diff options
Diffstat (limited to 'drivers/infiniband')
82 files changed, 2236 insertions, 1204 deletions
diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c index ce511d8748c..5be1bd4fc7e 100644 --- a/drivers/infiniband/core/addr.c +++ b/drivers/infiniband/core/addr.c @@ -514,7 +514,7 @@ static struct notifier_block nb = { .notifier_call = netevent_callback }; -static int addr_init(void) +static int __init addr_init(void) { addr_wq = create_singlethread_workqueue("ib_addr"); if (!addr_wq) @@ -524,7 +524,7 @@ static int addr_init(void) return 0; } -static void addr_cleanup(void) +static void __exit addr_cleanup(void) { unregister_netevent_notifier(&nb); destroy_workqueue(addr_wq); diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c index f1e82a92e61..5130fc55b8e 100644 --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -927,8 +927,7 @@ int ib_cm_listen(struct ib_cm_id *cm_id, __be64 service_id, __be64 service_mask, unsigned long flags; int ret = 0; - service_mask = service_mask ? service_mask : - __constant_cpu_to_be64(~0ULL); + service_mask = service_mask ? service_mask : ~cpu_to_be64(0); service_id &= service_mask; if ((service_id & IB_SERVICE_ID_AGN_MASK) == IB_CM_ASSIGN_SERVICE_ID && (service_id != IB_CM_ASSIGN_SERVICE_ID)) @@ -954,7 +953,7 @@ int ib_cm_listen(struct ib_cm_id *cm_id, __be64 service_id, __be64 service_mask, spin_lock_irqsave(&cm.lock, flags); if (service_id == IB_CM_ASSIGN_SERVICE_ID) { cm_id->service_id = cpu_to_be64(cm.listen_service_id++); - cm_id->service_mask = __constant_cpu_to_be64(~0ULL); + cm_id->service_mask = ~cpu_to_be64(0); } else { cm_id->service_id = service_id; cm_id->service_mask = service_mask; @@ -1134,7 +1133,7 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, goto error1; } cm_id->service_id = param->service_id; - cm_id->service_mask = __constant_cpu_to_be64(~0ULL); + cm_id->service_mask = ~cpu_to_be64(0); cm_id_priv->timeout_ms = cm_convert_to_ms( param->primary_path->packet_life_time) * 2 + cm_convert_to_ms( @@ -1545,7 +1544,7 @@ static int cm_req_handler(struct cm_work *work) cm_id_priv->id.cm_handler = listen_cm_id_priv->id.cm_handler; cm_id_priv->id.context = listen_cm_id_priv->id.context; cm_id_priv->id.service_id = req_msg->service_id; - cm_id_priv->id.service_mask = __constant_cpu_to_be64(~0ULL); + cm_id_priv->id.service_mask = ~cpu_to_be64(0); cm_process_routed_req(req_msg, work->mad_recv_wc->wc); cm_format_paths_from_req(req_msg, &work->path[0], &work->path[1]); @@ -2898,7 +2897,7 @@ int ib_send_cm_sidr_req(struct ib_cm_id *cm_id, goto out; cm_id->service_id = param->service_id; - cm_id->service_mask = __constant_cpu_to_be64(~0ULL); + cm_id->service_mask = ~cpu_to_be64(0); cm_id_priv->timeout_ms = param->timeout_ms; cm_id_priv->max_cm_retries = param->max_cm_retries; ret = cm_alloc_msg(cm_id_priv, &msg); @@ -2992,7 +2991,7 @@ static int cm_sidr_req_handler(struct cm_work *work) cm_id_priv->id.cm_handler = cur_cm_id_priv->id.cm_handler; cm_id_priv->id.context = cur_cm_id_priv->id.context; cm_id_priv->id.service_id = sidr_req_msg->service_id; - cm_id_priv->id.service_mask = __constant_cpu_to_be64(~0ULL); + cm_id_priv->id.service_mask = ~cpu_to_be64(0); cm_format_sidr_req_event(work, &cur_cm_id_priv->id); cm_process_work(cm_id_priv, work); @@ -3789,7 +3788,7 @@ static int __init ib_cm_init(void) rwlock_init(&cm.device_lock); spin_lock_init(&cm.lock); cm.listen_service_table = RB_ROOT; - cm.listen_service_id = __constant_be64_to_cpu(IB_CM_ASSIGN_SERVICE_ID); + cm.listen_service_id = be64_to_cpu(IB_CM_ASSIGN_SERVICE_ID); cm.remote_id_table = RB_ROOT; cm.remote_qp_table = RB_ROOT; cm.remote_sidr_table = RB_ROOT; diff --git a/drivers/infiniband/core/cm_msgs.h b/drivers/infiniband/core/cm_msgs.h index aec9c7af825..7e63c08f697 100644 --- a/drivers/infiniband/core/cm_msgs.h +++ b/drivers/infiniband/core/cm_msgs.h @@ -44,17 +44,17 @@ #define IB_CM_CLASS_VERSION 2 /* IB specification 1.2 */ -#define CM_REQ_ATTR_ID __constant_htons(0x0010) -#define CM_MRA_ATTR_ID __constant_htons(0x0011) -#define CM_REJ_ATTR_ID __constant_htons(0x0012) -#define CM_REP_ATTR_ID __constant_htons(0x0013) -#define CM_RTU_ATTR_ID __constant_htons(0x0014) -#define CM_DREQ_ATTR_ID __constant_htons(0x0015) -#define CM_DREP_ATTR_ID __constant_htons(0x0016) -#define CM_SIDR_REQ_ATTR_ID __constant_htons(0x0017) -#define CM_SIDR_REP_ATTR_ID __constant_htons(0x0018) -#define CM_LAP_ATTR_ID __constant_htons(0x0019) -#define CM_APR_ATTR_ID __constant_htons(0x001A) +#define CM_REQ_ATTR_ID cpu_to_be16(0x0010) +#define CM_MRA_ATTR_ID cpu_to_be16(0x0011) +#define CM_REJ_ATTR_ID cpu_to_be16(0x0012) +#define CM_REP_ATTR_ID cpu_to_be16(0x0013) +#define CM_RTU_ATTR_ID cpu_to_be16(0x0014) +#define CM_DREQ_ATTR_ID cpu_to_be16(0x0015) +#define CM_DREP_ATTR_ID cpu_to_be16(0x0016) +#define CM_SIDR_REQ_ATTR_ID cpu_to_be16(0x0017) +#define CM_SIDR_REP_ATTR_ID cpu_to_be16(0x0018) +#define CM_LAP_ATTR_ID cpu_to_be16(0x0019) +#define CM_APR_ATTR_ID cpu_to_be16(0x001A) enum cm_msg_sequence { CM_MSG_SEQUENCE_REQ, diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 2a2e50871b4..075317884b5 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -297,21 +297,25 @@ static void cma_detach_from_dev(struct rdma_id_private *id_priv) id_priv->cma_dev = NULL; } -static int cma_set_qkey(struct ib_device *device, u8 port_num, - enum rdma_port_space ps, - struct rdma_dev_addr *dev_addr, u32 *qkey) +static int cma_set_qkey(struct rdma_id_private *id_priv) { struct ib_sa_mcmember_rec rec; int ret = 0; - switch (ps) { + if (id_priv->qkey) + return 0; + + switch (id_priv->id.ps) { case RDMA_PS_UDP: - *qkey = RDMA_UDP_QKEY; + id_priv->qkey = RDMA_UDP_QKEY; break; case RDMA_PS_IPOIB: - ib_addr_get_mgid(dev_addr, &rec.mgid); - ret = ib_sa_get_mcmember_rec(device, port_num, &rec.mgid, &rec); - *qkey = be32_to_cpu(rec.qkey); + ib_addr_get_mgid(&id_priv->id.route.addr.dev_addr, &rec.mgid); + ret = ib_sa_get_mcmember_rec(id_priv->id.device, + id_priv->id.port_num, &rec.mgid, + &rec); + if (!ret) + id_priv->qkey = be32_to_cpu(rec.qkey); break; default: break; @@ -341,12 +345,7 @@ static int cma_acquire_dev(struct rdma_id_private *id_priv) ret = ib_find_cached_gid(cma_dev->device, &gid, &id_priv->id.port_num, NULL); if (!ret) { - ret = cma_set_qkey(cma_dev->device, - id_priv->id.port_num, - id_priv->id.ps, dev_addr, - &id_priv->qkey); - if (!ret) - cma_attach_to_dev(id_priv, cma_dev); + cma_attach_to_dev(id_priv, cma_dev); break; } } @@ -578,6 +577,10 @@ static int cma_ib_init_qp_attr(struct rdma_id_private *id_priv, *qp_attr_mask = IB_QP_STATE | IB_QP_PKEY_INDEX | IB_QP_PORT; if (cma_is_ud_ps(id_priv->id.ps)) { + ret = cma_set_qkey(id_priv); + if (ret) + return ret; + qp_attr->qkey = id_priv->qkey; *qp_attr_mask |= IB_QP_QKEY; } else { @@ -2201,6 +2204,12 @@ static int cma_sidr_rep_handler(struct ib_cm_id *cm_id, event.status = ib_event->param.sidr_rep_rcvd.status; break; } + ret = cma_set_qkey(id_priv); + if (ret) { + event.event = RDMA_CM_EVENT_ADDR_ERROR; + event.status = -EINVAL; + break; + } if (id_priv->qkey != rep->qkey) { event.event = RDMA_CM_EVENT_UNREACHABLE; event.status = -EINVAL; @@ -2480,10 +2489,14 @@ static int cma_send_sidr_rep(struct rdma_id_private *id_priv, const void *private_data, int private_data_len) { struct ib_cm_sidr_rep_param rep; + int ret; memset(&rep, 0, sizeof rep); rep.status = status; if (status == IB_SIDR_SUCCESS) { + ret = cma_set_qkey(id_priv); + if (ret) + return ret; rep.qp_num = id_priv->qp_num; rep.qkey = id_priv->qkey; } @@ -2713,6 +2726,10 @@ static int cma_join_ib_multicast(struct rdma_id_private *id_priv, IB_SA_MCMEMBER_REC_FLOW_LABEL | IB_SA_MCMEMBER_REC_TRAFFIC_CLASS; + if (id_priv->id.ps == RDMA_PS_IPOIB) + comp_mask |= IB_SA_MCMEMBER_REC_RATE | + IB_SA_MCMEMBER_REC_RATE_SELECTOR; + mc->multicast.ib = ib_sa_join_multicast(&sa_client, id_priv->id.device, id_priv->id.port_num, &rec, comp_mask, GFP_KERNEL, @@ -2943,7 +2960,7 @@ static void cma_remove_one(struct ib_device *device) kfree(cma_dev); } -static int cma_init(void) +static int __init cma_init(void) { int ret, low, high, remaining; @@ -2973,7 +2990,7 @@ err: return ret; } -static void cma_cleanup(void) +static void __exit cma_cleanup(void) { ib_unregister_client(&cma_client); unregister_netdevice_notifier(&cma_nb); diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c index 7913b804311..d1fba415333 100644 --- a/drivers/infiniband/core/device.c +++ b/drivers/infiniband/core/device.c @@ -193,7 +193,7 @@ void ib_dealloc_device(struct ib_device *device) BUG_ON(device->reg_state != IB_DEV_UNREGISTERED); - ib_device_unregister_sysfs(device); + kobject_put(&device->dev.kobj); } EXPORT_SYMBOL(ib_dealloc_device); @@ -348,6 +348,8 @@ void ib_unregister_device(struct ib_device *device) mutex_unlock(&device_mutex); + ib_device_unregister_sysfs(device); + spin_lock_irqsave(&device->client_data_lock, flags); list_for_each_entry_safe(context, tmp, &device->client_data_list, list) kfree(context); diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index 5c54fc2350b..de922a04ca2 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -301,6 +301,16 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, mad_agent_priv->agent.context = context; mad_agent_priv->agent.qp = port_priv->qp_info[qpn].qp; mad_agent_priv->agent.port_num = port_num; + spin_lock_init(&mad_agent_priv->lock); + INIT_LIST_HEAD(&mad_agent_priv->send_list); + INIT_LIST_HEAD(&mad_agent_priv->wait_list); + INIT_LIST_HEAD(&mad_agent_priv->done_list); + INIT_LIST_HEAD(&mad_agent_priv->rmpp_list); + INIT_DELAYED_WORK(&mad_agent_priv->timed_work, timeout_sends); + INIT_LIST_HEAD(&mad_agent_priv->local_list); + INIT_WORK(&mad_agent_priv->local_work, local_completions); + atomic_set(&mad_agent_priv->refcount, 1); + init_completion(&mad_agent_priv->comp); spin_lock_irqsave(&port_priv->reg_lock, flags); mad_agent_priv->agent.hi_tid = ++ib_mad_client_id; @@ -350,17 +360,6 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, list_add_tail(&mad_agent_priv->agent_list, &port_priv->agent_list); spin_unlock_irqrestore(&port_priv->reg_lock, flags); - spin_lock_init(&mad_agent_priv->lock); - INIT_LIST_HEAD(&mad_agent_priv->send_list); - INIT_LIST_HEAD(&mad_agent_priv->wait_list); - INIT_LIST_HEAD(&mad_agent_priv->done_list); - INIT_LIST_HEAD(&mad_agent_priv->rmpp_list); - INIT_DELAYED_WORK(&mad_agent_priv->timed_work, timeout_sends); - INIT_LIST_HEAD(&mad_agent_priv->local_list); - INIT_WORK(&mad_agent_priv->local_work, local_completions); - atomic_set(&mad_agent_priv->refcount, 1); - init_completion(&mad_agent_priv->comp); - return &mad_agent_priv->agent; error4: @@ -743,9 +742,7 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv, break; case IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_CONSUMED: kmem_cache_free(ib_mad_cache, mad_priv); - kfree(local); - ret = 1; - goto out; + break; case IB_MAD_RESULT_SUCCESS: /* Treat like an incoming receive MAD */ port_priv = ib_get_mad_port(mad_agent_priv->agent.device, @@ -756,10 +753,12 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv, &mad_priv->mad.mad); } if (!port_priv || !recv_mad_agent) { + /* + * No receiving agent so drop packet and + * generate send completion. + */ kmem_cache_free(ib_mad_cache, mad_priv); - kfree(local); - ret = 0; - goto out; + break; } local->mad_priv = mad_priv; local->recv_mad_agent = recv_mad_agent; @@ -2356,7 +2355,7 @@ static void local_completions(struct work_struct *work) struct ib_mad_local_private *local; struct ib_mad_agent_private *recv_mad_agent; unsigned long flags; - int recv = 0; + int free_mad; struct ib_wc wc; struct ib_mad_send_wc mad_send_wc; @@ -2370,14 +2369,15 @@ static void local_completions(struct work_struct *work) completion_list); list_del(&local->completion_list); spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + free_mad = 0; if (local->mad_priv) { recv_mad_agent = local->recv_mad_agent; if (!recv_mad_agent) { printk(KERN_ERR PFX "No receive MAD agent for local completion\n"); + free_mad = 1; goto local_send_completion; } - recv = 1; /* * Defined behavior is to complete response * before request @@ -2422,7 +2422,7 @@ local_send_completion: spin_lock_irqsave(&mad_agent_priv->lock, flags); atomic_dec(&mad_agent_priv->refcount); - if (!recv) + if (free_mad) kmem_cache_free(ib_mad_cache, local->mad_priv); kfree(local); } diff --git a/drivers/infiniband/core/mad_rmpp.c b/drivers/infiniband/core/mad_rmpp.c index 3af2b84cd83..57a3c6f947b 100644 --- a/drivers/infiniband/core/mad_rmpp.c +++ b/drivers/infiniband/core/mad_rmpp.c @@ -735,7 +735,7 @@ process_rmpp_data(struct ib_mad_agent_private *agent, goto bad; } - if (rmpp_hdr->seg_num == __constant_htonl(1)) { + if (rmpp_hdr->seg_num == cpu_to_be32(1)) { if (!(ib_get_rmpp_flags(rmpp_hdr) & IB_MGMT_RMPP_FLAG_FIRST)) { rmpp_status = IB_MGMT_RMPP_STATUS_BAD_SEG; goto bad; diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c index 7863a50d56f..1865049e80f 100644 --- a/drivers/infiniband/core/sa_query.c +++ b/drivers/infiniband/core/sa_query.c @@ -395,6 +395,8 @@ static void update_sm_ah(struct work_struct *work) } spin_lock_irq(&port->ah_lock); + if (port->sm_ah) + kref_put(&port->sm_ah->ref, free_sm_ah); port->sm_ah = new_ah; spin_unlock_irq(&port->ah_lock); diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c index b43f7d3682d..158a214da2f 100644 --- a/drivers/infiniband/core/sysfs.c +++ b/drivers/infiniband/core/sysfs.c @@ -66,11 +66,6 @@ struct port_table_attribute { int index; }; -static inline int ibdev_is_alive(const struct ib_device *dev) -{ - return dev->reg_state == IB_DEV_REGISTERED; -} - static ssize_t port_attr_show(struct kobject *kobj, struct attribute *attr, char *buf) { @@ -80,8 +75,6 @@ static ssize_t port_attr_show(struct kobject *kobj, if (!port_attr->show) return -EIO; - if (!ibdev_is_alive(p->ibdev)) - return -ENODEV; return port_attr->show(p, port_attr, buf); } @@ -562,9 +555,6 @@ static ssize_t show_node_type(struct device *device, { struct ib_device *dev = container_of(device, struct ib_device, dev); - if (!ibdev_is_alive(dev)) - return -ENODEV; - switch (dev->node_type) { case RDMA_NODE_IB_CA: return sprintf(buf, "%d: CA\n", dev->node_type); case RDMA_NODE_RNIC: return sprintf(buf, "%d: RNIC\n", dev->node_type); @@ -581,9 +571,6 @@ static ssize_t show_sys_image_guid(struct device *device, struct ib_device_attr attr; ssize_t ret; - if (!ibdev_is_alive(dev)) - return -ENODEV; - ret = ib_query_device(dev, &attr); if (ret) return ret; @@ -600,9 +587,6 @@ static ssize_t show_node_guid(struct device *device, { struct ib_device *dev = container_of(device, struct ib_device, dev); - if (!ibdev_is_alive(dev)) - return -ENODEV; - return sprintf(buf, "%04x:%04x:%04x:%04x\n", be16_to_cpu(((__be16 *) &dev->node_guid)[0]), be16_to_cpu(((__be16 *) &dev->node_guid)[1]), @@ -776,9 +760,9 @@ int ib_device_register_sysfs(struct ib_device *device) int i; class_dev->class = &ib_class; - class_dev->driver_data = device; class_dev->parent = device->dma_device; dev_set_name(class_dev, device->name); + dev_set_drvdata(class_dev, device); INIT_LIST_HEAD(&device->port_list); @@ -848,6 +832,9 @@ void ib_device_unregister_sysfs(struct ib_device *device) struct kobject *p, *t; struct ib_port *port; + /* Hold kobject until ib_dealloc_device() */ + kobject_get(&device->dev.kobj); + list_for_each_entry_safe(p, t, &device->port_list, entry) { list_del(&p->entry); port = container_of(p, struct ib_port, kobj); diff --git a/drivers/infiniband/hw/amso1100/c2.c b/drivers/infiniband/hw/amso1100/c2.c index 113f3c03c5b..0cfbb6d2f76 100644 --- a/drivers/infiniband/hw/amso1100/c2.c +++ b/drivers/infiniband/hw/amso1100/c2.c @@ -76,7 +76,6 @@ static irqreturn_t c2_interrupt(int irq, void *dev_id); static void c2_tx_timeout(struct net_device *netdev); static int c2_change_mtu(struct net_device *netdev, int new_mtu); static void c2_reset(struct c2_port *c2_port); -static struct net_device_stats *c2_get_stats(struct net_device *netdev); static struct pci_device_id c2_pci_table[] = { { PCI_DEVICE(0x18b8, 0xb001) }, @@ -349,7 +348,7 @@ static void c2_tx_clean(struct c2_port *c2_port) elem->hw_desc + C2_TXP_ADDR); __raw_writew((__force u16) cpu_to_be16(TXP_HTXD_DONE), elem->hw_desc + C2_TXP_FLAGS); - c2_port->netstats.tx_dropped++; + c2_port->netdev->stats.tx_dropped++; break; } else { __raw_writew(0, @@ -457,7 +456,7 @@ static void c2_rx_error(struct c2_port *c2_port, struct c2_element *elem) elem->hw_desc + C2_RXP_FLAGS); pr_debug("packet dropped\n"); - c2_port->netstats.rx_dropped++; + c2_port->netdev->stats.rx_dropped++; } static void c2_rx_interrupt(struct net_device *netdev) @@ -532,8 +531,8 @@ static void c2_rx_interrupt(struct net_device *netdev) netif_rx(skb); netdev->last_rx = jiffies; - c2_port->netstats.rx_packets++; - c2_port->netstats.rx_bytes += buflen; + netdev->stats.rx_packets++; + netdev->stats.rx_bytes += buflen; } /* Save where we left off */ @@ -797,8 +796,8 @@ static int c2_xmit_frame(struct sk_buff *skb, struct net_device *netdev) __raw_writew((__force u16) cpu_to_be16(TXP_HTXD_READY), elem->hw_desc + C2_TXP_FLAGS); - c2_port->netstats.tx_packets++; - c2_port->netstats.tx_bytes += maplen; + netdev->stats.tx_packets++; + netdev->stats.tx_bytes += maplen; /* Loop thru additional data fragments and queue them */ if (skb_shinfo(skb)->nr_frags) { @@ -823,8 +822,8 @@ static int c2_xmit_frame(struct sk_buff *skb, struct net_device *netdev) __raw_writew((__force u16) cpu_to_be16(TXP_HTXD_READY), elem->hw_desc + C2_TXP_FLAGS); - c2_port->netstats.tx_packets++; - c2_port->netstats.tx_bytes += maplen; + netdev->stats.tx_packets++; + netdev->stats.tx_bytes += maplen; } } @@ -845,13 +844,6 @@ static int c2_xmit_frame(struct sk_buff *skb, struct net_device *netdev) return NETDEV_TX_OK; } -static struct net_device_stats *c2_get_stats(struct net_device *netdev) -{ - struct c2_port *c2_port = netdev_priv(netdev); - - return &c2_port->netstats; -} - static void c2_tx_timeout(struct net_device *netdev) { struct c2_port *c2_port = netdev_priv(netdev); @@ -880,6 +872,16 @@ static int c2_change_mtu(struct net_device *netdev, int new_mtu) return ret; } +static const struct net_device_ops c2_netdev = { + .ndo_open = c2_up, + .ndo_stop = c2_down, + .ndo_start_xmit = c2_xmit_frame, + .ndo_tx_timeout = c2_tx_timeout, + .ndo_change_mtu = c2_change_mtu, + .ndo_set_mac_address = eth_mac_addr, + .ndo_validate_addr = eth_validate_addr, +}; + /* Initialize network device */ static struct net_device *c2_devinit(struct c2_dev *c2dev, void __iomem * mmio_addr) @@ -894,12 +896,7 @@ static struct net_device *c2_devinit(struct c2_dev *c2dev, SET_NETDEV_DEV(netdev, &c2dev->pcidev->dev); - netdev->open = c2_up; - netdev->stop = c2_down; - netdev->hard_start_xmit = c2_xmit_frame; - netdev->get_stats = c2_get_stats; - netdev->tx_timeout = c2_tx_timeout; - netdev->change_mtu = c2_change_mtu; + netdev->netdev_ops = &c2_netdev; netdev->watchdog_timeo = C2_TX_TIMEOUT; netdev->irq = c2dev->pcidev->irq; @@ -992,13 +989,13 @@ static int __devinit c2_probe(struct pci_dev *pcidev, } if ((sizeof(dma_addr_t) > 4)) { - ret = pci_set_dma_mask(pcidev, DMA_64BIT_MASK); + ret = pci_set_dma_mask(pcidev, DMA_BIT_MASK(64)); if (ret < 0) { printk(KERN_ERR PFX "64b DMA configuration failed\n"); goto bail2; } } else { - ret = pci_set_dma_mask(pcidev, DMA_32BIT_MASK); + ret = pci_set_dma_mask(pcidev, DMA_BIT_MASK(32)); if (ret < 0) { printk(KERN_ERR PFX "32b DMA configuration failed\n"); goto bail2; diff --git a/drivers/infiniband/hw/amso1100/c2.h b/drivers/infiniband/hw/amso1100/c2.h index d12a24a84fd..f7ff66f9836 100644 --- a/drivers/infiniband/hw/amso1100/c2.h +++ b/drivers/infiniband/hw/amso1100/c2.h @@ -369,8 +369,6 @@ struct c2_port { unsigned long mem_size; u32 rx_buf_size; - - struct net_device_stats netstats; }; /* diff --git a/drivers/infiniband/hw/amso1100/c2_cq.c b/drivers/infiniband/hw/amso1100/c2_cq.c index bb17cce3cb5..f5c45b194f5 100644 --- a/drivers/infiniband/hw/amso1100/c2_cq.c +++ b/drivers/infiniband/hw/amso1100/c2_cq.c @@ -133,7 +133,7 @@ static inline int c2_poll_one(struct c2_dev *c2dev, struct c2_qp *qp; int is_recv = 0; - ce = (struct c2wr_ce *) c2_mq_consume(&cq->mq); + ce = c2_mq_consume(&cq->mq); if (!ce) { return -EAGAIN; } @@ -146,7 +146,7 @@ static inline int c2_poll_one(struct c2_dev *c2dev, while ((qp = (struct c2_qp *) (unsigned long) ce->qp_user_context) == NULL) { c2_mq_free(&cq->mq); - ce = (struct c2wr_ce *) c2_mq_consume(&cq->mq); + ce = c2_mq_consume(&cq->mq); if (!ce) return -EAGAIN; } diff --git a/drivers/infiniband/hw/amso1100/c2_provider.c b/drivers/infiniband/hw/amso1100/c2_provider.c index 5119d650818..f1948fad85d 100644 --- a/drivers/infiniband/hw/amso1100/c2_provider.c +++ b/drivers/infiniband/hw/amso1100/c2_provider.c @@ -708,26 +708,27 @@ static int c2_pseudo_xmit_frame(struct sk_buff *skb, struct net_device *netdev) static int c2_pseudo_change_mtu(struct net_device *netdev, int new_mtu) { - int ret = 0; - if (new_mtu < ETH_ZLEN || new_mtu > ETH_JUMBO_MTU) return -EINVAL; netdev->mtu = new_mtu; /* TODO: Tell rnic about new rmda interface mtu */ - return ret; + return 0; } +static const struct net_device_ops c2_pseudo_netdev_ops = { + .ndo_open = c2_pseudo_up, + .ndo_stop = c2_pseudo_down, + .ndo_start_xmit = c2_pseudo_xmit_frame, + .ndo_change_mtu = c2_pseudo_change_mtu, + .ndo_validate_addr = eth_validate_addr, +}; + static void setup(struct net_device *netdev) { - netdev->open = c2_pseudo_up; - netdev->stop = c2_pseudo_down; - netdev->hard_start_xmit = c2_pseudo_xmit_frame; - netdev->get_stats = NULL; - netdev->tx_timeout = NULL; - netdev->set_mac_address = NULL; - netdev->change_mtu = c2_pseudo_change_mtu; + netdev->netdev_ops = &c2_pseudo_netdev_ops; + netdev->watchdog_timeo = 0; netdev->type = ARPHRD_ETHER; netdev->mtu = 1500; @@ -735,7 +736,6 @@ static void setup(struct net_device *netdev) netdev->addr_len = ETH_ALEN; netdev->tx_queue_len = 0; netdev->flags |= IFF_NOARP; - return; } static struct net_device *c2_pseudo_netdev_init(struct c2_dev *c2dev) diff --git a/drivers/infiniband/hw/cxgb3/cxio_hal.c b/drivers/infiniband/hw/cxgb3/cxio_hal.c index 4dcf08b3fd8..62f9cf2f94e 100644 --- a/drivers/infiniband/hw/cxgb3/cxio_hal.c +++ b/drivers/infiniband/hw/cxgb3/cxio_hal.c @@ -152,7 +152,7 @@ static int cxio_hal_clear_qp_ctx(struct cxio_rdev *rdev_p, u32 qpid) sge_cmd = qpid << 8 | 3; wqe->sge_cmd = cpu_to_be64(sge_cmd); skb->priority = CPL_PRIORITY_CONTROL; - return (cxgb3_ofld_send(rdev_p->t3cdev_p, skb)); + return iwch_cxgb3_ofld_send(rdev_p->t3cdev_p, skb); } int cxio_create_cq(struct cxio_rdev *rdev_p, struct t3_cq *cq) @@ -410,6 +410,7 @@ int cxio_flush_sq(struct t3_wq *wq, struct t3_cq *cq, int count) ptr = wq->sq_rptr + count; sqp = wq->sq + Q_PTR2IDX(ptr, wq->sq_size_log2); while (ptr != wq->sq_wptr) { + sqp->signaled = 0; insert_sq_cqe(wq, cq, sqp); ptr++; sqp = wq->sq + Q_PTR2IDX(ptr, wq->sq_size_log2); @@ -450,7 +451,7 @@ static int cqe_completes_wr(struct t3_cqe *cqe, struct t3_wq *wq) if ((CQE_OPCODE(*cqe) == T3_READ_RESP) && SQ_TYPE(*cqe)) return 0; - if ((CQE_OPCODE(*cqe) == T3_SEND) && RQ_TYPE(*cqe) && + if (CQE_SEND_OPCODE(*cqe) && RQ_TYPE(*cqe) && Q_EMPTY(wq->rq_rptr, wq->rq_wptr)) return 0; @@ -571,7 +572,7 @@ static int cxio_hal_init_ctrl_qp(struct cxio_rdev *rdev_p) (unsigned long long) rdev_p->ctrl_qp.dma_addr, rdev_p->ctrl_qp.workq, 1 << T3_CTRL_QP_SIZE_LOG2); skb->priority = CPL_PRIORITY_CONTROL; - return (cxgb3_ofld_send(rdev_p->t3cdev_p, skb)); + return iwch_cxgb3_ofld_send(rdev_p->t3cdev_p, skb); err: kfree_skb(skb); return err; @@ -701,6 +702,9 @@ static int __cxio_tpt_op(struct cxio_rdev *rdev_p, u32 reset_tpt_entry, u32 stag_idx; u32 wptr; + if (cxio_fatal_error(rdev_p)) + return -EIO; + stag_state = stag_state > 0; stag_idx = (*stag) >> 8; @@ -855,7 +859,7 @@ int cxio_rdma_init(struct cxio_rdev *rdev_p, struct t3_rdma_init_attr *attr) wqe->qp_dma_size = cpu_to_be32(attr->qp_dma_size); wqe->irs = cpu_to_be32(attr->irs); skb->priority = 0; /* 0=>ToeQ; 1=>CtrlQ */ - return (cxgb3_ofld_send(rdev_p->t3cdev_p, skb)); + return iwch_cxgb3_ofld_send(rdev_p->t3cdev_p, skb); } void cxio_register_ev_cb(cxio_hal_ev_callback_func_t ev_cb) @@ -938,6 +942,23 @@ int cxio_rdev_open(struct cxio_rdev *rdev_p) if (!rdev_p->t3cdev_p) rdev_p->t3cdev_p = dev2t3cdev(netdev_p); rdev_p->t3cdev_p->ulp = (void *) rdev_p; + + err = rdev_p->t3cdev_p->ctl(rdev_p->t3cdev_p, GET_EMBEDDED_INFO, + &(rdev_p->fw_info)); + if (err) { + printk(KERN_ERR "%s t3cdev_p(%p)->ctl returned error %d.\n", + __func__, rdev_p->t3cdev_p, err); + goto err1; + } + if (G_FW_VERSION_MAJOR(rdev_p->fw_info.fw_vers) != CXIO_FW_MAJ) { + printk(KERN_ERR MOD "fatal firmware version mismatch: " + "need version %u but adapter has version %u\n", + CXIO_FW_MAJ, + G_FW_VERSION_MAJOR(rdev_p->fw_info.fw_vers)); + err = -EINVAL; + goto err1; + } + err = rdev_p->t3cdev_p->ctl(rdev_p->t3cdev_p, RDMA_GET_PARAMS, &(rdev_p->rnic_info)); if (err) { @@ -1021,9 +1042,9 @@ void cxio_rdev_close(struct cxio_rdev *rdev_p) cxio_hal_pblpool_destroy(rdev_p); cxio_hal_rqtpool_destroy(rdev_p); list_del(&rdev_p->entry); - rdev_p->t3cdev_p->ulp = NULL; cxio_hal_destroy_ctrl_qp(rdev_p); cxio_hal_destroy_resource(rdev_p->rscp); + rdev_p->t3cdev_p->ulp = NULL; } } @@ -1204,11 +1225,12 @@ int cxio_poll_cq(struct t3_wq *wq, struct t3_cq *cq, struct t3_cqe *cqe, } /* incoming SEND with no receive posted failures */ - if ((CQE_OPCODE(*hw_cqe) == T3_SEND) && RQ_TYPE(*hw_cqe) && + if (CQE_SEND_OPCODE(*hw_cqe) && RQ_TYPE(*hw_cqe) && Q_EMPTY(wq->rq_rptr, wq->rq_wptr)) { ret = -1; goto skip_cqe; } + BUG_ON((*cqe_flushed == 0) && !SW_CQE(*hw_cqe)); goto proc_cqe; } @@ -1223,6 +1245,13 @@ int cxio_poll_cq(struct t3_wq *wq, struct t3_cq *cq, struct t3_cqe *cqe, * then we complete this with TPT_ERR_MSN and mark the wq in * error. */ + + if (Q_EMPTY(wq->rq_rptr, wq->rq_wptr)) { + wq->error = 1; + ret = -1; + goto skip_cqe; + } + if (unlikely((CQE_WRID_MSN(*hw_cqe) != (wq->rq_rptr + 1)))) { wq->error = 1; hw_cqe->header |= htonl(V_CQE_STATUS(TPT_ERR_MSN)); @@ -1277,6 +1306,7 @@ proc_cqe: cxio_hal_pblpool_free(wq->rdev, wq->rq[Q_PTR2IDX(wq->rq_rptr, wq->rq_size_log2)].pbl_addr, T3_STAG0_PBL_SIZE); + BUG_ON(Q_EMPTY(wq->rq_rptr, wq->rq_wptr)); wq->rq_rptr++; } diff --git a/drivers/infiniband/hw/cxgb3/cxio_hal.h b/drivers/infiniband/hw/cxgb3/cxio_hal.h index 656fe47bc84..bfd03bf8be5 100644 --- a/drivers/infiniband/hw/cxgb3/cxio_hal.h +++ b/drivers/infiniband/hw/cxgb3/cxio_hal.h @@ -61,6 +61,8 @@ #define T3_MAX_DEV_NAME_LEN 32 +#define CXIO_FW_MAJ 7 + struct cxio_hal_ctrl_qp { u32 wptr; u32 rptr; @@ -108,8 +110,16 @@ struct cxio_rdev { struct gen_pool *pbl_pool; struct gen_pool *rqt_pool; struct list_head entry; + struct ch_embedded_info fw_info; + u32 flags; +#define CXIO_ERROR_FATAL 1 }; +static inline int cxio_fatal_error(struct cxio_rdev *rdev_p) +{ + return rdev_p->flags & CXIO_ERROR_FATAL; +} + static inline int cxio_num_stags(struct cxio_rdev *rdev_p) { return min((int)T3_MAX_NUM_STAG, (int)((rdev_p->rnic_info.tpt_top - rdev_p->rnic_info.tpt_base) >> 5)); @@ -183,6 +193,7 @@ void cxio_count_scqes(struct t3_cq *cq, struct t3_wq *wq, int *count); void cxio_flush_hw_cq(struct t3_cq *cq); int cxio_poll_cq(struct t3_wq *wq, struct t3_cq *cq, struct t3_cqe *cqe, u8 *cqe_flushed, u64 *cookie, u32 *credit); +int iwch_cxgb3_ofld_send(struct t3cdev *tdev, struct sk_buff *skb); #define MOD "iw_cxgb3: " #define PDBG(fmt, args...) pr_debug(MOD fmt, ## args) diff --git a/drivers/infiniband/hw/cxgb3/cxio_wr.h b/drivers/infiniband/hw/cxgb3/cxio_wr.h index 04618f7bfbb..32e3b1461d8 100644 --- a/drivers/infiniband/hw/cxgb3/cxio_wr.h +++ b/drivers/infiniband/hw/cxgb3/cxio_wr.h @@ -176,7 +176,7 @@ struct t3_send_wr { struct t3_sge sgl[T3_MAX_SGE]; /* 4+ */ }; -#define T3_MAX_FASTREG_DEPTH 24 +#define T3_MAX_FASTREG_DEPTH 10 #define T3_MAX_FASTREG_FRAG 10 struct t3_fastreg_wr { @@ -604,6 +604,12 @@ struct t3_cqe { #define CQE_STATUS(x) (G_CQE_STATUS(be32_to_cpu((x).header))) #define CQE_OPCODE(x) (G_CQE_OPCODE(be32_to_cpu((x).header))) +#define CQE_SEND_OPCODE(x)( \ + (G_CQE_OPCODE(be32_to_cpu((x).header)) == T3_SEND) || \ + (G_CQE_OPCODE(be32_to_cpu((x).header)) == T3_SEND_WITH_SE) || \ + (G_CQE_OPCODE(be32_to_cpu((x).header)) == T3_SEND_WITH_INV) || \ + (G_CQE_OPCODE(be32_to_cpu((x).header)) == T3_SEND_WITH_SE_INV)) + #define CQE_LEN(x) (be32_to_cpu((x).len)) /* used for RQ completion processing */ diff --git a/drivers/infiniband/hw/cxgb3/iwch.c b/drivers/infiniband/hw/cxgb3/iwch.c index 4489c89d671..26fc0a4eaa7 100644 --- a/drivers/infiniband/hw/cxgb3/iwch.c +++ b/drivers/infiniband/hw/cxgb3/iwch.c @@ -51,13 +51,15 @@ cxgb3_cpl_handler_func t3c_handlers[NUM_CPL_CMDS]; static void open_rnic_dev(struct t3cdev *); static void close_rnic_dev(struct t3cdev *); +static void iwch_err_handler(struct t3cdev *, u32, u32); struct cxgb3_client t3c_client = { .name = "iw_cxgb3", .add = open_rnic_dev, .remove = close_rnic_dev, .handlers = t3c_handlers, - .redirect = iwch_ep_redirect + .redirect = iwch_ep_redirect, + .err_handler = iwch_err_handler }; static LIST_HEAD(dev_list); @@ -160,6 +162,24 @@ static void close_rnic_dev(struct t3cdev *tdev) mutex_unlock(&dev_mutex); } +static void iwch_err_handler(struct t3cdev *tdev, u32 status, u32 error) +{ + struct cxio_rdev *rdev = tdev->ulp; + struct iwch_dev *rnicp = rdev_to_iwch_dev(rdev); + struct ib_event event; + + if (status == OFFLOAD_STATUS_DOWN) { + rdev->flags = CXIO_ERROR_FATAL; + + event.device = &rnicp->ibdev; + event.event = IB_EVENT_DEVICE_FATAL; + event.element.port_num = 0; + ib_dispatch_event(&event); + } + + return; +} + static int __init iwch_init_module(void) { int err; diff --git a/drivers/infiniband/hw/cxgb3/iwch.h b/drivers/infiniband/hw/cxgb3/iwch.h index 3773453b2cf..84735506333 100644 --- a/drivers/infiniband/hw/cxgb3/iwch.h +++ b/drivers/infiniband/hw/cxgb3/iwch.h @@ -117,6 +117,11 @@ static inline struct iwch_dev *to_iwch_dev(struct ib_device *ibdev) return container_of(ibdev, struct iwch_dev, ibdev); } +static inline struct iwch_dev *rdev_to_iwch_dev(struct cxio_rdev *rdev) +{ + return container_of(rdev, struct iwch_dev, rdev); +} + static inline int t3b_device(const struct iwch_dev *rhp) { return rhp->rdev.t3cdev_p->type == T3B; diff --git a/drivers/infiniband/hw/cxgb3/iwch_cm.c b/drivers/infiniband/hw/cxgb3/iwch_cm.c index 44e936e48a3..52d7bb0c2a1 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_cm.c +++ b/drivers/infiniband/hw/cxgb3/iwch_cm.c @@ -139,6 +139,38 @@ static void stop_ep_timer(struct iwch_ep *ep) put_ep(&ep->com); } +int iwch_l2t_send(struct t3cdev *tdev, struct sk_buff *skb, struct l2t_entry *l2e) +{ + int error = 0; + struct cxio_rdev *rdev; + + rdev = (struct cxio_rdev *)tdev->ulp; + if (cxio_fatal_error(rdev)) { + kfree_skb(skb); + return -EIO; + } + error = l2t_send(tdev, skb, l2e); + if (error) + kfree_skb(skb); + return error; +} + +int iwch_cxgb3_ofld_send(struct t3cdev *tdev, struct sk_buff *skb) +{ + int error = 0; + struct cxio_rdev *rdev; + + rdev = (struct cxio_rdev *)tdev->ulp; + if (cxio_fatal_error(rdev)) { + kfree_skb(skb); + return -EIO; + } + error = cxgb3_ofld_send(tdev, skb); + if (error) + kfree_skb(skb); + return error; +} + static void release_tid(struct t3cdev *tdev, u32 hwtid, struct sk_buff *skb) { struct cpl_tid_release *req; @@ -150,7 +182,7 @@ static void release_tid(struct t3cdev *tdev, u32 hwtid, struct sk_buff *skb) req->wr.wr_hi = htonl(V_WR_OP(FW_WROPCODE_FORWARD)); OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_TID_RELEASE, hwtid)); skb->priority = CPL_PRIORITY_SETUP; - cxgb3_ofld_send(tdev, skb); + iwch_cxgb3_ofld_send(tdev, skb); return; } @@ -172,8 +204,7 @@ int iwch_quiesce_tid(struct iwch_ep *ep) req->val = cpu_to_be64(1 << S_TCB_RX_QUIESCE); skb->priority = CPL_PRIORITY_DATA; - cxgb3_ofld_send(ep->com.tdev, skb); - return 0; + return iwch_cxgb3_ofld_send(ep->com.tdev, skb); } int iwch_resume_tid(struct iwch_ep *ep) @@ -194,8 +225,7 @@ int iwch_resume_tid(struct iwch_ep *ep) req->val = 0; skb->priority = CPL_PRIORITY_DATA; - cxgb3_ofld_send(ep->com.tdev, skb); - return 0; + return iwch_cxgb3_ofld_send(ep->com.tdev, skb); } static void set_emss(struct iwch_ep *ep, u16 opt) @@ -252,18 +282,22 @@ static void *alloc_ep(int size, gfp_t gfp) void __free_ep(struct kref *kref) { - struct iwch_ep_common *epc; - epc = container_of(kref, struct iwch_ep_common, kref); - PDBG("%s ep %p state %s\n", __func__, epc, states[state_read(epc)]); - kfree(epc); + struct iwch_ep *ep; + ep = container_of(container_of(kref, struct iwch_ep_common, kref), + struct iwch_ep, com); + PDBG("%s ep %p state %s\n", __func__, ep, states[state_read(&ep->com)]); + if (ep->com.flags & RELEASE_RESOURCES) { + cxgb3_remove_tid(ep->com.tdev, (void *)ep, ep->hwtid); + dst_release(ep->dst); + l2t_release(L2DATA(ep->com.tdev), ep->l2t); + } + kfree(ep); } static void release_ep_resources(struct iwch_ep *ep) { PDBG("%s ep %p tid %d\n", __func__, ep, ep->hwtid); - cxgb3_remove_tid(ep->com.tdev, (void *)ep, ep->hwtid); - dst_release(ep->dst); - l2t_release(L2DATA(ep->com.tdev), ep->l2t); + ep->com.flags |= RELEASE_RESOURCES; put_ep(&ep->com); } @@ -382,7 +416,7 @@ static void abort_arp_failure(struct t3cdev *dev, struct sk_buff *skb) PDBG("%s t3cdev %p\n", __func__, dev); req->cmd = CPL_ABORT_NO_RST; - cxgb3_ofld_send(dev, skb); + iwch_cxgb3_ofld_send(dev, skb); } static int send_halfclose(struct iwch_ep *ep, gfp_t gfp) @@ -402,8 +436,7 @@ static int send_halfclose(struct iwch_ep *ep, gfp_t gfp) req->wr.wr_hi = htonl(V_WR_OP(FW_WROPCODE_OFLD_CLOSE_CON)); req->wr.wr_lo = htonl(V_WR_TID(ep->hwtid)); OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_CLOSE_CON_REQ, ep->hwtid)); - l2t_send(ep->com.tdev, skb, ep->l2t); - return 0; + return iwch_l2t_send(ep->com.tdev, skb, ep->l2t); } static int send_abort(struct iwch_ep *ep, struct sk_buff *skb, gfp_t gfp) @@ -424,8 +457,7 @@ static int send_abort(struct iwch_ep *ep, struct sk_buff *skb, gfp_t gfp) req->wr.wr_lo = htonl(V_WR_TID(ep->hwtid)); OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_ABORT_REQ, ep->hwtid)); req->cmd = CPL_ABORT_SEND_RST; - l2t_send(ep->com.tdev, skb, ep->l2t); - return 0; + return iwch_l2t_send(ep->com.tdev, skb, ep->l2t); } static int send_connect(struct iwch_ep *ep) @@ -469,8 +501,7 @@ static int send_connect(struct iwch_ep *ep) req->opt0l = htonl(opt0l); req->params = 0; req->opt2 = htonl(opt2); - l2t_send(ep->com.tdev, skb, ep->l2t); - return 0; + return iwch_l2t_send(ep->com.tdev, skb, ep->l2t); } static void send_mpa_req(struct iwch_ep *ep, struct sk_buff *skb) @@ -527,7 +558,7 @@ static void send_mpa_req(struct iwch_ep *ep, struct sk_buff *skb) req->sndseq = htonl(ep->snd_seq); BUG_ON(ep->mpa_skb); ep->mpa_skb = skb; - l2t_send(ep->com.tdev, skb, ep->l2t); + iwch_l2t_send(ep->com.tdev, skb, ep->l2t); start_ep_timer(ep); state_set(&ep->com, MPA_REQ_SENT); return; @@ -578,8 +609,7 @@ static int send_mpa_reject(struct iwch_ep *ep, const void *pdata, u8 plen) req->sndseq = htonl(ep->snd_seq); BUG_ON(ep->mpa_skb); ep->mpa_skb = skb; - l2t_send(ep->com.tdev, skb, ep->l2t); - return 0; + return iwch_l2t_send(ep->com.tdev, skb, ep->l2t); } static int send_mpa_reply(struct iwch_ep *ep, const void *pdata, u8 plen) @@ -630,8 +660,7 @@ static int send_mpa_reply(struct iwch_ep *ep, const void *pdata, u8 plen) req->sndseq = htonl(ep->snd_seq); ep->mpa_skb = skb; state_set(&ep->com, MPA_REP_SENT); - l2t_send(ep->com.tdev, skb, ep->l2t); - return 0; + return iwch_l2t_send(ep->com.tdev, skb, ep->l2t); } static int act_establish(struct t3cdev *tdev, struct sk_buff *skb, void *ctx) @@ -795,7 +824,7 @@ static int update_rx_credits(struct iwch_ep *ep, u32 credits) OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_RX_DATA_ACK, ep->hwtid)); req->credit_dack = htonl(V_RX_CREDITS(credits) | V_RX_FORCE_ACK(1)); skb->priority = CPL_PRIORITY_ACK; - cxgb3_ofld_send(ep->com.tdev, skb); + iwch_cxgb3_ofld_send(ep->com.tdev, skb); return credits; } @@ -1127,8 +1156,8 @@ static int abort_rpl(struct t3cdev *tdev, struct sk_buff *skb, void *ctx) * We get 2 abort replies from the HW. The first one must * be ignored except for scribbling that we need one more. */ - if (!(ep->flags & ABORT_REQ_IN_PROGRESS)) { - ep->flags |= ABORT_REQ_IN_PROGRESS; + if (!(ep->com.flags & ABORT_REQ_IN_PROGRESS)) { + ep->com.flags |= ABORT_REQ_IN_PROGRESS; return CPL_RET_BUF_DONE; } @@ -1203,8 +1232,7 @@ static int listen_start(struct iwch_listen_ep *ep) req->opt1 = htonl(V_CONN_POLICY(CPL_CONN_POLICY_ASK)); skb->priority = 1; - cxgb3_ofld_send(ep->com.tdev, skb); - return 0; + return iwch_cxgb3_ofld_send(ep->com.tdev, skb); } static int pass_open_rpl(struct t3cdev *tdev, struct sk_buff *skb, void *ctx) @@ -1237,8 +1265,7 @@ static int listen_stop(struct iwch_listen_ep *ep) req->cpu_idx = 0; OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_CLOSE_LISTSRV_REQ, ep->stid)); skb->priority = 1; - cxgb3_ofld_send(ep->com.tdev, skb); - return 0; + return iwch_cxgb3_ofld_send(ep->com.tdev, skb); } static int close_listsrv_rpl(struct t3cdev *tdev, struct sk_buff *skb, @@ -1286,7 +1313,7 @@ static void accept_cr(struct iwch_ep *ep, __be32 peer_ip, struct sk_buff *skb) rpl->opt2 = htonl(opt2); rpl->rsvd = rpl->opt2; /* workaround for HW bug */ skb->priority = CPL_PRIORITY_SETUP; - l2t_send(ep->com.tdev, skb, ep->l2t); + iwch_l2t_send(ep->com.tdev, skb, ep->l2t); return; } @@ -1315,7 +1342,7 @@ static void reject_cr(struct t3cdev *tdev, u32 hwtid, __be32 peer_ip, rpl->opt0l_status = htonl(CPL_PASS_OPEN_REJECT); rpl->opt2 = 0; rpl->rsvd = rpl->opt2; - cxgb3_ofld_send(tdev, skb); + iwch_cxgb3_ofld_send(tdev, skb); } } @@ -1534,8 +1561,8 @@ static int peer_abort(struct t3cdev *tdev, struct sk_buff *skb, void *ctx) * We get 2 peer aborts from the HW. The first one must * be ignored except for scribbling that we need one more. */ - if (!(ep->flags & PEER_ABORT_IN_PROGRESS)) { - ep->flags |= PEER_ABORT_IN_PROGRESS; + if (!(ep->com.flags & PEER_ABORT_IN_PROGRESS)) { + ep->com.flags |= PEER_ABORT_IN_PROGRESS; return CPL_RET_BUF_DONE; } @@ -1613,7 +1640,7 @@ static int peer_abort(struct t3cdev *tdev, struct sk_buff *skb, void *ctx) rpl->wr.wr_lo = htonl(V_WR_TID(ep->hwtid)); OPCODE_TID(rpl) = htonl(MK_OPCODE_TID(CPL_ABORT_RPL, ep->hwtid)); rpl->cmd = CPL_ABORT_NO_RST; - cxgb3_ofld_send(ep->com.tdev, rpl_skb); + iwch_cxgb3_ofld_send(ep->com.tdev, rpl_skb); out: if (release) release_ep_resources(ep); @@ -1678,6 +1705,9 @@ static int terminate(struct t3cdev *tdev, struct sk_buff *skb, void *ctx) { struct iwch_ep *ep = ctx; + if (state_read(&ep->com) != FPDU_MODE) + return CPL_RET_BUF_DONE; + PDBG("%s ep %p\n", __func__, ep); skb_pull(skb, sizeof(struct cpl_rdma_terminate)); PDBG("%s saving %d bytes of term msg\n", __func__, skb->len); @@ -1800,6 +1830,10 @@ int iwch_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) ep->com.rpl_err = 0; ep->ird = conn_param->ird; ep->ord = conn_param->ord; + + if (peer2peer && ep->ird == 0) + ep->ird = 1; + PDBG("%s %d ird %d ord %d\n", __func__, __LINE__, ep->ird, ep->ord); get_ep(&ep->com); @@ -1885,6 +1919,10 @@ int iwch_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) conn_param->private_data, ep->plen); ep->ird = conn_param->ird; ep->ord = conn_param->ord; + + if (peer2peer && ep->ord == 0) + ep->ord = 1; + ep->com.tdev = h->rdev.t3cdev_p; cm_id->add_ref(cm_id); @@ -2014,8 +2052,11 @@ int iwch_destroy_listen(struct iw_cm_id *cm_id) ep->com.rpl_done = 0; ep->com.rpl_err = 0; err = listen_stop(ep); + if (err) + goto done; wait_event(ep->com.waitq, ep->com.rpl_done); cxgb3_free_stid(ep->com.tdev, ep->stid); +done: err = ep->com.rpl_err; cm_id->rem_ref(cm_id); put_ep(&ep->com); @@ -2027,12 +2068,22 @@ int iwch_ep_disconnect(struct iwch_ep *ep, int abrupt, gfp_t gfp) int ret=0; unsigned long flags; int close = 0; + int fatal = 0; + struct t3cdev *tdev; + struct cxio_rdev *rdev; spin_lock_irqsave(&ep->com.lock, flags); PDBG("%s ep %p state %s, abrupt %d\n", __func__, ep, states[ep->com.state], abrupt); + tdev = (struct t3cdev *)ep->com.tdev; + rdev = (struct cxio_rdev *)tdev->ulp; + if (cxio_fatal_error(rdev)) { + fatal = 1; + close_complete_upcall(ep); + ep->com.state = DEAD; + } switch (ep->com.state) { case MPA_REQ_WAIT: case MPA_REQ_SENT: @@ -2072,7 +2123,11 @@ int iwch_ep_disconnect(struct iwch_ep *ep, int abrupt, gfp_t gfp) ret = send_abort(ep, NULL, gfp); else ret = send_halfclose(ep, gfp); + if (ret) + fatal = 1; } + if (fatal) + release_ep_resources(ep); return ret; } diff --git a/drivers/infiniband/hw/cxgb3/iwch_cm.h b/drivers/infiniband/hw/cxgb3/iwch_cm.h index d7c7e09f099..43c0aea7ead 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_cm.h +++ b/drivers/infiniband/hw/cxgb3/iwch_cm.h @@ -147,6 +147,7 @@ enum iwch_ep_state { enum iwch_ep_flags { PEER_ABORT_IN_PROGRESS = (1 << 0), ABORT_REQ_IN_PROGRESS = (1 << 1), + RELEASE_RESOURCES = (1 << 2), }; struct iwch_ep_common { @@ -161,6 +162,7 @@ struct iwch_ep_common { wait_queue_head_t waitq; int rpl_done; int rpl_err; + u32 flags; }; struct iwch_listen_ep { @@ -188,7 +190,6 @@ struct iwch_ep { u16 plen; u32 ird; u32 ord; - u32 flags; }; static inline struct iwch_ep *to_ep(struct iw_cm_id *cm_id) diff --git a/drivers/infiniband/hw/cxgb3/iwch_ev.c b/drivers/infiniband/hw/cxgb3/iwch_ev.c index 7b67a677172..743c5d8b880 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_ev.c +++ b/drivers/infiniband/hw/cxgb3/iwch_ev.c @@ -179,11 +179,6 @@ void iwch_ev_dispatch(struct cxio_rdev *rdev_p, struct sk_buff *skb) case TPT_ERR_BOUND: case TPT_ERR_INVALIDATE_SHARED_MR: case TPT_ERR_INVALIDATE_MR_WITH_MW_BOUND: - printk(KERN_ERR "%s - CQE Err qpid 0x%x opcode %d status 0x%x " - "type %d wrid.hi 0x%x wrid.lo 0x%x \n", __func__, - CQE_QPID(rsp_msg->cqe), CQE_OPCODE(rsp_msg->cqe), - CQE_STATUS(rsp_msg->cqe), CQE_TYPE(rsp_msg->cqe), - CQE_WRID_HI(rsp_msg->cqe), CQE_WRID_LOW(rsp_msg->cqe)); (*chp->ibcq.comp_handler)(&chp->ibcq, chp->ibcq.cq_context); post_qp_event(rnicp, chp, rsp_msg, IB_EVENT_QP_ACCESS_ERR, 1); break; diff --git a/drivers/infiniband/hw/cxgb3/iwch_provider.c b/drivers/infiniband/hw/cxgb3/iwch_provider.c index 160ef482712..e2a63214008 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_provider.c +++ b/drivers/infiniband/hw/cxgb3/iwch_provider.c @@ -40,6 +40,7 @@ #include <linux/spinlock.h> #include <linux/ethtool.h> #include <linux/rtnetlink.h> +#include <linux/inetdevice.h> #include <asm/io.h> #include <asm/irq.h> @@ -1152,12 +1153,39 @@ static int iwch_query_device(struct ib_device *ibdev, static int iwch_query_port(struct ib_device *ibdev, u8 port, struct ib_port_attr *props) { + struct iwch_dev *dev; + struct net_device *netdev; + struct in_device *inetdev; + PDBG("%s ibdev %p\n", __func__, ibdev); + dev = to_iwch_dev(ibdev); + netdev = dev->rdev.port_info.lldevs[port-1]; + memset(props, 0, sizeof(struct ib_port_attr)); props->max_mtu = IB_MTU_4096; - props->active_mtu = IB_MTU_2048; - props->state = IB_PORT_ACTIVE; + if (netdev->mtu >= 4096) + props->active_mtu = IB_MTU_4096; + else if (netdev->mtu >= 2048) + props->active_mtu = IB_MTU_2048; + else if (netdev->mtu >= 1024) + props->active_mtu = IB_MTU_1024; + else if (netdev->mtu >= 512) + props->active_mtu = IB_MTU_512; + else + props->active_mtu = IB_MTU_256; + + if (!netif_carrier_ok(netdev)) + props->state = IB_PORT_DOWN; + else { + inetdev = in_dev_get(netdev); + if (inetdev->ifa_list) + props->state = IB_PORT_ACTIVE; + else + props->state = IB_PORT_INIT; + in_dev_put(inetdev); + } + props->port_cap_flags = IB_PORT_CM_SUP | IB_PORT_SNMP_TUNNEL_SUP | diff --git a/drivers/infiniband/hw/cxgb3/iwch_qp.c b/drivers/infiniband/hw/cxgb3/iwch_qp.c index 19661b2f040..27bbdc8e773 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_qp.c +++ b/drivers/infiniband/hw/cxgb3/iwch_qp.c @@ -99,8 +99,8 @@ static int build_rdma_write(union t3_wr *wqe, struct ib_send_wr *wr, if (wr->opcode == IB_WR_RDMA_WRITE_WITH_IMM) { plen = 4; wqe->write.sgl[0].stag = wr->ex.imm_data; - wqe->write.sgl[0].len = __constant_cpu_to_be32(0); - wqe->write.num_sgle = __constant_cpu_to_be32(0); + wqe->write.sgl[0].len = cpu_to_be32(0); + wqe->write.num_sgle = cpu_to_be32(0); *flit_cnt = 6; } else { plen = 0; @@ -195,15 +195,12 @@ static int build_inv_stag(union t3_wr *wqe, struct ib_send_wr *wr, return 0; } -/* - * TBD: this is going to be moved to firmware. Missing pdid/qpid check for now. - */ static int iwch_sgl2pbl_map(struct iwch_dev *rhp, struct ib_sge *sg_list, u32 num_sgle, u32 * pbl_addr, u8 * page_size) { int i; struct iwch_mr *mhp; - u32 offset; + u64 offset; for (i = 0; i < num_sgle; i++) { mhp = get_mhp(rhp, (sg_list[i].lkey) >> 8); @@ -235,8 +232,8 @@ static int iwch_sgl2pbl_map(struct iwch_dev *rhp, struct ib_sge *sg_list, return -EINVAL; } offset = sg_list[i].addr - mhp->attr.va_fbo; - offset += ((u32) mhp->attr.va_fbo) % - (1UL << (12 + mhp->attr.page_size)); + offset += mhp->attr.va_fbo & + ((1UL << (12 + mhp->attr.page_size)) - 1); pbl_addr[i] = ((mhp->attr.pbl_addr - rhp->rdev.rnic_info.pbl_base) >> 3) + (offset >> (12 + mhp->attr.page_size)); @@ -266,8 +263,8 @@ static int build_rdma_recv(struct iwch_qp *qhp, union t3_wr *wqe, wqe->recv.sgl[i].len = cpu_to_be32(wr->sg_list[i].length); /* to in the WQE == the offset into the page */ - wqe->recv.sgl[i].to = cpu_to_be64(((u32) wr->sg_list[i].addr) % - (1UL << (12 + page_size[i]))); + wqe->recv.sgl[i].to = cpu_to_be64(((u32)wr->sg_list[i].addr) & + ((1UL << (12 + page_size[i])) - 1)); /* pbl_addr is the adapters address in the PBL */ wqe->recv.pbl_addr[i] = cpu_to_be32(pbl_addr[i]); @@ -754,7 +751,7 @@ int iwch_post_zb_read(struct iwch_qp *qhp) wqe->send.wrh.gen_tid_len = cpu_to_be32(V_FW_RIWR_TID(qhp->ep->hwtid)| V_FW_RIWR_LEN(flit_cnt)); skb->priority = CPL_PRIORITY_DATA; - return cxgb3_ofld_send(qhp->rhp->rdev.t3cdev_p, skb); + return iwch_cxgb3_ofld_send(qhp->rhp->rdev.t3cdev_p, skb); } /* @@ -786,7 +783,7 @@ int iwch_post_terminate(struct iwch_qp *qhp, struct respQ_msg_t *rsp_msg) V_FW_RIWR_FLAGS(T3_COMPLETION_FLAG | T3_NOTIFY_FLAG)); wqe->send.wrh.gen_tid_len = cpu_to_be32(V_FW_RIWR_TID(qhp->ep->hwtid)); skb->priority = CPL_PRIORITY_DATA; - return cxgb3_ofld_send(qhp->rhp->rdev.t3cdev_p, skb); + return iwch_cxgb3_ofld_send(qhp->rhp->rdev.t3cdev_p, skb); } /* @@ -1072,7 +1069,6 @@ int iwch_modify_qp(struct iwch_dev *rhp, struct iwch_qp *qhp, goto out; } qhp->attr.state = IWCH_QP_STATE_IDLE; - memset(&qhp->attr, 0, sizeof(qhp->attr)); break; case IWCH_QP_STATE_TERMINATE: if (!internal) { diff --git a/drivers/infiniband/hw/ehca/ehca_classes_pSeries.h b/drivers/infiniband/hw/ehca/ehca_classes_pSeries.h index 1798e6466bd..689c35786dd 100644 --- a/drivers/infiniband/hw/ehca/ehca_classes_pSeries.h +++ b/drivers/infiniband/hw/ehca/ehca_classes_pSeries.h @@ -165,7 +165,6 @@ struct hcp_modify_qp_control_block { #define MQPCB_MASK_ALT_P_KEY_IDX EHCA_BMASK_IBM( 7, 7) #define MQPCB_MASK_RDMA_ATOMIC_CTRL EHCA_BMASK_IBM( 8, 8) #define MQPCB_MASK_QP_STATE EHCA_BMASK_IBM( 9, 9) -#define MQPCB_QP_STATE EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_RDMA_NR_ATOMIC_RESP_RES EHCA_BMASK_IBM(11, 11) #define MQPCB_MASK_PATH_MIGRATION_STATE EHCA_BMASK_IBM(12, 12) #define MQPCB_MASK_RDMA_ATOMIC_OUTST_DEST_QP EHCA_BMASK_IBM(13, 13) @@ -176,60 +175,33 @@ struct hcp_modify_qp_control_block { #define MQPCB_MASK_RETRY_COUNT EHCA_BMASK_IBM(18, 18) #define MQPCB_MASK_TIMEOUT EHCA_BMASK_IBM(19, 19) #define MQPCB_MASK_PATH_MTU EHCA_BMASK_IBM(20, 20) -#define MQPCB_PATH_MTU EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_MAX_STATIC_RATE EHCA_BMASK_IBM(21, 21) -#define MQPCB_MAX_STATIC_RATE EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_DLID EHCA_BMASK_IBM(22, 22) -#define MQPCB_DLID EHCA_BMASK_IBM(16, 31) #define MQPCB_MASK_RNR_RETRY_COUNT EHCA_BMASK_IBM(23, 23) -#define MQPCB_RNR_RETRY_COUNT EHCA_BMASK_IBM(29, 31) #define MQPCB_MASK_SOURCE_PATH_BITS EHCA_BMASK_IBM(24, 24) -#define MQPCB_SOURCE_PATH_BITS EHCA_BMASK_IBM(25, 31) #define MQPCB_MASK_TRAFFIC_CLASS EHCA_BMASK_IBM(25, 25) -#define MQPCB_TRAFFIC_CLASS EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_HOP_LIMIT EHCA_BMASK_IBM(26, 26) -#define MQPCB_HOP_LIMIT EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_SOURCE_GID_IDX EHCA_BMASK_IBM(27, 27) -#define MQPCB_SOURCE_GID_IDX EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_FLOW_LABEL EHCA_BMASK_IBM(28, 28) -#define MQPCB_FLOW_LABEL EHCA_BMASK_IBM(12, 31) #define MQPCB_MASK_DEST_GID EHCA_BMASK_IBM(30, 30) #define MQPCB_MASK_SERVICE_LEVEL_AL EHCA_BMASK_IBM(31, 31) -#define MQPCB_SERVICE_LEVEL_AL EHCA_BMASK_IBM(28, 31) #define MQPCB_MASK_SEND_GRH_FLAG_AL EHCA_BMASK_IBM(32, 32) -#define MQPCB_SEND_GRH_FLAG_AL EHCA_BMASK_IBM(31, 31) #define MQPCB_MASK_RETRY_COUNT_AL EHCA_BMASK_IBM(33, 33) -#define MQPCB_RETRY_COUNT_AL EHCA_BMASK_IBM(29, 31) #define MQPCB_MASK_TIMEOUT_AL EHCA_BMASK_IBM(34, 34) -#define MQPCB_TIMEOUT_AL EHCA_BMASK_IBM(27, 31) #define MQPCB_MASK_MAX_STATIC_RATE_AL EHCA_BMASK_IBM(35, 35) -#define MQPCB_MAX_STATIC_RATE_AL EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_DLID_AL EHCA_BMASK_IBM(36, 36) -#define MQPCB_DLID_AL EHCA_BMASK_IBM(16, 31) #define MQPCB_MASK_RNR_RETRY_COUNT_AL EHCA_BMASK_IBM(37, 37) -#define MQPCB_RNR_RETRY_COUNT_AL EHCA_BMASK_IBM(29, 31) #define MQPCB_MASK_SOURCE_PATH_BITS_AL EHCA_BMASK_IBM(38, 38) -#define MQPCB_SOURCE_PATH_BITS_AL EHCA_BMASK_IBM(25, 31) #define MQPCB_MASK_TRAFFIC_CLASS_AL EHCA_BMASK_IBM(39, 39) -#define MQPCB_TRAFFIC_CLASS_AL EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_HOP_LIMIT_AL EHCA_BMASK_IBM(40, 40) -#define MQPCB_HOP_LIMIT_AL EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_SOURCE_GID_IDX_AL EHCA_BMASK_IBM(41, 41) -#define MQPCB_SOURCE_GID_IDX_AL EHCA_BMASK_IBM(24, 31) #define MQPCB_MASK_FLOW_LABEL_AL EHCA_BMASK_IBM(42, 42) -#define MQPCB_FLOW_LABEL_AL EHCA_BMASK_IBM(12, 31) #define MQPCB_MASK_DEST_GID_AL EHCA_BMASK_IBM(44, 44) #define MQPCB_MASK_MAX_NR_OUTST_SEND_WR EHCA_BMASK_IBM(45, 45) -#define MQPCB_MAX_NR_OUTST_SEND_WR EHCA_BMASK_IBM(16, 31) #define MQPCB_MASK_MAX_NR_OUTST_RECV_WR EHCA_BMASK_IBM(46, 46) -#define MQPCB_MAX_NR_OUTST_RECV_WR EHCA_BMASK_IBM(16, 31) #define MQPCB_MASK_DISABLE_ETE_CREDIT_CHECK EHCA_BMASK_IBM(47, 47) -#define MQPCB_DISABLE_ETE_CREDIT_CHECK EHCA_BMASK_IBM(31, 31) -#define MQPCB_QP_NUMBER EHCA_BMASK_IBM( 8, 31) #define MQPCB_MASK_QP_ENABLE EHCA_BMASK_IBM(48, 48) -#define MQPCB_QP_ENABLE EHCA_BMASK_IBM(31, 31) #define MQPCB_MASK_CURR_SRQ_LIMIT EHCA_BMASK_IBM(49, 49) -#define MQPCB_CURR_SRQ_LIMIT EHCA_BMASK_IBM(16, 31) #define MQPCB_MASK_QP_AFF_ASYN_EV_LOG_REG EHCA_BMASK_IBM(50, 50) #define MQPCB_MASK_SHARED_RQ_HNDL EHCA_BMASK_IBM(51, 51) diff --git a/drivers/infiniband/hw/ehca/ehca_hca.c b/drivers/infiniband/hw/ehca/ehca_hca.c index 9209c5332df..8b92f85d4dd 100644 --- a/drivers/infiniband/hw/ehca/ehca_hca.c +++ b/drivers/infiniband/hw/ehca/ehca_hca.c @@ -319,7 +319,7 @@ int ehca_query_gid(struct ib_device *ibdev, u8 port, ib_device); struct hipz_query_port *rblock; - if (index > 255) { + if (index < 0 || index > 255) { ehca_err(&shca->ib_device, "Invalid index: %x.", index); return -EINVAL; } diff --git a/drivers/infiniband/hw/ehca/ehca_irq.c b/drivers/infiniband/hw/ehca/ehca_irq.c index 99bcbd7ffb0..4b89b791be6 100644 --- a/drivers/infiniband/hw/ehca/ehca_irq.c +++ b/drivers/infiniband/hw/ehca/ehca_irq.c @@ -479,13 +479,13 @@ void ehca_tasklet_neq(unsigned long data) struct ehca_eqe *eqe; u64 ret; - eqe = (struct ehca_eqe *)ehca_poll_eq(shca, &shca->neq); + eqe = ehca_poll_eq(shca, &shca->neq); while (eqe) { if (!EHCA_BMASK_GET(NEQE_COMPLETION_EVENT, eqe->entry)) parse_ec(shca, eqe->entry); - eqe = (struct ehca_eqe *)ehca_poll_eq(shca, &shca->neq); + eqe = ehca_poll_eq(shca, &shca->neq); } ret = hipz_h_reset_event(shca->ipz_hca_handle, @@ -572,8 +572,7 @@ void ehca_process_eq(struct ehca_shca *shca, int is_irq) eqe_cnt = 0; do { u32 token; - eqe_cache[eqe_cnt].eqe = - (struct ehca_eqe *)ehca_poll_eq(shca, eq); + eqe_cache[eqe_cnt].eqe = ehca_poll_eq(shca, eq); if (!eqe_cache[eqe_cnt].eqe) break; eqe_value = eqe_cache[eqe_cnt].eqe->entry; @@ -637,7 +636,7 @@ void ehca_process_eq(struct ehca_shca *shca, int is_irq) goto unlock_irq_spinlock; do { struct ehca_eqe *eqe; - eqe = (struct ehca_eqe *)ehca_poll_eq(shca, &shca->eq); + eqe = ehca_poll_eq(shca, &shca->eq); if (!eqe) break; process_eqe(shca, eqe); diff --git a/drivers/infiniband/hw/ehca/ehca_main.c b/drivers/infiniband/hw/ehca/ehca_main.c index 368311ce332..fab18a2c74a 100644 --- a/drivers/infiniband/hw/ehca/ehca_main.c +++ b/drivers/infiniband/hw/ehca/ehca_main.c @@ -52,7 +52,7 @@ #include "ehca_tools.h" #include "hcp_if.h" -#define HCAD_VERSION "0026" +#define HCAD_VERSION "0028" MODULE_LICENSE("Dual BSD/GPL"); MODULE_AUTHOR("Christoph Raisch <raisch@de.ibm.com>"); @@ -506,6 +506,7 @@ static int ehca_init_device(struct ehca_shca *shca) shca->ib_device.detach_mcast = ehca_detach_mcast; shca->ib_device.process_mad = ehca_process_mad; shca->ib_device.mmap = ehca_mmap; + shca->ib_device.dma_ops = &ehca_dma_mapping_ops; if (EHCA_BMASK_GET(HCA_CAP_SRQ, shca->hca_cap)) { shca->ib_device.uverbs_cmd_mask |= @@ -636,7 +637,7 @@ static ssize_t ehca_show_##name(struct device *dev, \ struct hipz_query_hca *rblock; \ int data; \ \ - shca = dev->driver_data; \ + shca = dev_get_drvdata(dev); \ \ rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL); \ if (!rblock) { \ @@ -680,7 +681,7 @@ static ssize_t ehca_show_adapter_handle(struct device *dev, struct device_attribute *attr, char *buf) { - struct ehca_shca *shca = dev->driver_data; + struct ehca_shca *shca = dev_get_drvdata(dev); return sprintf(buf, "%llx\n", shca->ipz_hca_handle.handle); @@ -749,7 +750,7 @@ static int __devinit ehca_probe(struct of_device *dev, shca->ofdev = dev; shca->ipz_hca_handle.handle = *handle; - dev->dev.driver_data = shca; + dev_set_drvdata(&dev->dev, shca); ret = ehca_sense_attributes(shca); if (ret < 0) { @@ -878,7 +879,7 @@ probe1: static int __devexit ehca_remove(struct of_device *dev) { - struct ehca_shca *shca = dev->dev.driver_data; + struct ehca_shca *shca = dev_get_drvdata(&dev->dev); unsigned long flags; int ret; @@ -1028,17 +1029,23 @@ static int __init ehca_module_init(void) goto module_init1; } + ret = ehca_create_busmap(); + if (ret) { + ehca_gen_err("Cannot create busmap."); + goto module_init2; + } + ret = ibmebus_register_driver(&ehca_driver); if (ret) { ehca_gen_err("Cannot register eHCA device driver"); ret = -EINVAL; - goto module_init2; + goto module_init3; } ret = register_memory_notifier(&ehca_mem_nb); if (ret) { ehca_gen_err("Failed registering memory add/remove notifier"); - goto module_init3; + goto module_init4; } if (ehca_poll_all_eqs != 1) { @@ -1053,9 +1060,12 @@ static int __init ehca_module_init(void) return 0; -module_init3: +module_init4: ibmebus_unregister_driver(&ehca_driver); +module_init3: + ehca_destroy_busmap(); + module_init2: ehca_destroy_slab_caches(); @@ -1073,6 +1083,8 @@ static void __exit ehca_module_exit(void) unregister_memory_notifier(&ehca_mem_nb); + ehca_destroy_busmap(); + ehca_destroy_slab_caches(); ehca_destroy_comp_pool(); diff --git a/drivers/infiniband/hw/ehca/ehca_mrmw.c b/drivers/infiniband/hw/ehca/ehca_mrmw.c index 72f83f7df61..7663a2a9f13 100644 --- a/drivers/infiniband/hw/ehca/ehca_mrmw.c +++ b/drivers/infiniband/hw/ehca/ehca_mrmw.c @@ -53,6 +53,38 @@ /* max number of rpages (per hcall register_rpages) */ #define MAX_RPAGES 512 +/* DMEM toleration management */ +#define EHCA_SECTSHIFT SECTION_SIZE_BITS +#define EHCA_SECTSIZE (1UL << EHCA_SECTSHIFT) +#define EHCA_HUGEPAGESHIFT 34 +#define EHCA_HUGEPAGE_SIZE (1UL << EHCA_HUGEPAGESHIFT) +#define EHCA_HUGEPAGE_PFN_MASK ((EHCA_HUGEPAGE_SIZE - 1) >> PAGE_SHIFT) +#define EHCA_INVAL_ADDR 0xFFFFFFFFFFFFFFFFULL +#define EHCA_DIR_INDEX_SHIFT 13 /* 8k Entries in 64k block */ +#define EHCA_TOP_INDEX_SHIFT (EHCA_DIR_INDEX_SHIFT * 2) +#define EHCA_MAP_ENTRIES (1 << EHCA_DIR_INDEX_SHIFT) +#define EHCA_TOP_MAP_SIZE (0x10000) /* currently fixed map size */ +#define EHCA_DIR_MAP_SIZE (0x10000) +#define EHCA_ENT_MAP_SIZE (0x10000) +#define EHCA_INDEX_MASK (EHCA_MAP_ENTRIES - 1) + +static unsigned long ehca_mr_len; + +/* + * Memory map data structures + */ +struct ehca_dir_bmap { + u64 ent[EHCA_MAP_ENTRIES]; +}; +struct ehca_top_bmap { + struct ehca_dir_bmap *dir[EHCA_MAP_ENTRIES]; +}; +struct ehca_bmap { + struct ehca_top_bmap *top[EHCA_MAP_ENTRIES]; +}; + +static struct ehca_bmap *ehca_bmap; + static struct kmem_cache *mr_cache; static struct kmem_cache *mw_cache; @@ -68,6 +100,8 @@ enum ehca_mr_pgsize { #define EHCA_MR_PGSHIFT1M 20 #define EHCA_MR_PGSHIFT16M 24 +static u64 ehca_map_vaddr(void *caddr); + static u32 ehca_encode_hwpage_size(u32 pgsize) { int log = ilog2(pgsize); @@ -135,7 +169,8 @@ struct ib_mr *ehca_get_dma_mr(struct ib_pd *pd, int mr_access_flags) goto get_dma_mr_exit0; } - ret = ehca_reg_maxmr(shca, e_maxmr, (u64 *)KERNELBASE, + ret = ehca_reg_maxmr(shca, e_maxmr, + (void *)ehca_map_vaddr((void *)KERNELBASE), mr_access_flags, e_pd, &e_maxmr->ib.ib_mr.lkey, &e_maxmr->ib.ib_mr.rkey); @@ -251,7 +286,7 @@ struct ib_mr *ehca_reg_phys_mr(struct ib_pd *pd, ret = ehca_reg_mr(shca, e_mr, iova_start, size, mr_access_flags, e_pd, &pginfo, &e_mr->ib.ib_mr.lkey, - &e_mr->ib.ib_mr.rkey); + &e_mr->ib.ib_mr.rkey, EHCA_REG_MR); if (ret) { ib_mr = ERR_PTR(ret); goto reg_phys_mr_exit1; @@ -370,7 +405,7 @@ reg_user_mr_fallback: ret = ehca_reg_mr(shca, e_mr, (u64 *)virt, length, mr_access_flags, e_pd, &pginfo, &e_mr->ib.ib_mr.lkey, - &e_mr->ib.ib_mr.rkey); + &e_mr->ib.ib_mr.rkey, EHCA_REG_MR); if (ret == -EINVAL && pginfo.hwpage_size > PAGE_SIZE) { ehca_warn(pd->device, "failed to register mr " "with hwpage_size=%llx", hwpage_size); @@ -794,7 +829,7 @@ struct ib_fmr *ehca_alloc_fmr(struct ib_pd *pd, ret = ehca_reg_mr(shca, e_fmr, NULL, fmr_attr->max_pages * (1 << fmr_attr->page_shift), mr_access_flags, e_pd, &pginfo, - &tmp_lkey, &tmp_rkey); + &tmp_lkey, &tmp_rkey, EHCA_REG_MR); if (ret) { ib_fmr = ERR_PTR(ret); goto alloc_fmr_exit1; @@ -983,6 +1018,10 @@ free_fmr_exit0: /*----------------------------------------------------------------------*/ +static int ehca_reg_bmap_mr_rpages(struct ehca_shca *shca, + struct ehca_mr *e_mr, + struct ehca_mr_pginfo *pginfo); + int ehca_reg_mr(struct ehca_shca *shca, struct ehca_mr *e_mr, u64 *iova_start, @@ -991,7 +1030,8 @@ int ehca_reg_mr(struct ehca_shca *shca, struct ehca_pd *e_pd, struct ehca_mr_pginfo *pginfo, u32 *lkey, /*OUT*/ - u32 *rkey) /*OUT*/ + u32 *rkey, /*OUT*/ + enum ehca_reg_type reg_type) { int ret; u64 h_ret; @@ -1015,7 +1055,13 @@ int ehca_reg_mr(struct ehca_shca *shca, e_mr->ipz_mr_handle = hipzout.handle; - ret = ehca_reg_mr_rpages(shca, e_mr, pginfo); + if (reg_type == EHCA_REG_BUSMAP_MR) + ret = ehca_reg_bmap_mr_rpages(shca, e_mr, pginfo); + else if (reg_type == EHCA_REG_MR) + ret = ehca_reg_mr_rpages(shca, e_mr, pginfo); + else + ret = -EINVAL; + if (ret) goto ehca_reg_mr_exit1; @@ -1316,7 +1362,7 @@ int ehca_rereg_mr(struct ehca_shca *shca, e_mr->fmr_map_cnt = save_mr.fmr_map_cnt; ret = ehca_reg_mr(shca, e_mr, iova_start, size, acl, - e_pd, pginfo, lkey, rkey); + e_pd, pginfo, lkey, rkey, EHCA_REG_MR); if (ret) { u32 offset = (u64)(&e_mr->flags) - (u64)e_mr; memcpy(&e_mr->flags, &(save_mr.flags), @@ -1409,7 +1455,7 @@ int ehca_unmap_one_fmr(struct ehca_shca *shca, ret = ehca_reg_mr(shca, e_fmr, NULL, (e_fmr->fmr_max_pages * e_fmr->fmr_page_size), e_fmr->acl, e_pd, &pginfo, &tmp_lkey, - &tmp_rkey); + &tmp_rkey, EHCA_REG_MR); if (ret) { u32 offset = (u64)(&e_fmr->flags) - (u64)e_fmr; memcpy(&e_fmr->flags, &(save_mr.flags), @@ -1478,6 +1524,90 @@ ehca_reg_smr_exit0: } /* end ehca_reg_smr() */ /*----------------------------------------------------------------------*/ +static inline void *ehca_calc_sectbase(int top, int dir, int idx) +{ + unsigned long ret = idx; + ret |= dir << EHCA_DIR_INDEX_SHIFT; + ret |= top << EHCA_TOP_INDEX_SHIFT; + return abs_to_virt(ret << SECTION_SIZE_BITS); +} + +#define ehca_bmap_valid(entry) \ + ((u64)entry != (u64)EHCA_INVAL_ADDR) + +static u64 ehca_reg_mr_section(int top, int dir, int idx, u64 *kpage, + struct ehca_shca *shca, struct ehca_mr *mr, + struct ehca_mr_pginfo *pginfo) +{ + u64 h_ret = 0; + unsigned long page = 0; + u64 rpage = virt_to_abs(kpage); + int page_count; + + void *sectbase = ehca_calc_sectbase(top, dir, idx); + if ((unsigned long)sectbase & (pginfo->hwpage_size - 1)) { + ehca_err(&shca->ib_device, "reg_mr_section will probably fail:" + "hwpage_size does not fit to " + "section start address"); + } + page_count = EHCA_SECTSIZE / pginfo->hwpage_size; + + while (page < page_count) { + u64 rnum; + for (rnum = 0; (rnum < MAX_RPAGES) && (page < page_count); + rnum++) { + void *pg = sectbase + ((page++) * pginfo->hwpage_size); + kpage[rnum] = virt_to_abs(pg); + } + + h_ret = hipz_h_register_rpage_mr(shca->ipz_hca_handle, mr, + ehca_encode_hwpage_size(pginfo->hwpage_size), + 0, rpage, rnum); + + if ((h_ret != H_SUCCESS) && (h_ret != H_PAGE_REGISTERED)) { + ehca_err(&shca->ib_device, "register_rpage_mr failed"); + return h_ret; + } + } + return h_ret; +} + +static u64 ehca_reg_mr_sections(int top, int dir, u64 *kpage, + struct ehca_shca *shca, struct ehca_mr *mr, + struct ehca_mr_pginfo *pginfo) +{ + u64 hret = H_SUCCESS; + int idx; + + for (idx = 0; idx < EHCA_MAP_ENTRIES; idx++) { + if (!ehca_bmap_valid(ehca_bmap->top[top]->dir[dir]->ent[idx])) + continue; + + hret = ehca_reg_mr_section(top, dir, idx, kpage, shca, mr, + pginfo); + if ((hret != H_SUCCESS) && (hret != H_PAGE_REGISTERED)) + return hret; + } + return hret; +} + +static u64 ehca_reg_mr_dir_sections(int top, u64 *kpage, struct ehca_shca *shca, + struct ehca_mr *mr, + struct ehca_mr_pginfo *pginfo) +{ + u64 hret = H_SUCCESS; + int dir; + + for (dir = 0; dir < EHCA_MAP_ENTRIES; dir++) { + if (!ehca_bmap_valid(ehca_bmap->top[top]->dir[dir])) + continue; + + hret = ehca_reg_mr_sections(top, dir, kpage, shca, mr, pginfo); + if ((hret != H_SUCCESS) && (hret != H_PAGE_REGISTERED)) + return hret; + } + return hret; +} /* register internal max-MR to internal SHCA */ int ehca_reg_internal_maxmr( @@ -1495,6 +1625,11 @@ int ehca_reg_internal_maxmr( u32 num_hwpages; u64 hw_pgsize; + if (!ehca_bmap) { + ret = -EFAULT; + goto ehca_reg_internal_maxmr_exit0; + } + e_mr = ehca_mr_new(); if (!e_mr) { ehca_err(&shca->ib_device, "out of memory"); @@ -1504,8 +1639,8 @@ int ehca_reg_internal_maxmr( e_mr->flags |= EHCA_MR_FLAG_MAXMR; /* register internal max-MR on HCA */ - size_maxmr = (u64)high_memory - PAGE_OFFSET; - iova_start = (u64 *)KERNELBASE; + size_maxmr = ehca_mr_len; + iova_start = (u64 *)ehca_map_vaddr((void *)KERNELBASE); ib_pbuf.addr = 0; ib_pbuf.size = size_maxmr; num_kpages = NUM_CHUNKS(((u64)iova_start % PAGE_SIZE) + size_maxmr, @@ -1524,7 +1659,7 @@ int ehca_reg_internal_maxmr( ret = ehca_reg_mr(shca, e_mr, iova_start, size_maxmr, 0, e_pd, &pginfo, &e_mr->ib.ib_mr.lkey, - &e_mr->ib.ib_mr.rkey); + &e_mr->ib.ib_mr.rkey, EHCA_REG_BUSMAP_MR); if (ret) { ehca_err(&shca->ib_device, "reg of internal max MR failed, " "e_mr=%p iova_start=%p size_maxmr=%llx num_kpages=%x " @@ -2077,8 +2212,8 @@ int ehca_mr_is_maxmr(u64 size, u64 *iova_start) { /* a MR is treated as max-MR only if it fits following: */ - if ((size == ((u64)high_memory - PAGE_OFFSET)) && - (iova_start == (void *)KERNELBASE)) { + if ((size == ehca_mr_len) && + (iova_start == (void *)ehca_map_vaddr((void *)KERNELBASE))) { ehca_gen_dbg("this is a max-MR"); return 1; } else @@ -2184,3 +2319,350 @@ void ehca_cleanup_mrmw_cache(void) if (mw_cache) kmem_cache_destroy(mw_cache); } + +static inline int ehca_init_top_bmap(struct ehca_top_bmap *ehca_top_bmap, + int dir) +{ + if (!ehca_bmap_valid(ehca_top_bmap->dir[dir])) { + ehca_top_bmap->dir[dir] = + kmalloc(sizeof(struct ehca_dir_bmap), GFP_KERNEL); + if (!ehca_top_bmap->dir[dir]) + return -ENOMEM; + /* Set map block to 0xFF according to EHCA_INVAL_ADDR */ + memset(ehca_top_bmap->dir[dir], 0xFF, EHCA_ENT_MAP_SIZE); + } + return 0; +} + +static inline int ehca_init_bmap(struct ehca_bmap *ehca_bmap, int top, int dir) +{ + if (!ehca_bmap_valid(ehca_bmap->top[top])) { + ehca_bmap->top[top] = + kmalloc(sizeof(struct ehca_top_bmap), GFP_KERNEL); + if (!ehca_bmap->top[top]) + return -ENOMEM; + /* Set map block to 0xFF according to EHCA_INVAL_ADDR */ + memset(ehca_bmap->top[top], 0xFF, EHCA_DIR_MAP_SIZE); + } + return ehca_init_top_bmap(ehca_bmap->top[top], dir); +} + +static inline int ehca_calc_index(unsigned long i, unsigned long s) +{ + return (i >> s) & EHCA_INDEX_MASK; +} + +void ehca_destroy_busmap(void) +{ + int top, dir; + + if (!ehca_bmap) + return; + + for (top = 0; top < EHCA_MAP_ENTRIES; top++) { + if (!ehca_bmap_valid(ehca_bmap->top[top])) + continue; + for (dir = 0; dir < EHCA_MAP_ENTRIES; dir++) { + if (!ehca_bmap_valid(ehca_bmap->top[top]->dir[dir])) + continue; + + kfree(ehca_bmap->top[top]->dir[dir]); + } + + kfree(ehca_bmap->top[top]); + } + + kfree(ehca_bmap); + ehca_bmap = NULL; +} + +static int ehca_update_busmap(unsigned long pfn, unsigned long nr_pages) +{ + unsigned long i, start_section, end_section; + int top, dir, idx; + + if (!nr_pages) + return 0; + + if (!ehca_bmap) { + ehca_bmap = kmalloc(sizeof(struct ehca_bmap), GFP_KERNEL); + if (!ehca_bmap) + return -ENOMEM; + /* Set map block to 0xFF according to EHCA_INVAL_ADDR */ + memset(ehca_bmap, 0xFF, EHCA_TOP_MAP_SIZE); + } + + start_section = phys_to_abs(pfn * PAGE_SIZE) / EHCA_SECTSIZE; + end_section = phys_to_abs((pfn + nr_pages) * PAGE_SIZE) / EHCA_SECTSIZE; + for (i = start_section; i < end_section; i++) { + int ret; + top = ehca_calc_index(i, EHCA_TOP_INDEX_SHIFT); + dir = ehca_calc_index(i, EHCA_DIR_INDEX_SHIFT); + idx = i & EHCA_INDEX_MASK; + + ret = ehca_init_bmap(ehca_bmap, top, dir); + if (ret) { + ehca_destroy_busmap(); + return ret; + } + ehca_bmap->top[top]->dir[dir]->ent[idx] = ehca_mr_len; + ehca_mr_len += EHCA_SECTSIZE; + } + return 0; +} + +static int ehca_is_hugepage(unsigned long pfn) +{ + int page_order; + + if (pfn & EHCA_HUGEPAGE_PFN_MASK) + return 0; + + page_order = compound_order(pfn_to_page(pfn)); + if (page_order + PAGE_SHIFT != EHCA_HUGEPAGESHIFT) + return 0; + + return 1; +} + +static int ehca_create_busmap_callback(unsigned long initial_pfn, + unsigned long total_nr_pages, void *arg) +{ + int ret; + unsigned long pfn, start_pfn, end_pfn, nr_pages; + + if ((total_nr_pages * PAGE_SIZE) < EHCA_HUGEPAGE_SIZE) + return ehca_update_busmap(initial_pfn, total_nr_pages); + + /* Given chunk is >= 16GB -> check for hugepages */ + start_pfn = initial_pfn; + end_pfn = initial_pfn + total_nr_pages; + pfn = start_pfn; + + while (pfn < end_pfn) { + if (ehca_is_hugepage(pfn)) { + /* Add mem found in front of the hugepage */ + nr_pages = pfn - start_pfn; + ret = ehca_update_busmap(start_pfn, nr_pages); + if (ret) + return ret; + /* Skip the hugepage */ + pfn += (EHCA_HUGEPAGE_SIZE / PAGE_SIZE); + start_pfn = pfn; + } else + pfn += (EHCA_SECTSIZE / PAGE_SIZE); + } + + /* Add mem found behind the hugepage(s) */ + nr_pages = pfn - start_pfn; + return ehca_update_busmap(start_pfn, nr_pages); +} + +int ehca_create_busmap(void) +{ + int ret; + + ehca_mr_len = 0; + ret = walk_memory_resource(0, 1ULL << MAX_PHYSMEM_BITS, NULL, + ehca_create_busmap_callback); + return ret; +} + +static int ehca_reg_bmap_mr_rpages(struct ehca_shca *shca, + struct ehca_mr *e_mr, + struct ehca_mr_pginfo *pginfo) +{ + int top; + u64 hret, *kpage; + + kpage = ehca_alloc_fw_ctrlblock(GFP_KERNEL); + if (!kpage) { + ehca_err(&shca->ib_device, "kpage alloc failed"); + return -ENOMEM; + } + for (top = 0; top < EHCA_MAP_ENTRIES; top++) { + if (!ehca_bmap_valid(ehca_bmap->top[top])) + continue; + hret = ehca_reg_mr_dir_sections(top, kpage, shca, e_mr, pginfo); + if ((hret != H_PAGE_REGISTERED) && (hret != H_SUCCESS)) + break; + } + + ehca_free_fw_ctrlblock(kpage); + + if (hret == H_SUCCESS) + return 0; /* Everything is fine */ + else { + ehca_err(&shca->ib_device, "ehca_reg_bmap_mr_rpages failed, " + "h_ret=%lli e_mr=%p top=%x lkey=%x " + "hca_hndl=%llx mr_hndl=%llx", hret, e_mr, top, + e_mr->ib.ib_mr.lkey, + shca->ipz_hca_handle.handle, + e_mr->ipz_mr_handle.handle); + return ehca2ib_return_code(hret); + } +} + +static u64 ehca_map_vaddr(void *caddr) +{ + int top, dir, idx; + unsigned long abs_addr, offset; + u64 entry; + + if (!ehca_bmap) + return EHCA_INVAL_ADDR; + + abs_addr = virt_to_abs(caddr); + top = ehca_calc_index(abs_addr, EHCA_TOP_INDEX_SHIFT + EHCA_SECTSHIFT); + if (!ehca_bmap_valid(ehca_bmap->top[top])) + return EHCA_INVAL_ADDR; + + dir = ehca_calc_index(abs_addr, EHCA_DIR_INDEX_SHIFT + EHCA_SECTSHIFT); + if (!ehca_bmap_valid(ehca_bmap->top[top]->dir[dir])) + return EHCA_INVAL_ADDR; + + idx = ehca_calc_index(abs_addr, EHCA_SECTSHIFT); + + entry = ehca_bmap->top[top]->dir[dir]->ent[idx]; + if (ehca_bmap_valid(entry)) { + offset = (unsigned long)caddr & (EHCA_SECTSIZE - 1); + return entry | offset; + } else + return EHCA_INVAL_ADDR; +} + +static int ehca_dma_mapping_error(struct ib_device *dev, u64 dma_addr) +{ + return dma_addr == EHCA_INVAL_ADDR; +} + +static u64 ehca_dma_map_single(struct ib_device *dev, void *cpu_addr, + size_t size, enum dma_data_direction direction) +{ + if (cpu_addr) + return ehca_map_vaddr(cpu_addr); + else + return EHCA_INVAL_ADDR; +} + +static void ehca_dma_unmap_single(struct ib_device *dev, u64 addr, size_t size, + enum dma_data_direction direction) +{ + /* This is only a stub; nothing to be done here */ +} + +static u64 ehca_dma_map_page(struct ib_device *dev, struct page *page, + unsigned long offset, size_t size, + enum dma_data_direction direction) +{ + u64 addr; + + if (offset + size > PAGE_SIZE) + return EHCA_INVAL_ADDR; + + addr = ehca_map_vaddr(page_address(page)); + if (!ehca_dma_mapping_error(dev, addr)) + addr += offset; + + return addr; +} + +static void ehca_dma_unmap_page(struct ib_device *dev, u64 addr, size_t size, + enum dma_data_direction direction) +{ + /* This is only a stub; nothing to be done here */ +} + +static int ehca_dma_map_sg(struct ib_device *dev, struct scatterlist *sgl, + int nents, enum dma_data_direction direction) +{ + struct scatterlist *sg; + int i; + + for_each_sg(sgl, sg, nents, i) { + u64 addr; + addr = ehca_map_vaddr(sg_virt(sg)); + if (ehca_dma_mapping_error(dev, addr)) + return 0; + + sg->dma_address = addr; + sg->dma_length = sg->length; + } + return nents; +} + +static void ehca_dma_unmap_sg(struct ib_device *dev, struct scatterlist *sg, + int nents, enum dma_data_direction direction) +{ + /* This is only a stub; nothing to be done here */ +} + +static u64 ehca_dma_address(struct ib_device *dev, struct scatterlist *sg) +{ + return sg->dma_address; +} + +static unsigned int ehca_dma_len(struct ib_device *dev, struct scatterlist *sg) +{ + return sg->length; +} + +static void ehca_dma_sync_single_for_cpu(struct ib_device *dev, u64 addr, + size_t size, + enum dma_data_direction dir) +{ + dma_sync_single_for_cpu(dev->dma_device, addr, size, dir); +} + +static void ehca_dma_sync_single_for_device(struct ib_device *dev, u64 addr, + size_t size, + enum dma_data_direction dir) +{ + dma_sync_single_for_device(dev->dma_device, addr, size, dir); +} + +static void *ehca_dma_alloc_coherent(struct ib_device *dev, size_t size, + u64 *dma_handle, gfp_t flag) +{ + struct page *p; + void *addr = NULL; + u64 dma_addr; + + p = alloc_pages(flag, get_order(size)); + if (p) { + addr = page_address(p); + dma_addr = ehca_map_vaddr(addr); + if (ehca_dma_mapping_error(dev, dma_addr)) { + free_pages((unsigned long)addr, get_order(size)); + return NULL; + } + if (dma_handle) + *dma_handle = dma_addr; + return addr; + } + return NULL; +} + +static void ehca_dma_free_coherent(struct ib_device *dev, size_t size, + void *cpu_addr, u64 dma_handle) +{ + if (cpu_addr && size) + free_pages((unsigned long)cpu_addr, get_order(size)); +} + + +struct ib_dma_mapping_ops ehca_dma_mapping_ops = { + .mapping_error = ehca_dma_mapping_error, + .map_single = ehca_dma_map_single, + .unmap_single = ehca_dma_unmap_single, + .map_page = ehca_dma_map_page, + .unmap_page = ehca_dma_unmap_page, + .map_sg = ehca_dma_map_sg, + .unmap_sg = ehca_dma_unmap_sg, + .dma_address = ehca_dma_address, + .dma_len = ehca_dma_len, + .sync_single_for_cpu = ehca_dma_sync_single_for_cpu, + .sync_single_for_device = ehca_dma_sync_single_for_device, + .alloc_coherent = ehca_dma_alloc_coherent, + .free_coherent = ehca_dma_free_coherent, +}; diff --git a/drivers/infiniband/hw/ehca/ehca_mrmw.h b/drivers/infiniband/hw/ehca/ehca_mrmw.h index bc8f4e31c12..50d8b51306d 100644 --- a/drivers/infiniband/hw/ehca/ehca_mrmw.h +++ b/drivers/infiniband/hw/ehca/ehca_mrmw.h @@ -42,6 +42,11 @@ #ifndef _EHCA_MRMW_H_ #define _EHCA_MRMW_H_ +enum ehca_reg_type { + EHCA_REG_MR, + EHCA_REG_BUSMAP_MR +}; + int ehca_reg_mr(struct ehca_shca *shca, struct ehca_mr *e_mr, u64 *iova_start, @@ -50,7 +55,8 @@ int ehca_reg_mr(struct ehca_shca *shca, struct ehca_pd *e_pd, struct ehca_mr_pginfo *pginfo, u32 *lkey, - u32 *rkey); + u32 *rkey, + enum ehca_reg_type reg_type); int ehca_reg_mr_rpages(struct ehca_shca *shca, struct ehca_mr *e_mr, @@ -118,4 +124,9 @@ void ehca_mrmw_reverse_map_acl(const u32 *hipz_acl, void ehca_mr_deletenew(struct ehca_mr *mr); +int ehca_create_busmap(void); + +void ehca_destroy_busmap(void); + +extern struct ib_dma_mapping_ops ehca_dma_mapping_ops; #endif /*_EHCA_MRMW_H_*/ diff --git a/drivers/infiniband/hw/ehca/ehca_qp.c b/drivers/infiniband/hw/ehca/ehca_qp.c index 00c10815971..0338f1fabe8 100644 --- a/drivers/infiniband/hw/ehca/ehca_qp.c +++ b/drivers/infiniband/hw/ehca/ehca_qp.c @@ -461,7 +461,7 @@ static struct ehca_qp *internal_create_qp( ib_device); struct ib_ucontext *context = NULL; u64 h_ret; - int is_llqp = 0, has_srq = 0; + int is_llqp = 0, has_srq = 0, is_user = 0; int qp_type, max_send_sge, max_recv_sge, ret; /* h_call's out parameters */ @@ -609,9 +609,6 @@ static struct ehca_qp *internal_create_qp( } } - if (pd->uobject && udata) - context = pd->uobject->context; - my_qp = kmem_cache_zalloc(qp_cache, GFP_KERNEL); if (!my_qp) { ehca_err(pd->device, "pd=%p not enough memory to alloc qp", pd); @@ -619,6 +616,11 @@ static struct ehca_qp *internal_create_qp( return ERR_PTR(-ENOMEM); } + if (pd->uobject && udata) { + is_user = 1; + context = pd->uobject->context; + } + atomic_set(&my_qp->nr_events, 0); init_waitqueue_head(&my_qp->wait_completion); spin_lock_init(&my_qp->spinlock_s); @@ -707,7 +709,7 @@ static struct ehca_qp *internal_create_qp( (parms.squeue.is_small || parms.rqueue.is_small); } - h_ret = hipz_h_alloc_resource_qp(shca->ipz_hca_handle, &parms); + h_ret = hipz_h_alloc_resource_qp(shca->ipz_hca_handle, &parms, is_user); if (h_ret != H_SUCCESS) { ehca_err(pd->device, "h_alloc_resource_qp() failed h_ret=%lli", h_ret); @@ -769,18 +771,20 @@ static struct ehca_qp *internal_create_qp( goto create_qp_exit2; } - my_qp->sq_map.entries = my_qp->ipz_squeue.queue_length / - my_qp->ipz_squeue.qe_size; - my_qp->sq_map.map = vmalloc(my_qp->sq_map.entries * - sizeof(struct ehca_qmap_entry)); - if (!my_qp->sq_map.map) { - ehca_err(pd->device, "Couldn't allocate squeue " - "map ret=%i", ret); - goto create_qp_exit3; + if (!is_user) { + my_qp->sq_map.entries = my_qp->ipz_squeue.queue_length / + my_qp->ipz_squeue.qe_size; + my_qp->sq_map.map = vmalloc(my_qp->sq_map.entries * + sizeof(struct ehca_qmap_entry)); + if (!my_qp->sq_map.map) { + ehca_err(pd->device, "Couldn't allocate squeue " + "map ret=%i", ret); + goto create_qp_exit3; + } + INIT_LIST_HEAD(&my_qp->sq_err_node); + /* to avoid the generation of bogus flush CQEs */ + reset_queue_map(&my_qp->sq_map); } - INIT_LIST_HEAD(&my_qp->sq_err_node); - /* to avoid the generation of bogus flush CQEs */ - reset_queue_map(&my_qp->sq_map); } if (HAS_RQ(my_qp)) { @@ -792,20 +796,21 @@ static struct ehca_qp *internal_create_qp( "and pages ret=%i", ret); goto create_qp_exit4; } - - my_qp->rq_map.entries = my_qp->ipz_rqueue.queue_length / - my_qp->ipz_rqueue.qe_size; - my_qp->rq_map.map = vmalloc(my_qp->rq_map.entries * - sizeof(struct ehca_qmap_entry)); - if (!my_qp->rq_map.map) { - ehca_err(pd->device, "Couldn't allocate squeue " - "map ret=%i", ret); - goto create_qp_exit5; + if (!is_user) { + my_qp->rq_map.entries = my_qp->ipz_rqueue.queue_length / + my_qp->ipz_rqueue.qe_size; + my_qp->rq_map.map = vmalloc(my_qp->rq_map.entries * + sizeof(struct ehca_qmap_entry)); + if (!my_qp->rq_map.map) { + ehca_err(pd->device, "Couldn't allocate squeue " + "map ret=%i", ret); + goto create_qp_exit5; + } + INIT_LIST_HEAD(&my_qp->rq_err_node); + /* to avoid the generation of bogus flush CQEs */ + reset_queue_map(&my_qp->rq_map); } - INIT_LIST_HEAD(&my_qp->rq_err_node); - /* to avoid the generation of bogus flush CQEs */ - reset_queue_map(&my_qp->rq_map); - } else if (init_attr->srq) { + } else if (init_attr->srq && !is_user) { /* this is a base QP, use the queue map of the SRQ */ my_qp->rq_map = my_srq->rq_map; INIT_LIST_HEAD(&my_qp->rq_err_node); @@ -918,7 +923,7 @@ create_qp_exit7: kfree(my_qp->mod_qp_parm); create_qp_exit6: - if (HAS_RQ(my_qp)) + if (HAS_RQ(my_qp) && !is_user) vfree(my_qp->rq_map.map); create_qp_exit5: @@ -926,7 +931,7 @@ create_qp_exit5: ipz_queue_dtor(my_pd, &my_qp->ipz_rqueue); create_qp_exit4: - if (HAS_SQ(my_qp)) + if (HAS_SQ(my_qp) && !is_user) vfree(my_qp->sq_map.map); create_qp_exit3: @@ -1244,6 +1249,7 @@ static int internal_modify_qp(struct ib_qp *ibqp, u64 update_mask; u64 h_ret; int bad_wqe_cnt = 0; + int is_user = 0; int squeue_locked = 0; unsigned long flags = 0; @@ -1266,6 +1272,8 @@ static int internal_modify_qp(struct ib_qp *ibqp, ret = ehca2ib_return_code(h_ret); goto modify_qp_exit1; } + if (ibqp->uobject) + is_user = 1; qp_cur_state = ehca2ib_qp_state(mqpcb->qp_state); @@ -1728,7 +1736,8 @@ static int internal_modify_qp(struct ib_qp *ibqp, goto modify_qp_exit2; } } - if ((qp_new_state == IB_QPS_ERR) && (qp_cur_state != IB_QPS_ERR)) { + if ((qp_new_state == IB_QPS_ERR) && (qp_cur_state != IB_QPS_ERR) + && !is_user) { ret = check_for_left_cqes(my_qp, shca); if (ret) goto modify_qp_exit2; @@ -1738,16 +1747,17 @@ static int internal_modify_qp(struct ib_qp *ibqp, ipz_qeit_reset(&my_qp->ipz_rqueue); ipz_qeit_reset(&my_qp->ipz_squeue); - if (qp_cur_state == IB_QPS_ERR) { + if (qp_cur_state == IB_QPS_ERR && !is_user) { del_from_err_list(my_qp->send_cq, &my_qp->sq_err_node); if (HAS_RQ(my_qp)) del_from_err_list(my_qp->recv_cq, &my_qp->rq_err_node); } - reset_queue_map(&my_qp->sq_map); + if (!is_user) + reset_queue_map(&my_qp->sq_map); - if (HAS_RQ(my_qp)) + if (HAS_RQ(my_qp) && !is_user) reset_queue_map(&my_qp->rq_map); } @@ -1952,19 +1962,13 @@ int ehca_query_qp(struct ib_qp *qp, qp_attr->cap.max_inline_data = my_qp->sq_max_inline_data_size; qp_attr->dest_qp_num = qpcb->dest_qp_nr; - qp_attr->pkey_index = - EHCA_BMASK_GET(MQPCB_PRIM_P_KEY_IDX, qpcb->prim_p_key_idx); - - qp_attr->port_num = - EHCA_BMASK_GET(MQPCB_PRIM_PHYS_PORT, qpcb->prim_phys_port); - + qp_attr->pkey_index = qpcb->prim_p_key_idx; + qp_attr->port_num = qpcb->prim_phys_port; qp_attr->timeout = qpcb->timeout; qp_attr->retry_cnt = qpcb->retry_count; qp_attr->rnr_retry = qpcb->rnr_retry_count; - qp_attr->alt_pkey_index = - EHCA_BMASK_GET(MQPCB_PRIM_P_KEY_IDX, qpcb->alt_p_key_idx); - + qp_attr->alt_pkey_index = qpcb->alt_p_key_idx; qp_attr->alt_port_num = qpcb->alt_phys_port; qp_attr->alt_timeout = qpcb->timeout_al; @@ -2051,8 +2055,7 @@ int ehca_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr, update_mask |= EHCA_BMASK_SET(MQPCB_MASK_CURR_SRQ_LIMIT, 1) | EHCA_BMASK_SET(MQPCB_MASK_QP_AFF_ASYN_EV_LOG_REG, 1); - mqpcb->curr_srq_limit = - EHCA_BMASK_SET(MQPCB_CURR_SRQ_LIMIT, attr->srq_limit); + mqpcb->curr_srq_limit = attr->srq_limit; mqpcb->qp_aff_asyn_ev_log_reg = EHCA_BMASK_SET(QPX_AAELOG_RESET_SRQ_LIMIT, 1); } @@ -2115,8 +2118,7 @@ int ehca_query_srq(struct ib_srq *srq, struct ib_srq_attr *srq_attr) srq_attr->max_wr = qpcb->max_nr_outst_recv_wr - 1; srq_attr->max_sge = 3; - srq_attr->srq_limit = EHCA_BMASK_GET( - MQPCB_CURR_SRQ_LIMIT, qpcb->curr_srq_limit); + srq_attr->srq_limit = qpcb->curr_srq_limit; if (ehca_debug_level >= 2) ehca_dmp(qpcb, 4*70, "qp_num=%x", my_qp->real_qp_num); @@ -2138,10 +2140,12 @@ static int internal_destroy_qp(struct ib_device *dev, struct ehca_qp *my_qp, int ret; u64 h_ret; u8 port_num; + int is_user = 0; enum ib_qp_type qp_type; unsigned long flags; if (uobject) { + is_user = 1; if (my_qp->mm_count_galpa || my_qp->mm_count_rqueue || my_qp->mm_count_squeue) { ehca_err(dev, "Resources still referenced in " @@ -2168,10 +2172,10 @@ static int internal_destroy_qp(struct ib_device *dev, struct ehca_qp *my_qp, * SRQs will never get into an error list and do not have a recv_cq, * so we need to skip them here. */ - if (HAS_RQ(my_qp) && !IS_SRQ(my_qp)) + if (HAS_RQ(my_qp) && !IS_SRQ(my_qp) && !is_user) del_from_err_list(my_qp->recv_cq, &my_qp->rq_err_node); - if (HAS_SQ(my_qp)) + if (HAS_SQ(my_qp) && !is_user) del_from_err_list(my_qp->send_cq, &my_qp->sq_err_node); /* now wait until all pending events have completed */ @@ -2209,13 +2213,13 @@ static int internal_destroy_qp(struct ib_device *dev, struct ehca_qp *my_qp, if (HAS_RQ(my_qp)) { ipz_queue_dtor(my_pd, &my_qp->ipz_rqueue); - - vfree(my_qp->rq_map.map); + if (!is_user) + vfree(my_qp->rq_map.map); } if (HAS_SQ(my_qp)) { ipz_queue_dtor(my_pd, &my_qp->ipz_squeue); - - vfree(my_qp->sq_map.map); + if (!is_user) + vfree(my_qp->sq_map.map); } kmem_cache_free(qp_cache, my_qp); atomic_dec(&shca->num_qps); diff --git a/drivers/infiniband/hw/ehca/ehca_sqp.c b/drivers/infiniband/hw/ehca/ehca_sqp.c index 44447aaa550..c568b28f4e2 100644 --- a/drivers/infiniband/hw/ehca/ehca_sqp.c +++ b/drivers/infiniband/hw/ehca/ehca_sqp.c @@ -46,11 +46,11 @@ #include "ehca_iverbs.h" #include "hcp_if.h" -#define IB_MAD_STATUS_REDIRECT __constant_htons(0x0002) -#define IB_MAD_STATUS_UNSUP_VERSION __constant_htons(0x0004) -#define IB_MAD_STATUS_UNSUP_METHOD __constant_htons(0x0008) +#define IB_MAD_STATUS_REDIRECT cpu_to_be16(0x0002) +#define IB_MAD_STATUS_UNSUP_VERSION cpu_to_be16(0x0004) +#define IB_MAD_STATUS_UNSUP_METHOD cpu_to_be16(0x0008) -#define IB_PMA_CLASS_PORT_INFO __constant_htons(0x0001) +#define IB_PMA_CLASS_PORT_INFO cpu_to_be16(0x0001) /** * ehca_define_sqp - Defines special queue pair 1 (GSI QP). When special queue diff --git a/drivers/infiniband/hw/ehca/hcp_if.c b/drivers/infiniband/hw/ehca/hcp_if.c index d0ab0c0d5e9..4d5dc3304d4 100644 --- a/drivers/infiniband/hw/ehca/hcp_if.c +++ b/drivers/infiniband/hw/ehca/hcp_if.c @@ -284,7 +284,7 @@ u64 hipz_h_alloc_resource_cq(const struct ipz_adapter_handle adapter_handle, param->act_pages = (u32)outs[4]; if (ret == H_SUCCESS) - hcp_galpas_ctor(&cq->galpas, outs[5], outs[6]); + hcp_galpas_ctor(&cq->galpas, 0, outs[5], outs[6]); if (ret == H_NOT_ENOUGH_RESOURCES) ehca_gen_err("Not enough resources. ret=%lli", ret); @@ -293,7 +293,7 @@ u64 hipz_h_alloc_resource_cq(const struct ipz_adapter_handle adapter_handle, } u64 hipz_h_alloc_resource_qp(const struct ipz_adapter_handle adapter_handle, - struct ehca_alloc_qp_parms *parms) + struct ehca_alloc_qp_parms *parms, int is_user) { u64 ret; u64 allocate_controls, max_r10_reg, r11, r12; @@ -359,7 +359,7 @@ u64 hipz_h_alloc_resource_qp(const struct ipz_adapter_handle adapter_handle, (u32)EHCA_BMASK_GET(H_ALL_RES_QP_RQUEUE_SIZE_PAGES, outs[4]); if (ret == H_SUCCESS) - hcp_galpas_ctor(&parms->galpas, outs[6], outs[6]); + hcp_galpas_ctor(&parms->galpas, is_user, outs[6], outs[6]); if (ret == H_NOT_ENOUGH_RESOURCES) ehca_gen_err("Not enough resources. ret=%lli", ret); diff --git a/drivers/infiniband/hw/ehca/hcp_if.h b/drivers/infiniband/hw/ehca/hcp_if.h index 2c3c6e0ea5c..39c1c3618ec 100644 --- a/drivers/infiniband/hw/ehca/hcp_if.h +++ b/drivers/infiniband/hw/ehca/hcp_if.h @@ -78,7 +78,7 @@ u64 hipz_h_alloc_resource_cq(const struct ipz_adapter_handle adapter_handle, * initialize resources, create empty QPPTs (2 rings). */ u64 hipz_h_alloc_resource_qp(const struct ipz_adapter_handle adapter_handle, - struct ehca_alloc_qp_parms *parms); + struct ehca_alloc_qp_parms *parms, int is_user); u64 hipz_h_query_port(const struct ipz_adapter_handle adapter_handle, const u8 port_id, diff --git a/drivers/infiniband/hw/ehca/hcp_phyp.c b/drivers/infiniband/hw/ehca/hcp_phyp.c index 214821095cb..b3e0e72e8a7 100644 --- a/drivers/infiniband/hw/ehca/hcp_phyp.c +++ b/drivers/infiniband/hw/ehca/hcp_phyp.c @@ -54,12 +54,15 @@ int hcall_unmap_page(u64 mapaddr) return 0; } -int hcp_galpas_ctor(struct h_galpas *galpas, +int hcp_galpas_ctor(struct h_galpas *galpas, int is_user, u64 paddr_kernel, u64 paddr_user) { - int ret = hcall_map_page(paddr_kernel, &galpas->kernel.fw_handle); - if (ret) - return ret; + if (!is_user) { + int ret = hcall_map_page(paddr_kernel, &galpas->kernel.fw_handle); + if (ret) + return ret; + } else + galpas->kernel.fw_handle = 0; galpas->user.fw_handle = paddr_user; diff --git a/drivers/infiniband/hw/ehca/hcp_phyp.h b/drivers/infiniband/hw/ehca/hcp_phyp.h index 5305c2a3ed9..204227d5303 100644 --- a/drivers/infiniband/hw/ehca/hcp_phyp.h +++ b/drivers/infiniband/hw/ehca/hcp_phyp.h @@ -78,7 +78,7 @@ static inline void hipz_galpa_store(struct h_galpa galpa, u32 offset, u64 value) *(volatile u64 __force *)addr = value; } -int hcp_galpas_ctor(struct h_galpas *galpas, +int hcp_galpas_ctor(struct h_galpas *galpas, int is_user, u64 paddr_kernel, u64 paddr_user); int hcp_galpas_dtor(struct h_galpas *galpas); diff --git a/drivers/infiniband/hw/ehca/ipz_pt_fn.c b/drivers/infiniband/hw/ehca/ipz_pt_fn.c index c3a32846543..1227c593627 100644 --- a/drivers/infiniband/hw/ehca/ipz_pt_fn.c +++ b/drivers/infiniband/hw/ehca/ipz_pt_fn.c @@ -220,10 +220,13 @@ int ipz_queue_ctor(struct ehca_pd *pd, struct ipz_queue *queue, queue->small_page = NULL; /* allocate queue page pointers */ - queue->queue_pages = vmalloc(nr_of_pages * sizeof(void *)); + queue->queue_pages = kmalloc(nr_of_pages * sizeof(void *), GFP_KERNEL); if (!queue->queue_pages) { - ehca_gen_err("Couldn't allocate queue page list"); - return 0; + queue->queue_pages = vmalloc(nr_of_pages * sizeof(void *)); + if (!queue->queue_pages) { + ehca_gen_err("Couldn't allocate queue page list"); + return 0; + } } memset(queue->queue_pages, 0, nr_of_pages * sizeof(void *)); @@ -240,7 +243,10 @@ int ipz_queue_ctor(struct ehca_pd *pd, struct ipz_queue *queue, ipz_queue_ctor_exit0: ehca_gen_err("Couldn't alloc pages queue=%p " "nr_of_pages=%x", queue, nr_of_pages); - vfree(queue->queue_pages); + if (is_vmalloc_addr(queue->queue_pages)) + vfree(queue->queue_pages); + else + kfree(queue->queue_pages); return 0; } @@ -262,7 +268,10 @@ int ipz_queue_dtor(struct ehca_pd *pd, struct ipz_queue *queue) free_page((unsigned long)queue->queue_pages[i]); } - vfree(queue->queue_pages); + if (is_vmalloc_addr(queue->queue_pages)) + vfree(queue->queue_pages); + else + kfree(queue->queue_pages); return 1; } diff --git a/drivers/infiniband/hw/ipath/ipath_driver.c b/drivers/infiniband/hw/ipath/ipath_driver.c index 69c0ce321b4..04e88b60055 100644 --- a/drivers/infiniband/hw/ipath/ipath_driver.c +++ b/drivers/infiniband/hw/ipath/ipath_driver.c @@ -470,14 +470,14 @@ static int __devinit ipath_init_one(struct pci_dev *pdev, goto bail_disable; } - ret = pci_set_dma_mask(pdev, DMA_64BIT_MASK); + ret = pci_set_dma_mask(pdev, DMA_BIT_MASK(64)); if (ret) { /* * if the 64 bit setup fails, try 32 bit. Some systems * do not setup 64 bit maps on systems with 2GB or less * memory installed. */ - ret = pci_set_dma_mask(pdev, DMA_32BIT_MASK); + ret = pci_set_dma_mask(pdev, DMA_BIT_MASK(32)); if (ret) { dev_info(&pdev->dev, "Unable to set DMA mask for unit %u: %d\n", @@ -486,7 +486,7 @@ static int __devinit ipath_init_one(struct pci_dev *pdev, } else { ipath_dbg("No 64bit DMA mask, used 32 bit mask\n"); - ret = pci_set_consistent_dma_mask(pdev, DMA_32BIT_MASK); + ret = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32)); if (ret) dev_info(&pdev->dev, "Unable to set DMA consistent mask " @@ -496,7 +496,7 @@ static int __devinit ipath_init_one(struct pci_dev *pdev, } } else { - ret = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK); + ret = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64)); if (ret) dev_info(&pdev->dev, "Unable to set DMA consistent mask " @@ -2715,7 +2715,7 @@ static void ipath_hol_signal_up(struct ipath_devdata *dd) * to prevent HoL blocking, then start the HoL timer that * periodically continues, then stop procs, so they can detect * link down if they want, and do something about it. - * Timer may already be running, so use __mod_timer, not add_timer. + * Timer may already be running, so use mod_timer, not add_timer. */ void ipath_hol_down(struct ipath_devdata *dd) { @@ -2724,7 +2724,7 @@ void ipath_hol_down(struct ipath_devdata *dd) dd->ipath_hol_next = IPATH_HOL_DOWNCONT; dd->ipath_hol_timer.expires = jiffies + msecs_to_jiffies(ipath_hol_timeout_ms); - __mod_timer(&dd->ipath_hol_timer, dd->ipath_hol_timer.expires); + mod_timer(&dd->ipath_hol_timer, dd->ipath_hol_timer.expires); } /* @@ -2763,7 +2763,7 @@ void ipath_hol_event(unsigned long opaque) else { dd->ipath_hol_timer.expires = jiffies + msecs_to_jiffies(ipath_hol_timeout_ms); - __mod_timer(&dd->ipath_hol_timer, + mod_timer(&dd->ipath_hol_timer, dd->ipath_hol_timer.expires); } } diff --git a/drivers/infiniband/hw/ipath/ipath_eeprom.c b/drivers/infiniband/hw/ipath/ipath_eeprom.c index dc37277f1c8..fc7181985e8 100644 --- a/drivers/infiniband/hw/ipath/ipath_eeprom.c +++ b/drivers/infiniband/hw/ipath/ipath_eeprom.c @@ -772,8 +772,8 @@ void ipath_get_eeprom_info(struct ipath_devdata *dd) "0x%x, not 0x%x\n", csum, ifp->if_csum); goto done; } - if (*(__be64 *) ifp->if_guid == 0ULL || - *(__be64 *) ifp->if_guid == __constant_cpu_to_be64(-1LL)) { + if (*(__be64 *) ifp->if_guid == cpu_to_be64(0) || + *(__be64 *) ifp->if_guid == ~cpu_to_be64(0)) { ipath_dev_err(dd, "Invalid GUID %llx from flash; " "ignoring\n", *(unsigned long long *) ifp->if_guid); diff --git a/drivers/infiniband/hw/ipath/ipath_fs.c b/drivers/infiniband/hw/ipath/ipath_fs.c index 8dc2bb78160..b3684060465 100644 --- a/drivers/infiniband/hw/ipath/ipath_fs.c +++ b/drivers/infiniband/hw/ipath/ipath_fs.c @@ -347,7 +347,7 @@ static int ipathfs_fill_super(struct super_block *sb, void *data, spin_unlock_irqrestore(&ipath_devs_lock, flags); ret = create_device_files(sb, dd); if (ret) { - deactivate_super(sb); + deactivate_locked_super(sb); goto bail; } spin_lock_irqsave(&ipath_devs_lock, flags); diff --git a/drivers/infiniband/hw/ipath/ipath_init_chip.c b/drivers/infiniband/hw/ipath/ipath_init_chip.c index 64aeefbd2a5..077879c0bdb 100644 --- a/drivers/infiniband/hw/ipath/ipath_init_chip.c +++ b/drivers/infiniband/hw/ipath/ipath_init_chip.c @@ -455,7 +455,7 @@ static void init_shadow_tids(struct ipath_devdata *dd) if (!addrs) { ipath_dev_err(dd, "failed to allocate shadow dma handle " "array, no expected sends!\n"); - vfree(dd->ipath_pageshadow); + vfree(pages); dd->ipath_pageshadow = NULL; return; } diff --git a/drivers/infiniband/hw/ipath/ipath_mad.c b/drivers/infiniband/hw/ipath/ipath_mad.c index 17a12319747..16a702d4601 100644 --- a/drivers/infiniband/hw/ipath/ipath_mad.c +++ b/drivers/infiniband/hw/ipath/ipath_mad.c @@ -37,10 +37,10 @@ #include "ipath_verbs.h" #include "ipath_common.h" -#define IB_SMP_UNSUP_VERSION __constant_htons(0x0004) -#define IB_SMP_UNSUP_METHOD __constant_htons(0x0008) -#define IB_SMP_UNSUP_METH_ATTR __constant_htons(0x000C) -#define IB_SMP_INVALID_FIELD __constant_htons(0x001C) +#define IB_SMP_UNSUP_VERSION cpu_to_be16(0x0004) +#define IB_SMP_UNSUP_METHOD cpu_to_be16(0x0008) +#define IB_SMP_UNSUP_METH_ATTR cpu_to_be16(0x000C) +#define IB_SMP_INVALID_FIELD cpu_to_be16(0x001C) static int reply(struct ib_smp *smp) { @@ -789,12 +789,12 @@ static int recv_subn_set_pkeytable(struct ib_smp *smp, return recv_subn_get_pkeytable(smp, ibdev); } -#define IB_PMA_CLASS_PORT_INFO __constant_htons(0x0001) -#define IB_PMA_PORT_SAMPLES_CONTROL __constant_htons(0x0010) -#define IB_PMA_PORT_SAMPLES_RESULT __constant_htons(0x0011) -#define IB_PMA_PORT_COUNTERS __constant_htons(0x0012) -#define IB_PMA_PORT_COUNTERS_EXT __constant_htons(0x001D) -#define IB_PMA_PORT_SAMPLES_RESULT_EXT __constant_htons(0x001E) +#define IB_PMA_CLASS_PORT_INFO cpu_to_be16(0x0001) +#define IB_PMA_PORT_SAMPLES_CONTROL cpu_to_be16(0x0010) +#define IB_PMA_PORT_SAMPLES_RESULT cpu_to_be16(0x0011) +#define IB_PMA_PORT_COUNTERS cpu_to_be16(0x0012) +#define IB_PMA_PORT_COUNTERS_EXT cpu_to_be16(0x001D) +#define IB_PMA_PORT_SAMPLES_RESULT_EXT cpu_to_be16(0x001E) struct ib_perf { u8 base_version; @@ -884,19 +884,19 @@ struct ib_pma_portcounters { __be32 port_rcv_packets; } __attribute__ ((packed)); -#define IB_PMA_SEL_SYMBOL_ERROR __constant_htons(0x0001) -#define IB_PMA_SEL_LINK_ERROR_RECOVERY __constant_htons(0x0002) -#define IB_PMA_SEL_LINK_DOWNED __constant_htons(0x0004) -#define IB_PMA_SEL_PORT_RCV_ERRORS __constant_htons(0x0008) -#define IB_PMA_SEL_PORT_RCV_REMPHYS_ERRORS __constant_htons(0x0010) -#define IB_PMA_SEL_PORT_XMIT_DISCARDS __constant_htons(0x0040) -#define IB_PMA_SEL_LOCAL_LINK_INTEGRITY_ERRORS __constant_htons(0x0200) -#define IB_PMA_SEL_EXCESSIVE_BUFFER_OVERRUNS __constant_htons(0x0400) -#define IB_PMA_SEL_PORT_VL15_DROPPED __constant_htons(0x0800) -#define IB_PMA_SEL_PORT_XMIT_DATA __constant_htons(0x1000) -#define IB_PMA_SEL_PORT_RCV_DATA __constant_htons(0x2000) -#define IB_PMA_SEL_PORT_XMIT_PACKETS __constant_htons(0x4000) -#define IB_PMA_SEL_PORT_RCV_PACKETS __constant_htons(0x8000) +#define IB_PMA_SEL_SYMBOL_ERROR cpu_to_be16(0x0001) +#define IB_PMA_SEL_LINK_ERROR_RECOVERY cpu_to_be16(0x0002) +#define IB_PMA_SEL_LINK_DOWNED cpu_to_be16(0x0004) +#define IB_PMA_SEL_PORT_RCV_ERRORS cpu_to_be16(0x0008) +#define IB_PMA_SEL_PORT_RCV_REMPHYS_ERRORS cpu_to_be16(0x0010) +#define IB_PMA_SEL_PORT_XMIT_DISCARDS cpu_to_be16(0x0040) +#define IB_PMA_SEL_LOCAL_LINK_INTEGRITY_ERRORS cpu_to_be16(0x0200) +#define IB_PMA_SEL_EXCESSIVE_BUFFER_OVERRUNS cpu_to_be16(0x0400) +#define IB_PMA_SEL_PORT_VL15_DROPPED cpu_to_be16(0x0800) +#define IB_PMA_SEL_PORT_XMIT_DATA cpu_to_be16(0x1000) +#define IB_PMA_SEL_PORT_RCV_DATA cpu_to_be16(0x2000) +#define IB_PMA_SEL_PORT_XMIT_PACKETS cpu_to_be16(0x4000) +#define IB_PMA_SEL_PORT_RCV_PACKETS cpu_to_be16(0x8000) struct ib_pma_portcounters_ext { u8 reserved; @@ -913,14 +913,14 @@ struct ib_pma_portcounters_ext { __be64 port_multicast_rcv_packets; } __attribute__ ((packed)); -#define IB_PMA_SELX_PORT_XMIT_DATA __constant_htons(0x0001) -#define IB_PMA_SELX_PORT_RCV_DATA __constant_htons(0x0002) -#define IB_PMA_SELX_PORT_XMIT_PACKETS __constant_htons(0x0004) -#define IB_PMA_SELX_PORT_RCV_PACKETS __constant_htons(0x0008) -#define IB_PMA_SELX_PORT_UNI_XMIT_PACKETS __constant_htons(0x0010) -#define IB_PMA_SELX_PORT_UNI_RCV_PACKETS __constant_htons(0x0020) -#define IB_PMA_SELX_PORT_MULTI_XMIT_PACKETS __constant_htons(0x0040) -#define IB_PMA_SELX_PORT_MULTI_RCV_PACKETS __constant_htons(0x0080) +#define IB_PMA_SELX_PORT_XMIT_DATA cpu_to_be16(0x0001) +#define IB_PMA_SELX_PORT_RCV_DATA cpu_to_be16(0x0002) +#define IB_PMA_SELX_PORT_XMIT_PACKETS cpu_to_be16(0x0004) +#define IB_PMA_SELX_PORT_RCV_PACKETS cpu_to_be16(0x0008) +#define IB_PMA_SELX_PORT_UNI_XMIT_PACKETS cpu_to_be16(0x0010) +#define IB_PMA_SELX_PORT_UNI_RCV_PACKETS cpu_to_be16(0x0020) +#define IB_PMA_SELX_PORT_MULTI_XMIT_PACKETS cpu_to_be16(0x0040) +#define IB_PMA_SELX_PORT_MULTI_RCV_PACKETS cpu_to_be16(0x0080) static int recv_pma_get_classportinfo(struct ib_perf *pmp) { @@ -933,7 +933,7 @@ static int recv_pma_get_classportinfo(struct ib_perf *pmp) pmp->status |= IB_SMP_INVALID_FIELD; /* Indicate AllPortSelect is valid (only one port anyway) */ - p->cap_mask = __constant_cpu_to_be16(1 << 8); + p->cap_mask = cpu_to_be16(1 << 8); p->base_version = 1; p->class_version = 1; /* @@ -951,12 +951,11 @@ static int recv_pma_get_classportinfo(struct ib_perf *pmp) * We support 5 counters which only count the mandatory quantities. */ #define COUNTER_MASK(q, n) (q << ((9 - n) * 3)) -#define COUNTER_MASK0_9 \ - __constant_cpu_to_be32(COUNTER_MASK(1, 0) | \ - COUNTER_MASK(1, 1) | \ - COUNTER_MASK(1, 2) | \ - COUNTER_MASK(1, 3) | \ - COUNTER_MASK(1, 4)) +#define COUNTER_MASK0_9 cpu_to_be32(COUNTER_MASK(1, 0) | \ + COUNTER_MASK(1, 1) | \ + COUNTER_MASK(1, 2) | \ + COUNTER_MASK(1, 3) | \ + COUNTER_MASK(1, 4)) static int recv_pma_get_portsamplescontrol(struct ib_perf *pmp, struct ib_device *ibdev, u8 port) @@ -1137,7 +1136,7 @@ static int recv_pma_get_portsamplesresult_ext(struct ib_perf *pmp, status = dev->pma_sample_status; p->sample_status = cpu_to_be16(status); /* 64 bits */ - p->extended_width = __constant_cpu_to_be32(0x80000000); + p->extended_width = cpu_to_be32(0x80000000); for (i = 0; i < ARRAY_SIZE(dev->pma_counter_select); i++) p->counter[i] = (status != IB_PMA_SAMPLE_STATUS_DONE) ? 0 : cpu_to_be64( @@ -1185,7 +1184,7 @@ static int recv_pma_get_portcounters(struct ib_perf *pmp, pmp->status |= IB_SMP_INVALID_FIELD; if (cntrs.symbol_error_counter > 0xFFFFUL) - p->symbol_error_counter = __constant_cpu_to_be16(0xFFFF); + p->symbol_error_counter = cpu_to_be16(0xFFFF); else p->symbol_error_counter = cpu_to_be16((u16)cntrs.symbol_error_counter); @@ -1199,17 +1198,17 @@ static int recv_pma_get_portcounters(struct ib_perf *pmp, else p->link_downed_counter = (u8)cntrs.link_downed_counter; if (cntrs.port_rcv_errors > 0xFFFFUL) - p->port_rcv_errors = __constant_cpu_to_be16(0xFFFF); + p->port_rcv_errors = cpu_to_be16(0xFFFF); else p->port_rcv_errors = cpu_to_be16((u16) cntrs.port_rcv_errors); if (cntrs.port_rcv_remphys_errors > 0xFFFFUL) - p->port_rcv_remphys_errors = __constant_cpu_to_be16(0xFFFF); + p->port_rcv_remphys_errors = cpu_to_be16(0xFFFF); else p->port_rcv_remphys_errors = cpu_to_be16((u16)cntrs.port_rcv_remphys_errors); if (cntrs.port_xmit_discards > 0xFFFFUL) - p->port_xmit_discards = __constant_cpu_to_be16(0xFFFF); + p->port_xmit_discards = cpu_to_be16(0xFFFF); else p->port_xmit_discards = cpu_to_be16((u16)cntrs.port_xmit_discards); @@ -1220,24 +1219,24 @@ static int recv_pma_get_portcounters(struct ib_perf *pmp, p->lli_ebor_errors = (cntrs.local_link_integrity_errors << 4) | cntrs.excessive_buffer_overrun_errors; if (cntrs.vl15_dropped > 0xFFFFUL) - p->vl15_dropped = __constant_cpu_to_be16(0xFFFF); + p->vl15_dropped = cpu_to_be16(0xFFFF); else p->vl15_dropped = cpu_to_be16((u16)cntrs.vl15_dropped); if (cntrs.port_xmit_data > 0xFFFFFFFFUL) - p->port_xmit_data = __constant_cpu_to_be32(0xFFFFFFFF); + p->port_xmit_data = cpu_to_be32(0xFFFFFFFF); else p->port_xmit_data = cpu_to_be32((u32)cntrs.port_xmit_data); if (cntrs.port_rcv_data > 0xFFFFFFFFUL) - p->port_rcv_data = __constant_cpu_to_be32(0xFFFFFFFF); + p->port_rcv_data = cpu_to_be32(0xFFFFFFFF); else p->port_rcv_data = cpu_to_be32((u32)cntrs.port_rcv_data); if (cntrs.port_xmit_packets > 0xFFFFFFFFUL) - p->port_xmit_packets = __constant_cpu_to_be32(0xFFFFFFFF); + p->port_xmit_packets = cpu_to_be32(0xFFFFFFFF); else p->port_xmit_packets = cpu_to_be32((u32)cntrs.port_xmit_packets); if (cntrs.port_rcv_packets > 0xFFFFFFFFUL) - p->port_rcv_packets = __constant_cpu_to_be32(0xFFFFFFFF); + p->port_rcv_packets = cpu_to_be32(0xFFFFFFFF); else p->port_rcv_packets = cpu_to_be32((u32) cntrs.port_rcv_packets); diff --git a/drivers/infiniband/hw/ipath/ipath_rc.c b/drivers/infiniband/hw/ipath/ipath_rc.c index 9170710b950..79b3dbc9717 100644 --- a/drivers/infiniband/hw/ipath/ipath_rc.c +++ b/drivers/infiniband/hw/ipath/ipath_rc.c @@ -1744,7 +1744,7 @@ void ipath_rc_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr, /* Signal completion event if the solicited bit is set. */ ipath_cq_enter(to_icq(qp->ibqp.recv_cq), &wc, (ohdr->bth[0] & - __constant_cpu_to_be32(1 << 23)) != 0); + cpu_to_be32(1 << 23)) != 0); break; case OP(RDMA_WRITE_FIRST): diff --git a/drivers/infiniband/hw/ipath/ipath_sdma.c b/drivers/infiniband/hw/ipath/ipath_sdma.c index 8e255adf5d9..4b069859085 100644 --- a/drivers/infiniband/hw/ipath/ipath_sdma.c +++ b/drivers/infiniband/hw/ipath/ipath_sdma.c @@ -781,10 +781,10 @@ retry: descqp = &dd->ipath_sdma_descq[dd->ipath_sdma_descq_cnt].qw[0]; descqp -= 2; /* SDmaLastDesc */ - descqp[0] |= __constant_cpu_to_le64(1ULL << 11); + descqp[0] |= cpu_to_le64(1ULL << 11); if (tx->txreq.flags & IPATH_SDMA_TXREQ_F_INTREQ) { /* SDmaIntReq */ - descqp[0] |= __constant_cpu_to_le64(1ULL << 15); + descqp[0] |= cpu_to_le64(1ULL << 15); } /* Commit writes to memory and advance the tail on the chip */ diff --git a/drivers/infiniband/hw/ipath/ipath_uc.c b/drivers/infiniband/hw/ipath/ipath_uc.c index 82cc588b8bf..22e60998f1a 100644 --- a/drivers/infiniband/hw/ipath/ipath_uc.c +++ b/drivers/infiniband/hw/ipath/ipath_uc.c @@ -419,7 +419,7 @@ void ipath_uc_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr, /* Signal completion event if the solicited bit is set. */ ipath_cq_enter(to_icq(qp->ibqp.recv_cq), &wc, (ohdr->bth[0] & - __constant_cpu_to_be32(1 << 23)) != 0); + cpu_to_be32(1 << 23)) != 0); break; case OP(RDMA_WRITE_FIRST): diff --git a/drivers/infiniband/hw/ipath/ipath_ud.c b/drivers/infiniband/hw/ipath/ipath_ud.c index 91c74cc797a..6076cb61bf6 100644 --- a/drivers/infiniband/hw/ipath/ipath_ud.c +++ b/drivers/infiniband/hw/ipath/ipath_ud.c @@ -370,7 +370,7 @@ int ipath_make_ud_req(struct ipath_qp *qp) */ ohdr->bth[1] = ah_attr->dlid >= IPATH_MULTICAST_LID_BASE && ah_attr->dlid != IPATH_PERMISSIVE_LID ? - __constant_cpu_to_be32(IPATH_MULTICAST_QPN) : + cpu_to_be32(IPATH_MULTICAST_QPN) : cpu_to_be32(wqe->wr.wr.ud.remote_qpn); ohdr->bth[2] = cpu_to_be32(qp->s_next_psn++ & IPATH_PSN_MASK); /* @@ -573,7 +573,7 @@ void ipath_ud_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr, /* Signal completion event if the solicited bit is set. */ ipath_cq_enter(to_icq(qp->ibqp.recv_cq), &wc, (ohdr->bth[0] & - __constant_cpu_to_be32(1 << 23)) != 0); + cpu_to_be32(1 << 23)) != 0); bail:; } diff --git a/drivers/infiniband/hw/ipath/ipath_user_pages.c b/drivers/infiniband/hw/ipath/ipath_user_pages.c index 0190edc8044..855911e7396 100644 --- a/drivers/infiniband/hw/ipath/ipath_user_pages.c +++ b/drivers/infiniband/hw/ipath/ipath_user_pages.c @@ -209,20 +209,20 @@ void ipath_release_user_pages_on_close(struct page **p, size_t num_pages) mm = get_task_mm(current); if (!mm) - goto bail; + return; work = kmalloc(sizeof(*work), GFP_KERNEL); if (!work) goto bail_mm; - goto bail; - INIT_WORK(&work->work, user_pages_account); work->mm = mm; work->num_pages = num_pages; + schedule_work(&work->work); + return; + bail_mm: mmput(mm); -bail: return; } diff --git a/drivers/infiniband/hw/ipath/ipath_user_sdma.c b/drivers/infiniband/hw/ipath/ipath_user_sdma.c index 82d9a0b5ca2..7bff4b9baa0 100644 --- a/drivers/infiniband/hw/ipath/ipath_user_sdma.c +++ b/drivers/infiniband/hw/ipath/ipath_user_sdma.c @@ -667,13 +667,13 @@ static inline __le64 ipath_sdma_make_desc0(struct ipath_devdata *dd, static inline __le64 ipath_sdma_make_first_desc0(__le64 descq) { - return descq | __constant_cpu_to_le64(1ULL << 12); + return descq | cpu_to_le64(1ULL << 12); } static inline __le64 ipath_sdma_make_last_desc0(__le64 descq) { /* last */ /* dma head */ - return descq | __constant_cpu_to_le64(1ULL << 11 | 1ULL << 13); + return descq | cpu_to_le64(1ULL << 11 | 1ULL << 13); } static inline __le64 ipath_sdma_make_desc1(u64 addr) @@ -763,7 +763,7 @@ static int ipath_user_sdma_push_pkts(struct ipath_devdata *dd, if (ofs >= IPATH_SMALLBUF_DWORDS) { for (i = 0; i < pkt->naddr; i++) { dd->ipath_sdma_descq[dtail].qw[0] |= - __constant_cpu_to_le64(1ULL << 14); + cpu_to_le64(1ULL << 14); if (++dtail == dd->ipath_sdma_descq_cnt) dtail = 0; } diff --git a/drivers/infiniband/hw/ipath/ipath_verbs.c b/drivers/infiniband/hw/ipath/ipath_verbs.c index cdf0e6abd34..9289ab4b0ae 100644 --- a/drivers/infiniband/hw/ipath/ipath_verbs.c +++ b/drivers/infiniband/hw/ipath/ipath_verbs.c @@ -1585,7 +1585,7 @@ static int ipath_query_port(struct ib_device *ibdev, u64 ibcstat; memset(props, 0, sizeof(*props)); - props->lid = lid ? lid : __constant_be16_to_cpu(IB_LID_PERMISSIVE); + props->lid = lid ? lid : be16_to_cpu(IB_LID_PERMISSIVE); props->lmc = dd->ipath_lmc; props->sm_lid = dev->sm_lid; props->sm_sl = dev->sm_sl; diff --git a/drivers/infiniband/hw/ipath/ipath_verbs.h b/drivers/infiniband/hw/ipath/ipath_verbs.h index 11e3f613df9..ae6cff4abff 100644 --- a/drivers/infiniband/hw/ipath/ipath_verbs.h +++ b/drivers/infiniband/hw/ipath/ipath_verbs.h @@ -86,11 +86,11 @@ #define IB_PMA_SAMPLE_STATUS_RUNNING 0x02 /* Mandatory IB performance counter select values. */ -#define IB_PMA_PORT_XMIT_DATA __constant_htons(0x0001) -#define IB_PMA_PORT_RCV_DATA __constant_htons(0x0002) -#define IB_PMA_PORT_XMIT_PKTS __constant_htons(0x0003) -#define IB_PMA_PORT_RCV_PKTS __constant_htons(0x0004) -#define IB_PMA_PORT_XMIT_WAIT __constant_htons(0x0005) +#define IB_PMA_PORT_XMIT_DATA cpu_to_be16(0x0001) +#define IB_PMA_PORT_RCV_DATA cpu_to_be16(0x0002) +#define IB_PMA_PORT_XMIT_PKTS cpu_to_be16(0x0003) +#define IB_PMA_PORT_RCV_PKTS cpu_to_be16(0x0004) +#define IB_PMA_PORT_XMIT_WAIT cpu_to_be16(0x0005) struct ib_reth { __be64 vaddr; diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c index 606f1e2ef28..19e68ab6616 100644 --- a/drivers/infiniband/hw/mlx4/mad.c +++ b/drivers/infiniband/hw/mlx4/mad.c @@ -147,7 +147,8 @@ static void update_sm_ah(struct mlx4_ib_dev *dev, u8 port_num, u16 lid, u8 sl) * Snoop SM MADs for port info and P_Key table sets, so we can * synthesize LID change and P_Key change events. */ -static void smp_snoop(struct ib_device *ibdev, u8 port_num, struct ib_mad *mad) +static void smp_snoop(struct ib_device *ibdev, u8 port_num, struct ib_mad *mad, + u16 prev_lid) { struct ib_event event; @@ -157,6 +158,7 @@ static void smp_snoop(struct ib_device *ibdev, u8 port_num, struct ib_mad *mad) if (mad->mad_hdr.attr_id == IB_SMP_ATTR_PORT_INFO) { struct ib_port_info *pinfo = (struct ib_port_info *) ((struct ib_smp *) mad)->data; + u16 lid = be16_to_cpu(pinfo->lid); update_sm_ah(to_mdev(ibdev), port_num, be16_to_cpu(pinfo->sm_lid), @@ -165,12 +167,15 @@ static void smp_snoop(struct ib_device *ibdev, u8 port_num, struct ib_mad *mad) event.device = ibdev; event.element.port_num = port_num; - if (pinfo->clientrereg_resv_subnetto & 0x80) + if (pinfo->clientrereg_resv_subnetto & 0x80) { event.event = IB_EVENT_CLIENT_REREGISTER; - else - event.event = IB_EVENT_LID_CHANGE; + ib_dispatch_event(&event); + } - ib_dispatch_event(&event); + if (prev_lid != lid) { + event.event = IB_EVENT_LID_CHANGE; + ib_dispatch_event(&event); + } } if (mad->mad_hdr.attr_id == IB_SMP_ATTR_PKEY_TABLE) { @@ -228,8 +233,9 @@ int mlx4_ib_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num, struct ib_wc *in_wc, struct ib_grh *in_grh, struct ib_mad *in_mad, struct ib_mad *out_mad) { - u16 slid; + u16 slid, prev_lid = 0; int err; + struct ib_port_attr pattr; slid = in_wc ? in_wc->slid : be16_to_cpu(IB_LID_PERMISSIVE); @@ -263,6 +269,13 @@ int mlx4_ib_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num, } else return IB_MAD_RESULT_SUCCESS; + if ((in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED || + in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) && + in_mad->mad_hdr.method == IB_MGMT_METHOD_SET && + in_mad->mad_hdr.attr_id == IB_SMP_ATTR_PORT_INFO && + !ib_query_port(ibdev, port_num, &pattr)) + prev_lid = pattr.lid; + err = mlx4_MAD_IFC(to_mdev(ibdev), mad_flags & IB_MAD_IGNORE_MKEY, mad_flags & IB_MAD_IGNORE_BKEY, @@ -271,7 +284,7 @@ int mlx4_ib_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num, return IB_MAD_RESULT_FAILURE; if (!out_mad->mad_hdr.status) { - smp_snoop(ibdev, port_num, in_mad); + smp_snoop(ibdev, port_num, in_mad, prev_lid); node_desc_override(ibdev, out_mad); } diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c index 61588bd273b..ae3d7590346 100644 --- a/drivers/infiniband/hw/mlx4/main.c +++ b/drivers/infiniband/hw/mlx4/main.c @@ -394,8 +394,7 @@ static int mlx4_ib_mmap(struct ib_ucontext *context, struct vm_area_struct *vma) PAGE_SIZE, vma->vm_page_prot)) return -EAGAIN; } else if (vma->vm_pgoff == 1 && dev->dev->caps.bf_reg_size != 0) { - /* FIXME want pgprot_writecombine() for BlueFlame pages */ - vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); + vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot); if (io_remap_pfn_range(vma, vma->vm_start, to_mucontext(context)->uar.pfn + @@ -699,11 +698,12 @@ static void mlx4_ib_remove(struct mlx4_dev *dev, void *ibdev_ptr) struct mlx4_ib_dev *ibdev = ibdev_ptr; int p; + mlx4_ib_mad_cleanup(ibdev); + ib_unregister_device(&ibdev->ib_dev); + for (p = 1; p <= ibdev->num_ports; ++p) mlx4_CLOSE_PORT(dev, p); - mlx4_ib_mad_cleanup(ibdev); - ib_unregister_device(&ibdev->ib_dev); iounmap(ibdev->uar_map); mlx4_uar_free(dev, &ibdev->priv_uar); mlx4_pd_free(dev, ibdev->priv_pdn); diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h index 9974e886b8d..8a7dd6795fa 100644 --- a/drivers/infiniband/hw/mlx4/mlx4_ib.h +++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h @@ -86,6 +86,7 @@ struct mlx4_ib_mr { struct mlx4_ib_fast_reg_page_list { struct ib_fast_reg_page_list ibfrpl; + __be64 *mapped_page_list; dma_addr_t map; }; diff --git a/drivers/infiniband/hw/mlx4/mr.c b/drivers/infiniband/hw/mlx4/mr.c index 8e4d26d56a9..8f3666b20ea 100644 --- a/drivers/infiniband/hw/mlx4/mr.c +++ b/drivers/infiniband/hw/mlx4/mr.c @@ -231,7 +231,11 @@ struct ib_fast_reg_page_list *mlx4_ib_alloc_fast_reg_page_list(struct ib_device if (!mfrpl) return ERR_PTR(-ENOMEM); - mfrpl->ibfrpl.page_list = dma_alloc_coherent(&dev->dev->pdev->dev, + mfrpl->ibfrpl.page_list = kmalloc(size, GFP_KERNEL); + if (!mfrpl->ibfrpl.page_list) + goto err_free; + + mfrpl->mapped_page_list = dma_alloc_coherent(&dev->dev->pdev->dev, size, &mfrpl->map, GFP_KERNEL); if (!mfrpl->ibfrpl.page_list) @@ -242,6 +246,7 @@ struct ib_fast_reg_page_list *mlx4_ib_alloc_fast_reg_page_list(struct ib_device return &mfrpl->ibfrpl; err_free: + kfree(mfrpl->ibfrpl.page_list); kfree(mfrpl); return ERR_PTR(-ENOMEM); } @@ -252,8 +257,9 @@ void mlx4_ib_free_fast_reg_page_list(struct ib_fast_reg_page_list *page_list) struct mlx4_ib_fast_reg_page_list *mfrpl = to_mfrpl(page_list); int size = page_list->max_page_list_len * sizeof (u64); - dma_free_coherent(&dev->dev->pdev->dev, size, page_list->page_list, + dma_free_coherent(&dev->dev->pdev->dev, size, mfrpl->mapped_page_list, mfrpl->map); + kfree(mfrpl->ibfrpl.page_list); kfree(mfrpl); } diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c index a91cb4c3fa5..c4a02648c8a 100644 --- a/drivers/infiniband/hw/mlx4/qp.c +++ b/drivers/infiniband/hw/mlx4/qp.c @@ -71,17 +71,17 @@ enum { }; static const __be32 mlx4_ib_opcode[] = { - [IB_WR_SEND] = __constant_cpu_to_be32(MLX4_OPCODE_SEND), - [IB_WR_LSO] = __constant_cpu_to_be32(MLX4_OPCODE_LSO), - [IB_WR_SEND_WITH_IMM] = __constant_cpu_to_be32(MLX4_OPCODE_SEND_IMM), - [IB_WR_RDMA_WRITE] = __constant_cpu_to_be32(MLX4_OPCODE_RDMA_WRITE), - [IB_WR_RDMA_WRITE_WITH_IMM] = __constant_cpu_to_be32(MLX4_OPCODE_RDMA_WRITE_IMM), - [IB_WR_RDMA_READ] = __constant_cpu_to_be32(MLX4_OPCODE_RDMA_READ), - [IB_WR_ATOMIC_CMP_AND_SWP] = __constant_cpu_to_be32(MLX4_OPCODE_ATOMIC_CS), - [IB_WR_ATOMIC_FETCH_AND_ADD] = __constant_cpu_to_be32(MLX4_OPCODE_ATOMIC_FA), - [IB_WR_SEND_WITH_INV] = __constant_cpu_to_be32(MLX4_OPCODE_SEND_INVAL), - [IB_WR_LOCAL_INV] = __constant_cpu_to_be32(MLX4_OPCODE_LOCAL_INVAL), - [IB_WR_FAST_REG_MR] = __constant_cpu_to_be32(MLX4_OPCODE_FMR), + [IB_WR_SEND] = cpu_to_be32(MLX4_OPCODE_SEND), + [IB_WR_LSO] = cpu_to_be32(MLX4_OPCODE_LSO), + [IB_WR_SEND_WITH_IMM] = cpu_to_be32(MLX4_OPCODE_SEND_IMM), + [IB_WR_RDMA_WRITE] = cpu_to_be32(MLX4_OPCODE_RDMA_WRITE), + [IB_WR_RDMA_WRITE_WITH_IMM] = cpu_to_be32(MLX4_OPCODE_RDMA_WRITE_IMM), + [IB_WR_RDMA_READ] = cpu_to_be32(MLX4_OPCODE_RDMA_READ), + [IB_WR_ATOMIC_CMP_AND_SWP] = cpu_to_be32(MLX4_OPCODE_ATOMIC_CS), + [IB_WR_ATOMIC_FETCH_AND_ADD] = cpu_to_be32(MLX4_OPCODE_ATOMIC_FA), + [IB_WR_SEND_WITH_INV] = cpu_to_be32(MLX4_OPCODE_SEND_INVAL), + [IB_WR_LOCAL_INV] = cpu_to_be32(MLX4_OPCODE_LOCAL_INVAL), + [IB_WR_FAST_REG_MR] = cpu_to_be32(MLX4_OPCODE_FMR), }; static struct mlx4_ib_sqp *to_msqp(struct mlx4_ib_qp *mqp) @@ -1365,7 +1365,7 @@ static void set_fmr_seg(struct mlx4_wqe_fmr_seg *fseg, struct ib_send_wr *wr) int i; for (i = 0; i < wr->wr.fast_reg.page_list_len; ++i) - wr->wr.fast_reg.page_list->page_list[i] = + mfrpl->mapped_page_list[i] = cpu_to_be64(wr->wr.fast_reg.page_list->page_list[i] | MLX4_MTT_FLAG_PRESENT); @@ -1585,12 +1585,16 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, break; case IB_WR_LOCAL_INV: + ctrl->srcrb_flags |= + cpu_to_be32(MLX4_WQE_CTRL_STRONG_ORDER); set_local_inv_seg(wqe, wr->ex.invalidate_rkey); wqe += sizeof (struct mlx4_wqe_local_inval_seg); size += sizeof (struct mlx4_wqe_local_inval_seg) / 16; break; case IB_WR_FAST_REG_MR: + ctrl->srcrb_flags |= + cpu_to_be32(MLX4_WQE_CTRL_STRONG_ORDER); set_fmr_seg(wqe, wr); wqe += sizeof (struct mlx4_wqe_fmr_seg); size += sizeof (struct mlx4_wqe_fmr_seg) / 16; diff --git a/drivers/infiniband/hw/mthca/mthca_cmd.c b/drivers/infiniband/hw/mthca/mthca_cmd.c index c33e1c53c79..8c2ed994d54 100644 --- a/drivers/infiniband/hw/mthca/mthca_cmd.c +++ b/drivers/infiniband/hw/mthca/mthca_cmd.c @@ -157,13 +157,15 @@ enum { enum { CMD_TIME_CLASS_A = (HZ + 999) / 1000 + 1, CMD_TIME_CLASS_B = (HZ + 99) / 100 + 1, - CMD_TIME_CLASS_C = (HZ + 9) / 10 + 1 + CMD_TIME_CLASS_C = (HZ + 9) / 10 + 1, + CMD_TIME_CLASS_D = 60 * HZ }; #else enum { CMD_TIME_CLASS_A = 60 * HZ, CMD_TIME_CLASS_B = 60 * HZ, - CMD_TIME_CLASS_C = 60 * HZ + CMD_TIME_CLASS_C = 60 * HZ, + CMD_TIME_CLASS_D = 60 * HZ }; #endif @@ -598,7 +600,7 @@ int mthca_SYS_EN(struct mthca_dev *dev, u8 *status) u64 out; int ret; - ret = mthca_cmd_imm(dev, 0, &out, 0, 0, CMD_SYS_EN, HZ, status); + ret = mthca_cmd_imm(dev, 0, &out, 0, 0, CMD_SYS_EN, CMD_TIME_CLASS_D, status); if (*status == MTHCA_CMD_STAT_DDR_MEM_ERR) mthca_warn(dev, "SYS_EN DDR error: syn=%x, sock=%d, " @@ -611,7 +613,7 @@ int mthca_SYS_EN(struct mthca_dev *dev, u8 *status) int mthca_SYS_DIS(struct mthca_dev *dev, u8 *status) { - return mthca_cmd(dev, 0, 0, 0, CMD_SYS_DIS, HZ, status); + return mthca_cmd(dev, 0, 0, 0, CMD_SYS_DIS, CMD_TIME_CLASS_C, status); } static int mthca_map_cmd(struct mthca_dev *dev, u16 op, struct mthca_icm *icm, @@ -1057,7 +1059,7 @@ int mthca_QUERY_DEV_LIM(struct mthca_dev *dev, MTHCA_GET(field, outbox, QUERY_DEV_LIM_RSVD_MTT_OFFSET); if (mthca_is_memfree(dev)) dev_lim->reserved_mtts = ALIGN((1 << (field >> 4)) * sizeof(u64), - MTHCA_MTT_SEG_SIZE) / MTHCA_MTT_SEG_SIZE; + dev->limits.mtt_seg_size) / dev->limits.mtt_seg_size; else dev_lim->reserved_mtts = 1 << (field >> 4); MTHCA_GET(field, outbox, QUERY_DEV_LIM_MAX_MRW_SZ_OFFSET); @@ -1390,7 +1392,7 @@ int mthca_INIT_HCA(struct mthca_dev *dev, MTHCA_PUT(inbox, param->uarc_base, INIT_HCA_UAR_CTX_BASE_OFFSET); } - err = mthca_cmd(dev, mailbox->dma, 0, 0, CMD_INIT_HCA, HZ, status); + err = mthca_cmd(dev, mailbox->dma, 0, 0, CMD_INIT_HCA, CMD_TIME_CLASS_D, status); mthca_free_mailbox(dev, mailbox); return err; @@ -1450,12 +1452,12 @@ int mthca_INIT_IB(struct mthca_dev *dev, int mthca_CLOSE_IB(struct mthca_dev *dev, int port, u8 *status) { - return mthca_cmd(dev, 0, port, 0, CMD_CLOSE_IB, HZ, status); + return mthca_cmd(dev, 0, port, 0, CMD_CLOSE_IB, CMD_TIME_CLASS_A, status); } int mthca_CLOSE_HCA(struct mthca_dev *dev, int panic, u8 *status) { - return mthca_cmd(dev, 0, 0, panic, CMD_CLOSE_HCA, HZ, status); + return mthca_cmd(dev, 0, 0, panic, CMD_CLOSE_HCA, CMD_TIME_CLASS_C, status); } int mthca_SET_IB(struct mthca_dev *dev, struct mthca_set_ib_param *param, diff --git a/drivers/infiniband/hw/mthca/mthca_dev.h b/drivers/infiniband/hw/mthca/mthca_dev.h index 252590116df..9ef611f6dd3 100644 --- a/drivers/infiniband/hw/mthca/mthca_dev.h +++ b/drivers/infiniband/hw/mthca/mthca_dev.h @@ -159,6 +159,7 @@ struct mthca_limits { int reserved_eqs; int num_mpts; int num_mtt_segs; + int mtt_seg_size; int fmr_reserved_mtts; int reserved_mtts; int reserved_mrws; diff --git a/drivers/infiniband/hw/mthca/mthca_eq.c b/drivers/infiniband/hw/mthca/mthca_eq.c index 28f0e0c40d7..90e4e450a12 100644 --- a/drivers/infiniband/hw/mthca/mthca_eq.c +++ b/drivers/infiniband/hw/mthca/mthca_eq.c @@ -641,9 +641,11 @@ static void mthca_free_irqs(struct mthca_dev *dev) if (dev->eq_table.have_irq) free_irq(dev->pdev->irq, dev); for (i = 0; i < MTHCA_NUM_EQ; ++i) - if (dev->eq_table.eq[i].have_irq) + if (dev->eq_table.eq[i].have_irq) { free_irq(dev->eq_table.eq[i].msi_x_vector, dev->eq_table.eq + i); + dev->eq_table.eq[i].have_irq = 0; + } } static int mthca_map_reg(struct mthca_dev *dev, diff --git a/drivers/infiniband/hw/mthca/mthca_mad.c b/drivers/infiniband/hw/mthca/mthca_mad.c index 640449582ab..5648659ff0b 100644 --- a/drivers/infiniband/hw/mthca/mthca_mad.c +++ b/drivers/infiniband/hw/mthca/mthca_mad.c @@ -104,7 +104,8 @@ static void update_sm_ah(struct mthca_dev *dev, */ static void smp_snoop(struct ib_device *ibdev, u8 port_num, - struct ib_mad *mad) + struct ib_mad *mad, + u16 prev_lid) { struct ib_event event; @@ -114,6 +115,7 @@ static void smp_snoop(struct ib_device *ibdev, if (mad->mad_hdr.attr_id == IB_SMP_ATTR_PORT_INFO) { struct ib_port_info *pinfo = (struct ib_port_info *) ((struct ib_smp *) mad)->data; + u16 lid = be16_to_cpu(pinfo->lid); mthca_update_rate(to_mdev(ibdev), port_num); update_sm_ah(to_mdev(ibdev), port_num, @@ -123,12 +125,15 @@ static void smp_snoop(struct ib_device *ibdev, event.device = ibdev; event.element.port_num = port_num; - if (pinfo->clientrereg_resv_subnetto & 0x80) + if (pinfo->clientrereg_resv_subnetto & 0x80) { event.event = IB_EVENT_CLIENT_REREGISTER; - else - event.event = IB_EVENT_LID_CHANGE; + ib_dispatch_event(&event); + } - ib_dispatch_event(&event); + if (prev_lid != lid) { + event.event = IB_EVENT_LID_CHANGE; + ib_dispatch_event(&event); + } } if (mad->mad_hdr.attr_id == IB_SMP_ATTR_PKEY_TABLE) { @@ -196,6 +201,8 @@ int mthca_process_mad(struct ib_device *ibdev, int err; u8 status; u16 slid = in_wc ? in_wc->slid : be16_to_cpu(IB_LID_PERMISSIVE); + u16 prev_lid = 0; + struct ib_port_attr pattr; /* Forward locally generated traps to the SM */ if (in_mad->mad_hdr.method == IB_MGMT_METHOD_TRAP && @@ -233,6 +240,12 @@ int mthca_process_mad(struct ib_device *ibdev, return IB_MAD_RESULT_SUCCESS; } else return IB_MAD_RESULT_SUCCESS; + if ((in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED || + in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) && + in_mad->mad_hdr.method == IB_MGMT_METHOD_SET && + in_mad->mad_hdr.attr_id == IB_SMP_ATTR_PORT_INFO && + !ib_query_port(ibdev, port_num, &pattr)) + prev_lid = pattr.lid; err = mthca_MAD_IFC(to_mdev(ibdev), mad_flags & IB_MAD_IGNORE_MKEY, @@ -252,7 +265,7 @@ int mthca_process_mad(struct ib_device *ibdev, } if (!out_mad->mad_hdr.status) { - smp_snoop(ibdev, port_num, in_mad); + smp_snoop(ibdev, port_num, in_mad, prev_lid); node_desc_override(ibdev, out_mad); } diff --git a/drivers/infiniband/hw/mthca/mthca_main.c b/drivers/infiniband/hw/mthca/mthca_main.c index 52f60f4eea0..13da9f1d24c 100644 --- a/drivers/infiniband/hw/mthca/mthca_main.c +++ b/drivers/infiniband/hw/mthca/mthca_main.c @@ -125,6 +125,10 @@ module_param_named(fmr_reserved_mtts, hca_profile.fmr_reserved_mtts, int, 0444); MODULE_PARM_DESC(fmr_reserved_mtts, "number of memory translation table segments reserved for FMR"); +static int log_mtts_per_seg = ilog2(MTHCA_MTT_SEG_SIZE / 8); +module_param_named(log_mtts_per_seg, log_mtts_per_seg, int, 0444); +MODULE_PARM_DESC(log_mtts_per_seg, "Log2 number of MTT entries per segment (1-5)"); + static char mthca_version[] __devinitdata = DRV_NAME ": Mellanox InfiniBand HCA driver v" DRV_VERSION " (" DRV_RELDATE ")\n"; @@ -162,6 +166,7 @@ static int mthca_dev_lim(struct mthca_dev *mdev, struct mthca_dev_lim *dev_lim) int err; u8 status; + mdev->limits.mtt_seg_size = (1 << log_mtts_per_seg) * 8; err = mthca_QUERY_DEV_LIM(mdev, dev_lim, &status); if (err) { mthca_err(mdev, "QUERY_DEV_LIM command failed, aborting.\n"); @@ -460,11 +465,11 @@ static int mthca_init_icm(struct mthca_dev *mdev, } /* CPU writes to non-reserved MTTs, while HCA might DMA to reserved mtts */ - mdev->limits.reserved_mtts = ALIGN(mdev->limits.reserved_mtts * MTHCA_MTT_SEG_SIZE, - dma_get_cache_alignment()) / MTHCA_MTT_SEG_SIZE; + mdev->limits.reserved_mtts = ALIGN(mdev->limits.reserved_mtts * mdev->limits.mtt_seg_size, + dma_get_cache_alignment()) / mdev->limits.mtt_seg_size; mdev->mr_table.mtt_table = mthca_alloc_icm_table(mdev, init_hca->mtt_base, - MTHCA_MTT_SEG_SIZE, + mdev->limits.mtt_seg_size, mdev->limits.num_mtt_segs, mdev->limits.reserved_mtts, 1, 0); @@ -1016,20 +1021,20 @@ static int __mthca_init_one(struct pci_dev *pdev, int hca_type) pci_set_master(pdev); - err = pci_set_dma_mask(pdev, DMA_64BIT_MASK); + err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64)); if (err) { dev_warn(&pdev->dev, "Warning: couldn't set 64-bit PCI DMA mask.\n"); - err = pci_set_dma_mask(pdev, DMA_32BIT_MASK); + err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32)); if (err) { dev_err(&pdev->dev, "Can't set PCI DMA mask, aborting.\n"); goto err_free_res; } } - err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK); + err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64)); if (err) { dev_warn(&pdev->dev, "Warning: couldn't set 64-bit " "consistent PCI DMA mask.\n"); - err = pci_set_consistent_dma_mask(pdev, DMA_32BIT_MASK); + err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32)); if (err) { dev_err(&pdev->dev, "Can't set consistent PCI DMA mask, " "aborting.\n"); @@ -1315,6 +1320,12 @@ static void __init mthca_validate_profile(void) printk(KERN_WARNING PFX "Corrected fmr_reserved_mtts to %d.\n", hca_profile.fmr_reserved_mtts); } + + if ((log_mtts_per_seg < 1) || (log_mtts_per_seg > 5)) { + printk(KERN_WARNING PFX "bad log_mtts_per_seg (%d). Using default - %d\n", + log_mtts_per_seg, ilog2(MTHCA_MTT_SEG_SIZE / 8)); + log_mtts_per_seg = ilog2(MTHCA_MTT_SEG_SIZE / 8); + } } static int __init mthca_init(void) diff --git a/drivers/infiniband/hw/mthca/mthca_mr.c b/drivers/infiniband/hw/mthca/mthca_mr.c index 882e6b73591..065b2089987 100644 --- a/drivers/infiniband/hw/mthca/mthca_mr.c +++ b/drivers/infiniband/hw/mthca/mthca_mr.c @@ -220,7 +220,7 @@ static struct mthca_mtt *__mthca_alloc_mtt(struct mthca_dev *dev, int size, mtt->buddy = buddy; mtt->order = 0; - for (i = MTHCA_MTT_SEG_SIZE / 8; i < size; i <<= 1) + for (i = dev->limits.mtt_seg_size / 8; i < size; i <<= 1) ++mtt->order; mtt->first_seg = mthca_alloc_mtt_range(dev, mtt->order, buddy); @@ -267,7 +267,7 @@ static int __mthca_write_mtt(struct mthca_dev *dev, struct mthca_mtt *mtt, while (list_len > 0) { mtt_entry[0] = cpu_to_be64(dev->mr_table.mtt_base + - mtt->first_seg * MTHCA_MTT_SEG_SIZE + + mtt->first_seg * dev->limits.mtt_seg_size + start_index * 8); mtt_entry[1] = 0; for (i = 0; i < list_len && i < MTHCA_MAILBOX_SIZE / 8 - 2; ++i) @@ -326,7 +326,7 @@ static void mthca_tavor_write_mtt_seg(struct mthca_dev *dev, u64 __iomem *mtts; int i; - mtts = dev->mr_table.tavor_fmr.mtt_base + mtt->first_seg * MTHCA_MTT_SEG_SIZE + + mtts = dev->mr_table.tavor_fmr.mtt_base + mtt->first_seg * dev->limits.mtt_seg_size + start_index * sizeof (u64); for (i = 0; i < list_len; ++i) mthca_write64_raw(cpu_to_be64(buffer_list[i] | MTHCA_MTT_FLAG_PRESENT), @@ -345,17 +345,21 @@ static void mthca_arbel_write_mtt_seg(struct mthca_dev *dev, /* For Arbel, all MTTs must fit in the same page. */ BUG_ON(s / PAGE_SIZE != (s + list_len * sizeof(u64) - 1) / PAGE_SIZE); /* Require full segments */ - BUG_ON(s % MTHCA_MTT_SEG_SIZE); + BUG_ON(s % dev->limits.mtt_seg_size); mtts = mthca_table_find(dev->mr_table.mtt_table, mtt->first_seg + - s / MTHCA_MTT_SEG_SIZE, &dma_handle); + s / dev->limits.mtt_seg_size, &dma_handle); BUG_ON(!mtts); + dma_sync_single_for_cpu(&dev->pdev->dev, dma_handle, + list_len * sizeof (u64), DMA_TO_DEVICE); + for (i = 0; i < list_len; ++i) mtts[i] = cpu_to_be64(buffer_list[i] | MTHCA_MTT_FLAG_PRESENT); - dma_sync_single(&dev->pdev->dev, dma_handle, list_len * sizeof (u64), DMA_TO_DEVICE); + dma_sync_single_for_device(&dev->pdev->dev, dma_handle, + list_len * sizeof (u64), DMA_TO_DEVICE); } int mthca_write_mtt(struct mthca_dev *dev, struct mthca_mtt *mtt, @@ -479,7 +483,7 @@ int mthca_mr_alloc(struct mthca_dev *dev, u32 pd, int buffer_size_shift, if (mr->mtt) mpt_entry->mtt_seg = cpu_to_be64(dev->mr_table.mtt_base + - mr->mtt->first_seg * MTHCA_MTT_SEG_SIZE); + mr->mtt->first_seg * dev->limits.mtt_seg_size); if (0) { mthca_dbg(dev, "Dumping MPT entry %08x:\n", mr->ibmr.lkey); @@ -626,7 +630,7 @@ int mthca_fmr_alloc(struct mthca_dev *dev, u32 pd, goto err_out_table; } - mtt_seg = mr->mtt->first_seg * MTHCA_MTT_SEG_SIZE; + mtt_seg = mr->mtt->first_seg * dev->limits.mtt_seg_size; if (mthca_is_memfree(dev)) { mr->mem.arbel.mtts = mthca_table_find(dev->mr_table.mtt_table, @@ -803,12 +807,15 @@ int mthca_arbel_map_phys_fmr(struct ib_fmr *ibfmr, u64 *page_list, wmb(); + dma_sync_single_for_cpu(&dev->pdev->dev, fmr->mem.arbel.dma_handle, + list_len * sizeof(u64), DMA_TO_DEVICE); + for (i = 0; i < list_len; ++i) fmr->mem.arbel.mtts[i] = cpu_to_be64(page_list[i] | MTHCA_MTT_FLAG_PRESENT); - dma_sync_single(&dev->pdev->dev, fmr->mem.arbel.dma_handle, - list_len * sizeof(u64), DMA_TO_DEVICE); + dma_sync_single_for_device(&dev->pdev->dev, fmr->mem.arbel.dma_handle, + list_len * sizeof(u64), DMA_TO_DEVICE); fmr->mem.arbel.mpt->key = cpu_to_be32(key); fmr->mem.arbel.mpt->lkey = cpu_to_be32(key); @@ -908,7 +915,7 @@ int mthca_init_mr_table(struct mthca_dev *dev) dev->mr_table.mtt_base); dev->mr_table.tavor_fmr.mtt_base = - ioremap(addr, mtts * MTHCA_MTT_SEG_SIZE); + ioremap(addr, mtts * dev->limits.mtt_seg_size); if (!dev->mr_table.tavor_fmr.mtt_base) { mthca_warn(dev, "MTT ioremap for FMR failed.\n"); err = -ENOMEM; diff --git a/drivers/infiniband/hw/mthca/mthca_profile.c b/drivers/infiniband/hw/mthca/mthca_profile.c index d168c254061..8edb28a9a0e 100644 --- a/drivers/infiniband/hw/mthca/mthca_profile.c +++ b/drivers/infiniband/hw/mthca/mthca_profile.c @@ -94,7 +94,7 @@ s64 mthca_make_profile(struct mthca_dev *dev, profile[MTHCA_RES_RDB].size = MTHCA_RDB_ENTRY_SIZE; profile[MTHCA_RES_MCG].size = MTHCA_MGM_ENTRY_SIZE; profile[MTHCA_RES_MPT].size = dev_lim->mpt_entry_sz; - profile[MTHCA_RES_MTT].size = MTHCA_MTT_SEG_SIZE; + profile[MTHCA_RES_MTT].size = dev->limits.mtt_seg_size; profile[MTHCA_RES_UAR].size = dev_lim->uar_scratch_entry_sz; profile[MTHCA_RES_UDAV].size = MTHCA_AV_SIZE; profile[MTHCA_RES_UARC].size = request->uarc_size; @@ -232,7 +232,7 @@ s64 mthca_make_profile(struct mthca_dev *dev, dev->limits.num_mtt_segs = profile[i].num; dev->mr_table.mtt_base = profile[i].start; init_hca->mtt_base = profile[i].start; - init_hca->mtt_seg_sz = ffs(MTHCA_MTT_SEG_SIZE) - 7; + init_hca->mtt_seg_sz = ffs(dev->limits.mtt_seg_size) - 7; break; case MTHCA_RES_UAR: dev->limits.num_uars = profile[i].num; diff --git a/drivers/infiniband/hw/nes/nes.c b/drivers/infiniband/hw/nes/nes.c index b9611ade9ea..cbde0cfe27e 100644 --- a/drivers/infiniband/hw/nes/nes.c +++ b/drivers/infiniband/hw/nes/nes.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -478,23 +478,23 @@ static int __devinit nes_probe(struct pci_dev *pcidev, const struct pci_device_i } if ((sizeof(dma_addr_t) > 4)) { - ret = pci_set_dma_mask(pcidev, DMA_64BIT_MASK); + ret = pci_set_dma_mask(pcidev, DMA_BIT_MASK(64)); if (ret < 0) { printk(KERN_ERR PFX "64b DMA mask configuration failed\n"); goto bail2; } - ret = pci_set_consistent_dma_mask(pcidev, DMA_64BIT_MASK); + ret = pci_set_consistent_dma_mask(pcidev, DMA_BIT_MASK(64)); if (ret) { printk(KERN_ERR PFX "64b DMA consistent mask configuration failed\n"); goto bail2; } } else { - ret = pci_set_dma_mask(pcidev, DMA_32BIT_MASK); + ret = pci_set_dma_mask(pcidev, DMA_BIT_MASK(32)); if (ret < 0) { printk(KERN_ERR PFX "32b DMA mask configuration failed\n"); goto bail2; } - ret = pci_set_consistent_dma_mask(pcidev, DMA_32BIT_MASK); + ret = pci_set_consistent_dma_mask(pcidev, DMA_BIT_MASK(32)); if (ret) { printk(KERN_ERR PFX "32b DMA consistent mask configuration failed\n"); goto bail2; diff --git a/drivers/infiniband/hw/nes/nes.h b/drivers/infiniband/hw/nes/nes.h index 13a5bb1a7bc..bf1720f7f35 100644 --- a/drivers/infiniband/hw/nes/nes.h +++ b/drivers/infiniband/hw/nes/nes.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -56,10 +56,8 @@ #define QUEUE_DISCONNECTS -#define DRV_BUILD "1" - #define DRV_NAME "iw_nes" -#define DRV_VERSION "1.0 KO Build " DRV_BUILD +#define DRV_VERSION "1.5.0.0" #define PFX DRV_NAME ": " /* @@ -289,8 +287,8 @@ static inline __le32 get_crc_value(struct nes_v4_quad *nes_quad) static inline void set_wqe_64bit_value(__le32 *wqe_words, u32 index, u64 value) { - wqe_words[index] = cpu_to_le32((u32) ((unsigned long)value)); - wqe_words[index + 1] = cpu_to_le32((u32)(upper_32_bits((unsigned long)value))); + wqe_words[index] = cpu_to_le32((u32) value); + wqe_words[index + 1] = cpu_to_le32(upper_32_bits(value)); } static inline void diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c index 4a65b96db2c..114b802771a 100644 --- a/drivers/infiniband/hw/nes/nes_cm.c +++ b/drivers/infiniband/hw/nes/nes_cm.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -56,6 +56,7 @@ #include <net/neighbour.h> #include <net/route.h> #include <net/ip_fib.h> +#include <net/tcp.h> #include "nes.h" @@ -103,6 +104,7 @@ static int nes_disconnect(struct nes_qp *nesqp, int abrupt); static void nes_disconnect_worker(struct work_struct *work); static int send_mpa_request(struct nes_cm_node *, struct sk_buff *); +static int send_mpa_reject(struct nes_cm_node *); static int send_syn(struct nes_cm_node *, u32, struct sk_buff *); static int send_reset(struct nes_cm_node *, struct sk_buff *); static int send_ack(struct nes_cm_node *cm_node, struct sk_buff *skb); @@ -113,8 +115,7 @@ static void process_packet(struct nes_cm_node *, struct sk_buff *, static void active_open_err(struct nes_cm_node *, struct sk_buff *, int); static void passive_open_err(struct nes_cm_node *, struct sk_buff *, int); static void cleanup_retrans_entry(struct nes_cm_node *); -static void handle_rcv_mpa(struct nes_cm_node *, struct sk_buff *, - enum nes_cm_event_type); +static void handle_rcv_mpa(struct nes_cm_node *, struct sk_buff *); static void free_retrans_entry(struct nes_cm_node *cm_node); static int handle_tcp_options(struct nes_cm_node *cm_node, struct tcphdr *tcph, struct sk_buff *skb, int optionsize, int passive); @@ -124,6 +125,8 @@ static void cm_event_connected(struct nes_cm_event *); static void cm_event_connect_error(struct nes_cm_event *); static void cm_event_reset(struct nes_cm_event *); static void cm_event_mpa_req(struct nes_cm_event *); +static void cm_event_mpa_reject(struct nes_cm_event *); +static void handle_recv_entry(struct nes_cm_node *cm_node, u32 rem_node); static void print_core(struct nes_cm_core *core); @@ -196,7 +199,6 @@ static struct nes_cm_event *create_event(struct nes_cm_node *cm_node, */ static int send_mpa_request(struct nes_cm_node *cm_node, struct sk_buff *skb) { - int ret; if (!skb) { nes_debug(NES_DBG_CM, "skb set to NULL\n"); return -1; @@ -206,11 +208,27 @@ static int send_mpa_request(struct nes_cm_node *cm_node, struct sk_buff *skb) form_cm_frame(skb, cm_node, NULL, 0, &cm_node->mpa_frame, cm_node->mpa_frame_size, SET_ACK); - ret = schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 1, 0); - if (ret < 0) - return ret; + return schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 1, 0); +} - return 0; + + +static int send_mpa_reject(struct nes_cm_node *cm_node) +{ + struct sk_buff *skb = NULL; + + skb = dev_alloc_skb(MAX_CM_BUFFER); + if (!skb) { + nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); + return -ENOMEM; + } + + /* send an MPA reject frame */ + form_cm_frame(skb, cm_node, NULL, 0, &cm_node->mpa_frame, + cm_node->mpa_frame_size, SET_ACK | SET_FIN); + + cm_node->state = NES_CM_STATE_FIN_WAIT1; + return schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 1, 0); } @@ -218,14 +236,17 @@ static int send_mpa_request(struct nes_cm_node *cm_node, struct sk_buff *skb) * recv_mpa - process a received TCP pkt, we are expecting an * IETF MPA frame */ -static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 len) +static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 *type, + u32 len) { struct ietf_mpa_frame *mpa_frame; + *type = NES_MPA_REQUEST_ACCEPT; + /* assume req frame is in tcp data payload */ if (len < sizeof(struct ietf_mpa_frame)) { nes_debug(NES_DBG_CM, "The received ietf buffer was too small (%x)\n", len); - return -1; + return -EINVAL; } mpa_frame = (struct ietf_mpa_frame *)buffer; @@ -234,14 +255,25 @@ static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 len) if (cm_node->mpa_frame_size + sizeof(struct ietf_mpa_frame) != len) { nes_debug(NES_DBG_CM, "The received ietf buffer was not right" " complete (%x + %x != %x)\n", - cm_node->mpa_frame_size, (u32)sizeof(struct ietf_mpa_frame), len); - return -1; + cm_node->mpa_frame_size, + (u32)sizeof(struct ietf_mpa_frame), len); + return -EINVAL; + } + /* make sure it does not exceed the max size */ + if (len > MAX_CM_BUFFER) { + nes_debug(NES_DBG_CM, "The received ietf buffer was too large" + " (%x + %x != %x)\n", + cm_node->mpa_frame_size, + (u32)sizeof(struct ietf_mpa_frame), len); + return -EINVAL; } /* copy entire MPA frame to our cm_node's frame */ memcpy(cm_node->mpa_frame_buf, buffer + sizeof(struct ietf_mpa_frame), cm_node->mpa_frame_size); + if (mpa_frame->flags & IETF_MPA_FLAGS_REJECT) + *type = NES_MPA_REQUEST_REJECT; return 0; } @@ -380,7 +412,7 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb, new_send = kzalloc(sizeof(*new_send), GFP_ATOMIC); if (!new_send) - return -1; + return -ENOMEM; /* new_send->timetosend = currenttime */ new_send->retrycount = NES_DEFAULT_RETRYS; @@ -394,9 +426,12 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb, if (type == NES_TIMER_TYPE_CLOSE) { new_send->timetosend += (HZ/10); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - list_add_tail(&new_send->list, &cm_node->recv_list); - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); + if (cm_node->recv_entry) { + kfree(new_send); + WARN_ON(1); + return -EINVAL; + } + cm_node->recv_entry = new_send; } if (type == NES_TIMER_TYPE_SEND) { @@ -412,8 +447,8 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb, if (ret != NETDEV_TX_OK) { nes_debug(NES_DBG_CM, "Error sending packet %p " "(jiffies = %lu)\n", new_send, jiffies); - atomic_dec(&new_send->skb->users); new_send->timetosend = jiffies; + ret = NETDEV_TX_OK; } else { cm_packets_sent++; if (!send_retrans) { @@ -435,24 +470,82 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb, return ret; } +static void nes_retrans_expired(struct nes_cm_node *cm_node) +{ + struct iw_cm_id *cm_id = cm_node->cm_id; + switch (cm_node->state) { + case NES_CM_STATE_SYN_RCVD: + case NES_CM_STATE_CLOSING: + rem_ref_cm_node(cm_node->cm_core, cm_node); + break; + case NES_CM_STATE_LAST_ACK: + case NES_CM_STATE_FIN_WAIT1: + if (cm_node->cm_id) + cm_id->rem_ref(cm_id); + cm_node->state = NES_CM_STATE_CLOSED; + send_reset(cm_node, NULL); + break; + default: + create_event(cm_node, NES_CM_EVENT_ABORTED); + } +} + +static void handle_recv_entry(struct nes_cm_node *cm_node, u32 rem_node) +{ + struct nes_timer_entry *recv_entry = cm_node->recv_entry; + struct iw_cm_id *cm_id = cm_node->cm_id; + struct nes_qp *nesqp; + unsigned long qplockflags; + + if (!recv_entry) + return; + nesqp = (struct nes_qp *)recv_entry->skb; + if (nesqp) { + spin_lock_irqsave(&nesqp->lock, qplockflags); + if (nesqp->cm_id) { + nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, " + "refcount = %d: HIT A " + "NES_TIMER_TYPE_CLOSE with something " + "to do!!!\n", nesqp->hwqp.qp_id, cm_id, + atomic_read(&nesqp->refcount)); + nesqp->hw_tcp_state = NES_AEQE_TCP_STATE_CLOSED; + nesqp->last_aeq = NES_AEQE_AEID_RESET_SENT; + nesqp->ibqp_state = IB_QPS_ERR; + spin_unlock_irqrestore(&nesqp->lock, qplockflags); + nes_cm_disconn(nesqp); + } else { + spin_unlock_irqrestore(&nesqp->lock, qplockflags); + nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, " + "refcount = %d: HIT A " + "NES_TIMER_TYPE_CLOSE with nothing " + "to do!!!\n", nesqp->hwqp.qp_id, cm_id, + atomic_read(&nesqp->refcount)); + } + } else if (rem_node) { + /* TIME_WAIT state */ + rem_ref_cm_node(cm_node->cm_core, cm_node); + } + if (cm_node->cm_id) + cm_id->rem_ref(cm_id); + kfree(recv_entry); + cm_node->recv_entry = NULL; +} /** * nes_cm_timer_tick */ static void nes_cm_timer_tick(unsigned long pass) { - unsigned long flags, qplockflags; + unsigned long flags; unsigned long nexttimeout = jiffies + NES_LONG_TIME; - struct iw_cm_id *cm_id; struct nes_cm_node *cm_node; struct nes_timer_entry *send_entry, *recv_entry; - struct list_head *list_core, *list_core_temp; - struct list_head *list_node, *list_node_temp; + struct list_head *list_core_temp; + struct list_head *list_node; struct nes_cm_core *cm_core = g_cm_core; - struct nes_qp *nesqp; u32 settimer = 0; + unsigned long timetosend; int ret = NETDEV_TX_OK; - enum nes_cm_node_state last_state; struct list_head timer_list; INIT_LIST_HEAD(&timer_list); @@ -461,7 +554,7 @@ static void nes_cm_timer_tick(unsigned long pass) list_for_each_safe(list_node, list_core_temp, &cm_core->connected_nodes) { cm_node = container_of(list_node, struct nes_cm_node, list); - if (!list_empty(&cm_node->recv_list) || (cm_node->send_entry)) { + if ((cm_node->recv_entry) || (cm_node->send_entry)) { add_ref_cm_node(cm_node); list_add(&cm_node->timer_entry, &timer_list); } @@ -471,54 +564,18 @@ static void nes_cm_timer_tick(unsigned long pass) list_for_each_safe(list_node, list_core_temp, &timer_list) { cm_node = container_of(list_node, struct nes_cm_node, timer_entry); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - list_for_each_safe(list_core, list_node_temp, - &cm_node->recv_list) { - recv_entry = container_of(list_core, - struct nes_timer_entry, list); - if (!recv_entry) - break; + recv_entry = cm_node->recv_entry; + + if (recv_entry) { if (time_after(recv_entry->timetosend, jiffies)) { if (nexttimeout > recv_entry->timetosend || - !settimer) { + !settimer) { nexttimeout = recv_entry->timetosend; settimer = 1; } - continue; - } - list_del(&recv_entry->list); - cm_id = cm_node->cm_id; - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); - nesqp = (struct nes_qp *)recv_entry->skb; - spin_lock_irqsave(&nesqp->lock, qplockflags); - if (nesqp->cm_id) { - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, " - "refcount = %d: HIT A " - "NES_TIMER_TYPE_CLOSE with something " - "to do!!!\n", nesqp->hwqp.qp_id, cm_id, - atomic_read(&nesqp->refcount)); - nesqp->hw_tcp_state = NES_AEQE_TCP_STATE_CLOSED; - nesqp->last_aeq = NES_AEQE_AEID_RESET_SENT; - nesqp->ibqp_state = IB_QPS_ERR; - spin_unlock_irqrestore(&nesqp->lock, - qplockflags); - nes_cm_disconn(nesqp); - } else { - spin_unlock_irqrestore(&nesqp->lock, - qplockflags); - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, " - "refcount = %d: HIT A " - "NES_TIMER_TYPE_CLOSE with nothing " - "to do!!!\n", nesqp->hwqp.qp_id, cm_id, - atomic_read(&nesqp->refcount)); - } - if (cm_id) - cm_id->rem_ref(cm_id); - - kfree(recv_entry); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); + } else + handle_recv_entry(cm_node, 1); } - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); spin_lock_irqsave(&cm_node->retrans_list_lock, flags); do { @@ -533,12 +590,11 @@ static void nes_cm_timer_tick(unsigned long pass) nexttimeout = send_entry->timetosend; settimer = 1; - break; } } else { free_retrans_entry(cm_node); - break; } + break; } if ((cm_node->state == NES_CM_STATE_TSA) || @@ -550,16 +606,12 @@ static void nes_cm_timer_tick(unsigned long pass) if (!send_entry->retranscount || !send_entry->retrycount) { cm_packets_dropped++; - last_state = cm_node->state; - cm_node->state = NES_CM_STATE_CLOSED; free_retrans_entry(cm_node); + spin_unlock_irqrestore( &cm_node->retrans_list_lock, flags); - if (last_state == NES_CM_STATE_SYN_RCVD) - rem_ref_cm_node(cm_core, cm_node); - else - create_event(cm_node, - NES_CM_EVENT_ABORTED); + nes_retrans_expired(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; spin_lock_irqsave(&cm_node->retrans_list_lock, flags); break; @@ -584,7 +636,6 @@ static void nes_cm_timer_tick(unsigned long pass) nes_debug(NES_DBG_CM, "rexmit failed for " "node=%p\n", cm_node); cm_packets_bounced++; - atomic_dec(&send_entry->skb->users); send_entry->retrycount--; nexttimeout = jiffies + NES_SHORT_TIME; settimer = 1; @@ -598,8 +649,11 @@ static void nes_cm_timer_tick(unsigned long pass) send_entry->retrycount); if (send_entry->send_retrans) { send_entry->retranscount--; + timetosend = (NES_RETRY_TIMEOUT << + (NES_DEFAULT_RETRANS - send_entry->retranscount)); + send_entry->timetosend = jiffies + - NES_RETRY_TIMEOUT; + min(timetosend, NES_MAX_TIMEOUT); if (nexttimeout > send_entry->timetosend || !settimer) { nexttimeout = send_entry->timetosend; @@ -620,11 +674,6 @@ static void nes_cm_timer_tick(unsigned long pass) spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); rem_ref_cm_node(cm_node->cm_core, cm_node); - if (ret != NETDEV_TX_OK) { - nes_debug(NES_DBG_CM, "rexmit failed for cm_node=%p\n", - cm_node); - break; - } } if (settimer) { @@ -714,7 +763,7 @@ static int send_reset(struct nes_cm_node *cm_node, struct sk_buff *skb) skb = dev_alloc_skb(MAX_CM_BUFFER); if (!skb) { nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); - return -1; + return -ENOMEM; } form_cm_frame(skb, cm_node, NULL, 0, NULL, 0, flags); @@ -778,14 +827,10 @@ static struct nes_cm_node *find_node(struct nes_cm_core *cm_core, unsigned long flags; struct list_head *hte; struct nes_cm_node *cm_node; - __be32 tmp_addr = cpu_to_be32(loc_addr); /* get a handle on the hte */ hte = &cm_core->connected_nodes; - nes_debug(NES_DBG_CM, "Searching for an owner node: %pI4:%x from core %p->%p\n", - &tmp_addr, loc_port, cm_core, hte); - /* walk list and find cm_node associated with this session ID */ spin_lock_irqsave(&cm_core->ht_lock, flags); list_for_each_entry(cm_node, hte, list) { @@ -817,7 +862,6 @@ static struct nes_cm_listener *find_listener(struct nes_cm_core *cm_core, { unsigned long flags; struct nes_cm_listener *listen_node; - __be32 tmp_addr = cpu_to_be32(dst_addr); /* walk list and find cm_node associated with this session ID */ spin_lock_irqsave(&cm_core->listen_list_lock, flags); @@ -834,9 +878,6 @@ static struct nes_cm_listener *find_listener(struct nes_cm_core *cm_core, } spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); - nes_debug(NES_DBG_CM, "Unable to find listener for %pI4:%x\n", - &tmp_addr, dst_port); - /* no listener */ return NULL; } @@ -875,7 +916,8 @@ static int add_hte_node(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node static int mini_cm_dec_refcnt_listen(struct nes_cm_core *cm_core, struct nes_cm_listener *listener, int free_hanging_nodes) { - int ret = 1; + int ret = -EINVAL; + int err = 0; unsigned long flags; struct list_head *list_pos = NULL; struct list_head *list_temp = NULL; @@ -904,10 +946,60 @@ static int mini_cm_dec_refcnt_listen(struct nes_cm_core *cm_core, list_for_each_safe(list_pos, list_temp, &reset_list) { cm_node = container_of(list_pos, struct nes_cm_node, - reset_entry); - cleanup_retrans_entry(cm_node); - send_reset(cm_node, NULL); - rem_ref_cm_node(cm_node->cm_core, cm_node); + reset_entry); + { + struct nes_cm_node *loopback = cm_node->loopbackpartner; + if (NES_CM_STATE_FIN_WAIT1 <= cm_node->state) { + rem_ref_cm_node(cm_node->cm_core, cm_node); + } else { + if (!loopback) { + cleanup_retrans_entry(cm_node); + err = send_reset(cm_node, NULL); + if (err) { + cm_node->state = + NES_CM_STATE_CLOSED; + WARN_ON(1); + } else { + cm_node->state = + NES_CM_STATE_CLOSED; + rem_ref_cm_node( + cm_node->cm_core, + cm_node); + } + } else { + struct nes_cm_event event; + + event.cm_node = loopback; + event.cm_info.rem_addr = + loopback->rem_addr; + event.cm_info.loc_addr = + loopback->loc_addr; + event.cm_info.rem_port = + loopback->rem_port; + event.cm_info.loc_port = + loopback->loc_port; + event.cm_info.cm_id = loopback->cm_id; + cm_event_connect_error(&event); + loopback->state = NES_CM_STATE_CLOSED; + + event.cm_node = cm_node; + event.cm_info.rem_addr = + cm_node->rem_addr; + event.cm_info.loc_addr = + cm_node->loc_addr; + event.cm_info.rem_port = + cm_node->rem_port; + event.cm_info.loc_port = + cm_node->loc_port; + event.cm_info.cm_id = cm_node->cm_id; + cm_event_reset(&event); + + rem_ref_cm_node(cm_node->cm_core, + cm_node); + + } + } + } } spin_lock_irqsave(&cm_core->listen_list_lock, flags); @@ -968,6 +1060,7 @@ static inline int mini_cm_accelerated(struct nes_cm_core *cm_core, if (cm_node->accept_pend) { BUG_ON(!cm_node->listener); atomic_dec(&cm_node->listener->pend_accepts_cnt); + cm_node->accept_pend = 0; BUG_ON(atomic_read(&cm_node->listener->pend_accepts_cnt) < 0); } @@ -994,7 +1087,7 @@ static int nes_addr_resolve_neigh(struct nes_vnic *nesvnic, u32 dst_ip) memset(&fl, 0, sizeof fl); fl.nl_u.ip4_u.daddr = htonl(dst_ip); if (ip_route_output_key(&init_net, &rt, &fl)) { - printk("%s: ip_route_output_key failed for 0x%08X\n", + printk(KERN_ERR "%s: ip_route_output_key failed for 0x%08X\n", __func__, dst_ip); return rc; } @@ -1057,8 +1150,6 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core, cm_node->cm_id); spin_lock_init(&cm_node->retrans_list_lock); - INIT_LIST_HEAD(&cm_node->recv_list); - spin_lock_init(&cm_node->recv_list_lock); cm_node->loopbackpartner = NULL; atomic_set(&cm_node->ref_count, 1); @@ -1126,10 +1217,7 @@ static int add_ref_cm_node(struct nes_cm_node *cm_node) static int rem_ref_cm_node(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) { - unsigned long flags, qplockflags; - struct nes_timer_entry *recv_entry; - struct iw_cm_id *cm_id; - struct list_head *list_core, *list_node_temp; + unsigned long flags; struct nes_qp *nesqp; if (!cm_node) @@ -1150,38 +1238,9 @@ static int rem_ref_cm_node(struct nes_cm_core *cm_core, atomic_dec(&cm_node->listener->pend_accepts_cnt); BUG_ON(atomic_read(&cm_node->listener->pend_accepts_cnt) < 0); } - BUG_ON(cm_node->send_entry); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - list_for_each_safe(list_core, list_node_temp, &cm_node->recv_list) { - recv_entry = container_of(list_core, struct nes_timer_entry, - list); - list_del(&recv_entry->list); - cm_id = cm_node->cm_id; - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); - nesqp = (struct nes_qp *)recv_entry->skb; - spin_lock_irqsave(&nesqp->lock, qplockflags); - if (nesqp->cm_id) { - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p: HIT A " - "NES_TIMER_TYPE_CLOSE with something to do!\n", - nesqp->hwqp.qp_id, cm_id); - nesqp->hw_tcp_state = NES_AEQE_TCP_STATE_CLOSED; - nesqp->last_aeq = NES_AEQE_AEID_RESET_SENT; - nesqp->ibqp_state = IB_QPS_ERR; - spin_unlock_irqrestore(&nesqp->lock, qplockflags); - nes_cm_disconn(nesqp); - } else { - spin_unlock_irqrestore(&nesqp->lock, qplockflags); - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p: HIT A " - "NES_TIMER_TYPE_CLOSE with nothing to do!\n", - nesqp->hwqp.qp_id, cm_id); - } - cm_id->rem_ref(cm_id); - - kfree(recv_entry); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - } - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); - + WARN_ON(cm_node->send_entry); + if (cm_node->recv_entry) + handle_recv_entry(cm_node, 0); if (cm_node->listener) { mini_cm_dec_refcnt_listen(cm_core, cm_node->listener, 0); } else { @@ -1202,7 +1261,6 @@ static int rem_ref_cm_node(struct nes_cm_core *cm_core, cm_node->nesqp = NULL; } - cm_node->freed = 1; kfree(cm_node); return 0; } @@ -1266,36 +1324,48 @@ static void drop_packet(struct sk_buff *skb) dev_kfree_skb_any(skb); } -static void handle_fin_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, - struct tcphdr *tcph) +static void handle_fin_pkt(struct nes_cm_node *cm_node) { nes_debug(NES_DBG_CM, "Received FIN, cm_node = %p, state = %u. " "refcnt=%d\n", cm_node, cm_node->state, atomic_read(&cm_node->ref_count)); - cm_node->tcp_cntxt.rcv_nxt++; - cleanup_retrans_entry(cm_node); switch (cm_node->state) { case NES_CM_STATE_SYN_RCVD: case NES_CM_STATE_SYN_SENT: case NES_CM_STATE_ESTABLISHED: case NES_CM_STATE_MPAREQ_SENT: + case NES_CM_STATE_MPAREJ_RCVD: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); cm_node->state = NES_CM_STATE_LAST_ACK; - send_fin(cm_node, skb); + send_fin(cm_node, NULL); break; case NES_CM_STATE_FIN_WAIT1: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); cm_node->state = NES_CM_STATE_CLOSING; - send_ack(cm_node, skb); + send_ack(cm_node, NULL); + /* Wait for ACK as this is simultanous close.. + * After we receive ACK, do not send anything.. + * Just rm the node.. Done.. */ break; case NES_CM_STATE_FIN_WAIT2: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); cm_node->state = NES_CM_STATE_TIME_WAIT; - send_ack(cm_node, skb); + send_ack(cm_node, NULL); + schedule_nes_timer(cm_node, NULL, NES_TIMER_TYPE_CLOSE, 1, 0); + break; + case NES_CM_STATE_TIME_WAIT: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); cm_node->state = NES_CM_STATE_CLOSED; + rem_ref_cm_node(cm_node->cm_core, cm_node); break; case NES_CM_STATE_TSA: default: nes_debug(NES_DBG_CM, "Error Rcvd FIN for node-%p state = %d\n", cm_node, cm_node->state); - drop_packet(skb); break; } } @@ -1324,7 +1394,6 @@ static void handle_rst_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, passive_state = atomic_add_return(1, &cm_node->passive_state); if (passive_state == NES_SEND_RESET_EVENT) create_event(cm_node, NES_CM_EVENT_RESET); - cleanup_retrans_entry(cm_node); cm_node->state = NES_CM_STATE_CLOSED; dev_kfree_skb_any(skb); break; @@ -1338,7 +1407,14 @@ static void handle_rst_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, active_open_err(cm_node, skb, reset); break; case NES_CM_STATE_CLOSED: - cleanup_retrans_entry(cm_node); + drop_packet(skb); + break; + case NES_CM_STATE_FIN_WAIT1: + case NES_CM_STATE_LAST_ACK: + cm_node->cm_id->rem_ref(cm_node->cm_id); + case NES_CM_STATE_TIME_WAIT: + cm_node->state = NES_CM_STATE_CLOSED; + rem_ref_cm_node(cm_node->cm_core, cm_node); drop_packet(skb); break; default: @@ -1347,17 +1423,20 @@ static void handle_rst_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, } } -static void handle_rcv_mpa(struct nes_cm_node *cm_node, struct sk_buff *skb, - enum nes_cm_event_type type) + +static void handle_rcv_mpa(struct nes_cm_node *cm_node, struct sk_buff *skb) { - int ret; + int ret = 0; int datasize = skb->len; u8 *dataloc = skb->data; - ret = parse_mpa(cm_node, dataloc, datasize); - if (ret < 0) { + + enum nes_cm_event_type type = NES_CM_EVENT_UNKNOWN; + u32 res_type; + ret = parse_mpa(cm_node, dataloc, &res_type, datasize); + if (ret) { nes_debug(NES_DBG_CM, "didn't like MPA Request\n"); - if (type == NES_CM_EVENT_CONNECTED) { + if (cm_node->state == NES_CM_STATE_MPAREQ_SENT) { nes_debug(NES_DBG_CM, "%s[%u] create abort for " "cm_node=%p listener=%p state=%d\n", __func__, __LINE__, cm_node, cm_node->listener, @@ -1366,18 +1445,39 @@ static void handle_rcv_mpa(struct nes_cm_node *cm_node, struct sk_buff *skb, } else { passive_open_err(cm_node, skb, 1); } - } else { + return; + } + + switch (cm_node->state) { + case NES_CM_STATE_ESTABLISHED: + if (res_type == NES_MPA_REQUEST_REJECT) { + /*BIG problem as we are receiving the MPA.. So should + * not be REJECT.. This is Passive Open.. We can + * only receive it Reject for Active Open...*/ + WARN_ON(1); + } + cm_node->state = NES_CM_STATE_MPAREQ_RCVD; + type = NES_CM_EVENT_MPA_REQ; + atomic_set(&cm_node->passive_state, + NES_PASSIVE_STATE_INDICATED); + break; + case NES_CM_STATE_MPAREQ_SENT: cleanup_retrans_entry(cm_node); - dev_kfree_skb_any(skb); - if (type == NES_CM_EVENT_CONNECTED) + if (res_type == NES_MPA_REQUEST_REJECT) { + type = NES_CM_EVENT_MPA_REJECT; + cm_node->state = NES_CM_STATE_MPAREJ_RCVD; + } else { + type = NES_CM_EVENT_CONNECTED; cm_node->state = NES_CM_STATE_TSA; - else - atomic_set(&cm_node->passive_state, - NES_PASSIVE_STATE_INDICATED); - create_event(cm_node, type); + } + break; + default: + WARN_ON(1); + break; } - return ; + dev_kfree_skb_any(skb); + create_event(cm_node, type); } static void indicate_pkt_err(struct nes_cm_node *cm_node, struct sk_buff *skb) @@ -1426,7 +1526,7 @@ static int check_seq(struct nes_cm_node *cm_node, struct tcphdr *tcph, rcv_wnd = cm_node->tcp_cntxt.rcv_wnd; if (ack_seq != loc_seq_num) err = 1; - else if ((seq + rcv_wnd) < rcv_nxt) + else if (!between(seq, rcv_nxt, (rcv_nxt+rcv_wnd))) err = 1; if (err) { nes_debug(NES_DBG_CM, "%s[%u] create abort for cm_node=%p " @@ -1465,8 +1565,6 @@ static void handle_syn_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, break; case NES_CM_STATE_LISTENING: /* Passive OPEN */ - cm_node->accept_pend = 1; - atomic_inc(&cm_node->listener->pend_accepts_cnt); if (atomic_read(&cm_node->listener->pend_accepts_cnt) > cm_node->listener->backlog) { nes_debug(NES_DBG_CM, "drop syn due to backlog " @@ -1484,6 +1582,9 @@ static void handle_syn_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, } cm_node->tcp_cntxt.rcv_nxt = inc_sequence + 1; BUG_ON(cm_node->send_entry); + cm_node->accept_pend = 1; + atomic_inc(&cm_node->listener->pend_accepts_cnt); + cm_node->state = NES_CM_STATE_SYN_RCVD; send_syn(cm_node, 1, skb); break; @@ -1518,6 +1619,7 @@ static void handle_synack_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, inc_sequence = ntohl(tcph->seq); switch (cm_node->state) { case NES_CM_STATE_SYN_SENT: + cleanup_retrans_entry(cm_node); /* active open */ if (check_syn(cm_node, tcph, skb)) return; @@ -1558,62 +1660,46 @@ static void handle_synack_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, } } -static void handle_ack_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, +static int handle_ack_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, struct tcphdr *tcph) { int datasize = 0; u32 inc_sequence; u32 rem_seq_ack; u32 rem_seq; - int ret; + int ret = 0; int optionsize; - u32 temp_seq = cm_node->tcp_cntxt.loc_seq_num; - optionsize = (tcph->doff << 2) - sizeof(struct tcphdr); - cm_node->tcp_cntxt.loc_seq_num = ntohl(tcph->ack_seq); if (check_seq(cm_node, tcph, skb)) - return; + return -EINVAL; skb_pull(skb, tcph->doff << 2); inc_sequence = ntohl(tcph->seq); rem_seq = ntohl(tcph->seq); rem_seq_ack = ntohl(tcph->ack_seq); datasize = skb->len; - switch (cm_node->state) { case NES_CM_STATE_SYN_RCVD: /* Passive OPEN */ + cleanup_retrans_entry(cm_node); ret = handle_tcp_options(cm_node, tcph, skb, optionsize, 1); if (ret) break; cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->ack_seq); - cm_node->tcp_cntxt.loc_seq_num = temp_seq; - if (cm_node->tcp_cntxt.rem_ack_num != - cm_node->tcp_cntxt.loc_seq_num) { - nes_debug(NES_DBG_CM, "rem_ack_num != loc_seq_num\n"); - cleanup_retrans_entry(cm_node); - send_reset(cm_node, skb); - return; - } cm_node->state = NES_CM_STATE_ESTABLISHED; if (datasize) { cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; - cm_node->state = NES_CM_STATE_MPAREQ_RCVD; - handle_rcv_mpa(cm_node, skb, NES_CM_EVENT_MPA_REQ); - } else { /* rcvd ACK only */ + handle_rcv_mpa(cm_node, skb); + } else /* rcvd ACK only */ dev_kfree_skb_any(skb); - cleanup_retrans_entry(cm_node); - } break; case NES_CM_STATE_ESTABLISHED: /* Passive OPEN */ - /* We expect mpa frame to be received only */ + cleanup_retrans_entry(cm_node); if (datasize) { cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; - cm_node->state = NES_CM_STATE_MPAREQ_RCVD; - handle_rcv_mpa(cm_node, skb, - NES_CM_EVENT_MPA_REQ); + handle_rcv_mpa(cm_node, skb); } else drop_packet(skb); break; @@ -1621,29 +1707,39 @@ static void handle_ack_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->ack_seq); if (datasize) { cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; - handle_rcv_mpa(cm_node, skb, NES_CM_EVENT_CONNECTED); - } else { /* Could be just an ack pkt.. */ - cleanup_retrans_entry(cm_node); + handle_rcv_mpa(cm_node, skb); + } else /* Could be just an ack pkt.. */ dev_kfree_skb_any(skb); - } break; case NES_CM_STATE_LISTENING: case NES_CM_STATE_CLOSED: cleanup_retrans_entry(cm_node); send_reset(cm_node, skb); break; + case NES_CM_STATE_LAST_ACK: + case NES_CM_STATE_CLOSING: + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + cm_node->cm_id->rem_ref(cm_node->cm_id); + rem_ref_cm_node(cm_node->cm_core, cm_node); + drop_packet(skb); + break; case NES_CM_STATE_FIN_WAIT1: + cleanup_retrans_entry(cm_node); + drop_packet(skb); + cm_node->state = NES_CM_STATE_FIN_WAIT2; + break; case NES_CM_STATE_SYN_SENT: case NES_CM_STATE_FIN_WAIT2: case NES_CM_STATE_TSA: case NES_CM_STATE_MPAREQ_RCVD: - case NES_CM_STATE_LAST_ACK: - case NES_CM_STATE_CLOSING: case NES_CM_STATE_UNKNOWN: default: + cleanup_retrans_entry(cm_node); drop_packet(skb); break; } + return ret; } @@ -1748,6 +1844,8 @@ static void process_packet(struct nes_cm_node *cm_node, struct sk_buff *skb, { enum nes_tcpip_pkt_type pkt_type = NES_PKT_TYPE_UNKNOWN; struct tcphdr *tcph = tcp_hdr(skb); + u32 fin_set = 0; + int ret = 0; skb_pull(skb, ip_hdr(skb)->ihl << 2); nes_debug(NES_DBG_CM, "process_packet: cm_node=%p state =%d syn=%d " @@ -1760,10 +1858,10 @@ static void process_packet(struct nes_cm_node *cm_node, struct sk_buff *skb, pkt_type = NES_PKT_TYPE_SYN; if (tcph->ack) pkt_type = NES_PKT_TYPE_SYNACK; - } else if (tcph->fin) - pkt_type = NES_PKT_TYPE_FIN; - else if (tcph->ack) + } else if (tcph->ack) pkt_type = NES_PKT_TYPE_ACK; + if (tcph->fin) + fin_set = 1; switch (pkt_type) { case NES_PKT_TYPE_SYN: @@ -1773,15 +1871,16 @@ static void process_packet(struct nes_cm_node *cm_node, struct sk_buff *skb, handle_synack_pkt(cm_node, skb, tcph); break; case NES_PKT_TYPE_ACK: - handle_ack_pkt(cm_node, skb, tcph); + ret = handle_ack_pkt(cm_node, skb, tcph); + if (fin_set && !ret) + handle_fin_pkt(cm_node); break; case NES_PKT_TYPE_RST: handle_rst_pkt(cm_node, skb, tcph); break; - case NES_PKT_TYPE_FIN: - handle_fin_pkt(cm_node, skb, tcph); - break; default: + if ((fin_set) && (!check_seq(cm_node, tcph, skb))) + handle_fin_pkt(cm_node); drop_packet(skb); break; } @@ -1891,13 +1990,17 @@ static struct nes_cm_node *mini_cm_connect(struct nes_cm_core *cm_core, if (loopbackremotelistener == NULL) { create_event(cm_node, NES_CM_EVENT_ABORTED); } else { - atomic_inc(&cm_loopbacks); loopback_cm_info = *cm_info; loopback_cm_info.loc_port = cm_info->rem_port; loopback_cm_info.rem_port = cm_info->loc_port; loopback_cm_info.cm_id = loopbackremotelistener->cm_id; loopbackremotenode = make_cm_node(cm_core, nesvnic, &loopback_cm_info, loopbackremotelistener); + if (!loopbackremotenode) { + rem_ref_cm_node(cm_node->cm_core, cm_node); + return NULL; + } + atomic_inc(&cm_loopbacks); loopbackremotenode->loopbackpartner = cm_node; loopbackremotenode->tcp_cntxt.rcv_wscale = NES_CM_DEFAULT_RCV_WND_SCALE; @@ -1925,7 +2028,7 @@ static struct nes_cm_node *mini_cm_connect(struct nes_cm_core *cm_core, loopbackremotenode->tcp_cntxt.rcv_wscale; loopbackremotenode->tcp_cntxt.snd_wscale = cm_node->tcp_cntxt.rcv_wscale; - + loopbackremotenode->state = NES_CM_STATE_MPAREQ_RCVD; create_event(loopbackremotenode, NES_CM_EVENT_MPA_REQ); } return cm_node; @@ -1980,7 +2083,11 @@ static int mini_cm_reject(struct nes_cm_core *cm_core, struct ietf_mpa_frame *mpa_frame, struct nes_cm_node *cm_node) { int ret = 0; + int err = 0; int passive_state; + struct nes_cm_event event; + struct iw_cm_id *cm_id = cm_node->cm_id; + struct nes_cm_node *loopback = cm_node->loopbackpartner; nes_debug(NES_DBG_CM, "%s cm_node=%p type=%d state=%d\n", __func__, cm_node, cm_node->tcp_cntxt.client, cm_node->state); @@ -1989,12 +2096,38 @@ static int mini_cm_reject(struct nes_cm_core *cm_core, return ret; cleanup_retrans_entry(cm_node); - passive_state = atomic_add_return(1, &cm_node->passive_state); - cm_node->state = NES_CM_STATE_CLOSED; - if (passive_state == NES_SEND_RESET_EVENT) + if (!loopback) { + passive_state = atomic_add_return(1, &cm_node->passive_state); + if (passive_state == NES_SEND_RESET_EVENT) { + cm_node->state = NES_CM_STATE_CLOSED; + rem_ref_cm_node(cm_core, cm_node); + } else { + ret = send_mpa_reject(cm_node); + if (ret) { + cm_node->state = NES_CM_STATE_CLOSED; + err = send_reset(cm_node, NULL); + if (err) + WARN_ON(1); + } else + cm_id->add_ref(cm_id); + } + } else { + cm_node->cm_id = NULL; + event.cm_node = loopback; + event.cm_info.rem_addr = loopback->rem_addr; + event.cm_info.loc_addr = loopback->loc_addr; + event.cm_info.rem_port = loopback->rem_port; + event.cm_info.loc_port = loopback->loc_port; + event.cm_info.cm_id = loopback->cm_id; + cm_event_mpa_reject(&event); rem_ref_cm_node(cm_core, cm_node); - else - ret = send_reset(cm_node, NULL); + loopback->state = NES_CM_STATE_CLOSING; + + cm_id = loopback->cm_id; + rem_ref_cm_node(cm_core, loopback); + cm_id->rem_ref(cm_id); + } + return ret; } @@ -2031,6 +2164,7 @@ static int mini_cm_close(struct nes_cm_core *cm_core, struct nes_cm_node *cm_nod case NES_CM_STATE_CLOSING: ret = -1; break; + case NES_CM_STATE_MPAREJ_RCVD: case NES_CM_STATE_LISTENING: case NES_CM_STATE_UNKNOWN: case NES_CM_STATE_INITED: @@ -2227,15 +2361,15 @@ static int mini_cm_set(struct nes_cm_core *cm_core, u32 type, u32 value) int ret = 0; switch (type) { - case NES_CM_SET_PKT_SIZE: - cm_core->mtu = value; - break; - case NES_CM_SET_FREE_PKT_Q_SIZE: - cm_core->free_tx_pkt_max = value; - break; - default: - /* unknown set option */ - ret = -EINVAL; + case NES_CM_SET_PKT_SIZE: + cm_core->mtu = value; + break; + case NES_CM_SET_FREE_PKT_Q_SIZE: + cm_core->free_tx_pkt_max = value; + break; + default: + /* unknown set option */ + ret = -EINVAL; } return ret; @@ -2551,6 +2685,7 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) struct ib_mr *ibmr = NULL; struct ib_phys_buf ibphysbuf; struct nes_pd *nespd; + u64 tagged_offset; @@ -2572,7 +2707,6 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) /* associate the node with the QP */ nesqp->cm_node = (void *)cm_node; cm_node->nesqp = nesqp; - nes_add_ref(&nesqp->ibqp); nes_debug(NES_DBG_CM, "QP%u, cm_node=%p, jiffies = %lu listener = %p\n", nesqp->hwqp.qp_id, cm_node, jiffies, cm_node->listener); @@ -2616,14 +2750,18 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) ibphysbuf.addr = nesqp->ietf_frame_pbase; ibphysbuf.size = conn_param->private_data_len + sizeof(struct ietf_mpa_frame); + tagged_offset = (u64)(unsigned long)nesqp->ietf_frame; ibmr = nesibdev->ibdev.reg_phys_mr((struct ib_pd *)nespd, &ibphysbuf, 1, IB_ACCESS_LOCAL_WRITE, - (u64 *)&nesqp->ietf_frame); + &tagged_offset); if (!ibmr) { nes_debug(NES_DBG_CM, "Unable to register memory region" "for lSMM for cm_node = %p \n", cm_node); + pci_free_consistent(nesdev->pcidev, + nesqp->private_data_len+sizeof(struct ietf_mpa_frame), + nesqp->ietf_frame, nesqp->ietf_frame_pbase); return -ENOMEM; } @@ -2643,7 +2781,7 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) sizeof(struct ietf_mpa_frame)); set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_FRAG0_LOW_IDX, - (u64)nesqp->ietf_frame); + (u64)(unsigned long)nesqp->ietf_frame); wqe->wqe_words[NES_IWARP_SQ_WQE_LENGTH0_IDX] = cpu_to_le32(conn_param->private_data_len + sizeof(struct ietf_mpa_frame)); @@ -2654,9 +2792,7 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) NES_QPCONTEXT_ORDIRD_WRPDU); } else { nesqp->nesqp_context->ird_ord_sizes |= - cpu_to_le32((NES_QPCONTEXT_ORDIRD_LSMM_PRESENT | - NES_QPCONTEXT_ORDIRD_WRPDU | - NES_QPCONTEXT_ORDIRD_ALSMM)); + cpu_to_le32(NES_QPCONTEXT_ORDIRD_WRPDU); } nesqp->skip_lsmm = 1; @@ -2742,6 +2878,7 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) /* notify OF layer that accept event was successful */ cm_id->add_ref(cm_id); + nes_add_ref(&nesqp->ibqp); cm_event.event = IW_CM_EVENT_ESTABLISHED; cm_event.status = IW_CM_EVENT_STATUS_ACCEPTED; @@ -2778,23 +2915,35 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) int nes_reject(struct iw_cm_id *cm_id, const void *pdata, u8 pdata_len) { struct nes_cm_node *cm_node; + struct nes_cm_node *loopback; + struct nes_cm_core *cm_core; atomic_inc(&cm_rejects); cm_node = (struct nes_cm_node *) cm_id->provider_data; + loopback = cm_node->loopbackpartner; cm_core = cm_node->cm_core; + cm_node->cm_id = cm_id; cm_node->mpa_frame_size = sizeof(struct ietf_mpa_frame) + pdata_len; + if (cm_node->mpa_frame_size > MAX_CM_BUFFER) + return -EINVAL; + strcpy(&cm_node->mpa_frame.key[0], IEFT_MPA_KEY_REP); - memcpy(&cm_node->mpa_frame.priv_data, pdata, pdata_len); + if (loopback) { + memcpy(&loopback->mpa_frame.priv_data, pdata, pdata_len); + loopback->mpa_frame.priv_data_len = pdata_len; + loopback->mpa_frame_size = sizeof(struct ietf_mpa_frame) + + pdata_len; + } else { + memcpy(&cm_node->mpa_frame.priv_data, pdata, pdata_len); + cm_node->mpa_frame.priv_data_len = cpu_to_be16(pdata_len); + } - cm_node->mpa_frame.priv_data_len = cpu_to_be16(pdata_len); cm_node->mpa_frame.rev = mpa_version; cm_node->mpa_frame.flags = IETF_MPA_FLAGS_CRC | IETF_MPA_FLAGS_REJECT; - cm_core->api->reject(cm_core, &cm_node->mpa_frame, cm_node); - - return 0; + return cm_core->api->reject(cm_core, &cm_node->mpa_frame, cm_node); } @@ -2810,6 +2959,7 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) struct nes_device *nesdev; struct nes_cm_node *cm_node; struct nes_cm_info cm_info; + int apbvt_set = 0; ibqp = nes_get_qp(cm_id->device, conn_param->qpn); if (!ibqp) @@ -2847,9 +2997,11 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) conn_param->private_data_len); if (cm_id->local_addr.sin_addr.s_addr != - cm_id->remote_addr.sin_addr.s_addr) + cm_id->remote_addr.sin_addr.s_addr) { nes_manage_apbvt(nesvnic, ntohs(cm_id->local_addr.sin_port), PCI_FUNC(nesdev->pcidev->devfn), NES_MANAGE_APBVT_ADD); + apbvt_set = 1; + } /* set up the connection params for the node */ cm_info.loc_addr = htonl(cm_id->local_addr.sin_addr.s_addr); @@ -2866,8 +3018,7 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) conn_param->private_data_len, (void *)conn_param->private_data, &cm_info); if (!cm_node) { - if (cm_id->local_addr.sin_addr.s_addr != - cm_id->remote_addr.sin_addr.s_addr) + if (apbvt_set) nes_manage_apbvt(nesvnic, ntohs(cm_id->local_addr.sin_port), PCI_FUNC(nesdev->pcidev->devfn), NES_MANAGE_APBVT_DEL); @@ -2876,7 +3027,7 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) return -ENOMEM; } - cm_node->apbvt_set = 1; + cm_node->apbvt_set = apbvt_set; nesqp->cm_node = cm_node; cm_node->nesqp = nesqp; nes_add_ref(&nesqp->ibqp); @@ -3303,13 +3454,56 @@ static void cm_event_mpa_req(struct nes_cm_event *event) cm_event.remote_addr.sin_family = AF_INET; cm_event.remote_addr.sin_port = htons(event->cm_info.rem_port); cm_event.remote_addr.sin_addr.s_addr = htonl(event->cm_info.rem_addr); + cm_event.private_data = cm_node->mpa_frame_buf; + cm_event.private_data_len = (u8) cm_node->mpa_frame_size; - cm_event.private_data = cm_node->mpa_frame_buf; - cm_event.private_data_len = (u8) cm_node->mpa_frame_size; + ret = cm_id->event_handler(cm_id, &cm_event); + if (ret) + printk(KERN_ERR "%s[%u] OFA CM event_handler returned, ret=%d\n", + __func__, __LINE__, ret); + return; +} + + +static void cm_event_mpa_reject(struct nes_cm_event *event) +{ + struct iw_cm_id *cm_id; + struct iw_cm_event cm_event; + struct nes_cm_node *cm_node; + int ret; + + cm_node = event->cm_node; + if (!cm_node) + return; + cm_id = cm_node->cm_id; + + atomic_inc(&cm_connect_reqs); + nes_debug(NES_DBG_CM, "cm_node = %p - cm_id = %p, jiffies = %lu\n", + cm_node, cm_id, jiffies); + + cm_event.event = IW_CM_EVENT_CONNECT_REPLY; + cm_event.status = -ECONNREFUSED; + cm_event.provider_data = cm_id->provider_data; + + cm_event.local_addr.sin_family = AF_INET; + cm_event.local_addr.sin_port = htons(event->cm_info.loc_port); + cm_event.local_addr.sin_addr.s_addr = htonl(event->cm_info.loc_addr); + + cm_event.remote_addr.sin_family = AF_INET; + cm_event.remote_addr.sin_port = htons(event->cm_info.rem_port); + cm_event.remote_addr.sin_addr.s_addr = htonl(event->cm_info.rem_addr); + + cm_event.private_data = cm_node->mpa_frame_buf; + cm_event.private_data_len = (u8) cm_node->mpa_frame_size; + + nes_debug(NES_DBG_CM, "call CM_EVENT_MPA_REJECTED, local_addr=%08x, " + "remove_addr=%08x\n", + cm_event.local_addr.sin_addr.s_addr, + cm_event.remote_addr.sin_addr.s_addr); ret = cm_id->event_handler(cm_id, &cm_event); if (ret) - printk("%s[%u] OFA CM event_handler returned, ret=%d\n", + printk(KERN_ERR "%s[%u] OFA CM event_handler returned, ret=%d\n", __func__, __LINE__, ret); return; @@ -3374,6 +3568,14 @@ static void nes_cm_event_handler(struct work_struct *work) cm_event_connected(event); nes_debug(NES_DBG_CM, "CM Event: CONNECTED\n"); break; + case NES_CM_EVENT_MPA_REJECT: + if ((!event->cm_node->cm_id) || + (event->cm_node->state == NES_CM_STATE_TSA)) + break; + cm_event_mpa_reject(event); + nes_debug(NES_DBG_CM, "CM Event: REJECT\n"); + break; + case NES_CM_EVENT_ABORTED: if ((!event->cm_node->cm_id) || (event->cm_node->state == NES_CM_STATE_TSA)) diff --git a/drivers/infiniband/hw/nes/nes_cm.h b/drivers/infiniband/hw/nes/nes_cm.h index fafa35042eb..8b7e7c0e496 100644 --- a/drivers/infiniband/hw/nes/nes_cm.h +++ b/drivers/infiniband/hw/nes/nes_cm.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -39,6 +39,9 @@ #define NES_MANAGE_APBVT_DEL 0 #define NES_MANAGE_APBVT_ADD 1 +#define NES_MPA_REQUEST_ACCEPT 1 +#define NES_MPA_REQUEST_REJECT 2 + /* IETF MPA -- defines, enums, structs */ #define IEFT_MPA_KEY_REQ "MPA ID Req Frame" #define IEFT_MPA_KEY_REP "MPA ID Rep Frame" @@ -146,6 +149,7 @@ struct nes_timer_entry { #endif #define NES_SHORT_TIME (10) #define NES_LONG_TIME (2000*HZ/1000) +#define NES_MAX_TIMEOUT ((unsigned long) (12*HZ)) #define NES_CM_HASHTABLE_SIZE 1024 #define NES_CM_TCP_TIMER_INTERVAL 3000 @@ -186,6 +190,7 @@ enum nes_cm_node_state { NES_CM_STATE_ACCEPTING, NES_CM_STATE_MPAREQ_SENT, NES_CM_STATE_MPAREQ_RCVD, + NES_CM_STATE_MPAREJ_RCVD, NES_CM_STATE_TSA, NES_CM_STATE_FIN_WAIT1, NES_CM_STATE_FIN_WAIT2, @@ -278,13 +283,12 @@ struct nes_cm_node { struct nes_timer_entry *send_entry; spinlock_t retrans_list_lock; - struct list_head recv_list; - spinlock_t recv_list_lock; + struct nes_timer_entry *recv_entry; int send_write0; union { struct ietf_mpa_frame mpa_frame; - u8 mpa_frame_buf[NES_CM_DEFAULT_MTU]; + u8 mpa_frame_buf[MAX_CM_BUFFER]; }; u16 mpa_frame_size; struct iw_cm_id *cm_id; @@ -295,7 +299,6 @@ struct nes_cm_node { struct nes_vnic *nesvnic; int apbvt_set; int accept_pend; - int freed; struct list_head timer_entry; struct list_head reset_entry; struct nes_qp *nesqp; @@ -326,6 +329,7 @@ enum nes_cm_event_type { NES_CM_EVENT_MPA_REQ, NES_CM_EVENT_MPA_CONNECT, NES_CM_EVENT_MPA_ACCEPT, + NES_CM_EVENT_MPA_REJECT, NES_CM_EVENT_MPA_ESTABLISHED, NES_CM_EVENT_CONNECTED, NES_CM_EVENT_CLOSED, diff --git a/drivers/infiniband/hw/nes/nes_context.h b/drivers/infiniband/hw/nes/nes_context.h index da9daba8e66..0fb8d81d9a6 100644 --- a/drivers/infiniband/hw/nes/nes_context.h +++ b/drivers/infiniband/hw/nes/nes_context.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU diff --git a/drivers/infiniband/hw/nes/nes_hw.c b/drivers/infiniband/hw/nes/nes_hw.c index 5d139db1b77..4a84d02ece0 100644 --- a/drivers/infiniband/hw/nes/nes_hw.c +++ b/drivers/infiniband/hw/nes/nes_hw.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -46,6 +46,10 @@ static unsigned int nes_lro_max_aggr = NES_LRO_MAX_AGGR; module_param(nes_lro_max_aggr, uint, 0444); MODULE_PARM_DESC(nes_lro_max_aggr, "NIC LRO max packet aggregation"); +static int wide_ppm_offset; +module_param(wide_ppm_offset, int, 0644); +MODULE_PARM_DESC(wide_ppm_offset, "Increase CX4 interface clock ppm offset, 0=100ppm (default), 1=300ppm"); + static u32 crit_err_count; u32 int_mod_timer_init; u32 int_mod_cq_depth_256; @@ -254,6 +258,7 @@ struct nes_adapter *nes_init_adapter(struct nes_device *nesdev, u8 hw_rev) { u32 adapter_size; u32 arp_table_size; u16 vendor_id; + u16 device_id; u8 OneG_Mode; u8 func_index; @@ -356,6 +361,13 @@ struct nes_adapter *nes_init_adapter(struct nes_device *nesdev, u8 hw_rev) { return NULL; } + nesadapter->vendor_id = (((u32) nesadapter->mac_addr_high) << 8) | + (nesadapter->mac_addr_low >> 24); + + pci_bus_read_config_word(nesdev->pcidev->bus, nesdev->pcidev->devfn, + PCI_DEVICE_ID, &device_id); + nesadapter->vendor_part_id = device_id; + if (nes_init_serdes(nesdev, hw_rev, port_count, nesadapter, OneG_Mode)) { kfree(nesadapter); @@ -539,7 +551,7 @@ struct nes_adapter *nes_init_adapter(struct nes_device *nesdev, u8 hw_rev) { } if (int_cnt > 1) { spin_lock_irqsave(&nesadapter->phy_lock, flags); - nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL1, 0x0000F088); + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL1, 0x0000F0C8); mh_detected++; reset_value = nes_read32(nesdev->regs+NES_SOFTWARE_RESET); reset_value |= 0x0000003d; @@ -564,7 +576,7 @@ struct nes_adapter *nes_init_adapter(struct nes_device *nesdev, u8 hw_rev) { if (++ext_cnt > int_cnt) { spin_lock_irqsave(&nesadapter->phy_lock, flags); nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL1, - 0x0000F0C8); + 0x0000F088); mh_detected++; reset_value = nes_read32(nesdev->regs+NES_SOFTWARE_RESET); reset_value |= 0x0000003d; @@ -655,7 +667,7 @@ static unsigned int nes_reset_adapter_ne020(struct nes_device *nesdev, u8 *OneG_ i = 0; while (((nes_read32(nesdev->regs+NES_SOFTWARE_RESET) & 0x00000040) == 0) && i++ < 10000) mdelay(1); - if (i >= 10000) { + if (i > 10000) { nes_debug(NES_DBG_INIT, "Did not see full soft reset done.\n"); return 0; } @@ -663,7 +675,7 @@ static unsigned int nes_reset_adapter_ne020(struct nes_device *nesdev, u8 *OneG_ i = 0; while ((nes_read_indexed(nesdev, NES_IDX_INT_CPU_STATUS) != 0x80) && i++ < 10000) mdelay(1); - if (i >= 10000) { + if (i > 10000) { printk(KERN_ERR PFX "Internal CPU not ready, status = %02X\n", nes_read_indexed(nesdev, NES_IDX_INT_CPU_STATUS)); return 0; @@ -689,7 +701,7 @@ static unsigned int nes_reset_adapter_ne020(struct nes_device *nesdev, u8 *OneG_ i = 0; while (((nes_read32(nesdev->regs+NES_SOFTWARE_RESET) & 0x00000040) == 0) && i++ < 10000) mdelay(1); - if (i >= 10000) { + if (i > 10000) { nes_debug(NES_DBG_INIT, "Did not see port soft reset done.\n"); return 0; } @@ -699,7 +711,7 @@ static unsigned int nes_reset_adapter_ne020(struct nes_device *nesdev, u8 *OneG_ while (((u32temp = (nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_STATUS0) & 0x0000000f)) != 0x0000000f) && i++ < 5000) mdelay(1); - if (i >= 5000) { + if (i > 5000) { nes_debug(NES_DBG_INIT, "Serdes 0 not ready, status=%x\n", u32temp); return 0; } @@ -710,7 +722,7 @@ static unsigned int nes_reset_adapter_ne020(struct nes_device *nesdev, u8 *OneG_ while (((u32temp = (nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_STATUS1) & 0x0000000f)) != 0x0000000f) && i++ < 5000) mdelay(1); - if (i >= 5000) { + if (i > 5000) { nes_debug(NES_DBG_INIT, "Serdes 1 not ready, status=%x\n", u32temp); return 0; } @@ -728,38 +740,50 @@ static int nes_init_serdes(struct nes_device *nesdev, u8 hw_rev, u8 port_count, { int i; u32 u32temp; - u32 serdes_common_control; + u32 sds; if (hw_rev != NE020_REV) { /* init serdes 0 */ + if (wide_ppm_offset && (nesadapter->phy_type[0] == NES_PHY_TYPE_CX4)) + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_CDR_CONTROL0, 0x000FFFAA); + else + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_CDR_CONTROL0, 0x000000FF); - nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_CDR_CONTROL0, 0x000000FF); if (nesadapter->phy_type[0] == NES_PHY_TYPE_PUMA_1G) { - serdes_common_control = nes_read_indexed(nesdev, - NES_IDX_ETH_SERDES_COMMON_CONTROL0); - serdes_common_control |= 0x000000100; - nes_write_indexed(nesdev, - NES_IDX_ETH_SERDES_COMMON_CONTROL0, - serdes_common_control); - } else if (!OneG_Mode) { - nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_TX_HIGHZ_LANE_MODE0, 0x11110000); + sds = nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0); + sds |= 0x00000100; + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0, sds); } - if (((port_count > 1) && - (nesadapter->phy_type[0] != NES_PHY_TYPE_PUMA_1G)) || - ((port_count > 2) && - (nesadapter->phy_type[0] == NES_PHY_TYPE_PUMA_1G))) { - /* init serdes 1 */ + if (!OneG_Mode) + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_TX_HIGHZ_LANE_MODE0, 0x11110000); + + if (port_count < 2) + return 0; + + /* init serdes 1 */ + if (!(OneG_Mode && (nesadapter->phy_type[1] != NES_PHY_TYPE_PUMA_1G))) nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_CDR_CONTROL1, 0x000000FF); - if (nesadapter->phy_type[0] == NES_PHY_TYPE_PUMA_1G) { - serdes_common_control = nes_read_indexed(nesdev, - NES_IDX_ETH_SERDES_COMMON_CONTROL1); - serdes_common_control |= 0x000000100; - nes_write_indexed(nesdev, - NES_IDX_ETH_SERDES_COMMON_CONTROL1, - serdes_common_control); - } else if (!OneG_Mode) { - nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_TX_HIGHZ_LANE_MODE1, 0x11110000); - } + + switch (nesadapter->phy_type[1]) { + case NES_PHY_TYPE_ARGUS: + case NES_PHY_TYPE_SFP_D: + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_TX_EMP0, 0x00000000); + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_TX_EMP1, 0x00000000); + break; + case NES_PHY_TYPE_CX4: + if (wide_ppm_offset) + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_CDR_CONTROL1, 0x000FFFAA); + break; + case NES_PHY_TYPE_PUMA_1G: + sds = nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL1); + sds |= 0x000000100; + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL1, sds); + } + if (!OneG_Mode) { + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_TX_HIGHZ_LANE_MODE1, 0x11110000); + sds = nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL1); + sds &= 0xFFFFFFBF; + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL1, sds); } } else { /* init serdes 0 */ @@ -768,7 +792,7 @@ static int nes_init_serdes(struct nes_device *nesdev, u8 hw_rev, u8 port_count, while (((u32temp = (nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_STATUS0) & 0x0000000f)) != 0x0000000f) && i++ < 5000) mdelay(1); - if (i >= 5000) { + if (i > 5000) { nes_debug(NES_DBG_PHY, "Init: serdes 0 not ready, status=%x\n", u32temp); return 1; } @@ -791,7 +815,7 @@ static int nes_init_serdes(struct nes_device *nesdev, u8 hw_rev, u8 port_count, while (((u32temp = (nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_STATUS1) & 0x0000000f)) != 0x0000000f) && (i++ < 5000)) mdelay(1); - if (i >= 5000) { + if (i > 5000) { printk("%s: Init: serdes 1 not ready, status=%x\n", __func__, u32temp); /* return 1; */ } @@ -888,6 +912,12 @@ static void nes_init_csr_ne020(struct nes_device *nesdev, u8 hw_rev, u8 port_cou u32temp &= 0x7fffffff; u32temp |= 0x7fff0010; nes_write_indexed(nesdev, 0x000021f8, u32temp); + if (port_count > 1) { + u32temp = nes_read_indexed(nesdev, 0x000023f8); + u32temp &= 0x7fffffff; + u32temp |= 0x7fff0010; + nes_write_indexed(nesdev, 0x000023f8, u32temp); + } } } @@ -1251,203 +1281,163 @@ int nes_init_phy(struct nes_device *nesdev) { struct nes_adapter *nesadapter = nesdev->nesadapter; u32 counter = 0; - u32 sds_common_control0; + u32 sds; u32 mac_index = nesdev->mac_index; u32 tx_config = 0; u16 phy_data; u32 temp_phy_data = 0; u32 temp_phy_data2 = 0; - u32 i = 0; + u8 phy_type = nesadapter->phy_type[mac_index]; + u8 phy_index = nesadapter->phy_index[mac_index]; if ((nesadapter->OneG_Mode) && - (nesadapter->phy_type[mac_index] != NES_PHY_TYPE_PUMA_1G)) { + (phy_type != NES_PHY_TYPE_PUMA_1G)) { nes_debug(NES_DBG_PHY, "1G PHY, mac_index = %d.\n", mac_index); - if (nesadapter->phy_type[mac_index] == NES_PHY_TYPE_1G) { - printk(PFX "%s: Programming mdc config for 1G\n", __func__); + if (phy_type == NES_PHY_TYPE_1G) { tx_config = nes_read_indexed(nesdev, NES_IDX_MAC_TX_CONFIG); tx_config &= 0xFFFFFFE3; tx_config |= 0x04; nes_write_indexed(nesdev, NES_IDX_MAC_TX_CONFIG, tx_config); } - nes_read_1G_phy_reg(nesdev, 1, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 1 phy address %u = 0x%X.\n", - nesadapter->phy_index[mac_index], phy_data); - nes_write_1G_phy_reg(nesdev, 23, nesadapter->phy_index[mac_index], 0xb000); + nes_read_1G_phy_reg(nesdev, 1, phy_index, &phy_data); + nes_write_1G_phy_reg(nesdev, 23, phy_index, 0xb000); /* Reset the PHY */ - nes_write_1G_phy_reg(nesdev, 0, nesadapter->phy_index[mac_index], 0x8000); + nes_write_1G_phy_reg(nesdev, 0, phy_index, 0x8000); udelay(100); counter = 0; do { - nes_read_1G_phy_reg(nesdev, 0, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0 = 0x%X.\n", phy_data); - if (counter++ > 100) break; + nes_read_1G_phy_reg(nesdev, 0, phy_index, &phy_data); + if (counter++ > 100) + break; } while (phy_data & 0x8000); /* Setting no phy loopback */ phy_data &= 0xbfff; phy_data |= 0x1140; - nes_write_1G_phy_reg(nesdev, 0, nesadapter->phy_index[mac_index], phy_data); - nes_read_1G_phy_reg(nesdev, 0, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0 = 0x%X.\n", phy_data); - - nes_read_1G_phy_reg(nesdev, 0x17, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x17 = 0x%X.\n", phy_data); - - nes_read_1G_phy_reg(nesdev, 0x1e, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x1e = 0x%X.\n", phy_data); + nes_write_1G_phy_reg(nesdev, 0, phy_index, phy_data); + nes_read_1G_phy_reg(nesdev, 0, phy_index, &phy_data); + nes_read_1G_phy_reg(nesdev, 0x17, phy_index, &phy_data); + nes_read_1G_phy_reg(nesdev, 0x1e, phy_index, &phy_data); /* Setting the interrupt mask */ - nes_read_1G_phy_reg(nesdev, 0x19, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x19 = 0x%X.\n", phy_data); - nes_write_1G_phy_reg(nesdev, 0x19, nesadapter->phy_index[mac_index], 0xffee); - - nes_read_1G_phy_reg(nesdev, 0x19, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x19 = 0x%X.\n", phy_data); + nes_read_1G_phy_reg(nesdev, 0x19, phy_index, &phy_data); + nes_write_1G_phy_reg(nesdev, 0x19, phy_index, 0xffee); + nes_read_1G_phy_reg(nesdev, 0x19, phy_index, &phy_data); /* turning on flow control */ - nes_read_1G_phy_reg(nesdev, 4, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x4 = 0x%X.\n", phy_data); - nes_write_1G_phy_reg(nesdev, 4, nesadapter->phy_index[mac_index], - (phy_data & ~(0x03E0)) | 0xc00); - /* nes_write_1G_phy_reg(nesdev, 4, nesadapter->phy_index[mac_index], - phy_data | 0xc00); */ - nes_read_1G_phy_reg(nesdev, 4, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x4 = 0x%X.\n", phy_data); - - nes_read_1G_phy_reg(nesdev, 9, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x9 = 0x%X.\n", phy_data); - /* Clear Half duplex */ - nes_write_1G_phy_reg(nesdev, 9, nesadapter->phy_index[mac_index], - phy_data & ~(0x0100)); - nes_read_1G_phy_reg(nesdev, 9, nesadapter->phy_index[mac_index], &phy_data); - nes_debug(NES_DBG_PHY, "Phy data from register 0x9 = 0x%X.\n", phy_data); + nes_read_1G_phy_reg(nesdev, 4, phy_index, &phy_data); + nes_write_1G_phy_reg(nesdev, 4, phy_index, (phy_data & ~(0x03E0)) | 0xc00); + nes_read_1G_phy_reg(nesdev, 4, phy_index, &phy_data); - nes_read_1G_phy_reg(nesdev, 0, nesadapter->phy_index[mac_index], &phy_data); - nes_write_1G_phy_reg(nesdev, 0, nesadapter->phy_index[mac_index], phy_data | 0x0300); - } else { - if ((nesadapter->phy_type[mac_index] == NES_PHY_TYPE_IRIS) || - (nesadapter->phy_type[mac_index] == NES_PHY_TYPE_ARGUS)) { - /* setup 10G MDIO operation */ - tx_config = nes_read_indexed(nesdev, NES_IDX_MAC_TX_CONFIG); - tx_config &= 0xFFFFFFE3; - tx_config |= 0x15; - nes_write_indexed(nesdev, NES_IDX_MAC_TX_CONFIG, tx_config); - } - if ((nesadapter->phy_type[mac_index] == NES_PHY_TYPE_ARGUS)) { - nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0xd7ee); - - temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); - mdelay(10); - nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0xd7ee); - temp_phy_data2 = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + /* Clear Half duplex */ + nes_read_1G_phy_reg(nesdev, 9, phy_index, &phy_data); + nes_write_1G_phy_reg(nesdev, 9, phy_index, phy_data & ~(0x0100)); + nes_read_1G_phy_reg(nesdev, 9, phy_index, &phy_data); - /* - * if firmware is already running (like from a - * driver un-load/load, don't do anything. - */ - if (temp_phy_data == temp_phy_data2) { - /* configure QT2505 AMCC PHY */ - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0x0000, 0x8000); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xc300, 0x0000); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xc302, 0x0044); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xc318, 0x0052); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xc319, 0x0008); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xc31a, 0x0098); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0x0026, 0x0E00); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0x0027, 0x0001); - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0x0028, 0xA528); + nes_read_1G_phy_reg(nesdev, 0, phy_index, &phy_data); + nes_write_1G_phy_reg(nesdev, 0, phy_index, phy_data | 0x0300); - /* - * remove micro from reset; chip boots from ROM, - * uploads EEPROM f/w image, uC executes f/w - */ - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xc300, 0x0002); + return 0; + } - /* - * wait for heart beat to start to - * know loading is done - */ - counter = 0; - do { - nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0xd7ee); - temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); - if (counter++ > 1000) { - nes_debug(NES_DBG_PHY, "AMCC PHY- breaking from heartbeat check <this is bad!!!> \n"); - break; - } - mdelay(100); - nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0xd7ee); - temp_phy_data2 = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); - } while ((temp_phy_data2 == temp_phy_data)); + if ((phy_type == NES_PHY_TYPE_IRIS) || + (phy_type == NES_PHY_TYPE_ARGUS) || + (phy_type == NES_PHY_TYPE_SFP_D)) { + /* setup 10G MDIO operation */ + tx_config = nes_read_indexed(nesdev, NES_IDX_MAC_TX_CONFIG); + tx_config &= 0xFFFFFFE3; + tx_config |= 0x15; + nes_write_indexed(nesdev, NES_IDX_MAC_TX_CONFIG, tx_config); + } + if ((phy_type == NES_PHY_TYPE_ARGUS) || + (phy_type == NES_PHY_TYPE_SFP_D)) { + /* Check firmware heartbeat */ + nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7ee); + temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + udelay(1500); + nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7ee); + temp_phy_data2 = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + + if (temp_phy_data != temp_phy_data2) + return 0; - /* - * wait for tracking to start to know - * f/w is good to go - */ - counter = 0; - do { - nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x3, 0xd7fd); - temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); - if (counter++ > 1000) { - nes_debug(NES_DBG_PHY, "AMCC PHY- breaking from status check <this is bad!!!> \n"); - break; - } - mdelay(1000); - /* - * nes_debug(NES_DBG_PHY, "AMCC PHY- phy_status not ready yet = 0x%02X\n", - * temp_phy_data); - */ - } while (((temp_phy_data & 0xff) != 0x50) && ((temp_phy_data & 0xff) != 0x70)); - - /* set LOS Control invert RXLOSB_I_PADINV */ - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xd003, 0x0000); - /* set LOS Control to mask of RXLOSB_I */ - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xc314, 0x0042); - /* set LED1 to input mode (LED1 and LED2 share same LED) */ - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xd006, 0x0007); - /* set LED2 to RX link_status and activity */ - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xd007, 0x000A); - /* set LED3 to RX link_status */ - nes_write_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 0x1, 0xd008, 0x0009); + /* no heartbeat, configure the PHY */ + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0x0000, 0x8000); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc300, 0x0000); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc316, 0x000A); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc318, 0x0052); + if (phy_type == NES_PHY_TYPE_ARGUS) { + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc302, 0x000C); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc319, 0x0008); + nes_write_10G_phy_reg(nesdev, phy_index, 0x3, 0x0027, 0x0001); + } else { + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc302, 0x0004); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc319, 0x0038); + nes_write_10G_phy_reg(nesdev, phy_index, 0x3, 0x0027, 0x0013); + } + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc31a, 0x0098); + nes_write_10G_phy_reg(nesdev, phy_index, 0x3, 0x0026, 0x0E00); - /* - * reset the res-calibration on t2 - * serdes; ensures it is stable after - * the amcc phy is stable - */ + /* setup LEDs */ + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xd006, 0x0007); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xd007, 0x000A); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xd008, 0x0009); - sds_common_control0 = nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0); - sds_common_control0 |= 0x1; - nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0, sds_common_control0); + nes_write_10G_phy_reg(nesdev, phy_index, 0x3, 0x0028, 0xA528); - /* release the res-calibration reset */ - sds_common_control0 &= 0xfffffffe; - nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0, sds_common_control0); + /* Bring PHY out of reset */ + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc300, 0x0002); - i = 0; - while (((nes_read32(nesdev->regs + NES_SOFTWARE_RESET) & 0x00000040) != 0x00000040) - && (i++ < 5000)) { - /* mdelay(1); */ - } + /* Check for heartbeat */ + counter = 0; + mdelay(690); + nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7ee); + temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + do { + if (counter++ > 150) { + nes_debug(NES_DBG_PHY, "No PHY heartbeat\n"); + break; + } + mdelay(1); + nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7ee); + temp_phy_data2 = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + } while ((temp_phy_data2 == temp_phy_data)); - /* - * wait for link train done before moving on, - * or will get an interupt storm - */ - counter = 0; - do { - temp_phy_data = nes_read_indexed(nesdev, NES_IDX_PHY_PCS_CONTROL_STATUS0 + - (0x200 * (nesdev->mac_index & 1))); - if (counter++ > 1000) { - nes_debug(NES_DBG_PHY, "AMCC PHY- breaking from link train wait <this is bad, link didnt train!!!>\n"); - break; - } - mdelay(1); - } while (((temp_phy_data & 0x0f1f0000) != 0x0f0f0000)); + /* wait for tracking */ + counter = 0; + do { + nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7fd); + temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + if (counter++ > 300) { + nes_debug(NES_DBG_PHY, "PHY did not track\n"); + break; } - } + mdelay(10); + } while (((temp_phy_data & 0xff) != 0x50) && ((temp_phy_data & 0xff) != 0x70)); + + /* setup signal integrity */ + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xd003, 0x0000); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xF00D, 0x00FE); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xF00E, 0x0032); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xF00F, 0x0002); + nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xc314, 0x0063); + + /* reset serdes */ + sds = nes_read_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0 + + mac_index * 0x200); + sds |= 0x1; + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0 + + mac_index * 0x200, sds); + sds &= 0xfffffffe; + nes_write_indexed(nesdev, NES_IDX_ETH_SERDES_COMMON_CONTROL0 + + mac_index * 0x200, sds); + + counter = 0; + while (((nes_read32(nesdev->regs + NES_SOFTWARE_RESET) & 0x00000040) != 0x00000040) + && (counter++ < 5000)) + ; } return 0; } @@ -1636,7 +1626,6 @@ int nes_init_nic_qp(struct nes_device *nesdev, struct net_device *netdev) nesvnic->post_cqp_request = nes_post_cqp_request; nesvnic->mcrq_mcast_filter = NULL; - spin_lock_init(&nesvnic->nic.sq_lock); spin_lock_init(&nesvnic->nic.rq_lock); /* setup the RQ */ @@ -2261,6 +2250,8 @@ static void nes_process_aeq(struct nes_device *nesdev, struct nes_hw_aeq *aeq) if (++head >= aeq_size) head = 0; + + nes_write32(nesdev->regs + NES_AEQ_ALLOC, 1 << 16); } while (1); aeq->aeq_head = head; @@ -2350,6 +2341,7 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number) u16 temp_phy_data; u32 pcs_val = 0x0f0f0000; u32 pcs_mask = 0x0f1f0000; + u32 cdr_ctrl; spin_lock_irqsave(&nesadapter->phy_lock, flags); if (nesadapter->mac_sw_state[mac_number] != NES_MAC_SW_IDLE) { @@ -2464,6 +2456,7 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number) break; case NES_PHY_TYPE_ARGUS: + case NES_PHY_TYPE_SFP_D: /* clear the alarms */ nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 4, 0x0008); nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 4, 0xc001); @@ -2474,19 +2467,18 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number) nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 1, 0x9004); nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 1, 0x9005); /* check link status */ - nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 1, 1); + nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 1, 0x9003); temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); - u32temp = 100; - do { - nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 1, 1); - phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); - if ((phy_data == temp_phy_data) || (!(--u32temp))) - break; - temp_phy_data = phy_data; - } while (1); + nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 3, 0x0021); + nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + nes_read_10G_phy_reg(nesdev, nesadapter->phy_index[mac_index], 3, 0x0021); + phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL); + + phy_data = (!temp_phy_data && (phy_data == 0x8000)) ? 0x4 : 0x0; + nes_debug(NES_DBG_PHY, "%s: Phy data = 0x%04X, link was %s.\n", - __func__, phy_data, nesadapter->mac_link_down ? "DOWN" : "UP"); + __func__, phy_data, nesadapter->mac_link_down[mac_index] ? "DOWN" : "UP"); break; case NES_PHY_TYPE_PUMA_1G: @@ -2502,6 +2494,17 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number) } if (phy_data & 0x0004) { + if (wide_ppm_offset && + (nesadapter->phy_type[mac_index] == NES_PHY_TYPE_CX4) && + (nesadapter->hw_rev != NE020_REV)) { + cdr_ctrl = nes_read_indexed(nesdev, + NES_IDX_ETH_SERDES_CDR_CONTROL0 + + mac_index * 0x200); + nes_write_indexed(nesdev, + NES_IDX_ETH_SERDES_CDR_CONTROL0 + + mac_index * 0x200, + cdr_ctrl | 0x000F0000); + } nesadapter->mac_link_down[mac_index] = 0; list_for_each_entry(nesvnic, &nesadapter->nesvnic_list[mac_index], list) { nes_debug(NES_DBG_PHY, "The Link is UP!!. linkup was %d\n", @@ -2516,6 +2519,17 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number) } } } else { + if (wide_ppm_offset && + (nesadapter->phy_type[mac_index] == NES_PHY_TYPE_CX4) && + (nesadapter->hw_rev != NE020_REV)) { + cdr_ctrl = nes_read_indexed(nesdev, + NES_IDX_ETH_SERDES_CDR_CONTROL0 + + mac_index * 0x200); + nes_write_indexed(nesdev, + NES_IDX_ETH_SERDES_CDR_CONTROL0 + + mac_index * 0x200, + cdr_ctrl & 0xFFF0FFFF); + } nesadapter->mac_link_down[mac_index] = 1; list_for_each_entry(nesvnic, &nesadapter->nesvnic_list[mac_index], list) { nes_debug(NES_DBG_PHY, "The Link is Down!!. linkup was %d\n", @@ -2541,7 +2555,7 @@ static void nes_nic_napi_ce_handler(struct nes_device *nesdev, struct nes_hw_nic { struct nes_vnic *nesvnic = container_of(cq, struct nes_vnic, nic_cq); - netif_rx_schedule(&nesvnic->napi); + napi_schedule(&nesvnic->napi); } @@ -2622,9 +2636,9 @@ void nes_nic_ce_handler(struct nes_device *nesdev, struct nes_hw_nic_cq *cq) } else break; } - if (skb) - dev_kfree_skb_any(skb); } + if (skb) + dev_kfree_skb_any(skb); nesnic->sq_tail++; nesnic->sq_tail &= nesnic->sq_size-1; if (sq_cqes > 128) { diff --git a/drivers/infiniband/hw/nes/nes_hw.h b/drivers/infiniband/hw/nes/nes_hw.h index bc0b4de0445..c3654c6383f 100644 --- a/drivers/infiniband/hw/nes/nes_hw.h +++ b/drivers/infiniband/hw/nes/nes_hw.h @@ -1,5 +1,5 @@ /* -* Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. +* Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -35,12 +35,14 @@ #include <linux/inet_lro.h> +#define NES_PHY_TYPE_CX4 1 #define NES_PHY_TYPE_1G 2 #define NES_PHY_TYPE_IRIS 3 #define NES_PHY_TYPE_ARGUS 4 #define NES_PHY_TYPE_PUMA_1G 5 #define NES_PHY_TYPE_PUMA_10G 6 #define NES_PHY_TYPE_GLADIUS 7 +#define NES_PHY_TYPE_SFP_D 8 #define NES_MULTICAST_PF_MAX 8 @@ -61,6 +63,7 @@ enum pci_regs { NES_CQ_ACK = 0x0034, NES_WQE_ALLOC = 0x0040, NES_CQE_ALLOC = 0x0044, + NES_AEQ_ALLOC = 0x0048 }; enum indexed_regs { @@ -875,7 +878,6 @@ struct nes_hw_nic { u8 replenishing_rq; u8 reserved; - spinlock_t sq_lock; spinlock_t rq_lock; }; @@ -1147,7 +1149,6 @@ struct nes_ib_device; struct nes_vnic { struct nes_ib_device *nesibdev; u64 sq_full; - u64 sq_locked; u64 tso_requests; u64 segmented_tso_requests; u64 linearized_skbs; diff --git a/drivers/infiniband/hw/nes/nes_nic.c b/drivers/infiniband/hw/nes/nes_nic.c index 57a47cf7e51..c6e6611d301 100644 --- a/drivers/infiniband/hw/nes/nes_nic.c +++ b/drivers/infiniband/hw/nes/nes_nic.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -111,7 +111,7 @@ static int nes_netdev_poll(struct napi_struct *napi, int budget) nes_nic_ce_handler(nesdev, nescq); if (nescq->cqes_pending == 0) { - netif_rx_complete(napi); + napi_complete(napi); /* clear out completed cqes and arm */ nes_write32(nesdev->regs+NES_CQE_ALLOC, NES_CQE_ALLOC_NOTIFY_NEXT | nescq->cq_number | (nescq->cqe_allocs_pending << 16)); @@ -400,8 +400,7 @@ static int nes_nic_send(struct sk_buff *skb, struct net_device *netdev) if (skb_headlen(skb) == skb->len) { if (skb_headlen(skb) <= NES_FIRST_FRAG_SIZE) { nic_sqe->wqe_words[NES_NIC_SQ_WQE_LENGTH_2_1_IDX] = 0; - nesnic->tx_skb[nesnic->sq_head] = NULL; - dev_kfree_skb(skb); + nesnic->tx_skb[nesnic->sq_head] = skb; } } else { /* Deal with Fragments */ @@ -453,7 +452,6 @@ static int nes_netdev_start_xmit(struct sk_buff *skb, struct net_device *netdev) u32 wqe_count=1; u32 send_rc; struct iphdr *iph; - unsigned long flags; __le16 *wqe_fragment_length; u32 nr_frags; u32 original_first_length; @@ -480,13 +478,6 @@ static int nes_netdev_start_xmit(struct sk_buff *skb, struct net_device *netdev) if (netif_queue_stopped(netdev)) return NETDEV_TX_BUSY; - local_irq_save(flags); - if (!spin_trylock(&nesnic->sq_lock)) { - local_irq_restore(flags); - nesvnic->sq_locked++; - return NETDEV_TX_LOCKED; - } - /* Check if SQ is full */ if ((((nesnic->sq_tail+(nesnic->sq_size*2))-nesnic->sq_head) & (nesnic->sq_size - 1)) == 1) { if (!netif_queue_stopped(netdev)) { @@ -498,7 +489,6 @@ static int nes_netdev_start_xmit(struct sk_buff *skb, struct net_device *netdev) } } nesvnic->sq_full++; - spin_unlock_irqrestore(&nesnic->sq_lock, flags); return NETDEV_TX_BUSY; } @@ -531,7 +521,6 @@ sq_no_longer_full: } } nesvnic->sq_full++; - spin_unlock_irqrestore(&nesnic->sq_lock, flags); nes_debug(NES_DBG_NIC_TX, "%s: HNIC SQ full- TSO request has too many frags!\n", netdev->name); return NETDEV_TX_BUSY; @@ -656,17 +645,13 @@ tso_sq_no_longer_full: skb_set_transport_header(skb, hoffset); skb_set_network_header(skb, nhoffset); send_rc = nes_nic_send(skb, netdev); - if (send_rc != NETDEV_TX_OK) { - spin_unlock_irqrestore(&nesnic->sq_lock, flags); + if (send_rc != NETDEV_TX_OK) return NETDEV_TX_OK; - } } } else { send_rc = nes_nic_send(skb, netdev); - if (send_rc != NETDEV_TX_OK) { - spin_unlock_irqrestore(&nesnic->sq_lock, flags); + if (send_rc != NETDEV_TX_OK) return NETDEV_TX_OK; - } } barrier(); @@ -676,7 +661,6 @@ tso_sq_no_longer_full: (wqe_count << 24) | (1 << 23) | nesvnic->nic.qp_id); netdev->trans_start = jiffies; - spin_unlock_irqrestore(&nesnic->sq_lock, flags); return NETDEV_TX_OK; } @@ -1012,7 +996,6 @@ static const char nes_ethtool_stringset[][ETH_GSTRING_LEN] = { "Pause Frames Received", "Internal Routing Errors", "SQ SW Dropped SKBs", - "SQ Locked", "SQ Full", "Segmented TSO Requests", "Rx Symbol Errors", @@ -1129,16 +1112,17 @@ static void nes_netdev_get_ethtool_stats(struct net_device *netdev, struct nes_device *nesdev = nesvnic->nesdev; u32 nic_count; u32 u32temp; + u32 index = 0; target_ethtool_stats->n_stats = NES_ETHTOOL_STAT_COUNT; - target_stat_values[0] = nesvnic->nesdev->link_status_interrupts; - target_stat_values[1] = nesvnic->linearized_skbs; - target_stat_values[2] = nesvnic->tso_requests; + target_stat_values[index] = nesvnic->nesdev->link_status_interrupts; + target_stat_values[++index] = nesvnic->linearized_skbs; + target_stat_values[++index] = nesvnic->tso_requests; u32temp = nes_read_indexed(nesdev, NES_IDX_MAC_TX_PAUSE_FRAMES + (nesvnic->nesdev->mac_index*0x200)); nesvnic->nesdev->mac_pause_frames_sent += u32temp; - target_stat_values[3] = nesvnic->nesdev->mac_pause_frames_sent; + target_stat_values[++index] = nesvnic->nesdev->mac_pause_frames_sent; u32temp = nes_read_indexed(nesdev, NES_IDX_MAC_RX_PAUSE_FRAMES + (nesvnic->nesdev->mac_index*0x200)); @@ -1209,60 +1193,59 @@ static void nes_netdev_get_ethtool_stats(struct net_device *netdev, nesvnic->endnode_ipv4_tcp_retransmits += u32temp; } - target_stat_values[4] = nesvnic->nesdev->mac_pause_frames_received; - target_stat_values[5] = nesdev->nesadapter->nic_rx_eth_route_err; - target_stat_values[6] = nesvnic->tx_sw_dropped; - target_stat_values[7] = nesvnic->sq_locked; - target_stat_values[8] = nesvnic->sq_full; - target_stat_values[9] = nesvnic->segmented_tso_requests; - target_stat_values[10] = nesvnic->nesdev->mac_rx_symbol_err_frames; - target_stat_values[11] = nesvnic->nesdev->mac_rx_jabber_frames; - target_stat_values[12] = nesvnic->nesdev->mac_rx_oversized_frames; - target_stat_values[13] = nesvnic->nesdev->mac_rx_short_frames; - target_stat_values[14] = nesvnic->endnode_nstat_rx_discard; - target_stat_values[15] = nesvnic->endnode_nstat_rx_octets; - target_stat_values[16] = nesvnic->endnode_nstat_rx_frames; - target_stat_values[17] = nesvnic->endnode_nstat_tx_octets; - target_stat_values[18] = nesvnic->endnode_nstat_tx_frames; - target_stat_values[19] = mh_detected; - target_stat_values[20] = mh_pauses_sent; - target_stat_values[21] = nesvnic->endnode_ipv4_tcp_retransmits; - target_stat_values[22] = atomic_read(&cm_connects); - target_stat_values[23] = atomic_read(&cm_accepts); - target_stat_values[24] = atomic_read(&cm_disconnects); - target_stat_values[25] = atomic_read(&cm_connecteds); - target_stat_values[26] = atomic_read(&cm_connect_reqs); - target_stat_values[27] = atomic_read(&cm_rejects); - target_stat_values[28] = atomic_read(&mod_qp_timouts); - target_stat_values[29] = atomic_read(&qps_created); - target_stat_values[30] = atomic_read(&sw_qps_destroyed); - target_stat_values[31] = atomic_read(&qps_destroyed); - target_stat_values[32] = atomic_read(&cm_closes); - target_stat_values[33] = cm_packets_sent; - target_stat_values[34] = cm_packets_bounced; - target_stat_values[35] = cm_packets_created; - target_stat_values[36] = cm_packets_received; - target_stat_values[37] = cm_packets_dropped; - target_stat_values[38] = cm_packets_retrans; - target_stat_values[39] = cm_listens_created; - target_stat_values[40] = cm_listens_destroyed; - target_stat_values[41] = cm_backlog_drops; - target_stat_values[42] = atomic_read(&cm_loopbacks); - target_stat_values[43] = atomic_read(&cm_nodes_created); - target_stat_values[44] = atomic_read(&cm_nodes_destroyed); - target_stat_values[45] = atomic_read(&cm_accel_dropped_pkts); - target_stat_values[46] = atomic_read(&cm_resets_recvd); - target_stat_values[47] = int_mod_timer_init; - target_stat_values[48] = int_mod_cq_depth_1; - target_stat_values[49] = int_mod_cq_depth_4; - target_stat_values[50] = int_mod_cq_depth_16; - target_stat_values[51] = int_mod_cq_depth_24; - target_stat_values[52] = int_mod_cq_depth_32; - target_stat_values[53] = int_mod_cq_depth_128; - target_stat_values[54] = int_mod_cq_depth_256; - target_stat_values[55] = nesvnic->lro_mgr.stats.aggregated; - target_stat_values[56] = nesvnic->lro_mgr.stats.flushed; - target_stat_values[57] = nesvnic->lro_mgr.stats.no_desc; + target_stat_values[++index] = nesvnic->nesdev->mac_pause_frames_received; + target_stat_values[++index] = nesdev->nesadapter->nic_rx_eth_route_err; + target_stat_values[++index] = nesvnic->tx_sw_dropped; + target_stat_values[++index] = nesvnic->sq_full; + target_stat_values[++index] = nesvnic->segmented_tso_requests; + target_stat_values[++index] = nesvnic->nesdev->mac_rx_symbol_err_frames; + target_stat_values[++index] = nesvnic->nesdev->mac_rx_jabber_frames; + target_stat_values[++index] = nesvnic->nesdev->mac_rx_oversized_frames; + target_stat_values[++index] = nesvnic->nesdev->mac_rx_short_frames; + target_stat_values[++index] = nesvnic->endnode_nstat_rx_discard; + target_stat_values[++index] = nesvnic->endnode_nstat_rx_octets; + target_stat_values[++index] = nesvnic->endnode_nstat_rx_frames; + target_stat_values[++index] = nesvnic->endnode_nstat_tx_octets; + target_stat_values[++index] = nesvnic->endnode_nstat_tx_frames; + target_stat_values[++index] = mh_detected; + target_stat_values[++index] = mh_pauses_sent; + target_stat_values[++index] = nesvnic->endnode_ipv4_tcp_retransmits; + target_stat_values[++index] = atomic_read(&cm_connects); + target_stat_values[++index] = atomic_read(&cm_accepts); + target_stat_values[++index] = atomic_read(&cm_disconnects); + target_stat_values[++index] = atomic_read(&cm_connecteds); + target_stat_values[++index] = atomic_read(&cm_connect_reqs); + target_stat_values[++index] = atomic_read(&cm_rejects); + target_stat_values[++index] = atomic_read(&mod_qp_timouts); + target_stat_values[++index] = atomic_read(&qps_created); + target_stat_values[++index] = atomic_read(&sw_qps_destroyed); + target_stat_values[++index] = atomic_read(&qps_destroyed); + target_stat_values[++index] = atomic_read(&cm_closes); + target_stat_values[++index] = cm_packets_sent; + target_stat_values[++index] = cm_packets_bounced; + target_stat_values[++index] = cm_packets_created; + target_stat_values[++index] = cm_packets_received; + target_stat_values[++index] = cm_packets_dropped; + target_stat_values[++index] = cm_packets_retrans; + target_stat_values[++index] = cm_listens_created; + target_stat_values[++index] = cm_listens_destroyed; + target_stat_values[++index] = cm_backlog_drops; + target_stat_values[++index] = atomic_read(&cm_loopbacks); + target_stat_values[++index] = atomic_read(&cm_nodes_created); + target_stat_values[++index] = atomic_read(&cm_nodes_destroyed); + target_stat_values[++index] = atomic_read(&cm_accel_dropped_pkts); + target_stat_values[++index] = atomic_read(&cm_resets_recvd); + target_stat_values[++index] = int_mod_timer_init; + target_stat_values[++index] = int_mod_cq_depth_1; + target_stat_values[++index] = int_mod_cq_depth_4; + target_stat_values[++index] = int_mod_cq_depth_16; + target_stat_values[++index] = int_mod_cq_depth_24; + target_stat_values[++index] = int_mod_cq_depth_32; + target_stat_values[++index] = int_mod_cq_depth_128; + target_stat_values[++index] = int_mod_cq_depth_256; + target_stat_values[++index] = nesvnic->lro_mgr.stats.aggregated; + target_stat_values[++index] = nesvnic->lro_mgr.stats.flushed; + target_stat_values[++index] = nesvnic->lro_mgr.stats.no_desc; } @@ -1443,49 +1426,55 @@ static int nes_netdev_get_settings(struct net_device *netdev, struct ethtool_cmd struct nes_vnic *nesvnic = netdev_priv(netdev); struct nes_device *nesdev = nesvnic->nesdev; struct nes_adapter *nesadapter = nesdev->nesadapter; + u32 mac_index = nesdev->mac_index; + u8 phy_type = nesadapter->phy_type[mac_index]; + u8 phy_index = nesadapter->phy_index[mac_index]; u16 phy_data; et_cmd->duplex = DUPLEX_FULL; et_cmd->port = PORT_MII; + et_cmd->maxtxpkt = 511; + et_cmd->maxrxpkt = 511; if (nesadapter->OneG_Mode) { et_cmd->speed = SPEED_1000; - if (nesadapter->phy_type[nesdev->mac_index] == NES_PHY_TYPE_PUMA_1G) { + if (phy_type == NES_PHY_TYPE_PUMA_1G) { et_cmd->supported = SUPPORTED_1000baseT_Full; et_cmd->advertising = ADVERTISED_1000baseT_Full; et_cmd->autoneg = AUTONEG_DISABLE; et_cmd->transceiver = XCVR_INTERNAL; - et_cmd->phy_address = nesdev->mac_index; + et_cmd->phy_address = mac_index; } else { - et_cmd->supported = SUPPORTED_1000baseT_Full | SUPPORTED_Autoneg; - et_cmd->advertising = ADVERTISED_1000baseT_Full | ADVERTISED_Autoneg; - nes_read_1G_phy_reg(nesdev, 0, nesadapter->phy_index[nesdev->mac_index], &phy_data); + et_cmd->supported = SUPPORTED_1000baseT_Full + | SUPPORTED_Autoneg; + et_cmd->advertising = ADVERTISED_1000baseT_Full + | ADVERTISED_Autoneg; + nes_read_1G_phy_reg(nesdev, 0, phy_index, &phy_data); if (phy_data & 0x1000) et_cmd->autoneg = AUTONEG_ENABLE; else et_cmd->autoneg = AUTONEG_DISABLE; et_cmd->transceiver = XCVR_EXTERNAL; - et_cmd->phy_address = nesadapter->phy_index[nesdev->mac_index]; + et_cmd->phy_address = phy_index; } + return 0; + } + if ((phy_type == NES_PHY_TYPE_IRIS) || + (phy_type == NES_PHY_TYPE_ARGUS) || + (phy_type == NES_PHY_TYPE_SFP_D)) { + et_cmd->transceiver = XCVR_EXTERNAL; + et_cmd->port = PORT_FIBRE; + et_cmd->supported = SUPPORTED_FIBRE; + et_cmd->advertising = ADVERTISED_FIBRE; + et_cmd->phy_address = phy_index; } else { - if ((nesadapter->phy_type[nesdev->mac_index] == NES_PHY_TYPE_IRIS) || - (nesadapter->phy_type[nesdev->mac_index] == NES_PHY_TYPE_ARGUS)) { - et_cmd->transceiver = XCVR_EXTERNAL; - et_cmd->port = PORT_FIBRE; - et_cmd->supported = SUPPORTED_FIBRE; - et_cmd->advertising = ADVERTISED_FIBRE; - et_cmd->phy_address = nesadapter->phy_index[nesdev->mac_index]; - } else { - et_cmd->transceiver = XCVR_INTERNAL; - et_cmd->supported = SUPPORTED_10000baseT_Full; - et_cmd->advertising = ADVERTISED_10000baseT_Full; - et_cmd->phy_address = nesdev->mac_index; - } - et_cmd->speed = SPEED_10000; - et_cmd->autoneg = AUTONEG_DISABLE; + et_cmd->transceiver = XCVR_INTERNAL; + et_cmd->supported = SUPPORTED_10000baseT_Full; + et_cmd->advertising = ADVERTISED_10000baseT_Full; + et_cmd->phy_address = mac_index; } - et_cmd->maxtxpkt = 511; - et_cmd->maxrxpkt = 511; + et_cmd->speed = SPEED_10000; + et_cmd->autoneg = AUTONEG_DISABLE; return 0; } @@ -1568,6 +1557,19 @@ static void nes_netdev_vlan_rx_register(struct net_device *netdev, struct vlan_g spin_unlock_irqrestore(&nesadapter->phy_lock, flags); } +static const struct net_device_ops nes_netdev_ops = { + .ndo_open = nes_netdev_open, + .ndo_stop = nes_netdev_stop, + .ndo_start_xmit = nes_netdev_start_xmit, + .ndo_get_stats = nes_netdev_get_stats, + .ndo_tx_timeout = nes_netdev_tx_timeout, + .ndo_set_mac_address = nes_netdev_set_mac_address, + .ndo_set_multicast_list = nes_netdev_set_multicast_list, + .ndo_change_mtu = nes_netdev_change_mtu, + .ndo_set_mac_address = eth_mac_addr, + .ndo_validate_addr = eth_validate_addr, + .ndo_vlan_rx_register = nes_netdev_vlan_rx_register, +}; /** * nes_netdev_init - initialize network device @@ -1576,7 +1578,7 @@ struct net_device *nes_netdev_init(struct nes_device *nesdev, void __iomem *mmio_addr) { u64 u64temp; - struct nes_vnic *nesvnic = NULL; + struct nes_vnic *nesvnic; struct net_device *netdev; struct nic_qp_map *curr_qp_map; u32 u32temp; @@ -1588,22 +1590,12 @@ struct net_device *nes_netdev_init(struct nes_device *nesdev, printk(KERN_ERR PFX "nesvnic etherdev alloc failed"); return NULL; } + nesvnic = netdev_priv(netdev); nes_debug(NES_DBG_INIT, "netdev = %p, %s\n", netdev, netdev->name); SET_NETDEV_DEV(netdev, &nesdev->pcidev->dev); - nesvnic = netdev_priv(netdev); - memset(nesvnic, 0, sizeof(*nesvnic)); - - netdev->open = nes_netdev_open; - netdev->stop = nes_netdev_stop; - netdev->hard_start_xmit = nes_netdev_start_xmit; - netdev->get_stats = nes_netdev_get_stats; - netdev->tx_timeout = nes_netdev_tx_timeout; - netdev->set_mac_address = nes_netdev_set_mac_address; - netdev->set_multicast_list = nes_netdev_set_multicast_list; - netdev->change_mtu = nes_netdev_change_mtu; netdev->watchdog_timeo = NES_TX_TIMEOUT; netdev->irq = nesdev->pcidev->irq; netdev->mtu = ETH_DATA_LEN; @@ -1611,12 +1603,11 @@ struct net_device *nes_netdev_init(struct nes_device *nesdev, netdev->addr_len = ETH_ALEN; netdev->type = ARPHRD_ETHER; netdev->features = NETIF_F_HIGHDMA; + netdev->netdev_ops = &nes_netdev_ops; netdev->ethtool_ops = &nes_ethtool_ops; netif_napi_add(netdev, &nesvnic->napi, nes_netdev_poll, 128); nes_debug(NES_DBG_INIT, "Enabling VLAN Insert/Delete.\n"); netdev->features |= NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_RX; - netdev->vlan_rx_register = nes_netdev_vlan_rx_register; - netdev->features |= NETIF_F_LLTX; /* Fill in the port structure */ nesvnic->netdev = netdev; diff --git a/drivers/infiniband/hw/nes/nes_user.h b/drivers/infiniband/hw/nes/nes_user.h index e64306bce80..cc90c14b49e 100644 --- a/drivers/infiniband/hw/nes/nes_user.h +++ b/drivers/infiniband/hw/nes/nes_user.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * Copyright (c) 2005 Topspin Communications. All rights reserved. * Copyright (c) 2005 Cisco Systems. All rights reserved. * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved. diff --git a/drivers/infiniband/hw/nes/nes_utils.c b/drivers/infiniband/hw/nes/nes_utils.c index 6f3bc1b6bf2..a282031d15c 100644 --- a/drivers/infiniband/hw/nes/nes_utils.c +++ b/drivers/infiniband/hw/nes/nes_utils.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c index d93a6562817..21e0fd336cf 100644 --- a/drivers/infiniband/hw/nes/nes_verbs.c +++ b/drivers/infiniband/hw/nes/nes_verbs.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -551,6 +551,7 @@ static int nes_dealloc_fmr(struct ib_fmr *ibfmr) struct nes_device *nesdev = nesvnic->nesdev; struct nes_adapter *nesadapter = nesdev->nesadapter; int i = 0; + int rc; /* free the resources */ if (nesfmr->leaf_pbl_cnt == 0) { @@ -572,7 +573,9 @@ static int nes_dealloc_fmr(struct ib_fmr *ibfmr) nesmr->ibmw.rkey = ibfmr->rkey; nesmr->ibmw.uobject = NULL; - if (nesfmr->nesmr.pbls_used != 0) { + rc = nes_dealloc_mw(&nesmr->ibmw); + + if ((rc == 0) && (nesfmr->nesmr.pbls_used != 0)) { spin_lock_irqsave(&nesadapter->pbl_lock, flags); if (nesfmr->nesmr.pbl_4k) { nesadapter->free_4kpbl += nesfmr->nesmr.pbls_used; @@ -584,7 +587,7 @@ static int nes_dealloc_fmr(struct ib_fmr *ibfmr) spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); } - return nes_dealloc_mw(&nesmr->ibmw); + return rc; } @@ -651,7 +654,7 @@ static int nes_query_device(struct ib_device *ibdev, struct ib_device_attr *prop default: props->max_qp_rd_atom = 0; } - props->max_qp_init_rd_atom = props->max_qp_wr; + props->max_qp_init_rd_atom = props->max_qp_rd_atom; props->atomic_cap = IB_ATOMIC_NONE; props->max_map_per_fmr = 1; @@ -1624,6 +1627,7 @@ static struct ib_cq *nes_create_cq(struct ib_device *ibdev, int entries, nescq->hw_cq.cq_number = nes_ucontext->mcrqf & 0xffff; else nescq->hw_cq.cq_number = nesvnic->mcrq_qp_id + nes_ucontext->mcrqf-1; + nescq->mcrqf = nes_ucontext->mcrqf; nes_free_resource(nesadapter, nesadapter->allocated_cqs, cq_num); } nes_debug(NES_DBG_CQ, "CQ Virtual Address = %08lX, size = %u.\n", @@ -1679,6 +1683,12 @@ static struct ib_cq *nes_create_cq(struct ib_device *ibdev, int entries, if (!context) pci_free_consistent(nesdev->pcidev, nescq->cq_mem_size, mem, nescq->hw_cq.cq_pbase); + else { + pci_free_consistent(nesdev->pcidev, nespbl->pbl_size, + nespbl->pbl_vbase, nespbl->pbl_pbase); + kfree(nespbl); + } + nes_free_resource(nesadapter, nesadapter->allocated_cqs, cq_num); kfree(nescq); return ERR_PTR(-ENOMEM); @@ -1702,6 +1712,11 @@ static struct ib_cq *nes_create_cq(struct ib_device *ibdev, int entries, if (!context) pci_free_consistent(nesdev->pcidev, nescq->cq_mem_size, mem, nescq->hw_cq.cq_pbase); + else { + pci_free_consistent(nesdev->pcidev, nespbl->pbl_size, + nespbl->pbl_vbase, nespbl->pbl_pbase); + kfree(nespbl); + } nes_free_resource(nesadapter, nesadapter->allocated_cqs, cq_num); kfree(nescq); return ERR_PTR(-ENOMEM); @@ -1719,6 +1734,11 @@ static struct ib_cq *nes_create_cq(struct ib_device *ibdev, int entries, if (!context) pci_free_consistent(nesdev->pcidev, nescq->cq_mem_size, mem, nescq->hw_cq.cq_pbase); + else { + pci_free_consistent(nesdev->pcidev, nespbl->pbl_size, + nespbl->pbl_vbase, nespbl->pbl_pbase); + kfree(nespbl); + } nes_free_resource(nesadapter, nesadapter->allocated_cqs, cq_num); kfree(nescq); return ERR_PTR(-ENOMEM); @@ -1771,6 +1791,11 @@ static struct ib_cq *nes_create_cq(struct ib_device *ibdev, int entries, if (!context) pci_free_consistent(nesdev->pcidev, nescq->cq_mem_size, mem, nescq->hw_cq.cq_pbase); + else { + pci_free_consistent(nesdev->pcidev, nespbl->pbl_size, + nespbl->pbl_vbase, nespbl->pbl_pbase); + kfree(nespbl); + } nes_free_resource(nesadapter, nesadapter->allocated_cqs, cq_num); kfree(nescq); return ERR_PTR(-EIO); @@ -1852,7 +1877,9 @@ static int nes_destroy_cq(struct ib_cq *ib_cq) set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_WQE_OPCODE_IDX, opcode); set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_WQE_ID_IDX, (nescq->hw_cq.cq_number | ((u32)PCI_FUNC(nesdev->pcidev->devfn) << 16))); - nes_free_resource(nesadapter, nesadapter->allocated_cqs, nescq->hw_cq.cq_number); + if (!nescq->mcrqf) + nes_free_resource(nesadapter, nesadapter->allocated_cqs, nescq->hw_cq.cq_number); + atomic_set(&cqp_request->refcount, 2); nes_post_cqp_request(nesdev, cqp_request); @@ -1886,21 +1913,74 @@ static int nes_destroy_cq(struct ib_cq *ib_cq) return ret; } +/** + * root_256 + */ +static u32 root_256(struct nes_device *nesdev, + struct nes_root_vpbl *root_vpbl, + struct nes_root_vpbl *new_root, + u16 pbl_count_4k) +{ + u64 leaf_pbl; + int i, j, k; + + if (pbl_count_4k == 1) { + new_root->pbl_vbase = pci_alloc_consistent(nesdev->pcidev, + 512, &new_root->pbl_pbase); + + if (new_root->pbl_vbase == NULL) + return 0; + + leaf_pbl = (u64)root_vpbl->pbl_pbase; + for (i = 0; i < 16; i++) { + new_root->pbl_vbase[i].pa_low = + cpu_to_le32((u32)leaf_pbl); + new_root->pbl_vbase[i].pa_high = + cpu_to_le32((u32)((((u64)leaf_pbl) >> 32))); + leaf_pbl += 256; + } + } else { + for (i = 3; i >= 0; i--) { + j = i * 16; + root_vpbl->pbl_vbase[j] = root_vpbl->pbl_vbase[i]; + leaf_pbl = le32_to_cpu(root_vpbl->pbl_vbase[j].pa_low) + + (((u64)le32_to_cpu(root_vpbl->pbl_vbase[j].pa_high)) + << 32); + for (k = 1; k < 16; k++) { + leaf_pbl += 256; + root_vpbl->pbl_vbase[j + k].pa_low = + cpu_to_le32((u32)leaf_pbl); + root_vpbl->pbl_vbase[j + k].pa_high = + cpu_to_le32((u32)((((u64)leaf_pbl) >> 32))); + } + } + } + + return 1; +} + /** * nes_reg_mr */ static int nes_reg_mr(struct nes_device *nesdev, struct nes_pd *nespd, u32 stag, u64 region_length, struct nes_root_vpbl *root_vpbl, - dma_addr_t single_buffer, u16 pbl_count, u16 residual_page_count, - int acc, u64 *iova_start) + dma_addr_t single_buffer, u16 pbl_count_4k, + u16 residual_page_count_4k, int acc, u64 *iova_start, + u16 *actual_pbl_cnt, u8 *used_4k_pbls) { struct nes_hw_cqp_wqe *cqp_wqe; struct nes_cqp_request *cqp_request; unsigned long flags; int ret; struct nes_adapter *nesadapter = nesdev->nesadapter; - /* int count; */ + uint pg_cnt = 0; + u16 pbl_count_256 = 0; + u16 pbl_count = 0; + u8 use_256_pbls = 0; + u8 use_4k_pbls = 0; + u16 use_two_level = (pbl_count_4k > 1) ? 1 : 0; + struct nes_root_vpbl new_root = {0, 0, 0}; u32 opcode = 0; u16 major_code; @@ -1913,41 +1993,70 @@ static int nes_reg_mr(struct nes_device *nesdev, struct nes_pd *nespd, cqp_request->waiting = 1; cqp_wqe = &cqp_request->cqp_wqe; - spin_lock_irqsave(&nesadapter->pbl_lock, flags); - /* track PBL resources */ - if (pbl_count != 0) { - if (pbl_count > 1) { - /* Two level PBL */ - if ((pbl_count+1) > nesadapter->free_4kpbl) { - nes_debug(NES_DBG_MR, "Out of 4KB Pbls for two level request.\n"); - spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); - nes_free_cqp_request(nesdev, cqp_request); - return -ENOMEM; - } else { - nesadapter->free_4kpbl -= pbl_count+1; - } - } else if (residual_page_count > 32) { - if (pbl_count > nesadapter->free_4kpbl) { - nes_debug(NES_DBG_MR, "Out of 4KB Pbls.\n"); - spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); - nes_free_cqp_request(nesdev, cqp_request); - return -ENOMEM; - } else { - nesadapter->free_4kpbl -= pbl_count; + if (pbl_count_4k) { + spin_lock_irqsave(&nesadapter->pbl_lock, flags); + + pg_cnt = ((pbl_count_4k - 1) * 512) + residual_page_count_4k; + pbl_count_256 = (pg_cnt + 31) / 32; + if (pg_cnt <= 32) { + if (pbl_count_256 <= nesadapter->free_256pbl) + use_256_pbls = 1; + else if (pbl_count_4k <= nesadapter->free_4kpbl) + use_4k_pbls = 1; + } else if (pg_cnt <= 2048) { + if (((pbl_count_4k + use_two_level) <= nesadapter->free_4kpbl) && + (nesadapter->free_4kpbl > (nesadapter->max_4kpbl >> 1))) { + use_4k_pbls = 1; + } else if ((pbl_count_256 + 1) <= nesadapter->free_256pbl) { + use_256_pbls = 1; + use_two_level = 1; + } else if ((pbl_count_4k + use_two_level) <= nesadapter->free_4kpbl) { + use_4k_pbls = 1; } } else { - if (pbl_count > nesadapter->free_256pbl) { - nes_debug(NES_DBG_MR, "Out of 256B Pbls.\n"); - spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); - nes_free_cqp_request(nesdev, cqp_request); - return -ENOMEM; - } else { - nesadapter->free_256pbl -= pbl_count; - } + if ((pbl_count_4k + 1) <= nesadapter->free_4kpbl) + use_4k_pbls = 1; } + + if (use_256_pbls) { + pbl_count = pbl_count_256; + nesadapter->free_256pbl -= pbl_count + use_two_level; + } else if (use_4k_pbls) { + pbl_count = pbl_count_4k; + nesadapter->free_4kpbl -= pbl_count + use_two_level; + } else { + spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); + nes_debug(NES_DBG_MR, "Out of Pbls\n"); + nes_free_cqp_request(nesdev, cqp_request); + return -ENOMEM; + } + + spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); } - spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); + if (use_256_pbls && use_two_level) { + if (root_256(nesdev, root_vpbl, &new_root, pbl_count_4k) == 1) { + if (new_root.pbl_pbase != 0) + root_vpbl = &new_root; + } else { + spin_lock_irqsave(&nesadapter->pbl_lock, flags); + nesadapter->free_256pbl += pbl_count_256 + use_two_level; + use_256_pbls = 0; + + if (pbl_count_4k == 1) + use_two_level = 0; + pbl_count = pbl_count_4k; + + if ((pbl_count_4k + use_two_level) <= nesadapter->free_4kpbl) { + nesadapter->free_4kpbl -= pbl_count + use_two_level; + use_4k_pbls = 1; + } + spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); + + if (use_4k_pbls == 0) + return -ENOMEM; + } + } opcode = NES_CQP_REGISTER_STAG | NES_CQP_STAG_RIGHTS_LOCAL_READ | NES_CQP_STAG_VA_TO | NES_CQP_STAG_MR; @@ -1976,10 +2085,9 @@ static int nes_reg_mr(struct nes_device *nesdev, struct nes_pd *nespd, } else { set_wqe_64bit_value(cqp_wqe->wqe_words, NES_CQP_STAG_WQE_PA_LOW_IDX, root_vpbl->pbl_pbase); set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_STAG_WQE_PBL_BLK_COUNT_IDX, pbl_count); - set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_STAG_WQE_PBL_LEN_IDX, - (((pbl_count - 1) * 4096) + (residual_page_count*8))); + set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_STAG_WQE_PBL_LEN_IDX, (pg_cnt * 8)); - if ((pbl_count > 1) || (residual_page_count > 32)) + if (use_4k_pbls) cqp_wqe->wqe_words[NES_CQP_WQE_OPCODE_IDX] |= cpu_to_le32(NES_CQP_STAG_PBL_BLK_SIZE); } barrier(); @@ -1996,13 +2104,25 @@ static int nes_reg_mr(struct nes_device *nesdev, struct nes_pd *nespd, major_code = cqp_request->major_code; nes_put_cqp_request(nesdev, cqp_request); + if ((!ret || major_code) && pbl_count != 0) { + spin_lock_irqsave(&nesadapter->pbl_lock, flags); + if (use_256_pbls) + nesadapter->free_256pbl += pbl_count + use_two_level; + else if (use_4k_pbls) + nesadapter->free_4kpbl += pbl_count + use_two_level; + spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); + } + if (new_root.pbl_pbase) + pci_free_consistent(nesdev->pcidev, 512, new_root.pbl_vbase, + new_root.pbl_pbase); + if (!ret) return -ETIME; else if (major_code) return -EIO; - else - return 0; + *actual_pbl_cnt = pbl_count + use_two_level; + *used_4k_pbls = use_4k_pbls; return 0; } @@ -2025,6 +2145,7 @@ static struct ib_mr *nes_reg_phys_mr(struct ib_pd *ib_pd, struct nes_root_vpbl root_vpbl; u32 stag; u32 i; + unsigned long mask; u32 stag_index = 0; u32 next_stag_index = 0; u32 driver_key = 0; @@ -2053,6 +2174,9 @@ static struct ib_mr *nes_reg_phys_mr(struct ib_pd *ib_pd, return ERR_PTR(-E2BIG); } + if ((buffer_list[0].addr ^ *iova_start) & ~PAGE_MASK) + return ERR_PTR(-EINVAL); + err = nes_alloc_resource(nesadapter, nesadapter->allocated_mrs, nesadapter->max_mr, &stag_index, &next_stag_index); if (err) { @@ -2118,19 +2242,16 @@ static struct ib_mr *nes_reg_phys_mr(struct ib_pd *ib_pd, root_pbl_index++; cur_pbl_index = 0; } - if (buffer_list[i].addr & ~PAGE_MASK) { - /* TODO: Unwind allocated buffers */ - nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index); - nes_debug(NES_DBG_MR, "Unaligned Memory Buffer: 0x%x\n", - (unsigned int) buffer_list[i].addr); - ibmr = ERR_PTR(-EINVAL); - kfree(nesmr); - goto reg_phys_err; - } - if (!buffer_list[i].size) { + mask = !buffer_list[i].size; + if (i != 0) + mask |= buffer_list[i].addr; + if (i != num_phys_buf - 1) + mask |= buffer_list[i].addr + buffer_list[i].size; + + if (mask & ~PAGE_MASK) { nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index); - nes_debug(NES_DBG_MR, "Invalid Buffer Size\n"); + nes_debug(NES_DBG_MR, "Invalid buffer addr or size\n"); ibmr = ERR_PTR(-EINVAL); kfree(nesmr); goto reg_phys_err; @@ -2141,7 +2262,7 @@ static struct ib_mr *nes_reg_phys_mr(struct ib_pd *ib_pd, if ((buffer_list[i-1].addr+PAGE_SIZE) != buffer_list[i].addr) single_page = 0; } - vpbl.pbl_vbase[cur_pbl_index].pa_low = cpu_to_le32((u32)buffer_list[i].addr); + vpbl.pbl_vbase[cur_pbl_index].pa_low = cpu_to_le32((u32)buffer_list[i].addr & PAGE_MASK); vpbl.pbl_vbase[cur_pbl_index++].pa_high = cpu_to_le32((u32)((((u64)buffer_list[i].addr) >> 32))); } @@ -2154,8 +2275,6 @@ static struct ib_mr *nes_reg_phys_mr(struct ib_pd *ib_pd, " length = 0x%016lX, index = 0x%08X\n", stag, (unsigned long)*iova_start, (unsigned long)region_length, stag_index); - region_length -= (*iova_start)&PAGE_MASK; - /* Make the leaf PBL the root if only one PBL */ if (root_pbl_index == 1) { root_vpbl.pbl_pbase = vpbl.pbl_pbase; @@ -2167,18 +2286,14 @@ static struct ib_mr *nes_reg_phys_mr(struct ib_pd *ib_pd, pbl_count = root_pbl_index; } ret = nes_reg_mr(nesdev, nespd, stag, region_length, &root_vpbl, - buffer_list[0].addr, pbl_count, (u16)cur_pbl_index, acc, iova_start); + buffer_list[0].addr, pbl_count, (u16)cur_pbl_index, acc, iova_start, + &nesmr->pbls_used, &nesmr->pbl_4k); if (ret == 0) { nesmr->ibmr.rkey = stag; nesmr->ibmr.lkey = stag; nesmr->mode = IWNES_MEMREG_TYPE_MEM; ibmr = &nesmr->ibmr; - nesmr->pbl_4k = ((pbl_count > 1) || (cur_pbl_index > 32)) ? 1 : 0; - nesmr->pbls_used = pbl_count; - if (pbl_count > 1) { - nesmr->pbls_used++; - } } else { kfree(nesmr); ibmr = ERR_PTR(-ENOMEM); @@ -2456,8 +2571,9 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length, stag, (unsigned int)iova_start, (unsigned int)region_length, stag_index, (unsigned long long)region->length, pbl_count); - ret = nes_reg_mr( nesdev, nespd, stag, region->length, &root_vpbl, - first_dma_addr, pbl_count, (u16)cur_pbl_index, acc, &iova_start); + ret = nes_reg_mr(nesdev, nespd, stag, region->length, &root_vpbl, + first_dma_addr, pbl_count, (u16)cur_pbl_index, acc, + &iova_start, &nesmr->pbls_used, &nesmr->pbl_4k); nes_debug(NES_DBG_MR, "ret=%d\n", ret); @@ -2466,11 +2582,6 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length, nesmr->ibmr.lkey = stag; nesmr->mode = IWNES_MEMREG_TYPE_MEM; ibmr = &nesmr->ibmr; - nesmr->pbl_4k = ((pbl_count > 1) || (cur_pbl_index > 32)) ? 1 : 0; - nesmr->pbls_used = pbl_count; - if (pbl_count > 1) { - nesmr->pbls_used++; - } } else { ib_umem_release(region); kfree(nesmr); @@ -2609,24 +2720,6 @@ static int nes_dereg_mr(struct ib_mr *ib_mr) cqp_request->waiting = 1; cqp_wqe = &cqp_request->cqp_wqe; - spin_lock_irqsave(&nesadapter->pbl_lock, flags); - if (nesmr->pbls_used != 0) { - if (nesmr->pbl_4k) { - nesadapter->free_4kpbl += nesmr->pbls_used; - if (nesadapter->free_4kpbl > nesadapter->max_4kpbl) { - printk(KERN_ERR PFX "free 4KB PBLs(%u) has exceeded the max(%u)\n", - nesadapter->free_4kpbl, nesadapter->max_4kpbl); - } - } else { - nesadapter->free_256pbl += nesmr->pbls_used; - if (nesadapter->free_256pbl > nesadapter->max_256pbl) { - printk(KERN_ERR PFX "free 256B PBLs(%u) has exceeded the max(%u)\n", - nesadapter->free_256pbl, nesadapter->max_256pbl); - } - } - } - - spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); nes_fill_init_cqp_wqe(cqp_wqe, nesdev); set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_WQE_OPCODE_IDX, NES_CQP_DEALLOCATE_STAG | NES_CQP_STAG_VA_TO | @@ -2644,11 +2737,6 @@ static int nes_dereg_mr(struct ib_mr *ib_mr) " CQP Major:Minor codes = 0x%04X:0x%04X\n", ib_mr->rkey, ret, cqp_request->major_code, cqp_request->minor_code); - nes_free_resource(nesadapter, nesadapter->allocated_mrs, - (ib_mr->rkey & 0x0fffff00) >> 8); - - kfree(nesmr); - major_code = cqp_request->major_code; minor_code = cqp_request->minor_code; @@ -2664,8 +2752,33 @@ static int nes_dereg_mr(struct ib_mr *ib_mr) " to destroy STag, ib_mr=%p, rkey = 0x%08X\n", major_code, minor_code, ib_mr, ib_mr->rkey); return -EIO; - } else - return 0; + } + + if (nesmr->pbls_used != 0) { + spin_lock_irqsave(&nesadapter->pbl_lock, flags); + if (nesmr->pbl_4k) { + nesadapter->free_4kpbl += nesmr->pbls_used; + if (nesadapter->free_4kpbl > nesadapter->max_4kpbl) + printk(KERN_ERR PFX "free 4KB PBLs(%u) has " + "exceeded the max(%u)\n", + nesadapter->free_4kpbl, + nesadapter->max_4kpbl); + } else { + nesadapter->free_256pbl += nesmr->pbls_used; + if (nesadapter->free_256pbl > nesadapter->max_256pbl) + printk(KERN_ERR PFX "free 256B PBLs(%u) has " + "exceeded the max(%u)\n", + nesadapter->free_256pbl, + nesadapter->max_256pbl); + } + spin_unlock_irqrestore(&nesadapter->pbl_lock, flags); + } + nes_free_resource(nesadapter, nesadapter->allocated_mrs, + (ib_mr->rkey & 0x0fffff00) >> 8); + + kfree(nesmr); + + return 0; } @@ -2695,10 +2808,9 @@ static ssize_t show_fw_ver(struct device *dev, struct device_attribute *attr, struct nes_vnic *nesvnic = nesibdev->nesvnic; nes_debug(NES_DBG_INIT, "\n"); - return sprintf(buf, "%x.%x.%x\n", - (int)(nesvnic->nesdev->nesadapter->fw_ver >> 32), - (int)(nesvnic->nesdev->nesadapter->fw_ver >> 16) & 0xffff, - (int)(nesvnic->nesdev->nesadapter->fw_ver & 0xffff)); + return sprintf(buf, "%u.%u\n", + (nesvnic->nesdev->nesadapter->firmware_version >> 16), + (nesvnic->nesdev->nesadapter->firmware_version & 0x000000ff)); } diff --git a/drivers/infiniband/hw/nes/nes_verbs.h b/drivers/infiniband/hw/nes/nes_verbs.h index ae0ca9bc83b..41c07f29f7c 100644 --- a/drivers/infiniband/hw/nes/nes_verbs.h +++ b/drivers/infiniband/hw/nes/nes_verbs.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2009 Intel-NE, Inc. All rights reserved. * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -112,6 +112,7 @@ struct nes_cq { spinlock_t lock; u8 virtual_cq; u8 pad[3]; + u32 mcrqf; }; struct nes_wq { diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c index 47d588ba2a7..181b1f32325 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c @@ -1394,8 +1394,8 @@ void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb, struct ipoib_dev_priv *priv = netdev_priv(dev); int e = skb_queue_empty(&priv->cm.skb_queue); - if (skb->dst) - skb->dst->ops->update_pmtu(skb->dst, mtu); + if (skb_dst(skb)) + skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu); skb_queue_tail(&priv->cm.skb_queue, skb); if (e) @@ -1455,13 +1455,15 @@ static ssize_t set_mode(struct device *d, struct device_attribute *attr, struct net_device *dev = to_net_dev(d); struct ipoib_dev_priv *priv = netdev_priv(dev); + if (!rtnl_trylock()) + return restart_syscall(); + /* flush paths if we switch modes so that connections are restarted */ if (IPOIB_CM_SUPPORTED(dev->dev_addr) && !strcmp(buf, "connected\n")) { set_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); ipoib_warn(priv, "enabling connected mode " "will cause multicast packet drops\n"); - rtnl_lock(); dev->features &= ~(NETIF_F_IP_CSUM | NETIF_F_SG | NETIF_F_TSO); rtnl_unlock(); priv->tx_wr.send_flags &= ~IB_SEND_IP_CSUM; @@ -1473,7 +1475,6 @@ static ssize_t set_mode(struct device *d, struct device_attribute *attr, if (!strcmp(buf, "datagram\n")) { clear_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); - rtnl_lock(); if (test_bit(IPOIB_FLAG_CSUM, &priv->flags)) { dev->features |= NETIF_F_IP_CSUM | NETIF_F_SG; if (priv->hca_caps & IB_DEVICE_UD_TSO) @@ -1485,6 +1486,7 @@ static ssize_t set_mode(struct device *d, struct device_attribute *attr, return count; } + rtnl_unlock(); return -EINVAL; } diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c index a1925810be3..e7e5adf84e8 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c @@ -446,11 +446,11 @@ poll_more: if (dev->features & NETIF_F_LRO) lro_flush_all(&priv->lro.lro_mgr); - netif_rx_complete(napi); + napi_complete(napi); if (unlikely(ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS)) && - netif_rx_reschedule(napi)) + napi_reschedule(napi)) goto poll_more; } @@ -462,7 +462,7 @@ void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr) struct net_device *dev = dev_ptr; struct ipoib_dev_priv *priv = netdev_priv(dev); - netif_rx_schedule(&priv->napi); + napi_schedule(&priv->napi); } static void drain_tx_cq(struct net_device *dev) @@ -685,7 +685,8 @@ int ipoib_ib_dev_open(struct net_device *dev) queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task, round_jiffies_relative(HZ)); - set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags); + if (!test_and_set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) + napi_enable(&priv->napi); return 0; } @@ -804,7 +805,8 @@ int ipoib_ib_dev_stop(struct net_device *dev, int flush) struct ipoib_tx_buf *tx_req; int i; - clear_bit(IPOIB_FLAG_INITIALIZED, &priv->flags); + if (test_and_clear_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) + napi_disable(&priv->napi); ipoib_cm_dev_stop(dev); diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index 0bd2a4ff084..e319d91f60a 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -106,8 +106,7 @@ int ipoib_open(struct net_device *dev) ipoib_dbg(priv, "bringing up interface\n"); - if (!test_and_set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) - napi_enable(&priv->napi); + set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); if (ipoib_pkey_dev_delay_open(dev)) return 0; @@ -143,7 +142,6 @@ err_stop: ipoib_ib_dev_stop(dev, 1); err_disable: - napi_disable(&priv->napi); clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); return -EINVAL; @@ -156,7 +154,6 @@ static int ipoib_stop(struct net_device *dev) ipoib_dbg(priv, "stopping interface\n"); clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); - napi_disable(&priv->napi); netif_stop_queue(dev); @@ -564,7 +561,7 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev) struct ipoib_neigh *neigh; unsigned long flags; - neigh = ipoib_neigh_alloc(skb->dst->neighbour, skb->dev); + neigh = ipoib_neigh_alloc(skb_dst(skb)->neighbour, skb->dev); if (!neigh) { ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); @@ -573,9 +570,9 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev) spin_lock_irqsave(&priv->lock, flags); - path = __path_find(dev, skb->dst->neighbour->ha + 4); + path = __path_find(dev, skb_dst(skb)->neighbour->ha + 4); if (!path) { - path = path_rec_create(dev, skb->dst->neighbour->ha + 4); + path = path_rec_create(dev, skb_dst(skb)->neighbour->ha + 4); if (!path) goto err_path; @@ -608,7 +605,7 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev) goto err_drop; } } else - ipoib_send(dev, skb, path->ah, IPOIB_QPN(skb->dst->neighbour->ha)); + ipoib_send(dev, skb, path->ah, IPOIB_QPN(skb_dst(skb)->neighbour->ha)); } else { neigh->ah = NULL; @@ -638,15 +635,15 @@ static void ipoib_path_lookup(struct sk_buff *skb, struct net_device *dev) struct ipoib_dev_priv *priv = netdev_priv(skb->dev); /* Look up path record for unicasts */ - if (skb->dst->neighbour->ha[4] != 0xff) { + if (skb_dst(skb)->neighbour->ha[4] != 0xff) { neigh_add_path(skb, dev); return; } /* Add in the P_Key for multicasts */ - skb->dst->neighbour->ha[8] = (priv->pkey >> 8) & 0xff; - skb->dst->neighbour->ha[9] = priv->pkey & 0xff; - ipoib_mcast_send(dev, skb->dst->neighbour->ha + 4, skb); + skb_dst(skb)->neighbour->ha[8] = (priv->pkey >> 8) & 0xff; + skb_dst(skb)->neighbour->ha[9] = priv->pkey & 0xff; + ipoib_mcast_send(dev, skb_dst(skb)->neighbour->ha + 4, skb); } static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, @@ -660,8 +657,12 @@ static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, path = __path_find(dev, phdr->hwaddr + 4); if (!path || !path->valid) { - if (!path) + int new_path = 0; + + if (!path) { path = path_rec_create(dev, phdr->hwaddr + 4); + new_path = 1; + } if (path) { /* put pseudoheader back on for next time */ skb_push(skb, sizeof *phdr); @@ -669,7 +670,8 @@ static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, if (!path->query && path_rec_start(dev, path)) { spin_unlock_irqrestore(&priv->lock, flags); - path_free(dev, path); + if (new_path) + path_free(dev, path); return; } else __path_add(dev, path); @@ -706,16 +708,16 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) struct ipoib_neigh *neigh; unsigned long flags; - if (likely(skb->dst && skb->dst->neighbour)) { - if (unlikely(!*to_ipoib_neigh(skb->dst->neighbour))) { + if (likely(skb_dst(skb) && skb_dst(skb)->neighbour)) { + if (unlikely(!*to_ipoib_neigh(skb_dst(skb)->neighbour))) { ipoib_path_lookup(skb, dev); return NETDEV_TX_OK; } - neigh = *to_ipoib_neigh(skb->dst->neighbour); + neigh = *to_ipoib_neigh(skb_dst(skb)->neighbour); if (unlikely((memcmp(&neigh->dgid.raw, - skb->dst->neighbour->ha + 4, + skb_dst(skb)->neighbour->ha + 4, sizeof(union ib_gid))) || (neigh->dev != dev))) { spin_lock_irqsave(&priv->lock, flags); @@ -741,7 +743,7 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) return NETDEV_TX_OK; } } else if (neigh->ah) { - ipoib_send(dev, skb, neigh->ah, IPOIB_QPN(skb->dst->neighbour->ha)); + ipoib_send(dev, skb, neigh->ah, IPOIB_QPN(skb_dst(skb)->neighbour->ha)); return NETDEV_TX_OK; } @@ -770,7 +772,7 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) if ((be16_to_cpup((__be16 *) skb->data) != ETH_P_ARP) && (be16_to_cpup((__be16 *) skb->data) != ETH_P_RARP)) { ipoib_warn(priv, "Unicast, no %s: type %04x, QPN %06x %pI6\n", - skb->dst ? "neigh" : "dst", + skb_dst(skb) ? "neigh" : "dst", be16_to_cpup((__be16 *) skb->data), IPOIB_QPN(phdr->hwaddr), phdr->hwaddr + 4); @@ -815,7 +817,7 @@ static int ipoib_hard_header(struct sk_buff *skb, * destination address onto the front of the skb so we can * figure out where to send the packet later. */ - if ((!skb->dst || !skb->dst->neighbour) && daddr) { + if ((!skb_dst(skb) || !skb_dst(skb)->neighbour) && daddr) { struct ipoib_pseudoheader *phdr = (struct ipoib_pseudoheader *) skb_push(skb, sizeof *phdr); memcpy(phdr->hwaddr, daddr, INFINIBAND_ALEN); @@ -1016,18 +1018,22 @@ static void ipoib_lro_setup(struct ipoib_dev_priv *priv) priv->lro.lro_mgr.ip_summed_aggr = CHECKSUM_UNNECESSARY; } +static const struct net_device_ops ipoib_netdev_ops = { + .ndo_open = ipoib_open, + .ndo_stop = ipoib_stop, + .ndo_change_mtu = ipoib_change_mtu, + .ndo_start_xmit = ipoib_start_xmit, + .ndo_tx_timeout = ipoib_timeout, + .ndo_set_multicast_list = ipoib_set_mcast_list, + .ndo_neigh_setup = ipoib_neigh_setup_dev, +}; + static void ipoib_setup(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); - dev->open = ipoib_open; - dev->stop = ipoib_stop; - dev->change_mtu = ipoib_change_mtu; - dev->hard_start_xmit = ipoib_start_xmit; - dev->tx_timeout = ipoib_timeout; + dev->netdev_ops = &ipoib_netdev_ops; dev->header_ops = &ipoib_header_ops; - dev->set_multicast_list = ipoib_set_mcast_list; - dev->neigh_setup = ipoib_neigh_setup_dev; ipoib_set_ethtool_ops(dev); @@ -1047,6 +1053,7 @@ static void ipoib_setup(struct net_device *dev) dev->tx_queue_len = ipoib_sendq_size * 2; dev->features = (NETIF_F_VLAN_CHALLENGED | NETIF_F_HIGHDMA); + dev->priv_flags &= ~IFF_XMIT_DST_RELEASE; memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN); diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c index 425e31112ed..a0e97532e71 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c @@ -261,7 +261,7 @@ static int ipoib_mcast_join_finish(struct ipoib_mcast *mcast, skb->dev = dev; - if (!skb->dst || !skb->dst->neighbour) { + if (!skb_dst(skb) || !skb_dst(skb)->neighbour) { /* put pseudoheader back on for next time */ skb_push(skb, sizeof (struct ipoib_pseudoheader)); } @@ -707,10 +707,10 @@ void ipoib_mcast_send(struct net_device *dev, void *mgid, struct sk_buff *skb) out: if (mcast && mcast->ah) { - if (skb->dst && - skb->dst->neighbour && - !*to_ipoib_neigh(skb->dst->neighbour)) { - struct ipoib_neigh *neigh = ipoib_neigh_alloc(skb->dst->neighbour, + if (skb_dst(skb) && + skb_dst(skb)->neighbour && + !*to_ipoib_neigh(skb_dst(skb)->neighbour)) { + struct ipoib_neigh *neigh = ipoib_neigh_alloc(skb_dst(skb)->neighbour, skb->dev); if (neigh) { diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c index 5a76a551035..e3bf00d8cd2 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c @@ -61,7 +61,8 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) ppriv = netdev_priv(pdev); - rtnl_lock(); + if (!rtnl_trylock()) + return restart_syscall(); mutex_lock(&ppriv->vlan_mutex); /* @@ -70,12 +71,14 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) */ if (ppriv->pkey == pkey) { result = -ENOTUNIQ; + priv = NULL; goto err; } list_for_each_entry(priv, &ppriv->child_intfs, list) { if (priv->pkey == pkey) { result = -ENOTUNIQ; + priv = NULL; goto err; } } @@ -96,7 +99,7 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) result = ipoib_set_dev_features(priv, ppriv->ca); if (result) - goto device_init_failed; + goto err; priv->pkey = pkey; @@ -109,7 +112,7 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) ipoib_warn(ppriv, "failed to initialize subinterface: " "device %s, port %d", ppriv->ca->name, ppriv->port); - goto device_init_failed; + goto err; } result = register_netdevice(priv->dev); @@ -146,40 +149,44 @@ sysfs_failed: register_failed: ipoib_dev_cleanup(priv->dev); -device_init_failed: - free_netdev(priv->dev); - err: mutex_unlock(&ppriv->vlan_mutex); rtnl_unlock(); + if (priv) + free_netdev(priv->dev); + return result; } int ipoib_vlan_delete(struct net_device *pdev, unsigned short pkey) { struct ipoib_dev_priv *ppriv, *priv, *tpriv; - int ret = -ENOENT; + struct net_device *dev = NULL; if (!capable(CAP_NET_ADMIN)) return -EPERM; ppriv = netdev_priv(pdev); - rtnl_lock(); + if (!rtnl_trylock()) + return restart_syscall(); mutex_lock(&ppriv->vlan_mutex); list_for_each_entry_safe(priv, tpriv, &ppriv->child_intfs, list) { if (priv->pkey == pkey) { unregister_netdevice(priv->dev); ipoib_dev_cleanup(priv->dev); list_del(&priv->list); - free_netdev(priv->dev); - - ret = 0; + dev = priv->dev; break; } } mutex_unlock(&ppriv->vlan_mutex); rtnl_unlock(); - return ret; + if (dev) { + free_netdev(dev); + return 0; + } + + return -ENODEV; } diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c index 12876392516..0ba6ec87629 100644 --- a/drivers/infiniband/ulp/iser/iscsi_iser.c +++ b/drivers/infiniband/ulp/iser/iscsi_iser.c @@ -168,7 +168,7 @@ iscsi_iser_mtask_xmit(struct iscsi_conn *conn, struct iscsi_task *task) { int error = 0; - debug_scsi("task deq [cid %d itt 0x%x]\n", conn->id, task->itt); + iser_dbg("task deq [cid %d itt 0x%x]\n", conn->id, task->itt); error = iser_send_control(conn, task); @@ -195,7 +195,7 @@ iscsi_iser_task_xmit_unsol_data(struct iscsi_conn *conn, /* Send data-out PDUs while there's still unsolicited data to send */ while (iscsi_task_has_unsol_data(task)) { iscsi_prep_data_out_pdu(task, r2t, &hdr); - debug_scsi("Sending data-out: itt 0x%x, data count %d\n", + iser_dbg("Sending data-out: itt 0x%x, data count %d\n", hdr.itt, r2t->data_count); /* the buffer description has been passed with the command */ @@ -206,7 +206,7 @@ iscsi_iser_task_xmit_unsol_data(struct iscsi_conn *conn, goto iscsi_iser_task_xmit_unsol_data_exit; } r2t->sent += r2t->data_count; - debug_scsi("Need to send %d more as data-out PDUs\n", + iser_dbg("Need to send %d more as data-out PDUs\n", r2t->data_length - r2t->sent); } @@ -227,12 +227,12 @@ iscsi_iser_task_xmit(struct iscsi_task *task) if (task->sc->sc_data_direction == DMA_TO_DEVICE) { BUG_ON(scsi_bufflen(task->sc) == 0); - debug_scsi("cmd [itt %x total %d imm %d unsol_data %d\n", + iser_dbg("cmd [itt %x total %d imm %d unsol_data %d\n", task->itt, scsi_bufflen(task->sc), task->imm_count, task->unsol_r2t.data_length); } - debug_scsi("task deq [cid %d itt 0x%x]\n", + iser_dbg("task deq [cid %d itt 0x%x]\n", conn->id, task->itt); /* Send the cmd PDU */ @@ -257,11 +257,8 @@ static void iscsi_iser_cleanup_task(struct iscsi_task *task) { struct iscsi_iser_task *iser_task = task->dd_data; - /* - * mgmt tasks do not need special cleanup and we do not - * allocate anything in the init task callout - */ - if (!task->sc || task->state == ISCSI_TASK_PENDING) + /* mgmt tasks do not need special cleanup */ + if (!task->sc) return; if (iser_task->status == ISER_TASK_STATUS_STARTED) { @@ -397,14 +394,14 @@ static void iscsi_iser_session_destroy(struct iscsi_cls_session *cls_session) static struct iscsi_cls_session * iscsi_iser_session_create(struct iscsi_endpoint *ep, uint16_t cmds_max, uint16_t qdepth, - uint32_t initial_cmdsn, uint32_t *hostno) + uint32_t initial_cmdsn) { struct iscsi_cls_session *cls_session; struct iscsi_session *session; struct Scsi_Host *shost; struct iser_conn *ib_conn; - shost = iscsi_host_alloc(&iscsi_iser_sht, 0, ISCSI_MAX_CMD_PER_LUN); + shost = iscsi_host_alloc(&iscsi_iser_sht, 0, 1); if (!shost) return NULL; shost->transportt = iscsi_iser_scsi_transport; @@ -423,7 +420,6 @@ iscsi_iser_session_create(struct iscsi_endpoint *ep, if (iscsi_host_add(shost, ep ? ib_conn->device->ib_device->dma_device : NULL)) goto free_host; - *hostno = shost->host_no; /* * we do not support setting can_queue cmd_per_lun from userspace yet @@ -518,7 +514,8 @@ iscsi_iser_conn_get_stats(struct iscsi_cls_conn *cls_conn, struct iscsi_stats *s } static struct iscsi_endpoint * -iscsi_iser_ep_connect(struct sockaddr *dst_addr, int non_blocking) +iscsi_iser_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr, + int non_blocking) { int err; struct iser_conn *ib_conn; @@ -596,10 +593,11 @@ static struct scsi_host_template iscsi_iser_sht = { .change_queue_depth = iscsi_change_queue_depth, .sg_tablesize = ISCSI_ISER_SG_TABLESIZE, .max_sectors = 1024, - .cmd_per_lun = ISCSI_MAX_CMD_PER_LUN, + .cmd_per_lun = ISER_DEF_CMD_PER_LUN, .eh_abort_handler = iscsi_eh_abort, .eh_device_reset_handler= iscsi_eh_device_reset, .eh_target_reset_handler= iscsi_eh_target_reset, + .target_alloc = iscsi_target_alloc, .use_clustering = DISABLE_CLUSTERING, .proc_name = "iscsi_iser", .this_id = -1, diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h index 861119593f2..9d529cae1f0 100644 --- a/drivers/infiniband/ulp/iser/iscsi_iser.h +++ b/drivers/infiniband/ulp/iser/iscsi_iser.h @@ -93,7 +93,7 @@ /* support upto 512KB in one RDMA */ #define ISCSI_ISER_SG_TABLESIZE (0x80000 >> SHIFT_4K) -#define ISCSI_ISER_MAX_LUN 256 +#define ISER_DEF_CMD_PER_LUN 128 /* QP settings */ /* Maximal bounds on received asynchronous PDUs */ diff --git a/drivers/infiniband/ulp/iser/iser_initiator.c b/drivers/infiniband/ulp/iser/iser_initiator.c index e209cb8dd94..9de640200ad 100644 --- a/drivers/infiniband/ulp/iser/iser_initiator.c +++ b/drivers/infiniband/ulp/iser/iser_initiator.c @@ -661,7 +661,7 @@ void iser_snd_completion(struct iser_desc *tx_desc) if (resume_tx) { iser_dbg("%ld resuming tx\n",jiffies); - scsi_queue_work(conn->session->host, &conn->xmitwork); + iscsi_conn_queue_work(conn); } if (tx_desc->type == ISCSI_TX_CONTROL) { diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c index 319b188145b..ea9e1556e0d 100644 --- a/drivers/infiniband/ulp/iser/iser_verbs.c +++ b/drivers/infiniband/ulp/iser/iser_verbs.c @@ -401,13 +401,6 @@ static void iser_route_handler(struct rdma_cm_id *cma_id) if (ret) goto failure; - iser_dbg("path.mtu is %d setting it to %d\n", - cma_id->route.path_rec->mtu, IB_MTU_1024); - - /* we must set the MTU to 1024 as this is what the target is assuming */ - if (cma_id->route.path_rec->mtu > IB_MTU_1024) - cma_id->route.path_rec->mtu = IB_MTU_1024; - memset(&conn_param, 0, sizeof conn_param); conn_param.responder_resources = 4; conn_param.initiator_depth = 1; |