diff options
Diffstat (limited to 'drivers/infiniband/core')
52 files changed, 5182 insertions, 2744 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 65e3e7df8a4b..61596cda2b65 100644 --- a/drivers/infiniband/core/addr.c +++ b/drivers/infiniband/core/addr.c @@ -37,7 +37,6 @@ #include <linux/inetdevice.h> #include <linux/slab.h> #include <linux/workqueue.h> -#include <linux/module.h> #include <net/arp.h> #include <net/neighbour.h> #include <net/route.h> @@ -349,16 +348,10 @@ static int dst_fetch_ha(const struct dst_entry *dst, static bool has_gateway(const struct dst_entry *dst, sa_family_t family) { - struct rtable *rt; - struct rt6_info *rt6; - - if (family == AF_INET) { - rt = container_of(dst, struct rtable, dst); - return rt->rt_uses_gateway; - } + if (family == AF_INET) + return dst_rtable(dst)->rt_uses_gateway; - rt6 = container_of(dst, struct rt6_info, dst); - return rt6->rt6i_flags & RTF_GATEWAY; + return dst_rt6_info(dst)->rt6i_flags & RTF_GATEWAY; } static int fetch_ha(const struct dst_entry *dst, struct rdma_dev_addr *dev_addr, @@ -453,63 +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 @@ -519,9 +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) @@ -558,7 +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; @@ -595,7 +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(); /* @@ -603,7 +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 f82b4260de42..25a060a28301 100644 --- a/drivers/infiniband/core/agent.c +++ b/drivers/infiniband/core/agent.c @@ -59,7 +59,16 @@ __ib_get_agent_port(const struct ib_device *device, int port_num) struct ib_agent_port_private *entry; list_for_each_entry(entry, &ib_agent_port_list, port_list) { - if (entry->agent[1]->device == device && + /* Need to check both agent[0] and agent[1], as an agent port + * may only have one of them + */ + if (entry->agent[0] && + entry->agent[0]->device == device && + entry->agent[0]->port_num == port_num) + return entry; + + if (entry->agent[1] && + entry->agent[1]->device == device && entry->agent[1]->port_num == port_num) return entry; } @@ -101,8 +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; } @@ -172,14 +180,16 @@ int ib_agent_port_open(struct ib_device *device, int port_num) } } - /* Obtain send only MAD agent for GSI QP */ - port_priv->agent[1] = ib_register_mad_agent(device, port_num, - IB_QPT_GSI, NULL, 0, - &agent_send_handler, - NULL, NULL, 0); - if (IS_ERR(port_priv->agent[1])) { - ret = PTR_ERR(port_priv->agent[1]); - goto error3; + if (rdma_cap_ib_cm(device, port_num)) { + /* Obtain send only MAD agent for GSI QP */ + port_priv->agent[1] = ib_register_mad_agent(device, port_num, + IB_QPT_GSI, NULL, 0, + &agent_send_handler, + NULL, NULL, 0); + if (IS_ERR(port_priv->agent[1])) { + ret = PTR_ERR(port_priv->agent[1]); + goto error3; + } } spin_lock_irqsave(&ib_agent_port_list_lock, flags); @@ -212,7 +222,8 @@ int ib_agent_port_close(struct ib_device *device, int port_num) list_del(&port_priv->port_list); spin_unlock_irqrestore(&ib_agent_port_list_lock, flags); - ib_unregister_mad_agent(port_priv->agent[1]); + if (port_priv->agent[1]) + ib_unregister_mad_agent(port_priv->agent[1]); if (port_priv->agent[0]) ib_unregister_mad_agent(port_priv->agent[0]); diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c index c9e9fc81447e..81cf3c902e81 100644 --- a/drivers/infiniband/core/cache.c +++ b/drivers/infiniband/core/cache.c @@ -33,7 +33,7 @@ * SOFTWARE. */ -#include <linux/module.h> +#include <linux/if_vlan.h> #include <linux/errno.h> #include <linux/slab.h> #include <linux/workqueue.h> @@ -46,7 +46,7 @@ struct ib_pkey_cache { int table_len; - u16 table[]; + u16 table[] __counted_by(table_len); }; struct ib_update_work { @@ -400,6 +400,9 @@ static void del_gid(struct ib_device *ib_dev, u32 port, table->data_vec[ix] = NULL; write_unlock_irq(&table->rwlock); + if (rdma_cap_roce_gid_table(ib_dev, port)) + ib_dev->ops.del_gid(&entry->attr, &entry->context); + ndev_storage = entry->ndev_storage; if (ndev_storage) { entry->ndev_storage = NULL; @@ -407,9 +410,6 @@ static void del_gid(struct ib_device *ib_dev, u32 port, call_rcu(&ndev_storage->rcu_head, put_gid_ndev); } - if (rdma_cap_roce_gid_table(ib_dev, port)) - ib_dev->ops.del_gid(&entry->attr, &entry->context); - put_gid_entry_locked(entry); } @@ -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; } @@ -794,7 +794,6 @@ err_free_table: static void release_gid_table(struct ib_device *device, struct ib_gid_table *table) { - bool leak = false; int i; if (!table) @@ -803,15 +802,12 @@ static void release_gid_table(struct ib_device *device, for (i = 0; i < table->sz; i++) { if (is_gid_entry_free(table->data_vec[i])) continue; - if (kref_read(&table->data_vec[i]->kref) > 1) { - dev_err(&device->dev, - "GID entry ref leak for index %d ref=%u\n", i, - kref_read(&table->data_vec[i]->kref)); - leak = true; - } + + WARN_ONCE(true, + "GID entry ref leak for dev %s index %d ref=%u\n", + dev_name(&device->dev), i, + kref_read(&table->data_vec[i]->kref)); } - if (leak) - return; mutex_destroy(&table->lock); kfree(table->data_vec); @@ -955,7 +951,7 @@ int rdma_query_gid(struct ib_device *device, u32 port_num, { struct ib_gid_table *table; unsigned long flags; - int res = -EINVAL; + int res; if (!rdma_is_port_valid(device, port_num)) return -EINVAL; @@ -963,9 +959,15 @@ int rdma_query_gid(struct ib_device *device, u32 port_num, table = rdma_gid_table(device, port_num); read_lock_irqsave(&table->rwlock, flags); - if (index < 0 || index >= table->sz || - !is_gid_entry_valid(table->data_vec[index])) + if (index < 0 || index >= table->sz) { + res = -EINVAL; + goto done; + } + + if (!is_gid_entry_valid(table->data_vec[index])) { + res = -ENOENT; goto done; + } memcpy(gid, &table->data_vec[index]->attr.gid, sizeof(*gid)); res = 0; @@ -1125,41 +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; @@ -1416,7 +1383,7 @@ int rdma_read_gid_l2_fields(const struct ib_gid_attr *attr, *vlan_id = vlan_dev_vlan_id(ndev); } else { /* If the netdev is upper device and if it's lower - * device is vlan device, consider vlan id of the + * device is vlan device, consider vlan id of * the lower vlan device for this gid entry. */ netdev_walk_all_lower_dev_rcu(attr->ndev, @@ -1429,7 +1396,7 @@ int rdma_read_gid_l2_fields(const struct ib_gid_attr *attr, EXPORT_SYMBOL(rdma_read_gid_l2_fields); static int config_non_roce_gid_cache(struct ib_device *device, - u32 port, int gid_tbl_len) + u32 port, struct ib_port_attr *tprops) { struct ib_gid_attr gid_attr = {}; struct ib_gid_table *table; @@ -1441,7 +1408,7 @@ static int config_non_roce_gid_cache(struct ib_device *device, table = rdma_gid_table(device, port); mutex_lock(&table->lock); - for (i = 0; i < gid_tbl_len; ++i) { + for (i = 0; i < tprops->gid_tbl_len; ++i) { if (!device->ops.query_gid) continue; ret = device->ops.query_gid(device, port, i, &gid_attr.gid); @@ -1451,7 +1418,20 @@ static int config_non_roce_gid_cache(struct ib_device *device, i); goto err; } + + if (rdma_protocol_iwarp(device, port)) { + struct net_device *ndev; + + ndev = ib_device_get_netdev(device, port); + if (!ndev) + continue; + RCU_INIT_POINTER(gid_attr.ndev, ndev); + dev_put(ndev); + } + gid_attr.index = i; + tprops->subnet_prefix = + be64_to_cpu(gid_attr.gid.global.subnet_prefix); add_modify_gid(table, &gid_attr); } err: @@ -1484,7 +1464,7 @@ ib_cache_update(struct ib_device *device, u32 port, bool update_gids, if (!rdma_protocol_roce(device, port) && update_gids) { ret = config_non_roce_gid_cache(device, port, - tprops->gid_tbl_len); + tprops); if (ret) goto err; } @@ -1521,6 +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; @@ -1619,16 +1605,16 @@ int ib_cache_setup_one(struct ib_device *device) u32 p; int err; - rwlock_init(&device->cache_lock); - err = gid_table_setup_one(device); if (err) return err; rdma_for_each_port (device, p) { err = ib_cache_update(device, p, true, true, true); - if (err) + if (err) { + gid_table_cleanup_one(device); return err; + } } return 0; diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c index c903b74f46a4..024df6ee239d 100644 --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -34,6 +34,9 @@ MODULE_AUTHOR("Sean Hefty"); MODULE_DESCRIPTION("InfiniBand CM"); MODULE_LICENSE("Dual BSD/GPL"); +#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", [IB_CM_REJ_NO_EEC] = "no EEC", @@ -92,8 +95,7 @@ static void cm_process_work(struct cm_id_private *cm_id_priv, struct cm_work *work); static int cm_send_sidr_rep_locked(struct cm_id_private *cm_id_priv, struct ib_cm_sidr_rep_param *param); -static int cm_send_dreq_locked(struct cm_id_private *cm_id_priv, - const void *private_data, u8 private_data_len); +static void cm_issue_dreq(struct cm_id_private *cm_id_priv); static int cm_send_drep_locked(struct cm_id_private *cm_id_priv, void *private_data, u8 private_data_len); static int cm_send_rej_locked(struct cm_id_private *cm_id_priv, @@ -158,6 +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]; }; @@ -165,7 +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; @@ -175,6 +178,7 @@ struct cm_device { struct cm_av { struct cm_port *port; struct rdma_ah_attr ah_attr; + u16 dlid_datapath; u16 pkey_index; u8 timeout; }; @@ -238,7 +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; @@ -271,7 +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; @@ -282,8 +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; @@ -305,42 +310,54 @@ static struct ib_mad_send_buf *cm_alloc_msg(struct cm_id_private *cm_id_priv) goto out; } - /* Timeout set by caller if response is expected. */ m->ah = ah; - m->retries = cm_id_priv->max_cm_retries; - - refcount_inc(&cm_id_priv->refcount); - m->context[0] = cm_id_priv; 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 void cm_free_msg(struct ib_mad_send_buf *msg) +static struct ib_mad_send_buf *cm_alloc_msg(struct cm_id_private *cm_id_priv) { - struct cm_id_private *cm_id_priv = msg->context[0]; + return cm_alloc_msg_agent(cm_id_priv, false); +} +static void cm_free_msg(struct ib_mad_send_buf *msg) +{ if (msg->ah) rdma_destroy_ah(msg->ah, 0); - cm_deref_id(cm_id_priv); ib_free_send_mad(msg); } static struct ib_mad_send_buf * -cm_alloc_priv_msg(struct cm_id_private *cm_id_priv) +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; + cm_id_priv->msg = msg; + refcount_inc(&cm_id_priv->refcount); + msg->context[0] = cm_id_priv; + msg->context[1] = (void *) (unsigned long) state; + + msg->retries = cm_id_priv->max_cm_retries; + msg->timeout_ms = cm_id_priv->timeout_ms; + 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]; @@ -356,13 +373,20 @@ static void cm_free_priv_msg(struct ib_mad_send_buf *msg) ib_free_send_mad(msg); } -static struct ib_mad_send_buf *cm_alloc_response_msg_no_ah(struct cm_port *port, - struct ib_mad_recv_wc *mad_recv_wc) +static struct ib_mad_send_buf * +cm_alloc_response_msg_no_ah(struct cm_port *port, + struct ib_mad_recv_wc *mad_recv_wc, + bool direct_retry) { - return ib_create_send_mad(port->mad_agent, 1, mad_recv_wc->wc->pkey_index, - 0, IB_MGMT_MAD_HDR, IB_MGMT_MAD_DATA, - GFP_ATOMIC, - IB_MGMT_BASE_VERSION); + struct ib_mad_send_buf *m; + + m = ib_create_send_mad(port->mad_agent, 1, mad_recv_wc->wc->pkey_index, + 0, IB_MGMT_MAD_HDR, IB_MGMT_MAD_DATA, + GFP_ATOMIC, IB_MGMT_BASE_VERSION); + if (!IS_ERR(m)) + m->context[0] = direct_retry ? CM_DIRECT_RETRY_CTX : NULL; + + return m; } static int cm_create_response_msg_ah(struct cm_port *port, @@ -382,12 +406,13 @@ static int cm_create_response_msg_ah(struct cm_port *port, static int cm_alloc_response_msg(struct cm_port *port, struct ib_mad_recv_wc *mad_recv_wc, + bool direct_retry, struct ib_mad_send_buf **msg) { struct ib_mad_send_buf *m; int ret; - m = cm_alloc_response_msg_no_ah(port, mad_recv_wc); + m = cm_alloc_response_msg_no_ah(port, mad_recv_wc, direct_retry); if (IS_ERR(m)) return PTR_ERR(m); @@ -401,13 +426,6 @@ static int cm_alloc_response_msg(struct cm_port *port, return 0; } -static void cm_free_response_msg(struct ib_mad_send_buf *msg) -{ - if (msg->ah) - rdma_destroy_ah(msg->ah, 0); - ib_free_send_mad(msg); -} - static void *cm_copy_private_data(const void *private_data, u8 private_data_len) { void *data; @@ -617,7 +635,6 @@ static struct cm_id_private *cm_insert_listen(struct cm_id_private *cm_id_priv, struct rb_node *parent = NULL; struct cm_id_private *cur_cm_id_priv; __be64 service_id = cm_id_priv->id.service_id; - __be64 service_mask = cm_id_priv->id.service_mask; unsigned long flags; spin_lock_irqsave(&cm.lock, flags); @@ -625,9 +642,16 @@ static struct cm_id_private *cm_insert_listen(struct cm_id_private *cm_id_priv, parent = *link; cur_cm_id_priv = rb_entry(parent, struct cm_id_private, service_node); - if ((cur_cm_id_priv->id.service_mask & service_id) == - (service_mask & cur_cm_id_priv->id.service_id) && - (cm_id_priv->id.device == cur_cm_id_priv->id.device)) { + + if (cm_id_priv->id.device < cur_cm_id_priv->id.device) + link = &(*link)->rb_left; + else if (cm_id_priv->id.device > cur_cm_id_priv->id.device) + link = &(*link)->rb_right; + else if (be64_lt(service_id, cur_cm_id_priv->id.service_id)) + link = &(*link)->rb_left; + else if (be64_gt(service_id, cur_cm_id_priv->id.service_id)) + link = &(*link)->rb_right; + else { /* * Sharing an ib_cm_id with different handlers is not * supported @@ -643,17 +667,6 @@ static struct cm_id_private *cm_insert_listen(struct cm_id_private *cm_id_priv, spin_unlock_irqrestore(&cm.lock, flags); return cur_cm_id_priv; } - - if (cm_id_priv->id.device < cur_cm_id_priv->id.device) - link = &(*link)->rb_left; - else if (cm_id_priv->id.device > cur_cm_id_priv->id.device) - link = &(*link)->rb_right; - else if (be64_lt(service_id, cur_cm_id_priv->id.service_id)) - link = &(*link)->rb_left; - else if (be64_gt(service_id, cur_cm_id_priv->id.service_id)) - link = &(*link)->rb_right; - else - link = &(*link)->rb_right; } cm_id_priv->listen_sharecount++; rb_link_node(&cm_id_priv->service_node, parent, link); @@ -670,12 +683,7 @@ static struct cm_id_private *cm_find_listen(struct ib_device *device, while (node) { cm_id_priv = rb_entry(node, struct cm_id_private, service_node); - if ((cm_id_priv->id.service_mask & service_id) == - cm_id_priv->id.service_id && - (cm_id_priv->id.device == device)) { - refcount_inc(&cm_id_priv->refcount); - return cm_id_priv; - } + if (device < cm_id_priv->id.device) node = node->rb_left; else if (device > cm_id_priv->id.device) @@ -684,8 +692,10 @@ static struct cm_id_private *cm_find_listen(struct ib_device *device, node = node->rb_left; else if (be64_gt(service_id, cm_id_priv->id.service_id)) node = node->rb_right; - else - node = node->rb_right; + else { + refcount_inc(&cm_id_priv->refcount); + return cm_id_priv; + } } return NULL; } @@ -1032,13 +1042,27 @@ static void cm_reset_to_idle(struct cm_id_private *cm_id_priv) } } +static noinline void cm_destroy_id_wait_timeout(struct ib_cm_id *cm_id, + enum ib_cm_state old_state) +{ + struct cm_id_private *cm_id_priv; + + cm_id_priv = container_of(cm_id, struct cm_id_private, id); + 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; cm_id_priv = container_of(cm_id, struct cm_id_private, id); spin_lock_irq(&cm_id_priv->lock); + old_state = cm_id->state; retest: switch (cm_id->state) { case IB_CM_LISTEN: @@ -1102,7 +1126,8 @@ retest: cm_id->state = IB_CM_IDLE; break; } - cm_send_dreq_locked(cm_id_priv, NULL, 0); + cm_issue_dreq(cm_id_priv); + cm_enter_timewait(cm_id_priv); goto retest; case IB_CM_DREQ_SENT: ib_cancel_mad(cm_id_priv->msg); @@ -1142,7 +1167,13 @@ retest: xa_erase(&cm.local_id_table, cm_local_id(cm_id->local_id)); cm_deref_id(cm_id_priv); - wait_for_completion(&cm_id_priv->comp); + 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, timeout); + if (!ret) /* timeout happened */ + cm_destroy_id_wait_timeout(cm_id, old_state); + } while (!ret); + while ((work = cm_dequeue_work(cm_id_priv)) != NULL) cm_free_work(work); @@ -1158,22 +1189,17 @@ void ib_destroy_cm_id(struct ib_cm_id *cm_id) } EXPORT_SYMBOL(ib_destroy_cm_id); -static int cm_init_listen(struct cm_id_private *cm_id_priv, __be64 service_id, - __be64 service_mask) +static int cm_init_listen(struct cm_id_private *cm_id_priv, __be64 service_id) { - service_mask = service_mask ? service_mask : ~cpu_to_be64(0); - service_id &= service_mask; if ((service_id & IB_SERVICE_ID_AGN_MASK) == IB_CM_ASSIGN_SERVICE_ID && (service_id != IB_CM_ASSIGN_SERVICE_ID)) return -EINVAL; - if (service_id == IB_CM_ASSIGN_SERVICE_ID) { + if (service_id == IB_CM_ASSIGN_SERVICE_ID) cm_id_priv->id.service_id = cpu_to_be64(cm.listen_service_id++); - cm_id_priv->id.service_mask = ~cpu_to_be64(0); - } else { + else cm_id_priv->id.service_id = service_id; - cm_id_priv->id.service_mask = service_mask; - } + return 0; } @@ -1185,12 +1211,8 @@ static int cm_init_listen(struct cm_id_private *cm_id_priv, __be64 service_id, * and service ID resolution requests. The service ID should be specified * network-byte order. If set to IB_CM_ASSIGN_SERVICE_ID, the CM will * assign a service ID to the caller. - * @service_mask: Mask applied to service ID used to listen across a - * range of service IDs. If set to 0, the service ID is matched - * exactly. This parameter is ignored if %service_id is set to - * IB_CM_ASSIGN_SERVICE_ID. */ -int ib_cm_listen(struct ib_cm_id *cm_id, __be64 service_id, __be64 service_mask) +int ib_cm_listen(struct ib_cm_id *cm_id, __be64 service_id) { struct cm_id_private *cm_id_priv = container_of(cm_id, struct cm_id_private, id); @@ -1203,7 +1225,7 @@ int ib_cm_listen(struct ib_cm_id *cm_id, __be64 service_id, __be64 service_mask) goto out; } - ret = cm_init_listen(cm_id_priv, service_id, service_mask); + ret = cm_init_listen(cm_id_priv, service_id); if (ret) goto out; @@ -1251,9 +1273,11 @@ struct ib_cm_id *ib_cm_insert_listen(struct ib_device *device, if (IS_ERR(cm_id_priv)) return ERR_CAST(cm_id_priv); - err = cm_init_listen(cm_id_priv, service_id, 0); - if (err) + err = cm_init_listen(cm_id_priv, service_id); + if (err) { + ib_destroy_cm_id(&cm_id_priv->id); return ERR_PTR(err); + } spin_lock_irq(&cm_id_priv->lock); listen_id_priv = cm_insert_listen(cm_id_priv, cm_handler); @@ -1287,10 +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); } @@ -1319,6 +1343,7 @@ static void cm_format_req(struct cm_req_msg *req_msg, struct sa_path_rec *pri_path = param->primary_path; struct sa_path_rec *alt_path = param->alternate_path; bool pri_ext = false; + __be16 lid; if (pri_path->rec_type == SA_PATH_REC_TYPE_OPA) pri_ext = opa_is_extended_lid(pri_path->opa.dlid, @@ -1378,9 +1403,16 @@ static void cm_format_req(struct cm_req_msg *req_msg, htons(ntohl(sa_path_get_dlid( pri_path))))); } else { + + if (param->primary_path_inbound) { + lid = param->primary_path_inbound->ib.dlid; + IBA_SET(CM_REQ_PRIMARY_LOCAL_PORT_LID, req_msg, + be16_to_cpu(lid)); + } else + IBA_SET(CM_REQ_PRIMARY_LOCAL_PORT_LID, req_msg, + be16_to_cpu(IB_LID_PERMISSIVE)); + /* Work-around until there's a way to obtain remote LID info */ - IBA_SET(CM_REQ_PRIMARY_LOCAL_PORT_LID, req_msg, - be16_to_cpu(IB_LID_PERMISSIVE)); IBA_SET(CM_REQ_PRIMARY_REMOTE_PORT_LID, req_msg, be16_to_cpu(IB_LID_PERMISSIVE)); } @@ -1520,7 +1552,6 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, } } cm_id->service_id = param->service_id; - cm_id->service_mask = ~cpu_to_be64(0); cm_id_priv->timeout_ms = cm_convert_to_ms( param->primary_path->packet_life_time) * 2 + cm_convert_to_ms( @@ -1536,10 +1567,14 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, spin_lock_irqsave(&cm_id_priv->lock, flags); cm_move_av_from_path(&cm_id_priv->av, &av); + if (param->primary_path_outbound) + cm_id_priv->av.dlid_datapath = + be16_to_cpu(param->primary_path_outbound->ib.dlid); + if (param->alternate_path) cm_move_av_from_path(&cm_id_priv->alt_av, &alt_av); - msg = cm_alloc_priv_msg(cm_id_priv); + msg = cm_alloc_priv_msg(cm_id_priv, IB_CM_REQ_SENT); if (IS_ERR(msg)) { ret = PTR_ERR(msg); goto out_unlock; @@ -1548,8 +1583,6 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, req_msg = (struct cm_req_msg *)msg->mad; cm_format_req(req_msg, cm_id_priv, param); cm_id_priv->tid = req_msg->hdr.tid; - msg->timeout_ms = cm_id_priv->timeout_ms; - msg->context[1] = (void *)(unsigned long)IB_CM_REQ_SENT; cm_id_priv->local_qpn = cpu_to_be32(IBA_GET(CM_REQ_LOCAL_QPN, req_msg)); cm_id_priv->rq_psn = cpu_to_be32(IBA_GET(CM_REQ_STARTING_PSN, req_msg)); @@ -1580,7 +1613,7 @@ static int cm_issue_rej(struct cm_port *port, struct cm_rej_msg *rej_msg, *rcv_msg; int ret; - ret = cm_alloc_response_msg(port, mad_recv_wc, &msg); + ret = cm_alloc_response_msg(port, mad_recv_wc, false, &msg); if (ret) return ret; @@ -1606,7 +1639,7 @@ static int cm_issue_rej(struct cm_port *port, IBA_GET(CM_REJ_REMOTE_COMM_ID, rcv_msg)); ret = ib_post_send_mad(msg, NULL); if (ret) - cm_free_response_msg(msg); + cm_free_msg(msg); return ret; } @@ -1630,14 +1663,13 @@ static void cm_path_set_rec_type(struct ib_device *ib_device, u32 port_num, static void cm_format_path_lid_from_req(struct cm_req_msg *req_msg, struct sa_path_rec *primary_path, - struct sa_path_rec *alt_path) + struct sa_path_rec *alt_path, + struct ib_wc *wc) { u32 lid; if (primary_path->rec_type != SA_PATH_REC_TYPE_OPA) { - sa_path_set_dlid(primary_path, - IBA_GET(CM_REQ_PRIMARY_LOCAL_PORT_LID, - req_msg)); + sa_path_set_dlid(primary_path, wc->slid); sa_path_set_slid(primary_path, IBA_GET(CM_REQ_PRIMARY_REMOTE_PORT_LID, req_msg)); @@ -1674,7 +1706,8 @@ static void cm_format_path_lid_from_req(struct cm_req_msg *req_msg, static void cm_format_paths_from_req(struct cm_req_msg *req_msg, struct sa_path_rec *primary_path, - struct sa_path_rec *alt_path) + struct sa_path_rec *alt_path, + struct ib_wc *wc) { primary_path->dgid = *IBA_GET_MEM_PTR(CM_REQ_PRIMARY_LOCAL_PORT_GID, req_msg); @@ -1732,7 +1765,7 @@ static void cm_format_paths_from_req(struct cm_req_msg *req_msg, if (sa_path_is_roce(alt_path)) alt_path->roce.route_resolved = false; } - cm_format_path_lid_from_req(req_msg, primary_path, alt_path); + cm_format_path_lid_from_req(req_msg, primary_path, alt_path, wc); } static u16 cm_get_bth_pkey(struct cm_work *work) @@ -1853,7 +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); @@ -1862,7 +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, @@ -1933,7 +1966,7 @@ static void cm_dup_req_handler(struct cm_work *work, } spin_unlock_irq(&cm_id_priv->lock); - ret = cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg); + ret = cm_alloc_response_msg(work->port, work->mad_recv_wc, true, &msg); if (ret) return; @@ -1941,7 +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; @@ -1962,7 +1995,7 @@ static void cm_dup_req_handler(struct cm_work *work, return; unlock: spin_unlock_irq(&cm_id_priv->lock); -free: cm_free_response_msg(msg); +free: cm_free_msg(msg); } static struct cm_id_private *cm_match_req(struct cm_work *work, @@ -2077,7 +2110,6 @@ static int cm_req_handler(struct cm_work *work) cpu_to_be32(IBA_GET(CM_REQ_LOCAL_COMM_ID, req_msg)); cm_id_priv->id.service_id = cpu_to_be64(IBA_GET(CM_REQ_SERVICE_ID, req_msg)); - cm_id_priv->id.service_mask = ~cpu_to_be64(0); cm_id_priv->tid = req_msg->hdr.tid; cm_id_priv->timeout_ms = cm_convert_to_ms( IBA_GET(CM_REQ_LOCAL_CM_RESPONSE_TIMEOUT, req_msg)); @@ -2146,7 +2178,7 @@ static int cm_req_handler(struct cm_work *work) if (cm_req_has_alt_path(req_msg)) work->path[1].rec_type = work->path[0].rec_type; cm_format_paths_from_req(req_msg, &work->path[0], - &work->path[1]); + &work->path[1], work->mad_recv_wc->wc); if (cm_id_priv->av.ah_attr.type == RDMA_AH_ATTR_TYPE_ROCE) sa_path_set_dmac(&work->path[0], cm_id_priv->av.ah_attr.roce.dmac); @@ -2171,6 +2203,10 @@ static int cm_req_handler(struct cm_work *work) NULL, 0); goto rejected; } + if (cm_id_priv->av.ah_attr.type == RDMA_AH_ATTR_TYPE_IB) + cm_id_priv->av.dlid_datapath = + IBA_GET(CM_REQ_PRIMARY_LOCAL_PORT_LID, req_msg); + if (cm_req_has_alt_path(req_msg)) { ret = cm_init_av_by_path(&work->path[1], NULL, &cm_id_priv->alt_av); @@ -2273,7 +2309,7 @@ int ib_send_cm_rep(struct ib_cm_id *cm_id, goto out; } - msg = cm_alloc_priv_msg(cm_id_priv); + msg = cm_alloc_priv_msg_rep(cm_id_priv, IB_CM_REP_SENT, true); if (IS_ERR(msg)) { ret = PTR_ERR(msg); goto out; @@ -2281,8 +2317,6 @@ int ib_send_cm_rep(struct ib_cm_id *cm_id, rep_msg = (struct cm_rep_msg *) msg->mad; cm_format_rep(rep_msg, cm_id_priv, param); - msg->timeout_ms = cm_id_priv->timeout_ms; - msg->context[1] = (void *) (unsigned long) IB_CM_REP_SENT; trace_icm_send_rep(cm_id); ret = ib_post_send_mad(msg, NULL); @@ -2423,7 +2457,7 @@ static void cm_dup_rep_handler(struct cm_work *work) atomic_long_inc( &work->port->counters[CM_RECV_DUPLICATES][CM_REP_COUNTER]); - ret = cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg); + ret = cm_alloc_response_msg(work->port, work->mad_recv_wc, true, &msg); if (ret) goto deref; @@ -2434,7 +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 @@ -2448,7 +2482,7 @@ static void cm_dup_rep_handler(struct cm_work *work) goto deref; unlock: spin_unlock_irq(&cm_id_priv->lock); -free: cm_free_response_msg(msg); +free: cm_free_msg(msg); deref: cm_deref_id(cm_id_priv); } @@ -2632,59 +2666,68 @@ static void cm_format_dreq(struct cm_dreq_msg *dreq_msg, private_data_len); } -static int cm_send_dreq_locked(struct cm_id_private *cm_id_priv, - const void *private_data, u8 private_data_len) +static void cm_issue_dreq(struct cm_id_private *cm_id_priv) { struct ib_mad_send_buf *msg; int ret; lockdep_assert_held(&cm_id_priv->lock); + msg = cm_alloc_msg(cm_id_priv); + if (IS_ERR(msg)) + return; + + cm_format_dreq((struct cm_dreq_msg *) msg->mad, cm_id_priv, NULL, 0); + + trace_icm_send_dreq(&cm_id_priv->id); + ret = ib_post_send_mad(msg, NULL); + if (ret) + cm_free_msg(msg); +} + +int ib_send_cm_dreq(struct ib_cm_id *cm_id, const void *private_data, + u8 private_data_len) +{ + struct cm_id_private *cm_id_priv = + container_of(cm_id, struct cm_id_private, id); + struct ib_mad_send_buf *msg; + unsigned long flags; + int ret; + if (private_data && private_data_len > IB_CM_DREQ_PRIVATE_DATA_SIZE) return -EINVAL; + spin_lock_irqsave(&cm_id_priv->lock, flags); if (cm_id_priv->id.state != IB_CM_ESTABLISHED) { trace_icm_dreq_skipped(&cm_id_priv->id); - return -EINVAL; + ret = -EINVAL; + goto unlock; } if (cm_id_priv->id.lap_state == IB_CM_LAP_SENT || cm_id_priv->id.lap_state == IB_CM_MRA_LAP_RCVD) ib_cancel_mad(cm_id_priv->msg); - msg = cm_alloc_priv_msg(cm_id_priv); + msg = cm_alloc_priv_msg(cm_id_priv, IB_CM_DREQ_SENT); if (IS_ERR(msg)) { cm_enter_timewait(cm_id_priv); - return PTR_ERR(msg); + ret = PTR_ERR(msg); + goto unlock; } cm_format_dreq((struct cm_dreq_msg *) msg->mad, cm_id_priv, private_data, private_data_len); - msg->timeout_ms = cm_id_priv->timeout_ms; - msg->context[1] = (void *) (unsigned long) IB_CM_DREQ_SENT; trace_icm_send_dreq(&cm_id_priv->id); ret = ib_post_send_mad(msg, NULL); if (ret) { cm_enter_timewait(cm_id_priv); cm_free_priv_msg(msg); - return ret; + goto unlock; } cm_id_priv->id.state = IB_CM_DREQ_SENT; - return 0; -} - -int ib_send_cm_dreq(struct ib_cm_id *cm_id, const void *private_data, - u8 private_data_len) -{ - struct cm_id_private *cm_id_priv = - container_of(cm_id, struct cm_id_private, id); - unsigned long flags; - int ret; - - spin_lock_irqsave(&cm_id_priv->lock, flags); - ret = cm_send_dreq_locked(cm_id_priv, private_data, private_data_len); +unlock: spin_unlock_irqrestore(&cm_id_priv->lock, flags); return ret; } @@ -2770,7 +2813,7 @@ static int cm_issue_drep(struct cm_port *port, struct cm_drep_msg *drep_msg; int ret; - ret = cm_alloc_response_msg(port, mad_recv_wc, &msg); + ret = cm_alloc_response_msg(port, mad_recv_wc, true, &msg); if (ret) return ret; @@ -2788,7 +2831,7 @@ static int cm_issue_drep(struct cm_port *port, IBA_GET(CM_DREQ_REMOTE_COMM_ID, dreq_msg)); ret = ib_post_send_mad(msg, NULL); if (ret) - cm_free_response_msg(msg); + cm_free_msg(msg); return ret; } @@ -2824,6 +2867,7 @@ static int cm_dreq_handler(struct cm_work *work) switch (cm_id_priv->id.state) { case IB_CM_REP_SENT: case IB_CM_DREQ_SENT: + case IB_CM_MRA_REP_RCVD: ib_cancel_mad(cm_id_priv->msg); break; case IB_CM_ESTABLISHED: @@ -2831,12 +2875,11 @@ static int cm_dreq_handler(struct cm_work *work) cm_id_priv->id.lap_state == IB_CM_MRA_LAP_RCVD) ib_cancel_mad(cm_id_priv->msg); break; - case IB_CM_MRA_REP_RCVD: - break; case IB_CM_TIMEWAIT: atomic_long_inc(&work->port->counters[CM_RECV_DUPLICATES] [CM_DREQ_COUNTER]); - msg = cm_alloc_response_msg_no_ah(work->port, work->mad_recv_wc); + msg = cm_alloc_response_msg_no_ah(work->port, work->mad_recv_wc, + true); if (IS_ERR(msg)) goto unlock; @@ -2847,7 +2890,7 @@ static int cm_dreq_handler(struct cm_work *work) if (cm_create_response_msg_ah(work->port, work->mad_recv_wc, msg) || ib_post_send_mad(msg, NULL)) - cm_free_response_msg(msg); + cm_free_msg(msg); goto deref; case IB_CM_DREQ_RCVD: atomic_long_inc(&work->port->counters[CM_RECV_DUPLICATES] @@ -2913,6 +2956,8 @@ static int cm_send_rej_locked(struct cm_id_private *cm_id_priv, (ari && ari_length > IB_CM_REJ_ARI_LENGTH)) return -EINVAL; + trace_icm_send_rej(&cm_id_priv->id, reason); + switch (state) { case IB_CM_REQ_SENT: case IB_CM_MRA_REQ_RCVD: @@ -2943,7 +2988,6 @@ static int cm_send_rej_locked(struct cm_id_private *cm_id_priv, return -EINVAL; } - trace_icm_send_rej(&cm_id_priv->id, reason); ret = ib_post_send_mad(msg, NULL); if (ret) { cm_free_msg(msg); @@ -3064,26 +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); @@ -3092,58 +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) { @@ -3322,7 +3328,7 @@ static int cm_lap_handler(struct cm_work *work) ret = cm_init_av_by_path(param->alternate_path, NULL, &alt_av); if (ret) { rdma_destroy_ah_attr(&ah_attr); - return -EINVAL; + goto deref; } spin_lock_irq(&cm_id_priv->lock); @@ -3340,20 +3346,20 @@ static int cm_lap_handler(struct cm_work *work) case IB_CM_MRA_LAP_SENT: atomic_long_inc(&work->port->counters[CM_RECV_DUPLICATES] [CM_LAP_COUNTER]); - msg = cm_alloc_response_msg_no_ah(work->port, work->mad_recv_wc); + msg = cm_alloc_response_msg_no_ah(work->port, work->mad_recv_wc, + true); if (IS_ERR(msg)) goto unlock; 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); if (cm_create_response_msg_ah(work->port, work->mad_recv_wc, msg) || ib_post_send_mad(msg, NULL)) - cm_free_response_msg(msg); + cm_free_msg(msg); goto deref; case IB_CM_LAP_RCVD: atomic_long_inc(&work->port->counters[CM_RECV_DUPLICATES] @@ -3485,7 +3491,6 @@ int ib_send_cm_sidr_req(struct ib_cm_id *cm_id, spin_lock_irqsave(&cm_id_priv->lock, flags); cm_move_av_from_path(&cm_id_priv->av, &av); cm_id->service_id = param->service_id; - cm_id->service_mask = ~cpu_to_be64(0); cm_id_priv->timeout_ms = param->timeout_ms; cm_id_priv->max_cm_retries = param->max_cm_retries; if (cm_id->state != IB_CM_IDLE) { @@ -3493,7 +3498,7 @@ int ib_send_cm_sidr_req(struct ib_cm_id *cm_id, goto out_unlock; } - msg = cm_alloc_priv_msg(cm_id_priv); + msg = cm_alloc_priv_msg(cm_id_priv, IB_CM_SIDR_REQ_SENT); if (IS_ERR(msg)) { ret = PTR_ERR(msg); goto out_unlock; @@ -3501,8 +3506,6 @@ int ib_send_cm_sidr_req(struct ib_cm_id *cm_id, cm_format_sidr_req((struct cm_sidr_req_msg *)msg->mad, cm_id_priv, param); - msg->timeout_ms = cm_id_priv->timeout_ms; - msg->context[1] = (void *)(unsigned long)IB_CM_SIDR_REQ_SENT; trace_icm_send_sidr_req(&cm_id_priv->id); ret = ib_post_send_mad(msg, NULL); @@ -3560,7 +3563,6 @@ static int cm_sidr_req_handler(struct cm_work *work) cpu_to_be32(IBA_GET(CM_SIDR_REQ_REQUESTID, sidr_req_msg)); cm_id_priv->id.service_id = cpu_to_be64(IBA_GET(CM_SIDR_REQ_SERVICEID, sidr_req_msg)); - cm_id_priv->id.service_mask = ~cpu_to_be64(0); cm_id_priv->tid = sidr_req_msg->hdr.tid; wc = work->mad_recv_wc->wc; @@ -3749,17 +3751,18 @@ out: static void cm_process_send_error(struct cm_id_private *cm_id_priv, struct ib_mad_send_buf *msg, - enum ib_cm_state state, enum ib_wc_status wc_status) { + enum ib_cm_state state = (unsigned long) msg->context[1]; struct ib_cm_event cm_event = {}; int ret; - /* Discard old sends or ones without a response. */ + /* Discard old sends. */ spin_lock_irq(&cm_id_priv->lock); if (msg != cm_id_priv->msg) { spin_unlock_irq(&cm_id_priv->lock); cm_free_msg(msg); + cm_deref_id(cm_id_priv); return; } cm_free_priv_msg(msg); @@ -3807,9 +3810,7 @@ static void cm_send_handler(struct ib_mad_agent *mad_agent, struct ib_mad_send_wc *mad_send_wc) { struct ib_mad_send_buf *msg = mad_send_wc->send_buf; - struct cm_id_private *cm_id_priv = msg->context[0]; - enum ib_cm_state state = - (enum ib_cm_state)(unsigned long)msg->context[1]; + struct cm_id_private *cm_id_priv; struct cm_port *port; u16 attr_index; @@ -3817,13 +3818,12 @@ static void cm_send_handler(struct ib_mad_agent *mad_agent, attr_index = be16_to_cpu(((struct ib_mad_hdr *) msg->mad)->attr_id) - CM_ATTR_ID_OFFSET; - /* - * If the send was in response to a received message (context[0] is not - * set to a cm_id), and is not a REJ, then it is a send that was - * manually retried. - */ - if (!cm_id_priv && (attr_index != CM_REJ_COUNTER)) + if (msg->context[0] == CM_DIRECT_RETRY_CTX) { msg->retries = 1; + cm_id_priv = NULL; + } else { + cm_id_priv = msg->context[0]; + } atomic_long_add(1 + msg->retries, &port->counters[CM_XMIT][attr_index]); if (msg->retries) @@ -3831,10 +3831,9 @@ static void cm_send_handler(struct ib_mad_agent *mad_agent, &port->counters[CM_XMIT_RETRIES][attr_index]); if (cm_id_priv) - cm_process_send_error(cm_id_priv, msg, state, - mad_send_wc->status); + cm_process_send_error(cm_id_priv, msg, mad_send_wc->status); else - cm_free_response_msg(msg); + cm_free_msg(msg); } static void cm_work_handler(struct work_struct *_work) @@ -4097,9 +4096,18 @@ static int cm_init_qp_init_attr(struct cm_id_private *cm_id_priv, *qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS | IB_QP_PKEY_INDEX | IB_QP_PORT; qp_attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE; - if (cm_id_priv->responder_resources) + if (cm_id_priv->responder_resources) { + struct ib_device *ib_dev = cm_id_priv->id.device; + u64 support_flush = ib_dev->attrs.device_cap_flags & + (IB_DEVICE_FLUSH_GLOBAL | IB_DEVICE_FLUSH_PERSISTENT); + u32 flushable = support_flush ? + (IB_ACCESS_FLUSH_GLOBAL | + IB_ACCESS_FLUSH_PERSISTENT) : 0; + qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_READ | - IB_ACCESS_REMOTE_ATOMIC; + IB_ACCESS_REMOTE_ATOMIC | + flushable; + } qp_attr->pkey_index = cm_id_priv->av.pkey_index; if (cm_id_priv->av.port) qp_attr->port_num = cm_id_priv->av.port->port_num; @@ -4133,6 +4141,10 @@ static int cm_init_qp_rtr_attr(struct cm_id_private *cm_id_priv, *qp_attr_mask = IB_QP_STATE | IB_QP_AV | IB_QP_PATH_MTU | IB_QP_DEST_QPN | IB_QP_RQ_PSN; qp_attr->ah_attr = cm_id_priv->av.ah_attr; + if ((qp_attr->ah_attr.type == RDMA_AH_ATTR_TYPE_IB) && + cm_id_priv->av.dlid_datapath && + (cm_id_priv->av.dlid_datapath != 0xffff)) + qp_attr->ah_attr.ib.dlid = cm_id_priv->av.dlid_datapath; qp_attr->path_mtu = cm_id_priv->path_mtu; qp_attr->dest_qp_num = be32_to_cpu(cm_id_priv->remote_qpn); qp_attr->rq_psn = be32_to_cpu(cm_id_priv->rq_psn); @@ -4342,7 +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; @@ -4382,9 +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++; } @@ -4399,6 +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: @@ -4412,6 +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); @@ -4441,12 +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 @@ -4458,10 +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); } @@ -4485,7 +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 e9d282679ef1..4a4987da69d4 100644 --- a/drivers/infiniband/core/cm_trace.h +++ b/drivers/infiniband/core/cm_trace.h @@ -16,7 +16,7 @@ #include <linux/tracepoint.h> #include <rdma/ib_cm.h> -#include <trace/events/rdma.h> +#include <trace/misc/rdma.h> /* * enum ib_cm_state, from include/rdma/ib_cm.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 515a7e95a421..95e89f5c147c 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -11,6 +11,7 @@ #include <linux/in6.h> #include <linux/mutex.h> #include <linux/random.h> +#include <linux/rbtree.h> #include <linux/igmp.h> #include <linux/xarray.h> #include <linux/inetdevice.h> @@ -20,6 +21,7 @@ #include <net/net_namespace.h> #include <net/netns/generic.h> +#include <net/netevent.h> #include <net/tcp.h> #include <net/ipv6.h> #include <net/ip_fib.h> @@ -44,8 +46,7 @@ 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 18 +#define CMA_IBOE_PACKET_LIFETIME 16 #define CMA_PREFERRED_ROCE_GID_TYPE IB_GID_TYPE_ROCE_UDP_ENCAP static const char * const cma_events[] = { @@ -67,8 +68,10 @@ static const char * const cma_events[] = { [RDMA_CM_EVENT_TIMEWAIT_EXIT] = "timewait exit", }; -static void cma_set_mgid(struct rdma_id_private *id_priv, struct sockaddr *addr, - union ib_gid *mgid); +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) { @@ -142,19 +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); @@ -168,6 +158,9 @@ static struct ib_sa_client sa_client; static LIST_HEAD(dev_list); static LIST_HEAD(listen_any_list); static DEFINE_MUTEX(lock); +static struct rb_root id_table = RB_ROOT; +/* Serialize operations of id_table tree */ +static DEFINE_SPINLOCK(id_table_lock); static struct workqueue_struct *cma_wq; static unsigned int cma_pernet_id; @@ -202,6 +195,11 @@ struct xarray *cma_pernet_xa(struct net *net, enum rdma_ucm_port_space ps) } } +struct id_table_entry { + struct list_head id_list; + struct rb_node rb_node; +}; + struct cma_device { struct list_head list; struct ib_device *device; @@ -420,11 +418,21 @@ static inline u8 cma_get_ip_ver(const struct cma_hdr *hdr) return hdr->ip_version >> 4; } -static inline void cma_set_ip_ver(struct cma_hdr *hdr, u8 ip_ver) +static void cma_set_ip_ver(struct cma_hdr *hdr, u8 ip_ver) { hdr->ip_version = (ip_ver << 4) | (hdr->ip_version & 0xF); } +static struct sockaddr *cma_src_addr(struct rdma_id_private *id_priv) +{ + return (struct sockaddr *)&id_priv->id.route.addr.src_addr; +} + +static inline struct sockaddr *cma_dst_addr(struct rdma_id_private *id_priv) +{ + return (struct sockaddr *)&id_priv->id.route.addr.dst_addr; +} + static int cma_igmp_send(struct net_device *ndev, union ib_gid *mgid, bool join) { struct in_device *in_dev = NULL; @@ -445,6 +453,124 @@ static int cma_igmp_send(struct net_device *ndev, union ib_gid *mgid, bool join) return (in_dev) ? 0 : -ENODEV; } +static int compare_netdev_and_ip(int ifindex_a, struct sockaddr *sa, + struct id_table_entry *entry_b) +{ + struct rdma_id_private *id_priv = list_first_entry( + &entry_b->id_list, struct rdma_id_private, id_list_entry); + int ifindex_b = id_priv->id.route.addr.dev_addr.bound_dev_if; + struct sockaddr *sb = cma_dst_addr(id_priv); + + if (ifindex_a != ifindex_b) + return (ifindex_a > ifindex_b) ? 1 : -1; + + if (sa->sa_family != sb->sa_family) + return sa->sa_family - sb->sa_family; + + if (sa->sa_family == AF_INET && + __builtin_object_size(sa, 0) >= sizeof(struct sockaddr_in)) { + return memcmp(&((struct sockaddr_in *)sa)->sin_addr, + &((struct sockaddr_in *)sb)->sin_addr, + sizeof(((struct sockaddr_in *)sa)->sin_addr)); + } + + if (sa->sa_family == AF_INET6 && + __builtin_object_size(sa, 0) >= sizeof(struct sockaddr_in6)) { + return ipv6_addr_cmp(&((struct sockaddr_in6 *)sa)->sin6_addr, + &((struct sockaddr_in6 *)sb)->sin6_addr); + } + + return -1; +} + +static int cma_add_id_to_tree(struct rdma_id_private *node_id_priv) +{ + struct rb_node **new, *parent = NULL; + struct id_table_entry *this, *node; + unsigned long flags; + int result; + + node = kzalloc(sizeof(*node), GFP_KERNEL); + if (!node) + return -ENOMEM; + + spin_lock_irqsave(&id_table_lock, flags); + new = &id_table.rb_node; + while (*new) { + this = container_of(*new, struct id_table_entry, rb_node); + result = compare_netdev_and_ip( + node_id_priv->id.route.addr.dev_addr.bound_dev_if, + cma_dst_addr(node_id_priv), this); + + parent = *new; + if (result < 0) + new = &((*new)->rb_left); + else if (result > 0) + new = &((*new)->rb_right); + else { + list_add_tail(&node_id_priv->id_list_entry, + &this->id_list); + kfree(node); + goto unlock; + } + } + + INIT_LIST_HEAD(&node->id_list); + list_add_tail(&node_id_priv->id_list_entry, &node->id_list); + + rb_link_node(&node->rb_node, parent, new); + rb_insert_color(&node->rb_node, &id_table); + +unlock: + spin_unlock_irqrestore(&id_table_lock, flags); + return 0; +} + +static struct id_table_entry * +node_from_ndev_ip(struct rb_root *root, int ifindex, struct sockaddr *sa) +{ + struct rb_node *node = root->rb_node; + struct id_table_entry *data; + int result; + + while (node) { + data = container_of(node, struct id_table_entry, rb_node); + result = compare_netdev_and_ip(ifindex, sa, data); + if (result < 0) + node = node->rb_left; + else if (result > 0) + node = node->rb_right; + else + return data; + } + + return NULL; +} + +static void cma_remove_id_from_tree(struct rdma_id_private *id_priv) +{ + struct id_table_entry *data; + unsigned long flags; + + spin_lock_irqsave(&id_table_lock, flags); + if (list_empty(&id_priv->id_list_entry)) + goto out; + + data = node_from_ndev_ip(&id_table, + id_priv->id.route.addr.dev_addr.bound_dev_if, + cma_dst_addr(id_priv)); + if (!data) + goto out; + + list_del_init(&id_priv->id_list_entry); + if (list_empty(&data->id_list)) { + rb_erase(&data->rb_node, &id_table); + kfree(data); + } +out: + spin_unlock_irqrestore(&id_table_lock, flags); +} + static void _cma_attach_to_dev(struct rdma_id_private *id_priv, struct cma_device *cma_dev) { @@ -453,7 +579,7 @@ static void _cma_attach_to_dev(struct rdma_id_private *id_priv, id_priv->id.device = cma_dev->device; id_priv->id.route.addr.dev_addr.transport = rdma_node_get_transport(cma_dev->device->node_type); - list_add_tail(&id_priv->list, &cma_dev->id_list); + list_add_tail(&id_priv->device_item, &cma_dev->id_list); trace_cm_id_attach(id_priv, cma_dev->device); } @@ -470,7 +596,7 @@ static void cma_attach_to_dev(struct rdma_id_private *id_priv, static void cma_release_dev(struct rdma_id_private *id_priv) { mutex_lock(&lock); - list_del(&id_priv->list); + list_del_init(&id_priv->device_item); cma_dev_put(id_priv->cma_dev); id_priv->cma_dev = NULL; id_priv->id.device = NULL; @@ -481,37 +607,16 @@ static void cma_release_dev(struct rdma_id_private *id_priv) mutex_unlock(&lock); } -static inline struct sockaddr *cma_src_addr(struct rdma_id_private *id_priv) -{ - return (struct sockaddr *) &id_priv->id.route.addr.src_addr; -} - -static inline struct sockaddr *cma_dst_addr(struct rdma_id_private *id_priv) -{ - return (struct sockaddr *) &id_priv->id.route.addr.dst_addr; -} - static inline unsigned short cma_family(struct rdma_id_private *id_priv) { return id_priv->id.route.addr.src_addr.ss_family; } -static int cma_set_qkey(struct rdma_id_private *id_priv, u32 qkey) +static int cma_set_default_qkey(struct rdma_id_private *id_priv) { struct ib_sa_mcmember_rec rec; int ret = 0; - if (id_priv->qkey) { - if (qkey && id_priv->qkey != qkey) - return -EINVAL; - return 0; - } - - if (qkey) { - id_priv->qkey = qkey; - return 0; - } - switch (id_priv->id.ps) { case RDMA_PS_UDP: case RDMA_PS_IB: @@ -531,6 +636,16 @@ static int cma_set_qkey(struct rdma_id_private *id_priv, u32 qkey) return ret; } +static int cma_set_qkey(struct rdma_id_private *id_priv, u32 qkey) +{ + if (!qkey || + (id_priv->qkey && (id_priv->qkey != qkey))) + return -EINVAL; + + id_priv->qkey = qkey; + return 0; +} + static void cma_translate_ib(struct sockaddr_ib *sib, struct rdma_dev_addr *dev_addr) { dev_addr->dev_type = ARPHRD_INFINIBAND; @@ -559,31 +674,84 @@ cma_validate_port(struct ib_device *device, u32 port, struct rdma_id_private *id_priv) { struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; + const struct ib_gid_attr *sgid_attr = ERR_PTR(-ENODEV); int bound_if_index = dev_addr->bound_dev_if; - const struct ib_gid_attr *sgid_attr; 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)) - return ERR_PTR(-ENODEV); + goto out; if ((dev_type == ARPHRD_INFINIBAND) && !rdma_protocol_ib(device, port)) - return ERR_PTR(-ENODEV); + goto out; if ((dev_type != ARPHRD_INFINIBAND) && rdma_protocol_ib(device, port)) - return ERR_PTR(-ENODEV); + goto out; + + /* + * For drivers that do not associate more than one net device with + * their gid tables, such as iWARP drivers, it is sufficient to + * return the first table entry. + * + * Other driver classes might be included in the future. + */ + if (rdma_protocol_iwarp(device, port)) { + sgid_attr = rdma_get_gid_attr(device, port, 0); + if (IS_ERR(sgid_attr)) + goto out; + + 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); + sgid_attr = ERR_PTR(-ENODEV); + } + rcu_read_unlock(); + 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) - return ERR_PTR(-ENODEV); + /* + * 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); - if (ndev) - dev_put(ndev); + dev_put(ndev); +out: return sgid_attr; } @@ -766,6 +934,7 @@ static int cma_resolve_ib_dev(struct rdma_id_private *id_priv) unsigned int p; u16 pkey, index; enum ib_port_state port_state; + int ret; int i; cma_dev = NULL; @@ -784,9 +953,14 @@ static int cma_resolve_ib_dev(struct rdma_id_private *id_priv) if (ib_get_cached_port_state(cur_dev->device, p, &port_state)) continue; - for (i = 0; !rdma_query_gid(cur_dev->device, - p, i, &gid); - i++) { + + for (i = 0; i < cur_dev->device->port_data[p].immutable.gid_tbl_len; + ++i) { + ret = rdma_query_gid(cur_dev->device, p, i, + &gid); + if (ret) + continue; + if (!memcmp(&gid, dgid, sizeof(gid))) { cma_dev = cur_dev; sgid = gid; @@ -854,11 +1028,14 @@ __rdma_create_id(struct net *net, rdma_cm_event_handler event_handler, init_completion(&id_priv->comp); refcount_set(&id_priv->refcount, 1); mutex_init(&id_priv->handler_mutex); + INIT_LIST_HEAD(&id_priv->device_item); + INIT_LIST_HEAD(&id_priv->id_list_entry); INIT_LIST_HEAD(&id_priv->listen_list); INIT_LIST_HEAD(&id_priv->mc_list); 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) @@ -926,12 +1103,25 @@ static int cma_init_ud_qp(struct rdma_id_private *id_priv, struct ib_qp *qp) return ret; } +static int cma_init_conn_qp(struct rdma_id_private *id_priv, struct ib_qp *qp) +{ + struct ib_qp_attr qp_attr; + int qp_attr_mask, ret; + + qp_attr.qp_state = IB_QPS_INIT; + ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); + if (ret) + return ret; + + return ib_modify_qp(qp, &qp_attr, qp_attr_mask); +} + int rdma_create_qp(struct rdma_cm_id *id, struct ib_pd *pd, struct ib_qp_init_attr *qp_init_attr) { struct rdma_id_private *id_priv; struct ib_qp *qp; - int ret = 0; + int ret; id_priv = container_of(id, struct rdma_id_private, id); if (id->device != pd->device) { @@ -948,6 +1138,8 @@ int rdma_create_qp(struct rdma_cm_id *id, struct ib_pd *pd, if (id->qp_type == IB_QPT_UD) ret = cma_init_ud_qp(id_priv, qp); + else + ret = cma_init_conn_qp(id_priv, qp); if (ret) goto out_destroy; @@ -1078,7 +1270,7 @@ static int cma_ib_init_qp_attr(struct rdma_id_private *id_priv, *qp_attr_mask = IB_QP_STATE | IB_QP_PKEY_INDEX | IB_QP_PORT; if (id_priv->id.qp_type == IB_QPT_UD) { - ret = cma_set_qkey(id_priv, 0); + ret = cma_set_default_qkey(id_priv); if (ret) return ret; @@ -1412,7 +1604,7 @@ static bool validate_ipv4_net_dev(struct net_device *net_dev, return false; memset(&fl4, 0, sizeof(fl4)); - fl4.flowi4_iif = net_dev->ifindex; + fl4.flowi4_oif = net_dev->ifindex; fl4.daddr = daddr; fl4.saddr = saddr; @@ -1632,7 +1824,7 @@ static struct rdma_id_private *cma_find_listener( return id_priv; list_for_each_entry(id_priv_dev, &id_priv->listen_list, - listen_list) { + listen_item) { if (id_priv_dev->id.device == cm_id->device && cma_match_net_dev(&id_priv_dev->id, net_dev, req)) @@ -1697,8 +1889,8 @@ cma_ib_id_from_event(struct ib_cm_id *cm_id, } if (!validate_net_dev(*net_dev, - (struct sockaddr *)&req->listen_addr_storage, - (struct sockaddr *)&req->src_addr_storage)) { + (struct sockaddr *)&req->src_addr_storage, + (struct sockaddr *)&req->listen_addr_storage)) { id_priv = ERR_PTR(-EHOSTUNREACH); goto err; } @@ -1731,28 +1923,36 @@ static void cma_cancel_route(struct rdma_id_private *id_priv) } } -static void cma_cancel_listens(struct rdma_id_private *id_priv) +static void _cma_cancel_listens(struct rdma_id_private *id_priv) { struct rdma_id_private *dev_id_priv; + lockdep_assert_held(&lock); + /* * Remove from listen_any_list to prevent added devices from spawning * additional listen requests. */ - mutex_lock(&lock); - list_del(&id_priv->list); + list_del_init(&id_priv->listen_any_item); while (!list_empty(&id_priv->listen_list)) { - dev_id_priv = list_entry(id_priv->listen_list.next, - struct rdma_id_private, listen_list); + dev_id_priv = + list_first_entry(&id_priv->listen_list, + struct rdma_id_private, listen_item); /* sync with device removal to avoid duplicate destruction */ - list_del_init(&dev_id_priv->list); - list_del(&dev_id_priv->listen_list); + list_del_init(&dev_id_priv->device_item); + list_del_init(&dev_id_priv->listen_item); mutex_unlock(&lock); rdma_destroy_id(&dev_id_priv->id); mutex_lock(&lock); } +} + +static void cma_cancel_listens(struct rdma_id_private *id_priv) +{ + mutex_lock(&lock); + _cma_cancel_listens(id_priv); mutex_unlock(&lock); } @@ -1761,6 +1961,14 @@ static void cma_cancel_operation(struct rdma_id_private *id_priv, { switch (state) { case RDMA_CM_ADDR_QUERY: + /* + * We can avoid doing the rdma_addr_cancel() based on state, + * only RDMA_CM_ADDR_QUERY has a work that could still execute. + * Notice that the addr_handler work could still be exiting + * outside this state, however due to the interaction with the + * handler_mutex the work is guaranteed not to touch id_priv + * during exit. + */ rdma_addr_cancel(&id_priv->id.route.addr.dev_addr); break; case RDMA_CM_ROUTE_QUERY: @@ -1795,6 +2003,8 @@ static void cma_release_port(struct rdma_id_private *id_priv) static void destroy_mc(struct rdma_id_private *id_priv, struct cma_multicast *mc) { + bool send_only = mc->join_state == BIT(SENDONLY_FULLMEMBER_JOIN); + if (rdma_cap_ib_mcast(id_priv->id.device, id_priv->id.port_num)) ib_sa_free_multicast(mc->sa_mc); @@ -1806,14 +2016,19 @@ static void destroy_mc(struct rdma_id_private *id_priv, if (dev_addr->bound_dev_if) ndev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if); - if (ndev) { + if (ndev && !send_only) { + enum ib_gid_type gid_type; union ib_gid mgid; - cma_set_mgid(id_priv, (struct sockaddr *)&mc->addr, - &mgid); + gid_type = id_priv->cma_dev->default_gid_type + [id_priv->id.port_num - + rdma_start_port( + id_priv->cma_dev->device)]; + cma_iboe_set_mgid((struct sockaddr *)&mc->addr, &mgid, + gid_type); cma_igmp_send(ndev, &mgid, false); - dev_put(ndev); } + dev_put(ndev); cancel_work_sync(&mc->iboe_join.work); } @@ -1838,6 +2053,7 @@ static void _destroy_id(struct rdma_id_private *id_priv, cma_cancel_operation(id_priv, state); rdma_restrack_del(&id_priv->res); + cma_remove_id_from_tree(id_priv); if (id_priv->cma_dev) { if (rdma_cap_ib_cm(id_priv->id.device, 1)) { if (id_priv->cm_id.ib) @@ -1858,6 +2074,9 @@ static void _destroy_id(struct rdma_id_private *id_priv, cma_id_put(id_priv->id.context); 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); @@ -1982,8 +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); @@ -2073,14 +2292,14 @@ cma_ib_new_conn_id(const struct rdma_cm_id *listen_id, goto err; rt = &id->route; - rt->num_paths = ib_event->param.req_rcvd.alternate_path ? 2 : 1; - rt->path_rec = kmalloc_array(rt->num_paths, sizeof(*rt->path_rec), - GFP_KERNEL); + rt->num_pri_alt_paths = ib_event->param.req_rcvd.alternate_path ? 2 : 1; + rt->path_rec = kmalloc_array(rt->num_pri_alt_paths, + sizeof(*rt->path_rec), GFP_KERNEL); if (!rt->path_rec) goto err; rt->path_rec[0] = *path; - if (rt->num_paths == 2) + if (rt->num_pri_alt_paths == 2) rt->path_rec[1] = *ib_event->param.req_rcvd.alternate_path; if (net_dev) { @@ -2244,8 +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); @@ -2253,8 +2472,7 @@ err_unlock: mutex_unlock(&listen_id->handler_mutex); net_dev_put: - if (net_dev) - dev_put(net_dev); + dev_put(net_dev); return ret; } @@ -2529,7 +2747,7 @@ static int cma_listen_on_dev(struct rdma_id_private *id_priv, ret = rdma_listen(&dev_id_priv->id, id_priv->backlog); if (ret) goto err_listen; - list_add_tail(&dev_id_priv->listen_list, &id_priv->listen_list); + list_add_tail(&dev_id_priv->listen_item, &id_priv->listen_list); return 0; err_listen: /* Caller must destroy this after releasing lock */ @@ -2545,13 +2763,13 @@ static int cma_listen_on_all(struct rdma_id_private *id_priv) int ret; mutex_lock(&lock); - list_add_tail(&id_priv->list, &listen_any_list); + list_add_tail(&id_priv->listen_any_item, &listen_any_list); list_for_each_entry(cma_dev, &dev_list, list) { ret = cma_listen_on_dev(id_priv, cma_dev, &to_destroy); if (ret) { /* Prevent racing with cma_process_remove() */ if (to_destroy) - list_del_init(&to_destroy->list); + list_del_init(&to_destroy->device_item); goto err_listen; } } @@ -2559,7 +2777,7 @@ static int cma_listen_on_all(struct rdma_id_private *id_priv) return 0; err_listen: - list_del(&id_priv->list); + _cma_cancel_listens(id_priv); mutex_unlock(&lock); if (to_destroy) rdma_destroy_id(&to_destroy->id); @@ -2597,7 +2815,7 @@ int rdma_set_ack_timeout(struct rdma_cm_id *id, u8 timeout) { struct rdma_id_private *id_priv; - if (id->qp_type != IB_QPT_RC) + if (id->qp_type != IB_QPT_RC && id->qp_type != IB_QPT_XRC_INI) return -EINVAL; id_priv = container_of(id, struct rdma_id_private, id); @@ -2649,26 +2867,76 @@ int rdma_set_min_rnr_timer(struct rdma_cm_id *id, u8 min_rnr_timer) } EXPORT_SYMBOL(rdma_set_min_rnr_timer); +static int route_set_path_rec_inbound(struct cma_work *work, + struct sa_path_rec *path_rec) +{ + struct rdma_route *route = &work->id->id.route; + + if (!route->path_rec_inbound) { + route->path_rec_inbound = + kzalloc(sizeof(*route->path_rec_inbound), GFP_KERNEL); + if (!route->path_rec_inbound) + return -ENOMEM; + } + + *route->path_rec_inbound = *path_rec; + return 0; +} + +static int route_set_path_rec_outbound(struct cma_work *work, + struct sa_path_rec *path_rec) +{ + struct rdma_route *route = &work->id->id.route; + + if (!route->path_rec_outbound) { + route->path_rec_outbound = + kzalloc(sizeof(*route->path_rec_outbound), GFP_KERNEL); + if (!route->path_rec_outbound) + return -ENOMEM; + } + + *route->path_rec_outbound = *path_rec; + return 0; +} + static void cma_query_handler(int status, struct sa_path_rec *path_rec, - void *context) + unsigned int num_prs, void *context) { struct cma_work *work = context; struct rdma_route *route; + int i; route = &work->id->id.route; - if (!status) { - route->num_paths = 1; - *route->path_rec = *path_rec; - } else { - work->old_state = RDMA_CM_ROUTE_QUERY; - work->new_state = RDMA_CM_ADDR_RESOLVED; - work->event.event = RDMA_CM_EVENT_ROUTE_ERROR; - work->event.status = status; - pr_debug_ratelimited("RDMA CM: ROUTE_ERROR: failed to query path. status %d\n", - status); + if (status) + goto fail; + + for (i = 0; i < num_prs; i++) { + if (!path_rec[i].flags || (path_rec[i].flags & IB_PATH_GMP)) + *route->path_rec = path_rec[i]; + else if (path_rec[i].flags & IB_PATH_INBOUND) + status = route_set_path_rec_inbound(work, &path_rec[i]); + else if (path_rec[i].flags & IB_PATH_OUTBOUND) + status = route_set_path_rec_outbound(work, + &path_rec[i]); + else + status = -EINVAL; + + if (status) + goto fail; } + route->num_pri_alt_paths = 1; + queue_work(cma_wq, &work->work); + return; + +fail: + work->old_state = RDMA_CM_ROUTE_QUERY; + work->new_state = RDMA_CM_ADDR_RESOLVED; + work->event.event = RDMA_CM_EVENT_ROUTE_ERROR; + work->event.status = status; + pr_debug_ratelimited("RDMA CM: ROUTE_ERROR: failed to query path. status %d\n", + status); queue_work(cma_wq, &work->work); } @@ -2913,7 +3181,7 @@ int rdma_set_ib_path(struct rdma_cm_id *id, dev_put(ndev); } - id->route.num_paths = 1; + id->route.num_pri_alt_paths = 1; return 0; err_free: @@ -3046,7 +3314,7 @@ static int cma_resolve_iboe_route(struct rdma_id_private *id_priv) goto err1; } - route->num_paths = 1; + route->num_pri_alt_paths = 1; ndev = cma_iboe_set_path_rec_l2_fields(id_priv); if (!ndev) { @@ -3071,7 +3339,7 @@ static int cma_resolve_iboe_route(struct rdma_id_private *id_priv) route->path_rec->traffic_class = tos; route->path_rec->mtu = iboe_get_mtu(ndev->mtu); route->path_rec->rate_selector = IB_SA_EQ; - route->path_rec->rate = iboe_get_rate(ndev); + route->path_rec->rate = IB_RATE_PORT_CURRENT; dev_put(ndev); route->path_rec->packet_life_time_selector = IB_SA_EQ; /* In case ACK timeout is set, use this value to calculate @@ -3106,7 +3374,7 @@ static int cma_resolve_iboe_route(struct rdma_id_private *id_priv) err2: kfree(route->path_rec); route->path_rec = NULL; - route->num_paths = 0; + route->num_pri_alt_paths = 0; err1: kfree(work); return ret; @@ -3115,17 +3383,28 @@ 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); if (rdma_cap_ib_sa(id->device, id->port_num)) ret = cma_resolve_ib_route(id_priv, timeout_ms); - else if (rdma_protocol_roce(id->device, id->port_num)) + else if (rdma_protocol_roce(id->device, id->port_num)) { ret = cma_resolve_iboe_route(id_priv); + if (!ret) + cma_add_id_to_tree(id_priv); + } else if (rdma_protocol_iwarp(id->device, id->port_num)) ret = cma_resolve_iw_route(id_priv); else @@ -3136,7 +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; } @@ -3319,98 +3598,6 @@ err: return ret; } -static int cma_bind_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, - const struct sockaddr *dst_addr) -{ - if (!src_addr || !src_addr->sa_family) { - src_addr = (struct sockaddr *) &id->route.addr.src_addr; - src_addr->sa_family = dst_addr->sa_family; - if (IS_ENABLED(CONFIG_IPV6) && - dst_addr->sa_family == AF_INET6) { - struct sockaddr_in6 *src_addr6 = (struct sockaddr_in6 *) src_addr; - struct sockaddr_in6 *dst_addr6 = (struct sockaddr_in6 *) dst_addr; - src_addr6->sin6_scope_id = dst_addr6->sin6_scope_id; - if (ipv6_addr_type(&dst_addr6->sin6_addr) & IPV6_ADDR_LINKLOCAL) - id->route.addr.dev_addr.bound_dev_if = dst_addr6->sin6_scope_id; - } else if (dst_addr->sa_family == AF_IB) { - ((struct sockaddr_ib *) src_addr)->sib_pkey = - ((struct sockaddr_ib *) dst_addr)->sib_pkey; - } - } - return rdma_bind_addr(id, src_addr); -} - -/* - * If required, resolve the source address for bind and leave the id_priv in - * state RDMA_CM_ADDR_BOUND. This oddly uses the state to determine the prior - * calls made by ULP, a previously bound ID will not be re-bound and src_addr is - * ignored. - */ -static int resolve_prepare_src(struct rdma_id_private *id_priv, - struct sockaddr *src_addr, - const struct sockaddr *dst_addr) -{ - int ret; - - memcpy(cma_dst_addr(id_priv), dst_addr, rdma_addr_size(dst_addr)); - if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_ADDR_QUERY)) { - /* For a well behaved ULP state will be RDMA_CM_IDLE */ - ret = cma_bind_addr(&id_priv->id, src_addr, dst_addr); - if (ret) - goto err_dst; - if (WARN_ON(!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, - RDMA_CM_ADDR_QUERY))) { - ret = -EINVAL; - goto err_dst; - } - } - - if (cma_family(id_priv) != dst_addr->sa_family) { - ret = -EINVAL; - goto err_state; - } - return 0; - -err_state: - cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, RDMA_CM_ADDR_BOUND); -err_dst: - memset(cma_dst_addr(id_priv), 0, rdma_addr_size(dst_addr)); - return ret; -} - -int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, - const struct sockaddr *dst_addr, unsigned long timeout_ms) -{ - struct rdma_id_private *id_priv = - container_of(id, struct rdma_id_private, id); - int ret; - - ret = resolve_prepare_src(id_priv, src_addr, dst_addr); - if (ret) - return ret; - - if (cma_any_addr(dst_addr)) { - ret = cma_resolve_loopback(id_priv); - } else { - if (dst_addr->sa_family == AF_IB) { - ret = cma_resolve_ib_addr(id_priv); - } else { - ret = rdma_resolve_ip(cma_src_addr(id_priv), dst_addr, - &id->route.addr.dev_addr, - timeout_ms, addr_handler, - false, id_priv); - } - } - if (ret) - goto err; - - return 0; -err: - cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, RDMA_CM_ADDR_BOUND); - return ret; -} -EXPORT_SYMBOL(rdma_resolve_addr); - int rdma_set_reuseaddr(struct rdma_cm_id *id, int reuse) { struct rdma_id_private *id_priv; @@ -3562,7 +3749,7 @@ static int cma_alloc_any_port(enum rdma_ucm_port_space ps, inet_get_local_port_range(net, &low, &high); remaining = (high - low) + 1; - rover = prandom_u32() % remaining + low; + rover = get_random_u32_inclusive(low, remaining + low - 1); retry: if (last_used_port != rover) { struct rdma_bind_list *bind_list; @@ -3753,9 +3940,13 @@ int rdma_listen(struct rdma_cm_id *id, int backlog) int ret; if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_LISTEN)) { + struct sockaddr_in any_in = { + .sin_family = AF_INET, + .sin_addr.s_addr = htonl(INADDR_ANY), + }; + /* For a well behaved ULP state will be RDMA_CM_IDLE */ - id->route.addr.src_addr.ss_family = AF_INET; - ret = rdma_bind_addr(id, cma_src_addr(id_priv)); + ret = rdma_bind_addr(id, (struct sockaddr *)&any_in); if (ret) return ret; if (WARN_ON(!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, @@ -3809,27 +4000,26 @@ err: } EXPORT_SYMBOL(rdma_listen); -int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr) +static int rdma_bind_addr_dst(struct rdma_id_private *id_priv, + struct sockaddr *addr, const struct sockaddr *daddr) { - struct rdma_id_private *id_priv; + struct sockaddr *id_daddr; int ret; - struct sockaddr *daddr; if (addr->sa_family != AF_INET && addr->sa_family != AF_INET6 && addr->sa_family != AF_IB) return -EAFNOSUPPORT; - id_priv = container_of(id, struct rdma_id_private, id); if (!cma_comp_exch(id_priv, RDMA_CM_IDLE, RDMA_CM_ADDR_BOUND)) return -EINVAL; - ret = cma_check_linklocal(&id->route.addr.dev_addr, addr); + ret = cma_check_linklocal(&id_priv->id.route.addr.dev_addr, addr); if (ret) goto err1; memcpy(cma_src_addr(id_priv), addr, rdma_addr_size(addr)); if (!cma_any_addr(addr)) { - ret = cma_translate_addr(addr, &id->route.addr.dev_addr); + ret = cma_translate_addr(addr, &id_priv->id.route.addr.dev_addr); if (ret) goto err1; @@ -3849,8 +4039,10 @@ int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr) } #endif } - daddr = cma_dst_addr(id_priv); - daddr->sa_family = addr->sa_family; + id_daddr = cma_dst_addr(id_priv); + if (daddr != id_daddr) + memcpy(id_daddr, daddr, rdma_addr_size(addr)); + id_daddr->sa_family = addr->sa_family; ret = cma_get_port(id_priv); if (ret) @@ -3866,6 +4058,129 @@ err1: cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_IDLE); return ret; } + +static int cma_bind_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, + const struct sockaddr *dst_addr) +{ + struct rdma_id_private *id_priv = + container_of(id, struct rdma_id_private, id); + struct sockaddr_storage zero_sock = {}; + + if (src_addr && src_addr->sa_family) + return rdma_bind_addr_dst(id_priv, src_addr, dst_addr); + + /* + * When the src_addr is not specified, automatically supply an any addr + */ + zero_sock.ss_family = dst_addr->sa_family; + if (IS_ENABLED(CONFIG_IPV6) && dst_addr->sa_family == AF_INET6) { + struct sockaddr_in6 *src_addr6 = + (struct sockaddr_in6 *)&zero_sock; + struct sockaddr_in6 *dst_addr6 = + (struct sockaddr_in6 *)dst_addr; + + src_addr6->sin6_scope_id = dst_addr6->sin6_scope_id; + if (ipv6_addr_type(&dst_addr6->sin6_addr) & IPV6_ADDR_LINKLOCAL) + id->route.addr.dev_addr.bound_dev_if = + dst_addr6->sin6_scope_id; + } else if (dst_addr->sa_family == AF_IB) { + ((struct sockaddr_ib *)&zero_sock)->sib_pkey = + ((struct sockaddr_ib *)dst_addr)->sib_pkey; + } + return rdma_bind_addr_dst(id_priv, (struct sockaddr *)&zero_sock, dst_addr); +} + +/* + * If required, resolve the source address for bind and leave the id_priv in + * state RDMA_CM_ADDR_BOUND. This oddly uses the state to determine the prior + * calls made by ULP, a previously bound ID will not be re-bound and src_addr is + * ignored. + */ +static int resolve_prepare_src(struct rdma_id_private *id_priv, + struct sockaddr *src_addr, + const struct sockaddr *dst_addr) +{ + int ret; + + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_ADDR_QUERY)) { + /* For a well behaved ULP state will be RDMA_CM_IDLE */ + ret = cma_bind_addr(&id_priv->id, src_addr, dst_addr); + if (ret) + return ret; + if (WARN_ON(!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, + RDMA_CM_ADDR_QUERY))) + return -EINVAL; + + } else { + memcpy(cma_dst_addr(id_priv), dst_addr, rdma_addr_size(dst_addr)); + } + + if (cma_family(id_priv) != dst_addr->sa_family) { + ret = -EINVAL; + goto err_state; + } + return 0; + +err_state: + cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, RDMA_CM_ADDR_BOUND); + return ret; +} + +int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, + const struct sockaddr *dst_addr, unsigned long timeout_ms) +{ + struct rdma_id_private *id_priv = + container_of(id, struct rdma_id_private, id); + int ret; + + ret = resolve_prepare_src(id_priv, src_addr, dst_addr); + if (ret) + return ret; + + if (cma_any_addr(dst_addr)) { + ret = cma_resolve_loopback(id_priv); + } else { + if (dst_addr->sa_family == AF_IB) { + ret = cma_resolve_ib_addr(id_priv); + } else { + /* + * The FSM can return back to RDMA_CM_ADDR_BOUND after + * rdma_resolve_ip() is called, eg through the error + * path in addr_handler(). If this happens the existing + * request must be canceled before issuing a new one. + * Since canceling a request is a bit slow and this + * oddball path is rare, keep track once a request has + * been issued. The track turns out to be a permanent + * state since this is the only cancel as it is + * immediately before rdma_resolve_ip(). + */ + if (id_priv->used_resolve_ip) + rdma_addr_cancel(&id->route.addr.dev_addr); + else + id_priv->used_resolve_ip = 1; + ret = rdma_resolve_ip(cma_src_addr(id_priv), dst_addr, + &id->route.addr.dev_addr, + timeout_ms, addr_handler, + false, id_priv); + } + } + if (ret) + goto err; + + return 0; +err: + cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, RDMA_CM_ADDR_BOUND); + return ret; +} +EXPORT_SYMBOL(rdma_resolve_addr); + +int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr) +{ + struct rdma_id_private *id_priv = + container_of(id, struct rdma_id_private, id); + + return rdma_bind_addr_dst(id_priv, addr, cma_dst_addr(id_priv)); +} EXPORT_SYMBOL(rdma_bind_addr); static int cma_format_hdr(void *hdr, struct rdma_id_private *id_priv) @@ -3974,8 +4289,7 @@ static int cma_resolve_ib_udp(struct rdma_id_private *id_priv, memset(&req, 0, sizeof req); offset = cma_user_data_offset(id_priv); - req.private_data_len = offset + conn_param->private_data_len; - if (req.private_data_len < conn_param->private_data_len) + if (check_add_overflow(offset, conn_param->private_data_len, &req.private_data_len)) return -EINVAL; if (req.private_data_len) { @@ -4034,8 +4348,7 @@ static int cma_connect_ib(struct rdma_id_private *id_priv, memset(&req, 0, sizeof req); offset = cma_user_data_offset(id_priv); - req.private_data_len = offset + conn_param->private_data_len; - if (req.private_data_len < conn_param->private_data_len) + if (check_add_overflow(offset, conn_param->private_data_len, &req.private_data_len)) return -EINVAL; if (req.private_data_len) { @@ -4066,7 +4379,9 @@ static int cma_connect_ib(struct rdma_id_private *id_priv, } req.primary_path = &route->path_rec[0]; - if (route->num_paths == 2) + req.primary_path_inbound = route->path_rec_inbound; + req.primary_path_outbound = route->path_rec_outbound; + if (route->num_pri_alt_paths == 2) req.alternate_path = &route->path_rec[1]; req.ppath_sgid_attr = id_priv->id.route.addr.dev_addr.sgid_attr; @@ -4160,6 +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; @@ -4302,7 +4619,10 @@ static int cma_send_sidr_rep(struct rdma_id_private *id_priv, memset(&rep, 0, sizeof rep); rep.status = status; if (status == IB_SIDR_SUCCESS) { - ret = cma_set_qkey(id_priv, qkey); + if (qkey) + ret = cma_set_qkey(id_priv, qkey); + else + ret = cma_set_default_qkey(id_priv); if (ret) return ret; rep.qp_num = id_priv->qp_num; @@ -4507,9 +4827,7 @@ static void cma_make_mc_event(int status, struct rdma_id_private *id_priv, enum ib_gid_type gid_type; struct net_device *ndev; - if (!status) - status = cma_set_qkey(id_priv, be32_to_cpu(multicast->rec.qkey)); - else + if (status) pr_debug_ratelimited("RDMA CM: MULTICAST_ERROR: failed to join multicast. status %d\n", status); @@ -4537,11 +4855,10 @@ static void cma_make_mc_event(int status, struct rdma_id_private *id_priv, } event->param.ud.qp_num = 0xFFFFFF; - event->param.ud.qkey = be32_to_cpu(multicast->rec.qkey); + event->param.ud.qkey = id_priv->qkey; out: - if (ndev) - dev_put(ndev); + dev_put(ndev); } static int cma_ib_mc_handler(int status, struct ib_sa_multicast *multicast) @@ -4556,8 +4873,11 @@ static int cma_ib_mc_handler(int status, struct ib_sa_multicast *multicast) READ_ONCE(id_priv->state) == RDMA_CM_DESTROYING) goto out; - cma_make_mc_event(status, id_priv, multicast, &event, mc); - ret = cma_cm_event_handler(id_priv, &event); + ret = cma_set_qkey(id_priv, be32_to_cpu(multicast->rec.qkey)); + if (!ret) { + cma_make_mc_event(status, id_priv, multicast, &event, mc); + ret = cma_cm_event_handler(id_priv, &event); + } rdma_destroy_ah_attr(&event.param.ud.ah_attr); WARN_ON(ret); @@ -4610,9 +4930,11 @@ static int cma_join_ib_multicast(struct rdma_id_private *id_priv, if (ret) return ret; - ret = cma_set_qkey(id_priv, 0); - if (ret) - return ret; + if (!id_priv->qkey) { + ret = cma_set_default_qkey(id_priv); + if (ret) + return ret; + } cma_set_mgid(id_priv, (struct sockaddr *) &mc->addr, &rec.mgid); rec.qkey = cpu_to_be32(id_priv->qkey); @@ -4675,7 +4997,7 @@ static int cma_iboe_join_multicast(struct rdma_id_private *id_priv, int err = 0; struct sockaddr *addr = (struct sockaddr *)&mc->addr; struct net_device *ndev = NULL; - struct ib_sa_multicast ib; + struct ib_sa_multicast ib = {}; enum ib_gid_type gid_type; bool send_only; @@ -4689,15 +5011,12 @@ static int cma_iboe_join_multicast(struct rdma_id_private *id_priv, cma_iboe_set_mgid(addr, &ib.rec.mgid, gid_type); ib.rec.pkey = cpu_to_be16(0xffff); - if (id_priv->id.ps == RDMA_PS_UDP) - ib.rec.qkey = cpu_to_be32(RDMA_UDP_QKEY); - if (dev_addr->bound_dev_if) ndev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if); if (!ndev) return -ENODEV; - ib.rec.rate = iboe_get_rate(ndev); + ib.rec.rate = IB_RATE_PORT_CURRENT; ib.rec.hop_limit = 1; ib.rec.mtu = iboe_get_mtu(ndev->mtu); @@ -4717,6 +5036,9 @@ static int cma_iboe_join_multicast(struct rdma_id_private *id_priv, if (err || !ib.rec.mtu) return err ?: -EINVAL; + if (!id_priv->qkey) + cma_set_default_qkey(id_priv); + rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.src_addr, &ib.rec.port_gid); INIT_WORK(&mc->iboe_join.work, cma_iboe_join_work_handler); @@ -4742,6 +5064,9 @@ int rdma_join_multicast(struct rdma_cm_id *id, struct sockaddr *addr, READ_ONCE(id_priv->state) != RDMA_CM_ADDR_RESOLVED)) return -EINVAL; + if (id_priv->id.qp_type != IB_QPT_UD) + return -EINVAL; + mc = kzalloc(sizeof(*mc), GFP_KERNEL); if (!mc) return -ENOMEM; @@ -4838,7 +5163,7 @@ static int cma_netdev_callback(struct notifier_block *self, unsigned long event, mutex_lock(&lock); list_for_each_entry(cma_dev, &dev_list, list) - list_for_each_entry(id_priv, &cma_dev->id_list, list) { + list_for_each_entry(id_priv, &cma_dev->id_list, device_item) { ret = cma_netdev_change(ndev, id_priv); if (ret) goto out; @@ -4849,10 +5174,87 @@ out: return ret; } +static void cma_netevent_work_handler(struct work_struct *_work) +{ + struct rdma_id_private *id_priv = + container_of(_work, struct rdma_id_private, id.net_work); + struct rdma_cm_event event = {}; + + mutex_lock(&id_priv->handler_mutex); + + if (READ_ONCE(id_priv->state) == RDMA_CM_DESTROYING || + READ_ONCE(id_priv->state) == RDMA_CM_DEVICE_REMOVAL) + goto out_unlock; + + event.event = RDMA_CM_EVENT_UNREACHABLE; + event.status = -ETIMEDOUT; + + if (cma_cm_event_handler(id_priv, &event)) { + __acquire(&id_priv->handler_mutex); + id_priv->cm_id.ib = NULL; + cma_id_put(id_priv); + destroy_id_handler_unlock(id_priv); + return; + } + +out_unlock: + mutex_unlock(&id_priv->handler_mutex); + cma_id_put(id_priv); +} + +static int cma_netevent_callback(struct notifier_block *self, + unsigned long event, void *ctx) +{ + struct id_table_entry *ips_node = NULL; + struct rdma_id_private *current_id; + struct neighbour *neigh = ctx; + unsigned long flags; + + if (event != NETEVENT_NEIGH_UPDATE) + return NOTIFY_DONE; + + spin_lock_irqsave(&id_table_lock, flags); + if (neigh->tbl->family == AF_INET6) { + struct sockaddr_in6 neigh_sock_6; + + neigh_sock_6.sin6_family = AF_INET6; + neigh_sock_6.sin6_addr = *(struct in6_addr *)neigh->primary_key; + ips_node = node_from_ndev_ip(&id_table, neigh->dev->ifindex, + (struct sockaddr *)&neigh_sock_6); + } else if (neigh->tbl->family == AF_INET) { + struct sockaddr_in neigh_sock_4; + + neigh_sock_4.sin_family = AF_INET; + neigh_sock_4.sin_addr.s_addr = *(__be32 *)(neigh->primary_key); + ips_node = node_from_ndev_ip(&id_table, neigh->dev->ifindex, + (struct sockaddr *)&neigh_sock_4); + } else + goto out; + + if (!ips_node) + goto out; + + list_for_each_entry(current_id, &ips_node->id_list, id_list_entry) { + if (!memcmp(current_id->id.route.addr.dev_addr.dst_dev_addr, + neigh->ha, ETH_ALEN)) + continue; + cma_id_get(current_id); + if (!queue_work(cma_wq, ¤t_id->id.net_work)) + cma_id_put(current_id); + } +out: + spin_unlock_irqrestore(&id_table_lock, flags); + return NOTIFY_DONE; +} + static struct notifier_block cma_nb = { .notifier_call = cma_netdev_callback }; +static struct notifier_block cma_netevent_cb = { + .notifier_call = cma_netevent_callback +}; + static void cma_send_device_removal_put(struct rdma_id_private *id_priv) { struct rdma_cm_event event = { .event = RDMA_CM_EVENT_DEVICE_REMOVAL }; @@ -4898,10 +5300,10 @@ static void cma_process_remove(struct cma_device *cma_dev) mutex_lock(&lock); while (!list_empty(&cma_dev->id_list)) { struct rdma_id_private *id_priv = list_first_entry( - &cma_dev->id_list, struct rdma_id_private, list); + &cma_dev->id_list, struct rdma_id_private, device_item); - list_del(&id_priv->listen_list); - list_del_init(&id_priv->list); + list_del_init(&id_priv->listen_item); + list_del_init(&id_priv->device_item); cma_id_get(id_priv); mutex_unlock(&lock); @@ -4978,7 +5380,7 @@ static int cma_add_one(struct ib_device *device) mutex_lock(&lock); list_add_tail(&cma_dev->list, &dev_list); - list_for_each_entry(id_priv, &listen_any_list, list) { + list_for_each_entry(id_priv, &listen_any_list, listen_any_item) { ret = cma_listen_on_dev(id_priv, cma_dev, &to_destroy); if (ret) goto free_listen; @@ -5075,6 +5477,7 @@ static int __init cma_init(void) ib_sa_register_client(&sa_client); register_netdevice_notifier(&cma_nb); + register_netevent_notifier(&cma_netevent_cb); ret = ib_register_client(&cma_client); if (ret) @@ -5089,6 +5492,7 @@ static int __init cma_init(void) err_ib: ib_unregister_client(&cma_client); err: + unregister_netevent_notifier(&cma_netevent_cb); unregister_netdevice_notifier(&cma_nb); ib_sa_unregister_client(&sa_client); unregister_pernet_subsys(&cma_pernet_operations); @@ -5101,6 +5505,7 @@ static void __exit cma_cleanup(void) { cma_configfs_exit(); ib_unregister_client(&cma_client); + unregister_netevent_notifier(&cma_netevent_cb); unregister_netdevice_notifier(&cma_nb); ib_sa_unregister_client(&sa_client); unregister_pernet_subsys(&cma_pernet_operations); @@ -5109,3 +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_configfs.c b/drivers/infiniband/core/cma_configfs.c index 9ac16e0db761..f2fb2d8a6597 100644 --- a/drivers/infiniband/core/cma_configfs.c +++ b/drivers/infiniband/core/cma_configfs.c @@ -30,7 +30,6 @@ * SOFTWARE. */ -#include <linux/module.h> #include <linux/configfs.h> #include <rdma/ib_verbs.h> #include <rdma/rdma_cm.h> @@ -218,7 +217,7 @@ static int make_cma_ports(struct cma_dev_group *cma_dev_group, return -ENOMEM; for (i = 0; i < ports_num; i++) { - char port_str[10]; + char port_str[11]; ports[i].port_num = i + 1; snprintf(port_str, sizeof(port_str), "%u", i + 1); @@ -293,7 +292,7 @@ static struct config_group *make_cma_dev(struct config_group *group, goto fail; } - strlcpy(cma_dev_group->name, name, sizeof(cma_dev_group->name)); + strscpy(cma_dev_group->name, name, sizeof(cma_dev_group->name)); config_group_init_type_name(&cma_dev_group->ports_group, "ports", &cma_ports_group_type); diff --git a/drivers/infiniband/core/cma_priv.h b/drivers/infiniband/core/cma_priv.h index 5c463da99845..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 { @@ -55,8 +57,16 @@ struct rdma_id_private { struct rdma_bind_list *bind_list; struct hlist_node node; - struct list_head list; /* listen_any_list or cma_device.list */ - struct list_head listen_list; /* per device listens */ + union { + struct list_head device_item; /* On cma_device->id_list */ + struct list_head listen_any_item; /* On listen_any_list */ + }; + union { + /* On rdma_id_private->listen_list */ + struct list_head listen_item; + struct list_head listen_list; + }; + struct list_head id_list_entry; struct cma_device *cma_dev; struct list_head mc_list; @@ -91,6 +101,7 @@ struct rdma_id_private { u8 afonly; u8 timeout; u8 min_rnr_timer; + u8 used_resolve_ip; enum ib_gid_type gid_type; /* diff --git a/drivers/infiniband/core/cma_trace.h b/drivers/infiniband/core/cma_trace.h index e45264267bcc..3456d5f3aa47 100644 --- a/drivers/infiniband/core/cma_trace.h +++ b/drivers/infiniband/core/cma_trace.h @@ -15,7 +15,7 @@ #define _TRACE_RDMA_CMA_H #include <linux/tracepoint.h> -#include <trace/events/rdma.h> +#include <trace/misc/rdma.h> DECLARE_EVENT_CLASS(cma_fsm_class, @@ -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); @@ -84,7 +84,7 @@ TRACE_EVENT(cm_id_attach, sizeof(struct sockaddr_in6)); memcpy(__entry->dstaddr, &id_priv->id.route.addr.dst_addr, sizeof(struct sockaddr_in6)); - __assign_str(devname, device->name); + __assign_str(devname); ), TP_printk("cm.id=%u src=%pISpc dst=%pISpc device=%s", @@ -334,7 +334,7 @@ DECLARE_EVENT_CLASS(cma_client_class, ), TP_fast_assign( - __assign_str(name, device->name); + __assign_str(name); ), TP_printk("device name=%s", diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h index 647cca4e0240..05102769a918 100644 --- a/drivers/infiniband/core/core_priv.h +++ b/drivers/infiniband/core/core_priv.h @@ -316,50 +316,15 @@ struct ib_device *ib_device_get_by_index(const struct net *net, u32 index); void nldev_init(void); void nldev_exit(void); -static inline struct ib_qp * -_ib_create_qp(struct ib_device *dev, struct ib_pd *pd, - struct ib_qp_init_attr *attr, struct ib_udata *udata, - struct ib_uqp_object *uobj, const char *caller) -{ - struct ib_qp *qp; - - if (!dev->ops.create_qp) - return ERR_PTR(-EOPNOTSUPP); - - qp = dev->ops.create_qp(pd, attr, udata); - if (IS_ERR(qp)) - return qp; - - qp->device = dev; - qp->pd = pd; - qp->uobject = uobj; - qp->real_qp = qp; - - qp->qp_type = attr->qp_type; - qp->rwq_ind_tbl = attr->rwq_ind_tbl; - qp->send_cq = attr->send_cq; - qp->recv_cq = attr->recv_cq; - qp->srq = attr->srq; - qp->rwq_ind_tbl = attr->rwq_ind_tbl; - qp->event_handler = attr->event_handler; - qp->port = attr->port_num; - - atomic_set(&qp->usecnt, 0); - spin_lock_init(&qp->mr_lock); - INIT_LIST_HEAD(&qp->rdma_mrs); - INIT_LIST_HEAD(&qp->sig_mrs); - - rdma_restrack_new(&qp->res, RDMA_RESTRACK_QP); - WARN_ONCE(!udata && !caller, "Missing kernel QP owner"); - rdma_restrack_set_name(&qp->res, udata ? NULL : caller); - rdma_restrack_add(&qp->res); - return qp; -} +struct ib_qp *ib_create_qp_user(struct ib_device *dev, struct ib_pd *pd, + struct ib_qp_init_attr *attr, + struct ib_udata *udata, + struct ib_uqp_object *uobj, const char *caller); + +void ib_qp_usecnt_inc(struct ib_qp *qp); +void ib_qp_usecnt_dec(struct ib_qp *qp); struct rdma_dev_addr; -int rdma_resolve_ip_route(struct sockaddr *src_addr, - const struct sockaddr *dst_addr, - struct rdma_dev_addr *addr); int rdma_addr_find_l2_eth_by_grh(const union ib_gid *sgid, const union ib_gid *dgid, @@ -405,4 +370,5 @@ void rdma_umap_priv_init(struct rdma_umap_priv *priv, void ib_cq_pool_cleanup(struct ib_device *dev); +bool rdma_nl_get_privileged_qkey(void); #endif /* _CORE_PRIV_H */ diff --git a/drivers/infiniband/core/counters.c b/drivers/infiniband/core/counters.c index df9e6c5e4ddf..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; @@ -100,15 +104,48 @@ 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; } +int rdma_counter_modify(struct ib_device *dev, u32 port, + unsigned int index, bool enable) +{ + struct rdma_hw_stats *stats; + int ret = 0; + + if (!dev->ops.modify_hw_stat) + return -EOPNOTSUPP; + + stats = ib_get_hw_stats_port(dev, port); + if (!stats || index >= stats->num_counters || + !(stats->descs[index].flags & IB_STAT_FLAG_OPTIONAL)) + return -EINVAL; + + mutex_lock(&stats->lock); + + if (enable != test_bit(index, stats->is_disabled)) + goto out; + + ret = dev->ops.modify_hw_stat(dev, port, index, enable); + if (ret) + goto out; + + if (enable) + clear_bit(index, stats->is_disabled); + else + set_bit(index, stats->is_disabled); +out: + mutex_unlock(&stats->lock); + return ret; +} + 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; @@ -117,13 +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) @@ -134,7 +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; @@ -153,10 +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; @@ -165,7 +205,7 @@ static struct rdma_counter *alloc_and_bind(struct ib_device *dev, u32 port, return counter; err_mode: - kfree(counter->stats); + rdma_free_hw_stats_struct(counter->stats); err_stats: rdma_restrack_put(&counter->res); kfree(counter); @@ -181,12 +221,13 @@ 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); rdma_restrack_del(&counter->res); - kfree(counter->stats); + rdma_free_hw_stats_struct(counter->stats); kfree(counter); } @@ -206,7 +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; @@ -215,7 +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; @@ -307,13 +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; } @@ -326,7 +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; @@ -334,7 +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; @@ -419,7 +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; @@ -481,7 +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; @@ -526,7 +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; @@ -572,7 +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); @@ -581,13 +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; } @@ -618,7 +662,7 @@ void rdma_counter_init(struct ib_device *dev) fail: for (i = port; i >= rdma_start_port(dev); i--) { port_counter = &dev->port_data[port].port_counter; - kfree(port_counter->hstats); + rdma_free_hw_stats_struct(port_counter->hstats); port_counter->hstats = NULL; mutex_destroy(&port_counter->lock); } @@ -631,7 +675,7 @@ void rdma_counter_release(struct ib_device *dev) rdma_for_each_port(dev, port) { port_counter = &dev->port_data[port].port_counter; - kfree(port_counter->hstats); + rdma_free_hw_stats_struct(port_counter->hstats); mutex_destroy(&port_counter->lock); } } diff --git a/drivers/infiniband/core/cq.c b/drivers/infiniband/core/cq.c index 433b426729d4..584537c71545 100644 --- a/drivers/infiniband/core/cq.c +++ b/drivers/infiniband/core/cq.c @@ -2,7 +2,6 @@ /* * Copyright (c) 2015 HGST, a Western Digital Company. */ -#include <linux/module.h> #include <linux/err.h> #include <linux/slab.h> #include <rdma/ib_verbs.h> @@ -318,13 +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; @@ -341,7 +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 fa20b1824fb8..13e8a1714bbd 100644 --- a/drivers/infiniband/core/device.c +++ b/drivers/infiniband/core/device.c @@ -58,6 +58,7 @@ struct workqueue_struct *ib_comp_wq; struct workqueue_struct *ib_comp_unbound_wq; struct workqueue_struct *ib_wq; EXPORT_SYMBOL_GPL(ib_wq); +static struct workqueue_struct *ib_unreg_wq; /* * Each of the three rwsem locks (devices, clients, client_data) protects the @@ -144,6 +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 @@ -208,23 +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, ...) \ { \ @@ -421,7 +432,7 @@ int ib_device_rename(struct ib_device *ibdev, const char *name) return ret; } - strlcpy(ibdev->name, name, IB_DEVICE_NAME_MAX); + strscpy(ibdev->name, name, IB_DEVICE_NAME_MAX); ret = rename_compat_devs(ibdev); downgrade_write(&devices_rwsem); @@ -436,6 +447,7 @@ int ib_device_rename(struct ib_device *ibdev, const char *name) client->rename(ibdev, client_data); } up_read(&ibdev->client_data_rwsem); + rdma_nl_notify_event(ibdev, 0, RDMA_RENAME_EVENT); up_read(&devices_rwsem); return 0; } @@ -502,6 +514,7 @@ static void ib_device_release(struct device *device) rcu_head); } + mutex_destroy(&dev->subdev_lock); mutex_destroy(&dev->unregistration_lock); mutex_destroy(&dev->compat_devs_mutex); @@ -510,7 +523,7 @@ static void ib_device_release(struct device *device) kfree_rcu(dev, rcu_head); } -static int ib_device_uevent(struct device *device, +static int ib_device_uevent(const struct device *device, struct kobj_uevent_env *env) { if (add_uevent_var(env, "NAME=%s", dev_name(device))) @@ -523,9 +536,9 @@ static int ib_device_uevent(struct device *device, return 0; } -static const void *net_namespace(struct device *d) +static const void *net_namespace(const struct device *d) { - struct ib_core_device *coredev = + const struct ib_core_device *coredev = container_of(d, struct ib_core_device, dev); return read_pnet(&coredev->rdma_net); @@ -542,6 +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 @@ -553,6 +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); @@ -562,6 +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, @@ -569,7 +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; @@ -586,7 +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); @@ -607,6 +639,8 @@ struct ib_device *_ib_alloc_device(size_t size) for (i = 0; i < ARRAY_SIZE(device->cq_pools); i++) INIT_LIST_HEAD(&device->cq_pools[i]); + rwlock_init(&device->cache_lock); + device->uverbs_cmd_mask = BIT_ULL(IB_USER_VERBS_CMD_ALLOC_MW) | BIT_ULL(IB_USER_VERBS_CMD_ALLOC_PD) | @@ -638,6 +672,11 @@ struct ib_device *_ib_alloc_device(size_t size) BIT_ULL(IB_USER_VERBS_CMD_REG_MR) | BIT_ULL(IB_USER_VERBS_CMD_REREG_MR) | BIT_ULL(IB_USER_VERBS_CMD_RESIZE_CQ); + + mutex_init(&device->subdev_lock); + INIT_LIST_HEAD(&device->subdev_list_head); + INIT_LIST_HEAD(&device->subdev_list); + return device; } EXPORT_SYMBOL(_ib_alloc_device); @@ -801,7 +840,7 @@ static int alloc_port_data(struct ib_device *device) * empty slots at the beginning. */ pdata_rcu = kzalloc(struct_size(pdata_rcu, pdata, - rdma_end_port(device) + 1), + size_add(rdma_end_port(device), 1)), GFP_KERNEL); if (!pdata_rcu) return -ENOMEM; @@ -1214,7 +1253,7 @@ static int assign_name(struct ib_device *device, const char *name) ret = -ENFILE; goto out; } - strlcpy(device->name, dev_name(&device->dev), IB_DEVICE_NAME_MAX); + strscpy(device->name, dev_name(&device->dev), IB_DEVICE_NAME_MAX); ret = xa_alloc_cyclic(&devices, &device->index, device, xa_limit_31b, &last_id, GFP_KERNEL); @@ -1342,6 +1381,37 @@ static void prevent_dealloc_device(struct ib_device *ib_dev) { } +static void ib_device_notify_register(struct ib_device *device) +{ + struct net_device *netdev; + 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) + goto out; + + rdma_for_each_port(device, port) { + netdev = ib_device_get_netdev(device, port); + if (!netdev) + continue; + + ret = rdma_nl_notify_event(device, port, + RDMA_NETDEV_ATTACH_EVENT); + dev_put(netdev); + if (ret) + goto out; + } + +out: + up_read(&devices_rwsem); +} + /** * ib_register_device - Register an IB device with IB core * @device: Device to register @@ -1438,8 +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; @@ -1458,9 +1529,21 @@ EXPORT_SYMBOL(ib_register_device); /* Callers must hold a get on the device. */ static void __ib_unregister_device(struct ib_device *ib_dev) { + struct ib_device *sub, *tmp; + + mutex_lock(&ib_dev->subdev_lock); + list_for_each_entry_safe_reverse(sub, tmp, + &ib_dev->subdev_list_head, + subdev_list) { + list_del(&sub->subdev_list); + ib_dev->ops.del_sub_dev(sub); + ib_device_put(ib_dev); + } + mutex_unlock(&ib_dev->subdev_lock); + /* * 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. @@ -1470,6 +1553,7 @@ static void __ib_unregister_device(struct ib_device *ib_dev) goto out; disable_device(ib_dev); + rdma_nl_notify_event(ib_dev, 0, RDMA_UNREGISTER_EVENT); /* Expedite removing unregistered pointers from the hash table */ free_netdevs(ib_dev); @@ -1600,7 +1684,7 @@ void ib_unregister_device_queued(struct ib_device *ib_dev) WARN_ON(!refcount_read(&ib_dev->refcount)); WARN_ON(!ib_dev->ops.dealloc_driver); get_device(&ib_dev->dev); - if (!queue_work(system_unbound_wq, &ib_dev->unregistration_work)) + if (!queue_work(ib_unreg_wq, &ib_dev->unregistration_work)) put_device(&ib_dev->dev); } EXPORT_SYMBOL(ib_unregister_device_queued); @@ -1727,7 +1811,7 @@ static int assign_client_id(struct ib_client *client) { int ret; - down_write(&clients_rwsem); + lockdep_assert_held(&clients_rwsem); /* * The add/remove callbacks must be called in FIFO/LIFO order. To * achieve this we assign client_ids so they are sorted in @@ -1736,14 +1820,11 @@ static int assign_client_id(struct ib_client *client) client->client_id = highest_client_id; ret = xa_insert(&clients, client->client_id, client, GFP_KERNEL); if (ret) - goto out; + return ret; highest_client_id++; xa_set_mark(&clients, client->client_id, CLIENT_REGISTERED); - -out: - up_write(&clients_rwsem); - return ret; + return 0; } static void remove_client_id(struct ib_client *client) @@ -1773,25 +1854,35 @@ int ib_register_client(struct ib_client *client) { struct ib_device *device; unsigned long index; + bool need_unreg = false; int ret; refcount_set(&client->uses, 1); init_completion(&client->uses_zero); + + /* + * The devices_rwsem is held in write mode to ensure that a racing + * ib_register_device() sees a consisent view of clients and devices. + */ + down_write(&devices_rwsem); + down_write(&clients_rwsem); ret = assign_client_id(client); if (ret) - return ret; + goto out; - down_read(&devices_rwsem); + need_unreg = true; xa_for_each_marked (&devices, index, device, DEVICE_REGISTERED) { ret = add_client_context(device, client); - if (ret) { - up_read(&devices_rwsem); - ib_unregister_client(client); - return ret; - } + if (ret) + goto out; } - up_read(&devices_rwsem); - return 0; + ret = 0; +out: + up_write(&clients_rwsem); + up_write(&devices_rwsem); + if (need_unreg && ret) + ib_unregister_client(client); + return ret; } EXPORT_SYMBOL(ib_register_client); @@ -2050,7 +2141,6 @@ static int __ib_query_port(struct ib_device *device, u32 port_num, struct ib_port_attr *port_attr) { - union ib_gid gid = {}; int err; memset(port_attr, 0, sizeof(*port_attr)); @@ -2063,11 +2153,8 @@ static int __ib_query_port(struct ib_device *device, IB_LINK_LAYER_INFINIBAND) return 0; - err = device->ops.query_gid(device, port_num, 0, &gid); - if (err) - return err; - - port_attr->subnet_prefix = be64_to_cpu(gid.global.subnet_prefix); + ib_get_cached_subnet_prefix(device, port_num, + &port_attr->subnet_prefix); return 0; } @@ -2135,11 +2222,15 @@ static void add_ndev_hash(struct ib_port_data *pdata) int ib_device_set_netdev(struct ib_device *ib_dev, struct net_device *ndev, u32 port) { + enum rdma_nl_notify_event_type etype; struct net_device *old_ndev; struct ib_port_data *pdata; unsigned long flags; int ret; + if (!rdma_is_port_valid(ib_dev, port)) + return -EINVAL; + /* * Drivers wish to call this before ib_register_driver, so we have to * setup the port data early. @@ -2148,9 +2239,6 @@ int ib_device_set_netdev(struct ib_device *ib_dev, struct net_device *ndev, if (ret) return ret; - if (!rdma_is_port_valid(ib_dev, port)) - return -EINVAL; - pdata = &ib_dev->port_data[port]; spin_lock_irqsave(&pdata->netdev_lock, flags); old_ndev = rcu_dereference_protected( @@ -2160,14 +2248,19 @@ int ib_device_set_netdev(struct ib_device *ib_dev, struct net_device *ndev, return 0; } - if (ndev) - dev_hold(ndev); rcu_assign_pointer(pdata->netdev, ndev); + netdev_put(old_ndev, &pdata->netdev_tracker); + netdev_hold(ndev, &pdata->netdev_tracker, GFP_ATOMIC); spin_unlock_irqrestore(&pdata->netdev_lock, flags); add_ndev_hash(pdata); - if (old_ndev) - dev_put(old_ndev); + + /* Make sure that the device is registered before we send events */ + if (xa_load(&devices, ib_dev->index) != ib_dev) + return 0; + + etype = ndev ? RDMA_NETDEV_ATTACH_EVENT : RDMA_NETDEV_DETACH_EVENT; + rdma_nl_notify_event(ib_dev, port, etype); return 0; } @@ -2200,7 +2293,7 @@ static void free_netdevs(struct ib_device *ib_dev) * comparisons after the put */ rcu_assign_pointer(pdata->netdev, NULL); - dev_put(ndev); + netdev_put(ndev, &pdata->netdev_tracker); } spin_unlock_irqrestore(&pdata->netdev_lock, flags); } @@ -2215,6 +2308,9 @@ struct net_device *ib_device_get_netdev(struct ib_device *ib_dev, if (!rdma_is_port_valid(ib_dev, port)) return NULL; + if (!ib_dev->port_data) + return NULL; + pdata = &ib_dev->port_data[port]; /* @@ -2227,22 +2323,40 @@ struct net_device *ib_device_get_netdev(struct ib_device *ib_dev, spin_lock(&pdata->netdev_lock); res = rcu_dereference_protected( pdata->netdev, lockdep_is_held(&pdata->netdev_lock)); - if (res) - dev_hold(res); + dev_hold(res); spin_unlock(&pdata->netdev_lock); } - /* - * If we are starting to unregister expedite things by preventing - * propagation of an unregistering netdev. - */ - if (res && res->reg_state != NETREG_REGISTERED) { - dev_put(res); - return NULL; + return res; +} +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 res; + return -ENOENT; } +EXPORT_SYMBOL(ib_query_netdev_port); /** * ib_device_get_by_netdev - Find an IB device associated with a netdev @@ -2303,9 +2417,7 @@ void ib_enum_roce_netdev(struct ib_device *ib_dev, if (filter(ib_dev, port, idev, filter_cookie)) cb(ib_dev, port, idev, cookie); - - if (idev) - dev_put(idev); + dev_put(idev); } } @@ -2463,7 +2575,8 @@ int ib_find_gid(struct ib_device *device, union ib_gid *gid, ++i) { ret = rdma_query_gid(device, port, i, &tmp_gid); if (ret) - return ret; + continue; + if (!memcmp(&tmp_gid, gid, sizeof *gid)) { *port_num = port; if (index) @@ -2592,8 +2705,10 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) ops->uverbs_no_driver_id_binding; SET_DEVICE_OP(dev_ops, add_gid); + 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); @@ -2608,25 +2723,28 @@ 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_flow_action_esp); SET_DEVICE_OP(dev_ops, create_qp); SET_DEVICE_OP(dev_ops, create_rwq_ind_table); SET_DEVICE_OP(dev_ops, create_srq); 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); SET_DEVICE_OP(dev_ops, dealloc_ucontext); SET_DEVICE_OP(dev_ops, dealloc_xrcd); SET_DEVICE_OP(dev_ops, del_gid); + SET_DEVICE_OP(dev_ops, del_sub_dev); SET_DEVICE_OP(dev_ops, dereg_mr); SET_DEVICE_OP(dev_ops, destroy_ah); SET_DEVICE_OP(dev_ops, destroy_counters); @@ -2650,12 +2768,15 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, fill_res_mr_entry_raw); SET_DEVICE_OP(dev_ops, fill_res_qp_entry); SET_DEVICE_OP(dev_ops, fill_res_qp_entry_raw); + SET_DEVICE_OP(dev_ops, fill_res_srq_entry); + SET_DEVICE_OP(dev_ops, fill_res_srq_entry_raw); SET_DEVICE_OP(dev_ops, fill_stat_mr_entry); SET_DEVICE_OP(dev_ops, get_dev_fw_str); SET_DEVICE_OP(dev_ops, get_dma_mr); SET_DEVICE_OP(dev_ops, get_hw_stats); SET_DEVICE_OP(dev_ops, get_link_layer); SET_DEVICE_OP(dev_ops, get_netdev); + SET_DEVICE_OP(dev_ops, get_numa_node); SET_DEVICE_OP(dev_ops, get_port_immutable); SET_DEVICE_OP(dev_ops, get_vector_affinity); SET_DEVICE_OP(dev_ops, get_vf_config); @@ -2676,14 +2797,16 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, modify_ah); SET_DEVICE_OP(dev_ops, modify_cq); SET_DEVICE_OP(dev_ops, modify_device); - SET_DEVICE_OP(dev_ops, modify_flow_action_esp); + SET_DEVICE_OP(dev_ops, modify_hw_stat); SET_DEVICE_OP(dev_ops, modify_port); SET_DEVICE_OP(dev_ops, modify_qp); 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); @@ -2706,19 +2829,73 @@ void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops) SET_DEVICE_OP(dev_ops, resize_cq); 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); SET_OBJ_SIZE(dev_ops, ib_rwq_ind_table); 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); +int ib_add_sub_device(struct ib_device *parent, + enum rdma_nl_dev_type type, + const char *name) +{ + struct ib_device *sub; + int ret = 0; + + if (!parent->ops.add_sub_dev || !parent->ops.del_sub_dev) + return -EOPNOTSUPP; + + if (!ib_device_try_get(parent)) + return -EINVAL; + + sub = parent->ops.add_sub_dev(parent, type, name); + if (IS_ERR(sub)) { + ib_device_put(parent); + return PTR_ERR(sub); + } + + sub->type = type; + sub->parent = parent; + + mutex_lock(&parent->subdev_lock); + list_add_tail(&parent->subdev_list_head, &sub->subdev_list); + mutex_unlock(&parent->subdev_lock); + + return ret; +} +EXPORT_SYMBOL(ib_add_sub_device); + +int ib_del_sub_device_and_put(struct ib_device *sub) +{ + struct ib_device *parent = sub->parent; + + if (!parent) + return -EOPNOTSUPP; + + mutex_lock(&parent->subdev_lock); + list_del(&sub->subdev_list); + mutex_unlock(&parent->subdev_lock); + + ib_device_put(sub); + parent->ops.del_sub_dev(sub); + ib_device_put(parent); + + return 0; +} +EXPORT_SYMBOL(ib_del_sub_device_and_put); + #ifdef CONFIG_INFINIBAND_VIRT_DMA int ib_dma_virt_map_sg(struct ib_device *dev, struct scatterlist *sg, int nents) { @@ -2749,29 +2926,121 @@ 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 ib_device *ibdev; + u32 port; + + switch (event) { + case NETDEV_CHANGENAME: + ibdev = ib_device_get_by_netdev(ndev, RDMA_DRIVER_UNKNOWN); + if (!ibdev) + return NOTIFY_DONE; + + 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; + } + + return NOTIFY_DONE; +} + +static struct notifier_block nb_netdevice = { + .notifier_call = ib_netdevice_event, +}; + static int __init ib_core_init(void) { - int ret; + int ret = -ENOMEM; - ib_wq = alloc_workqueue("infiniband", 0, 0); + ib_wq = alloc_workqueue("infiniband", WQ_PERCPU, 0); if (!ib_wq) return -ENOMEM; - ib_comp_wq = alloc_workqueue("ib-comp-wq", - WQ_HIGHPRI | WQ_MEM_RECLAIM | WQ_SYSFS, 0); - if (!ib_comp_wq) { - ret = -ENOMEM; + ib_unreg_wq = alloc_workqueue("ib-unreg-wq", WQ_UNBOUND, + WQ_UNBOUND_MAX_ACTIVE); + if (!ib_unreg_wq) goto err; - } + + ib_comp_wq = alloc_workqueue("ib-comp-wq", + WQ_HIGHPRI | WQ_MEM_RECLAIM | WQ_SYSFS | WQ_PERCPU, 0); + if (!ib_comp_wq) + goto err_unbound; ib_comp_unbound_wq = alloc_workqueue("ib-comp-unb-wq", WQ_UNBOUND | WQ_HIGHPRI | WQ_MEM_RECLAIM | WQ_SYSFS, WQ_UNBOUND_MAX_ACTIVE); - if (!ib_comp_unbound_wq) { - ret = -ENOMEM; + if (!ib_comp_unbound_wq) goto err_comp; - } ret = class_register(&ib_class); if (ret) { @@ -2813,10 +3082,20 @@ static int __init ib_core_init(void) nldev_init(); rdma_nl_register(RDMA_NL_LS, ibnl_ls_cb_table); - roce_gid_mgmt_init(); + ret = roce_gid_mgmt_init(); + if (ret) { + pr_warn("Couldn't init RoCE GID management\n"); + goto err_parent; + } + + register_netdevice_notifier(&nb_netdevice); return 0; +err_parent: + rdma_nl_unregister(RDMA_NL_LS); + nldev_exit(); + unregister_pernet_device(&rdma_dev_net_ops); err_compat: unregister_blocking_lsm_notifier(&ibdev_lsm_nb); err_sa: @@ -2831,6 +3110,8 @@ err_comp_unbound: destroy_workqueue(ib_comp_unbound_wq); err_comp: destroy_workqueue(ib_comp_wq); +err_unbound: + destroy_workqueue(ib_unreg_wq); err: destroy_workqueue(ib_wq); return ret; @@ -2838,9 +3119,10 @@ err: static void __exit ib_core_cleanup(void) { + unregister_netdevice_notifier(&nb_netdevice); roce_gid_mgmt_cleanup(); - nldev_exit(); rdma_nl_unregister(RDMA_NL_LS); + nldev_exit(); unregister_pernet_device(&rdma_dev_net_ops); unregister_blocking_lsm_notifier(&ibdev_lsm_nb); ib_sa_cleanup(); @@ -2852,7 +3134,7 @@ static void __exit ib_core_cleanup(void) destroy_workqueue(ib_comp_wq); /* Make sure that any pending umem accounting work is done. */ destroy_workqueue(ib_wq); - flush_workqueue(system_unbound_wq); + destroy_workqueue(ib_unreg_wq); WARN_ON(!xa_empty(&clients)); WARN_ON(!xa_empty(&devices)); } diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c index 42261152b489..62410578dec3 100644 --- a/drivers/infiniband/core/iwcm.c +++ b/drivers/infiniband/core/iwcm.c @@ -109,9 +109,10 @@ 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, }, - { } }; /* @@ -144,8 +145,8 @@ static struct iwcm_work *get_work(struct iwcm_id_private *cm_id_priv) if (list_empty(&cm_id_priv->work_free_list)) return NULL; - work = list_entry(cm_id_priv->work_free_list.next, struct iwcm_work, - free_list); + work = list_first_entry(&cm_id_priv->work_free_list, struct iwcm_work, + free_list); list_del_init(&work->free_list); return work; } @@ -207,17 +208,17 @@ static void free_cm_id(struct iwcm_id_private *cm_id_priv) /* * Release a reference on cm_id. If the last reference is being - * released, free the cm_id and return 1. + * released, free the cm_id and return 'true'. */ -static int iwcm_deref_id(struct iwcm_id_private *cm_id_priv) +static bool iwcm_deref_id(struct iwcm_id_private *cm_id_priv) { if (refcount_dec_and_test(&cm_id_priv->refcount)) { BUG_ON(!list_empty(&cm_id_priv->work_list)); free_cm_id(cm_id_priv); - return 1; + return true; } - return 0; + return false; } static void add_ref(struct iw_cm_id *cm_id) @@ -367,8 +368,7 @@ 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. + * Clean up all resources associated with the connection. */ static void destroy_cm_id(struct iw_cm_id *cm_id) { @@ -439,19 +439,22 @@ static void 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); } - - (void)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) { + 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); @@ -1018,30 +1021,27 @@ static void cm_work_handler(struct work_struct *_work) struct iw_cm_event levent; struct iwcm_id_private *cm_id_priv = work->cm_id; unsigned long flags; - int empty; int ret = 0; spin_lock_irqsave(&cm_id_priv->lock, flags); - empty = list_empty(&cm_id_priv->work_list); - while (!empty) { - work = list_entry(cm_id_priv->work_list.next, - struct iwcm_work, list); + while (!list_empty(&cm_id_priv->work_list)) { + work = list_first_entry(&cm_id_priv->work_list, + struct iwcm_work, list); list_del_init(&work->list); - empty = list_empty(&cm_id_priv->work_list); levent = work->event; put_work(work); spin_unlock_irqrestore(&cm_id_priv->lock, flags); if (!test_bit(IWCM_F_DROP_EVENTS, &cm_id_priv->flags)) { ret = process_event(cm_id_priv, &levent); - if (ret) + 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)) return; - if (empty) - return; spin_lock_irqsave(&cm_id_priv->lock, flags); } spin_unlock_irqrestore(&cm_id_priv->lock, flags); @@ -1094,11 +1094,8 @@ static int cm_event_handler(struct iw_cm_id *cm_id, } refcount_inc(&cm_id_priv->refcount); - if (list_empty(&cm_id_priv->work_list)) { - list_add_tail(&work->list, &cm_id_priv->work_list); - queue_work(iwcm_wq, &work->work); - } else - list_add_tail(&work->list, &cm_id_priv->work_list); + list_add_tail(&work->list, &cm_id_priv->work_list); + queue_work(iwcm_wq, &work->work); out: spin_unlock_irqrestore(&cm_id_priv->lock, flags); return ret; @@ -1186,29 +1183,34 @@ static int __init iw_cm_init(void) ret = iwpm_init(RDMA_NL_IWCM); if (ret) - pr_err("iw_cm: couldn't init iwpm\n"); - else - rdma_nl_register(RDMA_NL_IWCM, iwcm_nl_cb_table); - iwcm_wq = alloc_ordered_workqueue("iw_cm_wq", 0); + return ret; + + iwcm_wq = alloc_ordered_workqueue("iw_cm_wq", WQ_MEM_RECLAIM); if (!iwcm_wq) - return -ENOMEM; + goto err_alloc; iwcm_ctl_table_hdr = register_net_sysctl(&init_net, "net/iw_cm", iwcm_ctl_table); if (!iwcm_ctl_table_hdr) { pr_err("iw_cm: couldn't register sysctl paths\n"); - destroy_workqueue(iwcm_wq); - return -ENOMEM; + goto err_sysctl; } + rdma_nl_register(RDMA_NL_IWCM, iwcm_nl_cb_table); return 0; + +err_sysctl: + destroy_workqueue(iwcm_wq); +err_alloc: + iwpm_exit(RDMA_NL_IWCM); + return -ENOMEM; } static void __exit iw_cm_cleanup(void) { + rdma_nl_unregister(RDMA_NL_IWCM); unregister_net_sysctl_table(iwcm_ctl_table_hdr); destroy_workqueue(iwcm_wq); - rdma_nl_unregister(RDMA_NL_IWCM); iwpm_exit(RDMA_NL_IWCM); } diff --git a/drivers/infiniband/core/iwpm_msg.c b/drivers/infiniband/core/iwpm_msg.c index 12a9816fc0e2..3c9a9869212b 100644 --- a/drivers/infiniband/core/iwpm_msg.c +++ b/drivers/infiniband/core/iwpm_msg.c @@ -69,10 +69,6 @@ int iwpm_register_pid(struct iwpm_dev_data *pm_msg, u8 nl_client) const char *err_str = ""; int ret = -EINVAL; - if (!iwpm_valid_client(nl_client)) { - err_str = "Invalid port mapper client"; - goto pid_query_error; - } if (iwpm_check_registration(nl_client, IWPM_REG_VALID) || iwpm_user_pid == IWPM_PID_UNAVAILABLE) return 0; @@ -153,10 +149,6 @@ int iwpm_add_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client) const char *err_str = ""; int ret = -EINVAL; - if (!iwpm_valid_client(nl_client)) { - err_str = "Invalid port mapper client"; - goto add_mapping_error; - } if (!iwpm_valid_pid()) return 0; if (!iwpm_check_registration(nl_client, IWPM_REG_VALID)) { @@ -240,10 +232,6 @@ int iwpm_add_and_query_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client) const char *err_str = ""; int ret = -EINVAL; - if (!iwpm_valid_client(nl_client)) { - err_str = "Invalid port mapper client"; - goto query_mapping_error; - } if (!iwpm_valid_pid()) return 0; if (!iwpm_check_registration(nl_client, IWPM_REG_VALID)) { @@ -331,10 +319,6 @@ int iwpm_remove_mapping(struct sockaddr_storage *local_addr, u8 nl_client) const char *err_str = ""; int ret = -EINVAL; - if (!iwpm_valid_client(nl_client)) { - err_str = "Invalid port mapper client"; - goto remove_mapping_error; - } if (!iwpm_valid_pid()) return 0; if (iwpm_check_registration(nl_client, IWPM_REG_UNDEF)) { @@ -444,8 +428,7 @@ int iwpm_register_pid_cb(struct sk_buff *skb, struct netlink_callback *cb) atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); pr_debug("%s: iWarp Port Mapper (pid = %d) is available!\n", __func__, iwpm_user_pid); - if (iwpm_valid_client(nl_client)) - iwpm_set_registration(nl_client, IWPM_REG_VALID); + iwpm_set_registration(nl_client, IWPM_REG_VALID); register_pid_response_exit: nlmsg_request->request_done = 1; /* always for found nlmsg_request */ @@ -649,11 +632,6 @@ int iwpm_remote_info_cb(struct sk_buff *skb, struct netlink_callback *cb) return ret; nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type); - if (!iwpm_valid_client(nl_client)) { - pr_info("%s: Invalid port mapper client = %u\n", - __func__, nl_client); - return ret; - } atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); local_sockaddr = (struct sockaddr_storage *) @@ -736,11 +714,6 @@ int iwpm_mapping_info_cb(struct sk_buff *skb, struct netlink_callback *cb) return ret; } nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type); - if (!iwpm_valid_client(nl_client)) { - pr_info("%s: Invalid port mapper client = %u\n", - __func__, nl_client); - return ret; - } iwpm_set_registration(nl_client, IWPM_REG_INCOMPL); atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); iwpm_user_pid = cb->nlh->nlmsg_pid; @@ -863,11 +836,6 @@ int iwpm_hello_cb(struct sk_buff *skb, struct netlink_callback *cb) } abi_version = nla_get_u16(nltb[IWPM_NLA_HELLO_ABI_VERSION]); nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type); - if (!iwpm_valid_client(nl_client)) { - pr_info("%s: Invalid port mapper client = %u\n", - __func__, nl_client); - return ret; - } iwpm_set_registration(nl_client, IWPM_REG_INCOMPL); atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); iwpm_ulib_version = min_t(u16, IWPM_UABI_VERSION, abi_version); diff --git a/drivers/infiniband/core/iwpm_util.c b/drivers/infiniband/core/iwpm_util.c index 3f8c019c7260..eecb369898f5 100644 --- a/drivers/infiniband/core/iwpm_util.c +++ b/drivers/infiniband/core/iwpm_util.c @@ -48,7 +48,6 @@ static DEFINE_SPINLOCK(iwpm_mapinfo_lock); static struct hlist_head *iwpm_reminfo_bucket; static DEFINE_SPINLOCK(iwpm_reminfo_lock); -static DEFINE_MUTEX(iwpm_admin_lock); static struct iwpm_admin_data iwpm_admin; /** @@ -59,39 +58,21 @@ static struct iwpm_admin_data iwpm_admin; */ int iwpm_init(u8 nl_client) { - int ret = 0; - mutex_lock(&iwpm_admin_lock); - if (!refcount_read(&iwpm_admin.refcount)) { - iwpm_hash_bucket = kcalloc(IWPM_MAPINFO_HASH_SIZE, - sizeof(struct hlist_head), - GFP_KERNEL); - if (!iwpm_hash_bucket) { - ret = -ENOMEM; - goto init_exit; - } - iwpm_reminfo_bucket = kcalloc(IWPM_REMINFO_HASH_SIZE, - sizeof(struct hlist_head), - GFP_KERNEL); - if (!iwpm_reminfo_bucket) { - kfree(iwpm_hash_bucket); - ret = -ENOMEM; - goto init_exit; - } + iwpm_hash_bucket = kcalloc(IWPM_MAPINFO_HASH_SIZE, + sizeof(struct hlist_head), GFP_KERNEL); + if (!iwpm_hash_bucket) + return -ENOMEM; - refcount_set(&iwpm_admin.refcount, 1); - } else { - refcount_inc(&iwpm_admin.refcount); + iwpm_reminfo_bucket = kcalloc(IWPM_REMINFO_HASH_SIZE, + sizeof(struct hlist_head), GFP_KERNEL); + if (!iwpm_reminfo_bucket) { + kfree(iwpm_hash_bucket); + return -ENOMEM; } -init_exit: - mutex_unlock(&iwpm_admin_lock); - if (!ret) { - iwpm_set_valid(nl_client, 1); - iwpm_set_registration(nl_client, IWPM_REG_UNDEF); - pr_debug("%s: Mapinfo and reminfo tables are created\n", - __func__); - } - return ret; + iwpm_set_registration(nl_client, IWPM_REG_UNDEF); + pr_debug("%s: Mapinfo and reminfo tables are created\n", __func__); + return 0; } static void free_hash_bucket(void); @@ -105,22 +86,9 @@ static void free_reminfo_bucket(void); */ int iwpm_exit(u8 nl_client) { - - if (!iwpm_valid_client(nl_client)) - return -EINVAL; - mutex_lock(&iwpm_admin_lock); - if (!refcount_read(&iwpm_admin.refcount)) { - mutex_unlock(&iwpm_admin_lock); - pr_err("%s Incorrect usage - negative refcount\n", __func__); - return -EINVAL; - } - if (refcount_dec_and_test(&iwpm_admin.refcount)) { - free_hash_bucket(); - free_reminfo_bucket(); - pr_debug("%s: Resources are destroyed\n", __func__); - } - mutex_unlock(&iwpm_admin_lock); - iwpm_set_valid(nl_client, 0); + free_hash_bucket(); + free_reminfo_bucket(); + pr_debug("%s: Resources are destroyed\n", __func__); iwpm_set_registration(nl_client, IWPM_REG_UNDEF); return 0; } @@ -145,8 +113,6 @@ int iwpm_create_mapinfo(struct sockaddr_storage *local_sockaddr, unsigned long flags; int ret = -EINVAL; - if (!iwpm_valid_client(nl_client)) - return ret; map_info = kzalloc(sizeof(struct iwpm_mapping_info), GFP_KERNEL); if (!map_info) return -ENOMEM; @@ -306,10 +272,6 @@ int iwpm_get_remote_info(struct sockaddr_storage *mapped_loc_addr, unsigned long flags; int ret = -EINVAL; - if (!iwpm_valid_client(nl_client)) { - pr_info("%s: Invalid client = %u\n", __func__, nl_client); - return ret; - } spin_lock_irqsave(&iwpm_reminfo_lock, flags); if (iwpm_reminfo_bucket) { hash_bucket_head = get_reminfo_hash_bucket( @@ -345,7 +307,7 @@ get_remote_info_exit: struct iwpm_nlmsg_request *iwpm_get_nlmsg_request(__u32 nlmsg_seq, u8 nl_client, gfp_t gfp) { - struct iwpm_nlmsg_request *nlmsg_request = NULL; + struct iwpm_nlmsg_request *nlmsg_request; unsigned long flags; nlmsg_request = kzalloc(sizeof(struct iwpm_nlmsg_request), gfp); @@ -424,16 +386,6 @@ int iwpm_get_nlmsg_seq(void) return atomic_inc_return(&iwpm_admin.nlmsg_seq); } -int iwpm_valid_client(u8 nl_client) -{ - return iwpm_admin.client_list[nl_client]; -} - -void iwpm_set_valid(u8 nl_client, int valid) -{ - iwpm_admin.client_list[nl_client] = valid; -} - /* valid client */ u32 iwpm_get_registration(u8 nl_client) { @@ -810,7 +762,7 @@ int iwpm_send_hello(u8 nl_client, int iwpm_pid, u16 abi_version) { struct sk_buff *skb = NULL; struct nlmsghdr *nlh; - const char *err_str = ""; + const char *err_str; int ret = -EINVAL; skb = iwpm_create_nlmsg(RDMA_NL_IWPM_HELLO, &nlh, nl_client); diff --git a/drivers/infiniband/core/iwpm_util.h b/drivers/infiniband/core/iwpm_util.h index e201835de733..d6fc8402158a 100644 --- a/drivers/infiniband/core/iwpm_util.h +++ b/drivers/infiniband/core/iwpm_util.h @@ -33,7 +33,6 @@ #ifndef _IWPM_UTIL_H #define _IWPM_UTIL_H -#include <linux/module.h> #include <linux/io.h> #include <linux/in.h> #include <linux/in6.h> @@ -90,9 +89,7 @@ struct iwpm_remote_info { }; struct iwpm_admin_data { - refcount_t refcount; atomic_t nlmsg_seq; - int client_list[RDMA_NL_NUM_CLIENTS]; u32 reg_list[RDMA_NL_NUM_CLIENTS]; }; @@ -148,22 +145,6 @@ int iwpm_get_nlmsg_seq(void); void iwpm_add_remote_info(struct iwpm_remote_info *reminfo); /** - * iwpm_valid_client - Check if the port mapper client is valid - * @nl_client: The index of the netlink client - * - * Valid clients need to call iwpm_init() before using - * the port mapper - */ -int iwpm_valid_client(u8 nl_client); - -/** - * iwpm_set_valid - Set the port mapper client to valid or not - * @nl_client: The index of the netlink client - * @valid: 1 if valid or 0 if invalid - */ -void iwpm_set_valid(u8 nl_client, int valid); - -/** * iwpm_check_registration - Check if the client registration * matches the given one * @nl_client: The index of the netlink client diff --git a/drivers/infiniband/core/lag.c b/drivers/infiniband/core/lag.c index 7063e41eaf26..8fd80adfe833 100644 --- a/drivers/infiniband/core/lag.c +++ b/drivers/infiniband/core/lag.c @@ -7,8 +7,7 @@ #include <rdma/ib_cache.h> #include <rdma/lag.h> -static struct sk_buff *rdma_build_skb(struct ib_device *device, - struct net_device *netdev, +static struct sk_buff *rdma_build_skb(struct net_device *netdev, struct rdma_ah_attr *ah_attr, gfp_t flags) { @@ -86,7 +85,7 @@ static struct net_device *rdma_get_xmit_slave_udp(struct ib_device *device, struct net_device *slave; struct sk_buff *skb; - skb = rdma_build_skb(device, master, ah_attr, flags); + skb = rdma_build_skb(master, ah_attr, flags); if (!skb) return ERR_PTR(-ENOMEM); @@ -94,8 +93,7 @@ static struct net_device *rdma_get_xmit_slave_udp(struct ib_device *device, slave = netdev_get_xmit_slave(master, skb, !!(device->lag_flags & RDMA_LAG_FLAGS_HASH_ALL_SLAVES)); - if (slave) - dev_hold(slave); + dev_hold(slave); rcu_read_unlock(); kfree_skb(skb); return slave; @@ -103,8 +101,7 @@ static struct net_device *rdma_get_xmit_slave_udp(struct ib_device *device, void rdma_lag_put_ah_roce_slave(struct net_device *xmit_slave) { - if (xmit_slave) - dev_put(xmit_slave); + dev_put(xmit_slave); } struct net_device *rdma_lag_get_ah_roce_slave(struct ib_device *device, diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index 1893aa613ad7..8f26bfb69586 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -59,9 +59,6 @@ static void create_mad_addr_info(struct ib_mad_send_wr_private *mad_send_wr, struct ib_mad_qp_info *qp_info, struct trace_event_raw_ib_mad_send_template *entry) { - u16 pkey; - struct ib_device *dev = qp_info->port_priv->device; - u32 pnum = qp_info->port_priv->port_num; struct ib_ud_wr *wr = &mad_send_wr->send_wr; struct rdma_ah_attr attr = {}; @@ -69,8 +66,6 @@ static void create_mad_addr_info(struct ib_mad_send_wr_private *mad_send_wr, /* These are common */ entry->sl = attr.sl; - ib_query_pkey(dev, pnum, wr->pkey_index, &pkey); - entry->pkey = pkey; entry->rqpn = wr->remote_qpn; entry->rqkey = wr->remote_qkey; entry->dlid = rdma_ah_get_dlid(&attr); @@ -215,6 +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 * @@ -396,13 +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) { @@ -1060,6 +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 @@ -1085,9 +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; } @@ -1123,15 +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)) { @@ -1143,7 +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; @@ -1751,7 +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; } /* @@ -1770,17 +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, @@ -1789,6 +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); @@ -1837,6 +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 */ @@ -1846,10 +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, @@ -2177,30 +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) @@ -2234,32 +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); @@ -2401,40 +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); - } - } + &mad_agent_priv->send_list, agent_list) + change_mad_state(mad_send_wr, IB_MAD_STATE_CANCELED); - /* 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); - - /* 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* @@ -2456,6 +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; } @@ -2473,16 +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) @@ -2594,6 +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); @@ -2611,24 +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_agent_private *mad_agent_priv; struct ib_mad_send_wr_private *mad_send_wr; - struct ib_mad_send_wc mad_send_wc; + struct ib_mad_agent_private *mad_agent_priv; + 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(&timeout_list); + INIT_LIST_HEAD(&cancel_list); spin_lock_irqsave(&mad_agent_priv->lock, flags); while (!list_empty(&mad_agent_priv->wait_list)) { @@ -2646,25 +2902,22 @@ static void timeout_sends(struct work_struct *work) break; } - list_del(&mad_send_wr->agent_list); - if (mad_send_wr->status == IB_WC_SUCCESS && - !retry_send(mad_send_wr)) - continue; - - spin_unlock_irqrestore(&mad_agent_priv->lock, flags); - - if (mad_send_wr->status == IB_WC_SUCCESS) - mad_send_wc.status = IB_WC_RESP_TIMEOUT_ERR; + 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 - 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); + continue; - deref_mad_agent(mad_agent_priv); - spin_lock_irqsave(&mad_agent_priv->lock, flags); + 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); + 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); } /* @@ -2674,11 +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; @@ -2688,7 +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; @@ -2696,10 +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, @@ -2708,37 +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; } @@ -2942,7 +3197,6 @@ static int ib_mad_port_open(struct ib_device *device, int ret, cq_size; struct ib_mad_port_private *port_priv; unsigned long flags; - char name[sizeof "ib_mad123"]; int has_smi; if (WARN_ON(rdma_max_mad_size(device, port_num) < IB_MGMT_MAD_SIZE)) @@ -2988,12 +3242,15 @@ static int ib_mad_port_open(struct ib_device *device, if (ret) goto error6; } - ret = create_mad_qp(&port_priv->qp_info[1], IB_QPT_GSI); - if (ret) - goto error7; - snprintf(name, sizeof(name), "ib_mad%u", port_num); - port_priv->wq = alloc_ordered_workqueue(name, WQ_MEM_RECLAIM); + if (rdma_cap_ib_cm(device, port_num)) { + ret = create_mad_qp(&port_priv->qp_info[1], IB_QPT_GSI); + if (ret) + goto error7; + } + + port_priv->wq = alloc_ordered_workqueue("ib_mad%u", WQ_MEM_RECLAIM, + port_num); if (!port_priv->wq) { ret = -ENOMEM; goto error8; 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/netlink.c b/drivers/infiniband/core/netlink.c index 1b2cc9e45ade..def14c54b648 100644 --- a/drivers/infiniband/core/netlink.c +++ b/drivers/infiniband/core/netlink.c @@ -75,7 +75,7 @@ static bool is_nl_msg_valid(unsigned int type, unsigned int op) if (type >= RDMA_NL_NUM_CLIENTS) return false; - return (op < max_num_ops[type]) ? true : false; + return op < max_num_ops[type]; } static const struct rdma_nl_cbs * @@ -311,6 +311,7 @@ int rdma_nl_net_init(struct rdma_dev_net *rnet) struct net *net = read_pnet(&rnet->net); struct netlink_kernel_cfg cfg = { .input = rdma_nl_rcv, + .flags = NL_CFG_F_NONROOT_RECV, }; struct sock *nls; diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c index e9b4b2cccaa0..2220a2dfab24 100644 --- a/drivers/infiniband/core/nldev.c +++ b/drivers/infiniband/core/nldev.c @@ -43,6 +43,13 @@ #include "restrack.h" #include "uverbs.h" +/* + * This determines whether a non-privileged user is allowed to specify a + * controlled QKEY or not, when true non-privileged user is allowed to specify + * a controlled QKEY. + */ +static bool privileged_qkey; + typedef int (*res_fill_func_t)(struct sk_buff*, bool, struct rdma_restrack_entry*, uint32_t); @@ -130,6 +137,8 @@ static const struct nla_policy nldev_policy[RDMA_NLDEV_ATTR_MAX] = { [RDMA_NLDEV_ATTR_RES_SUMMARY_ENTRY_NAME]= { .type = NLA_NUL_STRING, .len = RDMA_NLDEV_ATTR_EMPTY_STRING }, [RDMA_NLDEV_ATTR_RES_TYPE] = { .type = NLA_U8 }, + [RDMA_NLDEV_ATTR_RES_SUBTYPE] = { .type = NLA_NUL_STRING, + .len = RDMA_NLDEV_ATTR_EMPTY_STRING }, [RDMA_NLDEV_ATTR_RES_UNSAFE_GLOBAL_RKEY]= { .type = NLA_U32 }, [RDMA_NLDEV_ATTR_RES_USECNT] = { .type = NLA_U64 }, [RDMA_NLDEV_ATTR_RES_SRQ] = { .type = NLA_NESTED }, @@ -154,6 +163,15 @@ static const struct nla_policy nldev_policy[RDMA_NLDEV_ATTR_MAX] = { [RDMA_NLDEV_NET_NS_FD] = { .type = NLA_U32 }, [RDMA_NLDEV_SYS_ATTR_NETNS_MODE] = { .type = NLA_U8 }, [RDMA_NLDEV_SYS_ATTR_COPY_ON_FORK] = { .type = NLA_U8 }, + [RDMA_NLDEV_ATTR_STAT_HWCOUNTER_INDEX] = { .type = NLA_U32 }, + [RDMA_NLDEV_ATTR_STAT_HWCOUNTER_DYNAMIC] = { .type = NLA_U8 }, + [RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE] = { .type = NLA_U8 }, + [RDMA_NLDEV_ATTR_DRIVER_DETAILS] = { .type = NLA_U8 }, + [RDMA_NLDEV_ATTR_DEV_TYPE] = { .type = NLA_U8 }, + [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, @@ -235,6 +253,12 @@ int rdma_nl_put_driver_u64_hex(struct sk_buff *msg, const char *name, u64 value) } EXPORT_SYMBOL(rdma_nl_put_driver_u64_hex); +bool rdma_nl_get_privileged_qkey(void) +{ + return privileged_qkey; +} +EXPORT_SYMBOL(rdma_nl_get_privileged_qkey); + static int fill_nldev_handle(struct sk_buff *msg, struct ib_device *device) { if (nla_put_u32(msg, RDMA_NLDEV_ATTR_DEV_INDEX, device->index)) @@ -282,6 +306,19 @@ static int fill_dev_info(struct sk_buff *msg, struct ib_device *device) if (nla_put_u8(msg, RDMA_NLDEV_ATTR_DEV_DIM, device->use_cq_dim)) return -EMSGSIZE; + if (device->type && + nla_put_u8(msg, RDMA_NLDEV_ATTR_DEV_TYPE, device->type)) + return -EMSGSIZE; + + if (device->parent && + nla_put_string(msg, RDMA_NLDEV_ATTR_PARENT_NAME, + dev_name(&device->parent->dev))) + return -EMSGSIZE; + + if (nla_put_u8(msg, RDMA_NLDEV_ATTR_NAME_ASSIGN_TYPE, + device->name_assign_type)) + return -EMSGSIZE; + /* * Link type is determined on first port and mlx4 device * which can potentially have two different link type for the same @@ -355,8 +392,7 @@ static int fill_port_info(struct sk_buff *msg, } out: - if (netdev) - dev_put(netdev); + dev_put(netdev); return ret; } @@ -384,7 +420,8 @@ err: return -EMSGSIZE; } -static int fill_res_info(struct sk_buff *msg, struct ib_device *device) +static int fill_res_info(struct sk_buff *msg, struct ib_device *device, + bool show_details) { static const char * const names[RDMA_RESTRACK_MAX] = { [RDMA_RESTRACK_PD] = "pd", @@ -409,7 +446,7 @@ static int fill_res_info(struct sk_buff *msg, struct ib_device *device) for (i = 0; i < RDMA_RESTRACK_MAX; i++) { if (!names[i]) continue; - curr = rdma_restrack_count(device, i); + curr = rdma_restrack_count(device, i, show_details); ret = fill_res_info_entry(msg, names[i], curr); if (ret) goto err; @@ -511,7 +548,7 @@ static int fill_res_qp_entry(struct sk_buff *msg, bool has_cap_net_admin, /* In create_qp() port is not set yet */ if (qp->port && nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, qp->port)) - return -EINVAL; + return -EMSGSIZE; ret = nla_put_u32(msg, RDMA_NLDEV_ATTR_RES_LQPN, qp->qp_num); if (ret) @@ -550,7 +587,7 @@ static int fill_res_cm_id_entry(struct sk_buff *msg, bool has_cap_net_admin, struct rdma_cm_id *cm_id = &id_priv->id; if (port && port != cm_id->port_num) - return 0; + return -EAGAIN; if (cm_id->port_num && nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, cm_id->port_num)) @@ -816,6 +853,7 @@ static int fill_res_srq_entry(struct sk_buff *msg, bool has_cap_net_admin, struct rdma_restrack_entry *res, uint32_t port) { struct ib_srq *srq = container_of(res, struct ib_srq, res); + struct ib_device *dev = srq->device; if (nla_put_u32(msg, RDMA_NLDEV_ATTR_RES_SRQN, srq->res.id)) goto err; @@ -835,12 +873,29 @@ static int fill_res_srq_entry(struct sk_buff *msg, bool has_cap_net_admin, if (fill_res_srq_qps(msg, srq)) goto err; - return fill_res_name_pid(msg, res); + if (fill_res_name_pid(msg, res)) + goto err; + + if (dev->ops.fill_res_srq_entry) + return dev->ops.fill_res_srq_entry(msg, srq); + + return 0; err: return -EMSGSIZE; } +static int fill_res_srq_raw_entry(struct sk_buff *msg, bool has_cap_net_admin, + struct rdma_restrack_entry *res, uint32_t port) +{ + struct ib_srq *srq = container_of(res, struct ib_srq, res); + struct ib_device *dev = srq->device; + + if (!dev->ops.fill_res_srq_entry_raw) + return -EINVAL; + return dev->ops.fill_res_srq_entry_raw(msg, srq); +} + static int fill_stat_counter_mode(struct sk_buff *msg, struct rdma_counter *counter) { @@ -892,6 +947,8 @@ static int fill_stat_counter_qps(struct sk_buff *msg, int ret = 0; table_attr = nla_nest_start(msg, RDMA_NLDEV_ATTR_RES_QP); + if (!table_attr) + return -EMSGSIZE; rt = &counter->device->res[RDMA_RESTRACK_QP]; xa_lock(&rt->xa); @@ -968,14 +1025,21 @@ static int fill_stat_counter_hwcounters(struct sk_buff *msg, if (!table_attr) return -EMSGSIZE; - for (i = 0; i < st->num_counters; i++) - if (rdma_nl_stat_hwcounter_entry(msg, st->names[i], st->value[i])) + mutex_lock(&st->lock); + for (i = 0; i < st->num_counters; i++) { + if (test_bit(i, st->is_disabled)) + continue; + if (rdma_nl_stat_hwcounter_entry(msg, st->descs[i].name, + st->value[i])) goto err; + } + mutex_unlock(&st->lock); nla_nest_end(msg, table_attr); return 0; err: + mutex_unlock(&st->lock); nla_nest_cancel(msg, table_attr); return -EMSGSIZE; } @@ -1012,8 +1076,8 @@ static int nldev_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, u32 index; int err; - err = nlmsg_parse_deprecated(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + err = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, extack); if (err || !tb[RDMA_NLDEV_ATTR_DEV_INDEX]) return -EINVAL; @@ -1032,6 +1096,10 @@ static int nldev_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET), 0, 0); + if (!nlh) { + err = -EMSGSIZE; + goto err_free; + } err = fill_dev_info(msg, device); if (err) @@ -1057,8 +1125,8 @@ static int nldev_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh, u32 index; int err; - err = nlmsg_parse_deprecated(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, extack); if (err || !tb[RDMA_NLDEV_ATTR_DEV_INDEX]) return -EINVAL; @@ -1117,7 +1185,7 @@ static int _nldev_get_dumpit(struct ib_device *device, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET), 0, NLM_F_MULTI); - if (fill_dev_info(skb, device)) { + if (!nlh || fill_dev_info(skb, device)) { nlmsg_cancel(skb, nlh); goto out; } @@ -1149,8 +1217,8 @@ static int nldev_port_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, u32 port; int err; - err = nlmsg_parse_deprecated(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + err = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, extack); if (err || !tb[RDMA_NLDEV_ATTR_DEV_INDEX] || !tb[RDMA_NLDEV_ATTR_PORT_INDEX]) @@ -1176,6 +1244,10 @@ static int nldev_port_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET), 0, 0); + if (!nlh) { + err = -EMSGSIZE; + goto err_free; + } err = fill_port_info(msg, device, port, sock_net(skb->sk)); if (err) @@ -1205,8 +1277,8 @@ static int nldev_port_get_dumpit(struct sk_buff *skb, int err; unsigned int p; - err = nlmsg_parse_deprecated(cb->nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, NULL); + err = __nlmsg_parse(cb->nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, NULL); if (err || !tb[RDMA_NLDEV_ATTR_DEV_INDEX]) return -EINVAL; @@ -1237,7 +1309,7 @@ static int nldev_port_get_dumpit(struct sk_buff *skb, RDMA_NLDEV_CMD_PORT_GET), 0, NLM_F_MULTI); - if (fill_port_info(skb, device, p, sock_net(skb->sk))) { + if (!nlh || fill_port_info(skb, device, p, sock_net(skb->sk))) { nlmsg_cancel(skb, nlh); goto out; } @@ -1255,13 +1327,14 @@ static int nldev_res_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, struct netlink_ext_ack *extack) { struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; + bool show_details = false; struct ib_device *device; struct sk_buff *msg; u32 index; int ret; - ret = nlmsg_parse_deprecated(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + ret = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, extack); if (ret || !tb[RDMA_NLDEV_ATTR_DEV_INDEX]) return -EINVAL; @@ -1270,6 +1343,9 @@ static int nldev_res_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, if (!device) return -EINVAL; + if (tb[RDMA_NLDEV_ATTR_DRIVER_DETAILS]) + show_details = nla_get_u8(tb[RDMA_NLDEV_ATTR_DRIVER_DETAILS]); + msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL); if (!msg) { ret = -ENOMEM; @@ -1279,8 +1355,12 @@ static int nldev_res_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_RES_GET), 0, 0); + if (!nlh) { + ret = -EMSGSIZE; + goto err_free; + } - ret = fill_res_info(msg, device); + ret = fill_res_info(msg, device, show_details); if (ret) goto err_free; @@ -1310,7 +1390,7 @@ static int _nldev_res_get_dumpit(struct ib_device *device, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_RES_GET), 0, NLM_F_MULTI); - if (fill_res_info(skb, device)) { + if (!nlh || fill_res_info(skb, device, false)) { nlmsg_cancel(skb, nlh); goto out; } @@ -1389,10 +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]; @@ -1403,8 +1484,8 @@ static int res_get_common_doit(struct sk_buff *skb, struct nlmsghdr *nlh, struct sk_buff *msg; int ret; - ret = nlmsg_parse_deprecated(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + ret = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, extack); if (ret || !tb[RDMA_NLDEV_ATTR_DEV_INDEX] || !fe->id || !tb[fe->id]) return -EINVAL; @@ -1445,7 +1526,7 @@ static int res_get_common_doit(struct sk_buff *skb, struct nlmsghdr *nlh, RDMA_NL_GET_OP(nlh->nlmsg_type)), 0, 0); - if (fill_nldev_handle(msg, device)) { + if (!nlh || fill_nldev_handle(msg, device)) { ret = -EMSGSIZE; goto err_free; } @@ -1480,6 +1561,7 @@ static int res_get_common_dumpit(struct sk_buff *skb, struct rdma_restrack_entry *res; struct rdma_restrack_root *rt; int err, ret = 0, idx = 0; + bool show_details = false; struct nlattr *table_attr; struct nlattr *entry_attr; struct ib_device *device; @@ -1490,8 +1572,8 @@ static int res_get_common_dumpit(struct sk_buff *skb, u32 index, port = 0; bool filled = false; - err = nlmsg_parse_deprecated(cb->nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, NULL); + err = __nlmsg_parse(cb->nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, NULL); /* * Right now, we are expecting the device index to get res information, * but it is possible to extend this code to return all devices in @@ -1508,6 +1590,9 @@ static int res_get_common_dumpit(struct sk_buff *skb, if (!device) return -EINVAL; + if (tb[RDMA_NLDEV_ATTR_DRIVER_DETAILS]) + show_details = nla_get_u8(tb[RDMA_NLDEV_ATTR_DRIVER_DETAILS]); + /* * If no PORT_INDEX is supplied, we will return all QPs from that device */ @@ -1524,7 +1609,7 @@ static int res_get_common_dumpit(struct sk_buff *skb, RDMA_NL_GET_OP(cb->nlh->nlmsg_type)), 0, NLM_F_MULTI); - if (fill_nldev_handle(skb, device)) { + if (!nlh || fill_nldev_handle(skb, device)) { ret = -EMSGSIZE; goto err; } @@ -1545,6 +1630,9 @@ static int res_get_common_dumpit(struct sk_buff *skb, * objects. */ xa_for_each(&rt->xa, id, res) { + if (xa_get_mark(&rt->xa, res->id, RESTRACK_DD) && !show_details) + goto next; + if (idx < start || !rdma_restrack_get(res)) goto next; @@ -1629,6 +1717,7 @@ RES_GET_FUNCS(mr_raw, RDMA_RESTRACK_MR); RES_GET_FUNCS(counter, RDMA_RESTRACK_COUNTER); RES_GET_FUNCS(ctx, RDMA_RESTRACK_CTX); RES_GET_FUNCS(srq, RDMA_RESTRACK_SRQ); +RES_GET_FUNCS(srq_raw, RDMA_RESTRACK_SRQ); static LIST_HEAD(link_ops); static DECLARE_RWSEM(link_ops_rwsem); @@ -1676,8 +1765,8 @@ static int nldev_newlink(struct sk_buff *skb, struct nlmsghdr *nlh, char type[IFNAMSIZ]; int err; - err = nlmsg_parse_deprecated(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, extack); if (err || !tb[RDMA_NLDEV_ATTR_DEV_NAME] || !tb[RDMA_NLDEV_ATTR_LINK_TYPE] || !tb[RDMA_NLDEV_ATTR_NDEV_NAME]) return -EINVAL; @@ -1720,8 +1809,8 @@ static int nldev_dellink(struct sk_buff *skb, struct nlmsghdr *nlh, u32 index; int err; - err = nlmsg_parse_deprecated(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, extack); if (err || !tb[RDMA_NLDEV_ATTR_DEV_INDEX]) return -EINVAL; @@ -1730,7 +1819,7 @@ static int nldev_dellink(struct sk_buff *skb, struct nlmsghdr *nlh, if (!device) return -EINVAL; - if (!(device->attrs.device_cap_flags & IB_DEVICE_ALLOW_USER_UNREG)) { + if (!(device->attrs.kernel_cap_flags & IBK_ALLOW_USER_UNREG)) { ib_device_put(device); return -EINVAL; } @@ -1750,8 +1839,8 @@ static int nldev_get_chardev(struct sk_buff *skb, struct nlmsghdr *nlh, u32 index; int err; - err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, nldev_policy, - extack); + err = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, nldev_policy, + NL_VALIDATE_LIBERAL, extack); if (err || !tb[RDMA_NLDEV_ATTR_CHARDEV_TYPE]) return -EINVAL; @@ -1786,6 +1875,10 @@ static int nldev_get_chardev(struct sk_buff *skb, struct nlmsghdr *nlh, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET_CHARDEV), 0, 0); + if (!nlh) { + err = -EMSGSIZE; + goto out_nlmsg; + } data.nl_msg = msg; err = ib_get_client_nl_info(ibdev, client_name, &data); @@ -1830,8 +1923,8 @@ static int nldev_sys_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, struct sk_buff *msg; int err; - err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + err = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, extack); if (err) return err; @@ -1843,6 +1936,10 @@ static int nldev_sys_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_SYS_GET), 0, 0); + if (!nlh) { + nlmsg_free(msg); + return -EMSGSIZE; + } err = nla_put_u8(msg, RDMA_NLDEV_SYS_ATTR_NETNS_MODE, (u8)ib_devices_shared_netns); @@ -1851,6 +1948,18 @@ static int nldev_sys_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, return err; } + err = nla_put_u8(msg, RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE, + (u8)privileged_qkey); + if (err) { + nlmsg_free(msg); + return err; + } + + err = nla_put_u8(msg, RDMA_NLDEV_SYS_ATTR_MONITOR_MODE, 1); + if (err) { + nlmsg_free(msg); + return err; + } /* * Copy-on-fork is supported. * See commits: @@ -1867,45 +1976,167 @@ static int nldev_sys_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid); } +static int nldev_set_sys_set_netns_doit(struct nlattr *tb[]) +{ + u8 enable; + int err; + + enable = nla_get_u8(tb[RDMA_NLDEV_SYS_ATTR_NETNS_MODE]); + /* Only 0 and 1 are supported */ + if (enable > 1) + return -EINVAL; + + err = rdma_compatdev_set(enable); + return err; +} + +static int nldev_set_sys_set_pqkey_doit(struct nlattr *tb[]) +{ + u8 enable; + + enable = nla_get_u8(tb[RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE]); + /* Only 0 and 1 are supported */ + if (enable > 1) + return -EINVAL; + + privileged_qkey = enable; + return 0; +} + static int nldev_set_sys_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh, struct netlink_ext_ack *extack) { struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; - u8 enable; int err; err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, nldev_policy, extack); - if (err || !tb[RDMA_NLDEV_SYS_ATTR_NETNS_MODE]) + if (err) return -EINVAL; - enable = nla_get_u8(tb[RDMA_NLDEV_SYS_ATTR_NETNS_MODE]); - /* Only 0 and 1 are supported */ - if (enable > 1) + if (tb[RDMA_NLDEV_SYS_ATTR_NETNS_MODE]) + return nldev_set_sys_set_netns_doit(tb); + + if (tb[RDMA_NLDEV_SYS_ATTR_PRIVILEGED_QKEY_MODE]) + return nldev_set_sys_set_pqkey_doit(tb); + + return -EINVAL; +} + + +static int nldev_stat_set_mode_doit(struct sk_buff *msg, + struct netlink_ext_ack *extack, + struct nlattr *tb[], + 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 */ + if (!tb[RDMA_NLDEV_ATTR_STAT_RES] || + nla_get_u32(tb[RDMA_NLDEV_ATTR_STAT_RES]) != RDMA_NLDEV_ATTR_RES_QP) return -EINVAL; - err = rdma_compatdev_set(enable); - return err; + 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, opcnt, + extack); + } + + if (!tb[RDMA_NLDEV_ATTR_RES_LQPN]) + return -EINVAL; + + qpn = nla_get_u32(tb[RDMA_NLDEV_ATTR_RES_LQPN]); + if (tb[RDMA_NLDEV_ATTR_STAT_COUNTER_ID]) { + cntn = nla_get_u32(tb[RDMA_NLDEV_ATTR_STAT_COUNTER_ID]); + ret = rdma_counter_bind_qpn(device, port, qpn, cntn); + if (ret) + return ret; + } else { + ret = rdma_counter_bind_qpn_alloc(device, port, qpn, &cntn); + if (ret) + return ret; + } + + if (nla_put_u32(msg, RDMA_NLDEV_ATTR_STAT_COUNTER_ID, cntn) || + nla_put_u32(msg, RDMA_NLDEV_ATTR_RES_LQPN, qpn)) { + ret = -EMSGSIZE; + goto err_fill; + } + + return 0; + +err_fill: + rdma_counter_unbind_qpn(device, port, qpn, cntn); + return ret; +} + +static int nldev_stat_set_counter_dynamic_doit(struct nlattr *tb[], + struct ib_device *device, + u32 port) +{ + struct rdma_hw_stats *stats; + struct nlattr *entry_attr; + unsigned long *target; + int rem, i, ret = 0; + u32 index; + + stats = ib_get_hw_stats_port(device, port); + if (!stats) + return -EINVAL; + + target = kcalloc(BITS_TO_LONGS(stats->num_counters), + sizeof(*stats->is_disabled), GFP_KERNEL); + if (!target) + return -ENOMEM; + + nla_for_each_nested(entry_attr, tb[RDMA_NLDEV_ATTR_STAT_HWCOUNTERS], + rem) { + index = nla_get_u32(entry_attr); + if ((index >= stats->num_counters) || + !(stats->descs[index].flags & IB_STAT_FLAG_OPTIONAL)) { + ret = -EINVAL; + goto out; + } + + set_bit(index, target); + } + + for (i = 0; i < stats->num_counters; i++) { + if (!(stats->descs[i].flags & IB_STAT_FLAG_OPTIONAL)) + continue; + + ret = rdma_counter_modify(device, port, i, test_bit(i, target)); + if (ret) + goto out; + } + +out: + kfree(target); + return ret; } static int nldev_stat_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh, struct netlink_ext_ack *extack) { - u32 index, port, mode, mask = 0, qpn, cntn = 0; struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; struct ib_device *device; struct sk_buff *msg; + u32 index, port; int ret; - ret = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); - /* Currently only counter for QP is supported */ - if (ret || !tb[RDMA_NLDEV_ATTR_STAT_RES] || - !tb[RDMA_NLDEV_ATTR_DEV_INDEX] || - !tb[RDMA_NLDEV_ATTR_PORT_INDEX] || !tb[RDMA_NLDEV_ATTR_STAT_MODE]) - return -EINVAL; - - if (nla_get_u32(tb[RDMA_NLDEV_ATTR_STAT_RES]) != RDMA_NLDEV_ATTR_RES_QP) + ret = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, nldev_policy, + extack); + if (ret || !tb[RDMA_NLDEV_ATTR_DEV_INDEX] || + !tb[RDMA_NLDEV_ATTR_PORT_INDEX]) return -EINVAL; index = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]); @@ -1916,59 +2147,49 @@ static int nldev_stat_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh, port = nla_get_u32(tb[RDMA_NLDEV_ATTR_PORT_INDEX]); if (!rdma_is_port_valid(device, port)) { ret = -EINVAL; - goto err; + goto err_put_device; + } + + if (!tb[RDMA_NLDEV_ATTR_STAT_MODE] && + !tb[RDMA_NLDEV_ATTR_STAT_HWCOUNTERS]) { + ret = -EINVAL; + goto err_put_device; } msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL); if (!msg) { ret = -ENOMEM; - goto err; + goto err_put_device; } nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_STAT_SET), 0, 0); + if (!nlh || fill_nldev_handle(msg, device) || + nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, port)) { + ret = -EMSGSIZE; + goto err_free_msg; + } - 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]); - ret = rdma_counter_set_auto_mode(device, port, mask, extack); - if (ret) - goto err_msg; - } else { - if (!tb[RDMA_NLDEV_ATTR_RES_LQPN]) - goto err_msg; - qpn = nla_get_u32(tb[RDMA_NLDEV_ATTR_RES_LQPN]); - if (tb[RDMA_NLDEV_ATTR_STAT_COUNTER_ID]) { - cntn = nla_get_u32(tb[RDMA_NLDEV_ATTR_STAT_COUNTER_ID]); - ret = rdma_counter_bind_qpn(device, port, qpn, cntn); - } else { - ret = rdma_counter_bind_qpn_alloc(device, port, - qpn, &cntn); - } + if (tb[RDMA_NLDEV_ATTR_STAT_MODE]) { + ret = nldev_stat_set_mode_doit(msg, extack, tb, device, port); if (ret) - goto err_msg; + goto err_free_msg; + } - if (fill_nldev_handle(msg, device) || - nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, port) || - nla_put_u32(msg, RDMA_NLDEV_ATTR_STAT_COUNTER_ID, cntn) || - nla_put_u32(msg, RDMA_NLDEV_ATTR_RES_LQPN, qpn)) { - ret = -EMSGSIZE; - goto err_fill; - } + if (tb[RDMA_NLDEV_ATTR_STAT_HWCOUNTERS]) { + ret = nldev_stat_set_counter_dynamic_doit(tb, device, port); + if (ret) + goto err_free_msg; } nlmsg_end(msg, nlh); ib_device_put(device); return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid); -err_fill: - rdma_counter_unbind_qpn(device, port, qpn, cntn); -err_msg: +err_free_msg: nlmsg_free(msg); -err: +err_put_device: ib_device_put(device); return ret; } @@ -2013,6 +2234,10 @@ static int nldev_stat_del_doit(struct sk_buff *skb, struct nlmsghdr *nlh, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_STAT_SET), 0, 0); + if (!nlh) { + ret = -EMSGSIZE; + goto err_fill; + } cntn = nla_get_u32(tb[RDMA_NLDEV_ATTR_STAT_COUNTER_ID]); qpn = nla_get_u32(tb[RDMA_NLDEV_ATTR_RES_LQPN]); @@ -2039,10 +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; @@ -2083,7 +2308,7 @@ static int stat_get_doit_default_counter(struct sk_buff *skb, RDMA_NLDEV_CMD_STAT_GET), 0, 0); - if (fill_nldev_handle(msg, device) || + if (!nlh || fill_nldev_handle(msg, device) || nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, port)) { ret = -EMSGSIZE; goto err_msg; @@ -2103,9 +2328,13 @@ static int stat_get_doit_default_counter(struct sk_buff *skb, goto err_stats; } for (i = 0; i < num_cnts; i++) { + if (test_bit(i, stats->is_disabled)) + continue; + v = stats->value[i] + rdma_counter_get_hwstat_value(device, port, i); - if (rdma_nl_stat_hwcounter_entry(msg, stats->names[i], v)) { + if (rdma_nl_stat_hwcounter_entry(msg, + stats->descs[i].name, v)) { ret = -EMSGSIZE; goto err_table; } @@ -2128,8 +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; @@ -2137,6 +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]) @@ -2167,8 +2398,12 @@ static int stat_get_doit_qp(struct sk_buff *skb, struct nlmsghdr *nlh, RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_STAT_GET), 0, 0); + if (!nlh) { + ret = -EMSGSIZE; + 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; @@ -2185,6 +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); @@ -2202,8 +2443,8 @@ static int nldev_stat_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh, struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; int ret; - ret = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, extack); + ret = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, extack); if (ret) return -EINVAL; @@ -2232,8 +2473,8 @@ static int nldev_stat_get_dumpit(struct sk_buff *skb, struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; int ret; - ret = nlmsg_parse(cb->nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, - nldev_policy, NULL); + ret = __nlmsg_parse(cb->nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, NULL); if (ret || !tb[RDMA_NLDEV_ATTR_STAT_RES]) return -EINVAL; @@ -2253,6 +2494,149 @@ static int nldev_stat_get_dumpit(struct sk_buff *skb, return ret; } +static int nldev_stat_get_counter_status_doit(struct sk_buff *skb, + struct nlmsghdr *nlh, + struct netlink_ext_ack *extack) +{ + struct nlattr *tb[RDMA_NLDEV_ATTR_MAX], *table, *entry; + struct rdma_hw_stats *stats; + struct ib_device *device; + struct sk_buff *msg; + u32 devid, port; + int ret, i; + + ret = __nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, NL_VALIDATE_LIBERAL, extack); + if (ret || !tb[RDMA_NLDEV_ATTR_DEV_INDEX] || + !tb[RDMA_NLDEV_ATTR_PORT_INDEX]) + return -EINVAL; + + devid = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]); + device = ib_device_get_by_index(sock_net(skb->sk), devid); + if (!device) + return -EINVAL; + + port = nla_get_u32(tb[RDMA_NLDEV_ATTR_PORT_INDEX]); + if (!rdma_is_port_valid(device, port)) { + ret = -EINVAL; + goto err; + } + + stats = ib_get_hw_stats_port(device, port); + if (!stats) { + ret = -EINVAL; + goto err; + } + + msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL); + if (!msg) { + ret = -ENOMEM; + goto err; + } + + nlh = nlmsg_put( + msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq, + RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_STAT_GET_STATUS), + 0, 0); + + ret = -EMSGSIZE; + if (!nlh || fill_nldev_handle(msg, device) || + nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, port)) + goto err_msg; + + table = nla_nest_start(msg, RDMA_NLDEV_ATTR_STAT_HWCOUNTERS); + if (!table) + goto err_msg; + + mutex_lock(&stats->lock); + for (i = 0; i < stats->num_counters; i++) { + entry = nla_nest_start(msg, + RDMA_NLDEV_ATTR_STAT_HWCOUNTER_ENTRY); + if (!entry) + goto err_msg_table; + + if (nla_put_string(msg, + RDMA_NLDEV_ATTR_STAT_HWCOUNTER_ENTRY_NAME, + stats->descs[i].name) || + nla_put_u32(msg, RDMA_NLDEV_ATTR_STAT_HWCOUNTER_INDEX, i)) + goto err_msg_entry; + + if ((stats->descs[i].flags & IB_STAT_FLAG_OPTIONAL) && + (nla_put_u8(msg, RDMA_NLDEV_ATTR_STAT_HWCOUNTER_DYNAMIC, + !test_bit(i, stats->is_disabled)))) + goto err_msg_entry; + + nla_nest_end(msg, entry); + } + mutex_unlock(&stats->lock); + + nla_nest_end(msg, table); + nlmsg_end(msg, nlh); + ib_device_put(device); + return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid); + +err_msg_entry: + nla_nest_cancel(msg, entry); +err_msg_table: + mutex_unlock(&stats->lock); + nla_nest_cancel(msg, table); +err_msg: + nlmsg_free(msg); +err: + ib_device_put(device); + return ret; +} + +static int nldev_newdev(struct sk_buff *skb, struct nlmsghdr *nlh, + struct netlink_ext_ack *extack) +{ + struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; + enum rdma_nl_dev_type type; + struct ib_device *parent; + char name[IFNAMSIZ] = {}; + u32 parentid; + int ret; + + ret = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, extack); + if (ret || !tb[RDMA_NLDEV_ATTR_DEV_INDEX] || + !tb[RDMA_NLDEV_ATTR_DEV_NAME] || !tb[RDMA_NLDEV_ATTR_DEV_TYPE]) + return -EINVAL; + + nla_strscpy(name, tb[RDMA_NLDEV_ATTR_DEV_NAME], sizeof(name)); + type = nla_get_u8(tb[RDMA_NLDEV_ATTR_DEV_TYPE]); + parentid = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]); + parent = ib_device_get_by_index(sock_net(skb->sk), parentid); + if (!parent) + return -EINVAL; + + ret = ib_add_sub_device(parent, type, name); + ib_device_put(parent); + + return ret; +} + +static int nldev_deldev(struct sk_buff *skb, struct nlmsghdr *nlh, + struct netlink_ext_ack *extack) +{ + struct nlattr *tb[RDMA_NLDEV_ATTR_MAX]; + struct ib_device *device; + u32 devid; + int ret; + + ret = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, + nldev_policy, extack); + if (ret || !tb[RDMA_NLDEV_ATTR_DEV_INDEX]) + return -EINVAL; + + devid = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]); + device = ib_device_get_by_index(sock_net(skb->sk), devid); + if (!device) + return -EINVAL; + + return ib_del_sub_device_and_put(device); +} + static const struct rdma_nl_cbs nldev_cb_table[RDMA_NLDEV_NUM_OPS] = { [RDMA_NLDEV_CMD_GET] = { .doit = nldev_get_doit, @@ -2314,6 +2698,7 @@ static const struct rdma_nl_cbs nldev_cb_table[RDMA_NLDEV_NUM_OPS] = { }, [RDMA_NLDEV_CMD_SYS_SET] = { .doit = nldev_set_sys_set_doit, + .flags = RDMA_NL_ADMIN_PERM, }, [RDMA_NLDEV_CMD_STAT_SET] = { .doit = nldev_stat_set_doit, @@ -2342,14 +2727,192 @@ static const struct rdma_nl_cbs nldev_cb_table[RDMA_NLDEV_NUM_OPS] = { .dump = nldev_res_get_mr_raw_dumpit, .flags = RDMA_NL_ADMIN_PERM, }, + [RDMA_NLDEV_CMD_RES_SRQ_GET_RAW] = { + .doit = nldev_res_get_srq_raw_doit, + .dump = nldev_res_get_srq_raw_dumpit, + .flags = RDMA_NL_ADMIN_PERM, + }, + [RDMA_NLDEV_CMD_STAT_GET_STATUS] = { + .doit = nldev_stat_get_counter_status_doit, + }, + [RDMA_NLDEV_CMD_NEWDEV] = { + .doit = nldev_newdev, + .flags = RDMA_NL_ADMIN_PERM, + }, + [RDMA_NLDEV_CMD_DELDEV] = { + .doit = nldev_deldev, + .flags = RDMA_NL_ADMIN_PERM, + }, }; +static int fill_mon_netdev_rename(struct sk_buff *msg, + struct ib_device *device, u32 port, + const struct net *net) +{ + struct net_device *netdev = ib_device_get_netdev(device, port); + int ret = 0; + + if (!netdev || !net_eq(dev_net(netdev), net)) + goto out; + + ret = nla_put_u32(msg, RDMA_NLDEV_ATTR_NDEV_INDEX, netdev->ifindex); + if (ret) + goto out; + ret = nla_put_string(msg, RDMA_NLDEV_ATTR_NDEV_NAME, netdev->name); +out: + dev_put(netdev); + return ret; +} + +static int fill_mon_netdev_association(struct sk_buff *msg, + struct ib_device *device, u32 port, + const struct net *net) +{ + struct net_device *netdev = ib_device_get_netdev(device, port); + int ret = 0; + + if (netdev && !net_eq(dev_net(netdev), net)) + goto out; + + ret = nla_put_u32(msg, RDMA_NLDEV_ATTR_DEV_INDEX, device->index); + if (ret) + goto out; + + ret = nla_put_string(msg, RDMA_NLDEV_ATTR_DEV_NAME, + dev_name(&device->dev)); + if (ret) + goto out; + + ret = nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, port); + if (ret) + goto out; + + if (netdev) { + ret = nla_put_u32(msg, + RDMA_NLDEV_ATTR_NDEV_INDEX, netdev->ifindex); + if (ret) + goto out; + + ret = nla_put_string(msg, + RDMA_NLDEV_ATTR_NDEV_NAME, netdev->name); + } + +out: + dev_put(netdev); + return ret; +} + +static void rdma_nl_notify_err_msg(struct ib_device *device, u32 port_num, + enum rdma_nl_notify_event_type type) +{ + struct net_device *netdev; + + switch (type) { + case RDMA_REGISTER_EVENT: + dev_warn_ratelimited(&device->dev, + "Failed to send RDMA monitor register device event\n"); + break; + case RDMA_UNREGISTER_EVENT: + dev_warn_ratelimited(&device->dev, + "Failed to send RDMA monitor unregister device event\n"); + break; + case RDMA_NETDEV_ATTACH_EVENT: + netdev = ib_device_get_netdev(device, port_num); + dev_warn_ratelimited(&device->dev, + "Failed to send RDMA monitor netdev attach event: port %d netdev %d\n", + port_num, netdev->ifindex); + dev_put(netdev); + break; + case RDMA_NETDEV_DETACH_EVENT: + dev_warn_ratelimited(&device->dev, + "Failed to send RDMA monitor netdev detach event: port %d\n", + port_num); + break; + case RDMA_RENAME_EVENT: + dev_warn_ratelimited(&device->dev, + "Failed to send RDMA monitor rename device event\n"); + break; + + case RDMA_NETDEV_RENAME_EVENT: + netdev = ib_device_get_netdev(device, port_num); + dev_warn_ratelimited(&device->dev, + "Failed to send RDMA monitor netdev rename event: port %d netdev %d\n", + port_num, netdev->ifindex); + dev_put(netdev); + break; + default: + break; + } +} + +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; + void *nlh; + + net = read_pnet(&device->coredev.rdma_net); + if (!net) + return -EINVAL; + + skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL); + if (!skb) + return -ENOMEM; + nlh = nlmsg_put(skb, 0, 0, + RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_MONITOR), + 0, 0); + if (!nlh) + goto err_free; + + switch (type) { + case RDMA_REGISTER_EVENT: + case RDMA_UNREGISTER_EVENT: + case RDMA_RENAME_EVENT: + ret = fill_nldev_handle(skb, device); + if (ret) + goto err_free; + break; + case RDMA_NETDEV_ATTACH_EVENT: + case RDMA_NETDEV_DETACH_EVENT: + ret = fill_mon_netdev_association(skb, device, port_num, net); + if (ret) + goto err_free; + break; + case RDMA_NETDEV_RENAME_EVENT: + ret = fill_mon_netdev_rename(skb, device, port_num, net); + if (ret) + goto err_free; + break; + default: + break; + } + + ret = nla_put_u8(skb, RDMA_NLDEV_ATTR_EVENT_TYPE, type); + if (ret) + goto err_free; + + nlmsg_end(skb, nlh); + ret = rdma_nl_multicast(net, skb, RDMA_NL_GROUP_NOTIFY, GFP_KERNEL); + if (ret && ret != -ESRCH) { + skb = NULL; /* skb is freed in the netlink send-op handling */ + goto err_free; + } + return 0; + +err_free: + rdma_nl_notify_err_msg(device, port_num, type); + nlmsg_free(skb); + return ret; +} + void __init nldev_init(void) { rdma_nl_register(RDMA_NL_NLDEV, nldev_cb_table); } -void __exit nldev_exit(void) +void nldev_exit(void) { rdma_nl_unregister(RDMA_NL_NLDEV); } diff --git a/drivers/infiniband/core/rdma_core.c b/drivers/infiniband/core/rdma_core.c index 94d83b665a2f..18918f463361 100644 --- a/drivers/infiniband/core/rdma_core.c +++ b/drivers/infiniband/core/rdma_core.c @@ -58,8 +58,8 @@ void uverbs_uobject_put(struct ib_uobject *uobject) } EXPORT_SYMBOL(uverbs_uobject_put); -static int uverbs_try_lock_object(struct ib_uobject *uobj, - enum rdma_lookup_mode mode) +int uverbs_try_lock_object(struct ib_uobject *uobj, + enum rdma_lookup_mode mode) { /* * When a shared access is required, we use a positive counter. Each @@ -68,7 +68,7 @@ static int uverbs_try_lock_object(struct ib_uobject *uobj, * In exclusive access mode, we check that the counter is zero (nobody * claimed this object) and we set it to -1. Releasing a shared access * lock is done simply by decreasing the counter. As for exclusive - * access locks, since only a single one of them is is allowed + * access locks, since only a single one of them is allowed * concurrently, setting the counter to zero is enough for releasing * this lock. */ @@ -84,6 +84,7 @@ static int uverbs_try_lock_object(struct ib_uobject *uobj, } return 0; } +EXPORT_SYMBOL(uverbs_try_lock_object); static void assert_uverbs_usecnt(struct ib_uobject *uobj, enum rdma_lookup_mode mode) @@ -880,9 +881,14 @@ static void ufile_destroy_ucontext(struct ib_uverbs_file *ufile, static int __uverbs_cleanup_ufile(struct ib_uverbs_file *ufile, enum rdma_remove_reason reason) { + struct uverbs_attr_bundle attrs = { .ufile = ufile }; + struct ib_ucontext *ucontext = ufile->ucontext; + struct ib_device *ib_dev = ucontext->device; struct ib_uobject *obj, *next_obj; int ret = -EINVAL; - struct uverbs_attr_bundle attrs = { .ufile = ufile }; + + if (ib_dev->ops.ufile_hw_cleanup) + ib_dev->ops.ufile_hw_cleanup(ufile); /* * This shouldn't run while executing other commands on this @@ -1013,3 +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 033207882c82..b097cfcade1c 100644 --- a/drivers/infiniband/core/restrack.c +++ b/drivers/infiniband/core/restrack.c @@ -37,22 +37,6 @@ int rdma_restrack_init(struct ib_device *dev) return 0; } -static const char *type2str(enum rdma_restrack_type type) -{ - static const char * const names[RDMA_RESTRACK_MAX] = { - [RDMA_RESTRACK_PD] = "PD", - [RDMA_RESTRACK_CQ] = "CQ", - [RDMA_RESTRACK_QP] = "QP", - [RDMA_RESTRACK_CM_ID] = "CM_ID", - [RDMA_RESTRACK_MR] = "MR", - [RDMA_RESTRACK_CTX] = "CTX", - [RDMA_RESTRACK_COUNTER] = "COUNTER", - [RDMA_RESTRACK_SRQ] = "SRQ", - }; - - return names[type]; -}; - /** * rdma_restrack_clean() - clean resource tracking * @dev: IB device @@ -60,47 +44,14 @@ static const char *type2str(enum rdma_restrack_type type) void rdma_restrack_clean(struct ib_device *dev) { struct rdma_restrack_root *rt = dev->res; - struct rdma_restrack_entry *e; - char buf[TASK_COMM_LEN]; - bool found = false; - const char *owner; int i; for (i = 0 ; i < RDMA_RESTRACK_MAX; i++) { struct xarray *xa = &dev->res[i].xa; - if (!xa_empty(xa)) { - unsigned long index; - - if (!found) { - pr_err("restrack: %s", CUT_HERE); - dev_err(&dev->dev, "BUG: RESTRACK detected leak of resources\n"); - } - xa_for_each(xa, index, e) { - if (rdma_is_kernel_res(e)) { - owner = e->kern_name; - } else { - /* - * There is no need to call get_task_struct here, - * because we can be here only if there are more - * get_task_struct() call than put_task_struct(). - */ - get_task_comm(buf, e->task); - owner = buf; - } - - pr_err("restrack: %s %s object allocated by %s is not freed\n", - rdma_is_kernel_res(e) ? "Kernel" : - "User", - type2str(e->type), owner); - } - found = true; - } + WARN_ON(!xa_empty(xa)); xa_destroy(xa); } - if (found) - pr_err("restrack: %s", CUT_HERE); - kfree(rt); } @@ -108,8 +59,10 @@ void rdma_restrack_clean(struct ib_device *dev) * rdma_restrack_count() - the current usage of specific object * @dev: IB device * @type: actual type of object to operate + * @show_details: count driver specific objects */ -int rdma_restrack_count(struct ib_device *dev, enum rdma_restrack_type type) +int rdma_restrack_count(struct ib_device *dev, enum rdma_restrack_type type, + bool show_details) { struct rdma_restrack_root *rt = &dev->res[type]; struct rdma_restrack_entry *e; @@ -117,8 +70,11 @@ int rdma_restrack_count(struct ib_device *dev, enum rdma_restrack_type type) u32 cnt = 0; xa_lock(&rt->xa); - xas_for_each(&xas, e, U32_MAX) + xas_for_each(&xas, e, U32_MAX) { + if (xa_get_mark(&rt->xa, e->id, RESTRACK_DD) && !show_details) + continue; cnt++; + } xa_unlock(&rt->xa); return cnt; } @@ -144,6 +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; @@ -217,7 +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) @@ -247,6 +205,9 @@ void rdma_restrack_add(struct rdma_restrack_entry *res) ret = xa_insert(&rt->xa, res->id, res, GFP_KERNEL); if (ret) res->id = 0; + + if (qp->qp_type >= IB_QPT_DRIVER) + xa_set_mark(&rt->xa, res->id, RESTRACK_DD); } else if (res->type == RDMA_RESTRACK_COUNTER) { /* Special case to ensure that cntn points to right counter */ struct rdma_counter *counter; @@ -316,7 +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) @@ -343,8 +304,6 @@ void rdma_restrack_del(struct rdma_restrack_entry *res) rt = &dev->res[res->type]; old = xa_erase(&rt->xa, res->id); - if (res->type == RDMA_RESTRACK_MR || res->type == RDMA_RESTRACK_QP) - return; WARN_ON(old != res); out: diff --git a/drivers/infiniband/core/roce_gid_mgmt.c b/drivers/infiniband/core/roce_gid_mgmt.c index 68197e576433..a9f2c6b1b29e 100644 --- a/drivers/infiniband/core/roce_gid_mgmt.c +++ b/drivers/infiniband/core/roce_gid_mgmt.c @@ -250,7 +250,7 @@ static bool upper_device_filter(struct ib_device *ib_dev, u32 port, /** * is_upper_ndev_bond_master_filter - Check if a given netdevice - * is bond master device of netdevice of the the RDMA device of port. + * is bond master device of netdevice of the RDMA device of port. * @ib_dev: IB device to check * @port: Port to consider for adding default GID * @rdma_ndev: Pointer to rdma netdevice @@ -515,6 +515,27 @@ void rdma_roce_rescan_device(struct ib_device *ib_dev) } EXPORT_SYMBOL(rdma_roce_rescan_device); +/** + * rdma_roce_rescan_port - Rescan all of the network devices in the system + * and add their gids if relevant to the port of the RoCE device. + * + * @ib_dev: IB device + * @port: Port number + */ +void rdma_roce_rescan_port(struct ib_device *ib_dev, u32 port) +{ + struct net_device *ndev = NULL; + + if (rdma_protocol_roce(ib_dev, port)) { + ndev = ib_device_get_netdev(ib_dev, port); + if (!ndev) + return; + enum_all_gids_of_dev_cb(ib_dev, port, ndev, ndev); + dev_put(ndev); + } +} +EXPORT_SYMBOL(rdma_roce_rescan_port); + static void callback_for_addr_gid_device_scan(struct ib_device *device, u32 port, struct net_device *rdma_ndev, @@ -575,16 +596,17 @@ static void handle_netdev_upper(struct ib_device *ib_dev, u32 port, } } -static void _roce_del_all_netdev_gids(struct ib_device *ib_dev, u32 port, - struct net_device *event_ndev) +void roce_del_all_netdev_gids(struct ib_device *ib_dev, + u32 port, struct net_device *ndev) { - ib_cache_gid_del_all_netdev_gids(ib_dev, port, event_ndev); + ib_cache_gid_del_all_netdev_gids(ib_dev, port, ndev); } +EXPORT_SYMBOL(roce_del_all_netdev_gids); static void del_netdev_upper_ips(struct ib_device *ib_dev, u32 port, struct net_device *rdma_ndev, void *cookie) { - handle_netdev_upper(ib_dev, port, cookie, _roce_del_all_netdev_gids); + handle_netdev_upper(ib_dev, port, cookie, roce_del_all_netdev_gids); } static void add_netdev_upper_ips(struct ib_device *ib_dev, u32 port, @@ -601,8 +623,7 @@ static void del_netdev_default_ips_join(struct ib_device *ib_dev, u32 port, rcu_read_lock(); master_ndev = netdev_master_upper_dev_get_rcu(rdma_ndev); - if (master_ndev) - dev_hold(master_ndev); + dev_hold(master_ndev); rcu_read_unlock(); if (master_ndev) { diff --git a/drivers/infiniband/core/rw.c b/drivers/infiniband/core/rw.c index 5221cce65675..6354ddf2a274 100644 --- a/drivers/infiniband/core/rw.c +++ b/drivers/infiniband/core/rw.c @@ -2,6 +2,7 @@ /* * Copyright (c) 2016 HGST, a Western Digital Company. */ +#include <linux/memremap.h> #include <linux/moduleparam.h> #include <linux/slab.h> #include <linux/pci-p2pdma.h> @@ -273,26 +274,6 @@ static int rdma_rw_init_single_wr(struct rdma_rw_ctx *ctx, struct ib_qp *qp, return 1; } -static void rdma_rw_unmap_sg(struct ib_device *dev, struct scatterlist *sg, - u32 sg_cnt, enum dma_data_direction dir) -{ - if (is_pci_p2pdma_page(sg_page(sg))) - pci_p2pdma_unmap_sg(dev->dma_device, sg, sg_cnt, dir); - else - ib_dma_unmap_sg(dev, sg, sg_cnt, dir); -} - -static int rdma_rw_map_sg(struct ib_device *dev, struct scatterlist *sg, - u32 sg_cnt, enum dma_data_direction dir) -{ - if (is_pci_p2pdma_page(sg_page(sg))) { - if (WARN_ON_ONCE(ib_uses_virt_dma(dev))) - return 0; - return pci_p2pdma_map_sg(dev->dma_device, sg, sg_cnt, dir); - } - return ib_dma_map_sg(dev, sg, sg_cnt, dir); -} - /** * rdma_rw_ctx_init - initialize a RDMA READ/WRITE context * @ctx: context to initialize @@ -313,12 +294,16 @@ int rdma_rw_ctx_init(struct rdma_rw_ctx *ctx, struct ib_qp *qp, u32 port_num, u64 remote_addr, u32 rkey, enum dma_data_direction dir) { struct ib_device *dev = qp->pd->device; + struct sg_table sgt = { + .sgl = sg, + .orig_nents = sg_cnt, + }; int ret; - ret = rdma_rw_map_sg(dev, sg, sg_cnt, dir); - if (!ret) - return -ENOMEM; - sg_cnt = ret; + ret = ib_dma_map_sgtable_attrs(dev, &sgt, dir, 0); + if (ret) + return ret; + sg_cnt = sgt.nents; /* * Skip to the S/G entry that sg_offset falls into: @@ -354,7 +339,7 @@ int rdma_rw_ctx_init(struct rdma_rw_ctx *ctx, struct ib_qp *qp, u32 port_num, return ret; out_unmap_sg: - rdma_rw_unmap_sg(dev, sg, sg_cnt, dir); + ib_dma_unmap_sgtable_attrs(dev, &sgt, dir, 0); return ret; } EXPORT_SYMBOL(rdma_rw_ctx_init); @@ -385,6 +370,14 @@ int rdma_rw_ctx_signature_init(struct rdma_rw_ctx *ctx, struct ib_qp *qp, struct ib_device *dev = qp->pd->device; u32 pages_per_mr = rdma_rw_fr_page_list_len(qp->pd->device, qp->integrity_en); + struct sg_table sgt = { + .sgl = sg, + .orig_nents = sg_cnt, + }; + struct sg_table prot_sgt = { + .sgl = prot_sg, + .orig_nents = prot_sg_cnt, + }; struct ib_rdma_wr *rdma_wr; int count = 0, ret; @@ -394,18 +387,14 @@ int rdma_rw_ctx_signature_init(struct rdma_rw_ctx *ctx, struct ib_qp *qp, return -EINVAL; } - ret = rdma_rw_map_sg(dev, sg, sg_cnt, dir); - if (!ret) - return -ENOMEM; - sg_cnt = ret; + ret = ib_dma_map_sgtable_attrs(dev, &sgt, dir, 0); + if (ret) + return ret; if (prot_sg_cnt) { - ret = rdma_rw_map_sg(dev, prot_sg, prot_sg_cnt, dir); - if (!ret) { - ret = -ENOMEM; + ret = ib_dma_map_sgtable_attrs(dev, &prot_sgt, dir, 0); + if (ret) goto out_unmap_sg; - } - prot_sg_cnt = ret; } ctx->type = RDMA_RW_SIG_MR; @@ -426,10 +415,11 @@ int rdma_rw_ctx_signature_init(struct rdma_rw_ctx *ctx, struct ib_qp *qp, memcpy(ctx->reg->mr->sig_attrs, sig_attrs, sizeof(struct ib_sig_attrs)); - ret = ib_map_mr_sg_pi(ctx->reg->mr, sg, sg_cnt, NULL, prot_sg, - prot_sg_cnt, NULL, SZ_4K); + ret = ib_map_mr_sg_pi(ctx->reg->mr, sg, sgt.nents, NULL, prot_sg, + prot_sgt.nents, NULL, SZ_4K); if (unlikely(ret)) { - pr_err("failed to map PI sg (%u)\n", sg_cnt + prot_sg_cnt); + pr_err("failed to map PI sg (%u)\n", + sgt.nents + prot_sgt.nents); goto out_destroy_sig_mr; } @@ -468,10 +458,10 @@ out_destroy_sig_mr: out_free_ctx: kfree(ctx->reg); out_unmap_prot_sg: - if (prot_sg_cnt) - rdma_rw_unmap_sg(dev, prot_sg, prot_sg_cnt, dir); + if (prot_sgt.nents) + ib_dma_unmap_sgtable_attrs(dev, &prot_sgt, dir, 0); out_unmap_sg: - rdma_rw_unmap_sg(dev, sg, sg_cnt, dir); + ib_dma_unmap_sgtable_attrs(dev, &sgt, dir, 0); return ret; } EXPORT_SYMBOL(rdma_rw_ctx_signature_init); @@ -604,7 +594,7 @@ void rdma_rw_ctx_destroy(struct rdma_rw_ctx *ctx, struct ib_qp *qp, break; } - rdma_rw_unmap_sg(qp->pd->device, sg, sg_cnt, dir); + ib_dma_unmap_sg(qp->pd->device, sg, sg_cnt, dir); } EXPORT_SYMBOL(rdma_rw_ctx_destroy); @@ -632,8 +622,8 @@ void rdma_rw_ctx_destroy_signature(struct rdma_rw_ctx *ctx, struct ib_qp *qp, kfree(ctx->reg); if (prot_sg_cnt) - rdma_rw_unmap_sg(qp->pd->device, prot_sg, prot_sg_cnt, dir); - rdma_rw_unmap_sg(qp->pd->device, sg, sg_cnt, dir); + ib_dma_unmap_sg(qp->pd->device, prot_sg, prot_sg_cnt, dir); + ib_dma_unmap_sg(qp->pd->device, sg, sg_cnt, dir); } EXPORT_SYMBOL(rdma_rw_ctx_destroy_signature); @@ -676,7 +666,7 @@ void rdma_rw_init_qp(struct ib_device *dev, struct ib_qp_init_attr *attr) factor = 1; /* - * If the devices needs MRs to perform RDMA READ or WRITE operations, + * If the device needs MRs to perform RDMA READ or WRITE operations, * we'll need two additional MRs for the registrations and the * invalidation. */ diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c index b61576f702b8..c23e9c847314 100644 --- a/drivers/infiniband/core/sa_query.c +++ b/drivers/infiniband/core/sa_query.c @@ -32,7 +32,6 @@ * SOFTWARE. */ -#include <linux/module.h> #include <linux/init.h> #include <linux/err.h> #include <linux/random.h> @@ -51,6 +50,7 @@ #include <rdma/ib_marshall.h> #include <rdma/ib_addr.h> #include <rdma/opa_addr.h> +#include <rdma/rdma_cm.h> #include "sa.h" #include "core_priv.h" @@ -105,7 +105,10 @@ struct ib_sa_device { }; struct ib_sa_query { - void (*callback)(struct ib_sa_query *, int, struct ib_sa_mad *); + 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; @@ -123,14 +126,9 @@ struct ib_sa_query { #define IB_SA_CANCEL 0x00000002 #define IB_SA_QUERY_OPA 0x00000004 -struct ib_sa_service_query { - void (*callback)(int, struct ib_sa_service_rec *, void *); - void *context; - struct ib_sa_query sa_query; -}; - struct ib_sa_path_query { - void (*callback)(int, struct sa_path_rec *, void *); + void (*callback)(int status, struct sa_path_rec *rec, + unsigned int num_paths, void *context); void *context; struct ib_sa_query sa_query; struct sa_path_rec *conv_pr; @@ -154,6 +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; @@ -502,54 +507,6 @@ static const struct ib_field mcmember_rec_table[] = { .size_bits = 23 }, }; -#define SERVICE_REC_FIELD(field) \ - .struct_offset_bytes = offsetof(struct ib_sa_service_rec, field), \ - .struct_size_bytes = sizeof_field(struct ib_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 }, - { 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 = 64*8 }, - { SERVICE_REC_FIELD(data8), - .offset_words = 28, - .offset_bits = 0, - .size_bits = 16*8 }, - { SERVICE_REC_FIELD(data16), - .offset_words = 32, - .offset_bits = 0, - .size_bits = 8*16 }, - { SERVICE_REC_FIELD(data32), - .offset_words = 36, - .offset_bits = 0, - .size_bits = 4*32 }, - { SERVICE_REC_FIELD(data64), - .offset_words = 40, - .offset_bits = 0, - .size_bits = 2*64 }, -}; - #define CLASSPORTINFO_REC_FIELD(field) \ .struct_offset_bytes = offsetof(struct ib_class_port_info, field), \ .struct_size_bytes = sizeof_field(struct ib_class_port_info, field), \ @@ -736,6 +693,60 @@ 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) { query->flags &= ~IB_SA_ENABLE_LOCAL_SERVICE; @@ -760,13 +771,14 @@ static void ib_nl_set_path_rec_attrs(struct sk_buff *skb, /* Construct the family header first */ header = skb_put(skb, NLMSG_ALIGN(sizeof(*header))); - memcpy(header->device_name, dev_name(&query->port->agent->device->dev), - LS_DEVICE_NAME_MAX); + strscpy_pad(header->device_name, + dev_name(&query->port->agent->device->dev), + LS_DEVICE_NAME_MAX); header->port_num = query->port->port_num; if ((comp_mask & IB_SA_PATH_REC_REVERSIBLE) && sa_rec->reversible != 0) - query->path_use = LS_RESOLVE_PATH_USE_GMP; + query->path_use = LS_RESOLVE_PATH_USE_ALL; else query->path_use = LS_RESOLVE_PATH_USE_UNIDIRECTIONAL; header->path_use = query->path_use; @@ -919,50 +931,77 @@ static void send_handler(struct ib_mad_agent *agent, static void ib_nl_process_good_resolve_rsp(struct ib_sa_query *query, const struct nlmsghdr *nlh) { + struct sa_path_rec recs[RDMA_PRIMARY_PATH_MAX_REC_NUM]; + struct ib_sa_path_query *path_query; + struct ib_path_rec_data *rec_data; struct ib_mad_send_wc mad_send_wc; - struct ib_sa_mad *mad = NULL; const struct nlattr *head, *curr; - struct ib_path_rec_data *rec; - int len, rem; + struct ib_sa_mad *mad = NULL; + int len, rem, status = -EIO; + unsigned int num_prs = 0; u32 mask = 0; - int status = -EIO; - - if (query->callback) { - head = (const struct nlattr *) nlmsg_data(nlh); - len = nlmsg_len(nlh); - switch (query->path_use) { - case LS_RESOLVE_PATH_USE_UNIDIRECTIONAL: - mask = IB_PATH_PRIMARY | IB_PATH_OUTBOUND; - break; - case LS_RESOLVE_PATH_USE_ALL: - case LS_RESOLVE_PATH_USE_GMP: - default: - mask = IB_PATH_PRIMARY | IB_PATH_GMP | - IB_PATH_BIDIRECTIONAL; - break; - } - nla_for_each_attr(curr, head, len, rem) { - if (curr->nla_type == LS_NLA_TYPE_PATH_RECORD) { - rec = nla_data(curr); - /* - * Get the first one. In the future, we may - * need to get up to 6 pathrecords. - */ - if ((rec->flags & mask) == mask) { - mad = query->mad_buf->mad; - mad->mad_hdr.method |= - IB_MGMT_METHOD_RESP; - memcpy(mad->data, rec->path_rec, - sizeof(rec->path_rec)); - status = 0; - break; - } - } + if (!query->callback) + goto out; + + path_query = container_of(query, struct ib_sa_path_query, sa_query); + mad = query->mad_buf->mad; + + head = (const struct nlattr *) nlmsg_data(nlh); + len = nlmsg_len(nlh); + switch (query->path_use) { + case LS_RESOLVE_PATH_USE_UNIDIRECTIONAL: + mask = IB_PATH_PRIMARY | IB_PATH_OUTBOUND; + break; + + case LS_RESOLVE_PATH_USE_ALL: + mask = IB_PATH_PRIMARY; + break; + + case LS_RESOLVE_PATH_USE_GMP: + default: + mask = IB_PATH_PRIMARY | IB_PATH_GMP | + IB_PATH_BIDIRECTIONAL; + break; + } + + nla_for_each_attr(curr, head, len, rem) { + if (curr->nla_type != LS_NLA_TYPE_PATH_RECORD) + continue; + + rec_data = nla_data(curr); + if ((rec_data->flags & mask) != mask) + continue; + + if ((query->flags & IB_SA_QUERY_OPA) || + path_query->conv_pr) { + mad->mad_hdr.method |= IB_MGMT_METHOD_RESP; + memcpy(mad->data, rec_data->path_rec, + sizeof(rec_data->path_rec)); + query->callback(query, 0, mad); + goto out; } - query->callback(query, status, mad); + + status = 0; + ib_unpack(path_rec_table, ARRAY_SIZE(path_rec_table), + rec_data->path_rec, &recs[num_prs]); + recs[num_prs].flags = rec_data->flags; + recs[num_prs].rec_type = SA_PATH_REC_TYPE_IB; + sa_path_set_dmac_zero(&recs[num_prs]); + + num_prs++; + if (num_prs >= RDMA_PRIMARY_PATH_MAX_REC_NUM) + break; } + if (!status) { + mad->mad_hdr.method |= IB_MGMT_METHOD_RESP; + path_query->callback(status, recs, num_prs, + path_query->context); + } else + query->callback(query, status, mad); + +out: mad_send_wc.send_buf = query->mad_buf; mad_send_wc.status = IB_WC_SUCCESS; send_handler(query->mad_buf->mad_agent, &mad_send_wc); @@ -1035,6 +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; @@ -1042,7 +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) @@ -1060,9 +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; } @@ -1088,10 +1129,9 @@ int ib_nl_handle_resolve_resp(struct sk_buff *skb, struct netlink_ext_ack *extack) { unsigned long flags; - struct ib_sa_query *query; + struct ib_sa_query *query = NULL, *iter; struct ib_mad_send_buf *send_buf; struct ib_mad_send_wc mad_send_wc; - int found = 0; int ret; if ((nlh->nlmsg_flags & NLM_F_REQUEST) || @@ -1099,20 +1139,21 @@ int ib_nl_handle_resolve_resp(struct sk_buff *skb, return -EPERM; spin_lock_irqsave(&ib_nl_request_lock, flags); - list_for_each_entry(query, &ib_nl_request_list, list) { + list_for_each_entry(iter, &ib_nl_request_list, list) { /* * If the query is cancelled, let the timeout routine * take care of it. */ - if (nlh->nlmsg_seq == query->seq) { - found = !ib_sa_query_cancelled(query); - if (found) - list_del(&query->list); + if (nlh->nlmsg_seq == iter->seq) { + if (!ib_sa_query_cancelled(iter)) { + list_del(&iter->list); + query = iter; + } break; } } - if (!found) { + if (!query) { spin_unlock_irqrestore(&ib_nl_request_lock, flags); goto resp_out; } @@ -1358,6 +1399,7 @@ static int send_mad(struct ib_sa_query *query, unsigned long timeout_ms, { unsigned long flags; int ret, id; + const int nmbr_sa_query_retries = 10; xa_lock_irqsave(&queries, flags); ret = __xa_alloc(&queries, &id, query, xa_limit_32b, gfp_mask); @@ -1365,7 +1407,13 @@ static int send_mad(struct ib_sa_query *query, unsigned long timeout_ms, if (ret < 0) return ret; - query->mad_buf->timeout_ms = timeout_ms; + query->mad_buf->timeout_ms = timeout_ms / nmbr_sa_query_retries; + query->mad_buf->retries = nmbr_sa_query_retries; + if (!query->mad_buf->timeout_ms) { + /* Special case, very small timeout_ms */ + query->mad_buf->timeout_ms = 1; + query->mad_buf->retries = timeout_ms; + } query->mad_buf->context[0] = query; query->id = id; @@ -1405,6 +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) @@ -1435,7 +1497,7 @@ enum opa_pr_supported { /* * opa_pr_query_possible - Check if current PR query can be an OPA query. * - * Retuns PR_NOT_SUPPORTED if a path record query is not + * Returns PR_NOT_SUPPORTED if a path record query is not * possible, PR_OPA_SUPPORTED if an OPA path record query * is possible and PR_IB_SUPPORTED if an IB path record * query is possible. @@ -1459,40 +1521,101 @@ static int opa_pr_query_possible(struct ib_sa_client *client, } static void ib_sa_path_rec_callback(struct ib_sa_query *sa_query, - int status, - struct ib_sa_mad *mad) + int status, struct ib_sa_mad *mad) { struct ib_sa_path_query *query = container_of(sa_query, struct ib_sa_path_query, sa_query); + struct sa_path_rec rec = {}; - if (mad) { - struct sa_path_rec rec; + if (!mad) { + query->callback(status, NULL, 0, query->context); + return; + } - if (sa_query->flags & IB_SA_QUERY_OPA) { - ib_unpack(opa_path_rec_table, - ARRAY_SIZE(opa_path_rec_table), - mad->data, &rec); - rec.rec_type = SA_PATH_REC_TYPE_OPA; - query->callback(status, &rec, query->context); - } else { - ib_unpack(path_rec_table, - ARRAY_SIZE(path_rec_table), - mad->data, &rec); - rec.rec_type = SA_PATH_REC_TYPE_IB; - sa_path_set_dmac_zero(&rec); + if (sa_query->flags & IB_SA_QUERY_OPA) { + ib_unpack(opa_path_rec_table, ARRAY_SIZE(opa_path_rec_table), + mad->data, &rec); + rec.rec_type = SA_PATH_REC_TYPE_OPA; + query->callback(status, &rec, 1, query->context); + return; + } + + ib_unpack(path_rec_table, ARRAY_SIZE(path_rec_table), + mad->data, &rec); + rec.rec_type = SA_PATH_REC_TYPE_IB; + sa_path_set_dmac_zero(&rec); + + if (query->conv_pr) { + struct sa_path_rec opa; + + memset(&opa, 0, sizeof(struct sa_path_rec)); + sa_convert_path_ib_to_opa(&opa, &rec); + query->callback(status, &opa, 1, query->context); + } else { + query->callback(status, &rec, 1, query->context); + } +} - if (query->conv_pr) { - struct sa_path_rec opa; +#define IB_SA_DATA_OFFS 56 +#define IB_SERVICE_REC_SZ 176 - memset(&opa, 0, sizeof(struct sa_path_rec)); - sa_convert_path_ib_to_opa(&opa, &rec); - query->callback(status, &opa, query->context); - } else { - query->callback(status, &rec, query->context); +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++; } } - } else - query->callback(status, NULL, query->context); + } +} + +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) @@ -1504,6 +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 @@ -1536,7 +1667,7 @@ int ib_sa_path_rec_get(struct ib_sa_client *client, unsigned long timeout_ms, gfp_t gfp_mask, void (*callback)(int status, struct sa_path_rec *resp, - void *context), + unsigned int num_paths, void *context), void *context, struct ib_sa_query **sa_query) { @@ -1634,89 +1765,61 @@ err1: } EXPORT_SYMBOL(ib_sa_path_rec_get); -static void ib_sa_service_rec_callback(struct ib_sa_query *sa_query, - int status, - struct ib_sa_mad *mad) -{ - struct ib_sa_service_query *query = - container_of(sa_query, struct ib_sa_service_query, sa_query); - - if (mad) { - struct ib_sa_service_rec rec; - - ib_unpack(service_rec_table, ARRAY_SIZE(service_rec_table), - mad->data, &rec); - query->callback(status, &rec, query->context); - } else - query->callback(status, NULL, query->context); -} - -static void ib_sa_service_rec_release(struct ib_sa_query *sa_query) -{ - kfree(container_of(sa_query, struct ib_sa_service_query, sa_query)); -} - /** - * ib_sa_service_rec_query - Start Service Record operation - * @client:SA client - * @device:device to send request on - * @port_num: port number to send request on - * @method:SA method - should be get, set, or delete - * @rec:Service Record to send in request - * @comp_mask:component mask to send in request - * @timeout_ms:time to wait for response - * @gfp_mask:GFP mask to use for internal allocations - * @callback:function called when request completes, times out or is + * 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:request context, used to cancel request + * @context: opaque user context passed to callback + * @sa_query: query context, used to cancel query * - * Send a Service Record set/get/delete to the SA to register, - * unregister or query a service record. - * The callback function will be called when the request completes (or + * 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_query() is negative, it is an - * error code. Otherwise it is a request ID that can be used to cancel + * 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_query(struct ib_sa_client *client, - struct ib_device *device, u32 port_num, u8 method, - struct ib_sa_service_rec *rec, - ib_sa_comp_mask comp_mask, - unsigned long timeout_ms, gfp_t gfp_mask, - void (*callback)(int status, - struct ib_sa_service_rec *resp, - void *context), - void *context, - struct ib_sa_query **sa_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_service_query *query; struct ib_sa_device *sa_dev = ib_get_client_data(device, &sa_client); - struct ib_sa_port *port; + 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]; + port = &sa_dev->port[port_num - sa_dev->start_port]; agent = port->agent; - if (method != IB_MGMT_METHOD_GET && - method != IB_MGMT_METHOD_SET && - method != IB_SA_METHOD_DELETE) - return -EINVAL; - query = kzalloc(sizeof(*query), gfp_mask); if (!query) return -ENOMEM; - query->sa_query.port = port; + query->sa_query.port = port; + ret = alloc_mad(&query->sa_query, gfp_mask); if (ret) goto err1; @@ -1729,16 +1832,17 @@ int ib_sa_service_rec_query(struct ib_sa_client *client, mad = query->sa_query.mad_buf->mad; init_mad(&query->sa_query, agent); - query->sa_query.callback = callback ? ib_sa_service_rec_callback : NULL; - query->sa_query.release = ib_sa_service_rec_release; - mad->mad_hdr.method = method; - mad->mad_hdr.attr_id = cpu_to_be16(IB_SA_ATTR_SERVICE_REC); - mad->sa_hdr.comp_mask = comp_mask; + 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_pack(service_rec_table, ARRAY_SIZE(service_rec_table), - rec, mad->data); + 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) @@ -1750,16 +1854,14 @@ 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_query); +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) + int status, struct ib_sa_mad *mad) { struct ib_sa_mcmember_query *query = container_of(sa_query, struct ib_sa_mcmember_query, sa_query); @@ -1850,8 +1952,7 @@ err1: /* Support GuidInfoRecord */ static void ib_sa_guidinfo_rec_callback(struct ib_sa_query *sa_query, - int status, - struct ib_sa_mad *mad) + int status, struct ib_sa_mad *mad) { struct ib_sa_guidinfo_query *query = container_of(sa_query, struct ib_sa_guidinfo_query, sa_query); @@ -1960,8 +2061,7 @@ static void ib_classportinfo_cb(void *context) } static void ib_sa_classport_info_rec_callback(struct ib_sa_query *sa_query, - int status, - struct ib_sa_mad *mad) + int status, struct ib_sa_mad *mad) { unsigned long flags; struct ib_sa_classport_info_query *query = @@ -2129,23 +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); @@ -2161,17 +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); } @@ -2301,7 +2415,9 @@ static int ib_sa_add_one(struct ib_device *device) s = rdma_start_port(device); e = rdma_end_port(device); - sa_dev = kzalloc(struct_size(sa_dev, port, e - s + 1), GFP_KERNEL); + sa_dev = kzalloc(struct_size(sa_dev, port, + size_add(size_sub(e, s), 1)), + GFP_KERNEL); if (!sa_dev) return -ENOMEM; @@ -2321,8 +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; @@ -2431,7 +2548,6 @@ err1: void ib_sa_cleanup(void) { cancel_delayed_work(&ib_nl_timed_work); - flush_workqueue(ib_nl_wq); destroy_workqueue(ib_nl_wq); mcast_cleanup(); ib_unregister_client(&sa_client); diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c index 6146c3c1cbe5..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, @@ -342,6 +330,10 @@ static ssize_t rate_show(struct ib_device *ibdev, u32 port_num, speed = " NDR"; rate = 1000; break; + case IB_SPEED_XDR: + speed = " XDR"; + rate = 2000; + break; case IB_SPEED_SDR: default: /* default to SDR for invalid rates */ speed = " SDR"; @@ -433,6 +425,7 @@ static struct attribute *port_default_attrs[] = { &ib_port_attr_link_layer.attr, NULL }; +ATTRIBUTE_GROUPS(port_default); static ssize_t print_ndev(const struct ib_gid_attr *gid_attr, char *buf) { @@ -755,9 +748,9 @@ static void ib_port_release(struct kobject *kobj) for (i = 0; i != ARRAY_SIZE(port->groups); i++) kfree(port->groups[i].attrs); if (port->hw_stats_data) - kfree(port->hw_stats_data->stats); + rdma_free_hw_stats_struct(port->hw_stats_data->stats); kfree(port->hw_stats_data); - kfree(port); + kvfree(port); } static void ib_port_gid_attr_release(struct kobject *kobj) @@ -774,7 +767,7 @@ static void ib_port_gid_attr_release(struct kobject *kobj) static struct kobj_type port_type = { .release = ib_port_release, .sysfs_ops = &port_sysfs_ops, - .default_attrs = port_default_attrs + .default_groups = port_default_groups, }; static struct kobj_type gid_attr_type = { @@ -895,14 +888,14 @@ alloc_hw_stats_device(struct ib_device *ibdev) stats = ibdev->ops.alloc_hw_device_stats(ibdev); if (!stats) return ERR_PTR(-ENOMEM); - if (!stats->names || stats->num_counters <= 0) + if (!stats->descs || stats->num_counters <= 0) goto err_free_stats; /* * Two extra attribue elements here, one for the lifespan entry and * one to NULL terminate the list for the sysfs core code */ - data = kzalloc(struct_size(data, attrs, stats->num_counters + 1), + data = kzalloc(struct_size(data, attrs, size_add(stats->num_counters, 1)), GFP_KERNEL); if (!data) goto err_free_stats; @@ -911,7 +904,6 @@ alloc_hw_stats_device(struct ib_device *ibdev) if (!data->group.attrs) goto err_free_data; - mutex_init(&stats->lock); data->group.name = "hw_counters"; data->stats = stats; return data; @@ -919,14 +911,14 @@ alloc_hw_stats_device(struct ib_device *ibdev) err_free_data: kfree(data); err_free_stats: - kfree(stats); + rdma_free_hw_stats_struct(stats); return ERR_PTR(-ENOMEM); } void ib_device_release_hw_stats(struct hw_stats_device_data *data) { kfree(data->group.attrs); - kfree(data->stats); + rdma_free_hw_stats_struct(data->stats); kfree(data); } @@ -934,7 +926,8 @@ int ib_setup_device_attrs(struct ib_device *ibdev) { struct hw_stats_device_attribute *attr; struct hw_stats_device_data *data; - int i, ret; + bool opstat_skipped = false; + int i, ret, pos = 0; data = alloc_hw_stats_device(ibdev); if (IS_ERR(data)) { @@ -955,16 +948,23 @@ int ib_setup_device_attrs(struct ib_device *ibdev) data->stats->timestamp = jiffies; for (i = 0; i < data->stats->num_counters; i++) { - attr = &data->attrs[i]; + if (data->stats->descs[i].flags & IB_STAT_FLAG_OPTIONAL) { + opstat_skipped = true; + continue; + } + + WARN_ON(opstat_skipped); + attr = &data->attrs[pos]; sysfs_attr_init(&attr->attr.attr); - attr->attr.attr.name = data->stats->names[i]; + attr->attr.attr.name = data->stats->descs[i].name; attr->attr.attr.mode = 0444; attr->attr.show = hw_stat_device_show; attr->show = show_hw_stats; - data->group.attrs[i] = &attr->attr.attr; + data->group.attrs[pos] = &attr->attr.attr; + pos++; } - attr = &data->attrs[i]; + attr = &data->attrs[pos]; sysfs_attr_init(&attr->attr.attr); attr->attr.attr.name = "lifespan"; attr->attr.attr.mode = 0644; @@ -972,10 +972,11 @@ int ib_setup_device_attrs(struct ib_device *ibdev) attr->show = show_stats_lifespan; attr->attr.store = hw_stat_device_store; attr->store = set_stats_lifespan; - data->group.attrs[i] = &attr->attr.attr; + data->group.attrs[pos] = &attr->attr.attr; 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"); @@ -994,14 +995,14 @@ alloc_hw_stats_port(struct ib_port *port, struct attribute_group *group) stats = ibdev->ops.alloc_hw_port_stats(port->ibdev, port->port_num); if (!stats) return ERR_PTR(-ENOMEM); - if (!stats->names || stats->num_counters <= 0) + if (!stats->descs || stats->num_counters <= 0) goto err_free_stats; /* * Two extra attribue elements here, one for the lifespan entry and * one to NULL terminate the list for the sysfs core code */ - data = kzalloc(struct_size(data, attrs, stats->num_counters + 1), + data = kzalloc(struct_size(data, attrs, size_add(stats->num_counters, 1)), GFP_KERNEL); if (!data) goto err_free_stats; @@ -1010,7 +1011,6 @@ alloc_hw_stats_port(struct ib_port *port, struct attribute_group *group) if (!group->attrs) goto err_free_data; - mutex_init(&stats->lock); group->name = "hw_counters"; data->stats = stats; return data; @@ -1018,7 +1018,7 @@ alloc_hw_stats_port(struct ib_port *port, struct attribute_group *group) err_free_data: kfree(data); err_free_stats: - kfree(stats); + rdma_free_hw_stats_struct(stats); return ERR_PTR(-ENOMEM); } @@ -1027,7 +1027,8 @@ static int setup_hw_port_stats(struct ib_port *port, { struct hw_stats_port_attribute *attr; struct hw_stats_port_data *data; - int i, ret; + bool opstat_skipped = false; + int i, ret, pos = 0; data = alloc_hw_stats_port(port, group); if (IS_ERR(data)) @@ -1045,16 +1046,23 @@ static int setup_hw_port_stats(struct ib_port *port, data->stats->timestamp = jiffies; for (i = 0; i < data->stats->num_counters; i++) { - attr = &data->attrs[i]; + if (data->stats->descs[i].flags & IB_STAT_FLAG_OPTIONAL) { + opstat_skipped = true; + continue; + } + + WARN_ON(opstat_skipped); + attr = &data->attrs[pos]; sysfs_attr_init(&attr->attr.attr); - attr->attr.attr.name = data->stats->names[i]; + attr->attr.attr.name = data->stats->descs[i].name; attr->attr.attr.mode = 0444; attr->attr.show = hw_stat_port_show; attr->show = show_hw_stats; - group->attrs[i] = &attr->attr.attr; + group->attrs[pos] = &attr->attr.attr; + pos++; } - attr = &data->attrs[i]; + attr = &data->attrs[pos]; sysfs_attr_init(&attr->attr.attr); attr->attr.attr.name = "lifespan"; attr->attr.attr.mode = 0644; @@ -1062,7 +1070,7 @@ static int setup_hw_port_stats(struct ib_port *port, attr->show = show_stats_lifespan; attr->attr.store = hw_stat_port_store; attr->store = set_stats_lifespan; - group->attrs[i] = &attr->attr.attr; + group->attrs[pos] = &attr->attr.attr; port->hw_stats_data = data; return 0; @@ -1125,7 +1133,7 @@ static int setup_gid_attrs(struct ib_port *port, int ret; gid_attr_group = kzalloc(struct_size(gid_attr_group, attrs_list, - attr->gid_tbl_len * 2), + size_mul(attr->gid_tbl_len, 2)), GFP_KERNEL); if (!gid_attr_group) return -ENOMEM; @@ -1189,15 +1197,18 @@ static struct ib_port *setup_port(struct ib_core_device *coredev, int port_num, struct ib_port *p; int ret; - p = kzalloc(struct_size(p, attrs_list, - attr->gid_tbl_len + attr->pkey_tbl_len), - GFP_KERNEL); + p = kvzalloc(struct_size(p, attrs_list, + size_add(attr->gid_tbl_len, attr->pkey_tbl_len)), + GFP_KERNEL); if (!p) return ERR_PTR(-ENOMEM); p->ibdev = device; p->port_num = port_num; kobject_init(&p->kobj, &port_type); + if (device->port_data && is_full_dev) + device->port_data[port_num].sysfs = p; + cur_group = p->groups_list; ret = alloc_port_table_group("gids", &p->groups[0], p->attrs_list, attr->gid_tbl_len, show_port_gid); @@ -1243,9 +1254,6 @@ static struct ib_port *setup_port(struct ib_core_device *coredev, int port_num, } list_add_tail(&p->kobj.entry, &coredev->port_list); - if (device->port_data && is_full_dev) - device->port_data[port_num].sysfs = p; - return p; err_groups: @@ -1253,6 +1261,8 @@ err_groups: err_del: kobject_del(&p->kobj); err_put: + if (device->port_data && is_full_dev) + device->port_data[port_num].sysfs = NULL; kobject_put(&p->kobj); return ERR_PTR(ret); } @@ -1261,14 +1271,17 @@ static void destroy_port(struct ib_core_device *coredev, struct ib_port *port) { bool is_full_dev = &port->ibdev->coredev == coredev; - if (port->ibdev->port_data && - port->ibdev->port_data[port->port_num].sysfs == port) - port->ibdev->port_data[port->port_num].sysfs = NULL; list_del(&port->kobj.entry); if (is_full_dev) sysfs_remove_groups(&port->kobj, port->ibdev->ops.port_groups); + sysfs_remove_groups(&port->kobj, port->groups_list); kobject_del(&port->kobj); + + if (port->ibdev->port_data && + port->ibdev->port_data[port->port_num].sysfs == port) + port->ibdev->port_data[port->port_num].sysfs = NULL; + kobject_put(&port->kobj); } 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 2b72c4fa9550..ec3be65a2b88 100644 --- a/drivers/infiniband/core/ucma.c +++ b/drivers/infiniband/core/ucma.c @@ -69,9 +69,10 @@ 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, }, - { } }; struct ucma_file { @@ -95,6 +96,7 @@ struct ucma_context { u64 uid; struct list_head list; + struct list_head mc_list; struct work_struct close_work; }; @@ -105,6 +107,7 @@ struct ucma_multicast { u64 uid; u8 join_state; + struct list_head list; struct sockaddr_storage addr; }; @@ -198,6 +201,7 @@ static struct ucma_context *ucma_alloc_ctx(struct ucma_file *file) INIT_WORK(&ctx->close_work, ucma_close_id); init_completion(&ctx->comp); + INIT_LIST_HEAD(&ctx->mc_list); /* So list_del() will work if we don't do ucma_finish_ctx() */ INIT_LIST_HEAD(&ctx->list); ctx->file = file; @@ -278,6 +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); @@ -285,6 +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; @@ -357,7 +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; @@ -484,19 +493,19 @@ err1: static void ucma_cleanup_multicast(struct ucma_context *ctx) { - struct ucma_multicast *mc; - unsigned long index; + struct ucma_multicast *mc, *tmp; - xa_for_each(&multicast_table, index, mc) { - if (mc->ctx != ctx) - continue; + xa_lock(&multicast_table); + list_for_each_entry_safe(mc, tmp, &ctx->mc_list, list) { + list_del(&mc->list); /* * At this point mc->ctx->ref is 0 so the mc cannot leave the * lock on the reader and this is enough serialization */ - xa_erase(&multicast_table, index); + __xa_erase(&multicast_table, mc->id); kfree(mc); } + xa_unlock(&multicast_table); } static void ucma_cleanup_mc_events(struct ucma_multicast *mc) @@ -724,6 +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) @@ -751,8 +782,8 @@ static void ucma_copy_ib_route(struct rdma_ucm_query_route_resp *resp, { struct rdma_dev_addr *dev_addr; - resp->num_paths = route->num_paths; - switch (route->num_paths) { + resp->num_paths = route->num_pri_alt_paths; + switch (route->num_pri_alt_paths) { case 0: dev_addr = &route->addr.dev_addr; rdma_addr_get_dgid(dev_addr, @@ -778,8 +809,8 @@ static void ucma_copy_iboe_route(struct rdma_ucm_query_route_resp *resp, struct rdma_route *route) { - resp->num_paths = route->num_paths; - switch (route->num_paths) { + resp->num_paths = route->num_pri_alt_paths; + switch (route->num_pri_alt_paths) { case 0: rdma_ip2gid((struct sockaddr *)&route->addr.dst_addr, (union ib_gid *)&resp->ib_route[0].dgid); @@ -918,7 +949,7 @@ static ssize_t ucma_query_path(struct ucma_context *ctx, if (!resp) return -ENOMEM; - resp->num_paths = ctx->cm_id->route.num_paths; + resp->num_paths = ctx->cm_id->route.num_pri_alt_paths; for (i = 0, out_len -= sizeof(*resp); i < resp->num_paths && out_len > sizeof(struct ib_path_rec_data); i++, out_len -= sizeof(struct ib_path_rec_data)) { @@ -990,6 +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) @@ -1018,6 +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; @@ -1469,12 +1540,16 @@ static ssize_t ucma_process_join(struct ucma_file *file, mc->uid = cmd->uid; memcpy(&mc->addr, addr, cmd->addr_size); - if (xa_alloc(&multicast_table, &mc->id, NULL, xa_limit_32b, + xa_lock(&multicast_table); + if (__xa_alloc(&multicast_table, &mc->id, NULL, xa_limit_32b, GFP_KERNEL)) { ret = -ENOMEM; goto err_free_mc; } + list_add_tail(&mc->list, &ctx->mc_list); + xa_unlock(&multicast_table); + mutex_lock(&ctx->mutex); ret = rdma_join_multicast(ctx->cm_id, (struct sockaddr *)&mc->addr, join_state, mc); @@ -1500,8 +1575,11 @@ err_leave_multicast: mutex_unlock(&ctx->mutex); ucma_cleanup_mc_events(mc); err_xa_erase: - xa_erase(&multicast_table, mc->id); + xa_lock(&multicast_table); + list_del(&mc->list); + __xa_erase(&multicast_table, mc->id); err_free_mc: + xa_unlock(&multicast_table); kfree(mc); err_put_ctx: ucma_put_ctx(ctx); @@ -1569,15 +1647,17 @@ static ssize_t ucma_leave_multicast(struct ucma_file *file, mc = ERR_PTR(-EINVAL); else if (!refcount_inc_not_zero(&mc->ctx->ref)) mc = ERR_PTR(-ENXIO); - else - __xa_erase(&multicast_table, mc->id); - xa_unlock(&multicast_table); if (IS_ERR(mc)) { + xa_unlock(&multicast_table); ret = PTR_ERR(mc); goto out; } + list_del(&mc->list); + __xa_erase(&multicast_table, mc->id); + xa_unlock(&multicast_table); + mutex_lock(&mc->ctx->mutex); rdma_leave_multicast(mc->ctx->cm_id, (struct sockaddr *) &mc->addr); mutex_unlock(&mc->ctx->mutex); @@ -1604,7 +1684,6 @@ static ssize_t ucma_migrate_id(struct ucma_file *new_file, struct ucma_event *uevent, *tmp; struct ucma_context *ctx; LIST_HEAD(event_list); - struct fd f; struct ucma_file *cur_file; int ret = 0; @@ -1612,21 +1691,17 @@ static ssize_t ucma_migrate_id(struct ucma_file *new_file, return -EFAULT; /* Get current fd to protect against it being closed */ - f = fdget(cmd.fd); - if (!f.file) + CLASS(fd, f)(cmd.fd); + if (fd_empty(f)) return -ENOENT; - if (f.file->f_op != &ucma_fops) { - ret = -EINVAL; - goto file_put; - } - cur_file = f.file->private_data; + if (fd_file(f)->f_op != &ucma_fops) + return -EINVAL; + cur_file = fd_file(f)->private_data; /* Validate current fd and prevent destruction of id. */ ctx = ucma_get_ctx(cur_file, cmd.id); - if (IS_ERR(ctx)) { - ret = PTR_ERR(ctx); - goto file_put; - } + if (IS_ERR(ctx)) + return PTR_ERR(ctx); rdma_lock_handler(ctx->cm_id); /* @@ -1667,8 +1742,55 @@ static ssize_t ucma_migrate_id(struct ucma_file *new_file, err_unlock: rdma_unlock_handler(ctx->cm_id); ucma_put_ctx(ctx); -file_put: - fdput(f); + 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; } @@ -1697,7 +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, @@ -1806,7 +1930,6 @@ static const struct file_operations ucma_fops = { .release = ucma_close, .write = ucma_write, .poll = ucma_poll, - .llseek = no_llseek, }; static struct miscdevice ucma_misc = { 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 0eb40025075f..8137031c2a65 100644 --- a/drivers/infiniband/core/umem.c +++ b/drivers/infiniband/core/umem.c @@ -45,21 +45,27 @@ #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; struct scatterlist *sg; unsigned int i; - if (umem->nmap > 0) - ib_dma_unmap_sg(dev, umem->sg_head.sgl, umem->sg_nents, - DMA_BIDIRECTIONAL); + if (dirty) + ib_dma_unmap_sgtable_attrs(dev, &umem->sgt_append.sgt, + DMA_BIDIRECTIONAL, 0); - for_each_sg(umem->sg_head.sgl, sg, umem->sg_nents, 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); - sg_free_table(&umem->sg_head); + if (i && !(i % RESCHED_LOOP_CNT_THRESHOLD)) + cond_resched(); + } + + sg_free_append_table(&umem->sgt_append); } /** @@ -80,11 +86,16 @@ 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; + if (umem->is_odp) { unsigned int page_size = BIT(to_ib_umem_odp(umem)->page_shift); @@ -94,13 +105,6 @@ unsigned long ib_umem_find_best_pgsz(struct ib_umem *umem, return page_size; } - /* rdma_for_each_block() has a bug if the page size is smaller than the - * page size used to build the umem. For now prevent smaller page sizes - * from being returned. - */ - pgsz_bitmap &= GENMASK(BITS_PER_LONG - 1, PAGE_SHIFT); - - umem->iova = va = virt; /* The best result is the smallest page size that results in the minimum * number of required pages. Compute the largest page size that could * work based on VA address bits that don't change. @@ -111,18 +115,31 @@ unsigned long ib_umem_find_best_pgsz(struct ib_umem *umem, /* offset into first SGL */ pgoff = umem->address & ~PAGE_MASK; - for_each_sg(umem->sg_head.sgl, sg, umem->nmap, i) { - /* Walk SGL and reduce max page size if VA/PA bits differ - * for any address. + for_each_sgtable_dma_sg(&umem->sgt_append.sgt, sg, i) { + /* 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->nmap - 1)) - mask |= va; + pgoff = 0; } @@ -155,9 +172,8 @@ struct ib_umem *ib_umem_get(struct ib_device *device, unsigned long addr, unsigned long dma_attr = 0; struct mm_struct *mm; unsigned long npages; - int ret; - struct scatterlist *sg = NULL; - unsigned int gup_flags = FOLL_WRITE; + int pinned, ret; + unsigned int gup_flags = FOLL_LONGTERM; /* * If the combination of the addr and size requested for this memory @@ -211,29 +227,29 @@ struct ib_umem *ib_umem_get(struct ib_device *device, unsigned long addr, cur_base = addr & PAGE_MASK; - if (!umem->writable) - gup_flags |= FOLL_FORCE; + if (umem->writable) + gup_flags |= FOLL_WRITE; while (npages) { cond_resched(); - ret = pin_user_pages_fast(cur_base, + pinned = pin_user_pages_fast(cur_base, min_t(unsigned long, npages, PAGE_SIZE / sizeof(struct page *)), - gup_flags | FOLL_LONGTERM, page_list); - if (ret < 0) + gup_flags, page_list); + if (pinned < 0) { + ret = pinned; goto umem_release; + } - cur_base += ret * PAGE_SIZE; - npages -= ret; - sg = __sg_alloc_table_from_pages(&umem->sg_head, page_list, ret, - 0, ret << PAGE_SHIFT, - ib_dma_max_seg_size(device), sg, npages, - GFP_KERNEL); - umem->sg_nents = umem->sg_head.nents; - if (IS_ERR(sg)) { - unpin_user_pages_dirty_lock(page_list, ret, 0); - ret = PTR_ERR(sg); + cur_base += pinned * PAGE_SIZE; + npages -= pinned; + ret = sg_alloc_append_table_from_pages( + &umem->sgt_append, page_list, pinned, 0, + pinned << PAGE_SHIFT, ib_dma_max_seg_size(device), + npages, GFP_KERNEL); + if (ret) { + unpin_user_pages_dirty_lock(page_list, pinned, 0); goto umem_release; } } @@ -241,16 +257,10 @@ struct ib_umem *ib_umem_get(struct ib_device *device, unsigned long addr, if (access & IB_ACCESS_RELAXED_ORDERING) dma_attr |= DMA_ATTR_WEAK_ORDERING; - umem->nmap = - ib_dma_map_sg_attrs(device, umem->sg_head.sgl, umem->sg_nents, - DMA_BIDIRECTIONAL, dma_attr); - - if (!umem->nmap) { - ret = -ENOMEM; + ret = ib_dma_map_sgtable_attrs(device, &umem->sgt_append.sgt, + DMA_BIDIRECTIONAL, dma_attr); + if (ret) goto umem_release; - } - - ret = 0; goto out; umem_release: @@ -310,7 +320,8 @@ int ib_umem_copy_from(void *dst, struct ib_umem *umem, size_t offset, return -EINVAL; } - ret = sg_pcopy_to_buffer(umem->sg_head.sgl, umem->sg_nents, dst, length, + ret = sg_pcopy_to_buffer(umem->sgt_append.sgt.sgl, + umem->sgt_append.sgt.orig_nents, dst, length, offset + ib_umem_offset(umem)); if (ret < 0) diff --git a/drivers/infiniband/core/umem_dmabuf.c b/drivers/infiniband/core/umem_dmabuf.c index c6e875619fac..0ec2e4120cc9 100644 --- a/drivers/infiniband/core/umem_dmabuf.c +++ b/drivers/infiniband/core/umem_dmabuf.c @@ -6,24 +6,31 @@ #include <linux/dma-buf.h> #include <linux/dma-resv.h> #include <linux/dma-mapping.h> +#include <linux/module.h> #include "uverbs.h" +MODULE_IMPORT_NS("DMA_BUF"); + int ib_umem_dmabuf_map_pages(struct ib_umem_dmabuf *umem_dmabuf) { struct sg_table *sgt; struct scatterlist *sg; - struct dma_fence *fence; unsigned long start, end, cur = 0; unsigned int nmap = 0; + long ret; int i; dma_resv_assert_held(umem_dmabuf->attach->dmabuf->resv); + if (umem_dmabuf->revoked) + return -EINVAL; + if (umem_dmabuf->sgt) goto wait_fence; - sgt = dma_buf_map_attachment(umem_dmabuf->attach, DMA_BIDIRECTIONAL); + sgt = dma_buf_map_attachment(umem_dmabuf->attach, + DMA_BIDIRECTIONAL); if (IS_ERR(sgt)) return PTR_ERR(sgt); @@ -55,9 +62,8 @@ int ib_umem_dmabuf_map_pages(struct ib_umem_dmabuf *umem_dmabuf) cur += sg_dma_len(sg); } - umem_dmabuf->umem.sg_head.sgl = umem_dmabuf->first_sg; - umem_dmabuf->umem.sg_head.nents = nmap; - umem_dmabuf->umem.nmap = nmap; + umem_dmabuf->umem.sgt_append.sgt.sgl = umem_dmabuf->first_sg; + umem_dmabuf->umem.sgt_append.sgt.nents = nmap; umem_dmabuf->sgt = sgt; wait_fence: @@ -66,10 +72,13 @@ wait_fence: * may be not up-to-date. Wait for the exporter to finish * the migration. */ - fence = dma_resv_excl_fence(umem_dmabuf->attach->dmabuf->resv); - if (fence) - return dma_fence_wait(fence, false); - + ret = dma_resv_wait_timeout(umem_dmabuf->attach->dmabuf->resv, + DMA_RESV_USAGE_KERNEL, + false, MAX_SCHEDULE_TIMEOUT); + if (ret < 0) + return ret; + if (ret == 0) + return -ETIMEDOUT; return 0; } EXPORT_SYMBOL(ib_umem_dmabuf_map_pages); @@ -104,10 +113,12 @@ void ib_umem_dmabuf_unmap_pages(struct ib_umem_dmabuf *umem_dmabuf) } EXPORT_SYMBOL(ib_umem_dmabuf_unmap_pages); -struct ib_umem_dmabuf *ib_umem_dmabuf_get(struct ib_device *device, - unsigned long offset, size_t size, - int fd, int access, - const struct dma_buf_attach_ops *ops) +static struct ib_umem_dmabuf * +ib_umem_dmabuf_get_with_dma_device(struct ib_device *device, + struct device *dma_device, + unsigned long offset, size_t size, + int fd, int access, + const struct dma_buf_attach_ops *ops) { struct dma_buf *dmabuf; struct ib_umem_dmabuf *umem_dmabuf; @@ -146,7 +157,7 @@ struct ib_umem_dmabuf *ib_umem_dmabuf_get(struct ib_device *device, umem_dmabuf->attach = dma_buf_dynamic_attach( dmabuf, - device->dma_device, + dma_device, ops, umem_dmabuf); if (IS_ERR(umem_dmabuf->attach)) { @@ -162,15 +173,101 @@ out_release_dmabuf: dma_buf_put(dmabuf); return ret; } + +struct ib_umem_dmabuf *ib_umem_dmabuf_get(struct ib_device *device, + unsigned long offset, size_t size, + int fd, int access, + const struct dma_buf_attach_ops *ops) +{ + return ib_umem_dmabuf_get_with_dma_device(device, device->dma_device, + offset, size, fd, access, ops); +} EXPORT_SYMBOL(ib_umem_dmabuf_get); -void ib_umem_dmabuf_release(struct ib_umem_dmabuf *umem_dmabuf) +static void +ib_umem_dmabuf_unsupported_move_notify(struct dma_buf_attachment *attach) +{ + struct ib_umem_dmabuf *umem_dmabuf = attach->importer_priv; + + ibdev_warn_ratelimited(umem_dmabuf->umem.ibdev, + "Invalidate callback should not be called when memory is pinned\n"); +} + +static struct dma_buf_attach_ops ib_umem_dmabuf_attach_pinned_ops = { + .allow_peer2peer = true, + .move_notify = ib_umem_dmabuf_unsupported_move_notify, +}; + +struct ib_umem_dmabuf * +ib_umem_dmabuf_get_pinned_with_dma_device(struct ib_device *device, + struct device *dma_device, + unsigned long offset, size_t size, + int fd, int access) +{ + struct ib_umem_dmabuf *umem_dmabuf; + int err; + + umem_dmabuf = ib_umem_dmabuf_get_with_dma_device(device, dma_device, offset, + size, fd, access, + &ib_umem_dmabuf_attach_pinned_ops); + if (IS_ERR(umem_dmabuf)) + return umem_dmabuf; + + dma_resv_lock(umem_dmabuf->attach->dmabuf->resv, NULL); + err = dma_buf_pin(umem_dmabuf->attach); + if (err) + goto err_release; + umem_dmabuf->pinned = 1; + + err = ib_umem_dmabuf_map_pages(umem_dmabuf); + if (err) + goto err_unpin; + dma_resv_unlock(umem_dmabuf->attach->dmabuf->resv); + + return umem_dmabuf; + +err_unpin: + dma_buf_unpin(umem_dmabuf->attach); +err_release: + dma_resv_unlock(umem_dmabuf->attach->dmabuf->resv); + ib_umem_release(&umem_dmabuf->umem); + return ERR_PTR(err); +} +EXPORT_SYMBOL(ib_umem_dmabuf_get_pinned_with_dma_device); + +struct ib_umem_dmabuf *ib_umem_dmabuf_get_pinned(struct ib_device *device, + unsigned long offset, + size_t size, int fd, + int access) +{ + return ib_umem_dmabuf_get_pinned_with_dma_device(device, device->dma_device, + offset, size, fd, access); +} +EXPORT_SYMBOL(ib_umem_dmabuf_get_pinned); + +void ib_umem_dmabuf_revoke(struct ib_umem_dmabuf *umem_dmabuf) { struct dma_buf *dmabuf = umem_dmabuf->attach->dmabuf; dma_resv_lock(dmabuf->resv, NULL); + if (umem_dmabuf->revoked) + goto end; ib_umem_dmabuf_unmap_pages(umem_dmabuf); + if (umem_dmabuf->pinned) { + dma_buf_unpin(umem_dmabuf->attach); + umem_dmabuf->pinned = 0; + } + umem_dmabuf->revoked = 1; +end: dma_resv_unlock(dmabuf->resv); +} +EXPORT_SYMBOL(ib_umem_dmabuf_revoke); + +void ib_umem_dmabuf_release(struct ib_umem_dmabuf *umem_dmabuf) +{ + struct dma_buf *dmabuf = umem_dmabuf->attach->dmabuf; + + ib_umem_dmabuf_revoke(umem_dmabuf); dma_buf_detach(dmabuf, umem_dmabuf->attach); dma_buf_put(dmabuf); diff --git a/drivers/infiniband/core/umem_odp.c b/drivers/infiniband/core/umem_odp.c index 9462dbe66014..572a91a62a7b 100644 --- a/drivers/infiniband/core/umem_odp.c +++ b/drivers/infiniband/core/umem_odp.c @@ -41,67 +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_verbs.h> -#include <rdma/ib_umem.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; } @@ -120,7 +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); @@ -132,16 +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); @@ -227,7 +236,6 @@ struct ib_umem_odp *ib_umem_odp_get(struct ib_device *device, const struct mmu_interval_notifier_ops *ops) { struct ib_umem_odp *umem_odp; - struct mm_struct *mm; int ret; if (WARN_ON_ONCE(!(access & IB_ACCESS_ON_DEMAND))) @@ -241,7 +249,7 @@ struct ib_umem_odp *ib_umem_odp_get(struct ib_device *device, umem_odp->umem.length = size; umem_odp->umem.address = addr; umem_odp->umem.writable = ib_access_writable(access); - umem_odp->umem.owning_mm = mm = current->mm; + umem_odp->umem.owning_mm = current->mm; umem_odp->notifier.ops = ops; umem_odp->page_shift = PAGE_SHIFT; @@ -263,77 +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. - * @current_seq: sequence number for synchronization with invalidations. - * the sequence number is taken from - * umem_odp->notifiers_seq. - * - * 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. * @@ -361,9 +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; @@ -389,11 +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: @@ -421,22 +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 @@ -449,24 +413,15 @@ 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 sucesss lock should stay on hold for the callee */ + /* upon success lock should stay on hold for the callee */ if (!ret) ret = dma_index - start_idx; else mutex_unlock(&umem_odp->umem_mutex); out_put_mm: - mmput(owning_mm); + mmput_async(owning_mm); out_put_task: if (owning_process) put_task_struct(owning_process); @@ -477,45 +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/user_mad.c b/drivers/infiniband/core/user_mad.c index 98cb594cd9a6..fd67fc9fe85a 100644 --- a/drivers/infiniband/core/user_mad.c +++ b/drivers/infiniband/core/user_mad.c @@ -63,6 +63,8 @@ MODULE_AUTHOR("Roland Dreier"); MODULE_DESCRIPTION("InfiniBand userspace MAD packet access"); MODULE_LICENSE("Dual BSD/GPL"); +#define MAX_UMAD_RECV_LIST_SIZE 200000 + enum { IB_UMAD_MAX_PORTS = RDMA_MAX_PORTS, IB_UMAD_MAX_AGENTS = 32, @@ -113,6 +115,7 @@ struct ib_umad_file { struct mutex mutex; struct ib_umad_port *port; struct list_head recv_list; + atomic_t recv_list_size; struct list_head send_list; struct list_head port_list; spinlock_t send_lock; @@ -131,6 +134,11 @@ struct ib_umad_packet { struct ib_user_mad mad; }; +struct ib_rmpp_mad_hdr { + struct ib_mad_hdr mad_hdr; + struct ib_rmpp_hdr rmpp_hdr; +} __packed; + #define CREATE_TRACE_POINTS #include <trace/events/ib_umad.h> @@ -175,24 +183,28 @@ static struct ib_mad_agent *__get_agent(struct ib_umad_file *file, int id) return file->agents_dead ? NULL : file->agent[id]; } -static int queue_packet(struct ib_umad_file *file, - struct ib_mad_agent *agent, - struct ib_umad_packet *packet) +static int queue_packet(struct ib_umad_file *file, struct ib_mad_agent *agent, + struct ib_umad_packet *packet, bool is_recv_mad) { int ret = 1; mutex_lock(&file->mutex); + if (is_recv_mad && + atomic_read(&file->recv_list_size) > MAX_UMAD_RECV_LIST_SIZE) + goto unlock; + for (packet->mad.hdr.id = 0; packet->mad.hdr.id < IB_UMAD_MAX_AGENTS; packet->mad.hdr.id++) if (agent == __get_agent(file, packet->mad.hdr.id)) { list_add_tail(&packet->list, &file->recv_list); + atomic_inc(&file->recv_list_size); wake_up_interruptible(&file->recv_wait); ret = 0; break; } - +unlock: mutex_unlock(&file->mutex); return ret; @@ -219,7 +231,7 @@ static void send_handler(struct ib_mad_agent *agent, if (send_wc->status == IB_WC_RESP_TIMEOUT_ERR) { packet->length = IB_MGMT_MAD_HDR; packet->mad.hdr.status = ETIMEDOUT; - if (!queue_packet(file, agent, packet)) + if (!queue_packet(file, agent, packet, false)) return; } kfree(packet); @@ -279,7 +291,7 @@ static void recv_handler(struct ib_mad_agent *agent, rdma_destroy_ah_attr(&ah_attr); } - if (queue_packet(file, agent, packet)) + if (queue_packet(file, agent, packet, true)) goto err2; return; @@ -404,6 +416,7 @@ static ssize_t ib_umad_read(struct file *filp, char __user *buf, packet = list_entry(file->recv_list.next, struct ib_umad_packet, list); list_del(&packet->list); + atomic_dec(&file->recv_list_size); mutex_unlock(&file->mutex); @@ -416,6 +429,7 @@ static ssize_t ib_umad_read(struct file *filp, char __user *buf, /* Requeue packet */ mutex_lock(&file->mutex); list_add(&packet->list, &file->recv_list); + atomic_inc(&file->recv_list_size); mutex_unlock(&file->mutex); } else { if (packet->recv_wc) @@ -494,11 +508,11 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, size_t count, loff_t *pos) { struct ib_umad_file *file = filp->private_data; + struct ib_rmpp_mad_hdr *rmpp_mad_hdr; struct ib_umad_packet *packet; struct ib_mad_agent *agent; struct rdma_ah_attr ah_attr; struct ib_ah *ah; - struct ib_rmpp_mad *rmpp_mad; __be64 *tid; int ret, data_len, hdr_len, copy_offset, rmpp_active; u8 base_version; @@ -506,7 +520,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, if (count < hdr_size(file) + IB_MGMT_RMPP_HDR) return -EINVAL; - packet = kzalloc(sizeof *packet + IB_MGMT_RMPP_HDR, GFP_KERNEL); + packet = kzalloc(sizeof(*packet) + IB_MGMT_RMPP_HDR, GFP_KERNEL); if (!packet) return -ENOMEM; @@ -560,13 +574,13 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, goto err_up; } - rmpp_mad = (struct ib_rmpp_mad *) packet->mad.data; - hdr_len = ib_get_mad_data_offset(rmpp_mad->mad_hdr.mgmt_class); + rmpp_mad_hdr = (struct ib_rmpp_mad_hdr *)packet->mad.data; + hdr_len = ib_get_mad_data_offset(rmpp_mad_hdr->mad_hdr.mgmt_class); - if (ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class) + if (ib_is_mad_class_rmpp(rmpp_mad_hdr->mad_hdr.mgmt_class) && ib_mad_kernel_rmpp_agent(agent)) { copy_offset = IB_MGMT_RMPP_HDR; - rmpp_active = ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & + rmpp_active = ib_get_rmpp_flags(&rmpp_mad_hdr->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE; } else { copy_offset = IB_MGMT_MAD_HDR; @@ -615,12 +629,12 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, tid = &((struct ib_mad_hdr *) packet->msg->mad)->tid; *tid = cpu_to_be64(((u64) agent->hi_tid) << 32 | (be64_to_cpup(tid) & 0xffffffff)); - rmpp_mad->mad_hdr.tid = *tid; + rmpp_mad_hdr->mad_hdr.tid = *tid; } if (!ib_mad_kernel_rmpp_agent(agent) - && ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class) - && (ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE)) { + && ib_is_mad_class_rmpp(rmpp_mad_hdr->mad_hdr.mgmt_class) + && (ib_get_rmpp_flags(&rmpp_mad_hdr->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE)) { spin_lock_irq(&file->send_lock); list_add_tail(&packet->list, &file->send_list); spin_unlock_irq(&file->send_lock); @@ -1068,7 +1082,6 @@ static const struct file_operations umad_fops = { #endif .open = ib_umad_open, .release = ib_umad_close, - .llseek = no_llseek, }; static int ib_umad_sm_open(struct inode *inode, struct file *filp) @@ -1136,7 +1149,6 @@ static const struct file_operations umad_sm_fops = { .owner = THIS_MODULE, .open = ib_umad_sm_open, .release = ib_umad_sm_close, - .llseek = no_llseek, }; static struct ib_umad_port *get_port(struct ib_device *ibdev, @@ -1224,13 +1236,13 @@ static struct attribute *umad_class_dev_attrs[] = { }; ATTRIBUTE_GROUPS(umad_class_dev); -static char *umad_devnode(struct device *dev, umode_t *mode) +static char *umad_devnode(const struct device *dev, umode_t *mode) { return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev)); } -static ssize_t abi_version_show(struct class *class, - struct class_attribute *attr, char *buf) +static ssize_t abi_version_show(const struct class *class, + const struct class_attribute *attr, char *buf) { return sysfs_emit(buf, "%d\n", IB_USER_MAD_ABI_VERSION); } @@ -1307,15 +1319,17 @@ static int ib_umad_init_port(struct ib_device *device, int port_num, if (ret) goto err_cdev; - ib_umad_init_port_dev(&port->sm_dev, port, device); - port->sm_dev.devt = base_issm; - dev_set_name(&port->sm_dev, "issm%d", port->dev_num); - cdev_init(&port->sm_cdev, &umad_sm_fops); - port->sm_cdev.owner = THIS_MODULE; + if (rdma_cap_ib_smi(device, port_num)) { + ib_umad_init_port_dev(&port->sm_dev, port, device); + port->sm_dev.devt = base_issm; + dev_set_name(&port->sm_dev, "issm%d", port->dev_num); + cdev_init(&port->sm_cdev, &umad_sm_fops); + port->sm_cdev.owner = THIS_MODULE; - ret = cdev_device_add(&port->sm_cdev, &port->sm_dev); - if (ret) - goto err_dev; + ret = cdev_device_add(&port->sm_cdev, &port->sm_dev); + if (ret) + goto err_dev; + } return 0; @@ -1331,9 +1345,13 @@ err_cdev: static void ib_umad_kill_port(struct ib_umad_port *port) { struct ib_umad_file *file; + bool has_smi = false; int id; - cdev_device_del(&port->sm_cdev, &port->sm_dev); + if (rdma_cap_ib_smi(port->ib_dev, port->port_num)) { + cdev_device_del(&port->sm_cdev, &port->sm_dev); + has_smi = true; + } cdev_device_del(&port->cdev, &port->dev); mutex_lock(&port->file_mutex); @@ -1359,7 +1377,8 @@ static void ib_umad_kill_port(struct ib_umad_port *port) ida_free(&umad_ida, port->dev_num); /* balances device_initialize() */ - put_device(&port->sm_dev); + if (has_smi) + put_device(&port->sm_dev); put_device(&port->dev); } @@ -1373,7 +1392,9 @@ static int ib_umad_add_one(struct ib_device *device) s = rdma_start_port(device); e = rdma_end_port(device); - umad_dev = kzalloc(struct_size(umad_dev, ports, e - s + 1), GFP_KERNEL); + umad_dev = kzalloc(struct_size(umad_dev, ports, + size_add(size_sub(e, s), 1)), + GFP_KERNEL); if (!umad_dev) return -ENOMEM; diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h index 821d93c8f712..797e2fcc8072 100644 --- a/drivers/infiniband/core/uverbs.h +++ b/drivers/infiniband/core/uverbs.h @@ -133,35 +133,6 @@ struct ib_uverbs_completion_event_file { struct ib_uverbs_event_queue ev_queue; }; -struct ib_uverbs_file { - struct kref ref; - struct ib_uverbs_device *device; - struct mutex ucontext_lock; - /* - * ucontext must be accessed via ib_uverbs_get_ucontext() or with - * ucontext_lock held - */ - struct ib_ucontext *ucontext; - struct ib_uverbs_async_event_file *default_async_file; - struct list_head list; - - /* - * To access the uobjects list hw_destroy_rwsem must be held for write - * OR hw_destroy_rwsem held for read AND uobjects_lock held. - * hw_destroy_rwsem should be called across any destruction of the HW - * object of an associated uobject. - */ - struct rw_semaphore hw_destroy_rwsem; - spinlock_t uobjects_lock; - struct list_head uobjects; - - struct mutex umap_lock; - struct list_head umaps; - struct page *disassociate_page; - - struct xarray idr; -}; - struct ib_uverbs_event { union { struct ib_uverbs_async_event_desc async; diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c index 8c8ca7bce3ca..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) @@ -337,7 +356,7 @@ static void copy_query_dev_fields(struct ib_ucontext *ucontext, resp->hw_ver = attr->hw_ver; resp->max_qp = attr->max_qp; resp->max_qp_wr = attr->max_qp_wr; - resp->device_cap_flags = lower_32_bits(attr->device_cap_flags); + resp->device_cap_flags = lower_32_bits(attr->device_cap_flags); resp->max_sge = min(attr->max_send_sge, attr->max_recv_sge); resp->max_sge_rd = attr->max_sge_rd; resp->max_cq = attr->max_cq; @@ -572,7 +591,7 @@ static int ib_uverbs_open_xrcd(struct uverbs_attr_bundle *attrs) struct inode *inode = NULL; int new_xrcd = 0; struct ib_device *ib_dev; - struct fd f = {}; + struct fd f = EMPTY_FD; int ret; ret = uverbs_request(attrs, &cmd, sizeof(cmd)); @@ -584,12 +603,12 @@ static int ib_uverbs_open_xrcd(struct uverbs_attr_bundle *attrs) if (cmd.fd != -1) { /* search for file descriptor */ f = fdget(cmd.fd); - if (!f.file) { + if (fd_empty(f)) { ret = -EBADF; goto err_tree_mutex_unlock; } - inode = file_inode(f.file); + inode = file_inode(fd_file(f)); xrcd = find_xrcd(ibudev, inode); if (!xrcd && !(cmd.oflags & O_CREAT)) { /* no file descriptor. Need CREATE flag */ @@ -632,8 +651,7 @@ static int ib_uverbs_open_xrcd(struct uverbs_attr_bundle *attrs) atomic_inc(&xrcd->usecnt); } - if (f.file) - fdput(f); + fdput(f); mutex_unlock(&ibudev->xrcd_tree_mutex); uobj_finalize_uobj_create(&obj->uobject, attrs); @@ -648,8 +666,7 @@ err: uobj_alloc_abort(&obj->uobject, attrs); err_tree_mutex_unlock: - if (f.file) - fdput(f); + fdput(f); mutex_unlock(&ibudev->xrcd_tree_mutex); @@ -718,13 +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); @@ -739,6 +756,7 @@ static int ib_uverbs_reg_mr(struct uverbs_attr_bundle *attrs) mr->uobject = uobj; atomic_inc(&pd->usecnt); mr->iova = cmd.hca_va; + mr->length = cmd.length; rdma_restrack_new(&mr->res, RDMA_RESTRACK_MR); rdma_restrack_set_name(&mr->res, NULL); @@ -808,8 +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 { @@ -837,11 +855,8 @@ static int ib_uverbs_rereg_mr(struct uverbs_attr_bundle *attrs) new_mr->device = new_pd->device; new_mr->pd = new_pd; new_mr->type = IB_MR_TYPE_USER; - new_mr->dm = NULL; - new_mr->sig_attrs = NULL; new_mr->uobject = uobj; atomic_inc(&new_pd->usecnt); - new_mr->iova = cmd.hca_va; new_uobj->object = new_mr; rdma_restrack_new(&new_mr->res, RDMA_RESTRACK_MR); @@ -864,8 +879,10 @@ static int ib_uverbs_rereg_mr(struct uverbs_attr_bundle *attrs) mr->pd = new_pd; atomic_inc(&new_pd->usecnt); } - if (cmd.flags & IB_MR_REREG_TRANS) + if (cmd.flags & IB_MR_REREG_TRANS) { mr->iova = cmd.hca_va; + mr->length = cmd.length; + } } memset(&resp, 0, sizeof(resp)); @@ -919,8 +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; } @@ -1051,7 +1068,7 @@ static int create_cq(struct uverbs_attr_bundle *attrs, rdma_restrack_new(&cq->res, RDMA_RESTRACK_CQ); rdma_restrack_set_name(&cq->res, NULL); - ret = ib_dev->ops.create_cq(cq, &attr, &attrs->driver_udata); + ret = ib_dev->ops.create_cq(cq, &attr, attrs); if (ret) goto err_free; rdma_restrack_add(&cq->res); @@ -1127,8 +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) @@ -1189,8 +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; @@ -1238,8 +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); @@ -1295,9 +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: @@ -1321,8 +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; } @@ -1360,8 +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; } } @@ -1371,23 +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; } @@ -1402,7 +1427,6 @@ static int create_qp(struct uverbs_attr_bundle *attrs, attr.sq_sig_type = cmd->sq_sig_all ? IB_SIGNAL_ALL_WR : IB_SIGNAL_REQ_WR; attr.qp_type = cmd->qp_type; - attr.create_flags = 0; attr.cap.max_send_wr = cmd->max_send_wr; attr.cap.max_recv_wr = cmd->max_recv_wr; @@ -1427,7 +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; } @@ -1435,35 +1459,13 @@ static int create_qp(struct uverbs_attr_bundle *attrs, attr.source_qpn = cmd->source_qpn; } - if (cmd->qp_type == IB_QPT_XRC_TGT) - qp = ib_create_qp(pd, &attr); - else - qp = _ib_create_qp(device, pd, &attr, &attrs->driver_udata, obj, - NULL); - + qp = ib_create_qp_user(device, pd, &attr, &attrs->driver_udata, obj, + KBUILD_MODNAME); if (IS_ERR(qp)) { ret = PTR_ERR(qp); goto err_put; } - - if (cmd->qp_type != IB_QPT_XRC_TGT) { - ret = ib_create_qp_security(qp, device); - if (ret) - goto err_cb; - - atomic_inc(&pd->usecnt); - if (attr.send_cq) - atomic_inc(&attr.send_cq->usecnt); - if (attr.recv_cq) - atomic_inc(&attr.recv_cq->usecnt); - if (attr.srq) - atomic_inc(&attr.srq->usecnt); - if (ind_tbl) - atomic_inc(&ind_tbl->usecnt); - } else { - /* It is done in _ib_create_qp for other QP types */ - qp->uobject = obj; - } + ib_qp_usecnt_inc(qp); obj->uevent.uobject.object = qp; obj->uevent.event_file = READ_ONCE(attrs->ufile->default_async_file); @@ -1502,24 +1504,21 @@ static int create_qp(struct uverbs_attr_bundle *attrs, resp.response_length = uverbs_response_length(attrs, sizeof(resp)); return uverbs_response(attrs, &resp, sizeof(resp)); -err_cb: - ib_destroy_qp_user(qp, uverbs_get_cleared_udata(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); @@ -1681,8 +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; } @@ -1787,8 +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; } @@ -1876,8 +1875,15 @@ static int modify_qp(struct uverbs_attr_bundle *attrs, attr->path_mtu = cmd->base.path_mtu; if (cmd->base.attr_mask & IB_QP_PATH_MIG_STATE) attr->path_mig_state = cmd->base.path_mig_state; - if (cmd->base.attr_mask & IB_QP_QKEY) + if (cmd->base.attr_mask & IB_QP_QKEY) { + if (cmd->base.qkey & IB_QP_SET_QKEY && + !(rdma_nl_get_privileged_qkey() || + rdma_uattrs_has_raw_cap(attrs))) { + ret = -EPERM; + goto release_qp; + } attr->qkey = cmd->base.qkey; + } if (cmd->base.attr_mask & IB_QP_RQ_PSN) attr->rq_psn = cmd->base.rq_psn; if (cmd->base.attr_mask & IB_QP_SQ_PSN) @@ -2030,11 +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); @@ -2046,8 +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; } @@ -2084,9 +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; @@ -2220,11 +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); @@ -2323,8 +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; } @@ -2374,8 +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; } @@ -2431,8 +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; } @@ -2501,8 +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; @@ -2551,8 +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); @@ -2686,8 +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, @@ -2704,8 +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, @@ -2757,7 +2765,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, switch (ib_spec->type & ~IB_FLOW_SPEC_INNER) { case IB_FLOW_SPEC_ETH: - ib_filter_sz = offsetof(struct ib_flow_eth_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_eth_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2768,7 +2776,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, memcpy(&ib_spec->eth.mask, kern_spec_mask, actual_filter_sz); break; case IB_FLOW_SPEC_IPV4: - ib_filter_sz = offsetof(struct ib_flow_ipv4_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_ipv4_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2779,7 +2787,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, memcpy(&ib_spec->ipv4.mask, kern_spec_mask, actual_filter_sz); break; case IB_FLOW_SPEC_IPV6: - ib_filter_sz = offsetof(struct ib_flow_ipv6_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_ipv6_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2795,7 +2803,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, break; case IB_FLOW_SPEC_TCP: case IB_FLOW_SPEC_UDP: - ib_filter_sz = offsetof(struct ib_flow_tcp_udp_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_tcp_udp_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2806,7 +2814,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, memcpy(&ib_spec->tcp_udp.mask, kern_spec_mask, actual_filter_sz); break; case IB_FLOW_SPEC_VXLAN_TUNNEL: - ib_filter_sz = offsetof(struct ib_flow_tunnel_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_tunnel_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2821,7 +2829,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, return -EINVAL; break; case IB_FLOW_SPEC_ESP: - ib_filter_sz = offsetof(struct ib_flow_esp_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_esp_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2832,7 +2840,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, memcpy(&ib_spec->esp.mask, kern_spec_mask, actual_filter_sz); break; case IB_FLOW_SPEC_GRE: - ib_filter_sz = offsetof(struct ib_flow_gre_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_gre_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2843,7 +2851,7 @@ int ib_uverbs_kern_spec_to_ib_spec_filter(enum ib_flow_spec_type type, memcpy(&ib_spec->gre.mask, kern_spec_mask, actual_filter_sz); break; case IB_FLOW_SPEC_MPLS: - ib_filter_sz = offsetof(struct ib_flow_mpls_filter, real_sz); + ib_filter_sz = sizeof(struct ib_flow_mpls_filter); actual_filter_sz = spec_filter_size(kern_spec_mask, kern_filter_sz, ib_filter_sz); @@ -2923,14 +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; } @@ -3031,8 +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; @@ -3115,8 +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; } @@ -3218,7 +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) @@ -3271,8 +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; } @@ -3418,15 +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; } @@ -3533,8 +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; @@ -3561,8 +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); @@ -3687,8 +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_ioctl.c b/drivers/infiniband/core/uverbs_ioctl.c index 990f0724acc6..f80da6a67e24 100644 --- a/drivers/infiniband/core/uverbs_ioctl.c +++ b/drivers/infiniband/core/uverbs_ioctl.c @@ -36,13 +36,15 @@ #include "uverbs.h" struct bundle_alloc_head { - struct bundle_alloc_head *next; + struct_group_tagged(bundle_alloc_head_hdr, hdr, + struct bundle_alloc_head *next; + ); u8 data[]; }; struct bundle_priv { /* Must be first */ - struct bundle_alloc_head alloc_head; + struct bundle_alloc_head_hdr alloc_head; struct bundle_alloc_head *allocated_mem; size_t internal_avail; size_t internal_used; @@ -64,7 +66,7 @@ struct bundle_priv { * Must be last. bundle ends in a flex array which overlaps * internal_buffer. */ - struct uverbs_attr_bundle bundle; + struct uverbs_attr_bundle_hdr bundle; u64 internal_buffer[32]; }; @@ -77,9 +79,10 @@ void uapi_compute_bundle_size(struct uverbs_api_ioctl_method *method_elm, unsigned int num_attrs) { struct bundle_priv *pbundle; + struct uverbs_attr_bundle *bundle; size_t bundle_size = offsetof(struct bundle_priv, internal_buffer) + - sizeof(*pbundle->bundle.attrs) * method_elm->key_bitmap_len + + sizeof(*bundle->attrs) * method_elm->key_bitmap_len + sizeof(*pbundle->uattrs) * num_attrs; method_elm->use_stack = bundle_size <= sizeof(*pbundle); @@ -107,7 +110,7 @@ __malloc void *_uverbs_alloc(struct uverbs_attr_bundle *bundle, size_t size, gfp_t flags) { struct bundle_priv *pbundle = - container_of(bundle, struct bundle_priv, bundle); + container_of(&bundle->hdr, struct bundle_priv, bundle); size_t new_used; void *res; @@ -149,7 +152,7 @@ static int uverbs_set_output(const struct uverbs_attr_bundle *bundle, const struct uverbs_attr *attr) { struct bundle_priv *pbundle = - container_of(bundle, struct bundle_priv, bundle); + container_of(&bundle->hdr, struct bundle_priv, bundle); u16 flags; flags = pbundle->uattrs[attr->ptr_attr.uattr_idx].flags | @@ -166,6 +169,8 @@ static int uverbs_process_idrs_array(struct bundle_priv *pbundle, struct ib_uverbs_attr *uattr, u32 attr_bkey) { + struct uverbs_attr_bundle *bundle = + container_of(&pbundle->bundle, struct uverbs_attr_bundle, hdr); const struct uverbs_attr_spec *spec = &attr_uapi->spec; size_t array_len; u32 *idr_vals; @@ -184,7 +189,7 @@ static int uverbs_process_idrs_array(struct bundle_priv *pbundle, return -EINVAL; attr->uobjects = - uverbs_alloc(&pbundle->bundle, + uverbs_alloc(bundle, array_size(array_len, sizeof(*attr->uobjects))); if (IS_ERR(attr->uobjects)) return PTR_ERR(attr->uobjects); @@ -209,7 +214,7 @@ static int uverbs_process_idrs_array(struct bundle_priv *pbundle, for (i = 0; i != array_len; i++) { attr->uobjects[i] = uverbs_get_uobject_from_file( spec->u2.objs_arr.obj_type, spec->u2.objs_arr.access, - idr_vals[i], &pbundle->bundle); + idr_vals[i], bundle); if (IS_ERR(attr->uobjects[i])) { ret = PTR_ERR(attr->uobjects[i]); break; @@ -240,7 +245,9 @@ static int uverbs_process_attr(struct bundle_priv *pbundle, struct ib_uverbs_attr *uattr, u32 attr_bkey) { const struct uverbs_attr_spec *spec = &attr_uapi->spec; - struct uverbs_attr *e = &pbundle->bundle.attrs[attr_bkey]; + struct uverbs_attr_bundle *bundle = + container_of(&pbundle->bundle, struct uverbs_attr_bundle, hdr); + struct uverbs_attr *e = &bundle->attrs[attr_bkey]; const struct uverbs_attr_spec *val_spec = spec; struct uverbs_obj_attr *o_attr; @@ -288,7 +295,7 @@ static int uverbs_process_attr(struct bundle_priv *pbundle, if (val_spec->alloc_and_copy && !uverbs_attr_ptr_is_inline(e)) { void *p; - p = uverbs_alloc(&pbundle->bundle, uattr->len); + p = uverbs_alloc(bundle, uattr->len); if (IS_ERR(p)) return PTR_ERR(p); @@ -321,7 +328,7 @@ static int uverbs_process_attr(struct bundle_priv *pbundle, */ o_attr->uobject = uverbs_get_uobject_from_file( spec->u.obj.obj_type, spec->u.obj.access, - uattr->data_s64, &pbundle->bundle); + uattr->data_s64, bundle); if (IS_ERR(o_attr->uobject)) return PTR_ERR(o_attr->uobject); __set_bit(attr_bkey, pbundle->uobj_finalize); @@ -337,6 +344,14 @@ static int uverbs_process_attr(struct bundle_priv *pbundle, break; + case UVERBS_ATTR_TYPE_RAW_FD: + if (uattr->attr_data.reserved || uattr->len != 0 || + uattr->data_s64 < INT_MIN || uattr->data_s64 > INT_MAX) + return -EINVAL; + /* _uverbs_get_const_signed() is the accessor */ + e->ptr_attr.data = uattr->data_s64; + break; + case UVERBS_ATTR_TYPE_IDRS_ARRAY: return uverbs_process_idrs_array(pbundle, attr_uapi, &e->objs_arr_attr, uattr, @@ -414,6 +429,8 @@ static int ib_uverbs_run_method(struct bundle_priv *pbundle, unsigned int num_attrs) { int (*handler)(struct uverbs_attr_bundle *attrs); + struct uverbs_attr_bundle *bundle = + container_of(&pbundle->bundle, struct uverbs_attr_bundle, hdr); size_t uattrs_size = array_size(sizeof(*pbundle->uattrs), num_attrs); unsigned int destroy_bkey = pbundle->method_elm->destroy_bkey; unsigned int i; @@ -426,7 +443,7 @@ static int ib_uverbs_run_method(struct bundle_priv *pbundle, if (!handler) return -EIO; - pbundle->uattrs = uverbs_alloc(&pbundle->bundle, uattrs_size); + pbundle->uattrs = uverbs_alloc(bundle, uattrs_size); if (IS_ERR(pbundle->uattrs)) return PTR_ERR(pbundle->uattrs); if (copy_from_user(pbundle->uattrs, pbundle->user_attrs, uattrs_size)) @@ -445,25 +462,23 @@ static int ib_uverbs_run_method(struct bundle_priv *pbundle, return -EINVAL; if (pbundle->method_elm->has_udata) - uverbs_fill_udata(&pbundle->bundle, - &pbundle->bundle.driver_udata, + uverbs_fill_udata(bundle, &pbundle->bundle.driver_udata, UVERBS_ATTR_UHW_IN, UVERBS_ATTR_UHW_OUT); else pbundle->bundle.driver_udata = (struct ib_udata){}; if (destroy_bkey != UVERBS_API_ATTR_BKEY_LEN) { - struct uverbs_obj_attr *destroy_attr = - &pbundle->bundle.attrs[destroy_bkey].obj_attr; + struct uverbs_obj_attr *destroy_attr = &bundle->attrs[destroy_bkey].obj_attr; - ret = uobj_destroy(destroy_attr->uobject, &pbundle->bundle); + ret = uobj_destroy(destroy_attr->uobject, bundle); if (ret) return ret; __clear_bit(destroy_bkey, pbundle->uobj_finalize); - ret = handler(&pbundle->bundle); + ret = handler(bundle); uobj_put_destroy(destroy_attr->uobject); } else { - ret = handler(&pbundle->bundle); + ret = handler(bundle); } /* @@ -473,10 +488,10 @@ static int ib_uverbs_run_method(struct bundle_priv *pbundle, */ if (!ret && pbundle->method_elm->has_udata) { const struct uverbs_attr *attr = - uverbs_attr_get(&pbundle->bundle, UVERBS_ATTR_UHW_OUT); + uverbs_attr_get(bundle, UVERBS_ATTR_UHW_OUT); if (!IS_ERR(attr)) - ret = uverbs_set_output(&pbundle->bundle, attr); + ret = uverbs_set_output(bundle, attr); } /* @@ -493,6 +508,8 @@ static int ib_uverbs_run_method(struct bundle_priv *pbundle, static void bundle_destroy(struct bundle_priv *pbundle, bool commit) { unsigned int key_bitmap_len = pbundle->method_elm->key_bitmap_len; + struct uverbs_attr_bundle *bundle = + container_of(&pbundle->bundle, struct uverbs_attr_bundle, hdr); struct bundle_alloc_head *memblock; unsigned int i; @@ -500,20 +517,19 @@ static void bundle_destroy(struct bundle_priv *pbundle, bool commit) i = -1; while ((i = find_next_bit(pbundle->uobj_finalize, key_bitmap_len, i + 1)) < key_bitmap_len) { - struct uverbs_attr *attr = &pbundle->bundle.attrs[i]; + struct uverbs_attr *attr = &bundle->attrs[i]; uverbs_finalize_object( attr->obj_attr.uobject, attr->obj_attr.attr_elm->spec.u.obj.access, test_bit(i, pbundle->uobj_hw_obj_valid), - commit, - &pbundle->bundle); + commit, bundle); } i = -1; while ((i = find_next_bit(pbundle->spec_finalize, key_bitmap_len, i + 1)) < key_bitmap_len) { - struct uverbs_attr *attr = &pbundle->bundle.attrs[i]; + struct uverbs_attr *attr = &bundle->attrs[i]; const struct uverbs_api_attr *attr_uapi; void __rcu **slot; @@ -527,7 +543,7 @@ static void bundle_destroy(struct bundle_priv *pbundle, bool commit) if (attr_uapi->spec.type == UVERBS_ATTR_TYPE_IDRS_ARRAY) { uverbs_free_idrs_array(attr_uapi, &attr->objs_arr_attr, - commit, &pbundle->bundle); + commit, bundle); } } @@ -570,7 +586,8 @@ static int ib_uverbs_cmd_verbs(struct ib_uverbs_file *ufile, method_elm->bundle_size - offsetof(struct bundle_priv, internal_buffer); pbundle->alloc_head.next = NULL; - pbundle->allocated_mem = &pbundle->alloc_head; + pbundle->allocated_mem = container_of(&pbundle->alloc_head, + struct bundle_alloc_head, hdr); } else { pbundle = &onstack; pbundle->internal_avail = sizeof(pbundle->internal_buffer); @@ -588,8 +605,9 @@ static int ib_uverbs_cmd_verbs(struct ib_uverbs_file *ufile, pbundle->user_attrs = user_attrs; pbundle->internal_used = ALIGN(pbundle->method_elm->key_bitmap_len * - sizeof(*pbundle->bundle.attrs), - sizeof(*pbundle->internal_buffer)); + sizeof(*container_of(&pbundle->bundle, + struct uverbs_attr_bundle, hdr)->attrs), + sizeof(*pbundle->internal_buffer)); memset(pbundle->bundle.attr_present, 0, sizeof(pbundle->bundle.attr_present)); memset(pbundle->uobj_finalize, 0, sizeof(pbundle->uobj_finalize)); @@ -692,11 +710,13 @@ void uverbs_fill_udata(struct uverbs_attr_bundle *bundle, unsigned int attr_out) { struct bundle_priv *pbundle = - container_of(bundle, struct bundle_priv, bundle); + container_of(&bundle->hdr, struct bundle_priv, bundle); + struct uverbs_attr_bundle *bundle_aux = + container_of(&pbundle->bundle, struct uverbs_attr_bundle, hdr); const struct uverbs_attr *in = - uverbs_attr_get(&pbundle->bundle, attr_in); + uverbs_attr_get(bundle_aux, attr_in); const struct uverbs_attr *out = - uverbs_attr_get(&pbundle->bundle, attr_out); + uverbs_attr_get(bundle_aux, attr_out); if (!IS_ERR(in)) { udata->inlen = in->ptr_attr.len; @@ -821,7 +841,7 @@ void uverbs_finalize_uobj_create(const struct uverbs_attr_bundle *bundle, u16 idx) { struct bundle_priv *pbundle = - container_of(bundle, struct bundle_priv, bundle); + container_of(&bundle->hdr, struct bundle_priv, bundle); __set_bit(uapi_bkey_attr(uapi_key_attr(idx)), pbundle->uobj_hw_obj_valid); diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c index d54434088727..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" @@ -72,11 +73,23 @@ enum { #define IB_UVERBS_BASE_DEV MKDEV(IB_UVERBS_MAJOR, IB_UVERBS_BASE_MINOR) static dev_t dynamic_uverbs_dev; -static struct class *uverbs_class; static DEFINE_IDA(uverbs_ida); static int ib_uverbs_add_one(struct ib_device *device); static void ib_uverbs_remove_one(struct ib_device *device, void *client_data); +static struct ib_client uverbs_client; + +static char *uverbs_devnode(const struct device *dev, umode_t *mode) +{ + if (mode) + *mode = 0666; + return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev)); +} + +static const struct class uverbs_class = { + .name = "infiniband_verbs", + .devnode = uverbs_devnode, +}; /* * Must be called with the ufile->device->disassociate_srcu held, and the lock @@ -206,6 +219,7 @@ void ib_uverbs_release_file(struct kref *ref) if (file->disassociate_page) __free_pages(file->disassociate_page, 0); + mutex_destroy(&file->disassociation_lock); mutex_destroy(&file->umap_lock); mutex_destroy(&file->ucontext_lock); kfree(file); @@ -222,8 +236,12 @@ static ssize_t ib_uverbs_event_read(struct ib_uverbs_event_queue *ev_queue, spin_lock_irq(&ev_queue->lock); while (list_empty(&ev_queue->event_list)) { - spin_unlock_irq(&ev_queue->lock); + if (ev_queue->is_closed) { + spin_unlock_irq(&ev_queue->lock); + return -EIO; + } + spin_unlock_irq(&ev_queue->lock); if (filp->f_flags & O_NONBLOCK) return -EAGAIN; @@ -233,12 +251,6 @@ static ssize_t ib_uverbs_event_read(struct ib_uverbs_event_queue *ev_queue, return -ERESTARTSYS; spin_lock_irq(&ev_queue->lock); - - /* If device was disassociated and no event exists set an error */ - if (list_empty(&ev_queue->event_list) && ev_queue->is_closed) { - spin_unlock_irq(&ev_queue->lock); - return -EIO; - } } event = list_entry(ev_queue->event_list.next, struct ib_uverbs_event, list); @@ -344,7 +356,6 @@ const struct file_operations uverbs_event_fops = { .poll = ib_uverbs_comp_event_poll, .release = uverbs_uobject_fd_release, .fasync = ib_uverbs_comp_event_fasync, - .llseek = no_llseek, }; const struct file_operations uverbs_async_event_fops = { @@ -353,7 +364,6 @@ const struct file_operations uverbs_async_event_fops = { .poll = ib_uverbs_async_event_poll, .release = uverbs_async_event_release, .fasync = ib_uverbs_async_event_fasync, - .llseek = no_llseek, }; void ib_uverbs_comp_handler(struct ib_cq *cq, void *cq_context) @@ -537,7 +547,7 @@ static ssize_t verify_hdr(struct ib_uverbs_cmd_hdr *hdr, if (hdr->in_words * 4 != count) return -EINVAL; - if (count < method_elm->req_size + sizeof(hdr)) { + if (count < method_elm->req_size + sizeof(*hdr)) { /* * rdma-core v18 and v19 have a bug where they send DESTROY_CQ * with a 16 byte write instead of 24. Old kernels didn't @@ -691,8 +701,13 @@ static int ib_uverbs_mmap(struct file *filp, struct vm_area_struct *vma) ret = PTR_ERR(ucontext); goto out; } + + mutex_lock(&file->disassociation_lock); + vma->vm_ops = &rdma_umap_ops; ret = ucontext->device->ops.mmap(ucontext, vma); + + mutex_unlock(&file->disassociation_lock); out: srcu_read_unlock(&file->device->disassociate_srcu, srcu_key); return ret; @@ -714,6 +729,8 @@ static void rdma_umap_open(struct vm_area_struct *vma) /* We are racing with disassociation */ if (!down_read_trylock(&ufile->hw_destroy_rwsem)) goto out_zap; + mutex_lock(&ufile->disassociation_lock); + /* * Disassociation already completed, the VMA should already be zapped. */ @@ -725,10 +742,12 @@ static void rdma_umap_open(struct vm_area_struct *vma) goto out_unlock; rdma_umap_priv_init(priv, vma, opriv->entry); + mutex_unlock(&ufile->disassociation_lock); up_read(&ufile->hw_destroy_rwsem); return; out_unlock: + mutex_unlock(&ufile->disassociation_lock); up_read(&ufile->hw_destroy_rwsem); out_zap: /* @@ -812,7 +831,7 @@ void uverbs_user_mmap_disassociate(struct ib_uverbs_file *ufile) { struct rdma_umap_priv *priv, *next_priv; - lockdep_assert_held(&ufile->hw_destroy_rwsem); + mutex_lock(&ufile->disassociation_lock); while (1) { struct mm_struct *mm = NULL; @@ -838,8 +857,10 @@ void uverbs_user_mmap_disassociate(struct ib_uverbs_file *ufile) break; } mutex_unlock(&ufile->umap_lock); - if (!mm) + if (!mm) { + mutex_unlock(&ufile->disassociation_lock); return; + } /* * The umap_lock is nested under mmap_lock since it used within @@ -869,7 +890,31 @@ void uverbs_user_mmap_disassociate(struct ib_uverbs_file *ufile) mmap_read_unlock(mm); mmput(mm); } + + mutex_unlock(&ufile->disassociation_lock); +} + +/** + * rdma_user_mmap_disassociate() - Revoke mmaps for a device + * @device: device to revoke + * + * This function should be called by drivers that need to disable mmaps for the + * device, for instance because it is going to be reset. + */ +void rdma_user_mmap_disassociate(struct ib_device *device) +{ + struct ib_uverbs_device *uverbs_dev = + ib_get_client_data(device, &uverbs_client); + struct ib_uverbs_file *ufile; + + mutex_lock(&uverbs_dev->lists_mutex); + list_for_each_entry(ufile, &uverbs_dev->uverbs_file_list, list) { + if (ufile->ucontext) + uverbs_user_mmap_disassociate(ufile); + } + mutex_unlock(&uverbs_dev->lists_mutex); } +EXPORT_SYMBOL(rdma_user_mmap_disassociate); /* * ib_uverbs_open() does not need the BKL: @@ -940,6 +985,8 @@ static int ib_uverbs_open(struct inode *inode, struct file *filp) mutex_init(&file->umap_lock); INIT_LIST_HEAD(&file->umaps); + mutex_init(&file->disassociation_lock); + filp->private_data = file; list_add_tail(&file->list, &dev->uverbs_file_list); mutex_unlock(&dev->lists_mutex); @@ -982,7 +1029,6 @@ static const struct file_operations uverbs_fops = { .write = ib_uverbs_write, .open = ib_uverbs_open, .release = ib_uverbs_close, - .llseek = no_llseek, .unlocked_ioctl = ib_uverbs_ioctl, .compat_ioctl = compat_ptr_ioctl, }; @@ -993,7 +1039,6 @@ static const struct file_operations uverbs_mmap_fops = { .mmap = ib_uverbs_mmap, .open = ib_uverbs_open, .release = ib_uverbs_close, - .llseek = no_llseek, .unlocked_ioctl = ib_uverbs_ioctl, .compat_ioctl = compat_ptr_ioctl, }; @@ -1105,7 +1150,8 @@ static int ib_uverbs_add_one(struct ib_device *device) struct ib_uverbs_device *uverbs_dev; int ret; - if (!device->ops.alloc_ucontext) + if (!device->ops.alloc_ucontext || + device->type == RDMA_DEVICE_TYPE_SMI) return -EOPNOTSUPP; uverbs_dev = kzalloc(sizeof(*uverbs_dev), GFP_KERNEL); @@ -1119,7 +1165,7 @@ static int ib_uverbs_add_one(struct ib_device *device) } device_initialize(&uverbs_dev->dev); - uverbs_dev->dev.class = uverbs_class; + uverbs_dev->dev.class = &uverbs_class; uverbs_dev->dev.parent = device->dev.parent; uverbs_dev->dev.release = ib_uverbs_release_dev; uverbs_dev->groups[0] = &dev_attr_group; @@ -1237,13 +1283,6 @@ static void ib_uverbs_remove_one(struct ib_device *device, void *client_data) put_device(&uverbs_dev->dev); } -static char *uverbs_devnode(struct device *dev, umode_t *mode) -{ - if (mode) - *mode = 0666; - return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev)); -} - static int __init ib_uverbs_init(void) { int ret; @@ -1264,16 +1303,13 @@ static int __init ib_uverbs_init(void) goto out_alloc; } - uverbs_class = class_create(THIS_MODULE, "infiniband_verbs"); - if (IS_ERR(uverbs_class)) { - ret = PTR_ERR(uverbs_class); + ret = class_register(&uverbs_class); + if (ret) { pr_err("user_verbs: couldn't create class infiniband_verbs\n"); goto out_chrdev; } - uverbs_class->devnode = uverbs_devnode; - - ret = class_create_file(uverbs_class, &class_attr_abi_version.attr); + ret = class_create_file(&uverbs_class, &class_attr_abi_version.attr); if (ret) { pr_err("user_verbs: couldn't create abi_version attribute\n"); goto out_class; @@ -1288,7 +1324,7 @@ static int __init ib_uverbs_init(void) return 0; out_class: - class_destroy(uverbs_class); + class_unregister(&uverbs_class); out_chrdev: unregister_chrdev_region(dynamic_uverbs_dev, @@ -1305,11 +1341,12 @@ out: static void __exit ib_uverbs_cleanup(void) { ib_unregister_client(&uverbs_client); - class_destroy(uverbs_class); + class_unregister(&uverbs_class); unregister_chrdev_region(IB_UVERBS_BASE_DEV, 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 b8d715c68ca4..e803f609ec87 100644 --- a/drivers/infiniband/core/uverbs_marshall.c +++ b/drivers/infiniband/core/uverbs_marshall.c @@ -66,7 +66,7 @@ void ib_copy_ah_attr_to_user(struct ib_device *device, struct rdma_ah_attr *src = ah_attr; struct rdma_ah_attr conv_ah; - memset(&dst->grh.reserved, 0, sizeof(dst->grh.reserved)); + memset(&dst->grh, 0, sizeof(dst->grh)); if ((ah_attr->type == RDMA_AH_ATTR_TYPE_OPA) && (rdma_ah_get_dlid(ah_attr) > be16_to_cpu(IB_LID_PERMISSIVE)) && @@ -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_counters.c b/drivers/infiniband/core/uverbs_std_types_counters.c index 999da9c79866..381aa5797641 100644 --- a/drivers/infiniband/core/uverbs_std_types_counters.c +++ b/drivers/infiniband/core/uverbs_std_types_counters.c @@ -107,6 +107,8 @@ static int UVERBS_HANDLER(UVERBS_METHOD_COUNTERS_READ)( return ret; uattr = uverbs_attr_get(attrs, UVERBS_ATTR_READ_COUNTERS_BUFF); + if (IS_ERR(uattr)) + return PTR_ERR(uattr); read_attr.ncounters = uattr->ptr_attr.len / sizeof(u64); read_attr.counters_buff = uverbs_zalloc( attrs, array_size(read_attr.ncounters, sizeof(u64))); diff --git a/drivers/infiniband/core/uverbs_std_types_cq.c b/drivers/infiniband/core/uverbs_std_types_cq.c index 370ad7c83f88..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->driver_udata); + 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 049684880ae0..c0fd283d9d6c 100644 --- a/drivers/infiniband/core/uverbs_std_types_device.c +++ b/drivers/infiniband/core/uverbs_std_types_device.c @@ -203,6 +203,7 @@ static int UVERBS_HANDLER(UVERBS_METHOD_QUERY_PORT)( copy_port_attr_to_resp(&attr, &resp.legacy_resp, ib_dev, port_num); resp.port_cap_flags2 = attr.port_cap_flags2; + resp.active_speed_ex = attr.active_speed; return uverbs_copy_to_struct_or_zero(attrs, UVERBS_ATTR_QUERY_PORT_RESP, &resp, sizeof(resp)); @@ -436,6 +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( @@ -461,7 +466,7 @@ DECLARE_UVERBS_NAMED_METHOD( UVERBS_ATTR_PTR_OUT( UVERBS_ATTR_QUERY_PORT_RESP, UVERBS_ATTR_STRUCT(struct ib_uverbs_query_port_resp_ex, - reserved), + active_speed_ex), UA_MANDATORY)); 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_flow_action.c b/drivers/infiniband/core/uverbs_std_types_flow_action.c index d42ed7ff223e..0ddcf6da66c4 100644 --- a/drivers/infiniband/core/uverbs_std_types_flow_action.c +++ b/drivers/infiniband/core/uverbs_std_types_flow_action.c @@ -46,385 +46,6 @@ static int uverbs_free_flow_action(struct ib_uobject *uobject, return action->device->ops.destroy_flow_action(action); } -static u64 esp_flags_uverbs_to_verbs(struct uverbs_attr_bundle *attrs, - u32 flags, bool is_modify) -{ - u64 verbs_flags = flags; - - if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_FLOW_ACTION_ESP_ESN)) - verbs_flags |= IB_FLOW_ACTION_ESP_FLAGS_ESN_TRIGGERED; - - if (is_modify && uverbs_attr_is_valid(attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_ATTRS)) - verbs_flags |= IB_FLOW_ACTION_ESP_FLAGS_MOD_ESP_ATTRS; - - return verbs_flags; -}; - -static int validate_flow_action_esp_keymat_aes_gcm(struct ib_flow_action_attrs_esp_keymats *keymat) -{ - struct ib_uverbs_flow_action_esp_keymat_aes_gcm *aes_gcm = - &keymat->keymat.aes_gcm; - - if (aes_gcm->iv_algo > IB_UVERBS_FLOW_ACTION_IV_ALGO_SEQ) - return -EOPNOTSUPP; - - if (aes_gcm->key_len != 32 && - aes_gcm->key_len != 24 && - aes_gcm->key_len != 16) - return -EINVAL; - - if (aes_gcm->icv_len != 16 && - aes_gcm->icv_len != 8 && - aes_gcm->icv_len != 12) - return -EINVAL; - - return 0; -} - -static int (* const flow_action_esp_keymat_validate[])(struct ib_flow_action_attrs_esp_keymats *keymat) = { - [IB_UVERBS_FLOW_ACTION_ESP_KEYMAT_AES_GCM] = validate_flow_action_esp_keymat_aes_gcm, -}; - -static int flow_action_esp_replay_none(struct ib_flow_action_attrs_esp_replays *replay, - bool is_modify) -{ - /* This is used in order to modify an esp flow action with an enabled - * replay protection to a disabled one. This is only supported via - * modify, as in create verb we can simply drop the REPLAY attribute and - * achieve the same thing. - */ - return is_modify ? 0 : -EINVAL; -} - -static int flow_action_esp_replay_def_ok(struct ib_flow_action_attrs_esp_replays *replay, - bool is_modify) -{ - /* Some replay protections could always be enabled without validating - * anything. - */ - return 0; -} - -static int (* const flow_action_esp_replay_validate[])(struct ib_flow_action_attrs_esp_replays *replay, - bool is_modify) = { - [IB_UVERBS_FLOW_ACTION_ESP_REPLAY_NONE] = flow_action_esp_replay_none, - [IB_UVERBS_FLOW_ACTION_ESP_REPLAY_BMP] = flow_action_esp_replay_def_ok, -}; - -static int parse_esp_ip(enum ib_flow_spec_type proto, - const void __user *val_ptr, - size_t len, union ib_flow_spec *out) -{ - int ret; - const struct ib_uverbs_flow_ipv4_filter ipv4 = { - .src_ip = cpu_to_be32(0xffffffffUL), - .dst_ip = cpu_to_be32(0xffffffffUL), - .proto = 0xff, - .tos = 0xff, - .ttl = 0xff, - .flags = 0xff, - }; - const struct ib_uverbs_flow_ipv6_filter ipv6 = { - .src_ip = {0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, - 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff}, - .dst_ip = {0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, - 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff}, - .flow_label = cpu_to_be32(0xffffffffUL), - .next_hdr = 0xff, - .traffic_class = 0xff, - .hop_limit = 0xff, - }; - union { - struct ib_uverbs_flow_ipv4_filter ipv4; - struct ib_uverbs_flow_ipv6_filter ipv6; - } user_val = {}; - const void *user_pmask; - size_t val_len; - - /* If the flow IPv4/IPv6 flow specifications are extended, the mask - * should be changed as well. - */ - BUILD_BUG_ON(offsetof(struct ib_uverbs_flow_ipv4_filter, flags) + - sizeof(ipv4.flags) != sizeof(ipv4)); - BUILD_BUG_ON(offsetof(struct ib_uverbs_flow_ipv6_filter, reserved) + - sizeof(ipv6.reserved) != sizeof(ipv6)); - - switch (proto) { - case IB_FLOW_SPEC_IPV4: - if (len > sizeof(user_val.ipv4) && - !ib_is_buffer_cleared(val_ptr + sizeof(user_val.ipv4), - len - sizeof(user_val.ipv4))) - return -EOPNOTSUPP; - - val_len = min_t(size_t, len, sizeof(user_val.ipv4)); - ret = copy_from_user(&user_val.ipv4, val_ptr, - val_len); - if (ret) - return -EFAULT; - - user_pmask = &ipv4; - break; - case IB_FLOW_SPEC_IPV6: - if (len > sizeof(user_val.ipv6) && - !ib_is_buffer_cleared(val_ptr + sizeof(user_val.ipv6), - len - sizeof(user_val.ipv6))) - return -EOPNOTSUPP; - - val_len = min_t(size_t, len, sizeof(user_val.ipv6)); - ret = copy_from_user(&user_val.ipv6, val_ptr, - val_len); - if (ret) - return -EFAULT; - - user_pmask = &ipv6; - break; - default: - return -EOPNOTSUPP; - } - - return ib_uverbs_kern_spec_to_ib_spec_filter(proto, user_pmask, - &user_val, - val_len, out); -} - -static int flow_action_esp_get_encap(struct ib_flow_spec_list *out, - struct uverbs_attr_bundle *attrs) -{ - struct ib_uverbs_flow_action_esp_encap uverbs_encap; - int ret; - - ret = uverbs_copy_from(&uverbs_encap, attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_ENCAP); - if (ret) - return ret; - - /* We currently support only one encap */ - if (uverbs_encap.next_ptr) - return -EOPNOTSUPP; - - if (uverbs_encap.type != IB_FLOW_SPEC_IPV4 && - uverbs_encap.type != IB_FLOW_SPEC_IPV6) - return -EOPNOTSUPP; - - return parse_esp_ip(uverbs_encap.type, - u64_to_user_ptr(uverbs_encap.val_ptr), - uverbs_encap.len, - &out->spec); -} - -struct ib_flow_action_esp_attr { - struct ib_flow_action_attrs_esp hdr; - struct ib_flow_action_attrs_esp_keymats keymat; - struct ib_flow_action_attrs_esp_replays replay; - /* We currently support only one spec */ - struct ib_flow_spec_list encap; -}; - -#define ESP_LAST_SUPPORTED_FLAG IB_UVERBS_FLOW_ACTION_ESP_FLAGS_ESN_NEW_WINDOW -static int parse_flow_action_esp(struct ib_device *ib_dev, - struct uverbs_attr_bundle *attrs, - struct ib_flow_action_esp_attr *esp_attr, - bool is_modify) -{ - struct ib_uverbs_flow_action_esp uverbs_esp = {}; - int ret; - - /* Optional param, if it doesn't exist, we get -ENOENT and skip it */ - ret = uverbs_copy_from(&esp_attr->hdr.esn, attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_ESN); - if (IS_UVERBS_COPY_ERR(ret)) - return ret; - - /* This can be called from FLOW_ACTION_ESP_MODIFY where - * UVERBS_ATTR_FLOW_ACTION_ESP_ATTRS is optional - */ - if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_FLOW_ACTION_ESP_ATTRS)) { - ret = uverbs_copy_from_or_zero(&uverbs_esp, attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_ATTRS); - if (ret) - return ret; - - if (uverbs_esp.flags & ~((ESP_LAST_SUPPORTED_FLAG << 1) - 1)) - return -EOPNOTSUPP; - - esp_attr->hdr.spi = uverbs_esp.spi; - esp_attr->hdr.seq = uverbs_esp.seq; - esp_attr->hdr.tfc_pad = uverbs_esp.tfc_pad; - esp_attr->hdr.hard_limit_pkts = uverbs_esp.hard_limit_pkts; - } - esp_attr->hdr.flags = esp_flags_uverbs_to_verbs(attrs, uverbs_esp.flags, - is_modify); - - if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_FLOW_ACTION_ESP_KEYMAT)) { - esp_attr->keymat.protocol = - uverbs_attr_get_enum_id(attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_KEYMAT); - ret = uverbs_copy_from_or_zero(&esp_attr->keymat.keymat, - attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_KEYMAT); - if (ret) - return ret; - - ret = flow_action_esp_keymat_validate[esp_attr->keymat.protocol](&esp_attr->keymat); - if (ret) - return ret; - - esp_attr->hdr.keymat = &esp_attr->keymat; - } - - if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_FLOW_ACTION_ESP_REPLAY)) { - esp_attr->replay.protocol = - uverbs_attr_get_enum_id(attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_REPLAY); - - ret = uverbs_copy_from_or_zero(&esp_attr->replay.replay, - attrs, - UVERBS_ATTR_FLOW_ACTION_ESP_REPLAY); - if (ret) - return ret; - - ret = flow_action_esp_replay_validate[esp_attr->replay.protocol](&esp_attr->replay, - is_modify); - if (ret) - return ret; - - esp_attr->hdr.replay = &esp_attr->replay; - } - - if (uverbs_attr_is_valid(attrs, UVERBS_ATTR_FLOW_ACTION_ESP_ENCAP)) { - ret = flow_action_esp_get_encap(&esp_attr->encap, attrs); - if (ret) - return ret; - - esp_attr->hdr.encap = &esp_attr->encap; - } - - return 0; -} - -static int UVERBS_HANDLER(UVERBS_METHOD_FLOW_ACTION_ESP_CREATE)( - struct uverbs_attr_bundle *attrs) -{ - struct ib_uobject *uobj = uverbs_attr_get_uobject( - attrs, UVERBS_ATTR_CREATE_FLOW_ACTION_ESP_HANDLE); - struct ib_device *ib_dev = attrs->context->device; - int ret; - struct ib_flow_action *action; - struct ib_flow_action_esp_attr esp_attr = {}; - - if (!ib_dev->ops.create_flow_action_esp) - return -EOPNOTSUPP; - - ret = parse_flow_action_esp(ib_dev, attrs, &esp_attr, false); - if (ret) - return ret; - - /* No need to check as this attribute is marked as MANDATORY */ - action = ib_dev->ops.create_flow_action_esp(ib_dev, &esp_attr.hdr, - attrs); - if (IS_ERR(action)) - return PTR_ERR(action); - - uverbs_flow_action_fill_action(action, uobj, ib_dev, - IB_FLOW_ACTION_ESP); - - return 0; -} - -static int UVERBS_HANDLER(UVERBS_METHOD_FLOW_ACTION_ESP_MODIFY)( - struct uverbs_attr_bundle *attrs) -{ - struct ib_uobject *uobj = uverbs_attr_get_uobject( - attrs, UVERBS_ATTR_MODIFY_FLOW_ACTION_ESP_HANDLE); - struct ib_flow_action *action = uobj->object; - int ret; - struct ib_flow_action_esp_attr esp_attr = {}; - - if (!action->device->ops.modify_flow_action_esp) - return -EOPNOTSUPP; - - ret = parse_flow_action_esp(action->device, attrs, &esp_attr, true); - if (ret) - return ret; - - if (action->type != IB_FLOW_ACTION_ESP) - return -EINVAL; - - return action->device->ops.modify_flow_action_esp(action, - &esp_attr.hdr, - attrs); -} - -static const struct uverbs_attr_spec uverbs_flow_action_esp_keymat[] = { - [IB_UVERBS_FLOW_ACTION_ESP_KEYMAT_AES_GCM] = { - .type = UVERBS_ATTR_TYPE_PTR_IN, - UVERBS_ATTR_STRUCT( - struct ib_uverbs_flow_action_esp_keymat_aes_gcm, - aes_key), - }, -}; - -static const struct uverbs_attr_spec uverbs_flow_action_esp_replay[] = { - [IB_UVERBS_FLOW_ACTION_ESP_REPLAY_NONE] = { - .type = UVERBS_ATTR_TYPE_PTR_IN, - UVERBS_ATTR_NO_DATA(), - }, - [IB_UVERBS_FLOW_ACTION_ESP_REPLAY_BMP] = { - .type = UVERBS_ATTR_TYPE_PTR_IN, - UVERBS_ATTR_STRUCT(struct ib_uverbs_flow_action_esp_replay_bmp, - size), - }, -}; - -DECLARE_UVERBS_NAMED_METHOD( - UVERBS_METHOD_FLOW_ACTION_ESP_CREATE, - UVERBS_ATTR_IDR(UVERBS_ATTR_CREATE_FLOW_ACTION_ESP_HANDLE, - UVERBS_OBJECT_FLOW_ACTION, - UVERBS_ACCESS_NEW, - UA_MANDATORY), - UVERBS_ATTR_PTR_IN(UVERBS_ATTR_FLOW_ACTION_ESP_ATTRS, - UVERBS_ATTR_STRUCT(struct ib_uverbs_flow_action_esp, - hard_limit_pkts), - UA_MANDATORY), - UVERBS_ATTR_PTR_IN(UVERBS_ATTR_FLOW_ACTION_ESP_ESN, - UVERBS_ATTR_TYPE(__u32), - UA_OPTIONAL), - UVERBS_ATTR_ENUM_IN(UVERBS_ATTR_FLOW_ACTION_ESP_KEYMAT, - uverbs_flow_action_esp_keymat, - UA_MANDATORY), - UVERBS_ATTR_ENUM_IN(UVERBS_ATTR_FLOW_ACTION_ESP_REPLAY, - uverbs_flow_action_esp_replay, - UA_OPTIONAL), - UVERBS_ATTR_PTR_IN( - UVERBS_ATTR_FLOW_ACTION_ESP_ENCAP, - UVERBS_ATTR_TYPE(struct ib_uverbs_flow_action_esp_encap), - UA_OPTIONAL)); - -DECLARE_UVERBS_NAMED_METHOD( - UVERBS_METHOD_FLOW_ACTION_ESP_MODIFY, - UVERBS_ATTR_IDR(UVERBS_ATTR_MODIFY_FLOW_ACTION_ESP_HANDLE, - UVERBS_OBJECT_FLOW_ACTION, - UVERBS_ACCESS_WRITE, - UA_MANDATORY), - UVERBS_ATTR_PTR_IN(UVERBS_ATTR_FLOW_ACTION_ESP_ATTRS, - UVERBS_ATTR_STRUCT(struct ib_uverbs_flow_action_esp, - hard_limit_pkts), - UA_OPTIONAL), - UVERBS_ATTR_PTR_IN(UVERBS_ATTR_FLOW_ACTION_ESP_ESN, - UVERBS_ATTR_TYPE(__u32), - UA_OPTIONAL), - UVERBS_ATTR_ENUM_IN(UVERBS_ATTR_FLOW_ACTION_ESP_KEYMAT, - uverbs_flow_action_esp_keymat, - UA_OPTIONAL), - UVERBS_ATTR_ENUM_IN(UVERBS_ATTR_FLOW_ACTION_ESP_REPLAY, - uverbs_flow_action_esp_replay, - UA_OPTIONAL), - UVERBS_ATTR_PTR_IN( - UVERBS_ATTR_FLOW_ACTION_ESP_ENCAP, - UVERBS_ATTR_TYPE(struct ib_uverbs_flow_action_esp_encap), - UA_OPTIONAL)); - DECLARE_UVERBS_NAMED_METHOD_DESTROY( UVERBS_METHOD_FLOW_ACTION_DESTROY, UVERBS_ATTR_IDR(UVERBS_ATTR_DESTROY_FLOW_ACTION_HANDLE, @@ -435,9 +56,7 @@ DECLARE_UVERBS_NAMED_METHOD_DESTROY( DECLARE_UVERBS_NAMED_OBJECT( UVERBS_OBJECT_FLOW_ACTION, UVERBS_TYPE_ALLOC_IDR(uverbs_free_flow_action), - &UVERBS_METHOD(UVERBS_METHOD_FLOW_ACTION_ESP_CREATE), - &UVERBS_METHOD(UVERBS_METHOD_FLOW_ACTION_DESTROY), - &UVERBS_METHOD(UVERBS_METHOD_FLOW_ACTION_ESP_MODIFY)); + &UVERBS_METHOD(UVERBS_METHOD_FLOW_ACTION_DESTROY)); const struct uapi_definition uverbs_def_obj_flow_action[] = { UAPI_DEF_CHAIN_OBJ_TREE_NAMED( diff --git a/drivers/infiniband/core/uverbs_std_types_mr.c b/drivers/infiniband/core/uverbs_std_types_mr.c index f782d5e1aa25..570b9656801d 100644 --- a/drivers/infiniband/core/uverbs_std_types_mr.c +++ b/drivers/infiniband/core/uverbs_std_types_mr.c @@ -238,8 +238,8 @@ 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, - &attrs->driver_udata); + access_flags, NULL, + attrs); if (IS_ERR(mr)) return PTR_ERR(mr); @@ -249,6 +249,9 @@ static int UVERBS_HANDLER(UVERBS_METHOD_REG_DMABUF_MR)( mr->uobject = uobj; atomic_inc(&pd->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_DMABUF_MR_HANDLE); @@ -263,6 +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, @@ -359,6 +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, @@ -373,7 +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 c00cfb5ed387..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: @@ -163,7 +163,7 @@ static int UVERBS_HANDLER(UVERBS_METHOD_QP_CREATE)( UVERBS_ATTR_CREATE_QP_SRQ_HANDLE)) return -EINVAL; - /* send_cq is optinal */ + /* send_cq is optional */ if (cap.max_send_wr) { send_cq = uverbs_attr_get_obj(attrs, UVERBS_ATTR_CREATE_QP_SEND_CQ_HANDLE); @@ -248,44 +248,23 @@ static int UVERBS_HANDLER(UVERBS_METHOD_QP_CREATE)( set_caps(&attr, &cap, true); mutex_init(&obj->mcast_lock); - if (attr.qp_type == IB_QPT_XRC_TGT) - qp = ib_create_qp(pd, &attr); - else - qp = _ib_create_qp(device, pd, &attr, &attrs->driver_udata, obj, - NULL); - + qp = ib_create_qp_user(device, pd, &attr, &attrs->driver_udata, obj, + KBUILD_MODNAME); if (IS_ERR(qp)) { ret = PTR_ERR(qp); goto err_put; } + ib_qp_usecnt_inc(qp); - if (attr.qp_type != IB_QPT_XRC_TGT) { - atomic_inc(&pd->usecnt); - if (attr.send_cq) - atomic_inc(&attr.send_cq->usecnt); - if (attr.recv_cq) - atomic_inc(&attr.recv_cq->usecnt); - if (attr.srq) - atomic_inc(&attr.srq->usecnt); - if (attr.rwq_ind_tbl) - atomic_inc(&attr.rwq_ind_tbl->usecnt); - } else { + if (attr.qp_type == IB_QPT_XRC_TGT) { obj->uxrcd = container_of(xrcd_uobj, struct ib_uxrcd_object, uobject); atomic_inc(&obj->uxrcd->refcnt); - /* It is done in _ib_create_qp for other QP types */ - qp->uobject = obj; } obj->uevent.uobject.object = qp; uverbs_finalize_uobj_create(attrs, UVERBS_ATTR_CREATE_QP_HANDLE); - if (attr.qp_type != IB_QPT_XRC_TGT) { - ret = ib_create_qp_security(qp, device); - if (ret) - return ret; - } - set_caps(&attr, &cap, false); ret = uverbs_copy_to_struct_or_zero(attrs, UVERBS_ATTR_CREATE_QP_RESP_CAP, &cap, diff --git a/drivers/infiniband/core/uverbs_uapi.c b/drivers/infiniband/core/uverbs_uapi.c index 2f2c7646fce1..e00ea63175bd 100644 --- a/drivers/infiniband/core/uverbs_uapi.c +++ b/drivers/infiniband/core/uverbs_uapi.c @@ -447,6 +447,9 @@ static int uapi_finalize(struct uverbs_api *uapi) uapi->num_write_ex = max_write_ex + 1; data = kmalloc_array(uapi->num_write + uapi->num_write_ex, sizeof(*uapi->write_methods), GFP_KERNEL); + if (!data) + return -ENOMEM; + for (i = 0; i != uapi->num_write + uapi->num_write_ex; i++) data[i] = &uapi->notsupp_method; uapi->write_methods = data; @@ -628,6 +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 7036967e4c0b..11b1a194de44 100644 --- a/drivers/infiniband/core/verbs.c +++ b/drivers/infiniband/core/verbs.c @@ -147,6 +147,8 @@ __attribute_const__ int ib_rate_to_mult(enum ib_rate rate) case IB_RATE_50_GBPS: return 20; 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; } } @@ -176,6 +178,8 @@ __attribute_const__ enum ib_rate mult_to_ib_rate(int mult) case 20: return IB_RATE_50_GBPS; 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; } } @@ -205,6 +209,8 @@ __attribute_const__ int ib_rate_to_mbps(enum ib_rate rate) case IB_RATE_50_GBPS: return 53125; 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; } } @@ -268,9 +274,6 @@ struct ib_pd *__ib_alloc_pd(struct ib_device *device, unsigned int flags, return ERR_PTR(-ENOMEM); pd->device = device; - pd->uobject = NULL; - pd->__internal_mr = NULL; - atomic_set(&pd->usecnt, 0); pd->flags = flags; rdma_restrack_new(&pd->res, RDMA_RESTRACK_PD); @@ -284,7 +287,7 @@ struct ib_pd *__ib_alloc_pd(struct ib_device *device, unsigned int flags, } rdma_restrack_add(&pd->res); - if (device->attrs.device_cap_flags & IB_DEVICE_LOCAL_DMA_LKEY) + if (device->attrs.kernel_cap_flags & IBK_LOCAL_DMA_LKEY) pd->local_dma_lkey = device->local_dma_lkey; else mr_access_flags |= IB_ACCESS_LOCAL_WRITE; @@ -311,7 +314,7 @@ struct ib_pd *__ib_alloc_pd(struct ib_device *device, unsigned int flags, pd->__internal_mr = mr; - if (!(device->attrs.device_cap_flags & IB_DEVICE_LOCAL_DMA_LKEY)) + if (!(device->attrs.kernel_cap_flags & IBK_LOCAL_DMA_LKEY)) pd->local_dma_lkey = pd->__internal_mr->lkey; if (flags & IB_PD_UNSAFE_GLOBAL_RKEY) @@ -341,11 +344,6 @@ int ib_dealloc_pd_user(struct ib_pd *pd, struct ib_udata *udata) pd->__internal_mr = NULL; } - /* uverbs manipulates usecnt with proper locking, while the kabi - * requires the caller to guarantee we can't race here. - */ - WARN_ON(atomic_read(&pd->usecnt)); - ret = pd->device->ops.dealloc_pd(pd, udata); if (ret) return ret; @@ -374,7 +372,7 @@ void rdma_copy_ah_attr(struct rdma_ah_attr *dest, EXPORT_SYMBOL(rdma_copy_ah_attr); /** - * rdma_replace_ah_attr - Replace valid ah_attr with new new one. + * rdma_replace_ah_attr - Replace valid ah_attr with new one. * @old: Pointer to existing ah_attr which needs to be replaced. * old is assumed to be valid or zero'd * @new: Pointer to the new ah_attr. @@ -540,6 +538,8 @@ static struct ib_ah *_rdma_create_ah(struct ib_pd *pd, else ret = device->ops.create_ah(ah, &init_attr, NULL); if (ret) { + if (ah->sgid_attr) + rdma_put_gid_attr(ah->sgid_attr); kfree(ah); return ERR_PTR(ret); } @@ -575,7 +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); @@ -750,7 +750,7 @@ EXPORT_SYMBOL(ib_get_gids_from_rdma_hdr); /* Resolve destination mac address and hop limit for unicast destination * GID entry, considering the source GID entry as well. - * ah_attribute must have have valid port_num, sgid_index. + * ah_attribute must have valid port_num, sgid_index. */ static int ib_resolve_unicast_gid_dmac(struct ib_device *device, struct rdma_ah_attr *ah_attr) @@ -1035,7 +1035,8 @@ struct ib_srq *ib_create_srq_user(struct ib_pd *pd, } if (srq->srq_type == IB_SRQT_XRC) { srq->ext.xrc.xrcd = srq_init_attr->ext.xrc.xrcd; - atomic_inc(&srq->ext.xrc.xrcd->usecnt); + if (srq->ext.xrc.xrcd) + atomic_inc(&srq->ext.xrc.xrcd->usecnt); } atomic_inc(&pd->usecnt); @@ -1045,8 +1046,8 @@ struct ib_srq *ib_create_srq_user(struct ib_pd *pd, ret = pd->device->ops.create_srq(srq, srq_init_attr, udata); if (ret) { rdma_restrack_put(&srq->res); - atomic_dec(&srq->pd->usecnt); - if (srq->srq_type == IB_SRQT_XRC) + atomic_dec(&pd->usecnt); + if (srq->srq_type == IB_SRQT_XRC && srq->ext.xrc.xrcd) atomic_dec(&srq->ext.xrc.xrcd->usecnt); if (ib_srq_has_cq(srq->srq_type)) atomic_dec(&srq->ext.cq->usecnt); @@ -1090,7 +1091,7 @@ int ib_destroy_srq_user(struct ib_srq *srq, struct ib_udata *udata) return ret; atomic_dec(&srq->pd->usecnt); - if (srq->srq_type == IB_SRQT_XRC) + if (srq->srq_type == IB_SRQT_XRC && srq->ext.xrc.xrcd) atomic_dec(&srq->ext.xrc.xrcd->usecnt); if (ib_srq_has_cq(srq->srq_type)) atomic_dec(&srq->ext.cq->usecnt); @@ -1103,6 +1104,16 @@ EXPORT_SYMBOL(ib_destroy_srq_user); /* Queue pairs */ +static void __ib_qp_event_handler(struct ib_event *event, void *context) +{ + struct ib_qp *qp = event->element.qp; + + if (event->event == IB_EVENT_QP_LAST_WQE_REACHED) + complete(&qp->srq_completion); + if (qp->registered_event_handler) + qp->registered_event_handler(event, qp->qp_context); +} + static void __ib_shared_qp_event_handler(struct ib_event *event, void *context) { struct ib_qp *qp = context; @@ -1199,34 +1210,147 @@ static struct ib_qp *create_xrc_qp_user(struct ib_qp *qp, return qp; } +static struct ib_qp *create_qp(struct ib_device *dev, struct ib_pd *pd, + struct ib_qp_init_attr *attr, + struct ib_udata *udata, + struct ib_uqp_object *uobj, const char *caller) +{ + struct ib_udata dummy = {}; + struct ib_qp *qp; + int ret; + + if (!dev->ops.create_qp) + return ERR_PTR(-EOPNOTSUPP); + + qp = rdma_zalloc_drv_obj_numa(dev, ib_qp); + if (!qp) + return ERR_PTR(-ENOMEM); + + qp->device = dev; + qp->pd = pd; + qp->uobject = uobj; + qp->real_qp = qp; + + qp->qp_type = attr->qp_type; + qp->rwq_ind_tbl = attr->rwq_ind_tbl; + qp->srq = attr->srq; + qp->event_handler = __ib_qp_event_handler; + qp->registered_event_handler = attr->event_handler; + qp->port = attr->port_num; + qp->qp_context = attr->qp_context; + + spin_lock_init(&qp->mr_lock); + INIT_LIST_HEAD(&qp->rdma_mrs); + INIT_LIST_HEAD(&qp->sig_mrs); + init_completion(&qp->srq_completion); + + qp->send_cq = attr->send_cq; + qp->recv_cq = attr->recv_cq; + + rdma_restrack_new(&qp->res, RDMA_RESTRACK_QP); + WARN_ONCE(!udata && !caller, "Missing kernel QP owner"); + rdma_restrack_set_name(&qp->res, udata ? NULL : caller); + ret = dev->ops.create_qp(qp, attr, udata); + if (ret) + goto err_create; + + /* + * TODO: The mlx4 internally overwrites send_cq and recv_cq. + * Unfortunately, it is not an easy task to fix that driver. + */ + qp->send_cq = attr->send_cq; + qp->recv_cq = attr->recv_cq; + + ret = ib_create_qp_security(qp, dev); + if (ret) + goto err_security; + + rdma_restrack_add(&qp->res); + return qp; + +err_security: + qp->device->ops.destroy_qp(qp, udata ? &dummy : NULL); +err_create: + rdma_restrack_put(&qp->res); + kfree(qp); + return ERR_PTR(ret); + +} + /** - * ib_create_named_qp - Creates a kernel QP associated with the specified protection + * ib_create_qp_user - Creates a QP associated with the specified protection * domain. + * @dev: IB device * @pd: The protection domain associated with the QP. - * @qp_init_attr: A list of initial attributes required to create the + * @attr: A list of initial attributes required to create the * QP. If QP creation succeeds, then the attributes are updated to * the actual capabilities of the created QP. + * @udata: User data + * @uobj: uverbs obect * @caller: caller's build-time module name - * - * NOTE: for user qp use ib_create_qp_user with valid udata! */ -struct ib_qp *ib_create_named_qp(struct ib_pd *pd, - struct ib_qp_init_attr *qp_init_attr, - const char *caller) +struct ib_qp *ib_create_qp_user(struct ib_device *dev, struct ib_pd *pd, + struct ib_qp_init_attr *attr, + struct ib_udata *udata, + struct ib_uqp_object *uobj, const char *caller) { - struct ib_device *device = pd ? pd->device : qp_init_attr->xrcd->device; - struct ib_qp *qp; - int ret; + struct ib_qp *qp, *xrc_qp; - if (qp_init_attr->rwq_ind_tbl && - (qp_init_attr->recv_cq || - qp_init_attr->srq || qp_init_attr->cap.max_recv_wr || - qp_init_attr->cap.max_recv_sge)) - return ERR_PTR(-EINVAL); + if (attr->qp_type == IB_QPT_XRC_TGT) + qp = create_qp(dev, pd, attr, NULL, NULL, caller); + else + qp = create_qp(dev, pd, attr, udata, uobj, NULL); + if (attr->qp_type != IB_QPT_XRC_TGT || IS_ERR(qp)) + return qp; - if ((qp_init_attr->create_flags & IB_QP_CREATE_INTEGRITY_EN) && - !(device->attrs.device_cap_flags & IB_DEVICE_INTEGRITY_HANDOVER)) - return ERR_PTR(-EINVAL); + xrc_qp = create_xrc_qp_user(qp, attr); + if (IS_ERR(xrc_qp)) { + ib_destroy_qp(qp); + return xrc_qp; + } + + xrc_qp->uobject = uobj; + return xrc_qp; +} +EXPORT_SYMBOL(ib_create_qp_user); + +void ib_qp_usecnt_inc(struct ib_qp *qp) +{ + if (qp->pd) + atomic_inc(&qp->pd->usecnt); + if (qp->send_cq) + atomic_inc(&qp->send_cq->usecnt); + if (qp->recv_cq) + atomic_inc(&qp->recv_cq->usecnt); + if (qp->srq) + atomic_inc(&qp->srq->usecnt); + if (qp->rwq_ind_tbl) + atomic_inc(&qp->rwq_ind_tbl->usecnt); +} +EXPORT_SYMBOL(ib_qp_usecnt_inc); + +void ib_qp_usecnt_dec(struct ib_qp *qp) +{ + if (qp->rwq_ind_tbl) + atomic_dec(&qp->rwq_ind_tbl->usecnt); + if (qp->srq) + atomic_dec(&qp->srq->usecnt); + if (qp->recv_cq) + atomic_dec(&qp->recv_cq->usecnt); + if (qp->send_cq) + atomic_dec(&qp->send_cq->usecnt); + if (qp->pd) + atomic_dec(&qp->pd->usecnt); +} +EXPORT_SYMBOL(ib_qp_usecnt_dec); + +struct ib_qp *ib_create_qp_kernel(struct ib_pd *pd, + struct ib_qp_init_attr *qp_init_attr, + const char *caller) +{ + struct ib_device *device = pd->device; + struct ib_qp *qp; + int ret; /* * If the callers is using the RDMA API calculate the resources @@ -1237,47 +1361,11 @@ struct ib_qp *ib_create_named_qp(struct ib_pd *pd, if (qp_init_attr->cap.max_rdma_ctxs) rdma_rw_init_qp(device, qp_init_attr); - qp = _ib_create_qp(device, pd, qp_init_attr, NULL, NULL, caller); + qp = create_qp(device, pd, qp_init_attr, NULL, NULL, caller); if (IS_ERR(qp)) return qp; - ret = ib_create_qp_security(qp, device); - if (ret) - goto err; - - if (qp_init_attr->qp_type == IB_QPT_XRC_TGT) { - struct ib_qp *xrc_qp = - create_xrc_qp_user(qp, qp_init_attr); - - if (IS_ERR(xrc_qp)) { - ret = PTR_ERR(xrc_qp); - goto err; - } - return xrc_qp; - } - - qp->event_handler = qp_init_attr->event_handler; - qp->qp_context = qp_init_attr->qp_context; - if (qp_init_attr->qp_type == IB_QPT_XRC_INI) { - qp->recv_cq = NULL; - qp->srq = NULL; - } else { - qp->recv_cq = qp_init_attr->recv_cq; - if (qp_init_attr->recv_cq) - atomic_inc(&qp_init_attr->recv_cq->usecnt); - qp->srq = qp_init_attr->srq; - if (qp->srq) - atomic_inc(&qp_init_attr->srq->usecnt); - } - - qp->send_cq = qp_init_attr->send_cq; - qp->xrcd = NULL; - - atomic_inc(&pd->usecnt); - if (qp_init_attr->send_cq) - atomic_inc(&qp_init_attr->send_cq->usecnt); - if (qp_init_attr->rwq_ind_tbl) - atomic_inc(&qp->rwq_ind_tbl->usecnt); + ib_qp_usecnt_inc(qp); if (qp_init_attr->cap.max_rdma_ctxs) { ret = rdma_rw_init_mrs(qp, qp_init_attr); @@ -1303,7 +1391,7 @@ err: return ERR_PTR(ret); } -EXPORT_SYMBOL(ib_create_named_qp); +EXPORT_SYMBOL(ib_create_qp_kernel); static const struct { int valid; @@ -1810,12 +1898,95 @@ int ib_modify_qp_with_udata(struct ib_qp *ib_qp, struct ib_qp_attr *attr, } EXPORT_SYMBOL(ib_modify_qp_with_udata); +static void ib_get_width_and_speed(u32 netdev_speed, u32 lanes, + u16 *speed, u8 *width) +{ + if (!lanes) { + if (netdev_speed <= SPEED_1000) { + *width = IB_WIDTH_1X; + *speed = IB_SPEED_SDR; + } else if (netdev_speed <= SPEED_10000) { + *width = IB_WIDTH_1X; + *speed = IB_SPEED_FDR10; + } else if (netdev_speed <= SPEED_20000) { + *width = IB_WIDTH_4X; + *speed = IB_SPEED_DDR; + } else if (netdev_speed <= SPEED_25000) { + *width = IB_WIDTH_1X; + *speed = IB_SPEED_EDR; + } else if (netdev_speed <= SPEED_40000) { + *width = IB_WIDTH_4X; + *speed = IB_SPEED_FDR10; + } else if (netdev_speed <= SPEED_50000) { + *width = IB_WIDTH_2X; + *speed = IB_SPEED_EDR; + } else if (netdev_speed <= SPEED_100000) { + *width = IB_WIDTH_4X; + *speed = IB_SPEED_EDR; + } else if (netdev_speed <= SPEED_200000) { + *width = IB_WIDTH_4X; + *speed = IB_SPEED_HDR; + } else { + *width = IB_WIDTH_4X; + *speed = IB_SPEED_NDR; + } + + return; + } + + switch (lanes) { + case 1: + *width = IB_WIDTH_1X; + break; + case 2: + *width = IB_WIDTH_2X; + break; + case 4: + *width = IB_WIDTH_4X; + break; + case 8: + *width = IB_WIDTH_8X; + break; + case 12: + *width = IB_WIDTH_12X; + break; + default: + *width = IB_WIDTH_1X; + } + + switch (netdev_speed / lanes) { + case SPEED_2500: + *speed = IB_SPEED_SDR; + break; + case SPEED_5000: + *speed = IB_SPEED_DDR; + break; + case SPEED_10000: + *speed = IB_SPEED_FDR10; + break; + case SPEED_14000: + *speed = IB_SPEED_FDR; + break; + case SPEED_25000: + *speed = IB_SPEED_EDR; + break; + case SPEED_50000: + *speed = IB_SPEED_HDR; + break; + case SPEED_100000: + *speed = IB_SPEED_NDR; + break; + default: + *speed = IB_SPEED_SDR; + } +} + int ib_get_eth_speed(struct ib_device *dev, u32 port_num, u16 *speed, u8 *width) { int rc; u32 netdev_speed; struct net_device *netdev; - struct ethtool_link_ksettings lksettings; + struct ethtool_link_ksettings lksettings = {}; if (rdma_port_get_link_layer(dev, port_num) != IB_LINK_LAYER_ETHERNET) return -EINVAL; @@ -1834,29 +2005,13 @@ int ib_get_eth_speed(struct ib_device *dev, u32 port_num, u16 *speed, u8 *width) netdev_speed = lksettings.base.speed; } else { netdev_speed = SPEED_1000; - pr_warn("%s speed is unknown, defaulting to %u\n", netdev->name, - netdev_speed); + if (rc) + pr_warn("%s speed is unknown, defaulting to %u\n", + netdev->name, netdev_speed); } - if (netdev_speed <= SPEED_1000) { - *width = IB_WIDTH_1X; - *speed = IB_SPEED_SDR; - } else if (netdev_speed <= SPEED_10000) { - *width = IB_WIDTH_1X; - *speed = IB_SPEED_FDR10; - } else if (netdev_speed <= SPEED_20000) { - *width = IB_WIDTH_4X; - *speed = IB_SPEED_DDR; - } else if (netdev_speed <= SPEED_25000) { - *width = IB_WIDTH_1X; - *speed = IB_SPEED_EDR; - } else if (netdev_speed <= SPEED_40000) { - *width = IB_WIDTH_4X; - *speed = IB_SPEED_FDR10; - } else { - *width = IB_WIDTH_4X; - *speed = IB_SPEED_EDR; - } + ib_get_width_and_speed(netdev_speed, lksettings.lanes, + speed, width); return 0; } @@ -1935,10 +2090,6 @@ int ib_destroy_qp_user(struct ib_qp *qp, struct ib_udata *udata) { const struct ib_gid_attr *alt_path_sgid_attr = qp->alt_path_sgid_attr; const struct ib_gid_attr *av_sgid_attr = qp->av_sgid_attr; - struct ib_pd *pd; - struct ib_cq *scq, *rcq; - struct ib_srq *srq; - struct ib_rwq_ind_table *ind_tbl; struct ib_qp_security *sec; int ret; @@ -1950,11 +2101,6 @@ int ib_destroy_qp_user(struct ib_qp *qp, struct ib_udata *udata) if (qp->real_qp != qp) return __ib_destroy_shared_qp(qp); - pd = qp->pd; - scq = qp->send_cq; - rcq = qp->recv_cq; - srq = qp->srq; - ind_tbl = qp->rwq_ind_tbl; sec = qp->qp_sec; if (sec) ib_destroy_qp_security_begin(sec); @@ -1962,31 +2108,25 @@ 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_restrack_del(&qp->res); + rdma_counter_unbind_qp(qp, qp->port, true); ret = qp->device->ops.destroy_qp(qp, udata); - if (!ret) { - if (alt_path_sgid_attr) - rdma_put_gid_attr(alt_path_sgid_attr); - if (av_sgid_attr) - rdma_put_gid_attr(av_sgid_attr); - if (pd) - atomic_dec(&pd->usecnt); - if (scq) - atomic_dec(&scq->usecnt); - if (rcq) - atomic_dec(&rcq->usecnt); - if (srq) - atomic_dec(&srq->usecnt); - if (ind_tbl) - atomic_dec(&ind_tbl->usecnt); - if (sec) - ib_destroy_qp_security_end(sec); - } else { + if (ret) { if (sec) ib_destroy_qp_security_abort(sec); + return ret; } + if (alt_path_sgid_attr) + rdma_put_gid_attr(alt_path_sgid_attr); + if (av_sgid_attr) + rdma_put_gid_attr(av_sgid_attr); + + ib_qp_usecnt_dec(qp); + if (sec) + ib_destroy_qp_security_end(sec); + + rdma_restrack_del(&qp->res); + kfree(qp); return ret; } EXPORT_SYMBOL(ib_destroy_qp_user); @@ -2078,23 +2218,26 @@ struct ib_mr *ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length, struct ib_mr *mr; if (access_flags & IB_ACCESS_ON_DEMAND) { - if (!(pd->device->attrs.device_cap_flags & - IB_DEVICE_ON_DEMAND_PAGING)) { + if (!(pd->device->attrs.kernel_cap_flags & + IBK_ON_DEMAND_PAGING)) { pr_debug("ODP support not available\n"); return ERR_PTR(-EINVAL); } } 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; mr->device = pd->device; + mr->type = IB_MR_TYPE_USER; mr->pd = pd; mr->dm = NULL; atomic_inc(&pd->usecnt); + mr->iova = virt_addr; + mr->length = length; rdma_restrack_new(&mr->res, RDMA_RESTRACK_MR); rdma_restrack_parent_name(&mr->res, &pd->res); @@ -2122,6 +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; @@ -2132,6 +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); } @@ -2756,6 +2902,72 @@ static void __ib_drain_rq(struct ib_qp *qp) wait_for_completion(&rdrain.done); } +/* + * __ib_drain_srq() - Block until Last WQE Reached event arrives, or timeout + * expires. + * @qp: queue pair associated with SRQ to drain + * + * Quoting 10.3.1 Queue Pair and EE Context States: + * + * Note, for QPs that are associated with an SRQ, the Consumer should take the + * QP through the Error State before invoking a Destroy QP or a Modify QP to the + * Reset State. The Consumer may invoke the Destroy QP without first performing + * a Modify QP to the Error State and waiting for the Affiliated Asynchronous + * Last WQE Reached Event. However, if the Consumer does not wait for the + * Affiliated Asynchronous Last WQE Reached Event, then WQE and Data Segment + * leakage may occur. Therefore, it is good programming practice to tear down a + * QP that is associated with an SRQ by using the following process: + * + * - Put the QP in the Error State + * - Wait for the Affiliated Asynchronous Last WQE Reached Event; + * - either: + * drain the CQ by invoking the Poll CQ verb and either wait for CQ + * to be empty or the number of Poll CQ operations has exceeded + * CQ capacity size; + * - or + * post another WR that completes on the same CQ and wait for this + * WR to return as a WC; + * - and then invoke a Destroy QP or Reset QP. + * + * We use the first option. + */ +static void __ib_drain_srq(struct ib_qp *qp) +{ + struct ib_qp_attr attr = { .qp_state = IB_QPS_ERR }; + struct ib_cq *cq; + int n, polled = 0; + int ret; + + if (!qp->srq) { + WARN_ONCE(1, "QP 0x%p is not associated with SRQ\n", qp); + return; + } + + ret = ib_modify_qp(qp, &attr, IB_QP_STATE); + if (ret) { + WARN_ONCE(ret, "failed to drain shared recv queue: %d\n", ret); + return; + } + + if (ib_srq_has_cq(qp->srq->srq_type)) { + cq = qp->srq->ext.cq; + } else if (qp->recv_cq) { + cq = qp->recv_cq; + } else { + WARN_ONCE(1, "QP 0x%p has no CQ associated with SRQ\n", qp); + return; + } + + if (wait_for_completion_timeout(&qp->srq_completion, 60 * HZ) > 0) { + while (polled != cq->cqe) { + n = ib_process_cq_direct(cq, cq->cqe - polled); + if (!n) + return; + polled += n; + } + } +} + /** * ib_drain_sq() - Block until all SQ CQEs have been consumed by the * application. @@ -2834,6 +3046,8 @@ void ib_drain_qp(struct ib_qp *qp) ib_drain_sq(qp); if (!qp->srq) ib_drain_rq(qp); + else + __ib_drain_srq(qp); } EXPORT_SYMBOL(ib_drain_qp); @@ -2901,20 +3115,73 @@ EXPORT_SYMBOL(__rdma_block_iter_start); bool __rdma_block_iter_next(struct ib_block_iter *biter) { unsigned int block_offset; + 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); - biter->__sg_advance += BIT_ULL(biter->__pg_bit) - block_offset; + delta = BIT_ULL(biter->__pg_bit) - block_offset; - if (biter->__sg_advance >= sg_dma_len(biter->__sg)) { + 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; } EXPORT_SYMBOL(__rdma_block_iter_next); + +/** + * rdma_alloc_hw_stats_struct - Helper function to allocate dynamic struct + * for the drivers. + * @descs: array of static descriptors + * @num_counters: number of elements in array + * @lifespan: milliseconds between updates + */ +struct rdma_hw_stats *rdma_alloc_hw_stats_struct( + const struct rdma_stat_desc *descs, int num_counters, + unsigned long lifespan) +{ + struct rdma_hw_stats *stats; + + stats = kzalloc(struct_size(stats, value, num_counters), GFP_KERNEL); + if (!stats) + return NULL; + + stats->is_disabled = kcalloc(BITS_TO_LONGS(num_counters), + sizeof(*stats->is_disabled), GFP_KERNEL); + if (!stats->is_disabled) + goto err; + + stats->descs = descs; + stats->num_counters = num_counters; + stats->lifespan = msecs_to_jiffies(lifespan); + mutex_init(&stats->lock); + + return stats; + +err: + kfree(stats); + return NULL; +} +EXPORT_SYMBOL(rdma_alloc_hw_stats_struct); + +/** + * rdma_free_hw_stats_struct - Helper function to release rdma_hw_stats + * @stats: statistics to release + */ +void rdma_free_hw_stats_struct(struct rdma_hw_stats *stats) +{ + if (!stats) + return; + + kfree(stats->is_disabled); + kfree(stats); +} +EXPORT_SYMBOL(rdma_free_hw_stats_struct); |