diff options
Diffstat (limited to 'drivers/infiniband/ulp/ipoib/ipoib_main.c')
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_main.c | 1408 |
1 files changed, 1050 insertions, 358 deletions
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index 6f60abbaebd..5786a78ff8b 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -1,5 +1,7 @@ /* * Copyright (c) 2004 Topspin Communications. All rights reserved. + * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright (c) 2004 Voltaire, 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 @@ -28,17 +30,15 @@ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. - * - * $Id: ipoib_main.c 1377 2004-12-23 19:57:12Z roland $ */ #include "ipoib.h" -#include <linux/version.h> #include <linux/module.h> #include <linux/init.h> #include <linux/slab.h> +#include <linux/kernel.h> #include <linux/vmalloc.h> #include <linux/if_arp.h> /* For ARPHRD_xxx */ @@ -46,9 +46,25 @@ #include <linux/ip.h> #include <linux/in.h> +#include <linux/jhash.h> +#include <net/arp.h> + +#define DRV_VERSION "1.0.0" + +const char ipoib_driver_version[] = DRV_VERSION; + MODULE_AUTHOR("Roland Dreier"); MODULE_DESCRIPTION("IP-over-InfiniBand net driver"); MODULE_LICENSE("Dual BSD/GPL"); +MODULE_VERSION(DRV_VERSION); + +int ipoib_sendq_size __read_mostly = IPOIB_TX_RING_SIZE; +int ipoib_recvq_size __read_mostly = IPOIB_RX_RING_SIZE; + +module_param_named(send_queue_size, ipoib_sendq_size, int, 0444); +MODULE_PARM_DESC(send_queue_size, "Number of descriptors in send queue"); +module_param_named(recv_queue_size, ipoib_recvq_size, int, 0444); +MODULE_PARM_DESC(recv_queue_size, "Number of descriptors in receive queue"); #ifdef CONFIG_INFINIBAND_IPOIB_DEBUG int ipoib_debug_level; @@ -57,6 +73,11 @@ module_param_named(debug_level, ipoib_debug_level, int, 0644); MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0"); #endif +struct ipoib_path_iter { + struct net_device *dev; + struct ipoib_path path; +}; + static const u8 ipv4_bcast_addr[] = { 0x00, 0xff, 0xff, 0xff, 0xff, 0x12, 0x40, 0x1b, 0x00, 0x00, 0x00, 0x00, @@ -65,8 +86,11 @@ static const u8 ipv4_bcast_addr[] = { struct workqueue_struct *ipoib_workqueue; +struct ib_sa_client ipoib_sa_client; + static void ipoib_add_one(struct ib_device *device); static void ipoib_remove_one(struct ib_device *device); +static void ipoib_neigh_reclaim(struct rcu_head *rp); static struct ib_client ipoib_client = { .name = "ipoib", @@ -80,22 +104,24 @@ int ipoib_open(struct net_device *dev) ipoib_dbg(priv, "bringing up interface\n"); + netif_carrier_off(dev); + set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); if (ipoib_pkey_dev_delay_open(dev)) return 0; if (ipoib_ib_dev_open(dev)) - return -EINVAL; + goto err_disable; if (ipoib_ib_dev_up(dev)) - return -EINVAL; + goto err_stop; if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { struct ipoib_dev_priv *cpriv; /* Bring up any child interfaces too */ - down(&priv->vlan_mutex); + down_read(&priv->vlan_rwsem); list_for_each_entry(cpriv, &priv->child_intfs, list) { int flags; @@ -105,12 +131,20 @@ int ipoib_open(struct net_device *dev) dev_change_flags(cpriv->dev, flags | IFF_UP); } - up(&priv->vlan_mutex); + up_read(&priv->vlan_rwsem); } netif_start_queue(dev); return 0; + +err_stop: + ipoib_ib_dev_stop(dev, 1); + +err_disable: + clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); + + return -EINVAL; } static int ipoib_stop(struct net_device *dev) @@ -123,14 +157,14 @@ static int ipoib_stop(struct net_device *dev) netif_stop_queue(dev); - ipoib_ib_dev_down(dev); - ipoib_ib_dev_stop(dev); + ipoib_ib_dev_down(dev, 1); + ipoib_ib_dev_stop(dev, 0); if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { struct ipoib_dev_priv *cpriv; /* Bring down any child interfaces too */ - down(&priv->vlan_mutex); + down_read(&priv->vlan_rwsem); list_for_each_entry(cpriv, &priv->child_intfs, list) { int flags; @@ -140,17 +174,45 @@ static int ipoib_stop(struct net_device *dev) dev_change_flags(cpriv->dev, flags & ~IFF_UP); } - up(&priv->vlan_mutex); + up_read(&priv->vlan_rwsem); } return 0; } +static void ipoib_uninit(struct net_device *dev) +{ + ipoib_dev_cleanup(dev); +} + +static netdev_features_t ipoib_fix_features(struct net_device *dev, netdev_features_t features) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + + if (test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags)) + features &= ~(NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_TSO); + + return features; +} + static int ipoib_change_mtu(struct net_device *dev, int new_mtu) { struct ipoib_dev_priv *priv = netdev_priv(dev); - if (new_mtu > IPOIB_PACKET_SIZE - IPOIB_ENCAP_LEN) + /* dev->mtu > 2K ==> connected mode */ + if (ipoib_cm_admin_enabled(dev)) { + if (new_mtu > ipoib_cm_max_mtu(dev)) + return -EINVAL; + + if (new_mtu > priv->mcast_mtu) + ipoib_warn(priv, "mtu > %d will cause multicast packet drops.\n", + priv->mcast_mtu); + + dev->mtu = new_mtu; + return 0; + } + + if (new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu)) return -EINVAL; priv->admin_mtu = new_mtu; @@ -160,8 +222,38 @@ static int ipoib_change_mtu(struct net_device *dev, int new_mtu) return 0; } -static struct ipoib_path *__path_find(struct net_device *dev, - union ib_gid *gid) +int ipoib_set_mode(struct net_device *dev, const char *buf) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + + /* 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"); + netdev_update_features(dev); + rtnl_unlock(); + priv->tx_wr.send_flags &= ~IB_SEND_IP_CSUM; + + ipoib_flush_paths(dev); + rtnl_lock(); + return 0; + } + + if (!strcmp(buf, "datagram\n")) { + clear_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); + netdev_update_features(dev); + dev_set_mtu(dev, min(priv->mcast_mtu, dev->mtu)); + rtnl_unlock(); + ipoib_flush_paths(dev); + rtnl_lock(); + return 0; + } + + return -EINVAL; +} + +static struct ipoib_path *__path_find(struct net_device *dev, void *gid) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct rb_node *n = priv->path_tree.rb_node; @@ -171,7 +263,7 @@ static struct ipoib_path *__path_find(struct net_device *dev, while (n) { path = rb_entry(n, struct ipoib_path, rb_node); - ret = memcmp(gid->raw, path->pathrec.dgid.raw, + ret = memcmp(gid, path->pathrec.dgid.raw, sizeof (union ib_gid)); if (ret < 0) @@ -217,31 +309,15 @@ static int __path_add(struct net_device *dev, struct ipoib_path *path) static void path_free(struct net_device *dev, struct ipoib_path *path) { - struct ipoib_dev_priv *priv = netdev_priv(dev); - struct ipoib_neigh *neigh, *tn; struct sk_buff *skb; - unsigned long flags; while ((skb = __skb_dequeue(&path->queue))) dev_kfree_skb_irq(skb); - spin_lock_irqsave(&priv->lock, flags); - - list_for_each_entry_safe(neigh, tn, &path->neigh_list, list) { - /* - * It's safe to call ipoib_put_ah() inside priv->lock - * here, because we know that path->ah will always - * hold one more reference, so ipoib_put_ah() will - * never do more than decrement the ref count. - */ - if (neigh->ah) - ipoib_put_ah(neigh->ah); - *to_ipoib_neigh(neigh->neighbour) = NULL; - neigh->neighbour->ops->destructor = NULL; - kfree(neigh); - } + ipoib_dbg(netdev_priv(dev), "path_free\n"); - spin_unlock_irqrestore(&priv->lock, flags); + /* remove all neigh connected to this path */ + ipoib_del_neighs_by_gid(dev, path->pathrec.dgid.raw); if (path->ah) ipoib_put_ah(path->ah); @@ -249,6 +325,81 @@ static void path_free(struct net_device *dev, struct ipoib_path *path) kfree(path); } +#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG + +struct ipoib_path_iter *ipoib_path_iter_init(struct net_device *dev) +{ + struct ipoib_path_iter *iter; + + iter = kmalloc(sizeof *iter, GFP_KERNEL); + if (!iter) + return NULL; + + iter->dev = dev; + memset(iter->path.pathrec.dgid.raw, 0, 16); + + if (ipoib_path_iter_next(iter)) { + kfree(iter); + return NULL; + } + + return iter; +} + +int ipoib_path_iter_next(struct ipoib_path_iter *iter) +{ + struct ipoib_dev_priv *priv = netdev_priv(iter->dev); + struct rb_node *n; + struct ipoib_path *path; + int ret = 1; + + spin_lock_irq(&priv->lock); + + n = rb_first(&priv->path_tree); + + while (n) { + path = rb_entry(n, struct ipoib_path, rb_node); + + if (memcmp(iter->path.pathrec.dgid.raw, path->pathrec.dgid.raw, + sizeof (union ib_gid)) < 0) { + iter->path = *path; + ret = 0; + break; + } + + n = rb_next(n); + } + + spin_unlock_irq(&priv->lock); + + return ret; +} + +void ipoib_path_iter_read(struct ipoib_path_iter *iter, + struct ipoib_path *path) +{ + *path = iter->path; +} + +#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */ + +void ipoib_mark_paths_invalid(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct ipoib_path *path, *tp; + + spin_lock_irq(&priv->lock); + + list_for_each_entry_safe(path, tp, &priv->path_list, list) { + ipoib_dbg(priv, "mark path LID 0x%04x GID %pI6 invalid\n", + be16_to_cpu(path->pathrec.dlid), + path->pathrec.dgid.raw); + path->valid = 0; + } + + spin_unlock_irq(&priv->lock); +} + void ipoib_flush_paths(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); @@ -256,22 +407,27 @@ void ipoib_flush_paths(struct net_device *dev) LIST_HEAD(remove_list); unsigned long flags; + netif_tx_lock_bh(dev); spin_lock_irqsave(&priv->lock, flags); - list_splice(&priv->path_list, &remove_list); - INIT_LIST_HEAD(&priv->path_list); + list_splice_init(&priv->path_list, &remove_list); list_for_each_entry(path, &remove_list, list) rb_erase(&path->rb_node, &priv->path_tree); - spin_unlock_irqrestore(&priv->lock, flags); - list_for_each_entry_safe(path, tp, &remove_list, list) { if (path->query) ib_sa_cancel_query(path->query_id, path->query); + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); wait_for_completion(&path->done); path_free(dev, path); + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); } + + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); } static void path_rec_completion(int status, @@ -282,65 +438,85 @@ static void path_rec_completion(int status, struct net_device *dev = path->dev; struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_ah *ah = NULL; - struct ipoib_neigh *neigh; + struct ipoib_ah *old_ah = NULL; + struct ipoib_neigh *neigh, *tn; struct sk_buff_head skqueue; struct sk_buff *skb; unsigned long flags; - if (pathrec) - ipoib_dbg(priv, "PathRec LID 0x%04x for GID " IPOIB_GID_FMT "\n", - be16_to_cpu(pathrec->dlid), IPOIB_GID_ARG(pathrec->dgid)); + if (!status) + ipoib_dbg(priv, "PathRec LID 0x%04x for GID %pI6\n", + be16_to_cpu(pathrec->dlid), pathrec->dgid.raw); else - ipoib_dbg(priv, "PathRec status %d for GID " IPOIB_GID_FMT "\n", - status, IPOIB_GID_ARG(path->pathrec.dgid)); + ipoib_dbg(priv, "PathRec status %d for GID %pI6\n", + status, path->pathrec.dgid.raw); skb_queue_head_init(&skqueue); if (!status) { - struct ib_ah_attr av = { - .dlid = be16_to_cpu(pathrec->dlid), - .sl = pathrec->sl, - .port_num = priv->port - }; - int path_rate = ib_sa_rate_enum_to_int(pathrec->rate); - - if (path_rate > 0 && priv->local_rate > path_rate) - av.static_rate = (priv->local_rate - 1) / path_rate; + struct ib_ah_attr av; - ipoib_dbg(priv, "static_rate %d for local port %dX, path %dX\n", - av.static_rate, priv->local_rate, - ib_sa_rate_enum_to_int(pathrec->rate)); - - ah = ipoib_create_ah(dev, priv->pd, &av); + if (!ib_init_ah_from_path(priv->ca, priv->port, pathrec, &av)) + ah = ipoib_create_ah(dev, priv->pd, &av); } spin_lock_irqsave(&priv->lock, flags); - path->ah = ah; - - if (ah) { + if (!IS_ERR_OR_NULL(ah)) { path->pathrec = *pathrec; + old_ah = path->ah; + path->ah = ah; + ipoib_dbg(priv, "created address handle %p for LID 0x%04x, SL %d\n", ah, be16_to_cpu(pathrec->dlid), pathrec->sl); while ((skb = __skb_dequeue(&path->queue))) __skb_queue_tail(&skqueue, skb); - list_for_each_entry(neigh, &path->neigh_list, list) { + list_for_each_entry_safe(neigh, tn, &path->neigh_list, list) { + if (neigh->ah) { + WARN_ON(neigh->ah != old_ah); + /* + * Dropping the ah reference inside + * priv->lock is safe here, because we + * will hold one more reference from + * the original value of path->ah (ie + * old_ah). + */ + ipoib_put_ah(neigh->ah); + } kref_get(&path->ah->ref); neigh->ah = path->ah; + if (ipoib_cm_enabled(dev, neigh->daddr)) { + if (!ipoib_cm_get(neigh)) + ipoib_cm_set(neigh, ipoib_cm_create_tx(dev, + path, + neigh)); + if (!ipoib_cm_get(neigh)) { + ipoib_neigh_free(neigh); + continue; + } + } + while ((skb = __skb_dequeue(&neigh->queue))) __skb_queue_tail(&skqueue, skb); } - } else - path->query = NULL; + path->valid = 1; + } + path->query = NULL; complete(&path->done); spin_unlock_irqrestore(&priv->lock, flags); + if (IS_ERR_OR_NULL(ah)) + ipoib_del_neighs_by_gid(dev, path->pathrec.dgid.raw); + + if (old_ah) + ipoib_put_ah(old_ah); + while ((skb = __skb_dequeue(&skqueue))) { skb->dev = dev; if (dev_queue_xmit(skb)) @@ -349,30 +525,29 @@ static void path_rec_completion(int status, } } -static struct ipoib_path *path_rec_create(struct net_device *dev, - union ib_gid *gid) +static struct ipoib_path *path_rec_create(struct net_device *dev, void *gid) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_path *path; - path = kmalloc(sizeof *path, GFP_ATOMIC); + if (!priv->broadcast) + return NULL; + + path = kzalloc(sizeof *path, GFP_ATOMIC); if (!path) return NULL; - path->dev = dev; - path->pathrec.dlid = 0; - path->ah = NULL; + path->dev = dev; skb_queue_head_init(&path->queue); INIT_LIST_HEAD(&path->neigh_list); - path->query = NULL; - init_completion(&path->done); - memcpy(path->pathrec.dgid.raw, gid->raw, sizeof (union ib_gid)); - path->pathrec.sgid = priv->local_gid; - path->pathrec.pkey = cpu_to_be16(priv->pkey); - path->pathrec.numb_path = 1; + memcpy(path->pathrec.dgid.raw, gid, sizeof (union ib_gid)); + path->pathrec.sgid = priv->local_gid; + path->pathrec.pkey = cpu_to_be16(priv->pkey); + path->pathrec.numb_path = 1; + path->pathrec.traffic_class = priv->broadcast->mcmember.traffic_class; return path; } @@ -382,270 +557,263 @@ static int path_rec_start(struct net_device *dev, { struct ipoib_dev_priv *priv = netdev_priv(dev); - ipoib_dbg(priv, "Start path record lookup for " IPOIB_GID_FMT "\n", - IPOIB_GID_ARG(path->pathrec.dgid)); + ipoib_dbg(priv, "Start path record lookup for %pI6\n", + path->pathrec.dgid.raw); + + init_completion(&path->done); path->query_id = - ib_sa_path_rec_get(priv->ca, priv->port, + ib_sa_path_rec_get(&ipoib_sa_client, priv->ca, priv->port, &path->pathrec, IB_SA_PATH_REC_DGID | IB_SA_PATH_REC_SGID | IB_SA_PATH_REC_NUMB_PATH | + IB_SA_PATH_REC_TRAFFIC_CLASS | IB_SA_PATH_REC_PKEY, 1000, GFP_ATOMIC, path_rec_completion, path, &path->query); if (path->query_id < 0) { - ipoib_warn(priv, "ib_sa_path_rec_get failed\n"); + ipoib_warn(priv, "ib_sa_path_rec_get failed: %d\n", path->query_id); path->query = NULL; + complete(&path->done); return path->query_id; } return 0; } -static void neigh_add_path(struct sk_buff *skb, struct net_device *dev) +static void neigh_add_path(struct sk_buff *skb, u8 *daddr, + struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_path *path; struct ipoib_neigh *neigh; + unsigned long flags; - neigh = kmalloc(sizeof *neigh, GFP_ATOMIC); + spin_lock_irqsave(&priv->lock, flags); + neigh = ipoib_neigh_alloc(daddr, dev); if (!neigh) { - ++priv->stats.tx_dropped; + spin_unlock_irqrestore(&priv->lock, flags); + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); return; } - skb_queue_head_init(&neigh->queue); - neigh->neighbour = skb->dst->neighbour; - *to_ipoib_neigh(skb->dst->neighbour) = neigh; - - /* - * We can only be called from ipoib_start_xmit, so we're - * inside tx_lock -- no need to save/restore flags. - */ - spin_lock(&priv->lock); - - path = __path_find(dev, (union ib_gid *) (skb->dst->neighbour->ha + 4)); + path = __path_find(dev, daddr + 4); if (!path) { - path = path_rec_create(dev, - (union ib_gid *) (skb->dst->neighbour->ha + 4)); + path = path_rec_create(dev, daddr + 4); if (!path) - goto err; + goto err_path; __path_add(dev, path); } list_add_tail(&neigh->list, &path->neigh_list); - if (path->pathrec.dlid) { + if (path->ah) { kref_get(&path->ah->ref); neigh->ah = path->ah; - ipoib_send(dev, skb, path->ah, - be32_to_cpup((__be32 *) skb->dst->neighbour->ha)); - } else { - neigh->ah = NULL; - if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) { - __skb_queue_tail(&neigh->queue, skb); + if (ipoib_cm_enabled(dev, neigh->daddr)) { + if (!ipoib_cm_get(neigh)) + ipoib_cm_set(neigh, ipoib_cm_create_tx(dev, path, neigh)); + if (!ipoib_cm_get(neigh)) { + ipoib_neigh_free(neigh); + goto err_drop; + } + if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) + __skb_queue_tail(&neigh->queue, skb); + else { + ipoib_warn(priv, "queue length limit %d. Packet drop.\n", + skb_queue_len(&neigh->queue)); + goto err_drop; + } } else { - ++priv->stats.tx_dropped; - dev_kfree_skb_any(skb); + spin_unlock_irqrestore(&priv->lock, flags); + ipoib_send(dev, skb, path->ah, IPOIB_QPN(daddr)); + ipoib_neigh_put(neigh); + return; } + } else { + neigh->ah = NULL; if (!path->query && path_rec_start(dev, path)) - goto err; + goto err_path; + + __skb_queue_tail(&neigh->queue, skb); } - spin_unlock(&priv->lock); + spin_unlock_irqrestore(&priv->lock, flags); + ipoib_neigh_put(neigh); return; -err: - *to_ipoib_neigh(skb->dst->neighbour) = NULL; - list_del(&neigh->list); - neigh->neighbour->ops->destructor = NULL; - kfree(neigh); - - ++priv->stats.tx_dropped; +err_path: + ipoib_neigh_free(neigh); +err_drop: + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); - spin_unlock(&priv->lock); -} - -static void 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) { - 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, (union ib_gid *) (skb->dst->neighbour->ha + 4), skb); + spin_unlock_irqrestore(&priv->lock, flags); + ipoib_neigh_put(neigh); } static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, - struct ipoib_pseudoheader *phdr) + struct ipoib_cb *cb) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_path *path; + unsigned long flags; - /* - * We can only be called from ipoib_start_xmit, so we're - * inside tx_lock -- no need to save/restore flags. - */ - spin_lock(&priv->lock); + spin_lock_irqsave(&priv->lock, flags); - path = __path_find(dev, (union ib_gid *) (phdr->hwaddr + 4)); - if (!path) { - path = path_rec_create(dev, - (union ib_gid *) (phdr->hwaddr + 4)); + path = __path_find(dev, cb->hwaddr + 4); + if (!path || !path->valid) { + int new_path = 0; + + if (!path) { + path = path_rec_create(dev, cb->hwaddr + 4); + new_path = 1; + } if (path) { - /* put pseudoheader back on for next time */ - skb_push(skb, sizeof *phdr); __skb_queue_tail(&path->queue, skb); - if (path_rec_start(dev, path)) { - spin_unlock(&priv->lock); - path_free(dev, path); + if (!path->query && path_rec_start(dev, path)) { + spin_unlock_irqrestore(&priv->lock, flags); + if (new_path) + path_free(dev, path); return; } else __path_add(dev, path); } else { - ++priv->stats.tx_dropped; + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); } - spin_unlock(&priv->lock); + spin_unlock_irqrestore(&priv->lock, flags); return; } - if (path->pathrec.dlid) { + if (path->ah) { ipoib_dbg(priv, "Send unicast ARP to %04x\n", be16_to_cpu(path->pathrec.dlid)); - ipoib_send(dev, skb, path->ah, - be32_to_cpup((__be32 *) phdr->hwaddr)); + spin_unlock_irqrestore(&priv->lock, flags); + ipoib_send(dev, skb, path->ah, IPOIB_QPN(cb->hwaddr)); + return; } else if ((path->query || !path_rec_start(dev, path)) && skb_queue_len(&path->queue) < IPOIB_MAX_PATH_REC_QUEUE) { - /* put pseudoheader back on for next time */ - skb_push(skb, sizeof *phdr); __skb_queue_tail(&path->queue, skb); } else { - ++priv->stats.tx_dropped; + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); } - spin_unlock(&priv->lock); + spin_unlock_irqrestore(&priv->lock, flags); } static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_neigh *neigh; + struct ipoib_cb *cb = (struct ipoib_cb *) skb->cb; + struct ipoib_header *header; unsigned long flags; - local_irq_save(flags); - if (!spin_trylock(&priv->tx_lock)) { - local_irq_restore(flags); - return NETDEV_TX_LOCKED; - } - - /* - * Check if our queue is stopped. Since we have the LLTX bit - * set, we can't rely on netif_stop_queue() preventing our - * xmit function from being called with a full queue. - */ - if (unlikely(netif_queue_stopped(dev))) { - spin_unlock_irqrestore(&priv->tx_lock, flags); - return NETDEV_TX_BUSY; + header = (struct ipoib_header *) skb->data; + + if (unlikely(cb->hwaddr[4] == 0xff)) { + /* multicast, arrange "if" according to probability */ + if ((header->proto != htons(ETH_P_IP)) && + (header->proto != htons(ETH_P_IPV6)) && + (header->proto != htons(ETH_P_ARP)) && + (header->proto != htons(ETH_P_RARP)) && + (header->proto != htons(ETH_P_TIPC))) { + /* ethertype not supported by IPoIB */ + ++dev->stats.tx_dropped; + dev_kfree_skb_any(skb); + return NETDEV_TX_OK; + } + /* Add in the P_Key for multicast*/ + cb->hwaddr[8] = (priv->pkey >> 8) & 0xff; + cb->hwaddr[9] = priv->pkey & 0xff; + + neigh = ipoib_neigh_get(dev, cb->hwaddr); + if (likely(neigh)) + goto send_using_neigh; + ipoib_mcast_send(dev, cb->hwaddr, skb); + return NETDEV_TX_OK; } - if (skb->dst && skb->dst->neighbour) { - if (unlikely(!*to_ipoib_neigh(skb->dst->neighbour))) { - path_lookup(skb, dev); - goto out; + /* unicast, arrange "switch" according to probability */ + switch (header->proto) { + case htons(ETH_P_IP): + case htons(ETH_P_IPV6): + case htons(ETH_P_TIPC): + neigh = ipoib_neigh_get(dev, cb->hwaddr); + if (unlikely(!neigh)) { + neigh_add_path(skb, cb->hwaddr, dev); + return NETDEV_TX_OK; } + break; + case htons(ETH_P_ARP): + case htons(ETH_P_RARP): + /* for unicast ARP and RARP should always perform path find */ + unicast_arp_send(skb, dev, cb); + return NETDEV_TX_OK; + default: + /* ethertype not supported by IPoIB */ + ++dev->stats.tx_dropped; + dev_kfree_skb_any(skb); + return NETDEV_TX_OK; + } - neigh = *to_ipoib_neigh(skb->dst->neighbour); - - if (likely(neigh->ah)) { - ipoib_send(dev, skb, neigh->ah, - be32_to_cpup((__be32 *) skb->dst->neighbour->ha)); - goto out; +send_using_neigh: + /* note we now hold a ref to neigh */ + if (ipoib_cm_get(neigh)) { + if (ipoib_cm_up(neigh)) { + ipoib_cm_send(dev, skb, ipoib_cm_get(neigh)); + goto unref; } + } else if (neigh->ah) { + ipoib_send(dev, skb, neigh->ah, IPOIB_QPN(cb->hwaddr)); + goto unref; + } - if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) { - spin_lock(&priv->lock); - __skb_queue_tail(&neigh->queue, skb); - spin_unlock(&priv->lock); - } else { - ++priv->stats.tx_dropped; - dev_kfree_skb_any(skb); - } + if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) { + spin_lock_irqsave(&priv->lock, flags); + __skb_queue_tail(&neigh->queue, skb); + spin_unlock_irqrestore(&priv->lock, flags); } else { - struct ipoib_pseudoheader *phdr = - (struct ipoib_pseudoheader *) skb->data; - skb_pull(skb, sizeof *phdr); - - if (phdr->hwaddr[4] == 0xff) { - /* Add in the P_Key for multicast*/ - phdr->hwaddr[8] = (priv->pkey >> 8) & 0xff; - phdr->hwaddr[9] = priv->pkey & 0xff; - - ipoib_mcast_send(dev, (union ib_gid *) (phdr->hwaddr + 4), skb); - } else { - /* unicast GID -- should be ARP reply */ - - if (be16_to_cpup((u16 *) skb->data) != ETH_P_ARP) { - ipoib_warn(priv, "Unicast, no %s: type %04x, QPN %06x " - IPOIB_GID_FMT "\n", - skb->dst ? "neigh" : "dst", - be16_to_cpup((u16 *) skb->data), - be32_to_cpup((u32 *) phdr->hwaddr), - IPOIB_GID_ARG(*(union ib_gid *) (phdr->hwaddr + 4))); - dev_kfree_skb_any(skb); - ++priv->stats.tx_dropped; - goto out; - } - - unicast_arp_send(skb, dev, phdr); - } + ++dev->stats.tx_dropped; + dev_kfree_skb_any(skb); } -out: - spin_unlock_irqrestore(&priv->tx_lock, flags); +unref: + ipoib_neigh_put(neigh); return NETDEV_TX_OK; } -static struct net_device_stats *ipoib_get_stats(struct net_device *dev) -{ - struct ipoib_dev_priv *priv = netdev_priv(dev); - - return &priv->stats; -} - static void ipoib_timeout(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); - ipoib_warn(priv, "transmit timeout: latency %ld\n", - jiffies - dev->trans_start); + ipoib_warn(priv, "transmit timeout: latency %d msecs\n", + jiffies_to_msecs(jiffies - dev->trans_start)); + ipoib_warn(priv, "queue stopped %d, tx_head %u, tx_tail %u\n", + netif_queue_stopped(dev), + priv->tx_head, priv->tx_tail); /* XXX reset QP, etc. */ } static int ipoib_hard_header(struct sk_buff *skb, struct net_device *dev, unsigned short type, - void *daddr, void *saddr, unsigned len) + const void *daddr, const void *saddr, unsigned len) { struct ipoib_header *header; + struct ipoib_cb *cb = (struct ipoib_cb *) skb->cb; header = (struct ipoib_header *) skb_push(skb, sizeof *header); @@ -653,101 +821,466 @@ static int ipoib_hard_header(struct sk_buff *skb, header->reserved = 0; /* - * If we don't have a neighbour structure, stuff the - * destination address onto the front of the skb so we can - * figure out where to send the packet later. + * we don't rely on dst_entry structure, always stuff the + * destination address into skb->cb so we can figure out where + * to send the packet later. */ - if (!skb->dst || !skb->dst->neighbour) { - struct ipoib_pseudoheader *phdr = - (struct ipoib_pseudoheader *) skb_push(skb, sizeof *phdr); - memcpy(phdr->hwaddr, daddr, INFINIBAND_ALEN); - } + memcpy(cb->hwaddr, daddr, INFINIBAND_ALEN); - return 0; + return sizeof *header; } static void ipoib_set_mcast_list(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); - schedule_work(&priv->restart_task); + if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)) { + ipoib_dbg(priv, "IPOIB_FLAG_OPER_UP not set"); + return; + } + + queue_work(ipoib_workqueue, &priv->restart_task); } -static void ipoib_neigh_destructor(struct neighbour *n) +static u32 ipoib_addr_hash(struct ipoib_neigh_hash *htbl, u8 *daddr) { - struct ipoib_neigh *neigh; - struct ipoib_dev_priv *priv = netdev_priv(n->dev); + /* + * Use only the address parts that contributes to spreading + * The subnet prefix is not used as one can not connect to + * same remote port (GUID) using the same remote QPN via two + * different subnets. + */ + /* qpn octets[1:4) & port GUID octets[12:20) */ + u32 *d32 = (u32 *) daddr; + u32 hv; + + hv = jhash_3words(d32[3], d32[4], IPOIB_QPN_MASK & d32[0], 0); + return hv & htbl->mask; +} + +struct ipoib_neigh *ipoib_neigh_get(struct net_device *dev, u8 *daddr) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct ipoib_neigh_table *ntbl = &priv->ntbl; + struct ipoib_neigh_hash *htbl; + struct ipoib_neigh *neigh = NULL; + u32 hash_val; + + rcu_read_lock_bh(); + + htbl = rcu_dereference_bh(ntbl->htbl); + + if (!htbl) + goto out_unlock; + + hash_val = ipoib_addr_hash(htbl, daddr); + for (neigh = rcu_dereference_bh(htbl->buckets[hash_val]); + neigh != NULL; + neigh = rcu_dereference_bh(neigh->hnext)) { + if (memcmp(daddr, neigh->daddr, INFINIBAND_ALEN) == 0) { + /* found, take one ref on behalf of the caller */ + if (!atomic_inc_not_zero(&neigh->refcnt)) { + /* deleted */ + neigh = NULL; + goto out_unlock; + } + neigh->alive = jiffies; + goto out_unlock; + } + } + +out_unlock: + rcu_read_unlock_bh(); + return neigh; +} + +static void __ipoib_reap_neigh(struct ipoib_dev_priv *priv) +{ + struct ipoib_neigh_table *ntbl = &priv->ntbl; + struct ipoib_neigh_hash *htbl; + unsigned long neigh_obsolete; + unsigned long dt; unsigned long flags; - struct ipoib_ah *ah = NULL; + int i; - ipoib_dbg(priv, - "neigh_destructor for %06x " IPOIB_GID_FMT "\n", - be32_to_cpup((__be32 *) n->ha), - IPOIB_GID_ARG(*((union ib_gid *) (n->ha + 4)))); + if (test_bit(IPOIB_STOP_NEIGH_GC, &priv->flags)) + return; spin_lock_irqsave(&priv->lock, flags); - neigh = *to_ipoib_neigh(n); - if (neigh) { - if (neigh->ah) - ah = neigh->ah; - list_del(&neigh->list); - *to_ipoib_neigh(n) = NULL; - kfree(neigh); + htbl = rcu_dereference_protected(ntbl->htbl, + lockdep_is_held(&priv->lock)); + + if (!htbl) + goto out_unlock; + + /* neigh is obsolete if it was idle for two GC periods */ + dt = 2 * arp_tbl.gc_interval; + neigh_obsolete = jiffies - dt; + /* handle possible race condition */ + if (test_bit(IPOIB_STOP_NEIGH_GC, &priv->flags)) + goto out_unlock; + + for (i = 0; i < htbl->size; i++) { + struct ipoib_neigh *neigh; + struct ipoib_neigh __rcu **np = &htbl->buckets[i]; + + while ((neigh = rcu_dereference_protected(*np, + lockdep_is_held(&priv->lock))) != NULL) { + /* was the neigh idle for two GC periods */ + if (time_after(neigh_obsolete, neigh->alive)) { + rcu_assign_pointer(*np, + rcu_dereference_protected(neigh->hnext, + lockdep_is_held(&priv->lock))); + /* remove from path/mc list */ + list_del(&neigh->list); + call_rcu(&neigh->rcu, ipoib_neigh_reclaim); + } else { + np = &neigh->hnext; + } + + } } +out_unlock: spin_unlock_irqrestore(&priv->lock, flags); +} - if (ah) - ipoib_put_ah(ah); +static void ipoib_reap_neigh(struct work_struct *work) +{ + struct ipoib_dev_priv *priv = + container_of(work, struct ipoib_dev_priv, neigh_reap_task.work); + + __ipoib_reap_neigh(priv); + + if (!test_bit(IPOIB_STOP_NEIGH_GC, &priv->flags)) + queue_delayed_work(ipoib_workqueue, &priv->neigh_reap_task, + arp_tbl.gc_interval); } -static int ipoib_neigh_setup(struct neighbour *neigh) + +static struct ipoib_neigh *ipoib_neigh_ctor(u8 *daddr, + struct net_device *dev) { - /* - * Is this kosher? I can't find anybody in the kernel that - * sets neigh->destructor, so we should be able to set it here - * without trouble. + struct ipoib_neigh *neigh; + + neigh = kzalloc(sizeof *neigh, GFP_ATOMIC); + if (!neigh) + return NULL; + + neigh->dev = dev; + memcpy(&neigh->daddr, daddr, sizeof(neigh->daddr)); + skb_queue_head_init(&neigh->queue); + INIT_LIST_HEAD(&neigh->list); + ipoib_cm_set(neigh, NULL); + /* one ref on behalf of the caller */ + atomic_set(&neigh->refcnt, 1); + + return neigh; +} + +struct ipoib_neigh *ipoib_neigh_alloc(u8 *daddr, + struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct ipoib_neigh_table *ntbl = &priv->ntbl; + struct ipoib_neigh_hash *htbl; + struct ipoib_neigh *neigh; + u32 hash_val; + + htbl = rcu_dereference_protected(ntbl->htbl, + lockdep_is_held(&priv->lock)); + if (!htbl) { + neigh = NULL; + goto out_unlock; + } + + /* need to add a new neigh, but maybe some other thread succeeded? + * recalc hash, maybe hash resize took place so we do a search */ - neigh->ops->destructor = ipoib_neigh_destructor; + hash_val = ipoib_addr_hash(htbl, daddr); + for (neigh = rcu_dereference_protected(htbl->buckets[hash_val], + lockdep_is_held(&priv->lock)); + neigh != NULL; + neigh = rcu_dereference_protected(neigh->hnext, + lockdep_is_held(&priv->lock))) { + if (memcmp(daddr, neigh->daddr, INFINIBAND_ALEN) == 0) { + /* found, take one ref on behalf of the caller */ + if (!atomic_inc_not_zero(&neigh->refcnt)) { + /* deleted */ + neigh = NULL; + break; + } + neigh->alive = jiffies; + goto out_unlock; + } + } - return 0; + neigh = ipoib_neigh_ctor(daddr, dev); + if (!neigh) + goto out_unlock; + + /* one ref on behalf of the hash table */ + atomic_inc(&neigh->refcnt); + neigh->alive = jiffies; + /* put in hash */ + rcu_assign_pointer(neigh->hnext, + rcu_dereference_protected(htbl->buckets[hash_val], + lockdep_is_held(&priv->lock))); + rcu_assign_pointer(htbl->buckets[hash_val], neigh); + atomic_inc(&ntbl->entries); + +out_unlock: + + return neigh; +} + +void ipoib_neigh_dtor(struct ipoib_neigh *neigh) +{ + /* neigh reference count was dropprd to zero */ + struct net_device *dev = neigh->dev; + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct sk_buff *skb; + if (neigh->ah) + ipoib_put_ah(neigh->ah); + while ((skb = __skb_dequeue(&neigh->queue))) { + ++dev->stats.tx_dropped; + dev_kfree_skb_any(skb); + } + if (ipoib_cm_get(neigh)) + ipoib_cm_destroy_tx(ipoib_cm_get(neigh)); + ipoib_dbg(netdev_priv(dev), + "neigh free for %06x %pI6\n", + IPOIB_QPN(neigh->daddr), + neigh->daddr + 4); + kfree(neigh); + if (atomic_dec_and_test(&priv->ntbl.entries)) { + if (test_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags)) + complete(&priv->ntbl.flushed); + } } -static int ipoib_neigh_setup_dev(struct net_device *dev, struct neigh_parms *parms) +static void ipoib_neigh_reclaim(struct rcu_head *rp) { - parms->neigh_setup = ipoib_neigh_setup; + /* Called as a result of removal from hash table */ + struct ipoib_neigh *neigh = container_of(rp, struct ipoib_neigh, rcu); + /* note TX context may hold another ref */ + ipoib_neigh_put(neigh); +} + +void ipoib_neigh_free(struct ipoib_neigh *neigh) +{ + struct net_device *dev = neigh->dev; + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct ipoib_neigh_table *ntbl = &priv->ntbl; + struct ipoib_neigh_hash *htbl; + struct ipoib_neigh __rcu **np; + struct ipoib_neigh *n; + u32 hash_val; + + htbl = rcu_dereference_protected(ntbl->htbl, + lockdep_is_held(&priv->lock)); + if (!htbl) + return; + + hash_val = ipoib_addr_hash(htbl, neigh->daddr); + np = &htbl->buckets[hash_val]; + for (n = rcu_dereference_protected(*np, + lockdep_is_held(&priv->lock)); + n != NULL; + n = rcu_dereference_protected(*np, + lockdep_is_held(&priv->lock))) { + if (n == neigh) { + /* found */ + rcu_assign_pointer(*np, + rcu_dereference_protected(neigh->hnext, + lockdep_is_held(&priv->lock))); + /* remove from parent list */ + list_del(&neigh->list); + call_rcu(&neigh->rcu, ipoib_neigh_reclaim); + return; + } else { + np = &n->hnext; + } + } +} + +static int ipoib_neigh_hash_init(struct ipoib_dev_priv *priv) +{ + struct ipoib_neigh_table *ntbl = &priv->ntbl; + struct ipoib_neigh_hash *htbl; + struct ipoib_neigh **buckets; + u32 size; + + clear_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags); + ntbl->htbl = NULL; + htbl = kzalloc(sizeof(*htbl), GFP_KERNEL); + if (!htbl) + return -ENOMEM; + set_bit(IPOIB_STOP_NEIGH_GC, &priv->flags); + size = roundup_pow_of_two(arp_tbl.gc_thresh3); + buckets = kzalloc(size * sizeof(*buckets), GFP_KERNEL); + if (!buckets) { + kfree(htbl); + return -ENOMEM; + } + htbl->size = size; + htbl->mask = (size - 1); + htbl->buckets = buckets; + ntbl->htbl = htbl; + htbl->ntbl = ntbl; + atomic_set(&ntbl->entries, 0); + + /* start garbage collection */ + clear_bit(IPOIB_STOP_NEIGH_GC, &priv->flags); + queue_delayed_work(ipoib_workqueue, &priv->neigh_reap_task, + arp_tbl.gc_interval); return 0; } +static void neigh_hash_free_rcu(struct rcu_head *head) +{ + struct ipoib_neigh_hash *htbl = container_of(head, + struct ipoib_neigh_hash, + rcu); + struct ipoib_neigh __rcu **buckets = htbl->buckets; + struct ipoib_neigh_table *ntbl = htbl->ntbl; + + kfree(buckets); + kfree(htbl); + complete(&ntbl->deleted); +} + +void ipoib_del_neighs_by_gid(struct net_device *dev, u8 *gid) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct ipoib_neigh_table *ntbl = &priv->ntbl; + struct ipoib_neigh_hash *htbl; + unsigned long flags; + int i; + + /* remove all neigh connected to a given path or mcast */ + spin_lock_irqsave(&priv->lock, flags); + + htbl = rcu_dereference_protected(ntbl->htbl, + lockdep_is_held(&priv->lock)); + + if (!htbl) + goto out_unlock; + + for (i = 0; i < htbl->size; i++) { + struct ipoib_neigh *neigh; + struct ipoib_neigh __rcu **np = &htbl->buckets[i]; + + while ((neigh = rcu_dereference_protected(*np, + lockdep_is_held(&priv->lock))) != NULL) { + /* delete neighs belong to this parent */ + if (!memcmp(gid, neigh->daddr + 4, sizeof (union ib_gid))) { + rcu_assign_pointer(*np, + rcu_dereference_protected(neigh->hnext, + lockdep_is_held(&priv->lock))); + /* remove from parent list */ + list_del(&neigh->list); + call_rcu(&neigh->rcu, ipoib_neigh_reclaim); + } else { + np = &neigh->hnext; + } + + } + } +out_unlock: + spin_unlock_irqrestore(&priv->lock, flags); +} + +static void ipoib_flush_neighs(struct ipoib_dev_priv *priv) +{ + struct ipoib_neigh_table *ntbl = &priv->ntbl; + struct ipoib_neigh_hash *htbl; + unsigned long flags; + int i, wait_flushed = 0; + + init_completion(&priv->ntbl.flushed); + + spin_lock_irqsave(&priv->lock, flags); + + htbl = rcu_dereference_protected(ntbl->htbl, + lockdep_is_held(&priv->lock)); + if (!htbl) + goto out_unlock; + + wait_flushed = atomic_read(&priv->ntbl.entries); + if (!wait_flushed) + goto free_htbl; + + for (i = 0; i < htbl->size; i++) { + struct ipoib_neigh *neigh; + struct ipoib_neigh __rcu **np = &htbl->buckets[i]; + + while ((neigh = rcu_dereference_protected(*np, + lockdep_is_held(&priv->lock))) != NULL) { + rcu_assign_pointer(*np, + rcu_dereference_protected(neigh->hnext, + lockdep_is_held(&priv->lock))); + /* remove from path/mc list */ + list_del(&neigh->list); + call_rcu(&neigh->rcu, ipoib_neigh_reclaim); + } + } + +free_htbl: + rcu_assign_pointer(ntbl->htbl, NULL); + call_rcu(&htbl->rcu, neigh_hash_free_rcu); + +out_unlock: + spin_unlock_irqrestore(&priv->lock, flags); + if (wait_flushed) + wait_for_completion(&priv->ntbl.flushed); +} + +static void ipoib_neigh_hash_uninit(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + int stopped; + + ipoib_dbg(priv, "ipoib_neigh_hash_uninit\n"); + init_completion(&priv->ntbl.deleted); + set_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags); + + /* Stop GC if called at init fail need to cancel work */ + stopped = test_and_set_bit(IPOIB_STOP_NEIGH_GC, &priv->flags); + if (!stopped) + cancel_delayed_work(&priv->neigh_reap_task); + + ipoib_flush_neighs(priv); + + wait_for_completion(&priv->ntbl.deleted); +} + + int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port) { struct ipoib_dev_priv *priv = netdev_priv(dev); + if (ipoib_neigh_hash_init(priv) < 0) + goto out; /* Allocate RX/TX "rings" to hold queued skbs */ - - priv->rx_ring = kmalloc(IPOIB_RX_RING_SIZE * sizeof (struct ipoib_buf), + priv->rx_ring = kzalloc(ipoib_recvq_size * sizeof *priv->rx_ring, GFP_KERNEL); if (!priv->rx_ring) { printk(KERN_WARNING "%s: failed to allocate RX ring (%d entries)\n", - ca->name, IPOIB_RX_RING_SIZE); - goto out; + ca->name, ipoib_recvq_size); + goto out_neigh_hash_cleanup; } - memset(priv->rx_ring, 0, - IPOIB_RX_RING_SIZE * sizeof (struct ipoib_buf)); - priv->tx_ring = kmalloc(IPOIB_TX_RING_SIZE * sizeof (struct ipoib_buf), - GFP_KERNEL); + priv->tx_ring = vzalloc(ipoib_sendq_size * sizeof *priv->tx_ring); if (!priv->tx_ring) { printk(KERN_WARNING "%s: failed to allocate TX ring (%d entries)\n", - ca->name, IPOIB_TX_RING_SIZE); + ca->name, ipoib_sendq_size); goto out_rx_ring_cleanup; } - memset(priv->tx_ring, 0, - IPOIB_TX_RING_SIZE * sizeof (struct ipoib_buf)); - /* priv->tx_head & tx_tail are already 0 */ + /* priv->tx_head, tx_tail & tx_outstanding are already 0 */ if (ipoib_ib_dev_init(dev, ca, port)) goto out_tx_ring_cleanup; @@ -755,11 +1288,13 @@ int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port) return 0; out_tx_ring_cleanup: - kfree(priv->tx_ring); + vfree(priv->tx_ring); out_rx_ring_cleanup: kfree(priv->rx_ring); +out_neigh_hash_cleanup: + ipoib_neigh_hash_uninit(dev); out: return -ENOMEM; } @@ -767,89 +1302,92 @@ out: void ipoib_dev_cleanup(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev), *cpriv, *tcpriv; + LIST_HEAD(head); + + ASSERT_RTNL(); - ipoib_delete_debug_file(dev); + ipoib_delete_debug_files(dev); /* Delete any child interfaces first */ list_for_each_entry_safe(cpriv, tcpriv, &priv->child_intfs, list) { - unregister_netdev(cpriv->dev); - ipoib_dev_cleanup(cpriv->dev); - free_netdev(cpriv->dev); + /* Stop GC on child */ + set_bit(IPOIB_STOP_NEIGH_GC, &cpriv->flags); + cancel_delayed_work(&cpriv->neigh_reap_task); + unregister_netdevice_queue(cpriv->dev, &head); } + unregister_netdevice_many(&head); ipoib_ib_dev_cleanup(dev); - if (priv->rx_ring) { - kfree(priv->rx_ring); - priv->rx_ring = NULL; - } + kfree(priv->rx_ring); + vfree(priv->tx_ring); - if (priv->tx_ring) { - kfree(priv->tx_ring); - priv->tx_ring = NULL; - } + priv->rx_ring = NULL; + priv->tx_ring = NULL; + + ipoib_neigh_hash_uninit(dev); } -static void ipoib_setup(struct net_device *dev) +static const struct header_ops ipoib_header_ops = { + .create = ipoib_hard_header, +}; + +static const struct net_device_ops ipoib_netdev_ops = { + .ndo_uninit = ipoib_uninit, + .ndo_open = ipoib_open, + .ndo_stop = ipoib_stop, + .ndo_change_mtu = ipoib_change_mtu, + .ndo_fix_features = ipoib_fix_features, + .ndo_start_xmit = ipoib_start_xmit, + .ndo_tx_timeout = ipoib_timeout, + .ndo_set_rx_mode = ipoib_set_mcast_list, +}; + +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->get_stats = ipoib_get_stats; - dev->tx_timeout = ipoib_timeout; - dev->hard_header = ipoib_hard_header; - dev->set_multicast_list = ipoib_set_mcast_list; - dev->neigh_setup = ipoib_neigh_setup_dev; + dev->netdev_ops = &ipoib_netdev_ops; + dev->header_ops = &ipoib_header_ops; - dev->watchdog_timeo = HZ; + ipoib_set_ethtool_ops(dev); - dev->rebuild_header = NULL; - dev->set_mac_address = NULL; - dev->header_cache_update = NULL; + netif_napi_add(dev, &priv->napi, ipoib_poll, NAPI_POLL_WEIGHT); - dev->flags |= IFF_BROADCAST | IFF_MULTICAST; + dev->watchdog_timeo = HZ; - /* - * We add in INFINIBAND_ALEN to allow for the destination - * address "pseudoheader" for skbs without neighbour struct. - */ - dev->hard_header_len = IPOIB_ENCAP_LEN + INFINIBAND_ALEN; - dev->addr_len = INFINIBAND_ALEN; - dev->type = ARPHRD_INFINIBAND; - dev->tx_queue_len = IPOIB_TX_RING_SIZE * 2; - dev->features = NETIF_F_VLAN_CHALLENGED | NETIF_F_LLTX; + dev->flags |= IFF_BROADCAST | IFF_MULTICAST; - /* MTU will be reset when mcast join happens */ - dev->mtu = IPOIB_PACKET_SIZE - IPOIB_ENCAP_LEN; - priv->mcast_mtu = priv->admin_mtu = dev->mtu; + dev->hard_header_len = IPOIB_ENCAP_LEN; + dev->addr_len = INFINIBAND_ALEN; + dev->type = ARPHRD_INFINIBAND; + 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); - netif_carrier_off(dev); - - SET_MODULE_OWNER(dev); - priv->dev = dev; spin_lock_init(&priv->lock); - spin_lock_init(&priv->tx_lock); - init_MUTEX(&priv->mcast_mutex); - init_MUTEX(&priv->vlan_mutex); + init_rwsem(&priv->vlan_rwsem); INIT_LIST_HEAD(&priv->path_list); INIT_LIST_HEAD(&priv->child_intfs); INIT_LIST_HEAD(&priv->dead_ahs); INIT_LIST_HEAD(&priv->multicast_list); - INIT_WORK(&priv->pkey_task, ipoib_pkey_poll, priv->dev); - INIT_WORK(&priv->mcast_task, ipoib_mcast_join_task, priv->dev); - INIT_WORK(&priv->flush_task, ipoib_ib_dev_flush, priv->dev); - INIT_WORK(&priv->restart_task, ipoib_mcast_restart_task, priv->dev); - INIT_WORK(&priv->ah_reap_task, ipoib_reap_ah, priv->dev); + INIT_DELAYED_WORK(&priv->pkey_poll_task, ipoib_pkey_poll); + INIT_DELAYED_WORK(&priv->mcast_task, ipoib_mcast_join_task); + INIT_WORK(&priv->carrier_on_task, ipoib_mcast_carrier_on_task); + INIT_WORK(&priv->flush_light, ipoib_ib_dev_flush_light); + INIT_WORK(&priv->flush_normal, ipoib_ib_dev_flush_normal); + INIT_WORK(&priv->flush_heavy, ipoib_ib_dev_flush_heavy); + INIT_WORK(&priv->restart_task, ipoib_mcast_restart_task); + INIT_DELAYED_WORK(&priv->ah_reap_task, ipoib_reap_ah); + INIT_DELAYED_WORK(&priv->neigh_reap_task, ipoib_reap_neigh); } struct ipoib_dev_priv *ipoib_intf_alloc(const char *name) @@ -864,16 +1402,54 @@ struct ipoib_dev_priv *ipoib_intf_alloc(const char *name) return netdev_priv(dev); } -static ssize_t show_pkey(struct class_device *cdev, char *buf) +static ssize_t show_pkey(struct device *dev, + struct device_attribute *attr, char *buf) { - struct ipoib_dev_priv *priv = - netdev_priv(container_of(cdev, struct net_device, class_dev)); + struct ipoib_dev_priv *priv = netdev_priv(to_net_dev(dev)); return sprintf(buf, "0x%04x\n", priv->pkey); } -static CLASS_DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL); +static DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL); + +static ssize_t show_umcast(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct ipoib_dev_priv *priv = netdev_priv(to_net_dev(dev)); + + return sprintf(buf, "%d\n", test_bit(IPOIB_FLAG_UMCAST, &priv->flags)); +} -static ssize_t create_child(struct class_device *cdev, +void ipoib_set_umcast(struct net_device *ndev, int umcast_val) +{ + struct ipoib_dev_priv *priv = netdev_priv(ndev); + + if (umcast_val > 0) { + set_bit(IPOIB_FLAG_UMCAST, &priv->flags); + ipoib_warn(priv, "ignoring multicast groups joined directly " + "by userspace\n"); + } else + clear_bit(IPOIB_FLAG_UMCAST, &priv->flags); +} + +static ssize_t set_umcast(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + unsigned long umcast_val = simple_strtoul(buf, NULL, 0); + + ipoib_set_umcast(to_net_dev(dev), umcast_val); + + return count; +} +static DEVICE_ATTR(umcast, S_IWUSR | S_IRUGO, show_umcast, set_umcast); + +int ipoib_add_umcast_attr(struct net_device *dev) +{ + return device_create_file(&dev->dev, &dev_attr_umcast); +} + +static ssize_t create_child(struct device *dev, + struct device_attribute *attr, const char *buf, size_t count) { int pkey; @@ -882,17 +1458,23 @@ static ssize_t create_child(struct class_device *cdev, if (sscanf(buf, "%i", &pkey) != 1) return -EINVAL; - if (pkey < 0 || pkey > 0xffff) + if (pkey <= 0 || pkey > 0xffff || pkey == 0x8000) return -EINVAL; - ret = ipoib_vlan_add(container_of(cdev, struct net_device, class_dev), - pkey); + /* + * Set the full membership bit, so that we join the right + * broadcast group, etc. + */ + pkey |= 0x8000; + + ret = ipoib_vlan_add(to_net_dev(dev), pkey); return ret ? ret : count; } -static CLASS_DEVICE_ATTR(create_child, S_IWUGO, NULL, create_child); +static DEVICE_ATTR(create_child, S_IWUSR, NULL, create_child); -static ssize_t delete_child(struct class_device *cdev, +static ssize_t delete_child(struct device *dev, + struct device_attribute *attr, const char *buf, size_t count) { int pkey; @@ -904,24 +1486,59 @@ static ssize_t delete_child(struct class_device *cdev, if (pkey < 0 || pkey > 0xffff) return -EINVAL; - ret = ipoib_vlan_delete(container_of(cdev, struct net_device, class_dev), - pkey); + ret = ipoib_vlan_delete(to_net_dev(dev), pkey); return ret ? ret : count; } -static CLASS_DEVICE_ATTR(delete_child, S_IWUGO, NULL, delete_child); +static DEVICE_ATTR(delete_child, S_IWUSR, NULL, delete_child); int ipoib_add_pkey_attr(struct net_device *dev) { - return class_device_create_file(&dev->class_dev, - &class_device_attr_pkey); + return device_create_file(&dev->dev, &dev_attr_pkey); +} + +int ipoib_set_dev_features(struct ipoib_dev_priv *priv, struct ib_device *hca) +{ + struct ib_device_attr *device_attr; + int result = -ENOMEM; + + device_attr = kmalloc(sizeof *device_attr, GFP_KERNEL); + if (!device_attr) { + printk(KERN_WARNING "%s: allocation of %zu bytes failed\n", + hca->name, sizeof *device_attr); + return result; + } + + result = ib_query_device(hca, device_attr); + if (result) { + printk(KERN_WARNING "%s: ib_query_device failed (ret = %d)\n", + hca->name, result); + kfree(device_attr); + return result; + } + priv->hca_caps = device_attr->device_cap_flags; + + kfree(device_attr); + + if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) { + priv->dev->hw_features = NETIF_F_SG | + NETIF_F_IP_CSUM | NETIF_F_RXCSUM; + + if (priv->hca_caps & IB_DEVICE_UD_TSO) + priv->dev->hw_features |= NETIF_F_TSO; + + priv->dev->features |= priv->dev->hw_features; + } + + return 0; } static struct net_device *ipoib_add_port(const char *format, struct ib_device *hca, u8 port) { struct ipoib_dev_priv *priv; + struct ib_port_attr attr; int result = -ENOMEM; priv = ipoib_intf_alloc(format); @@ -929,14 +1546,38 @@ static struct net_device *ipoib_add_port(const char *format, goto alloc_mem_failed; SET_NETDEV_DEV(priv->dev, hca->dma_device); + priv->dev->dev_id = port - 1; + + if (!ib_query_port(hca, port, &attr)) + priv->max_ib_mtu = ib_mtu_enum_to_int(attr.max_mtu); + else { + printk(KERN_WARNING "%s: ib_query_port %d failed\n", + hca->name, port); + goto device_init_failed; + } + + /* MTU will be reset when mcast join happens */ + priv->dev->mtu = IPOIB_UD_MTU(priv->max_ib_mtu); + priv->mcast_mtu = priv->admin_mtu = priv->dev->mtu; + + priv->dev->neigh_priv_len = sizeof(struct ipoib_neigh); result = ib_query_pkey(hca, port, 0, &priv->pkey); if (result) { printk(KERN_WARNING "%s: ib_query_pkey port %d failed (ret = %d)\n", hca->name, port, result); - goto alloc_mem_failed; + goto device_init_failed; } + if (ipoib_set_dev_features(priv, hca)) + goto device_init_failed; + + /* + * Set the full membership bit, so that we join the right + * broadcast group, etc. + */ + priv->pkey |= 0x8000; + priv->dev->broadcast[8] = priv->pkey >> 8; priv->dev->broadcast[9] = priv->pkey & 0xff; @@ -944,11 +1585,10 @@ static struct net_device *ipoib_add_port(const char *format, if (result) { printk(KERN_WARNING "%s: ib_query_gid port %d failed (ret = %d)\n", hca->name, port, result); - goto alloc_mem_failed; + goto device_init_failed; } else memcpy(priv->dev->dev_addr + 4, priv->local_gid.raw, sizeof (union ib_gid)); - result = ipoib_dev_init(priv->dev, hca, port); if (result < 0) { printk(KERN_WARNING "%s: failed to initialize port %d (ret = %d)\n", @@ -973,28 +1613,31 @@ static struct net_device *ipoib_add_port(const char *format, goto register_failed; } - if (ipoib_create_debug_file(priv->dev)) - goto debug_failed; + ipoib_create_debug_files(priv->dev); + if (ipoib_cm_add_mode_attr(priv->dev)) + goto sysfs_failed; if (ipoib_add_pkey_attr(priv->dev)) goto sysfs_failed; - if (class_device_create_file(&priv->dev->class_dev, - &class_device_attr_create_child)) + if (ipoib_add_umcast_attr(priv->dev)) + goto sysfs_failed; + if (device_create_file(&priv->dev->dev, &dev_attr_create_child)) goto sysfs_failed; - if (class_device_create_file(&priv->dev->class_dev, - &class_device_attr_delete_child)) + if (device_create_file(&priv->dev->dev, &dev_attr_delete_child)) goto sysfs_failed; return priv->dev; sysfs_failed: - ipoib_delete_debug_file(priv->dev); - -debug_failed: + ipoib_delete_debug_files(priv->dev); unregister_netdev(priv->dev); register_failed: ib_unregister_event_handler(&priv->event_handler); + /* Stop GC if started before flush */ + set_bit(IPOIB_STOP_NEIGH_GC, &priv->flags); + cancel_delayed_work(&priv->neigh_reap_task); + flush_workqueue(ipoib_workqueue); event_failed: ipoib_dev_cleanup(priv->dev); @@ -1013,13 +1656,16 @@ static void ipoib_add_one(struct ib_device *device) struct ipoib_dev_priv *priv; int s, e, p; + if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) + return; + dev_list = kmalloc(sizeof *dev_list, GFP_KERNEL); if (!dev_list) return; INIT_LIST_HEAD(dev_list); - if (device->node_type == IB_NODE_SWITCH) { + if (device->node_type == RDMA_NODE_IB_SWITCH) { s = 0; e = 0; } else { @@ -1028,6 +1674,8 @@ static void ipoib_add_one(struct ib_device *device) } for (p = s; p <= e; ++p) { + if (rdma_port_get_link_layer(device, p) != IB_LINK_LAYER_INFINIBAND) + continue; dev = ipoib_add_port("ib%d", device, p); if (!IS_ERR(dev)) { priv = netdev_priv(dev); @@ -1043,21 +1691,53 @@ static void ipoib_remove_one(struct ib_device *device) struct ipoib_dev_priv *priv, *tmp; struct list_head *dev_list; + if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) + return; + dev_list = ib_get_client_data(device, &ipoib_client); + if (!dev_list) + return; list_for_each_entry_safe(priv, tmp, dev_list, list) { ib_unregister_event_handler(&priv->event_handler); + rtnl_lock(); + dev_change_flags(priv->dev, priv->dev->flags & ~IFF_UP); + rtnl_unlock(); + + /* Stop GC */ + set_bit(IPOIB_STOP_NEIGH_GC, &priv->flags); + cancel_delayed_work(&priv->neigh_reap_task); + flush_workqueue(ipoib_workqueue); + unregister_netdev(priv->dev); - ipoib_dev_cleanup(priv->dev); free_netdev(priv->dev); } + + kfree(dev_list); } static int __init ipoib_init_module(void) { int ret; + ipoib_recvq_size = roundup_pow_of_two(ipoib_recvq_size); + ipoib_recvq_size = min(ipoib_recvq_size, IPOIB_MAX_QUEUE_SIZE); + ipoib_recvq_size = max(ipoib_recvq_size, IPOIB_MIN_QUEUE_SIZE); + + ipoib_sendq_size = roundup_pow_of_two(ipoib_sendq_size); + ipoib_sendq_size = min(ipoib_sendq_size, IPOIB_MAX_QUEUE_SIZE); + ipoib_sendq_size = max3(ipoib_sendq_size, 2 * MAX_SEND_CQE, IPOIB_MIN_QUEUE_SIZE); +#ifdef CONFIG_INFINIBAND_IPOIB_CM + ipoib_max_conn_qp = min(ipoib_max_conn_qp, IPOIB_CM_MAX_CONN_QP); +#endif + + /* + * When copying small received packets, we only copy from the + * linear data part of the SKB, so we rely on this condition. + */ + BUILD_BUG_ON(IPOIB_CM_COPYBREAK > IPOIB_CM_HEAD_SIZE); + ret = ipoib_register_debugfs(); if (ret) return ret; @@ -1076,13 +1756,23 @@ static int __init ipoib_init_module(void) goto err_fs; } + ib_sa_register_client(&ipoib_sa_client); + ret = ib_register_client(&ipoib_client); if (ret) - goto err_wq; + goto err_sa; + + ret = ipoib_netlink_init(); + if (ret) + goto err_client; return 0; -err_wq: +err_client: + ib_unregister_client(&ipoib_client); + +err_sa: + ib_sa_unregister_client(&ipoib_sa_client); destroy_workqueue(ipoib_workqueue); err_fs: @@ -1093,7 +1783,9 @@ err_fs: static void __exit ipoib_cleanup_module(void) { + ipoib_netlink_fini(); ib_unregister_client(&ipoib_client); + ib_sa_unregister_client(&ipoib_sa_client); ipoib_unregister_debugfs(); destroy_workqueue(ipoib_workqueue); } |