core/addr.c +++ b/drivers/infiniband/core/addr.c @@ -47,6 +47,7 @@ struct addr_req { struct sockaddr src_addr; struct sockaddr dst_addr; struct rdma_dev_addr *addr; + struct rdma_addr_client *client; void *context; void (*callback)(int status, struct sockaddr *src_addr, struct rdma_dev_addr *addr, void *context); @@ -61,6 +62,26 @@ static LIST_HEAD(req_list); static DECLARE_WORK(work, process_req, NULL); static struct workqueue_struct *addr_wq; +void rdma_addr_register_client(struct rdma_addr_client *client) +{ + atomic_set(&client->refcount, 1); + init_completion(&client->comp); +} +EXPORT_SYMBOL(rdma_addr_register_client); + +static inline void put_client(struct rdma_addr_client *client) +{ + if (atomic_dec_and_test(&client->refcount)) + complete(&client->comp); +} + +void rdma_addr_unregister_client(struct rdma_addr_client *client) +{ + put_client(client); + wait_for_completion(&client->comp); +} +EXPORT_SYMBOL(rdma_addr_unregister_client); + int rdma_copy_addr(struct rdma_dev_addr *dev_addr, struct net_device *dev, const unsigned char *dst_dev_addr) { @@ -118,7 +139,7 @@ static void queue_req(struct addr_req *req) mutex_lock(&lock); list_for_each_entry_reverse(temp_req, &req_list, list) { - if (time_after(req->timeout, temp_req->timeout)) + if (time_after_eq(req->timeout, temp_req->timeout)) break; } @@ -204,19 +225,17 @@ static void process_req(void *data) mutex_lock(&lock); list_for_each_entry_safe(req, temp_req, &req_list, list) { - if (req->status) { + if (req->status == -ENODATA) { src_in = (struct sockaddr_in *) &req->src_addr; dst_in = (struct sockaddr_in *) &req->dst_addr; req->status = addr_resolve_remote(src_in, dst_in, req->addr); + if (req->status && time_after_eq(jiffies, req->timeout)) + req->status = -ETIMEDOUT; + else if (req->status == -ENODATA) + continue; } - if (req->status && time_after(jiffies, req->timeout)) - req->status = -ETIMEDOUT; - else if (req->status == -ENODATA) - continue; - - list_del(&req->list); - list_add_tail(&req->list, &done_list); + list_move_tail(&req->list, &done_list); } if (!list_empty(&req_list)) { @@ -229,6 +248,7 @@ static void process_req(void *data) list_del(&req->list); req->callback(req->status, &req->src_addr, req->addr, req->context); + put_client(req->client); kfree(req); } } @@ -264,7 +284,8 @@ static int addr_resolve_local(struct sockaddr_in *src_in, return ret; } -int rdma_resolve_ip(struct sockaddr *src_addr, struct sockaddr *dst_addr, +int rdma_resolve_ip(struct rdma_addr_client *client, + struct sockaddr *src_addr, struct sockaddr *dst_addr, struct rdma_dev_addr *addr, int timeout_ms, void (*callback)(int status, struct sockaddr *src_addr, struct rdma_dev_addr *addr, void *context), @@ -285,6 +306,8 @@ int rdma_resolve_ip(struct sockaddr *src_addr, struct sockaddr *dst_addr, req->addr = addr; req->callback = callback; req->context = context; + req->client = client; + atomic_inc(&client->refcount); src_in = (struct sockaddr_in *) &req->src_addr; dst_in = (struct sockaddr_in *) &req->dst_addr; @@ -305,6 +328,7 @@ int rdma_resolve_ip(struct sockaddr *src_addr, struct sockaddr *dst_addr, break; default: ret = req->status; + atomic_dec(&client->refcount); kfree(req); break; } @@ -321,8 +345,7 @@ void rdma_addr_cancel(struct rdma_dev_addr *addr) if (req->addr == addr) { req->status = -ECANCELED; req->timeout = jiffies; - list_del(&req->list); - list_add(&req->list, &req_list); + list_move(&req->list, &req_list); set_timeout(req->timeout); break; } diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c index 25b1018a476..e5dc4530808 100644 --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -147,12 +147,12 @@ struct cm_id_private { __be32 rq_psn; int timeout_ms; enum ib_mtu path_mtu; + __be16 pkey; u8 private_data_len; u8 max_cm_retries; u8 peer_to_peer; u8 responder_resources; u8 initiator_depth; - u8 local_ack_timeout; u8 retry_count; u8 rnr_retry_count; u8 service_timeout; @@ -240,11 +240,10 @@ static void * cm_copy_private_data(const void *private_data, if (!private_data || !private_data_len) return NULL; - data = kmalloc(private_data_len, GFP_KERNEL); + data = kmemdup(private_data, private_data_len, GFP_KERNEL); if (!data) return ERR_PTR(-ENOMEM); - memcpy(data, private_data, private_data_len); return data; } @@ -691,7 +690,7 @@ static void cm_enter_timewait(struct cm_id_private *cm_id_priv) * timewait before notifying the user that we've exited timewait. */ cm_id_priv->id.state = IB_CM_TIMEWAIT; - wait_time = cm_convert_to_ms(cm_id_priv->local_ack_timeout); + wait_time = cm_convert_to_ms(cm_id_priv->av.packet_life_time + 1); queue_delayed_work(cm.wq, &cm_id_priv->timewait_info->work.work, msecs_to_jiffies(wait_time)); cm_id_priv->timewait_info = NULL; @@ -1010,6 +1009,7 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, cm_id_priv->responder_resources = param->responder_resources; cm_id_priv->retry_count = param->retry_count; cm_id_priv->path_mtu = param->primary_path->mtu; + cm_id_priv->pkey = param->primary_path->pkey; cm_id_priv->qp_type = param->qp_type; ret = cm_alloc_msg(cm_id_priv, &cm_id_priv->msg); @@ -1024,8 +1024,6 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, cm_id_priv->local_qpn = cm_req_get_local_qpn(req_msg); cm_id_priv->rq_psn = cm_req_get_starting_psn(req_msg); - cm_id_priv->local_ack_timeout = - cm_req_get_primary_local_ack_timeout(req_msg); spin_lock_irqsave(&cm_id_priv->lock, flags); ret = ib_post_send_mad(cm_id_priv->msg, NULL); @@ -1410,9 +1408,8 @@ static int cm_req_handler(struct cm_work *work) cm_id_priv->initiator_depth = cm_req_get_resp_res(req_msg); cm_id_priv->responder_resources = cm_req_get_init_depth(req_msg); cm_id_priv->path_mtu = cm_req_get_path_mtu(req_msg); + cm_id_priv->pkey = req_msg->pkey; cm_id_priv->sq_psn = cm_req_get_starting_psn(req_msg); - cm_id_priv->local_ack_timeout = - cm_req_get_primary_local_ack_timeout(req_msg); cm_id_priv->retry_count = cm_req_get_retry_count(req_msg); cm_id_priv->rnr_retry_count = cm_req_get_rnr_retry_count(req_msg); cm_id_priv->qp_type = cm_req_get_qp_type(req_msg); @@ -1716,7 +1713,7 @@ static int cm_establish_handler(struct cm_work *work) unsigned long flags; int ret; - /* See comment in ib_cm_establish about lookup. */ + /* See comment in cm_establish about lookup. */ cm_id_priv = cm_acquire_id(work->local_id, work->remote_id); if (!cm_id_priv) return -EINVAL; @@ -2402,11 +2399,16 @@ int ib_send_cm_lap(struct ib_cm_id *cm_id, cm_id_priv = container_of(cm_id, struct cm_id_private, id); spin_lock_irqsave(&cm_id_priv->lock, flags); if (cm_id->state != IB_CM_ESTABLISHED || - cm_id->lap_state != IB_CM_LAP_IDLE) { + (cm_id->lap_state != IB_CM_LAP_UNINIT && + cm_id->lap_state != IB_CM_LAP_IDLE)) { ret = -EINVAL; goto out; } + ret = cm_init_av_by_path(alternate_path, &cm_id_priv->alt_av); + if (ret) + goto out; + ret = cm_alloc_msg(cm_id_priv, &msg); if (ret) goto out; @@ -2431,7 +2433,8 @@ out: spin_unlock_irqrestore(&cm_id_priv->lock, flags); } EXPORT_SYMBOL(ib_send_cm_lap); -static void cm_format_path_from_lap(struct ib_sa_path_rec *path, +static void cm_format_path_from_lap(struct cm_id_private *cm_id_priv, + struct ib_sa_path_rec *path, struct cm_lap_msg *lap_msg) { memset(path, 0, sizeof *path); @@ -2443,10 +2446,10 @@ static void cm_format_path_from_lap(struct ib_sa_path_rec *path, path->hop_limit = lap_msg->alt_hop_limit; path->traffic_class = cm_lap_get_traffic_class(lap_msg); path->reversible = 1; - /* pkey is same as in REQ */ + path->pkey = cm_id_priv->pkey; path->sl = cm_lap_get_sl(lap_msg); path->mtu_selector = IB_SA_EQ; - /* mtu is same as in REQ */ + path->mtu = cm_id_priv->path_mtu; path->rate_selector = IB_SA_EQ; path->rate = cm_lap_get_packet_rate(lap_msg); path->packet_life_time_selector = IB_SA_EQ; @@ -2472,7 +2475,7 @@ static int cm_lap_handler(struct cm_work *work) param = &work->cm_event.param.lap_rcvd; param->alternate_path = &work->path[0]; - cm_format_path_from_lap(param->alternate_path, lap_msg); + cm_format_path_from_lap(cm_id_priv, param->alternate_path, lap_msg); work->cm_event.private_data = &lap_msg->private_data; spin_lock_irqsave(&cm_id_priv->lock, flags); @@ -2480,6 +2483,7 @@ static int cm_lap_handler(struct cm_work *work) goto unlock; switch (cm_id_priv->id.lap_state) { + case IB_CM_LAP_UNINIT: case IB_CM_LAP_IDLE: break; case IB_CM_MRA_LAP_SENT: @@ -2502,6 +2506,10 @@ static int cm_lap_handler(struct cm_work *work) cm_id_priv->id.lap_state = IB_CM_LAP_RCVD; cm_id_priv->tid = lap_msg->hdr.tid; + cm_init_av_for_response(work->port, work->mad_recv_wc->wc, + work->mad_recv_wc->recv_buf.grh, + &cm_id_priv->av); + cm_init_av_by_path(param->alternate_path, &cm_id_priv->alt_av); ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); @@ -3040,7 +3048,7 @@ static void cm_work_handler(void *data) cm_free_work(work); } -int ib_cm_establish(struct ib_cm_id *cm_id) +static int cm_establish(struct ib_cm_id *cm_id) { struct cm_id_private *cm_id_priv; struct cm_work *work; @@ -3088,7 +3096,44 @@ int ib_cm_establish(struct ib_cm_id *cm_id) out: return ret; } -EXPORT_SYMBOL(ib_cm_establish); + +static int cm_migrate(struct ib_cm_id *cm_id) +{ + struct cm_id_private *cm_id_priv; + unsigned long flags; + int ret = 0; + + cm_id_priv = container_of(cm_id, struct cm_id_private, id); + spin_lock_irqsave(&cm_id_priv->lock, flags); + if (cm_id->state == IB_CM_ESTABLISHED && + (cm_id->lap_state == IB_CM_LAP_UNINIT || + cm_id->lap_state == IB_CM_LAP_IDLE)) { + cm_id->lap_state = IB_CM_LAP_IDLE; + cm_id_priv->av = cm_id_priv->alt_av; + } else + ret = -EINVAL; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + return ret; +} + +int ib_cm_notify(struct ib_cm_id *cm_id, enum ib_event_type event) +{ + int ret; + + switch (event) { + case IB_EVENT_COMM_EST: + ret = cm_establish(cm_id); + break; + case IB_EVENT_PATH_MIG: + ret = cm_migrate(cm_id); + break; + default: + ret = -EINVAL; + } + return ret; +} +EXPORT_SYMBOL(ib_cm_notify); static void cm_recv_handler(struct ib_mad_agent *mad_agent, struct ib_mad_recv_wc *mad_recv_wc) @@ -3173,8 +3218,7 @@ static int cm_init_qp_init_attr(struct cm_id_private *cm_id_priv, case IB_CM_ESTABLISHED: *qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS | IB_QP_PKEY_INDEX | IB_QP_PORT; - qp_attr->qp_access_flags = IB_ACCESS_LOCAL_WRITE | - IB_ACCESS_REMOTE_WRITE; + qp_attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE; if (cm_id_priv->responder_resources) qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_READ | IB_ACCESS_REMOTE_ATOMIC; @@ -3222,6 +3266,9 @@ static int cm_init_qp_rtr_attr(struct cm_id_private *cm_id_priv, if (cm_id_priv->alt_av.ah_attr.dlid) { *qp_attr_mask |= IB_QP_ALT_PATH; qp_attr->alt_port_num = cm_id_priv->alt_av.port->port_num; + qp_attr->alt_pkey_index = cm_id_priv->alt_av.pkey_index; + qp_attr->alt_timeout = + cm_id_priv->alt_av.packet_life_time + 1; qp_attr->alt_ah_attr = cm_id_priv->alt_av.ah_attr; } ret = 0; @@ -3248,19 +3295,31 @@ static int cm_init_qp_rts_attr(struct cm_id_private *cm_id_priv, case IB_CM_REP_SENT: case IB_CM_MRA_REP_RCVD: case IB_CM_ESTABLISHED: - *qp_attr_mask = IB_QP_STATE | IB_QP_SQ_PSN; - qp_attr->sq_psn = be32_to_cpu(cm_id_priv->sq_psn); - if (cm_id_priv->qp_type == IB_QPT_RC) { - *qp_attr_mask |= IB_QP_TIMEOUT | IB_QP_RETRY_CNT | - IB_QP_RNR_RETRY | - IB_QP_MAX_QP_RD_ATOMIC; - qp_attr->timeout = cm_id_priv->local_ack_timeout; - qp_attr->retry_cnt = cm_id_priv->retry_count; - qp_attr->rnr_retry = cm_id_priv->rnr_retry_count; - qp_attr->max_rd_atomic = cm_id_priv->initiator_depth; - } - if (cm_id_priv->alt_av.ah_attr.dlid) { - *qp_attr_mask |= IB_QP_PATH_MIG_STATE; + if (cm_id_priv->id.lap_state == IB_CM_LAP_UNINIT) { + *qp_attr_mask = IB_QP_STATE | IB_QP_SQ_PSN; + qp_attr->sq_psn = be32_to_cpu(cm_id_priv->sq_psn); + if (cm_id_priv->qp_type == IB_QPT_RC) { + *qp_attr_mask |= IB_QP_TIMEOUT | IB_QP_RETRY_CNT | + IB_QP_RNR_RETRY | + IB_QP_MAX_QP_RD_ATOMIC; + qp_attr->timeout = + cm_id_priv->av.packet_life_time + 1; + qp_attr->retry_cnt = cm_id_priv->retry_count; + qp_attr->rnr_retry = cm_id_priv->rnr_retry_count; + qp_attr->max_rd_atomic = + cm_id_priv->initiator_depth; + } + if (cm_id_priv->alt_av.ah_attr.dlid) { + *qp_attr_mask |= IB_QP_PATH_MIG_STATE; + qp_attr->path_mig_state = IB_MIG_REARM; + } + } else { + *qp_attr_mask = IB_QP_ALT_PATH | IB_QP_PATH_MIG_STATE; + qp_attr->alt_port_num = cm_id_priv->alt_av.port->port_num; + qp_attr->alt_pkey_index = cm_id_priv->alt_av.pkey_index; + qp_attr->alt_timeout = + cm_id_priv->alt_av.packet_life_time + 1; + qp_attr->alt_ah_attr = cm_id_priv->alt_av.ah_attr; qp_attr->path_mig_state = IB_MIG_REARM; } ret = 0; diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 9ae4f3a67c7..cf48f269743 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -63,6 +63,7 @@ static struct ib_client cma_client = { }; static struct ib_sa_client sa_client; +static struct rdma_addr_client addr_client; static LIST_HEAD(dev_list); static LIST_HEAD(listen_any_list); static DEFINE_MUTEX(lock); @@ -343,7 +344,7 @@ static int cma_init_ib_qp(struct rdma_id_private *id_priv, struct ib_qp *qp) return ret; qp_attr.qp_state = IB_QPS_INIT; - qp_attr.qp_access_flags = IB_ACCESS_LOCAL_WRITE; + qp_attr.qp_access_flags = 0; qp_attr.port_num = id_priv->id.port_num; return ib_modify_qp(qp, &qp_attr, IB_QP_STATE | IB_QP_ACCESS_FLAGS | IB_QP_PKEY_INDEX | IB_QP_PORT); @@ -934,13 +935,8 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event) mutex_lock(&lock); ret = cma_acquire_dev(conn_id); mutex_unlock(&lock); - if (ret) { - ret = -ENODEV; - cma_exch(conn_id, CMA_DESTROYING); - cma_release_remove(conn_id); - rdma_destroy_id(&conn_id->id); - goto out; - } + if (ret) + goto release_conn_id; conn_id->cm_id.ib = cm_id; cm_id->context = conn_id; @@ -950,13 +946,17 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event) ret = cma_notify_user(conn_id, RDMA_CM_EVENT_CONNECT_REQUEST, 0, ib_event->private_data + offset, IB_CM_REQ_PRIVATE_DATA_SIZE - offset); - if (ret) { - /* Destroy the CM ID by returning a non-zero value. */ - conn_id->cm_id.ib = NULL; - cma_exch(conn_id, CMA_DESTROYING); - cma_release_remove(conn_id); - rdma_destroy_id(&conn_id->id); - } + if (!ret) + goto out; + + /* Destroy the CM ID by returning a non-zero value. */ + conn_id->cm_id.ib = NULL; + +release_conn_id: + cma_exch(conn_id, CMA_DESTROYING); + cma_release_remove(conn_id); + rdma_destroy_id(&conn_id->id); + out: cma_release_remove(listen_id); return ret; @@ -1480,19 +1480,18 @@ static int cma_bind_loopback(struct rdma_id_private *id_priv) u8 p; mutex_lock(&lock); + if (list_empty(&dev_list)) { + ret = -ENODEV; + goto out; + } list_for_each_entry(cma_dev, &dev_list, list) for (p = 1; p <= cma_dev->device->phys_port_cnt; ++p) - if (!ib_query_port (cma_dev->device, p, &port_attr) && + if (!ib_query_port(cma_dev->device, p, &port_attr) && port_attr.state == IB_PORT_ACTIVE) goto port_found; - if (!list_empty(&dev_list)) { - p = 1; - cma_dev = list_entry(dev_list.next, struct cma_device, list); - } else { - ret = -ENODEV; - goto out; - } + p = 1; + cma_dev = list_entry(dev_list.next, struct cma_device, list); port_found: ret = ib_get_cached_gid(cma_dev->device, p, 0, &gid); @@ -1625,8 +1624,8 @@ int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, if (cma_any_addr(dst_addr)) ret = cma_resolve_loopback(id_priv); else - ret = rdma_resolve_ip(&id->route.addr.src_addr, dst_addr, - &id->route.addr.dev_addr, + ret = rdma_resolve_ip(&addr_client, &id->route.addr.src_addr, + dst_addr, &id->route.addr.dev_addr, timeout_ms, addr_handler, id_priv); if (ret) goto err; @@ -1762,22 +1761,29 @@ int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr) if (!cma_any_addr(addr)) { ret = rdma_translate_ip(addr, &id->route.addr.dev_addr); - if (!ret) { - mutex_lock(&lock); - ret = cma_acquire_dev(id_priv); - mutex_unlock(&lock); - } if (ret) - goto err; + goto err1; + + mutex_lock(&lock); + ret = cma_acquire_dev(id_priv); + mutex_unlock(&lock); + if (ret) + goto err1; } memcpy(&id->route.addr.src_addr, addr, ip_addr_size(addr)); ret = cma_get_port(id_priv); if (ret) - goto err; + goto err2; return 0; -err: +err2: + if (!cma_any_addr(addr)) { + mutex_lock(&lock); + cma_detach_from_dev(id_priv); + mutex_unlock(&lock); + } +err1: cma_comp_exch(id_priv, CMA_ADDR_BOUND, CMA_IDLE); return ret; } @@ -2115,8 +2121,6 @@ static void cma_add_one(struct ib_device *device) cma_dev->device = device; cma_dev->node_guid = device->node_guid; - if (!cma_dev->node_guid) - goto err; init_completion(&cma_dev->comp); atomic_set(&cma_dev->refcount, 1); @@ -2128,9 +2132,6 @@ static void cma_add_one(struct ib_device *device) list_for_each_entry(id_priv, &listen_any_list, list) cma_listen_on_dev(id_priv, cma_dev); mutex_unlock(&lock); - return; -err: - kfree(cma_dev); } static int cma_remove_id_dev(struct rdma_id_private *id_priv) @@ -2210,6 +2211,7 @@ static int cma_init(void) return -ENOMEM; ib_sa_register_client(&sa_client); + rdma_addr_register_client(&addr_client); ret = ib_register_client(&cma_client); if (ret) @@ -2217,6 +2219,7 @@ static int cma_init(void) return 0; err: + rdma_addr_unregister_client(&addr_client); ib_sa_unregister_client(&sa_client); destroy_workqueue(cma_wq); return ret; @@ -2225,6 +2228,7 @@ err: static void cma_cleanup(void) { ib_unregister_client(&cma_client); + rdma_addr_unregister_client(&addr_client); ib_sa_unregister_client(&sa_client); destroy_workqueue(cma_wq); idr_destroy(&sdp_ps); diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c index c3fb304a4e8..cf797d7aea0 100644 --- a/drivers/infiniband/core/iwcm.c +++ b/drivers/infiniband/core/iwcm.c @@ -80,7 +80,7 @@ struct iwcm_work { * 1) in the event upcall, cm_event_handler(), for a listening cm_id. If * the backlog is exceeded, then no more connection request events will * be processed. cm_event_handler() returns -ENOMEM in this case. Its up - * to the provider to reject the connectino request. + * to the provider to reject the connection request. * 2) in the connection request workqueue handler, cm_conn_req_handler(). * If work elements cannot be allocated for the new connect request cm_id, * then IWCM will call the provider reject method. This is ok since @@ -131,26 +131,25 @@ static int alloc_work_entries(struct iwcm_id_private *cm_id_priv, int count) } /* - * Save private data from incoming connection requests in the - * cm_id_priv so the low level driver doesn't have to. Adjust + * Save private data from incoming connection requests to + * iw_cm_event, so the low level driver doesn't have to. Adjust * the event ptr to point to the local copy. */ -static int copy_private_data(struct iwcm_id_private *cm_id_priv, - struct iw_cm_event *event) +static int copy_private_data(struct iw_cm_event *event) { void *p; - p = kmalloc(event->private_data_len, GFP_ATOMIC); + p = kmemdup(event->private_data, event->private_data_len, GFP_ATOMIC); if (!p) return -ENOMEM; - memcpy(p, event->private_data, event->private_data_len); event->private_data = p; return 0; } /* - * Release a reference on cm_id. If the last reference is being removed - * and iw_destroy_cm_id is waiting, wake up the waiting thread. + * Release a reference on cm_id. If the last reference is being + * released, enable the waiting thread (in iw_destroy_cm_id) to + * get woken up, and return 1 if a thread is already waiting. */ static int iwcm_deref_id(struct iwcm_id_private *cm_id_priv) { @@ -243,7 +242,7 @@ static int iwcm_modify_qp_sqd(struct ib_qp *qp) /* * CM_ID <-- CLOSING * - * Block if a passive or active connection is currenlty being processed. Then + * Block if a passive or active connection is currently being processed. Then * process the event as follows: * - If we are ESTABLISHED, move to CLOSING and modify the QP state * based on the abrupt flag @@ -408,7 +407,7 @@ int iw_cm_listen(struct iw_cm_id *cm_id, int backlog) { struct iwcm_id_private *cm_id_priv; unsigned long flags; - int ret = 0; + int ret; cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); @@ -535,7 +534,7 @@ EXPORT_SYMBOL(iw_cm_accept); int iw_cm_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *iw_param) { struct iwcm_id_private *cm_id_priv; - int ret = 0; + int ret; unsigned long flags; struct ib_qp *qp; @@ -620,7 +619,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv, spin_lock_irqsave(&listen_id_priv->lock, flags); if (listen_id_priv->state != IW_CM_STATE_LISTEN) { spin_unlock_irqrestore(&listen_id_priv->lock, flags); - return; + goto out; } spin_unlock_irqrestore(&listen_id_priv->lock, flags); @@ -629,7 +628,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv, listen_id_priv->id.context); /* If the cm_id could not be created, ignore the request */ if (IS_ERR(cm_id)) - return; + goto out; cm_id->provider_data = iw_event->provider_data; cm_id->local_addr = iw_event->local_addr; @@ -642,7 +641,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv, if (ret) { iw_cm_reject(cm_id, NULL, 0); iw_destroy_cm_id(cm_id); - return; + goto out; } /* Call the client CM handler */ @@ -654,6 +653,7 @@ static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv, kfree(cm_id); } +out: if (iw_event->private_data_len) kfree(iw_event->private_data); } @@ -674,7 +674,7 @@ static int cm_conn_est_handler(struct iwcm_id_private *cm_id_priv, struct iw_cm_event *iw_event) { unsigned long flags; - int ret = 0; + int ret; spin_lock_irqsave(&cm_id_priv->lock, flags); @@ -704,7 +704,7 @@ static int cm_conn_rep_handler(struct iwcm_id_private *cm_id_priv, struct iw_cm_event *iw_event) { unsigned long flags; - int ret = 0; + int ret; spin_lock_irqsave(&cm_id_priv->lock, flags); /* @@ -830,7 +830,8 @@ static int process_event(struct iwcm_id_private *cm_id_priv, */ static void cm_work_handler(void *arg) { - struct iwcm_work *work = arg, lwork; + struct iwcm_work *work = arg; + struct iw_cm_event levent; struct iwcm_id_private *cm_id_priv = work->cm_id; unsigned long flags; int empty; @@ -843,11 +844,11 @@ static void cm_work_handler(void *arg) struct iwcm_work, list); list_del_init(&work->list); empty = list_empty(&cm_id_priv->work_list); - lwork = *work; + levent = work->event; put_work(work); spin_unlock_irqrestore(&cm_id_priv->lock, flags); - ret = process_event(cm_id_priv, &work->event); + ret = process_event(cm_id_priv, &levent); if (ret) { set_bit(IWCM_F_CALLBACK_DESTROY, &cm_id_priv->flags); destroy_cm_id(&cm_id_priv->id); @@ -906,7 +907,7 @@ static int cm_event_handler(struct iw_cm_id *cm_id, if ((work->event.event == IW_CM_EVENT_CONNECT_REQUEST || work->event.event == IW_CM_EVENT_CONNECT_REPLY) && work->event.private_data_len) { - ret = copy_private_data(cm_id_priv, &work->event); + ret = copy_private_data(&work->event); if (ret) { put_work(work); goto out; diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index 493f4c65c7a..3f9c16232c4 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -46,7 +46,7 @@ MODULE_DESCRIPTION("kernel IB MAD API"); MODULE_AUTHOR("Hal Rosenstock"); MODULE_AUTHOR("Sean Hefty"); -static kmem_cache_t *ib_mad_cache; +static struct kmem_cache *ib_mad_cache; static struct list_head ib_mad_port_list; static u32 ib_mad_client_id = 0; @@ -1750,7 +1750,7 @@ ib_find_send_mad(struct ib_mad_agent_private *mad_agent_priv, */ (is_direct(wc->recv_buf.mad->mad_hdr.mgmt_class) || rcv_has_same_gid(mad_agent_priv, wr, wc))) - return wr; + return (wr->status == IB_WC_SUCCESS) ? wr : NULL; } /* diff --git a/drivers/infiniband/core/ucm.c b/drivers/infiniband/core/ucm.c index ad4f4d5c292..f15220a0ee7 100644 --- a/drivers/infiniband/core/ucm.c +++ b/drivers/infiniband/core/ucm.c @@ -161,12 +161,14 @@ static void ib_ucm_cleanup_events(struct ib_ucm_context *ctx) struct ib_ucm_event, ctx_list); list_del(&uevent->file_list); list_del(&uevent->ctx_list); + mutex_unlock(&ctx->file->file_mutex); /* clear incoming connections. */ if (ib_ucm_new_cm_id(uevent->resp.event)) ib_destroy_cm_id(uevent->cm_id); kfree(uevent); + mutex_lock(&ctx->file->file_mutex); } mutex_unlock(&ctx->file->file_mutex); } @@ -328,20 +330,18 @@ static int ib_ucm_event_process(struct ib_cm_event *evt, } if (uvt->data_len) { - uvt->data = kmalloc(uvt->data_len, GFP_KERNEL); + uvt->data = kmemdup(evt->private_data, uvt->data_len, GFP_KERNEL); if (!uvt->data) goto err1; - memcpy(uvt->data, evt->private_data, uvt->data_len); uvt->resp.present |= IB_UCM_PRES_DATA; } if (uvt->info_len) { - uvt->info = kmalloc(uvt->info_len, GFP_KERNEL); + uvt->info = kmemdup(info, uvt->info_len, GFP_KERNEL); if (!uvt->info) goto err2; - memcpy(uvt->info, info, uvt->info_len); uvt->resp.present |= IB_UCM_PRES_INFO; } return 0; @@ -685,11 +685,11 @@ out: return result; } -static ssize_t ib_ucm_establish(struct ib_ucm_file *file, - const char __user *inbuf, - int in_len, int out_len) +static ssize_t ib_ucm_notify(struct ib_ucm_file *file, + const char __user *inbuf, + int in_len, int out_len) { - struct ib_ucm_establish cmd; + struct ib_ucm_notify cmd; struct ib_ucm_context *ctx; int result; @@ -700,7 +700,7 @@ static ssize_t ib_ucm_establish(struct ib_ucm_file *file, if (IS_ERR(ctx)) return PTR_ERR(ctx); - result = ib_cm_establish(ctx->cm_id); + result = ib_cm_notify(ctx->cm_id, (enum ib_event_type) cmd.event); ib_ucm_ctx_put(ctx); return result; } @@ -1107,7 +1107,7 @@ static ssize_t (*ucm_cmd_table[])(struct ib_ucm_file *file, [IB_USER_CM_CMD_DESTROY_ID] = ib_ucm_destroy_id, [IB_USER_CM_CMD_ATTR_ID] = ib_ucm_attr_id, [IB_USER_CM_CMD_LISTEN] = ib_ucm_listen, - [IB_USER_CM_CMD_ESTABLISH] = ib_ucm_establish, + [IB_USER_CM_CMD_NOTIFY] = ib_ucm_notify, [IB_USER_CM_CMD_SEND_REQ] = ib_ucm_send_req, [IB_USER_CM_CMD_SEND_REP] = ib_ucm_send_rep, [IB_USER_CM_CMD_SEND_RTU] = ib_ucm_send_rtu, diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c index b72c7f69ca9..743247ec065 100644 --- a/drivers/infiniband/core/uverbs_cmd.c +++ b/drivers/infiniband/core/uverbs_cmd.c @@ -1214,7 +1214,7 @@ ssize_t ib_uverbs_query_qp(struct ib_uverbs_file *file, resp.qp_access_flags = attr->qp_access_flags; resp.pkey_index = attr->pkey_index; resp.alt_pkey_index = attr->alt_pkey_index; - resp.en_sqd_async_notify = attr->en_sqd_async_notify; + resp.sq_draining = attr->sq_draining; resp.max_rd_atomic = attr->max_rd_atomic; resp.max_dest_rd_atomic = attr->max_dest_rd_atomic; resp.min_rnr_timer = attr->min_rnr_timer; diff --git a/drivers/infiniband/hw/amso1100/c2.c b/drivers/infiniband/hw/amso1100/c2.c index dc1ebeac35c..27fe242ed43 100644 --- a/drivers/infiniband/hw/amso1100/c2.c +++ b/drivers/infiniband/hw/amso1100/c2.c @@ -1155,7 +1155,8 @@ static int __devinit c2_probe(struct pci_dev *pcidev, goto bail10; } - c2_register_device(c2dev); + if (c2_register_device(c2dev)) + goto bail10; return 0; @@ -1243,7 +1244,7 @@ static struct pci_driver c2_pci_driver = { static int __init c2_init_module(void) { - return pci_module_init(&c2_pci_driver); + return pci_register_driver(&c2_pci_driver); } static void __exit c2_exit_module(void) diff --git a/drivers/infiniband/hw/amso1100/c2.h b/drivers/infiniband/hw/amso1100/c2.h index 1b17dcdd050..04a9db5de88 100644 --- a/drivers/infiniband/hw/amso1100/c2.h +++ b/drivers/infiniband/hw/amso1100/c2.h @@ -302,7 +302,7 @@ struct c2_dev { unsigned long pa; /* PA device memory */ void **qptr_array; - kmem_cache_t *host_msg_cache; + struct kmem_cache *host_msg_cache; struct list_head cca_link; /* adapter list */ struct list_head eh_wakeup_list; /* event wakeup list */ diff --git a/drivers/infiniband/hw/amso1100/c2_alloc.c b/drivers/infiniband/hw/amso1100/c2_alloc.c index 028a60bbfca..0315f99e419 100644 --- a/drivers/infiniband/hw/amso1100/c2_alloc.c +++ b/drivers/infiniband/hw/amso1100/c2_alloc.c @@ -42,13 +42,14 @@ static int c2_alloc_mqsp_chunk(struct c2_dev *c2dev, gfp_t gfp_mask, { int i; struct sp_chunk *new_head; + dma_addr_t dma_addr; - new_head = (struct sp_chunk *) __get_free_page(gfp_mask); + new_head = dma_alloc_coherent(&c2dev->pcidev->dev, PAGE_SIZE, + &dma_addr, gfp_mask); if (new_head == NULL) return -ENOMEM; - new_head->dma_addr = dma_map_single(c2dev->ibdev.dma_device, new_head, - PAGE_SIZE, DMA_FROM_DEVICE); + new_head->dma_addr = dma_addr; pci_unmap_addr_set(new_head, mapping, new_head->dma_addr); new_head->next = NULL; @@ -80,10 +81,8 @@ void c2_free_mqsp_pool(struct c2_dev *c2dev, struct sp_chunk *root) while (root) { next = root->next; - dma_unmap_single(c2dev->ibdev.dma_device, - pci_unmap_addr(root, mapping), PAGE_SIZE, - DMA_FROM_DEVICE); - __free_page((struct page *) root); + dma_free_coherent(&c2dev->pcidev->dev, PAGE_SIZE, root, + pci_unmap_addr(root, mapping)); root = next; } } diff --git a/drivers/infiniband/hw/amso1100/c2_cq.c b/drivers/infiniband/hw/amso1100/c2_cq.c index 9d7bcc5ade9..05c9154d46f 100644 --- a/drivers/infiniband/hw/amso1100/c2_cq.c +++ b/drivers/infiniband/hw/amso1100/c2_cq.c @@ -246,20 +246,17 @@ int c2_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify notify) static void c2_free_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq) { - - dma_unmap_single(c2dev->ibdev.dma_device, pci_unmap_addr(mq, mapping), - mq->q_size * mq->msg_size, DMA_FROM_DEVICE); - free_pages((unsigned long) mq->msg_pool.host, - get_order(mq->q_size * mq->msg_size)); + dma_free_coherent(&c2dev->pcidev->dev, mq->q_size * mq->msg_size, + mq->msg_pool.host, pci_unmap_addr(mq, mapping)); } static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq, int q_size, int msg_size) { - unsigned long pool_start; + u8 *pool_start; - pool_start = __get_free_pages(GFP_KERNEL, - get_order(q_size * msg_size)); + pool_start = dma_alloc_coherent(&c2dev->pcidev->dev, q_size * msg_size, + &mq->host_dma, GFP_KERNEL); if (!pool_start) return -ENOMEM; @@ -267,13 +264,10 @@ static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq, int q_size, 0, /* index (currently unknown) */ q_size, msg_size, - (u8 *) pool_start, + pool_start, NULL, /* peer (currently unknown) */ C2_MQ_HOST_TARGET); - mq->host_dma = dma_map_single(c2dev->ibdev.dma_device, - (void *)pool_start, - q_size * msg_size, DMA_FROM_DEVICE); pci_unmap_addr_set(mq, mapping, mq->host_dma); return 0; diff --git a/drivers/infiniband/hw/amso1100/c2_provider.c b/drivers/infiniband/hw/amso1100/c2_provider.c index da98d9f7142..fef97275291 100644 --- a/drivers/infiniband/hw/amso1100/c2_provider.c +++ b/drivers/infiniband/hw/amso1100/c2_provider.c @@ -757,20 +757,17 @@ static struct net_device *c2_pseudo_netdev_init(struct c2_dev *c2dev) int c2_register_device(struct c2_dev *dev) { - int ret; + int ret = -ENOMEM; int i; /* Register pseudo network device */ dev->pseudo_netdev = c2_pseudo_netdev_init(dev); - if (dev->pseudo_netdev) { - ret = register_netdev(dev->pseudo_netdev); - if (ret) { - printk(KERN_ERR PFX - "Unable to register netdev, ret = %d\n", ret); - free_netdev(dev->pseudo_netdev); - return ret; - } - } + if (!dev->pseudo_netdev) + goto out3; + + ret = register_netdev(dev->pseudo_netdev); + if (ret) + goto out2; pr_debug("%s:%u\n", __FUNCTION__, __LINE__); strlcpy(dev->ibdev.name, "amso%d", IB_DEVICE_NAME_MAX); @@ -848,21 +845,25 @@ int c2_register_device(struct c2_dev *dev) ret = ib_register_device(&dev->ibdev); if (ret) - return ret; + goto out1; for (i = 0; i < ARRAY_SIZE(c2_class_attributes); ++i) { ret = class_device_create_file(&dev->ibdev.class_dev, c2_class_attributes[i]); - if (ret) { - unregister_netdev(dev->pseudo_netdev); - free_netdev(dev->pseudo_netdev); - ib_unregister_device(&dev->ibdev); - return ret; - } + if (ret) + goto out0; } + goto out3; - pr_debug("%s:%u\n", __FUNCTION__, __LINE__); - return 0; +out0: + ib_unregister_device(&dev->ibdev); +out1: + unregister_netdev(dev->pseudo_netdev); +out2: + free_netdev(dev->pseudo_netdev); +out3: + pr_debug("%s:%u ret=%d\n", __FUNCTION__, __LINE__, ret); + return ret; } void c2_unregister_device(struct c2_dev *dev) diff --git a/drivers/infiniband/hw/amso1100/c2_qp.c b/drivers/infiniband/hw/amso1100/c2_qp.c index 5bcf697aa33..179d005ed4a 100644 --- a/drivers/infiniband/hw/amso1100/c2_qp.c +++ b/drivers/infiniband/hw/amso1100/c2_qp.c @@ -564,6 +564,32 @@ int c2_alloc_qp(struct c2_dev *c2dev, return err; } +static inline void c2_lock_cqs(struct c2_cq *send_cq, struct c2_cq *recv_cq) +{ + if (send_cq == recv_cq) + spin_lock_irq(&send_cq->lock); + else if (send_cq > recv_cq) { + spin_lock_irq(&send_cq->lock); + spin_lock_nested(&recv_cq->lock, SINGLE_DEPTH_NESTING); + } else { + spin_lock_irq(&recv_cq->lock); + spin_lock_nested(&send_cq->lock, SINGLE_DEPTH_NESTING); + } +} + +static inline void c2_unlock_cqs(struct c2_cq *send_cq, struct c2_cq *recv_cq) +{ + if (send_cq == recv_cq) + spin_unlock_irq(&send_cq->lock); + else if (send_cq > recv_cq) { + spin_unlock(&recv_cq->lock); + spin_unlock_irq(&send_cq->lock); + } else { + spin_unlock(&send_cq->lock); + spin_unlock_irq(&recv_cq->lock); + } +} + void c2_free_qp(struct c2_dev *c2dev, struct c2_qp *qp) { struct c2_cq *send_cq; @@ -576,15 +602,9 @@ void c2_free_qp(struct c2_dev *c2dev, struct c2_qp *qp) * Lock CQs here, so that CQ polling code can do QP lookup * without taking a lock. */ - spin_lock_irq(&send_cq->lock); - if (send_cq != recv_cq) - spin_lock(&recv_cq->lock); - + c2_lock_cqs(send_cq, recv_cq); c2_free_qpn(c2dev, qp->qpn); - - if (send_cq != recv_cq) - spin_unlock(&recv_cq->lock); - spin_unlock_irq(&send_cq->lock); + c2_unlock_cqs(send_cq, recv_cq); /* * Destory qp in the rnic... diff --git a/drivers/infiniband/hw/amso1100/c2_rnic.c b/drivers/infiniband/hw/amso1100/c2_rnic.c index e37c5688c21..1687c511cb2 100644 --- a/drivers/infiniband/hw/amso1100/c2_rnic.c +++ b/drivers/infiniband/hw/amso1100/c2_rnic.c @@ -150,15 +150,15 @@ static int c2_rnic_query(struct c2_dev *c2dev, struct ib_device_attr *props) (struct c2wr_rnic_query_rep *) (unsigned long) (vq_req->reply_msg); if (!reply) err = -ENOMEM; - - err = c2_errno(reply); + else + err = c2_errno(reply); if (err) goto bail2; props->fw_ver = ((u64)be32_to_cpu(reply->fw_ver_major) << 32) | - ((be32_to_cpu(reply->fw_ver_minor) && 0xFFFF) << 16) | - (be32_to_cpu(reply->fw_ver_patch) && 0xFFFF); + ((be32_to_cpu(reply->fw_ver_minor) & 0xFFFF) << 16) | + (be32_to_cpu(reply->fw_ver_patch) & 0xFFFF); memcpy(&props->sys_image_guid, c2dev->netdev->dev_addr, 6); props->max_mr_size = 0xFFFFFFFF; props->page_size_cap = ~(C2_MIN_PAGESIZE-1); @@ -441,7 +441,7 @@ static int c2_rnic_close(struct c2_dev *c2dev) * involves initalizing the various limits and resouce pools that * comprise the RNIC instance. */ -int c2_rnic_init(struct c2_dev *c2dev) +int __devinit c2_rnic_init(struct c2_dev *c2dev) { int err; u32 qsize, msgsize; @@ -517,14 +517,12 @@ int c2_rnic_init(struct c2_dev *c2dev) /* Initialize the Verbs Reply Queue */ qsize = be32_to_cpu(readl(mmio_regs + C2_REGS_Q1_QSIZE)); msgsize = be32_to_cpu(readl(mmio_regs + C2_REGS_Q1_MSGSIZE)); - q1_pages = kmalloc(qsize * msgsize, GFP_KERNEL); + q1_pages = dma_alloc_coherent(&c2dev->pcidev->dev, qsize * msgsize, + &c2dev->rep_vq.host_dma, GFP_KERNEL); if (!q1_pages) { err = -ENOMEM; goto bail1; } - c2dev->rep_vq.host_dma = dma_map_single(c2dev->ibdev.dma_device, - (void *)q1_pages, qsize * msgsize, - DMA_FROM_DEVICE); pci_unmap_addr_set(&c2dev->rep_vq, mapping, c2dev->rep_vq.host_dma); pr_debug("%s rep_vq va %p dma %llx\n", __FUNCTION__, q1_pages, (unsigned long long) c2dev->rep_vq.host_dma); @@ -540,17 +538,15 @@ int c2_rnic_init(struct c2_dev *c2dev) /* Initialize the Asynchronus Event Queue */ qsize = be32_to_cpu(readl(mmio_regs + C2_REGS_Q2_QSIZE)); msgsize = be32_to_cpu(readl(mmio_regs + C2_REGS_Q2_MSGSIZE)); - q2_pages = kmalloc(qsize * msgsize, GFP_KERNEL); + q2_pages = dma_alloc_coherent(&c2dev->pcidev->dev, qsize * msgsize, + &c2dev->aeq.host_dma, GFP_KERNEL); if (!q2_pages) { err = -ENOMEM; goto bail2; } - c2dev->aeq.host_dma = dma_map_single(c2dev->ibdev.dma_device, - (void *)q2_pages, qsize * msgsize, - DMA_FROM_DEVICE); pci_unmap_addr_set(&c2dev->aeq, mapping, c2dev->aeq.host_dma); - pr_debug("%s aeq va %p dma %llx\n", __FUNCTION__, q1_pages, - (unsigned long long) c2dev->rep_vq.host_dma); + pr_debug("%s aeq va %p dma %llx\n", __FUNCTION__, q2_pages, + (unsigned long long) c2dev->aeq.host_dma); c2_mq_rep_init(&c2dev->aeq, 2, qsize, @@ -597,17 +593,13 @@ int c2_rnic_init(struct c2_dev *c2dev) bail4: vq_term(c2dev); bail3: - dma_unmap_single(c2dev->ibdev.dma_device, - pci_unmap_addr(&c2dev->aeq, mapping), - c2dev->aeq.q_size * c2dev->aeq.msg_size, - DMA_FROM_DEVICE); - kfree(q2_pages); + dma_free_coherent(&c2dev->pcidev->dev, + c2dev->aeq.q_size * c2dev->aeq.msg_size, + q2_pages, pci_unmap_addr(&c2dev->aeq, mapping)); bail2: - dma_unmap_single(c2dev->ibdev.dma_device, - pci_unmap_addr(&c2dev->rep_vq, mapping), - c2dev->rep_vq.q_size * c2dev->rep_vq.msg_size, - DMA_FROM_DEVICE); - kfree(q1_pages); + dma_free_coherent(&c2dev->pcidev->dev, + c2dev->rep_vq.q_size * c2dev->rep_vq.msg_size, + q1_pages, pci_unmap_addr(&c2dev->rep_vq, mapping)); bail1: c2_free_mqsp_pool(c2dev, c2dev->kern_mqsp_pool); bail0: @@ -619,7 +611,7 @@ int c2_rnic_init(struct c2_dev *c2dev) /* * Called by c2_remove to cleanup the RNIC resources. */ -void c2_rnic_term(struct c2_dev *c2dev) +void __devexit c2_rnic_term(struct c2_dev *c2dev) { /* Close the open adapter instance */ @@ -640,19 +632,17 @@ void c2_rnic_term(struct c2_dev *c2dev) /* Free the verbs request allocator */ vq_term(c2dev); - /* Unmap and free the asynchronus event queue */ - dma_unmap_single(c2dev->ibdev.dma_device, - pci_unmap_addr(&c2dev->aeq, mapping), - c2dev->aeq.q_size * c2dev->aeq.msg_size, - DMA_FROM_DEVICE); - kfree(c2dev->aeq.msg_pool.host); - - /* Unmap and free the verbs reply queue */ - dma_unmap_single(c2dev->ibdev.dma_device, - pci_unmap_addr(&c2dev->rep_vq, mapping), - c2dev->rep_vq.q_size * c2dev->rep_vq.msg_size, - DMA_FROM_DEVICE); - kfree(c2dev->rep_vq.msg_pool.host); + /* Free the asynchronus event queue */ + dma_free_coherent(&c2dev->pcidev->dev, + c2dev->aeq.q_size * c2dev->aeq.msg_size, + c2dev->aeq.msg_pool.host, + pci_unmap_addr(&c2dev->aeq, mapping)); + + /* Free the verbs reply queue */ + dma_free_coherent(&c2dev->pcidev->dev, + c2dev->rep_vq.q_size * c2dev->rep_vq.msg_size, + c2dev->rep_vq.msg_pool.host, + pci_unmap_addr(&c2dev->rep_vq, mapping)); /* Free the MQ shared pointer pool */ c2_free_mqsp_pool(c2dev, c2dev->kern_mqsp_pool); diff --git a/drivers/infiniband/hw/ehca/Kconfig b/drivers/infiniband/hw/ehca/Kconfig index 922389b6439..727b10d8968 100644 --- a/drivers/infiniband/hw/ehca/Kconfig +++ b/drivers/infiniband/hw/ehca/Kconfig @@ -10,6 +10,7 @@ config INFINIBAND_EHCA config INFINIBAND_EHCA_SCALING bool "Scaling support (EXPERIMENTAL)" depends on IBMEBUS && INFINIBAND_EHCA && HOTPLUG_CPU && EXPERIMENTAL + default y ---help--- eHCA scaling support schedules the CQ callbacks to different CPUs. diff --git a/drivers/infiniband/hw/ehca/ehca_av.c b/drivers/infiniband/hw/ehca/ehca_av.c index 3bac197f901..214e2fdddee 100644 --- a/drivers/infiniband/hw/ehca/ehca_av.c +++ b/drivers/infiniband/hw/ehca/ehca_av.c @@ -118,8 +118,7 @@ struct ib_ah *ehca_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr) } memcpy(&av->av.grh.word_1, &gid, sizeof(gid)); } - /* for the time being we use a hard coded PMTU of 2048 Bytes */ - av->av.pmtu = 4; + av->av.pmtu = EHCA_MAX_MTU; /* dgid comes in grh.word_3 */ memcpy(&av->av.grh.word_3, &ah_attr->grh.dgid, @@ -193,7 +192,7 @@ int ehca_modify_ah(struct ib_ah *ah, struct ib_ah_attr *ah_attr) memcpy(&new_ehca_av.grh.word_1, &gid, sizeof(gid)); } - new_ehca_av.pmtu = 4; /* see also comment in create_ah() */ + new_ehca_av.pmtu = EHCA_MAX_MTU; memcpy(&new_ehca_av.grh.word_3, &ah_attr->grh.dgid, sizeof(ah_attr->grh.dgid)); diff --git a/drivers/infiniband/hw/ehca/ehca_hca.c b/drivers/infiniband/hw/ehca/ehca_hca.c index 5eae6ac4842..e1b618c5f68 100644 --- a/drivers/infiniband/hw/ehca/ehca_hca.c +++ b/drivers/infiniband/hw/ehca/ehca_hca.c @@ -40,6 +40,7 @@ */ #include "ehca_tools.h" +#include "ehca_iverbs.h" #include "hcp_if.h" int ehca_query_device(struct ib_device *ibdev, struct ib_device_attr *props) @@ -49,7 +50,7 @@ int ehca_query_device(struct ib_device *ibdev, struct ib_device_attr *props) ib_device); struct hipz_query_hca *rblock; - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + rblock = ehca_alloc_fw_ctrlblock(); if (!rblock) { ehca_err(&shca->ib_device, "Can't allocate rblock memory."); return -ENOMEM; @@ -96,7 +97,7 @@ int ehca_query_device(struct ib_device *ibdev, struct ib_device_attr *props) = min_t(int, rblock->max_total_mcast_qp_attach, INT_MAX); query_device1: - kfree(rblock); + ehca_free_fw_ctrlblock(rblock); return ret; } @@ -109,7 +110,7 @@ int ehca_query_port(struct ib_device *ibdev, ib_device); struct hipz_query_port *rblock; - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + rblock = ehca_alloc_fw_ctrlblock(); if (!rblock) { ehca_err(&shca->ib_device, "Can't allocate rblock memory."); return -ENOMEM; @@ -162,7 +163,7 @@ int ehca_query_port(struct ib_device *ibdev, props->active_speed = 0x1; query_port1: - kfree(rblock); + ehca_free_fw_ctrlblock(rblock); return ret; } @@ -178,7 +179,7 @@ int ehca_query_pkey(struct ib_device *ibdev, u8 port, u16 index, u16 *pkey) return -EINVAL; } - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + rblock = ehca_alloc_fw_ctrlblock(); if (!rblock) { ehca_err(&shca->ib_device, "Can't allocate rblock memory."); return -ENOMEM; @@ -193,7 +194,7 @@ int ehca_query_pkey(struct ib_device *ibdev, u8 port, u16 index, u16 *pkey) memcpy(pkey, &rblock->pkey_entries + index, sizeof(u16)); query_pkey1: - kfree(rblock); + ehca_free_fw_ctrlblock(rblock); return ret; } @@ -211,7 +212,7 @@ int ehca_query_gid(struct ib_device *ibdev, u8 port, return -EINVAL; } - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + rblock = ehca_alloc_fw_ctrlblock(); if (!rblock) { ehca_err(&shca->ib_device, "Can't allocate rblock memory."); return -ENOMEM; @@ -227,7 +228,7 @@ int ehca_query_gid(struct ib_device *ibdev, u8 port, memcpy(&gid->raw[8], &rblock->guid_entries[index], sizeof(u64)); query_gid1: - kfree(rblock); + ehca_free_fw_ctrlblock(rblock); return ret; } diff --git a/drivers/infiniband/hw/ehca/ehca_irq.c b/drivers/infiniband/hw/ehca/ehca_irq.c index 048cc443d1e..c3ea746e904 100644 --- a/drivers/infiniband/hw/ehca/ehca_irq.c +++ b/drivers/infiniband/hw/ehca/ehca_irq.c @@ -45,6 +45,7 @@ #include "ehca_tools.h" #include "hcp_if.h" #include "hipz_fns.h" +#include "ipz_pt_fn.h" #define EQE_COMPLETION_EVENT EHCA_BMASK_IBM(1,1) #define EQE_CQ_QP_NUMBER EHCA_BMASK_IBM(8,31) @@ -137,38 +138,36 @@ int ehca_error_data(struct ehca_shca *shca, void *data, u64 *rblock; unsigned long block_count; - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + rblock = ehca_alloc_fw_ctrlblock(); if (!rblock) { ehca_err(&shca->ib_device, "Cannot allocate rblock memory."); ret = -ENOMEM; goto error_data1; } + /* rblock must be 4K aligned and should be 4K large */ ret = hipz_h_error_data(shca->ipz_hca_handle, resource, rblock, &block_count); - if (ret == H_R_STATE) { + if (ret == H_R_STATE) ehca_err(&shca->ib_device, "No error data is available: %lx.", resource); - } else if (ret == H_SUCCESS) { int length; length = EHCA_BMASK_GET(ERROR_DATA_LENGTH, rblock[0]); - if (length > PAGE_SIZE) - length = PAGE_SIZE; + if (length > EHCA_PAGESIZE) + length = EHCA_PAGESIZE; print_error_data(shca, data, rblock, length); - } - else { + } else ehca_err(&shca->ib_device, "Error data could not be fetched: %lx", resource); - } - kfree(rblock); + ehca_free_fw_ctrlblock(rblock); error_data1: return ret; diff --git a/drivers/infiniband/hw/ehca/ehca_iverbs.h b/drivers/infiniband/hw/ehca/ehca_iverbs.h index 319c39d47f3..3720e3032cc 100644 --- a/drivers/infiniband/hw/ehca/ehca_iverbs.h +++ b/drivers/infiniband/hw/ehca/ehca_iverbs.h @@ -179,4 +179,12 @@ int ehca_mmap_register(u64 physical,void **mapped, int ehca_munmap(unsigned long addr, size_t len); +#ifdef CONFIG_PPC_64K_PAGES +void *ehca_alloc_fw_ctrlblock(void); +void ehca_free_fw_ctrlblock(void *ptr); +#else +#define ehca_alloc_fw_ctrlblock() ((void *) get_zeroed_page(GFP_KERNEL)) +#define ehca_free_fw_ctrlblock(ptr) free_page((unsigned long)(ptr)) +#endif + #endif diff --git a/drivers/infiniband/hw/ehca/ehca_main.c b/drivers/infiniband/hw/ehca/ehca_main.c index 024d511c4b5..3d1c1c53503 100644 --- a/drivers/infiniband/hw/ehca/ehca_main.c +++ b/drivers/infiniband/hw/ehca/ehca_main.c @@ -40,6 +40,9 @@ * POSSIBILITY OF SUCH DAMAGE. */ +#ifdef CONFIG_PPC_64K_PAGES +#include <linux/slab.h> +#endif #include "ehca_classes.h" #include "ehca_iverbs.h" #include "ehca_mrmw.h" @@ -49,7 +52,7 @@ MODULE_LICENSE("Dual BSD/GPL"); MODULE_AUTHOR("Christoph Raisch <raisch@de.ibm.com>"); MODULE_DESCRIPTION("IBM eServer HCA InfiniBand Device Driver"); -MODULE_VERSION("SVNEHCA_0017"); +MODULE_VERSION("SVNEHCA_0019"); int ehca_open_aqp1 = 0; int ehca_debug_level = 0; @@ -94,11 +97,31 @@ spinlock_t ehca_cq_idr_lock; DEFINE_IDR(ehca_qp_idr); DEFINE_IDR(ehca_cq_idr); + static struct list_head shca_list; /* list of all registered ehcas */ static spinlock_t shca_list_lock; static struct timer_list poll_eqs_timer; +#ifdef CONFIG_PPC_64K_PAGES +static struct kmem_cache *ctblk_cache = NULL; + +void *ehca_alloc_fw_ctrlblock(void) +{ + void *ret = kmem_cache_zalloc(ctblk_cache, SLAB_KERNEL); + if (!ret) + ehca_gen_err("Out of memory for ctblk"); + return ret; +} + +void ehca_free_fw_ctrlblock(void *ptr) +{ + if (ptr) + kmem_cache_free(ctblk_cache, ptr); + +} +#endif + static int ehca_create_slab_caches(void) { int ret; @@ -133,6 +156,17 @@ static int ehca_create_slab_caches(void) goto create_slab_caches5; } +#ifdef CONFIG_PPC_64K_PAGES + ctblk_cache = kmem_cache_create("ehca_cache_ctblk", + EHCA_PAGESIZE, H_CB_ALIGNMENT, + SLAB_HWCACHE_ALIGN, + NULL, NULL); + if (!ctblk_cache) { + ehca_gen_err("Cannot create ctblk SLAB cache."); + ehca_cleanup_mrmw_cache(); + goto create_slab_caches5; + } +#endif return 0; create_slab_caches5: @@ -157,6 +191,10 @@ static void ehca_destroy_slab_caches(void) ehca_cleanup_qp_cache(); ehca_cleanup_cq_cache(); ehca_cleanup_pd_cache(); +#ifdef CONFIG_PPC_64K_PAGES + if (ctblk_cache) + kmem_cache_destroy(ctblk_cache); +#endif } #define EHCA_HCAAVER EHCA_BMASK_IBM(32,39) @@ -168,7 +206,7 @@ int ehca_sense_attributes(struct ehca_shca *shca) u64 h_ret; struct hipz_query_hca *rblock; - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + rblock = ehca_alloc_fw_ctrlblock(); if (!rblock) { ehca_gen_err("Cannot allocate rblock memory."); return -ENOMEM; @@ -211,7 +249,7 @@ int ehca_sense_attributes(struct ehca_shca *shca) shca->sport[1].rate = IB_RATE_30_GBPS; num_ports1: - kfree(rblock); + ehca_free_fw_ctrlblock(rblock); return ret; } @@ -220,7 +258,7 @@ static int init_node_guid(struct ehca_shca *shca) int ret = 0; struct hipz_query_hca *rblock; - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + rblock = ehca_alloc_fw_ctrlblock(); if (!rblock) { ehca_err(&shca->ib_device, "Can't allocate rblock memory."); return -ENOMEM; @@ -235,7 +273,7 @@ static int init_node_guid(struct ehca_shca *shca) memcpy(&shca->ib_device.node_guid, &rblock->node_guid, sizeof(u64)); init_node_guid1: - kfree(rblock); + ehca_free_fw_ctrlblock(rblock); return ret; } @@ -431,7 +469,7 @@ static ssize_t ehca_show_##name(struct device *dev, \ \ shca = dev->driver_data; \ \ - rblock = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); \ + rblock = ehca_alloc_fw_ctrlblock(); \ if (!rblock) { \ dev_err(dev, "Can't allocate rblock memory."); \ return 0; \ @@ -439,12 +477,12 @@ static ssize_t ehca_show_##name(struct device *dev, \ \ if (hipz_h_query_hca(shca->ipz_hca_handle, rblock) != H_SUCCESS) { \ dev_err(dev, "Can't query device properties"); \ - kfree(rblock); \ + ehca_free_fw_ctrlblock(rblock); \ return 0; \ } \ \ data = rblock->name; \ - kfree(rblock); \ + ehca_free_fw_ctrlblock(rblock); \ \ if ((strcmp(#name, "num_ports") == 0) && (ehca_nr_ports == 1)) \ return snprintf(buf, 256, "1\n"); \ @@ -752,7 +790,7 @@ int __init ehca_module_init(void) int ret; printk(KERN_INFO "eHCA Infiniband Device Driver " - "(Rel.: SVNEHCA_0017)\n"); + "(Rel.: SVNEHCA_0019)\n"); idr_init(&ehca_qp_idr); idr_init(&ehca_cq_idr); spin_lock_init(&ehca_qp_idr_lock); diff --git a/drivers/infiniband/hw/ehca/ehca_mrmw.c b/drivers/infiniband/hw/ehca/ehca_mrmw.c index 5ca65441e1d..abce676c0ae 100644 --- a/drivers/infiniband/hw/ehca/ehca_mrmw.c +++ b/drivers/infiniband/hw/ehca/ehca_mrmw.c @@ -1013,7 +1013,7 @@ int ehca_reg_mr_rpages(struct ehca_shca *shca, u32 i; u64 *kpage; - kpage = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + kpage = ehca_alloc_fw_ctrlblock(); if (!kpage) { ehca_err(&shca->ib_device, "kpage alloc failed"); ret = -ENOMEM; @@ -1092,7 +1092,7 @@ int ehca_reg_mr_rpages(struct ehca_shca *shca, ehca_reg_mr_rpages_exit1: - kfree(kpage); + ehca_free_fw_ctrlblock(kpage); ehca_reg_mr_rpages_exit0: if (ret) ehca_err(&shca->ib_device, "ret=%x shca=%p e_mr=%p pginfo=%p " @@ -1124,7 +1124,7 @@ inline int ehca_rereg_mr_rereg1(struct ehca_shca *shca, ehca_mrmw_map_acl(acl, &hipz_acl); ehca_mrmw_set_pgsize_hipz_acl(&hipz_acl); - kpage = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); + kpage = ehca_alloc_fw_ctrlblock(); if (!kpage) { ehca_err(&shca->ib_device, "kpage alloc failed"); ret = -ENOMEM; @@ -1181,7 +1181,7 @@ inline int ehca_rereg_mr_rereg1(struct ehca_shca *shca, } ehca_rereg_mr_rereg1_exit1: - kfree(kpage); + ehca_free_fw_ctrlblock(kpage); ehca_rereg_mr_rereg1_exit0: if ( ret && (ret != -EAGAIN) ) ehca_err(&shca->ib_device, "ret=%x lkey=%x rkey=%x " diff --git a/drivers/infiniband/hw/ehca/ehca_qp.c b/drivers/infiniband/hw/ehca/ehca_qp.c index 4394123cdbd..8682aa50c70 100644 --- a/drivers/infiniband/hw/ehca/ehca_qp.c +++ b/drivers/infiniband/hw/ehca/ehca_qp.c @@ -732,8 +732,7 @@ static int prepare_sqe_rts(struct ehca_qp *my_qp, struct ehca_shca *shca, u64 h_ret; struct ipz_queue *squeue; void *bad_send_wqe_p, *bad_send_wqe_v; - void *squeue_start_p, *squeue_end_p; - void *squeue_start_v, *squeue_end_v; + u64 q_ofs; struct ehca_wqe *wqe; int qp_num = my_qp->ib_qp.qp_num; @@ -755,26 +754,23 @@ static int prepare_sqe_rts(struct ehca_qp *my_qp, struct ehca_shca *shca, if (ehca_debug_level) ehca_dmp(bad_send_wqe_v, 32, "qp_num=%x bad_wqe", qp_num); squeue = &my_qp->ipz_squeue; - squeue_start_p = (void*)virt_to_abs(ipz_qeit_calc(squeue, 0L)); - squeue_end_p = squeue_start_p+squeue->queue_length; - squeue_start_v = abs_to_virt((u64)squeue_start_p); - squeue_end_v = abs_to_virt((u64)squeue_end_p); - ehca_dbg(&shca->ib_device, "qp_num=%x squeue_start_v=%p squeue_end_v=%p", - qp_num, squeue_start_v, squeue_end_v); + if (ipz_queue_abs_to_offset(squeue, (u64)bad_send_wqe_p, &q_ofs)) { + ehca_err(&shca->ib_device, "failed to get wqe offset qp_num=%x" + " bad_send_wqe_p=%p", qp_num, bad_send_wqe_p); + return -EFAULT; + } /* loop sets wqe's purge bit */ - wqe = (struct ehca_wqe*)bad_send_wqe_v; + wqe = (struct ehca_wqe*)ipz_qeit_calc(squeue, q_ofs); *bad_wqe_cnt = 0; while (wqe->optype != 0xff && wqe->wqef != 0xff) { if (ehca_debug_level) ehca_dmp(wqe, 32, "qp_num=%x wqe", qp_num); wqe->nr_of_data_seg = 0; /* suppress data access */ wqe->wqef = WQEF_PURGE; /* WQE to be purged */ - wqe = (struct ehca_wqe*)((u8*)wqe+squeue->qe_size); + q_ofs = ipz_queue_advance_offset(squeue, q_ofs); + wqe = (struct ehca_wqe*)ipz_qeit_calc(squeue, q_ofs); *bad_wqe_cnt = (*bad_wqe_cnt)+1; - if ((void*)wqe >= squeue_end_v) { - wqe = squeue_start_v; - } } /* * bad wqe will be reprocessed and ignored when pol_cq() is called, @@ -811,8 +807,8 @@ static int internal_modify_qp(struct ib_qp *ibqp, unsigned long spl_flags = 0; /* do query_qp to obtain current attr values */ - mqpcb = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL); - if (mqpcb == NULL) { + mqpcb = ehca_alloc_fw_ctrlblock(); + if (!mqpcb) { ehca_err(ibqp->device, "Could not get zeroed page for mqpcb " "ehca_qp=%p qp_num=%x ", my_qp, ibqp->qp_num); return -ENOMEM; @@ -1225,7 +1221,7 @@ modify_qp_exit2: } modify_qp_exit1: - kfree(mqpcb); + ehca_free_fw_ctrlblock(mqpcb); return ret; } @@ -1277,7 +1273,7 @@ int ehca_query_qp(struct ib_qp *qp, return -EINVAL; } - qpcb = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL ); + qpcb = ehca_alloc_fw_ctrlblock(); if (!qpcb) { ehca_err(qp->device,"Out of memory for qpcb " "ehca_qp=%p qp_num=%x", my_qp, qp->qp_num); @@ -1401,7 +1397,7 @@ int ehca_query_qp(struct ib_qp *qp, ehca_dmp(qpcb, 4*70, "qp_num=%x", qp->qp_num); query_qp_exit1: - kfree(qpcb); + ehca_free_fw_ctrlblock(qpcb); return ret; } diff --git a/drivers/infiniband/hw/ehca/ehca_tools.h b/drivers/infiniband/hw/ehca/ehca_tools.h index 809da3ef706..973c4b59154 100644 --- a/drivers/infiniband/hw/ehca/ehca_tools.h +++ b/drivers/infiniband/hw/ehca/ehca_tools.h @@ -63,6 +63,7 @@ #include <asm/ibmebus.h> #include <asm/io.h> #include <asm/pgtable.h> +#include <asm/hvcall.h> extern int ehca_debug_level; diff --git a/drivers/infiniband/hw/ehca/hipz_hw.h b/drivers/infiniband/hw/ehca/hipz_hw.h index 3fc92b031c5..fad91368dc5 100644 --- a/drivers/infiniband/hw/ehca/hipz_hw.h +++ b/drivers/infiniband/hw/ehca/hipz_hw.h @@ -45,6 +45,8 @@ #include "ehca_tools.h" +#define EHCA_MAX_MTU 4 + /* QP Table Entry Memory Map */ struct hipz_qptemm { u64 qpx_hcr; diff --git a/drivers/infiniband/hw/ehca/ipz_pt_fn.c b/drivers/infiniband/hw/ehca/ipz_pt_fn.c index e028ff1588c..bf7a40088f6 100644 --- a/drivers/infiniband/hw/ehca/ipz_pt_fn.c +++ b/drivers/infiniband/hw/ehca/ipz_pt_fn.c @@ -70,6 +70,19 @@ void *ipz_qeit_eq_get_inc(struct ipz_queue *queue) return ret; } +int ipz_queue_abs_to_offset(struct ipz_queue *queue, u64 addr, u64 *q_offset) +{ + int i; + for (i = 0; i < queue->queue_length / queue->pagesize; i++) { + u64 page = (u64)virt_to_abs(queue->queue_pages[i]); + if (addr >= page && addr < page + queue->pagesize) { + *q_offset = addr - page + i * queue->pagesize; + return 0; + } + } + return -EINVAL; +} + int ipz_queue_ctor(struct ipz_queue *queue, const u32 nr_of_pages, const u32 pagesize, const u32 qe_size, const u32 nr_of_sg) diff --git a/drivers/infiniband/hw/ehca/ipz_pt_fn.h b/drivers/infiniband/hw/ehca/ipz_pt_fn.h index 2f13509d525..dc3bda2634b 100644 --- a/drivers/infiniband/hw/ehca/ipz_pt_fn.h +++ b/drivers/infiniband/hw/ehca/ipz_pt_fn.h @@ -150,6 +150,21 @@ static inline void *ipz_qeit_reset(struct ipz_queue *queue) return ipz_qeit_get(queue); } +/* + * return the q_offset corresponding to an absolute address + */ +int ipz_queue_abs_to_offset(struct ipz_queue *queue, u64 addr, u64 *q_offset); + +/* + * return the next queue offset. don't modify the queue. + */ +static inline u64 ipz_queue_advance_offset(struct ipz_queue *queue, u64 offset) +{ + offset += queue->qe_size; + if (offset >= queue->queue_length) offset = 0; + return offset; +} + /* struct generic page table */ struct ipz_pt { u64 entries[EHCA_PT_ENTRIES]; diff --git a/drivers/infiniband/hw/ipath/Kconfig b/drivers/infiniband/hw/ipath/Kconfig index 574a678e7fd..90c14543677 100644 --- a/drivers/infiniband/hw/ipath/Kconfig +++ b/drivers/infiniband/hw/ipath/Kconfig @@ -1,6 +1,6 @@ config INFINIBAND_IPATH tristate "QLogic InfiniPath Driver" - depends on PCI_MSI && 64BIT && INFINIBAND + depends on (PCI_MSI || HT_IRQ) && 64BIT && INFINIBAND && NET ---help--- This is a driver for QLogic InfiniPath host channel adapters, including InfiniBand verbs support. This driver allows these diff --git a/drivers/infiniband/hw/ipath/Makefile b/drivers/infiniband/hw/ipath/Makefile index 5e29cb0095e..7dc10551cf1 100644 --- a/drivers/infiniband/hw/ipath/Makefile +++ b/drivers/infiniband/hw/ipath/Makefile @@ -10,8 +10,6 @@ ib_ipath-y := \ ipath_eeprom.o \ ipath_file_ops.o \ ipath_fs.o \ - ipath_iba6110.o \ - ipath_iba6120.o \ ipath_init_chip.o \ ipath_intr.o \ ipath_keys.o \ @@ -31,5 +29,8 @@ ib_ipath-y := \ ipath_verbs_mcast.o \ ipath_verbs.o +ib_ipath-$(CONFIG_HT_IRQ) += ipath_iba6110.o +ib_ipath-$(CONFIG_PCI_MSI) += ipath_iba6120.o + ib_ipath-$(CONFIG_X86_64) += ipath_wc_x86_64.o ib_ipath-$(CONFIG_PPC64) += ipath_wc_ppc64.o diff --git a/drivers/infiniband/hw/ipath/ipath_diag.c b/drivers/infiniband/hw/ipath/ipath_diag.c index 29958b6e021..28c087b824c 100644 --- a/drivers/infiniband/hw/ipath/ipath_diag.c +++ b/drivers/infiniband/hw/ipath/ipath_diag.c @@ -67,19 +67,54 @@ static struct file_operations diag_file_ops = { .release = ipath_diag_release }; +static ssize_t ipath_diagpkt_write(struct file *fp, + const char __user *data, + size_t count, loff_t *off); + +static struct file_operations diagpkt_file_ops = { + .owner = THIS_MODULE, + .write = ipath_diagpkt_write, +}; + +static atomic_t diagpkt_count = ATOMIC_INIT(0); +static struct cdev *diagpkt_cdev; +static struct class_device *diagpkt_class_dev; + int ipath_diag_add(struct ipath_devdata *dd) { char name[16]; + int ret = 0; + + if (atomic_inc_return(&diagpkt_count) == 1) { + ret = ipath_cdev_init(IPATH_DIAGPKT_MINOR, + "ipath_diagpkt", &diagpkt_file_ops, + &diagpkt_cdev, &diagpkt_class_dev); + + if (ret) { + ipath_dev_err(dd, "Couldn't create ipath_diagpkt " + "device: %d", ret); + goto done; + } + } snprintf(name, sizeof(name), "ipath_diag%d", dd->ipath_unit); - return ipath_cdev_init(IPATH_DIAG_MINOR_BASE + dd->ipath_unit, name, - &diag_file_ops, &dd->diag_cdev, - &dd->diag_class_dev); + ret = ipath_cdev_init(IPATH_DIAG_MINOR_BASE + dd->ipath_unit, name, + &diag_file_ops, &dd->diag_cdev, + &dd->diag_class_dev); + if (ret) + ipath_dev_err(dd, "Couldn't create %s device: %d", + name, ret); + +done: + return ret; } void ipath_diag_remove(struct ipath_devdata *dd) { + if (atomic_dec_and_test(&diagpkt_count)) + ipath_cdev_cleanup(&diagpkt_cdev, &diagpkt_class_dev); + ipath_cdev_cleanup(&dd->diag_cdev, &dd->diag_class_dev); } @@ -275,30 +310,6 @@ bail: return ret; } -static ssize_t ipath_diagpkt_write(struct file *fp, - const char __user *data, - size_t count, loff_t *off); - -static struct file_operations diagpkt_file_ops = { - .owner = THIS_MODULE, - .write = ipath_diagpkt_write, -}; - -static struct cdev *diagpkt_cdev; -static struct class_device *diagpkt_class_dev; - -int __init ipath_diagpkt_add(void) -{ - return ipath_cdev_init(IPATH_DIAGPKT_MINOR, - "ipath_diagpkt", &diagpkt_file_ops, - &diagpkt_cdev, &diagpkt_class_dev); -} - -void __exit ipath_diagpkt_remove(void) -{ - ipath_cdev_cleanup(&diagpkt_cdev, &diagpkt_class_dev); -} - /** * ipath_diagpkt_write - write an IB packet * @fp: the diag data device file pointer diff --git a/drivers/infiniband/hw/ipath/ipath_driver.c b/drivers/infiniband/hw/ipath/ipath_driver.c index 12cefa658f3..1aeddb48e35 100644 --- a/drivers/infiniband/hw/ipath/ipath_driver.c +++ b/drivers/infiniband/hw/ipath/ipath_driver.c @@ -304,7 +304,7 @@ static int __devinit ipath_init_one(struct pci_dev *pdev, } addr = pci_resource_start(pdev, 0); len = pci_resource_len(pdev, 0); - ipath_cdbg(VERBOSE, "regbase (0) %llx len %d irq %x, vend %x/%x " + ipath_cdbg(VERBOSE, "regbase (0) %llx len %d pdev->irq %d, vend %x/%x " "driver_data %lx\n", addr, len, pdev->irq, ent->vendor, ent->device, ent->driver_data); @@ -390,12 +390,16 @@ static int __devinit ipath_init_one(struct pci_dev *pdev, /* setup the chip-specific functions, as early as possible. */ switch (ent->device) { +#ifdef CONFIG_HT_IRQ case PCI_DEVICE_ID_INFINIPATH_HT: ipath_init_iba6110_funcs(dd); break; +#endif +#ifdef CONFIG_PCI_MSI case PCI_DEVICE_ID_INFINIPATH_PE800: ipath_init_iba6120_funcs(dd); break; +#endif default: ipath_dev_err(dd, "Found unknown QLogic deviceid 0x%x, " "failing\n", ent->device); @@ -467,15 +471,15 @@ static int __devinit ipath_init_one(struct pci_dev *pdev, * check 0 irq after we return from chip-specific bus setup, since * that can affect this due to setup */ - if (!pdev->irq) + if (!dd->ipath_irq) ipath_dev_err(dd, "irq is 0, BIOS error? Interrupts won't " "work\n"); else { - ret = request_irq(pdev->irq, ipath_intr, IRQF_SHARED, + ret = request_irq(dd->ipath_irq, ipath_intr, IRQF_SHARED, IPATH_DRV_NAME, dd); if (ret) { ipath_dev_err(dd, "Couldn't setup irq handler, " - "irq=%u: %d\n", pdev->irq, ret); + "irq=%d: %d\n", dd->ipath_irq, ret); goto bail_iounmap; } } @@ -637,11 +641,10 @@ static void __devexit ipath_remove_one(struct pci_dev *pdev) * free up port 0 (kernel) rcvhdr, egr bufs, and eventually tid bufs * for all versions of the driver, if they were allocated */ - if (pdev->irq) { - ipath_cdbg(VERBOSE, - "unit %u free_irq of irq %x\n", - dd->ipath_unit, pdev->irq); - free_irq(pdev->irq, dd); + if (dd->ipath_irq) { + ipath_cdbg(VERBOSE, "unit %u free irq %d\n", + dd->ipath_unit, dd->ipath_irq); + dd->ipath_f_free_irq(dd); } else ipath_dbg("irq is 0, not doing free_irq " "for unit %u\n", dd->ipath_unit); @@ -2005,18 +2008,8 @@ static int __init infinipath_init(void) goto bail_group; } - ret = ipath_diagpkt_add(); - if (ret < 0) { - printk(KERN_ERR IPATH_DRV_NAME ": Unable to create " - "diag data device: error %d\n", -ret); - goto bail_ipathfs; - } - goto bail; -bail_ipathfs: - ipath_exit_ipathfs(); - bail_group: ipath_driver_remove_group(&ipath_driver.driver); diff --git a/drivers/infiniband/hw/ipath/ipath_iba6110.c b/drivers/infiniband/hw/ipath/ipath_iba6110.c index 9e4e8d4c6e2..e57c7a351cb 100644 --- a/drivers/infiniband/hw/ipath/ipath_iba6110.c +++ b/drivers/infiniband/hw/ipath/ipath_iba6110.c @@ -38,6 +38,7 @@ #include <linux/pci.h> #include <linux/delay.h> +#include <linux/htirq.h> #include "ipath_kernel.h" #include "ipath_registers.h" @@ -913,49 +914,40 @@ static void slave_or_pri_blk(struct ipath_devdata *dd, struct pci_dev *pdev, } } -static int set_int_handler(struct ipath_devdata *dd, struct pci_dev *pdev, - int pos) +static int ipath_ht_intconfig(struct ipath_devdata *dd) { - u32 int_handler_addr_lower; - u32 int_handler_addr_upper; - u64 ihandler; - u32 intvec; + int ret; - /* use indirection register to get the intr handler */ - pci_write_config_byte(pdev, pos + HT_INTR_REG_INDEX, 0x10); - pci_read_config_dword(pdev, pos + 4, &int_handler_addr_lower); - pci_write_config_byte(pdev, pos + HT_INTR_REG_INDEX, 0x11); - pci_read_config_dword(pdev, pos + 4, &int_handler_addr_upper); + if (dd->ipath_intconfig) { + ipath_write_kreg(dd, dd->ipath_kregs->kr_interruptconfig, + dd->ipath_intconfig); /* interrupt address */ + ret = 0; + } else { + ipath_dev_err(dd, "No interrupts enabled, couldn't setup " + "interrupt address\n"); + ret = -EINVAL; + } - ihandler = (u64) int_handler_addr_lower | - ((u64) int_handler_addr_upper << 32); + return ret; +} + +static void ipath_ht_irq_update(struct pci_dev *dev, int irq, + struct ht_irq_msg *msg) +{ + struct ipath_devdata *dd = pci_get_drvdata(dev); + u64 prev_intconfig = dd->ipath_intconfig; + + dd->ipath_intconfig = msg->address_lo; + dd->ipath_intconfig |= ((u64) msg->address_hi) << 32; /* - * kernels with CONFIG_PCI_MSI set the vector in the irq field of - * struct pci_device, so we use that to program the internal - * interrupt register (not config space) with that value. The BIOS - * must still have done the basic MSI setup. - */ - intvec = pdev->irq; - /* - * clear any vector bits there; normally not set but we'll overload - * this for some debug purposes (setting the HTC debug register - * value from software, rather than GPIOs), so it might be set on a - * driver reload. + * If the previous value of dd->ipath_intconfig is zero, we're + * getting configured for the first time, and must not program the + * intconfig register here (it will be programmed later, when the + * hardware is ready). Otherwise, we should. */ - ihandler &= ~0xff0000; - /* x86 vector goes in intrinfo[23:16] */ - ihandler |= intvec << 16; - ipath_cdbg(VERBOSE, "ihandler lower %x, upper %x, intvec %x, " - "interruptconfig %llx\n", int_handler_addr_lower, - int_handler_addr_upper, intvec, - (unsigned long long) ihandler); - - /* can't program yet, so save for interrupt setup */ - dd->ipath_intconfig = ihandler; - /* keep going, so we find link control stuff also */ - - return ihandler != 0; + if (prev_intconfig) + ipath_ht_intconfig(dd); } /** @@ -971,12 +963,19 @@ static int set_int_handler(struct ipath_devdata *dd, struct pci_dev *pdev, static int ipath_setup_ht_config(struct ipath_devdata *dd, struct pci_dev *pdev) { - int pos, ret = 0; - int ihandler = 0; + int pos, ret; + + ret = __ht_create_irq(pdev, 0, ipath_ht_irq_update); + if (ret < 0) { + ipath_dev_err(dd, "Couldn't create interrupt handler: " + "err %d\n", ret); + goto bail; + } + dd->ipath_irq = ret; + ret = 0; /* - * Read the capability info to find the interrupt info, and also - * handle clearing CRC errors in linkctrl register if necessary. We + * Handle clearing CRC errors in linkctrl register if necessary. We * do this early, before we ever enable errors or hardware errors, * mostly to avoid causing the chip to enter freeze mode. */ @@ -1000,17 +999,9 @@ static int ipath_setup_ht_config(struct ipath_devdata *dd, } if (!(cap_type & 0xE0)) slave_or_pri_blk(dd, pdev, pos, cap_type); - else if (cap_type == HT_INTR_DISC_CONFIG) - ihandler = set_int_handler(dd, pdev, pos); } while ((pos = pci_find_next_capability(pdev, pos, PCI_CAP_ID_HT))); - if (!ihandler) { - ipath_dev_err(dd, "Couldn't find interrupt handler in " - "config space\n"); - ret = -ENODEV; - } - bail: return ret; } @@ -1360,25 +1351,6 @@ static void ipath_ht_quiet_serdes(struct ipath_devdata *dd) ipath_write_kreg(dd, dd->ipath_kregs->kr_serdesconfig0, val); } -static int ipath_ht_intconfig(struct ipath_devdata *dd) -{ - int ret; - - if (!dd->ipath_intconfig) { - ipath_dev_err(dd, "No interrupts enabled, couldn't setup " - "interrupt address\n"); - ret = 1; - goto bail; - } - - ipath_write_kreg(dd, dd->ipath_kregs->kr_interruptconfig, - dd->ipath_intconfig); /* interrupt address */ - ret = 0; - -bail: - return ret; -} - /** * ipath_pe_put_tid - write a TID in chip * @dd: the infinipath device @@ -1575,6 +1547,14 @@ static int ipath_ht_get_base_info(struct ipath_portdata *pd, void *kbase) return 0; } +static void ipath_ht_free_irq(struct ipath_devdata *dd) +{ + free_irq(dd->ipath_irq, dd); + ht_destroy_irq(dd->ipath_irq); + dd->ipath_irq = 0; + dd->ipath_intconfig = 0; +} + /** * ipath_init_iba6110_funcs - set up the chip-specific function pointers * @dd: the infinipath device @@ -1598,6 +1578,7 @@ void ipath_init_iba6110_funcs(struct ipath_devdata *dd) dd->ipath_f_cleanup = ipath_setup_ht_cleanup; dd->ipath_f_setextled = ipath_setup_ht_setextled; dd->ipath_f_get_base_info = ipath_ht_get_base_info; + dd->ipath_f_free_irq = ipath_ht_free_irq; /* * initialize chip-specific variables diff --git a/drivers/infiniband/hw/ipath/ipath_iba6120.c b/drivers/infiniband/hw/ipath/ipath_iba6120.c index a72ab9de386..6af89683f71 100644 --- a/drivers/infiniband/hw/ipath/ipath_iba6120.c +++ b/drivers/infiniband/hw/ipath/ipath_iba6120.c @@ -851,6 +851,7 @@ static int ipath_setup_pe_config(struct ipath_devdata *dd, int pos, ret; dd->ipath_msi_lo = 0; /* used as a flag during reset processing */ + dd->ipath_irq = pdev->irq; ret = pci_enable_msi(dd->pcidev); if (ret) ipath_dev_err(dd, "pci_enable_msi failed: %d, " @@ -1323,6 +1324,12 @@ done: return 0; } +static void ipath_pe_free_irq(struct ipath_devdata *dd) +{ + free_irq(dd->ipath_irq, dd); + dd->ipath_irq = 0; +} + /** * ipath_init_iba6120_funcs - set up the chip-specific function pointers * @dd: the infinipath device @@ -1349,6 +1356,7 @@ void ipath_init_iba6120_funcs(struct ipath_devdata *dd) dd->ipath_f_cleanup = ipath_setup_pe_cleanup; dd->ipath_f_setextled = ipath_setup_pe_setextled; dd->ipath_f_get_base_info = ipath_pe_get_base_info; + dd->ipath_f_free_irq = ipath_pe_free_irq; /* initialize chip-specific variables */ dd->ipath_f_tidtemplate = ipath_pe_tidtemplate; diff --git a/drivers/infiniband/hw/ipath/ipath_intr.c b/drivers/infiniband/hw/ipath/ipath_intr.c index d9079ee1203..5652a550d44 100644 --- a/drivers/infiniband/hw/ipath/ipath_intr.c +++ b/drivers/infiniband/hw/ipath/ipath_intr.c @@ -710,14 +710,14 @@ static void ipath_bad_intr(struct ipath_devdata *dd, u32 * unexpectp) * linuxbios development work, and it may happen in * the future again. */ - if (dd->pcidev && dd->pcidev->irq) { + if (dd->pcidev && dd->ipath_irq) { ipath_dev_err(dd, "Now %u unexpected " "interrupts, unregistering " "interrupt handler\n", *unexpectp); - ipath_dbg("free_irq of irq %x\n", - dd->pcidev->irq); - free_irq(dd->pcidev->irq, dd); + ipath_dbg("free_irq of irq %d\n", + dd->ipath_irq); + dd->ipath_f_free_irq(dd); } } if (ipath_read_kreg32(dd, dd->ipath_kregs->kr_intmask)) { @@ -753,7 +753,7 @@ static void ipath_bad_regread(struct ipath_devdata *dd) if (allbits == 2) { ipath_dev_err(dd, "Still bad interrupt status, " "unregistering interrupt\n"); - free_irq(dd->pcidev->irq, dd); + dd->ipath_f_free_irq(dd); } else if (allbits > 2) { if ((allbits % 10000) == 0) printk("."); diff --git a/drivers/infiniband/hw/ipath/ipath_kernel.h b/drivers/infiniband/hw/ipath/ipath_kernel.h index 7c436697d0e..986b2125b8f 100644 --- a/drivers/infiniband/hw/ipath/ipath_kernel.h +++ b/drivers/infiniband/hw/ipath/ipath_kernel.h @@ -213,6 +213,8 @@ struct ipath_devdata { void (*ipath_f_setextled)(struct ipath_devdata *, u64, u64); /* fill out chip-specific fields */ int (*ipath_f_get_base_info)(struct ipath_portdata *, void *); + /* free irq */ + void (*ipath_f_free_irq)(struct ipath_devdata *); struct ipath_ibdev *verbs_dev; struct timer_list verbs_timer; /* total dwords sent (summed from counter) */ @@ -328,6 +330,8 @@ struct ipath_devdata { /* so we can rewrite it after a chip reset */ u32 ipath_pcibar1; + /* interrupt number */ + int ipath_irq; /* HT/PCI Vendor ID (here for NodeInfo) */ u16 ipath_vendorid; /* HT/PCI Device ID (here for NodeInfo) */ @@ -869,9 +873,6 @@ int ipath_device_create_group(struct device *, struct ipath_devdata *); void ipath_device_remove_group(struct device *, struct ipath_devdata *); int ipath_expose_reset(struct device *); -int ipath_diagpkt_add(void); -void ipath_diagpkt_remove(void); - int ipath_init_ipathfs(void); void ipath_exit_ipathfs(void); int ipathfs_add_device(struct ipath_devdata *); diff --git a/drivers/infiniband/hw/ipath/ipath_verbs.c b/drivers/infiniband/hw/ipath/ipath_verbs.c index a5456108dba..acdee33ee1f 100644 --- a/drivers/infiniband/hw/ipath/ipath_verbs.c +++ b/drivers/infiniband/hw/ipath/ipath_verbs.c @@ -1487,7 +1487,7 @@ int ipath_register_ib_device(struct ipath_devdata *dd) idev->pma_counter_select[1] = IB_PMA_PORT_RCV_DATA; idev->pma_counter_select[2] = IB_PMA_PORT_XMIT_PKTS; idev->pma_counter_select[3] = IB_PMA_PORT_RCV_PKTS; - idev->pma_counter_select[5] = IB_PMA_PORT_XMIT_WAIT; + idev->pma_counter_select[4] = IB_PMA_PORT_XMIT_WAIT; idev->link_width_enabled = 3; /* 1x or 4x */ /* Snapshot current HW counters to "clear" them. */ diff --git a/drivers/infiniband/hw/mthca/mthca_av.c b/drivers/infiniband/hw/mthca/mthca_av.c index 69599455aca..57cdc1bc5f5 100644 --- a/drivers/infiniband/hw/mthca/mthca_av.c +++ b/drivers/infiniband/hw/mthca/mthca_av.c @@ -33,7 +33,6 @@ * $Id: mthca_av.c 1349 2004-12-16 21:09:43Z roland $ */ -#include <linux/init.h> #include <linux/string.h> #include <linux/slab.h> @@ -323,7 +322,7 @@ int mthca_ah_query(struct ib_ah *ibah, struct ib_ah_attr *attr) return 0; } -int __devinit mthca_init_av_table(struct mthca_dev *dev) +int mthca_init_av_table(struct mthca_dev *dev) { int err; diff --git a/drivers/infiniband/hw/mthca/mthca_cmd.c b/drivers/infiniband/hw/mthca/mthca_cmd.c index 99a94d71093..768df7265b8 100644 --- a/drivers/infiniband/hw/mthca/mthca_cmd.c +++ b/drivers/infiniband/hw/mthca/mthca_cmd.c @@ -1820,11 +1820,11 @@ int mthca_MAD_IFC(struct mthca_dev *dev, int ignore_mkey, int ignore_bkey, #define MAD_IFC_BOX_SIZE 0x400 #define MAD_IFC_MY_QPN_OFFSET 0x100 -#define MAD_IFC_RQPN_OFFSET 0x104 -#define MAD_IFC_SL_OFFSET 0x108 -#define MAD_IFC_G_PATH_OFFSET 0x109 -#define MAD_IFC_RLID_OFFSET 0x10a -#define MAD_IFC_PKEY_OFFSET 0x10e +#define MAD_IFC_RQPN_OFFSET 0x108 +#define MAD_IFC_SL_OFFSET 0x10c +#define MAD_IFC_G_PATH_OFFSET 0x10d +#define MAD_IFC_RLID_OFFSET 0x10e +#define MAD_IFC_PKEY_OFFSET 0x112 #define MAD_IFC_GRH_OFFSET 0x140 inmailbox = mthca_alloc_mailbox(dev, GFP_KERNEL); @@ -1862,7 +1862,7 @@ int mthca_MAD_IFC(struct mthca_dev *dev, int ignore_mkey, int ignore_bkey, val = in_wc->dlid_path_bits | (in_wc->wc_flags & IB_WC_GRH ? 0x80 : 0); - MTHCA_PUT(inbox, val, MAD_IFC_GRH_OFFSET); + MTHCA_PUT(inbox, val, MAD_IFC_G_PATH_OFFSET); MTHCA_PUT(inbox, in_wc->slid, MAD_IFC_RLID_OFFSET); MTHCA_PUT(inbox, in_wc->pkey_index, MAD_IFC_PKEY_OFFSET); @@ -1870,7 +1870,7 @@ int mthca_MAD_IFC(struct mthca_dev *dev, int ignore_mkey, int ignore_bkey, if (in_grh) memcpy(inbox + MAD_IFC_GRH_OFFSET, in_grh, 40); - op_modifier |= 0x10; + op_modifier |= 0x4; in_modifier |= in_wc->slid << 16; } diff --git a/drivers/infiniband/hw/mthca/mthca_cq.c b/drivers/infiniband/hw/mthca/mthca_cq.c index e393681ba7d..283d50b76c3 100644 --- a/drivers/infiniband/hw/mthca/mthca_cq.c +++ b/drivers/infiniband/hw/mthca/mthca_cq.c @@ -36,9 +36,10 @@ * $Id: mthca_cq.c 1369 2004-12-20 16:17:07Z roland $ */ -#include <linux/init.h> #include <linux/hardirq.h> +#include <asm/io.h> + #include <rdma/ib_pack.h> #include "mthca_dev.h" @@ -210,6 +211,11 @@ static inline void update_cons_index(struct mthca_dev *dev, struct mthca_cq *cq, mthca_write64(doorbell, dev->kar + MTHCA_CQ_DOORBELL, MTHCA_GET_DOORBELL_LOCK(&dev->doorbell_lock)); + /* + * Make sure doorbells don't leak out of CQ spinlock + * and reach the HCA out of order: + */ + mmiowb(); } } @@ -963,7 +969,7 @@ void mthca_free_cq(struct mthca_dev *dev, mthca_free_mailbox(dev, mailbox); } -int __devinit mthca_init_cq_table(struct mthca_dev *dev) +int mthca_init_cq_table(struct mthca_dev *dev) { int err; diff --git a/drivers/infiniband/hw/mthca/mthca_eq.c b/drivers/infiniband/hw/mthca/mthca_eq.c index e284e0613a9..8ec9fa1ff9e 100644 --- a/drivers/infiniband/hw/mthca/mthca_eq.c +++ b/drivers/infiniband/hw/mthca/mthca_eq.c @@ -33,7 +33,6 @@ * $Id: mthca_eq.c 1382 2004-12-24 02:21:02Z roland $ */ -#include <linux/init.h> #include <linux/errno.h> #include <linux/interrupt.h> #include <linux/pci.h> @@ -479,10 +478,10 @@ static irqreturn_t mthca_arbel_msi_x_interrupt(int irq, void *eq_ptr) return IRQ_HANDLED; } -static int __devinit mthca_create_eq(struct mthca_dev *dev, - int nent, - u8 intr, - struct mthca_eq *eq) +static int mthca_create_eq(struct mthca_dev *dev, + int nent, + u8 intr, + struct mthca_eq *eq) { int npages; u64 *dma_list = NULL; @@ -664,9 +663,9 @@ static void mthca_free_irqs(struct mthca_dev *dev) dev->eq_table.eq + i); } -static int __devinit mthca_map_reg(struct mthca_dev *dev, - unsigned long offset, unsigned long size, - void __iomem **map) +static int mthca_map_reg(struct mthca_dev *dev, + unsigned long offset, unsigned long size, + void __iomem **map) { unsigned long base = pci_resource_start(dev->pdev, 0); @@ -691,7 +690,7 @@ static void mthca_unmap_reg(struct mthca_dev *dev, unsigned long offset, iounmap(map); } -static int __devinit mthca_map_eq_regs(struct mthca_dev *dev) +static int mthca_map_eq_regs(struct mthca_dev *dev) { if (mthca_is_memfree(dev)) { /* @@ -781,7 +780,7 @@ static void mthca_unmap_eq_regs(struct mthca_dev *dev) } } -int __devinit mthca_map_eq_icm(struct mthca_dev *dev, u64 icm_virt) +int mthca_map_eq_icm(struct mthca_dev *dev, u64 icm_virt) { int ret; u8 status; @@ -825,7 +824,7 @@ void mthca_unmap_eq_icm(struct mthca_dev *dev) __free_page(dev->eq_table.icm_page); } -int __devinit mthca_init_eq_table(struct mthca_dev *dev) +int mthca_init_eq_table(struct mthca_dev *dev) { int err; u8 status; diff --git a/drivers/infiniband/hw/mthca/mthca_mad.c b/drivers/infiniband/hw/mthca/mthca_mad.c index 45e106f1480..acfa41d968e 100644 --- a/drivers/infiniband/hw/mthca/mthca_mad.c +++ b/drivers/infiniband/hw/mthca/mthca_mad.c @@ -317,7 +317,7 @@ err: return ret; } -void __devexit mthca_free_agents(struct mthca_dev *dev) +void mthca_free_agents(struct mthca_dev *dev) { struct ib_mad_agent *agent; int p, q; diff --git a/drivers/infiniband/hw/mthca/mthca_main.c b/drivers/infiniband/hw/mthca/mthca_main.c index 47ea0214836..0491ec7a7c0 100644 --- a/drivers/infiniband/hw/mthca/mthca_main.c +++ b/drivers/infiniband/hw/mthca/mthca_main.c @@ -98,7 +98,7 @@ static struct mthca_profile default_profile = { .uarc_size = 1 << 18, /* Arbel only */ }; -static int __devinit mthca_tune_pci(struct mthca_dev *mdev) +static int mthca_tune_pci(struct mthca_dev *mdev) { int cap; u16 val; @@ -143,7 +143,7 @@ static int __devinit mthca_tune_pci(struct mthca_dev *mdev) return 0; } -static int __devinit mthca_dev_lim(struct mthca_dev *mdev, struct mthca_dev_lim *dev_lim) +static int mthca_dev_lim(struct mthca_dev *mdev, struct mthca_dev_lim *dev_lim) { int err; u8 status; @@ -255,7 +255,7 @@ static int __devinit mthca_dev_lim(struct mthca_dev *mdev, struct mthca_dev_lim return 0; } -static int __devinit mthca_init_tavor(struct mthca_dev *mdev) +static int mthca_init_tavor(struct mthca_dev *mdev) { u8 status; int err; @@ -333,7 +333,7 @@ err_disable: return err; } -static int __devinit mthca_load_fw(struct mthca_dev *mdev) +static int mthca_load_fw(struct mthca_dev *mdev) { u8 status; int err; @@ -379,10 +379,10 @@ err_free: return err; } -static int __devinit mthca_init_icm(struct mthca_dev *mdev, - struct mthca_dev_lim *dev_lim, - struct mthca_init_hca_param *init_hca, - u64 icm_size) +static int mthca_init_icm(struct mthca_dev *mdev, + struct mthca_dev_lim *dev_lim, + struct mthca_init_hca_param *init_hca, + u64 icm_size) { u64 aux_pages; u8 status; @@ -575,7 +575,7 @@ static void mthca_free_icms(struct mthca_dev *mdev) mthca_free_icm(mdev, mdev->fw.arbel.aux_icm); } -static int __devinit mthca_init_arbel(struct mthca_dev *mdev) +static int mthca_init_arbel(struct mthca_dev *mdev) { struct mthca_dev_lim dev_lim; struct mthca_profile profile; @@ -683,7 +683,7 @@ static void mthca_close_hca(struct mthca_dev *mdev) mthca_SYS_DIS(mdev, &status); } -static int __devinit mthca_init_hca(struct mthca_dev *mdev) +static int mthca_init_hca(struct mthca_dev *mdev) { u8 status; int err; @@ -720,7 +720,7 @@ err_close: return err; } -static int __devinit mthca_setup_hca(struct mthca_dev *dev) +static int mthca_setup_hca(struct mthca_dev *dev) { int err; u8 status; @@ -875,8 +875,7 @@ err_uar_table_free: return err; } -static int __devinit mthca_request_regions(struct pci_dev *pdev, - int ddr_hidden) +static int mthca_request_regions(struct pci_dev *pdev, int ddr_hidden) { int err; @@ -928,7 +927,7 @@ static void mthca_release_regions(struct pci_dev *pdev, MTHCA_HCR_SIZE); } -static int __devinit mthca_enable_msi_x(struct mthca_dev *mdev) +static int mthca_enable_msi_x(struct mthca_dev *mdev) { struct msix_entry entries[3]; int err; @@ -1213,7 +1212,7 @@ int __mthca_restart_one(struct pci_dev *pdev) } static int __devinit mthca_init_one(struct pci_dev *pdev, - const struct pci_device_id *id) + const struct pci_device_id *id) { static int mthca_version_printed = 0; int ret; diff --git a/drivers/infiniband/hw/mthca/mthca_mcg.c b/drivers/infiniband/hw/mthca/mthca_mcg.c index 47ca8a9b724..a8ad072be07 100644 --- a/drivers/infiniband/hw/mthca/mthca_mcg.c +++ b/drivers/infiniband/hw/mthca/mthca_mcg.c @@ -32,7 +32,6 @@ * $Id: mthca_mcg.c 1349 2004-12-16 21:09:43Z roland $ */ -#include <linux/init.h> #include <linux/string.h> #include <linux/slab.h> @@ -371,7 +370,7 @@ int mthca_multicast_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid) return err; } -int __devinit mthca_init_mcg_table(struct mthca_dev *dev) +int mthca_init_mcg_table(struct mthca_dev *dev) { int err; int table_size = dev->limits.num_mgms + dev->limits.num_amgms; diff --git a/drivers/infiniband/hw/mthca/mthca_mr.c b/drivers/infiniband/hw/mthca/mthca_mr.c index a486dec1707..f71ffa88db3 100644 --- a/drivers/infiniband/hw/mthca/mthca_mr.c +++ b/drivers/infiniband/hw/mthca/mthca_mr.c @@ -34,7 +34,6 @@ */ #include <linux/slab.h> -#include <linux/init.h> #include <linux/errno.h> #include "mthca_dev.h" @@ -135,7 +134,7 @@ static void mthca_buddy_free(struct mthca_buddy *buddy, u32 seg, int order) spin_unlock(&buddy->lock); } -static int __devinit mthca_buddy_init(struct mthca_buddy *buddy, int max_order) +static int mthca_buddy_init(struct mthca_buddy *buddy, int max_order) { int i, s; @@ -759,7 +758,7 @@ void mthca_arbel_fmr_unmap(struct mthca_dev *dev, struct mthca_fmr *fmr) *(u8 *) fmr->mem.arbel.mpt = MTHCA_MPT_STATUS_SW; } -int __devinit mthca_init_mr_table(struct mthca_dev *dev) +int mthca_init_mr_table(struct mthca_dev *dev) { unsigned long addr; int err, i; diff --git a/drivers/infiniband/hw/mthca/mthca_pd.c b/drivers/infiniband/hw/mthca/mthca_pd.c index 59df51614c8..c1e950764bd 100644 --- a/drivers/infiniband/hw/mthca/mthca_pd.c +++ b/drivers/infiniband/hw/mthca/mthca_pd.c @@ -34,7 +34,6 @@ * $Id: mthca_pd.c 1349 2004-12-16 21:09:43Z roland $ */ -#include <linux/init.h> #include <linux/errno.h> #include "mthca_dev.h" @@ -69,7 +68,7 @@ void mthca_pd_free(struct mthca_dev *dev, struct mthca_pd *pd) mthca_free(&dev->pd_table.alloc, pd->pd_num); } -int __devinit mthca_init_pd_table(struct mthca_dev *dev) +int mthca_init_pd_table(struct mthca_dev *dev) { return mthca_alloc_init(&dev->pd_table.alloc, dev->limits.num_pds, diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c index fc67f780581..21422a3336a 100644 --- a/drivers/infiniband/hw/mthca/mthca_provider.c +++ b/drivers/infiniband/hw/mthca/mthca_provider.c @@ -1100,11 +1100,10 @@ static struct ib_fmr *mthca_alloc_fmr(struct ib_pd *pd, int mr_access_flags, struct mthca_fmr *fmr; int err; - fmr = kmalloc(sizeof *fmr, GFP_KERNEL); + fmr = kmemdup(fmr_attr, sizeof *fmr, GFP_KERNEL); if (!fmr) return ERR_PTR(-ENOMEM); - memcpy(&fmr->attr, fmr_attr, sizeof *fmr_attr); err = mthca_fmr_alloc(to_mdev(pd->device), to_mpd(pd)->pd_num, convert_access(mr_access_flags), fmr); diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c b/drivers/infiniband/hw/mthca/mthca_qp.c index 5e5c58b9920..33e3ba7937f 100644 --- a/drivers/infiniband/hw/mthca/mthca_qp.c +++ b/drivers/infiniband/hw/mthca/mthca_qp.c @@ -35,10 +35,11 @@ * $Id: mthca_qp.c 1355 2004-12-17 15:23:43Z roland $ */ -#include <linux/init.h> #include <linux/string.h> #include <linux/slab.h> +#include <asm/io.h> + #include <rdma/ib_verbs.h> #include <rdma/ib_cache.h> #include <rdma/ib_pack.h> @@ -1732,6 +1733,11 @@ out: mthca_write64(doorbell, dev->kar + MTHCA_SEND_DOORBELL, MTHCA_GET_DOORBELL_LOCK(&dev->doorbell_lock)); + /* + * Make sure doorbells don't leak out of SQ spinlock + * and reach the HCA out of order: + */ + mmiowb(); } qp->sq.next_ind = ind; @@ -1851,6 +1857,12 @@ out: qp->rq.next_ind = ind; qp->rq.head += nreq; + /* + * Make sure doorbells don't leak out of RQ spinlock and reach + * the HCA out of order: + */ + mmiowb(); + spin_unlock_irqrestore(&qp->rq.lock, flags); return err; } @@ -2112,6 +2124,12 @@ out: MTHCA_GET_DOORBELL_LOCK(&dev->doorbell_lock)); } + /* + * Make sure doorbells don't leak out of SQ spinlock and reach + * the HCA out of order: + */ + mmiowb(); + spin_unlock_irqrestore(&qp->sq.lock, flags); return err; } @@ -2222,7 +2240,7 @@ void mthca_free_err_wqe(struct mthca_dev *dev, struct mthca_qp *qp, int is_send, *new_wqe = 0; } -int __devinit mthca_init_qp_table(struct mthca_dev *dev) +int mthca_init_qp_table(struct mthca_dev *dev) { int err; u8 status; diff --git a/drivers/infiniband/hw/mthca/mthca_srq.c b/drivers/infiniband/hw/mthca/mthca_srq.c index 92a72f52152..34d2c476896 100644 --- a/drivers/infiniband/hw/mthca/mthca_srq.c +++ b/drivers/infiniband/hw/mthca/mthca_srq.c @@ -35,6 +35,8 @@ #include <linux/slab.h> #include <linux/string.h> +#include <asm/io.h> + #include "mthca_dev.h" #include "mthca_cmd.h" #include "mthca_memfree.h" @@ -118,7 +120,7 @@ static void mthca_arbel_init_srq_context(struct mthca_dev *dev, memset(context, 0, sizeof *context); - logsize = long_log2(srq->max) + srq->wqe_shift; + logsize = long_log2(srq->max); context->state_logsize_srqn = cpu_to_be32(logsize << 24 | srq->srqn); context->lkey = cpu_to_be32(srq->mr.ibmr.lkey); context->db_index = cpu_to_be32(srq->db_index); @@ -595,6 +597,12 @@ int mthca_tavor_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr, MTHCA_GET_DOORBELL_LOCK(&dev->doorbell_lock)); } + /* + * Make sure doorbells don't leak out of SRQ spinlock and + * reach the HCA out of order: + */ + mmiowb(); + spin_unlock_irqrestore(&srq->lock, flags); return err; } @@ -707,7 +715,7 @@ int mthca_max_srq_sge(struct mthca_dev *dev) sizeof (struct mthca_data_seg)); } -int __devinit mthca_init_srq_table(struct mthca_dev *dev) +int mthca_init_srq_table(struct mthca_dev *dev) { int err; diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h index 0b8a79d53a0..f2b61851a49 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib.h +++ b/drivers/infiniband/ulp/ipoib/ipoib.h @@ -233,7 +233,7 @@ static inline struct ipoib_neigh **to_ipoib_neigh(struct neighbour *neigh) } struct ipoib_neigh *ipoib_neigh_alloc(struct neighbour *neigh); -void ipoib_neigh_free(struct ipoib_neigh *neigh); +void ipoib_neigh_free(struct net_device *dev, struct ipoib_neigh *neigh); extern struct workqueue_struct *ipoib_workqueue; diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index 1eaf00e9862..5ba3154320b 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -49,6 +49,8 @@ #include <net/dst.h> +#define IPOIB_QPN(ha) (be32_to_cpup((__be32 *) ha) & 0xffffff) + MODULE_AUTHOR("Roland Dreier"); MODULE_DESCRIPTION("IP-over-InfiniBand net driver"); MODULE_LICENSE("Dual BSD/GPL"); @@ -262,7 +264,7 @@ static void path_free(struct net_device *dev, struct ipoib_path *path) if (neigh->ah) ipoib_put_ah(neigh->ah); - ipoib_neigh_free(neigh); + ipoib_neigh_free(dev, neigh); } spin_unlock_irqrestore(&priv->lock, flags); @@ -520,14 +522,14 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev) memcpy(&neigh->dgid.raw, &path->pathrec.dgid.raw, sizeof(union ib_gid)); - ipoib_send(dev, skb, path->ah, - be32_to_cpup((__be32 *) skb->dst->neighbour->ha)); + ipoib_send(dev, skb, path->ah, IPOIB_QPN(skb->dst->neighbour->ha)); } else { neigh->ah = NULL; - __skb_queue_tail(&neigh->queue, skb); if (!path->query && path_rec_start(dev, path)) goto err_list; + + __skb_queue_tail(&neigh->queue, skb); } spin_unlock(&priv->lock); @@ -537,7 +539,7 @@ err_list: list_del(&neigh->list); err_path: - ipoib_neigh_free(neigh); + ipoib_neigh_free(dev, neigh); ++priv->stats.tx_dropped; dev_kfree_skb_any(skb); @@ -599,8 +601,7 @@ static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, ipoib_dbg(priv, "Send unicast ARP to %04x\n", be16_to_cpu(path->pathrec.dlid)); - ipoib_send(dev, skb, path->ah, - be32_to_cpup((__be32 *) phdr->hwaddr)); + ipoib_send(dev, skb, path->ah, IPOIB_QPN(phdr->hwaddr)); } else if ((path->query || !path_rec_start(dev, path)) && skb_queue_len(&path->queue) < IPOIB_MAX_PATH_REC_QUEUE) { /* put pseudoheader back on for next time */ @@ -655,14 +656,13 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) */ ipoib_put_ah(neigh->ah); list_del(&neigh->list); - ipoib_neigh_free(neigh); + ipoib_neigh_free(dev, neigh); spin_unlock(&priv->lock); ipoib_path_lookup(skb, dev); goto out; } - ipoib_send(dev, skb, neigh->ah, - be32_to_cpup((__be32 *) skb->dst->neighbour->ha)); + ipoib_send(dev, skb, neigh->ah, IPOIB_QPN(skb->dst->neighbour->ha)); goto out; } @@ -694,7 +694,7 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) IPOIB_GID_FMT "\n", skb->dst ? "neigh" : "dst", be16_to_cpup((__be16 *) skb->data), - be32_to_cpup((__be32 *) phdr->hwaddr), + IPOIB_QPN(phdr->hwaddr), IPOIB_GID_RAW_ARG(phdr->hwaddr + 4)); dev_kfree_skb_any(skb); ++priv->stats.tx_dropped; @@ -777,7 +777,7 @@ static void ipoib_neigh_destructor(struct neighbour *n) ipoib_dbg(priv, "neigh_destructor for %06x " IPOIB_GID_FMT "\n", - be32_to_cpup((__be32 *) n->ha), + IPOIB_QPN(n->ha), IPOIB_GID_RAW_ARG(n->ha + 4)); spin_lock_irqsave(&priv->lock, flags); @@ -787,7 +787,7 @@ static void ipoib_neigh_destructor(struct neighbour *n) if (neigh->ah) ah = neigh->ah; list_del(&neigh->list); - ipoib_neigh_free(neigh); + ipoib_neigh_free(n->dev, neigh); } spin_unlock_irqrestore(&priv->lock, flags); @@ -810,9 +810,15 @@ struct ipoib_neigh *ipoib_neigh_alloc(struct neighbour *neighbour) return neigh; } -void ipoib_neigh_free(struct ipoib_neigh *neigh) +void ipoib_neigh_free(struct net_device *dev, struct ipoib_neigh *neigh) { + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct sk_buff *skb; *to_ipoib_neigh(neigh->neighbour) = NULL; + while ((skb = __skb_dequeue(&neigh->queue))) { + ++priv->stats.tx_dropped; + dev_kfree_skb_any(skb); + } kfree(neigh); } diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c index 3faa1820f0e..d282d65e3ee 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c @@ -114,7 +114,7 @@ static void ipoib_mcast_free(struct ipoib_mcast *mcast) */ if (neigh->ah) ipoib_put_ah(neigh->ah); - ipoib_neigh_free(neigh); + ipoib_neigh_free(dev, neigh); } spin_unlock_irqrestore(&priv->lock, flags); diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c index eb6f98d8228..9b2041e25d5 100644 --- a/drivers/infiniband/ulp/iser/iscsi_iser.c +++ b/drivers/infiniband/ulp/iser/iscsi_iser.c @@ -363,11 +363,11 @@ iscsi_iser_conn_start(struct iscsi_cls_conn *cls_conn) struct iscsi_conn *conn = cls_conn->dd_data; int err; - err = iscsi_conn_start(cls_conn); + err = iser_conn_set_full_featured_mode(conn); if (err) return err; - return iser_conn_set_full_featured_mode(conn); + return iscsi_conn_start(cls_conn); } static struct iscsi_transport iscsi_iser_transport; diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h index 9c53916f28c..234e5b061a7 100644 --- a/drivers/infiniband/ulp/iser/iscsi_iser.h +++ b/drivers/infiniband/ulp/iser/iscsi_iser.h @@ -283,7 +283,7 @@ struct iser_global { struct mutex connlist_mutex; struct list_head connlist; /* all iSER IB connections */ - kmem_cache_t *desc_cache; + struct kmem_cache *desc_cache; }; extern struct iser_global ig; diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index 4b09147f438..64ab5fc7cca 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -1176,9 +1176,11 @@ static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event) break; } - target->status = srp_alloc_iu_bufs(target); - if (target->status) - break; + if (!target->rx_ring[0]) { + target->status = srp_alloc_iu_bufs(target); + if (target->status) + break; + } qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL); if (!qp_attr) { @@ -1716,7 +1718,8 @@ static ssize_t srp_create_target(struct class_device *class_dev, if (!target_host) return -ENOMEM; - target_host->max_lun = SRP_MAX_LUN; + target_host->max_lun = SRP_MAX_LUN; + target_host->max_cmd_len = sizeof ((struct srp_cmd *) (void *) 0L)->cdb; target = host_to_target(target_host); |