Index: head/sys/ofed/drivers/infiniband/core/ib_cma.c =================================================================== --- head/sys/ofed/drivers/infiniband/core/ib_cma.c (revision 338525) +++ head/sys/ofed/drivers/infiniband/core/ib_cma.c (revision 338526) @@ -1,4376 +1,4387 @@ /*- * SPDX-License-Identifier: BSD-2-Clause OR GPL-2.0 * * Copyright (c) 2005 Voltaire Inc. All rights reserved. * Copyright (c) 2002-2005, Network Appliance, Inc. All rights reserved. * Copyright (c) 1999-2005, Mellanox Technologies, Inc. All rights reserved. * Copyright (c) 2005-2006 Intel Corporation. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU * General Public License (GPL) Version 2, available from the file * COPYING in the main directory of this source tree, or the * OpenIB.org BSD license below: * * Redistribution and use in source and binary forms, with or * without modification, are permitted provided that the following * conditions are met: * * - Redistributions of source code must retain the above * copyright notice, this list of conditions and the following * disclaimer. * * - Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following * disclaimer in the documentation and/or other materials * provided with the distribution. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. */ #include __FBSDID("$FreeBSD$"); #define LINUXKPI_PARAM_PREFIX ibcore_ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "core_priv.h" MODULE_AUTHOR("Sean Hefty"); MODULE_DESCRIPTION("Generic RDMA CM Agent"); MODULE_LICENSE("Dual BSD/GPL"); #define CMA_CM_RESPONSE_TIMEOUT 20 #define CMA_QUERY_CLASSPORT_INFO_TIMEOUT 3000 #define CMA_MAX_CM_RETRIES 15 #define CMA_CM_MRA_SETTING (IB_CM_MRA_FLAG_DELAY | 24) #define CMA_IBOE_PACKET_LIFETIME 18 static const char * const cma_events[] = { [RDMA_CM_EVENT_ADDR_RESOLVED] = "address resolved", [RDMA_CM_EVENT_ADDR_ERROR] = "address error", [RDMA_CM_EVENT_ROUTE_RESOLVED] = "route resolved ", [RDMA_CM_EVENT_ROUTE_ERROR] = "route error", [RDMA_CM_EVENT_CONNECT_REQUEST] = "connect request", [RDMA_CM_EVENT_CONNECT_RESPONSE] = "connect response", [RDMA_CM_EVENT_CONNECT_ERROR] = "connect error", [RDMA_CM_EVENT_UNREACHABLE] = "unreachable", [RDMA_CM_EVENT_REJECTED] = "rejected", [RDMA_CM_EVENT_ESTABLISHED] = "established", [RDMA_CM_EVENT_DISCONNECTED] = "disconnected", [RDMA_CM_EVENT_DEVICE_REMOVAL] = "device removal", [RDMA_CM_EVENT_MULTICAST_JOIN] = "multicast join", [RDMA_CM_EVENT_MULTICAST_ERROR] = "multicast error", [RDMA_CM_EVENT_ADDR_CHANGE] = "address change", [RDMA_CM_EVENT_TIMEWAIT_EXIT] = "timewait exit", }; const char *__attribute_const__ rdma_event_msg(enum rdma_cm_event_type event) { size_t index = event; return (index < ARRAY_SIZE(cma_events) && cma_events[index]) ? cma_events[index] : "unrecognized event"; } EXPORT_SYMBOL(rdma_event_msg); static int cma_check_linklocal(struct rdma_dev_addr *, struct sockaddr *); static void cma_add_one(struct ib_device *device); static void cma_remove_one(struct ib_device *device, void *client_data); static struct ib_client cma_client = { .name = "cma", .add = cma_add_one, .remove = cma_remove_one }; static struct ib_sa_client sa_client; static struct rdma_addr_client addr_client; static LIST_HEAD(dev_list); static LIST_HEAD(listen_any_list); static DEFINE_MUTEX(lock); static struct workqueue_struct *cma_wq; struct cma_pernet { struct idr tcp_ps; struct idr udp_ps; struct idr ipoib_ps; struct idr ib_ps; }; VNET_DEFINE(struct cma_pernet, cma_pernet); static struct cma_pernet *cma_pernet_ptr(struct vnet *vnet) { struct cma_pernet *retval; CURVNET_SET_QUIET(vnet); retval = &VNET(cma_pernet); CURVNET_RESTORE(); return (retval); } static struct idr *cma_pernet_idr(struct vnet *net, enum rdma_port_space ps) { struct cma_pernet *pernet = cma_pernet_ptr(net); switch (ps) { case RDMA_PS_TCP: return &pernet->tcp_ps; case RDMA_PS_UDP: return &pernet->udp_ps; case RDMA_PS_IPOIB: return &pernet->ipoib_ps; case RDMA_PS_IB: return &pernet->ib_ps; default: return NULL; } } struct cma_device { struct list_head list; struct ib_device *device; struct completion comp; atomic_t refcount; struct list_head id_list; struct sysctl_ctx_list sysctl_ctx; enum ib_gid_type *default_gid_type; }; struct rdma_bind_list { enum rdma_port_space ps; struct hlist_head owners; unsigned short port; }; struct class_port_info_context { struct ib_class_port_info *class_port_info; struct ib_device *device; struct completion done; struct ib_sa_query *sa_query; u8 port_num; }; static int cma_ps_alloc(struct vnet *vnet, enum rdma_port_space ps, struct rdma_bind_list *bind_list, int snum) { struct idr *idr = cma_pernet_idr(vnet, ps); return idr_alloc(idr, bind_list, snum, snum + 1, GFP_KERNEL); } static struct rdma_bind_list *cma_ps_find(struct vnet *net, enum rdma_port_space ps, int snum) { struct idr *idr = cma_pernet_idr(net, ps); return idr_find(idr, snum); } static void cma_ps_remove(struct vnet *net, enum rdma_port_space ps, int snum) { struct idr *idr = cma_pernet_idr(net, ps); idr_remove(idr, snum); } enum { CMA_OPTION_AFONLY, }; void cma_ref_dev(struct cma_device *cma_dev) { atomic_inc(&cma_dev->refcount); } struct cma_device *cma_enum_devices_by_ibdev(cma_device_filter filter, void *cookie) { struct cma_device *cma_dev; struct cma_device *found_cma_dev = NULL; mutex_lock(&lock); list_for_each_entry(cma_dev, &dev_list, list) if (filter(cma_dev->device, cookie)) { found_cma_dev = cma_dev; break; } if (found_cma_dev) cma_ref_dev(found_cma_dev); mutex_unlock(&lock); return found_cma_dev; } int cma_get_default_gid_type(struct cma_device *cma_dev, unsigned int port) { if (port < rdma_start_port(cma_dev->device) || port > rdma_end_port(cma_dev->device)) return -EINVAL; return cma_dev->default_gid_type[port - rdma_start_port(cma_dev->device)]; } int cma_set_default_gid_type(struct cma_device *cma_dev, unsigned int port, enum ib_gid_type default_gid_type) { unsigned long supported_gids; if (port < rdma_start_port(cma_dev->device) || port > rdma_end_port(cma_dev->device)) return -EINVAL; supported_gids = roce_gid_type_mask_support(cma_dev->device, port); if (!(supported_gids & 1 << default_gid_type)) return -EINVAL; cma_dev->default_gid_type[port - rdma_start_port(cma_dev->device)] = default_gid_type; return 0; } struct ib_device *cma_get_ib_dev(struct cma_device *cma_dev) { return cma_dev->device; } /* * Device removal can occur at anytime, so we need extra handling to * serialize notifying the user of device removal with other callbacks. * We do this by disabling removal notification while a callback is in process, * and reporting it after the callback completes. */ struct rdma_id_private { struct rdma_cm_id id; 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 */ struct cma_device *cma_dev; struct list_head mc_list; int internal_id; enum rdma_cm_state state; spinlock_t lock; struct mutex qp_mutex; struct completion comp; atomic_t refcount; struct mutex handler_mutex; int backlog; int timeout_ms; struct ib_sa_query *query; int query_id; union { struct ib_cm_id *ib; struct iw_cm_id *iw; } cm_id; u32 seq_num; u32 qkey; u32 qp_num; pid_t owner; u32 options; u8 srq; u8 tos; u8 reuseaddr; u8 afonly; enum ib_gid_type gid_type; }; struct cma_multicast { struct rdma_id_private *id_priv; union { struct ib_sa_multicast *ib; } multicast; struct list_head list; void *context; struct sockaddr_storage addr; struct kref mcref; bool igmp_joined; u8 join_state; }; struct cma_work { struct work_struct work; struct rdma_id_private *id; enum rdma_cm_state old_state; enum rdma_cm_state new_state; struct rdma_cm_event event; }; struct cma_ndev_work { struct work_struct work; struct rdma_id_private *id; struct rdma_cm_event event; }; struct iboe_mcast_work { struct work_struct work; struct rdma_id_private *id; struct cma_multicast *mc; }; union cma_ip_addr { struct in6_addr ip6; struct { __be32 pad[3]; __be32 addr; } ip4; }; struct cma_hdr { u8 cma_version; u8 ip_version; /* IP version: 7:4 */ __be16 port; union cma_ip_addr src_addr; union cma_ip_addr dst_addr; }; #define CMA_VERSION 0x00 struct cma_req_info { struct ib_device *device; int port; union ib_gid local_gid; __be64 service_id; u16 pkey; bool has_gid:1; }; static int cma_comp(struct rdma_id_private *id_priv, enum rdma_cm_state comp) { unsigned long flags; int ret; spin_lock_irqsave(&id_priv->lock, flags); ret = (id_priv->state == comp); spin_unlock_irqrestore(&id_priv->lock, flags); return ret; } static int cma_comp_exch(struct rdma_id_private *id_priv, enum rdma_cm_state comp, enum rdma_cm_state exch) { unsigned long flags; int ret; spin_lock_irqsave(&id_priv->lock, flags); if ((ret = (id_priv->state == comp))) id_priv->state = exch; spin_unlock_irqrestore(&id_priv->lock, flags); return ret; } static enum rdma_cm_state cma_exch(struct rdma_id_private *id_priv, enum rdma_cm_state exch) { unsigned long flags; enum rdma_cm_state old; spin_lock_irqsave(&id_priv->lock, flags); old = id_priv->state; id_priv->state = exch; spin_unlock_irqrestore(&id_priv->lock, flags); return old; } 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) { hdr->ip_version = (ip_ver << 4) | (hdr->ip_version & 0xF); } static int cma_igmp_send(struct net_device *ndev, const union ib_gid *mgid, bool join) { int retval; if (ndev) { union { struct sockaddr sock; struct sockaddr_storage storage; } addr; rdma_gid2ip(&addr.sock, mgid); CURVNET_SET_QUIET(ndev->if_vnet); if (join) retval = -if_addmulti(ndev, &addr.sock, NULL); else retval = -if_delmulti(ndev, &addr.sock); CURVNET_RESTORE(); } else { retval = -ENODEV; } return retval; } static void _cma_attach_to_dev(struct rdma_id_private *id_priv, struct cma_device *cma_dev) { cma_ref_dev(cma_dev); id_priv->cma_dev = cma_dev; id_priv->gid_type = 0; 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); } static void cma_attach_to_dev(struct rdma_id_private *id_priv, struct cma_device *cma_dev) { _cma_attach_to_dev(id_priv, cma_dev); id_priv->gid_type = cma_dev->default_gid_type[id_priv->id.port_num - rdma_start_port(cma_dev->device)]; } void cma_deref_dev(struct cma_device *cma_dev) { if (atomic_dec_and_test(&cma_dev->refcount)) complete(&cma_dev->comp); } static inline void release_mc(struct kref *kref) { struct cma_multicast *mc = container_of(kref, struct cma_multicast, mcref); kfree(mc->multicast.ib); kfree(mc); } static void cma_release_dev(struct rdma_id_private *id_priv) { mutex_lock(&lock); list_del(&id_priv->list); cma_deref_dev(id_priv->cma_dev); id_priv->cma_dev = NULL; 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) { 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: id_priv->qkey = RDMA_UDP_QKEY; break; case RDMA_PS_IPOIB: ib_addr_get_mgid(&id_priv->id.route.addr.dev_addr, &rec.mgid); ret = ib_sa_get_mcmember_rec(id_priv->id.device, id_priv->id.port_num, &rec.mgid, &rec); if (!ret) id_priv->qkey = be32_to_cpu(rec.qkey); break; default: break; } return ret; } static void cma_translate_ib(struct sockaddr_ib *sib, struct rdma_dev_addr *dev_addr) { dev_addr->dev_type = ARPHRD_INFINIBAND; rdma_addr_set_sgid(dev_addr, (union ib_gid *) &sib->sib_addr); ib_addr_set_pkey(dev_addr, ntohs(sib->sib_pkey)); } static int cma_translate_addr(struct sockaddr *addr, struct rdma_dev_addr *dev_addr) { int ret; if (addr->sa_family != AF_IB) { ret = rdma_translate_ip(addr, dev_addr); } else { cma_translate_ib((struct sockaddr_ib *) addr, dev_addr); ret = 0; } return ret; } static inline int cma_validate_port(struct ib_device *device, u8 port, enum ib_gid_type gid_type, union ib_gid *gid, const struct rdma_dev_addr *dev_addr) { const int dev_type = dev_addr->dev_type; struct net_device *ndev; int ret = -ENODEV; if ((dev_type == ARPHRD_INFINIBAND) && !rdma_protocol_ib(device, port)) return ret; if ((dev_type != ARPHRD_INFINIBAND) && rdma_protocol_ib(device, port)) return ret; if (dev_type == ARPHRD_ETHER && rdma_protocol_roce(device, port)) { ndev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if); } else { ndev = NULL; gid_type = IB_GID_TYPE_IB; } ret = ib_find_cached_gid_by_port(device, gid, gid_type, port, ndev, NULL); if (ndev) dev_put(ndev); return ret; } static int cma_acquire_dev(struct rdma_id_private *id_priv, struct rdma_id_private *listen_id_priv) { struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; struct cma_device *cma_dev; union ib_gid gid, iboe_gid, *gidp; int ret = -ENODEV; u8 port; if (dev_addr->dev_type != ARPHRD_INFINIBAND && id_priv->id.ps == RDMA_PS_IPOIB) return -EINVAL; mutex_lock(&lock); rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.src_addr, &iboe_gid); memcpy(&gid, dev_addr->src_dev_addr + rdma_addr_gid_offset(dev_addr), sizeof gid); if (listen_id_priv) { cma_dev = listen_id_priv->cma_dev; port = listen_id_priv->id.port_num; gidp = rdma_protocol_roce(cma_dev->device, port) ? &iboe_gid : &gid; ret = cma_validate_port(cma_dev->device, port, rdma_protocol_ib(cma_dev->device, port) ? IB_GID_TYPE_IB : listen_id_priv->gid_type, gidp, dev_addr); if (!ret) { id_priv->id.port_num = port; goto out; } } list_for_each_entry(cma_dev, &dev_list, list) { for (port = 1; port <= cma_dev->device->phys_port_cnt; ++port) { if (listen_id_priv && listen_id_priv->cma_dev == cma_dev && listen_id_priv->id.port_num == port) continue; gidp = rdma_protocol_roce(cma_dev->device, port) ? &iboe_gid : &gid; ret = cma_validate_port(cma_dev->device, port, rdma_protocol_ib(cma_dev->device, port) ? IB_GID_TYPE_IB : cma_dev->default_gid_type[port - 1], gidp, dev_addr); if (!ret) { id_priv->id.port_num = port; goto out; } } } out: if (!ret) cma_attach_to_dev(id_priv, cma_dev); mutex_unlock(&lock); return ret; } /* * Select the source IB device and address to reach the destination IB address. */ static int cma_resolve_ib_dev(struct rdma_id_private *id_priv) { struct cma_device *cma_dev, *cur_dev; struct sockaddr_ib *addr; union ib_gid gid, sgid, *dgid; u16 pkey, index; u8 p; int i; cma_dev = NULL; addr = (struct sockaddr_ib *) cma_dst_addr(id_priv); dgid = (union ib_gid *) &addr->sib_addr; pkey = ntohs(addr->sib_pkey); list_for_each_entry(cur_dev, &dev_list, list) { for (p = 1; p <= cur_dev->device->phys_port_cnt; ++p) { if (!rdma_cap_af_ib(cur_dev->device, p)) continue; if (ib_find_cached_pkey(cur_dev->device, p, pkey, &index)) continue; for (i = 0; !ib_get_cached_gid(cur_dev->device, p, i, &gid, NULL); i++) { if (!memcmp(&gid, dgid, sizeof(gid))) { cma_dev = cur_dev; sgid = gid; id_priv->id.port_num = p; goto found; } if (!cma_dev && (gid.global.subnet_prefix == dgid->global.subnet_prefix)) { cma_dev = cur_dev; sgid = gid; id_priv->id.port_num = p; } } } } if (!cma_dev) return -ENODEV; found: cma_attach_to_dev(id_priv, cma_dev); addr = (struct sockaddr_ib *) cma_src_addr(id_priv); memcpy(&addr->sib_addr, &sgid, sizeof sgid); cma_translate_ib(addr, &id_priv->id.route.addr.dev_addr); return 0; } static void cma_deref_id(struct rdma_id_private *id_priv) { if (atomic_dec_and_test(&id_priv->refcount)) complete(&id_priv->comp); } struct rdma_cm_id *rdma_create_id(struct vnet *net, rdma_cm_event_handler event_handler, void *context, enum rdma_port_space ps, enum ib_qp_type qp_type) { struct rdma_id_private *id_priv; id_priv = kzalloc(sizeof *id_priv, GFP_KERNEL); if (!id_priv) return ERR_PTR(-ENOMEM); id_priv->owner = task_pid_nr(current); id_priv->state = RDMA_CM_IDLE; id_priv->id.context = context; id_priv->id.event_handler = event_handler; id_priv->id.ps = ps; id_priv->id.qp_type = qp_type; spin_lock_init(&id_priv->lock); mutex_init(&id_priv->qp_mutex); init_completion(&id_priv->comp); atomic_set(&id_priv->refcount, 1); mutex_init(&id_priv->handler_mutex); 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 = TD_TO_VNET(curthread); return &id_priv->id; } EXPORT_SYMBOL(rdma_create_id); static int cma_init_ud_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; ret = ib_modify_qp(qp, &qp_attr, qp_attr_mask); if (ret) return ret; qp_attr.qp_state = IB_QPS_RTR; ret = ib_modify_qp(qp, &qp_attr, IB_QP_STATE); if (ret) return ret; qp_attr.qp_state = IB_QPS_RTS; qp_attr.sq_psn = 0; ret = ib_modify_qp(qp, &qp_attr, IB_QP_STATE | IB_QP_SQ_PSN); 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; id_priv = container_of(id, struct rdma_id_private, id); if (id->device != pd->device) return -EINVAL; qp_init_attr->port_num = id->port_num; qp = ib_create_qp(pd, qp_init_attr); if (IS_ERR(qp)) return PTR_ERR(qp); 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 err; id->qp = qp; id_priv->qp_num = qp->qp_num; id_priv->srq = (qp->srq != NULL); return 0; err: ib_destroy_qp(qp); return ret; } EXPORT_SYMBOL(rdma_create_qp); void rdma_destroy_qp(struct rdma_cm_id *id) { struct rdma_id_private *id_priv; id_priv = container_of(id, struct rdma_id_private, id); mutex_lock(&id_priv->qp_mutex); ib_destroy_qp(id_priv->id.qp); id_priv->id.qp = NULL; mutex_unlock(&id_priv->qp_mutex); } EXPORT_SYMBOL(rdma_destroy_qp); static int cma_modify_qp_rtr(struct rdma_id_private *id_priv, struct rdma_conn_param *conn_param) { struct ib_qp_attr qp_attr; int qp_attr_mask, ret; union ib_gid sgid; mutex_lock(&id_priv->qp_mutex); if (!id_priv->id.qp) { ret = 0; goto out; } /* Need to update QP attributes from default values. */ qp_attr.qp_state = IB_QPS_INIT; ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); if (ret) goto out; ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask); if (ret) goto out; qp_attr.qp_state = IB_QPS_RTR; ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); if (ret) goto out; ret = ib_query_gid(id_priv->id.device, id_priv->id.port_num, qp_attr.ah_attr.grh.sgid_index, &sgid, NULL); if (ret) goto out; BUG_ON(id_priv->cma_dev->device != id_priv->id.device); if (conn_param) qp_attr.max_dest_rd_atomic = conn_param->responder_resources; ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask); out: mutex_unlock(&id_priv->qp_mutex); return ret; } static int cma_modify_qp_rts(struct rdma_id_private *id_priv, struct rdma_conn_param *conn_param) { struct ib_qp_attr qp_attr; int qp_attr_mask, ret; mutex_lock(&id_priv->qp_mutex); if (!id_priv->id.qp) { ret = 0; goto out; } qp_attr.qp_state = IB_QPS_RTS; ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); if (ret) goto out; if (conn_param) qp_attr.max_rd_atomic = conn_param->initiator_depth; ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask); out: mutex_unlock(&id_priv->qp_mutex); return ret; } static int cma_modify_qp_err(struct rdma_id_private *id_priv) { struct ib_qp_attr qp_attr; int ret; mutex_lock(&id_priv->qp_mutex); if (!id_priv->id.qp) { ret = 0; goto out; } qp_attr.qp_state = IB_QPS_ERR; ret = ib_modify_qp(id_priv->id.qp, &qp_attr, IB_QP_STATE); out: mutex_unlock(&id_priv->qp_mutex); return ret; } static int cma_ib_init_qp_attr(struct rdma_id_private *id_priv, struct ib_qp_attr *qp_attr, int *qp_attr_mask) { struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; int ret; u16 pkey; if (rdma_cap_eth_ah(id_priv->id.device, id_priv->id.port_num)) pkey = 0xffff; else pkey = ib_addr_get_pkey(dev_addr); ret = ib_find_cached_pkey(id_priv->id.device, id_priv->id.port_num, pkey, &qp_attr->pkey_index); if (ret) return ret; qp_attr->port_num = id_priv->id.port_num; *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); if (ret) return ret; qp_attr->qkey = id_priv->qkey; *qp_attr_mask |= IB_QP_QKEY; } else { qp_attr->qp_access_flags = 0; *qp_attr_mask |= IB_QP_ACCESS_FLAGS; } return 0; } int rdma_init_qp_attr(struct rdma_cm_id *id, struct ib_qp_attr *qp_attr, int *qp_attr_mask) { struct rdma_id_private *id_priv; int ret = 0; id_priv = container_of(id, struct rdma_id_private, id); if (rdma_cap_ib_cm(id->device, id->port_num)) { if (!id_priv->cm_id.ib || (id_priv->id.qp_type == IB_QPT_UD)) ret = cma_ib_init_qp_attr(id_priv, qp_attr, qp_attr_mask); else ret = ib_cm_init_qp_attr(id_priv->cm_id.ib, qp_attr, qp_attr_mask); if (qp_attr->qp_state == IB_QPS_RTR) qp_attr->rq_psn = id_priv->seq_num; } else if (rdma_cap_iw_cm(id->device, id->port_num)) { if (!id_priv->cm_id.iw) { qp_attr->qp_access_flags = 0; *qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS; } else ret = iw_cm_init_qp_attr(id_priv->cm_id.iw, qp_attr, qp_attr_mask); qp_attr->port_num = id_priv->id.port_num; *qp_attr_mask |= IB_QP_PORT; } else ret = -ENOSYS; return ret; } EXPORT_SYMBOL(rdma_init_qp_attr); static inline int cma_zero_addr(struct sockaddr *addr) { switch (addr->sa_family) { case AF_INET: return ipv4_is_zeronet(((struct sockaddr_in *)addr)->sin_addr.s_addr); case AF_INET6: return ipv6_addr_any(&((struct sockaddr_in6 *) addr)->sin6_addr); case AF_IB: return ib_addr_any(&((struct sockaddr_ib *) addr)->sib_addr); default: return 0; } } static inline int cma_loopback_addr(struct sockaddr *addr) { switch (addr->sa_family) { case AF_INET: return ipv4_is_loopback(((struct sockaddr_in *) addr)->sin_addr.s_addr); case AF_INET6: return ipv6_addr_loopback(&((struct sockaddr_in6 *) addr)->sin6_addr); case AF_IB: return ib_addr_loopback(&((struct sockaddr_ib *) addr)->sib_addr); default: return 0; } } static inline int cma_any_addr(struct sockaddr *addr) { return cma_zero_addr(addr) || cma_loopback_addr(addr); } static int cma_addr_cmp(struct sockaddr *src, struct sockaddr *dst) { if (src->sa_family != dst->sa_family) return -1; switch (src->sa_family) { case AF_INET: return ((struct sockaddr_in *) src)->sin_addr.s_addr != ((struct sockaddr_in *) dst)->sin_addr.s_addr; case AF_INET6: return ipv6_addr_cmp(&((struct sockaddr_in6 *) src)->sin6_addr, &((struct sockaddr_in6 *) dst)->sin6_addr); default: return ib_addr_cmp(&((struct sockaddr_ib *) src)->sib_addr, &((struct sockaddr_ib *) dst)->sib_addr); } } static __be16 cma_port(struct sockaddr *addr) { struct sockaddr_ib *sib; switch (addr->sa_family) { case AF_INET: return ((struct sockaddr_in *) addr)->sin_port; case AF_INET6: return ((struct sockaddr_in6 *) addr)->sin6_port; case AF_IB: sib = (struct sockaddr_ib *) addr; return htons((u16) (be64_to_cpu(sib->sib_sid) & be64_to_cpu(sib->sib_sid_mask))); default: return 0; } } static inline int cma_any_port(struct sockaddr *addr) { return !cma_port(addr); } static void cma_save_ib_info(struct sockaddr *src_addr, struct sockaddr *dst_addr, struct rdma_cm_id *listen_id, struct ib_sa_path_rec *path) { struct sockaddr_ib *listen_ib, *ib; listen_ib = (struct sockaddr_ib *) &listen_id->route.addr.src_addr; if (src_addr) { ib = (struct sockaddr_ib *)src_addr; ib->sib_family = AF_IB; if (path) { ib->sib_pkey = path->pkey; ib->sib_flowinfo = path->flow_label; memcpy(&ib->sib_addr, &path->sgid, 16); ib->sib_sid = path->service_id; ib->sib_scope_id = 0; } else { ib->sib_pkey = listen_ib->sib_pkey; ib->sib_flowinfo = listen_ib->sib_flowinfo; ib->sib_addr = listen_ib->sib_addr; ib->sib_sid = listen_ib->sib_sid; ib->sib_scope_id = listen_ib->sib_scope_id; } ib->sib_sid_mask = cpu_to_be64(0xffffffffffffffffULL); } if (dst_addr) { ib = (struct sockaddr_ib *)dst_addr; ib->sib_family = AF_IB; if (path) { ib->sib_pkey = path->pkey; ib->sib_flowinfo = path->flow_label; memcpy(&ib->sib_addr, &path->dgid, 16); } } } static void cma_save_ip4_info(struct sockaddr_in *src_addr, struct sockaddr_in *dst_addr, struct cma_hdr *hdr, __be16 local_port) { if (src_addr) { *src_addr = (struct sockaddr_in) { .sin_len = sizeof(struct sockaddr_in), .sin_family = AF_INET, .sin_addr.s_addr = hdr->dst_addr.ip4.addr, .sin_port = local_port, }; } if (dst_addr) { *dst_addr = (struct sockaddr_in) { .sin_len = sizeof(struct sockaddr_in), .sin_family = AF_INET, .sin_addr.s_addr = hdr->src_addr.ip4.addr, .sin_port = hdr->port, }; } } static void cma_ip6_clear_scope_id(struct in6_addr *addr) { /* make sure link local scope ID gets zeroed */ if (IN6_IS_SCOPE_LINKLOCAL(addr) || IN6_IS_ADDR_MC_INTFACELOCAL(addr)) { /* use byte-access to be alignment safe */ addr->s6_addr[2] = 0; addr->s6_addr[3] = 0; } } static void cma_save_ip6_info(struct sockaddr_in6 *src_addr, struct sockaddr_in6 *dst_addr, struct cma_hdr *hdr, __be16 local_port) { if (src_addr) { *src_addr = (struct sockaddr_in6) { .sin6_len = sizeof(struct sockaddr_in6), .sin6_family = AF_INET6, .sin6_addr = hdr->dst_addr.ip6, .sin6_port = local_port, }; cma_ip6_clear_scope_id(&src_addr->sin6_addr); } if (dst_addr) { *dst_addr = (struct sockaddr_in6) { .sin6_len = sizeof(struct sockaddr_in6), .sin6_family = AF_INET6, .sin6_addr = hdr->src_addr.ip6, .sin6_port = hdr->port, }; cma_ip6_clear_scope_id(&dst_addr->sin6_addr); } } static u16 cma_port_from_service_id(__be64 service_id) { return (u16)be64_to_cpu(service_id); } static int cma_save_ip_info(struct sockaddr *src_addr, struct sockaddr *dst_addr, struct ib_cm_event *ib_event, __be64 service_id) { struct cma_hdr *hdr; __be16 port; hdr = ib_event->private_data; if (hdr->cma_version != CMA_VERSION) return -EINVAL; port = htons(cma_port_from_service_id(service_id)); switch (cma_get_ip_ver(hdr)) { case 4: cma_save_ip4_info((struct sockaddr_in *)src_addr, (struct sockaddr_in *)dst_addr, hdr, port); break; case 6: cma_save_ip6_info((struct sockaddr_in6 *)src_addr, (struct sockaddr_in6 *)dst_addr, hdr, port); break; default: return -EAFNOSUPPORT; } return 0; } static int cma_save_net_info(struct sockaddr *src_addr, struct sockaddr *dst_addr, struct rdma_cm_id *listen_id, struct ib_cm_event *ib_event, sa_family_t sa_family, __be64 service_id) { if (sa_family == AF_IB) { if (ib_event->event == IB_CM_REQ_RECEIVED) cma_save_ib_info(src_addr, dst_addr, listen_id, ib_event->param.req_rcvd.primary_path); else if (ib_event->event == IB_CM_SIDR_REQ_RECEIVED) cma_save_ib_info(src_addr, dst_addr, listen_id, NULL); return 0; } return cma_save_ip_info(src_addr, dst_addr, ib_event, service_id); } static int cma_save_req_info(const struct ib_cm_event *ib_event, struct cma_req_info *req) { const struct ib_cm_req_event_param *req_param = &ib_event->param.req_rcvd; const struct ib_cm_sidr_req_event_param *sidr_param = &ib_event->param.sidr_req_rcvd; switch (ib_event->event) { case IB_CM_REQ_RECEIVED: req->device = req_param->listen_id->device; req->port = req_param->port; memcpy(&req->local_gid, &req_param->primary_path->sgid, sizeof(req->local_gid)); req->has_gid = true; req->service_id = req_param->primary_path->service_id; req->pkey = be16_to_cpu(req_param->primary_path->pkey); if (req->pkey != req_param->bth_pkey) pr_warn_ratelimited("RDMA CMA: got different BTH P_Key (0x%x) and primary path P_Key (0x%x)\n" "RDMA CMA: in the future this may cause the request to be dropped\n", req_param->bth_pkey, req->pkey); break; case IB_CM_SIDR_REQ_RECEIVED: req->device = sidr_param->listen_id->device; req->port = sidr_param->port; req->has_gid = false; req->service_id = sidr_param->service_id; req->pkey = sidr_param->pkey; if (req->pkey != sidr_param->bth_pkey) pr_warn_ratelimited("RDMA CMA: got different BTH P_Key (0x%x) and SIDR request payload P_Key (0x%x)\n" "RDMA CMA: in the future this may cause the request to be dropped\n", sidr_param->bth_pkey, req->pkey); break; default: return -EINVAL; } return 0; } static bool validate_ipv4_net_dev(struct net_device *net_dev, const struct sockaddr_in *dst_addr, const struct sockaddr_in *src_addr) { #ifdef INET - struct sockaddr_in dst_tmp = *dst_addr; + struct sockaddr_in src_tmp = *src_addr; __be32 daddr = dst_addr->sin_addr.s_addr, saddr = src_addr->sin_addr.s_addr; - struct net_device *src_dev; + struct net_device *dst_dev; struct rtentry *rte; bool ret; if (ipv4_is_multicast(saddr) || ipv4_is_lbcast(saddr) || ipv4_is_lbcast(daddr) || ipv4_is_zeronet(saddr) || ipv4_is_zeronet(daddr) || ipv4_is_loopback(daddr) || ipv4_is_loopback(saddr)) return false; - src_dev = ip_dev_find(net_dev->if_vnet, saddr); - if (src_dev != net_dev) { - if (src_dev != NULL) - dev_put(src_dev); + dst_dev = ip_dev_find(net_dev->if_vnet, daddr); + if (dst_dev != net_dev) { + if (dst_dev != NULL) + dev_put(dst_dev); return false; } - dev_put(src_dev); + dev_put(dst_dev); /* * Make sure the socket address length field * is set, else rtalloc1() will fail. */ - dst_tmp.sin_len = sizeof(dst_tmp); + src_tmp.sin_len = sizeof(src_tmp); CURVNET_SET(net_dev->if_vnet); - rte = rtalloc1((struct sockaddr *)&dst_tmp, 1, 0); - CURVNET_RESTORE(); + rte = rtalloc1((struct sockaddr *)&src_tmp, 1, 0); if (rte != NULL) { ret = (rte->rt_ifp == net_dev); RTFREE_LOCKED(rte); } else { ret = false; } + CURVNET_RESTORE(); return ret; #else return false; #endif } static bool validate_ipv6_net_dev(struct net_device *net_dev, const struct sockaddr_in6 *dst_addr, const struct sockaddr_in6 *src_addr) { #ifdef INET6 - struct sockaddr_in6 dst_tmp = *dst_addr; - struct in6_addr in6_addr = src_addr->sin6_addr; - struct net_device *src_dev; + struct sockaddr_in6 src_tmp = *src_addr; + struct in6_addr in6_addr = dst_addr->sin6_addr; + struct net_device *dst_dev; struct rtentry *rte; bool ret; - src_dev = ip6_dev_find(net_dev->if_vnet, in6_addr); - if (src_dev != net_dev) + dst_dev = ip6_dev_find(net_dev->if_vnet, in6_addr); + if (dst_dev != net_dev) { + if (dst_dev != NULL) + dev_put(dst_dev); return false; + } + CURVNET_SET(net_dev->if_vnet); + /* * Make sure the socket address length field * is set, else rtalloc1() will fail. */ - dst_tmp.sin6_len = sizeof(dst_tmp); + src_tmp.sin6_len = sizeof(src_tmp); - CURVNET_SET(net_dev->if_vnet); - rte = rtalloc1((struct sockaddr *)&dst_tmp, 1, 0); - CURVNET_RESTORE(); + /* + * Make sure the scope ID gets embedded, else rtalloc1() will + * resolve to the loopback interface. + */ + src_tmp.sin6_scope_id = net_dev->if_index; + sa6_embedscope(&src_tmp, 0); + + rte = rtalloc1((struct sockaddr *)&src_tmp, 1, 0); if (rte != NULL) { ret = (rte->rt_ifp == net_dev); RTFREE_LOCKED(rte); } else { ret = false; } + CURVNET_RESTORE(); return ret; #else return false; #endif } static bool validate_net_dev(struct net_device *net_dev, const struct sockaddr *daddr, const struct sockaddr *saddr) { const struct sockaddr_in *daddr4 = (const struct sockaddr_in *)daddr; const struct sockaddr_in *saddr4 = (const struct sockaddr_in *)saddr; const struct sockaddr_in6 *daddr6 = (const struct sockaddr_in6 *)daddr; const struct sockaddr_in6 *saddr6 = (const struct sockaddr_in6 *)saddr; switch (daddr->sa_family) { case AF_INET: return saddr->sa_family == AF_INET && validate_ipv4_net_dev(net_dev, daddr4, saddr4); case AF_INET6: return saddr->sa_family == AF_INET6 && validate_ipv6_net_dev(net_dev, daddr6, saddr6); default: return false; } } static struct net_device *cma_get_net_dev(struct ib_cm_event *ib_event, const struct cma_req_info *req) { struct sockaddr_storage listen_addr_storage, src_addr_storage; struct sockaddr *listen_addr = (struct sockaddr *)&listen_addr_storage, *src_addr = (struct sockaddr *)&src_addr_storage; struct net_device *net_dev; const union ib_gid *gid = req->has_gid ? &req->local_gid : NULL; int err; err = cma_save_ip_info(listen_addr, src_addr, ib_event, req->service_id); if (err) return ERR_PTR(err); net_dev = ib_get_net_dev_by_params(req->device, req->port, req->pkey, gid, listen_addr); if (!net_dev) return ERR_PTR(-ENODEV); if (!validate_net_dev(net_dev, listen_addr, src_addr)) { dev_put(net_dev); return ERR_PTR(-EHOSTUNREACH); } return net_dev; } static enum rdma_port_space rdma_ps_from_service_id(__be64 service_id) { return (be64_to_cpu(service_id) >> 16) & 0xffff; } static bool cma_match_private_data(struct rdma_id_private *id_priv, const struct cma_hdr *hdr) { struct sockaddr *addr = cma_src_addr(id_priv); __be32 ip4_addr; struct in6_addr ip6_addr; if (cma_any_addr(addr) && !id_priv->afonly) return true; switch (addr->sa_family) { case AF_INET: ip4_addr = ((struct sockaddr_in *)addr)->sin_addr.s_addr; if (cma_get_ip_ver(hdr) != 4) return false; if (!cma_any_addr(addr) && hdr->dst_addr.ip4.addr != ip4_addr) return false; break; case AF_INET6: ip6_addr = ((struct sockaddr_in6 *)addr)->sin6_addr; if (cma_get_ip_ver(hdr) != 6) return false; cma_ip6_clear_scope_id(&ip6_addr); if (!cma_any_addr(addr) && memcmp(&hdr->dst_addr.ip6, &ip6_addr, sizeof(ip6_addr))) return false; break; case AF_IB: return true; default: return false; } return true; } static bool cma_protocol_roce_dev_port(struct ib_device *device, int port_num) { enum rdma_link_layer ll = rdma_port_get_link_layer(device, port_num); enum rdma_transport_type transport = rdma_node_get_transport(device->node_type); return ll == IB_LINK_LAYER_ETHERNET && transport == RDMA_TRANSPORT_IB; } static bool cma_protocol_roce(const struct rdma_cm_id *id) { struct ib_device *device = id->device; const int port_num = id->port_num ?: rdma_start_port(device); return cma_protocol_roce_dev_port(device, port_num); } static bool cma_match_net_dev(const struct rdma_cm_id *id, const struct net_device *net_dev, u8 port_num) { const struct rdma_addr *addr = &id->route.addr; if (!net_dev) /* This request is an AF_IB request or a RoCE request */ return (!id->port_num || id->port_num == port_num) && (addr->src_addr.ss_family == AF_IB || cma_protocol_roce_dev_port(id->device, port_num)); return !addr->dev_addr.bound_dev_if || (net_eq(dev_net(net_dev), addr->dev_addr.net) && addr->dev_addr.bound_dev_if == net_dev->if_index); } static struct rdma_id_private *cma_find_listener( const struct rdma_bind_list *bind_list, const struct ib_cm_id *cm_id, const struct ib_cm_event *ib_event, const struct cma_req_info *req, const struct net_device *net_dev) { struct rdma_id_private *id_priv, *id_priv_dev; if (!bind_list) return ERR_PTR(-EINVAL); hlist_for_each_entry(id_priv, &bind_list->owners, node) { if (cma_match_private_data(id_priv, ib_event->private_data)) { if (id_priv->id.device == cm_id->device && cma_match_net_dev(&id_priv->id, net_dev, req->port)) return id_priv; list_for_each_entry(id_priv_dev, &id_priv->listen_list, listen_list) { if (id_priv_dev->id.device == cm_id->device && cma_match_net_dev(&id_priv_dev->id, net_dev, req->port)) return id_priv_dev; } } } return ERR_PTR(-EINVAL); } static struct rdma_id_private *cma_id_from_event(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event, struct net_device **net_dev) { struct cma_req_info req; struct rdma_bind_list *bind_list; struct rdma_id_private *id_priv; int err; err = cma_save_req_info(ib_event, &req); if (err) return ERR_PTR(err); *net_dev = cma_get_net_dev(ib_event, &req); if (IS_ERR(*net_dev)) { if (PTR_ERR(*net_dev) == -EAFNOSUPPORT) { /* Assuming the protocol is AF_IB */ *net_dev = NULL; } else if (cma_protocol_roce_dev_port(req.device, req.port)) { /* TODO find the net dev matching the request parameters * through the RoCE GID table */ *net_dev = NULL; } else { return ERR_CAST(*net_dev); } } bind_list = cma_ps_find(*net_dev ? dev_net(*net_dev) : &init_net, rdma_ps_from_service_id(req.service_id), cma_port_from_service_id(req.service_id)); id_priv = cma_find_listener(bind_list, cm_id, ib_event, &req, *net_dev); if (IS_ERR(id_priv) && *net_dev) { dev_put(*net_dev); *net_dev = NULL; } return id_priv; } static inline int cma_user_data_offset(struct rdma_id_private *id_priv) { return cma_family(id_priv) == AF_IB ? 0 : sizeof(struct cma_hdr); } static void cma_cancel_route(struct rdma_id_private *id_priv) { if (rdma_cap_ib_sa(id_priv->id.device, id_priv->id.port_num)) { if (id_priv->query) ib_sa_cancel_query(id_priv->query_id, id_priv->query); } } static void cma_cancel_listens(struct rdma_id_private *id_priv) { struct rdma_id_private *dev_id_priv; /* * Remove from listen_any_list to prevent added devices from spawning * additional listen requests. */ mutex_lock(&lock); list_del(&id_priv->list); while (!list_empty(&id_priv->listen_list)) { dev_id_priv = list_entry(id_priv->listen_list.next, struct rdma_id_private, listen_list); /* sync with device removal to avoid duplicate destruction */ list_del_init(&dev_id_priv->list); list_del(&dev_id_priv->listen_list); mutex_unlock(&lock); rdma_destroy_id(&dev_id_priv->id); mutex_lock(&lock); } mutex_unlock(&lock); } static void cma_cancel_operation(struct rdma_id_private *id_priv, enum rdma_cm_state state) { switch (state) { case RDMA_CM_ADDR_QUERY: rdma_addr_cancel(&id_priv->id.route.addr.dev_addr); break; case RDMA_CM_ROUTE_QUERY: cma_cancel_route(id_priv); break; case RDMA_CM_LISTEN: if (cma_any_addr(cma_src_addr(id_priv)) && !id_priv->cma_dev) cma_cancel_listens(id_priv); break; default: break; } } static void cma_release_port(struct rdma_id_private *id_priv) { struct rdma_bind_list *bind_list = id_priv->bind_list; struct vnet *net = id_priv->id.route.addr.dev_addr.net; if (!bind_list) return; mutex_lock(&lock); hlist_del(&id_priv->node); if (hlist_empty(&bind_list->owners)) { cma_ps_remove(net, bind_list->ps, bind_list->port); kfree(bind_list); } mutex_unlock(&lock); } static void cma_leave_mc_groups(struct rdma_id_private *id_priv) { struct cma_multicast *mc; while (!list_empty(&id_priv->mc_list)) { mc = container_of(id_priv->mc_list.next, struct cma_multicast, list); list_del(&mc->list); if (rdma_cap_ib_mcast(id_priv->cma_dev->device, id_priv->id.port_num)) { ib_sa_free_multicast(mc->multicast.ib); kfree(mc); } else { if (mc->igmp_joined) { struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; struct net_device *ndev = NULL; if (dev_addr->bound_dev_if) ndev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if); if (ndev) { cma_igmp_send(ndev, &mc->multicast.ib->rec.mgid, false); dev_put(ndev); } } kref_put(&mc->mcref, release_mc); } } } void rdma_destroy_id(struct rdma_cm_id *id) { struct rdma_id_private *id_priv; enum rdma_cm_state state; id_priv = container_of(id, struct rdma_id_private, id); state = cma_exch(id_priv, RDMA_CM_DESTROYING); cma_cancel_operation(id_priv, state); /* * Wait for any active callback to finish. New callbacks will find * the id_priv state set to destroying and abort. */ mutex_lock(&id_priv->handler_mutex); mutex_unlock(&id_priv->handler_mutex); if (id_priv->cma_dev) { if (rdma_cap_ib_cm(id_priv->id.device, 1)) { if (id_priv->cm_id.ib) ib_destroy_cm_id(id_priv->cm_id.ib); } else if (rdma_cap_iw_cm(id_priv->id.device, 1)) { if (id_priv->cm_id.iw) iw_destroy_cm_id(id_priv->cm_id.iw); } cma_leave_mc_groups(id_priv); cma_release_dev(id_priv); } cma_release_port(id_priv); cma_deref_id(id_priv); wait_for_completion(&id_priv->comp); if (id_priv->internal_id) cma_deref_id(id_priv->id.context); kfree(id_priv->id.route.path_rec); kfree(id_priv); } EXPORT_SYMBOL(rdma_destroy_id); static int cma_rep_recv(struct rdma_id_private *id_priv) { int ret; ret = cma_modify_qp_rtr(id_priv, NULL); if (ret) goto reject; ret = cma_modify_qp_rts(id_priv, NULL); if (ret) goto reject; ret = ib_send_cm_rtu(id_priv->cm_id.ib, NULL, 0); if (ret) goto reject; return 0; reject: cma_modify_qp_err(id_priv); ib_send_cm_rej(id_priv->cm_id.ib, IB_CM_REJ_CONSUMER_DEFINED, NULL, 0, NULL, 0); return ret; } static void cma_set_rep_event_data(struct rdma_cm_event *event, struct ib_cm_rep_event_param *rep_data, void *private_data) { event->param.conn.private_data = private_data; event->param.conn.private_data_len = IB_CM_REP_PRIVATE_DATA_SIZE; event->param.conn.responder_resources = rep_data->responder_resources; event->param.conn.initiator_depth = rep_data->initiator_depth; event->param.conn.flow_control = rep_data->flow_control; event->param.conn.rnr_retry_count = rep_data->rnr_retry_count; event->param.conn.srq = rep_data->srq; event->param.conn.qp_num = rep_data->remote_qpn; } static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event) { struct rdma_id_private *id_priv = cm_id->context; struct rdma_cm_event event; int ret = 0; mutex_lock(&id_priv->handler_mutex); if ((ib_event->event != IB_CM_TIMEWAIT_EXIT && id_priv->state != RDMA_CM_CONNECT) || (ib_event->event == IB_CM_TIMEWAIT_EXIT && id_priv->state != RDMA_CM_DISCONNECT)) goto out; memset(&event, 0, sizeof event); switch (ib_event->event) { case IB_CM_REQ_ERROR: case IB_CM_REP_ERROR: event.event = RDMA_CM_EVENT_UNREACHABLE; event.status = -ETIMEDOUT; break; case IB_CM_REP_RECEIVED: if (id_priv->id.qp) { event.status = cma_rep_recv(id_priv); event.event = event.status ? RDMA_CM_EVENT_CONNECT_ERROR : RDMA_CM_EVENT_ESTABLISHED; } else { event.event = RDMA_CM_EVENT_CONNECT_RESPONSE; } cma_set_rep_event_data(&event, &ib_event->param.rep_rcvd, ib_event->private_data); break; case IB_CM_RTU_RECEIVED: case IB_CM_USER_ESTABLISHED: event.event = RDMA_CM_EVENT_ESTABLISHED; break; case IB_CM_DREQ_ERROR: event.status = -ETIMEDOUT; /* fall through */ case IB_CM_DREQ_RECEIVED: case IB_CM_DREP_RECEIVED: if (!cma_comp_exch(id_priv, RDMA_CM_CONNECT, RDMA_CM_DISCONNECT)) goto out; event.event = RDMA_CM_EVENT_DISCONNECTED; break; case IB_CM_TIMEWAIT_EXIT: event.event = RDMA_CM_EVENT_TIMEWAIT_EXIT; break; case IB_CM_MRA_RECEIVED: /* ignore event */ goto out; case IB_CM_REJ_RECEIVED: cma_modify_qp_err(id_priv); event.status = ib_event->param.rej_rcvd.reason; event.event = RDMA_CM_EVENT_REJECTED; event.param.conn.private_data = ib_event->private_data; event.param.conn.private_data_len = IB_CM_REJ_PRIVATE_DATA_SIZE; break; default: pr_err("RDMA CMA: unexpected IB CM event: %d\n", ib_event->event); goto out; } ret = id_priv->id.event_handler(&id_priv->id, &event); if (ret) { /* Destroy the CM ID by returning a non-zero value. */ id_priv->cm_id.ib = NULL; cma_exch(id_priv, RDMA_CM_DESTROYING); mutex_unlock(&id_priv->handler_mutex); rdma_destroy_id(&id_priv->id); return ret; } out: mutex_unlock(&id_priv->handler_mutex); return ret; } static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id, struct ib_cm_event *ib_event, struct net_device *net_dev) { struct rdma_id_private *id_priv; struct rdma_cm_id *id; struct rdma_route *rt; const sa_family_t ss_family = listen_id->route.addr.src_addr.ss_family; const __be64 service_id = ib_event->param.req_rcvd.primary_path->service_id; int ret; id = rdma_create_id(listen_id->route.addr.dev_addr.net, listen_id->event_handler, listen_id->context, listen_id->ps, ib_event->param.req_rcvd.qp_type); if (IS_ERR(id)) return NULL; id_priv = container_of(id, struct rdma_id_private, id); if (cma_save_net_info((struct sockaddr *)&id->route.addr.src_addr, (struct sockaddr *)&id->route.addr.dst_addr, listen_id, ib_event, ss_family, service_id)) goto err; rt = &id->route; rt->num_paths = ib_event->param.req_rcvd.alternate_path ? 2 : 1; rt->path_rec = kmalloc(sizeof *rt->path_rec * rt->num_paths, GFP_KERNEL); if (!rt->path_rec) goto err; rt->path_rec[0] = *ib_event->param.req_rcvd.primary_path; if (rt->num_paths == 2) rt->path_rec[1] = *ib_event->param.req_rcvd.alternate_path; if (net_dev) { ret = rdma_copy_addr(&rt->addr.dev_addr, net_dev, NULL); if (ret) goto err; } else { if (!cma_protocol_roce(listen_id) && cma_any_addr(cma_src_addr(id_priv))) { rt->addr.dev_addr.dev_type = ARPHRD_INFINIBAND; rdma_addr_set_sgid(&rt->addr.dev_addr, &rt->path_rec[0].sgid); ib_addr_set_pkey(&rt->addr.dev_addr, be16_to_cpu(rt->path_rec[0].pkey)); } else if (!cma_any_addr(cma_src_addr(id_priv))) { ret = cma_translate_addr(cma_src_addr(id_priv), &rt->addr.dev_addr); if (ret) goto err; } } rdma_addr_set_dgid(&rt->addr.dev_addr, &rt->path_rec[0].dgid); id_priv->state = RDMA_CM_CONNECT; return id_priv; err: rdma_destroy_id(id); return NULL; } static struct rdma_id_private *cma_new_udp_id(struct rdma_cm_id *listen_id, struct ib_cm_event *ib_event, struct net_device *net_dev) { struct rdma_id_private *id_priv; struct rdma_cm_id *id; const sa_family_t ss_family = listen_id->route.addr.src_addr.ss_family; struct vnet *net = listen_id->route.addr.dev_addr.net; int ret; id = rdma_create_id(net, listen_id->event_handler, listen_id->context, listen_id->ps, IB_QPT_UD); if (IS_ERR(id)) return NULL; id_priv = container_of(id, struct rdma_id_private, id); if (cma_save_net_info((struct sockaddr *)&id->route.addr.src_addr, (struct sockaddr *)&id->route.addr.dst_addr, listen_id, ib_event, ss_family, ib_event->param.sidr_req_rcvd.service_id)) goto err; if (net_dev) { ret = rdma_copy_addr(&id->route.addr.dev_addr, net_dev, NULL); if (ret) goto err; } else { if (!cma_any_addr(cma_src_addr(id_priv))) { ret = cma_translate_addr(cma_src_addr(id_priv), &id->route.addr.dev_addr); if (ret) goto err; } } id_priv->state = RDMA_CM_CONNECT; return id_priv; err: rdma_destroy_id(id); return NULL; } static void cma_set_req_event_data(struct rdma_cm_event *event, struct ib_cm_req_event_param *req_data, void *private_data, int offset) { event->param.conn.private_data = (char *)private_data + offset; event->param.conn.private_data_len = IB_CM_REQ_PRIVATE_DATA_SIZE - offset; event->param.conn.responder_resources = req_data->responder_resources; event->param.conn.initiator_depth = req_data->initiator_depth; event->param.conn.flow_control = req_data->flow_control; event->param.conn.retry_count = req_data->retry_count; event->param.conn.rnr_retry_count = req_data->rnr_retry_count; event->param.conn.srq = req_data->srq; event->param.conn.qp_num = req_data->remote_qpn; } static int cma_check_req_qp_type(struct rdma_cm_id *id, struct ib_cm_event *ib_event) { return (((ib_event->event == IB_CM_REQ_RECEIVED) && (ib_event->param.req_rcvd.qp_type == id->qp_type)) || ((ib_event->event == IB_CM_SIDR_REQ_RECEIVED) && (id->qp_type == IB_QPT_UD)) || (!id->qp_type)); } static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event) { struct rdma_id_private *listen_id, *conn_id = NULL; struct rdma_cm_event event; struct net_device *net_dev; int offset, ret; listen_id = cma_id_from_event(cm_id, ib_event, &net_dev); if (IS_ERR(listen_id)) return PTR_ERR(listen_id); if (!cma_check_req_qp_type(&listen_id->id, ib_event)) { ret = -EINVAL; goto net_dev_put; } mutex_lock(&listen_id->handler_mutex); if (listen_id->state != RDMA_CM_LISTEN) { ret = -ECONNABORTED; goto err1; } memset(&event, 0, sizeof event); offset = cma_user_data_offset(listen_id); event.event = RDMA_CM_EVENT_CONNECT_REQUEST; if (ib_event->event == IB_CM_SIDR_REQ_RECEIVED) { conn_id = cma_new_udp_id(&listen_id->id, ib_event, net_dev); event.param.ud.private_data = (char *)ib_event->private_data + offset; event.param.ud.private_data_len = IB_CM_SIDR_REQ_PRIVATE_DATA_SIZE - offset; } else { conn_id = cma_new_conn_id(&listen_id->id, ib_event, net_dev); cma_set_req_event_data(&event, &ib_event->param.req_rcvd, ib_event->private_data, offset); } if (!conn_id) { ret = -ENOMEM; goto err1; } mutex_lock_nested(&conn_id->handler_mutex, SINGLE_DEPTH_NESTING); ret = cma_acquire_dev(conn_id, listen_id); if (ret) goto err2; conn_id->cm_id.ib = cm_id; cm_id->context = conn_id; cm_id->cm_handler = cma_ib_handler; /* * Protect against the user destroying conn_id from another thread * until we're done accessing it. */ atomic_inc(&conn_id->refcount); ret = conn_id->id.event_handler(&conn_id->id, &event); if (ret) goto err3; /* * Acquire mutex to prevent user executing rdma_destroy_id() * while we're accessing the cm_id. */ mutex_lock(&lock); if (cma_comp(conn_id, RDMA_CM_CONNECT) && (conn_id->id.qp_type != IB_QPT_UD)) ib_send_cm_mra(cm_id, CMA_CM_MRA_SETTING, NULL, 0); mutex_unlock(&lock); mutex_unlock(&conn_id->handler_mutex); mutex_unlock(&listen_id->handler_mutex); cma_deref_id(conn_id); if (net_dev) dev_put(net_dev); return 0; err3: cma_deref_id(conn_id); /* Destroy the CM ID by returning a non-zero value. */ conn_id->cm_id.ib = NULL; err2: cma_exch(conn_id, RDMA_CM_DESTROYING); mutex_unlock(&conn_id->handler_mutex); err1: mutex_unlock(&listen_id->handler_mutex); if (conn_id) rdma_destroy_id(&conn_id->id); net_dev_put: if (net_dev) dev_put(net_dev); return ret; } __be64 rdma_get_service_id(struct rdma_cm_id *id, struct sockaddr *addr) { if (addr->sa_family == AF_IB) return ((struct sockaddr_ib *) addr)->sib_sid; return cpu_to_be64(((u64)id->ps << 16) + be16_to_cpu(cma_port(addr))); } EXPORT_SYMBOL(rdma_get_service_id); static int cma_iw_handler(struct iw_cm_id *iw_id, struct iw_cm_event *iw_event) { struct rdma_id_private *id_priv = iw_id->context; struct rdma_cm_event event; int ret = 0; struct sockaddr *laddr = (struct sockaddr *)&iw_event->local_addr; struct sockaddr *raddr = (struct sockaddr *)&iw_event->remote_addr; mutex_lock(&id_priv->handler_mutex); if (id_priv->state != RDMA_CM_CONNECT) goto out; memset(&event, 0, sizeof event); switch (iw_event->event) { case IW_CM_EVENT_CLOSE: event.event = RDMA_CM_EVENT_DISCONNECTED; break; case IW_CM_EVENT_CONNECT_REPLY: memcpy(cma_src_addr(id_priv), laddr, rdma_addr_size(laddr)); memcpy(cma_dst_addr(id_priv), raddr, rdma_addr_size(raddr)); switch (iw_event->status) { case 0: event.event = RDMA_CM_EVENT_ESTABLISHED; event.param.conn.initiator_depth = iw_event->ird; event.param.conn.responder_resources = iw_event->ord; break; case -ECONNRESET: case -ECONNREFUSED: event.event = RDMA_CM_EVENT_REJECTED; break; case -ETIMEDOUT: event.event = RDMA_CM_EVENT_UNREACHABLE; break; default: event.event = RDMA_CM_EVENT_CONNECT_ERROR; break; } break; case IW_CM_EVENT_ESTABLISHED: event.event = RDMA_CM_EVENT_ESTABLISHED; event.param.conn.initiator_depth = iw_event->ird; event.param.conn.responder_resources = iw_event->ord; break; default: BUG_ON(1); } event.status = iw_event->status; event.param.conn.private_data = iw_event->private_data; event.param.conn.private_data_len = iw_event->private_data_len; ret = id_priv->id.event_handler(&id_priv->id, &event); if (ret) { /* Destroy the CM ID by returning a non-zero value. */ id_priv->cm_id.iw = NULL; cma_exch(id_priv, RDMA_CM_DESTROYING); mutex_unlock(&id_priv->handler_mutex); rdma_destroy_id(&id_priv->id); return ret; } out: mutex_unlock(&id_priv->handler_mutex); return ret; } static int iw_conn_req_handler(struct iw_cm_id *cm_id, struct iw_cm_event *iw_event) { struct rdma_cm_id *new_cm_id; struct rdma_id_private *listen_id, *conn_id; struct rdma_cm_event event; int ret = -ECONNABORTED; struct sockaddr *laddr = (struct sockaddr *)&iw_event->local_addr; struct sockaddr *raddr = (struct sockaddr *)&iw_event->remote_addr; listen_id = cm_id->context; mutex_lock(&listen_id->handler_mutex); if (listen_id->state != RDMA_CM_LISTEN) goto out; /* Create a new RDMA id for the new IW CM ID */ new_cm_id = rdma_create_id(listen_id->id.route.addr.dev_addr.net, listen_id->id.event_handler, listen_id->id.context, RDMA_PS_TCP, IB_QPT_RC); if (IS_ERR(new_cm_id)) { ret = -ENOMEM; goto out; } conn_id = container_of(new_cm_id, struct rdma_id_private, id); mutex_lock_nested(&conn_id->handler_mutex, SINGLE_DEPTH_NESTING); conn_id->state = RDMA_CM_CONNECT; ret = rdma_translate_ip(laddr, &conn_id->id.route.addr.dev_addr); if (ret) { mutex_unlock(&conn_id->handler_mutex); rdma_destroy_id(new_cm_id); goto out; } ret = cma_acquire_dev(conn_id, listen_id); if (ret) { mutex_unlock(&conn_id->handler_mutex); rdma_destroy_id(new_cm_id); goto out; } conn_id->cm_id.iw = cm_id; cm_id->context = conn_id; cm_id->cm_handler = cma_iw_handler; memcpy(cma_src_addr(conn_id), laddr, rdma_addr_size(laddr)); memcpy(cma_dst_addr(conn_id), raddr, rdma_addr_size(raddr)); memset(&event, 0, sizeof event); event.event = RDMA_CM_EVENT_CONNECT_REQUEST; event.param.conn.private_data = iw_event->private_data; event.param.conn.private_data_len = iw_event->private_data_len; event.param.conn.initiator_depth = iw_event->ird; event.param.conn.responder_resources = iw_event->ord; /* * Protect against the user destroying conn_id from another thread * until we're done accessing it. */ atomic_inc(&conn_id->refcount); ret = conn_id->id.event_handler(&conn_id->id, &event); if (ret) { /* User wants to destroy the CM ID */ conn_id->cm_id.iw = NULL; cma_exch(conn_id, RDMA_CM_DESTROYING); mutex_unlock(&conn_id->handler_mutex); cma_deref_id(conn_id); rdma_destroy_id(&conn_id->id); goto out; } mutex_unlock(&conn_id->handler_mutex); cma_deref_id(conn_id); out: mutex_unlock(&listen_id->handler_mutex); return ret; } static int cma_ib_listen(struct rdma_id_private *id_priv) { struct sockaddr *addr; struct ib_cm_id *id; __be64 svc_id; addr = cma_src_addr(id_priv); svc_id = rdma_get_service_id(&id_priv->id, addr); id = ib_cm_insert_listen(id_priv->id.device, cma_req_handler, svc_id); if (IS_ERR(id)) return PTR_ERR(id); id_priv->cm_id.ib = id; return 0; } static int cma_iw_listen(struct rdma_id_private *id_priv, int backlog) { int ret; struct iw_cm_id *id; id = iw_create_cm_id(id_priv->id.device, iw_conn_req_handler, id_priv); if (IS_ERR(id)) return PTR_ERR(id); id->tos = id_priv->tos; id_priv->cm_id.iw = id; memcpy(&id_priv->cm_id.iw->local_addr, cma_src_addr(id_priv), rdma_addr_size(cma_src_addr(id_priv))); ret = iw_cm_listen(id_priv->cm_id.iw, backlog); if (ret) { iw_destroy_cm_id(id_priv->cm_id.iw); id_priv->cm_id.iw = NULL; } return ret; } static int cma_listen_handler(struct rdma_cm_id *id, struct rdma_cm_event *event) { struct rdma_id_private *id_priv = id->context; id->context = id_priv->id.context; id->event_handler = id_priv->id.event_handler; return id_priv->id.event_handler(id, event); } static void cma_listen_on_dev(struct rdma_id_private *id_priv, struct cma_device *cma_dev) { struct rdma_id_private *dev_id_priv; struct rdma_cm_id *id; struct vnet *net = id_priv->id.route.addr.dev_addr.net; int ret; if (cma_family(id_priv) == AF_IB && !rdma_cap_ib_cm(cma_dev->device, 1)) return; id = rdma_create_id(net, cma_listen_handler, id_priv, id_priv->id.ps, id_priv->id.qp_type); if (IS_ERR(id)) return; dev_id_priv = container_of(id, struct rdma_id_private, id); dev_id_priv->state = RDMA_CM_ADDR_BOUND; memcpy(cma_src_addr(dev_id_priv), cma_src_addr(id_priv), rdma_addr_size(cma_src_addr(id_priv))); _cma_attach_to_dev(dev_id_priv, cma_dev); list_add_tail(&dev_id_priv->listen_list, &id_priv->listen_list); atomic_inc(&id_priv->refcount); dev_id_priv->internal_id = 1; dev_id_priv->afonly = id_priv->afonly; ret = rdma_listen(id, id_priv->backlog); if (ret) pr_warn("RDMA CMA: cma_listen_on_dev, error %d, listening on device %s\n", ret, cma_dev->device->name); } static void cma_listen_on_all(struct rdma_id_private *id_priv) { struct cma_device *cma_dev; mutex_lock(&lock); list_add_tail(&id_priv->list, &listen_any_list); list_for_each_entry(cma_dev, &dev_list, list) cma_listen_on_dev(id_priv, cma_dev); mutex_unlock(&lock); } void rdma_set_service_type(struct rdma_cm_id *id, int tos) { struct rdma_id_private *id_priv; id_priv = container_of(id, struct rdma_id_private, id); id_priv->tos = (u8) tos; } EXPORT_SYMBOL(rdma_set_service_type); static void cma_query_handler(int status, struct ib_sa_path_rec *path_rec, void *context) { struct cma_work *work = context; struct rdma_route *route; 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; } queue_work(cma_wq, &work->work); } static int cma_query_ib_route(struct rdma_id_private *id_priv, int timeout_ms, struct cma_work *work) { struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; struct ib_sa_path_rec path_rec; ib_sa_comp_mask comp_mask; struct sockaddr_in6 *sin6; struct sockaddr_ib *sib; memset(&path_rec, 0, sizeof path_rec); rdma_addr_get_sgid(dev_addr, &path_rec.sgid); rdma_addr_get_dgid(dev_addr, &path_rec.dgid); path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr)); path_rec.numb_path = 1; path_rec.reversible = 1; path_rec.service_id = rdma_get_service_id(&id_priv->id, cma_dst_addr(id_priv)); comp_mask = IB_SA_PATH_REC_DGID | IB_SA_PATH_REC_SGID | IB_SA_PATH_REC_PKEY | IB_SA_PATH_REC_NUMB_PATH | IB_SA_PATH_REC_REVERSIBLE | IB_SA_PATH_REC_SERVICE_ID; switch (cma_family(id_priv)) { case AF_INET: path_rec.qos_class = cpu_to_be16((u16) id_priv->tos); comp_mask |= IB_SA_PATH_REC_QOS_CLASS; break; case AF_INET6: sin6 = (struct sockaddr_in6 *) cma_src_addr(id_priv); path_rec.traffic_class = (u8) (be32_to_cpu(sin6->sin6_flowinfo) >> 20); comp_mask |= IB_SA_PATH_REC_TRAFFIC_CLASS; break; case AF_IB: sib = (struct sockaddr_ib *) cma_src_addr(id_priv); path_rec.traffic_class = (u8) (be32_to_cpu(sib->sib_flowinfo) >> 20); comp_mask |= IB_SA_PATH_REC_TRAFFIC_CLASS; break; } id_priv->query_id = ib_sa_path_rec_get(&sa_client, id_priv->id.device, id_priv->id.port_num, &path_rec, comp_mask, timeout_ms, GFP_KERNEL, cma_query_handler, work, &id_priv->query); return (id_priv->query_id < 0) ? id_priv->query_id : 0; } static void cma_work_handler(struct work_struct *_work) { struct cma_work *work = container_of(_work, struct cma_work, work); struct rdma_id_private *id_priv = work->id; int destroy = 0; mutex_lock(&id_priv->handler_mutex); if (!cma_comp_exch(id_priv, work->old_state, work->new_state)) goto out; if (id_priv->id.event_handler(&id_priv->id, &work->event)) { cma_exch(id_priv, RDMA_CM_DESTROYING); destroy = 1; } out: mutex_unlock(&id_priv->handler_mutex); cma_deref_id(id_priv); if (destroy) rdma_destroy_id(&id_priv->id); kfree(work); } static int cma_resolve_ib_route(struct rdma_id_private *id_priv, int timeout_ms) { struct rdma_route *route = &id_priv->id.route; struct cma_work *work; int ret; work = kzalloc(sizeof *work, GFP_KERNEL); if (!work) return -ENOMEM; work->id = id_priv; INIT_WORK(&work->work, cma_work_handler); work->old_state = RDMA_CM_ROUTE_QUERY; work->new_state = RDMA_CM_ROUTE_RESOLVED; work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED; route->path_rec = kmalloc(sizeof *route->path_rec, GFP_KERNEL); if (!route->path_rec) { ret = -ENOMEM; goto err1; } ret = cma_query_ib_route(id_priv, timeout_ms, work); if (ret) goto err2; return 0; err2: kfree(route->path_rec); route->path_rec = NULL; err1: kfree(work); return ret; } int rdma_set_ib_paths(struct rdma_cm_id *id, struct ib_sa_path_rec *path_rec, int num_paths) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, RDMA_CM_ROUTE_RESOLVED)) return -EINVAL; id->route.path_rec = kmemdup(path_rec, sizeof *path_rec * num_paths, GFP_KERNEL); if (!id->route.path_rec) { ret = -ENOMEM; goto err; } id->route.num_paths = num_paths; return 0; err: cma_comp_exch(id_priv, RDMA_CM_ROUTE_RESOLVED, RDMA_CM_ADDR_RESOLVED); return ret; } EXPORT_SYMBOL(rdma_set_ib_paths); static int cma_resolve_iw_route(struct rdma_id_private *id_priv, int timeout_ms) { struct cma_work *work; work = kzalloc(sizeof *work, GFP_KERNEL); if (!work) return -ENOMEM; work->id = id_priv; INIT_WORK(&work->work, cma_work_handler); work->old_state = RDMA_CM_ROUTE_QUERY; work->new_state = RDMA_CM_ROUTE_RESOLVED; work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED; queue_work(cma_wq, &work->work); return 0; } static int iboe_tos_to_sl(struct net_device *ndev, int tos) { /* get service level, SL, from IPv4 type of service, TOS */ int sl = (tos >> 5) & 0x7; /* final mappings are done by the vendor specific drivers */ return sl; } static enum ib_gid_type cma_route_gid_type(enum rdma_network_type network_type, unsigned long supported_gids, enum ib_gid_type default_gid) { if ((network_type == RDMA_NETWORK_IPV4 || network_type == RDMA_NETWORK_IPV6) && test_bit(IB_GID_TYPE_ROCE_UDP_ENCAP, &supported_gids)) return IB_GID_TYPE_ROCE_UDP_ENCAP; return default_gid; } static int cma_resolve_iboe_route(struct rdma_id_private *id_priv) { struct rdma_route *route = &id_priv->id.route; struct rdma_addr *addr = &route->addr; struct cma_work *work; int ret; struct net_device *ndev = NULL; work = kzalloc(sizeof *work, GFP_KERNEL); if (!work) return -ENOMEM; work->id = id_priv; INIT_WORK(&work->work, cma_work_handler); route->path_rec = kzalloc(sizeof *route->path_rec, GFP_KERNEL); if (!route->path_rec) { ret = -ENOMEM; goto err1; } route->num_paths = 1; if (addr->dev_addr.bound_dev_if) { unsigned long supported_gids; ndev = dev_get_by_index(addr->dev_addr.net, addr->dev_addr.bound_dev_if); if (!ndev) { ret = -ENODEV; goto err2; } route->path_rec->net = ndev->if_vnet; route->path_rec->ifindex = ndev->if_index; supported_gids = roce_gid_type_mask_support(id_priv->id.device, id_priv->id.port_num); route->path_rec->gid_type = cma_route_gid_type(addr->dev_addr.network, supported_gids, id_priv->gid_type); } if (!ndev) { ret = -ENODEV; goto err2; } memcpy(route->path_rec->dmac, addr->dev_addr.dst_dev_addr, ETH_ALEN); rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.src_addr, &route->path_rec->sgid); rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.dst_addr, &route->path_rec->dgid); /* Use the hint from IP Stack to select GID Type */ if (route->path_rec->gid_type < ib_network_to_gid_type(addr->dev_addr.network)) route->path_rec->gid_type = ib_network_to_gid_type(addr->dev_addr.network); if (((struct sockaddr *)&id_priv->id.route.addr.dst_addr)->sa_family != AF_IB) /* TODO: get the hoplimit from the inet/inet6 device */ route->path_rec->hop_limit = addr->dev_addr.hoplimit; else route->path_rec->hop_limit = 1; route->path_rec->reversible = 1; route->path_rec->pkey = cpu_to_be16(0xffff); route->path_rec->mtu_selector = IB_SA_EQ; route->path_rec->sl = iboe_tos_to_sl(ndev, id_priv->tos); route->path_rec->traffic_class = id_priv->tos; route->path_rec->mtu = iboe_get_mtu(ndev->if_mtu); route->path_rec->rate_selector = IB_SA_EQ; route->path_rec->rate = iboe_get_rate(ndev); dev_put(ndev); route->path_rec->packet_life_time_selector = IB_SA_EQ; route->path_rec->packet_life_time = CMA_IBOE_PACKET_LIFETIME; if (!route->path_rec->mtu) { ret = -EINVAL; goto err2; } work->old_state = RDMA_CM_ROUTE_QUERY; work->new_state = RDMA_CM_ROUTE_RESOLVED; work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED; work->event.status = 0; queue_work(cma_wq, &work->work); return 0; err2: kfree(route->path_rec); route->path_rec = NULL; err1: kfree(work); return ret; } int rdma_resolve_route(struct rdma_cm_id *id, int timeout_ms) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, RDMA_CM_ROUTE_QUERY)) return -EINVAL; atomic_inc(&id_priv->refcount); 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)) ret = cma_resolve_iboe_route(id_priv); else if (rdma_protocol_iwarp(id->device, id->port_num)) ret = cma_resolve_iw_route(id_priv, timeout_ms); else ret = -ENOSYS; if (ret) goto err; return 0; err: cma_comp_exch(id_priv, RDMA_CM_ROUTE_QUERY, RDMA_CM_ADDR_RESOLVED); cma_deref_id(id_priv); return ret; } EXPORT_SYMBOL(rdma_resolve_route); static void cma_set_loopback(struct sockaddr *addr) { switch (addr->sa_family) { case AF_INET: ((struct sockaddr_in *) addr)->sin_addr.s_addr = htonl(INADDR_LOOPBACK); break; case AF_INET6: ipv6_addr_set(&((struct sockaddr_in6 *) addr)->sin6_addr, 0, 0, 0, htonl(1)); break; default: ib_addr_set(&((struct sockaddr_ib *) addr)->sib_addr, 0, 0, 0, htonl(1)); break; } } static int cma_bind_loopback(struct rdma_id_private *id_priv) { struct cma_device *cma_dev, *cur_dev; struct ib_port_attr port_attr; union ib_gid gid; u16 pkey; int ret; u8 p; cma_dev = NULL; mutex_lock(&lock); list_for_each_entry(cur_dev, &dev_list, list) { if (cma_family(id_priv) == AF_IB && !rdma_cap_ib_cm(cur_dev->device, 1)) continue; if (!cma_dev) cma_dev = cur_dev; for (p = 1; p <= cur_dev->device->phys_port_cnt; ++p) { if (!ib_query_port(cur_dev->device, p, &port_attr) && port_attr.state == IB_PORT_ACTIVE) { cma_dev = cur_dev; goto port_found; } } } if (!cma_dev) { ret = -ENODEV; goto out; } p = 1; port_found: ret = ib_get_cached_gid(cma_dev->device, p, 0, &gid, NULL); if (ret) goto out; ret = ib_get_cached_pkey(cma_dev->device, p, 0, &pkey); if (ret) goto out; id_priv->id.route.addr.dev_addr.dev_type = (rdma_protocol_ib(cma_dev->device, p)) ? ARPHRD_INFINIBAND : ARPHRD_ETHER; rdma_addr_set_sgid(&id_priv->id.route.addr.dev_addr, &gid); ib_addr_set_pkey(&id_priv->id.route.addr.dev_addr, pkey); id_priv->id.port_num = p; cma_attach_to_dev(id_priv, cma_dev); cma_set_loopback(cma_src_addr(id_priv)); out: mutex_unlock(&lock); return ret; } static void addr_handler(int status, struct sockaddr *src_addr, struct rdma_dev_addr *dev_addr, void *context) { struct rdma_id_private *id_priv = context; struct rdma_cm_event event; memset(&event, 0, sizeof event); mutex_lock(&id_priv->handler_mutex); if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, RDMA_CM_ADDR_RESOLVED)) goto out; memcpy(cma_src_addr(id_priv), src_addr, rdma_addr_size(src_addr)); if (!status && !id_priv->cma_dev) status = cma_acquire_dev(id_priv, NULL); if (status) { if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, RDMA_CM_ADDR_BOUND)) goto out; event.event = RDMA_CM_EVENT_ADDR_ERROR; event.status = status; } else event.event = RDMA_CM_EVENT_ADDR_RESOLVED; if (id_priv->id.event_handler(&id_priv->id, &event)) { cma_exch(id_priv, RDMA_CM_DESTROYING); mutex_unlock(&id_priv->handler_mutex); cma_deref_id(id_priv); rdma_destroy_id(&id_priv->id); return; } out: mutex_unlock(&id_priv->handler_mutex); cma_deref_id(id_priv); } static int cma_resolve_loopback(struct rdma_id_private *id_priv) { struct cma_work *work; union ib_gid gid; int ret; work = kzalloc(sizeof *work, GFP_KERNEL); if (!work) return -ENOMEM; if (!id_priv->cma_dev) { ret = cma_bind_loopback(id_priv); if (ret) goto err; } rdma_addr_get_sgid(&id_priv->id.route.addr.dev_addr, &gid); rdma_addr_set_dgid(&id_priv->id.route.addr.dev_addr, &gid); work->id = id_priv; INIT_WORK(&work->work, cma_work_handler); work->old_state = RDMA_CM_ADDR_QUERY; work->new_state = RDMA_CM_ADDR_RESOLVED; work->event.event = RDMA_CM_EVENT_ADDR_RESOLVED; queue_work(cma_wq, &work->work); return 0; err: kfree(work); return ret; } static int cma_resolve_ib_addr(struct rdma_id_private *id_priv) { struct cma_work *work; int ret; work = kzalloc(sizeof *work, GFP_KERNEL); if (!work) return -ENOMEM; if (!id_priv->cma_dev) { ret = cma_resolve_ib_dev(id_priv); if (ret) goto err; } rdma_addr_set_dgid(&id_priv->id.route.addr.dev_addr, (union ib_gid *) &(((struct sockaddr_ib *) &id_priv->id.route.addr.dst_addr)->sib_addr)); work->id = id_priv; INIT_WORK(&work->work, cma_work_handler); work->old_state = RDMA_CM_ADDR_QUERY; work->new_state = RDMA_CM_ADDR_RESOLVED; work->event.event = RDMA_CM_EVENT_ADDR_RESOLVED; queue_work(cma_wq, &work->work); return 0; err: kfree(work); return ret; } static int cma_bind_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, 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 (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 (IN6_IS_SCOPE_LINKLOCAL(&dst_addr6->sin6_addr) || IN6_IS_ADDR_MC_INTFACELOCAL(&dst_addr6->sin6_addr)) 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); } int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, struct sockaddr *dst_addr, int timeout_ms) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (id_priv->state == RDMA_CM_IDLE) { ret = cma_bind_addr(id, src_addr, dst_addr); if (ret) return ret; } if (cma_family(id_priv) != dst_addr->sa_family) return -EINVAL; if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_ADDR_QUERY)) return -EINVAL; atomic_inc(&id_priv->refcount); memcpy(cma_dst_addr(id_priv), dst_addr, rdma_addr_size(dst_addr)); 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 = cma_check_linklocal(&id->route.addr.dev_addr, dst_addr); if (ret) goto err; ret = rdma_resolve_ip(&addr_client, cma_src_addr(id_priv), dst_addr, &id->route.addr.dev_addr, timeout_ms, addr_handler, id_priv); } } if (ret) goto err; return 0; err: cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, RDMA_CM_ADDR_BOUND); cma_deref_id(id_priv); return ret; } EXPORT_SYMBOL(rdma_resolve_addr); int rdma_set_reuseaddr(struct rdma_cm_id *id, int reuse) { struct rdma_id_private *id_priv; unsigned long flags; int ret; id_priv = container_of(id, struct rdma_id_private, id); spin_lock_irqsave(&id_priv->lock, flags); if (reuse || id_priv->state == RDMA_CM_IDLE) { id_priv->reuseaddr = reuse; ret = 0; } else { ret = -EINVAL; } spin_unlock_irqrestore(&id_priv->lock, flags); return ret; } EXPORT_SYMBOL(rdma_set_reuseaddr); int rdma_set_afonly(struct rdma_cm_id *id, int afonly) { struct rdma_id_private *id_priv; unsigned long flags; int ret; id_priv = container_of(id, struct rdma_id_private, id); spin_lock_irqsave(&id_priv->lock, flags); if (id_priv->state == RDMA_CM_IDLE || id_priv->state == RDMA_CM_ADDR_BOUND) { id_priv->options |= (1 << CMA_OPTION_AFONLY); id_priv->afonly = afonly; ret = 0; } else { ret = -EINVAL; } spin_unlock_irqrestore(&id_priv->lock, flags); return ret; } EXPORT_SYMBOL(rdma_set_afonly); static void cma_bind_port(struct rdma_bind_list *bind_list, struct rdma_id_private *id_priv) { struct sockaddr *addr; struct sockaddr_ib *sib; u64 sid, mask; __be16 port; addr = cma_src_addr(id_priv); port = htons(bind_list->port); switch (addr->sa_family) { case AF_INET: ((struct sockaddr_in *) addr)->sin_port = port; break; case AF_INET6: ((struct sockaddr_in6 *) addr)->sin6_port = port; break; case AF_IB: sib = (struct sockaddr_ib *) addr; sid = be64_to_cpu(sib->sib_sid); mask = be64_to_cpu(sib->sib_sid_mask); sib->sib_sid = cpu_to_be64((sid & mask) | (u64) ntohs(port)); sib->sib_sid_mask = cpu_to_be64(~0ULL); break; } id_priv->bind_list = bind_list; hlist_add_head(&id_priv->node, &bind_list->owners); } static int cma_alloc_port(enum rdma_port_space ps, struct rdma_id_private *id_priv, unsigned short snum) { struct rdma_bind_list *bind_list; int ret; bind_list = kzalloc(sizeof *bind_list, GFP_KERNEL); if (!bind_list) return -ENOMEM; ret = cma_ps_alloc(id_priv->id.route.addr.dev_addr.net, ps, bind_list, snum); if (ret < 0) goto err; bind_list->ps = ps; bind_list->port = (unsigned short)ret; cma_bind_port(bind_list, id_priv); return 0; err: kfree(bind_list); return ret == -ENOSPC ? -EADDRNOTAVAIL : ret; } static int cma_alloc_any_port(enum rdma_port_space ps, struct rdma_id_private *id_priv) { static unsigned int last_used_port; int low, high, remaining; unsigned int rover; struct vnet *net = id_priv->id.route.addr.dev_addr.net; u32 rand; inet_get_local_port_range(net, &low, &high); remaining = (high - low) + 1; get_random_bytes(&rand, sizeof(rand)); rover = rand % remaining + low; retry: if (last_used_port != rover && !cma_ps_find(net, ps, (unsigned short)rover)) { int ret = cma_alloc_port(ps, id_priv, rover); /* * Remember previously used port number in order to avoid * re-using same port immediately after it is closed. */ if (!ret) last_used_port = rover; if (ret != -EADDRNOTAVAIL) return ret; } if (--remaining) { rover++; if ((rover < low) || (rover > high)) rover = low; goto retry; } return -EADDRNOTAVAIL; } /* * Check that the requested port is available. This is called when trying to * bind to a specific port, or when trying to listen on a bound port. In * the latter case, the provided id_priv may already be on the bind_list, but * we still need to check that it's okay to start listening. */ static int cma_check_port(struct rdma_bind_list *bind_list, struct rdma_id_private *id_priv, uint8_t reuseaddr) { struct rdma_id_private *cur_id; struct sockaddr *addr, *cur_addr; addr = cma_src_addr(id_priv); hlist_for_each_entry(cur_id, &bind_list->owners, node) { if (id_priv == cur_id) continue; if ((cur_id->state != RDMA_CM_LISTEN) && reuseaddr && cur_id->reuseaddr) continue; cur_addr = cma_src_addr(cur_id); if (id_priv->afonly && cur_id->afonly && (addr->sa_family != cur_addr->sa_family)) continue; if (cma_any_addr(addr) || cma_any_addr(cur_addr)) return -EADDRNOTAVAIL; if (!cma_addr_cmp(addr, cur_addr)) return -EADDRINUSE; } return 0; } static int cma_use_port(enum rdma_port_space ps, struct rdma_id_private *id_priv) { struct rdma_bind_list *bind_list; unsigned short snum; int ret; snum = ntohs(cma_port(cma_src_addr(id_priv))); if (snum < IPPORT_RESERVED && priv_check(curthread, PRIV_NETINET_BINDANY) != 0) return -EACCES; bind_list = cma_ps_find(id_priv->id.route.addr.dev_addr.net, ps, snum); if (!bind_list) { ret = cma_alloc_port(ps, id_priv, snum); } else { ret = cma_check_port(bind_list, id_priv, id_priv->reuseaddr); if (!ret) cma_bind_port(bind_list, id_priv); } return ret; } static int cma_bind_listen(struct rdma_id_private *id_priv) { struct rdma_bind_list *bind_list = id_priv->bind_list; int ret = 0; mutex_lock(&lock); if (bind_list->owners.first->next) ret = cma_check_port(bind_list, id_priv, 0); mutex_unlock(&lock); return ret; } static enum rdma_port_space cma_select_inet_ps( struct rdma_id_private *id_priv) { switch (id_priv->id.ps) { case RDMA_PS_TCP: case RDMA_PS_UDP: case RDMA_PS_IPOIB: case RDMA_PS_IB: return id_priv->id.ps; default: return 0; } } static enum rdma_port_space cma_select_ib_ps(struct rdma_id_private *id_priv) { enum rdma_port_space ps = 0; struct sockaddr_ib *sib; u64 sid_ps, mask, sid; sib = (struct sockaddr_ib *) cma_src_addr(id_priv); mask = be64_to_cpu(sib->sib_sid_mask) & RDMA_IB_IP_PS_MASK; sid = be64_to_cpu(sib->sib_sid) & mask; if ((id_priv->id.ps == RDMA_PS_IB) && (sid == (RDMA_IB_IP_PS_IB & mask))) { sid_ps = RDMA_IB_IP_PS_IB; ps = RDMA_PS_IB; } else if (((id_priv->id.ps == RDMA_PS_IB) || (id_priv->id.ps == RDMA_PS_TCP)) && (sid == (RDMA_IB_IP_PS_TCP & mask))) { sid_ps = RDMA_IB_IP_PS_TCP; ps = RDMA_PS_TCP; } else if (((id_priv->id.ps == RDMA_PS_IB) || (id_priv->id.ps == RDMA_PS_UDP)) && (sid == (RDMA_IB_IP_PS_UDP & mask))) { sid_ps = RDMA_IB_IP_PS_UDP; ps = RDMA_PS_UDP; } if (ps) { sib->sib_sid = cpu_to_be64(sid_ps | ntohs(cma_port((struct sockaddr *) sib))); sib->sib_sid_mask = cpu_to_be64(RDMA_IB_IP_PS_MASK | be64_to_cpu(sib->sib_sid_mask)); } return ps; } static int cma_get_port(struct rdma_id_private *id_priv) { enum rdma_port_space ps; int ret; if (cma_family(id_priv) != AF_IB) ps = cma_select_inet_ps(id_priv); else ps = cma_select_ib_ps(id_priv); if (!ps) return -EPROTONOSUPPORT; mutex_lock(&lock); if (cma_any_port(cma_src_addr(id_priv))) ret = cma_alloc_any_port(ps, id_priv); else ret = cma_use_port(ps, id_priv); mutex_unlock(&lock); return ret; } static int cma_check_linklocal(struct rdma_dev_addr *dev_addr, struct sockaddr *addr) { #ifdef INET6 struct sockaddr_in6 sin6; if (addr->sa_family != AF_INET6) return 0; sin6 = *(struct sockaddr_in6 *)addr; if (IN6_IS_SCOPE_LINKLOCAL(&sin6.sin6_addr) || IN6_IS_ADDR_MC_INTFACELOCAL(&sin6.sin6_addr)) { bool failure; CURVNET_SET_QUIET(dev_addr->net); failure = sa6_recoverscope(&sin6) || sin6.sin6_scope_id == 0; CURVNET_RESTORE(); /* check if IPv6 scope ID is not set */ if (failure) return -EINVAL; dev_addr->bound_dev_if = sin6.sin6_scope_id; } #endif return 0; } int rdma_listen(struct rdma_cm_id *id, int backlog) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (id_priv->state == RDMA_CM_IDLE) { id->route.addr.src_addr.ss_family = AF_INET; ret = rdma_bind_addr(id, cma_src_addr(id_priv)); if (ret) return ret; } if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_LISTEN)) return -EINVAL; if (id_priv->reuseaddr) { ret = cma_bind_listen(id_priv); if (ret) goto err; } id_priv->backlog = backlog; if (id->device) { if (rdma_cap_ib_cm(id->device, 1)) { ret = cma_ib_listen(id_priv); if (ret) goto err; } else if (rdma_cap_iw_cm(id->device, 1)) { ret = cma_iw_listen(id_priv, backlog); if (ret) goto err; } else { ret = -ENOSYS; goto err; } } else cma_listen_on_all(id_priv); return 0; err: id_priv->backlog = 0; cma_comp_exch(id_priv, RDMA_CM_LISTEN, RDMA_CM_ADDR_BOUND); return ret; } EXPORT_SYMBOL(rdma_listen); int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr) { struct rdma_id_private *id_priv; int ret; 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); 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); if (ret) goto err1; ret = cma_acquire_dev(id_priv, NULL); if (ret) goto err1; } if (!(id_priv->options & (1 << CMA_OPTION_AFONLY))) { if (addr->sa_family == AF_INET) id_priv->afonly = 1; #ifdef INET6 else if (addr->sa_family == AF_INET6) { CURVNET_SET_QUIET(id_priv->id.route.addr.dev_addr.net); id_priv->afonly = V_ip6_v6only; CURVNET_RESTORE(); } #endif } ret = cma_get_port(id_priv); if (ret) goto err2; return 0; err2: if (id_priv->cma_dev) cma_release_dev(id_priv); err1: cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_IDLE); return ret; } EXPORT_SYMBOL(rdma_bind_addr); static int cma_format_hdr(void *hdr, struct rdma_id_private *id_priv) { struct cma_hdr *cma_hdr; cma_hdr = hdr; cma_hdr->cma_version = CMA_VERSION; if (cma_family(id_priv) == AF_INET) { struct sockaddr_in *src4, *dst4; src4 = (struct sockaddr_in *) cma_src_addr(id_priv); dst4 = (struct sockaddr_in *) cma_dst_addr(id_priv); cma_set_ip_ver(cma_hdr, 4); cma_hdr->src_addr.ip4.addr = src4->sin_addr.s_addr; cma_hdr->dst_addr.ip4.addr = dst4->sin_addr.s_addr; cma_hdr->port = src4->sin_port; } else if (cma_family(id_priv) == AF_INET6) { struct sockaddr_in6 *src6, *dst6; src6 = (struct sockaddr_in6 *) cma_src_addr(id_priv); dst6 = (struct sockaddr_in6 *) cma_dst_addr(id_priv); cma_set_ip_ver(cma_hdr, 6); cma_hdr->src_addr.ip6 = src6->sin6_addr; cma_hdr->dst_addr.ip6 = dst6->sin6_addr; cma_hdr->port = src6->sin6_port; cma_ip6_clear_scope_id(&cma_hdr->src_addr.ip6); cma_ip6_clear_scope_id(&cma_hdr->dst_addr.ip6); } return 0; } static int cma_sidr_rep_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event) { struct rdma_id_private *id_priv = cm_id->context; struct rdma_cm_event event; struct ib_cm_sidr_rep_event_param *rep = &ib_event->param.sidr_rep_rcvd; int ret = 0; mutex_lock(&id_priv->handler_mutex); if (id_priv->state != RDMA_CM_CONNECT) goto out; memset(&event, 0, sizeof event); switch (ib_event->event) { case IB_CM_SIDR_REQ_ERROR: event.event = RDMA_CM_EVENT_UNREACHABLE; event.status = -ETIMEDOUT; break; case IB_CM_SIDR_REP_RECEIVED: event.param.ud.private_data = ib_event->private_data; event.param.ud.private_data_len = IB_CM_SIDR_REP_PRIVATE_DATA_SIZE; if (rep->status != IB_SIDR_SUCCESS) { event.event = RDMA_CM_EVENT_UNREACHABLE; event.status = ib_event->param.sidr_rep_rcvd.status; break; } ret = cma_set_qkey(id_priv, rep->qkey); if (ret) { event.event = RDMA_CM_EVENT_ADDR_ERROR; event.status = ret; break; } ret = ib_init_ah_from_path(id_priv->id.device, id_priv->id.port_num, id_priv->id.route.path_rec, &event.param.ud.ah_attr); if (ret) { event.event = RDMA_CM_EVENT_ADDR_ERROR; event.status = ret; break; } event.param.ud.qp_num = rep->qpn; event.param.ud.qkey = rep->qkey; event.event = RDMA_CM_EVENT_ESTABLISHED; event.status = 0; break; default: pr_err("RDMA CMA: unexpected IB CM event: %d\n", ib_event->event); goto out; } ret = id_priv->id.event_handler(&id_priv->id, &event); if (ret) { /* Destroy the CM ID by returning a non-zero value. */ id_priv->cm_id.ib = NULL; cma_exch(id_priv, RDMA_CM_DESTROYING); mutex_unlock(&id_priv->handler_mutex); rdma_destroy_id(&id_priv->id); return ret; } out: mutex_unlock(&id_priv->handler_mutex); return ret; } static int cma_resolve_ib_udp(struct rdma_id_private *id_priv, struct rdma_conn_param *conn_param) { struct ib_cm_sidr_req_param req; struct ib_cm_id *id; void *private_data; int offset, ret; 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) return -EINVAL; if (req.private_data_len) { private_data = kzalloc(req.private_data_len, GFP_ATOMIC); if (!private_data) return -ENOMEM; } else { private_data = NULL; } if (conn_param->private_data && conn_param->private_data_len) memcpy((char *)private_data + offset, conn_param->private_data, conn_param->private_data_len); if (private_data) { ret = cma_format_hdr(private_data, id_priv); if (ret) goto out; req.private_data = private_data; } id = ib_create_cm_id(id_priv->id.device, cma_sidr_rep_handler, id_priv); if (IS_ERR(id)) { ret = PTR_ERR(id); goto out; } id_priv->cm_id.ib = id; req.path = id_priv->id.route.path_rec; req.service_id = rdma_get_service_id(&id_priv->id, cma_dst_addr(id_priv)); req.timeout_ms = 1 << (CMA_CM_RESPONSE_TIMEOUT - 8); req.max_cm_retries = CMA_MAX_CM_RETRIES; ret = ib_send_cm_sidr_req(id_priv->cm_id.ib, &req); if (ret) { ib_destroy_cm_id(id_priv->cm_id.ib); id_priv->cm_id.ib = NULL; } out: kfree(private_data); return ret; } static int cma_connect_ib(struct rdma_id_private *id_priv, struct rdma_conn_param *conn_param) { struct ib_cm_req_param req; struct rdma_route *route; void *private_data; struct ib_cm_id *id; int offset, ret; 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) return -EINVAL; if (req.private_data_len) { private_data = kzalloc(req.private_data_len, GFP_ATOMIC); if (!private_data) return -ENOMEM; } else { private_data = NULL; } if (conn_param->private_data && conn_param->private_data_len) memcpy((char *)private_data + offset, conn_param->private_data, conn_param->private_data_len); id = ib_create_cm_id(id_priv->id.device, cma_ib_handler, id_priv); if (IS_ERR(id)) { ret = PTR_ERR(id); goto out; } id_priv->cm_id.ib = id; route = &id_priv->id.route; if (private_data) { ret = cma_format_hdr(private_data, id_priv); if (ret) goto out; req.private_data = private_data; } req.primary_path = &route->path_rec[0]; if (route->num_paths == 2) req.alternate_path = &route->path_rec[1]; req.service_id = rdma_get_service_id(&id_priv->id, cma_dst_addr(id_priv)); req.qp_num = id_priv->qp_num; req.qp_type = id_priv->id.qp_type; req.starting_psn = id_priv->seq_num; req.responder_resources = conn_param->responder_resources; req.initiator_depth = conn_param->initiator_depth; req.flow_control = conn_param->flow_control; req.retry_count = min_t(u8, 7, conn_param->retry_count); req.rnr_retry_count = min_t(u8, 7, conn_param->rnr_retry_count); req.remote_cm_response_timeout = CMA_CM_RESPONSE_TIMEOUT; req.local_cm_response_timeout = CMA_CM_RESPONSE_TIMEOUT; req.max_cm_retries = CMA_MAX_CM_RETRIES; req.srq = id_priv->srq ? 1 : 0; ret = ib_send_cm_req(id_priv->cm_id.ib, &req); out: if (ret && !IS_ERR(id)) { ib_destroy_cm_id(id); id_priv->cm_id.ib = NULL; } kfree(private_data); return ret; } static int cma_connect_iw(struct rdma_id_private *id_priv, struct rdma_conn_param *conn_param) { struct iw_cm_id *cm_id; int ret; struct iw_cm_conn_param iw_param; cm_id = iw_create_cm_id(id_priv->id.device, cma_iw_handler, id_priv); if (IS_ERR(cm_id)) return PTR_ERR(cm_id); cm_id->tos = id_priv->tos; id_priv->cm_id.iw = cm_id; memcpy(&cm_id->local_addr, cma_src_addr(id_priv), rdma_addr_size(cma_src_addr(id_priv))); memcpy(&cm_id->remote_addr, cma_dst_addr(id_priv), rdma_addr_size(cma_dst_addr(id_priv))); ret = cma_modify_qp_rtr(id_priv, conn_param); if (ret) goto out; if (conn_param) { iw_param.ord = conn_param->initiator_depth; iw_param.ird = conn_param->responder_resources; iw_param.private_data = conn_param->private_data; iw_param.private_data_len = conn_param->private_data_len; iw_param.qpn = id_priv->id.qp ? id_priv->qp_num : conn_param->qp_num; } else { memset(&iw_param, 0, sizeof iw_param); iw_param.qpn = id_priv->qp_num; } ret = iw_cm_connect(cm_id, &iw_param); out: if (ret) { iw_destroy_cm_id(cm_id); id_priv->cm_id.iw = NULL; } return ret; } int rdma_connect(struct rdma_cm_id *id, struct rdma_conn_param *conn_param) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (!cma_comp_exch(id_priv, RDMA_CM_ROUTE_RESOLVED, RDMA_CM_CONNECT)) return -EINVAL; if (!id->qp) { id_priv->qp_num = conn_param->qp_num; id_priv->srq = conn_param->srq; } if (rdma_cap_ib_cm(id->device, id->port_num)) { if (id->qp_type == IB_QPT_UD) ret = cma_resolve_ib_udp(id_priv, conn_param); else ret = cma_connect_ib(id_priv, conn_param); } else if (rdma_cap_iw_cm(id->device, id->port_num)) ret = cma_connect_iw(id_priv, conn_param); else ret = -ENOSYS; if (ret) goto err; return 0; err: cma_comp_exch(id_priv, RDMA_CM_CONNECT, RDMA_CM_ROUTE_RESOLVED); return ret; } EXPORT_SYMBOL(rdma_connect); static int cma_accept_ib(struct rdma_id_private *id_priv, struct rdma_conn_param *conn_param) { struct ib_cm_rep_param rep; int ret; ret = cma_modify_qp_rtr(id_priv, conn_param); if (ret) goto out; ret = cma_modify_qp_rts(id_priv, conn_param); if (ret) goto out; memset(&rep, 0, sizeof rep); rep.qp_num = id_priv->qp_num; rep.starting_psn = id_priv->seq_num; rep.private_data = conn_param->private_data; rep.private_data_len = conn_param->private_data_len; rep.responder_resources = conn_param->responder_resources; rep.initiator_depth = conn_param->initiator_depth; rep.failover_accepted = 0; rep.flow_control = conn_param->flow_control; rep.rnr_retry_count = min_t(u8, 7, conn_param->rnr_retry_count); rep.srq = id_priv->srq ? 1 : 0; ret = ib_send_cm_rep(id_priv->cm_id.ib, &rep); out: return ret; } static int cma_accept_iw(struct rdma_id_private *id_priv, struct rdma_conn_param *conn_param) { struct iw_cm_conn_param iw_param; int ret; ret = cma_modify_qp_rtr(id_priv, conn_param); if (ret) return ret; iw_param.ord = conn_param->initiator_depth; iw_param.ird = conn_param->responder_resources; iw_param.private_data = conn_param->private_data; iw_param.private_data_len = conn_param->private_data_len; if (id_priv->id.qp) { iw_param.qpn = id_priv->qp_num; } else iw_param.qpn = conn_param->qp_num; return iw_cm_accept(id_priv->cm_id.iw, &iw_param); } static int cma_send_sidr_rep(struct rdma_id_private *id_priv, enum ib_cm_sidr_status status, u32 qkey, const void *private_data, int private_data_len) { struct ib_cm_sidr_rep_param rep; int ret; memset(&rep, 0, sizeof rep); rep.status = status; if (status == IB_SIDR_SUCCESS) { ret = cma_set_qkey(id_priv, qkey); if (ret) return ret; rep.qp_num = id_priv->qp_num; rep.qkey = id_priv->qkey; } rep.private_data = private_data; rep.private_data_len = private_data_len; return ib_send_cm_sidr_rep(id_priv->cm_id.ib, &rep); } int rdma_accept(struct rdma_cm_id *id, struct rdma_conn_param *conn_param) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); id_priv->owner = task_pid_nr(current); if (!cma_comp(id_priv, RDMA_CM_CONNECT)) return -EINVAL; if (!id->qp && conn_param) { id_priv->qp_num = conn_param->qp_num; id_priv->srq = conn_param->srq; } if (rdma_cap_ib_cm(id->device, id->port_num)) { if (id->qp_type == IB_QPT_UD) { if (conn_param) ret = cma_send_sidr_rep(id_priv, IB_SIDR_SUCCESS, conn_param->qkey, conn_param->private_data, conn_param->private_data_len); else ret = cma_send_sidr_rep(id_priv, IB_SIDR_SUCCESS, 0, NULL, 0); } else { if (conn_param) ret = cma_accept_ib(id_priv, conn_param); else ret = cma_rep_recv(id_priv); } } else if (rdma_cap_iw_cm(id->device, id->port_num)) ret = cma_accept_iw(id_priv, conn_param); else ret = -ENOSYS; if (ret) goto reject; return 0; reject: cma_modify_qp_err(id_priv); rdma_reject(id, NULL, 0); return ret; } EXPORT_SYMBOL(rdma_accept); int rdma_notify(struct rdma_cm_id *id, enum ib_event_type event) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (!id_priv->cm_id.ib) return -EINVAL; switch (id->device->node_type) { case RDMA_NODE_IB_CA: ret = ib_cm_notify(id_priv->cm_id.ib, event); break; default: ret = 0; break; } return ret; } EXPORT_SYMBOL(rdma_notify); int rdma_reject(struct rdma_cm_id *id, const void *private_data, u8 private_data_len) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (!id_priv->cm_id.ib) return -EINVAL; if (rdma_cap_ib_cm(id->device, id->port_num)) { if (id->qp_type == IB_QPT_UD) ret = cma_send_sidr_rep(id_priv, IB_SIDR_REJECT, 0, private_data, private_data_len); else ret = ib_send_cm_rej(id_priv->cm_id.ib, IB_CM_REJ_CONSUMER_DEFINED, NULL, 0, private_data, private_data_len); } else if (rdma_cap_iw_cm(id->device, id->port_num)) { ret = iw_cm_reject(id_priv->cm_id.iw, private_data, private_data_len); } else ret = -ENOSYS; return ret; } EXPORT_SYMBOL(rdma_reject); int rdma_disconnect(struct rdma_cm_id *id) { struct rdma_id_private *id_priv; int ret; id_priv = container_of(id, struct rdma_id_private, id); if (!id_priv->cm_id.ib) return -EINVAL; if (rdma_cap_ib_cm(id->device, id->port_num)) { ret = cma_modify_qp_err(id_priv); if (ret) goto out; /* Initiate or respond to a disconnect. */ if (ib_send_cm_dreq(id_priv->cm_id.ib, NULL, 0)) ib_send_cm_drep(id_priv->cm_id.ib, NULL, 0); } else if (rdma_cap_iw_cm(id->device, id->port_num)) { ret = iw_cm_disconnect(id_priv->cm_id.iw, 0); } else ret = -EINVAL; out: return ret; } EXPORT_SYMBOL(rdma_disconnect); static int cma_ib_mc_handler(int status, struct ib_sa_multicast *multicast) { struct rdma_id_private *id_priv; struct cma_multicast *mc = multicast->context; struct rdma_cm_event event; int ret = 0; id_priv = mc->id_priv; mutex_lock(&id_priv->handler_mutex); if (id_priv->state != RDMA_CM_ADDR_BOUND && id_priv->state != RDMA_CM_ADDR_RESOLVED) goto out; if (!status) status = cma_set_qkey(id_priv, be32_to_cpu(multicast->rec.qkey)); mutex_lock(&id_priv->qp_mutex); if (!status && id_priv->id.qp) status = ib_attach_mcast(id_priv->id.qp, &multicast->rec.mgid, be16_to_cpu(multicast->rec.mlid)); mutex_unlock(&id_priv->qp_mutex); memset(&event, 0, sizeof event); event.status = status; event.param.ud.private_data = mc->context; if (!status) { struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; struct net_device *ndev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if); enum ib_gid_type gid_type = id_priv->cma_dev->default_gid_type[id_priv->id.port_num - rdma_start_port(id_priv->cma_dev->device)]; event.event = RDMA_CM_EVENT_MULTICAST_JOIN; ret = ib_init_ah_from_mcmember(id_priv->id.device, id_priv->id.port_num, &multicast->rec, ndev, gid_type, &event.param.ud.ah_attr); if (ret) event.event = RDMA_CM_EVENT_MULTICAST_ERROR; event.param.ud.qp_num = 0xFFFFFF; event.param.ud.qkey = be32_to_cpu(multicast->rec.qkey); if (ndev) dev_put(ndev); } else event.event = RDMA_CM_EVENT_MULTICAST_ERROR; ret = id_priv->id.event_handler(&id_priv->id, &event); if (ret) { cma_exch(id_priv, RDMA_CM_DESTROYING); mutex_unlock(&id_priv->handler_mutex); rdma_destroy_id(&id_priv->id); return 0; } out: mutex_unlock(&id_priv->handler_mutex); return 0; } static void cma_set_mgid(struct rdma_id_private *id_priv, struct sockaddr *addr, union ib_gid *mgid) { unsigned char mc_map[MAX_ADDR_LEN]; struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; struct sockaddr_in *sin = (struct sockaddr_in *) addr; struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) addr; if (cma_any_addr(addr)) { memset(mgid, 0, sizeof *mgid); } else if ((addr->sa_family == AF_INET6) && ((be32_to_cpu(sin6->sin6_addr.s6_addr32[0]) & 0xFFF0FFFF) == 0xFF10A01B)) { /* IPv6 address is an SA assigned MGID. */ memcpy(mgid, &sin6->sin6_addr, sizeof *mgid); } else if (addr->sa_family == AF_IB) { memcpy(mgid, &((struct sockaddr_ib *) addr)->sib_addr, sizeof *mgid); } else if (addr->sa_family == AF_INET6) { ipv6_ib_mc_map(&sin6->sin6_addr, dev_addr->broadcast, mc_map); if (id_priv->id.ps == RDMA_PS_UDP) mc_map[7] = 0x01; /* Use RDMA CM signature */ *mgid = *(union ib_gid *) (mc_map + 4); } else { ip_ib_mc_map(sin->sin_addr.s_addr, dev_addr->broadcast, mc_map); if (id_priv->id.ps == RDMA_PS_UDP) mc_map[7] = 0x01; /* Use RDMA CM signature */ *mgid = *(union ib_gid *) (mc_map + 4); } } static void cma_query_sa_classport_info_cb(int status, struct ib_class_port_info *rec, void *context) { struct class_port_info_context *cb_ctx = context; WARN_ON(!context); if (status || !rec) { pr_debug("RDMA CM: %s port %u failed query ClassPortInfo status: %d\n", cb_ctx->device->name, cb_ctx->port_num, status); goto out; } memcpy(cb_ctx->class_port_info, rec, sizeof(struct ib_class_port_info)); out: complete(&cb_ctx->done); } static int cma_query_sa_classport_info(struct ib_device *device, u8 port_num, struct ib_class_port_info *class_port_info) { struct class_port_info_context *cb_ctx; int ret; cb_ctx = kmalloc(sizeof(*cb_ctx), GFP_KERNEL); if (!cb_ctx) return -ENOMEM; cb_ctx->device = device; cb_ctx->class_port_info = class_port_info; cb_ctx->port_num = port_num; init_completion(&cb_ctx->done); ret = ib_sa_classport_info_rec_query(&sa_client, device, port_num, CMA_QUERY_CLASSPORT_INFO_TIMEOUT, GFP_KERNEL, cma_query_sa_classport_info_cb, cb_ctx, &cb_ctx->sa_query); if (ret < 0) { pr_err("RDMA CM: %s port %u failed to send ClassPortInfo query, ret: %d\n", device->name, port_num, ret); goto out; } wait_for_completion(&cb_ctx->done); out: kfree(cb_ctx); return ret; } static int cma_join_ib_multicast(struct rdma_id_private *id_priv, struct cma_multicast *mc) { struct ib_sa_mcmember_rec rec; struct ib_class_port_info class_port_info; struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; ib_sa_comp_mask comp_mask; int ret; ib_addr_get_mgid(dev_addr, &rec.mgid); ret = ib_sa_get_mcmember_rec(id_priv->id.device, id_priv->id.port_num, &rec.mgid, &rec); if (ret) return ret; ret = cma_set_qkey(id_priv, 0); if (ret) return ret; cma_set_mgid(id_priv, (struct sockaddr *) &mc->addr, &rec.mgid); rec.qkey = cpu_to_be32(id_priv->qkey); rdma_addr_get_sgid(dev_addr, &rec.port_gid); rec.pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr)); rec.join_state = mc->join_state; if (rec.join_state == BIT(SENDONLY_FULLMEMBER_JOIN)) { ret = cma_query_sa_classport_info(id_priv->id.device, id_priv->id.port_num, &class_port_info); if (ret) return ret; if (!(ib_get_cpi_capmask2(&class_port_info) & IB_SA_CAP_MASK2_SENDONLY_FULL_MEM_SUPPORT)) { pr_warn("RDMA CM: %s port %u Unable to multicast join\n" "RDMA CM: SM doesn't support Send Only Full Member option\n", id_priv->id.device->name, id_priv->id.port_num); return -EOPNOTSUPP; } } comp_mask = IB_SA_MCMEMBER_REC_MGID | IB_SA_MCMEMBER_REC_PORT_GID | IB_SA_MCMEMBER_REC_PKEY | IB_SA_MCMEMBER_REC_JOIN_STATE | IB_SA_MCMEMBER_REC_QKEY | IB_SA_MCMEMBER_REC_SL | IB_SA_MCMEMBER_REC_FLOW_LABEL | IB_SA_MCMEMBER_REC_TRAFFIC_CLASS; if (id_priv->id.ps == RDMA_PS_IPOIB) comp_mask |= IB_SA_MCMEMBER_REC_RATE | IB_SA_MCMEMBER_REC_RATE_SELECTOR | IB_SA_MCMEMBER_REC_MTU_SELECTOR | IB_SA_MCMEMBER_REC_MTU | IB_SA_MCMEMBER_REC_HOP_LIMIT; mc->multicast.ib = ib_sa_join_multicast(&sa_client, id_priv->id.device, id_priv->id.port_num, &rec, comp_mask, GFP_KERNEL, cma_ib_mc_handler, mc); return PTR_ERR_OR_ZERO(mc->multicast.ib); } static void iboe_mcast_work_handler(struct work_struct *work) { struct iboe_mcast_work *mw = container_of(work, struct iboe_mcast_work, work); struct cma_multicast *mc = mw->mc; struct ib_sa_multicast *m = mc->multicast.ib; mc->multicast.ib->context = mc; cma_ib_mc_handler(0, m); kref_put(&mc->mcref, release_mc); kfree(mw); } static void cma_iboe_set_mgid(struct sockaddr *addr, union ib_gid *mgid, enum ib_gid_type gid_type) { struct sockaddr_in *sin = (struct sockaddr_in *)addr; struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)addr; if (cma_any_addr(addr)) { memset(mgid, 0, sizeof *mgid); } else if (addr->sa_family == AF_INET6) { memcpy(mgid, &sin6->sin6_addr, sizeof *mgid); } else { mgid->raw[0] = (gid_type == IB_GID_TYPE_ROCE_UDP_ENCAP) ? 0 : 0xff; mgid->raw[1] = (gid_type == IB_GID_TYPE_ROCE_UDP_ENCAP) ? 0 : 0x0e; mgid->raw[2] = 0; mgid->raw[3] = 0; mgid->raw[4] = 0; mgid->raw[5] = 0; mgid->raw[6] = 0; mgid->raw[7] = 0; mgid->raw[8] = 0; mgid->raw[9] = 0; mgid->raw[10] = 0xff; mgid->raw[11] = 0xff; *(__be32 *)(&mgid->raw[12]) = sin->sin_addr.s_addr; } } static int cma_iboe_join_multicast(struct rdma_id_private *id_priv, struct cma_multicast *mc) { struct iboe_mcast_work *work; struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; int err = 0; struct sockaddr *addr = (struct sockaddr *)&mc->addr; struct net_device *ndev = NULL; enum ib_gid_type gid_type; bool send_only; send_only = mc->join_state == BIT(SENDONLY_FULLMEMBER_JOIN); if (cma_zero_addr((struct sockaddr *)&mc->addr)) return -EINVAL; work = kzalloc(sizeof *work, GFP_KERNEL); if (!work) return -ENOMEM; mc->multicast.ib = kzalloc(sizeof(struct ib_sa_multicast), GFP_KERNEL); if (!mc->multicast.ib) { err = -ENOMEM; goto out1; } 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(addr, &mc->multicast.ib->rec.mgid, gid_type); mc->multicast.ib->rec.pkey = cpu_to_be16(0xffff); if (id_priv->id.ps == RDMA_PS_UDP) mc->multicast.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) { err = -ENODEV; goto out2; } mc->multicast.ib->rec.rate = iboe_get_rate(ndev); mc->multicast.ib->rec.hop_limit = 1; mc->multicast.ib->rec.mtu = iboe_get_mtu(ndev->if_mtu); if (addr->sa_family == AF_INET || addr->sa_family == AF_INET6) { if (gid_type == IB_GID_TYPE_ROCE_UDP_ENCAP) { mc->multicast.ib->rec.hop_limit = IPV6_DEFAULT_HOPLIMIT; if (!send_only) { err = cma_igmp_send(ndev, &mc->multicast.ib->rec.mgid, true); if (!err) mc->igmp_joined = true; } } } else { if (gid_type == IB_GID_TYPE_ROCE_UDP_ENCAP) err = -ENOTSUPP; } dev_put(ndev); if (err || !mc->multicast.ib->rec.mtu) { if (!err) err = -EINVAL; goto out2; } rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.src_addr, &mc->multicast.ib->rec.port_gid); work->id = id_priv; work->mc = mc; INIT_WORK(&work->work, iboe_mcast_work_handler); kref_get(&mc->mcref); queue_work(cma_wq, &work->work); return 0; out2: kfree(mc->multicast.ib); out1: kfree(work); return err; } int rdma_join_multicast(struct rdma_cm_id *id, struct sockaddr *addr, u8 join_state, void *context) { struct rdma_id_private *id_priv; struct cma_multicast *mc; int ret; if (!id->device) return -EINVAL; id_priv = container_of(id, struct rdma_id_private, id); if (!cma_comp(id_priv, RDMA_CM_ADDR_BOUND) && !cma_comp(id_priv, RDMA_CM_ADDR_RESOLVED)) return -EINVAL; mc = kmalloc(sizeof *mc, GFP_KERNEL); if (!mc) return -ENOMEM; memcpy(&mc->addr, addr, rdma_addr_size(addr)); mc->context = context; mc->id_priv = id_priv; mc->igmp_joined = false; mc->join_state = join_state; spin_lock(&id_priv->lock); list_add(&mc->list, &id_priv->mc_list); spin_unlock(&id_priv->lock); if (rdma_protocol_roce(id->device, id->port_num)) { kref_init(&mc->mcref); ret = cma_iboe_join_multicast(id_priv, mc); } else if (rdma_cap_ib_mcast(id->device, id->port_num)) ret = cma_join_ib_multicast(id_priv, mc); else ret = -ENOSYS; if (ret) { spin_lock_irq(&id_priv->lock); list_del(&mc->list); spin_unlock_irq(&id_priv->lock); kfree(mc); } return ret; } EXPORT_SYMBOL(rdma_join_multicast); void rdma_leave_multicast(struct rdma_cm_id *id, struct sockaddr *addr) { struct rdma_id_private *id_priv; struct cma_multicast *mc; id_priv = container_of(id, struct rdma_id_private, id); spin_lock_irq(&id_priv->lock); list_for_each_entry(mc, &id_priv->mc_list, list) { if (!memcmp(&mc->addr, addr, rdma_addr_size(addr))) { list_del(&mc->list); spin_unlock_irq(&id_priv->lock); if (id->qp) ib_detach_mcast(id->qp, &mc->multicast.ib->rec.mgid, be16_to_cpu(mc->multicast.ib->rec.mlid)); BUG_ON(id_priv->cma_dev->device != id->device); if (rdma_cap_ib_mcast(id->device, id->port_num)) { ib_sa_free_multicast(mc->multicast.ib); kfree(mc); } else if (rdma_protocol_roce(id->device, id->port_num)) { if (mc->igmp_joined) { struct rdma_dev_addr *dev_addr = &id->route.addr.dev_addr; struct net_device *ndev = NULL; if (dev_addr->bound_dev_if) ndev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if); if (ndev) { cma_igmp_send(ndev, &mc->multicast.ib->rec.mgid, false); dev_put(ndev); } mc->igmp_joined = false; } kref_put(&mc->mcref, release_mc); } return; } } spin_unlock_irq(&id_priv->lock); } EXPORT_SYMBOL(rdma_leave_multicast); static int sysctl_cma_default_roce_mode(SYSCTL_HANDLER_ARGS) { struct cma_device *cma_dev = arg1; const int port = arg2; char buf[64]; int error; strlcpy(buf, ib_cache_gid_type_str( cma_get_default_gid_type(cma_dev, port)), sizeof(buf)); error = sysctl_handle_string(oidp, buf, sizeof(buf), req); if (error != 0 || req->newptr == NULL) goto done; error = ib_cache_gid_parse_type_str(buf); if (error < 0) { error = EINVAL; goto done; } cma_set_default_gid_type(cma_dev, port, error); error = 0; done: return (error); } static void cma_add_one(struct ib_device *device) { struct cma_device *cma_dev; struct rdma_id_private *id_priv; unsigned int i; cma_dev = kmalloc(sizeof *cma_dev, GFP_KERNEL); if (!cma_dev) return; sysctl_ctx_init(&cma_dev->sysctl_ctx); cma_dev->device = device; cma_dev->default_gid_type = kcalloc(device->phys_port_cnt, sizeof(*cma_dev->default_gid_type), GFP_KERNEL); if (!cma_dev->default_gid_type) { kfree(cma_dev); return; } for (i = rdma_start_port(device); i <= rdma_end_port(device); i++) { unsigned long supported_gids; unsigned int default_gid_type; supported_gids = roce_gid_type_mask_support(device, i); if (WARN_ON(!supported_gids)) { /* set something valid */ default_gid_type = 0; } else if (test_bit(IB_GID_TYPE_ROCE_UDP_ENCAP, &supported_gids)) { /* prefer RoCEv2, if supported */ default_gid_type = IB_GID_TYPE_ROCE_UDP_ENCAP; } else { default_gid_type = find_first_bit(&supported_gids, BITS_PER_LONG); } cma_dev->default_gid_type[i - rdma_start_port(device)] = default_gid_type; } init_completion(&cma_dev->comp); atomic_set(&cma_dev->refcount, 1); INIT_LIST_HEAD(&cma_dev->id_list); ib_set_client_data(device, &cma_client, cma_dev); mutex_lock(&lock); list_add_tail(&cma_dev->list, &dev_list); list_for_each_entry(id_priv, &listen_any_list, list) cma_listen_on_dev(id_priv, cma_dev); mutex_unlock(&lock); for (i = rdma_start_port(device); i <= rdma_end_port(device); i++) { char buf[64]; snprintf(buf, sizeof(buf), "default_roce_mode_port%d", i); (void) SYSCTL_ADD_PROC(&cma_dev->sysctl_ctx, SYSCTL_CHILDREN(device->ports_parent->parent->oidp), OID_AUTO, buf, CTLTYPE_STRING | CTLFLAG_RWTUN | CTLFLAG_MPSAFE, cma_dev, i, &sysctl_cma_default_roce_mode, "A", "Default RoCE mode. Valid values: IB/RoCE v1 and RoCE v2"); } } static int cma_remove_id_dev(struct rdma_id_private *id_priv) { struct rdma_cm_event event; enum rdma_cm_state state; int ret = 0; /* Record that we want to remove the device */ state = cma_exch(id_priv, RDMA_CM_DEVICE_REMOVAL); if (state == RDMA_CM_DESTROYING) return 0; cma_cancel_operation(id_priv, state); mutex_lock(&id_priv->handler_mutex); /* Check for destruction from another callback. */ if (!cma_comp(id_priv, RDMA_CM_DEVICE_REMOVAL)) goto out; memset(&event, 0, sizeof event); event.event = RDMA_CM_EVENT_DEVICE_REMOVAL; ret = id_priv->id.event_handler(&id_priv->id, &event); out: mutex_unlock(&id_priv->handler_mutex); return ret; } static void cma_process_remove(struct cma_device *cma_dev) { struct rdma_id_private *id_priv; int ret; mutex_lock(&lock); while (!list_empty(&cma_dev->id_list)) { id_priv = list_entry(cma_dev->id_list.next, struct rdma_id_private, list); list_del(&id_priv->listen_list); list_del_init(&id_priv->list); atomic_inc(&id_priv->refcount); mutex_unlock(&lock); ret = id_priv->internal_id ? 1 : cma_remove_id_dev(id_priv); cma_deref_id(id_priv); if (ret) rdma_destroy_id(&id_priv->id); mutex_lock(&lock); } mutex_unlock(&lock); cma_deref_dev(cma_dev); wait_for_completion(&cma_dev->comp); } static void cma_remove_one(struct ib_device *device, void *client_data) { struct cma_device *cma_dev = client_data; if (!cma_dev) return; mutex_lock(&lock); list_del(&cma_dev->list); mutex_unlock(&lock); cma_process_remove(cma_dev); sysctl_ctx_free(&cma_dev->sysctl_ctx); kfree(cma_dev->default_gid_type); kfree(cma_dev); } static void cma_init_vnet(void *arg) { struct cma_pernet *pernet = &VNET(cma_pernet); idr_init(&pernet->tcp_ps); idr_init(&pernet->udp_ps); idr_init(&pernet->ipoib_ps); idr_init(&pernet->ib_ps); } VNET_SYSINIT(cma_init_vnet, SI_SUB_OFED_MODINIT - 1, SI_ORDER_FIRST, cma_init_vnet, NULL); static void cma_destroy_vnet(void *arg) { struct cma_pernet *pernet = &VNET(cma_pernet); idr_destroy(&pernet->tcp_ps); idr_destroy(&pernet->udp_ps); idr_destroy(&pernet->ipoib_ps); idr_destroy(&pernet->ib_ps); } VNET_SYSUNINIT(cma_destroy_vnet, SI_SUB_OFED_MODINIT - 1, SI_ORDER_SECOND, cma_destroy_vnet, NULL); static int __init cma_init(void) { int ret; cma_wq = alloc_ordered_workqueue("rdma_cm", WQ_MEM_RECLAIM); if (!cma_wq) return -ENOMEM; ib_sa_register_client(&sa_client); rdma_addr_register_client(&addr_client); ret = ib_register_client(&cma_client); if (ret) goto err; cma_configfs_init(); return 0; err: rdma_addr_unregister_client(&addr_client); ib_sa_unregister_client(&sa_client); destroy_workqueue(cma_wq); return ret; } static void __exit cma_cleanup(void) { cma_configfs_exit(); ib_unregister_client(&cma_client); rdma_addr_unregister_client(&addr_client); ib_sa_unregister_client(&sa_client); destroy_workqueue(cma_wq); } module_init(cma_init); module_exit(cma_cleanup); Index: head/sys/ofed/drivers/infiniband/core/ib_roce_gid_mgmt.c =================================================================== --- head/sys/ofed/drivers/infiniband/core/ib_roce_gid_mgmt.c (revision 338525) +++ head/sys/ofed/drivers/infiniband/core/ib_roce_gid_mgmt.c (revision 338526) @@ -1,486 +1,476 @@ /*- * SPDX-License-Identifier: BSD-2-Clause OR GPL-2.0 * * Copyright (c) 2015-2017, Mellanox Technologies inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU * General Public License (GPL) Version 2, available from the file * COPYING in the main directory of this source tree, or the * OpenIB.org BSD license below: * * Redistribution and use in source and binary forms, with or * without modification, are permitted provided that the following * conditions are met: * * - Redistributions of source code must retain the above * copyright notice, this list of conditions and the following * disclaimer. * * - Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following * disclaimer in the documentation and/or other materials * provided with the distribution. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. */ #include __FBSDID("$FreeBSD$"); #include "core_priv.h" #include #include #include #include #include #include static struct workqueue_struct *roce_gid_mgmt_wq; enum gid_op_type { GID_DEL = 0, GID_ADD }; struct roce_netdev_event_work { struct work_struct work; struct net_device *ndev; }; struct roce_rescan_work { struct work_struct work; struct ib_device *ib_dev; }; static const struct { bool (*is_supported)(const struct ib_device *device, u8 port_num); enum ib_gid_type gid_type; } PORT_CAP_TO_GID_TYPE[] = { {rdma_protocol_roce_eth_encap, IB_GID_TYPE_ROCE}, {rdma_protocol_roce_udp_encap, IB_GID_TYPE_ROCE_UDP_ENCAP}, }; #define CAP_TO_GID_TABLE_SIZE ARRAY_SIZE(PORT_CAP_TO_GID_TYPE) unsigned long roce_gid_type_mask_support(struct ib_device *ib_dev, u8 port) { int i; unsigned int ret_flags = 0; if (!rdma_protocol_roce(ib_dev, port)) return 1UL << IB_GID_TYPE_IB; for (i = 0; i < CAP_TO_GID_TABLE_SIZE; i++) if (PORT_CAP_TO_GID_TYPE[i].is_supported(ib_dev, port)) ret_flags |= 1UL << PORT_CAP_TO_GID_TYPE[i].gid_type; return ret_flags; } EXPORT_SYMBOL(roce_gid_type_mask_support); static void update_gid(enum gid_op_type gid_op, struct ib_device *ib_dev, u8 port, union ib_gid *gid, struct net_device *ndev) { int i; unsigned long gid_type_mask = roce_gid_type_mask_support(ib_dev, port); struct ib_gid_attr gid_attr; memset(&gid_attr, 0, sizeof(gid_attr)); gid_attr.ndev = ndev; for (i = 0; i != IB_GID_TYPE_SIZE; i++) { if ((1UL << i) & gid_type_mask) { gid_attr.gid_type = i; switch (gid_op) { case GID_ADD: ib_cache_gid_add(ib_dev, port, gid, &gid_attr); break; case GID_DEL: ib_cache_gid_del(ib_dev, port, gid, &gid_attr); break; } } } } static int roce_gid_match_netdev(struct ib_device *ib_dev, u8 port, struct net_device *idev, void *cookie) { struct net_device *ndev = (struct net_device *)cookie; if (idev == NULL) return (0); return (ndev == idev); } static int roce_gid_match_all(struct ib_device *ib_dev, u8 port, struct net_device *idev, void *cookie) { if (idev == NULL) return (0); return (1); } static int roce_gid_enum_netdev_default(struct ib_device *ib_dev, u8 port, struct net_device *idev) { unsigned long gid_type_mask; gid_type_mask = roce_gid_type_mask_support(ib_dev, port); ib_cache_gid_set_default_gid(ib_dev, port, idev, gid_type_mask, IB_CACHE_GID_DEFAULT_MODE_SET); return (hweight_long(gid_type_mask)); } -#define ETH_IPOIB_DRV_NAME "ib" - -static inline int -is_eth_ipoib_intf(struct net_device *dev) -{ - if (strcmp(dev->if_dname, ETH_IPOIB_DRV_NAME)) - return 0; - return 1; -} - static void roce_gid_update_addr_callback(struct ib_device *device, u8 port, struct net_device *ndev, void *cookie) { struct ipx_entry { STAILQ_ENTRY(ipx_entry) entry; union ipx_addr { struct sockaddr sa[0]; struct sockaddr_in v4; struct sockaddr_in6 v6; } ipx_addr; struct net_device *ndev; }; struct ipx_entry *entry; struct net_device *idev; #if defined(INET) || defined(INET6) struct ifaddr *ifa; #endif struct ib_gid_attr gid_attr; union ib_gid gid; int default_gids; u16 index_num; int i; STAILQ_HEAD(, ipx_entry) ipx_head; STAILQ_INIT(&ipx_head); /* make sure default GIDs are in */ default_gids = roce_gid_enum_netdev_default(device, port, ndev); CURVNET_SET(ndev->if_vnet); IFNET_RLOCK(); CK_STAILQ_FOREACH(idev, &V_ifnet, if_link) { if (idev != ndev) { if (idev->if_type != IFT_L2VLAN) continue; if (ndev != rdma_vlan_dev_real_dev(idev)) continue; } /* clone address information for IPv4 and IPv6 */ IF_ADDR_RLOCK(idev); #if defined(INET) CK_STAILQ_FOREACH(ifa, &idev->if_addrhead, ifa_link) { if (ifa->ifa_addr == NULL || ifa->ifa_addr->sa_family != AF_INET) continue; entry = kzalloc(sizeof(*entry), GFP_ATOMIC); if (entry == NULL) { pr_warn("roce_gid_update_addr_callback: " "couldn't allocate entry for IPv4 update\n"); continue; } entry->ipx_addr.v4 = *((struct sockaddr_in *)ifa->ifa_addr); entry->ndev = idev; STAILQ_INSERT_TAIL(&ipx_head, entry, entry); } #endif #if defined(INET6) CK_STAILQ_FOREACH(ifa, &idev->if_addrhead, ifa_link) { if (ifa->ifa_addr == NULL || ifa->ifa_addr->sa_family != AF_INET6) continue; entry = kzalloc(sizeof(*entry), GFP_ATOMIC); if (entry == NULL) { pr_warn("roce_gid_update_addr_callback: " "couldn't allocate entry for IPv6 update\n"); continue; } entry->ipx_addr.v6 = *((struct sockaddr_in6 *)ifa->ifa_addr); entry->ndev = idev; /* trash IPv6 scope ID */ sa6_recoverscope(&entry->ipx_addr.v6); entry->ipx_addr.v6.sin6_scope_id = 0; STAILQ_INSERT_TAIL(&ipx_head, entry, entry); } #endif IF_ADDR_RUNLOCK(idev); } IFNET_RUNLOCK(); CURVNET_RESTORE(); /* add missing GIDs, if any */ STAILQ_FOREACH(entry, &ipx_head, entry) { unsigned long gid_type_mask = roce_gid_type_mask_support(device, port); if (rdma_ip2gid(&entry->ipx_addr.sa[0], &gid) != 0) continue; for (i = 0; i != IB_GID_TYPE_SIZE; i++) { if (!((1UL << i) & gid_type_mask)) continue; /* check if entry found */ if (ib_find_cached_gid_by_port(device, &gid, i, port, entry->ndev, &index_num) == 0) break; } if (i != IB_GID_TYPE_SIZE) continue; /* add new GID */ update_gid(GID_ADD, device, port, &gid, entry->ndev); } /* remove stale GIDs, if any */ for (i = default_gids; ib_get_cached_gid(device, port, i, &gid, &gid_attr) == 0; i++) { union ipx_addr ipx; /* check for valid network device pointer */ ndev = gid_attr.ndev; if (ndev == NULL) continue; dev_put(ndev); /* don't delete empty entries */ if (memcmp(&gid, &zgid, sizeof(zgid)) == 0) continue; /* zero default */ memset(&ipx, 0, sizeof(ipx)); rdma_gid2ip(&ipx.sa[0], &gid); STAILQ_FOREACH(entry, &ipx_head, entry) { if (entry->ndev == ndev && memcmp(&entry->ipx_addr, &ipx, sizeof(ipx)) == 0) break; } /* check if entry found */ if (entry != NULL) continue; /* remove GID */ update_gid(GID_DEL, device, port, &gid, ndev); } while ((entry = STAILQ_FIRST(&ipx_head))) { STAILQ_REMOVE_HEAD(&ipx_head, entry); kfree(entry); } } static void roce_gid_queue_scan_event_handler(struct work_struct *_work) { struct roce_netdev_event_work *work = container_of(_work, struct roce_netdev_event_work, work); ib_enum_all_roce_netdevs(roce_gid_match_netdev, work->ndev, roce_gid_update_addr_callback, NULL); dev_put(work->ndev); kfree(work); } static void roce_gid_queue_scan_event(struct net_device *ndev) { struct roce_netdev_event_work *work; retry: - if (is_eth_ipoib_intf(ndev)) - return; - - if (ndev->if_type != IFT_ETHER) { - if (ndev->if_type == IFT_L2VLAN) { - ndev = rdma_vlan_dev_real_dev(ndev); - if (ndev != NULL) - goto retry; - } + switch (ndev->if_type) { + case IFT_ETHER: + break; + case IFT_L2VLAN: + ndev = rdma_vlan_dev_real_dev(ndev); + if (ndev != NULL) + goto retry; + /* FALLTHROUGH */ + default: return; } work = kmalloc(sizeof(*work), GFP_ATOMIC); if (!work) { pr_warn("roce_gid_mgmt: Couldn't allocate work for addr_event\n"); return; } INIT_WORK(&work->work, roce_gid_queue_scan_event_handler); dev_hold(ndev); work->ndev = ndev; queue_work(roce_gid_mgmt_wq, &work->work); } static void roce_gid_delete_all_event_handler(struct work_struct *_work) { struct roce_netdev_event_work *work = container_of(_work, struct roce_netdev_event_work, work); ib_cache_gid_del_all_by_netdev(work->ndev); dev_put(work->ndev); kfree(work); } static void roce_gid_delete_all_event(struct net_device *ndev) { struct roce_netdev_event_work *work; work = kmalloc(sizeof(*work), GFP_ATOMIC); if (!work) { pr_warn("roce_gid_mgmt: Couldn't allocate work for addr_event\n"); return; } INIT_WORK(&work->work, roce_gid_delete_all_event_handler); dev_hold(ndev); work->ndev = ndev; queue_work(roce_gid_mgmt_wq, &work->work); /* make sure job is complete before returning */ flush_workqueue(roce_gid_mgmt_wq); } static int inetaddr_event(struct notifier_block *this, unsigned long event, void *ptr) { struct net_device *ndev = ptr; switch (event) { case NETDEV_UNREGISTER: roce_gid_delete_all_event(ndev); break; case NETDEV_REGISTER: case NETDEV_CHANGEADDR: case NETDEV_CHANGEIFADDR: roce_gid_queue_scan_event(ndev); break; default: break; } return NOTIFY_DONE; } static struct notifier_block nb_inetaddr = { .notifier_call = inetaddr_event }; static eventhandler_tag eh_ifnet_event; static void roce_ifnet_event(void *arg, struct ifnet *ifp, int event) { if (event != IFNET_EVENT_PCP || is_vlan_dev(ifp)) return; /* make sure GID table is reloaded */ roce_gid_delete_all_event(ifp); roce_gid_queue_scan_event(ifp); } static void roce_rescan_device_handler(struct work_struct *_work) { struct roce_rescan_work *work = container_of(_work, struct roce_rescan_work, work); ib_enum_roce_netdev(work->ib_dev, roce_gid_match_all, NULL, roce_gid_update_addr_callback, NULL); kfree(work); } /* Caller must flush system workqueue before removing the ib_device */ int roce_rescan_device(struct ib_device *ib_dev) { struct roce_rescan_work *work = kmalloc(sizeof(*work), GFP_KERNEL); if (!work) return -ENOMEM; work->ib_dev = ib_dev; INIT_WORK(&work->work, roce_rescan_device_handler); queue_work(roce_gid_mgmt_wq, &work->work); return 0; } int __init roce_gid_mgmt_init(void) { roce_gid_mgmt_wq = alloc_ordered_workqueue("roce_gid_mgmt_wq", 0); if (!roce_gid_mgmt_wq) { pr_warn("roce_gid_mgmt: can't allocate work queue\n"); return -ENOMEM; } register_inetaddr_notifier(&nb_inetaddr); /* * We rely on the netdevice notifier to enumerate all existing * devices in the system. Register to this notifier last to * make sure we will not miss any IP add/del callbacks. */ register_netdevice_notifier(&nb_inetaddr); eh_ifnet_event = EVENTHANDLER_REGISTER(ifnet_event, roce_ifnet_event, NULL, EVENTHANDLER_PRI_ANY); return 0; } void __exit roce_gid_mgmt_cleanup(void) { if (eh_ifnet_event != NULL) EVENTHANDLER_DEREGISTER(ifnet_event, eh_ifnet_event); unregister_inetaddr_notifier(&nb_inetaddr); unregister_netdevice_notifier(&nb_inetaddr); /* * Ensure all gid deletion tasks complete before we go down, * to avoid any reference to free'd memory. By the time * ib-core is removed, all physical devices have been removed, * so no issue with remaining hardware contexts. */ synchronize_rcu(); drain_workqueue(roce_gid_mgmt_wq); destroy_workqueue(roce_gid_mgmt_wq); } Index: head/sys/ofed/drivers/infiniband/ulp/ipoib/ipoib_main.c =================================================================== --- head/sys/ofed/drivers/infiniband/ulp/ipoib/ipoib_main.c (revision 338525) +++ head/sys/ofed/drivers/infiniband/ulp/ipoib/ipoib_main.c (revision 338526) @@ -1,1565 +1,1722 @@ /*- * SPDX-License-Identifier: BSD-2-Clause OR GPL-2.0 * * Copyright (c) 2004 Topspin Communications. All rights reserved. * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. * Copyright (c) 2004 Voltaire, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU * General Public License (GPL) Version 2, available from the file * COPYING in the main directory of this source tree, or the * OpenIB.org BSD license below: * * Redistribution and use in source and binary forms, with or * without modification, are permitted provided that the following * conditions are met: * * - Redistributions of source code must retain the above * copyright notice, this list of conditions and the following * disclaimer. * * - Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following * disclaimer in the documentation and/or other materials * provided with the distribution. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. */ #include __FBSDID("$FreeBSD$"); #include "ipoib.h" static int ipoib_resolvemulti(struct ifnet *, struct sockaddr **, struct sockaddr *); #include #include #include #include #include /* For ARPHRD_xxx */ #include #include #include +#include + MODULE_AUTHOR("Roland Dreier"); MODULE_DESCRIPTION("IP-over-InfiniBand net driver"); MODULE_LICENSE("Dual BSD/GPL"); int ipoib_sendq_size = IPOIB_TX_RING_SIZE; int ipoib_recvq_size = IPOIB_RX_RING_SIZE; module_param_named(send_queue_size, ipoib_sendq_size, int, 0444); MODULE_PARM_DESC(send_queue_size, "Number of descriptors in send queue"); module_param_named(recv_queue_size, ipoib_recvq_size, int, 0444); MODULE_PARM_DESC(recv_queue_size, "Number of descriptors in receive queue"); #ifdef CONFIG_INFINIBAND_IPOIB_DEBUG int ipoib_debug_level = 1; module_param_named(debug_level, ipoib_debug_level, int, 0644); MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0"); #endif struct ipoib_path_iter { struct ipoib_dev_priv *priv; struct ipoib_path path; }; static const u8 ipv4_bcast_addr[] = { 0x00, 0xff, 0xff, 0xff, 0xff, 0x12, 0x40, 0x1b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0xff, 0xff, 0xff }; struct workqueue_struct *ipoib_workqueue; struct ib_sa_client ipoib_sa_client; static void ipoib_add_one(struct ib_device *device); static void ipoib_remove_one(struct ib_device *device, void *client_data); +static struct net_device *ipoib_get_net_dev_by_params( + struct ib_device *dev, u8 port, u16 pkey, + const union ib_gid *gid, const struct sockaddr *addr, + void *client_data); static void ipoib_start(struct ifnet *dev); static int ipoib_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, struct route *ro); static int ipoib_ioctl(struct ifnet *ifp, u_long command, caddr_t data); static void ipoib_input(struct ifnet *ifp, struct mbuf *m); #define IPOIB_MTAP(_ifp, _m) \ do { \ if (bpf_peers_present((_ifp)->if_bpf)) { \ M_ASSERTVALID(_m); \ ipoib_mtap_mb((_ifp), (_m)); \ } \ } while (0) static struct unrhdr *ipoib_unrhdr; static void ipoib_unrhdr_init(void *arg) { ipoib_unrhdr = new_unrhdr(0, 65535, NULL); } SYSINIT(ipoib_unrhdr_init, SI_SUB_KLD - 1, SI_ORDER_ANY, ipoib_unrhdr_init, NULL); static void ipoib_unrhdr_uninit(void *arg) { if (ipoib_unrhdr != NULL) { struct unrhdr *hdr; hdr = ipoib_unrhdr; ipoib_unrhdr = NULL; delete_unrhdr(hdr); } } SYSUNINIT(ipoib_unrhdr_uninit, SI_SUB_KLD - 1, SI_ORDER_ANY, ipoib_unrhdr_uninit, NULL); /* * This is for clients that have an ipoib_header in the mbuf. */ static void ipoib_mtap_mb(struct ifnet *ifp, struct mbuf *mb) { struct ipoib_header *ih; struct ether_header eh; ih = mtod(mb, struct ipoib_header *); eh.ether_type = ih->proto; bcopy(ih->hwaddr, &eh.ether_dhost, ETHER_ADDR_LEN); bzero(&eh.ether_shost, ETHER_ADDR_LEN); mb->m_data += sizeof(struct ipoib_header); mb->m_len -= sizeof(struct ipoib_header); bpf_mtap2(ifp->if_bpf, &eh, sizeof(eh), mb); mb->m_data -= sizeof(struct ipoib_header); mb->m_len += sizeof(struct ipoib_header); } void ipoib_mtap_proto(struct ifnet *ifp, struct mbuf *mb, uint16_t proto) { struct ether_header eh; eh.ether_type = proto; bzero(&eh.ether_shost, ETHER_ADDR_LEN); bzero(&eh.ether_dhost, ETHER_ADDR_LEN); bpf_mtap2(ifp->if_bpf, &eh, sizeof(eh), mb); } static struct ib_client ipoib_client = { .name = "ipoib", .add = ipoib_add_one, - .remove = ipoib_remove_one + .remove = ipoib_remove_one, + .get_net_dev_by_params = ipoib_get_net_dev_by_params, }; int ipoib_open(struct ipoib_dev_priv *priv) { struct ifnet *dev = priv->dev; ipoib_dbg(priv, "bringing up interface\n"); set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); if (ipoib_pkey_dev_delay_open(priv)) return 0; if (ipoib_ib_dev_open(priv)) goto err_disable; if (ipoib_ib_dev_up(priv)) goto err_stop; if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { struct ipoib_dev_priv *cpriv; /* Bring up any child interfaces too */ mutex_lock(&priv->vlan_mutex); list_for_each_entry(cpriv, &priv->child_intfs, list) if ((cpriv->dev->if_drv_flags & IFF_DRV_RUNNING) == 0) ipoib_open(cpriv); mutex_unlock(&priv->vlan_mutex); } dev->if_drv_flags |= IFF_DRV_RUNNING; dev->if_drv_flags &= ~IFF_DRV_OACTIVE; return 0; err_stop: ipoib_ib_dev_stop(priv, 1); err_disable: clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); return -EINVAL; } static void ipoib_init(void *arg) { struct ifnet *dev; struct ipoib_dev_priv *priv; priv = arg; dev = priv->dev; if ((dev->if_drv_flags & IFF_DRV_RUNNING) == 0) ipoib_open(priv); queue_work(ipoib_workqueue, &priv->flush_light); } static int ipoib_stop(struct ipoib_dev_priv *priv) { struct ifnet *dev = priv->dev; ipoib_dbg(priv, "stopping interface\n"); clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); dev->if_drv_flags &= ~(IFF_DRV_RUNNING | IFF_DRV_OACTIVE); ipoib_ib_dev_down(priv, 0); ipoib_ib_dev_stop(priv, 0); if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { struct ipoib_dev_priv *cpriv; /* Bring down any child interfaces too */ mutex_lock(&priv->vlan_mutex); list_for_each_entry(cpriv, &priv->child_intfs, list) if ((cpriv->dev->if_drv_flags & IFF_DRV_RUNNING) != 0) ipoib_stop(cpriv); mutex_unlock(&priv->vlan_mutex); } return 0; } int ipoib_change_mtu(struct ipoib_dev_priv *priv, int new_mtu) { struct ifnet *dev = priv->dev; /* dev->if_mtu > 2K ==> connected mode */ if (ipoib_cm_admin_enabled(priv)) { if (new_mtu > IPOIB_CM_MTU(ipoib_cm_max_mtu(priv))) return -EINVAL; if (new_mtu > priv->mcast_mtu) ipoib_warn(priv, "mtu > %d will cause multicast packet drops.\n", priv->mcast_mtu); dev->if_mtu = new_mtu; return 0; } if (new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu)) return -EINVAL; priv->admin_mtu = new_mtu; dev->if_mtu = min(priv->mcast_mtu, priv->admin_mtu); queue_work(ipoib_workqueue, &priv->flush_light); return 0; } static int ipoib_ioctl(struct ifnet *ifp, u_long command, caddr_t data) { struct ipoib_dev_priv *priv = ifp->if_softc; struct ifaddr *ifa = (struct ifaddr *) data; struct ifreq *ifr = (struct ifreq *) data; int error = 0; /* check if detaching */ if (priv == NULL || priv->gone != 0) return (ENXIO); switch (command) { case SIOCSIFFLAGS: if (ifp->if_flags & IFF_UP) { if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) error = -ipoib_open(priv); } else if (ifp->if_drv_flags & IFF_DRV_RUNNING) ipoib_stop(priv); break; case SIOCADDMULTI: case SIOCDELMULTI: if (ifp->if_drv_flags & IFF_DRV_RUNNING) queue_work(ipoib_workqueue, &priv->restart_task); break; case SIOCSIFADDR: ifp->if_flags |= IFF_UP; switch (ifa->ifa_addr->sa_family) { #ifdef INET case AF_INET: ifp->if_init(ifp->if_softc); /* before arpwhohas */ arp_ifinit(ifp, ifa); break; #endif default: ifp->if_init(ifp->if_softc); break; } break; case SIOCGIFADDR: bcopy(IF_LLADDR(ifp), &ifr->ifr_addr.sa_data[0], INFINIBAND_ALEN); break; case SIOCSIFMTU: /* * Set the interface MTU. */ error = -ipoib_change_mtu(priv, ifr->ifr_mtu); break; default: error = EINVAL; break; } return (error); } static struct ipoib_path * __path_find(struct ipoib_dev_priv *priv, void *gid) { struct rb_node *n = priv->path_tree.rb_node; struct ipoib_path *path; int ret; while (n) { path = rb_entry(n, struct ipoib_path, rb_node); ret = memcmp(gid, path->pathrec.dgid.raw, sizeof (union ib_gid)); if (ret < 0) n = n->rb_left; else if (ret > 0) n = n->rb_right; else return path; } return NULL; } static int __path_add(struct ipoib_dev_priv *priv, struct ipoib_path *path) { struct rb_node **n = &priv->path_tree.rb_node; struct rb_node *pn = NULL; struct ipoib_path *tpath; int ret; while (*n) { pn = *n; tpath = rb_entry(pn, struct ipoib_path, rb_node); ret = memcmp(path->pathrec.dgid.raw, tpath->pathrec.dgid.raw, sizeof (union ib_gid)); if (ret < 0) n = &pn->rb_left; else if (ret > 0) n = &pn->rb_right; else return -EEXIST; } rb_link_node(&path->rb_node, pn, n); rb_insert_color(&path->rb_node, &priv->path_tree); list_add_tail(&path->list, &priv->path_list); return 0; } void ipoib_path_free(struct ipoib_dev_priv *priv, struct ipoib_path *path) { _IF_DRAIN(&path->queue); if (path->ah) ipoib_put_ah(path->ah); if (ipoib_cm_get(path)) ipoib_cm_destroy_tx(ipoib_cm_get(path)); kfree(path); } #ifdef CONFIG_INFINIBAND_IPOIB_DEBUG struct ipoib_path_iter * ipoib_path_iter_init(struct ipoib_dev_priv *priv) { struct ipoib_path_iter *iter; iter = kmalloc(sizeof *iter, GFP_KERNEL); if (!iter) return NULL; iter->priv = priv; memset(iter->path.pathrec.dgid.raw, 0, 16); if (ipoib_path_iter_next(iter)) { kfree(iter); return NULL; } return iter; } int ipoib_path_iter_next(struct ipoib_path_iter *iter) { struct ipoib_dev_priv *priv = iter->priv; struct rb_node *n; struct ipoib_path *path; int ret = 1; spin_lock_irq(&priv->lock); n = rb_first(&priv->path_tree); while (n) { path = rb_entry(n, struct ipoib_path, rb_node); if (memcmp(iter->path.pathrec.dgid.raw, path->pathrec.dgid.raw, sizeof (union ib_gid)) < 0) { iter->path = *path; ret = 0; break; } n = rb_next(n); } spin_unlock_irq(&priv->lock); return ret; } void ipoib_path_iter_read(struct ipoib_path_iter *iter, struct ipoib_path *path) { *path = iter->path; } #endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */ void ipoib_mark_paths_invalid(struct ipoib_dev_priv *priv) { struct ipoib_path *path, *tp; spin_lock_irq(&priv->lock); list_for_each_entry_safe(path, tp, &priv->path_list, list) { ipoib_dbg(priv, "mark path LID 0x%04x GID %16D invalid\n", be16_to_cpu(path->pathrec.dlid), path->pathrec.dgid.raw, ":"); path->valid = 0; } spin_unlock_irq(&priv->lock); } void ipoib_flush_paths(struct ipoib_dev_priv *priv) { struct ipoib_path *path, *tp; LIST_HEAD(remove_list); unsigned long flags; spin_lock_irqsave(&priv->lock, flags); list_splice_init(&priv->path_list, &remove_list); list_for_each_entry(path, &remove_list, list) rb_erase(&path->rb_node, &priv->path_tree); list_for_each_entry_safe(path, tp, &remove_list, list) { if (path->query) ib_sa_cancel_query(path->query_id, path->query); spin_unlock_irqrestore(&priv->lock, flags); wait_for_completion(&path->done); ipoib_path_free(priv, path); spin_lock_irqsave(&priv->lock, flags); } spin_unlock_irqrestore(&priv->lock, flags); } static void path_rec_completion(int status, struct ib_sa_path_rec *pathrec, void *path_ptr) { struct ipoib_path *path = path_ptr; struct ipoib_dev_priv *priv = path->priv; struct ifnet *dev = priv->dev; struct ipoib_ah *ah = NULL; struct ipoib_ah *old_ah = NULL; struct ifqueue mbqueue; struct mbuf *mb; unsigned long flags; if (!status) ipoib_dbg(priv, "PathRec LID 0x%04x for GID %16D\n", be16_to_cpu(pathrec->dlid), pathrec->dgid.raw, ":"); else ipoib_dbg(priv, "PathRec status %d for GID %16D\n", status, path->pathrec.dgid.raw, ":"); bzero(&mbqueue, sizeof(mbqueue)); if (!status) { struct ib_ah_attr av; if (!ib_init_ah_from_path(priv->ca, priv->port, pathrec, &av)) ah = ipoib_create_ah(priv, priv->pd, &av); } spin_lock_irqsave(&priv->lock, flags); if (ah) { path->pathrec = *pathrec; old_ah = path->ah; path->ah = ah; ipoib_dbg(priv, "created address handle %p for LID 0x%04x, SL %d\n", ah, be16_to_cpu(pathrec->dlid), pathrec->sl); for (;;) { _IF_DEQUEUE(&path->queue, mb); if (mb == NULL) break; _IF_ENQUEUE(&mbqueue, mb); } #ifdef CONFIG_INFINIBAND_IPOIB_CM if (ipoib_cm_enabled(priv, path->hwaddr) && !ipoib_cm_get(path)) ipoib_cm_set(path, ipoib_cm_create_tx(priv, path)); #endif path->valid = 1; } path->query = NULL; complete(&path->done); spin_unlock_irqrestore(&priv->lock, flags); if (old_ah) ipoib_put_ah(old_ah); for (;;) { _IF_DEQUEUE(&mbqueue, mb); if (mb == NULL) break; mb->m_pkthdr.rcvif = dev; if (dev->if_transmit(dev, mb)) ipoib_warn(priv, "dev_queue_xmit failed " "to requeue packet\n"); } } static struct ipoib_path * path_rec_create(struct ipoib_dev_priv *priv, uint8_t *hwaddr) { struct ipoib_path *path; if (!priv->broadcast) return NULL; path = kzalloc(sizeof *path, GFP_ATOMIC); if (!path) return NULL; path->priv = priv; bzero(&path->queue, sizeof(path->queue)); #ifdef CONFIG_INFINIBAND_IPOIB_CM memcpy(&path->hwaddr, hwaddr, INFINIBAND_ALEN); #endif memcpy(path->pathrec.dgid.raw, &hwaddr[4], sizeof (union ib_gid)); path->pathrec.sgid = priv->local_gid; path->pathrec.pkey = cpu_to_be16(priv->pkey); path->pathrec.numb_path = 1; path->pathrec.traffic_class = priv->broadcast->mcmember.traffic_class; return path; } static int path_rec_start(struct ipoib_dev_priv *priv, struct ipoib_path *path) { struct ifnet *dev = priv->dev; ib_sa_comp_mask comp_mask = IB_SA_PATH_REC_MTU_SELECTOR | IB_SA_PATH_REC_MTU; struct ib_sa_path_rec p_rec; p_rec = path->pathrec; p_rec.mtu_selector = IB_SA_GT; switch (roundup_pow_of_two(dev->if_mtu + IPOIB_ENCAP_LEN)) { case 512: p_rec.mtu = IB_MTU_256; break; case 1024: p_rec.mtu = IB_MTU_512; break; case 2048: p_rec.mtu = IB_MTU_1024; break; case 4096: p_rec.mtu = IB_MTU_2048; break; default: /* Wildcard everything */ comp_mask = 0; p_rec.mtu = 0; p_rec.mtu_selector = 0; } ipoib_dbg(priv, "Start path record lookup for %16D MTU > %d\n", p_rec.dgid.raw, ":", comp_mask ? ib_mtu_enum_to_int(p_rec.mtu) : 0); init_completion(&path->done); path->query_id = ib_sa_path_rec_get(&ipoib_sa_client, priv->ca, priv->port, &p_rec, comp_mask | IB_SA_PATH_REC_DGID | IB_SA_PATH_REC_SGID | IB_SA_PATH_REC_NUMB_PATH | IB_SA_PATH_REC_TRAFFIC_CLASS | IB_SA_PATH_REC_PKEY, 1000, GFP_ATOMIC, path_rec_completion, path, &path->query); if (path->query_id < 0) { ipoib_warn(priv, "ib_sa_path_rec_get failed: %d\n", path->query_id); path->query = NULL; complete(&path->done); return path->query_id; } return 0; } static void ipoib_unicast_send(struct mbuf *mb, struct ipoib_dev_priv *priv, struct ipoib_header *eh) { struct ipoib_path *path; path = __path_find(priv, eh->hwaddr + 4); if (!path || !path->valid) { int new_path = 0; if (!path) { path = path_rec_create(priv, eh->hwaddr); new_path = 1; } if (path) { if (_IF_QLEN(&path->queue) < IPOIB_MAX_PATH_REC_QUEUE) _IF_ENQUEUE(&path->queue, mb); else { if_inc_counter(priv->dev, IFCOUNTER_OERRORS, 1); m_freem(mb); } if (!path->query && path_rec_start(priv, path)) { spin_unlock_irqrestore(&priv->lock, flags); if (new_path) ipoib_path_free(priv, path); return; } else __path_add(priv, path); } else { if_inc_counter(priv->dev, IFCOUNTER_OERRORS, 1); m_freem(mb); } return; } if (ipoib_cm_get(path) && ipoib_cm_up(path)) { ipoib_cm_send(priv, mb, ipoib_cm_get(path)); } else if (path->ah) { ipoib_send(priv, mb, path->ah, IPOIB_QPN(eh->hwaddr)); } else if ((path->query || !path_rec_start(priv, path)) && path->queue.ifq_len < IPOIB_MAX_PATH_REC_QUEUE) { _IF_ENQUEUE(&path->queue, mb); } else { if_inc_counter(priv->dev, IFCOUNTER_OERRORS, 1); m_freem(mb); } } static int ipoib_send_one(struct ipoib_dev_priv *priv, struct mbuf *mb) { struct ipoib_header *eh; eh = mtod(mb, struct ipoib_header *); if (IPOIB_IS_MULTICAST(eh->hwaddr)) { /* Add in the P_Key for multicast*/ eh->hwaddr[8] = (priv->pkey >> 8) & 0xff; eh->hwaddr[9] = priv->pkey & 0xff; ipoib_mcast_send(priv, eh->hwaddr + 4, mb); } else ipoib_unicast_send(mb, priv, eh); return 0; } static void _ipoib_start(struct ifnet *dev, struct ipoib_dev_priv *priv) { struct mbuf *mb; if ((dev->if_drv_flags & (IFF_DRV_RUNNING|IFF_DRV_OACTIVE)) != IFF_DRV_RUNNING) return; spin_lock(&priv->lock); while (!IFQ_DRV_IS_EMPTY(&dev->if_snd) && (dev->if_drv_flags & IFF_DRV_OACTIVE) == 0) { IFQ_DRV_DEQUEUE(&dev->if_snd, mb); if (mb == NULL) break; IPOIB_MTAP(dev, mb); ipoib_send_one(priv, mb); } spin_unlock(&priv->lock); } static void ipoib_start(struct ifnet *dev) { _ipoib_start(dev, dev->if_softc); } static void ipoib_vlan_start(struct ifnet *dev) { struct ipoib_dev_priv *priv; struct mbuf *mb; priv = VLAN_COOKIE(dev); if (priv != NULL) return _ipoib_start(dev, priv); while (!IFQ_DRV_IS_EMPTY(&dev->if_snd)) { IFQ_DRV_DEQUEUE(&dev->if_snd, mb); if (mb == NULL) break; m_freem(mb); if_inc_counter(dev, IFCOUNTER_OERRORS, 1); } } int ipoib_dev_init(struct ipoib_dev_priv *priv, struct ib_device *ca, int port) { /* Allocate RX/TX "rings" to hold queued mbs */ priv->rx_ring = kzalloc(ipoib_recvq_size * sizeof *priv->rx_ring, GFP_KERNEL); if (!priv->rx_ring) { printk(KERN_WARNING "%s: failed to allocate RX ring (%d entries)\n", ca->name, ipoib_recvq_size); goto out; } priv->tx_ring = kzalloc(ipoib_sendq_size * sizeof *priv->tx_ring, GFP_KERNEL); if (!priv->tx_ring) { printk(KERN_WARNING "%s: failed to allocate TX ring (%d entries)\n", ca->name, ipoib_sendq_size); goto out_rx_ring_cleanup; } memset(priv->tx_ring, 0, ipoib_sendq_size * sizeof *priv->tx_ring); /* priv->tx_head, tx_tail & tx_outstanding are already 0 */ if (ipoib_ib_dev_init(priv, ca, port)) goto out_tx_ring_cleanup; return 0; out_tx_ring_cleanup: kfree(priv->tx_ring); out_rx_ring_cleanup: kfree(priv->rx_ring); out: return -ENOMEM; } static void ipoib_detach(struct ipoib_dev_priv *priv) { struct ifnet *dev; dev = priv->dev; if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { priv->gone = 1; bpfdetach(dev); if_detach(dev); if_free(dev); free_unr(ipoib_unrhdr, priv->unit); } else VLAN_SETCOOKIE(priv->dev, NULL); free(priv, M_TEMP); } void ipoib_dev_cleanup(struct ipoib_dev_priv *priv) { struct ipoib_dev_priv *cpriv, *tcpriv; /* Delete any child interfaces first */ list_for_each_entry_safe(cpriv, tcpriv, &priv->child_intfs, list) { ipoib_dev_cleanup(cpriv); ipoib_detach(cpriv); } ipoib_ib_dev_cleanup(priv); kfree(priv->rx_ring); kfree(priv->tx_ring); priv->rx_ring = NULL; priv->tx_ring = NULL; } static struct ipoib_dev_priv * ipoib_priv_alloc(void) { struct ipoib_dev_priv *priv; priv = malloc(sizeof(struct ipoib_dev_priv), M_TEMP, M_ZERO|M_WAITOK); spin_lock_init(&priv->lock); spin_lock_init(&priv->drain_lock); mutex_init(&priv->vlan_mutex); INIT_LIST_HEAD(&priv->path_list); INIT_LIST_HEAD(&priv->child_intfs); INIT_LIST_HEAD(&priv->dead_ahs); INIT_LIST_HEAD(&priv->multicast_list); INIT_DELAYED_WORK(&priv->pkey_poll_task, ipoib_pkey_poll); INIT_DELAYED_WORK(&priv->mcast_task, ipoib_mcast_join_task); INIT_WORK(&priv->carrier_on_task, ipoib_mcast_carrier_on_task); INIT_WORK(&priv->flush_light, ipoib_ib_dev_flush_light); INIT_WORK(&priv->flush_normal, ipoib_ib_dev_flush_normal); INIT_WORK(&priv->flush_heavy, ipoib_ib_dev_flush_heavy); INIT_WORK(&priv->restart_task, ipoib_mcast_restart_task); INIT_DELAYED_WORK(&priv->ah_reap_task, ipoib_reap_ah); memcpy(priv->broadcastaddr, ipv4_bcast_addr, INFINIBAND_ALEN); return (priv); } struct ipoib_dev_priv * ipoib_intf_alloc(const char *name) { struct ipoib_dev_priv *priv; struct sockaddr_dl *sdl; struct ifnet *dev; priv = ipoib_priv_alloc(); dev = priv->dev = if_alloc(IFT_INFINIBAND); if (!dev) { free(priv, M_TEMP); return NULL; } dev->if_softc = priv; priv->unit = alloc_unr(ipoib_unrhdr); if (priv->unit == -1) { if_free(dev); free(priv, M_TEMP); return NULL; } if_initname(dev, name, priv->unit); dev->if_flags = IFF_BROADCAST | IFF_MULTICAST; dev->if_addrlen = INFINIBAND_ALEN; dev->if_hdrlen = IPOIB_HEADER_LEN; if_attach(dev); dev->if_init = ipoib_init; dev->if_ioctl = ipoib_ioctl; dev->if_start = ipoib_start; dev->if_output = ipoib_output; dev->if_input = ipoib_input; dev->if_resolvemulti = ipoib_resolvemulti; dev->if_baudrate = IF_Gbps(10); dev->if_broadcastaddr = priv->broadcastaddr; dev->if_snd.ifq_maxlen = ipoib_sendq_size * 2; sdl = (struct sockaddr_dl *)dev->if_addr->ifa_addr; sdl->sdl_type = IFT_INFINIBAND; sdl->sdl_alen = dev->if_addrlen; priv->dev = dev; if_link_state_change(dev, LINK_STATE_DOWN); bpfattach(dev, DLT_EN10MB, ETHER_HDR_LEN); return dev->if_softc; } int ipoib_set_dev_features(struct ipoib_dev_priv *priv, struct ib_device *hca) { struct ib_device_attr *device_attr = &hca->attrs; priv->hca_caps = device_attr->device_cap_flags; priv->dev->if_hwassist = 0; priv->dev->if_capabilities = 0; #ifndef CONFIG_INFINIBAND_IPOIB_CM if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) { set_bit(IPOIB_FLAG_CSUM, &priv->flags); priv->dev->if_hwassist = CSUM_IP | CSUM_TCP | CSUM_UDP; priv->dev->if_capabilities = IFCAP_HWCSUM | IFCAP_VLAN_HWCSUM; } #if 0 if (priv->dev->features & NETIF_F_SG && priv->hca_caps & IB_DEVICE_UD_TSO) { priv->dev->if_capabilities |= IFCAP_TSO4; priv->dev->if_hwassist |= CSUM_TSO; } #endif #endif priv->dev->if_capabilities |= IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_MTU | IFCAP_LINKSTATE; priv->dev->if_capenable = priv->dev->if_capabilities; return 0; } static struct ifnet * ipoib_add_port(const char *format, struct ib_device *hca, u8 port) { struct ipoib_dev_priv *priv; struct ib_port_attr attr; int result = -ENOMEM; priv = ipoib_intf_alloc(format); if (!priv) goto alloc_mem_failed; if (!ib_query_port(hca, port, &attr)) priv->max_ib_mtu = ib_mtu_enum_to_int(attr.max_mtu); else { printk(KERN_WARNING "%s: ib_query_port %d failed\n", hca->name, port); goto device_init_failed; } /* MTU will be reset when mcast join happens */ priv->dev->if_mtu = IPOIB_UD_MTU(priv->max_ib_mtu); priv->mcast_mtu = priv->admin_mtu = priv->dev->if_mtu; result = ib_query_pkey(hca, port, 0, &priv->pkey); if (result) { printk(KERN_WARNING "%s: ib_query_pkey port %d failed (ret = %d)\n", hca->name, port, result); goto device_init_failed; } if (ipoib_set_dev_features(priv, hca)) goto device_init_failed; /* * Set the full membership bit, so that we join the right * broadcast group, etc. */ priv->pkey |= 0x8000; priv->broadcastaddr[8] = priv->pkey >> 8; priv->broadcastaddr[9] = priv->pkey & 0xff; result = ib_query_gid(hca, port, 0, &priv->local_gid, NULL); if (result) { printk(KERN_WARNING "%s: ib_query_gid port %d failed (ret = %d)\n", hca->name, port, result); goto device_init_failed; } memcpy(IF_LLADDR(priv->dev) + 4, priv->local_gid.raw, sizeof (union ib_gid)); result = ipoib_dev_init(priv, hca, port); if (result < 0) { printk(KERN_WARNING "%s: failed to initialize port %d (ret = %d)\n", hca->name, port, result); goto device_init_failed; } if (ipoib_cm_admin_enabled(priv)) priv->dev->if_mtu = IPOIB_CM_MTU(ipoib_cm_max_mtu(priv)); INIT_IB_EVENT_HANDLER(&priv->event_handler, priv->ca, ipoib_event); result = ib_register_event_handler(&priv->event_handler); if (result < 0) { printk(KERN_WARNING "%s: ib_register_event_handler failed for " "port %d (ret = %d)\n", hca->name, port, result); goto event_failed; } if_printf(priv->dev, "Attached to %s port %d\n", hca->name, port); return priv->dev; event_failed: ipoib_dev_cleanup(priv); device_init_failed: ipoib_detach(priv); alloc_mem_failed: return ERR_PTR(result); } static void ipoib_add_one(struct ib_device *device) { struct list_head *dev_list; struct ifnet *dev; struct ipoib_dev_priv *priv; int s, e, p; if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) return; dev_list = kmalloc(sizeof *dev_list, GFP_KERNEL); if (!dev_list) return; INIT_LIST_HEAD(dev_list); if (device->node_type == RDMA_NODE_IB_SWITCH) { s = 0; e = 0; } else { s = 1; e = device->phys_port_cnt; } for (p = s; p <= e; ++p) { if (rdma_port_get_link_layer(device, p) != IB_LINK_LAYER_INFINIBAND) continue; dev = ipoib_add_port("ib", device, p); if (!IS_ERR(dev)) { priv = dev->if_softc; list_add_tail(&priv->list, dev_list); } } ib_set_client_data(device, &ipoib_client, dev_list); } static void ipoib_remove_one(struct ib_device *device, void *client_data) { struct ipoib_dev_priv *priv, *tmp; struct list_head *dev_list = client_data; if (!dev_list) return; if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) return; list_for_each_entry_safe(priv, tmp, dev_list, list) { if (rdma_port_get_link_layer(device, priv->port) != IB_LINK_LAYER_INFINIBAND) continue; ipoib_stop(priv); ib_unregister_event_handler(&priv->event_handler); /* dev_change_flags(priv->dev, priv->dev->flags & ~IFF_UP); */ flush_workqueue(ipoib_workqueue); ipoib_dev_cleanup(priv); ipoib_detach(priv); } kfree(dev_list); +} + +static int +ipoib_match_dev_addr(const struct sockaddr *addr, struct net_device *dev) +{ + struct ifaddr *ifa; + int retval = 0; + + CURVNET_SET(dev->if_vnet); + IF_ADDR_RLOCK(dev); + CK_STAILQ_FOREACH(ifa, &dev->if_addrhead, ifa_link) { + if (ifa->ifa_addr == NULL || + ifa->ifa_addr->sa_family != addr->sa_family || + ifa->ifa_addr->sa_len != addr->sa_len) { + continue; + } + if (memcmp(ifa->ifa_addr, addr, addr->sa_len) == 0) { + retval = 1; + break; + } + } + IF_ADDR_RUNLOCK(dev); + CURVNET_RESTORE(); + + return (retval); +} + +/* + * ipoib_match_gid_pkey_addr - returns the number of IPoIB netdevs on + * top a given ipoib device matching a pkey_index and address, if one + * exists. + * + * @found_net_dev: contains a matching net_device if the return value + * >= 1, with a reference held. + */ +static int +ipoib_match_gid_pkey_addr(struct ipoib_dev_priv *priv, + const union ib_gid *gid, u16 pkey_index, const struct sockaddr *addr, + struct net_device **found_net_dev) +{ + struct ipoib_dev_priv *child_priv; + int matches = 0; + + if (priv->pkey_index == pkey_index && + (!gid || !memcmp(gid, &priv->local_gid, sizeof(*gid)))) { + if (addr == NULL || ipoib_match_dev_addr(addr, priv->dev) != 0) { + if (*found_net_dev == NULL) { + struct net_device *net_dev; + + if (priv->parent != NULL) + net_dev = priv->parent; + else + net_dev = priv->dev; + *found_net_dev = net_dev; + dev_hold(net_dev); + } + matches++; + } + } + + /* Check child interfaces */ + mutex_lock(&priv->vlan_mutex); + list_for_each_entry(child_priv, &priv->child_intfs, list) { + matches += ipoib_match_gid_pkey_addr(child_priv, gid, + pkey_index, addr, found_net_dev); + if (matches > 1) + break; + } + mutex_unlock(&priv->vlan_mutex); + + return matches; +} + +/* + * __ipoib_get_net_dev_by_params - returns the number of matching + * net_devs found (between 0 and 2). Also return the matching + * net_device in the @net_dev parameter, holding a reference to the + * net_device, if the number of matches >= 1 + */ +static int +__ipoib_get_net_dev_by_params(struct list_head *dev_list, u8 port, + u16 pkey_index, const union ib_gid *gid, + const struct sockaddr *addr, struct net_device **net_dev) +{ + struct ipoib_dev_priv *priv; + int matches = 0; + + *net_dev = NULL; + + list_for_each_entry(priv, dev_list, list) { + if (priv->port != port) + continue; + + matches += ipoib_match_gid_pkey_addr(priv, gid, pkey_index, + addr, net_dev); + + if (matches > 1) + break; + } + + return matches; +} + +static struct net_device * +ipoib_get_net_dev_by_params(struct ib_device *dev, u8 port, u16 pkey, + const union ib_gid *gid, const struct sockaddr *addr, void *client_data) +{ + struct net_device *net_dev; + struct list_head *dev_list = client_data; + u16 pkey_index; + int matches; + int ret; + + if (!rdma_protocol_ib(dev, port)) + return NULL; + + ret = ib_find_cached_pkey(dev, port, pkey, &pkey_index); + if (ret) + return NULL; + + if (!dev_list) + return NULL; + + /* See if we can find a unique device matching the L2 parameters */ + matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index, + gid, NULL, &net_dev); + + switch (matches) { + case 0: + return NULL; + case 1: + return net_dev; + } + + dev_put(net_dev); + + /* Couldn't find a unique device with L2 parameters only. Use L3 + * address to uniquely match the net device */ + matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index, + gid, addr, &net_dev); + switch (matches) { + case 0: + return NULL; + default: + dev_warn_ratelimited(&dev->dev, + "duplicate IP address detected\n"); + /* Fall through */ + case 1: + return net_dev; + } } static void ipoib_config_vlan(void *arg, struct ifnet *ifp, u_int16_t vtag) { struct ipoib_dev_priv *parent; struct ipoib_dev_priv *priv; struct ifnet *dev; uint16_t pkey; int error; if (ifp->if_type != IFT_INFINIBAND) return; dev = VLAN_DEVAT(ifp, vtag); if (dev == NULL) return; priv = NULL; error = 0; parent = ifp->if_softc; /* We only support 15 bits of pkey. */ if (vtag & 0x8000) return; pkey = vtag | 0x8000; /* Set full membership bit. */ if (pkey == parent->pkey) return; /* Check for dups */ mutex_lock(&parent->vlan_mutex); list_for_each_entry(priv, &parent->child_intfs, list) { if (priv->pkey == pkey) { priv = NULL; error = EBUSY; goto out; } } priv = ipoib_priv_alloc(); priv->dev = dev; priv->max_ib_mtu = parent->max_ib_mtu; priv->mcast_mtu = priv->admin_mtu = parent->dev->if_mtu; set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags); error = ipoib_set_dev_features(priv, parent->ca); if (error) goto out; priv->pkey = pkey; priv->broadcastaddr[8] = pkey >> 8; priv->broadcastaddr[9] = pkey & 0xff; dev->if_broadcastaddr = priv->broadcastaddr; error = ipoib_dev_init(priv, parent->ca, parent->port); if (error) goto out; priv->parent = parent->dev; list_add_tail(&priv->list, &parent->child_intfs); VLAN_SETCOOKIE(dev, priv); dev->if_start = ipoib_vlan_start; dev->if_drv_flags &= ~IFF_DRV_RUNNING; dev->if_hdrlen = IPOIB_HEADER_LEN; if (ifp->if_drv_flags & IFF_DRV_RUNNING) ipoib_open(priv); mutex_unlock(&parent->vlan_mutex); return; out: mutex_unlock(&parent->vlan_mutex); if (priv) free(priv, M_TEMP); if (error) ipoib_warn(parent, "failed to initialize subinterface: device %s, port %d vtag 0x%X", parent->ca->name, parent->port, vtag); return; } static void ipoib_unconfig_vlan(void *arg, struct ifnet *ifp, u_int16_t vtag) { struct ipoib_dev_priv *parent; struct ipoib_dev_priv *priv; struct ifnet *dev; uint16_t pkey; if (ifp->if_type != IFT_INFINIBAND) return; dev = VLAN_DEVAT(ifp, vtag); if (dev) VLAN_SETCOOKIE(dev, NULL); pkey = vtag | 0x8000; parent = ifp->if_softc; mutex_lock(&parent->vlan_mutex); list_for_each_entry(priv, &parent->child_intfs, list) { if (priv->pkey == pkey) { ipoib_dev_cleanup(priv); list_del(&priv->list); break; } } mutex_unlock(&parent->vlan_mutex); } eventhandler_tag ipoib_vlan_attach; eventhandler_tag ipoib_vlan_detach; static int __init ipoib_init_module(void) { int ret; ipoib_recvq_size = roundup_pow_of_two(ipoib_recvq_size); ipoib_recvq_size = min(ipoib_recvq_size, IPOIB_MAX_QUEUE_SIZE); ipoib_recvq_size = max(ipoib_recvq_size, IPOIB_MIN_QUEUE_SIZE); ipoib_sendq_size = roundup_pow_of_two(ipoib_sendq_size); ipoib_sendq_size = min(ipoib_sendq_size, IPOIB_MAX_QUEUE_SIZE); ipoib_sendq_size = max(ipoib_sendq_size, max(2 * MAX_SEND_CQE, IPOIB_MIN_QUEUE_SIZE)); #ifdef CONFIG_INFINIBAND_IPOIB_CM ipoib_max_conn_qp = min(ipoib_max_conn_qp, IPOIB_CM_MAX_CONN_QP); #endif ipoib_vlan_attach = EVENTHANDLER_REGISTER(vlan_config, ipoib_config_vlan, NULL, EVENTHANDLER_PRI_FIRST); ipoib_vlan_detach = EVENTHANDLER_REGISTER(vlan_unconfig, ipoib_unconfig_vlan, NULL, EVENTHANDLER_PRI_FIRST); /* * We create our own workqueue mainly because we want to be * able to flush it when devices are being removed. We can't * use schedule_work()/flush_scheduled_work() because both * unregister_netdev() and linkwatch_event take the rtnl lock, * so flush_scheduled_work() can deadlock during device * removal. */ ipoib_workqueue = create_singlethread_workqueue("ipoib"); if (!ipoib_workqueue) { ret = -ENOMEM; goto err_fs; } ib_sa_register_client(&ipoib_sa_client); ret = ib_register_client(&ipoib_client); if (ret) goto err_sa; return 0; err_sa: ib_sa_unregister_client(&ipoib_sa_client); destroy_workqueue(ipoib_workqueue); err_fs: return ret; } static void __exit ipoib_cleanup_module(void) { EVENTHANDLER_DEREGISTER(vlan_config, ipoib_vlan_attach); EVENTHANDLER_DEREGISTER(vlan_unconfig, ipoib_vlan_detach); ib_unregister_client(&ipoib_client); ib_sa_unregister_client(&ipoib_sa_client); destroy_workqueue(ipoib_workqueue); } /* * Infiniband output routine. */ static int ipoib_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, struct route *ro) { u_char edst[INFINIBAND_ALEN]; #if defined(INET) || defined(INET6) struct llentry *lle = NULL; #endif struct ipoib_header *eh; int error = 0, is_gw = 0; short type; if (ro != NULL) is_gw = (ro->ro_flags & RT_HAS_GW) != 0; #ifdef MAC error = mac_ifnet_check_transmit(ifp, m); if (error) goto bad; #endif M_PROFILE(m); if (ifp->if_flags & IFF_MONITOR) { error = ENETDOWN; goto bad; } if (!((ifp->if_flags & IFF_UP) && (ifp->if_drv_flags & IFF_DRV_RUNNING))) { error = ENETDOWN; goto bad; } switch (dst->sa_family) { #ifdef INET case AF_INET: if (lle != NULL && (lle->la_flags & LLE_VALID)) memcpy(edst, lle->ll_addr, sizeof(edst)); else if (m->m_flags & M_MCAST) ip_ib_mc_map(((struct sockaddr_in *)dst)->sin_addr.s_addr, ifp->if_broadcastaddr, edst); else error = arpresolve(ifp, is_gw, m, dst, edst, NULL, NULL); if (error) return (error == EWOULDBLOCK ? 0 : error); type = htons(ETHERTYPE_IP); break; case AF_ARP: { struct arphdr *ah; ah = mtod(m, struct arphdr *); ah->ar_hrd = htons(ARPHRD_INFINIBAND); switch(ntohs(ah->ar_op)) { case ARPOP_REVREQUEST: case ARPOP_REVREPLY: type = htons(ETHERTYPE_REVARP); break; case ARPOP_REQUEST: case ARPOP_REPLY: default: type = htons(ETHERTYPE_ARP); break; } if (m->m_flags & M_BCAST) bcopy(ifp->if_broadcastaddr, edst, INFINIBAND_ALEN); else bcopy(ar_tha(ah), edst, INFINIBAND_ALEN); } break; #endif #ifdef INET6 case AF_INET6: if (lle != NULL && (lle->la_flags & LLE_VALID)) memcpy(edst, lle->ll_addr, sizeof(edst)); else if (m->m_flags & M_MCAST) ipv6_ib_mc_map(&((struct sockaddr_in6 *)dst)->sin6_addr, ifp->if_broadcastaddr, edst); else error = nd6_resolve(ifp, is_gw, m, dst, edst, NULL, NULL); if (error) return error; type = htons(ETHERTYPE_IPV6); break; #endif default: if_printf(ifp, "can't handle af%d\n", dst->sa_family); error = EAFNOSUPPORT; goto bad; } /* * Add local net header. If no space in first mbuf, * allocate another. */ M_PREPEND(m, IPOIB_HEADER_LEN, M_NOWAIT); if (m == NULL) { error = ENOBUFS; goto bad; } eh = mtod(m, struct ipoib_header *); (void)memcpy(&eh->proto, &type, sizeof(eh->proto)); (void)memcpy(&eh->hwaddr, edst, sizeof (edst)); /* * Queue message on interface, update output statistics if * successful, and start output if interface not yet active. */ return ((ifp->if_transmit)(ifp, m)); bad: if (m != NULL) m_freem(m); return (error); } /* * Upper layer processing for a received Infiniband packet. */ void ipoib_demux(struct ifnet *ifp, struct mbuf *m, u_short proto) { int isr; #ifdef MAC /* * Tag the mbuf with an appropriate MAC label before any other * consumers can get to it. */ mac_ifnet_create_mbuf(ifp, m); #endif /* Allow monitor mode to claim this frame, after stats are updated. */ if (ifp->if_flags & IFF_MONITOR) { if_printf(ifp, "discard frame at IFF_MONITOR\n"); m_freem(m); return; } /* * Dispatch frame to upper layer. */ switch (proto) { #ifdef INET case ETHERTYPE_IP: isr = NETISR_IP; break; case ETHERTYPE_ARP: if (ifp->if_flags & IFF_NOARP) { /* Discard packet if ARP is disabled on interface */ m_freem(m); return; } isr = NETISR_ARP; break; #endif #ifdef INET6 case ETHERTYPE_IPV6: isr = NETISR_IPV6; break; #endif default: goto discard; } netisr_dispatch(isr, m); return; discard: m_freem(m); } /* * Process a received Infiniband packet. */ static void ipoib_input(struct ifnet *ifp, struct mbuf *m) { struct ipoib_header *eh; if ((ifp->if_flags & IFF_UP) == 0) { m_freem(m); return; } CURVNET_SET_QUIET(ifp->if_vnet); /* Let BPF have it before we strip the header. */ IPOIB_MTAP(ifp, m); eh = mtod(m, struct ipoib_header *); /* * Reset layer specific mbuf flags to avoid confusing upper layers. * Strip off Infiniband header. */ m->m_flags &= ~M_VLANTAG; m_clrprotoflags(m); m_adj(m, IPOIB_HEADER_LEN); if (IPOIB_IS_MULTICAST(eh->hwaddr)) { if (memcmp(eh->hwaddr, ifp->if_broadcastaddr, ifp->if_addrlen) == 0) m->m_flags |= M_BCAST; else m->m_flags |= M_MCAST; if_inc_counter(ifp, IFCOUNTER_IMCASTS, 1); } ipoib_demux(ifp, m, ntohs(eh->proto)); CURVNET_RESTORE(); } static int ipoib_resolvemulti(struct ifnet *ifp, struct sockaddr **llsa, struct sockaddr *sa) { struct sockaddr_dl *sdl; #ifdef INET struct sockaddr_in *sin; #endif #ifdef INET6 struct sockaddr_in6 *sin6; #endif u_char *e_addr; switch(sa->sa_family) { case AF_LINK: /* * No mapping needed. Just check that it's a valid MC address. */ sdl = (struct sockaddr_dl *)sa; e_addr = LLADDR(sdl); if (!IPOIB_IS_MULTICAST(e_addr)) return EADDRNOTAVAIL; *llsa = NULL; return 0; #ifdef INET case AF_INET: sin = (struct sockaddr_in *)sa; if (!IN_MULTICAST(ntohl(sin->sin_addr.s_addr))) return EADDRNOTAVAIL; sdl = link_init_sdl(ifp, *llsa, IFT_INFINIBAND); sdl->sdl_alen = INFINIBAND_ALEN; e_addr = LLADDR(sdl); ip_ib_mc_map(sin->sin_addr.s_addr, ifp->if_broadcastaddr, e_addr); *llsa = (struct sockaddr *)sdl; return 0; #endif #ifdef INET6 case AF_INET6: sin6 = (struct sockaddr_in6 *)sa; /* * An IP6 address of 0 means listen to all * of the multicast address used for IP6. * This has no meaning in ipoib. */ if (IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) return EADDRNOTAVAIL; if (!IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr)) return EADDRNOTAVAIL; sdl = link_init_sdl(ifp, *llsa, IFT_INFINIBAND); sdl->sdl_alen = INFINIBAND_ALEN; e_addr = LLADDR(sdl); ipv6_ib_mc_map(&sin6->sin6_addr, ifp->if_broadcastaddr, e_addr); *llsa = (struct sockaddr *)sdl; return 0; #endif default: return EAFNOSUPPORT; } } module_init(ipoib_init_module); module_exit(ipoib_cleanup_module); static int ipoib_evhand(module_t mod, int event, void *arg) { return (0); } static moduledata_t ipoib_mod = { .name = "ipoib", .evhand = ipoib_evhand, }; DECLARE_MODULE(ipoib, ipoib_mod, SI_SUB_LAST, SI_ORDER_ANY); MODULE_DEPEND(ipoib, ibcore, 1, 1, 1); MODULE_DEPEND(ipoib, linuxkpi, 1, 1, 1);