diff options
Diffstat (limited to 'drivers/infiniband/core')
38 files changed, 2415 insertions, 845 deletions
diff --git a/drivers/infiniband/core/Makefile b/drivers/infiniband/core/Makefile index 8ab4eea5a0a5..f483e0c12444 100644 --- a/drivers/infiniband/core/Makefile +++ b/drivers/infiniband/core/Makefile @@ -33,12 +33,14 @@ ib_umad-y := user_mad.o ib_uverbs-y := uverbs_main.o uverbs_cmd.o uverbs_marshall.o \ rdma_core.o uverbs_std_types.o uverbs_ioctl.o \ uverbs_std_types_cq.o \ + uverbs_std_types_dmah.o \ uverbs_std_types_flow_action.o uverbs_std_types_dm.o \ uverbs_std_types_mr.o uverbs_std_types_counters.o \ uverbs_uapi.o uverbs_std_types_device.o \ uverbs_std_types_async_fd.o \ uverbs_std_types_srq.o \ uverbs_std_types_wq.o \ - uverbs_std_types_qp.o + uverbs_std_types_qp.o \ + ucaps.o ib_uverbs-$(CONFIG_INFINIBAND_USER_MEM) += umem.o umem_dmabuf.o ib_uverbs-$(CONFIG_INFINIBAND_ON_DEMAND_PAGING) += umem_odp.o diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c index be0743dac3ff..61596cda2b65 100644 --- a/drivers/infiniband/core/addr.c +++ b/drivers/infiniband/core/addr.c @@ -446,63 +446,41 @@ static int addr6_resolve(struct sockaddr *src_sock, } #endif +static bool is_dst_local(const struct dst_entry *dst) +{ + if (dst->ops->family == AF_INET) + return !!(dst_rtable(dst)->rt_type & RTN_LOCAL); + else if (dst->ops->family == AF_INET6) + return !!(dst_rt6_info(dst)->rt6i_flags & RTF_LOCAL); + else + return false; +} + static int addr_resolve_neigh(const struct dst_entry *dst, const struct sockaddr *dst_in, struct rdma_dev_addr *addr, - unsigned int ndev_flags, u32 seq) { - int ret = 0; - - if (ndev_flags & IFF_LOOPBACK) { + if (is_dst_local(dst)) { + /* When the destination is local entry, source and destination + * are same. Skip the neighbour lookup. + */ memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN); - } else { - if (!(ndev_flags & IFF_NOARP)) { - /* If the device doesn't do ARP internally */ - ret = fetch_ha(dst, addr, dst_in, seq); - } + return 0; } - return ret; -} - -static int copy_src_l2_addr(struct rdma_dev_addr *dev_addr, - const struct sockaddr *dst_in, - const struct dst_entry *dst, - const struct net_device *ndev) -{ - int ret = 0; - - if (dst->dev->flags & IFF_LOOPBACK) - ret = rdma_translate_ip(dst_in, dev_addr); - else - rdma_copy_src_l2_addr(dev_addr, dst->dev); - - /* - * If there's a gateway and type of device not ARPHRD_INFINIBAND, - * we're definitely in RoCE v2 (as RoCE v1 isn't routable) set the - * network type accordingly. - */ - if (has_gateway(dst, dst_in->sa_family) && - ndev->type != ARPHRD_INFINIBAND) - dev_addr->network = dst_in->sa_family == AF_INET ? - RDMA_NETWORK_IPV4 : - RDMA_NETWORK_IPV6; - else - dev_addr->network = RDMA_NETWORK_IB; - return ret; + return fetch_ha(dst, addr, dst_in, seq); } static int rdma_set_src_addr_rcu(struct rdma_dev_addr *dev_addr, - unsigned int *ndev_flags, const struct sockaddr *dst_in, const struct dst_entry *dst) { struct net_device *ndev = READ_ONCE(dst->dev); - *ndev_flags = ndev->flags; /* A physical device must be the RDMA device to use */ - if (ndev->flags & IFF_LOOPBACK) { + if (is_dst_local(dst)) { + int ret; /* * RDMA (IB/RoCE, iWarp) doesn't run on lo interface or * loopback IP address. So if route is resolved to loopback @@ -512,9 +490,27 @@ static int rdma_set_src_addr_rcu(struct rdma_dev_addr *dev_addr, ndev = rdma_find_ndev_for_src_ip_rcu(dev_net(ndev), dst_in); if (IS_ERR(ndev)) return -ENODEV; + ret = rdma_translate_ip(dst_in, dev_addr); + if (ret) + return ret; + } else { + rdma_copy_src_l2_addr(dev_addr, dst->dev); } - return copy_src_l2_addr(dev_addr, dst_in, dst, ndev); + /* + * If there's a gateway and type of device not ARPHRD_INFINIBAND, + * we're definitely in RoCE v2 (as RoCE v1 isn't routable) set the + * network type accordingly. + */ + if (has_gateway(dst, dst_in->sa_family) && + ndev->type != ARPHRD_INFINIBAND) + dev_addr->network = dst_in->sa_family == AF_INET ? + RDMA_NETWORK_IPV4 : + RDMA_NETWORK_IPV6; + else + dev_addr->network = RDMA_NETWORK_IB; + + return 0; } static int set_addr_netns_by_gid_rcu(struct rdma_dev_addr *addr) @@ -551,7 +547,6 @@ static int addr_resolve(struct sockaddr *src_in, u32 seq) { struct dst_entry *dst = NULL; - unsigned int ndev_flags = 0; struct rtable *rt = NULL; int ret; @@ -588,7 +583,7 @@ static int addr_resolve(struct sockaddr *src_in, rcu_read_unlock(); goto done; } - ret = rdma_set_src_addr_rcu(addr, &ndev_flags, dst_in, dst); + ret = rdma_set_src_addr_rcu(addr, dst_in, dst); rcu_read_unlock(); /* @@ -596,7 +591,7 @@ static int addr_resolve(struct sockaddr *src_in, * only if src addr translation didn't fail. */ if (!ret && resolve_neigh) - ret = addr_resolve_neigh(dst, dst_in, addr, ndev_flags, seq); + ret = addr_resolve_neigh(dst, dst_in, addr, seq); if (src_in->sa_family == AF_INET) ip_rt_put(rt); diff --git a/drivers/infiniband/core/agent.c b/drivers/infiniband/core/agent.c index 3bb46696731e..25a060a28301 100644 --- a/drivers/infiniband/core/agent.c +++ b/drivers/infiniband/core/agent.c @@ -110,8 +110,7 @@ void agent_send_response(const struct ib_mad_hdr *mad_hdr, const struct ib_grh * agent = port_priv->agent[qpn]; ah = ib_create_ah_from_wc(agent->qp->pd, wc, grh, port_num); if (IS_ERR(ah)) { - dev_err(&device->dev, "ib_create_ah_from_wc error %ld\n", - PTR_ERR(ah)); + dev_err(&device->dev, "ib_create_ah_from_wc error %pe\n", ah); return; } diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c index b7c078b7f7cf..81cf3c902e81 100644 --- a/drivers/infiniband/core/cache.c +++ b/drivers/infiniband/core/cache.c @@ -582,8 +582,8 @@ static int __ib_cache_gid_add(struct ib_device *ib_dev, u32 port, out_unlock: mutex_unlock(&table->lock); if (ret) - pr_warn("%s: unable to add gid %pI6 error=%d\n", - __func__, gid->raw, ret); + pr_warn_ratelimited("%s: unable to add gid %pI6 error=%d\n", + __func__, gid->raw, ret); return ret; } @@ -1127,41 +1127,6 @@ err: } EXPORT_SYMBOL(ib_find_cached_pkey); -int ib_find_exact_cached_pkey(struct ib_device *device, u32 port_num, - u16 pkey, u16 *index) -{ - struct ib_pkey_cache *cache; - unsigned long flags; - int i; - int ret = -ENOENT; - - if (!rdma_is_port_valid(device, port_num)) - return -EINVAL; - - read_lock_irqsave(&device->cache_lock, flags); - - cache = device->port_data[port_num].cache.pkey; - if (!cache) { - ret = -EINVAL; - goto err; - } - - *index = -1; - - for (i = 0; i < cache->table_len; ++i) - if (cache->table[i] == pkey) { - *index = i; - ret = 0; - break; - } - -err: - read_unlock_irqrestore(&device->cache_lock, flags); - - return ret; -} -EXPORT_SYMBOL(ib_find_exact_cached_pkey); - int ib_get_cached_lmc(struct ib_device *device, u32 port_num, u8 *lmc) { unsigned long flags; @@ -1536,6 +1501,12 @@ ib_cache_update(struct ib_device *device, u32 port, bool update_gids, device->port_data[port].cache.pkey = pkey_cache; } device->port_data[port].cache.lmc = tprops->lmc; + + if (device->port_data[port].cache.port_state != IB_PORT_NOP && + device->port_data[port].cache.port_state != tprops->state) + ibdev_info(device, "Port: %d Link %s\n", port, + ib_port_state_to_str(tprops->state)); + device->port_data[port].cache.port_state = tprops->state; device->port_data[port].cache.subnet_prefix = tprops->subnet_prefix; diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c index 142170473e75..024df6ee239d 100644 --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -34,8 +34,8 @@ MODULE_AUTHOR("Sean Hefty"); MODULE_DESCRIPTION("InfiniBand CM"); MODULE_LICENSE("Dual BSD/GPL"); -#define CM_DESTROY_ID_WAIT_TIMEOUT 10000 /* msecs */ #define CM_DIRECT_RETRY_CTX ((void *) 1UL) +#define CM_MRA_SETTING 24 /* 4.096us * 2^24 = ~68.7 seconds */ static const char * const ibcm_rej_reason_strs[] = { [IB_CM_REJ_NO_QP] = "no QP", @@ -160,6 +160,7 @@ struct cm_counter_attribute { struct cm_port { struct cm_device *cm_dev; struct ib_mad_agent *mad_agent; + struct ib_mad_agent *rep_agent; u32 port_num; atomic_long_t counters[CM_COUNTER_GROUPS][CM_ATTR_COUNT]; }; @@ -167,7 +168,7 @@ struct cm_port { struct cm_device { struct kref kref; struct list_head list; - spinlock_t mad_agent_lock; + rwlock_t mad_agent_lock; struct ib_device *ib_device; u8 ack_delay; int going_down; @@ -241,7 +242,6 @@ struct cm_id_private { u8 initiator_depth; u8 retry_count; u8 rnr_retry_count; - u8 service_timeout; u8 target_ack_delay; struct list_head work_list; @@ -274,7 +274,8 @@ static inline void cm_deref_id(struct cm_id_private *cm_id_priv) complete(&cm_id_priv->comp); } -static struct ib_mad_send_buf *cm_alloc_msg(struct cm_id_private *cm_id_priv) +static struct ib_mad_send_buf * +cm_alloc_msg_agent(struct cm_id_private *cm_id_priv, bool rep_agent) { struct ib_mad_agent *mad_agent; struct ib_mad_send_buf *m; @@ -285,8 +286,9 @@ static struct ib_mad_send_buf *cm_alloc_msg(struct cm_id_private *cm_id_priv) if (!cm_id_priv->av.port) return ERR_PTR(-EINVAL); - spin_lock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); - mad_agent = cm_id_priv->av.port->mad_agent; + read_lock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); + mad_agent = rep_agent ? cm_id_priv->av.port->rep_agent : + cm_id_priv->av.port->mad_agent; if (!mad_agent) { m = ERR_PTR(-EINVAL); goto out; @@ -311,10 +313,15 @@ static struct ib_mad_send_buf *cm_alloc_msg(struct cm_id_private *cm_id_priv) m->ah = ah; out: - spin_unlock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); + read_unlock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); return m; } +static struct ib_mad_send_buf *cm_alloc_msg(struct cm_id_private *cm_id_priv) +{ + return cm_alloc_msg_agent(cm_id_priv, false); +} + static void cm_free_msg(struct ib_mad_send_buf *msg) { if (msg->ah) @@ -323,13 +330,14 @@ static void cm_free_msg(struct ib_mad_send_buf *msg) } static struct ib_mad_send_buf * -cm_alloc_priv_msg(struct cm_id_private *cm_id_priv, enum ib_cm_state state) +cm_alloc_priv_msg_rep(struct cm_id_private *cm_id_priv, enum ib_cm_state state, + bool rep_agent) { struct ib_mad_send_buf *msg; lockdep_assert_held(&cm_id_priv->lock); - msg = cm_alloc_msg(cm_id_priv); + msg = cm_alloc_msg_agent(cm_id_priv, rep_agent); if (IS_ERR(msg)) return msg; @@ -344,6 +352,12 @@ cm_alloc_priv_msg(struct cm_id_private *cm_id_priv, enum ib_cm_state state) return msg; } +static struct ib_mad_send_buf * +cm_alloc_priv_msg(struct cm_id_private *cm_id_priv, enum ib_cm_state state) +{ + return cm_alloc_priv_msg_rep(cm_id_priv, state, false); +} + static void cm_free_priv_msg(struct ib_mad_send_buf *msg) { struct cm_id_private *cm_id_priv = msg->context[0]; @@ -1034,14 +1048,15 @@ static noinline void cm_destroy_id_wait_timeout(struct ib_cm_id *cm_id, struct cm_id_private *cm_id_priv; cm_id_priv = container_of(cm_id, struct cm_id_private, id); - pr_err("%s: cm_id=%p timed out. state %d -> %d, refcnt=%d\n", __func__, - cm_id, old_state, cm_id->state, refcount_read(&cm_id_priv->refcount)); + pr_err_ratelimited("%s: cm_id=%p timed out. state %d -> %d, refcnt=%d\n", __func__, + cm_id, old_state, cm_id->state, refcount_read(&cm_id_priv->refcount)); } static void cm_destroy_id(struct ib_cm_id *cm_id, int err) { struct cm_id_private *cm_id_priv; enum ib_cm_state old_state; + unsigned long timeout; struct cm_work *work; int ret; @@ -1152,10 +1167,9 @@ retest: xa_erase(&cm.local_id_table, cm_local_id(cm_id->local_id)); cm_deref_id(cm_id_priv); + timeout = msecs_to_jiffies((cm_id_priv->max_cm_retries * cm_id_priv->timeout_ms * 5) / 4); do { - ret = wait_for_completion_timeout(&cm_id_priv->comp, - msecs_to_jiffies( - CM_DESTROY_ID_WAIT_TIMEOUT)); + ret = wait_for_completion_timeout(&cm_id_priv->comp, timeout); if (!ret) /* timeout happened */ cm_destroy_id_wait_timeout(cm_id, old_state); } while (!ret); @@ -1297,10 +1311,10 @@ static __be64 cm_form_tid(struct cm_id_private *cm_id_priv) if (!cm_id_priv->av.port) return cpu_to_be64(low_tid); - spin_lock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); + read_lock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); if (cm_id_priv->av.port->mad_agent) hi_tid = ((u64)cm_id_priv->av.port->mad_agent->hi_tid) << 32; - spin_unlock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); + read_unlock(&cm_id_priv->av.port->cm_dev->mad_agent_lock); return cpu_to_be64(hi_tid | low_tid); } @@ -1872,7 +1886,7 @@ static void cm_process_work(struct cm_id_private *cm_id_priv, static void cm_format_mra(struct cm_mra_msg *mra_msg, struct cm_id_private *cm_id_priv, - enum cm_msg_response msg_mraed, u8 service_timeout, + enum cm_msg_response msg_mraed, const void *private_data, u8 private_data_len) { cm_format_mad_hdr(&mra_msg->hdr, CM_MRA_ATTR_ID, cm_id_priv->tid); @@ -1881,7 +1895,7 @@ static void cm_format_mra(struct cm_mra_msg *mra_msg, be32_to_cpu(cm_id_priv->id.local_id)); IBA_SET(CM_MRA_REMOTE_COMM_ID, mra_msg, be32_to_cpu(cm_id_priv->id.remote_id)); - IBA_SET(CM_MRA_SERVICE_TIMEOUT, mra_msg, service_timeout); + IBA_SET(CM_MRA_SERVICE_TIMEOUT, mra_msg, CM_MRA_SETTING); if (private_data && private_data_len) IBA_SET_MEM(CM_MRA_PRIVATE_DATA, mra_msg, private_data, @@ -1960,7 +1974,7 @@ static void cm_dup_req_handler(struct cm_work *work, switch (cm_id_priv->id.state) { case IB_CM_MRA_REQ_SENT: cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, - CM_MSG_RESPONSE_REQ, cm_id_priv->service_timeout, + CM_MSG_RESPONSE_REQ, cm_id_priv->private_data, cm_id_priv->private_data_len); break; @@ -2295,7 +2309,7 @@ int ib_send_cm_rep(struct ib_cm_id *cm_id, goto out; } - msg = cm_alloc_priv_msg(cm_id_priv, IB_CM_REP_SENT); + msg = cm_alloc_priv_msg_rep(cm_id_priv, IB_CM_REP_SENT, true); if (IS_ERR(msg)) { ret = PTR_ERR(msg); goto out; @@ -2454,7 +2468,7 @@ static void cm_dup_rep_handler(struct cm_work *work) cm_id_priv->private_data_len); else if (cm_id_priv->id.state == IB_CM_MRA_REP_SENT) cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, - CM_MSG_RESPONSE_REP, cm_id_priv->service_timeout, + CM_MSG_RESPONSE_REP, cm_id_priv->private_data, cm_id_priv->private_data_len); else @@ -3094,26 +3108,13 @@ out: return -EINVAL; } -int ib_send_cm_mra(struct ib_cm_id *cm_id, - u8 service_timeout, - const void *private_data, - u8 private_data_len) +int ib_prepare_cm_mra(struct ib_cm_id *cm_id) { struct cm_id_private *cm_id_priv; - struct ib_mad_send_buf *msg; enum ib_cm_state cm_state; enum ib_cm_lap_state lap_state; - enum cm_msg_response msg_response; - void *data; unsigned long flags; - int ret; - - if (private_data && private_data_len > IB_CM_MRA_PRIVATE_DATA_SIZE) - return -EINVAL; - - data = cm_copy_private_data(private_data, private_data_len); - if (IS_ERR(data)) - return PTR_ERR(data); + int ret = 0; cm_id_priv = container_of(cm_id, struct cm_id_private, id); @@ -3122,58 +3123,33 @@ int ib_send_cm_mra(struct ib_cm_id *cm_id, case IB_CM_REQ_RCVD: cm_state = IB_CM_MRA_REQ_SENT; lap_state = cm_id->lap_state; - msg_response = CM_MSG_RESPONSE_REQ; break; case IB_CM_REP_RCVD: cm_state = IB_CM_MRA_REP_SENT; lap_state = cm_id->lap_state; - msg_response = CM_MSG_RESPONSE_REP; break; case IB_CM_ESTABLISHED: if (cm_id->lap_state == IB_CM_LAP_RCVD) { cm_state = cm_id->state; lap_state = IB_CM_MRA_LAP_SENT; - msg_response = CM_MSG_RESPONSE_OTHER; break; } fallthrough; default: - trace_icm_send_mra_unknown_err(&cm_id_priv->id); + trace_icm_prepare_mra_unknown_err(&cm_id_priv->id); ret = -EINVAL; goto error_unlock; } - if (!(service_timeout & IB_CM_MRA_FLAG_DELAY)) { - msg = cm_alloc_msg(cm_id_priv); - if (IS_ERR(msg)) { - ret = PTR_ERR(msg); - goto error_unlock; - } - - cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, - msg_response, service_timeout, - private_data, private_data_len); - trace_icm_send_mra(cm_id); - ret = ib_post_send_mad(msg, NULL); - if (ret) - goto error_free_msg; - } - cm_id->state = cm_state; cm_id->lap_state = lap_state; - cm_id_priv->service_timeout = service_timeout; - cm_set_private_data(cm_id_priv, data, private_data_len); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); - return 0; + cm_set_private_data(cm_id_priv, NULL, 0); -error_free_msg: - cm_free_msg(msg); error_unlock: spin_unlock_irqrestore(&cm_id_priv->lock, flags); - kfree(data); return ret; } -EXPORT_SYMBOL(ib_send_cm_mra); +EXPORT_SYMBOL(ib_prepare_cm_mra); static struct cm_id_private *cm_acquire_mraed_id(struct cm_mra_msg *mra_msg) { @@ -3377,7 +3353,6 @@ static int cm_lap_handler(struct cm_work *work) cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, CM_MSG_RESPONSE_OTHER, - cm_id_priv->service_timeout, cm_id_priv->private_data, cm_id_priv->private_data_len); spin_unlock_irq(&cm_id_priv->lock); @@ -3786,7 +3761,8 @@ static void cm_process_send_error(struct cm_id_private *cm_id_priv, spin_lock_irq(&cm_id_priv->lock); if (msg != cm_id_priv->msg) { spin_unlock_irq(&cm_id_priv->lock); - cm_free_priv_msg(msg); + cm_free_msg(msg); + cm_deref_id(cm_id_priv); return; } cm_free_priv_msg(msg); @@ -4378,7 +4354,7 @@ static int cm_add_one(struct ib_device *ib_device) return -ENOMEM; kref_init(&cm_dev->kref); - spin_lock_init(&cm_dev->mad_agent_lock); + rwlock_init(&cm_dev->mad_agent_lock); cm_dev->ib_device = ib_device; cm_dev->ack_delay = ib_device->attrs.local_ca_ack_delay; cm_dev->going_down = 0; @@ -4418,9 +4394,22 @@ static int cm_add_one(struct ib_device *ib_device) goto error2; } + port->rep_agent = ib_register_mad_agent(ib_device, i, + IB_QPT_GSI, + NULL, + 0, + cm_send_handler, + NULL, + port, + 0); + if (IS_ERR(port->rep_agent)) { + ret = PTR_ERR(port->rep_agent); + goto error3; + } + ret = ib_modify_port(ib_device, i, 0, &port_modify); if (ret) - goto error3; + goto error4; count++; } @@ -4435,6 +4424,8 @@ static int cm_add_one(struct ib_device *ib_device) write_unlock_irqrestore(&cm.device_lock, flags); return 0; +error4: + ib_unregister_mad_agent(port->rep_agent); error3: ib_unregister_mad_agent(port->mad_agent); error2: @@ -4448,6 +4439,7 @@ error1: port = cm_dev->port[i-1]; ib_modify_port(ib_device, port->port_num, 0, &port_modify); + ib_unregister_mad_agent(port->rep_agent); ib_unregister_mad_agent(port->mad_agent); ib_port_unregister_client_groups(ib_device, i, cm_counter_groups); @@ -4477,12 +4469,14 @@ static void cm_remove_one(struct ib_device *ib_device, void *client_data) rdma_for_each_port (ib_device, i) { struct ib_mad_agent *mad_agent; + struct ib_mad_agent *rep_agent; if (!rdma_cap_ib_cm(ib_device, i)) continue; port = cm_dev->port[i-1]; mad_agent = port->mad_agent; + rep_agent = port->rep_agent; ib_modify_port(ib_device, port->port_num, 0, &port_modify); /* * We flush the queue here after the going_down set, this @@ -4494,10 +4488,12 @@ static void cm_remove_one(struct ib_device *ib_device, void *client_data) * The above ensures no call paths from the work are running, * the remaining paths all take the mad_agent_lock. */ - spin_lock(&cm_dev->mad_agent_lock); + write_lock(&cm_dev->mad_agent_lock); port->mad_agent = NULL; - spin_unlock(&cm_dev->mad_agent_lock); + port->rep_agent = NULL; + write_unlock(&cm_dev->mad_agent_lock); ib_unregister_mad_agent(mad_agent); + ib_unregister_mad_agent(rep_agent); ib_port_unregister_client_groups(ib_device, i, cm_counter_groups); } @@ -4521,7 +4517,7 @@ static int __init ib_cm_init(void) get_random_bytes(&cm.random_id_operand, sizeof cm.random_id_operand); INIT_LIST_HEAD(&cm.timewait_list); - cm.wq = alloc_workqueue("ib_cm", 0, 1); + cm.wq = alloc_workqueue("ib_cm", WQ_PERCPU, 1); if (!cm.wq) { ret = -ENOMEM; goto error2; diff --git a/drivers/infiniband/core/cm_trace.h b/drivers/infiniband/core/cm_trace.h index 944d9071245d..4a4987da69d4 100644 --- a/drivers/infiniband/core/cm_trace.h +++ b/drivers/infiniband/core/cm_trace.h @@ -229,7 +229,7 @@ DEFINE_CM_ERR_EVENT(send_drep); DEFINE_CM_ERR_EVENT(dreq_unknown); DEFINE_CM_ERR_EVENT(send_unknown_rej); DEFINE_CM_ERR_EVENT(rej_unknown); -DEFINE_CM_ERR_EVENT(send_mra_unknown); +DEFINE_CM_ERR_EVENT(prepare_mra_unknown); DEFINE_CM_ERR_EVENT(mra_unknown); DEFINE_CM_ERR_EVENT(qp_init); DEFINE_CM_ERR_EVENT(qp_rtr); diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 64ace0b968f0..95e89f5c147c 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -46,7 +46,6 @@ MODULE_LICENSE("Dual BSD/GPL"); #define CMA_CM_RESPONSE_TIMEOUT 20 #define CMA_MAX_CM_RETRIES 15 -#define CMA_CM_MRA_SETTING (IB_CM_MRA_FLAG_DELAY | 24) #define CMA_IBOE_PACKET_LIFETIME 16 #define CMA_PREFERRED_ROCE_GID_TYPE IB_GID_TYPE_ROCE_UDP_ENCAP @@ -72,6 +71,8 @@ static const char * const cma_events[] = { static void cma_iboe_set_mgid(struct sockaddr *addr, union ib_gid *mgid, enum ib_gid_type gid_type); +static void cma_netevent_work_handler(struct work_struct *_work); + const char *__attribute_const__ rdma_event_msg(enum rdma_cm_event_type event) { size_t index = event; @@ -144,19 +145,6 @@ struct iw_cm_id *rdma_iw_cm_id(struct rdma_cm_id *id) } EXPORT_SYMBOL(rdma_iw_cm_id); -/** - * rdma_res_to_id() - return the rdma_cm_id pointer for this restrack. - * @res: rdma resource tracking entry pointer - */ -struct rdma_cm_id *rdma_res_to_id(struct rdma_restrack_entry *res) -{ - struct rdma_id_private *id_priv = - container_of(res, struct rdma_id_private, res); - - return &id_priv->id; -} -EXPORT_SYMBOL(rdma_res_to_id); - static int cma_add_one(struct ib_device *device); static void cma_remove_one(struct ib_device *device, void *client_data); @@ -690,6 +678,7 @@ cma_validate_port(struct ib_device *device, u32 port, int bound_if_index = dev_addr->bound_dev_if; int dev_type = dev_addr->dev_type; struct net_device *ndev = NULL; + struct net_device *pdev = NULL; if (!rdma_dev_access_netns(device, id_priv->id.route.addr.dev_addr.net)) goto out; @@ -714,6 +703,21 @@ cma_validate_port(struct ib_device *device, u32 port, rcu_read_lock(); ndev = rcu_dereference(sgid_attr->ndev); + if (ndev->ifindex != bound_if_index) { + pdev = dev_get_by_index_rcu(dev_addr->net, bound_if_index); + if (pdev) { + if (is_vlan_dev(pdev)) { + pdev = vlan_dev_real_dev(pdev); + if (ndev->ifindex == pdev->ifindex) + bound_if_index = pdev->ifindex; + } + if (is_vlan_dev(ndev)) { + pdev = vlan_dev_real_dev(ndev); + if (bound_if_index == pdev->ifindex) + bound_if_index = ndev->ifindex; + } + } + } if (!net_eq(dev_net(ndev), dev_addr->net) || ndev->ifindex != bound_if_index) { rdma_put_gid_attr(sgid_attr); @@ -723,12 +727,26 @@ cma_validate_port(struct ib_device *device, u32 port, goto out; } - if (dev_type == ARPHRD_ETHER && rdma_protocol_roce(device, port)) { - ndev = dev_get_by_index(dev_addr->net, bound_if_index); - if (!ndev) - goto out; + /* + * For a RXE device, it should work with TUN device and normal ethernet + * devices. Use driver_id to check if a device is a RXE device or not. + * ARPHDR_NONE means a TUN device. + */ + if (device->ops.driver_id == RDMA_DRIVER_RXE) { + if ((dev_type == ARPHRD_NONE || dev_type == ARPHRD_ETHER) + && rdma_protocol_roce(device, port)) { + ndev = dev_get_by_index(dev_addr->net, bound_if_index); + if (!ndev) + goto out; + } } else { - gid_type = IB_GID_TYPE_IB; + if (dev_type == ARPHRD_ETHER && rdma_protocol_roce(device, port)) { + ndev = dev_get_by_index(dev_addr->net, bound_if_index); + if (!ndev) + goto out; + } else { + gid_type = IB_GID_TYPE_IB; + } } sgid_attr = rdma_find_gid_by_port(device, gid, gid_type, port, ndev); @@ -1017,6 +1035,7 @@ __rdma_create_id(struct net *net, rdma_cm_event_handler event_handler, get_random_bytes(&id_priv->seq_num, sizeof id_priv->seq_num); id_priv->id.route.addr.dev_addr.net = get_net(net); id_priv->seq_num &= 0x00ffffff; + INIT_WORK(&id_priv->id.net_work, cma_netevent_work_handler); rdma_restrack_new(&id_priv->res, RDMA_RESTRACK_CM_ID); if (parent) @@ -2057,6 +2076,7 @@ static void _destroy_id(struct rdma_id_private *id_priv, kfree(id_priv->id.route.path_rec); kfree(id_priv->id.route.path_rec_inbound); kfree(id_priv->id.route.path_rec_outbound); + kfree(id_priv->id.route.service_recs); put_net(id_priv->id.route.addr.dev_addr.net); kfree(id_priv); @@ -2181,8 +2201,8 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, case IB_CM_REP_RECEIVED: if (state == RDMA_CM_CONNECT && (id_priv->id.qp_type != IB_QPT_UD)) { - trace_cm_send_mra(id_priv); - ib_send_cm_mra(cm_id, CMA_CM_MRA_SETTING, NULL, 0); + trace_cm_prepare_mra(id_priv); + ib_prepare_cm_mra(cm_id); } if (id_priv->id.qp) { event.status = cma_rep_recv(id_priv); @@ -2443,8 +2463,8 @@ static int cma_ib_req_handler(struct ib_cm_id *cm_id, if (READ_ONCE(conn_id->state) == RDMA_CM_CONNECT && conn_id->id.qp_type != IB_QPT_UD) { - trace_cm_send_mra(cm_id->context); - ib_send_cm_mra(cm_id, CMA_CM_MRA_SETTING, NULL, 0); + trace_cm_prepare_mra(cm_id->context); + ib_prepare_cm_mra(cm_id); } mutex_unlock(&conn_id->handler_mutex); @@ -3363,13 +3383,18 @@ err1: int rdma_resolve_route(struct rdma_cm_id *id, unsigned long timeout_ms) { struct rdma_id_private *id_priv; + enum rdma_cm_state state; int ret; if (!timeout_ms) return -EINVAL; id_priv = container_of(id, struct rdma_id_private, id); - if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, RDMA_CM_ROUTE_QUERY)) + state = id_priv->state; + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, + RDMA_CM_ROUTE_QUERY) && + !cma_comp_exch(id_priv, RDMA_CM_ADDRINFO_RESOLVED, + RDMA_CM_ROUTE_QUERY)) return -EINVAL; cma_id_get(id_priv); @@ -3390,7 +3415,7 @@ int rdma_resolve_route(struct rdma_cm_id *id, unsigned long timeout_ms) return 0; err: - cma_comp_exch(id_priv, RDMA_CM_ROUTE_QUERY, RDMA_CM_ADDR_RESOLVED); + cma_comp_exch(id_priv, RDMA_CM_ROUTE_QUERY, state); cma_id_put(id_priv); return ret; } @@ -4450,6 +4475,8 @@ int rdma_connect_locked(struct rdma_cm_id *id, container_of(id, struct rdma_id_private, id); int ret; + lockdep_assert_held(&id_priv->handler_mutex); + if (!cma_comp_exch(id_priv, RDMA_CM_ROUTE_RESOLVED, RDMA_CM_CONNECT)) return -EINVAL; @@ -5211,9 +5238,9 @@ static int cma_netevent_callback(struct notifier_block *self, if (!memcmp(current_id->id.route.addr.dev_addr.dst_dev_addr, neigh->ha, ETH_ALEN)) continue; - INIT_WORK(¤t_id->id.net_work, cma_netevent_work_handler); cma_id_get(current_id); - queue_work(cma_wq, ¤t_id->id.net_work); + if (!queue_work(cma_wq, ¤t_id->id.net_work)) + cma_id_put(current_id); } out: spin_unlock_irqrestore(&id_table_lock, flags); @@ -5487,3 +5514,129 @@ static void __exit cma_cleanup(void) module_init(cma_init); module_exit(cma_cleanup); + +static void cma_query_ib_service_handler(int status, + struct sa_service_rec *recs, + unsigned int num_recs, void *context) +{ + struct cma_work *work = context; + struct rdma_id_private *id_priv = work->id; + struct sockaddr_ib *addr; + + if (status) + goto fail; + + if (!num_recs) { + status = -ENOENT; + goto fail; + } + + if (id_priv->id.route.service_recs) { + status = -EALREADY; + goto fail; + } + + id_priv->id.route.service_recs = + kmalloc_array(num_recs, sizeof(*recs), GFP_KERNEL); + if (!id_priv->id.route.service_recs) { + status = -ENOMEM; + goto fail; + } + + id_priv->id.route.num_service_recs = num_recs; + memcpy(id_priv->id.route.service_recs, recs, sizeof(*recs) * num_recs); + + addr = (struct sockaddr_ib *)&id_priv->id.route.addr.dst_addr; + addr->sib_family = AF_IB; + addr->sib_addr = *(struct ib_addr *)&recs->gid; + addr->sib_pkey = recs->pkey; + addr->sib_sid = recs->id; + rdma_addr_set_dgid(&id_priv->id.route.addr.dev_addr, + (union ib_gid *)&addr->sib_addr); + ib_addr_set_pkey(&id_priv->id.route.addr.dev_addr, + ntohs(addr->sib_pkey)); + + queue_work(cma_wq, &work->work); + return; + +fail: + work->old_state = RDMA_CM_ADDRINFO_QUERY; + work->new_state = RDMA_CM_ADDR_BOUND; + work->event.event = RDMA_CM_EVENT_ADDRINFO_ERROR; + work->event.status = status; + pr_debug_ratelimited( + "RDMA CM: SERVICE_ERROR: failed to query service record. status %d\n", + status); + queue_work(cma_wq, &work->work); +} + +static int cma_resolve_ib_service(struct rdma_id_private *id_priv, + struct rdma_ucm_ib_service *ibs) +{ + struct sa_service_rec sr = {}; + ib_sa_comp_mask mask = 0; + struct cma_work *work; + + work = kzalloc(sizeof(*work), GFP_KERNEL); + if (!work) + return -ENOMEM; + + cma_id_get(id_priv); + + work->id = id_priv; + INIT_WORK(&work->work, cma_work_handler); + work->old_state = RDMA_CM_ADDRINFO_QUERY; + work->new_state = RDMA_CM_ADDRINFO_RESOLVED; + work->event.event = RDMA_CM_EVENT_ADDRINFO_RESOLVED; + + if (ibs->flags & RDMA_USER_CM_IB_SERVICE_FLAG_ID) { + sr.id = cpu_to_be64(ibs->service_id); + mask |= IB_SA_SERVICE_REC_SERVICE_ID; + } + if (ibs->flags & RDMA_USER_CM_IB_SERVICE_FLAG_NAME) { + strscpy(sr.name, ibs->service_name, sizeof(sr.name)); + mask |= IB_SA_SERVICE_REC_SERVICE_NAME; + } + + id_priv->query_id = ib_sa_service_rec_get(&sa_client, + id_priv->id.device, + id_priv->id.port_num, + &sr, mask, + 2000, GFP_KERNEL, + cma_query_ib_service_handler, + work, &id_priv->query); + + if (id_priv->query_id < 0) { + cma_id_put(id_priv); + kfree(work); + return id_priv->query_id; + } + + return 0; +} + +int rdma_resolve_ib_service(struct rdma_cm_id *id, + struct rdma_ucm_ib_service *ibs) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!id_priv->cma_dev || + !cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_ADDRINFO_QUERY)) + return -EINVAL; + + if (rdma_cap_ib_sa(id->device, id->port_num)) + ret = cma_resolve_ib_service(id_priv, ibs); + else + ret = -EOPNOTSUPP; + + if (ret) + goto err; + + return 0; +err: + cma_comp_exch(id_priv, RDMA_CM_ADDRINFO_QUERY, RDMA_CM_ADDR_BOUND); + return ret; +} +EXPORT_SYMBOL(rdma_resolve_ib_service); diff --git a/drivers/infiniband/core/cma_priv.h b/drivers/infiniband/core/cma_priv.h index b7354c94cf1b..c604b601f4d9 100644 --- a/drivers/infiniband/core/cma_priv.h +++ b/drivers/infiniband/core/cma_priv.h @@ -47,7 +47,9 @@ enum rdma_cm_state { RDMA_CM_ADDR_BOUND, RDMA_CM_LISTEN, RDMA_CM_DEVICE_REMOVAL, - RDMA_CM_DESTROYING + RDMA_CM_DESTROYING, + RDMA_CM_ADDRINFO_QUERY, + RDMA_CM_ADDRINFO_RESOLVED }; struct rdma_id_private { diff --git a/drivers/infiniband/core/cma_trace.h b/drivers/infiniband/core/cma_trace.h index dc622f3778be..3456d5f3aa47 100644 --- a/drivers/infiniband/core/cma_trace.h +++ b/drivers/infiniband/core/cma_trace.h @@ -55,7 +55,7 @@ DECLARE_EVENT_CLASS(cma_fsm_class, DEFINE_CMA_FSM_EVENT(send_rtu); DEFINE_CMA_FSM_EVENT(send_rej); -DEFINE_CMA_FSM_EVENT(send_mra); +DEFINE_CMA_FSM_EVENT(prepare_mra); DEFINE_CMA_FSM_EVENT(send_sidr_req); DEFINE_CMA_FSM_EVENT(send_sidr_rep); DEFINE_CMA_FSM_EVENT(disconnect); diff --git a/drivers/infiniband/core/counters.c b/drivers/infiniband/core/counters.c index af59486fe418..c3aa6d7fc66b 100644 --- a/drivers/infiniband/core/counters.c +++ b/drivers/infiniband/core/counters.c @@ -12,7 +12,8 @@ static int __counter_set_mode(struct rdma_port_counter *port_counter, enum rdma_nl_counter_mode new_mode, - enum rdma_nl_counter_mask new_mask) + enum rdma_nl_counter_mask new_mask, + bool bind_opcnt) { if (new_mode == RDMA_COUNTER_MODE_AUTO) { if (new_mask & (~ALL_AUTO_MODE_MASKS)) @@ -23,6 +24,7 @@ static int __counter_set_mode(struct rdma_port_counter *port_counter, port_counter->mode.mode = new_mode; port_counter->mode.mask = new_mask; + port_counter->mode.bind_opcnt = bind_opcnt; return 0; } @@ -41,6 +43,7 @@ static int __counter_set_mode(struct rdma_port_counter *port_counter, */ int rdma_counter_set_auto_mode(struct ib_device *dev, u32 port, enum rdma_nl_counter_mask mask, + bool bind_opcnt, struct netlink_ext_ack *extack) { struct rdma_port_counter *port_counter; @@ -59,12 +62,13 @@ int rdma_counter_set_auto_mode(struct ib_device *dev, u32 port, RDMA_COUNTER_MODE_NONE; if (port_counter->mode.mode == mode && - port_counter->mode.mask == mask) { + port_counter->mode.mask == mask && + port_counter->mode.bind_opcnt == bind_opcnt) { ret = 0; goto out; } - ret = __counter_set_mode(port_counter, mode, mask); + ret = __counter_set_mode(port_counter, mode, mask, bind_opcnt); out: mutex_unlock(&port_counter->lock); @@ -89,7 +93,7 @@ static void auto_mode_init_counter(struct rdma_counter *counter, } static int __rdma_counter_bind_qp(struct rdma_counter *counter, - struct ib_qp *qp) + struct ib_qp *qp, u32 port) { int ret; @@ -100,7 +104,7 @@ static int __rdma_counter_bind_qp(struct rdma_counter *counter, return -EOPNOTSUPP; mutex_lock(&counter->lock); - ret = qp->device->ops.counter_bind_qp(counter, qp); + ret = qp->device->ops.counter_bind_qp(counter, qp, port); mutex_unlock(&counter->lock); return ret; @@ -140,7 +144,8 @@ out: static struct rdma_counter *alloc_and_bind(struct ib_device *dev, u32 port, struct ib_qp *qp, - enum rdma_nl_counter_mode mode) + enum rdma_nl_counter_mode mode, + bool bind_opcnt) { struct rdma_port_counter *port_counter; struct rdma_counter *counter; @@ -149,13 +154,15 @@ static struct rdma_counter *alloc_and_bind(struct ib_device *dev, u32 port, if (!dev->ops.counter_dealloc || !dev->ops.counter_alloc_stats) return NULL; - counter = kzalloc(sizeof(*counter), GFP_KERNEL); + counter = rdma_zalloc_drv_obj(dev, rdma_counter); if (!counter) return NULL; counter->device = dev; counter->port = port; + dev->ops.counter_init(counter); + rdma_restrack_new(&counter->res, RDMA_RESTRACK_COUNTER); counter->stats = dev->ops.counter_alloc_stats(counter); if (!counter->stats) @@ -166,7 +173,7 @@ static struct rdma_counter *alloc_and_bind(struct ib_device *dev, u32 port, switch (mode) { case RDMA_COUNTER_MODE_MANUAL: ret = __counter_set_mode(port_counter, RDMA_COUNTER_MODE_MANUAL, - 0); + 0, bind_opcnt); if (ret) { mutex_unlock(&port_counter->lock); goto err_mode; @@ -185,10 +192,11 @@ static struct rdma_counter *alloc_and_bind(struct ib_device *dev, u32 port, mutex_unlock(&port_counter->lock); counter->mode.mode = mode; + counter->mode.bind_opcnt = bind_opcnt; kref_init(&counter->kref); mutex_init(&counter->lock); - ret = __rdma_counter_bind_qp(counter, qp); + ret = __rdma_counter_bind_qp(counter, qp, port); if (ret) goto err_mode; @@ -213,7 +221,8 @@ static void rdma_counter_free(struct rdma_counter *counter) port_counter->num_counters--; if (!port_counter->num_counters && (port_counter->mode.mode == RDMA_COUNTER_MODE_MANUAL)) - __counter_set_mode(port_counter, RDMA_COUNTER_MODE_NONE, 0); + __counter_set_mode(port_counter, RDMA_COUNTER_MODE_NONE, 0, + false); mutex_unlock(&port_counter->lock); @@ -238,7 +247,7 @@ static bool auto_mode_match(struct ib_qp *qp, struct rdma_counter *counter, return match; } -static int __rdma_counter_unbind_qp(struct ib_qp *qp) +static int __rdma_counter_unbind_qp(struct ib_qp *qp, u32 port) { struct rdma_counter *counter = qp->counter; int ret; @@ -247,7 +256,7 @@ static int __rdma_counter_unbind_qp(struct ib_qp *qp) return -EOPNOTSUPP; mutex_lock(&counter->lock); - ret = qp->device->ops.counter_unbind_qp(qp); + ret = qp->device->ops.counter_unbind_qp(qp, port); mutex_unlock(&counter->lock); return ret; @@ -339,13 +348,14 @@ int rdma_counter_bind_qp_auto(struct ib_qp *qp, u32 port) counter = rdma_get_counter_auto_mode(qp, port); if (counter) { - ret = __rdma_counter_bind_qp(counter, qp); + ret = __rdma_counter_bind_qp(counter, qp, port); if (ret) { kref_put(&counter->kref, counter_release); return ret; } } else { - counter = alloc_and_bind(dev, port, qp, RDMA_COUNTER_MODE_AUTO); + counter = alloc_and_bind(dev, port, qp, RDMA_COUNTER_MODE_AUTO, + port_counter->mode.bind_opcnt); if (!counter) return -ENOMEM; } @@ -358,7 +368,7 @@ int rdma_counter_bind_qp_auto(struct ib_qp *qp, u32 port) * @force: * true - Decrease the counter ref-count anyway (e.g., qp destroy) */ -int rdma_counter_unbind_qp(struct ib_qp *qp, bool force) +int rdma_counter_unbind_qp(struct ib_qp *qp, u32 port, bool force) { struct rdma_counter *counter = qp->counter; int ret; @@ -366,7 +376,7 @@ int rdma_counter_unbind_qp(struct ib_qp *qp, bool force) if (!counter) return -EINVAL; - ret = __rdma_counter_unbind_qp(qp); + ret = __rdma_counter_unbind_qp(qp, port); if (ret && !force) return ret; @@ -451,7 +461,7 @@ static struct ib_qp *rdma_counter_get_qp(struct ib_device *dev, u32 qp_num) return NULL; qp = container_of(res, struct ib_qp, res); - if (qp->qp_type == IB_QPT_RAW_PACKET && !capable(CAP_NET_RAW)) + if (qp->qp_type == IB_QPT_RAW_PACKET && !rdma_dev_has_raw_cap(dev)) goto err; return qp; @@ -513,7 +523,7 @@ int rdma_counter_bind_qpn(struct ib_device *dev, u32 port, goto err_task; } - ret = __rdma_counter_bind_qp(counter, qp); + ret = __rdma_counter_bind_qp(counter, qp, port); if (ret) goto err_task; @@ -558,7 +568,7 @@ int rdma_counter_bind_qpn_alloc(struct ib_device *dev, u32 port, goto err; } - counter = alloc_and_bind(dev, port, qp, RDMA_COUNTER_MODE_MANUAL); + counter = alloc_and_bind(dev, port, qp, RDMA_COUNTER_MODE_MANUAL, true); if (!counter) { ret = -ENOMEM; goto err; @@ -604,7 +614,7 @@ int rdma_counter_unbind_qpn(struct ib_device *dev, u32 port, goto out; } - ret = rdma_counter_unbind_qp(qp, false); + ret = rdma_counter_unbind_qp(qp, port, false); out: rdma_restrack_put(&qp->res); @@ -613,13 +623,15 @@ out: int rdma_counter_get_mode(struct ib_device *dev, u32 port, enum rdma_nl_counter_mode *mode, - enum rdma_nl_counter_mask *mask) + enum rdma_nl_counter_mask *mask, + bool *opcnt) { struct rdma_port_counter *port_counter; port_counter = &dev->port_data[port].port_counter; *mode = port_counter->mode.mode; *mask = port_counter->mode.mask; + *opcnt = port_counter->mode.bind_opcnt; return 0; } diff --git a/drivers/infiniband/core/cq.c b/drivers/infiniband/core/cq.c index a70876a0a231..584537c71545 100644 --- a/drivers/infiniband/core/cq.c +++ b/drivers/infiniband/core/cq.c @@ -317,13 +317,18 @@ EXPORT_SYMBOL(__ib_alloc_cq_any); */ void ib_free_cq(struct ib_cq *cq) { - int ret; + int ret = 0; if (WARN_ON_ONCE(atomic_read(&cq->usecnt))) return; if (WARN_ON_ONCE(cq->cqe_used)) return; + if (cq->device->ops.pre_destroy_cq) { + ret = cq->device->ops.pre_destroy_cq(cq); + WARN_ONCE(ret, "Disable of kernel CQ shouldn't fail"); + } + switch (cq->poll_ctx) { case IB_POLL_DIRECT: break; @@ -340,7 +345,10 @@ void ib_free_cq(struct ib_cq *cq) rdma_dim_destroy(cq); trace_cq_free(cq); - ret = cq->device->ops.destroy_cq(cq, NULL); + if (cq->device->ops.post_destroy_cq) + cq->device->ops.post_destroy_cq(cq); + else + ret = cq->device->ops.destroy_cq(cq, NULL); WARN_ONCE(ret, "Destroy of kernel CQ shouldn't fail"); rdma_restrack_del(&cq->res); kfree(cq->wc); diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c index ca9b956c034d..13e8a1714bbd 100644 --- a/drivers/infiniband/core/device.c +++ b/drivers/infiniband/core/device.c @@ -145,6 +145,33 @@ bool rdma_dev_access_netns(const struct ib_device *dev, const struct net *net) } EXPORT_SYMBOL(rdma_dev_access_netns); +/** + * rdma_dev_has_raw_cap() - Returns whether a specified rdma device has + * CAP_NET_RAW capability or not. + * + * @dev: Pointer to rdma device whose capability to be checked + * + * Returns true if a rdma device's owning user namespace has CAP_NET_RAW + * capability, otherwise false. When rdma subsystem is in legacy shared network, + * namespace mode, the default net namespace is considered. + */ +bool rdma_dev_has_raw_cap(const struct ib_device *dev) +{ + const struct net *net; + + /* Network namespace is the resource whose user namespace + * to be considered. When in shared mode, there is no reliable + * network namespace resource, so consider the default net namespace. + */ + if (ib_devices_shared_netns) + net = &init_net; + else + net = read_pnet(&dev->coredev.rdma_net); + + return ns_capable(net->user_ns, CAP_NET_RAW); +} +EXPORT_SYMBOL(rdma_dev_has_raw_cap); + /* * xarray has this behavior where it won't iterate over NULL values stored in * allocated arrays. So we need our own iterator to see all values stored in @@ -209,23 +236,6 @@ static void __ibdev_printk(const char *level, const struct ib_device *ibdev, printk("%s(NULL ib_device): %pV", level, vaf); } -void ibdev_printk(const char *level, const struct ib_device *ibdev, - const char *format, ...) -{ - struct va_format vaf; - va_list args; - - va_start(args, format); - - vaf.fmt = format; - vaf.va = &args; - - __ibdev_printk(level, ibdev, &vaf); - - va_end(args); -} -EXPORT_SYMBOL(ibdev_printk); - #define define_ibdev_printk_level(func, level) \ void func(const struct ib_device *ibdev, const char *fmt, ...) \ { \ @@ -545,6 +555,8 @@ static struct class ib_class = { static void rdma_init_coredev(struct ib_core_device *coredev, struct ib_device *dev, struct net *net) { + bool is_full_dev = &dev->coredev == coredev; + /* This BUILD_BUG_ON is intended to catch layout change * of union of ib_core_device and device. * dev must be the first element as ib_core and providers @@ -556,6 +568,13 @@ static void rdma_init_coredev(struct ib_core_device *coredev, coredev->dev.class = &ib_class; coredev->dev.groups = dev->groups; + + /* + * Don't expose hw counters outside of the init namespace. + */ + if (!is_full_dev && dev->hw_stats_attr_index) + coredev->dev.groups[dev->hw_stats_attr_index] = NULL; + device_initialize(&coredev->dev); coredev->owner = dev; INIT_LIST_HEAD(&coredev->port_list); @@ -565,6 +584,8 @@ static void rdma_init_coredev(struct ib_core_device *coredev, /** * _ib_alloc_device - allocate an IB device struct * @size:size of structure to allocate + * @net: network namespace device should be located in, namespace + * must stay valid until ib_register_device() is completed. * * Low-level drivers should use ib_alloc_device() to allocate &struct * ib_device. @size is the size of the structure to be allocated, @@ -572,7 +593,7 @@ static void rdma_init_coredev(struct ib_core_device *coredev, * ib_dealloc_device() must be used to free structures allocated with * ib_alloc_device(). */ -struct ib_device *_ib_alloc_device(size_t size) +struct ib_device *_ib_alloc_device(size_t size, struct net *net) { struct ib_device *device; unsigned int i; @@ -589,7 +610,15 @@ struct ib_device *_ib_alloc_device(size_t size) return NULL; } - rdma_init_coredev(&device->coredev, device, &init_net); + /* ib_devices_shared_netns can't change while we have active namespaces + * in the system which means either init_net is passed or the user has + * no idea what they are doing. + * + * To avoid breaking backward compatibility, when in shared mode, + * force to init the device in the init_net. + */ + net = ib_devices_shared_netns ? &init_net : net; + rdma_init_coredev(&device->coredev, device, net); INIT_LIST_HEAD(&device->event_handler_list); spin_lock_init(&device->qp_open_list_lock); @@ -1358,9 +1387,14 @@ static void ib_device_notify_register(struct ib_device *device) u32 port; int ret; + down_read(&devices_rwsem); + + /* Mark for userspace that device is ready */ + kobject_uevent(&device->dev.kobj, KOBJ_ADD); + ret = rdma_nl_notify_event(device, 0, RDMA_REGISTER_EVENT); if (ret) - return; + goto out; rdma_for_each_port(device, port) { netdev = ib_device_get_netdev(device, port); @@ -1371,8 +1405,11 @@ static void ib_device_notify_register(struct ib_device *device) RDMA_NETDEV_ATTACH_EVENT); dev_put(netdev); if (ret) - return; + goto out; } + +out: + up_read(&devices_rwsem); } /** @@ -1471,10 +1508,9 @@ int ib_register_device(struct ib_device *device, const char *name, return ret; } dev_set_uevent_suppress(&device->dev, false); - /* Mark for userspace that device is ready */ - kobject_uevent(&device->dev.kobj, KOBJ_ADD); ib_device_notify_register(device); + ib_device_put(device); return 0; @@ -1507,7 +1543,7 @@ static void __ib_unregister_device(struct ib_device *ib_dev) /* * We have a registration lock so that all the calls to unregister are - * fully fenced, once any unregister returns the device is truely + * fully fenced, once any unregister returns the device is truly * unregistered even if multiple callers are unregistering it at the * same time. This also interacts with the registration flow and * provides sane semantics if register and unregister are racing. @@ -2296,6 +2332,33 @@ struct net_device *ib_device_get_netdev(struct ib_device *ib_dev, EXPORT_SYMBOL(ib_device_get_netdev); /** + * ib_query_netdev_port - Query the port number of a net_device + * associated with an ibdev + * @ibdev: IB device + * @ndev: Network device + * @port: IB port the net_device is connected to + */ +int ib_query_netdev_port(struct ib_device *ibdev, struct net_device *ndev, + u32 *port) +{ + struct net_device *ib_ndev; + u32 port_num; + + rdma_for_each_port(ibdev, port_num) { + ib_ndev = ib_device_get_netdev(ibdev, port_num); + if (ndev == ib_ndev) { + *port = port_num; + dev_put(ib_ndev); + return 0; + } + dev_put(ib_ndev); + } + + return -ENOENT; +} +EXPORT_SYMBOL(ib_query_netdev_port); + +/** * ib_device_get_by_netdev - Find an IB device associated with a netdev * @ndev: netdev to locate * @driver_id: The driver ID that must match (RDMA_DRIVER_UNKNOWN matches all) @@ -2645,6 +2708,7 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, add_sub_dev); SET_DEVICE_OP(dev_ops, advise_mr); SET_DEVICE_OP(dev_ops, alloc_dm); + SET_DEVICE_OP(dev_ops, alloc_dmah); SET_DEVICE_OP(dev_ops, alloc_hw_device_stats); SET_DEVICE_OP(dev_ops, alloc_hw_port_stats); SET_DEVICE_OP(dev_ops, alloc_mr); @@ -2659,11 +2723,13 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, counter_alloc_stats); SET_DEVICE_OP(dev_ops, counter_bind_qp); SET_DEVICE_OP(dev_ops, counter_dealloc); + SET_DEVICE_OP(dev_ops, counter_init); SET_DEVICE_OP(dev_ops, counter_unbind_qp); SET_DEVICE_OP(dev_ops, counter_update_stats); SET_DEVICE_OP(dev_ops, create_ah); SET_DEVICE_OP(dev_ops, create_counters); SET_DEVICE_OP(dev_ops, create_cq); + SET_DEVICE_OP(dev_ops, create_cq_umem); SET_DEVICE_OP(dev_ops, create_flow); SET_DEVICE_OP(dev_ops, create_qp); SET_DEVICE_OP(dev_ops, create_rwq_ind_table); @@ -2671,6 +2737,7 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, create_user_ah); SET_DEVICE_OP(dev_ops, create_wq); SET_DEVICE_OP(dev_ops, dealloc_dm); + SET_DEVICE_OP(dev_ops, dealloc_dmah); SET_DEVICE_OP(dev_ops, dealloc_driver); SET_DEVICE_OP(dev_ops, dealloc_mw); SET_DEVICE_OP(dev_ops, dealloc_pd); @@ -2736,8 +2803,10 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, modify_srq); SET_DEVICE_OP(dev_ops, modify_wq); SET_DEVICE_OP(dev_ops, peek_cq); + SET_DEVICE_OP(dev_ops, pre_destroy_cq); SET_DEVICE_OP(dev_ops, poll_cq); SET_DEVICE_OP(dev_ops, port_groups); + SET_DEVICE_OP(dev_ops, post_destroy_cq); SET_DEVICE_OP(dev_ops, post_recv); SET_DEVICE_OP(dev_ops, post_send); SET_DEVICE_OP(dev_ops, post_srq_recv); @@ -2761,10 +2830,12 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, set_vf_guid); SET_DEVICE_OP(dev_ops, set_vf_link_state); SET_DEVICE_OP(dev_ops, ufile_hw_cleanup); + SET_DEVICE_OP(dev_ops, report_port_event); SET_OBJ_SIZE(dev_ops, ib_ah); SET_OBJ_SIZE(dev_ops, ib_counters); SET_OBJ_SIZE(dev_ops, ib_cq); + SET_OBJ_SIZE(dev_ops, ib_dmah); SET_OBJ_SIZE(dev_ops, ib_mw); SET_OBJ_SIZE(dev_ops, ib_pd); SET_OBJ_SIZE(dev_ops, ib_qp); @@ -2772,6 +2843,7 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_OBJ_SIZE(dev_ops, ib_srq); SET_OBJ_SIZE(dev_ops, ib_ucontext); SET_OBJ_SIZE(dev_ops, ib_xrcd); + SET_OBJ_SIZE(dev_ops, rdma_counter); } EXPORT_SYMBOL(ib_set_device_ops); @@ -2854,11 +2926,62 @@ static const struct rdma_nl_cbs ibnl_ls_cb_table[RDMA_NL_LS_NUM_OPS] = { }, }; +void ib_dispatch_port_state_event(struct ib_device *ibdev, struct net_device *ndev) +{ + enum ib_port_state curr_state; + struct ib_event ibevent = {}; + u32 port; + + if (ib_query_netdev_port(ibdev, ndev, &port)) + return; + + curr_state = ib_get_curr_port_state(ndev); + + write_lock_irq(&ibdev->cache_lock); + if (ibdev->port_data[port].cache.last_port_state == curr_state) { + write_unlock_irq(&ibdev->cache_lock); + return; + } + ibdev->port_data[port].cache.last_port_state = curr_state; + write_unlock_irq(&ibdev->cache_lock); + + ibevent.event = (curr_state == IB_PORT_DOWN) ? + IB_EVENT_PORT_ERR : IB_EVENT_PORT_ACTIVE; + ibevent.device = ibdev; + ibevent.element.port_num = port; + ib_dispatch_event(&ibevent); +} +EXPORT_SYMBOL(ib_dispatch_port_state_event); + +static void handle_port_event(struct net_device *ndev, unsigned long event) +{ + struct ib_device *ibdev; + + /* Currently, link events in bonding scenarios are still + * reported by drivers that support bonding. + */ + if (netif_is_lag_master(ndev) || netif_is_lag_port(ndev)) + return; + + ibdev = ib_device_get_by_netdev(ndev, RDMA_DRIVER_UNKNOWN); + if (!ibdev) + return; + + if (ibdev->ops.report_port_event) { + ibdev->ops.report_port_event(ibdev, ndev, event); + goto put_ibdev; + } + + ib_dispatch_port_state_event(ibdev, ndev); + +put_ibdev: + ib_device_put(ibdev); +}; + static int ib_netdevice_event(struct notifier_block *this, unsigned long event, void *ptr) { struct net_device *ndev = netdev_notifier_info_to_dev(ptr); - struct net_device *ib_ndev; struct ib_device *ibdev; u32 port; @@ -2868,15 +2991,21 @@ static int ib_netdevice_event(struct notifier_block *this, if (!ibdev) return NOTIFY_DONE; - rdma_for_each_port(ibdev, port) { - ib_ndev = ib_device_get_netdev(ibdev, port); - if (ndev == ib_ndev) - rdma_nl_notify_event(ibdev, port, - RDMA_NETDEV_RENAME_EVENT); - dev_put(ib_ndev); + if (ib_query_netdev_port(ibdev, ndev, &port)) { + ib_device_put(ibdev); + break; } + + rdma_nl_notify_event(ibdev, port, RDMA_NETDEV_RENAME_EVENT); ib_device_put(ibdev); break; + + case NETDEV_UP: + case NETDEV_CHANGE: + case NETDEV_DOWN: + handle_port_event(ndev, event); + break; + default: break; } @@ -2892,7 +3021,7 @@ static int __init ib_core_init(void) { int ret = -ENOMEM; - ib_wq = alloc_workqueue("infiniband", 0, 0); + ib_wq = alloc_workqueue("infiniband", WQ_PERCPU, 0); if (!ib_wq) return -ENOMEM; @@ -2902,7 +3031,7 @@ static int __init ib_core_init(void) goto err; ib_comp_wq = alloc_workqueue("ib-comp-wq", - WQ_HIGHPRI | WQ_MEM_RECLAIM | WQ_SYSFS, 0); + WQ_HIGHPRI | WQ_MEM_RECLAIM | WQ_SYSFS | WQ_PERCPU, 0); if (!ib_comp_wq) goto err_unbound; diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c index 7e3a55349e10..62410578dec3 100644 --- a/drivers/infiniband/core/iwcm.c +++ b/drivers/infiniband/core/iwcm.c @@ -109,7 +109,9 @@ static struct ctl_table iwcm_ctl_table[] = { .data = &default_backlog, .maxlen = sizeof(default_backlog), .mode = 0644, - .proc_handler = proc_dointvec, + .proc_handler = proc_dointvec_minmax, + .extra1 = SYSCTL_ZERO, + .extra2 = SYSCTL_INT_MAX, }, }; @@ -366,12 +368,9 @@ EXPORT_SYMBOL(iw_cm_disconnect); /* * CM_ID <-- DESTROYING * - * Clean up all resources associated with the connection and release - * the initial reference taken by iw_create_cm_id. - * - * Returns true if and only if the last cm_id_priv reference has been dropped. + * Clean up all resources associated with the connection. */ -static bool destroy_cm_id(struct iw_cm_id *cm_id) +static void destroy_cm_id(struct iw_cm_id *cm_id) { struct iwcm_id_private *cm_id_priv; struct ib_qp *qp; @@ -440,20 +439,22 @@ static bool destroy_cm_id(struct iw_cm_id *cm_id) iwpm_remove_mapinfo(&cm_id->local_addr, &cm_id->m_local_addr); iwpm_remove_mapping(&cm_id->local_addr, RDMA_NL_IWCM); } - - return iwcm_deref_id(cm_id_priv); } /* - * This function is only called by the application thread and cannot - * be called by the event thread. The function will wait for all - * references to be released on the cm_id and then kfree the cm_id - * object. + * Destroy cm_id. If the cm_id still has other references, wait for all + * references to be released on the cm_id and then release the initial + * reference taken by iw_create_cm_id. */ void iw_destroy_cm_id(struct iw_cm_id *cm_id) { - if (!destroy_cm_id(cm_id)) + struct iwcm_id_private *cm_id_priv; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + destroy_cm_id(cm_id); + if (refcount_read(&cm_id_priv->refcount) > 1) flush_workqueue(iwcm_wq); + iwcm_deref_id(cm_id_priv); } EXPORT_SYMBOL(iw_destroy_cm_id); @@ -1033,8 +1034,10 @@ static void cm_work_handler(struct work_struct *_work) if (!test_bit(IWCM_F_DROP_EVENTS, &cm_id_priv->flags)) { ret = process_event(cm_id_priv, &levent); - if (ret) - WARN_ON_ONCE(destroy_cm_id(&cm_id_priv->id)); + if (ret) { + destroy_cm_id(&cm_id_priv->id); + WARN_ON_ONCE(iwcm_deref_id(cm_id_priv)); + } } else pr_debug("dropping event %d\n", levent.event); if (iwcm_deref_id(cm_id_priv)) diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index 1fd54d5c4dd8..8f26bfb69586 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -210,6 +210,29 @@ int ib_response_mad(const struct ib_mad_hdr *hdr) } EXPORT_SYMBOL(ib_response_mad); +#define SOL_FC_MAX_DEFAULT_FRAC 4 +#define SOL_FC_MAX_SA_FRAC 32 + +static int get_sol_fc_max_outstanding(struct ib_mad_reg_req *mad_reg_req) +{ + if (!mad_reg_req) + /* Send only agent */ + return mad_recvq_size / SOL_FC_MAX_DEFAULT_FRAC; + + switch (mad_reg_req->mgmt_class) { + case IB_MGMT_CLASS_CM: + return mad_recvq_size / SOL_FC_MAX_DEFAULT_FRAC; + case IB_MGMT_CLASS_SUBN_ADM: + return mad_recvq_size / SOL_FC_MAX_SA_FRAC; + case IB_MGMT_CLASS_SUBN_LID_ROUTED: + case IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE: + return min(mad_recvq_size, IB_MAD_QP_RECV_SIZE) / + SOL_FC_MAX_DEFAULT_FRAC; + default: + return 0; + } +} + /* * ib_register_mad_agent - Register to send/receive MADs * @@ -391,13 +414,17 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, spin_lock_init(&mad_agent_priv->lock); INIT_LIST_HEAD(&mad_agent_priv->send_list); INIT_LIST_HEAD(&mad_agent_priv->wait_list); - INIT_LIST_HEAD(&mad_agent_priv->done_list); INIT_LIST_HEAD(&mad_agent_priv->rmpp_list); + INIT_LIST_HEAD(&mad_agent_priv->backlog_list); INIT_DELAYED_WORK(&mad_agent_priv->timed_work, timeout_sends); INIT_LIST_HEAD(&mad_agent_priv->local_list); INIT_WORK(&mad_agent_priv->local_work, local_completions); refcount_set(&mad_agent_priv->refcount, 1); init_completion(&mad_agent_priv->comp); + mad_agent_priv->sol_fc_send_count = 0; + mad_agent_priv->sol_fc_wait_count = 0; + mad_agent_priv->sol_fc_max = + recv_handler ? get_sol_fc_max_outstanding(mad_reg_req) : 0; ret2 = ib_mad_agent_security_setup(&mad_agent_priv->agent, qp_type); if (ret2) { @@ -1055,6 +1082,180 @@ int ib_send_mad(struct ib_mad_send_wr_private *mad_send_wr) return ret; } +static void handle_queued_state(struct ib_mad_send_wr_private *mad_send_wr, + struct ib_mad_agent_private *mad_agent_priv) +{ + if (mad_send_wr->state == IB_MAD_STATE_WAIT_RESP) { + mad_agent_priv->sol_fc_wait_count--; + list_move_tail(&mad_send_wr->agent_list, + &mad_agent_priv->backlog_list); + } else { + expect_mad_state(mad_send_wr, IB_MAD_STATE_INIT); + list_add_tail(&mad_send_wr->agent_list, + &mad_agent_priv->backlog_list); + } +} + +static void handle_send_state(struct ib_mad_send_wr_private *mad_send_wr, + struct ib_mad_agent_private *mad_agent_priv) +{ + if (mad_send_wr->state == IB_MAD_STATE_INIT) { + list_add_tail(&mad_send_wr->agent_list, + &mad_agent_priv->send_list); + } else { + expect_mad_state2(mad_send_wr, IB_MAD_STATE_WAIT_RESP, + IB_MAD_STATE_QUEUED); + list_move_tail(&mad_send_wr->agent_list, + &mad_agent_priv->send_list); + } + + if (mad_send_wr->is_solicited_fc) { + if (mad_send_wr->state == IB_MAD_STATE_WAIT_RESP) + mad_agent_priv->sol_fc_wait_count--; + mad_agent_priv->sol_fc_send_count++; + } +} + +static void handle_wait_state(struct ib_mad_send_wr_private *mad_send_wr, + struct ib_mad_agent_private *mad_agent_priv) +{ + struct ib_mad_send_wr_private *temp_mad_send_wr; + struct list_head *list_item; + unsigned long delay; + + expect_mad_state3(mad_send_wr, IB_MAD_STATE_SEND_START, + IB_MAD_STATE_WAIT_RESP, IB_MAD_STATE_CANCELED); + if (mad_send_wr->state == IB_MAD_STATE_SEND_START && + mad_send_wr->is_solicited_fc) { + mad_agent_priv->sol_fc_send_count--; + mad_agent_priv->sol_fc_wait_count++; + } + + list_del_init(&mad_send_wr->agent_list); + delay = mad_send_wr->timeout; + mad_send_wr->timeout += jiffies; + + if (delay) { + list_for_each_prev(list_item, + &mad_agent_priv->wait_list) { + temp_mad_send_wr = list_entry( + list_item, + struct ib_mad_send_wr_private, + agent_list); + if (time_after(mad_send_wr->timeout, + temp_mad_send_wr->timeout)) + break; + } + } else { + list_item = &mad_agent_priv->wait_list; + } + + list_add(&mad_send_wr->agent_list, list_item); +} + +static void handle_early_resp_state(struct ib_mad_send_wr_private *mad_send_wr, + struct ib_mad_agent_private *mad_agent_priv) +{ + expect_mad_state(mad_send_wr, IB_MAD_STATE_SEND_START); + mad_agent_priv->sol_fc_send_count -= mad_send_wr->is_solicited_fc; +} + +static void handle_canceled_state(struct ib_mad_send_wr_private *mad_send_wr, + struct ib_mad_agent_private *mad_agent_priv) +{ + not_expect_mad_state(mad_send_wr, IB_MAD_STATE_DONE); + if (mad_send_wr->is_solicited_fc) { + if (mad_send_wr->state == IB_MAD_STATE_SEND_START) + mad_agent_priv->sol_fc_send_count--; + else if (mad_send_wr->state == IB_MAD_STATE_WAIT_RESP) + mad_agent_priv->sol_fc_wait_count--; + } +} + +static void handle_done_state(struct ib_mad_send_wr_private *mad_send_wr, + struct ib_mad_agent_private *mad_agent_priv) +{ + if (mad_send_wr->is_solicited_fc) { + if (mad_send_wr->state == IB_MAD_STATE_SEND_START) + mad_agent_priv->sol_fc_send_count--; + else if (mad_send_wr->state == IB_MAD_STATE_WAIT_RESP) + mad_agent_priv->sol_fc_wait_count--; + } + + list_del_init(&mad_send_wr->agent_list); +} + +void change_mad_state(struct ib_mad_send_wr_private *mad_send_wr, + enum ib_mad_state new_state) +{ + struct ib_mad_agent_private *mad_agent_priv = + mad_send_wr->mad_agent_priv; + + switch (new_state) { + case IB_MAD_STATE_INIT: + break; + case IB_MAD_STATE_QUEUED: + handle_queued_state(mad_send_wr, mad_agent_priv); + break; + case IB_MAD_STATE_SEND_START: + handle_send_state(mad_send_wr, mad_agent_priv); + break; + case IB_MAD_STATE_WAIT_RESP: + handle_wait_state(mad_send_wr, mad_agent_priv); + if (mad_send_wr->state == IB_MAD_STATE_CANCELED) + return; + break; + case IB_MAD_STATE_EARLY_RESP: + handle_early_resp_state(mad_send_wr, mad_agent_priv); + break; + case IB_MAD_STATE_CANCELED: + handle_canceled_state(mad_send_wr, mad_agent_priv); + break; + case IB_MAD_STATE_DONE: + handle_done_state(mad_send_wr, mad_agent_priv); + break; + } + + mad_send_wr->state = new_state; +} + +static bool is_solicited_fc_mad(struct ib_mad_send_wr_private *mad_send_wr) +{ + struct ib_rmpp_mad *rmpp_mad; + u8 mgmt_class; + + if (!mad_send_wr->timeout) + return 0; + + rmpp_mad = mad_send_wr->send_buf.mad; + if (mad_send_wr->mad_agent_priv->agent.rmpp_version && + (ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE)) + return 0; + + mgmt_class = + ((struct ib_mad_hdr *)mad_send_wr->send_buf.mad)->mgmt_class; + return mgmt_class == IB_MGMT_CLASS_CM || + mgmt_class == IB_MGMT_CLASS_SUBN_ADM || + mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED || + mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE; +} + +static bool mad_is_for_backlog(struct ib_mad_send_wr_private *mad_send_wr) +{ + struct ib_mad_agent_private *mad_agent_priv = + mad_send_wr->mad_agent_priv; + + if (!mad_send_wr->is_solicited_fc || !mad_agent_priv->sol_fc_max) + return false; + + if (!list_empty(&mad_agent_priv->backlog_list)) + return true; + + return mad_agent_priv->sol_fc_send_count + + mad_agent_priv->sol_fc_wait_count >= + mad_agent_priv->sol_fc_max; +} + /* * ib_post_send_mad - Posts MAD(s) to the send queue of the QP associated * with the registered client @@ -1080,9 +1281,7 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf, if (ret) goto error; - if (!send_buf->mad_agent->send_handler || - (send_buf->timeout_ms && - !send_buf->mad_agent->recv_handler)) { + if (!send_buf->mad_agent->send_handler) { ret = -EINVAL; goto error; } @@ -1118,15 +1317,19 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf, mad_send_wr->max_retries = send_buf->retries; mad_send_wr->retries_left = send_buf->retries; send_buf->retries = 0; - /* Reference for work request to QP + response */ - mad_send_wr->refcount = 1 + (mad_send_wr->timeout > 0); - mad_send_wr->status = IB_WC_SUCCESS; + change_mad_state(mad_send_wr, IB_MAD_STATE_INIT); /* Reference MAD agent until send completes */ refcount_inc(&mad_agent_priv->refcount); spin_lock_irqsave(&mad_agent_priv->lock, flags); - list_add_tail(&mad_send_wr->agent_list, - &mad_agent_priv->send_list); + mad_send_wr->is_solicited_fc = is_solicited_fc_mad(mad_send_wr); + if (mad_is_for_backlog(mad_send_wr)) { + change_mad_state(mad_send_wr, IB_MAD_STATE_QUEUED); + spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + return 0; + } + + change_mad_state(mad_send_wr, IB_MAD_STATE_SEND_START); spin_unlock_irqrestore(&mad_agent_priv->lock, flags); if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) { @@ -1138,7 +1341,7 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf, if (ret < 0) { /* Fail send request */ spin_lock_irqsave(&mad_agent_priv->lock, flags); - list_del(&mad_send_wr->agent_list); + change_mad_state(mad_send_wr, IB_MAD_STATE_DONE); spin_unlock_irqrestore(&mad_agent_priv->lock, flags); deref_mad_agent(mad_agent_priv); goto error; @@ -1746,7 +1949,19 @@ ib_find_send_mad(const struct ib_mad_agent_private *mad_agent_priv, */ (is_direct(mad_hdr->mgmt_class) || rcv_has_same_gid(mad_agent_priv, wr, wc))) - return (wr->status == IB_WC_SUCCESS) ? wr : NULL; + return (wr->state != IB_MAD_STATE_CANCELED) ? wr : NULL; + } + + list_for_each_entry(wr, &mad_agent_priv->backlog_list, agent_list) { + if ((wr->tid == mad_hdr->tid) && + rcv_has_same_class(wr, wc) && + /* + * Don't check GID for direct routed MADs. + * These might have permissive LIDs. + */ + (is_direct(mad_hdr->mgmt_class) || + rcv_has_same_gid(mad_agent_priv, wr, wc))) + return (wr->state != IB_MAD_STATE_CANCELED) ? wr : NULL; } /* @@ -1765,17 +1980,55 @@ ib_find_send_mad(const struct ib_mad_agent_private *mad_agent_priv, (is_direct(mad_hdr->mgmt_class) || rcv_has_same_gid(mad_agent_priv, wr, wc))) /* Verify request has not been canceled */ - return (wr->status == IB_WC_SUCCESS) ? wr : NULL; + return (wr->state != IB_MAD_STATE_CANCELED) ? wr : NULL; } return NULL; } +static void +process_backlog_mads(struct ib_mad_agent_private *mad_agent_priv) +{ + struct ib_mad_send_wr_private *mad_send_wr; + struct ib_mad_send_wc mad_send_wc = {}; + unsigned long flags; + int ret; + + spin_lock_irqsave(&mad_agent_priv->lock, flags); + while (!list_empty(&mad_agent_priv->backlog_list) && + (mad_agent_priv->sol_fc_send_count + + mad_agent_priv->sol_fc_wait_count < + mad_agent_priv->sol_fc_max)) { + mad_send_wr = list_entry(mad_agent_priv->backlog_list.next, + struct ib_mad_send_wr_private, + agent_list); + change_mad_state(mad_send_wr, IB_MAD_STATE_SEND_START); + spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + ret = ib_send_mad(mad_send_wr); + if (ret) { + spin_lock_irqsave(&mad_agent_priv->lock, flags); + deref_mad_agent(mad_agent_priv); + change_mad_state(mad_send_wr, IB_MAD_STATE_DONE); + spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + mad_send_wc.send_buf = &mad_send_wr->send_buf; + mad_send_wc.status = IB_WC_LOC_QP_OP_ERR; + mad_agent_priv->agent.send_handler( + &mad_agent_priv->agent, &mad_send_wc); + } + + spin_lock_irqsave(&mad_agent_priv->lock, flags); + } + + spin_unlock_irqrestore(&mad_agent_priv->lock, flags); +} + void ib_mark_mad_done(struct ib_mad_send_wr_private *mad_send_wr) { mad_send_wr->timeout = 0; - if (mad_send_wr->refcount == 1) - list_move_tail(&mad_send_wr->agent_list, - &mad_send_wr->mad_agent_priv->done_list); + if (mad_send_wr->state == IB_MAD_STATE_WAIT_RESP || + mad_send_wr->state == IB_MAD_STATE_QUEUED) + change_mad_state(mad_send_wr, IB_MAD_STATE_DONE); + else + change_mad_state(mad_send_wr, IB_MAD_STATE_EARLY_RESP); } static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, @@ -1784,6 +2037,7 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, struct ib_mad_send_wr_private *mad_send_wr; struct ib_mad_send_wc mad_send_wc; unsigned long flags; + bool is_mad_done; int ret; INIT_LIST_HEAD(&mad_recv_wc->rmpp_list); @@ -1832,6 +2086,7 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, } } else { ib_mark_mad_done(mad_send_wr); + is_mad_done = (mad_send_wr->state == IB_MAD_STATE_DONE); spin_unlock_irqrestore(&mad_agent_priv->lock, flags); /* Defined behavior is to complete response before request */ @@ -1841,10 +2096,13 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, mad_recv_wc); deref_mad_agent(mad_agent_priv); - mad_send_wc.status = IB_WC_SUCCESS; - mad_send_wc.vendor_err = 0; - mad_send_wc.send_buf = &mad_send_wr->send_buf; - ib_mad_complete_send_wr(mad_send_wr, &mad_send_wc); + if (is_mad_done) { + mad_send_wc.status = IB_WC_SUCCESS; + mad_send_wc.vendor_err = 0; + mad_send_wc.send_buf = &mad_send_wr->send_buf; + ib_mad_complete_send_wr(mad_send_wr, + &mad_send_wc); + } } } else { mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent, NULL, @@ -2172,30 +2430,11 @@ static void adjust_timeout(struct ib_mad_agent_private *mad_agent_priv) static void wait_for_response(struct ib_mad_send_wr_private *mad_send_wr) { struct ib_mad_agent_private *mad_agent_priv; - struct ib_mad_send_wr_private *temp_mad_send_wr; - struct list_head *list_item; unsigned long delay; mad_agent_priv = mad_send_wr->mad_agent_priv; - list_del(&mad_send_wr->agent_list); - delay = mad_send_wr->timeout; - mad_send_wr->timeout += jiffies; - - if (delay) { - list_for_each_prev(list_item, &mad_agent_priv->wait_list) { - temp_mad_send_wr = list_entry(list_item, - struct ib_mad_send_wr_private, - agent_list); - if (time_after(mad_send_wr->timeout, - temp_mad_send_wr->timeout)) - break; - } - } else { - list_item = &mad_agent_priv->wait_list; - } - - list_add(&mad_send_wr->agent_list, list_item); + change_mad_state(mad_send_wr, IB_MAD_STATE_WAIT_RESP); /* Reschedule a work item if we have a shorter timeout */ if (mad_agent_priv->wait_list.next == &mad_send_wr->agent_list) @@ -2229,32 +2468,28 @@ void ib_mad_complete_send_wr(struct ib_mad_send_wr_private *mad_send_wr, } else ret = IB_RMPP_RESULT_UNHANDLED; - if (mad_send_wc->status != IB_WC_SUCCESS && - mad_send_wr->status == IB_WC_SUCCESS) { - mad_send_wr->status = mad_send_wc->status; - mad_send_wr->refcount -= (mad_send_wr->timeout > 0); - } - - if (--mad_send_wr->refcount > 0) { - if (mad_send_wr->refcount == 1 && mad_send_wr->timeout && - mad_send_wr->status == IB_WC_SUCCESS) { - wait_for_response(mad_send_wr); - } + if (mad_send_wr->state == IB_MAD_STATE_CANCELED) + mad_send_wc->status = IB_WC_WR_FLUSH_ERR; + else if (mad_send_wr->state == IB_MAD_STATE_SEND_START && + mad_send_wr->timeout) { + wait_for_response(mad_send_wr); goto done; } /* Remove send from MAD agent and notify client of completion */ - list_del(&mad_send_wr->agent_list); + if (mad_send_wr->state != IB_MAD_STATE_DONE) + change_mad_state(mad_send_wr, IB_MAD_STATE_DONE); adjust_timeout(mad_agent_priv); spin_unlock_irqrestore(&mad_agent_priv->lock, flags); - if (mad_send_wr->status != IB_WC_SUCCESS) - mad_send_wc->status = mad_send_wr->status; - if (ret == IB_RMPP_RESULT_INTERNAL) + if (ret == IB_RMPP_RESULT_INTERNAL) { ib_rmpp_send_handler(mad_send_wc); - else + } else { + if (mad_send_wr->is_solicited_fc) + process_backlog_mads(mad_agent_priv); mad_agent_priv->agent.send_handler(&mad_agent_priv->agent, mad_send_wc); + } /* Release reference on agent taken when sending */ deref_mad_agent(mad_agent_priv); @@ -2396,40 +2631,53 @@ static bool ib_mad_send_error(struct ib_mad_port_private *port_priv, return true; } +static void clear_mad_error_list(struct list_head *list, + enum ib_wc_status wc_status, + struct ib_mad_agent_private *mad_agent_priv) +{ + struct ib_mad_send_wr_private *mad_send_wr, *n; + struct ib_mad_send_wc mad_send_wc; + + mad_send_wc.status = wc_status; + mad_send_wc.vendor_err = 0; + + list_for_each_entry_safe(mad_send_wr, n, list, agent_list) { + mad_send_wc.send_buf = &mad_send_wr->send_buf; + mad_agent_priv->agent.send_handler(&mad_agent_priv->agent, + &mad_send_wc); + deref_mad_agent(mad_agent_priv); + } +} + static void cancel_mads(struct ib_mad_agent_private *mad_agent_priv) { unsigned long flags; struct ib_mad_send_wr_private *mad_send_wr, *temp_mad_send_wr; - struct ib_mad_send_wc mad_send_wc; struct list_head cancel_list; INIT_LIST_HEAD(&cancel_list); spin_lock_irqsave(&mad_agent_priv->lock, flags); list_for_each_entry_safe(mad_send_wr, temp_mad_send_wr, - &mad_agent_priv->send_list, agent_list) { - if (mad_send_wr->status == IB_WC_SUCCESS) { - mad_send_wr->status = IB_WC_WR_FLUSH_ERR; - mad_send_wr->refcount -= (mad_send_wr->timeout > 0); - } - } - - /* Empty wait list to prevent receives from finding a request */ - list_splice_init(&mad_agent_priv->wait_list, &cancel_list); - spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + &mad_agent_priv->send_list, agent_list) + change_mad_state(mad_send_wr, IB_MAD_STATE_CANCELED); - /* Report all cancelled requests */ - mad_send_wc.status = IB_WC_WR_FLUSH_ERR; - mad_send_wc.vendor_err = 0; + /* Empty wait & backlog list to prevent receives from finding request */ + list_for_each_entry_safe(mad_send_wr, temp_mad_send_wr, + &mad_agent_priv->wait_list, agent_list) { + change_mad_state(mad_send_wr, IB_MAD_STATE_DONE); + list_add_tail(&mad_send_wr->agent_list, &cancel_list); + } list_for_each_entry_safe(mad_send_wr, temp_mad_send_wr, - &cancel_list, agent_list) { - mad_send_wc.send_buf = &mad_send_wr->send_buf; - list_del(&mad_send_wr->agent_list); - mad_agent_priv->agent.send_handler(&mad_agent_priv->agent, - &mad_send_wc); - deref_mad_agent(mad_agent_priv); + &mad_agent_priv->backlog_list, agent_list) { + change_mad_state(mad_send_wr, IB_MAD_STATE_DONE); + list_add_tail(&mad_send_wr->agent_list, &cancel_list); } + + spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + /* Report all cancelled requests */ + clear_mad_error_list(&cancel_list, IB_WC_WR_FLUSH_ERR, mad_agent_priv); } static struct ib_mad_send_wr_private* @@ -2451,6 +2699,13 @@ find_send_wr(struct ib_mad_agent_private *mad_agent_priv, &mad_send_wr->send_buf == send_buf) return mad_send_wr; } + + list_for_each_entry(mad_send_wr, &mad_agent_priv->backlog_list, + agent_list) { + if (&mad_send_wr->send_buf == send_buf) + return mad_send_wr; + } + return NULL; } @@ -2468,16 +2723,16 @@ int ib_modify_mad(struct ib_mad_send_buf *send_buf, u32 timeout_ms) struct ib_mad_agent_private, agent); spin_lock_irqsave(&mad_agent_priv->lock, flags); mad_send_wr = find_send_wr(mad_agent_priv, send_buf); - if (!mad_send_wr || mad_send_wr->status != IB_WC_SUCCESS) { + if (!mad_send_wr || mad_send_wr->state == IB_MAD_STATE_CANCELED) { spin_unlock_irqrestore(&mad_agent_priv->lock, flags); return -EINVAL; } - active = (!mad_send_wr->timeout || mad_send_wr->refcount > 1); - if (!timeout_ms) { - mad_send_wr->status = IB_WC_WR_FLUSH_ERR; - mad_send_wr->refcount -= (mad_send_wr->timeout > 0); - } + active = ((mad_send_wr->state == IB_MAD_STATE_SEND_START) || + (mad_send_wr->state == IB_MAD_STATE_EARLY_RESP) || + (mad_send_wr->state == IB_MAD_STATE_QUEUED && timeout_ms)); + if (!timeout_ms) + change_mad_state(mad_send_wr, IB_MAD_STATE_CANCELED); mad_send_wr->send_buf.timeout_ms = timeout_ms; if (active) @@ -2589,6 +2844,11 @@ static int retry_send(struct ib_mad_send_wr_private *mad_send_wr) mad_send_wr->send_buf.retries++; mad_send_wr->timeout = msecs_to_jiffies(mad_send_wr->send_buf.timeout_ms); + if (mad_send_wr->is_solicited_fc && + !list_empty(&mad_send_wr->mad_agent_priv->backlog_list)) { + change_mad_state(mad_send_wr, IB_MAD_STATE_QUEUED); + return 0; + } if (ib_mad_kernel_rmpp_agent(&mad_send_wr->mad_agent_priv->agent)) { ret = ib_retry_rmpp(mad_send_wr); @@ -2606,26 +2866,25 @@ static int retry_send(struct ib_mad_send_wr_private *mad_send_wr) } else ret = ib_send_mad(mad_send_wr); - if (!ret) { - mad_send_wr->refcount++; - list_add_tail(&mad_send_wr->agent_list, - &mad_send_wr->mad_agent_priv->send_list); - } + if (!ret) + change_mad_state(mad_send_wr, IB_MAD_STATE_SEND_START); + return ret; } static void timeout_sends(struct work_struct *work) { - struct ib_mad_send_wr_private *mad_send_wr, *n; + struct ib_mad_send_wr_private *mad_send_wr; struct ib_mad_agent_private *mad_agent_priv; - struct ib_mad_send_wc mad_send_wc; - struct list_head local_list; + struct list_head timeout_list; + struct list_head cancel_list; + struct list_head *list_item; unsigned long flags, delay; mad_agent_priv = container_of(work, struct ib_mad_agent_private, timed_work.work); - mad_send_wc.vendor_err = 0; - INIT_LIST_HEAD(&local_list); + INIT_LIST_HEAD(&timeout_list); + INIT_LIST_HEAD(&cancel_list); spin_lock_irqsave(&mad_agent_priv->lock, flags); while (!list_empty(&mad_agent_priv->wait_list)) { @@ -2643,25 +2902,22 @@ static void timeout_sends(struct work_struct *work) break; } - list_del_init(&mad_send_wr->agent_list); - if (mad_send_wr->status == IB_WC_SUCCESS && - !retry_send(mad_send_wr)) + if (mad_send_wr->state == IB_MAD_STATE_CANCELED) + list_item = &cancel_list; + else if (retry_send(mad_send_wr)) + list_item = &timeout_list; + else continue; - list_add_tail(&mad_send_wr->agent_list, &local_list); + change_mad_state(mad_send_wr, IB_MAD_STATE_DONE); + list_add_tail(&mad_send_wr->agent_list, list_item); } - spin_unlock_irqrestore(&mad_agent_priv->lock, flags); - list_for_each_entry_safe(mad_send_wr, n, &local_list, agent_list) { - if (mad_send_wr->status == IB_WC_SUCCESS) - mad_send_wc.status = IB_WC_RESP_TIMEOUT_ERR; - else - mad_send_wc.status = mad_send_wr->status; - mad_send_wc.send_buf = &mad_send_wr->send_buf; - mad_agent_priv->agent.send_handler(&mad_agent_priv->agent, - &mad_send_wc); - deref_mad_agent(mad_agent_priv); - } + spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + process_backlog_mads(mad_agent_priv); + clear_mad_error_list(&timeout_list, IB_WC_RESP_TIMEOUT_ERR, + mad_agent_priv); + clear_mad_error_list(&cancel_list, IB_WC_WR_FLUSH_ERR, mad_agent_priv); } /* @@ -2671,11 +2927,11 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info, struct ib_mad_private *mad) { unsigned long flags; - int post, ret; struct ib_mad_private *mad_priv; struct ib_sge sg_list; struct ib_recv_wr recv_wr; struct ib_mad_queue *recv_queue = &qp_info->recv_queue; + int ret = 0; /* Initialize common scatter list fields */ sg_list.lkey = qp_info->port_priv->pd->local_dma_lkey; @@ -2685,7 +2941,7 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info, recv_wr.sg_list = &sg_list; recv_wr.num_sge = 1; - do { + while (true) { /* Allocate and map receive buffer */ if (mad) { mad_priv = mad; @@ -2693,10 +2949,8 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info, } else { mad_priv = alloc_mad_private(port_mad_size(qp_info->port_priv), GFP_ATOMIC); - if (!mad_priv) { - ret = -ENOMEM; - break; - } + if (!mad_priv) + return -ENOMEM; } sg_list.length = mad_priv_dma_size(mad_priv); sg_list.addr = ib_dma_map_single(qp_info->port_priv->device, @@ -2705,37 +2959,41 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info, DMA_FROM_DEVICE); if (unlikely(ib_dma_mapping_error(qp_info->port_priv->device, sg_list.addr))) { - kfree(mad_priv); ret = -ENOMEM; - break; + goto free_mad_priv; } mad_priv->header.mapping = sg_list.addr; mad_priv->header.mad_list.mad_queue = recv_queue; mad_priv->header.mad_list.cqe.done = ib_mad_recv_done; recv_wr.wr_cqe = &mad_priv->header.mad_list.cqe; - - /* Post receive WR */ spin_lock_irqsave(&recv_queue->lock, flags); - post = (++recv_queue->count < recv_queue->max_active); - list_add_tail(&mad_priv->header.mad_list.list, &recv_queue->list); + if (recv_queue->count >= recv_queue->max_active) { + /* Fully populated the receive queue */ + spin_unlock_irqrestore(&recv_queue->lock, flags); + break; + } + recv_queue->count++; + list_add_tail(&mad_priv->header.mad_list.list, + &recv_queue->list); spin_unlock_irqrestore(&recv_queue->lock, flags); + ret = ib_post_recv(qp_info->qp, &recv_wr, NULL); if (ret) { spin_lock_irqsave(&recv_queue->lock, flags); list_del(&mad_priv->header.mad_list.list); recv_queue->count--; spin_unlock_irqrestore(&recv_queue->lock, flags); - ib_dma_unmap_single(qp_info->port_priv->device, - mad_priv->header.mapping, - mad_priv_dma_size(mad_priv), - DMA_FROM_DEVICE); - kfree(mad_priv); dev_err(&qp_info->port_priv->device->dev, "ib_post_recv failed: %d\n", ret); break; } - } while (post); + } + ib_dma_unmap_single(qp_info->port_priv->device, + mad_priv->header.mapping, + mad_priv_dma_size(mad_priv), DMA_FROM_DEVICE); +free_mad_priv: + kfree(mad_priv); return ret; } diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h index 1b7445a6f671..f444357d33f4 100644 --- a/drivers/infiniband/core/mad_priv.h +++ b/drivers/infiniband/core/mad_priv.h @@ -95,13 +95,16 @@ struct ib_mad_agent_private { spinlock_t lock; struct list_head send_list; + unsigned int sol_fc_send_count; struct list_head wait_list; - struct list_head done_list; + unsigned int sol_fc_wait_count; struct delayed_work timed_work; unsigned long timeout; struct list_head local_list; struct work_struct local_work; struct list_head rmpp_list; + unsigned int sol_fc_max; + struct list_head backlog_list; refcount_t refcount; union { @@ -118,6 +121,32 @@ struct ib_mad_snoop_private { struct completion comp; }; +enum ib_mad_state { + /* MAD is in the making and is not yet in any list */ + IB_MAD_STATE_INIT, + /* MAD is in backlog list */ + IB_MAD_STATE_QUEUED, + /* + * MAD was sent to the QP and is waiting for completion + * notification in send list. + */ + IB_MAD_STATE_SEND_START, + /* + * MAD send completed successfully, waiting for a response + * in wait list. + */ + IB_MAD_STATE_WAIT_RESP, + /* + * Response came early, before send completion notification, + * in send list. + */ + IB_MAD_STATE_EARLY_RESP, + /* MAD was canceled while in wait or send list */ + IB_MAD_STATE_CANCELED, + /* MAD processing completed, MAD in no list */ + IB_MAD_STATE_DONE +}; + struct ib_mad_send_wr_private { struct ib_mad_list_head mad_list; struct list_head agent_list; @@ -132,8 +161,6 @@ struct ib_mad_send_wr_private { int max_retries; int retries_left; int retry; - int refcount; - enum ib_wc_status status; /* RMPP control */ struct list_head rmpp_list; @@ -143,8 +170,48 @@ struct ib_mad_send_wr_private { int seg_num; int newwin; int pad; + + enum ib_mad_state state; + + /* Solicited MAD flow control */ + bool is_solicited_fc; }; +static inline void expect_mad_state(struct ib_mad_send_wr_private *mad_send_wr, + enum ib_mad_state expected_state) +{ + if (IS_ENABLED(CONFIG_LOCKDEP)) + WARN_ON(mad_send_wr->state != expected_state); +} + +static inline void expect_mad_state2(struct ib_mad_send_wr_private *mad_send_wr, + enum ib_mad_state expected_state1, + enum ib_mad_state expected_state2) +{ + if (IS_ENABLED(CONFIG_LOCKDEP)) + WARN_ON(mad_send_wr->state != expected_state1 && + mad_send_wr->state != expected_state2); +} + +static inline void expect_mad_state3(struct ib_mad_send_wr_private *mad_send_wr, + enum ib_mad_state expected_state1, + enum ib_mad_state expected_state2, + enum ib_mad_state expected_state3) +{ + if (IS_ENABLED(CONFIG_LOCKDEP)) + WARN_ON(mad_send_wr->state != expected_state1 && + mad_send_wr->state != expected_state2 && + mad_send_wr->state != expected_state3); +} + +static inline void +not_expect_mad_state(struct ib_mad_send_wr_private *mad_send_wr, + enum ib_mad_state wrong_state) +{ + if (IS_ENABLED(CONFIG_LOCKDEP)) + WARN_ON(mad_send_wr->state == wrong_state); +} + struct ib_mad_local_private { struct list_head completion_list; struct ib_mad_private *mad_priv; @@ -222,4 +289,7 @@ void ib_mark_mad_done(struct ib_mad_send_wr_private *mad_send_wr); void ib_reset_mad_timeout(struct ib_mad_send_wr_private *mad_send_wr, unsigned long timeout_ms); +void change_mad_state(struct ib_mad_send_wr_private *mad_send_wr, + enum ib_mad_state new_state); + #endif /* __IB_MAD_PRIV_H__ */ diff --git a/drivers/infiniband/core/mad_rmpp.c b/drivers/infiniband/core/mad_rmpp.c index 8af0619a39cd..1c5e0eaf1c94 100644 --- a/drivers/infiniband/core/mad_rmpp.c +++ b/drivers/infiniband/core/mad_rmpp.c @@ -158,7 +158,7 @@ static struct ib_mad_send_buf *alloc_response_msg(struct ib_mad_agent *agent, ah = ib_create_ah_from_wc(agent->qp->pd, recv_wc->wc, recv_wc->recv_buf.grh, agent->port_num); if (IS_ERR(ah)) - return (void *) ah; + return ERR_CAST(ah); hdr_len = ib_get_mad_data_offset(recv_wc->recv_buf.mad->mad_hdr.mgmt_class); msg = ib_create_send_mad(agent, recv_wc->wc->src_qp, @@ -608,16 +608,20 @@ static void abort_send(struct ib_mad_agent_private *agent, goto out; /* Unmatched send */ if ((mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) || - (!mad_send_wr->timeout) || (mad_send_wr->status != IB_WC_SUCCESS)) + (!mad_send_wr->timeout) || + (mad_send_wr->state == IB_MAD_STATE_CANCELED)) goto out; /* Send is already done */ ib_mark_mad_done(mad_send_wr); + if (mad_send_wr->state == IB_MAD_STATE_DONE) { + spin_unlock_irqrestore(&agent->lock, flags); + wc.status = IB_WC_REM_ABORT_ERR; + wc.vendor_err = rmpp_status; + wc.send_buf = &mad_send_wr->send_buf; + ib_mad_complete_send_wr(mad_send_wr, &wc); + return; + } spin_unlock_irqrestore(&agent->lock, flags); - - wc.status = IB_WC_REM_ABORT_ERR; - wc.vendor_err = rmpp_status; - wc.send_buf = &mad_send_wr->send_buf; - ib_mad_complete_send_wr(mad_send_wr, &wc); return; out: spin_unlock_irqrestore(&agent->lock, flags); @@ -684,7 +688,8 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent, } if ((mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) || - (!mad_send_wr->timeout) || (mad_send_wr->status != IB_WC_SUCCESS)) + (!mad_send_wr->timeout) || + (mad_send_wr->state == IB_MAD_STATE_CANCELED)) goto out; /* Send is already done */ if (seg_num > mad_send_wr->send_buf.seg_count || @@ -709,21 +714,24 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent, struct ib_mad_send_wc wc; ib_mark_mad_done(mad_send_wr); + if (mad_send_wr->state == IB_MAD_STATE_DONE) { + spin_unlock_irqrestore(&agent->lock, flags); + wc.status = IB_WC_SUCCESS; + wc.vendor_err = 0; + wc.send_buf = &mad_send_wr->send_buf; + ib_mad_complete_send_wr(mad_send_wr, &wc); + return; + } spin_unlock_irqrestore(&agent->lock, flags); - - wc.status = IB_WC_SUCCESS; - wc.vendor_err = 0; - wc.send_buf = &mad_send_wr->send_buf; - ib_mad_complete_send_wr(mad_send_wr, &wc); return; } - if (mad_send_wr->refcount == 1) + if (mad_send_wr->state == IB_MAD_STATE_WAIT_RESP) ib_reset_mad_timeout(mad_send_wr, mad_send_wr->send_buf.timeout_ms); spin_unlock_irqrestore(&agent->lock, flags); ack_ds_ack(agent, mad_recv_wc); return; - } else if (mad_send_wr->refcount == 1 && + } else if (mad_send_wr->state == IB_MAD_STATE_WAIT_RESP && mad_send_wr->seg_num < mad_send_wr->newwin && mad_send_wr->seg_num < mad_send_wr->send_buf.seg_count) { /* Send failure will just result in a timeout/retry */ @@ -731,7 +739,7 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent, if (ret) goto out; - mad_send_wr->refcount++; + change_mad_state(mad_send_wr, IB_MAD_STATE_SEND_START); list_move_tail(&mad_send_wr->agent_list, &mad_send_wr->mad_agent_priv->send_list); } @@ -890,7 +898,6 @@ int ib_send_rmpp_mad(struct ib_mad_send_wr_private *mad_send_wr) mad_send_wr->newwin = init_newwin(mad_send_wr); /* We need to wait for the final ACK even if there isn't a response */ - mad_send_wr->refcount += (mad_send_wr->timeout == 0); ret = send_next_seg(mad_send_wr); if (!ret) return IB_RMPP_RESULT_CONSUMED; @@ -912,7 +919,7 @@ int ib_process_rmpp_send_wc(struct ib_mad_send_wr_private *mad_send_wr, return IB_RMPP_RESULT_INTERNAL; /* ACK, STOP, or ABORT */ if (mad_send_wc->status != IB_WC_SUCCESS || - mad_send_wr->status != IB_WC_SUCCESS) + mad_send_wr->state == IB_MAD_STATE_CANCELED) return IB_RMPP_RESULT_PROCESSED; /* Canceled or send error */ if (!mad_send_wr->timeout) diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c index ff121e59b9c0..2220a2dfab24 100644 --- a/drivers/infiniband/core/nldev.c +++ b/drivers/infiniband/core/nldev.c @@ -171,6 +171,7 @@ static const struct nla_policy nldev_policy[RDMA_NLDEV_ATTR_MAX] = { [RDMA_NLDEV_ATTR_PARENT_NAME] = { .type = NLA_NUL_STRING }, [RDMA_NLDEV_ATTR_NAME_ASSIGN_TYPE] = { .type = NLA_U8 }, [RDMA_NLDEV_ATTR_EVENT_TYPE] = { .type = NLA_U8 }, + [RDMA_NLDEV_ATTR_STAT_OPCOUNTER_ENABLED] = { .type = NLA_U8 }, }; static int put_driver_name_print_type(struct sk_buff *msg, const char *name, @@ -254,7 +255,7 @@ EXPORT_SYMBOL(rdma_nl_put_driver_u64_hex); bool rdma_nl_get_privileged_qkey(void) { - return privileged_qkey || capable(CAP_NET_RAW); + return privileged_qkey; } EXPORT_SYMBOL(rdma_nl_get_privileged_qkey); @@ -1468,10 +1469,11 @@ static const struct nldev_fill_res_entry fill_entries[RDMA_RESTRACK_MAX] = { }; -static int res_get_common_doit(struct sk_buff *skb, struct nlmsghdr *nlh, - struct netlink_ext_ack *extack, - enum rdma_restrack_type res_type, - res_fill_func_t fill_func) +static noinline_for_stack int +res_get_common_doit(struct sk_buff *skb, struct nlmsghdr *nlh, + struct netlink_ext_ack *extack, + enum rdma_restrack_type res_type, + res_fill_func_t fill_func) { const struct nldev_fill_res_entry *fe = &fill_entries[res_type]; struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; @@ -2028,6 +2030,7 @@ static int nldev_stat_set_mode_doit(struct sk_buff *msg, struct ib_device *device, u32 port) { u32 mode, mask = 0, qpn, cntn = 0; + bool opcnt = false; int ret; /* Currently only counter for QP is supported */ @@ -2035,12 +2038,17 @@ static int nldev_stat_set_mode_doit(struct sk_buff *msg, nla_get_u32(tb[RDMA_NLDEV_ATTR_STAT_RES]) != RDMA_NLDEV_ATTR_RES_QP) return -EINVAL; + if (tb[RDMA_NLDEV_ATTR_STAT_OPCOUNTER_ENABLED]) + opcnt = !!nla_get_u8( + tb[RDMA_NLDEV_ATTR_STAT_OPCOUNTER_ENABLED]); + mode = nla_get_u32(tb[RDMA_NLDEV_ATTR_STAT_MODE]); if (mode == RDMA_COUNTER_MODE_AUTO) { if (tb[RDMA_NLDEV_ATTR_STAT_AUTO_MODE_MASK]) mask = nla_get_u32( tb[RDMA_NLDEV_ATTR_STAT_AUTO_MODE_MASK]); - return rdma_counter_set_auto_mode(device, port, mask, extack); + return rdma_counter_set_auto_mode(device, port, mask, opcnt, + extack); } if (!tb[RDMA_NLDEV_ATTR_RES_LQPN]) @@ -2256,10 +2264,10 @@ err: return ret; } -static int stat_get_doit_default_counter(struct sk_buff *skb, - struct nlmsghdr *nlh, - struct netlink_ext_ack *extack, - struct nlattr *tb[]) +static noinline_for_stack int +stat_get_doit_default_counter(struct sk_buff *skb, struct nlmsghdr *nlh, + struct netlink_ext_ack *extack, + struct nlattr *tb[]) { struct rdma_hw_stats *stats; struct nlattr *table_attr; @@ -2349,8 +2357,9 @@ err: return ret; } -static int stat_get_doit_qp(struct sk_buff *skb, struct nlmsghdr *nlh, - struct netlink_ext_ack *extack, struct nlattr *tb[]) +static noinline_for_stack int +stat_get_doit_qp(struct sk_buff *skb, struct nlmsghdr *nlh, + struct netlink_ext_ack *extack, struct nlattr *tb[]) { static enum rdma_nl_counter_mode mode; @@ -2358,6 +2367,7 @@ static int stat_get_doit_qp(struct sk_buff *skb, struct nlmsghdr *nlh, struct ib_device *device; struct sk_buff *msg; u32 index, port; + bool opcnt; int ret; if (tb[RDMA_NLDEV_ATTR_STAT_COUNTER_ID]) @@ -2393,7 +2403,7 @@ static int stat_get_doit_qp(struct sk_buff *skb, struct nlmsghdr *nlh, goto err_msg; } - ret = rdma_counter_get_mode(device, port, &mode, &mask); + ret = rdma_counter_get_mode(device, port, &mode, &mask, &opcnt); if (ret) goto err_msg; @@ -2410,6 +2420,12 @@ static int stat_get_doit_qp(struct sk_buff *skb, struct nlmsghdr *nlh, goto err_msg; } + if ((mode == RDMA_COUNTER_MODE_AUTO) && + nla_put_u8(msg, RDMA_NLDEV_ATTR_STAT_OPCOUNTER_ENABLED, opcnt)) { + ret = -EMSGSIZE; + goto err_msg; + } + nlmsg_end(msg, nlh); ib_device_put(device); return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid); @@ -2833,8 +2849,8 @@ int rdma_nl_notify_event(struct ib_device *device, u32 port_num, enum rdma_nl_notify_event_type type) { struct sk_buff *skb; + int ret = -EMSGSIZE; struct net *net; - int ret = 0; void *nlh; net = read_pnet(&device->coredev.rdma_net); diff --git a/drivers/infiniband/core/rdma_core.c b/drivers/infiniband/core/rdma_core.c index 90c177edf9b0..18918f463361 100644 --- a/drivers/infiniband/core/rdma_core.c +++ b/drivers/infiniband/core/rdma_core.c @@ -1019,3 +1019,32 @@ void uverbs_finalize_object(struct ib_uobject *uobj, WARN_ON(true); } } + +/** + * rdma_uattrs_has_raw_cap() - Returns whether a rdma device linked to the + * uverbs attributes file has CAP_NET_RAW + * capability or not. + * + * @attrs: Pointer to uverbs attributes + * + * Returns true if a rdma device's owning user namespace has CAP_NET_RAW + * capability, otherwise false. + */ +bool rdma_uattrs_has_raw_cap(const struct uverbs_attr_bundle *attrs) +{ + struct ib_uverbs_file *ufile = attrs->ufile; + struct ib_ucontext *ucontext; + bool has_cap = false; + int srcu_key; + + srcu_key = srcu_read_lock(&ufile->device->disassociate_srcu); + ucontext = ib_uverbs_get_ucontext_file(ufile); + if (IS_ERR(ucontext)) + goto out; + has_cap = rdma_dev_has_raw_cap(ucontext->device); + +out: + srcu_read_unlock(&ufile->device->disassociate_srcu, srcu_key); + return has_cap; +} +EXPORT_SYMBOL(rdma_uattrs_has_raw_cap); diff --git a/drivers/infiniband/core/rdma_core.h b/drivers/infiniband/core/rdma_core.h index 33706dad6c0f..a59b087611cb 100644 --- a/drivers/infiniband/core/rdma_core.h +++ b/drivers/infiniband/core/rdma_core.h @@ -156,6 +156,7 @@ extern const struct uapi_definition uverbs_def_obj_counters[]; extern const struct uapi_definition uverbs_def_obj_cq[]; extern const struct uapi_definition uverbs_def_obj_device[]; extern const struct uapi_definition uverbs_def_obj_dm[]; +extern const struct uapi_definition uverbs_def_obj_dmah[]; extern const struct uapi_definition uverbs_def_obj_flow_action[]; extern const struct uapi_definition uverbs_def_obj_intf[]; extern const struct uapi_definition uverbs_def_obj_mr[]; diff --git a/drivers/infiniband/core/restrack.c b/drivers/infiniband/core/restrack.c index 3313410014cd..b097cfcade1c 100644 --- a/drivers/infiniband/core/restrack.c +++ b/drivers/infiniband/core/restrack.c @@ -100,6 +100,8 @@ static struct ib_device *res_to_dev(struct rdma_restrack_entry *res) return container_of(res, struct rdma_counter, res)->device; case RDMA_RESTRACK_SRQ: return container_of(res, struct ib_srq, res)->device; + case RDMA_RESTRACK_DMAH: + return container_of(res, struct ib_dmah, res)->device; default: WARN_ONCE(true, "Wrong resource tracking type %u\n", res->type); return NULL; @@ -173,7 +175,7 @@ void rdma_restrack_new(struct rdma_restrack_entry *res, EXPORT_SYMBOL(rdma_restrack_new); /** - * rdma_restrack_add() - add object to the reource tracking database + * rdma_restrack_add() - add object to the resource tracking database * @res: resource entry */ void rdma_restrack_add(struct rdma_restrack_entry *res) @@ -275,7 +277,7 @@ int rdma_restrack_put(struct rdma_restrack_entry *res) EXPORT_SYMBOL(rdma_restrack_put); /** - * rdma_restrack_del() - delete object from the reource tracking database + * rdma_restrack_del() - delete object from the resource tracking database * @res: resource entry */ void rdma_restrack_del(struct rdma_restrack_entry *res) diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c index 53571e6b3162..c23e9c847314 100644 --- a/drivers/infiniband/core/sa_query.c +++ b/drivers/infiniband/core/sa_query.c @@ -107,6 +107,8 @@ struct ib_sa_device { struct ib_sa_query { void (*callback)(struct ib_sa_query *sa_query, int status, struct ib_sa_mad *mad); + void (*rmpp_callback)(struct ib_sa_query *sa_query, int status, + struct ib_mad_recv_wc *mad); void (*release)(struct ib_sa_query *); struct ib_sa_client *client; struct ib_sa_port *port; @@ -150,6 +152,13 @@ struct ib_sa_mcmember_query { struct ib_sa_query sa_query; }; +struct ib_sa_service_query { + void (*callback)(int status, struct sa_service_rec *rec, + unsigned int num_services, void *context); + void *context; + struct ib_sa_query sa_query; +}; + static LIST_HEAD(ib_nl_request_list); static DEFINE_SPINLOCK(ib_nl_request_lock); static atomic_t ib_nl_sa_request_seq; @@ -684,6 +693,58 @@ static const struct ib_field guidinfo_rec_table[] = { .size_bits = 512 }, }; +#define SERVICE_REC_FIELD(field) \ + .struct_offset_bytes = offsetof(struct sa_service_rec, field), \ + .struct_size_bytes = sizeof_field(struct sa_service_rec, field), \ + .field_name = "sa_service_rec:" #field + +static const struct ib_field service_rec_table[] = { + { SERVICE_REC_FIELD(id), + .offset_words = 0, + .offset_bits = 0, + .size_bits = 64 }, + { SERVICE_REC_FIELD(gid), + .offset_words = 2, + .offset_bits = 0, + .size_bits = 128 }, + { SERVICE_REC_FIELD(pkey), + .offset_words = 6, + .offset_bits = 0, + .size_bits = 16 }, + { RESERVED, + .offset_words = 6, + .offset_bits = 16, + .size_bits = 16 }, + { SERVICE_REC_FIELD(lease), + .offset_words = 7, + .offset_bits = 0, + .size_bits = 32 }, + { SERVICE_REC_FIELD(key), + .offset_words = 8, + .offset_bits = 0, + .size_bits = 128 }, + { SERVICE_REC_FIELD(name), + .offset_words = 12, + .offset_bits = 0, + .size_bits = 512 }, + { SERVICE_REC_FIELD(data_8), + .offset_words = 28, + .offset_bits = 0, + .size_bits = 128 }, + { SERVICE_REC_FIELD(data_16), + .offset_words = 32, + .offset_bits = 0, + .size_bits = 128 }, + { SERVICE_REC_FIELD(data_32), + .offset_words = 36, + .offset_bits = 0, + .size_bits = 128 }, + { SERVICE_REC_FIELD(data_64), + .offset_words = 40, + .offset_bits = 0, + .size_bits = 128 }, +}; + #define RDMA_PRIMARY_PATH_MAX_REC_NUM 3 static inline void ib_sa_disable_local_svc(struct ib_sa_query *query) @@ -1013,6 +1074,8 @@ int ib_nl_handle_set_timeout(struct sk_buff *skb, if (timeout > IB_SA_LOCAL_SVC_TIMEOUT_MAX) timeout = IB_SA_LOCAL_SVC_TIMEOUT_MAX; + spin_lock_irqsave(&ib_nl_request_lock, flags); + delta = timeout - sa_local_svc_timeout_ms; if (delta < 0) abs_delta = -delta; @@ -1020,7 +1083,6 @@ int ib_nl_handle_set_timeout(struct sk_buff *skb, abs_delta = delta; if (delta != 0) { - spin_lock_irqsave(&ib_nl_request_lock, flags); sa_local_svc_timeout_ms = timeout; list_for_each_entry(query, &ib_nl_request_list, list) { if (delta < 0 && abs_delta > query->timeout) @@ -1038,9 +1100,10 @@ int ib_nl_handle_set_timeout(struct sk_buff *skb, if (delay) mod_delayed_work(ib_nl_wq, &ib_nl_timed_work, (unsigned long)delay); - spin_unlock_irqrestore(&ib_nl_request_lock, flags); } + spin_unlock_irqrestore(&ib_nl_request_lock, flags); + settimeout_out: return 0; } @@ -1390,6 +1453,20 @@ void ib_sa_pack_path(struct sa_path_rec *rec, void *attribute) } EXPORT_SYMBOL(ib_sa_pack_path); +void ib_sa_pack_service(struct sa_service_rec *rec, void *attribute) +{ + ib_pack(service_rec_table, ARRAY_SIZE(service_rec_table), rec, + attribute); +} +EXPORT_SYMBOL(ib_sa_pack_service); + +void ib_sa_unpack_service(void *attribute, struct sa_service_rec *rec) +{ + ib_unpack(service_rec_table, ARRAY_SIZE(service_rec_table), attribute, + rec); +} +EXPORT_SYMBOL(ib_sa_unpack_service); + static bool ib_sa_opa_pathrecord_support(struct ib_sa_client *client, struct ib_sa_device *sa_dev, u32 port_num) @@ -1479,6 +1556,68 @@ static void ib_sa_path_rec_callback(struct ib_sa_query *sa_query, } } +#define IB_SA_DATA_OFFS 56 +#define IB_SERVICE_REC_SZ 176 + +static void ib_unpack_service_rmpp(struct sa_service_rec *rec, + struct ib_mad_recv_wc *mad_wc, + int num_services) +{ + unsigned int cp_sz, data_i, data_size, rec_i = 0, buf_i = 0; + struct ib_mad_recv_buf *mad_buf; + u8 buf[IB_SERVICE_REC_SZ]; + u8 *data; + + data_size = sizeof(((struct ib_sa_mad *) mad_buf->mad)->data); + + list_for_each_entry(mad_buf, &mad_wc->rmpp_list, list) { + data = ((struct ib_sa_mad *) mad_buf->mad)->data; + data_i = 0; + while (data_i < data_size && rec_i < num_services) { + cp_sz = min(IB_SERVICE_REC_SZ - buf_i, + data_size - data_i); + memcpy(buf + buf_i, data + data_i, cp_sz); + data_i += cp_sz; + buf_i += cp_sz; + if (buf_i == IB_SERVICE_REC_SZ) { + ib_sa_unpack_service(buf, rec + rec_i); + buf_i = 0; + rec_i++; + } + } + } +} + +static void ib_sa_service_rec_callback(struct ib_sa_query *sa_query, int status, + struct ib_mad_recv_wc *mad_wc) +{ + struct ib_sa_service_query *query = + container_of(sa_query, struct ib_sa_service_query, sa_query); + struct sa_service_rec *rec; + int num_services; + + if (!mad_wc || !mad_wc->recv_buf.mad) { + query->callback(status, NULL, 0, query->context); + return; + } + + num_services = (mad_wc->mad_len - IB_SA_DATA_OFFS) / IB_SERVICE_REC_SZ; + if (!num_services) { + query->callback(-ENODATA, NULL, 0, query->context); + return; + } + + rec = kmalloc_array(num_services, sizeof(*rec), GFP_KERNEL); + if (!rec) { + query->callback(-ENOMEM, NULL, 0, query->context); + return; + } + + ib_unpack_service_rmpp(rec, mad_wc, num_services); + query->callback(status, rec, num_services, query->context); + kfree(rec); +} + static void ib_sa_path_rec_release(struct ib_sa_query *sa_query) { struct ib_sa_path_query *query = @@ -1488,6 +1627,14 @@ static void ib_sa_path_rec_release(struct ib_sa_query *sa_query) kfree(query); } +static void ib_sa_service_rec_release(struct ib_sa_query *sa_query) +{ + struct ib_sa_service_query *query = + container_of(sa_query, struct ib_sa_service_query, sa_query); + + kfree(query); +} + /** * ib_sa_path_rec_get - Start a Path get query * @client:SA client @@ -1618,6 +1765,101 @@ err1: } EXPORT_SYMBOL(ib_sa_path_rec_get); +/** + * ib_sa_service_rec_get - Start a Service get query + * @client: SA client + * @device: device to send query on + * @port_num: port number to send query on + * @rec: Service Record to send in query + * @comp_mask: component mask to send in query + * @timeout_ms: time to wait for response + * @gfp_mask: GFP mask to use for internal allocations + * @callback: function called when query completes, times out or is + * canceled + * @context: opaque user context passed to callback + * @sa_query: query context, used to cancel query + * + * Send a Service Record Get query to the SA to look up a path. The + * callback function will be called when the query completes (or + * fails); status is 0 for a successful response, -EINTR if the query + * is canceled, -ETIMEDOUT is the query timed out, or -EIO if an error + * occurred sending the query. The resp parameter of the callback is + * only valid if status is 0. + * + * If the return value of ib_sa_service_rec_get() is negative, it is an + * error code. Otherwise it is a query ID that can be used to cancel + * the query. + */ +int ib_sa_service_rec_get(struct ib_sa_client *client, + struct ib_device *device, u32 port_num, + struct sa_service_rec *rec, + ib_sa_comp_mask comp_mask, + unsigned long timeout_ms, gfp_t gfp_mask, + void (*callback)(int status, + struct sa_service_rec *resp, + unsigned int num_services, + void *context), + void *context, struct ib_sa_query **sa_query) +{ + struct ib_sa_device *sa_dev = ib_get_client_data(device, &sa_client); + struct ib_sa_service_query *query; + struct ib_mad_agent *agent; + struct ib_sa_port *port; + struct ib_sa_mad *mad; + int ret; + + if (!sa_dev) + return -ENODEV; + + port = &sa_dev->port[port_num - sa_dev->start_port]; + agent = port->agent; + + query = kzalloc(sizeof(*query), gfp_mask); + if (!query) + return -ENOMEM; + + query->sa_query.port = port; + + ret = alloc_mad(&query->sa_query, gfp_mask); + if (ret) + goto err1; + + ib_sa_client_get(client); + query->sa_query.client = client; + query->callback = callback; + query->context = context; + + mad = query->sa_query.mad_buf->mad; + init_mad(&query->sa_query, agent); + + query->sa_query.rmpp_callback = callback ? ib_sa_service_rec_callback : + NULL; + query->sa_query.release = ib_sa_service_rec_release; + mad->mad_hdr.method = IB_MGMT_METHOD_GET_TABLE; + mad->mad_hdr.attr_id = cpu_to_be16(IB_SA_ATTR_SERVICE_REC); + mad->sa_hdr.comp_mask = comp_mask; + + ib_sa_pack_service(rec, mad->data); + + *sa_query = &query->sa_query; + query->sa_query.mad_buf->context[1] = rec; + + ret = send_mad(&query->sa_query, timeout_ms, gfp_mask); + if (ret < 0) + goto err2; + + return ret; + +err2: + *sa_query = NULL; + ib_sa_client_put(query->sa_query.client); + free_mad(&query->sa_query); +err1: + kfree(query); + return ret; +} +EXPORT_SYMBOL(ib_sa_service_rec_get); + static void ib_sa_mcmember_rec_callback(struct ib_sa_query *sa_query, int status, struct ib_sa_mad *mad) { @@ -1987,23 +2229,29 @@ static void send_handler(struct ib_mad_agent *agent, { struct ib_sa_query *query = mad_send_wc->send_buf->context[0]; unsigned long flags; + int status = 0; - if (query->callback) + if (query->callback || query->rmpp_callback) { switch (mad_send_wc->status) { case IB_WC_SUCCESS: /* No callback -- already got recv */ break; case IB_WC_RESP_TIMEOUT_ERR: - query->callback(query, -ETIMEDOUT, NULL); + status = -ETIMEDOUT; break; case IB_WC_WR_FLUSH_ERR: - query->callback(query, -EINTR, NULL); + status = -EINTR; break; default: - query->callback(query, -EIO, NULL); + status = -EIO; break; } + if (status) + query->callback ? query->callback(query, status, NULL) : + query->rmpp_callback(query, status, NULL); + } + xa_lock_irqsave(&queries, flags); __xa_erase(&queries, query->id); xa_unlock_irqrestore(&queries, flags); @@ -2019,17 +2267,25 @@ static void recv_handler(struct ib_mad_agent *mad_agent, struct ib_mad_recv_wc *mad_recv_wc) { struct ib_sa_query *query; + struct ib_mad *mad; + if (!send_buf) return; query = send_buf->context[0]; - if (query->callback) { + mad = mad_recv_wc->recv_buf.mad; + + if (query->rmpp_callback) { + if (mad_recv_wc->wc->status == IB_WC_SUCCESS) + query->rmpp_callback(query, mad->mad_hdr.status ? + -EINVAL : 0, mad_recv_wc); + else + query->rmpp_callback(query, -EIO, NULL); + } else if (query->callback) { if (mad_recv_wc->wc->status == IB_WC_SUCCESS) - query->callback(query, - mad_recv_wc->recv_buf.mad->mad_hdr.status ? - -EINVAL : 0, - (struct ib_sa_mad *) mad_recv_wc->recv_buf.mad); + query->callback(query, mad->mad_hdr.status ? + -EINVAL : 0, (struct ib_sa_mad *)mad); else query->callback(query, -EIO, NULL); } @@ -2181,8 +2437,9 @@ static int ib_sa_add_one(struct ib_device *device) sa_dev->port[i].agent = ib_register_mad_agent(device, i + s, IB_QPT_GSI, - NULL, 0, send_handler, - recv_handler, sa_dev, 0); + NULL, IB_MGMT_RMPP_VERSION, + send_handler, recv_handler, + sa_dev, 0); if (IS_ERR(sa_dev->port[i].agent)) { ret = PTR_ERR(sa_dev->port[i].agent); goto err; diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c index 9f97bef02149..0ed862b38b44 100644 --- a/drivers/infiniband/core/sysfs.c +++ b/drivers/infiniband/core/sysfs.c @@ -216,24 +216,12 @@ static ssize_t state_show(struct ib_device *ibdev, u32 port_num, struct ib_port_attr attr; ssize_t ret; - static const char *state_name[] = { - [IB_PORT_NOP] = "NOP", - [IB_PORT_DOWN] = "DOWN", - [IB_PORT_INIT] = "INIT", - [IB_PORT_ARMED] = "ARMED", - [IB_PORT_ACTIVE] = "ACTIVE", - [IB_PORT_ACTIVE_DEFER] = "ACTIVE_DEFER" - }; - ret = ib_query_port(ibdev, port_num, &attr); if (ret) return ret; return sysfs_emit(buf, "%d: %s\n", attr.state, - attr.state >= 0 && - attr.state < ARRAY_SIZE(state_name) ? - state_name[attr.state] : - "UNKNOWN"); + ib_port_state_to_str(attr.state)); } static ssize_t lid_show(struct ib_device *ibdev, u32 port_num, @@ -988,6 +976,7 @@ int ib_setup_device_attrs(struct ib_device *ibdev) for (i = 0; i != ARRAY_SIZE(ibdev->groups); i++) if (!ibdev->groups[i]) { ibdev->groups[i] = &data->group; + ibdev->hw_stats_attr_index = i; return 0; } WARN(true, "struct ib_device->groups is too small"); diff --git a/drivers/infiniband/core/ucaps.c b/drivers/infiniband/core/ucaps.c new file mode 100644 index 000000000000..de5cb8bf0a61 --- /dev/null +++ b/drivers/infiniband/core/ucaps.c @@ -0,0 +1,267 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* + * Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved + */ + +#include <linux/kref.h> +#include <linux/cdev.h> +#include <linux/mutex.h> +#include <linux/file.h> +#include <linux/fs.h> +#include <rdma/ib_ucaps.h> + +#define RDMA_UCAP_FIRST RDMA_UCAP_MLX5_CTRL_LOCAL + +static DEFINE_MUTEX(ucaps_mutex); +static struct ib_ucap *ucaps_list[RDMA_UCAP_MAX]; +static bool ucaps_class_is_registered; +static dev_t ucaps_base_dev; + +struct ib_ucap { + struct cdev cdev; + struct device dev; + struct kref ref; +}; + +static const char *ucap_names[RDMA_UCAP_MAX] = { + [RDMA_UCAP_MLX5_CTRL_LOCAL] = "mlx5_perm_ctrl_local", + [RDMA_UCAP_MLX5_CTRL_OTHER_VHCA] = "mlx5_perm_ctrl_other_vhca" +}; + +static char *ucaps_devnode(const struct device *dev, umode_t *mode) +{ + if (mode) + *mode = 0600; + + return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev)); +} + +static const struct class ucaps_class = { + .name = "infiniband_ucaps", + .devnode = ucaps_devnode, +}; + +static const struct file_operations ucaps_cdev_fops = { + .owner = THIS_MODULE, + .open = simple_open, +}; + +/** + * ib_cleanup_ucaps - cleanup all API resources and class. + * + * This is called once, when removing the ib_uverbs module. + */ +void ib_cleanup_ucaps(void) +{ + mutex_lock(&ucaps_mutex); + if (!ucaps_class_is_registered) { + mutex_unlock(&ucaps_mutex); + return; + } + + for (int i = RDMA_UCAP_FIRST; i < RDMA_UCAP_MAX; i++) + WARN_ON(ucaps_list[i]); + + class_unregister(&ucaps_class); + ucaps_class_is_registered = false; + unregister_chrdev_region(ucaps_base_dev, RDMA_UCAP_MAX); + mutex_unlock(&ucaps_mutex); +} + +static int get_ucap_from_devt(dev_t devt, u64 *idx_mask) +{ + for (int type = RDMA_UCAP_FIRST; type < RDMA_UCAP_MAX; type++) { + if (ucaps_list[type] && ucaps_list[type]->dev.devt == devt) { + *idx_mask |= 1 << type; + return 0; + } + } + + return -EINVAL; +} + +static int get_devt_from_fd(unsigned int fd, dev_t *ret_dev) +{ + struct file *file; + + file = fget(fd); + if (!file) + return -EBADF; + + *ret_dev = file_inode(file)->i_rdev; + fput(file); + return 0; +} + +/** + * ib_ucaps_init - Initialization required before ucap creation. + * + * Return: 0 on success, or a negative errno value on failure + */ +static int ib_ucaps_init(void) +{ + int ret = 0; + + if (ucaps_class_is_registered) + return ret; + + ret = class_register(&ucaps_class); + if (ret) + return ret; + + ret = alloc_chrdev_region(&ucaps_base_dev, 0, RDMA_UCAP_MAX, + ucaps_class.name); + if (ret < 0) { + class_unregister(&ucaps_class); + return ret; + } + + ucaps_class_is_registered = true; + + return 0; +} + +static void ucap_dev_release(struct device *device) +{ + struct ib_ucap *ucap = container_of(device, struct ib_ucap, dev); + + kfree(ucap); +} + +/** + * ib_create_ucap - Add a ucap character device + * @type: UCAP type + * + * Creates a ucap character device in the /dev/infiniband directory. By default, + * the device has root-only read-write access. + * + * A driver may call this multiple times with the same UCAP type. A reference + * count tracks creations and deletions. + * + * Return: 0 on success, or a negative errno value on failure + */ +int ib_create_ucap(enum rdma_user_cap type) +{ + struct ib_ucap *ucap; + int ret; + + if (type >= RDMA_UCAP_MAX) + return -EINVAL; + + mutex_lock(&ucaps_mutex); + ret = ib_ucaps_init(); + if (ret) + goto unlock; + + ucap = ucaps_list[type]; + if (ucap) { + kref_get(&ucap->ref); + mutex_unlock(&ucaps_mutex); + return 0; + } + + ucap = kzalloc(sizeof(*ucap), GFP_KERNEL); + if (!ucap) { + ret = -ENOMEM; + goto unlock; + } + + device_initialize(&ucap->dev); + ucap->dev.class = &ucaps_class; + ucap->dev.devt = MKDEV(MAJOR(ucaps_base_dev), type); + ucap->dev.release = ucap_dev_release; + ret = dev_set_name(&ucap->dev, "%s", ucap_names[type]); + if (ret) + goto err_device; + + cdev_init(&ucap->cdev, &ucaps_cdev_fops); + ucap->cdev.owner = THIS_MODULE; + + ret = cdev_device_add(&ucap->cdev, &ucap->dev); + if (ret) + goto err_device; + + kref_init(&ucap->ref); + ucaps_list[type] = ucap; + mutex_unlock(&ucaps_mutex); + + return 0; + +err_device: + put_device(&ucap->dev); +unlock: + mutex_unlock(&ucaps_mutex); + return ret; +} +EXPORT_SYMBOL(ib_create_ucap); + +static void ib_release_ucap(struct kref *ref) +{ + struct ib_ucap *ucap = container_of(ref, struct ib_ucap, ref); + enum rdma_user_cap type; + + for (type = RDMA_UCAP_FIRST; type < RDMA_UCAP_MAX; type++) { + if (ucaps_list[type] == ucap) + break; + } + WARN_ON(type == RDMA_UCAP_MAX); + + ucaps_list[type] = NULL; + cdev_device_del(&ucap->cdev, &ucap->dev); + put_device(&ucap->dev); +} + +/** + * ib_remove_ucap - Remove a ucap character device + * @type: User cap type + * + * Removes the ucap character device according to type. The device is completely + * removed from the filesystem when its reference count reaches 0. + */ +void ib_remove_ucap(enum rdma_user_cap type) +{ + struct ib_ucap *ucap; + + mutex_lock(&ucaps_mutex); + ucap = ucaps_list[type]; + if (WARN_ON(!ucap)) + goto end; + + kref_put(&ucap->ref, ib_release_ucap); +end: + mutex_unlock(&ucaps_mutex); +} +EXPORT_SYMBOL(ib_remove_ucap); + +/** + * ib_get_ucaps - Get bitmask of ucap types from file descriptors + * @fds: Array of file descriptors + * @fd_count: Number of file descriptors in the array + * @idx_mask: Bitmask to be updated based on the ucaps in the fd list + * + * Given an array of file descriptors, this function returns a bitmask of + * the ucaps where a bit is set if an FD for that ucap type was in the array. + * + * Return: 0 on success, or a negative errno value on failure + */ +int ib_get_ucaps(int *fds, int fd_count, uint64_t *idx_mask) +{ + int ret = 0; + dev_t dev; + + *idx_mask = 0; + mutex_lock(&ucaps_mutex); + for (int i = 0; i < fd_count; i++) { + ret = get_devt_from_fd(fds[i], &dev); + if (ret) + goto end; + + ret = get_ucap_from_devt(dev, idx_mask); + if (ret) + goto end; + } + +end: + mutex_unlock(&ucaps_mutex); + return ret; +} diff --git a/drivers/infiniband/core/ucma.c b/drivers/infiniband/core/ucma.c index 02f1666f3cba..ec3be65a2b88 100644 --- a/drivers/infiniband/core/ucma.c +++ b/drivers/infiniband/core/ucma.c @@ -69,7 +69,9 @@ static struct ctl_table ucma_ctl_table[] = { .data = &max_backlog, .maxlen = sizeof max_backlog, .mode = 0644, - .proc_handler = proc_dointvec, + .proc_handler = proc_dointvec_minmax, + .extra1 = SYSCTL_ZERO, + .extra2 = SYSCTL_INT_MAX, }, }; @@ -280,6 +282,10 @@ static struct ucma_event *ucma_create_uevent(struct ucma_context *ctx, } uevent->resp.event = event->event; uevent->resp.status = event->status; + + if (event->event == RDMA_CM_EVENT_ADDRINFO_RESOLVED) + goto out; + if (ctx->cm_id->qp_type == IB_QPT_UD) ucma_copy_ud_event(ctx->cm_id->device, &uevent->resp.param.ud, &event->param.ud); @@ -287,6 +293,7 @@ static struct ucma_event *ucma_create_uevent(struct ucma_context *ctx, ucma_copy_conn_event(&uevent->resp.param.conn, &event->param.conn); +out: uevent->resp.ece.vendor_id = event->ece.vendor_id; uevent->resp.ece.attr_mod = event->ece.attr_mod; return uevent; @@ -359,7 +366,7 @@ static int ucma_event_handler(struct rdma_cm_id *cm_id, if (event->event == RDMA_CM_EVENT_DEVICE_REMOVAL) { xa_lock(&ctx_table); if (xa_load(&ctx_table, ctx->id) == ctx) - queue_work(system_unbound_wq, &ctx->close_work); + queue_work(system_dfl_wq, &ctx->close_work); xa_unlock(&ctx_table); } return 0; @@ -726,6 +733,28 @@ static ssize_t ucma_resolve_addr(struct ucma_file *file, return ret; } +static ssize_t ucma_resolve_ib_service(struct ucma_file *file, + const char __user *inbuf, int in_len, + int out_len) +{ + struct rdma_ucm_resolve_ib_service cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + mutex_lock(&ctx->mutex); + ret = rdma_resolve_ib_service(ctx->cm_id, &cmd.ibs); + mutex_unlock(&ctx->mutex); + ucma_put_ctx(ctx); + return ret; +} + static ssize_t ucma_resolve_route(struct ucma_file *file, const char __user *inbuf, int in_len, int out_len) @@ -992,6 +1021,43 @@ static ssize_t ucma_query_gid(struct ucma_context *ctx, return ret; } +static ssize_t ucma_query_ib_service(struct ucma_context *ctx, + void __user *response, int out_len) +{ + struct rdma_ucm_query_ib_service_resp *resp; + int n, ret = 0; + + if (out_len < sizeof(struct rdma_ucm_query_ib_service_resp)) + return -ENOSPC; + + if (!ctx->cm_id->route.service_recs) + return -ENODATA; + + resp = kzalloc(out_len, GFP_KERNEL); + if (!resp) + return -ENOMEM; + + resp->num_service_recs = ctx->cm_id->route.num_service_recs; + + n = (out_len - sizeof(struct rdma_ucm_query_ib_service_resp)) / + sizeof(struct ib_user_service_rec); + + if (!n) + goto out; + + if (n > ctx->cm_id->route.num_service_recs) + n = ctx->cm_id->route.num_service_recs; + + memcpy(resp->recs, ctx->cm_id->route.service_recs, + sizeof(*resp->recs) * n); + if (copy_to_user(response, resp, struct_size(resp, recs, n))) + ret = -EFAULT; + +out: + kfree(resp); + return ret; +} + static ssize_t ucma_query(struct ucma_file *file, const char __user *inbuf, int in_len, int out_len) @@ -1020,6 +1086,9 @@ static ssize_t ucma_query(struct ucma_file *file, case RDMA_USER_CM_QUERY_GID: ret = ucma_query_gid(ctx, response, out_len); break; + case RDMA_USER_CM_QUERY_IB_SERVICE: + ret = ucma_query_ib_service(ctx, response, out_len); + break; default: ret = -ENOSYS; break; @@ -1676,6 +1745,55 @@ err_unlock: return ret; } +static ssize_t ucma_write_cm_event(struct ucma_file *file, + const char __user *inbuf, int in_len, + int out_len) +{ + struct rdma_ucm_write_cm_event cmd; + struct rdma_cm_event event = {}; + struct ucma_event *uevent; + struct ucma_context *ctx; + int ret = 0; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + if ((cmd.event != RDMA_CM_EVENT_USER) && + (cmd.event != RDMA_CM_EVENT_INTERNAL)) + return -EINVAL; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + event.event = cmd.event; + event.status = cmd.status; + event.param.arg = cmd.param.arg; + + uevent = kzalloc(sizeof(*uevent), GFP_KERNEL); + if (!uevent) { + ret = -ENOMEM; + goto out; + } + + uevent->ctx = ctx; + uevent->resp.uid = ctx->uid; + uevent->resp.id = ctx->id; + uevent->resp.event = event.event; + uevent->resp.status = event.status; + memcpy(uevent->resp.param.arg32, &event.param.arg, + sizeof(event.param.arg)); + + mutex_lock(&ctx->file->mut); + list_add_tail(&uevent->list, &ctx->file->event_list); + mutex_unlock(&ctx->file->mut); + wake_up_interruptible(&ctx->file->poll_wait); + +out: + ucma_put_ctx(ctx); + return ret; +} + static ssize_t (*ucma_cmd_table[])(struct ucma_file *file, const char __user *inbuf, int in_len, int out_len) = { @@ -1701,7 +1819,9 @@ static ssize_t (*ucma_cmd_table[])(struct ucma_file *file, [RDMA_USER_CM_CMD_QUERY] = ucma_query, [RDMA_USER_CM_CMD_BIND] = ucma_bind, [RDMA_USER_CM_CMD_RESOLVE_ADDR] = ucma_resolve_addr, - [RDMA_USER_CM_CMD_JOIN_MCAST] = ucma_join_multicast + [RDMA_USER_CM_CMD_JOIN_MCAST] = ucma_join_multicast, + [RDMA_USER_CM_CMD_RESOLVE_IB_SERVICE] = ucma_resolve_ib_service, + [RDMA_USER_CM_CMD_WRITE_CM_EVENT] = ucma_write_cm_event, }; static ssize_t ucma_write(struct file *filp, const char __user *buf, diff --git a/drivers/infiniband/core/ud_header.c b/drivers/infiniband/core/ud_header.c index 64d9c492de64..8d3dfef9ebaa 100644 --- a/drivers/infiniband/core/ud_header.c +++ b/drivers/infiniband/core/ud_header.c @@ -462,86 +462,3 @@ int ib_ud_header_pack(struct ib_ud_header *header, return len; } EXPORT_SYMBOL(ib_ud_header_pack); - -/** - * ib_ud_header_unpack - Unpack UD header struct from wire format - * @header:UD header struct - * @buf:Buffer to pack into - * - * ib_ud_header_pack() unpacks the UD header structure @header from wire - * format in the buffer @buf. - */ -int ib_ud_header_unpack(void *buf, - struct ib_ud_header *header) -{ - ib_unpack(lrh_table, ARRAY_SIZE(lrh_table), - buf, &header->lrh); - buf += IB_LRH_BYTES; - - if (header->lrh.link_version != 0) { - pr_warn("Invalid LRH.link_version %u\n", - header->lrh.link_version); - return -EINVAL; - } - - switch (header->lrh.link_next_header) { - case IB_LNH_IBA_LOCAL: - header->grh_present = 0; - break; - - case IB_LNH_IBA_GLOBAL: - header->grh_present = 1; - ib_unpack(grh_table, ARRAY_SIZE(grh_table), - buf, &header->grh); - buf += IB_GRH_BYTES; - - if (header->grh.ip_version != 6) { - pr_warn("Invalid GRH.ip_version %u\n", - header->grh.ip_version); - return -EINVAL; - } - if (header->grh.next_header != 0x1b) { - pr_warn("Invalid GRH.next_header 0x%02x\n", - header->grh.next_header); - return -EINVAL; - } - break; - - default: - pr_warn("Invalid LRH.link_next_header %u\n", - header->lrh.link_next_header); - return -EINVAL; - } - - ib_unpack(bth_table, ARRAY_SIZE(bth_table), - buf, &header->bth); - buf += IB_BTH_BYTES; - - switch (header->bth.opcode) { - case IB_OPCODE_UD_SEND_ONLY: - header->immediate_present = 0; - break; - case IB_OPCODE_UD_SEND_ONLY_WITH_IMMEDIATE: - header->immediate_present = 1; - break; - default: - pr_warn("Invalid BTH.opcode 0x%02x\n", header->bth.opcode); - return -EINVAL; - } - - if (header->bth.transport_header_version != 0) { - pr_warn("Invalid BTH.transport_header_version %u\n", - header->bth.transport_header_version); - return -EINVAL; - } - - ib_unpack(deth_table, ARRAY_SIZE(deth_table), - buf, &header->deth); - buf += IB_DETH_BYTES; - - if (header->immediate_present) - memcpy(&header->immediate_data, buf, sizeof header->immediate_data); - - return 0; -} -EXPORT_SYMBOL(ib_ud_header_unpack); diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c index 07c571c7b699..8137031c2a65 100644 --- a/drivers/infiniband/core/umem.c +++ b/drivers/infiniband/core/umem.c @@ -45,6 +45,8 @@ #include "uverbs.h" +#define RESCHED_LOOP_CNT_THRESHOLD 0x1000 + static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int dirty) { bool make_dirty = umem->writable && dirty; @@ -55,10 +57,14 @@ static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int d ib_dma_unmap_sgtable_attrs(dev, &umem->sgt_append.sgt, DMA_BIDIRECTIONAL, 0); - for_each_sgtable_sg(&umem->sgt_append.sgt, sg, i) + for_each_sgtable_sg(&umem->sgt_append.sgt, sg, i) { unpin_user_page_range_dirty_lock(sg_page(sg), DIV_ROUND_UP(sg->length, PAGE_SIZE), make_dirty); + if (i && !(i % RESCHED_LOOP_CNT_THRESHOLD)) + cond_resched(); + } + sg_free_append_table(&umem->sgt_append); } @@ -80,9 +86,12 @@ unsigned long ib_umem_find_best_pgsz(struct ib_umem *umem, unsigned long pgsz_bitmap, unsigned long virt) { - struct scatterlist *sg; + unsigned long curr_len = 0; + dma_addr_t curr_base = ~0; unsigned long va, pgoff; + struct scatterlist *sg; dma_addr_t mask; + dma_addr_t end; int i; umem->iova = va = virt; @@ -107,17 +116,30 @@ unsigned long ib_umem_find_best_pgsz(struct ib_umem *umem, pgoff = umem->address & ~PAGE_MASK; for_each_sgtable_dma_sg(&umem->sgt_append.sgt, sg, i) { - /* Walk SGL and reduce max page size if VA/PA bits differ - * for any address. + /* If the current entry is physically contiguous with the previous + * one, no need to take its start addresses into consideration. */ - mask |= (sg_dma_address(sg) + pgoff) ^ va; + if (check_add_overflow(curr_base, curr_len, &end) || + end != sg_dma_address(sg)) { + + curr_base = sg_dma_address(sg); + curr_len = 0; + + /* Reduce max page size if VA/PA bits differ */ + mask |= (curr_base + pgoff) ^ va; + + /* The alignment of any VA matching a discontinuity point + * in the physical memory sets the maximum possible page + * size as this must be a starting point of a new page that + * needs to be aligned. + */ + if (i != 0) + mask |= va; + } + + curr_len += sg_dma_len(sg); va += sg_dma_len(sg) - pgoff; - /* Except for the last entry, the ending iova alignment sets - * the maximum possible page size as the low bits of the iova - * must be zero when starting the next chunk. - */ - if (i != (umem->sgt_append.sgt.nents - 1)) - mask |= va; + pgoff = 0; } diff --git a/drivers/infiniband/core/umem_dmabuf.c b/drivers/infiniband/core/umem_dmabuf.c index 9fcd37761264..0ec2e4120cc9 100644 --- a/drivers/infiniband/core/umem_dmabuf.c +++ b/drivers/infiniband/core/umem_dmabuf.c @@ -10,7 +10,7 @@ #include "uverbs.h" -MODULE_IMPORT_NS(DMA_BUF); +MODULE_IMPORT_NS("DMA_BUF"); int ib_umem_dmabuf_map_pages(struct ib_umem_dmabuf *umem_dmabuf) { diff --git a/drivers/infiniband/core/umem_odp.c b/drivers/infiniband/core/umem_odp.c index e9fa22d31c23..572a91a62a7b 100644 --- a/drivers/infiniband/core/umem_odp.c +++ b/drivers/infiniband/core/umem_odp.c @@ -41,65 +41,83 @@ #include <linux/hugetlb.h> #include <linux/interval_tree.h> #include <linux/hmm.h> +#include <linux/hmm-dma.h> #include <linux/pagemap.h> #include <rdma/ib_umem_odp.h> #include "uverbs.h" -static inline int ib_init_umem_odp(struct ib_umem_odp *umem_odp, - const struct mmu_interval_notifier_ops *ops) +static void ib_init_umem_implicit_odp(struct ib_umem_odp *umem_odp) { - int ret; + umem_odp->is_implicit_odp = 1; + umem_odp->umem.is_odp = 1; + mutex_init(&umem_odp->umem_mutex); +} + +static int ib_init_umem_odp(struct ib_umem_odp *umem_odp, + const struct mmu_interval_notifier_ops *ops) +{ + struct ib_device *dev = umem_odp->umem.ibdev; + size_t page_size = 1UL << umem_odp->page_shift; + struct hmm_dma_map *map; + unsigned long start; + unsigned long end; + size_t nr_entries; + int ret = 0; umem_odp->umem.is_odp = 1; mutex_init(&umem_odp->umem_mutex); - if (!umem_odp->is_implicit_odp) { - size_t page_size = 1UL << umem_odp->page_shift; - unsigned long start; - unsigned long end; - size_t ndmas, npfns; - - start = ALIGN_DOWN(umem_odp->umem.address, page_size); - if (check_add_overflow(umem_odp->umem.address, - (unsigned long)umem_odp->umem.length, - &end)) - return -EOVERFLOW; - end = ALIGN(end, page_size); - if (unlikely(end < page_size)) - return -EOVERFLOW; - - ndmas = (end - start) >> umem_odp->page_shift; - if (!ndmas) - return -EINVAL; - - npfns = (end - start) >> PAGE_SHIFT; - umem_odp->pfn_list = kvcalloc( - npfns, sizeof(*umem_odp->pfn_list), GFP_KERNEL); - if (!umem_odp->pfn_list) - return -ENOMEM; - - umem_odp->dma_list = kvcalloc( - ndmas, sizeof(*umem_odp->dma_list), GFP_KERNEL); - if (!umem_odp->dma_list) { + start = ALIGN_DOWN(umem_odp->umem.address, page_size); + if (check_add_overflow(umem_odp->umem.address, + (unsigned long)umem_odp->umem.length, &end)) + return -EOVERFLOW; + end = ALIGN(end, page_size); + if (unlikely(end < page_size)) + return -EOVERFLOW; + /* + * The mmu notifier can be called within reclaim contexts and takes the + * umem_mutex. This is rare to trigger in testing, teach lockdep about + * it. + */ + if (IS_ENABLED(CONFIG_LOCKDEP)) { + fs_reclaim_acquire(GFP_KERNEL); + mutex_lock(&umem_odp->umem_mutex); + mutex_unlock(&umem_odp->umem_mutex); + fs_reclaim_release(GFP_KERNEL); + } + + nr_entries = (end - start) >> PAGE_SHIFT; + if (!(nr_entries * PAGE_SIZE / page_size)) + return -EINVAL; + + map = &umem_odp->map; + if (ib_uses_virt_dma(dev)) { + map->pfn_list = kvcalloc(nr_entries, sizeof(*map->pfn_list), + GFP_KERNEL | __GFP_NOWARN); + if (!map->pfn_list) ret = -ENOMEM; - goto out_pfn_list; - } + } else + ret = hmm_dma_map_alloc(dev->dma_device, map, + (end - start) >> PAGE_SHIFT, + 1 << umem_odp->page_shift); + if (ret) + return ret; - ret = mmu_interval_notifier_insert(&umem_odp->notifier, - umem_odp->umem.owning_mm, - start, end - start, ops); - if (ret) - goto out_dma_list; - } + ret = mmu_interval_notifier_insert(&umem_odp->notifier, + umem_odp->umem.owning_mm, start, + end - start, ops); + if (ret) + goto out_free_map; return 0; -out_dma_list: - kvfree(umem_odp->dma_list); -out_pfn_list: - kvfree(umem_odp->pfn_list); +out_free_map: + if (ib_uses_virt_dma(dev)) + kvfree(map->pfn_list); + else + hmm_dma_map_free(dev->dma_device, map); return ret; } @@ -118,7 +136,6 @@ struct ib_umem_odp *ib_umem_odp_alloc_implicit(struct ib_device *device, { struct ib_umem *umem; struct ib_umem_odp *umem_odp; - int ret; if (access & IB_ACCESS_HUGETLB) return ERR_PTR(-EINVAL); @@ -130,16 +147,10 @@ struct ib_umem_odp *ib_umem_odp_alloc_implicit(struct ib_device *device, umem->ibdev = device; umem->writable = ib_access_writable(access); umem->owning_mm = current->mm; - umem_odp->is_implicit_odp = 1; umem_odp->page_shift = PAGE_SHIFT; umem_odp->tgid = get_task_pid(current->group_leader, PIDTYPE_PID); - ret = ib_init_umem_odp(umem_odp, NULL); - if (ret) { - put_pid(umem_odp->tgid); - kfree(umem_odp); - return ERR_PTR(ret); - } + ib_init_umem_implicit_odp(umem_odp); return umem_odp; } EXPORT_SYMBOL(ib_umem_odp_alloc_implicit); @@ -260,74 +271,41 @@ err_put_pid: } EXPORT_SYMBOL(ib_umem_odp_get); -void ib_umem_odp_release(struct ib_umem_odp *umem_odp) +static void ib_umem_odp_free(struct ib_umem_odp *umem_odp) { + struct ib_device *dev = umem_odp->umem.ibdev; + /* * Ensure that no more pages are mapped in the umem. * * It is the driver's responsibility to ensure, before calling us, * that the hardware will not attempt to access the MR any more. */ - if (!umem_odp->is_implicit_odp) { - mutex_lock(&umem_odp->umem_mutex); - ib_umem_odp_unmap_dma_pages(umem_odp, ib_umem_start(umem_odp), - ib_umem_end(umem_odp)); - mutex_unlock(&umem_odp->umem_mutex); - mmu_interval_notifier_remove(&umem_odp->notifier); - kvfree(umem_odp->dma_list); - kvfree(umem_odp->pfn_list); - } - put_pid(umem_odp->tgid); - kfree(umem_odp); + mutex_lock(&umem_odp->umem_mutex); + ib_umem_odp_unmap_dma_pages(umem_odp, ib_umem_start(umem_odp), + ib_umem_end(umem_odp)); + mutex_unlock(&umem_odp->umem_mutex); + mmu_interval_notifier_remove(&umem_odp->notifier); + if (ib_uses_virt_dma(dev)) + kvfree(umem_odp->map.pfn_list); + else + hmm_dma_map_free(dev->dma_device, &umem_odp->map); } -EXPORT_SYMBOL(ib_umem_odp_release); -/* - * Map for DMA and insert a single page into the on-demand paging page tables. - * - * @umem: the umem to insert the page to. - * @dma_index: index in the umem to add the dma to. - * @page: the page struct to map and add. - * @access_mask: access permissions needed for this page. - * - * The function returns -EFAULT if the DMA mapping operation fails. - * - */ -static int ib_umem_odp_map_dma_single_page( - struct ib_umem_odp *umem_odp, - unsigned int dma_index, - struct page *page, - u64 access_mask) +void ib_umem_odp_release(struct ib_umem_odp *umem_odp) { - struct ib_device *dev = umem_odp->umem.ibdev; - dma_addr_t *dma_addr = &umem_odp->dma_list[dma_index]; - - if (*dma_addr) { - /* - * If the page is already dma mapped it means it went through - * a non-invalidating trasition, like read-only to writable. - * Resync the flags. - */ - *dma_addr = (*dma_addr & ODP_DMA_ADDR_MASK) | access_mask; - return 0; - } + if (!umem_odp->is_implicit_odp) + ib_umem_odp_free(umem_odp); - *dma_addr = ib_dma_map_page(dev, page, 0, 1 << umem_odp->page_shift, - DMA_BIDIRECTIONAL); - if (ib_dma_mapping_error(dev, *dma_addr)) { - *dma_addr = 0; - return -EFAULT; - } - umem_odp->npages++; - *dma_addr |= access_mask; - return 0; + put_pid(umem_odp->tgid); + kfree(umem_odp); } +EXPORT_SYMBOL(ib_umem_odp_release); /** * ib_umem_odp_map_dma_and_lock - DMA map userspace memory in an ODP MR and lock it. * * Maps the range passed in the argument to DMA addresses. - * The DMA addresses of the mapped pages is updated in umem_odp->dma_list. * Upon success the ODP MR will be locked to let caller complete its device * page table update. * @@ -355,9 +333,6 @@ int ib_umem_odp_map_dma_and_lock(struct ib_umem_odp *umem_odp, u64 user_virt, struct hmm_range range = {}; unsigned long timeout; - if (access_mask == 0) - return -EINVAL; - if (user_virt < ib_umem_start(umem_odp) || user_virt + bcnt > ib_umem_end(umem_odp)) return -EFAULT; @@ -383,11 +358,11 @@ int ib_umem_odp_map_dma_and_lock(struct ib_umem_odp *umem_odp, u64 user_virt, if (fault) { range.default_flags = HMM_PFN_REQ_FAULT; - if (access_mask & ODP_WRITE_ALLOWED_BIT) + if (access_mask & HMM_PFN_WRITE) range.default_flags |= HMM_PFN_REQ_WRITE; } - range.hmm_pfns = &(umem_odp->pfn_list[pfn_start_idx]); + range.hmm_pfns = &(umem_odp->map.pfn_list[pfn_start_idx]); timeout = jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT); retry: @@ -415,22 +390,17 @@ retry: for (pfn_index = 0; pfn_index < num_pfns; pfn_index += 1 << (page_shift - PAGE_SHIFT), dma_index++) { - if (fault) { - /* - * Since we asked for hmm_range_fault() to populate - * pages it shouldn't return an error entry on success. - */ - WARN_ON(range.hmm_pfns[pfn_index] & HMM_PFN_ERROR); - WARN_ON(!(range.hmm_pfns[pfn_index] & HMM_PFN_VALID)); - } else { - if (!(range.hmm_pfns[pfn_index] & HMM_PFN_VALID)) { - WARN_ON(umem_odp->dma_list[dma_index]); - continue; - } - access_mask = ODP_READ_ALLOWED_BIT; - if (range.hmm_pfns[pfn_index] & HMM_PFN_WRITE) - access_mask |= ODP_WRITE_ALLOWED_BIT; - } + /* + * Since we asked for hmm_range_fault() to populate + * pages it shouldn't return an error entry on success. + */ + WARN_ON(fault && range.hmm_pfns[pfn_index] & HMM_PFN_ERROR); + WARN_ON(fault && !(range.hmm_pfns[pfn_index] & HMM_PFN_VALID)); + if (!(range.hmm_pfns[pfn_index] & HMM_PFN_VALID)) + continue; + + if (range.hmm_pfns[pfn_index] & HMM_PFN_DMA_MAPPED) + continue; hmm_order = hmm_pfn_to_map_order(range.hmm_pfns[pfn_index]); /* If a hugepage was detected and ODP wasn't set for, the umem @@ -443,15 +413,6 @@ retry: __func__, hmm_order, page_shift); break; } - - ret = ib_umem_odp_map_dma_single_page( - umem_odp, dma_index, hmm_pfn_to_page(range.hmm_pfns[pfn_index]), - access_mask); - if (ret < 0) { - ibdev_dbg(umem_odp->umem.ibdev, - "ib_umem_odp_map_dma_single_page failed with error %d\n", ret); - break; - } } /* upon success lock should stay on hold for the callee */ if (!ret) @@ -471,45 +432,38 @@ EXPORT_SYMBOL(ib_umem_odp_map_dma_and_lock); void ib_umem_odp_unmap_dma_pages(struct ib_umem_odp *umem_odp, u64 virt, u64 bound) { - dma_addr_t dma_addr; - dma_addr_t dma; - int idx; - u64 addr; struct ib_device *dev = umem_odp->umem.ibdev; + u64 addr; lockdep_assert_held(&umem_odp->umem_mutex); virt = max_t(u64, virt, ib_umem_start(umem_odp)); bound = min_t(u64, bound, ib_umem_end(umem_odp)); for (addr = virt; addr < bound; addr += BIT(umem_odp->page_shift)) { - idx = (addr - ib_umem_start(umem_odp)) >> umem_odp->page_shift; - dma = umem_odp->dma_list[idx]; - - /* The access flags guaranteed a valid DMA address in case was NULL */ - if (dma) { - unsigned long pfn_idx = (addr - ib_umem_start(umem_odp)) >> PAGE_SHIFT; - struct page *page = hmm_pfn_to_page(umem_odp->pfn_list[pfn_idx]); - - dma_addr = dma & ODP_DMA_ADDR_MASK; - ib_dma_unmap_page(dev, dma_addr, - BIT(umem_odp->page_shift), - DMA_BIDIRECTIONAL); - if (dma & ODP_WRITE_ALLOWED_BIT) { - struct page *head_page = compound_head(page); - /* - * set_page_dirty prefers being called with - * the page lock. However, MMU notifiers are - * called sometimes with and sometimes without - * the lock. We rely on the umem_mutex instead - * to prevent other mmu notifiers from - * continuing and allowing the page mapping to - * be removed. - */ - set_page_dirty(head_page); - } - umem_odp->dma_list[idx] = 0; - umem_odp->npages--; + u64 offset = addr - ib_umem_start(umem_odp); + size_t idx = offset >> umem_odp->page_shift; + unsigned long pfn = umem_odp->map.pfn_list[idx]; + + if (!hmm_dma_unmap_pfn(dev->dma_device, &umem_odp->map, idx)) + goto clear; + + if (pfn & HMM_PFN_WRITE) { + struct page *page = hmm_pfn_to_page(pfn); + struct page *head_page = compound_head(page); + /* + * set_page_dirty prefers being called with + * the page lock. However, MMU notifiers are + * called sometimes with and sometimes without + * the lock. We rely on the umem_mutex instead + * to prevent other mmu notifiers from + * continuing and allowing the page mapping to + * be removed. + */ + set_page_dirty(head_page); } + umem_odp->npages--; +clear: + umem_odp->map.pfn_list[idx] &= ~HMM_PFN_FLAGS; } } EXPORT_SYMBOL(ib_umem_odp_unmap_dma_pages); diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c index 66b02fbf077a..ce16404cdfb8 100644 --- a/drivers/infiniband/core/uverbs_cmd.c +++ b/drivers/infiniband/core/uverbs_cmd.c @@ -42,6 +42,7 @@ #include <rdma/uverbs_types.h> #include <rdma/uverbs_std_types.h> +#include <rdma/ib_ucaps.h> #include "rdma_core.h" #include "uverbs.h" @@ -161,7 +162,7 @@ static const void __user *uverbs_request_next_ptr(struct uverbs_req_iter *iter, { const void __user *res = iter->cur; - if (iter->cur + len > iter->end) + if (len > iter->end - iter->cur) return (void __force __user *)ERR_PTR(-ENOSPC); iter->cur += len; return res; @@ -192,7 +193,7 @@ _ib_uverbs_lookup_comp_file(s32 fd, struct uverbs_attr_bundle *attrs) fd, attrs); if (IS_ERR(uobj)) - return (void *)uobj; + return ERR_CAST(uobj); uverbs_uobject_get(uobj); uobj_put_read(uobj); @@ -232,6 +233,8 @@ int ib_init_ucontext(struct uverbs_attr_bundle *attrs) { struct ib_ucontext *ucontext = attrs->context; struct ib_uverbs_file *file = attrs->ufile; + int *fd_array; + int fd_count; int ret; if (!down_read_trylock(&file->hw_destroy_rwsem)) @@ -247,6 +250,22 @@ int ib_init_ucontext(struct uverbs_attr_bundle *attrs) if (ret) goto err; + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_GET_CONTEXT_FD_ARR)) { + fd_count = uverbs_attr_ptr_get_array_size(attrs, + UVERBS_ATTR_GET_CONTEXT_FD_ARR, + sizeof(int)); + if (fd_count < 0) { + ret = fd_count; + goto err_uncharge; + } + + fd_array = uverbs_attr_get_alloced_ptr(attrs, + UVERBS_ATTR_GET_CONTEXT_FD_ARR); + ret = ib_get_ucaps(fd_array, fd_count, &ucontext->enabled_caps); + if (ret) + goto err_uncharge; + } + ret = ucontext->device->ops.alloc_ucontext(ucontext, &attrs->driver_udata); if (ret) @@ -716,13 +735,13 @@ static int ib_uverbs_reg_mr(struct uverbs_attr_bundle *attrs) goto err_free; pd = uobj_get_obj_read(pd, UVERBS_OBJECT_PD, cmd.pd_handle, attrs); - if (!pd) { - ret = -EINVAL; + if (IS_ERR(pd)) { + ret = PTR_ERR(pd); goto err_free; } mr = pd->device->ops.reg_user_mr(pd, cmd.start, cmd.length, cmd.hca_va, - cmd.access_flags, + cmd.access_flags, NULL, &attrs->driver_udata); if (IS_ERR(mr)) { ret = PTR_ERR(mr); @@ -807,8 +826,8 @@ static int ib_uverbs_rereg_mr(struct uverbs_attr_bundle *attrs) if (cmd.flags & IB_MR_REREG_PD) { new_pd = uobj_get_obj_read(pd, UVERBS_OBJECT_PD, cmd.pd_handle, attrs); - if (!new_pd) { - ret = -EINVAL; + if (IS_ERR(new_pd)) { + ret = PTR_ERR(new_pd); goto put_uobjs; } } else { @@ -917,8 +936,8 @@ static int ib_uverbs_alloc_mw(struct uverbs_attr_bundle *attrs) return PTR_ERR(uobj); pd = uobj_get_obj_read(pd, UVERBS_OBJECT_PD, cmd.pd_handle, attrs); - if (!pd) { - ret = -EINVAL; + if (IS_ERR(pd)) { + ret = PTR_ERR(pd); goto err_free; } @@ -1125,8 +1144,8 @@ static int ib_uverbs_resize_cq(struct uverbs_attr_bundle *attrs) return ret; cq = uobj_get_obj_read(cq, UVERBS_OBJECT_CQ, cmd.cq_handle, attrs); - if (!cq) - return -EINVAL; + if (IS_ERR(cq)) + return PTR_ERR(cq); ret = cq->device->ops.resize_cq(cq, cmd.cqe, &attrs->driver_udata); if (ret) @@ -1187,8 +1206,8 @@ static int ib_uverbs_poll_cq(struct uverbs_attr_bundle *attrs) return ret; cq = uobj_get_obj_read(cq, UVERBS_OBJECT_CQ, cmd.cq_handle, attrs); - if (!cq) - return -EINVAL; + if (IS_ERR(cq)) + return PTR_ERR(cq); /* we copy a struct ib_uverbs_poll_cq_resp to user space */ header_ptr = attrs->ucore.outbuf; @@ -1236,8 +1255,8 @@ static int ib_uverbs_req_notify_cq(struct uverbs_attr_bundle *attrs) return ret; cq = uobj_get_obj_read(cq, UVERBS_OBJECT_CQ, cmd.cq_handle, attrs); - if (!cq) - return -EINVAL; + if (IS_ERR(cq)) + return PTR_ERR(cq); ib_req_notify_cq(cq, cmd.solicited_only ? IB_CQ_SOLICITED : IB_CQ_NEXT_COMP); @@ -1293,9 +1312,9 @@ static int create_qp(struct uverbs_attr_bundle *attrs, switch (cmd->qp_type) { case IB_QPT_RAW_PACKET: - if (!capable(CAP_NET_RAW)) + if (!rdma_uattrs_has_raw_cap(attrs)) return -EPERM; - break; + fallthrough; case IB_QPT_RC: case IB_QPT_UC: case IB_QPT_UD: @@ -1319,8 +1338,8 @@ static int create_qp(struct uverbs_attr_bundle *attrs, ind_tbl = uobj_get_obj_read(rwq_ind_table, UVERBS_OBJECT_RWQ_IND_TBL, cmd->rwq_ind_tbl_handle, attrs); - if (!ind_tbl) { - ret = -EINVAL; + if (IS_ERR(ind_tbl)) { + ret = PTR_ERR(ind_tbl); goto err_put; } @@ -1358,8 +1377,10 @@ static int create_qp(struct uverbs_attr_bundle *attrs, if (cmd->is_srq) { srq = uobj_get_obj_read(srq, UVERBS_OBJECT_SRQ, cmd->srq_handle, attrs); - if (!srq || srq->srq_type == IB_SRQT_XRC) { - ret = -EINVAL; + if (IS_ERR(srq) || + srq->srq_type == IB_SRQT_XRC) { + ret = IS_ERR(srq) ? PTR_ERR(srq) : + -EINVAL; goto err_put; } } @@ -1369,23 +1390,29 @@ static int create_qp(struct uverbs_attr_bundle *attrs, rcq = uobj_get_obj_read( cq, UVERBS_OBJECT_CQ, cmd->recv_cq_handle, attrs); - if (!rcq) { - ret = -EINVAL; + if (IS_ERR(rcq)) { + ret = PTR_ERR(rcq); goto err_put; } } } } - if (has_sq) + if (has_sq) { scq = uobj_get_obj_read(cq, UVERBS_OBJECT_CQ, cmd->send_cq_handle, attrs); + if (IS_ERR(scq)) { + ret = PTR_ERR(scq); + goto err_put; + } + } + if (!ind_tbl && cmd->qp_type != IB_QPT_XRC_INI) rcq = rcq ?: scq; pd = uobj_get_obj_read(pd, UVERBS_OBJECT_PD, cmd->pd_handle, attrs); - if (!pd || (!scq && has_sq)) { - ret = -EINVAL; + if (IS_ERR(pd)) { + ret = PTR_ERR(pd); goto err_put; } @@ -1424,7 +1451,7 @@ static int create_qp(struct uverbs_attr_bundle *attrs, } if (attr.create_flags & IB_QP_CREATE_SOURCE_QPN) { - if (!capable(CAP_NET_RAW)) { + if (!rdma_uattrs_has_raw_cap(attrs)) { ret = -EPERM; goto err_put; } @@ -1480,18 +1507,18 @@ static int create_qp(struct uverbs_attr_bundle *attrs, err_put: if (!IS_ERR(xrcd_uobj)) uobj_put_read(xrcd_uobj); - if (pd) + if (!IS_ERR_OR_NULL(pd)) uobj_put_obj_read(pd); - if (scq) + if (!IS_ERR_OR_NULL(scq)) rdma_lookup_put_uobject(&scq->uobject->uevent.uobject, UVERBS_LOOKUP_READ); - if (rcq && rcq != scq) + if (!IS_ERR_OR_NULL(rcq) && rcq != scq) rdma_lookup_put_uobject(&rcq->uobject->uevent.uobject, UVERBS_LOOKUP_READ); - if (srq) + if (!IS_ERR_OR_NULL(srq)) rdma_lookup_put_uobject(&srq->uobject->uevent.uobject, UVERBS_LOOKUP_READ); - if (ind_tbl) + if (!IS_ERR_OR_NULL(ind_tbl)) uobj_put_obj_read(ind_tbl); uobj_alloc_abort(&obj->uevent.uobject, attrs); @@ -1653,8 +1680,8 @@ static int ib_uverbs_query_qp(struct uverbs_attr_bundle *attrs) } qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd.qp_handle, attrs); - if (!qp) { - ret = -EINVAL; + if (IS_ERR(qp)) { + ret = PTR_ERR(qp); goto out; } @@ -1759,8 +1786,8 @@ static int modify_qp(struct uverbs_attr_bundle *attrs, qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd->base.qp_handle, attrs); - if (!qp) { - ret = -EINVAL; + if (IS_ERR(qp)) { + ret = PTR_ERR(qp); goto out; } @@ -1850,7 +1877,8 @@ static int modify_qp(struct uverbs_attr_bundle *attrs, attr->path_mig_state = cmd->base.path_mig_state; if (cmd->base.attr_mask & IB_QP_QKEY) { if (cmd->base.qkey & IB_QP_SET_QKEY && - !rdma_nl_get_privileged_qkey()) { + !(rdma_nl_get_privileged_qkey() || + rdma_uattrs_has_raw_cap(attrs))) { ret = -EPERM; goto release_qp; } @@ -2008,11 +2036,13 @@ static int ib_uverbs_post_send(struct uverbs_attr_bundle *attrs) ret = uverbs_request_start(attrs, &iter, &cmd, sizeof(cmd)); if (ret) return ret; - wqes = uverbs_request_next_ptr(&iter, cmd.wqe_size * cmd.wr_count); + wqes = uverbs_request_next_ptr(&iter, size_mul(cmd.wqe_size, + cmd.wr_count)); if (IS_ERR(wqes)) return PTR_ERR(wqes); - sgls = uverbs_request_next_ptr( - &iter, cmd.sge_count * sizeof(struct ib_uverbs_sge)); + sgls = uverbs_request_next_ptr(&iter, + size_mul(cmd.sge_count, + sizeof(struct ib_uverbs_sge))); if (IS_ERR(sgls)) return PTR_ERR(sgls); ret = uverbs_request_finish(&iter); @@ -2024,8 +2054,8 @@ static int ib_uverbs_post_send(struct uverbs_attr_bundle *attrs) return -ENOMEM; qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd.qp_handle, attrs); - if (!qp) { - ret = -EINVAL; + if (IS_ERR(qp)) { + ret = PTR_ERR(qp); goto out; } @@ -2062,9 +2092,9 @@ static int ib_uverbs_post_send(struct uverbs_attr_bundle *attrs) ud->ah = uobj_get_obj_read(ah, UVERBS_OBJECT_AH, user_wr->wr.ud.ah, attrs); - if (!ud->ah) { + if (IS_ERR(ud->ah)) { + ret = PTR_ERR(ud->ah); kfree(ud); - ret = -EINVAL; goto out_put; } ud->remote_qpn = user_wr->wr.ud.remote_qpn; @@ -2198,11 +2228,11 @@ ib_uverbs_unmarshall_recv(struct uverbs_req_iter *iter, u32 wr_count, if (wqe_size < sizeof(struct ib_uverbs_recv_wr)) return ERR_PTR(-EINVAL); - wqes = uverbs_request_next_ptr(iter, wqe_size * wr_count); + wqes = uverbs_request_next_ptr(iter, size_mul(wqe_size, wr_count)); if (IS_ERR(wqes)) return ERR_CAST(wqes); - sgls = uverbs_request_next_ptr( - iter, sge_count * sizeof(struct ib_uverbs_sge)); + sgls = uverbs_request_next_ptr(iter, size_mul(sge_count, + sizeof(struct ib_uverbs_sge))); if (IS_ERR(sgls)) return ERR_CAST(sgls); ret = uverbs_request_finish(iter); @@ -2301,8 +2331,8 @@ static int ib_uverbs_post_recv(struct uverbs_attr_bundle *attrs) return PTR_ERR(wr); qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd.qp_handle, attrs); - if (!qp) { - ret = -EINVAL; + if (IS_ERR(qp)) { + ret = PTR_ERR(qp); goto out; } @@ -2352,8 +2382,8 @@ static int ib_uverbs_post_srq_recv(struct uverbs_attr_bundle *attrs) return PTR_ERR(wr); srq = uobj_get_obj_read(srq, UVERBS_OBJECT_SRQ, cmd.srq_handle, attrs); - if (!srq) { - ret = -EINVAL; + if (IS_ERR(srq)) { + ret = PTR_ERR(srq); goto out; } @@ -2409,8 +2439,8 @@ static int ib_uverbs_create_ah(struct uverbs_attr_bundle *attrs) } pd = uobj_get_obj_read(pd, UVERBS_OBJECT_PD, cmd.pd_handle, attrs); - if (!pd) { - ret = -EINVAL; + if (IS_ERR(pd)) { + ret = PTR_ERR(pd); goto err; } @@ -2479,8 +2509,8 @@ static int ib_uverbs_attach_mcast(struct uverbs_attr_bundle *attrs) return ret; qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd.qp_handle, attrs); - if (!qp) - return -EINVAL; + if (IS_ERR(qp)) + return PTR_ERR(qp); obj = qp->uobject; @@ -2529,8 +2559,8 @@ static int ib_uverbs_detach_mcast(struct uverbs_attr_bundle *attrs) return ret; qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd.qp_handle, attrs); - if (!qp) - return -EINVAL; + if (IS_ERR(qp)) + return PTR_ERR(qp); obj = qp->uobject; mutex_lock(&obj->mcast_lock); @@ -2664,8 +2694,8 @@ static int kern_spec_to_ib_spec_action(struct uverbs_attr_bundle *attrs, UVERBS_OBJECT_FLOW_ACTION, kern_spec->action.handle, attrs); - if (!ib_spec->action.act) - return -EINVAL; + if (IS_ERR(ib_spec->action.act)) + return PTR_ERR(ib_spec->action.act); ib_spec->action.size = sizeof(struct ib_flow_spec_action_handle); flow_resources_add(uflow_res, @@ -2682,8 +2712,8 @@ static int kern_spec_to_ib_spec_action(struct uverbs_attr_bundle *attrs, UVERBS_OBJECT_COUNTERS, kern_spec->flow_count.handle, attrs); - if (!ib_spec->flow_count.counters) - return -EINVAL; + if (IS_ERR(ib_spec->flow_count.counters)) + return PTR_ERR(ib_spec->flow_count.counters); ib_spec->flow_count.size = sizeof(struct ib_flow_spec_action_count); flow_resources_add(uflow_res, @@ -2901,14 +2931,14 @@ static int ib_uverbs_ex_create_wq(struct uverbs_attr_bundle *attrs) return PTR_ERR(obj); pd = uobj_get_obj_read(pd, UVERBS_OBJECT_PD, cmd.pd_handle, attrs); - if (!pd) { - err = -EINVAL; + if (IS_ERR(pd)) { + err = PTR_ERR(pd); goto err_uobj; } cq = uobj_get_obj_read(cq, UVERBS_OBJECT_CQ, cmd.cq_handle, attrs); - if (!cq) { - err = -EINVAL; + if (IS_ERR(cq)) { + err = PTR_ERR(cq); goto err_put_pd; } @@ -3009,8 +3039,8 @@ static int ib_uverbs_ex_modify_wq(struct uverbs_attr_bundle *attrs) return -EINVAL; wq = uobj_get_obj_read(wq, UVERBS_OBJECT_WQ, cmd.wq_handle, attrs); - if (!wq) - return -EINVAL; + if (IS_ERR(wq)) + return PTR_ERR(wq); if (cmd.attr_mask & IB_WQ_FLAGS) { wq_attr.flags = cmd.flags; @@ -3093,8 +3123,8 @@ static int ib_uverbs_ex_create_rwq_ind_table(struct uverbs_attr_bundle *attrs) num_read_wqs++) { wq = uobj_get_obj_read(wq, UVERBS_OBJECT_WQ, wqs_handles[num_read_wqs], attrs); - if (!wq) { - err = -EINVAL; + if (IS_ERR(wq)) { + err = PTR_ERR(wq); goto put_wqs; } @@ -3196,7 +3226,7 @@ static int ib_uverbs_ex_create_flow(struct uverbs_attr_bundle *attrs) if (cmd.comp_mask) return -EINVAL; - if (!capable(CAP_NET_RAW)) + if (!rdma_uattrs_has_raw_cap(attrs)) return -EPERM; if (cmd.flow_attr.flags >= IB_FLOW_ATTR_FLAGS_RESERVED) @@ -3249,8 +3279,8 @@ static int ib_uverbs_ex_create_flow(struct uverbs_attr_bundle *attrs) } qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd.qp_handle, attrs); - if (!qp) { - err = -EINVAL; + if (IS_ERR(qp)) { + err = PTR_ERR(qp); goto err_uobj; } @@ -3396,15 +3426,15 @@ static int __uverbs_create_xsrq(struct uverbs_attr_bundle *attrs, if (ib_srq_has_cq(cmd->srq_type)) { attr.ext.cq = uobj_get_obj_read(cq, UVERBS_OBJECT_CQ, cmd->cq_handle, attrs); - if (!attr.ext.cq) { - ret = -EINVAL; + if (IS_ERR(attr.ext.cq)) { + ret = PTR_ERR(attr.ext.cq); goto err_put_xrcd; } } pd = uobj_get_obj_read(pd, UVERBS_OBJECT_PD, cmd->pd_handle, attrs); - if (!pd) { - ret = -EINVAL; + if (IS_ERR(pd)) { + ret = PTR_ERR(pd); goto err_put_cq; } @@ -3511,8 +3541,8 @@ static int ib_uverbs_modify_srq(struct uverbs_attr_bundle *attrs) return ret; srq = uobj_get_obj_read(srq, UVERBS_OBJECT_SRQ, cmd.srq_handle, attrs); - if (!srq) - return -EINVAL; + if (IS_ERR(srq)) + return PTR_ERR(srq); attr.max_wr = cmd.max_wr; attr.srq_limit = cmd.srq_limit; @@ -3539,8 +3569,8 @@ static int ib_uverbs_query_srq(struct uverbs_attr_bundle *attrs) return ret; srq = uobj_get_obj_read(srq, UVERBS_OBJECT_SRQ, cmd.srq_handle, attrs); - if (!srq) - return -EINVAL; + if (IS_ERR(srq)) + return PTR_ERR(srq); ret = ib_query_srq(srq, &attr); @@ -3665,8 +3695,8 @@ static int ib_uverbs_ex_modify_cq(struct uverbs_attr_bundle *attrs) return -EOPNOTSUPP; cq = uobj_get_obj_read(cq, UVERBS_OBJECT_CQ, cmd.cq_handle, attrs); - if (!cq) - return -EINVAL; + if (IS_ERR(cq)) + return PTR_ERR(cq); ret = rdma_set_cq_moderation(cq, cmd.attr.cq_count, cmd.attr.cq_period); diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c index 85cfc790a7bb..973fe2c7ef53 100644 --- a/drivers/infiniband/core/uverbs_main.c +++ b/drivers/infiniband/core/uverbs_main.c @@ -52,6 +52,7 @@ #include <rdma/ib.h> #include <rdma/uverbs_std_types.h> #include <rdma/rdma_netlink.h> +#include <rdma/ib_ucaps.h> #include "uverbs.h" #include "core_priv.h" @@ -1345,6 +1346,7 @@ static void __exit ib_uverbs_cleanup(void) IB_UVERBS_NUM_FIXED_MINOR); unregister_chrdev_region(dynamic_uverbs_dev, IB_UVERBS_NUM_DYNAMIC_MINOR); + ib_cleanup_ucaps(); mmu_notifier_synchronize(); } diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c index 11a080646916..e803f609ec87 100644 --- a/drivers/infiniband/core/uverbs_marshall.c +++ b/drivers/infiniband/core/uverbs_marshall.c @@ -171,45 +171,3 @@ void ib_copy_path_rec_to_user(struct ib_user_path_rec *dst, __ib_copy_path_rec_to_user(dst, src); } EXPORT_SYMBOL(ib_copy_path_rec_to_user); - -void ib_copy_path_rec_from_user(struct sa_path_rec *dst, - struct ib_user_path_rec *src) -{ - u32 slid, dlid; - - memset(dst, 0, sizeof(*dst)); - if ((ib_is_opa_gid((union ib_gid *)src->sgid)) || - (ib_is_opa_gid((union ib_gid *)src->dgid))) { - dst->rec_type = SA_PATH_REC_TYPE_OPA; - slid = opa_get_lid_from_gid((union ib_gid *)src->sgid); - dlid = opa_get_lid_from_gid((union ib_gid *)src->dgid); - } else { - dst->rec_type = SA_PATH_REC_TYPE_IB; - slid = ntohs(src->slid); - dlid = ntohs(src->dlid); - } - memcpy(dst->dgid.raw, src->dgid, sizeof dst->dgid); - memcpy(dst->sgid.raw, src->sgid, sizeof dst->sgid); - - sa_path_set_dlid(dst, dlid); - sa_path_set_slid(dst, slid); - sa_path_set_raw_traffic(dst, src->raw_traffic); - dst->flow_label = src->flow_label; - dst->hop_limit = src->hop_limit; - dst->traffic_class = src->traffic_class; - dst->reversible = src->reversible; - dst->numb_path = src->numb_path; - dst->pkey = src->pkey; - dst->sl = src->sl; - dst->mtu_selector = src->mtu_selector; - dst->mtu = src->mtu; - dst->rate_selector = src->rate_selector; - dst->rate = src->rate; - dst->packet_life_time = src->packet_life_time; - dst->preference = src->preference; - dst->packet_life_time_selector = src->packet_life_time_selector; - - /* TODO: No need to set this */ - sa_path_set_dmac_zero(dst); -} -EXPORT_SYMBOL(ib_copy_path_rec_from_user); diff --git a/drivers/infiniband/core/uverbs_std_types_cq.c b/drivers/infiniband/core/uverbs_std_types_cq.c index 432054f0a8a4..fab5d914029d 100644 --- a/drivers/infiniband/core/uverbs_std_types_cq.c +++ b/drivers/infiniband/core/uverbs_std_types_cq.c @@ -64,15 +64,21 @@ static int UVERBS_HANDLER(UVERBS_METHOD_CQ_CREATE)( struct ib_ucq_object *obj = container_of( uverbs_attr_get_uobject(attrs, UVERBS_ATTR_CREATE_CQ_HANDLE), typeof(*obj), uevent.uobject); + struct ib_uverbs_completion_event_file *ev_file = NULL; struct ib_device *ib_dev = attrs->context->device; - int ret; - u64 user_handle; + struct ib_umem_dmabuf *umem_dmabuf; struct ib_cq_init_attr attr = {}; - struct ib_cq *cq; - struct ib_uverbs_completion_event_file *ev_file = NULL; struct ib_uobject *ev_file_uobj; + struct ib_umem *umem = NULL; + u64 buffer_length; + u64 buffer_offset; + struct ib_cq *cq; + u64 user_handle; + u64 buffer_va; + int buffer_fd; + int ret; - if (!ib_dev->ops.create_cq || !ib_dev->ops.destroy_cq) + if ((!ib_dev->ops.create_cq && !ib_dev->ops.create_cq_umem) || !ib_dev->ops.destroy_cq) return -EOPNOTSUPP; ret = uverbs_copy_from(&attr.comp_vector, attrs, @@ -112,9 +118,66 @@ static int UVERBS_HANDLER(UVERBS_METHOD_CQ_CREATE)( INIT_LIST_HEAD(&obj->comp_list); INIT_LIST_HEAD(&obj->uevent.event_list); + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_VA)) { + + ret = uverbs_copy_from(&buffer_va, attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_VA); + if (ret) + goto err_event_file; + + ret = uverbs_copy_from(&buffer_length, attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_LENGTH); + if (ret) + goto err_event_file; + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_FD) || + uverbs_attr_is_valid(attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_OFFSET) || + !ib_dev->ops.create_cq_umem) { + ret = -EINVAL; + goto err_event_file; + } + + umem = ib_umem_get(ib_dev, buffer_va, buffer_length, IB_ACCESS_LOCAL_WRITE); + if (IS_ERR(umem)) { + ret = PTR_ERR(umem); + goto err_event_file; + } + } else if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_FD)) { + + ret = uverbs_get_raw_fd(&buffer_fd, attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_FD); + if (ret) + goto err_event_file; + + ret = uverbs_copy_from(&buffer_offset, attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_OFFSET); + if (ret) + goto err_event_file; + + ret = uverbs_copy_from(&buffer_length, attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_LENGTH); + if (ret) + goto err_event_file; + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_VA) || + !ib_dev->ops.create_cq_umem) { + ret = -EINVAL; + goto err_event_file; + } + + umem_dmabuf = ib_umem_dmabuf_get_pinned(ib_dev, buffer_offset, buffer_length, + buffer_fd, IB_ACCESS_LOCAL_WRITE); + if (IS_ERR(umem_dmabuf)) { + ret = PTR_ERR(umem_dmabuf); + goto err_event_file; + } + umem = &umem_dmabuf->umem; + } else if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_OFFSET) || + uverbs_attr_is_valid(attrs, UVERBS_ATTR_CREATE_CQ_BUFFER_LENGTH) || + !ib_dev->ops.create_cq) { + ret = -EINVAL; + goto err_event_file; + } + cq = rdma_zalloc_drv_obj(ib_dev, ib_cq); if (!cq) { ret = -ENOMEM; + ib_umem_release(umem); goto err_event_file; } @@ -128,7 +191,8 @@ static int UVERBS_HANDLER(UVERBS_METHOD_CQ_CREATE)( rdma_restrack_new(&cq->res, RDMA_RESTRACK_CQ); rdma_restrack_set_name(&cq->res, NULL); - ret = ib_dev->ops.create_cq(cq, &attr, attrs); + ret = umem ? ib_dev->ops.create_cq_umem(cq, &attr, umem, attrs) : + ib_dev->ops.create_cq(cq, &attr, attrs); if (ret) goto err_free; @@ -142,6 +206,7 @@ static int UVERBS_HANDLER(UVERBS_METHOD_CQ_CREATE)( return ret; err_free: + ib_umem_release(umem); rdma_restrack_put(&cq->res); kfree(cq); err_event_file: @@ -180,6 +245,17 @@ DECLARE_UVERBS_NAMED_METHOD( UVERBS_OBJECT_ASYNC_EVENT, UVERBS_ACCESS_READ, UA_OPTIONAL), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_CREATE_CQ_BUFFER_VA, + UVERBS_ATTR_TYPE(u64), + UA_OPTIONAL), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_CREATE_CQ_BUFFER_LENGTH, + UVERBS_ATTR_TYPE(u64), + UA_OPTIONAL), + UVERBS_ATTR_RAW_FD(UVERBS_ATTR_CREATE_CQ_BUFFER_FD, + UA_OPTIONAL), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_CREATE_CQ_BUFFER_OFFSET, + UVERBS_ATTR_TYPE(u64), + UA_OPTIONAL), UVERBS_ATTR_UHW()); static int UVERBS_HANDLER(UVERBS_METHOD_CQ_DESTROY)( diff --git a/drivers/infiniband/core/uverbs_std_types_device.c b/drivers/infiniband/core/uverbs_std_types_device.c index fb0555647336..c0fd283d9d6c 100644 --- a/drivers/infiniband/core/uverbs_std_types_device.c +++ b/drivers/infiniband/core/uverbs_std_types_device.c @@ -437,6 +437,10 @@ DECLARE_UVERBS_NAMED_METHOD( UVERBS_ATTR_TYPE(u32), UA_OPTIONAL), UVERBS_ATTR_PTR_OUT(UVERBS_ATTR_GET_CONTEXT_CORE_SUPPORT, UVERBS_ATTR_TYPE(u64), UA_OPTIONAL), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_GET_CONTEXT_FD_ARR, + UVERBS_ATTR_MIN_SIZE(sizeof(int)), + UA_OPTIONAL, + UA_ALLOC_AND_COPY), UVERBS_ATTR_UHW()); DECLARE_UVERBS_NAMED_METHOD( diff --git a/drivers/infiniband/core/uverbs_std_types_dmah.c b/drivers/infiniband/core/uverbs_std_types_dmah.c new file mode 100644 index 000000000000..453ce656c6f2 --- /dev/null +++ b/drivers/infiniband/core/uverbs_std_types_dmah.c @@ -0,0 +1,145 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* + * Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved + */ + +#include "rdma_core.h" +#include "uverbs.h" +#include <rdma/uverbs_std_types.h> +#include "restrack.h" + +static int uverbs_free_dmah(struct ib_uobject *uobject, + enum rdma_remove_reason why, + struct uverbs_attr_bundle *attrs) +{ + struct ib_dmah *dmah = uobject->object; + int ret; + + if (atomic_read(&dmah->usecnt)) + return -EBUSY; + + ret = dmah->device->ops.dealloc_dmah(dmah, attrs); + if (ret) + return ret; + + rdma_restrack_del(&dmah->res); + kfree(dmah); + return 0; +} + +static int UVERBS_HANDLER(UVERBS_METHOD_DMAH_ALLOC)( + struct uverbs_attr_bundle *attrs) +{ + struct ib_uobject *uobj = + uverbs_attr_get(attrs, UVERBS_ATTR_ALLOC_DMAH_HANDLE) + ->obj_attr.uobject; + struct ib_device *ib_dev = attrs->context->device; + struct ib_dmah *dmah; + int ret; + + dmah = rdma_zalloc_drv_obj(ib_dev, ib_dmah); + if (!dmah) + return -ENOMEM; + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_ALLOC_DMAH_CPU_ID)) { + ret = uverbs_copy_from(&dmah->cpu_id, attrs, + UVERBS_ATTR_ALLOC_DMAH_CPU_ID); + if (ret) + goto err; + + if (!cpumask_test_cpu(dmah->cpu_id, current->cpus_ptr)) { + ret = -EPERM; + goto err; + } + + dmah->valid_fields |= BIT(IB_DMAH_CPU_ID_EXISTS); + } + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_ALLOC_DMAH_TPH_MEM_TYPE)) { + dmah->mem_type = uverbs_attr_get_enum_id(attrs, + UVERBS_ATTR_ALLOC_DMAH_TPH_MEM_TYPE); + dmah->valid_fields |= BIT(IB_DMAH_MEM_TYPE_EXISTS); + } + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_ALLOC_DMAH_PH)) { + ret = uverbs_copy_from(&dmah->ph, attrs, + UVERBS_ATTR_ALLOC_DMAH_PH); + if (ret) + goto err; + + /* Per PCIe spec 6.2-1.0, only the lowest two bits are applicable */ + if (dmah->ph & 0xFC) { + ret = -EINVAL; + goto err; + } + + dmah->valid_fields |= BIT(IB_DMAH_PH_EXISTS); + } + + dmah->device = ib_dev; + dmah->uobject = uobj; + atomic_set(&dmah->usecnt, 0); + + rdma_restrack_new(&dmah->res, RDMA_RESTRACK_DMAH); + rdma_restrack_set_name(&dmah->res, NULL); + + ret = ib_dev->ops.alloc_dmah(dmah, attrs); + if (ret) { + rdma_restrack_put(&dmah->res); + goto err; + } + + uobj->object = dmah; + rdma_restrack_add(&dmah->res); + uverbs_finalize_uobj_create(attrs, UVERBS_ATTR_ALLOC_DMAH_HANDLE); + return 0; +err: + kfree(dmah); + return ret; +} + +static const struct uverbs_attr_spec uverbs_dmah_mem_type[] = { + [TPH_MEM_TYPE_VM] = { + .type = UVERBS_ATTR_TYPE_PTR_IN, + UVERBS_ATTR_NO_DATA(), + }, + [TPH_MEM_TYPE_PM] = { + .type = UVERBS_ATTR_TYPE_PTR_IN, + UVERBS_ATTR_NO_DATA(), + }, +}; + +DECLARE_UVERBS_NAMED_METHOD( + UVERBS_METHOD_DMAH_ALLOC, + UVERBS_ATTR_IDR(UVERBS_ATTR_ALLOC_DMAH_HANDLE, + UVERBS_OBJECT_DMAH, + UVERBS_ACCESS_NEW, + UA_MANDATORY), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_ALLOC_DMAH_CPU_ID, + UVERBS_ATTR_TYPE(u32), + UA_OPTIONAL), + UVERBS_ATTR_ENUM_IN(UVERBS_ATTR_ALLOC_DMAH_TPH_MEM_TYPE, + uverbs_dmah_mem_type, + UA_OPTIONAL), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_ALLOC_DMAH_PH, + UVERBS_ATTR_TYPE(u8), + UA_OPTIONAL)); + +DECLARE_UVERBS_NAMED_METHOD_DESTROY( + UVERBS_METHOD_DMAH_FREE, + UVERBS_ATTR_IDR(UVERBS_ATTR_FREE_DMA_HANDLE, + UVERBS_OBJECT_DMAH, + UVERBS_ACCESS_DESTROY, + UA_MANDATORY)); + +DECLARE_UVERBS_NAMED_OBJECT(UVERBS_OBJECT_DMAH, + UVERBS_TYPE_ALLOC_IDR(uverbs_free_dmah), + &UVERBS_METHOD(UVERBS_METHOD_DMAH_ALLOC), + &UVERBS_METHOD(UVERBS_METHOD_DMAH_FREE)); + +const struct uapi_definition uverbs_def_obj_dmah[] = { + UAPI_DEF_CHAIN_OBJ_TREE_NAMED(UVERBS_OBJECT_DMAH, + UAPI_DEF_OBJ_NEEDS_FN(dealloc_dmah), + UAPI_DEF_OBJ_NEEDS_FN(alloc_dmah)), + {} +}; diff --git a/drivers/infiniband/core/uverbs_std_types_mr.c b/drivers/infiniband/core/uverbs_std_types_mr.c index 7ebc7bd3caae..570b9656801d 100644 --- a/drivers/infiniband/core/uverbs_std_types_mr.c +++ b/drivers/infiniband/core/uverbs_std_types_mr.c @@ -238,7 +238,7 @@ static int UVERBS_HANDLER(UVERBS_METHOD_REG_DMABUF_MR)( return ret; mr = pd->device->ops.reg_user_mr_dmabuf(pd, offset, length, iova, fd, - access_flags, + access_flags, NULL, attrs); if (IS_ERR(mr)) return PTR_ERR(mr); @@ -266,6 +266,135 @@ static int UVERBS_HANDLER(UVERBS_METHOD_REG_DMABUF_MR)( return ret; } +static int UVERBS_HANDLER(UVERBS_METHOD_REG_MR)( + struct uverbs_attr_bundle *attrs) +{ + struct ib_uobject *uobj = + uverbs_attr_get_uobject(attrs, UVERBS_ATTR_REG_MR_HANDLE); + struct ib_pd *pd = + uverbs_attr_get_obj(attrs, UVERBS_ATTR_REG_MR_PD_HANDLE); + u32 valid_access_flags = IB_ACCESS_SUPPORTED; + u64 length, iova, fd_offset = 0, addr = 0; + struct ib_device *ib_dev = pd->device; + struct ib_dmah *dmah = NULL; + bool has_fd_offset = false; + bool has_addr = false; + bool has_fd = false; + u32 access_flags; + struct ib_mr *mr; + int fd; + int ret; + + ret = uverbs_copy_from(&iova, attrs, UVERBS_ATTR_REG_MR_IOVA); + if (ret) + return ret; + + ret = uverbs_copy_from(&length, attrs, UVERBS_ATTR_REG_MR_LENGTH); + if (ret) + return ret; + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_REG_MR_ADDR)) { + ret = uverbs_copy_from(&addr, attrs, + UVERBS_ATTR_REG_MR_ADDR); + if (ret) + return ret; + has_addr = true; + } + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_REG_MR_FD_OFFSET)) { + ret = uverbs_copy_from(&fd_offset, attrs, + UVERBS_ATTR_REG_MR_FD_OFFSET); + if (ret) + return ret; + has_fd_offset = true; + } + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_REG_MR_FD)) { + ret = uverbs_get_raw_fd(&fd, attrs, + UVERBS_ATTR_REG_MR_FD); + if (ret) + return ret; + has_fd = true; + } + + if (has_fd) { + if (!ib_dev->ops.reg_user_mr_dmabuf) + return -EOPNOTSUPP; + + /* FD requires offset and can't come with addr */ + if (!has_fd_offset || has_addr) + return -EINVAL; + + if ((fd_offset & ~PAGE_MASK) != (iova & ~PAGE_MASK)) + return -EINVAL; + + valid_access_flags = IB_ACCESS_LOCAL_WRITE | + IB_ACCESS_REMOTE_READ | + IB_ACCESS_REMOTE_WRITE | + IB_ACCESS_REMOTE_ATOMIC | + IB_ACCESS_RELAXED_ORDERING; + } else { + if (!has_addr || has_fd_offset) + return -EINVAL; + + if ((addr & ~PAGE_MASK) != (iova & ~PAGE_MASK)) + return -EINVAL; + } + + if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_REG_MR_DMA_HANDLE)) { + dmah = uverbs_attr_get_obj(attrs, + UVERBS_ATTR_REG_MR_DMA_HANDLE); + if (IS_ERR(dmah)) + return PTR_ERR(dmah); + } + + ret = uverbs_get_flags32(&access_flags, attrs, + UVERBS_ATTR_REG_MR_ACCESS_FLAGS, + valid_access_flags); + if (ret) + return ret; + + ret = ib_check_mr_access(ib_dev, access_flags); + if (ret) + return ret; + + if (has_fd) + mr = pd->device->ops.reg_user_mr_dmabuf(pd, fd_offset, length, + iova, fd, access_flags, + dmah, attrs); + else + mr = pd->device->ops.reg_user_mr(pd, addr, length, iova, + access_flags, dmah, NULL); + + if (IS_ERR(mr)) + return PTR_ERR(mr); + + mr->device = pd->device; + mr->pd = pd; + mr->type = IB_MR_TYPE_USER; + mr->uobject = uobj; + atomic_inc(&pd->usecnt); + if (dmah) { + mr->dmah = dmah; + atomic_inc(&dmah->usecnt); + } + rdma_restrack_new(&mr->res, RDMA_RESTRACK_MR); + rdma_restrack_set_name(&mr->res, NULL); + rdma_restrack_add(&mr->res); + uobj->object = mr; + + uverbs_finalize_uobj_create(attrs, UVERBS_ATTR_REG_MR_HANDLE); + + ret = uverbs_copy_to(attrs, UVERBS_ATTR_REG_MR_RESP_LKEY, + &mr->lkey, sizeof(mr->lkey)); + if (ret) + return ret; + + ret = uverbs_copy_to(attrs, UVERBS_ATTR_REG_MR_RESP_RKEY, + &mr->rkey, sizeof(mr->rkey)); + return ret; +} + DECLARE_UVERBS_NAMED_METHOD( UVERBS_METHOD_ADVISE_MR, UVERBS_ATTR_IDR(UVERBS_ATTR_ADVISE_MR_PD_HANDLE, @@ -362,6 +491,44 @@ DECLARE_UVERBS_NAMED_METHOD( UVERBS_ATTR_TYPE(u32), UA_MANDATORY)); +DECLARE_UVERBS_NAMED_METHOD( + UVERBS_METHOD_REG_MR, + UVERBS_ATTR_IDR(UVERBS_ATTR_REG_MR_HANDLE, + UVERBS_OBJECT_MR, + UVERBS_ACCESS_NEW, + UA_MANDATORY), + UVERBS_ATTR_IDR(UVERBS_ATTR_REG_MR_PD_HANDLE, + UVERBS_OBJECT_PD, + UVERBS_ACCESS_READ, + UA_MANDATORY), + UVERBS_ATTR_IDR(UVERBS_ATTR_REG_MR_DMA_HANDLE, + UVERBS_OBJECT_DMAH, + UVERBS_ACCESS_READ, + UA_OPTIONAL), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_REG_MR_IOVA, + UVERBS_ATTR_TYPE(u64), + UA_MANDATORY), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_REG_MR_LENGTH, + UVERBS_ATTR_TYPE(u64), + UA_MANDATORY), + UVERBS_ATTR_FLAGS_IN(UVERBS_ATTR_REG_MR_ACCESS_FLAGS, + enum ib_access_flags, + UA_MANDATORY), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_REG_MR_ADDR, + UVERBS_ATTR_TYPE(u64), + UA_OPTIONAL), + UVERBS_ATTR_PTR_IN(UVERBS_ATTR_REG_MR_FD_OFFSET, + UVERBS_ATTR_TYPE(u64), + UA_OPTIONAL), + UVERBS_ATTR_RAW_FD(UVERBS_ATTR_REG_MR_FD, + UA_OPTIONAL), + UVERBS_ATTR_PTR_OUT(UVERBS_ATTR_REG_MR_RESP_LKEY, + UVERBS_ATTR_TYPE(u32), + UA_MANDATORY), + UVERBS_ATTR_PTR_OUT(UVERBS_ATTR_REG_MR_RESP_RKEY, + UVERBS_ATTR_TYPE(u32), + UA_MANDATORY)); + DECLARE_UVERBS_NAMED_METHOD_DESTROY( UVERBS_METHOD_MR_DESTROY, UVERBS_ATTR_IDR(UVERBS_ATTR_DESTROY_MR_HANDLE, @@ -376,7 +543,8 @@ DECLARE_UVERBS_NAMED_OBJECT( &UVERBS_METHOD(UVERBS_METHOD_DM_MR_REG), &UVERBS_METHOD(UVERBS_METHOD_MR_DESTROY), &UVERBS_METHOD(UVERBS_METHOD_QUERY_MR), - &UVERBS_METHOD(UVERBS_METHOD_REG_DMABUF_MR)); + &UVERBS_METHOD(UVERBS_METHOD_REG_DMABUF_MR), + &UVERBS_METHOD(UVERBS_METHOD_REG_MR)); const struct uapi_definition uverbs_def_obj_mr[] = { UAPI_DEF_CHAIN_OBJ_TREE_NAMED(UVERBS_OBJECT_MR, diff --git a/drivers/infiniband/core/uverbs_std_types_qp.c b/drivers/infiniband/core/uverbs_std_types_qp.c index 7b4773fa4bc0..be0730e8509e 100644 --- a/drivers/infiniband/core/uverbs_std_types_qp.c +++ b/drivers/infiniband/core/uverbs_std_types_qp.c @@ -133,7 +133,7 @@ static int UVERBS_HANDLER(UVERBS_METHOD_QP_CREATE)( device = xrcd->device; break; case IB_UVERBS_QPT_RAW_PACKET: - if (!capable(CAP_NET_RAW)) + if (!rdma_uattrs_has_raw_cap(attrs)) return -EPERM; fallthrough; case IB_UVERBS_QPT_RC: diff --git a/drivers/infiniband/core/uverbs_uapi.c b/drivers/infiniband/core/uverbs_uapi.c index a02916a3a79c..e00ea63175bd 100644 --- a/drivers/infiniband/core/uverbs_uapi.c +++ b/drivers/infiniband/core/uverbs_uapi.c @@ -631,6 +631,7 @@ static const struct uapi_definition uverbs_core_api[] = { UAPI_DEF_CHAIN(uverbs_def_obj_cq), UAPI_DEF_CHAIN(uverbs_def_obj_device), UAPI_DEF_CHAIN(uverbs_def_obj_dm), + UAPI_DEF_CHAIN(uverbs_def_obj_dmah), UAPI_DEF_CHAIN(uverbs_def_obj_flow_action), UAPI_DEF_CHAIN(uverbs_def_obj_intf), UAPI_DEF_CHAIN(uverbs_def_obj_mr), diff --git a/drivers/infiniband/core/verbs.c b/drivers/infiniband/core/verbs.c index 473ee0831307..11b1a194de44 100644 --- a/drivers/infiniband/core/verbs.c +++ b/drivers/infiniband/core/verbs.c @@ -148,6 +148,7 @@ __attribute_const__ int ib_rate_to_mult(enum ib_rate rate) case IB_RATE_400_GBPS: return 160; case IB_RATE_600_GBPS: return 240; case IB_RATE_800_GBPS: return 320; + case IB_RATE_1600_GBPS: return 640; default: return -1; } } @@ -178,6 +179,7 @@ __attribute_const__ enum ib_rate mult_to_ib_rate(int mult) case 160: return IB_RATE_400_GBPS; case 240: return IB_RATE_600_GBPS; case 320: return IB_RATE_800_GBPS; + case 640: return IB_RATE_1600_GBPS; default: return IB_RATE_PORT_CURRENT; } } @@ -208,6 +210,7 @@ __attribute_const__ int ib_rate_to_mbps(enum ib_rate rate) case IB_RATE_400_GBPS: return 425000; case IB_RATE_600_GBPS: return 637500; case IB_RATE_800_GBPS: return 850000; + case IB_RATE_1600_GBPS: return 1700000; default: return -1; } } @@ -572,7 +575,7 @@ struct ib_ah *rdma_create_ah(struct ib_pd *pd, struct rdma_ah_attr *ah_attr, GFP_KERNEL : GFP_ATOMIC); if (IS_ERR(slave)) { rdma_unfill_sgid_attr(ah_attr, old_sgid_attr); - return (void *)slave; + return ERR_CAST(slave); } ah = _rdma_create_ah(pd, ah_attr, flags, NULL, slave); rdma_lag_put_ah_roce_slave(slave); @@ -2105,7 +2108,7 @@ int ib_destroy_qp_user(struct ib_qp *qp, struct ib_udata *udata) if (!qp->uobject) rdma_rw_cleanup_mrs(qp); - rdma_counter_unbind_qp(qp, true); + rdma_counter_unbind_qp(qp, qp->port, true); ret = qp->device->ops.destroy_qp(qp, udata); if (ret) { if (sec) @@ -2223,7 +2226,7 @@ struct ib_mr *ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length, } mr = pd->device->ops.reg_user_mr(pd, start, length, virt_addr, - access_flags, NULL); + access_flags, NULL, NULL); if (IS_ERR(mr)) return mr; @@ -2262,6 +2265,7 @@ int ib_dereg_mr_user(struct ib_mr *mr, struct ib_udata *udata) { struct ib_pd *pd = mr->pd; struct ib_dm *dm = mr->dm; + struct ib_dmah *dmah = mr->dmah; struct ib_sig_attrs *sig_attrs = mr->sig_attrs; int ret; @@ -2272,6 +2276,8 @@ int ib_dereg_mr_user(struct ib_mr *mr, struct ib_udata *udata) atomic_dec(&pd->usecnt); if (dm) atomic_dec(&dm->usecnt); + if (dmah) + atomic_dec(&dmah->usecnt); kfree(sig_attrs); } @@ -3109,22 +3115,23 @@ EXPORT_SYMBOL(__rdma_block_iter_start); bool __rdma_block_iter_next(struct ib_block_iter *biter) { unsigned int block_offset; - unsigned int sg_delta; + unsigned int delta; if (!biter->__sg_nents || !biter->__sg) return false; biter->__dma_addr = sg_dma_address(biter->__sg) + biter->__sg_advance; block_offset = biter->__dma_addr & (BIT_ULL(biter->__pg_bit) - 1); - sg_delta = BIT_ULL(biter->__pg_bit) - block_offset; + delta = BIT_ULL(biter->__pg_bit) - block_offset; - if (sg_dma_len(biter->__sg) - biter->__sg_advance > sg_delta) { - biter->__sg_advance += sg_delta; - } else { + while (biter->__sg_nents && biter->__sg && + sg_dma_len(biter->__sg) - biter->__sg_advance <= delta) { + delta -= sg_dma_len(biter->__sg) - biter->__sg_advance; biter->__sg_advance = 0; biter->__sg = sg_next(biter->__sg); biter->__sg_nents--; } + biter->__sg_advance += delta; return true; } |