diff options
Diffstat (limited to 'drivers/infiniband/ulp/ipoib')
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/Kconfig | 24 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/Makefile | 4 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib.h | 405 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_cm.c | 819 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_ethtool.c | 109 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_fs.c | 15 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_ib.c | 684 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_main.c | 1148 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_multicast.c | 323 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_netlink.c | 182 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_verbs.c | 154 | ||||
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_vlan.c | 161 |
12 files changed, 2801 insertions, 1227 deletions
diff --git a/drivers/infiniband/ulp/ipoib/Kconfig b/drivers/infiniband/ulp/ipoib/Kconfig index 1f76bad020f..cda8eac55ff 100644 --- a/drivers/infiniband/ulp/ipoib/Kconfig +++ b/drivers/infiniband/ulp/ipoib/Kconfig @@ -10,20 +10,21 @@ config INFINIBAND_IPOIB config INFINIBAND_IPOIB_CM bool "IP-over-InfiniBand Connected Mode support" - depends on INFINIBAND_IPOIB && EXPERIMENTAL + depends on INFINIBAND_IPOIB default n ---help--- - This option enables experimental support for IPoIB connected mode. - After enabling this option, you need to switch to connected mode through - /sys/class/net/ibXXX/mode to actually create connections, and then increase - the interface MTU with e.g. ifconfig ib0 mtu 65520. + This option enables support for IPoIB connected mode. After + enabling this option, you need to switch to connected mode + through /sys/class/net/ibXXX/mode to actually create + connections, and then increase the interface MTU with + e.g. ifconfig ib0 mtu 65520. - WARNING: Enabling connected mode will trigger some - packet drops for multicast and UD mode traffic from this interface, + WARNING: Enabling connected mode will trigger some packet + drops for multicast and UD mode traffic from this interface, unless you limit mtu for these destinations to 2044. config INFINIBAND_IPOIB_DEBUG - bool "IP-over-InfiniBand debugging" if EMBEDDED + bool "IP-over-InfiniBand debugging" if EXPERT depends on INFINIBAND_IPOIB default y ---help--- @@ -32,9 +33,10 @@ config INFINIBAND_IPOIB_DEBUG debug_level and mcast_debug_level module parameters (which can also be set after the driver is loaded through sysfs). - This option also creates an "ipoib_debugfs," which can be - mounted to expose debugging information about IB multicast - groups used by the IPoIB driver. + This option also creates a directory tree under ipoib/ in + debugfs, which contains files that expose debugging + information about IB multicast groups used by the IPoIB + driver. config INFINIBAND_IPOIB_DEBUG_DATA bool "IP-over-InfiniBand data path debugging" diff --git a/drivers/infiniband/ulp/ipoib/Makefile b/drivers/infiniband/ulp/ipoib/Makefile index 98ee38e8c2c..e5430dd5076 100644 --- a/drivers/infiniband/ulp/ipoib/Makefile +++ b/drivers/infiniband/ulp/ipoib/Makefile @@ -4,7 +4,9 @@ ib_ipoib-y := ipoib_main.o \ ipoib_ib.o \ ipoib_multicast.o \ ipoib_verbs.o \ - ipoib_vlan.o + ipoib_vlan.o \ + ipoib_ethtool.o \ + ipoib_netlink.o ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_CM) += ipoib_cm.o ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_DEBUG) += ipoib_fs.o diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h index 285c143115c..c639f90cfda 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib.h +++ b/drivers/infiniband/ulp/ipoib/ipoib.h @@ -30,8 +30,6 @@ * 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.h 1358 2004-12-17 22:00:11Z roland $ */ #ifndef _IPOIB_H @@ -46,62 +44,82 @@ #include <linux/mutex.h> #include <net/neighbour.h> +#include <net/sch_generic.h> -#include <asm/atomic.h> +#include <linux/atomic.h> #include <rdma/ib_verbs.h> #include <rdma/ib_pack.h> #include <rdma/ib_sa.h> +#include <linux/sched.h> /* constants */ +enum ipoib_flush_level { + IPOIB_FLUSH_LIGHT, + IPOIB_FLUSH_NORMAL, + IPOIB_FLUSH_HEAVY +}; + enum { - IPOIB_PACKET_SIZE = 2048, - IPOIB_BUF_SIZE = IPOIB_PACKET_SIZE + IB_GRH_BYTES, + IPOIB_ENCAP_LEN = 4, - IPOIB_ENCAP_LEN = 4, + IPOIB_UD_HEAD_SIZE = IB_GRH_BYTES + IPOIB_ENCAP_LEN, + IPOIB_UD_RX_SG = 2, /* max buffer needed for 4K mtu */ - IPOIB_CM_MTU = 0x10000 - 0x10, /* padding to align header to 16 */ - IPOIB_CM_BUF_SIZE = IPOIB_CM_MTU + IPOIB_ENCAP_LEN, - IPOIB_CM_HEAD_SIZE = IPOIB_CM_BUF_SIZE % PAGE_SIZE, - IPOIB_CM_RX_SG = ALIGN(IPOIB_CM_BUF_SIZE, PAGE_SIZE) / PAGE_SIZE, - IPOIB_RX_RING_SIZE = 128, - IPOIB_TX_RING_SIZE = 64, + IPOIB_CM_MTU = 0x10000 - 0x10, /* padding to align header to 16 */ + IPOIB_CM_BUF_SIZE = IPOIB_CM_MTU + IPOIB_ENCAP_LEN, + IPOIB_CM_HEAD_SIZE = IPOIB_CM_BUF_SIZE % PAGE_SIZE, + IPOIB_CM_RX_SG = ALIGN(IPOIB_CM_BUF_SIZE, PAGE_SIZE) / PAGE_SIZE, + IPOIB_RX_RING_SIZE = 256, + IPOIB_TX_RING_SIZE = 128, IPOIB_MAX_QUEUE_SIZE = 8192, IPOIB_MIN_QUEUE_SIZE = 2, + IPOIB_CM_MAX_CONN_QP = 4096, - IPOIB_NUM_WC = 4, + IPOIB_NUM_WC = 4, IPOIB_MAX_PATH_REC_QUEUE = 3, - IPOIB_MAX_MCAST_QUEUE = 3, - - IPOIB_FLAG_OPER_UP = 0, - IPOIB_FLAG_INITIALIZED = 1, - IPOIB_FLAG_ADMIN_UP = 2, - IPOIB_PKEY_ASSIGNED = 3, - IPOIB_PKEY_STOP = 4, - IPOIB_FLAG_SUBINTERFACE = 5, - IPOIB_MCAST_RUN = 6, - IPOIB_STOP_REAPER = 7, - IPOIB_MCAST_STARTED = 8, - IPOIB_FLAG_NETIF_STOPPED = 9, - IPOIB_FLAG_ADMIN_CM = 10, + IPOIB_MAX_MCAST_QUEUE = 3, + + IPOIB_FLAG_OPER_UP = 0, + IPOIB_FLAG_INITIALIZED = 1, + IPOIB_FLAG_ADMIN_UP = 2, + IPOIB_PKEY_ASSIGNED = 3, + IPOIB_PKEY_STOP = 4, + IPOIB_FLAG_SUBINTERFACE = 5, + IPOIB_MCAST_RUN = 6, + IPOIB_STOP_REAPER = 7, + IPOIB_FLAG_ADMIN_CM = 9, + IPOIB_FLAG_UMCAST = 10, + IPOIB_STOP_NEIGH_GC = 11, + IPOIB_NEIGH_TBL_FLUSH = 12, IPOIB_MAX_BACKOFF_SECONDS = 16, - IPOIB_MCAST_FLAG_FOUND = 0, /* used in set_multicast_list */ + IPOIB_MCAST_FLAG_FOUND = 0, /* used in set_multicast_list */ IPOIB_MCAST_FLAG_SENDONLY = 1, - IPOIB_MCAST_FLAG_BUSY = 2, /* joining or already joined */ + IPOIB_MCAST_FLAG_BUSY = 2, /* joining or already joined */ IPOIB_MCAST_FLAG_ATTACHED = 3, + IPOIB_MCAST_JOIN_STARTED = 4, + + MAX_SEND_CQE = 16, + IPOIB_CM_COPYBREAK = 256, + + IPOIB_NON_CHILD = 0, + IPOIB_LEGACY_CHILD = 1, + IPOIB_RTNL_CHILD = 2, }; #define IPOIB_OP_RECV (1ul << 31) #ifdef CONFIG_INFINIBAND_IPOIB_CM -#define IPOIB_CM_OP_SRQ (1ul << 30) +#define IPOIB_OP_CM (1ul << 30) #else -#define IPOIB_CM_OP_SRQ (0) +#define IPOIB_OP_CM (0) #endif +#define IPOIB_QPN_MASK ((__force u32) cpu_to_be32(0xFFFFFF)) + /* structs */ struct ipoib_header { @@ -109,19 +127,46 @@ struct ipoib_header { u16 reserved; }; -struct ipoib_pseudoheader { - u8 hwaddr[INFINIBAND_ALEN]; +struct ipoib_cb { + struct qdisc_skb_cb qdisc_cb; + u8 hwaddr[INFINIBAND_ALEN]; }; -struct ipoib_mcast; +/* Used for all multicast joins (broadcast, IPv4 mcast and IPv6 mcast) */ +struct ipoib_mcast { + struct ib_sa_mcmember_rec mcmember; + struct ib_sa_multicast *mc; + struct ipoib_ah *ah; + + struct rb_node rb_node; + struct list_head list; + + unsigned long created; + unsigned long backoff; + + unsigned long flags; + unsigned char logcount; + + struct list_head neigh_list; + + struct sk_buff_head pkt_queue; + + struct net_device *dev; + struct completion done; +}; struct ipoib_rx_buf { struct sk_buff *skb; - u64 mapping; + u64 mapping[IPOIB_UD_RX_SG]; }; struct ipoib_tx_buf { struct sk_buff *skb; + u64 mapping[MAX_SKB_FRAGS + 1]; +}; + +struct ipoib_cm_tx_buf { + struct sk_buff *skb; u64 mapping; }; @@ -166,28 +211,28 @@ enum ipoib_cm_state { }; struct ipoib_cm_rx { - struct ib_cm_id *id; - struct ib_qp *qp; - struct list_head list; - struct net_device *dev; - unsigned long jiffies; - enum ipoib_cm_state state; + struct ib_cm_id *id; + struct ib_qp *qp; + struct ipoib_cm_rx_buf *rx_ring; + struct list_head list; + struct net_device *dev; + unsigned long jiffies; + enum ipoib_cm_state state; + int recv_count; }; struct ipoib_cm_tx { - struct ib_cm_id *id; - struct ib_cq *cq; - struct ib_qp *qp; + struct ib_cm_id *id; + struct ib_qp *qp; struct list_head list; struct net_device *dev; struct ipoib_neigh *neigh; struct ipoib_path *path; - struct ipoib_tx_buf *tx_ring; - unsigned tx_head; - unsigned tx_tail; - unsigned long flags; - u32 mtu; - struct ib_wc ibwc[IPOIB_NUM_WC]; + struct ipoib_cm_tx_buf *tx_ring; + unsigned tx_head; + unsigned tx_tail; + unsigned long flags; + u32 mtu; }; struct ipoib_cm_rx_buf { @@ -196,81 +241,116 @@ struct ipoib_cm_rx_buf { }; struct ipoib_cm_dev_priv { - struct ib_srq *srq; + struct ib_srq *srq; struct ipoib_cm_rx_buf *srq_ring; - struct ib_cm_id *id; - struct list_head passive_ids; /* state: LIVE */ - struct list_head rx_error_list; /* state: ERROR */ - struct list_head rx_flush_list; /* state: FLUSH, drain not started */ - struct list_head rx_drain_list; /* state: FLUSH, drain started */ - struct list_head rx_reap_list; /* state: FLUSH, drain done */ + struct ib_cm_id *id; + struct list_head passive_ids; /* state: LIVE */ + struct list_head rx_error_list; /* state: ERROR */ + struct list_head rx_flush_list; /* state: FLUSH, drain not started */ + struct list_head rx_drain_list; /* state: FLUSH, drain started */ + struct list_head rx_reap_list; /* state: FLUSH, drain done */ struct work_struct start_task; struct work_struct reap_task; struct work_struct skb_task; struct work_struct rx_reap_task; struct delayed_work stale_task; struct sk_buff_head skb_queue; - struct list_head start_list; - struct list_head reap_list; - struct ib_wc ibwc[IPOIB_NUM_WC]; - struct ib_sge rx_sge[IPOIB_CM_RX_SG]; + struct list_head start_list; + struct list_head reap_list; + struct ib_wc ibwc[IPOIB_NUM_WC]; + struct ib_sge rx_sge[IPOIB_CM_RX_SG]; struct ib_recv_wr rx_wr; + int nonsrq_conn_qp; + int max_cm_mtu; + int num_frags; +}; + +struct ipoib_ethtool_st { + u16 coalesce_usecs; + u16 max_coalesced_frames; +}; + +struct ipoib_neigh_table; + +struct ipoib_neigh_hash { + struct ipoib_neigh_table *ntbl; + struct ipoib_neigh __rcu **buckets; + struct rcu_head rcu; + u32 mask; + u32 size; +}; + +struct ipoib_neigh_table { + struct ipoib_neigh_hash __rcu *htbl; + atomic_t entries; + struct completion flushed; + struct completion deleted; }; /* - * Device private locking: tx_lock protects members used in TX fast - * path (and we use LLTX so upper layers don't do extra locking). - * lock protects everything else. lock nests inside of tx_lock (ie - * tx_lock must be acquired first if needed). + * Device private locking: network stack tx_lock protects members used + * in TX fast path, lock protects everything else. lock nests inside + * of tx_lock (ie tx_lock must be acquired first if needed). */ struct ipoib_dev_priv { spinlock_t lock; struct net_device *dev; + struct napi_struct napi; + unsigned long flags; - struct mutex mcast_mutex; - struct mutex vlan_mutex; + struct rw_semaphore vlan_rwsem; struct rb_root path_tree; struct list_head path_list; + struct ipoib_neigh_table ntbl; + struct ipoib_mcast *broadcast; struct list_head multicast_list; struct rb_root multicast_tree; struct delayed_work pkey_poll_task; struct delayed_work mcast_task; - struct work_struct flush_task; + struct work_struct carrier_on_task; + struct work_struct flush_light; + struct work_struct flush_normal; + struct work_struct flush_heavy; struct work_struct restart_task; struct delayed_work ah_reap_task; - struct work_struct pkey_event_task; - + struct delayed_work neigh_reap_task; struct ib_device *ca; - u8 port; - u16 pkey; - u16 pkey_index; - struct ib_pd *pd; - struct ib_mr *mr; - struct ib_cq *cq; - struct ib_qp *qp; - u32 qkey; + u8 port; + 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; + u32 qkey; union ib_gid local_gid; - u16 local_lid; + u16 local_lid; unsigned int admin_mtu; unsigned int mcast_mtu; + unsigned int max_ib_mtu; struct ipoib_rx_buf *rx_ring; - spinlock_t tx_lock; struct ipoib_tx_buf *tx_ring; - unsigned tx_head; - unsigned tx_tail; - struct ib_sge tx_sge; + unsigned tx_head; + unsigned tx_tail; + struct ib_sge tx_sge[MAX_SKB_FRAGS + 1]; struct ib_send_wr tx_wr; + unsigned tx_outstanding; + struct ib_wc send_wc[MAX_SEND_CQE]; + + struct ib_recv_wr rx_wr; + struct ib_sge rx_sge[IPOIB_UD_RX_SG]; struct ib_wc ibwc[IPOIB_NUM_WC]; @@ -278,11 +358,10 @@ struct ipoib_dev_priv { struct ib_event_handler event_handler; - struct net_device_stats stats; - struct net_device *parent; struct list_head child_intfs; struct list_head list; + int child_type; #ifdef CONFIG_INFINIBAND_IPOIB_CM struct ipoib_cm_dev_priv cm; @@ -293,14 +372,17 @@ struct ipoib_dev_priv { struct dentry *mcg_dentry; struct dentry *path_dentry; #endif + int hca_caps; + struct ipoib_ethtool_st ethtool; + struct timer_list poll_timer; }; struct ipoib_ah { struct net_device *dev; - struct ib_ah *ah; + struct ib_ah *ah; struct list_head list; - struct kref ref; - unsigned last_send; + struct kref ref; + unsigned last_send; }; struct ipoib_path { @@ -311,12 +393,13 @@ struct ipoib_path { struct list_head neigh_list; - int query_id; + int query_id; struct ib_sa_query *query; struct completion done; - struct rb_node rb_node; + struct rb_node rb_node; struct list_head list; + int valid; }; struct ipoib_neigh { @@ -324,35 +407,45 @@ struct ipoib_neigh { #ifdef CONFIG_INFINIBAND_IPOIB_CM struct ipoib_cm_tx *cm; #endif - union ib_gid dgid; + u8 daddr[INFINIBAND_ALEN]; struct sk_buff_head queue; - struct neighbour *neighbour; + struct net_device *dev; struct list_head list; + struct ipoib_neigh __rcu *hnext; + struct rcu_head rcu; + atomic_t refcnt; + unsigned long alive; }; -/* - * We stash a pointer to our private neighbour information after our - * hardware address in neigh->ha. The ALIGN() expression here makes - * sure that this pointer is stored aligned so that an unaligned - * load is not needed to dereference it. - */ -static inline struct ipoib_neigh **to_ipoib_neigh(struct neighbour *neigh) +#define IPOIB_UD_MTU(ib_mtu) (ib_mtu - IPOIB_ENCAP_LEN) +#define IPOIB_UD_BUF_SIZE(ib_mtu) (ib_mtu + IB_GRH_BYTES) + +static inline int ipoib_ud_need_sg(unsigned int ib_mtu) { - return (void*) neigh + ALIGN(offsetof(struct neighbour, ha) + - INFINIBAND_ALEN, sizeof(void *)); + return IPOIB_UD_BUF_SIZE(ib_mtu) > PAGE_SIZE; } -struct ipoib_neigh *ipoib_neigh_alloc(struct neighbour *neigh); -void ipoib_neigh_free(struct net_device *dev, struct ipoib_neigh *neigh); +void ipoib_neigh_dtor(struct ipoib_neigh *neigh); +static inline void ipoib_neigh_put(struct ipoib_neigh *neigh) +{ + if (atomic_dec_and_test(&neigh->refcnt)) + ipoib_neigh_dtor(neigh); +} +struct ipoib_neigh *ipoib_neigh_get(struct net_device *dev, u8 *daddr); +struct ipoib_neigh *ipoib_neigh_alloc(u8 *daddr, + struct net_device *dev); +void ipoib_neigh_free(struct ipoib_neigh *neigh); +void ipoib_del_neighs_by_gid(struct net_device *dev, u8 *gid); extern struct workqueue_struct *ipoib_workqueue; /* functions */ -int ipoib_poll(struct net_device *dev, int *budget); +int ipoib_poll(struct napi_struct *napi, int budget); void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr); +void ipoib_send_comp_handler(struct ib_cq *cq, void *dev_ptr); struct ipoib_ah *ipoib_create_ah(struct net_device *dev, struct ib_pd *pd, struct ib_ah_attr *attr); @@ -361,19 +454,22 @@ static inline void ipoib_put_ah(struct ipoib_ah *ah) { kref_put(&ah->ref, ipoib_free_ah); } - int ipoib_open(struct net_device *dev); int ipoib_add_pkey_attr(struct net_device *dev); +int ipoib_add_umcast_attr(struct net_device *dev); void ipoib_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_ah *address, u32 qpn); void ipoib_reap_ah(struct work_struct *work); +void ipoib_mark_paths_invalid(struct net_device *dev); void ipoib_flush_paths(struct net_device *dev); struct ipoib_dev_priv *ipoib_intf_alloc(const char *format); int ipoib_ib_dev_init(struct net_device *dev, struct ib_device *ca, int port); -void ipoib_ib_dev_flush(struct work_struct *work); +void ipoib_ib_dev_flush_light(struct work_struct *work); +void ipoib_ib_dev_flush_normal(struct work_struct *work); +void ipoib_ib_dev_flush_heavy(struct work_struct *work); void ipoib_pkey_event(struct work_struct *work); void ipoib_ib_dev_cleanup(struct net_device *dev); @@ -386,7 +482,8 @@ int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port); void ipoib_dev_cleanup(struct net_device *dev); void ipoib_mcast_join_task(struct work_struct *work); -void ipoib_mcast_send(struct net_device *dev, void *mgid, struct sk_buff *skb); +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_restart_task(struct work_struct *work); int ipoib_mcast_start_thread(struct net_device *dev); @@ -412,9 +509,7 @@ void ipoib_path_iter_read(struct ipoib_path_iter *iter, #endif int ipoib_mcast_attach(struct net_device *dev, u16 mlid, - union ib_gid *mgid); -int ipoib_mcast_detach(struct net_device *dev, u16 mlid, - union ib_gid *mgid); + union ib_gid *mgid, int set_qkey); int ipoib_init_qp(struct net_device *dev); int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca); @@ -426,18 +521,34 @@ void ipoib_event(struct ib_event_handler *handler, int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey); int ipoib_vlan_delete(struct net_device *pdev, unsigned short pkey); +int __ipoib_vlan_add(struct ipoib_dev_priv *ppriv, struct ipoib_dev_priv *priv, + u16 pkey, int child_type); + +int __init ipoib_netlink_init(void); +void __exit ipoib_netlink_fini(void); + +void ipoib_set_umcast(struct net_device *ndev, int umcast_val); +int ipoib_set_mode(struct net_device *dev, const char *buf); + +void ipoib_setup(struct net_device *dev); + void ipoib_pkey_poll(struct work_struct *work); int ipoib_pkey_dev_delay_open(struct net_device *dev); void ipoib_drain_cq(struct net_device *dev); -#ifdef CONFIG_INFINIBAND_IPOIB_CM +void ipoib_set_ethtool_ops(struct net_device *dev); +int ipoib_set_dev_features(struct ipoib_dev_priv *priv, struct ib_device *hca); -#define IPOIB_FLAGS_RC 0x80 -#define IPOIB_FLAGS_UC 0x40 +#define IPOIB_FLAGS_RC 0x80 +#define IPOIB_FLAGS_UC 0x40 /* We don't support UC connections at the moment */ #define IPOIB_CM_SUPPORTED(ha) (ha[0] & (IPOIB_FLAGS_RC)) +#ifdef CONFIG_INFINIBAND_IPOIB_CM + +extern int ipoib_max_conn_qp; + static inline int ipoib_cm_admin_enabled(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); @@ -445,10 +556,10 @@ static inline int ipoib_cm_admin_enabled(struct net_device *dev) test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); } -static inline int ipoib_cm_enabled(struct net_device *dev, struct neighbour *n) +static inline int ipoib_cm_enabled(struct net_device *dev, u8 *hwaddr) { struct ipoib_dev_priv *priv = netdev_priv(dev); - return IPOIB_CM_SUPPORTED(n->ha) && + return IPOIB_CM_SUPPORTED(hwaddr) && test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); } @@ -468,6 +579,18 @@ static inline void ipoib_cm_set(struct ipoib_neigh *neigh, struct ipoib_cm_tx *t neigh->cm = tx; } +static inline int ipoib_cm_has_srq(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + return !!priv->cm.srq; +} + +static inline unsigned int ipoib_cm_max_mtu(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + return priv->cm.max_cm_mtu; +} + void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx); int ipoib_cm_dev_open(struct net_device *dev); void ipoib_cm_dev_stop(struct net_device *dev); @@ -477,18 +600,21 @@ void ipoib_cm_dev_cleanup(struct net_device *dev); struct ipoib_cm_tx *ipoib_cm_create_tx(struct net_device *dev, struct ipoib_path *path, struct ipoib_neigh *neigh); void ipoib_cm_destroy_tx(struct ipoib_cm_tx *tx); -void ipoib_cm_skb_too_long(struct net_device* dev, struct sk_buff *skb, +void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb, unsigned int mtu); void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc); +void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc); #else struct ipoib_cm_tx; +#define ipoib_max_conn_qp 0 + static inline int ipoib_cm_admin_enabled(struct net_device *dev) { return 0; } -static inline int ipoib_cm_enabled(struct net_device *dev, struct neighbour *n) +static inline int ipoib_cm_enabled(struct net_device *dev, u8 *hwaddr) { return 0; @@ -509,6 +635,16 @@ static inline void ipoib_cm_set(struct ipoib_neigh *neigh, struct ipoib_cm_tx *t { } +static inline int ipoib_cm_has_srq(struct net_device *dev) +{ + return 0; +} + +static inline unsigned int ipoib_cm_max_mtu(struct net_device *dev) +{ + return 0; +} + static inline void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx) { @@ -558,7 +694,7 @@ int ipoib_cm_add_mode_attr(struct net_device *dev) return 0; } -static inline void ipoib_cm_skb_too_long(struct net_device* dev, struct sk_buff *skb, +static inline void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb, unsigned int mtu) { dev_kfree_skb_any(skb); @@ -568,6 +704,9 @@ static inline void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *w { } +static inline void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) +{ +} #endif #ifdef CONFIG_INFINIBAND_IPOIB_DEBUG @@ -582,7 +721,6 @@ static inline int ipoib_register_debugfs(void) { return 0; } static inline void ipoib_unregister_debugfs(void) { } #endif - #define ipoib_printk(level, priv, format, arg...) \ printk(level "%s: " format, ((struct ipoib_dev_priv *) priv)->dev->name , ## arg) #define ipoib_warn(priv, format, arg...) \ @@ -597,12 +735,12 @@ extern struct ib_sa_client ipoib_sa_client; extern int ipoib_debug_level; #define ipoib_dbg(priv, format, arg...) \ - do { \ + do { \ if (ipoib_debug_level > 0) \ ipoib_printk(KERN_DEBUG, priv, format , ## arg); \ } while (0) #define ipoib_dbg_mcast(priv, format, arg...) \ - do { \ + do { \ if (mcast_debug_level > 0) \ ipoib_printk(KERN_DEBUG, priv, format , ## arg); \ } while (0) @@ -615,7 +753,7 @@ extern int ipoib_debug_level; #ifdef CONFIG_INFINIBAND_IPOIB_DEBUG_DATA #define ipoib_dbg_data(priv, format, arg...) \ - do { \ + do { \ if (data_debug_level > 0) \ ipoib_printk(KERN_DEBUG, priv, format , ## arg); \ } while (0) @@ -624,29 +762,8 @@ extern int ipoib_debug_level; do { (void) (priv); } while (0) #endif /* CONFIG_INFINIBAND_IPOIB_DEBUG_DATA */ - -#define IPOIB_GID_FMT "%2.2x%2.2x:%2.2x%2.2x:%2.2x%2.2x:%2.2x%2.2x:" \ - "%2.2x%2.2x:%2.2x%2.2x:%2.2x%2.2x:%2.2x%2.2x" - -#define IPOIB_GID_RAW_ARG(gid) ((u8 *)(gid))[0], \ - ((u8 *)(gid))[1], \ - ((u8 *)(gid))[2], \ - ((u8 *)(gid))[3], \ - ((u8 *)(gid))[4], \ - ((u8 *)(gid))[5], \ - ((u8 *)(gid))[6], \ - ((u8 *)(gid))[7], \ - ((u8 *)(gid))[8], \ - ((u8 *)(gid))[9], \ - ((u8 *)(gid))[10],\ - ((u8 *)(gid))[11],\ - ((u8 *)(gid))[12],\ - ((u8 *)(gid))[13],\ - ((u8 *)(gid))[14],\ - ((u8 *)(gid))[15] - -#define IPOIB_GID_ARG(gid) IPOIB_GID_RAW_ARG((gid).raw) - #define IPOIB_QPN(ha) (be32_to_cpup((__be32 *) ha) & 0xffffff) +extern const char ipoib_driver_version[]; + #endif /* _IPOIB_H */ diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c index 08b4676a382..933efcea0d0 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c @@ -28,16 +28,25 @@ * 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. - * - * $Id$ */ #include <rdma/ib_cm.h> -#include <rdma/ib_cache.h> #include <net/dst.h> #include <net/icmp.h> #include <linux/icmpv6.h> #include <linux/delay.h> +#include <linux/slab.h> +#include <linux/vmalloc.h> +#include <linux/moduleparam.h> + +#include "ipoib.h" + +int ipoib_max_conn_qp = 128; + +module_param_named(max_nonsrq_conn_qp, ipoib_max_conn_qp, int, 0444); +MODULE_PARM_DESC(max_nonsrq_conn_qp, + "Max number of connected-mode QPs per interface " + "(applied only if shared receive queue is not available)"); #ifdef CONFIG_INFINIBAND_IPOIB_DEBUG_DATA static int data_debug_level; @@ -47,8 +56,6 @@ MODULE_PARM_DESC(cm_data_debug_level, "Enable data path debug tracing for connected mode if > 0"); #endif -#include "ipoib.h" - #define IPOIB_CM_IETF_ID 0x1000000000000000ULL #define IPOIB_CM_RX_UPDATE_TIME (256 * HZ) @@ -60,7 +67,7 @@ static struct ib_qp_attr ipoib_cm_err_attr = { .qp_state = IB_QPS_ERR }; -#define IPOIB_CM_RX_DRAIN_WRID 0x7fffffff +#define IPOIB_CM_RX_DRAIN_WRID 0xffffffff static struct ib_send_wr ipoib_cm_rx_drain_wr = { .wr_id = IPOIB_CM_RX_DRAIN_WRID, @@ -78,24 +85,24 @@ static void ipoib_cm_dma_unmap_rx(struct ipoib_dev_priv *priv, int frags, ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_CM_HEAD_SIZE, DMA_FROM_DEVICE); for (i = 0; i < frags; ++i) - ib_dma_unmap_single(priv->ca, mapping[i + 1], PAGE_SIZE, DMA_FROM_DEVICE); + ib_dma_unmap_page(priv->ca, mapping[i + 1], PAGE_SIZE, DMA_FROM_DEVICE); } -static int ipoib_cm_post_receive(struct net_device *dev, int id) +static int ipoib_cm_post_receive_srq(struct net_device *dev, int id) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_recv_wr *bad_wr; int i, ret; - priv->cm.rx_wr.wr_id = id | IPOIB_CM_OP_SRQ; + priv->cm.rx_wr.wr_id = id | IPOIB_OP_CM | IPOIB_OP_RECV; - for (i = 0; i < IPOIB_CM_RX_SG; ++i) + for (i = 0; i < priv->cm.num_frags; ++i) priv->cm.rx_sge[i].addr = priv->cm.srq_ring[id].mapping[i]; ret = ib_post_srq_recv(priv->cm.srq, &priv->cm.rx_wr, &bad_wr); if (unlikely(ret)) { ipoib_warn(priv, "post srq failed for buf %d (%d)\n", id, ret); - ipoib_cm_dma_unmap_rx(priv, IPOIB_CM_RX_SG - 1, + ipoib_cm_dma_unmap_rx(priv, priv->cm.num_frags - 1, priv->cm.srq_ring[id].mapping); dev_kfree_skb_any(priv->cm.srq_ring[id].skb); priv->cm.srq_ring[id].skb = NULL; @@ -104,8 +111,37 @@ static int ipoib_cm_post_receive(struct net_device *dev, int id) return ret; } -static struct sk_buff *ipoib_cm_alloc_rx_skb(struct net_device *dev, int id, int frags, - u64 mapping[IPOIB_CM_RX_SG]) +static int ipoib_cm_post_receive_nonsrq(struct net_device *dev, + struct ipoib_cm_rx *rx, + struct ib_recv_wr *wr, + struct ib_sge *sge, int id) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct ib_recv_wr *bad_wr; + int i, ret; + + wr->wr_id = id | IPOIB_OP_CM | IPOIB_OP_RECV; + + for (i = 0; i < IPOIB_CM_RX_SG; ++i) + sge[i].addr = rx->rx_ring[id].mapping[i]; + + ret = ib_post_recv(rx->qp, wr, &bad_wr); + if (unlikely(ret)) { + ipoib_warn(priv, "post recv failed for buf %d (%d)\n", id, ret); + ipoib_cm_dma_unmap_rx(priv, IPOIB_CM_RX_SG - 1, + rx->rx_ring[id].mapping); + dev_kfree_skb_any(rx->rx_ring[id].skb); + rx->rx_ring[id].skb = NULL; + } + + return ret; +} + +static struct sk_buff *ipoib_cm_alloc_rx_skb(struct net_device *dev, + struct ipoib_cm_rx_buf *rx_ring, + int id, int frags, + u64 mapping[IPOIB_CM_RX_SG], + gfp_t gfp) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct sk_buff *skb; @@ -129,19 +165,19 @@ static struct sk_buff *ipoib_cm_alloc_rx_skb(struct net_device *dev, int id, int } for (i = 0; i < frags; i++) { - struct page *page = alloc_page(GFP_ATOMIC); + struct page *page = alloc_page(gfp); if (!page) goto partial_error; skb_fill_page_desc(skb, i, page, 0, PAGE_SIZE); - mapping[i + 1] = ib_dma_map_page(priv->ca, skb_shinfo(skb)->frags[i].page, + mapping[i + 1] = ib_dma_map_page(priv->ca, page, 0, PAGE_SIZE, DMA_FROM_DEVICE); if (unlikely(ib_dma_mapping_error(priv->ca, mapping[i + 1]))) goto partial_error; } - priv->cm.srq_ring[id].skb = skb; + rx_ring[id].skb = skb; return skb; partial_error: @@ -149,13 +185,29 @@ partial_error: ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_CM_HEAD_SIZE, DMA_FROM_DEVICE); for (; i > 0; --i) - ib_dma_unmap_single(priv->ca, mapping[i], PAGE_SIZE, DMA_FROM_DEVICE); + ib_dma_unmap_page(priv->ca, mapping[i], PAGE_SIZE, DMA_FROM_DEVICE); dev_kfree_skb_any(skb); return NULL; } -static void ipoib_cm_start_rx_drain(struct ipoib_dev_priv* priv) +static void ipoib_cm_free_rx_ring(struct net_device *dev, + struct ipoib_cm_rx_buf *rx_ring) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + int i; + + for (i = 0; i < ipoib_recvq_size; ++i) + if (rx_ring[i].skb) { + ipoib_cm_dma_unmap_rx(priv, IPOIB_CM_RX_SG - 1, + rx_ring[i].mapping); + dev_kfree_skb_any(rx_ring[i].skb); + } + + vfree(rx_ring); +} + +static void ipoib_cm_start_rx_drain(struct ipoib_dev_priv *priv) { struct ib_send_wr *bad_wr; struct ipoib_cm_rx *p; @@ -199,8 +251,8 @@ static struct ib_qp *ipoib_cm_create_rx_qp(struct net_device *dev, struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_qp_init_attr attr = { .event_handler = ipoib_cm_rx_event_handler, - .send_cq = priv->cq, /* For drain WR */ - .recv_cq = priv->cq, + .send_cq = priv->recv_cq, /* For drain WR */ + .recv_cq = priv->recv_cq, .srq = priv->cm.srq, .cap.max_send_wr = 1, /* For drain WR */ .cap.max_send_sge = 1, /* FIXME: 0 Seems not to work */ @@ -208,12 +260,18 @@ static struct ib_qp *ipoib_cm_create_rx_qp(struct net_device *dev, .qp_type = IB_QPT_RC, .qp_context = p, }; + + if (!ipoib_cm_has_srq(dev)) { + attr.cap.max_recv_wr = ipoib_recvq_size; + attr.cap.max_recv_sge = IPOIB_CM_RX_SG; + } + return ib_create_qp(priv->pd, &attr); } static int ipoib_cm_modify_rx_qp(struct net_device *dev, - struct ib_cm_id *cm_id, struct ib_qp *qp, - unsigned psn) + struct ib_cm_id *cm_id, struct ib_qp *qp, + unsigned psn) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_qp_attr qp_attr; @@ -266,6 +324,98 @@ static int ipoib_cm_modify_rx_qp(struct net_device *dev, return 0; } +static void ipoib_cm_init_rx_wr(struct net_device *dev, + struct ib_recv_wr *wr, + struct ib_sge *sge) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + int i; + + for (i = 0; i < priv->cm.num_frags; ++i) + sge[i].lkey = priv->mr->lkey; + + sge[0].length = IPOIB_CM_HEAD_SIZE; + for (i = 1; i < priv->cm.num_frags; ++i) + sge[i].length = PAGE_SIZE; + + wr->next = NULL; + wr->sg_list = sge; + wr->num_sge = priv->cm.num_frags; +} + +static int ipoib_cm_nonsrq_init_rx(struct net_device *dev, struct ib_cm_id *cm_id, + struct ipoib_cm_rx *rx) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct { + struct ib_recv_wr wr; + struct ib_sge sge[IPOIB_CM_RX_SG]; + } *t; + int ret; + int i; + + rx->rx_ring = vzalloc(ipoib_recvq_size * sizeof *rx->rx_ring); + if (!rx->rx_ring) { + printk(KERN_WARNING "%s: failed to allocate CM non-SRQ ring (%d entries)\n", + priv->ca->name, ipoib_recvq_size); + return -ENOMEM; + } + + t = kmalloc(sizeof *t, GFP_KERNEL); + if (!t) { + ret = -ENOMEM; + goto err_free; + } + + ipoib_cm_init_rx_wr(dev, &t->wr, t->sge); + + spin_lock_irq(&priv->lock); + + if (priv->cm.nonsrq_conn_qp >= ipoib_max_conn_qp) { + spin_unlock_irq(&priv->lock); + ib_send_cm_rej(cm_id, IB_CM_REJ_NO_QP, NULL, 0, NULL, 0); + ret = -EINVAL; + goto err_free; + } else + ++priv->cm.nonsrq_conn_qp; + + spin_unlock_irq(&priv->lock); + + for (i = 0; i < ipoib_recvq_size; ++i) { + if (!ipoib_cm_alloc_rx_skb(dev, rx->rx_ring, i, IPOIB_CM_RX_SG - 1, + rx->rx_ring[i].mapping, + GFP_KERNEL)) { + ipoib_warn(priv, "failed to allocate receive buffer %d\n", i); + ret = -ENOMEM; + goto err_count; + } + ret = ipoib_cm_post_receive_nonsrq(dev, rx, &t->wr, t->sge, i); + if (ret) { + ipoib_warn(priv, "ipoib_cm_post_receive_nonsrq " + "failed for buf %d\n", i); + ret = -EIO; + goto err_count; + } + } + + rx->recv_count = ipoib_recvq_size; + + kfree(t); + + return 0; + +err_count: + spin_lock_irq(&priv->lock); + --priv->cm.nonsrq_conn_qp; + spin_unlock_irq(&priv->lock); + +err_free: + kfree(t); + ipoib_cm_free_rx_ring(dev, rx->rx_ring); + + return ret; +} + static int ipoib_cm_send_rep(struct net_device *dev, struct ib_cm_id *cm_id, struct ib_qp *qp, struct ib_cm_req_event_param *req, unsigned psn) @@ -281,7 +431,7 @@ static int ipoib_cm_send_rep(struct net_device *dev, struct ib_cm_id *cm_id, rep.private_data_len = sizeof data; rep.flow_control = 0; rep.rnr_retry_count = req->rnr_retry_count; - rep.srq = 1; + rep.srq = ipoib_cm_has_srq(dev); rep.qp_num = qp->qp_num; rep.starting_psn = psn; return ib_send_cm_rep(cm_id, &rep); @@ -312,11 +462,17 @@ static int ipoib_cm_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *even goto err_qp; } - psn = random32() & 0xffffff; + psn = prandom_u32() & 0xffffff; ret = ipoib_cm_modify_rx_qp(dev, cm_id, p->qp, psn); if (ret) goto err_modify; + if (!ipoib_cm_has_srq(dev)) { + ret = ipoib_cm_nonsrq_init_rx(dev, cm_id, p); + if (ret) + goto err_modify; + } + spin_lock_irq(&priv->lock); queue_delayed_work(ipoib_workqueue, &priv->cm.stale_task, IPOIB_CM_RX_DELAY); @@ -384,12 +540,13 @@ static void skb_put_frags(struct sk_buff *skb, unsigned int hdr_space, if (length == 0) { /* don't need this page */ - skb_fill_page_desc(toskb, i, frag->page, 0, PAGE_SIZE); + skb_fill_page_desc(toskb, i, skb_frag_page(frag), + 0, PAGE_SIZE); --skb_shinfo(skb)->nr_frags; } else { size = min(length, (unsigned) PAGE_SIZE); - frag->size = size; + skb_frag_size_set(frag, size); skb->data_len += size; skb->truesize += size; skb->len += size; @@ -401,18 +558,21 @@ static void skb_put_frags(struct sk_buff *skb, unsigned int hdr_space, void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) { struct ipoib_dev_priv *priv = netdev_priv(dev); - unsigned int wr_id = wc->wr_id & ~IPOIB_CM_OP_SRQ; + struct ipoib_cm_rx_buf *rx_ring; + unsigned int wr_id = wc->wr_id & ~(IPOIB_OP_CM | IPOIB_OP_RECV); struct sk_buff *skb, *newskb; struct ipoib_cm_rx *p; unsigned long flags; u64 mapping[IPOIB_CM_RX_SG]; int frags; + int has_srq; + struct sk_buff *small_skb; ipoib_dbg_data(priv, "cm recv completion: id %d, status: %d\n", wr_id, wc->status); if (unlikely(wr_id >= ipoib_recvq_size)) { - if (wr_id == (IPOIB_CM_RX_DRAIN_WRID & ~IPOIB_CM_OP_SRQ)) { + if (wr_id == (IPOIB_CM_RX_DRAIN_WRID & ~(IPOIB_OP_CM | IPOIB_OP_RECV))) { spin_lock_irqsave(&priv->lock, flags); list_splice_init(&priv->cm.rx_drain_list, &priv->cm.rx_reap_list); ipoib_cm_start_rx_drain(priv); @@ -424,18 +584,32 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) return; } - skb = priv->cm.srq_ring[wr_id].skb; + p = wc->qp->qp_context; + + has_srq = ipoib_cm_has_srq(dev); + rx_ring = has_srq ? priv->cm.srq_ring : p->rx_ring; + + skb = rx_ring[wr_id].skb; if (unlikely(wc->status != IB_WC_SUCCESS)) { ipoib_dbg(priv, "cm recv error " "(status=%d, wrid=%d vend_err %x)\n", wc->status, wr_id, wc->vendor_err); - ++priv->stats.rx_dropped; - goto repost; + ++dev->stats.rx_dropped; + if (has_srq) + goto repost; + else { + if (!--p->recv_count) { + spin_lock_irqsave(&priv->lock, flags); + list_move(&p->list, &priv->cm.rx_reap_list); + spin_unlock_irqrestore(&priv->lock, flags); + queue_work(ipoib_workqueue, &priv->cm.rx_reap_task); + } + return; + } } - if (!likely(wr_id & IPOIB_CM_RX_UPDATE_MASK)) { - p = wc->qp->qp_context; + if (unlikely(!(wr_id & IPOIB_CM_RX_UPDATE_MASK))) { if (p && time_after_eq(jiffies, p->jiffies + IPOIB_CM_RX_UPDATE_TIME)) { spin_lock_irqsave(&priv->lock, flags); p->jiffies = jiffies; @@ -447,35 +621,53 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) } } + if (wc->byte_len < IPOIB_CM_COPYBREAK) { + int dlen = wc->byte_len; + + small_skb = dev_alloc_skb(dlen + 12); + if (small_skb) { + skb_reserve(small_skb, 12); + ib_dma_sync_single_for_cpu(priv->ca, rx_ring[wr_id].mapping[0], + dlen, DMA_FROM_DEVICE); + skb_copy_from_linear_data(skb, small_skb->data, dlen); + ib_dma_sync_single_for_device(priv->ca, rx_ring[wr_id].mapping[0], + dlen, DMA_FROM_DEVICE); + skb_put(small_skb, dlen); + skb = small_skb; + goto copied; + } + } + frags = PAGE_ALIGN(wc->byte_len - min(wc->byte_len, (unsigned)IPOIB_CM_HEAD_SIZE)) / PAGE_SIZE; - newskb = ipoib_cm_alloc_rx_skb(dev, wr_id, frags, mapping); + newskb = ipoib_cm_alloc_rx_skb(dev, rx_ring, wr_id, frags, + mapping, GFP_ATOMIC); if (unlikely(!newskb)) { /* * If we can't allocate a new RX buffer, dump * this packet and reuse the old buffer. */ ipoib_dbg(priv, "failed to allocate receive buffer %d\n", wr_id); - ++priv->stats.rx_dropped; + ++dev->stats.rx_dropped; goto repost; } - ipoib_cm_dma_unmap_rx(priv, frags, priv->cm.srq_ring[wr_id].mapping); - memcpy(priv->cm.srq_ring[wr_id].mapping, mapping, (frags + 1) * sizeof *mapping); + ipoib_cm_dma_unmap_rx(priv, frags, rx_ring[wr_id].mapping); + memcpy(rx_ring[wr_id].mapping, mapping, (frags + 1) * sizeof *mapping); ipoib_dbg_data(priv, "received %d bytes, SLID 0x%04x\n", wc->byte_len, wc->slid); skb_put_frags(skb, IPOIB_CM_HEAD_SIZE, wc->byte_len, newskb); +copied: skb->protocol = ((struct ipoib_header *) skb->data)->proto; skb_reset_mac_header(skb); skb_pull(skb, IPOIB_ENCAP_LEN); - dev->last_rx = jiffies; - ++priv->stats.rx_packets; - priv->stats.rx_bytes += skb->len; + ++dev->stats.rx_packets; + dev->stats.rx_bytes += skb->len; skb->dev = dev; /* XXX get correct PACKET_ type here */ @@ -483,9 +675,20 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) netif_receive_skb(skb); repost: - if (unlikely(ipoib_cm_post_receive(dev, wr_id))) - ipoib_warn(priv, "ipoib_cm_post_receive failed " - "for buf %d\n", wr_id); + if (has_srq) { + if (unlikely(ipoib_cm_post_receive_srq(dev, wr_id))) + ipoib_warn(priv, "ipoib_cm_post_receive_srq failed " + "for buf %d\n", wr_id); + } else { + if (unlikely(ipoib_cm_post_receive_nonsrq(dev, p, + &priv->cm.rx_wr, + priv->cm.rx_sge, + wr_id))) { + --p->recv_count; + ipoib_warn(priv, "ipoib_cm_post_receive_nonsrq failed " + "for buf %d\n", wr_id); + } + } } static inline int post_send(struct ipoib_dev_priv *priv, @@ -495,10 +698,11 @@ static inline int post_send(struct ipoib_dev_priv *priv, { struct ib_send_wr *bad_wr; - priv->tx_sge.addr = addr; - priv->tx_sge.length = len; + priv->tx_sge[0].addr = addr; + priv->tx_sge[0].length = len; - priv->tx_wr.wr_id = wr_id; + priv->tx_wr.num_sge = 1; + priv->tx_wr.wr_id = wr_id | IPOIB_OP_CM; return ib_post_send(tx->qp, &priv->tx_wr, &bad_wr); } @@ -506,14 +710,15 @@ static inline int post_send(struct ipoib_dev_priv *priv, 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_tx_buf *tx_req; + struct ipoib_cm_tx_buf *tx_req; u64 addr; + int rc; if (unlikely(skb->len > tx->mtu)) { ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n", skb->len, tx->mtu); - ++priv->stats.tx_dropped; - ++priv->stats.tx_errors; + ++dev->stats.tx_dropped; + ++dev->stats.tx_errors; ipoib_cm_skb_too_long(dev, skb, tx->mtu - IPOIB_ENCAP_LEN); return; } @@ -532,38 +737,47 @@ void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_ 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))) { - ++priv->stats.tx_errors; + ++dev->stats.tx_errors; dev_kfree_skb_any(skb); return; } tx_req->mapping = addr; - if (unlikely(post_send(priv, tx, tx->tx_head & (ipoib_sendq_size - 1), - addr, skb->len))) { - ipoib_warn(priv, "post_send failed\n"); - ++priv->stats.tx_errors; + skb_orphan(skb); + skb_dst_drop(skb); + + rc = post_send(priv, tx, tx->tx_head & (ipoib_sendq_size - 1), + addr, skb->len); + 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); dev_kfree_skb_any(skb); } else { dev->trans_start = jiffies; ++tx->tx_head; - if (tx->tx_head - tx->tx_tail == ipoib_sendq_size) { + if (++priv->tx_outstanding == ipoib_sendq_size) { ipoib_dbg(priv, "TX ring 0x%x full, stopping kernel net queue\n", tx->qp->qp_num); netif_stop_queue(dev); - set_bit(IPOIB_FLAG_NETIF_STOPPED, &tx->flags); + rc = ib_req_notify_cq(priv->send_cq, + IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS); + if (rc < 0) + ipoib_warn(priv, "request notify on send CQ failed\n"); + else if (rc) + ipoib_send_comp_handler(priv->send_cq, dev); } } } -static void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ipoib_cm_tx *tx, - struct ib_wc *wc) +void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) { struct ipoib_dev_priv *priv = netdev_priv(dev); - unsigned int wr_id = wc->wr_id; - struct ipoib_tx_buf *tx_req; + 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; unsigned long flags; ipoib_dbg_data(priv, "cm send completion: id %d, status: %d\n", @@ -580,18 +794,18 @@ static void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ipoib_cm_tx *tx ib_dma_unmap_single(priv->ca, tx_req->mapping, tx_req->skb->len, DMA_TO_DEVICE); /* FIXME: is this right? Shouldn't we only increment on success? */ - ++priv->stats.tx_packets; - priv->stats.tx_bytes += tx_req->skb->len; + ++dev->stats.tx_packets; + dev->stats.tx_bytes += tx_req->skb->len; dev_kfree_skb_any(tx_req->skb); - spin_lock_irqsave(&priv->tx_lock, flags); + netif_tx_lock(dev); + ++tx->tx_tail; - if (unlikely(test_bit(IPOIB_FLAG_NETIF_STOPPED, &tx->flags)) && - tx->tx_head - tx->tx_tail <= ipoib_sendq_size >> 1) { - clear_bit(IPOIB_FLAG_NETIF_STOPPED, &tx->flags); + if (unlikely(--priv->tx_outstanding == ipoib_sendq_size >> 1) && + netif_queue_stopped(dev) && + test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) netif_wake_queue(dev); - } if (wc->status != IB_WC_SUCCESS && wc->status != IB_WC_WR_FLUSH_ERR) { @@ -601,24 +815,16 @@ static void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ipoib_cm_tx *tx "(status=%d, wrid=%d vend_err %x)\n", wc->status, wr_id, wc->vendor_err); - spin_lock(&priv->lock); + spin_lock_irqsave(&priv->lock, flags); neigh = tx->neigh; if (neigh) { neigh->cm = NULL; - list_del(&neigh->list); - if (neigh->ah) - ipoib_put_ah(neigh->ah); - ipoib_neigh_free(dev, neigh); + ipoib_neigh_free(neigh); tx->neigh = NULL; } - /* queue would be re-started anyway when TX is destroyed, - * but it makes sense to do it ASAP here. */ - if (test_and_clear_bit(IPOIB_FLAG_NETIF_STOPPED, &tx->flags)) - netif_wake_queue(dev); - if (test_and_clear_bit(IPOIB_FLAG_INITIALIZED, &tx->flags)) { list_move(&tx->list, &priv->cm.reap_list); queue_work(ipoib_workqueue, &priv->cm.reap_task); @@ -626,23 +832,10 @@ static void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ipoib_cm_tx *tx clear_bit(IPOIB_FLAG_OPER_UP, &tx->flags); - spin_unlock(&priv->lock); + spin_unlock_irqrestore(&priv->lock, flags); } - spin_unlock_irqrestore(&priv->tx_lock, flags); -} - -static void ipoib_cm_tx_completion(struct ib_cq *cq, void *tx_ptr) -{ - struct ipoib_cm_tx *tx = tx_ptr; - int n, i; - - ib_req_notify_cq(cq, IB_CQ_NEXT_COMP); - do { - n = ib_poll_cq(cq, IPOIB_NUM_WC, tx->ibwc); - for (i = 0; i < n; ++i) - ipoib_cm_handle_tx_wc(tx->dev, tx, tx->ibwc + i); - } while (n == IPOIB_NUM_WC); + netif_tx_unlock(dev); } int ipoib_cm_dev_open(struct net_device *dev) @@ -677,12 +870,34 @@ err_cm: return ret; } +static void ipoib_cm_free_rx_reap_list(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + struct ipoib_cm_rx *rx, *n; + LIST_HEAD(list); + + spin_lock_irq(&priv->lock); + list_splice_init(&priv->cm.rx_reap_list, &list); + spin_unlock_irq(&priv->lock); + + list_for_each_entry_safe(rx, n, &list, list) { + ib_destroy_cm_id(rx->id); + ib_destroy_qp(rx->qp); + if (!ipoib_cm_has_srq(dev)) { + ipoib_cm_free_rx_ring(priv->dev, rx->rx_ring); + spin_lock_irq(&priv->lock); + --priv->cm.nonsrq_conn_qp; + spin_unlock_irq(&priv->lock); + } + kfree(rx); + } +} + void ipoib_cm_dev_stop(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); - struct ipoib_cm_rx *p, *n; + struct ipoib_cm_rx *p; unsigned long begin; - LIST_HEAD(list); int ret; if (!IPOIB_CM_SUPPORTED(dev->dev_addr) || !priv->cm.id) @@ -715,9 +930,12 @@ void ipoib_cm_dev_stop(struct net_device *dev) /* * assume the HW is wedged and just free up everything. */ - list_splice_init(&priv->cm.rx_flush_list, &list); - list_splice_init(&priv->cm.rx_error_list, &list); - list_splice_init(&priv->cm.rx_drain_list, &list); + list_splice_init(&priv->cm.rx_flush_list, + &priv->cm.rx_reap_list); + list_splice_init(&priv->cm.rx_error_list, + &priv->cm.rx_reap_list); + list_splice_init(&priv->cm.rx_drain_list, + &priv->cm.rx_reap_list); break; } spin_unlock_irq(&priv->lock); @@ -726,15 +944,9 @@ void ipoib_cm_dev_stop(struct net_device *dev) spin_lock_irq(&priv->lock); } - list_splice_init(&priv->cm.rx_reap_list, &list); - spin_unlock_irq(&priv->lock); - list_for_each_entry_safe(p, n, &list, list) { - ib_destroy_cm_id(p->id); - ib_destroy_qp(p->qp); - kfree(p); - } + ipoib_cm_free_rx_reap_list(dev); cancel_delayed_work(&priv->cm.stale_task); } @@ -807,18 +1019,31 @@ static int ipoib_cm_rep_handler(struct ib_cm_id *cm_id, struct ib_cm_event *even return 0; } -static struct ib_qp *ipoib_cm_create_tx_qp(struct net_device *dev, struct ib_cq *cq) +static struct ib_qp *ipoib_cm_create_tx_qp(struct net_device *dev, struct ipoib_cm_tx *tx) { struct ipoib_dev_priv *priv = netdev_priv(dev); - struct ib_qp_init_attr attr = {}; - attr.recv_cq = priv->cq; - attr.srq = priv->cm.srq; - attr.cap.max_send_wr = ipoib_sendq_size; - attr.cap.max_send_sge = 1; - attr.sq_sig_type = IB_SIGNAL_ALL_WR; - attr.qp_type = IB_QPT_RC; - attr.send_cq = cq; - return ib_create_qp(priv->pd, &attr); + struct ib_qp_init_attr attr = { + .send_cq = priv->recv_cq, + .recv_cq = priv->recv_cq, + .srq = priv->cm.srq, + .cap.max_send_wr = ipoib_sendq_size, + .cap.max_send_sge = 1, + .sq_sig_type = IB_SIGNAL_ALL_WR, + .qp_type = IB_QPT_RC, + .qp_context = tx, + .create_flags = IB_QP_CREATE_USE_GFP_NOIO + }; + + struct ib_qp *tx_qp; + + 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", + priv->ca->name); + attr.create_flags &= ~IB_QP_CREATE_USE_GFP_NOIO; + tx_qp = ib_create_qp(priv->pd, &attr); + } + return tx_qp; } static int ipoib_cm_send_req(struct net_device *dev, @@ -833,28 +1058,28 @@ static int ipoib_cm_send_req(struct net_device *dev, data.qpn = cpu_to_be32(priv->qp->qp_num); data.mtu = cpu_to_be32(IPOIB_CM_BUF_SIZE); - req.primary_path = pathrec; - req.alternate_path = NULL; - req.service_id = cpu_to_be64(IPOIB_CM_IETF_ID | qpn); - req.qp_num = qp->qp_num; - req.qp_type = qp->qp_type; - req.private_data = &data; - req.private_data_len = sizeof data; - req.flow_control = 0; + req.primary_path = pathrec; + req.alternate_path = NULL; + req.service_id = cpu_to_be64(IPOIB_CM_IETF_ID | qpn); + req.qp_num = qp->qp_num; + req.qp_type = qp->qp_type; + req.private_data = &data; + req.private_data_len = sizeof data; + req.flow_control = 0; - req.starting_psn = 0; /* FIXME */ + req.starting_psn = 0; /* FIXME */ /* * Pick some arbitrary defaults here; we could make these * module parameters if anyone cared about setting them. */ - req.responder_resources = 4; - req.remote_cm_response_timeout = 20; - req.local_cm_response_timeout = 20; - req.retry_count = 0; /* RFC draft warns against retries */ - req.rnr_retry_count = 0; /* RFC draft warns against retries */ - req.max_cm_retries = 15; - req.srq = 1; + req.responder_resources = 4; + req.remote_cm_response_timeout = 20; + req.local_cm_response_timeout = 20; + req.retry_count = 0; /* RFC draft warns against retries */ + req.rnr_retry_count = 0; /* RFC draft warns against retries */ + req.max_cm_retries = 15; + req.srq = ipoib_cm_has_srq(dev); return ib_send_cm_req(id, &req); } @@ -864,9 +1089,9 @@ static int ipoib_cm_modify_tx_init(struct net_device *dev, struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_qp_attr qp_attr; int qp_attr_mask, ret; - ret = ib_find_cached_pkey(priv->ca, priv->port, priv->pkey, &qp_attr.pkey_index); + ret = ib_find_pkey(priv->ca, priv->port, priv->pkey, &qp_attr.pkey_index); if (ret) { - ipoib_warn(priv, "pkey 0x%x not in cache: %d\n", priv->pkey, ret); + ipoib_warn(priv, "pkey 0x%x not found: %d\n", priv->pkey, ret); return ret; } @@ -889,29 +1114,16 @@ static int ipoib_cm_tx_init(struct ipoib_cm_tx *p, u32 qpn, struct ipoib_dev_priv *priv = netdev_priv(p->dev); int ret; - p->tx_ring = kzalloc(ipoib_sendq_size * sizeof *p->tx_ring, - GFP_KERNEL); + p->tx_ring = __vmalloc(ipoib_sendq_size * sizeof *p->tx_ring, + GFP_NOIO, PAGE_KERNEL); if (!p->tx_ring) { ipoib_warn(priv, "failed to allocate tx ring\n"); ret = -ENOMEM; goto err_tx; } + memset(p->tx_ring, 0, ipoib_sendq_size * sizeof *p->tx_ring); - p->cq = ib_create_cq(priv->ca, ipoib_cm_tx_completion, NULL, p, - ipoib_sendq_size + 1, 0); - if (IS_ERR(p->cq)) { - ret = PTR_ERR(p->cq); - ipoib_warn(priv, "failed to allocate tx cq: %d\n", ret); - goto err_cq; - } - - ret = ib_req_notify_cq(p->cq, IB_CQ_NEXT_COMP); - if (ret) { - ipoib_warn(priv, "failed to request completion notification: %d\n", ret); - goto err_req_notify; - } - - p->qp = ipoib_cm_create_tx_qp(p->dev, p->cq); + p->qp = ipoib_cm_create_tx_qp(p->dev, p); if (IS_ERR(p->qp)) { ret = PTR_ERR(p->qp); ipoib_warn(priv, "failed to allocate tx qp: %d\n", ret); @@ -937,8 +1149,8 @@ static int ipoib_cm_tx_init(struct ipoib_cm_tx *p, u32 qpn, goto err_send_cm; } - ipoib_dbg(priv, "Request connection 0x%x for gid " IPOIB_GID_FMT " qpn 0x%x\n", - p->qp->qp_num, IPOIB_GID_ARG(pathrec->dgid), qpn); + ipoib_dbg(priv, "Request connection 0x%x for gid %pI6 qpn 0x%x\n", + p->qp->qp_num, pathrec->dgid.raw, qpn); return 0; @@ -948,12 +1160,9 @@ err_modify: err_id: p->id = NULL; ib_destroy_qp(p->qp); -err_req_notify: err_qp: p->qp = NULL; - ib_destroy_cq(p->cq); -err_cq: - p->cq = NULL; + vfree(p->tx_ring); err_tx: return ret; } @@ -961,7 +1170,8 @@ err_tx: static void ipoib_cm_tx_destroy(struct ipoib_cm_tx *p) { struct ipoib_dev_priv *priv = netdev_priv(p->dev); - struct ipoib_tx_buf *tx_req; + struct ipoib_cm_tx_buf *tx_req; + unsigned long begin; ipoib_dbg(priv, "Destroy active connection 0x%x head 0x%x tail 0x%x\n", p->qp ? p->qp->qp_num : 0, p->tx_head, p->tx_tail); @@ -969,27 +1179,40 @@ static void ipoib_cm_tx_destroy(struct ipoib_cm_tx *p) if (p->id) ib_destroy_cm_id(p->id); - if (p->qp) - ib_destroy_qp(p->qp); - - if (p->cq) - ib_destroy_cq(p->cq); - - if (test_bit(IPOIB_FLAG_NETIF_STOPPED, &p->flags)) - netif_wake_queue(p->dev); - if (p->tx_ring) { + /* Wait for all sends to complete */ + begin = jiffies; 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); - dev_kfree_skb_any(tx_req->skb); - ++p->tx_tail; + if (time_after(jiffies, begin + 5 * HZ)) { + ipoib_warn(priv, "timing out; %d sends not completed\n", + p->tx_head - p->tx_tail); + goto timeout; + } + + msleep(1); } + } - kfree(p->tx_ring); +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); + dev_kfree_skb_any(tx_req->skb); + ++p->tx_tail; + netif_tx_lock_bh(p->dev); + if (unlikely(--priv->tx_outstanding == ipoib_sendq_size >> 1) && + netif_queue_stopped(p->dev) && + test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) + netif_wake_queue(p->dev); + netif_tx_unlock_bh(p->dev); } + if (p->qp) + ib_destroy_qp(p->qp); + + vfree(p->tx_ring); kfree(p); } @@ -1000,6 +1223,7 @@ static int ipoib_cm_tx_handler(struct ib_cm_id *cm_id, struct ipoib_dev_priv *priv = netdev_priv(tx->dev); struct net_device *dev = priv->dev; struct ipoib_neigh *neigh; + unsigned long flags; int ret; switch (event->event) { @@ -1018,16 +1242,13 @@ static int ipoib_cm_tx_handler(struct ib_cm_id *cm_id, case IB_CM_REJ_RECEIVED: case IB_CM_TIMEWAIT_EXIT: ipoib_dbg(priv, "CM error %d.\n", event->event); - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); neigh = tx->neigh; if (neigh) { neigh->cm = NULL; - list_del(&neigh->list); - if (neigh->ah) - ipoib_put_ah(neigh->ah); - ipoib_neigh_free(dev, neigh); + ipoib_neigh_free(neigh); tx->neigh = NULL; } @@ -1037,8 +1258,8 @@ static int ipoib_cm_tx_handler(struct ib_cm_id *cm_id, queue_work(ipoib_workqueue, &priv->cm.reap_task); } - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); break; default: break; @@ -1070,12 +1291,15 @@ struct ipoib_cm_tx *ipoib_cm_create_tx(struct net_device *dev, struct ipoib_path void ipoib_cm_destroy_tx(struct ipoib_cm_tx *tx) { struct ipoib_dev_priv *priv = netdev_priv(tx->dev); + unsigned long flags; if (test_and_clear_bit(IPOIB_FLAG_INITIALIZED, &tx->flags)) { + spin_lock_irqsave(&priv->lock, flags); list_move(&tx->list, &priv->cm.reap_list); queue_work(ipoib_workqueue, &priv->cm.reap_task); - ipoib_dbg(priv, "Reap connection for gid " IPOIB_GID_FMT "\n", - IPOIB_GID_ARG(tx->neigh->dgid)); + ipoib_dbg(priv, "Reap connection for gid %pI6\n", + tx->neigh->daddr + 4); tx->neigh = NULL; + spin_unlock_irqrestore(&priv->lock, flags); } } @@ -1092,92 +1316,104 @@ static void ipoib_cm_tx_start(struct work_struct *work) struct ib_sa_path_rec pathrec; u32 qpn; - spin_lock_irqsave(&priv->tx_lock, flags); - spin_lock(&priv->lock); + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); + while (!list_empty(&priv->cm.start_list)) { p = list_entry(priv->cm.start_list.next, typeof(*p), list); list_del_init(&p->list); neigh = p->neigh; - qpn = IPOIB_QPN(neigh->neighbour->ha); + qpn = IPOIB_QPN(neigh->daddr); memcpy(&pathrec, &p->path->pathrec, sizeof pathrec); - spin_unlock(&priv->lock); - spin_unlock_irqrestore(&priv->tx_lock, flags); + + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); + ret = ipoib_cm_tx_init(p, qpn, &pathrec); - spin_lock_irqsave(&priv->tx_lock, flags); - spin_lock(&priv->lock); + + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); + if (ret) { neigh = p->neigh; if (neigh) { neigh->cm = NULL; - list_del(&neigh->list); - if (neigh->ah) - ipoib_put_ah(neigh->ah); - ipoib_neigh_free(dev, neigh); + ipoib_neigh_free(neigh); } list_del(&p->list); kfree(p); } } - spin_unlock(&priv->lock); - spin_unlock_irqrestore(&priv->tx_lock, flags); + + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); } static void ipoib_cm_tx_reap(struct work_struct *work) { struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, cm.reap_task); + struct net_device *dev = priv->dev; struct ipoib_cm_tx *p; + unsigned long flags; + + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); while (!list_empty(&priv->cm.reap_list)) { p = list_entry(priv->cm.reap_list.next, typeof(*p), list); list_del(&p->list); - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); ipoib_cm_tx_destroy(p); - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); } - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); } static void ipoib_cm_skb_reap(struct work_struct *work) { struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, cm.skb_task); + struct net_device *dev = priv->dev; struct sk_buff *skb; - + unsigned long flags; unsigned mtu = priv->mcast_mtu; - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); + while ((skb = skb_dequeue(&priv->cm.skb_queue))) { - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); + if (skb->protocol == htons(ETH_P_IP)) icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(mtu)); -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) +#if IS_ENABLED(CONFIG_IPV6) else if (skb->protocol == htons(ETH_P_IPV6)) - icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu, priv->dev); + icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu); #endif dev_kfree_skb_any(skb); - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); + + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); } - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); } -void ipoib_cm_skb_too_long(struct net_device* dev, struct sk_buff *skb, +void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb, unsigned int mtu) { 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), NULL, skb, mtu); skb_queue_tail(&priv->cm.skb_queue, skb); if (e) @@ -1186,20 +1422,8 @@ void ipoib_cm_skb_too_long(struct net_device* dev, struct sk_buff *skb, static void ipoib_cm_rx_reap(struct work_struct *work) { - struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, - cm.rx_reap_task); - struct ipoib_cm_rx *p, *n; - LIST_HEAD(list); - - spin_lock_irq(&priv->lock); - list_splice_init(&priv->cm.rx_reap_list, &list); - spin_unlock_irq(&priv->lock); - - list_for_each_entry_safe(p, n, &list, list) { - ib_destroy_cm_id(p->id); - ib_destroy_qp(p->qp); - kfree(p); - } + ipoib_cm_free_rx_reap_list(container_of(work, struct ipoib_dev_priv, + cm.rx_reap_task)->dev); } static void ipoib_cm_stale_task(struct work_struct *work) @@ -1232,7 +1456,7 @@ static void ipoib_cm_stale_task(struct work_struct *work) } -static ssize_t show_mode(struct device *d, struct device_attribute *attr, +static ssize_t show_mode(struct device *d, struct device_attribute *attr, char *buf) { struct ipoib_dev_priv *priv = netdev_priv(to_net_dev(d)); @@ -1247,25 +1471,19 @@ static ssize_t set_mode(struct device *d, struct device_attribute *attr, const char *buf, size_t count) { struct net_device *dev = to_net_dev(d); - struct ipoib_dev_priv *priv = netdev_priv(dev); + int ret; - /* 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"); - ipoib_flush_paths(dev); - return count; - } + if (!rtnl_trylock()) + return restart_syscall(); + + ret = ipoib_set_mode(dev, buf); - if (!strcmp(buf, "datagram\n")) { - clear_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); - dev->mtu = min(priv->mcast_mtu, dev->mtu); - ipoib_flush_paths(dev); + rtnl_unlock(); + + if (!ret) return count; - } - return -EINVAL; + return ret; } static DEVICE_ATTR(mode, S_IWUSR | S_IRUGO, show_mode, set_mode); @@ -1275,16 +1493,42 @@ int ipoib_cm_add_mode_attr(struct net_device *dev) return device_create_file(&dev->dev, &dev_attr_mode); } -int ipoib_cm_dev_init(struct net_device *dev) +static void ipoib_cm_create_srq(struct net_device *dev, int max_sge) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_srq_init_attr srq_init_attr = { + .srq_type = IB_SRQT_BASIC, .attr = { .max_wr = ipoib_recvq_size, - .max_sge = IPOIB_CM_RX_SG + .max_sge = max_sge } }; - int ret, i; + + priv->cm.srq = ib_create_srq(priv->pd, &srq_init_attr); + if (IS_ERR(priv->cm.srq)) { + if (PTR_ERR(priv->cm.srq) != -ENOSYS) + printk(KERN_WARNING "%s: failed to allocate SRQ, error %ld\n", + priv->ca->name, PTR_ERR(priv->cm.srq)); + priv->cm.srq = NULL; + return; + } + + priv->cm.srq_ring = vzalloc(ipoib_recvq_size * sizeof *priv->cm.srq_ring); + if (!priv->cm.srq_ring) { + printk(KERN_WARNING "%s: failed to allocate CM SRQ ring (%d entries)\n", + priv->ca->name, ipoib_recvq_size); + ib_destroy_srq(priv->cm.srq); + priv->cm.srq = NULL; + return; + } + +} + +int ipoib_cm_dev_init(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + int i, ret; + struct ib_device_attr attr; INIT_LIST_HEAD(&priv->cm.passive_ids); INIT_LIST_HEAD(&priv->cm.reap_list); @@ -1301,43 +1545,46 @@ int ipoib_cm_dev_init(struct net_device *dev) skb_queue_head_init(&priv->cm.skb_queue); - priv->cm.srq = ib_create_srq(priv->pd, &srq_init_attr); - if (IS_ERR(priv->cm.srq)) { - ret = PTR_ERR(priv->cm.srq); - priv->cm.srq = NULL; + ret = ib_query_device(priv->ca, &attr); + if (ret) { + printk(KERN_WARNING "ib_query_device() failed with %d\n", ret); return ret; } - priv->cm.srq_ring = kzalloc(ipoib_recvq_size * sizeof *priv->cm.srq_ring, - GFP_KERNEL); - if (!priv->cm.srq_ring) { - printk(KERN_WARNING "%s: failed to allocate CM ring (%d entries)\n", - priv->ca->name, ipoib_recvq_size); - ipoib_cm_dev_cleanup(dev); - return -ENOMEM; - } + ipoib_dbg(priv, "max_srq_sge=%d\n", attr.max_srq_sge); - for (i = 0; i < IPOIB_CM_RX_SG; ++i) - priv->cm.rx_sge[i].lkey = priv->mr->lkey; + attr.max_srq_sge = min_t(int, IPOIB_CM_RX_SG, attr.max_srq_sge); + ipoib_cm_create_srq(dev, attr.max_srq_sge); + if (ipoib_cm_has_srq(dev)) { + priv->cm.max_cm_mtu = attr.max_srq_sge * PAGE_SIZE - 0x10; + priv->cm.num_frags = attr.max_srq_sge; + ipoib_dbg(priv, "max_cm_mtu = 0x%x, num_frags=%d\n", + priv->cm.max_cm_mtu, priv->cm.num_frags); + } else { + priv->cm.max_cm_mtu = IPOIB_CM_MTU; + priv->cm.num_frags = IPOIB_CM_RX_SG; + } - priv->cm.rx_sge[0].length = IPOIB_CM_HEAD_SIZE; - for (i = 1; i < IPOIB_CM_RX_SG; ++i) - priv->cm.rx_sge[i].length = PAGE_SIZE; - priv->cm.rx_wr.next = NULL; - priv->cm.rx_wr.sg_list = priv->cm.rx_sge; - priv->cm.rx_wr.num_sge = IPOIB_CM_RX_SG; + ipoib_cm_init_rx_wr(dev, &priv->cm.rx_wr, priv->cm.rx_sge); + + if (ipoib_cm_has_srq(dev)) { + for (i = 0; i < ipoib_recvq_size; ++i) { + if (!ipoib_cm_alloc_rx_skb(dev, priv->cm.srq_ring, i, + priv->cm.num_frags - 1, + priv->cm.srq_ring[i].mapping, + GFP_KERNEL)) { + ipoib_warn(priv, "failed to allocate " + "receive buffer %d\n", i); + ipoib_cm_dev_cleanup(dev); + return -ENOMEM; + } - for (i = 0; i < ipoib_recvq_size; ++i) { - if (!ipoib_cm_alloc_rx_skb(dev, i, IPOIB_CM_RX_SG - 1, - priv->cm.srq_ring[i].mapping)) { - ipoib_warn(priv, "failed to allocate receive buffer %d\n", i); - ipoib_cm_dev_cleanup(dev); - return -ENOMEM; - } - if (ipoib_cm_post_receive(dev, i)) { - ipoib_warn(priv, "ipoib_ib_post_receive failed for buf %d\n", i); - ipoib_cm_dev_cleanup(dev); - return -EIO; + if (ipoib_cm_post_receive_srq(dev, i)) { + ipoib_warn(priv, "ipoib_cm_post_receive_srq " + "failed for buf %d\n", i); + ipoib_cm_dev_cleanup(dev); + return -EIO; + } } } @@ -1348,7 +1595,7 @@ int ipoib_cm_dev_init(struct net_device *dev) void ipoib_cm_dev_cleanup(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); - int i, ret; + int ret; if (!priv->cm.srq) return; @@ -1362,13 +1609,7 @@ void ipoib_cm_dev_cleanup(struct net_device *dev) priv->cm.srq = NULL; if (!priv->cm.srq_ring) return; - for (i = 0; i < ipoib_recvq_size; ++i) - if (priv->cm.srq_ring[i].skb) { - ipoib_cm_dma_unmap_rx(priv, IPOIB_CM_RX_SG - 1, - priv->cm.srq_ring[i].mapping); - dev_kfree_skb_any(priv->cm.srq_ring[i].skb); - priv->cm.srq_ring[i].skb = NULL; - } - kfree(priv->cm.srq_ring); + + ipoib_cm_free_rx_ring(dev, priv->cm.srq_ring); priv->cm.srq_ring = NULL; } diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c new file mode 100644 index 00000000000..078cadd6c79 --- /dev/null +++ b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c @@ -0,0 +1,109 @@ +/* + * Copyright (c) 2007 Mellanox Technologies. 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 + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * 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. + */ + +#include <linux/kernel.h> +#include <linux/ethtool.h> +#include <linux/netdevice.h> + +#include "ipoib.h" + +static void ipoib_get_drvinfo(struct net_device *netdev, + struct ethtool_drvinfo *drvinfo) +{ + struct ipoib_dev_priv *priv = netdev_priv(netdev); + struct ib_device_attr *attr; + + attr = kmalloc(sizeof(*attr), GFP_KERNEL); + if (attr && !ib_query_device(priv->ca, attr)) + snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version), + "%d.%d.%d", (int)(attr->fw_ver >> 32), + (int)(attr->fw_ver >> 16) & 0xffff, + (int)attr->fw_ver & 0xffff); + kfree(attr); + + strlcpy(drvinfo->bus_info, dev_name(priv->ca->dma_device), + sizeof(drvinfo->bus_info)); + + strlcpy(drvinfo->version, ipoib_driver_version, + sizeof(drvinfo->version)); + + strlcpy(drvinfo->driver, "ib_ipoib", sizeof(drvinfo->driver)); +} + +static int ipoib_get_coalesce(struct net_device *dev, + struct ethtool_coalesce *coal) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + + coal->rx_coalesce_usecs = priv->ethtool.coalesce_usecs; + coal->rx_max_coalesced_frames = priv->ethtool.max_coalesced_frames; + + return 0; +} + +static int ipoib_set_coalesce(struct net_device *dev, + struct ethtool_coalesce *coal) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + int ret; + + /* + * These values are saved in the private data and returned + * when ipoib_get_coalesce() is called + */ + if (coal->rx_coalesce_usecs > 0xffff || + coal->rx_max_coalesced_frames > 0xffff) + return -EINVAL; + + ret = ib_modify_cq(priv->recv_cq, coal->rx_max_coalesced_frames, + coal->rx_coalesce_usecs); + if (ret && ret != -ENOSYS) { + ipoib_warn(priv, "failed modifying CQ (%d)\n", ret); + return ret; + } + + priv->ethtool.coalesce_usecs = coal->rx_coalesce_usecs; + priv->ethtool.max_coalesced_frames = coal->rx_max_coalesced_frames; + + return 0; +} + +static const struct ethtool_ops ipoib_ethtool_ops = { + .get_drvinfo = ipoib_get_drvinfo, + .get_coalesce = ipoib_get_coalesce, + .set_coalesce = ipoib_set_coalesce, +}; + +void ipoib_set_ethtool_ops(struct net_device *dev) +{ + dev->ethtool_ops = &ipoib_ethtool_ops; +} diff --git a/drivers/infiniband/ulp/ipoib/ipoib_fs.c b/drivers/infiniband/ulp/ipoib/ipoib_fs.c index 44c174182a8..50061854616 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_fs.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_fs.c @@ -28,16 +28,16 @@ * 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_fs.c 1389 2004-12-27 22:56:47Z roland $ */ #include <linux/err.h> #include <linux/seq_file.h> +#include <linux/slab.h> struct file_operations; #include <linux/debugfs.h> +#include <linux/export.h> #include "ipoib.h" @@ -124,7 +124,7 @@ static int ipoib_mcg_seq_show(struct seq_file *file, void *iter_ptr) return 0; } -static struct seq_operations ipoib_mcg_seq_ops = { +static const struct seq_operations ipoib_mcg_seq_ops = { .start = ipoib_mcg_seq_start, .next = ipoib_mcg_seq_next, .stop = ipoib_mcg_seq_stop, @@ -213,16 +213,15 @@ static int ipoib_path_seq_show(struct seq_file *file, void *iter_ptr) gid_buf, path.pathrec.dlid ? "yes" : "no"); if (path.pathrec.dlid) { - rate = ib_rate_to_mult(path.pathrec.rate) * 25; + rate = ib_rate_to_mbps(path.pathrec.rate); seq_printf(file, " DLID: 0x%04x\n" " SL: %12d\n" - " rate: %*d%s Gb/sec\n", + " rate: %8d.%d Gb/sec\n", be16_to_cpu(path.pathrec.dlid), path.pathrec.sl, - 10 - ((rate % 10) ? 2 : 0), - rate / 10, rate % 10 ? ".5" : ""); + rate / 1000, rate % 1000); } seq_putc(file, '\n'); @@ -230,7 +229,7 @@ static int ipoib_path_seq_show(struct seq_file *file, void *iter_ptr) return 0; } -static struct seq_operations ipoib_path_seq_ops = { +static const struct seq_operations ipoib_path_seq_ops = { .start = ipoib_path_seq_start, .next = ipoib_path_seq_next, .stop = ipoib_path_seq_stop, diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c index 10944888cff..6a7003ddb0b 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c @@ -31,14 +31,15 @@ * 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_ib.c 1386 2004-12-27 16:23:17Z roland $ */ #include <linux/delay.h> +#include <linux/moduleparam.h> #include <linux/dma-mapping.h> +#include <linux/slab.h> -#include <rdma/ib_cache.h> +#include <linux/ip.h> +#include <linux/tcp.h> #include "ipoib.h" @@ -56,21 +57,24 @@ struct ipoib_ah *ipoib_create_ah(struct net_device *dev, struct ib_pd *pd, struct ib_ah_attr *attr) { struct ipoib_ah *ah; + struct ib_ah *vah; ah = kmalloc(sizeof *ah, GFP_KERNEL); if (!ah) - return NULL; + return ERR_PTR(-ENOMEM); ah->dev = dev; ah->last_send = 0; kref_init(&ah->ref); - ah->ah = ib_create_ah(pd, attr); - if (IS_ERR(ah->ah)) { + vah = ib_create_ah(pd, attr); + if (IS_ERR(vah)) { kfree(ah); - ah = NULL; - } else + ah = (struct ipoib_ah *)vah; + } else { + ah->ah = vah; ipoib_dbg(netdev_priv(dev), "Created ah %p\n", ah->ah); + } return ah; } @@ -87,28 +91,59 @@ void ipoib_free_ah(struct kref *kref) spin_unlock_irqrestore(&priv->lock, flags); } +static void ipoib_ud_dma_unmap_rx(struct ipoib_dev_priv *priv, + u64 mapping[IPOIB_UD_RX_SG]) +{ + if (ipoib_ud_need_sg(priv->max_ib_mtu)) { + ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_UD_HEAD_SIZE, + DMA_FROM_DEVICE); + ib_dma_unmap_page(priv->ca, mapping[1], PAGE_SIZE, + DMA_FROM_DEVICE); + } else + ib_dma_unmap_single(priv->ca, mapping[0], + IPOIB_UD_BUF_SIZE(priv->max_ib_mtu), + DMA_FROM_DEVICE); +} + +static void ipoib_ud_skb_put_frags(struct ipoib_dev_priv *priv, + struct sk_buff *skb, + unsigned int length) +{ + if (ipoib_ud_need_sg(priv->max_ib_mtu)) { + skb_frag_t *frag = &skb_shinfo(skb)->frags[0]; + unsigned int size; + /* + * There is only two buffers needed for max_payload = 4K, + * first buf size is IPOIB_UD_HEAD_SIZE + */ + skb->tail += IPOIB_UD_HEAD_SIZE; + skb->len += length; + + size = length - IPOIB_UD_HEAD_SIZE; + + skb_frag_size_set(frag, size); + skb->data_len += size; + skb->truesize += PAGE_SIZE; + } else + skb_put(skb, length); + +} + static int ipoib_ib_post_receive(struct net_device *dev, int id) { struct ipoib_dev_priv *priv = netdev_priv(dev); - struct ib_sge list; - struct ib_recv_wr param; struct ib_recv_wr *bad_wr; int ret; - list.addr = priv->rx_ring[id].mapping; - list.length = IPOIB_BUF_SIZE; - list.lkey = priv->mr->lkey; + priv->rx_wr.wr_id = id | IPOIB_OP_RECV; + priv->rx_sge[0].addr = priv->rx_ring[id].mapping[0]; + priv->rx_sge[1].addr = priv->rx_ring[id].mapping[1]; - param.next = NULL; - param.wr_id = id | IPOIB_OP_RECV; - param.sg_list = &list; - param.num_sge = 1; - ret = ib_post_recv(priv->qp, ¶m, &bad_wr); + ret = ib_post_recv(priv->qp, &priv->rx_wr, &bad_wr); if (unlikely(ret)) { ipoib_warn(priv, "receive failed for buf %d (%d)\n", id, ret); - ib_dma_unmap_single(priv->ca, priv->rx_ring[id].mapping, - IPOIB_BUF_SIZE, DMA_FROM_DEVICE); + ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[id].mapping); dev_kfree_skb_any(priv->rx_ring[id].skb); priv->rx_ring[id].skb = NULL; } @@ -116,15 +151,25 @@ static int ipoib_ib_post_receive(struct net_device *dev, int id) return ret; } -static int ipoib_alloc_rx_skb(struct net_device *dev, int id) +static struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev, int id) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct sk_buff *skb; - u64 addr; + int buf_size; + int tailroom; + u64 *mapping; - skb = dev_alloc_skb(IPOIB_BUF_SIZE + 4); - if (!skb) - return -ENOMEM; + if (ipoib_ud_need_sg(priv->max_ib_mtu)) { + buf_size = IPOIB_UD_HEAD_SIZE; + tailroom = 128; /* reserve some tailroom for IP/TCP headers */ + } else { + buf_size = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu); + tailroom = 0; + } + + skb = dev_alloc_skb(buf_size + tailroom + 4); + if (unlikely(!skb)) + return NULL; /* * IB will leave a 40 byte gap for a GRH and IPoIB adds a 4 byte @@ -133,17 +178,32 @@ static int ipoib_alloc_rx_skb(struct net_device *dev, int id) */ skb_reserve(skb, 4); - addr = ib_dma_map_single(priv->ca, skb->data, IPOIB_BUF_SIZE, - DMA_FROM_DEVICE); - if (unlikely(ib_dma_mapping_error(priv->ca, addr))) { - dev_kfree_skb_any(skb); - return -EIO; + mapping = priv->rx_ring[id].mapping; + mapping[0] = ib_dma_map_single(priv->ca, skb->data, buf_size, + DMA_FROM_DEVICE); + if (unlikely(ib_dma_mapping_error(priv->ca, mapping[0]))) + goto error; + + if (ipoib_ud_need_sg(priv->max_ib_mtu)) { + struct page *page = alloc_page(GFP_ATOMIC); + if (!page) + goto partial_error; + skb_fill_page_desc(skb, 0, page, 0, PAGE_SIZE); + mapping[1] = + ib_dma_map_page(priv->ca, page, + 0, PAGE_SIZE, DMA_FROM_DEVICE); + if (unlikely(ib_dma_mapping_error(priv->ca, mapping[1]))) + goto partial_error; } - priv->rx_ring[id].skb = skb; - priv->rx_ring[id].mapping = addr; + priv->rx_ring[id].skb = skb; + return skb; - return 0; +partial_error: + ib_dma_unmap_single(priv->ca, mapping[0], buf_size, DMA_FROM_DEVICE); +error: + dev_kfree_skb_any(skb); + return NULL; } static int ipoib_ib_post_receives(struct net_device *dev) @@ -152,7 +212,7 @@ static int ipoib_ib_post_receives(struct net_device *dev) int i; for (i = 0; i < ipoib_recvq_size; ++i) { - if (ipoib_alloc_rx_skb(dev, i)) { + if (!ipoib_alloc_rx_skb(dev, i)) { ipoib_warn(priv, "failed to allocate receive buffer %d\n", i); return -ENOMEM; } @@ -170,7 +230,8 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) struct ipoib_dev_priv *priv = netdev_priv(dev); unsigned int wr_id = wc->wr_id & ~IPOIB_OP_RECV; struct sk_buff *skb; - u64 addr; + u64 mapping[IPOIB_UD_RX_SG]; + union ib_gid *dgid; ipoib_dbg_data(priv, "recv completion: id %d, status: %d\n", wr_id, wc->status); @@ -182,15 +243,13 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) } skb = priv->rx_ring[wr_id].skb; - addr = priv->rx_ring[wr_id].mapping; if (unlikely(wc->status != IB_WC_SUCCESS)) { if (wc->status != IB_WC_WR_FLUSH_ERR) ipoib_warn(priv, "failed recv event " "(status=%d, wrid=%d vend_err %x)\n", wc->status, wr_id, wc->vendor_err); - ib_dma_unmap_single(priv->ca, addr, - IPOIB_BUF_SIZE, DMA_FROM_DEVICE); + ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[wr_id].mapping); dev_kfree_skb_any(skb); priv->rx_ring[wr_id].skb = NULL; return; @@ -203,35 +262,49 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) if (wc->slid == priv->local_lid && wc->src_qp == priv->qp->qp_num) goto repost; + memcpy(mapping, priv->rx_ring[wr_id].mapping, + IPOIB_UD_RX_SG * sizeof *mapping); + /* * If we can't allocate a new RX buffer, dump * this packet and reuse the old buffer. */ - if (unlikely(ipoib_alloc_rx_skb(dev, wr_id))) { - ++priv->stats.rx_dropped; + if (unlikely(!ipoib_alloc_rx_skb(dev, wr_id))) { + ++dev->stats.rx_dropped; goto repost; } ipoib_dbg_data(priv, "received %d bytes, SLID 0x%04x\n", wc->byte_len, wc->slid); - ib_dma_unmap_single(priv->ca, addr, IPOIB_BUF_SIZE, DMA_FROM_DEVICE); + ipoib_ud_dma_unmap_rx(priv, mapping); + ipoib_ud_skb_put_frags(priv, skb, wc->byte_len); + + /* First byte of dgid signals multicast when 0xff */ + dgid = &((struct ib_grh *)skb->data)->dgid; + + if (!(wc->wc_flags & IB_WC_GRH) || dgid->raw[0] != 0xff) + skb->pkt_type = PACKET_HOST; + else if (memcmp(dgid, dev->broadcast + 4, sizeof(union ib_gid)) == 0) + skb->pkt_type = PACKET_BROADCAST; + else + skb->pkt_type = PACKET_MULTICAST; - skb_put(skb, wc->byte_len); skb_pull(skb, IB_GRH_BYTES); skb->protocol = ((struct ipoib_header *) skb->data)->proto; skb_reset_mac_header(skb); skb_pull(skb, IPOIB_ENCAP_LEN); - dev->last_rx = jiffies; - ++priv->stats.rx_packets; - priv->stats.rx_bytes += skb->len; + ++dev->stats.rx_packets; + dev->stats.rx_bytes += skb->len; skb->dev = dev; - /* XXX get correct PACKET_ type here */ - skb->pkt_type = PACKET_HOST; - netif_receive_skb(skb); + if ((dev->features & NETIF_F_RXCSUM) && + likely(wc->wc_flags & IB_WC_IP_CSUM_OK)) + skb->ip_summed = CHECKSUM_UNNECESSARY; + + napi_gro_receive(&priv->napi, skb); repost: if (unlikely(ipoib_ib_post_receive(dev, wr_id))) @@ -239,12 +312,75 @@ repost: "for buf %d\n", wr_id); } +static 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; + int i; + int off; + + if (skb_headlen(skb)) { + mapping[0] = ib_dma_map_single(ca, skb->data, skb_headlen(skb), + DMA_TO_DEVICE); + if (unlikely(ib_dma_mapping_error(ca, mapping[0]))) + return -EIO; + + off = 1; + } else + off = 0; + + for (i = 0; i < skb_shinfo(skb)->nr_frags; ++i) { + const skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; + mapping[i + off] = ib_dma_map_page(ca, + skb_frag_page(frag), + frag->page_offset, skb_frag_size(frag), + DMA_TO_DEVICE); + if (unlikely(ib_dma_mapping_error(ca, mapping[i + off]))) + goto partial_error; + } + return 0; + +partial_error: + for (; i > 0; --i) { + const skb_frag_t *frag = &skb_shinfo(skb)->frags[i - 1]; + + ib_dma_unmap_page(ca, mapping[i - !off], skb_frag_size(frag), DMA_TO_DEVICE); + } + + if (off) + ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE); + + return -EIO; +} + +static void ipoib_dma_unmap_tx(struct ib_device *ca, + struct ipoib_tx_buf *tx_req) +{ + struct sk_buff *skb = tx_req->skb; + u64 *mapping = tx_req->mapping; + int i; + int off; + + if (skb_headlen(skb)) { + ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE); + off = 1; + } else + off = 0; + + 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); + } +} + static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) { struct ipoib_dev_priv *priv = netdev_priv(dev); unsigned int wr_id = wc->wr_id; struct ipoib_tx_buf *tx_req; - unsigned long flags; ipoib_dbg_data(priv, "send completion: id %d, status: %d\n", wr_id, wc->status); @@ -257,22 +393,18 @@ static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) tx_req = &priv->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->ca, tx_req); - ++priv->stats.tx_packets; - priv->stats.tx_bytes += tx_req->skb->len; + ++dev->stats.tx_packets; + dev->stats.tx_bytes += tx_req->skb->len; dev_kfree_skb_any(tx_req->skb); - spin_lock_irqsave(&priv->tx_lock, flags); ++priv->tx_tail; - if (unlikely(test_bit(IPOIB_FLAG_NETIF_STOPPED, &priv->flags)) && - priv->tx_head - priv->tx_tail <= ipoib_sendq_size >> 1) { - clear_bit(IPOIB_FLAG_NETIF_STOPPED, &priv->flags); + if (unlikely(--priv->tx_outstanding == ipoib_sendq_size >> 1) && + netif_queue_stopped(dev) && + test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) netif_wake_queue(dev); - } - spin_unlock_irqrestore(&priv->tx_lock, flags); if (wc->status != IB_WC_SUCCESS && wc->status != IB_WC_WR_FLUSH_ERR) @@ -281,78 +413,128 @@ static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) wc->status, wr_id, wc->vendor_err); } -int ipoib_poll(struct net_device *dev, int *budget) +static int poll_tx(struct ipoib_dev_priv *priv) { - struct ipoib_dev_priv *priv = netdev_priv(dev); - int max = min(*budget, dev->quota); + int n, i; + + n = ib_poll_cq(priv->send_cq, MAX_SEND_CQE, priv->send_wc); + for (i = 0; i < n; ++i) + ipoib_ib_handle_tx_wc(priv->dev, priv->send_wc + i); + + return n == MAX_SEND_CQE; +} + +int ipoib_poll(struct napi_struct *napi, int budget) +{ + struct ipoib_dev_priv *priv = container_of(napi, struct ipoib_dev_priv, napi); + struct net_device *dev = priv->dev; int done; int t; - int empty; int n, i; done = 0; - empty = 0; - while (max) { +poll_more: + while (done < budget) { + int max = (budget - done); + t = min(IPOIB_NUM_WC, max); - n = ib_poll_cq(priv->cq, t, priv->ibwc); + n = ib_poll_cq(priv->recv_cq, t, priv->ibwc); - for (i = 0; i < n; ++i) { + for (i = 0; i < n; i++) { struct ib_wc *wc = priv->ibwc + i; - if (wc->wr_id & IPOIB_CM_OP_SRQ) { - ++done; - --max; - ipoib_cm_handle_rx_wc(dev, wc); - } else if (wc->wr_id & IPOIB_OP_RECV) { + if (wc->wr_id & IPOIB_OP_RECV) { ++done; - --max; - ipoib_ib_handle_rx_wc(dev, wc); + if (wc->wr_id & IPOIB_OP_CM) + ipoib_cm_handle_rx_wc(dev, wc); + else + ipoib_ib_handle_rx_wc(dev, wc); } else - ipoib_ib_handle_tx_wc(dev, wc); + ipoib_cm_handle_tx_wc(priv->dev, wc); } - if (n != t) { - empty = 1; + if (n != t) break; - } } - dev->quota -= done; - *budget -= done; - - if (empty) { - netif_rx_complete(dev); - if (unlikely(ib_req_notify_cq(priv->cq, + if (done < budget) { + napi_complete(napi); + if (unlikely(ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS)) && - netif_rx_reschedule(dev, 0)) - return 1; - - return 0; + napi_reschedule(napi)) + goto poll_more; } - return 1; + return done; } void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr) { - netif_rx_schedule(dev_ptr); + struct net_device *dev = dev_ptr; + struct ipoib_dev_priv *priv = netdev_priv(dev); + + napi_schedule(&priv->napi); +} + +static void drain_tx_cq(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + + netif_tx_lock(dev); + while (poll_tx(priv)) + ; /* nothing */ + + if (netif_queue_stopped(dev)) + mod_timer(&priv->poll_timer, jiffies + 1); + + netif_tx_unlock(dev); +} + +void ipoib_send_comp_handler(struct ib_cq *cq, void *dev_ptr) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev_ptr); + + mod_timer(&priv->poll_timer, jiffies); } static inline int post_send(struct ipoib_dev_priv *priv, unsigned int wr_id, struct ib_ah *address, u32 qpn, - u64 addr, int len) + struct ipoib_tx_buf *tx_req, + 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; - priv->tx_sge.addr = addr; - priv->tx_sge.length = len; - - priv->tx_wr.wr_id = wr_id; - priv->tx_wr.wr.ud.remote_qpn = qpn; - priv->tx_wr.wr.ud.ah = address; + 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; + + 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; + } else + priv->tx_wr.opcode = IB_WR_SEND; return ib_post_send(priv->qp, &priv->tx_wr, &bad_wr); } @@ -362,15 +544,30 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb, { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_tx_buf *tx_req; - u64 addr; - - if (unlikely(skb->len > priv->mcast_mtu + IPOIB_ENCAP_LEN)) { - ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n", - skb->len, priv->mcast_mtu + IPOIB_ENCAP_LEN); - ++priv->stats.tx_dropped; - ++priv->stats.tx_errors; - ipoib_cm_skb_too_long(dev, skb, priv->mcast_mtu); - return; + int hlen, rc; + void *phead; + + if (skb_is_gso(skb)) { + hlen = skb_transport_offset(skb) + tcp_hdrlen(skb); + phead = skb->data; + if (unlikely(!skb_pull(skb, hlen))) { + ipoib_warn(priv, "linear data too small\n"); + ++dev->stats.tx_dropped; + ++dev->stats.tx_errors; + dev_kfree_skb_any(skb); + return; + } + } else { + if (unlikely(skb->len > priv->mcast_mtu + IPOIB_ENCAP_LEN)) { + ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n", + skb->len, priv->mcast_mtu + IPOIB_ENCAP_LEN); + ++dev->stats.tx_dropped; + ++dev->stats.tx_errors; + ipoib_cm_skb_too_long(dev, skb, priv->mcast_mtu); + return; + } + phead = NULL; + hlen = 0; } ipoib_dbg_data(priv, "sending packet, length=%d address=%p qpn=0x%06x\n", @@ -385,33 +582,47 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb, */ tx_req = &priv->tx_ring[priv->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))) { - ++priv->stats.tx_errors; + if (unlikely(ipoib_dma_map_tx(priv->ca, tx_req))) { + ++dev->stats.tx_errors; dev_kfree_skb_any(skb); return; } - tx_req->mapping = addr; - if (unlikely(post_send(priv, priv->tx_head & (ipoib_sendq_size - 1), - address->ah, qpn, addr, skb->len))) { - ipoib_warn(priv, "post_send failed\n"); - ++priv->stats.tx_errors; - ib_dma_unmap_single(priv->ca, addr, skb->len, DMA_TO_DEVICE); + if (skb->ip_summed == CHECKSUM_PARTIAL) + priv->tx_wr.send_flags |= IB_SEND_IP_CSUM; + else + priv->tx_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"); + if (ib_req_notify_cq(priv->send_cq, IB_CQ_NEXT_COMP)) + ipoib_warn(priv, "request notify on send CQ failed\n"); + netif_stop_queue(dev); + } + + skb_orphan(skb); + skb_dst_drop(skb); + + rc = post_send(priv, priv->tx_head & (ipoib_sendq_size - 1), + address->ah, qpn, tx_req, phead, hlen); + if (unlikely(rc)) { + 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); dev_kfree_skb_any(skb); + if (netif_queue_stopped(dev)) + netif_wake_queue(dev); } else { dev->trans_start = jiffies; address->last_send = priv->tx_head; ++priv->tx_head; - - if (priv->tx_head - priv->tx_tail == ipoib_sendq_size) { - ipoib_dbg(priv, "TX ring full, stopping kernel net queue\n"); - netif_stop_queue(dev); - set_bit(IPOIB_FLAG_NETIF_STOPPED, &priv->flags); - } } + + if (unlikely(priv->tx_outstanding > MAX_SEND_CQE)) + while (poll_tx(priv)) + ; /* nothing */ } static void __ipoib_reap_ah(struct net_device *dev) @@ -419,17 +630,20 @@ static void __ipoib_reap_ah(struct net_device *dev) struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_ah *ah, *tah; LIST_HEAD(remove_list); + unsigned long flags; + + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); list_for_each_entry_safe(ah, tah, &priv->dead_ahs, list) if ((int) priv->tx_tail - (int) ah->last_send >= 0) { list_del(&ah->list); ib_destroy_ah(ah->ah); kfree(ah); } - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); } void ipoib_reap_ah(struct work_struct *work) @@ -441,7 +655,13 @@ void ipoib_reap_ah(struct work_struct *work) __ipoib_reap_ah(dev); if (!test_bit(IPOIB_STOP_REAPER, &priv->flags)) - queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task, HZ); + queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task, + round_jiffies_relative(HZ)); +} + +static void ipoib_ib_tx_timer_func(unsigned long ctx) +{ + drain_tx_cq((struct net_device *)ctx); } int ipoib_ib_dev_open(struct net_device *dev) @@ -465,23 +685,28 @@ int ipoib_ib_dev_open(struct net_device *dev) ret = ipoib_ib_post_receives(dev); if (ret) { ipoib_warn(priv, "ipoib_ib_post_receives returned %d\n", ret); - ipoib_ib_dev_stop(dev, 1); - return -1; + goto dev_stop; } ret = ipoib_cm_dev_open(dev); if (ret) { ipoib_warn(priv, "ipoib_cm_dev_open returned %d\n", ret); - ipoib_ib_dev_stop(dev, 1); - return -1; + goto dev_stop; } clear_bit(IPOIB_STOP_REAPER, &priv->flags); - queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task, HZ); + 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; +dev_stop: + if (!test_and_set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) + napi_enable(&priv->napi); + ipoib_ib_dev_stop(dev, 1); + return -1; } static void ipoib_pkey_dev_check_presence(struct net_device *dev) @@ -489,7 +714,7 @@ static void ipoib_pkey_dev_check_presence(struct net_device *dev) struct ipoib_dev_priv *priv = netdev_priv(dev); u16 pkey_index = 0; - if (ib_find_cached_pkey(priv->ca, priv->port, priv->pkey, &pkey_index)) + if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &pkey_index)) clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); else set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); @@ -524,10 +749,8 @@ int ipoib_ib_dev_down(struct net_device *dev, int flush) if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) { mutex_lock(&pkey_mutex); set_bit(IPOIB_PKEY_STOP, &priv->flags); - cancel_delayed_work(&priv->pkey_poll_task); + cancel_delayed_work_sync(&priv->pkey_poll_task); mutex_unlock(&pkey_mutex); - if (flush) - flush_workqueue(ipoib_workqueue); } ipoib_mcast_stop_thread(dev, flush); @@ -555,17 +778,39 @@ void ipoib_drain_cq(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); int i, n; + + /* + * We call completion handling routines that expect to be + * called from the BH-disabled NAPI poll context, so disable + * BHs here too. + */ + local_bh_disable(); + do { - n = ib_poll_cq(priv->cq, IPOIB_NUM_WC, priv->ibwc); + n = ib_poll_cq(priv->recv_cq, IPOIB_NUM_WC, priv->ibwc); for (i = 0; i < n; ++i) { - if (priv->ibwc[i].wr_id & IPOIB_CM_OP_SRQ) - ipoib_cm_handle_rx_wc(dev, priv->ibwc + i); - else if (priv->ibwc[i].wr_id & IPOIB_OP_RECV) - ipoib_ib_handle_rx_wc(dev, priv->ibwc + i); - else - ipoib_ib_handle_tx_wc(dev, priv->ibwc + i); + /* + * Convert any successful completions to flush + * errors to avoid passing packets up the + * stack after bringing the device down. + */ + if (priv->ibwc[i].status == IB_WC_SUCCESS) + priv->ibwc[i].status = IB_WC_WR_FLUSH_ERR; + + if (priv->ibwc[i].wr_id & IPOIB_OP_RECV) { + if (priv->ibwc[i].wr_id & IPOIB_OP_CM) + ipoib_cm_handle_rx_wc(dev, priv->ibwc + i); + else + ipoib_ib_handle_rx_wc(dev, priv->ibwc + i); + } else + ipoib_cm_handle_tx_wc(dev, priv->ibwc + i); } } while (n == IPOIB_NUM_WC); + + while (poll_tx(priv)) + ; /* nothing */ + + local_bh_enable(); } int ipoib_ib_dev_stop(struct net_device *dev, int flush) @@ -576,8 +821,8 @@ 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); - netif_poll_disable(dev); + if (test_and_clear_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) + napi_disable(&priv->napi); ipoib_cm_dev_stop(dev); @@ -604,12 +849,10 @@ int ipoib_ib_dev_stop(struct net_device *dev, int flush) while ((int) priv->tx_tail - (int) priv->tx_head < 0) { tx_req = &priv->tx_ring[priv->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->ca, tx_req); dev_kfree_skb_any(tx_req->skb); ++priv->tx_tail; + --priv->tx_outstanding; } for (i = 0; i < ipoib_recvq_size; ++i) { @@ -618,10 +861,8 @@ int ipoib_ib_dev_stop(struct net_device *dev, int flush) rx_req = &priv->rx_ring[i]; if (!rx_req->skb) continue; - ib_dma_unmap_single(priv->ca, - rx_req->mapping, - IPOIB_BUF_SIZE, - DMA_FROM_DEVICE); + ipoib_ud_dma_unmap_rx(priv, + priv->rx_ring[i].mapping); dev_kfree_skb_any(rx_req->skb); rx_req->skb = NULL; } @@ -637,6 +878,7 @@ int ipoib_ib_dev_stop(struct net_device *dev, int flush) ipoib_dbg(priv, "All sends and receives done.\n"); timeout: + del_timer_sync(&priv->poll_timer); qp_attr.qp_state = IB_QPS_RESET; if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE)) ipoib_warn(priv, "Failed to modify QP to RESET state\n"); @@ -660,8 +902,7 @@ timeout: msleep(1); } - netif_poll_enable(dev); - ib_req_notify_cq(priv->cq, IB_CQ_NEXT_COMP); + ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP); return 0; } @@ -679,6 +920,9 @@ int ipoib_ib_dev_init(struct net_device *dev, struct ib_device *ca, int port) return -ENODEV; } + setup_timer(&priv->poll_timer, ipoib_ib_tx_timer_func, + (unsigned long) dev); + if (dev->flags & IFF_UP) { if (ipoib_ib_dev_open(dev)) { ipoib_transport_dev_cleanup(dev); @@ -689,24 +933,64 @@ int ipoib_ib_dev_init(struct net_device *dev, struct ib_device *ca, int port) return 0; } -static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv, int pkey_event) +/* + * Takes whatever value which is in pkey index 0 and updates priv->pkey + * returns 0 if the pkey value was changed. + */ +static inline int update_parent_pkey(struct ipoib_dev_priv *priv) +{ + int result; + u16 prev_pkey; + + prev_pkey = priv->pkey; + result = ib_query_pkey(priv->ca, priv->port, 0, &priv->pkey); + if (result) { + ipoib_warn(priv, "ib_query_pkey port %d failed (ret = %d)\n", + priv->port, result); + return result; + } + + priv->pkey |= 0x8000; + + if (prev_pkey != priv->pkey) { + ipoib_dbg(priv, "pkey changed from 0x%x to 0x%x\n", + prev_pkey, priv->pkey); + /* + * Update the pkey in the broadcast address, while making sure to set + * the full membership bit, so that we join the right broadcast group. + */ + priv->dev->broadcast[8] = priv->pkey >> 8; + priv->dev->broadcast[9] = priv->pkey & 0xff; + return 0; + } + + return 1; +} + +static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv, + enum ipoib_flush_level level) { struct ipoib_dev_priv *cpriv; struct net_device *dev = priv->dev; u16 new_index; + int result; - mutex_lock(&priv->vlan_mutex); + down_read(&priv->vlan_rwsem); /* * 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, pkey_event); + __ipoib_ib_dev_flush(cpriv, level); - mutex_unlock(&priv->vlan_mutex); + up_read(&priv->vlan_rwsem); if (!test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) { + /* for non-child devices must check/update the pkey value here */ + if (level == IPOIB_FLUSH_HEAVY && + !test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) + update_parent_pkey(priv); ipoib_dbg(priv, "Not flushing - IPOIB_FLAG_INITIALIZED not set.\n"); return; } @@ -716,28 +1000,44 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv, int pkey_event) return; } - if (pkey_event) { - if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &new_index)) { - clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); - ipoib_ib_dev_down(dev, 0); - ipoib_pkey_dev_delay_open(dev); - return; - } - set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); - - /* restart QP only if P_Key index is changed */ - if (new_index == priv->pkey_index) { - ipoib_dbg(priv, "Not flushing - P_Key index not changed.\n"); - return; + if (level == IPOIB_FLUSH_HEAVY) { + /* child devices chase their origin pkey value, while non-child + * (parent) devices should always takes what present in pkey index 0 + */ + if (test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { + if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &new_index)) { + clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); + ipoib_ib_dev_down(dev, 0); + ipoib_ib_dev_stop(dev, 0); + if (ipoib_pkey_dev_delay_open(dev)) + return; + } + /* restart QP only if P_Key index is changed */ + if (test_and_set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags) && + new_index == priv->pkey_index) { + ipoib_dbg(priv, "Not flushing - P_Key index not changed.\n"); + return; + } + priv->pkey_index = new_index; + } else { + result = update_parent_pkey(priv); + /* restart QP only if P_Key value changed */ + if (result) { + ipoib_dbg(priv, "Not flushing - P_Key value not changed.\n"); + return; + } } - priv->pkey_index = new_index; } - ipoib_dbg(priv, "flushing\n"); + if (level == IPOIB_FLUSH_LIGHT) { + ipoib_mark_paths_invalid(dev); + ipoib_mcast_dev_flush(dev); + } - ipoib_ib_dev_down(dev, 0); + if (level >= IPOIB_FLUSH_NORMAL) + ipoib_ib_dev_down(dev, 0); - if (pkey_event) { + if (level == IPOIB_FLUSH_HEAVY) { ipoib_ib_dev_stop(dev, 0); ipoib_ib_dev_open(dev); } @@ -747,27 +1047,34 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv, int pkey_event) * we get here, don't bring it back up if it's not configured up */ if (test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) { - ipoib_ib_dev_up(dev); + if (level >= IPOIB_FLUSH_NORMAL) + ipoib_ib_dev_up(dev); ipoib_mcast_restart_task(&priv->restart_task); } } -void ipoib_ib_dev_flush(struct work_struct *work) +void ipoib_ib_dev_flush_light(struct work_struct *work) { struct ipoib_dev_priv *priv = - container_of(work, struct ipoib_dev_priv, flush_task); + container_of(work, struct ipoib_dev_priv, flush_light); - ipoib_dbg(priv, "Flushing %s\n", priv->dev->name); - __ipoib_ib_dev_flush(priv, 0); + __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_LIGHT); } -void ipoib_pkey_event(struct work_struct *work) +void ipoib_ib_dev_flush_normal(struct work_struct *work) { struct ipoib_dev_priv *priv = - container_of(work, struct ipoib_dev_priv, pkey_event_task); + container_of(work, struct ipoib_dev_priv, flush_normal); - ipoib_dbg(priv, "Flushing %s and restarting its QP\n", priv->dev->name); - __ipoib_ib_dev_flush(priv, 1); + __ipoib_ib_dev_flush(priv, IPOIB_FLUSH_NORMAL); +} + +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); } void ipoib_ib_dev_cleanup(struct net_device *dev) @@ -775,6 +1082,11 @@ void ipoib_ib_dev_cleanup(struct net_device *dev) struct ipoib_dev_priv *priv = netdev_priv(dev); ipoib_dbg(priv, "cleaning up ib_dev\n"); + /* + * We must make sure there are no more (path) completions + * that may wish to touch priv fields that are no longer valid + */ + ipoib_flush_paths(dev); ipoib_mcast_stop_thread(dev, 1); ipoib_mcast_dev_flush(dev); diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index 894b1dcdf3e..5786a78ff8b 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -30,8 +30,6 @@ * 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" @@ -41,17 +39,24 @@ #include <linux/init.h> #include <linux/slab.h> #include <linux/kernel.h> +#include <linux/vmalloc.h> #include <linux/if_arp.h> /* For ARPHRD_xxx */ #include <linux/ip.h> #include <linux/in.h> -#include <net/dst.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; @@ -85,6 +90,7 @@ 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", @@ -98,24 +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)) { - ipoib_ib_dev_stop(dev, 1); - return -EINVAL; - } + if (ipoib_ib_dev_up(dev)) + goto err_stop; if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { struct ipoib_dev_priv *cpriv; /* Bring up any child interfaces too */ - mutex_lock(&priv->vlan_mutex); + down_read(&priv->vlan_rwsem); list_for_each_entry(cpriv, &priv->child_intfs, list) { int flags; @@ -125,12 +131,20 @@ int ipoib_open(struct net_device *dev) dev_change_flags(cpriv->dev, flags | IFF_UP); } - mutex_unlock(&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) @@ -143,22 +157,14 @@ static int ipoib_stop(struct net_device *dev) netif_stop_queue(dev); - clear_bit(IPOIB_FLAG_NETIF_STOPPED, &priv->flags); - - /* - * Now flush workqueue to make sure a scheduled task doesn't - * bring our internal state back up. - */ - flush_workqueue(ipoib_workqueue); - ipoib_ib_dev_down(dev, 1); - ipoib_ib_dev_stop(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 */ - mutex_lock(&priv->vlan_mutex); + down_read(&priv->vlan_rwsem); list_for_each_entry(cpriv, &priv->child_intfs, list) { int flags; @@ -168,28 +174,46 @@ static int ipoib_stop(struct net_device *dev) dev_change_flags(cpriv->dev, flags & ~IFF_UP); } - mutex_unlock(&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); /* dev->mtu > 2K ==> connected mode */ - if (ipoib_cm_admin_enabled(dev) && new_mtu <= IPOIB_CM_MTU) { + 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_PACKET_SIZE - IPOIB_ENCAP_LEN) { + if (new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu)) return -EINVAL; - } priv->admin_mtu = new_mtu; @@ -198,6 +222,37 @@ static int ipoib_change_mtu(struct net_device *dev, int new_mtu) return 0; } +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); @@ -254,30 +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); + ipoib_dbg(netdev_priv(dev), "path_free\n"); - 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); - - ipoib_neigh_free(dev, neigh); - } - - 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); @@ -343,17 +383,34 @@ void ipoib_path_iter_read(struct ipoib_path_iter *iter, #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); struct ipoib_path *path, *tp; LIST_HEAD(remove_list); + unsigned long flags; - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); + 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); @@ -361,15 +418,16 @@ void ipoib_flush_paths(struct net_device *dev) list_for_each_entry_safe(path, tp, &remove_list, list) { if (path->query) ib_sa_cancel_query(path->query_id, path->query); - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); wait_for_completion(&path->done); path_free(dev, path); - spin_lock_irq(&priv->tx_lock); - spin_lock(&priv->lock); + netif_tx_lock_bh(dev); + spin_lock_irqsave(&priv->lock, flags); } - spin_unlock(&priv->lock); - spin_unlock_irq(&priv->tx_lock); + + spin_unlock_irqrestore(&priv->lock, flags); + netif_tx_unlock_bh(dev); } static void path_rec_completion(int status, @@ -380,17 +438,18 @@ 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_ah *old_ah = NULL; struct ipoib_neigh *neigh, *tn; struct sk_buff_head skqueue; struct sk_buff *skb; unsigned long flags; if (!status) - ipoib_dbg(priv, "PathRec LID 0x%04x for GID " IPOIB_GID_FMT "\n", - be16_to_cpu(pathrec->dlid), IPOIB_GID_ARG(pathrec->dgid)); + 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); @@ -403,11 +462,12 @@ static void path_rec_completion(int status, 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); @@ -415,21 +475,27 @@ static void path_rec_completion(int status, __skb_queue_tail(&skqueue, skb); 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; - memcpy(&neigh->dgid.raw, &path->pathrec.dgid.raw, - sizeof(union ib_gid)); - if (ipoib_cm_enabled(dev, neigh->neighbour)) { + 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)) { - list_del(&neigh->list); - if (neigh->ah) - ipoib_put_ah(neigh->ah); - ipoib_neigh_free(dev, neigh); + ipoib_neigh_free(neigh); continue; } } @@ -437,6 +503,7 @@ static void path_rec_completion(int status, while ((skb = __skb_dequeue(&neigh->queue))) __skb_queue_tail(&skqueue, skb); } + path->valid = 1; } path->query = NULL; @@ -444,6 +511,12 @@ static void path_rec_completion(int status, 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)) @@ -457,6 +530,9 @@ 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; + if (!priv->broadcast) + return NULL; + path = kzalloc(sizeof *path, GFP_ATOMIC); if (!path) return NULL; @@ -468,9 +544,10 @@ static struct ipoib_path *path_rec_create(struct net_device *dev, void *gid) INIT_LIST_HEAD(&path->neigh_list); 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.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; } @@ -480,8 +557,8 @@ 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); @@ -491,41 +568,41 @@ static int path_rec_start(struct net_device *dev, 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 = ipoib_neigh_alloc(skb->dst->neighbour); + 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; } - /* - * 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, skb->dst->neighbour->ha + 4); + path = __path_find(dev, daddr + 4); if (!path) { - path = path_rec_create(dev, skb->dst->neighbour->ha + 4); + path = path_rec_create(dev, daddr + 4); if (!path) goto err_path; @@ -537,17 +614,12 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev) if (path->ah) { kref_get(&path->ah->ref); neigh->ah = path->ah; - memcpy(&neigh->dgid.raw, &path->pathrec.dgid.raw, - sizeof(union ib_gid)); - if (ipoib_cm_enabled(dev, neigh->neighbour)) { + 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)) { - list_del(&neigh->list); - if (neigh->ah) - ipoib_put_ah(neigh->ah); - ipoib_neigh_free(dev, neigh); + ipoib_neigh_free(neigh); goto err_drop; } if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) @@ -557,80 +629,68 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev) skb_queue_len(&neigh->queue)); goto err_drop; } - } else - ipoib_send(dev, skb, path->ah, IPOIB_QPN(skb->dst->neighbour->ha)); + } else { + 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_list; + 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_list: - list_del(&neigh->list); - err_path: - ipoib_neigh_free(dev, neigh); + ipoib_neigh_free(neigh); err_drop: - ++priv->stats.tx_dropped; + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); - spin_unlock(&priv->lock); -} - -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) { - 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); + 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, phdr->hwaddr + 4); - if (!path) { - path = path_rec_create(dev, 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; } @@ -638,128 +698,103 @@ static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, ipoib_dbg(priv, "Send unicast ARP to %04x\n", be16_to_cpu(path->pathrec.dlid)); - ipoib_send(dev, skb, path->ah, IPOIB_QPN(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; - if (unlikely(!spin_trylock_irqsave(&priv->tx_lock, 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 (likely(skb->dst && skb->dst->neighbour)) { - if (unlikely(!*to_ipoib_neigh(skb->dst->neighbour))) { - ipoib_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 (ipoib_cm_get(neigh)) { - if (ipoib_cm_up(neigh)) { - ipoib_cm_send(dev, skb, ipoib_cm_get(neigh)); - goto out; - } - } else if (neigh->ah) { - if (unlikely(memcmp(&neigh->dgid.raw, - skb->dst->neighbour->ha + 4, - sizeof(union ib_gid)))) { - spin_lock(&priv->lock); - /* - * 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. - */ - ipoib_put_ah(neigh->ah); - list_del(&neigh->list); - ipoib_neigh_free(dev, neigh); - spin_unlock(&priv->lock); - ipoib_path_lookup(skb, dev); - goto out; - } - - ipoib_send(dev, skb, neigh->ah, IPOIB_QPN(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, phdr->hwaddr + 4, skb); - } else { - /* unicast GID -- should be ARP or RARP reply */ - - 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 " - IPOIB_GID_FMT "\n", - skb->dst ? "neigh" : "dst", - be16_to_cpup((__be16 *) skb->data), - IPOIB_QPN(phdr->hwaddr), - IPOIB_GID_RAW_ARG(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); @@ -775,9 +810,10 @@ static void ipoib_timeout(struct net_device *dev) 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); @@ -785,17 +821,13 @@ 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) && daddr) { - 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) @@ -810,93 +842,445 @@ static void ipoib_set_mcast_list(struct net_device *dev) queue_work(ipoib_workqueue, &priv->restart_task); } -static void ipoib_neigh_cleanup(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_cleanup for %06x " IPOIB_GID_FMT "\n", - IPOIB_QPN(n->ha), - IPOIB_GID_RAW_ARG(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); - ipoib_neigh_free(n->dev, 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); +} + +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); - if (ah) - ipoib_put_ah(ah); + __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); } -struct ipoib_neigh *ipoib_neigh_alloc(struct neighbour *neighbour) + +static struct ipoib_neigh *ipoib_neigh_ctor(u8 *daddr, + struct net_device *dev) { struct ipoib_neigh *neigh; - neigh = kmalloc(sizeof *neigh, GFP_ATOMIC); + neigh = kzalloc(sizeof *neigh, GFP_ATOMIC); if (!neigh) return NULL; - neigh->neighbour = neighbour; - *to_ipoib_neigh(neighbour) = neigh; + 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 + */ + 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; + } + } + + 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_free(struct net_device *dev, struct ipoib_neigh *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; - *to_ipoib_neigh(neigh->neighbour) = NULL; + if (neigh->ah) + ipoib_put_ah(neigh->ah); while ((skb = __skb_dequeue(&neigh->queue))) { - ++priv->stats.tx_dropped; + ++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 void ipoib_neigh_reclaim(struct rcu_head *rp) +{ + /* 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_setup_dev(struct net_device *dev, struct neigh_parms *parms) +static int ipoib_neigh_hash_init(struct ipoib_dev_priv *priv) { - parms->neigh_cleanup = ipoib_neigh_cleanup; + 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 = 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_recvq_size); - goto out; + goto out_neigh_hash_cleanup; } - priv->tx_ring = kzalloc(ipoib_sendq_size * sizeof *priv->tx_ring, - 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_sendq_size); goto out_rx_ring_cleanup; } - /* 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; @@ -904,11 +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; } @@ -916,72 +1302,77 @@ 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_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); kfree(priv->rx_ring); - kfree(priv->tx_ring); + vfree(priv->tx_ring); 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->poll = ipoib_poll; - dev->weight = 100; + dev->netdev_ops = &ipoib_netdev_ops; + dev->header_ops = &ipoib_header_ops; - dev->watchdog_timeo = HZ; + ipoib_set_ethtool_ops(dev); - dev->flags |= IFF_BROADCAST | IFF_MULTICAST; + netif_napi_add(dev, &priv->napi, ipoib_poll, NAPI_POLL_WEIGHT); - /* - * 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_sendq_size * 2; - dev->features = NETIF_F_VLAN_CHALLENGED | NETIF_F_LLTX; + dev->watchdog_timeo = HZ; - /* MTU will be reset when mcast join happens */ - dev->mtu = IPOIB_PACKET_SIZE - IPOIB_ENCAP_LEN; - priv->mcast_mtu = priv->admin_mtu = dev->mtu; - - memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN); + dev->flags |= IFF_BROADCAST | IFF_MULTICAST; - netif_carrier_off(dev); + 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; - SET_MODULE_OWNER(dev); + memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN); priv->dev = dev; spin_lock_init(&priv->lock); - spin_lock_init(&priv->tx_lock); - mutex_init(&priv->mcast_mutex); - mutex_init(&priv->vlan_mutex); + init_rwsem(&priv->vlan_rwsem); INIT_LIST_HEAD(&priv->path_list); INIT_LIST_HEAD(&priv->child_intfs); @@ -989,11 +1380,14 @@ static void ipoib_setup(struct net_device *dev) INIT_LIST_HEAD(&priv->multicast_list); INIT_DELAYED_WORK(&priv->pkey_poll_task, ipoib_pkey_poll); - INIT_WORK(&priv->pkey_event_task, ipoib_pkey_event); INIT_DELAYED_WORK(&priv->mcast_task, ipoib_mcast_join_task); - INIT_WORK(&priv->flush_task, ipoib_ib_dev_flush); + 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) @@ -1017,6 +1411,43 @@ static ssize_t show_pkey(struct device *dev, } 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)); +} + +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) @@ -1027,7 +1458,7 @@ static ssize_t create_child(struct device *dev, if (sscanf(buf, "%i", &pkey) != 1) return -EINVAL; - if (pkey < 0 || pkey > 0xffff) + if (pkey <= 0 || pkey > 0xffff || pkey == 0x8000) return -EINVAL; /* @@ -1040,7 +1471,7 @@ static ssize_t create_child(struct device *dev, return ret ? ret : count; } -static 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 device *dev, struct device_attribute *attr, @@ -1060,17 +1491,54 @@ static ssize_t delete_child(struct device *dev, return ret ? ret : count; } -static 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 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); @@ -1078,14 +1546,32 @@ 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. @@ -1099,11 +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", @@ -1134,6 +1619,8 @@ static struct net_device *ipoib_add_port(const char *format, goto sysfs_failed; if (ipoib_add_pkey_attr(priv->dev)) goto sysfs_failed; + 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 (device_create_file(&priv->dev->dev, &dev_attr_delete_child)) @@ -1147,7 +1634,10 @@ sysfs_failed: register_failed: ib_unregister_event_handler(&priv->event_handler); - flush_scheduled_work(); + /* 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); @@ -1184,6 +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); @@ -1203,13 +1695,22 @@ static void ipoib_remove_one(struct ib_device *device) 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); - flush_scheduled_work(); + + 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); } @@ -1226,7 +1727,16 @@ static int __init ipoib_init_module(void) ipoib_sendq_size = roundup_pow_of_two(ipoib_sendq_size); ipoib_sendq_size = min(ipoib_sendq_size, IPOIB_MAX_QUEUE_SIZE); - ipoib_sendq_size = max(ipoib_sendq_size, IPOIB_MIN_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) @@ -1252,8 +1762,15 @@ static int __init ipoib_init_module(void) if (ret) goto err_sa; + ret = ipoib_netlink_init(); + if (ret) + goto err_client; + return 0; +err_client: + ib_unregister_client(&ipoib_client); + err_sa: ib_sa_unregister_client(&ipoib_sa_client); destroy_workqueue(ipoib_workqueue); @@ -1266,6 +1783,7 @@ 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(); diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c index aae367057a5..d4e005720d0 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c @@ -30,18 +30,18 @@ * 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_multicast.c 1362 2004-12-18 15:56:29Z roland $ */ #include <linux/skbuff.h> #include <linux/rtnetlink.h> +#include <linux/moduleparam.h> #include <linux/ip.h> #include <linux/in.h> #include <linux/igmp.h> #include <linux/inetdevice.h> #include <linux/delay.h> #include <linux/completion.h> +#include <linux/slab.h> #include <net/dst.h> @@ -57,28 +57,6 @@ MODULE_PARM_DESC(mcast_debug_level, static DEFINE_MUTEX(mcast_mutex); -/* Used for all multicast joins (broadcast, IPv4 mcast and IPv6 mcast) */ -struct ipoib_mcast { - struct ib_sa_mcmember_rec mcmember; - struct ib_sa_multicast *mc; - struct ipoib_ah *ah; - - struct rb_node rb_node; - struct list_head list; - - unsigned long created; - unsigned long backoff; - - unsigned long flags; - unsigned char logcount; - - struct list_head neigh_list; - - struct sk_buff_head pkt_queue; - - struct net_device *dev; -}; - struct ipoib_mcast_iter { struct net_device *dev; union ib_gid mgid; @@ -91,30 +69,13 @@ struct ipoib_mcast_iter { static void ipoib_mcast_free(struct ipoib_mcast *mcast) { struct net_device *dev = mcast->dev; - struct ipoib_dev_priv *priv = netdev_priv(dev); - struct ipoib_neigh *neigh, *tmp; - unsigned long flags; int tx_dropped = 0; - ipoib_dbg_mcast(netdev_priv(dev), - "deleting multicast group " IPOIB_GID_FMT "\n", - IPOIB_GID_ARG(mcast->mcmember.mgid)); - - spin_lock_irqsave(&priv->lock, flags); - - list_for_each_entry_safe(neigh, tmp, &mcast->neigh_list, list) { - /* - * It's safe to call ipoib_put_ah() inside priv->lock - * here, because we know that mcast->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); - ipoib_neigh_free(dev, neigh); - } + ipoib_dbg_mcast(netdev_priv(dev), "deleting multicast group %pI6\n", + mcast->mcmember.mgid.raw); - spin_unlock_irqrestore(&priv->lock, flags); + /* remove all neigh connected to this mcast */ + ipoib_del_neighs_by_gid(dev, mcast->mcmember.mgid.raw); if (mcast->ah) ipoib_put_ah(mcast->ah); @@ -124,9 +85,9 @@ static void ipoib_mcast_free(struct ipoib_mcast *mcast) dev_kfree_skb_any(skb_dequeue(&mcast->pkt_queue)); } - spin_lock_irqsave(&priv->tx_lock, flags); - priv->stats.tx_dropped += tx_dropped; - spin_unlock_irqrestore(&priv->tx_lock, flags); + netif_tx_lock_bh(dev); + dev->stats.tx_dropped += tx_dropped; + netif_tx_unlock_bh(dev); kfree(mcast); } @@ -210,31 +171,46 @@ static int ipoib_mcast_join_finish(struct ipoib_mcast *mcast, struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_ah *ah; int ret; + int set_qkey = 0; mcast->mcmember = *mcmember; - /* Set the cached Q_Key before we attach if it's the broadcast group */ + /* Set the multicast MTU and cached Q_Key before we attach if it's + * the broadcast group. + */ if (!memcmp(mcast->mcmember.mgid.raw, priv->dev->broadcast + 4, sizeof (union ib_gid))) { + spin_lock_irq(&priv->lock); + if (!priv->broadcast) { + spin_unlock_irq(&priv->lock); + return -EAGAIN; + } + priv->mcast_mtu = IPOIB_UD_MTU(ib_mtu_enum_to_int(priv->broadcast->mcmember.mtu)); priv->qkey = be32_to_cpu(priv->broadcast->mcmember.qkey); + spin_unlock_irq(&priv->lock); priv->tx_wr.wr.ud.remote_qkey = priv->qkey; + set_qkey = 1; + + if (!ipoib_cm_admin_enabled(dev)) { + rtnl_lock(); + dev_set_mtu(dev, min(priv->mcast_mtu, priv->admin_mtu)); + rtnl_unlock(); + } } if (!test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) { if (test_and_set_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags)) { - ipoib_warn(priv, "multicast group " IPOIB_GID_FMT - " already attached\n", - IPOIB_GID_ARG(mcast->mcmember.mgid)); + ipoib_warn(priv, "multicast group %pI6 already attached\n", + mcast->mcmember.mgid.raw); return 0; } ret = ipoib_mcast_attach(dev, be16_to_cpu(mcast->mcmember.mlid), - &mcast->mcmember.mgid); + &mcast->mcmember.mgid, set_qkey); if (ret < 0) { - ipoib_warn(priv, "couldn't attach QP to multicast group " - IPOIB_GID_FMT "\n", - IPOIB_GID_ARG(mcast->mcmember.mgid)); + ipoib_warn(priv, "couldn't attach QP to multicast group %pI6\n", + mcast->mcmember.mgid.raw); clear_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags); return ret; @@ -258,16 +234,18 @@ static int ipoib_mcast_join_finish(struct ipoib_mcast *mcast, av.grh.dgid = mcast->mcmember.mgid; ah = ipoib_create_ah(dev, priv->pd, &av); - if (!ah) { - ipoib_warn(priv, "ib_address_create failed\n"); + if (IS_ERR(ah)) { + ipoib_warn(priv, "ib_address_create failed %ld\n", + -PTR_ERR(ah)); + /* use original error */ + return PTR_ERR(ah); } else { spin_lock_irq(&priv->lock); mcast->ah = ah; spin_unlock_irq(&priv->lock); - ipoib_dbg_mcast(priv, "MGID " IPOIB_GID_FMT - " AV %p, LID 0x%04x, SL %d\n", - IPOIB_GID_ARG(mcast->mcmember.mgid), + ipoib_dbg_mcast(priv, "MGID %pI6 AV %p, LID 0x%04x, SL %d\n", + mcast->mcmember.mgid.raw, mcast->ah->ah, be16_to_cpu(mcast->mcmember.mlid), mcast->mcmember.sl); @@ -275,23 +253,19 @@ static int ipoib_mcast_join_finish(struct ipoib_mcast *mcast, } /* actually send any queued packets */ - spin_lock_irq(&priv->tx_lock); + netif_tx_lock_bh(dev); while (!skb_queue_empty(&mcast->pkt_queue)) { struct sk_buff *skb = skb_dequeue(&mcast->pkt_queue); - spin_unlock_irq(&priv->tx_lock); - skb->dev = dev; - - if (!skb->dst || !skb->dst->neighbour) { - /* put pseudoheader back on for next time */ - skb_push(skb, sizeof (struct ipoib_pseudoheader)); - } + netif_tx_unlock_bh(dev); + skb->dev = dev; if (dev_queue_xmit(skb)) ipoib_warn(priv, "dev_queue_xmit failed to requeue packet\n"); - spin_lock_irq(&priv->tx_lock); + + netif_tx_lock_bh(dev); } - spin_unlock_irq(&priv->tx_lock); + netif_tx_unlock_bh(dev); return 0; } @@ -302,7 +276,6 @@ ipoib_mcast_sendonly_join_complete(int status, { struct ipoib_mcast *mcast = multicast->context; struct net_device *dev = mcast->dev; - struct ipoib_dev_priv *priv = netdev_priv(dev); /* We trap for port events ourselves. */ if (status == -ENETRESET) @@ -313,17 +286,16 @@ ipoib_mcast_sendonly_join_complete(int status, if (status) { if (mcast->logcount++ < 20) - ipoib_dbg_mcast(netdev_priv(dev), "multicast join failed for " - IPOIB_GID_FMT ", status %d\n", - IPOIB_GID_ARG(mcast->mcmember.mgid), status); + ipoib_dbg_mcast(netdev_priv(dev), "multicast join failed for %pI6, status %d\n", + mcast->mcmember.mgid.raw, status); /* Flush out any queued packets */ - spin_lock_irq(&priv->tx_lock); + netif_tx_lock_bh(dev); while (!skb_queue_empty(&mcast->pkt_queue)) { - ++priv->stats.tx_dropped; + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb_dequeue(&mcast->pkt_queue)); } - spin_unlock_irq(&priv->tx_lock); + netif_tx_unlock_bh(dev); /* Clear the busy flag so we try again */ status = test_and_clear_bit(IPOIB_MCAST_FLAG_BUSY, @@ -374,14 +346,35 @@ static int ipoib_mcast_sendonly_join(struct ipoib_mcast *mcast) ipoib_warn(priv, "ib_sa_join_multicast failed (ret = %d)\n", ret); } else { - ipoib_dbg_mcast(priv, "no multicast record for " IPOIB_GID_FMT - ", starting join\n", - IPOIB_GID_ARG(mcast->mcmember.mgid)); + ipoib_dbg_mcast(priv, "no multicast record for %pI6, starting join\n", + mcast->mcmember.mgid.raw); } return ret; } +void ipoib_mcast_carrier_on_task(struct work_struct *work) +{ + struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, + carrier_on_task); + struct ib_port_attr attr; + + /* + * Take rtnl_lock to avoid racing with ipoib_stop() and + * turning the carrier back on while a device is being + * removed. + */ + if (ib_query_port(priv->ca, priv->port, &attr) || + attr.state != IB_PORT_ACTIVE) { + ipoib_dbg(priv, "Keeping carrier off until IB port is active\n"); + return; + } + + rtnl_lock(); + netif_carrier_on(priv->dev); + rtnl_unlock(); +} + static int ipoib_mcast_join_complete(int status, struct ib_sa_multicast *multicast) { @@ -389,13 +382,14 @@ static int ipoib_mcast_join_complete(int status, struct net_device *dev = mcast->dev; struct ipoib_dev_priv *priv = netdev_priv(dev); - ipoib_dbg_mcast(priv, "join completion for " IPOIB_GID_FMT - " (status %d)\n", - IPOIB_GID_ARG(mcast->mcmember.mgid), status); + ipoib_dbg_mcast(priv, "join completion for %pI6 (status %d)\n", + mcast->mcmember.mgid.raw, status); /* We trap for port events ourselves. */ - if (status == -ENETRESET) - return 0; + if (status == -ENETRESET) { + status = 0; + goto out; + } if (!status) status = ipoib_mcast_join_finish(mcast, &multicast->rec); @@ -408,23 +402,24 @@ static int ipoib_mcast_join_complete(int status, &priv->mcast_task, 0); mutex_unlock(&mcast_mutex); + /* + * Defer carrier on work to ipoib_workqueue to avoid a + * deadlock on rtnl_lock here. + */ if (mcast == priv->broadcast) - netif_carrier_on(dev); + queue_work(ipoib_workqueue, &priv->carrier_on_task); - return 0; + status = 0; + goto out; } if (mcast->logcount++ < 20) { - if (status == -ETIMEDOUT) { - ipoib_dbg_mcast(priv, "multicast join failed for " IPOIB_GID_FMT - ", status %d\n", - IPOIB_GID_ARG(mcast->mcmember.mgid), - status); + if (status == -ETIMEDOUT || status == -EAGAIN) { + ipoib_dbg_mcast(priv, "multicast join failed for %pI6, status %d\n", + mcast->mcmember.mgid.raw, status); } else { - ipoib_warn(priv, "multicast join failed for " - IPOIB_GID_FMT ", status %d\n", - IPOIB_GID_ARG(mcast->mcmember.mgid), - status); + ipoib_warn(priv, "multicast join failed for %pI6, status %d\n", + mcast->mcmember.mgid.raw, status); } } @@ -442,7 +437,8 @@ static int ipoib_mcast_join_complete(int status, mcast->backoff * HZ); spin_unlock_irq(&priv->lock); mutex_unlock(&mcast_mutex); - +out: + complete(&mcast->done); return status; } @@ -456,8 +452,7 @@ static void ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast, ib_sa_comp_mask comp_mask; int ret = 0; - ipoib_dbg_mcast(priv, "joining MGID " IPOIB_GID_FMT "\n", - IPOIB_GID_ARG(mcast->mcmember.mgid)); + ipoib_dbg_mcast(priv, "joining MGID %pI6\n", mcast->mcmember.mgid.raw); rec.mgid = mcast->mcmember.mgid; rec.port_gid = priv->local_gid; @@ -493,11 +488,15 @@ static void ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast, } set_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags); + init_completion(&mcast->done); + set_bit(IPOIB_MCAST_JOIN_STARTED, &mcast->flags); + mcast->mc = ib_sa_join_multicast(&ipoib_sa_client, priv->ca, priv->port, &rec, comp_mask, GFP_KERNEL, ipoib_mcast_join_complete, mcast); if (IS_ERR(mcast->mc)) { clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags); + complete(&mcast->done); ret = PTR_ERR(mcast->mc); ipoib_warn(priv, "ib_sa_join_multicast failed, status %d\n", ret); @@ -519,10 +518,18 @@ void ipoib_mcast_join_task(struct work_struct *work) struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, mcast_task.work); struct net_device *dev = priv->dev; + struct ib_port_attr port_attr; if (!test_bit(IPOIB_MCAST_RUN, &priv->flags)) return; + if (ib_query_port(priv->ca, priv->port, &port_attr) || + port_attr.state != IB_PORT_ACTIVE) { + ipoib_dbg(priv, "port state is not ACTIVE (state = %d) suspending join task\n", + port_attr.state); + return; + } + if (ib_query_gid(priv->ca, priv->port, 0, &priv->local_gid)) ipoib_warn(priv, "ib_query_gid() failed\n"); else @@ -540,6 +547,9 @@ void ipoib_mcast_join_task(struct work_struct *work) if (!priv->broadcast) { struct ipoib_mcast *broadcast; + if (!test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) + return; + broadcast = ipoib_mcast_alloc(dev, 1); if (!broadcast) { ipoib_warn(priv, "failed to allocate broadcast group\n"); @@ -589,12 +599,6 @@ void ipoib_mcast_join_task(struct work_struct *work) return; } - priv->mcast_mtu = ib_mtu_enum_to_int(priv->broadcast->mcmember.mtu) - - IPOIB_ENCAP_LEN; - - if (!ipoib_cm_admin_enabled(dev)) - dev->mtu = min(priv->mcast_mtu, priv->admin_mtu); - ipoib_dbg_mcast(priv, "successfully joined all multicast groups\n"); clear_bit(IPOIB_MCAST_RUN, &priv->flags); @@ -611,10 +615,6 @@ int ipoib_mcast_start_thread(struct net_device *dev) queue_delayed_work(ipoib_workqueue, &priv->mcast_task, 0); mutex_unlock(&mcast_mutex); - spin_lock_irq(&priv->lock); - set_bit(IPOIB_MCAST_STARTED, &priv->flags); - spin_unlock_irq(&priv->lock); - return 0; } @@ -624,10 +624,6 @@ int ipoib_mcast_stop_thread(struct net_device *dev, int flush) ipoib_dbg_mcast(priv, "stopping multicast thread\n"); - spin_lock_irq(&priv->lock); - clear_bit(IPOIB_MCAST_STARTED, &priv->flags); - spin_unlock_irq(&priv->lock); - mutex_lock(&mcast_mutex); clear_bit(IPOIB_MCAST_RUN, &priv->flags); cancel_delayed_work(&priv->mcast_task); @@ -648,34 +644,32 @@ static int ipoib_mcast_leave(struct net_device *dev, struct ipoib_mcast *mcast) ib_sa_free_multicast(mcast->mc); if (test_and_clear_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags)) { - ipoib_dbg_mcast(priv, "leaving MGID " IPOIB_GID_FMT "\n", - IPOIB_GID_ARG(mcast->mcmember.mgid)); + ipoib_dbg_mcast(priv, "leaving MGID %pI6\n", + mcast->mcmember.mgid.raw); /* Remove ourselves from the multicast group */ - ret = ipoib_mcast_detach(dev, be16_to_cpu(mcast->mcmember.mlid), - &mcast->mcmember.mgid); + ret = ib_detach_mcast(priv->qp, &mcast->mcmember.mgid, + be16_to_cpu(mcast->mcmember.mlid)); if (ret) - ipoib_warn(priv, "ipoib_mcast_detach failed (result = %d)\n", ret); + ipoib_warn(priv, "ib_detach_mcast failed (result = %d)\n", ret); } return 0; } -void ipoib_mcast_send(struct net_device *dev, void *mgid, struct sk_buff *skb) +void ipoib_mcast_send(struct net_device *dev, u8 *daddr, struct sk_buff *skb) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_mcast *mcast; + unsigned long flags; + void *mgid = daddr + 4; - /* - * 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); - if (!test_bit(IPOIB_MCAST_STARTED, &priv->flags) || + if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags) || !priv->broadcast || !test_bit(IPOIB_MCAST_FLAG_ATTACHED, &priv->broadcast->flags)) { - ++priv->stats.tx_dropped; + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); goto unlock; } @@ -683,14 +677,14 @@ void ipoib_mcast_send(struct net_device *dev, void *mgid, struct sk_buff *skb) mcast = __ipoib_mcast_find(dev, mgid); if (!mcast) { /* Let's create a new send only group now */ - ipoib_dbg_mcast(priv, "setting up send only multicast group for " - IPOIB_GID_FMT "\n", IPOIB_GID_RAW_ARG(mgid)); + ipoib_dbg_mcast(priv, "setting up send only multicast group for %pI6\n", + mgid); mcast = ipoib_mcast_alloc(dev, 0); if (!mcast) { ipoib_warn(priv, "unable to allocate memory for " "multicast structure\n"); - ++priv->stats.tx_dropped; + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); goto out; } @@ -705,7 +699,7 @@ void ipoib_mcast_send(struct net_device *dev, void *mgid, struct sk_buff *skb) if (skb_queue_len(&mcast->pkt_queue) < IPOIB_MAX_MCAST_QUEUE) skb_queue_tail(&mcast->pkt_queue, skb); else { - ++priv->stats.tx_dropped; + ++dev->stats.tx_dropped; dev_kfree_skb_any(skb); } @@ -724,23 +718,28 @@ 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); + struct ipoib_neigh *neigh; + spin_unlock_irqrestore(&priv->lock, flags); + neigh = ipoib_neigh_get(dev, daddr); + spin_lock_irqsave(&priv->lock, flags); + if (!neigh) { + neigh = ipoib_neigh_alloc(daddr, dev); if (neigh) { kref_get(&mcast->ah->ref); - neigh->ah = mcast->ah; + neigh->ah = mcast->ah; list_add_tail(&neigh->list, &mcast->neigh_list); } } - + spin_unlock_irqrestore(&priv->lock, flags); ipoib_send(dev, skb, mcast->ah, IB_MULTICAST_QPN); + if (neigh) + ipoib_neigh_put(neigh); + return; } unlock: - spin_unlock(&priv->lock); + spin_unlock_irqrestore(&priv->lock, flags); } void ipoib_mcast_dev_flush(struct net_device *dev) @@ -768,28 +767,45 @@ void ipoib_mcast_dev_flush(struct net_device *dev) spin_unlock_irqrestore(&priv->lock, flags); + /* seperate between the wait to the leave*/ + list_for_each_entry_safe(mcast, tmcast, &remove_list, list) + if (test_bit(IPOIB_MCAST_JOIN_STARTED, &mcast->flags)) + wait_for_completion(&mcast->done); + list_for_each_entry_safe(mcast, tmcast, &remove_list, list) { ipoib_mcast_leave(dev, mcast); ipoib_mcast_free(mcast); } } +static int ipoib_mcast_addr_is_valid(const u8 *addr, const u8 *broadcast) +{ + /* reserved QPN, prefix, scope */ + if (memcmp(addr, broadcast, 6)) + return 0; + /* signature lower, pkey */ + if (memcmp(addr + 7, broadcast + 7, 3)) + return 0; + return 1; +} + void ipoib_mcast_restart_task(struct work_struct *work) { struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv, restart_task); struct net_device *dev = priv->dev; - struct dev_mc_list *mclist; + struct netdev_hw_addr *ha; struct ipoib_mcast *mcast, *tmcast; LIST_HEAD(remove_list); unsigned long flags; + struct ib_sa_mcmember_rec rec; ipoib_dbg_mcast(priv, "restarting multicast task\n"); ipoib_mcast_stop_thread(dev, 0); local_irq_save(flags); - netif_tx_lock(dev); + netif_addr_lock(dev); spin_lock(&priv->lock); /* @@ -803,22 +819,29 @@ void ipoib_mcast_restart_task(struct work_struct *work) clear_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags); /* Mark all of the entries that are found or don't exist */ - for (mclist = dev->mc_list; mclist; mclist = mclist->next) { + netdev_for_each_mc_addr(ha, dev) { union ib_gid mgid; - memcpy(mgid.raw, mclist->dmi_addr + 4, sizeof mgid); + if (!ipoib_mcast_addr_is_valid(ha->addr, dev->broadcast)) + continue; - /* Add in the P_Key */ - mgid.raw[4] = (priv->pkey >> 8) & 0xff; - mgid.raw[5] = priv->pkey & 0xff; + memcpy(mgid.raw, ha->addr + 4, sizeof mgid); mcast = __ipoib_mcast_find(dev, &mgid); if (!mcast || test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) { struct ipoib_mcast *nmcast; + /* ignore group which is directly joined by userspace */ + if (test_bit(IPOIB_FLAG_UMCAST, &priv->flags) && + !ib_sa_get_mcmember_rec(priv->ca, priv->port, &mgid, &rec)) { + ipoib_dbg_mcast(priv, "ignoring multicast entry for mgid %pI6\n", + mgid.raw); + continue; + } + /* Not found or send-only group, let's add a new entry */ - ipoib_dbg_mcast(priv, "adding multicast entry for mgid " - IPOIB_GID_FMT "\n", IPOIB_GID_ARG(mgid)); + ipoib_dbg_mcast(priv, "adding multicast entry for mgid %pI6\n", + mgid.raw); nmcast = ipoib_mcast_alloc(dev, 0); if (!nmcast) { @@ -851,8 +874,8 @@ void ipoib_mcast_restart_task(struct work_struct *work) list_for_each_entry_safe(mcast, tmcast, &priv->multicast_list, list) { if (!test_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags) && !test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) { - ipoib_dbg_mcast(priv, "deleting multicast group " IPOIB_GID_FMT "\n", - IPOIB_GID_ARG(mcast->mcmember.mgid)); + ipoib_dbg_mcast(priv, "deleting multicast group %pI6\n", + mcast->mcmember.mgid.raw); rb_erase(&mcast->rb_node, &priv->multicast_tree); @@ -862,7 +885,7 @@ void ipoib_mcast_restart_task(struct work_struct *work) } spin_unlock(&priv->lock); - netif_tx_unlock(dev); + netif_addr_unlock(dev); local_irq_restore(flags); /* We have to cancel outside of the spinlock */ diff --git a/drivers/infiniband/ulp/ipoib/ipoib_netlink.c b/drivers/infiniband/ulp/ipoib/ipoib_netlink.c new file mode 100644 index 00000000000..cdc7df4fdb8 --- /dev/null +++ b/drivers/infiniband/ulp/ipoib/ipoib_netlink.c @@ -0,0 +1,182 @@ +/* + * Copyright (c) 2012 Mellanox Technologies. - 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 + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * 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. + */ + +#include <linux/netdevice.h> +#include <linux/if_arp.h> /* For ARPHRD_xxx */ +#include <linux/module.h> +#include <net/rtnetlink.h> +#include "ipoib.h" + +static const struct nla_policy ipoib_policy[IFLA_IPOIB_MAX + 1] = { + [IFLA_IPOIB_PKEY] = { .type = NLA_U16 }, + [IFLA_IPOIB_MODE] = { .type = NLA_U16 }, + [IFLA_IPOIB_UMCAST] = { .type = NLA_U16 }, +}; + +static int ipoib_fill_info(struct sk_buff *skb, const struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + u16 val; + + if (nla_put_u16(skb, IFLA_IPOIB_PKEY, priv->pkey)) + goto nla_put_failure; + + val = test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); + if (nla_put_u16(skb, IFLA_IPOIB_MODE, val)) + goto nla_put_failure; + + val = test_bit(IPOIB_FLAG_UMCAST, &priv->flags); + if (nla_put_u16(skb, IFLA_IPOIB_UMCAST, val)) + goto nla_put_failure; + + return 0; + +nla_put_failure: + return -EMSGSIZE; +} + +static int ipoib_changelink(struct net_device *dev, + struct nlattr *tb[], struct nlattr *data[]) +{ + u16 mode, umcast; + int ret = 0; + + if (data[IFLA_IPOIB_MODE]) { + mode = nla_get_u16(data[IFLA_IPOIB_MODE]); + if (mode == IPOIB_MODE_DATAGRAM) + ret = ipoib_set_mode(dev, "datagram\n"); + else if (mode == IPOIB_MODE_CONNECTED) + ret = ipoib_set_mode(dev, "connected\n"); + else + ret = -EINVAL; + + if (ret < 0) + goto out_err; + } + + if (data[IFLA_IPOIB_UMCAST]) { + umcast = nla_get_u16(data[IFLA_IPOIB_UMCAST]); + ipoib_set_umcast(dev, umcast); + } + +out_err: + return ret; +} + +static int ipoib_new_child_link(struct net *src_net, struct net_device *dev, + struct nlattr *tb[], struct nlattr *data[]) +{ + struct net_device *pdev; + struct ipoib_dev_priv *ppriv; + u16 child_pkey; + int err; + + if (!tb[IFLA_LINK]) + return -EINVAL; + + pdev = __dev_get_by_index(src_net, nla_get_u32(tb[IFLA_LINK])); + if (!pdev || pdev->type != ARPHRD_INFINIBAND) + return -ENODEV; + + ppriv = netdev_priv(pdev); + + if (test_bit(IPOIB_FLAG_SUBINTERFACE, &ppriv->flags)) { + ipoib_warn(ppriv, "child creation disallowed for child devices\n"); + return -EINVAL; + } + + if (!data || !data[IFLA_IPOIB_PKEY]) { + ipoib_dbg(ppriv, "no pkey specified, using parent pkey\n"); + child_pkey = ppriv->pkey; + } else + child_pkey = nla_get_u16(data[IFLA_IPOIB_PKEY]); + + if (child_pkey == 0 || child_pkey == 0x8000) + return -EINVAL; + + /* + * Set the full membership bit, so that we join the right + * broadcast group, etc. + */ + child_pkey |= 0x8000; + + err = __ipoib_vlan_add(ppriv, netdev_priv(dev), child_pkey, IPOIB_RTNL_CHILD); + + if (!err && data) + err = ipoib_changelink(dev, tb, data); + return err; +} + +static void ipoib_unregister_child_dev(struct net_device *dev, struct list_head *head) +{ + struct ipoib_dev_priv *priv, *ppriv; + + priv = netdev_priv(dev); + ppriv = netdev_priv(priv->parent); + + down_write(&ppriv->vlan_rwsem); + unregister_netdevice_queue(dev, head); + list_del(&priv->list); + up_write(&ppriv->vlan_rwsem); +} + +static size_t ipoib_get_size(const struct net_device *dev) +{ + return nla_total_size(2) + /* IFLA_IPOIB_PKEY */ + nla_total_size(2) + /* IFLA_IPOIB_MODE */ + nla_total_size(2); /* IFLA_IPOIB_UMCAST */ +} + +static struct rtnl_link_ops ipoib_link_ops __read_mostly = { + .kind = "ipoib", + .maxtype = IFLA_IPOIB_MAX, + .policy = ipoib_policy, + .priv_size = sizeof(struct ipoib_dev_priv), + .setup = ipoib_setup, + .newlink = ipoib_new_child_link, + .changelink = ipoib_changelink, + .dellink = ipoib_unregister_child_dev, + .get_size = ipoib_get_size, + .fill_info = ipoib_fill_info, +}; + +int __init ipoib_netlink_init(void) +{ + return rtnl_link_register(&ipoib_link_ops); +} + +void __exit ipoib_netlink_fini(void) +{ + rtnl_link_unregister(&ipoib_link_ops); +} + +MODULE_ALIAS_RTNL_LINK("ipoib"); diff --git a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c index 563aeacf9e1..c56d5d44c53 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c @@ -29,24 +29,19 @@ * 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_verbs.c 1349 2004-12-16 21:09:43Z roland $ */ +#include <linux/slab.h> + #include "ipoib.h" -int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid) +int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid, int set_qkey) { struct ipoib_dev_priv *priv = netdev_priv(dev); - struct ib_qp_attr *qp_attr; + struct ib_qp_attr *qp_attr = NULL; int ret; u16 pkey_index; - ret = -ENOMEM; - qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL); - if (!qp_attr) - goto out; - if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &pkey_index)) { clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); ret = -ENXIO; @@ -54,18 +49,23 @@ int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid) } set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); - /* set correct QKey for QP */ - qp_attr->qkey = priv->qkey; - ret = ib_modify_qp(priv->qp, qp_attr, IB_QP_QKEY); - if (ret) { - ipoib_warn(priv, "failed to modify QP, ret = %d\n", ret); - goto out; + if (set_qkey) { + ret = -ENOMEM; + qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL); + if (!qp_attr) + goto out; + + /* set correct QKey for QP */ + qp_attr->qkey = priv->qkey; + ret = ib_modify_qp(priv->qp, qp_attr, IB_QP_QKEY); + if (ret) { + ipoib_warn(priv, "failed to modify QP, ret = %d\n", ret); + goto out; + } } /* attach QP to multicast group */ - mutex_lock(&priv->mcast_mutex); ret = ib_attach_mcast(priv->qp, mgid, mlid); - mutex_unlock(&priv->mcast_mutex); if (ret) ipoib_warn(priv, "failed to attach to multicast group, ret = %d\n", ret); @@ -74,20 +74,6 @@ out: return ret; } -int ipoib_mcast_detach(struct net_device *dev, u16 mlid, union ib_gid *mgid) -{ - struct ipoib_dev_priv *priv = netdev_priv(dev); - int ret; - - mutex_lock(&priv->mcast_mutex); - ret = ib_detach_mcast(priv->qp, mgid, mlid); - mutex_unlock(&priv->mcast_mutex); - if (ret) - ipoib_warn(priv, "ib_detach_mcast failed (result = %d)\n", ret); - - return ret; -} - int ipoib_init_qp(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); @@ -150,13 +136,14 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca) .max_send_wr = ipoib_sendq_size, .max_recv_wr = ipoib_recvq_size, .max_send_sge = 1, - .max_recv_sge = 1 + .max_recv_sge = IPOIB_UD_RX_SG }, .sq_sig_type = IB_SIGNAL_ALL_WR, .qp_type = IB_QPT_UD }; int ret, size; + int i; priv->pd = ib_alloc_pd(priv->ca); if (IS_ERR(priv->pd)) { @@ -170,44 +157,84 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca) goto out_free_pd; } - size = ipoib_sendq_size + ipoib_recvq_size + 1; + size = ipoib_recvq_size + 1; ret = ipoib_cm_dev_init(dev); - if (!ret) - size += ipoib_recvq_size + 1 /* 1 extra for rx_drain_qp */; + if (!ret) { + size += ipoib_sendq_size; + if (ipoib_cm_has_srq(dev)) + size += ipoib_recvq_size + 1; /* 1 extra for rx_drain_qp */ + else + size += ipoib_recvq_size * ipoib_max_conn_qp; + } - priv->cq = ib_create_cq(priv->ca, ipoib_ib_completion, NULL, dev, size, 0); - if (IS_ERR(priv->cq)) { - printk(KERN_WARNING "%s: failed to create CQ\n", ca->name); + priv->recv_cq = ib_create_cq(priv->ca, ipoib_ib_completion, NULL, dev, size, 0); + if (IS_ERR(priv->recv_cq)) { + printk(KERN_WARNING "%s: failed to create receive CQ\n", ca->name); goto out_free_mr; } - if (ib_req_notify_cq(priv->cq, IB_CQ_NEXT_COMP)) - goto out_free_cq; + priv->send_cq = ib_create_cq(priv->ca, ipoib_send_comp_handler, NULL, + dev, ipoib_sendq_size, 0); + if (IS_ERR(priv->send_cq)) { + printk(KERN_WARNING "%s: failed to create send CQ\n", ca->name); + goto out_free_recv_cq; + } + + if (ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP)) + goto out_free_send_cq; + + init_attr.send_cq = priv->send_cq; + init_attr.recv_cq = priv->recv_cq; + + if (priv->hca_caps & IB_DEVICE_UD_TSO) + init_attr.create_flags |= IB_QP_CREATE_IPOIB_UD_LSO; + + if (priv->hca_caps & IB_DEVICE_BLOCK_MULTICAST_LOOPBACK) + init_attr.create_flags |= IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK; - init_attr.send_cq = priv->cq; - init_attr.recv_cq = priv->cq, + if (priv->hca_caps & IB_DEVICE_MANAGED_FLOW_STEERING) + init_attr.create_flags |= IB_QP_CREATE_NETIF_QP; + + if (dev->features & NETIF_F_SG) + init_attr.cap.max_send_sge = MAX_SKB_FRAGS + 1; priv->qp = ib_create_qp(priv->pd, &init_attr); if (IS_ERR(priv->qp)) { printk(KERN_WARNING "%s: failed to create QP\n", ca->name); - goto out_free_cq; + goto out_free_send_cq; } priv->dev->dev_addr[1] = (priv->qp->qp_num >> 16) & 0xff; priv->dev->dev_addr[2] = (priv->qp->qp_num >> 8) & 0xff; priv->dev->dev_addr[3] = (priv->qp->qp_num ) & 0xff; - priv->tx_sge.lkey = priv->mr->lkey; - - priv->tx_wr.opcode = IB_WR_SEND; - priv->tx_wr.sg_list = &priv->tx_sge; - priv->tx_wr.num_sge = 1; - priv->tx_wr.send_flags = IB_SEND_SIGNALED; + for (i = 0; i < MAX_SKB_FRAGS + 1; ++i) + priv->tx_sge[i].lkey = priv->mr->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->rx_sge[0].lkey = priv->mr->lkey; + if (ipoib_ud_need_sg(priv->max_ib_mtu)) { + priv->rx_sge[0].length = IPOIB_UD_HEAD_SIZE; + priv->rx_sge[1].length = PAGE_SIZE; + priv->rx_sge[1].lkey = priv->mr->lkey; + priv->rx_wr.num_sge = IPOIB_UD_RX_SG; + } else { + priv->rx_sge[0].length = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu); + priv->rx_wr.num_sge = 1; + } + priv->rx_wr.next = NULL; + priv->rx_wr.sg_list = priv->rx_sge; return 0; -out_free_cq: - ib_destroy_cq(priv->cq); +out_free_send_cq: + ib_destroy_cq(priv->send_cq); + +out_free_recv_cq: + ib_destroy_cq(priv->recv_cq); out_free_mr: ib_dereg_mr(priv->mr); @@ -230,8 +257,11 @@ void ipoib_transport_dev_cleanup(struct net_device *dev) clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags); } - if (ib_destroy_cq(priv->cq)) - ipoib_warn(priv, "ib_cq_destroy failed\n"); + if (ib_destroy_cq(priv->send_cq)) + ipoib_warn(priv, "ib_cq_destroy (send) failed\n"); + + if (ib_destroy_cq(priv->recv_cq)) + ipoib_warn(priv, "ib_cq_destroy (recv) failed\n"); ipoib_cm_dev_cleanup(dev); @@ -251,15 +281,17 @@ void ipoib_event(struct ib_event_handler *handler, if (record->element.port_num != priv->port) return; - if (record->event == IB_EVENT_PORT_ERR || - record->event == IB_EVENT_PORT_ACTIVE || - record->event == IB_EVENT_LID_CHANGE || - record->event == IB_EVENT_SM_CHANGE || + ipoib_dbg(priv, "Event %d on device %s port %d\n", record->event, + record->device->name, record->element.port_num); + + if (record->event == IB_EVENT_SM_CHANGE || record->event == IB_EVENT_CLIENT_REREGISTER) { - ipoib_dbg(priv, "Port state change event\n"); - queue_work(ipoib_workqueue, &priv->flush_task); + queue_work(ipoib_workqueue, &priv->flush_light); + } else if (record->event == IB_EVENT_PORT_ERR || + record->event == IB_EVENT_PORT_ACTIVE || + record->event == IB_EVENT_LID_CHANGE) { + queue_work(ipoib_workqueue, &priv->flush_normal); } else if (record->event == IB_EVENT_PKEY_CHANGE) { - ipoib_dbg(priv, "P_Key change event on port:%d\n", priv->port); - queue_work(ipoib_workqueue, &priv->pkey_event_task); + queue_work(ipoib_workqueue, &priv->flush_heavy); } } diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c index 6762988439d..9fad7b5ac8b 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c @@ -28,14 +28,11 @@ * 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_vlan.c 1349 2004-12-16 21:09:43Z roland $ */ #include <linux/module.h> #include <linux/init.h> -#include <linux/slab.h> #include <linux/seq_file.h> #include <asm/uaccess.h> @@ -52,44 +49,20 @@ static ssize_t show_parent(struct device *d, struct device_attribute *attr, } static DEVICE_ATTR(parent, S_IRUGO, show_parent, NULL); -int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) +int __ipoib_vlan_add(struct ipoib_dev_priv *ppriv, struct ipoib_dev_priv *priv, + u16 pkey, int type) { - struct ipoib_dev_priv *ppriv, *priv; - char intf_name[IFNAMSIZ]; int result; - if (!capable(CAP_NET_ADMIN)) - return -EPERM; - - ppriv = netdev_priv(pdev); - - mutex_lock(&ppriv->vlan_mutex); - - /* - * First ensure this isn't a duplicate. We check the parent device and - * then all of the child interfaces to make sure the Pkey doesn't match. - */ - if (ppriv->pkey == pkey) { - result = -ENOTUNIQ; - goto err; - } - - list_for_each_entry(priv, &ppriv->child_intfs, list) { - if (priv->pkey == pkey) { - result = -ENOTUNIQ; - goto err; - } - } + priv->max_ib_mtu = ppriv->max_ib_mtu; + /* 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; + set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags); - snprintf(intf_name, sizeof intf_name, "%s.%04x", - ppriv->dev->name, pkey); - priv = ipoib_intf_alloc(intf_name); - if (!priv) { - result = -ENOMEM; + result = ipoib_set_dev_features(priv, ppriv->ca); + if (result) goto err; - } - - set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags); priv->pkey = pkey; @@ -102,10 +75,10 @@ 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_netdev(priv->dev); + result = register_netdevice(priv->dev); if (result) { ipoib_warn(priv, "failed to initialize; error %i", result); goto register_failed; @@ -115,58 +88,122 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) 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 (device_create_file(&priv->dev->dev, &dev_attr_parent)) - goto sysfs_failed; + /* RTNL childs don't need proprietary sysfs entries */ + if (type == IPOIB_LEGACY_CHILD) { + if (ipoib_cm_add_mode_attr(priv->dev)) + goto sysfs_failed; + if (ipoib_add_pkey_attr(priv->dev)) + goto sysfs_failed; + if (ipoib_add_umcast_attr(priv->dev)) + goto sysfs_failed; + + if (device_create_file(&priv->dev->dev, &dev_attr_parent)) + goto sysfs_failed; + } + priv->child_type = type; + priv->dev->iflink = ppriv->dev->ifindex; list_add_tail(&priv->list, &ppriv->child_intfs); - mutex_unlock(&ppriv->vlan_mutex); - return 0; sysfs_failed: + result = -ENOMEM; ipoib_delete_debug_files(priv->dev); - unregister_netdev(priv->dev); + unregister_netdevice(priv->dev); register_failed: ipoib_dev_cleanup(priv->dev); -device_init_failed: - free_netdev(priv->dev); - err: - mutex_unlock(&ppriv->vlan_mutex); + return result; +} + +int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey) +{ + struct ipoib_dev_priv *ppriv, *priv; + char intf_name[IFNAMSIZ]; + struct ipoib_dev_priv *tpriv; + int result; + + if (!capable(CAP_NET_ADMIN)) + return -EPERM; + + ppriv = netdev_priv(pdev); + + snprintf(intf_name, sizeof intf_name, "%s.%04x", + ppriv->dev->name, pkey); + priv = ipoib_intf_alloc(intf_name); + if (!priv) + return -ENOMEM; + + if (!rtnl_trylock()) + return restart_syscall(); + + down_write(&ppriv->vlan_rwsem); + + /* + * First ensure this isn't a duplicate. We check the parent device and + * then all of the legacy child interfaces to make sure the Pkey + * doesn't match. + */ + if (ppriv->pkey == pkey) { + result = -ENOTUNIQ; + goto out; + } + + list_for_each_entry(tpriv, &ppriv->child_intfs, list) { + if (tpriv->pkey == pkey && + tpriv->child_type == IPOIB_LEGACY_CHILD) { + result = -ENOTUNIQ; + goto out; + } + } + + result = __ipoib_vlan_add(ppriv, priv, pkey, IPOIB_LEGACY_CHILD); + +out: + up_write(&ppriv->vlan_rwsem); + + if (result) + free_netdev(priv->dev); + + rtnl_unlock(); + 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); - mutex_lock(&ppriv->vlan_mutex); + if (!rtnl_trylock()) + return restart_syscall(); + + down_write(&ppriv->vlan_rwsem); list_for_each_entry_safe(priv, tpriv, &ppriv->child_intfs, list) { - if (priv->pkey == pkey) { - unregister_netdev(priv->dev); - ipoib_dev_cleanup(priv->dev); + if (priv->pkey == pkey && + priv->child_type == IPOIB_LEGACY_CHILD) { + unregister_netdevice(priv->dev); list_del(&priv->list); - free_netdev(priv->dev); - - ret = 0; + dev = priv->dev; break; } } - mutex_unlock(&ppriv->vlan_mutex); + up_write(&ppriv->vlan_rwsem); + + rtnl_unlock(); + + if (dev) { + free_netdev(dev); + return 0; + } - return ret; + return -ENODEV; } |