diff options
Diffstat (limited to 'kernel/drivers/infiniband/ulp/ipoib')
| -rw-r--r-- | kernel/drivers/infiniband/ulp/ipoib/ipoib.h | 37 | ||||
| -rw-r--r-- | kernel/drivers/infiniband/ulp/ipoib/ipoib_cm.c | 41 | ||||
| -rw-r--r-- | kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c | 71 | ||||
| -rw-r--r-- | kernel/drivers/infiniband/ulp/ipoib/ipoib_main.c | 300 | ||||
| -rw-r--r-- | kernel/drivers/infiniband/ulp/ipoib/ipoib_multicast.c | 62 | ||||
| -rw-r--r-- | kernel/drivers/infiniband/ulp/ipoib/ipoib_verbs.c | 36 |
6 files changed, 408 insertions, 139 deletions
diff --git a/kernel/drivers/infiniband/ulp/ipoib/ipoib.h b/kernel/drivers/infiniband/ulp/ipoib/ipoib.h index bd94b0a6e..3ede10309 100644 --- a/kernel/drivers/infiniband/ulp/ipoib/ipoib.h +++ b/kernel/drivers/infiniband/ulp/ipoib/ipoib.h @@ -80,7 +80,7 @@ enum { IPOIB_NUM_WC = 4, IPOIB_MAX_PATH_REC_QUEUE = 3, - IPOIB_MAX_MCAST_QUEUE = 3, + IPOIB_MAX_MCAST_QUEUE = 64, IPOIB_FLAG_OPER_UP = 0, IPOIB_FLAG_INITIALIZED = 1, @@ -239,7 +239,7 @@ struct ipoib_cm_tx { struct net_device *dev; struct ipoib_neigh *neigh; struct ipoib_path *path; - struct ipoib_cm_tx_buf *tx_ring; + struct ipoib_tx_buf *tx_ring; unsigned tx_head; unsigned tx_tail; unsigned long flags; @@ -342,7 +342,6 @@ struct ipoib_dev_priv { u16 pkey; u16 pkey_index; struct ib_pd *pd; - struct ib_mr *mr; struct ib_cq *recv_cq; struct ib_cq *send_cq; struct ib_qp *qp; @@ -361,7 +360,7 @@ struct ipoib_dev_priv { unsigned tx_head; unsigned tx_tail; struct ib_sge tx_sge[MAX_SKB_FRAGS + 1]; - struct ib_send_wr tx_wr; + struct ib_ud_wr tx_wr; unsigned tx_outstanding; struct ib_wc send_wc[MAX_SEND_CQE]; @@ -496,6 +495,7 @@ void ipoib_dev_cleanup(struct net_device *dev); void ipoib_mcast_join_task(struct work_struct *work); void ipoib_mcast_carrier_on_task(struct work_struct *work); void ipoib_mcast_send(struct net_device *dev, u8 *daddr, struct sk_buff *skb); +void ipoib_mcast_free(struct ipoib_mcast *mc); void ipoib_mcast_restart_task(struct work_struct *work); int ipoib_mcast_start_thread(struct net_device *dev); @@ -504,6 +504,33 @@ int ipoib_mcast_stop_thread(struct net_device *dev); void ipoib_mcast_dev_down(struct net_device *dev); void ipoib_mcast_dev_flush(struct net_device *dev); +int ipoib_dma_map_tx(struct ib_device *ca, struct ipoib_tx_buf *tx_req); +void ipoib_dma_unmap_tx(struct ipoib_dev_priv *priv, + struct ipoib_tx_buf *tx_req); + +static inline void ipoib_build_sge(struct ipoib_dev_priv *priv, + struct ipoib_tx_buf *tx_req) +{ + int i, off; + struct sk_buff *skb = tx_req->skb; + skb_frag_t *frags = skb_shinfo(skb)->frags; + int nr_frags = skb_shinfo(skb)->nr_frags; + u64 *mapping = tx_req->mapping; + + if (skb_headlen(skb)) { + priv->tx_sge[0].addr = mapping[0]; + priv->tx_sge[0].length = skb_headlen(skb); + off = 1; + } else + off = 0; + + for (i = 0; i < nr_frags; ++i) { + priv->tx_sge[i + off].addr = mapping[i + off]; + priv->tx_sge[i + off].length = skb_frag_size(&frags[i]); + } + priv->tx_wr.wr.num_sge = nr_frags + off; +} + #ifdef CONFIG_INFINIBAND_IPOIB_DEBUG struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev); int ipoib_mcast_iter_next(struct ipoib_mcast_iter *iter); @@ -522,6 +549,8 @@ void ipoib_path_iter_read(struct ipoib_path_iter *iter, int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid, int set_qkey); +int ipoib_mcast_leave(struct net_device *dev, struct ipoib_mcast *mcast); +struct ipoib_mcast *__ipoib_mcast_find(struct net_device *dev, void *mgid); int ipoib_init_qp(struct net_device *dev); int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca); diff --git a/kernel/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/kernel/drivers/infiniband/ulp/ipoib/ipoib_cm.c index cf32a778e..3ae9726ef 100644 --- a/kernel/drivers/infiniband/ulp/ipoib/ipoib_cm.c +++ b/kernel/drivers/infiniband/ulp/ipoib/ipoib_cm.c @@ -332,7 +332,7 @@ static void ipoib_cm_init_rx_wr(struct net_device *dev, int i; for (i = 0; i < priv->cm.num_frags; ++i) - sge[i].lkey = priv->mr->lkey; + sge[i].lkey = priv->pd->local_dma_lkey; sge[0].length = IPOIB_CM_HEAD_SIZE; for (i = 1; i < priv->cm.num_frags; ++i) @@ -694,24 +694,21 @@ repost: static inline int post_send(struct ipoib_dev_priv *priv, struct ipoib_cm_tx *tx, unsigned int wr_id, - u64 addr, int len) + struct ipoib_tx_buf *tx_req) { struct ib_send_wr *bad_wr; - priv->tx_sge[0].addr = addr; - priv->tx_sge[0].length = len; + ipoib_build_sge(priv, tx_req); - priv->tx_wr.num_sge = 1; - priv->tx_wr.wr_id = wr_id | IPOIB_OP_CM; + priv->tx_wr.wr.wr_id = wr_id | IPOIB_OP_CM; - return ib_post_send(tx->qp, &priv->tx_wr, &bad_wr); + return ib_post_send(tx->qp, &priv->tx_wr.wr, &bad_wr); } void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx) { struct ipoib_dev_priv *priv = netdev_priv(dev); - struct ipoib_cm_tx_buf *tx_req; - u64 addr; + struct ipoib_tx_buf *tx_req; int rc; if (unlikely(skb->len > tx->mtu)) { @@ -735,24 +732,21 @@ void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_ */ tx_req = &tx->tx_ring[tx->tx_head & (ipoib_sendq_size - 1)]; tx_req->skb = skb; - addr = ib_dma_map_single(priv->ca, skb->data, skb->len, DMA_TO_DEVICE); - if (unlikely(ib_dma_mapping_error(priv->ca, addr))) { + + if (unlikely(ipoib_dma_map_tx(priv->ca, tx_req))) { ++dev->stats.tx_errors; dev_kfree_skb_any(skb); return; } - tx_req->mapping = addr; - skb_orphan(skb); skb_dst_drop(skb); - rc = post_send(priv, tx, tx->tx_head & (ipoib_sendq_size - 1), - addr, skb->len); + rc = post_send(priv, tx, tx->tx_head & (ipoib_sendq_size - 1), tx_req); if (unlikely(rc)) { ipoib_warn(priv, "post_send failed, error %d\n", rc); ++dev->stats.tx_errors; - ib_dma_unmap_single(priv->ca, addr, skb->len, DMA_TO_DEVICE); + ipoib_dma_unmap_tx(priv, tx_req); dev_kfree_skb_any(skb); } else { dev->trans_start = jiffies; @@ -777,7 +771,7 @@ void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_cm_tx *tx = wc->qp->qp_context; unsigned int wr_id = wc->wr_id & ~IPOIB_OP_CM; - struct ipoib_cm_tx_buf *tx_req; + struct ipoib_tx_buf *tx_req; unsigned long flags; ipoib_dbg_data(priv, "cm send completion: id %d, status: %d\n", @@ -791,7 +785,7 @@ void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) tx_req = &tx->tx_ring[wr_id]; - ib_dma_unmap_single(priv->ca, tx_req->mapping, tx_req->skb->len, DMA_TO_DEVICE); + ipoib_dma_unmap_tx(priv, tx_req); /* FIXME: is this right? Shouldn't we only increment on success? */ ++dev->stats.tx_packets; @@ -854,7 +848,7 @@ int ipoib_cm_dev_open(struct net_device *dev) } ret = ib_cm_listen(priv->cm.id, cpu_to_be64(IPOIB_CM_IETF_ID | priv->qp->qp_num), - 0, NULL); + 0); if (ret) { printk(KERN_WARNING "%s: failed to listen on ID 0x%llx\n", priv->ca->name, IPOIB_CM_IETF_ID | priv->qp->qp_num); @@ -1036,6 +1030,9 @@ static struct ib_qp *ipoib_cm_create_tx_qp(struct net_device *dev, struct ipoib_ struct ib_qp *tx_qp; + if (dev->features & NETIF_F_SG) + attr.cap.max_send_sge = MAX_SKB_FRAGS + 1; + tx_qp = ib_create_qp(priv->pd, &attr); if (PTR_ERR(tx_qp) == -EINVAL) { ipoib_warn(priv, "can't use GFP_NOIO for QPs on device %s, using GFP_KERNEL\n", @@ -1170,7 +1167,7 @@ err_tx: static void ipoib_cm_tx_destroy(struct ipoib_cm_tx *p) { struct ipoib_dev_priv *priv = netdev_priv(p->dev); - struct ipoib_cm_tx_buf *tx_req; + struct ipoib_tx_buf *tx_req; unsigned long begin; ipoib_dbg(priv, "Destroy active connection 0x%x head 0x%x tail 0x%x\n", @@ -1197,8 +1194,7 @@ timeout: while ((int) p->tx_tail - (int) p->tx_head < 0) { tx_req = &p->tx_ring[p->tx_tail & (ipoib_sendq_size - 1)]; - ib_dma_unmap_single(priv->ca, tx_req->mapping, tx_req->skb->len, - DMA_TO_DEVICE); + ipoib_dma_unmap_tx(priv, tx_req); dev_kfree_skb_any(tx_req->skb); ++p->tx_tail; netif_tx_lock_bh(p->dev); @@ -1455,7 +1451,6 @@ static void ipoib_cm_stale_task(struct work_struct *work) spin_unlock_irq(&priv->lock); } - static ssize_t show_mode(struct device *d, struct device_attribute *attr, char *buf) { diff --git a/kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c index 63b92cbb2..5ea0c1407 100644 --- a/kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c +++ b/kernel/drivers/infiniband/ulp/ipoib/ipoib_ib.c @@ -263,8 +263,7 @@ repost: "for buf %d\n", wr_id); } -static int ipoib_dma_map_tx(struct ib_device *ca, - struct ipoib_tx_buf *tx_req) +int ipoib_dma_map_tx(struct ib_device *ca, struct ipoib_tx_buf *tx_req) { struct sk_buff *skb = tx_req->skb; u64 *mapping = tx_req->mapping; @@ -305,8 +304,8 @@ partial_error: return -EIO; } -static void ipoib_dma_unmap_tx(struct ib_device *ca, - struct ipoib_tx_buf *tx_req) +void ipoib_dma_unmap_tx(struct ipoib_dev_priv *priv, + struct ipoib_tx_buf *tx_req) { struct sk_buff *skb = tx_req->skb; u64 *mapping = tx_req->mapping; @@ -314,7 +313,8 @@ static void ipoib_dma_unmap_tx(struct ib_device *ca, int off; if (skb_headlen(skb)) { - ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE); + ib_dma_unmap_single(priv->ca, mapping[0], skb_headlen(skb), + DMA_TO_DEVICE); off = 1; } else off = 0; @@ -322,8 +322,8 @@ static void ipoib_dma_unmap_tx(struct ib_device *ca, for (i = 0; i < skb_shinfo(skb)->nr_frags; ++i) { const skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; - ib_dma_unmap_page(ca, mapping[i + off], skb_frag_size(frag), - DMA_TO_DEVICE); + ib_dma_unmap_page(priv->ca, mapping[i + off], + skb_frag_size(frag), DMA_TO_DEVICE); } } @@ -389,7 +389,7 @@ static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) tx_req = &priv->tx_ring[wr_id]; - ipoib_dma_unmap_tx(priv->ca, tx_req); + ipoib_dma_unmap_tx(priv, tx_req); ++dev->stats.tx_packets; dev->stats.tx_bytes += tx_req->skb->len; @@ -514,37 +514,23 @@ static inline int post_send(struct ipoib_dev_priv *priv, void *head, int hlen) { struct ib_send_wr *bad_wr; - int i, off; struct sk_buff *skb = tx_req->skb; - skb_frag_t *frags = skb_shinfo(skb)->frags; - int nr_frags = skb_shinfo(skb)->nr_frags; - u64 *mapping = tx_req->mapping; - if (skb_headlen(skb)) { - priv->tx_sge[0].addr = mapping[0]; - priv->tx_sge[0].length = skb_headlen(skb); - off = 1; - } else - off = 0; + ipoib_build_sge(priv, tx_req); - for (i = 0; i < nr_frags; ++i) { - priv->tx_sge[i + off].addr = mapping[i + off]; - priv->tx_sge[i + off].length = skb_frag_size(&frags[i]); - } - priv->tx_wr.num_sge = nr_frags + off; - priv->tx_wr.wr_id = wr_id; - priv->tx_wr.wr.ud.remote_qpn = qpn; - priv->tx_wr.wr.ud.ah = address; + priv->tx_wr.wr.wr_id = wr_id; + priv->tx_wr.remote_qpn = qpn; + priv->tx_wr.ah = address; if (head) { - priv->tx_wr.wr.ud.mss = skb_shinfo(skb)->gso_size; - priv->tx_wr.wr.ud.header = head; - priv->tx_wr.wr.ud.hlen = hlen; - priv->tx_wr.opcode = IB_WR_LSO; + priv->tx_wr.mss = skb_shinfo(skb)->gso_size; + priv->tx_wr.header = head; + priv->tx_wr.hlen = hlen; + priv->tx_wr.wr.opcode = IB_WR_LSO; } else - priv->tx_wr.opcode = IB_WR_SEND; + priv->tx_wr.wr.opcode = IB_WR_SEND; - return ib_post_send(priv->qp, &priv->tx_wr, &bad_wr); + return ib_post_send(priv->qp, &priv->tx_wr.wr, &bad_wr); } void ipoib_send(struct net_device *dev, struct sk_buff *skb, @@ -597,9 +583,9 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb, } if (skb->ip_summed == CHECKSUM_PARTIAL) - priv->tx_wr.send_flags |= IB_SEND_IP_CSUM; + priv->tx_wr.wr.send_flags |= IB_SEND_IP_CSUM; else - priv->tx_wr.send_flags &= ~IB_SEND_IP_CSUM; + priv->tx_wr.wr.send_flags &= ~IB_SEND_IP_CSUM; if (++priv->tx_outstanding == ipoib_sendq_size) { ipoib_dbg(priv, "TX ring full, stopping kernel net queue\n"); @@ -617,7 +603,7 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb, ipoib_warn(priv, "post_send failed, error %d\n", rc); ++dev->stats.tx_errors; --priv->tx_outstanding; - ipoib_dma_unmap_tx(priv->ca, tx_req); + ipoib_dma_unmap_tx(priv, tx_req); dev_kfree_skb_any(skb); if (netif_queue_stopped(dev)) netif_wake_queue(dev); @@ -868,7 +854,7 @@ int ipoib_ib_dev_stop(struct net_device *dev) while ((int) priv->tx_tail - (int) priv->tx_head < 0) { tx_req = &priv->tx_ring[priv->tx_tail & (ipoib_sendq_size - 1)]; - ipoib_dma_unmap_tx(priv->ca, tx_req); + ipoib_dma_unmap_tx(priv, tx_req); dev_kfree_skb_any(tx_req->skb); ++priv->tx_tail; --priv->tx_outstanding; @@ -985,20 +971,21 @@ static inline int update_child_pkey(struct ipoib_dev_priv *priv) } static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv, - enum ipoib_flush_level level) + enum ipoib_flush_level level, + int nesting) { struct ipoib_dev_priv *cpriv; struct net_device *dev = priv->dev; int result; - down_read(&priv->vlan_rwsem); + down_read_nested(&priv->vlan_rwsem, nesting); /* * Flush any child interfaces too -- they might be up even if * the parent is down. */ list_for_each_entry(cpriv, &priv->child_intfs, list) - __ipoib_ib_dev_flush(cpriv, level); + __ipoib_ib_dev_flush(cpriv, level, nesting + 1); up_read(&priv->vlan_rwsem); @@ -1076,7 +1063,7 @@ void ipoib_ib_dev_flush_light(struct work_struct *work) struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, flush_light); - __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_LIGHT); + __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_LIGHT, 0); } void ipoib_ib_dev_flush_normal(struct work_struct *work) @@ -1084,7 +1071,7 @@ void ipoib_ib_dev_flush_normal(struct work_struct *work) struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, flush_normal); - __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_NORMAL); + __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_NORMAL, 0); } void ipoib_ib_dev_flush_heavy(struct work_struct *work) @@ -1092,7 +1079,7 @@ void ipoib_ib_dev_flush_heavy(struct work_struct *work) struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, flush_heavy); - __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_HEAVY); + __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_HEAVY, 0); } void ipoib_ib_dev_cleanup(struct net_device *dev) diff --git a/kernel/drivers/infiniband/ulp/ipoib/ipoib_main.c b/kernel/drivers/infiniband/ulp/ipoib/ipoib_main.c index 9e1b203d7..7d3281866 100644 --- a/kernel/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/kernel/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -48,6 +48,9 @@ #include <linux/jhash.h> #include <net/arp.h> +#include <net/addrconf.h> +#include <linux/inetdevice.h> +#include <rdma/ib_cache.h> #define DRV_VERSION "1.0.0" @@ -89,13 +92,18 @@ 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_remove_one(struct ib_device *device, void *client_data); static void ipoib_neigh_reclaim(struct rcu_head *rp); +static struct net_device *ipoib_get_net_dev_by_params( + struct ib_device *dev, u8 port, u16 pkey, + const union ib_gid *gid, const struct sockaddr *addr, + void *client_data); static struct ib_client ipoib_client = { .name = "ipoib", .add = ipoib_add_one, - .remove = ipoib_remove_one + .remove = ipoib_remove_one, + .get_net_dev_by_params = ipoib_get_net_dev_by_params, }; int ipoib_open(struct net_device *dev) @@ -190,7 +198,7 @@ static netdev_features_t ipoib_fix_features(struct net_device *dev, netdev_featu 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); + features &= ~(NETIF_F_IP_CSUM | NETIF_F_TSO); return features; } @@ -222,6 +230,225 @@ static int ipoib_change_mtu(struct net_device *dev, int new_mtu) return 0; } +/* Called with an RCU read lock taken */ +static bool ipoib_is_dev_match_addr_rcu(const struct sockaddr *addr, + struct net_device *dev) +{ + struct net *net = dev_net(dev); + struct in_device *in_dev; + struct sockaddr_in *addr_in = (struct sockaddr_in *)addr; + struct sockaddr_in6 *addr_in6 = (struct sockaddr_in6 *)addr; + __be32 ret_addr; + + switch (addr->sa_family) { + case AF_INET: + in_dev = in_dev_get(dev); + if (!in_dev) + return false; + + ret_addr = inet_confirm_addr(net, in_dev, 0, + addr_in->sin_addr.s_addr, + RT_SCOPE_HOST); + in_dev_put(in_dev); + if (ret_addr) + return true; + + break; + case AF_INET6: + if (IS_ENABLED(CONFIG_IPV6) && + ipv6_chk_addr(net, &addr_in6->sin6_addr, dev, 1)) + return true; + + break; + } + return false; +} + +/** + * Find the master net_device on top of the given net_device. + * @dev: base IPoIB net_device + * + * Returns the master net_device with a reference held, or the same net_device + * if no master exists. + */ +static struct net_device *ipoib_get_master_net_dev(struct net_device *dev) +{ + struct net_device *master; + + rcu_read_lock(); + master = netdev_master_upper_dev_get_rcu(dev); + if (master) + dev_hold(master); + rcu_read_unlock(); + + if (master) + return master; + + dev_hold(dev); + return dev; +} + +/** + * Find a net_device matching the given address, which is an upper device of + * the given net_device. + * @addr: IP address to look for. + * @dev: base IPoIB net_device + * + * If found, returns the net_device with a reference held. Otherwise return + * NULL. + */ +static struct net_device *ipoib_get_net_dev_match_addr( + const struct sockaddr *addr, struct net_device *dev) +{ + struct net_device *upper, + *result = NULL; + struct list_head *iter; + + rcu_read_lock(); + if (ipoib_is_dev_match_addr_rcu(addr, dev)) { + dev_hold(dev); + result = dev; + goto out; + } + + netdev_for_each_all_upper_dev_rcu(dev, upper, iter) { + if (ipoib_is_dev_match_addr_rcu(addr, upper)) { + dev_hold(upper); + result = upper; + break; + } + } +out: + rcu_read_unlock(); + return result; +} + +/* returns the number of IPoIB netdevs on top a given ipoib device matching a + * pkey_index and address, if one exists. + * + * @found_net_dev: contains a matching net_device if the return value >= 1, + * with a reference held. */ +static int ipoib_match_gid_pkey_addr(struct ipoib_dev_priv *priv, + const union ib_gid *gid, + u16 pkey_index, + const struct sockaddr *addr, + int nesting, + struct net_device **found_net_dev) +{ + struct ipoib_dev_priv *child_priv; + struct net_device *net_dev = NULL; + int matches = 0; + + if (priv->pkey_index == pkey_index && + (!gid || !memcmp(gid, &priv->local_gid, sizeof(*gid)))) { + if (!addr) { + net_dev = ipoib_get_master_net_dev(priv->dev); + } else { + /* Verify the net_device matches the IP address, as + * IPoIB child devices currently share a GID. */ + net_dev = ipoib_get_net_dev_match_addr(addr, priv->dev); + } + if (net_dev) { + if (!*found_net_dev) + *found_net_dev = net_dev; + else + dev_put(net_dev); + ++matches; + } + } + + /* Check child interfaces */ + down_read_nested(&priv->vlan_rwsem, nesting); + list_for_each_entry(child_priv, &priv->child_intfs, list) { + matches += ipoib_match_gid_pkey_addr(child_priv, gid, + pkey_index, addr, + nesting + 1, + found_net_dev); + if (matches > 1) + break; + } + up_read(&priv->vlan_rwsem); + + return matches; +} + +/* Returns the number of matching net_devs found (between 0 and 2). Also + * return the matching net_device in the @net_dev parameter, holding a + * reference to the net_device, if the number of matches >= 1 */ +static int __ipoib_get_net_dev_by_params(struct list_head *dev_list, u8 port, + u16 pkey_index, + const union ib_gid *gid, + const struct sockaddr *addr, + struct net_device **net_dev) +{ + struct ipoib_dev_priv *priv; + int matches = 0; + + *net_dev = NULL; + + list_for_each_entry(priv, dev_list, list) { + if (priv->port != port) + continue; + + matches += ipoib_match_gid_pkey_addr(priv, gid, pkey_index, + addr, 0, net_dev); + if (matches > 1) + break; + } + + return matches; +} + +static struct net_device *ipoib_get_net_dev_by_params( + struct ib_device *dev, u8 port, u16 pkey, + const union ib_gid *gid, const struct sockaddr *addr, + void *client_data) +{ + struct net_device *net_dev; + struct list_head *dev_list = client_data; + u16 pkey_index; + int matches; + int ret; + + if (!rdma_protocol_ib(dev, port)) + return NULL; + + ret = ib_find_cached_pkey(dev, port, pkey, &pkey_index); + if (ret) + return NULL; + + if (!dev_list) + return NULL; + + /* See if we can find a unique device matching the L2 parameters */ + matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index, + gid, NULL, &net_dev); + + switch (matches) { + case 0: + return NULL; + case 1: + return net_dev; + } + + dev_put(net_dev); + + /* Couldn't find a unique device with L2 parameters only. Use L3 + * address to uniquely match the net device */ + matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index, + gid, addr, &net_dev); + switch (matches) { + case 0: + return NULL; + default: + dev_warn_ratelimited(&dev->dev, + "duplicate IP address detected\n"); + /* Fall through */ + case 1: + return net_dev; + } +} + int ipoib_set_mode(struct net_device *dev, const char *buf) { struct ipoib_dev_priv *priv = netdev_priv(dev); @@ -232,8 +459,9 @@ int ipoib_set_mode(struct net_device *dev, const char *buf) ipoib_warn(priv, "enabling connected mode " "will cause multicast packet drops\n"); netdev_update_features(dev); + dev_set_mtu(dev, ipoib_cm_max_mtu(dev)); rtnl_unlock(); - priv->tx_wr.send_flags &= ~IB_SEND_IP_CSUM; + priv->tx_wr.wr.send_flags &= ~IB_SEND_IP_CSUM; ipoib_flush_paths(dev); rtnl_lock(); @@ -921,6 +1149,9 @@ static void __ipoib_reap_neigh(struct ipoib_dev_priv *priv) unsigned long dt; unsigned long flags; int i; + LIST_HEAD(remove_list); + struct ipoib_mcast *mcast, *tmcast; + struct net_device *dev = priv->dev; if (test_bit(IPOIB_STOP_NEIGH_GC, &priv->flags)) return; @@ -948,6 +1179,19 @@ static void __ipoib_reap_neigh(struct ipoib_dev_priv *priv) lockdep_is_held(&priv->lock))) != NULL) { /* was the neigh idle for two GC periods */ if (time_after(neigh_obsolete, neigh->alive)) { + u8 *mgid = neigh->daddr + 4; + + /* Is this multicast ? */ + if (*mgid == 0xff) { + mcast = __ipoib_mcast_find(dev, mgid); + + if (mcast && test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) { + list_del(&mcast->list); + rb_erase(&mcast->rb_node, &priv->multicast_tree); + list_add_tail(&mcast->list, &remove_list); + } + } + rcu_assign_pointer(*np, rcu_dereference_protected(neigh->hnext, lockdep_is_held(&priv->lock))); @@ -963,6 +1207,10 @@ static void __ipoib_reap_neigh(struct ipoib_dev_priv *priv) out_unlock: spin_unlock_irqrestore(&priv->lock, flags); + list_for_each_entry_safe(mcast, tmcast, &remove_list, list) { + ipoib_mcast_leave(dev, mcast); + ipoib_mcast_free(mcast); + } } static void ipoib_reap_neigh(struct work_struct *work) @@ -1128,7 +1376,7 @@ 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; + struct ipoib_neigh __rcu **buckets; u32 size; clear_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags); @@ -1146,7 +1394,7 @@ static int ipoib_neigh_hash_init(struct ipoib_dev_priv *priv) htbl->size = size; htbl->mask = (size - 1); htbl->buckets = buckets; - ntbl->htbl = htbl; + RCU_INIT_POINTER(ntbl->htbl, htbl); htbl->ntbl = ntbl; atomic_set(&ntbl->entries, 0); @@ -1577,7 +1825,8 @@ static struct net_device *ipoib_add_port(const char *format, SET_NETDEV_DEV(priv->dev, hca->dma_device); priv->dev->dev_id = port - 1; - if (!ib_query_port(hca, port, &attr)) + result = ib_query_port(hca, port, &attr); + if (!result) priv->max_ib_mtu = ib_mtu_enum_to_int(attr.max_mtu); else { printk(KERN_WARNING "%s: ib_query_port %d failed\n", @@ -1598,7 +1847,8 @@ static struct net_device *ipoib_add_port(const char *format, goto device_init_failed; } - if (ipoib_set_dev_features(priv, hca)) + result = ipoib_set_dev_features(priv, hca); + if (result) goto device_init_failed; /* @@ -1610,7 +1860,7 @@ static struct net_device *ipoib_add_port(const char *format, priv->dev->broadcast[8] = priv->pkey >> 8; priv->dev->broadcast[9] = priv->pkey & 0xff; - result = ib_query_gid(hca, port, 0, &priv->local_gid); + result = ib_query_gid(hca, port, 0, &priv->local_gid, NULL); if (result) { printk(KERN_WARNING "%s: ib_query_gid port %d failed (ret = %d)\n", hca->name, port, result); @@ -1684,10 +1934,8 @@ static void ipoib_add_one(struct ib_device *device) struct list_head *dev_list; struct net_device *dev; struct ipoib_dev_priv *priv; - int s, e, p; - - if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) - return; + int p; + int count = 0; dev_list = kmalloc(sizeof *dev_list, GFP_KERNEL); if (!dev_list) @@ -1695,36 +1943,30 @@ static void ipoib_add_one(struct ib_device *device) INIT_LIST_HEAD(dev_list); - if (device->node_type == RDMA_NODE_IB_SWITCH) { - s = 0; - e = 0; - } else { - s = 1; - e = device->phys_port_cnt; - } - - for (p = s; p <= e; ++p) { - if (rdma_port_get_link_layer(device, p) != IB_LINK_LAYER_INFINIBAND) + for (p = rdma_start_port(device); p <= rdma_end_port(device); ++p) { + if (!rdma_protocol_ib(device, p)) continue; dev = ipoib_add_port("ib%d", device, p); if (!IS_ERR(dev)) { priv = netdev_priv(dev); list_add_tail(&priv->list, dev_list); + count++; } } + if (!count) { + kfree(dev_list); + return; + } + ib_set_client_data(device, &ipoib_client, dev_list); } -static void ipoib_remove_one(struct ib_device *device) +static void ipoib_remove_one(struct ib_device *device, void *client_data) { struct ipoib_dev_priv *priv, *tmp; - struct list_head *dev_list; - - if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) - return; + struct list_head *dev_list = client_data; - dev_list = ib_get_client_data(device, &ipoib_client); if (!dev_list) return; diff --git a/kernel/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/kernel/drivers/infiniband/ulp/ipoib/ipoib_multicast.c index 140c94ce7..6dcef673d 100644 --- a/kernel/drivers/infiniband/ulp/ipoib/ipoib_multicast.c +++ b/kernel/drivers/infiniband/ulp/ipoib/ipoib_multicast.c @@ -106,7 +106,7 @@ static void __ipoib_mcast_schedule_join_thread(struct ipoib_dev_priv *priv, queue_delayed_work(priv->wq, &priv->mcast_task, 0); } -static void ipoib_mcast_free(struct ipoib_mcast *mcast) +void ipoib_mcast_free(struct ipoib_mcast *mcast) { struct net_device *dev = mcast->dev; int tx_dropped = 0; @@ -153,7 +153,7 @@ static struct ipoib_mcast *ipoib_mcast_alloc(struct net_device *dev, return mcast; } -static struct ipoib_mcast *__ipoib_mcast_find(struct net_device *dev, void *mgid) +struct ipoib_mcast *__ipoib_mcast_find(struct net_device *dev, void *mgid) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct rb_node *n = priv->multicast_tree.rb_node; @@ -245,7 +245,7 @@ static int ipoib_mcast_join_finish(struct ipoib_mcast *mcast, priv->qkey = be32_to_cpu(priv->broadcast->mcmember.qkey); spin_unlock_irq(&priv->lock); - priv->tx_wr.wr.ud.remote_qkey = priv->qkey; + priv->tx_wr.remote_qkey = priv->qkey; set_qkey = 1; } @@ -393,8 +393,13 @@ static int ipoib_mcast_join_complete(int status, goto out_locked; } } else { - if (mcast->logcount++ < 20) { - if (status == -ETIMEDOUT || status == -EAGAIN) { + bool silent_fail = + test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) && + status == -EINVAL; + + if (mcast->logcount < 20) { + if (status == -ETIMEDOUT || status == -EAGAIN || + silent_fail) { ipoib_dbg_mcast(priv, "%smulticast join failed for %pI6, status %d\n", test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) ? "sendonly " : "", mcast->mcmember.mgid.raw, status); @@ -403,6 +408,9 @@ static int ipoib_mcast_join_complete(int status, test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) ? "sendonly " : "", mcast->mcmember.mgid.raw, status); } + + if (!silent_fail) + mcast->logcount++; } if (test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) && @@ -448,8 +456,7 @@ out_locked: return status; } -static void ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast, - int create) +static void ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_sa_multicast *multicast; @@ -471,7 +478,14 @@ static void ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast, IB_SA_MCMEMBER_REC_PKEY | IB_SA_MCMEMBER_REC_JOIN_STATE; - if (create) { + if (mcast != priv->broadcast) { + /* + * RFC 4391: + * The MGID MUST use the same P_Key, Q_Key, SL, MTU, + * and HopLimit as those used in the broadcast-GID. The rest + * of attributes SHOULD follow the values used in the + * broadcast-GID as well. + */ comp_mask |= IB_SA_MCMEMBER_REC_QKEY | IB_SA_MCMEMBER_REC_MTU_SELECTOR | @@ -492,6 +506,24 @@ static void ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast, rec.sl = priv->broadcast->mcmember.sl; rec.flow_label = priv->broadcast->mcmember.flow_label; rec.hop_limit = priv->broadcast->mcmember.hop_limit; + + /* + * Send-only IB Multicast joins do not work at the core + * IB layer yet, so we can't use them here. However, + * we are emulating an Ethernet multicast send, which + * does not require a multicast subscription and will + * still send properly. The most appropriate thing to + * do is to create the group if it doesn't exist as that + * most closely emulates the behavior, from a user space + * application perspecitive, of Ethernet multicast + * operation. For now, we do a full join, maybe later + * when the core IB layers support send only joins we + * will use them. + */ +#if 0 + if (test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) + rec.join_state = 4; +#endif } multicast = ib_sa_join_multicast(&ipoib_sa_client, priv->ca, priv->port, @@ -517,7 +549,6 @@ void ipoib_mcast_join_task(struct work_struct *work) struct ib_port_attr port_attr; unsigned long delay_until = 0; struct ipoib_mcast *mcast = NULL; - int create = 1; if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)) return; @@ -530,7 +561,7 @@ void ipoib_mcast_join_task(struct work_struct *work) } priv->local_lid = port_attr.lid; - if (ib_query_gid(priv->ca, priv->port, 0, &priv->local_gid)) + if (ib_query_gid(priv->ca, priv->port, 0, &priv->local_gid, NULL)) ipoib_warn(priv, "ib_query_gid() failed\n"); else memcpy(priv->dev->dev_addr + 4, priv->local_gid.raw, sizeof (union ib_gid)); @@ -566,7 +597,6 @@ void ipoib_mcast_join_task(struct work_struct *work) if (IS_ERR_OR_NULL(priv->broadcast->mc) && !test_bit(IPOIB_MCAST_FLAG_BUSY, &priv->broadcast->flags)) { mcast = priv->broadcast; - create = 0; if (mcast->backoff > 1 && time_before(jiffies, mcast->delay_until)) { delay_until = mcast->delay_until; @@ -590,12 +620,8 @@ void ipoib_mcast_join_task(struct work_struct *work) /* Found the next unjoined group */ init_completion(&mcast->done); set_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags); - if (test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) - create = 0; - else - create = 1; spin_unlock_irq(&priv->lock); - ipoib_mcast_join(dev, mcast, create); + ipoib_mcast_join(dev, mcast); spin_lock_irq(&priv->lock); } else if (!delay_until || time_before(mcast->delay_until, delay_until)) @@ -618,7 +644,7 @@ out: } spin_unlock_irq(&priv->lock); if (mcast) - ipoib_mcast_join(dev, mcast, create); + ipoib_mcast_join(dev, mcast); } int ipoib_mcast_start_thread(struct net_device *dev) @@ -651,7 +677,7 @@ int ipoib_mcast_stop_thread(struct net_device *dev) return 0; } -static int ipoib_mcast_leave(struct net_device *dev, struct ipoib_mcast *mcast) +int ipoib_mcast_leave(struct net_device *dev, struct ipoib_mcast *mcast) { struct ipoib_dev_priv *priv = netdev_priv(dev); int ret = 0; diff --git a/kernel/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/kernel/drivers/infiniband/ulp/ipoib/ipoib_verbs.c index 2d13fd08c..d48c5bae7 100644 --- a/kernel/drivers/infiniband/ulp/ipoib/ipoib_verbs.c +++ b/kernel/drivers/infiniband/ulp/ipoib/ipoib_verbs.c @@ -141,6 +141,7 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca) .sq_sig_type = IB_SIGNAL_ALL_WR, .qp_type = IB_QPT_UD }; + struct ib_cq_init_attr cq_attr = {}; int ret, size; int i; @@ -151,12 +152,6 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca) return -ENODEV; } - priv->mr = ib_get_dma_mr(priv->pd, IB_ACCESS_LOCAL_WRITE); - if (IS_ERR(priv->mr)) { - printk(KERN_WARNING "%s: ib_get_dma_mr failed\n", ca->name); - goto out_free_pd; - } - /* * the various IPoIB tasks assume they will never race against * themselves, so always use a single thread workqueue @@ -164,7 +159,7 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca) priv->wq = create_singlethread_workqueue("ipoib_wq"); if (!priv->wq) { printk(KERN_WARNING "ipoib: failed to allocate device WQ\n"); - goto out_free_mr; + goto out_free_pd; } size = ipoib_recvq_size + 1; @@ -179,14 +174,17 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca) if (ret != -ENOSYS) goto out_free_wq; - priv->recv_cq = ib_create_cq(priv->ca, ipoib_ib_completion, NULL, dev, size, 0); + cq_attr.cqe = size; + priv->recv_cq = ib_create_cq(priv->ca, ipoib_ib_completion, NULL, + dev, &cq_attr); if (IS_ERR(priv->recv_cq)) { printk(KERN_WARNING "%s: failed to create receive CQ\n", ca->name); goto out_cm_dev_cleanup; } + cq_attr.cqe = ipoib_sendq_size; priv->send_cq = ib_create_cq(priv->ca, ipoib_send_comp_handler, NULL, - dev, ipoib_sendq_size, 0); + dev, &cq_attr); if (IS_ERR(priv->send_cq)) { printk(KERN_WARNING "%s: failed to create send CQ\n", ca->name); goto out_free_recv_cq; @@ -221,13 +219,13 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca) priv->dev->dev_addr[3] = (priv->qp->qp_num ) & 0xff; for (i = 0; i < MAX_SKB_FRAGS + 1; ++i) - priv->tx_sge[i].lkey = priv->mr->lkey; + priv->tx_sge[i].lkey = priv->pd->local_dma_lkey; - priv->tx_wr.opcode = IB_WR_SEND; - priv->tx_wr.sg_list = priv->tx_sge; - priv->tx_wr.send_flags = IB_SEND_SIGNALED; + priv->tx_wr.wr.opcode = IB_WR_SEND; + priv->tx_wr.wr.sg_list = priv->tx_sge; + priv->tx_wr.wr.send_flags = IB_SEND_SIGNALED; - priv->rx_sge[0].lkey = priv->mr->lkey; + priv->rx_sge[0].lkey = priv->pd->local_dma_lkey; priv->rx_sge[0].length = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu); priv->rx_wr.num_sge = 1; @@ -250,9 +248,6 @@ out_free_wq: destroy_workqueue(priv->wq); priv->wq = NULL; -out_free_mr: - ib_dereg_mr(priv->mr); - out_free_pd: ib_dealloc_pd(priv->pd); @@ -285,12 +280,7 @@ void ipoib_transport_dev_cleanup(struct net_device *dev) priv->wq = NULL; } - if (ib_dereg_mr(priv->mr)) - ipoib_warn(priv, "ib_dereg_mr failed\n"); - - if (ib_dealloc_pd(priv->pd)) - ipoib_warn(priv, "ib_dealloc_pd failed\n"); - + ib_dealloc_pd(priv->pd); } void ipoib_event(struct ib_event_handler *handler, |