diff --git a/sys/dev/irdma/fbsd_kcompat.c b/sys/dev/irdma/fbsd_kcompat.c index 5c11f7a35544..19c3943b5151 100644 --- a/sys/dev/irdma/fbsd_kcompat.c +++ b/sys/dev/irdma/fbsd_kcompat.c @@ -1,1126 +1,1124 @@ /*- * SPDX-License-Identifier: GPL-2.0 or Linux-OpenIB * * Copyright (c) 2021 - 2022 Intel Corporation * * 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 * OpenFabrics.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 "osdep.h" #include "ice_rdma.h" #include "irdma_di_if.h" #include "irdma_main.h" #include #include #include #include #include /* additional QP debuging option. Keep false unless needed */ bool irdma_upload_context = false; inline u32 irdma_rd32(struct irdma_dev_ctx *dev_ctx, u32 reg){ KASSERT(reg < dev_ctx->mem_bus_space_size, ("irdma: register offset %#jx too large (max is %#jx)", (uintmax_t)reg, (uintmax_t)dev_ctx->mem_bus_space_size)); return (bus_space_read_4(dev_ctx->mem_bus_space_tag, dev_ctx->mem_bus_space_handle, reg)); } inline void irdma_wr32(struct irdma_dev_ctx *dev_ctx, u32 reg, u32 value) { KASSERT(reg < dev_ctx->mem_bus_space_size, ("irdma: register offset %#jx too large (max is %#jx)", (uintmax_t)reg, (uintmax_t)dev_ctx->mem_bus_space_size)); bus_space_write_4(dev_ctx->mem_bus_space_tag, dev_ctx->mem_bus_space_handle, reg, value); } inline u64 irdma_rd64(struct irdma_dev_ctx *dev_ctx, u32 reg){ KASSERT(reg < dev_ctx->mem_bus_space_size, ("irdma: register offset %#jx too large (max is %#jx)", (uintmax_t)reg, (uintmax_t)dev_ctx->mem_bus_space_size)); return (bus_space_read_8(dev_ctx->mem_bus_space_tag, dev_ctx->mem_bus_space_handle, reg)); } inline void irdma_wr64(struct irdma_dev_ctx *dev_ctx, u32 reg, u64 value) { KASSERT(reg < dev_ctx->mem_bus_space_size, ("irdma: register offset %#jx too large (max is %#jx)", (uintmax_t)reg, (uintmax_t)dev_ctx->mem_bus_space_size)); bus_space_write_8(dev_ctx->mem_bus_space_tag, dev_ctx->mem_bus_space_handle, reg, value); } void irdma_request_reset(struct irdma_pci_f *rf) { struct ice_rdma_peer *peer = rf->peer_info; struct ice_rdma_request req = {0}; req.type = ICE_RDMA_EVENT_RESET; printf("%s:%d requesting pf-reset\n", __func__, __LINE__); IRDMA_DI_REQ_HANDLER(peer, &req); } int irdma_register_qset(struct irdma_sc_vsi *vsi, struct irdma_ws_node *tc_node) { struct irdma_device *iwdev = vsi->back_vsi; struct ice_rdma_peer *peer = iwdev->rf->peer_info; struct ice_rdma_request req = {0}; struct ice_rdma_qset_update *res = &req.res; req.type = ICE_RDMA_EVENT_QSET_REGISTER; res->cnt_req = 1; res->res_type = ICE_RDMA_QSET_ALLOC; res->qsets.qs_handle = tc_node->qs_handle; res->qsets.tc = tc_node->traffic_class; res->qsets.vsi_id = vsi->vsi_idx; IRDMA_DI_REQ_HANDLER(peer, &req); tc_node->l2_sched_node_id = res->qsets.teid; vsi->qos[tc_node->user_pri].l2_sched_node_id = res->qsets.teid; return 0; } void irdma_unregister_qset(struct irdma_sc_vsi *vsi, struct irdma_ws_node *tc_node) { struct irdma_device *iwdev = vsi->back_vsi; struct ice_rdma_peer *peer = iwdev->rf->peer_info; struct ice_rdma_request req = {0}; struct ice_rdma_qset_update *res = &req.res; req.type = ICE_RDMA_EVENT_QSET_REGISTER; res->res_allocated = 1; res->res_type = ICE_RDMA_QSET_FREE; res->qsets.vsi_id = vsi->vsi_idx; res->qsets.teid = tc_node->l2_sched_node_id; res->qsets.qs_handle = tc_node->qs_handle; IRDMA_DI_REQ_HANDLER(peer, &req); } void * hw_to_dev(struct irdma_hw *hw) { struct irdma_pci_f *rf; rf = container_of(hw, struct irdma_pci_f, hw); return rf->pcidev; } void irdma_free_hash_desc(void *desc) { return; } int irdma_init_hash_desc(void **desc) { return 0; } int irdma_ieq_check_mpacrc(void *desc, void *addr, u32 len, u32 val) { u32 crc = calculate_crc32c(0xffffffff, addr, len) ^ 0xffffffff; int ret_code = 0; if (crc != val) { irdma_pr_err("mpa crc check fail %x %x\n", crc, val); ret_code = -EINVAL; } printf("%s: result crc=%x value=%x\n", __func__, crc, val); return ret_code; } static u_int irdma_add_ipv6_cb(void *arg, struct ifaddr *addr, u_int count __unused){ struct irdma_device *iwdev = arg; struct sockaddr_in6 *sin6; u32 local_ipaddr6[4] = {}; char ip6buf[INET6_ADDRSTRLEN]; u8 *mac_addr; sin6 = (struct sockaddr_in6 *)addr->ifa_addr; irdma_copy_ip_ntohl(local_ipaddr6, (u32 *)&sin6->sin6_addr); mac_addr = if_getlladdr(addr->ifa_ifp); printf("%s:%d IP=%s, MAC=%02x:%02x:%02x:%02x:%02x:%02x\n", __func__, __LINE__, ip6_sprintf(ip6buf, &sin6->sin6_addr), mac_addr[0], mac_addr[1], mac_addr[2], mac_addr[3], mac_addr[4], mac_addr[5]); irdma_manage_arp_cache(iwdev->rf, mac_addr, local_ipaddr6, IRDMA_ARP_ADD); return (0); } /** * irdma_add_ipv6_addr - add ipv6 address to the hw arp table * @iwdev: irdma device * @ifp: interface network device pointer */ static void irdma_add_ipv6_addr(struct irdma_device *iwdev, struct ifnet *ifp) { if_addr_rlock(ifp); if_foreach_addr_type(ifp, AF_INET6, irdma_add_ipv6_cb, iwdev); if_addr_runlock(ifp); } static u_int irdma_add_ipv4_cb(void *arg, struct ifaddr *addr, u_int count __unused){ struct irdma_device *iwdev = arg; struct sockaddr_in *sin; u32 ip_addr[4] = {}; uint8_t *mac_addr; sin = (struct sockaddr_in *)addr->ifa_addr; ip_addr[0] = ntohl(sin->sin_addr.s_addr); mac_addr = if_getlladdr(addr->ifa_ifp); printf("%s:%d IP=%d.%d.%d.%d, MAC=%02x:%02x:%02x:%02x:%02x:%02x\n", __func__, __LINE__, ip_addr[0] >> 24, (ip_addr[0] >> 16) & 0xFF, (ip_addr[0] >> 8) & 0xFF, ip_addr[0] & 0xFF, mac_addr[0], mac_addr[1], mac_addr[2], mac_addr[3], mac_addr[4], mac_addr[5]); irdma_manage_arp_cache(iwdev->rf, mac_addr, ip_addr, IRDMA_ARP_ADD); return (0); } /** * irdma_add_ipv4_addr - add ipv4 address to the hw arp table * @iwdev: irdma device * @ifp: interface network device pointer */ static void irdma_add_ipv4_addr(struct irdma_device *iwdev, struct ifnet *ifp) { if_addr_rlock(ifp); if_foreach_addr_type(ifp, AF_INET, irdma_add_ipv4_cb, iwdev); if_addr_runlock(ifp); } /** * irdma_add_ip - add ip addresses * @iwdev: irdma device * * Add ipv4/ipv6 addresses to the arp cache */ void irdma_add_ip(struct irdma_device *iwdev) { struct ifnet *ifp = iwdev->netdev; struct ifnet *ifv; struct epoch_tracker et; int i; irdma_add_ipv4_addr(iwdev, ifp); irdma_add_ipv6_addr(iwdev, ifp); for (i = 0; if_getvlantrunk(ifp) != NULL && i < VLAN_N_VID; ++i) { NET_EPOCH_ENTER(et); ifv = VLAN_DEVAT(ifp, i); NET_EPOCH_EXIT(et); if (!ifv) continue; irdma_add_ipv4_addr(iwdev, ifv); irdma_add_ipv6_addr(iwdev, ifv); } } static void irdma_ifaddrevent_handler(void *arg, struct ifnet *ifp, struct ifaddr *ifa, int event) { struct irdma_pci_f *rf = arg; struct ifnet *ifv = NULL; struct sockaddr_in *sin; struct epoch_tracker et; int arp_index = 0, i = 0; u32 ip[4] = {}; if (!ifa || !ifa->ifa_addr || !ifp) return; if (rf->iwdev->netdev != ifp) { for (i = 0; if_getvlantrunk(rf->iwdev->netdev) != NULL && i < VLAN_N_VID; ++i) { NET_EPOCH_ENTER(et); ifv = VLAN_DEVAT(rf->iwdev->netdev, i); NET_EPOCH_EXIT(et); if (ifv == ifp) break; } if (ifv != ifp) return; } sin = (struct sockaddr_in *)ifa->ifa_addr; switch (event) { case IFADDR_EVENT_ADD: if (sin->sin_family == AF_INET) irdma_add_ipv4_addr(rf->iwdev, ifp); else if (sin->sin_family == AF_INET6) irdma_add_ipv6_addr(rf->iwdev, ifp); break; case IFADDR_EVENT_DEL: if (sin->sin_family == AF_INET) { ip[0] = ntohl(sin->sin_addr.s_addr); } else if (sin->sin_family == AF_INET6) { irdma_copy_ip_ntohl(ip, (u32 *)&((struct sockaddr_in6 *)sin)->sin6_addr); } else { break; } for_each_set_bit(arp_index, rf->allocated_arps, rf->arp_table_size) { if (!memcmp(rf->arp_table[arp_index].ip_addr, ip, sizeof(ip))) { irdma_manage_arp_cache(rf, rf->arp_table[arp_index].mac_addr, rf->arp_table[arp_index].ip_addr, IRDMA_ARP_DELETE); } } break; default: break; } } void irdma_reg_ipaddr_event_cb(struct irdma_pci_f *rf) { rf->irdma_ifaddr_event = EVENTHANDLER_REGISTER(ifaddr_event_ext, irdma_ifaddrevent_handler, rf, EVENTHANDLER_PRI_ANY); } void irdma_dereg_ipaddr_event_cb(struct irdma_pci_f *rf) { EVENTHANDLER_DEREGISTER(ifaddr_event_ext, rf->irdma_ifaddr_event); } static int irdma_get_route_ifp(struct sockaddr *dst_sin, struct ifnet *netdev, struct ifnet **ifp, struct sockaddr **nexthop, bool *gateway) { struct nhop_object *nh; if (dst_sin->sa_family == AF_INET6) nh = fib6_lookup(RT_DEFAULT_FIB, &((struct sockaddr_in6 *)dst_sin)->sin6_addr, ((struct sockaddr_in6 *)dst_sin)->sin6_scope_id, NHR_NONE, 0); else nh = fib4_lookup(RT_DEFAULT_FIB, ((struct sockaddr_in *)dst_sin)->sin_addr, 0, NHR_NONE, 0); if (!nh || (nh->nh_ifp != netdev && rdma_vlan_dev_real_dev(nh->nh_ifp) != netdev)) goto rt_not_found; *gateway = (nh->nh_flags & NHF_GATEWAY) ? true : false; *nexthop = (*gateway) ? &nh->gw_sa : dst_sin; *ifp = nh->nh_ifp; return 0; rt_not_found: pr_err("irdma: route not found\n"); return -ENETUNREACH; } /** * irdma_get_dst_mac - get destination mac address * @cm_node: connection's node * @dst_sin: destination address information * @dst_mac: mac address array to return */ int irdma_get_dst_mac(struct irdma_cm_node *cm_node, struct sockaddr *dst_sin, u8 *dst_mac) { struct ifnet *netdev = cm_node->iwdev->netdev; #ifdef VIMAGE - struct rdma_cm_id *rdma_id = (struct rdma_cm_id *)cm_node->cm_id->context; - struct vnet *vnet = rdma_id->route.addr.dev_addr.net; + struct vnet *vnet = irdma_cmid_to_vnet(cm_node->cm_id); #endif struct ifnet *ifp; struct llentry *lle; struct sockaddr *nexthop; struct epoch_tracker et; int err; bool gateway; NET_EPOCH_ENTER(et); CURVNET_SET_QUIET(vnet); err = irdma_get_route_ifp(dst_sin, netdev, &ifp, &nexthop, &gateway); if (err) goto get_route_fail; if (dst_sin->sa_family == AF_INET) { err = arpresolve(ifp, gateway, NULL, nexthop, dst_mac, NULL, &lle); } else if (dst_sin->sa_family == AF_INET6) { err = nd6_resolve(ifp, LLE_SF(AF_INET6, gateway), NULL, nexthop, dst_mac, NULL, &lle); } else { err = -EPROTONOSUPPORT; } get_route_fail: CURVNET_RESTORE(); NET_EPOCH_EXIT(et); if (err) { pr_err("failed to resolve neighbor address (err=%d)\n", err); return -ENETUNREACH; } return 0; } /** * irdma_addr_resolve_neigh - resolve neighbor address * @cm_node: connection's node * @dst_ip: remote ip address * @arpindex: if there is an arp entry */ int irdma_addr_resolve_neigh(struct irdma_cm_node *cm_node, u32 dst_ip, int arpindex) { struct irdma_device *iwdev = cm_node->iwdev; struct sockaddr_in dst_sin = {}; int err; u32 ip[4] = {}; u8 dst_mac[MAX_ADDR_LEN]; dst_sin.sin_len = sizeof(dst_sin); dst_sin.sin_family = AF_INET; dst_sin.sin_port = 0; dst_sin.sin_addr.s_addr = htonl(dst_ip); err = irdma_get_dst_mac(cm_node, (struct sockaddr *)&dst_sin, dst_mac); if (err) return arpindex; ip[0] = dst_ip; return irdma_add_arp(iwdev->rf, ip, dst_mac); } /** * irdma_addr_resolve_neigh_ipv6 - resolve neighbor ipv6 address * @cm_node: connection's node * @dest: remote ip address * @arpindex: if there is an arp entry */ int irdma_addr_resolve_neigh_ipv6(struct irdma_cm_node *cm_node, u32 *dest, int arpindex) { struct irdma_device *iwdev = cm_node->iwdev; struct sockaddr_in6 dst_addr = {}; int err; u8 dst_mac[MAX_ADDR_LEN]; dst_addr.sin6_family = AF_INET6; dst_addr.sin6_len = sizeof(dst_addr); dst_addr.sin6_scope_id = if_getindex(iwdev->netdev); irdma_copy_ip_htonl(dst_addr.sin6_addr.__u6_addr.__u6_addr32, dest); err = irdma_get_dst_mac(cm_node, (struct sockaddr *)&dst_addr, dst_mac); if (err) return arpindex; return irdma_add_arp(iwdev->rf, dest, dst_mac); } int irdma_resolve_neigh_lpb_chk(struct irdma_device *iwdev, struct irdma_cm_node *cm_node, struct irdma_cm_info *cm_info) { #ifdef VIMAGE - struct rdma_cm_id *rdma_id = (struct rdma_cm_id *)cm_node->cm_id->context; - struct vnet *vnet = rdma_id->route.addr.dev_addr.net; + struct vnet *vnet = irdma_cmid_to_vnet(cm_node->cm_id); #endif int arpindex; int oldarpindex; bool is_lpb = false; CURVNET_SET_QUIET(vnet); is_lpb = cm_node->ipv4 ? irdma_ipv4_is_lpb(cm_node->loc_addr[0], cm_node->rem_addr[0]) : irdma_ipv6_is_lpb(cm_node->loc_addr, cm_node->rem_addr); CURVNET_RESTORE(); if (is_lpb) { cm_node->do_lpb = true; arpindex = irdma_arp_table(iwdev->rf, cm_node->rem_addr, NULL, IRDMA_ARP_RESOLVE); } else { oldarpindex = irdma_arp_table(iwdev->rf, cm_node->rem_addr, NULL, IRDMA_ARP_RESOLVE); if (cm_node->ipv4) arpindex = irdma_addr_resolve_neigh(cm_node, cm_info->rem_addr[0], oldarpindex); else arpindex = irdma_addr_resolve_neigh_ipv6(cm_node, cm_info->rem_addr, oldarpindex); } return arpindex; } /** * irdma_add_handler - add a handler to the list * @hdl: handler to be added to the handler list */ void irdma_add_handler(struct irdma_handler *hdl) { unsigned long flags; spin_lock_irqsave(&irdma_handler_lock, flags); list_add(&hdl->list, &irdma_handlers); spin_unlock_irqrestore(&irdma_handler_lock, flags); } /** * irdma_del_handler - delete a handler from the list * @hdl: handler to be deleted from the handler list */ void irdma_del_handler(struct irdma_handler *hdl) { unsigned long flags; spin_lock_irqsave(&irdma_handler_lock, flags); list_del(&hdl->list); spin_unlock_irqrestore(&irdma_handler_lock, flags); } /** * irdma_set_rf_user_cfg_params - apply user configurable settings * @rf: RDMA PCI function */ void irdma_set_rf_user_cfg_params(struct irdma_pci_f *rf) { int en_rem_endpoint_trk = 0; int limits_sel = 4; rf->en_rem_endpoint_trk = en_rem_endpoint_trk; rf->limits_sel = limits_sel; rf->rst_to = IRDMA_RST_TIMEOUT_HZ; /* Enable DCQCN algorithm by default */ rf->dcqcn_ena = true; } /** * irdma_sysctl_dcqcn_update - handle dcqcn_ena sysctl update * @arg1: pointer to rf * @arg2: unused * @oidp: sysctl oid structure * @req: sysctl request pointer */ static int irdma_sysctl_dcqcn_update(SYSCTL_HANDLER_ARGS) { struct irdma_pci_f *rf = (struct irdma_pci_f *)arg1; int ret; u8 dcqcn_ena = rf->dcqcn_ena; ret = sysctl_handle_8(oidp, &dcqcn_ena, 0, req); if ((ret) || (req->newptr == NULL)) return ret; if (dcqcn_ena == 0) rf->dcqcn_ena = false; else rf->dcqcn_ena = true; return 0; } enum irdma_cqp_stats_info { IRDMA_CQP_REQ_CMDS = 28, IRDMA_CQP_CMPL_CMDS = 29 }; static int irdma_sysctl_cqp_stats(SYSCTL_HANDLER_ARGS) { struct irdma_sc_cqp *cqp = (struct irdma_sc_cqp *)arg1; char rslt[192] = "no cqp available yet"; int rslt_size = sizeof(rslt) - 1; int option = (int)arg2; if (!cqp) { return sysctl_handle_string(oidp, rslt, sizeof(rslt), req); } snprintf(rslt, sizeof(rslt), ""); switch (option) { case IRDMA_CQP_REQ_CMDS: snprintf(rslt, rslt_size, "%lu", cqp->requested_ops); break; case IRDMA_CQP_CMPL_CMDS: snprintf(rslt, rslt_size, "%lu", atomic64_read(&cqp->completed_ops)); break; } return sysctl_handle_string(oidp, rslt, sizeof(rslt), req); } struct irdma_sw_stats_tunable_info { u8 op_type; const char name[32]; const char desc[32]; uintptr_t value; }; static const struct irdma_sw_stats_tunable_info irdma_sws_list[] = { {IRDMA_OP_CEQ_DESTROY, "ceq_destroy", "ceq_destroy", 0}, {IRDMA_OP_AEQ_DESTROY, "aeq_destroy", "aeq_destroy", 0}, {IRDMA_OP_DELETE_ARP_CACHE_ENTRY, "delete_arp_cache_entry", "delete_arp_cache_entry", 0}, {IRDMA_OP_MANAGE_APBVT_ENTRY, "manage_apbvt_entry", "manage_apbvt_entry", 0}, {IRDMA_OP_CEQ_CREATE, "ceq_create", "ceq_create", 0}, {IRDMA_OP_AEQ_CREATE, "aeq_create", "aeq_create", 0}, {IRDMA_OP_MANAGE_QHASH_TABLE_ENTRY, "manage_qhash_table_entry", "manage_qhash_table_entry", 0}, {IRDMA_OP_QP_MODIFY, "qp_modify", "qp_modify", 0}, {IRDMA_OP_QP_UPLOAD_CONTEXT, "qp_upload_context", "qp_upload_context", 0}, {IRDMA_OP_CQ_CREATE, "cq_create", "cq_create", 0}, {IRDMA_OP_CQ_DESTROY, "cq_destroy", "cq_destroy", 0}, {IRDMA_OP_QP_CREATE, "qp_create", "qp_create", 0}, {IRDMA_OP_QP_DESTROY, "qp_destroy", "qp_destroy", 0}, {IRDMA_OP_ALLOC_STAG, "alloc_stag", "alloc_stag", 0}, {IRDMA_OP_MR_REG_NON_SHARED, "mr_reg_non_shared", "mr_reg_non_shared", 0}, {IRDMA_OP_DEALLOC_STAG, "dealloc_stag", "dealloc_stag", 0}, {IRDMA_OP_MW_ALLOC, "mw_alloc", "mw_alloc", 0}, {IRDMA_OP_QP_FLUSH_WQES, "qp_flush_wqes", "qp_flush_wqes", 0}, {IRDMA_OP_ADD_ARP_CACHE_ENTRY, "add_arp_cache_entry", "add_arp_cache_entry", 0}, {IRDMA_OP_MANAGE_PUSH_PAGE, "manage_push_page", "manage_push_page", 0}, {IRDMA_OP_UPDATE_PE_SDS, "update_pe_sds", "update_pe_sds", 0}, {IRDMA_OP_MANAGE_HMC_PM_FUNC_TABLE, "manage_hmc_pm_func_table", "manage_hmc_pm_func_table", 0}, {IRDMA_OP_SUSPEND, "suspend", "suspend", 0}, {IRDMA_OP_RESUME, "resume", "resume", 0}, {IRDMA_OP_MANAGE_VCHNL_REQ_PBLE_BP, "manage_vchnl_req_pble_bp", "manage_vchnl_req_pble_bp", 0}, {IRDMA_OP_QUERY_FPM_VAL, "query_fpm_val", "query_fpm_val", 0}, {IRDMA_OP_COMMIT_FPM_VAL, "commit_fpm_val", "commit_fpm_val", 0}, {IRDMA_OP_AH_CREATE, "ah_create", "ah_create", 0}, {IRDMA_OP_AH_MODIFY, "ah_modify", "ah_modify", 0}, {IRDMA_OP_AH_DESTROY, "ah_destroy", "ah_destroy", 0}, {IRDMA_OP_MC_CREATE, "mc_create", "mc_create", 0}, {IRDMA_OP_MC_DESTROY, "mc_destroy", "mc_destroy", 0}, {IRDMA_OP_MC_MODIFY, "mc_modify", "mc_modify", 0}, {IRDMA_OP_STATS_ALLOCATE, "stats_allocate", "stats_allocate", 0}, {IRDMA_OP_STATS_FREE, "stats_free", "stats_free", 0}, {IRDMA_OP_STATS_GATHER, "stats_gather", "stats_gather", 0}, {IRDMA_OP_WS_ADD_NODE, "ws_add_node", "ws_add_node", 0}, {IRDMA_OP_WS_MODIFY_NODE, "ws_modify_node", "ws_modify_node", 0}, {IRDMA_OP_WS_DELETE_NODE, "ws_delete_node", "ws_delete_node", 0}, {IRDMA_OP_WS_FAILOVER_START, "ws_failover_start", "ws_failover_start", 0}, {IRDMA_OP_WS_FAILOVER_COMPLETE, "ws_failover_complete", "ws_failover_complete", 0}, {IRDMA_OP_SET_UP_MAP, "set_up_map", "set_up_map", 0}, {IRDMA_OP_GEN_AE, "gen_ae", "gen_ae", 0}, {IRDMA_OP_QUERY_RDMA_FEATURES, "query_rdma_features", "query_rdma_features", 0}, {IRDMA_OP_ALLOC_LOCAL_MAC_ENTRY, "alloc_local_mac_entry", "alloc_local_mac_entry", 0}, {IRDMA_OP_ADD_LOCAL_MAC_ENTRY, "add_local_mac_entry", "add_local_mac_entry", 0}, {IRDMA_OP_DELETE_LOCAL_MAC_ENTRY, "delete_local_mac_entry", "delete_local_mac_entry", 0}, {IRDMA_OP_CQ_MODIFY, "cq_modify", "cq_modify", 0} }; static const struct irdma_sw_stats_tunable_info irdma_cmcs_list[] = { {0, "cm_nodes_created", "cm_nodes_created", offsetof(struct irdma_cm_core, stats_nodes_created)}, {0, "cm_nodes_destroyed", "cm_nodes_destroyed", offsetof(struct irdma_cm_core, stats_nodes_destroyed)}, {0, "cm_listen_created", "cm_listen_created", offsetof(struct irdma_cm_core, stats_listen_created)}, {0, "cm_listen_destroyed", "cm_listen_destroyed", offsetof(struct irdma_cm_core, stats_listen_destroyed)}, {0, "cm_listen_nodes_created", "cm_listen_nodes_created", offsetof(struct irdma_cm_core, stats_listen_nodes_created)}, {0, "cm_listen_nodes_destroyed", "cm_listen_nodes_destroyed", offsetof(struct irdma_cm_core, stats_listen_nodes_destroyed)}, {0, "cm_lpbs", "cm_lpbs", offsetof(struct irdma_cm_core, stats_lpbs)}, {0, "cm_accepts", "cm_accepts", offsetof(struct irdma_cm_core, stats_accepts)}, {0, "cm_rejects", "cm_rejects", offsetof(struct irdma_cm_core, stats_rejects)}, {0, "cm_connect_errs", "cm_connect_errs", offsetof(struct irdma_cm_core, stats_connect_errs)}, {0, "cm_passive_errs", "cm_passive_errs", offsetof(struct irdma_cm_core, stats_passive_errs)}, {0, "cm_pkt_retrans", "cm_pkt_retrans", offsetof(struct irdma_cm_core, stats_pkt_retrans)}, {0, "cm_backlog_drops", "cm_backlog_drops", offsetof(struct irdma_cm_core, stats_backlog_drops)}, }; static const struct irdma_sw_stats_tunable_info irdma_ilqs32_list[] = { {0, "ilq_avail_buf_count", "ilq_avail_buf_count", offsetof(struct irdma_puda_rsrc, avail_buf_count)}, {0, "ilq_alloc_buf_count", "ilq_alloc_buf_count", offsetof(struct irdma_puda_rsrc, alloc_buf_count)} }; static const struct irdma_sw_stats_tunable_info irdma_ilqs_list[] = { {0, "ilq_stats_buf_alloc_fail", "ilq_stats_buf_alloc_fail", offsetof(struct irdma_puda_rsrc, stats_buf_alloc_fail)}, {0, "ilq_stats_pkt_rcvd", "ilq_stats_pkt_rcvd", offsetof(struct irdma_puda_rsrc, stats_pkt_rcvd)}, {0, "ilq_stats_pkt_sent", "ilq_stats_pkt_sent", offsetof(struct irdma_puda_rsrc, stats_pkt_sent)}, {0, "ilq_stats_rcvd_pkt_err", "ilq_stats_rcvd_pkt_err", offsetof(struct irdma_puda_rsrc, stats_rcvd_pkt_err)}, {0, "ilq_stats_sent_pkt_q", "ilq_stats_sent_pkt_q", offsetof(struct irdma_puda_rsrc, stats_sent_pkt_q)} }; static const struct irdma_sw_stats_tunable_info irdma_ieqs32_list[] = { {0, "ieq_avail_buf_count", "ieq_avail_buf_count", offsetof(struct irdma_puda_rsrc, avail_buf_count)}, {0, "ieq_alloc_buf_count", "ieq_alloc_buf_count", offsetof(struct irdma_puda_rsrc, alloc_buf_count)} }; static const struct irdma_sw_stats_tunable_info irdma_ieqs_list[] = { {0, "ieq_stats_buf_alloc_fail", "ieq_stats_buf_alloc_fail", offsetof(struct irdma_puda_rsrc, stats_buf_alloc_fail)}, {0, "ieq_stats_pkt_rcvd", "ieq_stats_pkt_rcvd", offsetof(struct irdma_puda_rsrc, stats_pkt_rcvd)}, {0, "ieq_stats_pkt_sent", "ieq_stats_pkt_sent", offsetof(struct irdma_puda_rsrc, stats_pkt_sent)}, {0, "ieq_stats_rcvd_pkt_err", "ieq_stats_rcvd_pkt_err", offsetof(struct irdma_puda_rsrc, stats_rcvd_pkt_err)}, {0, "ieq_stats_sent_pkt_q", "ieq_stats_sent_pkt_q", offsetof(struct irdma_puda_rsrc, stats_sent_pkt_q)}, {0, "ieq_stats_bad_qp_id", "ieq_stats_bad_qp_id", offsetof(struct irdma_puda_rsrc, stats_bad_qp_id)}, {0, "ieq_fpdu_processed", "ieq_fpdu_processed", offsetof(struct irdma_puda_rsrc, fpdu_processed)}, {0, "ieq_bad_seq_num", "ieq_bad_seq_num", offsetof(struct irdma_puda_rsrc, bad_seq_num)}, {0, "ieq_crc_err", "ieq_crc_err", offsetof(struct irdma_puda_rsrc, crc_err)}, {0, "ieq_pmode_count", "ieq_pmode_count", offsetof(struct irdma_puda_rsrc, pmode_count)}, {0, "ieq_partials_handled", "ieq_partials_handled", offsetof(struct irdma_puda_rsrc, partials_handled)}, }; /** * irdma_dcqcn_tunables_init - create tunables for dcqcn settings * @rf: RDMA PCI function * * Create DCQCN related sysctls for the driver. * dcqcn_ena is writeable settings and applicable to next QP creation or * context setting. * all other settings are of RDTUN type (read on driver load) and are * applicable only to CQP creation. */ void irdma_dcqcn_tunables_init(struct irdma_pci_f *rf) { struct sysctl_oid_list *irdma_sysctl_oid_list; irdma_sysctl_oid_list = SYSCTL_CHILDREN(rf->tun_info.irdma_sysctl_tree); SYSCTL_ADD_PROC(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_enable", CTLFLAG_RW | CTLTYPE_U8, rf, 0, irdma_sysctl_dcqcn_update, "A", "enables DCQCN algorithm for RoCEv2 on all ports, default=true"); SYSCTL_ADD_U8(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_cc_cfg_valid", CTLFLAG_RDTUN, &rf->dcqcn_params.cc_cfg_valid, 0, "set DCQCN parameters to be valid, default=false"); rf->dcqcn_params.min_dec_factor = 1; SYSCTL_ADD_U8(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_min_dec_factor", CTLFLAG_RDTUN, &rf->dcqcn_params.min_dec_factor, 0, "set minimum percentage factor by which tx rate can be changed for CNP, Range: 1-100, default=1"); SYSCTL_ADD_U8(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_min_rate_MBps", CTLFLAG_RDTUN, &rf->dcqcn_params.min_rate, 0, "set minimum rate limit value, in MBits per second, default=0"); rf->dcqcn_params.dcqcn_f = 5; SYSCTL_ADD_U8(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_F", CTLFLAG_RDTUN, &rf->dcqcn_params.dcqcn_f, 0, "set number of times to stay in each stage of bandwidth recovery, default=5"); rf->dcqcn_params.dcqcn_t = 0x37; SYSCTL_ADD_U16(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_T", CTLFLAG_RDTUN, &rf->dcqcn_params.dcqcn_t, 0, "number of us to elapse before increasing the CWND in DCQCN mode, default=0x37"); rf->dcqcn_params.dcqcn_b = 0x249f0; SYSCTL_ADD_U32(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_B", CTLFLAG_RDTUN, &rf->dcqcn_params.dcqcn_b, 0, "set number of MSS to add to the congestion window in additive increase mode, default=0x249f0"); rf->dcqcn_params.rai_factor = 1; SYSCTL_ADD_U16(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_rai_factor", CTLFLAG_RDTUN, &rf->dcqcn_params.rai_factor, 0, "set number of MSS to add to the congestion window in additive increase mode, default=1"); rf->dcqcn_params.hai_factor = 5; SYSCTL_ADD_U16(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_hai_factor", CTLFLAG_RDTUN, &rf->dcqcn_params.hai_factor, 0, "set number of MSS to add to the congestion window in hyperactive increase mode, default=5"); rf->dcqcn_params.rreduce_mperiod = 50; SYSCTL_ADD_U32(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "dcqcn_rreduce_mperiod", CTLFLAG_RDTUN, &rf->dcqcn_params.rreduce_mperiod, 0, "set minimum time between 2 consecutive rate reductions for a single flow, default=50"); } /** * irdma_sysctl_settings - sysctl runtime settings init * @rf: RDMA PCI function */ void irdma_sysctl_settings(struct irdma_pci_f *rf) { struct sysctl_oid_list *irdma_sysctl_oid_list; irdma_sysctl_oid_list = SYSCTL_CHILDREN(rf->tun_info.irdma_sysctl_tree); SYSCTL_ADD_BOOL(&rf->tun_info.irdma_sysctl_ctx, irdma_sysctl_oid_list, OID_AUTO, "upload_context", CTLFLAG_RWTUN, &irdma_upload_context, 0, "allow for generating QP's upload context, default=0"); } void irdma_sw_stats_tunables_init(struct irdma_pci_f *rf) { struct sysctl_oid_list *sws_oid_list; struct sysctl_ctx_list *irdma_ctx = &rf->tun_info.irdma_sysctl_ctx; struct irdma_sc_dev *dev = &rf->sc_dev; struct irdma_cm_core *cm_core = &rf->iwdev->cm_core; struct irdma_puda_rsrc *ilq = rf->iwdev->vsi.ilq; struct irdma_puda_rsrc *ieq = rf->iwdev->vsi.ieq; u64 *ll_ptr; u32 *l_ptr; int cqp_stat_cnt = sizeof(irdma_sws_list) / sizeof(struct irdma_sw_stats_tunable_info); int cmcore_stat_cnt = sizeof(irdma_cmcs_list) / sizeof(struct irdma_sw_stats_tunable_info); int ilqs_stat_cnt = sizeof(irdma_ilqs_list) / sizeof(struct irdma_sw_stats_tunable_info); int ilqs32_stat_cnt = sizeof(irdma_ilqs32_list) / sizeof(struct irdma_sw_stats_tunable_info); int ieqs_stat_cnt = sizeof(irdma_ieqs_list) / sizeof(struct irdma_sw_stats_tunable_info); int ieqs32_stat_cnt = sizeof(irdma_ieqs32_list) / sizeof(struct irdma_sw_stats_tunable_info); int i; sws_oid_list = SYSCTL_CHILDREN(rf->tun_info.sws_sysctl_tree); for (i = 0; i < cqp_stat_cnt; ++i) { SYSCTL_ADD_U64(irdma_ctx, sws_oid_list, OID_AUTO, irdma_sws_list[i].name, CTLFLAG_RD, &dev->cqp_cmd_stats[irdma_sws_list[i].op_type], 0, irdma_sws_list[i].desc); } SYSCTL_ADD_PROC(irdma_ctx, sws_oid_list, OID_AUTO, "req_cmds", CTLFLAG_RD | CTLTYPE_STRING, dev->cqp, IRDMA_CQP_REQ_CMDS, irdma_sysctl_cqp_stats, "A", "req_cmds"); SYSCTL_ADD_PROC(irdma_ctx, sws_oid_list, OID_AUTO, "cmpl_cmds", CTLFLAG_RD | CTLTYPE_STRING, dev->cqp, IRDMA_CQP_CMPL_CMDS, irdma_sysctl_cqp_stats, "A", "cmpl_cmds"); for (i = 0; i < cmcore_stat_cnt; ++i) { ll_ptr = (u64 *)((uintptr_t)cm_core + irdma_cmcs_list[i].value); SYSCTL_ADD_U64(irdma_ctx, sws_oid_list, OID_AUTO, irdma_cmcs_list[i].name, CTLFLAG_RD, ll_ptr, 0, irdma_cmcs_list[i].desc); } for (i = 0; ilq && i < ilqs_stat_cnt; ++i) { ll_ptr = (u64 *)((uintptr_t)ilq + irdma_ilqs_list[i].value); SYSCTL_ADD_U64(irdma_ctx, sws_oid_list, OID_AUTO, irdma_ilqs_list[i].name, CTLFLAG_RD, ll_ptr, 0, irdma_ilqs_list[i].desc); } for (i = 0; ilq && i < ilqs32_stat_cnt; ++i) { l_ptr = (u32 *)((uintptr_t)ilq + irdma_ilqs32_list[i].value); SYSCTL_ADD_U32(irdma_ctx, sws_oid_list, OID_AUTO, irdma_ilqs32_list[i].name, CTLFLAG_RD, l_ptr, 0, irdma_ilqs32_list[i].desc); } for (i = 0; ieq && i < ieqs_stat_cnt; ++i) { ll_ptr = (u64 *)((uintptr_t)ieq + irdma_ieqs_list[i].value); SYSCTL_ADD_U64(irdma_ctx, sws_oid_list, OID_AUTO, irdma_ieqs_list[i].name, CTLFLAG_RD, ll_ptr, 0, irdma_ieqs_list[i].desc); } for (i = 0; ieq && i < ieqs32_stat_cnt; ++i) { l_ptr = (u32 *)((uintptr_t)ieq + irdma_ieqs32_list[i].value); SYSCTL_ADD_U32(irdma_ctx, sws_oid_list, OID_AUTO, irdma_ieqs32_list[i].name, CTLFLAG_RD, l_ptr, 0, irdma_ieqs32_list[i].desc); } } /** * irdma_dmamap_cb - callback for bus_dmamap_load */ static void irdma_dmamap_cb(void *arg, bus_dma_segment_t * segs, int nseg, int error) { if (error) return; *(bus_addr_t *) arg = segs->ds_addr; return; } /** * irdma_allocate_dma_mem - allocate dma memory * @hw: pointer to hw structure * @mem: structure holding memory information * @size: requested size * @alignment: requested alignment */ void * irdma_allocate_dma_mem(struct irdma_hw *hw, struct irdma_dma_mem *mem, u64 size, u32 alignment) { struct irdma_dev_ctx *dev_ctx = (struct irdma_dev_ctx *)hw->dev_context; device_t dev = dev_ctx->dev; void *va; int ret; ret = bus_dma_tag_create(bus_get_dma_tag(dev), /* parent */ alignment, 0, /* alignment, bounds */ BUS_SPACE_MAXADDR, /* lowaddr */ BUS_SPACE_MAXADDR, /* highaddr */ NULL, NULL, /* filter, filterarg */ size, /* maxsize */ 1, /* nsegments */ size, /* maxsegsize */ BUS_DMA_ALLOCNOW, /* flags */ NULL, /* lockfunc */ NULL, /* lockfuncarg */ &mem->tag); if (ret != 0) { device_printf(dev, "%s: bus_dma_tag_create failed, error %u\n", __func__, ret); goto fail_0; } ret = bus_dmamem_alloc(mem->tag, (void **)&va, BUS_DMA_NOWAIT | BUS_DMA_ZERO, &mem->map); if (ret != 0) { device_printf(dev, "%s: bus_dmamem_alloc failed, error %u\n", __func__, ret); goto fail_1; } ret = bus_dmamap_load(mem->tag, mem->map, va, size, irdma_dmamap_cb, &mem->pa, BUS_DMA_NOWAIT); if (ret != 0) { device_printf(dev, "%s: bus_dmamap_load failed, error %u\n", __func__, ret); goto fail_2; } mem->nseg = 1; mem->size = size; bus_dmamap_sync(mem->tag, mem->map, BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); return va; fail_2: bus_dmamem_free(mem->tag, va, mem->map); fail_1: bus_dma_tag_destroy(mem->tag); fail_0: mem->map = NULL; mem->tag = NULL; return NULL; } /** * irdma_free_dma_mem - Memory free helper fn * @hw: pointer to hw structure * @mem: ptr to mem struct to free */ int irdma_free_dma_mem(struct irdma_hw *hw, struct irdma_dma_mem *mem) { if (!mem) return -EINVAL; bus_dmamap_sync(mem->tag, mem->map, BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE); bus_dmamap_unload(mem->tag, mem->map); if (!mem->va) return -ENOMEM; bus_dmamem_free(mem->tag, mem->va, mem->map); bus_dma_tag_destroy(mem->tag); mem->va = NULL; return 0; } u_int if_foreach_addr_type(if_t ifp, int type, if_addr_cb_t cb, void *cb_arg){ struct epoch_tracker et; struct ifaddr *ifa; u_int count; MPASS(cb); count = 0; NET_EPOCH_ENTER(et); CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != type) continue; count += (*cb) (cb_arg, ifa, count); } NET_EPOCH_EXIT(et); return (count); } int if_foreach(if_foreach_cb_t cb, void *cb_arg) { if_t ifp; int error; NET_EPOCH_ASSERT(); MPASS(cb); error = 0; CK_STAILQ_FOREACH(ifp, &V_ifnet, if_link) { error = cb(ifp, cb_arg); if (error != 0) break; } return (error); } if_t if_iter_start(struct if_iter *iter){ if_t ifp; NET_EPOCH_ASSERT(); bzero(iter, sizeof(*iter)); ifp = CK_STAILQ_FIRST(&V_ifnet); if (ifp != NULL) iter->context[0] = CK_STAILQ_NEXT(ifp, if_link); else iter->context[0] = NULL; return (ifp); } if_t if_iter_next(struct if_iter *iter){ if_t cur_ifp = iter->context[0]; if (cur_ifp != NULL) iter->context[0] = CK_STAILQ_NEXT(cur_ifp, if_link); return (cur_ifp); } void if_iter_finish(struct if_iter *iter) { /* NOP */ } void irdma_cleanup_dead_qps(struct irdma_sc_vsi *vsi) { struct irdma_sc_qp *qp = NULL; struct irdma_qp *iwqp; struct irdma_pci_f *rf; u8 i; for (i = 0; i < IRDMA_MAX_USER_PRIORITY; i++) { qp = irdma_get_qp_from_list(&vsi->qos[i].qplist, qp); while (qp) { if (qp->qp_uk.qp_type == IRDMA_QP_TYPE_UDA) { qp = irdma_get_qp_from_list(&vsi->qos[i].qplist, qp); continue; } iwqp = qp->qp_uk.back_qp; rf = iwqp->iwdev->rf; irdma_free_dma_mem(rf->sc_dev.hw, &iwqp->q2_ctx_mem); irdma_free_dma_mem(rf->sc_dev.hw, &iwqp->kqp.dma_mem); kfree(iwqp->kqp.sq_wrid_mem); kfree(iwqp->kqp.rq_wrid_mem); qp = irdma_get_qp_from_list(&vsi->qos[i].qplist, qp); kfree(iwqp); } } } diff --git a/sys/dev/irdma/fbsd_kcompat.h b/sys/dev/irdma/fbsd_kcompat.h index 630e49068a74..5e656cd9be9f 100644 --- a/sys/dev/irdma/fbsd_kcompat.h +++ b/sys/dev/irdma/fbsd_kcompat.h @@ -1,297 +1,312 @@ /*- * SPDX-License-Identifier: GPL-2.0 or Linux-OpenIB * * Copyright (c) 2021 - 2023 Intel Corporation * * 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 * OpenFabrics.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. */ #ifndef FBSD_KCOMPAT_H #define FBSD_KCOMPAT_H #include "ice_rdma.h" #define TASKLET_DATA_TYPE unsigned long #define TASKLET_FUNC_TYPE void (*)(TASKLET_DATA_TYPE) #ifndef tasklet_setup #define tasklet_setup(tasklet, callback) \ tasklet_init((tasklet), (TASKLET_FUNC_TYPE)(callback), \ (TASKLET_DATA_TYPE)(tasklet)) #endif #ifndef from_tasklet #define from_tasklet(var, callback_tasklet, tasklet_fieldname) \ container_of(callback_tasklet, typeof(*var), tasklet_fieldname) #endif #define set_ibdev_dma_device(ibdev, dev) \ ibdev.dma_device = (dev) #define set_max_sge(props, rf) \ ((props)->max_sge = (rf)->sc_dev.hw_attrs.uk_attrs.max_hw_wq_frags) #define rdma_query_gid(ibdev, port, index, gid) \ ib_get_cached_gid(ibdev, port, index, gid, NULL) #define kmap(pg) page_address(pg) #define kmap_local_page(pg) page_address(pg) #define kunmap(pg) #define kunmap_local(pg) #define IB_UVERBS_CQ_FLAGS_TIMESTAMP_COMPLETION IB_CQ_FLAGS_TIMESTAMP_COMPLETION #define kc_irdma_destroy_qp(ibqp, udata) irdma_destroy_qp(ibqp) #ifndef IB_QP_ATTR_STANDARD_BITS #define IB_QP_ATTR_STANDARD_BITS GENMASK(20, 0) #endif #define IRDMA_QOS_MODE_VLAN 0x0 #define IRDMA_QOS_MODE_DSCP 0x1 #define IRDMA_VER_LEN 24 #ifndef EVNT_HNDLR_CRITERR #if ICE_RDMA_MAJOR_VERSION == 1 && ICE_RDMA_MINOR_VERSION == 1 #define EVNT_HNDLR_CRITERR #else #undef EVNT_HNDLR_CRITERR #endif #endif void kc_set_roce_uverbs_cmd_mask(struct irdma_device *iwdev); void kc_set_rdma_uverbs_cmd_mask(struct irdma_device *iwdev); struct irdma_tunable_info { struct sysctl_ctx_list irdma_sysctl_ctx; struct sysctl_oid *irdma_sysctl_tree; struct sysctl_oid *sws_sysctl_tree; char drv_ver[IRDMA_VER_LEN]; u8 roce_ena; }; typedef u_int if_addr_cb_t(void *, struct ifaddr *, u_int); u_int if_foreach_addr_type(if_t ifp, int type, if_addr_cb_t cb, void *cb_arg); typedef int (*if_foreach_cb_t)(if_t, void *); int if_foreach(if_foreach_cb_t cb, void *cb_arg); #ifndef if_iter struct if_iter { void *context[4]; }; #endif if_t if_iter_start(struct if_iter *iter); if_t if_iter_next(struct if_iter *iter); void if_iter_finish(struct if_iter *iter); #define if_getdunit(ifp) ifp->if_dunit #define if_getindex(ifp) ifp->if_index #define if_getlinkstate(ndev) ndev->if_link_state #define if_getvlantrunk(ifp) ifp->if_vlantrunk #define if_getvnet(ndev) ndev->if_vnet static inline int irdma_iw_query_pkey(struct ib_device *ibdev, u8 port, u16 index, u16 *pkey) { *pkey = 0; return 0; } static inline int cq_validate_flags(u32 flags, u8 hw_rev) { /* GEN1 does not support CQ create flags */ if (hw_rev == IRDMA_GEN_1) return flags ? -EOPNOTSUPP : 0; return flags & ~IB_UVERBS_CQ_FLAGS_TIMESTAMP_COMPLETION ? -EOPNOTSUPP : 0; } + static inline u64 *irdma_next_pbl_addr(u64 *pbl, struct irdma_pble_info **pinfo, u32 *idx) { *idx += 1; if (!(*pinfo) || *idx != (*pinfo)->cnt) return ++pbl; *idx = 0; (*pinfo)++; return (*pinfo)->addr; } + +static inline struct vnet * +irdma_cmid_to_vnet(struct iw_cm_id *cm_id) +{ + struct rdma_cm_id *rdma_id; + + if (!cm_id) + return &init_net; + + rdma_id = (struct rdma_cm_id *)cm_id->context; + + return rdma_id->route.addr.dev_addr.net; +} + struct ib_cq *irdma_create_cq(struct ib_device *ibdev, const struct ib_cq_init_attr *attr, struct ib_ucontext *context, struct ib_udata *udata); struct ib_qp *irdma_create_qp(struct ib_pd *ibpd, struct ib_qp_init_attr *init_attr, struct ib_udata *udata); struct ib_ah *irdma_create_ah(struct ib_pd *ibpd, struct ib_ah_attr *attr, struct ib_udata *udata); struct ib_ah *irdma_create_ah_stub(struct ib_pd *ibpd, struct ib_ah_attr *attr, struct ib_udata *udata); void irdma_ether_copy(u8 *dmac, struct ib_ah_attr *attr); int irdma_destroy_ah(struct ib_ah *ibah); int irdma_destroy_ah_stub(struct ib_ah *ibah); int irdma_destroy_qp(struct ib_qp *ibqp); int irdma_dereg_mr(struct ib_mr *ib_mr); int ib_get_eth_speed(struct ib_device *dev, u32 port_num, u8 *speed, u8 *width); enum rdma_link_layer irdma_get_link_layer(struct ib_device *ibdev, u8 port_num); int irdma_roce_port_immutable(struct ib_device *ibdev, u8 port_num, struct ib_port_immutable *immutable); int irdma_iw_port_immutable(struct ib_device *ibdev, u8 port_num, struct ib_port_immutable *immutable); int irdma_query_gid(struct ib_device *ibdev, u8 port, int index, union ib_gid *gid); int irdma_query_gid_roce(struct ib_device *ibdev, u8 port, int index, union ib_gid *gid); int irdma_query_pkey(struct ib_device *ibdev, u8 port, u16 index, u16 *pkey); int irdma_query_port(struct ib_device *ibdev, u8 port, struct ib_port_attr *props); struct rdma_hw_stats *irdma_alloc_hw_stats(struct ib_device *ibdev, u8 port_num); int irdma_get_hw_stats(struct ib_device *ibdev, struct rdma_hw_stats *stats, u8 port_num, int index); void irdma_request_reset(struct irdma_pci_f *rf); int irdma_register_qset(struct irdma_sc_vsi *vsi, struct irdma_ws_node *tc_node); void irdma_unregister_qset(struct irdma_sc_vsi *vsi, struct irdma_ws_node *tc_node); void ib_unregister_device(struct ib_device *ibdev); int rdma_user_mmap_io(struct ib_ucontext *ucontext, struct vm_area_struct *vma, unsigned long pfn, unsigned long size, pgprot_t prot); void irdma_disassociate_ucontext(struct ib_ucontext *context); int kc_irdma_set_roce_cm_info(struct irdma_qp *iwqp, struct ib_qp_attr *attr, u16 *vlan_id); void kc_set_loc_seq_num_mss(struct irdma_cm_node *cm_node); u16 kc_rdma_get_udp_sport(u32 fl, u32 lqpn, u32 rqpn); void irdma_get_dev_fw_str(struct ib_device *dev, char *str, size_t str_len); int irdma_modify_port(struct ib_device *ibdev, u8 port, int mask, struct ib_port_modify *props); int irdma_get_dst_mac(struct irdma_cm_node *cm_node, struct sockaddr *dst_sin, u8 *dst_mac); int irdma_resolve_neigh_lpb_chk(struct irdma_device *iwdev, struct irdma_cm_node *cm_node, struct irdma_cm_info *cm_info); int irdma_addr_resolve_neigh(struct irdma_cm_node *cm_node, u32 dst_ip, int arpindex); int irdma_addr_resolve_neigh_ipv6(struct irdma_cm_node *cm_node, u32 *dest, int arpindex); void irdma_dcqcn_tunables_init(struct irdma_pci_f *rf); void irdma_sysctl_settings(struct irdma_pci_f *rf); void irdma_sw_stats_tunables_init(struct irdma_pci_f *rf); u32 irdma_create_stag(struct irdma_device *iwdev); void irdma_free_stag(struct irdma_device *iwdev, u32 stag); int irdma_hwdereg_mr(struct ib_mr *ib_mr); int irdma_rereg_user_mr(struct ib_mr *ib_mr, int flags, u64 start, u64 len, u64 virt, int new_access, struct ib_pd *new_pd, struct ib_udata *udata); struct irdma_mr; struct irdma_cq; struct irdma_cq_buf; struct ib_mr *irdma_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type, u32 max_num_sg); int irdma_hwreg_mr(struct irdma_device *iwdev, struct irdma_mr *iwmr, u16 access); struct ib_mr *irdma_rereg_mr_trans(struct irdma_mr *iwmr, u64 start, u64 len, u64 virt, struct ib_udata *udata); int irdma_hw_alloc_mw(struct irdma_device *iwdev, struct irdma_mr *iwmr); struct ib_mw *irdma_alloc_mw(struct ib_pd *pd, enum ib_mw_type type, struct ib_udata *udata); int irdma_hw_alloc_stag(struct irdma_device *iwdev, struct irdma_mr *iwmr); void irdma_cq_free_rsrc(struct irdma_pci_f *rf, struct irdma_cq *iwcq); int irdma_validate_qp_attrs(struct ib_qp_init_attr *init_attr, struct irdma_device *iwdev); void irdma_setup_virt_qp(struct irdma_device *iwdev, struct irdma_qp *iwqp, struct irdma_qp_init_info *init_info); int irdma_setup_kmode_qp(struct irdma_device *iwdev, struct irdma_qp *iwqp, struct irdma_qp_init_info *info, struct ib_qp_init_attr *init_attr); int irdma_setup_umode_qp(struct ib_udata *udata, struct irdma_device *iwdev, struct irdma_qp *iwqp, struct irdma_qp_init_info *info, struct ib_qp_init_attr *init_attr); void irdma_roce_fill_and_set_qpctx_info(struct irdma_qp *iwqp, struct irdma_qp_host_ctx_info *ctx_info); void irdma_iw_fill_and_set_qpctx_info(struct irdma_qp *iwqp, struct irdma_qp_host_ctx_info *ctx_info); int irdma_cqp_create_qp_cmd(struct irdma_qp *iwqp); void irdma_dealloc_push_page(struct irdma_pci_f *rf, struct irdma_sc_qp *qp); int irdma_process_resize_list(struct irdma_cq *iwcq, struct irdma_device *iwdev, struct irdma_cq_buf *lcqe_buf); int irdma_destroy_cq(struct ib_cq *ib_cq); struct ib_ucontext *irdma_alloc_ucontext(struct ib_device *, struct ib_udata *); int irdma_dealloc_ucontext(struct ib_ucontext *); struct ib_pd *irdma_alloc_pd(struct ib_device *, struct ib_ucontext *, struct ib_udata *); int irdma_dealloc_pd(struct ib_pd *); int irdma_add_gid(struct ib_device *, u8, unsigned int, const union ib_gid *, const struct ib_gid_attr *, void **); int irdma_del_gid(struct ib_device *, u8, unsigned int, void **); struct ib_device *ib_device_get_by_netdev(struct ifnet *ndev, int driver_id); void ib_device_put(struct ib_device *device); void ib_unregister_device_put(struct ib_device *device); enum ib_mtu ib_mtu_int_to_enum(int mtu); struct irdma_pbl *irdma_get_pbl(unsigned long va, struct list_head *pbl_list); void irdma_clean_cqes(struct irdma_qp *iwqp, struct irdma_cq *iwcq); void irdma_remove_push_mmap_entries(struct irdma_qp *iwqp); struct irdma_ucontext; void irdma_del_memlist(struct irdma_mr *iwmr, struct irdma_ucontext *ucontext); void irdma_copy_user_pgaddrs(struct irdma_mr *iwmr, u64 *pbl, enum irdma_pble_level level); void irdma_reg_ipaddr_event_cb(struct irdma_pci_f *rf); void irdma_dereg_ipaddr_event_cb(struct irdma_pci_f *rf); /* Introduced in this series https://lore.kernel.org/linux-rdma/0-v2-270386b7e60b+28f4-umem_1_jgg@nvidia.com/ * An irdma version helper doing same for older functions with difference that iova is passed in * as opposed to derived from umem->iova. */ static inline size_t irdma_ib_umem_num_dma_blocks(struct ib_umem *umem, unsigned long pgsz, u64 iova) { /* some older OFED distros do not have ALIGN_DOWN */ #ifndef ALIGN_DOWN #define ALIGN_DOWN(x, a) ALIGN((x) - ((a) - 1), (a)) #endif return (size_t)((ALIGN(iova + umem->length, pgsz) - ALIGN_DOWN(iova, pgsz))) / pgsz; } static inline void addrconf_addr_eui48(u8 *deui, const char *const addr) { memcpy(deui, addr, 3); deui[3] = 0xFF; deui[4] = 0xFE; memcpy(deui + 5, addr + 3, 3); deui[0] ^= 2; } #endif /* FBSD_KCOMPAT_H */ diff --git a/sys/dev/irdma/irdma_cm.c b/sys/dev/irdma/irdma_cm.c index cb508bf52cb8..d34da57cc84d 100644 --- a/sys/dev/irdma/irdma_cm.c +++ b/sys/dev/irdma/irdma_cm.c @@ -1,4261 +1,4274 @@ /*- * SPDX-License-Identifier: GPL-2.0 or Linux-OpenIB * * Copyright (c) 2015 - 2023 Intel Corporation * * 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 * OpenFabrics.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 "irdma_main.h" static void irdma_cm_post_event(struct irdma_cm_event *event); static void irdma_disconnect_worker(struct work_struct *work); /** * irdma_free_sqbuf - put back puda buffer if refcount is 0 * @vsi: The VSI structure of the device * @bufp: puda buffer to free */ void irdma_free_sqbuf(struct irdma_sc_vsi *vsi, void *bufp) { struct irdma_puda_buf *buf = bufp; struct irdma_puda_rsrc *ilq = vsi->ilq; if (atomic_dec_and_test(&buf->refcount)) irdma_puda_ret_bufpool(ilq, buf); } /** * irdma_record_ird_ord - Record IRD/ORD passed in * @cm_node: connection's node * @conn_ird: connection IRD * @conn_ord: connection ORD */ static void irdma_record_ird_ord(struct irdma_cm_node *cm_node, u32 conn_ird, u32 conn_ord) { if (conn_ird > cm_node->dev->hw_attrs.max_hw_ird) conn_ird = cm_node->dev->hw_attrs.max_hw_ird; if (conn_ord > cm_node->dev->hw_attrs.max_hw_ord) conn_ord = cm_node->dev->hw_attrs.max_hw_ord; else if (!conn_ord && cm_node->send_rdma0_op == SEND_RDMA_READ_ZERO) conn_ord = 1; cm_node->ird_size = conn_ird; cm_node->ord_size = conn_ord; } /** * irdma_copy_ip_ntohl - copy IP address from network to host * @dst: IP address in host order * @src: IP address in network order (big endian) */ void irdma_copy_ip_ntohl(u32 *dst, __be32 *src) { *dst++ = ntohl(*src++); *dst++ = ntohl(*src++); *dst++ = ntohl(*src++); *dst = ntohl(*src); } /** * irdma_copy_ip_htonl - copy IP address from host to network order * @dst: IP address in network order (big endian) * @src: IP address in host order */ void irdma_copy_ip_htonl(__be32 *dst, u32 *src) { *dst++ = htonl(*src++); *dst++ = htonl(*src++); *dst++ = htonl(*src++); *dst = htonl(*src); } /** * irdma_get_addr_info * @cm_node: contains ip/tcp info * @cm_info: to get a copy of the cm_node ip/tcp info */ static void irdma_get_addr_info(struct irdma_cm_node *cm_node, struct irdma_cm_info *cm_info) { memset(cm_info, 0, sizeof(*cm_info)); cm_info->ipv4 = cm_node->ipv4; cm_info->vlan_id = cm_node->vlan_id; memcpy(cm_info->loc_addr, cm_node->loc_addr, sizeof(cm_info->loc_addr)); memcpy(cm_info->rem_addr, cm_node->rem_addr, sizeof(cm_info->rem_addr)); cm_info->loc_port = cm_node->loc_port; cm_info->rem_port = cm_node->rem_port; } /** * irdma_fill_sockaddr4 - fill in addr info for IPv4 connection * @cm_node: connection's node * @event: upper layer's cm event */ static inline void irdma_fill_sockaddr4(struct irdma_cm_node *cm_node, struct iw_cm_event *event) { struct sockaddr_in *laddr = (struct sockaddr_in *)&event->local_addr; struct sockaddr_in *raddr = (struct sockaddr_in *)&event->remote_addr; laddr->sin_family = AF_INET; raddr->sin_family = AF_INET; laddr->sin_port = htons(cm_node->loc_port); raddr->sin_port = htons(cm_node->rem_port); laddr->sin_addr.s_addr = htonl(cm_node->loc_addr[0]); raddr->sin_addr.s_addr = htonl(cm_node->rem_addr[0]); } /** * irdma_fill_sockaddr6 - fill in addr info for IPv6 connection * @cm_node: connection's node * @event: upper layer's cm event */ static inline void irdma_fill_sockaddr6(struct irdma_cm_node *cm_node, struct iw_cm_event *event) { struct sockaddr_in6 *laddr6 = (struct sockaddr_in6 *)&event->local_addr; struct sockaddr_in6 *raddr6 = (struct sockaddr_in6 *)&event->remote_addr; laddr6->sin6_family = AF_INET6; raddr6->sin6_family = AF_INET6; laddr6->sin6_port = htons(cm_node->loc_port); raddr6->sin6_port = htons(cm_node->rem_port); irdma_copy_ip_htonl(laddr6->sin6_addr.__u6_addr.__u6_addr32, cm_node->loc_addr); irdma_copy_ip_htonl(raddr6->sin6_addr.__u6_addr.__u6_addr32, cm_node->rem_addr); } /** * irdma_get_cmevent_info - for cm event upcall * @cm_node: connection's node * @cm_id: upper layers cm struct for the event * @event: upper layer's cm event */ static inline void irdma_get_cmevent_info(struct irdma_cm_node *cm_node, struct iw_cm_id *cm_id, struct iw_cm_event *event) { memcpy(&event->local_addr, &cm_id->m_local_addr, sizeof(event->local_addr)); memcpy(&event->remote_addr, &cm_id->m_remote_addr, sizeof(event->remote_addr)); if (cm_node) { event->private_data = cm_node->pdata_buf; event->private_data_len = (u8)cm_node->pdata.size; event->ird = cm_node->ird_size; event->ord = cm_node->ord_size; } } /** * irdma_send_cm_event - upcall cm's event handler * @cm_node: connection's node * @cm_id: upper layer's cm info struct * @type: Event type to indicate * @status: status for the event type */ static int irdma_send_cm_event(struct irdma_cm_node *cm_node, struct iw_cm_id *cm_id, enum iw_cm_event_type type, int status) { struct iw_cm_event event = {0}; event.event = type; event.status = status; irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "cm_node %p cm_id=%p state=%d accel=%d event_type=%d status=%d\n", cm_node, cm_id, cm_node->accelerated, cm_node->state, type, status); switch (type) { case IW_CM_EVENT_CONNECT_REQUEST: if (cm_node->ipv4) irdma_fill_sockaddr4(cm_node, &event); else irdma_fill_sockaddr6(cm_node, &event); event.provider_data = cm_node; event.private_data = cm_node->pdata_buf; event.private_data_len = (u8)cm_node->pdata.size; event.ird = cm_node->ird_size; break; case IW_CM_EVENT_CONNECT_REPLY: irdma_get_cmevent_info(cm_node, cm_id, &event); break; case IW_CM_EVENT_ESTABLISHED: event.ird = cm_node->ird_size; event.ord = cm_node->ord_size; break; case IW_CM_EVENT_DISCONNECT: case IW_CM_EVENT_CLOSE: /* Wait if we are in RTS but havent issued the iwcm event upcall */ if (!cm_node->accelerated) wait_for_completion(&cm_node->establish_comp); break; default: return -EINVAL; } return cm_id->event_handler(cm_id, &event); } /** * irdma_timer_list_prep - add connection nodes to a list to perform timer tasks * @cm_core: cm's core * @timer_list: a timer list to which cm_node will be selected */ static void irdma_timer_list_prep(struct irdma_cm_core *cm_core, struct list_head *timer_list) { struct irdma_cm_node *cm_node; int bkt; HASH_FOR_EACH_RCU(cm_core->cm_hash_tbl, bkt, cm_node, list) { if ((cm_node->close_entry || cm_node->send_entry) && atomic_inc_not_zero(&cm_node->refcnt)) list_add(&cm_node->timer_entry, timer_list); } } /** * irdma_create_event - create cm event * @cm_node: connection's node * @type: Event type to generate */ static struct irdma_cm_event * irdma_create_event(struct irdma_cm_node *cm_node, enum irdma_cm_event_type type) { struct irdma_cm_event *event; if (!cm_node->cm_id) return NULL; event = kzalloc(sizeof(*event), GFP_ATOMIC); if (!event) return NULL; event->type = type; event->cm_node = cm_node; memcpy(event->cm_info.rem_addr, cm_node->rem_addr, sizeof(event->cm_info.rem_addr)); memcpy(event->cm_info.loc_addr, cm_node->loc_addr, sizeof(event->cm_info.loc_addr)); event->cm_info.rem_port = cm_node->rem_port; event->cm_info.loc_port = cm_node->loc_port; event->cm_info.cm_id = cm_node->cm_id; irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "node=%p event=%p type=%u dst=%x src=%x\n", cm_node, event, type, event->cm_info.loc_addr[0], event->cm_info.rem_addr[0]); irdma_cm_post_event(event); return event; } /** * irdma_free_retrans_entry - free send entry * @cm_node: connection's node */ static void irdma_free_retrans_entry(struct irdma_cm_node *cm_node) { struct irdma_device *iwdev = cm_node->iwdev; struct irdma_timer_entry *send_entry; send_entry = cm_node->send_entry; if (!send_entry) return; cm_node->send_entry = NULL; irdma_free_sqbuf(&iwdev->vsi, send_entry->sqbuf); kfree(send_entry); atomic_dec(&cm_node->refcnt); } /** * irdma_cleanup_retrans_entry - free send entry with lock * @cm_node: connection's node */ static void irdma_cleanup_retrans_entry(struct irdma_cm_node *cm_node) { unsigned long flags; spin_lock_irqsave(&cm_node->retrans_list_lock, flags); irdma_free_retrans_entry(cm_node); spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); } /** * irdma_form_ah_cm_frame - get a free packet and build frame with address handle * @cm_node: connection's node ionfo to use in frame * @options: pointer to options info * @hdr: pointer mpa header * @pdata: pointer to private data * @flags: indicates FIN or ACK */ static struct irdma_puda_buf * irdma_form_ah_cm_frame(struct irdma_cm_node *cm_node, struct irdma_kmem_info *options, struct irdma_kmem_info *hdr, struct irdma_mpa_priv_info *pdata, u8 flags) { struct irdma_puda_buf *sqbuf; struct irdma_sc_vsi *vsi = &cm_node->iwdev->vsi; u8 *buf; struct tcphdr *tcph; u16 pktsize; u32 opts_len = 0; u32 pd_len = 0; u32 hdr_len = 0; if (!cm_node->ah || !cm_node->ah->ah_info.ah_valid) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "AH invalid\n"); return NULL; } sqbuf = irdma_puda_get_bufpool(vsi->ilq); if (!sqbuf) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "SQ buf NULL\n"); return NULL; } sqbuf->ah_id = cm_node->ah->ah_info.ah_idx; buf = sqbuf->mem.va; if (options) opts_len = (u32)options->size; if (hdr) hdr_len = hdr->size; if (pdata) pd_len = pdata->size; pktsize = sizeof(*tcph) + opts_len + hdr_len + pd_len; memset(buf, 0, pktsize); sqbuf->totallen = pktsize; sqbuf->tcphlen = sizeof(*tcph) + opts_len; sqbuf->scratch = cm_node; tcph = (struct tcphdr *)buf; buf += sizeof(*tcph); tcph->th_sport = htons(cm_node->loc_port); tcph->th_dport = htons(cm_node->rem_port); tcph->th_seq = htonl(cm_node->tcp_cntxt.loc_seq_num); if (flags & SET_ACK) { cm_node->tcp_cntxt.loc_ack_num = cm_node->tcp_cntxt.rcv_nxt; tcph->th_ack = htonl(cm_node->tcp_cntxt.loc_ack_num); tcph->th_flags |= TH_ACK; } else { tcph->th_ack = 0; } if (flags & SET_SYN) { cm_node->tcp_cntxt.loc_seq_num++; tcph->th_flags |= TH_SYN; } else { cm_node->tcp_cntxt.loc_seq_num += hdr_len + pd_len; } if (flags & SET_FIN) { cm_node->tcp_cntxt.loc_seq_num++; tcph->th_flags |= TH_FIN; } if (flags & SET_RST) tcph->th_flags |= TH_RST; tcph->th_off = (u16)((sizeof(*tcph) + opts_len + 3) >> 2); sqbuf->tcphlen = tcph->th_off << 2; tcph->th_win = htons(cm_node->tcp_cntxt.rcv_wnd); tcph->th_urp = 0; if (opts_len) { memcpy(buf, options->addr, opts_len); buf += opts_len; } if (hdr_len) { memcpy(buf, hdr->addr, hdr_len); buf += hdr_len; } if (pdata && pdata->addr) memcpy(buf, pdata->addr, pdata->size); atomic_set(&sqbuf->refcount, 1); irdma_debug_buf(vsi->dev, IRDMA_DEBUG_ILQ, "TRANSMIT ILQ BUFFER", sqbuf->mem.va, sqbuf->totallen); return sqbuf; } /** * irdma_form_uda_cm_frame - get a free packet and build frame full tcpip packet * @cm_node: connection's node ionfo to use in frame * @options: pointer to options info * @hdr: pointer mpa header * @pdata: pointer to private data * @flags: indicates FIN or ACK */ static struct irdma_puda_buf * irdma_form_uda_cm_frame(struct irdma_cm_node *cm_node, struct irdma_kmem_info *options, struct irdma_kmem_info *hdr, struct irdma_mpa_priv_info *pdata, u8 flags) { struct irdma_puda_buf *sqbuf; struct irdma_sc_vsi *vsi = &cm_node->iwdev->vsi; u8 *buf; struct tcphdr *tcph; struct ip *iph; struct ip6_hdr *ip6h; struct ether_header *ethh; u16 pktsize; u16 eth_hlen = ETH_HLEN; u32 opts_len = 0; u32 pd_len = 0; u32 hdr_len = 0; u16 vtag; sqbuf = irdma_puda_get_bufpool(vsi->ilq); if (!sqbuf) return NULL; buf = sqbuf->mem.va; if (options) opts_len = (u32)options->size; if (hdr) hdr_len = hdr->size; if (pdata) pd_len = pdata->size; if (cm_node->vlan_id < VLAN_N_VID) eth_hlen += 4; if (cm_node->ipv4) pktsize = sizeof(*iph) + sizeof(*tcph); else pktsize = sizeof(*ip6h) + sizeof(*tcph); pktsize += opts_len + hdr_len + pd_len; memset(buf, 0, eth_hlen + pktsize); sqbuf->totallen = pktsize + eth_hlen; sqbuf->maclen = eth_hlen; sqbuf->tcphlen = sizeof(*tcph) + opts_len; sqbuf->scratch = cm_node; ethh = (struct ether_header *)buf; buf += eth_hlen; if (cm_node->do_lpb) sqbuf->do_lpb = true; if (cm_node->ipv4) { sqbuf->ipv4 = true; iph = (struct ip *)buf; buf += sizeof(*iph); tcph = (struct tcphdr *)buf; buf += sizeof(*tcph); ether_addr_copy(ethh->ether_dhost, cm_node->rem_mac); ether_addr_copy(ethh->ether_shost, cm_node->loc_mac); if (cm_node->vlan_id < VLAN_N_VID) { ((struct ether_vlan_header *)ethh)->evl_proto = htons(ETH_P_8021Q); vtag = (cm_node->user_pri << VLAN_PRIO_SHIFT) | cm_node->vlan_id; ((struct ether_vlan_header *)ethh)->evl_tag = htons(vtag); ((struct ether_vlan_header *)ethh)->evl_encap_proto = htons(ETH_P_IP); } else { ethh->ether_type = htons(ETH_P_IP); } iph->ip_v = IPVERSION; iph->ip_hl = 5; /* 5 * 4Byte words, IP headr len */ iph->ip_tos = cm_node->tos; iph->ip_len = htons(pktsize); iph->ip_id = htons(++cm_node->tcp_cntxt.loc_id); iph->ip_off = htons(0x4000); iph->ip_ttl = 0x40; iph->ip_p = IPPROTO_TCP; iph->ip_src.s_addr = htonl(cm_node->loc_addr[0]); iph->ip_dst.s_addr = htonl(cm_node->rem_addr[0]); } else { sqbuf->ipv4 = false; ip6h = (struct ip6_hdr *)buf; buf += sizeof(*ip6h); tcph = (struct tcphdr *)buf; buf += sizeof(*tcph); ether_addr_copy(ethh->ether_dhost, cm_node->rem_mac); ether_addr_copy(ethh->ether_shost, cm_node->loc_mac); if (cm_node->vlan_id < VLAN_N_VID) { ((struct ether_vlan_header *)ethh)->evl_proto = htons(ETH_P_8021Q); vtag = (cm_node->user_pri << VLAN_PRIO_SHIFT) | cm_node->vlan_id; ((struct ether_vlan_header *)ethh)->evl_tag = htons(vtag); ((struct ether_vlan_header *)ethh)->evl_encap_proto = htons(ETH_P_IPV6); } else { ethh->ether_type = htons(ETH_P_IPV6); } ip6h->ip6_vfc = 6 << 4; ip6h->ip6_vfc |= cm_node->tos >> 4; ip6h->ip6_flow = cm_node->tos << 20; ip6h->ip6_plen = htons(pktsize - sizeof(*ip6h)); ip6h->ip6_nxt = 6; ip6h->ip6_hops = 128; irdma_copy_ip_htonl(ip6h->ip6_src.__u6_addr.__u6_addr32, cm_node->loc_addr); irdma_copy_ip_htonl(ip6h->ip6_dst.__u6_addr.__u6_addr32, cm_node->rem_addr); } tcph->th_sport = htons(cm_node->loc_port); tcph->th_dport = htons(cm_node->rem_port); tcph->th_seq = htonl(cm_node->tcp_cntxt.loc_seq_num); if (flags & SET_ACK) { cm_node->tcp_cntxt.loc_ack_num = cm_node->tcp_cntxt.rcv_nxt; tcph->th_ack = htonl(cm_node->tcp_cntxt.loc_ack_num); tcph->th_flags |= TH_ACK; } else { tcph->th_ack = 0; } if (flags & SET_SYN) { cm_node->tcp_cntxt.loc_seq_num++; tcph->th_flags |= TH_SYN; } else { cm_node->tcp_cntxt.loc_seq_num += hdr_len + pd_len; } if (flags & SET_FIN) { cm_node->tcp_cntxt.loc_seq_num++; tcph->th_flags |= TH_FIN; } if (flags & SET_RST) tcph->th_flags |= TH_RST; tcph->th_off = (u16)((sizeof(*tcph) + opts_len + 3) >> 2); sqbuf->tcphlen = tcph->th_off << 2; tcph->th_win = htons(cm_node->tcp_cntxt.rcv_wnd); tcph->th_urp = 0; if (opts_len) { memcpy(buf, options->addr, opts_len); buf += opts_len; } if (hdr_len) { memcpy(buf, hdr->addr, hdr_len); buf += hdr_len; } if (pdata && pdata->addr) memcpy(buf, pdata->addr, pdata->size); atomic_set(&sqbuf->refcount, 1); irdma_debug_buf(vsi->dev, IRDMA_DEBUG_ILQ, "TRANSMIT ILQ BUFFER", sqbuf->mem.va, sqbuf->totallen); return sqbuf; } /** * irdma_send_reset - Send RST packet * @cm_node: connection's node */ int irdma_send_reset(struct irdma_cm_node *cm_node) { struct irdma_puda_buf *sqbuf; int flags = SET_RST | SET_ACK; sqbuf = cm_node->cm_core->form_cm_frame(cm_node, NULL, NULL, NULL, flags); if (!sqbuf) return -ENOMEM; irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "caller: %pS cm_node %p cm_id=%p accel=%d state=%d rem_port=0x%04x, loc_port=0x%04x rem_addr=%x loc_addr=%x\n", __builtin_return_address(0), cm_node, cm_node->cm_id, cm_node->accelerated, cm_node->state, cm_node->rem_port, cm_node->loc_port, cm_node->rem_addr[0], cm_node->loc_addr[0]); return irdma_schedule_cm_timer(cm_node, sqbuf, IRDMA_TIMER_TYPE_SEND, 0, 1); } /** * irdma_active_open_err - send event for active side cm error * @cm_node: connection's node * @reset: Flag to send reset or not */ static void irdma_active_open_err(struct irdma_cm_node *cm_node, bool reset) { irdma_cleanup_retrans_entry(cm_node); cm_node->cm_core->stats_connect_errs++; if (reset) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "cm_node=%p state=%d\n", cm_node, cm_node->state); atomic_inc(&cm_node->refcnt); irdma_send_reset(cm_node); } cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_create_event(cm_node, IRDMA_CM_EVENT_ABORTED); } /** * irdma_passive_open_err - handle passive side cm error * @cm_node: connection's node * @reset: send reset or just free cm_node */ static void irdma_passive_open_err(struct irdma_cm_node *cm_node, bool reset) { irdma_cleanup_retrans_entry(cm_node); cm_node->cm_core->stats_passive_errs++; cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "cm_node=%p state=%d\n", cm_node, cm_node->state); if (reset) irdma_send_reset(cm_node); else irdma_rem_ref_cm_node(cm_node); } /** * irdma_event_connect_error - to create connect error event * @event: cm information for connect event */ static void irdma_event_connect_error(struct irdma_cm_event *event) { struct irdma_qp *iwqp; struct iw_cm_id *cm_id; cm_id = event->cm_node->cm_id; if (!cm_id) return; iwqp = cm_id->provider_data; if (!iwqp || !iwqp->iwdev) return; iwqp->cm_id = NULL; cm_id->provider_data = NULL; irdma_send_cm_event(event->cm_node, cm_id, IW_CM_EVENT_CONNECT_REPLY, -ECONNRESET); irdma_rem_ref_cm_node(event->cm_node); } /** * irdma_process_options - process options from TCP header * @cm_node: connection's node * @optionsloc: point to start of options * @optionsize: size of all options * @syn_pkt: flag if syn packet */ static int irdma_process_options(struct irdma_cm_node *cm_node, u8 *optionsloc, u32 optionsize, u32 syn_pkt) { u32 tmp; u32 offset = 0; union all_known_options *all_options; char got_mss_option = 0; while (offset < optionsize) { all_options = (union all_known_options *)(optionsloc + offset); switch (all_options->base.optionnum) { case OPTION_NUM_EOL: offset = optionsize; break; case OPTION_NUM_NONE: offset += 1; continue; case OPTION_NUM_MSS: irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "MSS Length: %d Offset: %d Size: %d\n", all_options->mss.len, offset, optionsize); got_mss_option = 1; if (all_options->mss.len != 4) return -EINVAL; tmp = ntohs(all_options->mss.mss); if ((cm_node->ipv4 && (tmp + IRDMA_MTU_TO_MSS_IPV4) < IRDMA_MIN_MTU_IPV4) || (!cm_node->ipv4 && (tmp + IRDMA_MTU_TO_MSS_IPV6) < IRDMA_MIN_MTU_IPV6)) return -EINVAL; if (tmp < cm_node->tcp_cntxt.mss) cm_node->tcp_cntxt.mss = tmp; break; case OPTION_NUM_WINDOW_SCALE: cm_node->tcp_cntxt.snd_wscale = all_options->windowscale.shiftcount; break; default: irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Unsupported TCP Option: %x\n", all_options->base.optionnum); break; } offset += all_options->base.len; } if (!got_mss_option && syn_pkt) cm_node->tcp_cntxt.mss = IRDMA_CM_DEFAULT_MSS; return 0; } /** * irdma_handle_tcp_options - setup TCP context info after parsing TCP options * @cm_node: connection's node * @tcph: pointer tcp header * @optionsize: size of options rcvd * @passive: active or passive flag */ static int irdma_handle_tcp_options(struct irdma_cm_node *cm_node, struct tcphdr *tcph, int optionsize, int passive) { u8 *optionsloc = (u8 *)&tcph[1]; int ret; if (optionsize) { ret = irdma_process_options(cm_node, optionsloc, optionsize, (u32)tcph->th_flags & TH_SYN); if (ret) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Node %p, Sending Reset\n", cm_node); if (passive) irdma_passive_open_err(cm_node, true); else irdma_active_open_err(cm_node, true); return ret; } } cm_node->tcp_cntxt.snd_wnd = ntohs(tcph->th_win) << cm_node->tcp_cntxt.snd_wscale; if (cm_node->tcp_cntxt.snd_wnd > cm_node->tcp_cntxt.max_snd_wnd) cm_node->tcp_cntxt.max_snd_wnd = cm_node->tcp_cntxt.snd_wnd; return 0; } /** * irdma_build_mpa_v1 - build a MPA V1 frame * @cm_node: connection's node * @start_addr: address where to build frame * @mpa_key: to do read0 or write0 */ static void irdma_build_mpa_v1(struct irdma_cm_node *cm_node, void *start_addr, u8 mpa_key) { struct ietf_mpa_v1 *mpa_frame = start_addr; switch (mpa_key) { case MPA_KEY_REQUEST: memcpy(mpa_frame->key, IEFT_MPA_KEY_REQ, IETF_MPA_KEY_SIZE); break; case MPA_KEY_REPLY: memcpy(mpa_frame->key, IEFT_MPA_KEY_REP, IETF_MPA_KEY_SIZE); break; default: break; } mpa_frame->flags = IETF_MPA_FLAGS_CRC; mpa_frame->rev = cm_node->mpa_frame_rev; mpa_frame->priv_data_len = htons(cm_node->pdata.size); } /** * irdma_build_mpa_v2 - build a MPA V2 frame * @cm_node: connection's node * @start_addr: buffer start address * @mpa_key: to do read0 or write0 */ static void irdma_build_mpa_v2(struct irdma_cm_node *cm_node, void *start_addr, u8 mpa_key) { struct ietf_mpa_v2 *mpa_frame = start_addr; struct ietf_rtr_msg *rtr_msg = &mpa_frame->rtr_msg; u16 ctrl_ird, ctrl_ord; /* initialize the upper 5 bytes of the frame */ irdma_build_mpa_v1(cm_node, start_addr, mpa_key); mpa_frame->flags |= IETF_MPA_V2_FLAG; if (cm_node->iwdev->iw_ooo) { mpa_frame->flags |= IETF_MPA_FLAGS_MARKERS; cm_node->rcv_mark_en = true; } mpa_frame->priv_data_len = cpu_to_be16(be16_to_cpu(mpa_frame->priv_data_len) + IETF_RTR_MSG_SIZE); /* initialize RTR msg */ if (cm_node->mpav2_ird_ord == IETF_NO_IRD_ORD) { ctrl_ird = IETF_NO_IRD_ORD; ctrl_ord = IETF_NO_IRD_ORD; } else { ctrl_ird = (cm_node->ird_size > IETF_NO_IRD_ORD) ? IETF_NO_IRD_ORD : cm_node->ird_size; ctrl_ord = (cm_node->ord_size > IETF_NO_IRD_ORD) ? IETF_NO_IRD_ORD : cm_node->ord_size; } ctrl_ird |= IETF_PEER_TO_PEER; switch (mpa_key) { case MPA_KEY_REQUEST: ctrl_ord |= IETF_RDMA0_WRITE; ctrl_ord |= IETF_RDMA0_READ; break; case MPA_KEY_REPLY: switch (cm_node->send_rdma0_op) { case SEND_RDMA_WRITE_ZERO: ctrl_ord |= IETF_RDMA0_WRITE; break; case SEND_RDMA_READ_ZERO: ctrl_ord |= IETF_RDMA0_READ; break; } break; default: break; } rtr_msg->ctrl_ird = htons(ctrl_ird); rtr_msg->ctrl_ord = htons(ctrl_ord); } /** * irdma_cm_build_mpa_frame - build mpa frame for mpa version 1 or version 2 * @cm_node: connection's node * @mpa: mpa: data buffer * @mpa_key: to do read0 or write0 */ static int irdma_cm_build_mpa_frame(struct irdma_cm_node *cm_node, struct irdma_kmem_info *mpa, u8 mpa_key) { int hdr_len = 0; switch (cm_node->mpa_frame_rev) { case IETF_MPA_V1: hdr_len = sizeof(struct ietf_mpa_v1); irdma_build_mpa_v1(cm_node, mpa->addr, mpa_key); break; case IETF_MPA_V2: hdr_len = sizeof(struct ietf_mpa_v2); irdma_build_mpa_v2(cm_node, mpa->addr, mpa_key); break; default: break; } return hdr_len; } /** * irdma_send_mpa_request - active node send mpa request to passive node * @cm_node: connection's node */ static int irdma_send_mpa_request(struct irdma_cm_node *cm_node) { struct irdma_puda_buf *sqbuf; cm_node->mpa_hdr.addr = &cm_node->mpa_v2_frame; cm_node->mpa_hdr.size = irdma_cm_build_mpa_frame(cm_node, &cm_node->mpa_hdr, MPA_KEY_REQUEST); if (!cm_node->mpa_hdr.size) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "mpa size = %d\n", cm_node->mpa_hdr.size); return -EINVAL; } sqbuf = cm_node->cm_core->form_cm_frame(cm_node, NULL, &cm_node->mpa_hdr, &cm_node->pdata, SET_ACK); if (!sqbuf) return -ENOMEM; return irdma_schedule_cm_timer(cm_node, sqbuf, IRDMA_TIMER_TYPE_SEND, 1, 0); } /** * irdma_send_mpa_reject - * @cm_node: connection's node * @pdata: reject data for connection * @plen: length of reject data */ static int irdma_send_mpa_reject(struct irdma_cm_node *cm_node, const void *pdata, u8 plen) { struct irdma_puda_buf *sqbuf; struct irdma_mpa_priv_info priv_info; cm_node->mpa_hdr.addr = &cm_node->mpa_v2_frame; cm_node->mpa_hdr.size = irdma_cm_build_mpa_frame(cm_node, &cm_node->mpa_hdr, MPA_KEY_REPLY); cm_node->mpa_v2_frame.flags |= IETF_MPA_FLAGS_REJECT; priv_info.addr = pdata; priv_info.size = plen; sqbuf = cm_node->cm_core->form_cm_frame(cm_node, NULL, &cm_node->mpa_hdr, &priv_info, SET_ACK | SET_FIN); if (!sqbuf) return -ENOMEM; cm_node->state = IRDMA_CM_STATE_FIN_WAIT1; return irdma_schedule_cm_timer(cm_node, sqbuf, IRDMA_TIMER_TYPE_SEND, 1, 0); } /** * irdma_negotiate_mpa_v2_ird_ord - negotiate MPAv2 IRD/ORD * @cm_node: connection's node * @buf: Data pointer */ static int irdma_negotiate_mpa_v2_ird_ord(struct irdma_cm_node *cm_node, u8 *buf) { struct ietf_mpa_v2 *mpa_v2_frame; struct ietf_rtr_msg *rtr_msg; u16 ird_size; u16 ord_size; u16 ctrl_ord; u16 ctrl_ird; mpa_v2_frame = (struct ietf_mpa_v2 *)buf; rtr_msg = &mpa_v2_frame->rtr_msg; /* parse rtr message */ ctrl_ord = ntohs(rtr_msg->ctrl_ord); ctrl_ird = ntohs(rtr_msg->ctrl_ird); ird_size = ctrl_ird & IETF_NO_IRD_ORD; ord_size = ctrl_ord & IETF_NO_IRD_ORD; if (!(ctrl_ird & IETF_PEER_TO_PEER)) return -EOPNOTSUPP; if (ird_size == IETF_NO_IRD_ORD || ord_size == IETF_NO_IRD_ORD) { cm_node->mpav2_ird_ord = IETF_NO_IRD_ORD; goto negotiate_done; } if (cm_node->state != IRDMA_CM_STATE_MPAREQ_SENT) { /* responder */ if (!ord_size && (ctrl_ord & IETF_RDMA0_READ)) cm_node->ird_size = 1; if (cm_node->ord_size > ird_size) cm_node->ord_size = ird_size; } else { /* initiator */ if (!ird_size && (ctrl_ord & IETF_RDMA0_READ)) /* Remote peer doesn't support RDMA0_READ */ return -EOPNOTSUPP; if (cm_node->ord_size > ird_size) cm_node->ord_size = ird_size; if (cm_node->ird_size < ord_size) /* no resources available */ return -EINVAL; } negotiate_done: if (ctrl_ord & IETF_RDMA0_READ) cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO; else if (ctrl_ord & IETF_RDMA0_WRITE) cm_node->send_rdma0_op = SEND_RDMA_WRITE_ZERO; else /* Not supported RDMA0 operation */ return -EOPNOTSUPP; irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "MPAV2 Negotiated ORD: %d, IRD: %d\n", cm_node->ord_size, cm_node->ird_size); return 0; } /** * irdma_parse_mpa - process an IETF MPA frame * @cm_node: connection's node * @buf: Data pointer * @type: to return accept or reject * @len: Len of mpa buffer */ static int irdma_parse_mpa(struct irdma_cm_node *cm_node, u8 *buf, u32 *type, u32 len) { struct ietf_mpa_v1 *mpa_frame; int mpa_hdr_len, priv_data_len, ret; *type = IRDMA_MPA_REQUEST_ACCEPT; if (len < sizeof(struct ietf_mpa_v1)) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "ietf buffer small (%x)\n", len); return -EINVAL; } mpa_frame = (struct ietf_mpa_v1 *)buf; mpa_hdr_len = sizeof(struct ietf_mpa_v1); priv_data_len = ntohs(mpa_frame->priv_data_len); if (priv_data_len > IETF_MAX_PRIV_DATA_LEN) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "private_data too big %d\n", priv_data_len); return -EOVERFLOW; } if (mpa_frame->rev != IETF_MPA_V1 && mpa_frame->rev != IETF_MPA_V2) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "unsupported mpa rev = %d\n", mpa_frame->rev); return -EINVAL; } if (mpa_frame->rev > cm_node->mpa_frame_rev) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "rev %d\n", mpa_frame->rev); return -EINVAL; } cm_node->mpa_frame_rev = mpa_frame->rev; if (cm_node->state != IRDMA_CM_STATE_MPAREQ_SENT) { if (memcmp(mpa_frame->key, IEFT_MPA_KEY_REQ, IETF_MPA_KEY_SIZE)) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Unexpected MPA Key received\n"); return -EINVAL; } } else { if (memcmp(mpa_frame->key, IEFT_MPA_KEY_REP, IETF_MPA_KEY_SIZE)) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Unexpected MPA Key received\n"); return -EINVAL; } } if (priv_data_len + mpa_hdr_len > len) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "ietf buffer len(%x + %x != %x)\n", priv_data_len, mpa_hdr_len, len); return -EOVERFLOW; } if (len > IRDMA_MAX_CM_BUF) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "ietf buffer large len = %d\n", len); return -EOVERFLOW; } switch (mpa_frame->rev) { case IETF_MPA_V2: mpa_hdr_len += IETF_RTR_MSG_SIZE; ret = irdma_negotiate_mpa_v2_ird_ord(cm_node, buf); if (ret) return ret; break; case IETF_MPA_V1: default: break; } memcpy(cm_node->pdata_buf, buf + mpa_hdr_len, priv_data_len); cm_node->pdata.size = priv_data_len; if (mpa_frame->flags & IETF_MPA_FLAGS_REJECT) *type = IRDMA_MPA_REQUEST_REJECT; if (mpa_frame->flags & IETF_MPA_FLAGS_MARKERS) cm_node->snd_mark_en = true; return 0; } /** * irdma_schedule_cm_timer * @cm_node: connection's node * @sqbuf: buffer to send * @type: if it is send or close * @send_retrans: if rexmits to be done * @close_when_complete: is cm_node to be removed * * note - cm_node needs to be protected before calling this. Encase in: * irdma_rem_ref_cm_node(cm_core, cm_node); * irdma_schedule_cm_timer(...) * atomic_inc(&cm_node->refcnt); */ int irdma_schedule_cm_timer(struct irdma_cm_node *cm_node, struct irdma_puda_buf *sqbuf, enum irdma_timer_type type, int send_retrans, int close_when_complete) { struct irdma_sc_vsi *vsi = &cm_node->iwdev->vsi; struct irdma_cm_core *cm_core = cm_node->cm_core; struct irdma_timer_entry *new_send; u32 was_timer_set; unsigned long flags; new_send = kzalloc(sizeof(*new_send), GFP_ATOMIC); if (!new_send) { if (type != IRDMA_TIMER_TYPE_CLOSE) irdma_free_sqbuf(vsi, sqbuf); return -ENOMEM; } new_send->retrycount = IRDMA_DEFAULT_RETRYS; new_send->retranscount = IRDMA_DEFAULT_RETRANS; new_send->sqbuf = sqbuf; new_send->timetosend = jiffies; new_send->type = type; new_send->send_retrans = send_retrans; new_send->close_when_complete = close_when_complete; if (type == IRDMA_TIMER_TYPE_CLOSE) { new_send->timetosend += (HZ / 10); if (cm_node->close_entry) { kfree(new_send); irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "already close entry\n"); return -EINVAL; } cm_node->close_entry = new_send; } else { /* type == IRDMA_TIMER_TYPE_SEND */ spin_lock_irqsave(&cm_node->retrans_list_lock, flags); cm_node->send_entry = new_send; atomic_inc(&cm_node->refcnt); spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); new_send->timetosend = jiffies + IRDMA_RETRY_TIMEOUT; atomic_inc(&sqbuf->refcount); irdma_puda_send_buf(vsi->ilq, sqbuf); if (!send_retrans) { irdma_cleanup_retrans_entry(cm_node); if (close_when_complete) irdma_rem_ref_cm_node(cm_node); return 0; } } spin_lock_irqsave(&cm_core->ht_lock, flags); was_timer_set = timer_pending(&cm_core->tcp_timer); if (!was_timer_set) { cm_core->tcp_timer.expires = new_send->timetosend; add_timer(&cm_core->tcp_timer); } spin_unlock_irqrestore(&cm_core->ht_lock, flags); return 0; } /** * irdma_retrans_expired - Could not rexmit the packet * @cm_node: connection's node */ static void irdma_retrans_expired(struct irdma_cm_node *cm_node) { enum irdma_cm_node_state state = cm_node->state; cm_node->state = IRDMA_CM_STATE_CLOSED; switch (state) { case IRDMA_CM_STATE_SYN_RCVD: case IRDMA_CM_STATE_CLOSING: irdma_rem_ref_cm_node(cm_node); break; case IRDMA_CM_STATE_FIN_WAIT1: case IRDMA_CM_STATE_LAST_ACK: irdma_send_reset(cm_node); break; default: atomic_inc(&cm_node->refcnt); irdma_send_reset(cm_node); irdma_create_event(cm_node, IRDMA_CM_EVENT_ABORTED); break; } } /** * irdma_handle_close_entry - for handling retry/timeouts * @cm_node: connection's node * @rem_node: flag for remove cm_node */ static void irdma_handle_close_entry(struct irdma_cm_node *cm_node, u32 rem_node) { struct irdma_timer_entry *close_entry = cm_node->close_entry; struct irdma_qp *iwqp; unsigned long flags; if (!close_entry) return; iwqp = (struct irdma_qp *)close_entry->sqbuf; if (iwqp) { spin_lock_irqsave(&iwqp->lock, flags); if (iwqp->cm_id) { iwqp->hw_tcp_state = IRDMA_TCP_STATE_CLOSED; iwqp->hw_iwarp_state = IRDMA_QP_STATE_ERROR; iwqp->last_aeq = IRDMA_AE_RESET_SENT; iwqp->ibqp_state = IB_QPS_ERR; spin_unlock_irqrestore(&iwqp->lock, flags); irdma_cm_disconn(iwqp); } else { spin_unlock_irqrestore(&iwqp->lock, flags); } } else if (rem_node) { /* TIME_WAIT state */ irdma_rem_ref_cm_node(cm_node); } kfree(close_entry); cm_node->close_entry = NULL; } /** * irdma_cm_timer_tick - system's timer expired callback * @t: Pointer to timer_list */ static void irdma_cm_timer_tick(struct timer_list *t) { unsigned long nexttimeout = jiffies + IRDMA_LONG_TIME; struct irdma_cm_node *cm_node; struct irdma_timer_entry *send_entry, *close_entry; struct list_head *list_core_temp; struct list_head *list_node; struct irdma_cm_core *cm_core = from_timer(cm_core, t, tcp_timer); struct irdma_sc_vsi *vsi; u32 settimer = 0; unsigned long timetosend; unsigned long flags; struct list_head timer_list; INIT_LIST_HEAD(&timer_list); rcu_read_lock(); irdma_timer_list_prep(cm_core, &timer_list); rcu_read_unlock(); list_for_each_safe(list_node, list_core_temp, &timer_list) { cm_node = container_of(list_node, struct irdma_cm_node, timer_entry); close_entry = cm_node->close_entry; if (close_entry) { if (time_after(close_entry->timetosend, jiffies)) { if (nexttimeout > close_entry->timetosend || !settimer) { nexttimeout = close_entry->timetosend; settimer = 1; } } else { irdma_handle_close_entry(cm_node, 1); } } spin_lock_irqsave(&cm_node->retrans_list_lock, flags); send_entry = cm_node->send_entry; if (!send_entry) goto done; if (time_after(send_entry->timetosend, jiffies)) { if (cm_node->state != IRDMA_CM_STATE_OFFLOADED) { if (nexttimeout > send_entry->timetosend || !settimer) { nexttimeout = send_entry->timetosend; settimer = 1; } } else { irdma_free_retrans_entry(cm_node); } goto done; } if (cm_node->state == IRDMA_CM_STATE_OFFLOADED || cm_node->state == IRDMA_CM_STATE_CLOSED) { irdma_free_retrans_entry(cm_node); goto done; } if (!send_entry->retranscount || !send_entry->retrycount) { irdma_free_retrans_entry(cm_node); spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); irdma_retrans_expired(cm_node); cm_node->state = IRDMA_CM_STATE_CLOSED; spin_lock_irqsave(&cm_node->retrans_list_lock, flags); goto done; } spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); vsi = &cm_node->iwdev->vsi; if (!cm_node->ack_rcvd) { atomic_inc(&send_entry->sqbuf->refcount); irdma_puda_send_buf(vsi->ilq, send_entry->sqbuf); cm_node->cm_core->stats_pkt_retrans++; } spin_lock_irqsave(&cm_node->retrans_list_lock, flags); if (send_entry->send_retrans) { send_entry->retranscount--; timetosend = (IRDMA_RETRY_TIMEOUT << (IRDMA_DEFAULT_RETRANS - send_entry->retranscount)); send_entry->timetosend = jiffies + min(timetosend, IRDMA_MAX_TIMEOUT); if (nexttimeout > send_entry->timetosend || !settimer) { nexttimeout = send_entry->timetosend; settimer = 1; } } else { int close_when_complete; close_when_complete = send_entry->close_when_complete; irdma_free_retrans_entry(cm_node); if (close_when_complete) irdma_rem_ref_cm_node(cm_node); } done: spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); irdma_rem_ref_cm_node(cm_node); } if (settimer) { spin_lock_irqsave(&cm_core->ht_lock, flags); if (!timer_pending(&cm_core->tcp_timer)) { cm_core->tcp_timer.expires = nexttimeout; add_timer(&cm_core->tcp_timer); } spin_unlock_irqrestore(&cm_core->ht_lock, flags); } } /** * irdma_send_syn - send SYN packet * @cm_node: connection's node * @sendack: flag to set ACK bit or not */ int irdma_send_syn(struct irdma_cm_node *cm_node, u32 sendack) { struct irdma_puda_buf *sqbuf; int flags = SET_SYN; char optionsbuf[sizeof(struct option_mss) + sizeof(struct option_windowscale) + sizeof(struct option_base) + TCP_OPTIONS_PADDING]; struct irdma_kmem_info opts; int optionssize = 0; /* Sending MSS option */ union all_known_options *options; opts.addr = optionsbuf; if (!cm_node) return -EINVAL; options = (union all_known_options *)&optionsbuf[optionssize]; options->mss.optionnum = OPTION_NUM_MSS; options->mss.len = sizeof(struct option_mss); options->mss.mss = htons(cm_node->tcp_cntxt.mss); optionssize += sizeof(struct option_mss); options = (union all_known_options *)&optionsbuf[optionssize]; options->windowscale.optionnum = OPTION_NUM_WINDOW_SCALE; options->windowscale.len = sizeof(struct option_windowscale); options->windowscale.shiftcount = cm_node->tcp_cntxt.rcv_wscale; optionssize += sizeof(struct option_windowscale); options = (union all_known_options *)&optionsbuf[optionssize]; options->eol = OPTION_NUM_EOL; optionssize += 1; if (sendack) flags |= SET_ACK; opts.size = optionssize; sqbuf = cm_node->cm_core->form_cm_frame(cm_node, &opts, NULL, NULL, flags); if (!sqbuf) return -ENOMEM; return irdma_schedule_cm_timer(cm_node, sqbuf, IRDMA_TIMER_TYPE_SEND, 1, 0); } /** * irdma_send_ack - Send ACK packet * @cm_node: connection's node */ void irdma_send_ack(struct irdma_cm_node *cm_node) { struct irdma_puda_buf *sqbuf; struct irdma_sc_vsi *vsi = &cm_node->iwdev->vsi; sqbuf = cm_node->cm_core->form_cm_frame(cm_node, NULL, NULL, NULL, SET_ACK); if (sqbuf) irdma_puda_send_buf(vsi->ilq, sqbuf); } /** * irdma_send_fin - Send FIN pkt * @cm_node: connection's node */ static int irdma_send_fin(struct irdma_cm_node *cm_node) { struct irdma_puda_buf *sqbuf; sqbuf = cm_node->cm_core->form_cm_frame(cm_node, NULL, NULL, NULL, SET_ACK | SET_FIN); if (!sqbuf) return -ENOMEM; return irdma_schedule_cm_timer(cm_node, sqbuf, IRDMA_TIMER_TYPE_SEND, 1, 0); } /** * irdma_find_listener - find a cm node listening on this addr-port pair * @cm_core: cm's core * @dst_addr: listener ip addr * @ipv4: flag indicating IPv4 when true * @dst_port: listener tcp port num * @vlan_id: virtual LAN ID * @listener_state: state to match with listen node's */ static struct irdma_cm_listener * irdma_find_listener(struct irdma_cm_core *cm_core, u32 *dst_addr, bool ipv4, u16 dst_port, u16 vlan_id, enum irdma_cm_listener_state listener_state) { struct irdma_cm_listener *listen_node; static const u32 ip_zero[4] = {0, 0, 0, 0}; u32 listen_addr[4]; u16 listen_port; unsigned long flags; /* walk list and find cm_node associated with this session ID */ spin_lock_irqsave(&cm_core->listen_list_lock, flags); list_for_each_entry(listen_node, &cm_core->listen_list, list) { memcpy(listen_addr, listen_node->loc_addr, sizeof(listen_addr)); listen_port = listen_node->loc_port; if (listen_node->ipv4 != ipv4 || listen_port != dst_port || !(listener_state & listen_node->listener_state)) continue; /* compare node pair, return node handle if a match */ if (!memcmp(listen_addr, ip_zero, sizeof(listen_addr)) || (!memcmp(listen_addr, dst_addr, sizeof(listen_addr)) && vlan_id == listen_node->vlan_id)) { atomic_inc(&listen_node->refcnt); spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); return listen_node; } } spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); return NULL; } /** * irdma_del_multiple_qhash - Remove qhash and child listens * @iwdev: iWarp device * @cm_info: CM info for parent listen node * @cm_parent_listen_node: The parent listen node */ static int irdma_del_multiple_qhash(struct irdma_device *iwdev, struct irdma_cm_info *cm_info, struct irdma_cm_listener *cm_parent_listen_node) { struct irdma_cm_listener *child_listen_node; struct list_head *pos, *tpos; unsigned long flags; int ret = -EINVAL; spin_lock_irqsave(&iwdev->cm_core.listen_list_lock, flags); list_for_each_safe(pos, tpos, &cm_parent_listen_node->child_listen_list) { child_listen_node = list_entry(pos, struct irdma_cm_listener, child_listen_list); if (child_listen_node->ipv4) irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "removing child listen for IP=%x, port=%d, vlan=%d\n", child_listen_node->loc_addr[0], child_listen_node->loc_port, child_listen_node->vlan_id); else irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "removing child listen for IP=%x:%x:%x:%x, port=%d, vlan=%d\n", IRDMA_PRINT_IP6(child_listen_node->loc_addr), child_listen_node->loc_port, child_listen_node->vlan_id); list_del(pos); memcpy(cm_info->loc_addr, child_listen_node->loc_addr, sizeof(cm_info->loc_addr)); cm_info->vlan_id = child_listen_node->vlan_id; if (child_listen_node->qhash_set) { ret = irdma_manage_qhash(iwdev, cm_info, IRDMA_QHASH_TYPE_TCP_SYN, IRDMA_QHASH_MANAGE_TYPE_DELETE, NULL, false); child_listen_node->qhash_set = false; } else { ret = 0; } irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Child listen node freed = %p\n", child_listen_node); kfree(child_listen_node); cm_parent_listen_node->cm_core->stats_listen_nodes_destroyed++; } spin_unlock_irqrestore(&iwdev->cm_core.listen_list_lock, flags); return ret; } static u8 irdma_iw_get_vlan_prio(u32 *loc_addr, u8 prio, bool ipv4) { return prio; } /** * irdma_netdev_vlan_ipv6 - Gets the netdev and mac * @addr: local IPv6 address * @vlan_id: vlan id for the given IPv6 address * @mac: mac address for the given IPv6 address * * Returns the net_device of the IPv6 address and also sets the * vlan id and mac for that address. */ if_t -irdma_netdev_vlan_ipv6(u32 *addr, u16 *vlan_id, u8 *mac) +irdma_netdev_vlan_ipv6(struct iw_cm_id *cm_id, u32 *addr, u16 *vlan_id, u8 *mac) { if_t ip_dev = NULL; struct in6_addr laddr6; +#ifdef VIMAGE + struct vnet *vnet = irdma_cmid_to_vnet(cm_id); +#endif struct ifaddr *ifa; u16 scope_id = 0; irdma_copy_ip_htonl(laddr6.__u6_addr.__u6_addr32, addr); if (vlan_id) *vlan_id = 0xFFFF; /* Match rdma_vlan_dev_vlan_id() */ if (mac) eth_zero_addr(mac); if (IN6_IS_SCOPE_LINKLOCAL(&laddr6) || IN6_IS_ADDR_MC_INTFACELOCAL(&laddr6)) scope_id = ntohs(laddr6.__u6_addr.__u6_addr16[1]); +#ifdef VIMAGE + ip_dev = ip6_ifp_find(vnet, laddr6, scope_id); +#else ip_dev = ip6_ifp_find(&init_net, laddr6, scope_id); +#endif if (ip_dev) { if (vlan_id) *vlan_id = rdma_vlan_dev_vlan_id(ip_dev); ifa = if_getifaddr(ip_dev); if (ifa && ifa->ifa_addr && mac) ether_addr_copy(mac, if_getlladdr(ip_dev)); } return ip_dev; } /** * irdma_get_vlan_ipv4 - Returns the vlan_id for IPv4 address * @addr: local IPv4 address */ u16 -irdma_get_vlan_ipv4(u32 *addr) +irdma_get_vlan_ipv4(struct iw_cm_id *cm_id, u32 *addr) { if_t netdev; +#ifdef VIMAGE + struct vnet *vnet = irdma_cmid_to_vnet(cm_id); +#endif u16 vlan_id = 0xFFFF; +#ifdef VIMAGE + netdev = ip_ifp_find(vnet, htonl(addr[0])); +#else netdev = ip_ifp_find(&init_net, htonl(addr[0])); +#endif if (netdev) { vlan_id = rdma_vlan_dev_vlan_id(netdev); dev_put(netdev); } return vlan_id; } static int irdma_manage_qhash_wait(struct irdma_pci_f *rf, struct irdma_cm_info *cm_info) { struct irdma_cqp_request *cqp_request = cm_info->cqp_request; int cnt = rf->sc_dev.hw_attrs.max_cqp_compl_wait_time_ms * CQP_TIMEOUT_THRESHOLD; u32 ret_val; if (!cqp_request) return -ENOMEM; do { irdma_cqp_ce_handler(rf, &rf->ccq.sc_cq); mdelay(1); } while (!READ_ONCE(cqp_request->request_done) && --cnt); ret_val = cqp_request->compl_info.op_ret_val; irdma_put_cqp_request(&rf->cqp, cqp_request); if (cnt) { if (!ret_val) return 0; return -EINVAL; } return -ETIMEDOUT; } /** * irdma_add_mqh_ifa_cb - Adds multiple qhashes for IPv4/IPv6 * @arg: Calback argument structure from irdma_add_mqh * @ifa: Current address to compute against * @count: Current cumulative output of all callbacks in this iteration * * Adds a qhash and a child listen node for a single IPv4/IPv6 address * on the adapter and adds the associated qhash filter */ static u_int irdma_add_mqh_ifa_cb(void *arg, struct ifaddr *ifa, u_int count){ struct irdma_add_mqh_cbs *cbs = arg; struct irdma_cm_listener *child_listen_node; struct irdma_cm_info *cm_info = cbs->cm_info; struct irdma_device *iwdev = cbs->iwdev; struct irdma_cm_listener *cm_parent_listen_node = cbs->cm_listen_node; if_t ip_dev = ifa->ifa_ifp; unsigned long flags; int ret; if (count) return 0; child_listen_node = kzalloc(sizeof(*child_listen_node), GFP_ATOMIC); if (!child_listen_node) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "listener memory allocation\n"); return -ENOMEM; } memcpy(child_listen_node, cm_parent_listen_node, sizeof(*child_listen_node)); cm_info->vlan_id = rdma_vlan_dev_vlan_id(ip_dev); child_listen_node->vlan_id = cm_info->vlan_id; if (cm_info->ipv4) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Allocating child CM Listener forIP=%x, vlan_id=%d, MAC=%x:%x:%x:%x:%x:%x\n", ((struct sockaddr_in *)&ifa->ifa_addr)->sin_addr.s_addr, rdma_vlan_dev_vlan_id(ip_dev), if_getlladdr(ip_dev)[0], if_getlladdr(ip_dev)[1], if_getlladdr(ip_dev)[2], if_getlladdr(ip_dev)[3], if_getlladdr(ip_dev)[4], if_getlladdr(ip_dev)[5]); child_listen_node->loc_addr[0] = ntohl(((struct sockaddr_in *)ifa->ifa_addr)->sin_addr.s_addr); } else { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "IP=%x:%x:%x:%x, vlan_id=%d, MAC=%x:%x:%x:%x:%x:%x\n", IRDMA_PRINT_IP6(&((struct sockaddr_in6 *)ifa->ifa_addr)->sin6_addr), rdma_vlan_dev_vlan_id(ip_dev), if_getlladdr(ip_dev)[0], if_getlladdr(ip_dev)[1], if_getlladdr(ip_dev)[2], if_getlladdr(ip_dev)[3], if_getlladdr(ip_dev)[4], if_getlladdr(ip_dev)[5]); irdma_copy_ip_ntohl(child_listen_node->loc_addr, ((struct sockaddr_in6 *)ifa->ifa_addr)->sin6_addr.__u6_addr.__u6_addr32); } memcpy(cm_info->loc_addr, child_listen_node->loc_addr, sizeof(cm_info->loc_addr)); if (!iwdev->vsi.dscp_mode) cm_info->user_pri = irdma_iw_get_vlan_prio(child_listen_node->loc_addr, cm_info->user_pri, cm_info->ipv4); ret = irdma_manage_qhash(iwdev, cm_info, IRDMA_QHASH_TYPE_TCP_SYN, IRDMA_QHASH_MANAGE_TYPE_ADD, NULL, false); if (ret) { kfree(child_listen_node); return ret; } /* wait for qhash finish */ ret = irdma_manage_qhash_wait(iwdev->rf, cm_info); if (ret) { kfree(child_listen_node); return ret; } child_listen_node->qhash_set = true; spin_lock_irqsave(&iwdev->cm_core.listen_list_lock, flags); list_add(&child_listen_node->child_listen_list, &cm_parent_listen_node->child_listen_list); spin_unlock_irqrestore(&iwdev->cm_core.listen_list_lock, flags); cm_parent_listen_node->cm_core->stats_listen_nodes_created++; return 0; } /** * irdma_add_mqh - Adds multiple qhashes * @iwdev: iWarp device * @cm_info: CM info for parent listen node * @cm_listen_node: The parent listen node */ static int irdma_add_mqh(struct irdma_device *iwdev, struct irdma_cm_info *cm_info, struct irdma_cm_listener *cm_listen_node) { struct epoch_tracker et; struct irdma_add_mqh_cbs cbs; struct if_iter iter; if_t ifp; int err = -ENOENT; cbs.iwdev = iwdev; cbs.cm_info = cm_info; cbs.cm_listen_node = cm_listen_node; VNET_ITERATOR_DECL(vnet_iter); VNET_LIST_RLOCK(); NET_EPOCH_ENTER(et); VNET_FOREACH(vnet_iter) { CURVNET_SET_QUIET(vnet_iter); for (ifp = if_iter_start(&iter); ifp != NULL; ifp = if_iter_next(&iter)) { if (!(if_getflags(ifp) & IFF_UP)) continue; if (((rdma_vlan_dev_vlan_id(ifp) >= VLAN_N_VID) || (rdma_vlan_dev_real_dev(ifp) != iwdev->netdev)) && ifp != iwdev->netdev) continue; if_addr_rlock(ifp); if (cm_info->ipv4) err = if_foreach_addr_type(ifp, AF_INET, irdma_add_mqh_ifa_cb, &cbs); else err = if_foreach_addr_type(ifp, AF_INET6, irdma_add_mqh_ifa_cb, &cbs); if_addr_runlock(ifp); } if_iter_finish(&iter); CURVNET_RESTORE(); } NET_EPOCH_EXIT(et); VNET_LIST_RUNLOCK(); return err; } /** * irdma_reset_list_prep - add connection nodes slated for reset to list * @cm_core: cm's core * @listener: pointer to listener node * @reset_list: a list to which cm_node will be selected */ static void irdma_reset_list_prep(struct irdma_cm_core *cm_core, struct irdma_cm_listener *listener, struct list_head *reset_list) { struct irdma_cm_node *cm_node; int bkt; HASH_FOR_EACH_RCU(cm_core->cm_hash_tbl, bkt, cm_node, list) { if (cm_node->listener == listener && !cm_node->accelerated && atomic_inc_not_zero(&cm_node->refcnt)) list_add(&cm_node->reset_entry, reset_list); } } /** * irdma_dec_refcnt_listen - delete listener and associated cm nodes * @cm_core: cm's core * @listener: pointer to listener node * @free_hanging_nodes: to free associated cm_nodes * @apbvt_del: flag to delete the apbvt */ static int irdma_dec_refcnt_listen(struct irdma_cm_core *cm_core, struct irdma_cm_listener *listener, int free_hanging_nodes, bool apbvt_del) { struct list_head *list_pos; struct list_head *list_temp; struct irdma_cm_node *cm_node; struct list_head reset_list; struct irdma_cm_info nfo; enum irdma_cm_node_state old_state; unsigned long flags; int err; /* free non-accelerated child nodes for this listener */ INIT_LIST_HEAD(&reset_list); if (free_hanging_nodes) { rcu_read_lock(); irdma_reset_list_prep(cm_core, listener, &reset_list); rcu_read_unlock(); } list_for_each_safe(list_pos, list_temp, &reset_list) { cm_node = container_of(list_pos, struct irdma_cm_node, reset_entry); if (cm_node->state >= IRDMA_CM_STATE_FIN_WAIT1) { irdma_rem_ref_cm_node(cm_node); continue; } irdma_cleanup_retrans_entry(cm_node); err = irdma_send_reset(cm_node); if (err) { cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "send reset failed\n"); } else { old_state = cm_node->state; cm_node->state = IRDMA_CM_STATE_LISTENER_DESTROYED; if (old_state != IRDMA_CM_STATE_MPAREQ_RCVD) irdma_rem_ref_cm_node(cm_node); } } if (atomic_dec_and_test(&listener->refcnt)) { spin_lock_irqsave(&cm_core->listen_list_lock, flags); list_del(&listener->list); spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); if (apbvt_del) irdma_del_apbvt(listener->iwdev, listener->apbvt_entry); memcpy(nfo.loc_addr, listener->loc_addr, sizeof(nfo.loc_addr)); nfo.loc_port = listener->loc_port; nfo.ipv4 = listener->ipv4; nfo.vlan_id = listener->vlan_id; nfo.user_pri = listener->user_pri; nfo.qh_qpid = listener->iwdev->vsi.ilq->qp_id; if (!list_empty(&listener->child_listen_list)) { irdma_del_multiple_qhash(listener->iwdev, &nfo, listener); } else { if (listener->qhash_set) irdma_manage_qhash(listener->iwdev, &nfo, IRDMA_QHASH_TYPE_TCP_SYN, IRDMA_QHASH_MANAGE_TYPE_DELETE, NULL, false); } cm_core->stats_listen_destroyed++; cm_core->stats_listen_nodes_destroyed++; irdma_debug(&listener->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "loc_port=0x%04x loc_addr=%x cm_listen_node=%p cm_id=%p qhash_set=%d vlan_id=%d apbvt_del=%d\n", listener->loc_port, listener->loc_addr[0], listener, listener->cm_id, listener->qhash_set, listener->vlan_id, apbvt_del); kfree(listener); listener = NULL; return 0; } return -EINVAL; } /** * irdma_cm_del_listen - delete a listener * @cm_core: cm's core * @listener: passive connection's listener * @apbvt_del: flag to delete apbvt */ static int irdma_cm_del_listen(struct irdma_cm_core *cm_core, struct irdma_cm_listener *listener, bool apbvt_del) { listener->listener_state = IRDMA_CM_LISTENER_PASSIVE_STATE; listener->cm_id = NULL; return irdma_dec_refcnt_listen(cm_core, listener, 1, apbvt_del); } /** * irdma_find_node - find a cm node that matches the reference cm node * @cm_core: cm's core * @rem_port: remote tcp port num * @rem_addr: remote ip addr * @loc_port: local tcp port num * @loc_addr: local ip addr * @vlan_id: local VLAN ID */ struct irdma_cm_node * irdma_find_node(struct irdma_cm_core *cm_core, u16 rem_port, u32 *rem_addr, u16 loc_port, u32 *loc_addr, u16 vlan_id) { struct irdma_cm_node *cm_node; u32 key = (rem_port << 16) | loc_port; rcu_read_lock(); HASH_FOR_EACH_POSSIBLE_RCU(cm_core->cm_hash_tbl, cm_node, list, key) { if (cm_node->vlan_id == vlan_id && cm_node->loc_port == loc_port && cm_node->rem_port == rem_port && !memcmp(cm_node->loc_addr, loc_addr, sizeof(cm_node->loc_addr)) && !memcmp(cm_node->rem_addr, rem_addr, sizeof(cm_node->rem_addr))) { if (!atomic_inc_not_zero(&cm_node->refcnt)) goto exit; rcu_read_unlock(); return cm_node; } } exit: rcu_read_unlock(); /* no owner node */ return NULL; } /** * irdma_add_hte_node - add a cm node to the hash table * @cm_core: cm's core * @cm_node: connection's node */ static void irdma_add_hte_node(struct irdma_cm_core *cm_core, struct irdma_cm_node *cm_node) { unsigned long flags; u32 key = (cm_node->rem_port << 16) | cm_node->loc_port; spin_lock_irqsave(&cm_core->ht_lock, flags); HASH_ADD_RCU(cm_core->cm_hash_tbl, &cm_node->list, key); spin_unlock_irqrestore(&cm_core->ht_lock, flags); } /** * irdma_ipv4_is_lpb - check if loopback * @loc_addr: local addr to compare * @rem_addr: remote address */ bool irdma_ipv4_is_lpb(u32 loc_addr, u32 rem_addr) { return ipv4_is_loopback(htonl(rem_addr)) || (loc_addr == rem_addr); } /** * irdma_ipv6_is_lpb - check if loopback * @loc_addr: local addr to compare * @rem_addr: remote address */ bool irdma_ipv6_is_lpb(u32 *loc_addr, u32 *rem_addr) { struct in6_addr raddr6; irdma_copy_ip_htonl(raddr6.__u6_addr.__u6_addr32, rem_addr); return !memcmp(loc_addr, rem_addr, 16) || ipv6_addr_loopback(&raddr6); } /** * irdma_cm_create_ah - create a cm address handle * @cm_node: The connection manager node to create AH for * @wait: Provides option to wait for ah creation or not */ static int irdma_cm_create_ah(struct irdma_cm_node *cm_node, bool wait) { struct irdma_ah_info ah_info = {0}; struct irdma_device *iwdev = cm_node->iwdev; #ifdef VIMAGE - struct rdma_cm_id *rdma_id = (struct rdma_cm_id *)cm_node->cm_id->context; - struct vnet *vnet = rdma_id->route.addr.dev_addr.net; + struct vnet *vnet = irdma_cmid_to_vnet(cm_node->cm_id); #endif ether_addr_copy(ah_info.mac_addr, if_getlladdr(iwdev->netdev)); ah_info.hop_ttl = 0x40; ah_info.tc_tos = cm_node->tos; ah_info.vsi = &iwdev->vsi; if (cm_node->ipv4) { ah_info.ipv4_valid = true; ah_info.dest_ip_addr[0] = cm_node->rem_addr[0]; ah_info.src_ip_addr[0] = cm_node->loc_addr[0]; CURVNET_SET_QUIET(vnet); ah_info.do_lpbk = irdma_ipv4_is_lpb(ah_info.src_ip_addr[0], ah_info.dest_ip_addr[0]); CURVNET_RESTORE(); } else { memcpy(ah_info.dest_ip_addr, cm_node->rem_addr, sizeof(ah_info.dest_ip_addr)); memcpy(ah_info.src_ip_addr, cm_node->loc_addr, sizeof(ah_info.src_ip_addr)); ah_info.do_lpbk = irdma_ipv6_is_lpb(ah_info.src_ip_addr, ah_info.dest_ip_addr); } ah_info.vlan_tag = cm_node->vlan_id; if (cm_node->vlan_id < VLAN_N_VID) { ah_info.insert_vlan_tag = 1; ah_info.vlan_tag |= cm_node->user_pri << VLAN_PRIO_SHIFT; } ah_info.dst_arpindex = irdma_arp_table(iwdev->rf, ah_info.dest_ip_addr, NULL, IRDMA_ARP_RESOLVE); if (irdma_puda_create_ah(&iwdev->rf->sc_dev, &ah_info, wait, IRDMA_PUDA_RSRC_TYPE_ILQ, cm_node, &cm_node->ah)) return -ENOMEM; return 0; } /** * irdma_cm_free_ah - free a cm address handle * @cm_node: The connection manager node to create AH for */ static void irdma_cm_free_ah(struct irdma_cm_node *cm_node) { struct irdma_device *iwdev = cm_node->iwdev; irdma_puda_free_ah(&iwdev->rf->sc_dev, cm_node->ah); cm_node->ah = NULL; } /** * irdma_make_cm_node - create a new instance of a cm node * @cm_core: cm's core * @iwdev: iwarp device structure * @cm_info: quad info for connection * @listener: passive connection's listener */ static struct irdma_cm_node * irdma_make_cm_node(struct irdma_cm_core *cm_core, struct irdma_device *iwdev, struct irdma_cm_info *cm_info, struct irdma_cm_listener *listener) { struct irdma_cm_node *cm_node; int arpindex; if_t netdev = iwdev->netdev; /* create an hte and cm_node for this instance */ cm_node = kzalloc(sizeof(*cm_node), GFP_ATOMIC); if (!cm_node) return NULL; /* set our node specific transport info */ cm_node->ipv4 = cm_info->ipv4; cm_node->vlan_id = cm_info->vlan_id; if (cm_node->vlan_id >= VLAN_N_VID && iwdev->dcb_vlan_mode) cm_node->vlan_id = 0; cm_node->tos = cm_info->tos; cm_node->user_pri = cm_info->user_pri; if (listener) { if (listener->tos != cm_info->tos) irdma_dev_warn(&iwdev->ibdev, "application TOS[%d] and remote client TOS[%d] mismatch\n", listener->tos, cm_info->tos); if (iwdev->vsi.dscp_mode) { cm_node->user_pri = listener->user_pri; } else { cm_node->tos = max(listener->tos, cm_info->tos); cm_node->user_pri = rt_tos2priority(cm_node->tos); cm_node->user_pri = irdma_iw_get_vlan_prio(cm_info->loc_addr, cm_node->user_pri, cm_info->ipv4); } irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_DCB, "listener: TOS:[%d] UP:[%d]\n", cm_node->tos, cm_node->user_pri); } memcpy(cm_node->loc_addr, cm_info->loc_addr, sizeof(cm_node->loc_addr)); memcpy(cm_node->rem_addr, cm_info->rem_addr, sizeof(cm_node->rem_addr)); cm_node->loc_port = cm_info->loc_port; cm_node->rem_port = cm_info->rem_port; cm_node->mpa_frame_rev = IRDMA_CM_DEFAULT_MPA_VER; cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO; cm_node->iwdev = iwdev; cm_node->dev = &iwdev->rf->sc_dev; cm_node->ird_size = cm_node->dev->hw_attrs.max_hw_ird; cm_node->ord_size = cm_node->dev->hw_attrs.max_hw_ord; cm_node->listener = listener; cm_node->cm_id = cm_info->cm_id; ether_addr_copy(cm_node->loc_mac, if_getlladdr(netdev)); spin_lock_init(&cm_node->retrans_list_lock); cm_node->ack_rcvd = false; init_completion(&cm_node->establish_comp); atomic_set(&cm_node->refcnt, 1); /* associate our parent CM core */ cm_node->cm_core = cm_core; cm_node->tcp_cntxt.loc_id = IRDMA_CM_DEFAULT_LOCAL_ID; cm_node->tcp_cntxt.rcv_wscale = iwdev->rcv_wscale; cm_node->tcp_cntxt.rcv_wnd = iwdev->rcv_wnd >> cm_node->tcp_cntxt.rcv_wscale; kc_set_loc_seq_num_mss(cm_node); arpindex = irdma_resolve_neigh_lpb_chk(iwdev, cm_node, cm_info); if (arpindex < 0) goto err; ether_addr_copy(cm_node->rem_mac, iwdev->rf->arp_table[arpindex].mac_addr); irdma_add_hte_node(cm_core, cm_node); cm_core->stats_nodes_created++; return cm_node; err: kfree(cm_node); return NULL; } static void irdma_destroy_connection(struct irdma_cm_node *cm_node) { struct irdma_cm_core *cm_core = cm_node->cm_core; struct irdma_qp *iwqp; struct irdma_cm_info nfo; /* if the node is destroyed before connection was accelerated */ if (!cm_node->accelerated && cm_node->accept_pend) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "node destroyed before established\n"); atomic_dec(&cm_node->listener->pend_accepts_cnt); } if (cm_node->close_entry) irdma_handle_close_entry(cm_node, 0); if (cm_node->listener) { irdma_dec_refcnt_listen(cm_core, cm_node->listener, 0, true); } else { if (cm_node->apbvt_set) { irdma_del_apbvt(cm_node->iwdev, cm_node->apbvt_entry); cm_node->apbvt_set = 0; } irdma_get_addr_info(cm_node, &nfo); if (cm_node->qhash_set) { nfo.qh_qpid = cm_node->iwdev->vsi.ilq->qp_id; irdma_manage_qhash(cm_node->iwdev, &nfo, IRDMA_QHASH_TYPE_TCP_ESTABLISHED, IRDMA_QHASH_MANAGE_TYPE_DELETE, NULL, false); cm_node->qhash_set = 0; } } iwqp = cm_node->iwqp; if (iwqp) { cm_node->cm_id->rem_ref(cm_node->cm_id); cm_node->cm_id = NULL; iwqp->cm_id = NULL; irdma_qp_rem_ref(&iwqp->ibqp); cm_node->iwqp = NULL; } else if (cm_node->qhash_set) { irdma_get_addr_info(cm_node, &nfo); nfo.qh_qpid = cm_node->iwdev->vsi.ilq->qp_id; irdma_manage_qhash(cm_node->iwdev, &nfo, IRDMA_QHASH_TYPE_TCP_ESTABLISHED, IRDMA_QHASH_MANAGE_TYPE_DELETE, NULL, false); cm_node->qhash_set = 0; } cm_core->cm_free_ah(cm_node); } /** * irdma_rem_ref_cm_node - destroy an instance of a cm node * @cm_node: connection's node */ void irdma_rem_ref_cm_node(struct irdma_cm_node *cm_node) { struct irdma_cm_core *cm_core = cm_node->cm_core; unsigned long flags; spin_lock_irqsave(&cm_core->ht_lock, flags); if (!atomic_dec_and_test(&cm_node->refcnt)) { spin_unlock_irqrestore(&cm_core->ht_lock, flags); return; } if (cm_node->iwqp) { cm_node->iwqp->cm_node = NULL; cm_node->iwqp->cm_id = NULL; } HASH_DEL_RCU(cm_core->cm_hash_tbl, &cm_node->list); cm_node->cm_core->stats_nodes_destroyed++; spin_unlock_irqrestore(&cm_core->ht_lock, flags); irdma_destroy_connection(cm_node); kfree_rcu(cm_node, rcu_head); } /** * irdma_handle_fin_pkt - FIN packet received * @cm_node: connection's node */ static void irdma_handle_fin_pkt(struct irdma_cm_node *cm_node) { switch (cm_node->state) { case IRDMA_CM_STATE_SYN_RCVD: case IRDMA_CM_STATE_SYN_SENT: case IRDMA_CM_STATE_ESTABLISHED: case IRDMA_CM_STATE_MPAREJ_RCVD: cm_node->tcp_cntxt.rcv_nxt++; irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_LAST_ACK; irdma_send_fin(cm_node); break; case IRDMA_CM_STATE_MPAREQ_SENT: irdma_create_event(cm_node, IRDMA_CM_EVENT_ABORTED); cm_node->tcp_cntxt.rcv_nxt++; irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_CLOSED; atomic_inc(&cm_node->refcnt); irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_FIN_WAIT1: cm_node->tcp_cntxt.rcv_nxt++; irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_CLOSING; irdma_send_ack(cm_node); /* * Wait for ACK as this is simultaneous close. After we receive ACK, do not send anything. Just rm the * node. */ break; case IRDMA_CM_STATE_FIN_WAIT2: cm_node->tcp_cntxt.rcv_nxt++; irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_TIME_WAIT; irdma_send_ack(cm_node); irdma_schedule_cm_timer(cm_node, NULL, IRDMA_TIMER_TYPE_CLOSE, 1, 0); break; case IRDMA_CM_STATE_TIME_WAIT: cm_node->tcp_cntxt.rcv_nxt++; irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_rem_ref_cm_node(cm_node); break; case IRDMA_CM_STATE_OFFLOADED: default: irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "bad state node state = %d\n", cm_node->state); break; } } /** * irdma_handle_rst_pkt - process received RST packet * @cm_node: connection's node * @rbuf: receive buffer */ static void irdma_handle_rst_pkt(struct irdma_cm_node *cm_node, struct irdma_puda_buf *rbuf) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "caller: %pS cm_node=%p state=%d rem_port=0x%04x loc_port=0x%04x rem_addr=%x loc_addr=%x\n", __builtin_return_address(0), cm_node, cm_node->state, cm_node->rem_port, cm_node->loc_port, cm_node->rem_addr[0], cm_node->loc_addr[0]); irdma_cleanup_retrans_entry(cm_node); switch (cm_node->state) { case IRDMA_CM_STATE_SYN_SENT: case IRDMA_CM_STATE_MPAREQ_SENT: switch (cm_node->mpa_frame_rev) { case IETF_MPA_V2: /* Drop down to MPA_V1 */ cm_node->mpa_frame_rev = IETF_MPA_V1; /* send a syn and goto syn sent state */ cm_node->state = IRDMA_CM_STATE_SYN_SENT; if (irdma_send_syn(cm_node, 0)) irdma_active_open_err(cm_node, false); break; case IETF_MPA_V1: default: irdma_active_open_err(cm_node, false); break; } break; case IRDMA_CM_STATE_MPAREQ_RCVD: atomic_inc(&cm_node->passive_state); break; case IRDMA_CM_STATE_ESTABLISHED: case IRDMA_CM_STATE_SYN_RCVD: case IRDMA_CM_STATE_LISTENING: irdma_passive_open_err(cm_node, false); break; case IRDMA_CM_STATE_OFFLOADED: irdma_active_open_err(cm_node, false); break; case IRDMA_CM_STATE_CLOSED: break; case IRDMA_CM_STATE_FIN_WAIT2: case IRDMA_CM_STATE_FIN_WAIT1: case IRDMA_CM_STATE_LAST_ACK: case IRDMA_CM_STATE_TIME_WAIT: cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_rem_ref_cm_node(cm_node); break; default: break; } } /** * irdma_handle_rcv_mpa - Process a recv'd mpa buffer * @cm_node: connection's node * @rbuf: receive buffer */ static void irdma_handle_rcv_mpa(struct irdma_cm_node *cm_node, struct irdma_puda_buf *rbuf) { int err; int datasize = rbuf->datalen; u8 *dataloc = rbuf->data; enum irdma_cm_event_type type = IRDMA_CM_EVENT_UNKNOWN; u32 res_type; err = irdma_parse_mpa(cm_node, dataloc, &res_type, datasize); if (err) { if (cm_node->state == IRDMA_CM_STATE_MPAREQ_SENT) irdma_active_open_err(cm_node, true); else irdma_passive_open_err(cm_node, true); return; } switch (cm_node->state) { case IRDMA_CM_STATE_ESTABLISHED: if (res_type == IRDMA_MPA_REQUEST_REJECT) irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "state for reject\n"); cm_node->state = IRDMA_CM_STATE_MPAREQ_RCVD; type = IRDMA_CM_EVENT_MPA_REQ; irdma_send_ack(cm_node); /* ACK received MPA request */ atomic_set(&cm_node->passive_state, IRDMA_PASSIVE_STATE_INDICATED); break; case IRDMA_CM_STATE_MPAREQ_SENT: irdma_cleanup_retrans_entry(cm_node); if (res_type == IRDMA_MPA_REQUEST_REJECT) { type = IRDMA_CM_EVENT_MPA_REJECT; cm_node->state = IRDMA_CM_STATE_MPAREJ_RCVD; } else { type = IRDMA_CM_EVENT_CONNECTED; cm_node->state = IRDMA_CM_STATE_OFFLOADED; } irdma_send_ack(cm_node); break; default: irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "wrong cm_node state=%d\n", cm_node->state); break; } irdma_create_event(cm_node, type); } /** * irdma_check_syn - Check for error on received syn ack * @cm_node: connection's node * @tcph: pointer tcp header */ static int irdma_check_syn(struct irdma_cm_node *cm_node, struct tcphdr *tcph) { if (ntohl(tcph->th_ack) != cm_node->tcp_cntxt.loc_seq_num) { irdma_active_open_err(cm_node, true); return 1; } return 0; } /** * irdma_check_seq - check seq numbers if OK * @cm_node: connection's node * @tcph: pointer tcp header */ static int irdma_check_seq(struct irdma_cm_node *cm_node, struct tcphdr *tcph) { u32 seq; u32 ack_seq; u32 loc_seq_num = cm_node->tcp_cntxt.loc_seq_num; u32 rcv_nxt = cm_node->tcp_cntxt.rcv_nxt; u32 rcv_wnd; int err = 0; seq = ntohl(tcph->th_seq); ack_seq = ntohl(tcph->th_ack); rcv_wnd = cm_node->tcp_cntxt.rcv_wnd; if (ack_seq != loc_seq_num || !between(seq, rcv_nxt, (rcv_nxt + rcv_wnd))) err = -1; if (err) irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "seq number err\n"); return err; } void irdma_add_conn_est_qh(struct irdma_cm_node *cm_node) { struct irdma_cm_info nfo; irdma_get_addr_info(cm_node, &nfo); nfo.qh_qpid = cm_node->iwdev->vsi.ilq->qp_id; irdma_manage_qhash(cm_node->iwdev, &nfo, IRDMA_QHASH_TYPE_TCP_ESTABLISHED, IRDMA_QHASH_MANAGE_TYPE_ADD, cm_node, false); cm_node->qhash_set = true; } /** * irdma_handle_syn_pkt - is for Passive node * @cm_node: connection's node * @rbuf: receive buffer */ static void irdma_handle_syn_pkt(struct irdma_cm_node *cm_node, struct irdma_puda_buf *rbuf) { struct tcphdr *tcph = (struct tcphdr *)rbuf->tcph; int err; u32 inc_sequence; int optionsize; optionsize = (tcph->th_off << 2) - sizeof(struct tcphdr); inc_sequence = ntohl(tcph->th_seq); switch (cm_node->state) { case IRDMA_CM_STATE_SYN_SENT: case IRDMA_CM_STATE_MPAREQ_SENT: /* Rcvd syn on active open connection */ irdma_active_open_err(cm_node, 1); break; case IRDMA_CM_STATE_LISTENING: /* Passive OPEN */ if (atomic_read(&cm_node->listener->pend_accepts_cnt) > cm_node->listener->backlog) { cm_node->cm_core->stats_backlog_drops++; irdma_passive_open_err(cm_node, false); break; } err = irdma_handle_tcp_options(cm_node, tcph, optionsize, 1); if (err) { irdma_passive_open_err(cm_node, false); /* drop pkt */ break; } err = cm_node->cm_core->cm_create_ah(cm_node, false); if (err) { irdma_passive_open_err(cm_node, false); /* drop pkt */ break; } cm_node->tcp_cntxt.rcv_nxt = inc_sequence + 1; cm_node->accept_pend = 1; atomic_inc(&cm_node->listener->pend_accepts_cnt); cm_node->state = IRDMA_CM_STATE_SYN_RCVD; break; case IRDMA_CM_STATE_CLOSED: irdma_cleanup_retrans_entry(cm_node); atomic_inc(&cm_node->refcnt); irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_OFFLOADED: case IRDMA_CM_STATE_ESTABLISHED: case IRDMA_CM_STATE_FIN_WAIT1: case IRDMA_CM_STATE_FIN_WAIT2: case IRDMA_CM_STATE_MPAREQ_RCVD: case IRDMA_CM_STATE_LAST_ACK: case IRDMA_CM_STATE_CLOSING: case IRDMA_CM_STATE_UNKNOWN: default: break; } } /** * irdma_handle_synack_pkt - Process SYN+ACK packet (active side) * @cm_node: connection's node * @rbuf: receive buffer */ static void irdma_handle_synack_pkt(struct irdma_cm_node *cm_node, struct irdma_puda_buf *rbuf) { struct tcphdr *tcph = (struct tcphdr *)rbuf->tcph; int err; u32 inc_sequence; int optionsize; optionsize = (tcph->th_off << 2) - sizeof(struct tcphdr); inc_sequence = ntohl(tcph->th_seq); switch (cm_node->state) { case IRDMA_CM_STATE_SYN_SENT: irdma_cleanup_retrans_entry(cm_node); /* active open */ if (irdma_check_syn(cm_node, tcph)) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "check syn fail\n"); return; } cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->th_ack); /* setup options */ err = irdma_handle_tcp_options(cm_node, tcph, optionsize, 0); if (err) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "cm_node=%p tcp_options failed\n", cm_node); break; } irdma_cleanup_retrans_entry(cm_node); cm_node->tcp_cntxt.rcv_nxt = inc_sequence + 1; irdma_send_ack(cm_node); /* ACK for the syn_ack */ err = irdma_send_mpa_request(cm_node); if (err) { irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "cm_node=%p irdma_send_mpa_request failed\n", cm_node); break; } cm_node->state = IRDMA_CM_STATE_MPAREQ_SENT; break; case IRDMA_CM_STATE_MPAREQ_RCVD: irdma_passive_open_err(cm_node, true); break; case IRDMA_CM_STATE_LISTENING: cm_node->tcp_cntxt.loc_seq_num = ntohl(tcph->th_ack); irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_CLOSED: cm_node->tcp_cntxt.loc_seq_num = ntohl(tcph->th_ack); irdma_cleanup_retrans_entry(cm_node); atomic_inc(&cm_node->refcnt); irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_ESTABLISHED: case IRDMA_CM_STATE_FIN_WAIT1: case IRDMA_CM_STATE_FIN_WAIT2: case IRDMA_CM_STATE_LAST_ACK: case IRDMA_CM_STATE_OFFLOADED: case IRDMA_CM_STATE_CLOSING: case IRDMA_CM_STATE_UNKNOWN: case IRDMA_CM_STATE_MPAREQ_SENT: default: break; } } /** * irdma_handle_ack_pkt - process packet with ACK * @cm_node: connection's node * @rbuf: receive buffer */ static int irdma_handle_ack_pkt(struct irdma_cm_node *cm_node, struct irdma_puda_buf *rbuf) { struct tcphdr *tcph = (struct tcphdr *)rbuf->tcph; u32 inc_sequence; int ret; int optionsize; u32 datasize = rbuf->datalen; optionsize = (tcph->th_off << 2) - sizeof(struct tcphdr); if (irdma_check_seq(cm_node, tcph)) return -EINVAL; inc_sequence = ntohl(tcph->th_seq); switch (cm_node->state) { case IRDMA_CM_STATE_SYN_RCVD: irdma_cleanup_retrans_entry(cm_node); ret = irdma_handle_tcp_options(cm_node, tcph, optionsize, 1); if (ret) return ret; cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->th_ack); cm_node->state = IRDMA_CM_STATE_ESTABLISHED; if (datasize) { cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; irdma_handle_rcv_mpa(cm_node, rbuf); } break; case IRDMA_CM_STATE_ESTABLISHED: irdma_cleanup_retrans_entry(cm_node); if (datasize) { cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; irdma_handle_rcv_mpa(cm_node, rbuf); } break; case IRDMA_CM_STATE_MPAREQ_SENT: cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->th_ack); if (datasize) { cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; cm_node->ack_rcvd = false; irdma_handle_rcv_mpa(cm_node, rbuf); } else { cm_node->ack_rcvd = true; } break; case IRDMA_CM_STATE_LISTENING: irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_CLOSED: irdma_cleanup_retrans_entry(cm_node); atomic_inc(&cm_node->refcnt); irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_LAST_ACK: case IRDMA_CM_STATE_CLOSING: irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_rem_ref_cm_node(cm_node); break; case IRDMA_CM_STATE_FIN_WAIT1: irdma_cleanup_retrans_entry(cm_node); cm_node->state = IRDMA_CM_STATE_FIN_WAIT2; break; case IRDMA_CM_STATE_SYN_SENT: case IRDMA_CM_STATE_FIN_WAIT2: case IRDMA_CM_STATE_OFFLOADED: case IRDMA_CM_STATE_MPAREQ_RCVD: case IRDMA_CM_STATE_UNKNOWN: default: irdma_cleanup_retrans_entry(cm_node); break; } return 0; } /** * irdma_process_pkt - process cm packet * @cm_node: connection's node * @rbuf: receive buffer */ static void irdma_process_pkt(struct irdma_cm_node *cm_node, struct irdma_puda_buf *rbuf) { enum irdma_tcpip_pkt_type pkt_type = IRDMA_PKT_TYPE_UNKNOWN; struct tcphdr *tcph = (struct tcphdr *)rbuf->tcph; u32 fin_set = 0; int err; if (tcph->th_flags & TH_RST) { pkt_type = IRDMA_PKT_TYPE_RST; } else if (tcph->th_flags & TH_SYN) { pkt_type = IRDMA_PKT_TYPE_SYN; if (tcph->th_flags & TH_ACK) pkt_type = IRDMA_PKT_TYPE_SYNACK; } else if (tcph->th_flags & TH_ACK) { pkt_type = IRDMA_PKT_TYPE_ACK; } if (tcph->th_flags & TH_FIN) fin_set = 1; switch (pkt_type) { case IRDMA_PKT_TYPE_SYN: irdma_handle_syn_pkt(cm_node, rbuf); break; case IRDMA_PKT_TYPE_SYNACK: irdma_handle_synack_pkt(cm_node, rbuf); break; case IRDMA_PKT_TYPE_ACK: err = irdma_handle_ack_pkt(cm_node, rbuf); if (fin_set && !err) irdma_handle_fin_pkt(cm_node); break; case IRDMA_PKT_TYPE_RST: irdma_handle_rst_pkt(cm_node, rbuf); break; default: if (fin_set && (!irdma_check_seq(cm_node, (struct tcphdr *)rbuf->tcph))) irdma_handle_fin_pkt(cm_node); break; } } /** * irdma_make_listen_node - create a listen node with params * @cm_core: cm's core * @iwdev: iwarp device structure * @cm_info: quad info for connection */ static struct irdma_cm_listener * irdma_make_listen_node(struct irdma_cm_core *cm_core, struct irdma_device *iwdev, struct irdma_cm_info *cm_info) { struct irdma_cm_listener *listener; unsigned long flags; /* cannot have multiple matching listeners */ listener = irdma_find_listener(cm_core, cm_info->loc_addr, cm_info->ipv4, cm_info->loc_port, cm_info->vlan_id, IRDMA_CM_LISTENER_EITHER_STATE); if (listener && listener->listener_state == IRDMA_CM_LISTENER_ACTIVE_STATE) { atomic_dec(&listener->refcnt); return NULL; } if (!listener) { /* * create a CM listen node 1/2 node to compare incoming traffic to */ listener = kzalloc(sizeof(*listener), GFP_KERNEL); if (!listener) return NULL; cm_core->stats_listen_nodes_created++; memcpy(listener->loc_addr, cm_info->loc_addr, sizeof(listener->loc_addr)); listener->loc_port = cm_info->loc_port; INIT_LIST_HEAD(&listener->child_listen_list); atomic_set(&listener->refcnt, 1); } else { listener->reused_node = 1; } listener->cm_id = cm_info->cm_id; listener->ipv4 = cm_info->ipv4; listener->vlan_id = cm_info->vlan_id; atomic_set(&listener->pend_accepts_cnt, 0); listener->cm_core = cm_core; listener->iwdev = iwdev; listener->backlog = cm_info->backlog; listener->listener_state = IRDMA_CM_LISTENER_ACTIVE_STATE; if (!listener->reused_node) { spin_lock_irqsave(&cm_core->listen_list_lock, flags); list_add(&listener->list, &cm_core->listen_list); spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); } return listener; } /** * irdma_create_cm_node - make a connection node with params * @cm_core: cm's core * @iwdev: iwarp device structure * @conn_param: connection parameters * @cm_info: quad info for connection * @caller_cm_node: pointer to cm_node structure to return */ static int irdma_create_cm_node(struct irdma_cm_core *cm_core, struct irdma_device *iwdev, struct iw_cm_conn_param *conn_param, struct irdma_cm_info *cm_info, struct irdma_cm_node **caller_cm_node) { struct irdma_cm_node *cm_node; u16 private_data_len = conn_param->private_data_len; const void *private_data = conn_param->private_data; /* create a CM connection node */ cm_node = irdma_make_cm_node(cm_core, iwdev, cm_info, NULL); if (!cm_node) return -ENOMEM; /* set our node side to client (active) side */ cm_node->tcp_cntxt.client = 1; cm_node->tcp_cntxt.rcv_wscale = IRDMA_CM_DEFAULT_RCV_WND_SCALE; irdma_record_ird_ord(cm_node, conn_param->ird, conn_param->ord); cm_node->pdata.size = private_data_len; cm_node->pdata.addr = cm_node->pdata_buf; memcpy(cm_node->pdata_buf, private_data, private_data_len); *caller_cm_node = cm_node; return 0; } /** * irdma_cm_reject - reject and teardown a connection * @cm_node: connection's node * @pdata: ptr to private data for reject * @plen: size of private data */ static int irdma_cm_reject(struct irdma_cm_node *cm_node, const void *pdata, u8 plen) { int ret; int passive_state; if (cm_node->tcp_cntxt.client) return 0; irdma_cleanup_retrans_entry(cm_node); passive_state = atomic_add_return(1, &cm_node->passive_state); if (passive_state == IRDMA_SEND_RESET_EVENT) { cm_node->state = IRDMA_CM_STATE_CLOSED; irdma_rem_ref_cm_node(cm_node); return 0; } if (cm_node->state == IRDMA_CM_STATE_LISTENER_DESTROYED) { irdma_rem_ref_cm_node(cm_node); return 0; } ret = irdma_send_mpa_reject(cm_node, pdata, plen); if (!ret) return 0; cm_node->state = IRDMA_CM_STATE_CLOSED; if (irdma_send_reset(cm_node)) irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "send reset failed\n"); return ret; } /** * irdma_cm_close - close of cm connection * @cm_node: connection's node */ static int irdma_cm_close(struct irdma_cm_node *cm_node) { switch (cm_node->state) { case IRDMA_CM_STATE_SYN_RCVD: case IRDMA_CM_STATE_SYN_SENT: case IRDMA_CM_STATE_ONE_SIDE_ESTABLISHED: case IRDMA_CM_STATE_ESTABLISHED: case IRDMA_CM_STATE_ACCEPTING: case IRDMA_CM_STATE_MPAREQ_SENT: case IRDMA_CM_STATE_MPAREQ_RCVD: irdma_cleanup_retrans_entry(cm_node); irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_CLOSE_WAIT: cm_node->state = IRDMA_CM_STATE_LAST_ACK; irdma_send_fin(cm_node); break; case IRDMA_CM_STATE_FIN_WAIT1: case IRDMA_CM_STATE_FIN_WAIT2: case IRDMA_CM_STATE_LAST_ACK: case IRDMA_CM_STATE_TIME_WAIT: case IRDMA_CM_STATE_CLOSING: return -EINVAL; case IRDMA_CM_STATE_LISTENING: irdma_cleanup_retrans_entry(cm_node); irdma_send_reset(cm_node); break; case IRDMA_CM_STATE_MPAREJ_RCVD: case IRDMA_CM_STATE_UNKNOWN: case IRDMA_CM_STATE_INITED: case IRDMA_CM_STATE_CLOSED: case IRDMA_CM_STATE_LISTENER_DESTROYED: irdma_rem_ref_cm_node(cm_node); break; case IRDMA_CM_STATE_OFFLOADED: if (cm_node->send_entry) irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "CM send_entry in OFFLOADED state\n"); irdma_rem_ref_cm_node(cm_node); break; } return 0; } /** * irdma_receive_ilq - recv an ETHERNET packet, and process it * through CM * @vsi: VSI structure of dev * @rbuf: receive buffer */ void irdma_receive_ilq(struct irdma_sc_vsi *vsi, struct irdma_puda_buf *rbuf) { struct irdma_cm_node *cm_node; struct irdma_cm_listener *listener; struct ip *iph; struct ip6_hdr *ip6h; struct tcphdr *tcph; struct irdma_cm_info cm_info = {0}; struct irdma_device *iwdev = vsi->back_vsi; struct irdma_cm_core *cm_core = &iwdev->cm_core; struct ether_vlan_header *ethh; u16 vtag; /* if vlan, then maclen = 18 else 14 */ iph = (struct ip *)rbuf->iph; irdma_debug_buf(vsi->dev, IRDMA_DEBUG_ILQ, "RECEIVE ILQ BUFFER", rbuf->mem.va, rbuf->totallen); if (iwdev->rf->sc_dev.hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) { if (rbuf->vlan_valid) { vtag = rbuf->vlan_id; cm_info.user_pri = (vtag & EVL_PRI_MASK) >> VLAN_PRIO_SHIFT; cm_info.vlan_id = vtag & EVL_VLID_MASK; } else { cm_info.vlan_id = 0xFFFF; } } else { ethh = rbuf->mem.va; if (ethh->evl_proto == htons(ETH_P_8021Q)) { vtag = ntohs(ethh->evl_tag); cm_info.user_pri = (vtag & EVL_PRI_MASK) >> VLAN_PRIO_SHIFT; cm_info.vlan_id = vtag & EVL_VLID_MASK; irdma_debug(&cm_core->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "vlan_id=%d\n", cm_info.vlan_id); } else { cm_info.vlan_id = 0xFFFF; } } tcph = (struct tcphdr *)rbuf->tcph; if (rbuf->ipv4) { cm_info.loc_addr[0] = ntohl(iph->ip_dst.s_addr); cm_info.rem_addr[0] = ntohl(iph->ip_src.s_addr); cm_info.ipv4 = true; cm_info.tos = iph->ip_tos; } else { ip6h = (struct ip6_hdr *)rbuf->iph; irdma_copy_ip_ntohl(cm_info.loc_addr, ip6h->ip6_dst.__u6_addr.__u6_addr32); irdma_copy_ip_ntohl(cm_info.rem_addr, ip6h->ip6_src.__u6_addr.__u6_addr32); cm_info.ipv4 = false; cm_info.tos = (ip6h->ip6_vfc << 4) | ip6h->ip6_flow; } cm_info.loc_port = ntohs(tcph->th_dport); cm_info.rem_port = ntohs(tcph->th_sport); cm_node = irdma_find_node(cm_core, cm_info.rem_port, cm_info.rem_addr, cm_info.loc_port, cm_info.loc_addr, cm_info.vlan_id); if (!cm_node) { /* * Only type of packet accepted are for the PASSIVE open (syn only) */ if (!(tcph->th_flags & TH_SYN) || tcph->th_flags & TH_ACK) return; listener = irdma_find_listener(cm_core, cm_info.loc_addr, cm_info.ipv4, cm_info.loc_port, cm_info.vlan_id, IRDMA_CM_LISTENER_ACTIVE_STATE); if (!listener) { cm_info.cm_id = NULL; irdma_debug(&cm_core->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "no listener found\n"); return; } cm_info.cm_id = listener->cm_id; cm_node = irdma_make_cm_node(cm_core, iwdev, &cm_info, listener); if (!cm_node) { irdma_debug(&cm_core->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "allocate node failed\n"); atomic_dec(&listener->refcnt); return; } if (!(tcph->th_flags & (TH_RST | TH_FIN))) { cm_node->state = IRDMA_CM_STATE_LISTENING; } else { irdma_rem_ref_cm_node(cm_node); return; } atomic_inc(&cm_node->refcnt); } else if (cm_node->state == IRDMA_CM_STATE_OFFLOADED) { irdma_rem_ref_cm_node(cm_node); return; } irdma_process_pkt(cm_node, rbuf); irdma_rem_ref_cm_node(cm_node); } static int irdma_add_qh(struct irdma_cm_node *cm_node, bool active) { if (!active) irdma_add_conn_est_qh(cm_node); return 0; } static void irdma_cm_free_ah_nop(struct irdma_cm_node *cm_node) { } /** * irdma_setup_cm_core - setup top level instance of a cm core * @iwdev: iwarp device structure * @rdma_ver: HW version */ int irdma_setup_cm_core(struct irdma_device *iwdev, u8 rdma_ver) { struct irdma_cm_core *cm_core = &iwdev->cm_core; cm_core->iwdev = iwdev; cm_core->dev = &iwdev->rf->sc_dev; /* Handles CM event work items send to Iwarp core */ cm_core->event_wq = alloc_ordered_workqueue("iwarp-event-wq", 0); if (!cm_core->event_wq) return -ENOMEM; INIT_LIST_HEAD(&cm_core->listen_list); timer_setup(&cm_core->tcp_timer, irdma_cm_timer_tick, 0); spin_lock_init(&cm_core->ht_lock); spin_lock_init(&cm_core->listen_list_lock); spin_lock_init(&cm_core->apbvt_lock); switch (rdma_ver) { case IRDMA_GEN_1: cm_core->form_cm_frame = irdma_form_uda_cm_frame; cm_core->cm_create_ah = irdma_add_qh; cm_core->cm_free_ah = irdma_cm_free_ah_nop; break; case IRDMA_GEN_2: default: cm_core->form_cm_frame = irdma_form_ah_cm_frame; cm_core->cm_create_ah = irdma_cm_create_ah; cm_core->cm_free_ah = irdma_cm_free_ah; } return 0; } /** * irdma_cleanup_cm_core - deallocate a top level instance of a * cm core * @cm_core: cm's core */ void irdma_cleanup_cm_core(struct irdma_cm_core *cm_core) { if (!cm_core) return; del_timer_sync(&cm_core->tcp_timer); destroy_workqueue(cm_core->event_wq); cm_core->dev->ws_reset(&cm_core->iwdev->vsi); } /** * irdma_init_tcp_ctx - setup qp context * @cm_node: connection's node * @tcp_info: offload info for tcp * @iwqp: associate qp for the connection */ static void irdma_init_tcp_ctx(struct irdma_cm_node *cm_node, struct irdma_tcp_offload_info *tcp_info, struct irdma_qp *iwqp) { tcp_info->ipv4 = cm_node->ipv4; tcp_info->drop_ooo_seg = !iwqp->iwdev->iw_ooo; tcp_info->wscale = true; tcp_info->ignore_tcp_opt = true; tcp_info->ignore_tcp_uns_opt = true; tcp_info->no_nagle = false; tcp_info->ttl = IRDMA_DEFAULT_TTL; tcp_info->rtt_var = IRDMA_DEFAULT_RTT_VAR; tcp_info->ss_thresh = IRDMA_DEFAULT_SS_THRESH; tcp_info->rexmit_thresh = IRDMA_DEFAULT_REXMIT_THRESH; tcp_info->tcp_state = IRDMA_TCP_STATE_ESTABLISHED; tcp_info->snd_wscale = cm_node->tcp_cntxt.snd_wscale; tcp_info->rcv_wscale = cm_node->tcp_cntxt.rcv_wscale; tcp_info->snd_nxt = cm_node->tcp_cntxt.loc_seq_num; tcp_info->snd_wnd = cm_node->tcp_cntxt.snd_wnd; tcp_info->rcv_nxt = cm_node->tcp_cntxt.rcv_nxt; tcp_info->snd_max = cm_node->tcp_cntxt.loc_seq_num; tcp_info->snd_una = cm_node->tcp_cntxt.loc_seq_num; tcp_info->cwnd = 2 * cm_node->tcp_cntxt.mss; tcp_info->snd_wl1 = cm_node->tcp_cntxt.rcv_nxt; tcp_info->snd_wl2 = cm_node->tcp_cntxt.loc_seq_num; tcp_info->max_snd_window = cm_node->tcp_cntxt.max_snd_wnd; tcp_info->rcv_wnd = cm_node->tcp_cntxt.rcv_wnd << cm_node->tcp_cntxt.rcv_wscale; tcp_info->flow_label = 0; tcp_info->snd_mss = (u32)cm_node->tcp_cntxt.mss; tcp_info->tos = cm_node->tos; if (cm_node->vlan_id < VLAN_N_VID) { tcp_info->insert_vlan_tag = true; tcp_info->vlan_tag = cm_node->vlan_id; tcp_info->vlan_tag |= cm_node->user_pri << VLAN_PRIO_SHIFT; } tcp_info->src_port = cm_node->loc_port; tcp_info->dst_port = cm_node->rem_port; tcp_info->arp_idx = (u16)irdma_arp_table(iwqp->iwdev->rf, cm_node->rem_addr, NULL, IRDMA_ARP_RESOLVE); if (cm_node->ipv4) { tcp_info->dest_ip_addr[3] = cm_node->rem_addr[0]; tcp_info->local_ipaddr[3] = cm_node->loc_addr[0]; } else { memcpy(tcp_info->dest_ip_addr, cm_node->rem_addr, sizeof(tcp_info->dest_ip_addr)); memcpy(tcp_info->local_ipaddr, cm_node->loc_addr, sizeof(tcp_info->local_ipaddr)); } } /** * irdma_cm_init_tsa_conn - setup qp for RTS * @iwqp: associate qp for the connection * @cm_node: connection's node */ static void irdma_cm_init_tsa_conn(struct irdma_qp *iwqp, struct irdma_cm_node *cm_node) { struct irdma_iwarp_offload_info *iwarp_info; struct irdma_qp_host_ctx_info *ctx_info; iwarp_info = &iwqp->iwarp_info; ctx_info = &iwqp->ctx_info; ctx_info->tcp_info = &iwqp->tcp_info; ctx_info->send_cq_num = iwqp->iwscq->sc_cq.cq_uk.cq_id; ctx_info->rcv_cq_num = iwqp->iwrcq->sc_cq.cq_uk.cq_id; iwarp_info->ord_size = cm_node->ord_size; iwarp_info->ird_size = cm_node->ird_size; iwarp_info->rd_en = true; iwarp_info->rdmap_ver = 1; iwarp_info->ddp_ver = 1; iwarp_info->pd_id = iwqp->iwpd->sc_pd.pd_id; ctx_info->tcp_info_valid = true; ctx_info->iwarp_info_valid = true; ctx_info->user_pri = cm_node->user_pri; irdma_init_tcp_ctx(cm_node, &iwqp->tcp_info, iwqp); if (cm_node->snd_mark_en) { iwarp_info->snd_mark_en = true; iwarp_info->snd_mark_offset = (iwqp->tcp_info.snd_nxt & SNDMARKER_SEQNMASK) + cm_node->lsmm_size; } cm_node->state = IRDMA_CM_STATE_OFFLOADED; iwqp->tcp_info.tcp_state = IRDMA_TCP_STATE_ESTABLISHED; iwqp->tcp_info.src_mac_addr_idx = iwqp->iwdev->mac_ip_table_idx; if (cm_node->rcv_mark_en) { iwarp_info->rcv_mark_en = true; iwarp_info->align_hdrs = true; } irdma_sc_qp_setctx(&iwqp->sc_qp, iwqp->host_ctx.va, ctx_info); /* once tcp_info is set, no need to do it again */ ctx_info->tcp_info_valid = false; ctx_info->iwarp_info_valid = false; } /** * irdma_cm_disconn - when a connection is being closed * @iwqp: associated qp for the connection */ void irdma_cm_disconn(struct irdma_qp *iwqp) { struct irdma_device *iwdev = iwqp->iwdev; struct disconn_work *work; unsigned long flags; work = kzalloc(sizeof(*work), GFP_ATOMIC); if (!work) return; spin_lock_irqsave(&iwdev->rf->qptable_lock, flags); if (!iwdev->rf->qp_table[iwqp->ibqp.qp_num]) { spin_unlock_irqrestore(&iwdev->rf->qptable_lock, flags); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "qp_id %d is already freed\n", iwqp->ibqp.qp_num); kfree(work); return; } irdma_qp_add_ref(&iwqp->ibqp); spin_unlock_irqrestore(&iwdev->rf->qptable_lock, flags); work->iwqp = iwqp; INIT_WORK(&work->work, irdma_disconnect_worker); queue_work(iwdev->cleanup_wq, &work->work); } /** * irdma_qp_disconnect - free qp and close cm * @iwqp: associate qp for the connection */ static void irdma_qp_disconnect(struct irdma_qp *iwqp) { struct irdma_device *iwdev = iwqp->iwdev; iwqp->active_conn = 0; /* close the CM node down if it is still active */ irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Call close API\n"); irdma_cm_close(iwqp->cm_node); } /** * irdma_cm_disconn_true - called by worker thread to disconnect qp * @iwqp: associate qp for the connection */ static void irdma_cm_disconn_true(struct irdma_qp *iwqp) { struct iw_cm_id *cm_id; struct irdma_device *iwdev; struct irdma_sc_qp *qp = &iwqp->sc_qp; u16 last_ae; u8 original_hw_tcp_state; u8 original_ibqp_state; int disconn_status = 0; int issue_disconn = 0; int issue_close = 0; int issue_flush = 0; unsigned long flags; int err; iwdev = iwqp->iwdev; spin_lock_irqsave(&iwqp->lock, flags); if (rdma_protocol_roce(&iwdev->ibdev, 1)) { struct ib_qp_attr attr; if (iwqp->flush_issued || iwqp->sc_qp.qp_uk.destroy_pending) { spin_unlock_irqrestore(&iwqp->lock, flags); return; } spin_unlock_irqrestore(&iwqp->lock, flags); attr.qp_state = IB_QPS_ERR; irdma_modify_qp_roce(&iwqp->ibqp, &attr, IB_QP_STATE, NULL); irdma_ib_qp_event(iwqp, qp->event_type); return; } cm_id = iwqp->cm_id; original_hw_tcp_state = iwqp->hw_tcp_state; original_ibqp_state = iwqp->ibqp_state; last_ae = iwqp->last_aeq; if (qp->term_flags) { issue_disconn = 1; issue_close = 1; iwqp->cm_id = NULL; irdma_terminate_del_timer(qp); if (!iwqp->flush_issued) { iwqp->flush_issued = 1; issue_flush = 1; } } else if ((original_hw_tcp_state == IRDMA_TCP_STATE_CLOSE_WAIT) || ((original_ibqp_state == IB_QPS_RTS) && (last_ae == IRDMA_AE_LLP_CONNECTION_RESET))) { issue_disconn = 1; if (last_ae == IRDMA_AE_LLP_CONNECTION_RESET) disconn_status = -ECONNRESET; } if (original_hw_tcp_state == IRDMA_TCP_STATE_CLOSED || original_hw_tcp_state == IRDMA_TCP_STATE_TIME_WAIT || last_ae == IRDMA_AE_RDMAP_ROE_BAD_LLP_CLOSE || last_ae == IRDMA_AE_BAD_CLOSE || last_ae == IRDMA_AE_LLP_CONNECTION_RESET || iwdev->rf->reset || !cm_id) { issue_close = 1; iwqp->cm_id = NULL; qp->term_flags = 0; if (!iwqp->flush_issued) { iwqp->flush_issued = 1; issue_flush = 1; } } spin_unlock_irqrestore(&iwqp->lock, flags); if (issue_flush && !iwqp->sc_qp.qp_uk.destroy_pending) { irdma_flush_wqes(iwqp, IRDMA_FLUSH_SQ | IRDMA_FLUSH_RQ | IRDMA_FLUSH_WAIT); if (qp->term_flags) irdma_ib_qp_event(iwqp, qp->event_type); } if (!cm_id || !cm_id->event_handler) return; spin_lock_irqsave(&iwdev->cm_core.ht_lock, flags); if (!iwqp->cm_node) { spin_unlock_irqrestore(&iwdev->cm_core.ht_lock, flags); return; } atomic_inc(&iwqp->cm_node->refcnt); spin_unlock_irqrestore(&iwdev->cm_core.ht_lock, flags); if (issue_disconn) { err = irdma_send_cm_event(iwqp->cm_node, cm_id, IW_CM_EVENT_DISCONNECT, disconn_status); if (err) irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "disconnect event failed: - cm_id = %p\n", cm_id); } if (issue_close) { cm_id->provider_data = iwqp; err = irdma_send_cm_event(iwqp->cm_node, cm_id, IW_CM_EVENT_CLOSE, 0); if (err) irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "close event failed: - cm_id = %p\n", cm_id); irdma_qp_disconnect(iwqp); } irdma_rem_ref_cm_node(iwqp->cm_node); } /** * irdma_disconnect_worker - worker for connection close * @work: points or disconn structure */ static void irdma_disconnect_worker(struct work_struct *work) { struct disconn_work *dwork = container_of(work, struct disconn_work, work); struct irdma_qp *iwqp = dwork->iwqp; kfree(dwork); irdma_cm_disconn_true(iwqp); irdma_qp_rem_ref(&iwqp->ibqp); } /** * irdma_free_lsmm_rsrc - free lsmm memory and deregister * @iwqp: associate qp for the connection */ void irdma_free_lsmm_rsrc(struct irdma_qp *iwqp) { struct irdma_device *iwdev; iwdev = iwqp->iwdev; if (iwqp->ietf_mem.va) { if (iwqp->lsmm_mr) iwdev->ibdev.dereg_mr(iwqp->lsmm_mr); irdma_free_dma_mem(iwdev->rf->sc_dev.hw, &iwqp->ietf_mem); iwqp->ietf_mem.va = NULL; } } /** * irdma_accept - registered call for connection to be accepted * @cm_id: cm information for passive connection * @conn_param: accpet parameters */ int irdma_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) { struct ib_qp *ibqp; struct irdma_qp *iwqp; struct irdma_device *iwdev; struct irdma_sc_dev *dev; struct irdma_cm_node *cm_node; struct ib_qp_attr attr = {0}; int passive_state; struct ib_mr *ibmr; struct irdma_pd *iwpd; u16 buf_len = 0; struct irdma_kmem_info accept; u64 tagged_offset; int wait_ret; int ret = 0; ibqp = irdma_get_qp(cm_id->device, conn_param->qpn); if (!ibqp) return -EINVAL; iwqp = to_iwqp(ibqp); iwdev = iwqp->iwdev; dev = &iwdev->rf->sc_dev; cm_node = cm_id->provider_data; if (((struct sockaddr_in *)&cm_id->local_addr)->sin_family == AF_INET) { cm_node->ipv4 = true; - cm_node->vlan_id = irdma_get_vlan_ipv4(cm_node->loc_addr); + cm_node->vlan_id = irdma_get_vlan_ipv4(cm_id, cm_node->loc_addr); } else { cm_node->ipv4 = false; - irdma_netdev_vlan_ipv6(cm_node->loc_addr, &cm_node->vlan_id, - NULL); + irdma_netdev_vlan_ipv6(cm_id, cm_node->loc_addr, + &cm_node->vlan_id, NULL); } irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Accept vlan_id=%d\n", cm_node->vlan_id); if (cm_node->state == IRDMA_CM_STATE_LISTENER_DESTROYED) { ret = -EINVAL; goto error; } passive_state = atomic_add_return(1, &cm_node->passive_state); if (passive_state == IRDMA_SEND_RESET_EVENT) { ret = -ECONNRESET; goto error; } buf_len = conn_param->private_data_len + IRDMA_MAX_IETF_SIZE; iwqp->ietf_mem.size = buf_len; iwqp->ietf_mem.va = irdma_allocate_dma_mem(dev->hw, &iwqp->ietf_mem, iwqp->ietf_mem.size, 1); if (!iwqp->ietf_mem.va) { ret = -ENOMEM; goto error; } cm_node->pdata.size = conn_param->private_data_len; accept.addr = iwqp->ietf_mem.va; accept.size = irdma_cm_build_mpa_frame(cm_node, &accept, MPA_KEY_REPLY); memcpy((u8 *)accept.addr + accept.size, conn_param->private_data, conn_param->private_data_len); if (cm_node->dev->ws_add(iwqp->sc_qp.vsi, cm_node->user_pri)) { ret = -ENOMEM; goto error; } iwqp->sc_qp.user_pri = cm_node->user_pri; irdma_qp_add_qos(&iwqp->sc_qp); if (cm_node->dev->hw_attrs.uk_attrs.hw_rev == IRDMA_GEN_2) iwdev->rf->check_fc(&iwdev->vsi, &iwqp->sc_qp); /* setup our first outgoing iWarp send WQE (the IETF frame response) */ iwpd = iwqp->iwpd; tagged_offset = (uintptr_t)iwqp->ietf_mem.va; ibmr = irdma_reg_phys_mr(&iwpd->ibpd, iwqp->ietf_mem.pa, buf_len, IB_ACCESS_LOCAL_WRITE, &tagged_offset); if (IS_ERR(ibmr)) { ret = -ENOMEM; goto error; } ibmr->pd = &iwpd->ibpd; ibmr->device = iwpd->ibpd.device; iwqp->lsmm_mr = ibmr; if (iwqp->page) iwqp->sc_qp.qp_uk.sq_base = kmap_local_page(iwqp->page); cm_node->lsmm_size = accept.size + conn_param->private_data_len; irdma_sc_send_lsmm(&iwqp->sc_qp, iwqp->ietf_mem.va, cm_node->lsmm_size, ibmr->lkey); if (iwqp->page) kunmap_local(iwqp->sc_qp.qp_uk.sq_base); iwqp->cm_id = cm_id; cm_node->cm_id = cm_id; cm_id->provider_data = iwqp; iwqp->active_conn = 0; iwqp->cm_node = cm_node; cm_node->iwqp = iwqp; irdma_cm_init_tsa_conn(iwqp, cm_node); irdma_qp_add_ref(&iwqp->ibqp); cm_id->add_ref(cm_id); attr.qp_state = IB_QPS_RTS; cm_node->qhash_set = false; cm_node->cm_core->cm_free_ah(cm_node); irdma_modify_qp(&iwqp->ibqp, &attr, IB_QP_STATE, NULL); if (dev->hw_attrs.uk_attrs.feature_flags & IRDMA_FEATURE_RTS_AE) { wait_ret = wait_event_interruptible_timeout(iwqp->waitq, iwqp->rts_ae_rcvd, IRDMA_MAX_TIMEOUT); if (!wait_ret) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Slow Connection: cm_node=%p, loc_port=%d, rem_port=%d, cm_id=%p\n", cm_node, cm_node->loc_port, cm_node->rem_port, cm_node->cm_id); ret = -ECONNRESET; goto error; } } irdma_send_cm_event(cm_node, cm_id, IW_CM_EVENT_ESTABLISHED, 0); cm_node->accelerated = true; complete(&cm_node->establish_comp); if (cm_node->accept_pend) { atomic_dec(&cm_node->listener->pend_accepts_cnt); cm_node->accept_pend = 0; } irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "rem_port=0x%04x, loc_port=0x%04x rem_addr=%x loc_addr=%x cm_node=%p cm_id=%p qp_id=%d\n\n", cm_node->rem_port, cm_node->loc_port, cm_node->rem_addr[0], cm_node->loc_addr[0], cm_node, cm_id, ibqp->qp_num); cm_node->cm_core->stats_accepts++; return 0; error: irdma_free_lsmm_rsrc(iwqp); irdma_rem_ref_cm_node(cm_node); return ret; } /** * irdma_reject - registered call for connection to be rejected * @cm_id: cm information for passive connection * @pdata: private data to be sent * @pdata_len: private data length */ int irdma_reject(struct iw_cm_id *cm_id, const void *pdata, u8 pdata_len) { struct irdma_device *iwdev; struct irdma_cm_node *cm_node; cm_node = cm_id->provider_data; cm_node->pdata.size = pdata_len; iwdev = to_iwdev(cm_id->device); if (!iwdev) return -EINVAL; cm_node->cm_core->stats_rejects++; if (pdata_len + sizeof(struct ietf_mpa_v2) > IRDMA_MAX_CM_BUF) return -EINVAL; return irdma_cm_reject(cm_node, pdata, pdata_len); } /** * irdma_connect - registered call for connection to be established * @cm_id: cm information for passive connection * @conn_param: Information about the connection */ int irdma_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) { struct ib_qp *ibqp; struct irdma_qp *iwqp; struct irdma_device *iwdev; struct irdma_cm_node *cm_node; struct irdma_cm_info cm_info; struct sockaddr_in *laddr; struct sockaddr_in *raddr; struct sockaddr_in6 *laddr6; struct sockaddr_in6 *raddr6; int ret = 0; ibqp = irdma_get_qp(cm_id->device, conn_param->qpn); if (!ibqp) return -EINVAL; iwqp = to_iwqp(ibqp); if (!iwqp) return -EINVAL; iwdev = iwqp->iwdev; if (!iwdev) return -EINVAL; laddr = (struct sockaddr_in *)&cm_id->m_local_addr; raddr = (struct sockaddr_in *)&cm_id->m_remote_addr; laddr6 = (struct sockaddr_in6 *)&cm_id->m_local_addr; raddr6 = (struct sockaddr_in6 *)&cm_id->m_remote_addr; if (!(laddr->sin_port) || !(raddr->sin_port)) return -EINVAL; iwqp->active_conn = 1; iwqp->cm_id = NULL; cm_id->provider_data = iwqp; /* set up the connection params for the node */ if (cm_id->remote_addr.ss_family == AF_INET) { if (iwdev->vsi.mtu < IRDMA_MIN_MTU_IPV4) return -EINVAL; cm_info.ipv4 = true; memset(cm_info.loc_addr, 0, sizeof(cm_info.loc_addr)); memset(cm_info.rem_addr, 0, sizeof(cm_info.rem_addr)); cm_info.loc_addr[0] = ntohl(laddr->sin_addr.s_addr); cm_info.rem_addr[0] = ntohl(raddr->sin_addr.s_addr); cm_info.loc_port = ntohs(laddr->sin_port); cm_info.rem_port = ntohs(raddr->sin_port); - cm_info.vlan_id = irdma_get_vlan_ipv4(cm_info.loc_addr); + cm_info.vlan_id = irdma_get_vlan_ipv4(cm_id, cm_info.loc_addr); } else { if (iwdev->vsi.mtu < IRDMA_MIN_MTU_IPV6) return -EINVAL; cm_info.ipv4 = false; irdma_copy_ip_ntohl(cm_info.loc_addr, laddr6->sin6_addr.__u6_addr.__u6_addr32); irdma_copy_ip_ntohl(cm_info.rem_addr, raddr6->sin6_addr.__u6_addr.__u6_addr32); cm_info.loc_port = ntohs(laddr6->sin6_port); cm_info.rem_port = ntohs(raddr6->sin6_port); - irdma_netdev_vlan_ipv6(cm_info.loc_addr, &cm_info.vlan_id, NULL); + irdma_netdev_vlan_ipv6(cm_id, cm_info.loc_addr, &cm_info.vlan_id, NULL); } cm_info.cm_id = cm_id; cm_info.qh_qpid = iwdev->vsi.ilq->qp_id; cm_info.tos = cm_id->tos; if (iwdev->vsi.dscp_mode) { cm_info.user_pri = iwqp->sc_qp.vsi->dscp_map[irdma_tos2dscp(cm_info.tos)]; } else { cm_info.user_pri = rt_tos2priority(cm_id->tos); cm_info.user_pri = irdma_iw_get_vlan_prio(cm_info.loc_addr, cm_info.user_pri, cm_info.ipv4); } if (iwqp->sc_qp.dev->ws_add(iwqp->sc_qp.vsi, cm_info.user_pri)) return -ENOMEM; iwqp->sc_qp.user_pri = cm_info.user_pri; irdma_qp_add_qos(&iwqp->sc_qp); if (iwdev->rf->sc_dev.hw_attrs.uk_attrs.hw_rev == IRDMA_GEN_2) iwdev->rf->check_fc(&iwdev->vsi, &iwqp->sc_qp); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_DCB, "TOS:[%d] UP:[%d]\n", cm_id->tos, cm_info.user_pri); ret = irdma_create_cm_node(&iwdev->cm_core, iwdev, conn_param, &cm_info, &cm_node); if (ret) return ret; ret = cm_node->cm_core->cm_create_ah(cm_node, true); if (ret) goto err; if (irdma_manage_qhash(iwdev, &cm_info, IRDMA_QHASH_TYPE_TCP_ESTABLISHED, IRDMA_QHASH_MANAGE_TYPE_ADD, NULL, true)) { ret = -EINVAL; goto err; } cm_node->qhash_set = true; cm_node->apbvt_entry = irdma_add_apbvt(iwdev, cm_info.loc_port); if (!cm_node->apbvt_entry) { ret = -EINVAL; goto err; } cm_node->apbvt_set = true; iwqp->cm_node = cm_node; cm_node->iwqp = iwqp; iwqp->cm_id = cm_id; irdma_qp_add_ref(&iwqp->ibqp); cm_id->add_ref(cm_id); if (cm_node->state != IRDMA_CM_STATE_OFFLOADED) { cm_node->state = IRDMA_CM_STATE_SYN_SENT; ret = irdma_send_syn(cm_node, 0); if (ret) goto err; } irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "rem_port=0x%04x, loc_port=0x%04x rem_addr=%x loc_addr=%x cm_node=%p cm_id=%p qp_id = %d\n\n", cm_node->rem_port, cm_node->loc_port, cm_node->rem_addr[0], cm_node->loc_addr[0], cm_node, cm_id, ibqp->qp_num); return 0; err: if (cm_info.ipv4) irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "connect() FAILED: dest addr=%x", cm_info.rem_addr[0]); else irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "connect() FAILED: dest addr=%x:%x:%x:%x", IRDMA_PRINT_IP6(cm_info.rem_addr)); irdma_rem_ref_cm_node(cm_node); iwdev->cm_core.stats_connect_errs++; return ret; } /** * irdma_create_listen - registered call creating listener * @cm_id: cm information for passive connection * @backlog: to max accept pending count */ int irdma_create_listen(struct iw_cm_id *cm_id, int backlog) { struct irdma_device *iwdev; struct irdma_cm_listener *cm_listen_node; struct irdma_cm_info cm_info = {0}; struct sockaddr_in *laddr; struct sockaddr_in6 *laddr6; bool wildcard = false; int err; iwdev = to_iwdev(cm_id->device); if (!iwdev) return -EINVAL; laddr = (struct sockaddr_in *)&cm_id->m_local_addr; laddr6 = (struct sockaddr_in6 *)&cm_id->m_local_addr; cm_info.qh_qpid = iwdev->vsi.ilq->qp_id; if (laddr->sin_family == AF_INET) { if (iwdev->vsi.mtu < IRDMA_MIN_MTU_IPV4) return -EINVAL; cm_info.ipv4 = true; cm_info.loc_addr[0] = ntohl(laddr->sin_addr.s_addr); cm_info.loc_port = ntohs(laddr->sin_port); if (laddr->sin_addr.s_addr != htonl(INADDR_ANY)) { - cm_info.vlan_id = irdma_get_vlan_ipv4(cm_info.loc_addr); + cm_info.vlan_id = irdma_get_vlan_ipv4(cm_id, cm_info.loc_addr); } else { cm_info.vlan_id = 0xFFFF; wildcard = true; } } else { if (iwdev->vsi.mtu < IRDMA_MIN_MTU_IPV6) return -EINVAL; cm_info.ipv4 = false; irdma_copy_ip_ntohl(cm_info.loc_addr, laddr6->sin6_addr.__u6_addr.__u6_addr32); cm_info.loc_port = ntohs(laddr6->sin6_port); if (!IN6_IS_ADDR_UNSPECIFIED(&laddr6->sin6_addr)) { - irdma_netdev_vlan_ipv6(cm_info.loc_addr, + irdma_netdev_vlan_ipv6(cm_id, cm_info.loc_addr, &cm_info.vlan_id, NULL); } else { cm_info.vlan_id = 0xFFFF; wildcard = true; } } if (cm_info.vlan_id >= VLAN_N_VID && iwdev->dcb_vlan_mode) cm_info.vlan_id = 0; cm_info.backlog = backlog; cm_info.cm_id = cm_id; cm_listen_node = irdma_make_listen_node(&iwdev->cm_core, iwdev, &cm_info); if (!cm_listen_node) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "cm_listen_node == NULL\n"); return -ENOMEM; } cm_id->provider_data = cm_listen_node; cm_listen_node->tos = cm_id->tos; if (iwdev->vsi.dscp_mode) cm_listen_node->user_pri = iwdev->vsi.dscp_map[irdma_tos2dscp(cm_id->tos)]; else cm_listen_node->user_pri = rt_tos2priority(cm_id->tos); cm_info.user_pri = cm_listen_node->user_pri; if (!cm_listen_node->reused_node) { if (wildcard) { err = irdma_add_mqh(iwdev, &cm_info, cm_listen_node); if (err) goto error; } else { if (!iwdev->vsi.dscp_mode) cm_info.user_pri = cm_listen_node->user_pri = irdma_iw_get_vlan_prio(cm_info.loc_addr, cm_info.user_pri, cm_info.ipv4); err = irdma_manage_qhash(iwdev, &cm_info, IRDMA_QHASH_TYPE_TCP_SYN, IRDMA_QHASH_MANAGE_TYPE_ADD, NULL, true); if (err) goto error; cm_listen_node->qhash_set = true; } cm_listen_node->apbvt_entry = irdma_add_apbvt(iwdev, cm_info.loc_port); if (!cm_listen_node->apbvt_entry) goto error; } cm_id->add_ref(cm_id); cm_listen_node->cm_core->stats_listen_created++; irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "loc_port=0x%04x loc_addr=%x cm_listen_node=%p cm_id=%p qhash_set=%d vlan_id=%d\n", cm_listen_node->loc_port, cm_listen_node->loc_addr[0], cm_listen_node, cm_listen_node->cm_id, cm_listen_node->qhash_set, cm_listen_node->vlan_id); return 0; error: irdma_cm_del_listen(&iwdev->cm_core, cm_listen_node, false); return -EINVAL; } /** * irdma_destroy_listen - registered call to destroy listener * @cm_id: cm information for passive connection */ int irdma_destroy_listen(struct iw_cm_id *cm_id) { struct irdma_device *iwdev; iwdev = to_iwdev(cm_id->device); if (cm_id->provider_data) irdma_cm_del_listen(&iwdev->cm_core, cm_id->provider_data, true); else irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "cm_id->provider_data was NULL\n"); cm_id->rem_ref(cm_id); return 0; } /** * irdma_iw_teardown_list_prep - add conn nodes slated for tear * down to list * @cm_core: cm's core * @teardown_list: a list to which cm_node will be selected * @ipaddr: pointer to ip address * @nfo: pointer to cm_info structure instance * @disconnect_all: flag indicating disconnect all QPs */ static void irdma_iw_teardown_list_prep(struct irdma_cm_core *cm_core, struct list_head *teardown_list, u32 *ipaddr, struct irdma_cm_info *nfo, bool disconnect_all) { struct irdma_cm_node *cm_node; int bkt; HASH_FOR_EACH_RCU(cm_core->cm_hash_tbl, bkt, cm_node, list) { if ((disconnect_all || (nfo->vlan_id == cm_node->vlan_id && !memcmp(cm_node->loc_addr, ipaddr, nfo->ipv4 ? 4 : 16))) && atomic_inc_not_zero(&cm_node->refcnt)) list_add(&cm_node->teardown_entry, teardown_list); } } static inline bool irdma_ip_vlan_match(u32 *ip1, u16 vlan_id1, bool check_vlan, u32 *ip2, u16 vlan_id2, bool ipv4) { return (!check_vlan || vlan_id1 == vlan_id2) && !memcmp(ip1, ip2, ipv4 ? 4 : 16); } /** * irdma_roce_teardown_list_prep - add conn nodes slated for * tear down to list * @iwdev: RDMA device * @teardown_list: a list to which cm_node will be selected * @ipaddr: pointer to ip address * @nfo: pointer to cm_info structure instance * @disconnect_all: flag indicating disconnect all QPs */ static void irdma_roce_teardown_list_prep(struct irdma_device *iwdev, struct list_head *teardown_list, u32 *ipaddr, struct irdma_cm_info *nfo, bool disconnect_all) { struct irdma_sc_vsi *vsi = &iwdev->vsi; struct irdma_sc_qp *sc_qp; struct list_head *list_node; struct irdma_qp *qp; unsigned long flags; int i; for (i = 0; i < IRDMA_MAX_USER_PRIORITY; i++) { mutex_lock(&vsi->qos[i].qos_mutex); list_for_each(list_node, &vsi->qos[i].qplist) { u32 qp_ip[4]; sc_qp = container_of(list_node, struct irdma_sc_qp, list); if (sc_qp->qp_uk.qp_type != IRDMA_QP_TYPE_ROCE_RC) continue; qp = sc_qp->qp_uk.back_qp; if (!disconnect_all) { if (nfo->ipv4) qp_ip[0] = qp->udp_info.local_ipaddr[3]; else memcpy(qp_ip, &qp->udp_info.local_ipaddr[0], sizeof(qp_ip)); } if (disconnect_all || irdma_ip_vlan_match(qp_ip, qp->udp_info.vlan_tag & EVL_VLID_MASK, qp->udp_info.insert_vlan_tag, ipaddr, nfo->vlan_id, nfo->ipv4)) { spin_lock_irqsave(&iwdev->rf->qptable_lock, flags); if (iwdev->rf->qp_table[sc_qp->qp_uk.qp_id]) { irdma_qp_add_ref(&qp->ibqp); list_add(&qp->teardown_entry, teardown_list); } spin_unlock_irqrestore(&iwdev->rf->qptable_lock, flags); } } mutex_unlock(&vsi->qos[i].qos_mutex); } } /** * irdma_cm_event_connected - handle connected active node * @event: the info for cm_node of connection */ static void irdma_cm_event_connected(struct irdma_cm_event *event) { struct irdma_qp *iwqp; struct irdma_device *iwdev; struct irdma_cm_node *cm_node; struct irdma_sc_dev *dev; struct ib_qp_attr attr = {0}; struct iw_cm_id *cm_id; int status; bool read0; int wait_ret = 0; cm_node = event->cm_node; cm_id = cm_node->cm_id; iwqp = cm_id->provider_data; iwdev = iwqp->iwdev; dev = &iwdev->rf->sc_dev; if (iwqp->sc_qp.qp_uk.destroy_pending) { status = -ETIMEDOUT; goto error; } irdma_cm_init_tsa_conn(iwqp, cm_node); read0 = (cm_node->send_rdma0_op == SEND_RDMA_READ_ZERO); if (iwqp->page) iwqp->sc_qp.qp_uk.sq_base = kmap_local_page(iwqp->page); irdma_sc_send_rtt(&iwqp->sc_qp, read0); if (iwqp->page) kunmap_local(iwqp->sc_qp.qp_uk.sq_base); attr.qp_state = IB_QPS_RTS; cm_node->qhash_set = false; irdma_modify_qp(&iwqp->ibqp, &attr, IB_QP_STATE, NULL); if (dev->hw_attrs.uk_attrs.feature_flags & IRDMA_FEATURE_RTS_AE) { wait_ret = wait_event_interruptible_timeout(iwqp->waitq, iwqp->rts_ae_rcvd, IRDMA_MAX_TIMEOUT); if (!wait_ret) irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "Slow Connection: cm_node=%p, loc_port=%d, rem_port=%d, cm_id=%p\n", cm_node, cm_node->loc_port, cm_node->rem_port, cm_node->cm_id); } irdma_send_cm_event(cm_node, cm_id, IW_CM_EVENT_CONNECT_REPLY, 0); cm_node->accelerated = true; complete(&cm_node->establish_comp); cm_node->cm_core->cm_free_ah(cm_node); return; error: iwqp->cm_id = NULL; cm_id->provider_data = NULL; irdma_send_cm_event(event->cm_node, cm_id, IW_CM_EVENT_CONNECT_REPLY, status); irdma_rem_ref_cm_node(event->cm_node); } /** * irdma_cm_event_reset - handle reset * @event: the info for cm_node of connection */ static void irdma_cm_event_reset(struct irdma_cm_event *event) { struct irdma_cm_node *cm_node = event->cm_node; struct iw_cm_id *cm_id = cm_node->cm_id; struct irdma_qp *iwqp; if (!cm_id) return; iwqp = cm_id->provider_data; if (!iwqp) return; irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "reset event %p - cm_id = %p\n", event->cm_node, cm_id); iwqp->cm_id = NULL; irdma_send_cm_event(cm_node, cm_node->cm_id, IW_CM_EVENT_DISCONNECT, -ECONNRESET); irdma_send_cm_event(cm_node, cm_node->cm_id, IW_CM_EVENT_CLOSE, 0); } /** * irdma_cm_event_handler - send event to cm upper layer * @work: pointer of cm event info. */ static void irdma_cm_event_handler(struct work_struct *work) { struct irdma_cm_event *event = container_of(work, struct irdma_cm_event, event_work); struct irdma_cm_node *cm_node; if (!event || !event->cm_node || !event->cm_node->cm_core) return; cm_node = event->cm_node; switch (event->type) { case IRDMA_CM_EVENT_MPA_REQ: irdma_send_cm_event(cm_node, cm_node->cm_id, IW_CM_EVENT_CONNECT_REQUEST, 0); break; case IRDMA_CM_EVENT_RESET: irdma_cm_event_reset(event); break; case IRDMA_CM_EVENT_CONNECTED: if (!event->cm_node->cm_id || event->cm_node->state != IRDMA_CM_STATE_OFFLOADED) break; irdma_cm_event_connected(event); break; case IRDMA_CM_EVENT_MPA_REJECT: if (!event->cm_node->cm_id || cm_node->state == IRDMA_CM_STATE_OFFLOADED) break; irdma_send_cm_event(cm_node, cm_node->cm_id, IW_CM_EVENT_CONNECT_REPLY, -ECONNREFUSED); break; case IRDMA_CM_EVENT_ABORTED: if (!event->cm_node->cm_id || event->cm_node->state == IRDMA_CM_STATE_OFFLOADED) break; irdma_event_connect_error(event); break; default: irdma_debug(&cm_node->iwdev->rf->sc_dev, IRDMA_DEBUG_CM, "bad event type = %d\n", event->type); break; } irdma_rem_ref_cm_node(event->cm_node); kfree(event); } /** * irdma_cm_post_event - queue event request for worker thread * @event: cm node's info for up event call */ static void irdma_cm_post_event(struct irdma_cm_event *event) { atomic_inc(&event->cm_node->refcnt); INIT_WORK(&event->event_work, irdma_cm_event_handler); queue_work(event->cm_node->cm_core->event_wq, &event->event_work); } /** * irdma_cm_teardown_connections - teardown QPs * @iwdev: device pointer * @ipaddr: Pointer to IPv4 or IPv6 address * @nfo: Connection info * @disconnect_all: flag indicating disconnect all QPs * * teardown QPs where source or destination addr matches ip addr */ static void __unused irdma_cm_teardown_connections(struct irdma_device *iwdev, u32 *ipaddr, struct irdma_cm_info *nfo, bool disconnect_all) { struct irdma_cm_core *cm_core = &iwdev->cm_core; struct list_head *list_core_temp; struct list_head *list_node; struct irdma_cm_node *cm_node; struct list_head teardown_list; struct ib_qp_attr attr; struct irdma_qp *qp; INIT_LIST_HEAD(&teardown_list); rcu_read_lock(); irdma_iw_teardown_list_prep(cm_core, &teardown_list, ipaddr, nfo, disconnect_all); rcu_read_unlock(); attr.qp_state = IB_QPS_ERR; list_for_each_safe(list_node, list_core_temp, &teardown_list) { cm_node = container_of(list_node, struct irdma_cm_node, teardown_entry); irdma_modify_qp(&cm_node->iwqp->ibqp, &attr, IB_QP_STATE, NULL); if (iwdev->rf->reset) irdma_cm_disconn(cm_node->iwqp); irdma_rem_ref_cm_node(cm_node); } if (!rdma_protocol_roce(&iwdev->ibdev, 1)) return; INIT_LIST_HEAD(&teardown_list); irdma_roce_teardown_list_prep(iwdev, &teardown_list, ipaddr, nfo, disconnect_all); list_for_each_safe(list_node, list_core_temp, &teardown_list) { qp = container_of(list_node, struct irdma_qp, teardown_entry); irdma_modify_qp_roce(&qp->ibqp, &attr, IB_QP_STATE, NULL); irdma_ib_qp_event(qp, IRDMA_QP_EVENT_CATASTROPHIC); irdma_qp_rem_ref(&qp->ibqp); } } diff --git a/sys/dev/irdma/irdma_main.h b/sys/dev/irdma/irdma_main.h index 8e304cae551f..3d6dd4e3f9db 100644 --- a/sys/dev/irdma/irdma_main.h +++ b/sys/dev/irdma/irdma_main.h @@ -1,601 +1,602 @@ /*- * SPDX-License-Identifier: GPL-2.0 or Linux-OpenIB * * Copyright (c) 2015 - 2022 Intel Corporation * * 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 * OpenFabrics.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. */ #ifndef IRDMA_MAIN_H #define IRDMA_MAIN_H #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "osdep.h" #include "irdma_defs.h" #include "irdma_hmc.h" #include "irdma_type.h" #include "irdma_ws.h" #include "irdma_protos.h" #include "irdma_pble.h" #include "irdma_cm.h" #include "fbsd_kcompat.h" #include "irdma-abi.h" #include "irdma_verbs.h" #include "irdma_user.h" #include "irdma_puda.h" extern struct list_head irdma_handlers; extern spinlock_t irdma_handler_lock; extern bool irdma_upload_context; #define IRDMA_FW_VER_DEFAULT 2 #define IRDMA_HW_VER 2 #define IRDMA_ARP_ADD 1 #define IRDMA_ARP_DELETE 2 #define IRDMA_ARP_RESOLVE 3 #define IRDMA_MACIP_ADD 1 #define IRDMA_MACIP_DELETE 2 #define IW_CCQ_SIZE (IRDMA_CQP_SW_SQSIZE_2048 + 1) #define IW_CEQ_SIZE 2048 #define IW_AEQ_SIZE 2048 #define RX_BUF_SIZE (1536 + 8) #define IW_REG0_SIZE (4 * 1024) #define IW_TX_TIMEOUT (6 * HZ) #define IW_FIRST_QPN 1 #define IW_SW_CONTEXT_ALIGN 1024 #define MAX_DPC_ITERATIONS 128 #define IRDMA_EVENT_TIMEOUT_MS 5000 #define IRDMA_VCHNL_EVENT_TIMEOUT_MS 10000 #define IRDMA_RST_TIMEOUT_HZ 4 #define IRDMA_NO_QSET 0xffff #define IW_CFG_FPM_QP_COUNT 32768 #define IRDMA_MAX_PAGES_PER_FMR 262144 #define IRDMA_MIN_PAGES_PER_FMR 1 #define IRDMA_CQP_COMPL_RQ_WQE_FLUSHED 2 #define IRDMA_CQP_COMPL_SQ_WQE_FLUSHED 3 #define IRDMA_Q_TYPE_PE_AEQ 0x80 #define IRDMA_Q_INVALID_IDX 0xffff #define IRDMA_REM_ENDPOINT_TRK_QPID 3 #define IRDMA_DRV_OPT_ENA_MPA_VER_0 0x00000001 #define IRDMA_DRV_OPT_DISABLE_MPA_CRC 0x00000002 #define IRDMA_DRV_OPT_DISABLE_FIRST_WRITE 0x00000004 #define IRDMA_DRV_OPT_DISABLE_INTF 0x00000008 #define IRDMA_DRV_OPT_ENA_MSI 0x00000010 #define IRDMA_DRV_OPT_DUAL_LOGICAL_PORT 0x00000020 #define IRDMA_DRV_OPT_NO_INLINE_DATA 0x00000080 #define IRDMA_DRV_OPT_DISABLE_INT_MOD 0x00000100 #define IRDMA_DRV_OPT_DISABLE_VIRT_WQ 0x00000200 #define IRDMA_DRV_OPT_ENA_PAU 0x00000400 #define IRDMA_DRV_OPT_MCAST_LOGPORT_MAP 0x00000800 #define IW_HMC_OBJ_TYPE_NUM ARRAY_SIZE(iw_hmc_obj_types) #define IRDMA_ROCE_CWND_DEFAULT 0x400 #define IRDMA_ROCE_ACKCREDS_DEFAULT 0x1E #define IRDMA_FLUSH_SQ BIT(0) #define IRDMA_FLUSH_RQ BIT(1) #define IRDMA_REFLUSH BIT(2) #define IRDMA_FLUSH_WAIT BIT(3) #define IRDMA_IRQ_NAME_STR_LEN 64 enum init_completion_state { INVALID_STATE = 0, INITIAL_STATE, CQP_CREATED, HMC_OBJS_CREATED, HW_RSRC_INITIALIZED, CCQ_CREATED, CEQ0_CREATED, /* Last state of probe */ ILQ_CREATED, IEQ_CREATED, REM_ENDPOINT_TRK_CREATED, CEQS_CREATED, PBLE_CHUNK_MEM, AEQ_CREATED, IP_ADDR_REGISTERED, /* Last state of open */ }; struct ae_desc { u16 id; const char *desc; }; struct irdma_rsrc_limits { u32 qplimit; u32 mrlimit; u32 cqlimit; }; struct irdma_cqp_err_info { u16 maj; u16 min; const char *desc; }; struct irdma_cqp_compl_info { u32 op_ret_val; u16 maj_err_code; u16 min_err_code; bool error; u8 op_code; }; struct irdma_cqp_request { struct cqp_cmds_info info; wait_queue_head_t waitq; struct list_head list; atomic_t refcnt; void (*callback_fcn)(struct irdma_cqp_request *cqp_request); void *param; struct irdma_cqp_compl_info compl_info; bool request_done; /* READ/WRITE_ONCE macros operate on it */ bool waiting:1; bool dynamic:1; }; struct irdma_cqp { struct irdma_sc_cqp sc_cqp; spinlock_t req_lock; /* protect CQP request list */ spinlock_t compl_lock; /* protect CQP completion processing */ wait_queue_head_t waitq; wait_queue_head_t remove_wq; struct irdma_dma_mem sq; struct irdma_dma_mem host_ctx; u64 *scratch_array; struct irdma_cqp_request *cqp_requests; struct list_head cqp_avail_reqs; struct list_head cqp_pending_reqs; }; struct irdma_ccq { struct irdma_sc_cq sc_cq; struct irdma_dma_mem mem_cq; struct irdma_dma_mem shadow_area; }; struct irdma_ceq { struct irdma_sc_ceq sc_ceq; struct irdma_dma_mem mem; u32 irq; u32 msix_idx; struct irdma_pci_f *rf; struct tasklet_struct dpc_tasklet; spinlock_t ce_lock; /* sync cq destroy with cq completion event notification */ }; struct irdma_aeq { struct irdma_sc_aeq sc_aeq; struct irdma_dma_mem mem; struct irdma_pble_alloc palloc; bool virtual_map; }; struct irdma_arp_entry { u32 ip_addr[4]; u8 mac_addr[ETHER_ADDR_LEN]; }; struct irdma_msix_vector { u32 idx; u32 irq; u32 cpu_affinity; u32 ceq_id; char name[IRDMA_IRQ_NAME_STR_LEN]; struct resource *res; void *tag; }; struct irdma_mc_table_info { u32 mgn; u32 dest_ip[4]; bool lan_fwd:1; bool ipv4_valid:1; }; struct mc_table_list { struct list_head list; struct irdma_mc_table_info mc_info; struct irdma_mcast_grp_info mc_grp_ctx; }; struct irdma_qv_info { u32 v_idx; /* msix_vector */ u16 ceq_idx; u16 aeq_idx; u8 itr_idx; }; struct irdma_qvlist_info { u32 num_vectors; struct irdma_qv_info qv_info[1]; }; struct irdma_gen_ops { void (*request_reset)(struct irdma_pci_f *rf); int (*register_qset)(struct irdma_sc_vsi *vsi, struct irdma_ws_node *tc_node); void (*unregister_qset)(struct irdma_sc_vsi *vsi, struct irdma_ws_node *tc_node); }; struct irdma_pci_f { bool reset:1; bool rsrc_created:1; bool msix_shared:1; bool ftype:1; u8 rsrc_profile; u8 *hmc_info_mem; u8 *mem_rsrc; u8 rdma_ver; u8 rst_to; /* Not used in SRIOV VF mode */ u8 pf_id; enum irdma_protocol_used protocol_used; bool en_rem_endpoint_trk:1; bool dcqcn_ena:1; u32 sd_type; u32 msix_count; u32 max_mr; u32 max_qp; u32 max_cq; u32 max_ah; u32 next_ah; u32 max_mcg; u32 next_mcg; u32 max_pd; u32 next_qp; u32 next_cq; u32 next_pd; u32 max_mr_size; u32 max_cqe; u32 mr_stagmask; u32 used_pds; u32 used_cqs; u32 used_mrs; u32 used_qps; u32 arp_table_size; u32 next_arp_index; u32 ceqs_count; u32 next_ws_node_id; u32 max_ws_node_id; u32 limits_sel; unsigned long *allocated_ws_nodes; unsigned long *allocated_qps; unsigned long *allocated_cqs; unsigned long *allocated_mrs; unsigned long *allocated_pds; unsigned long *allocated_mcgs; unsigned long *allocated_ahs; unsigned long *allocated_arps; enum init_completion_state init_state; struct irdma_sc_dev sc_dev; struct irdma_dev_ctx dev_ctx; struct irdma_tunable_info tun_info; eventhandler_tag irdma_ifaddr_event; struct irdma_handler *hdl; struct pci_dev *pcidev; struct ice_rdma_peer *peer_info; struct irdma_hw hw; struct irdma_cqp cqp; struct irdma_ccq ccq; struct irdma_aeq aeq; struct irdma_ceq *ceqlist; struct irdma_hmc_pble_rsrc *pble_rsrc; struct irdma_arp_entry *arp_table; spinlock_t arp_lock; /*protect ARP table access*/ spinlock_t rsrc_lock; /* protect HW resource array access */ spinlock_t qptable_lock; /*protect QP table access*/ spinlock_t cqtable_lock; /*protect CQ table access*/ struct irdma_qp **qp_table; struct irdma_cq **cq_table; spinlock_t qh_list_lock; /* protect mc_qht_list */ struct mc_table_list mc_qht_list; struct irdma_msix_vector *iw_msixtbl; struct irdma_qvlist_info *iw_qvlist; struct tasklet_struct dpc_tasklet; struct msix_entry msix_info; struct irdma_dma_mem obj_mem; struct irdma_dma_mem obj_next; atomic_t vchnl_msgs; wait_queue_head_t vchnl_waitq; struct workqueue_struct *cqp_cmpl_wq; struct work_struct cqp_cmpl_work; struct irdma_sc_vsi default_vsi; void *back_fcn; struct irdma_gen_ops gen_ops; void (*check_fc)(struct irdma_sc_vsi *vsi, struct irdma_sc_qp *sc_qp); struct irdma_dcqcn_cc_params dcqcn_params; struct irdma_device *iwdev; }; struct irdma_device { struct ib_device ibdev; struct irdma_pci_f *rf; if_t netdev; struct notifier_block nb_netdevice_event; struct irdma_handler *hdl; struct workqueue_struct *cleanup_wq; struct irdma_sc_vsi vsi; struct irdma_cm_core cm_core; u32 roce_cwnd; u32 roce_ackcreds; u32 vendor_id; u32 vendor_part_id; u32 push_mode; u32 rcv_wnd; u16 mac_ip_table_idx; u16 vsi_num; u8 rcv_wscale; u8 iw_status; u8 roce_rtomin; u8 rd_fence_rate; bool override_rcv_wnd:1; bool override_cwnd:1; bool override_ackcreds:1; bool override_ooo:1; bool override_rd_fence_rate:1; bool override_rtomin:1; bool roce_mode:1; bool roce_dcqcn_en:1; bool dcb_vlan_mode:1; bool iw_ooo:1; enum init_completion_state init_state; wait_queue_head_t suspend_wq; }; struct irdma_handler { struct list_head list; struct irdma_device *iwdev; struct task deferred_task; struct taskqueue *deferred_tq; bool shared_res_created; }; static inline struct irdma_device *to_iwdev(struct ib_device *ibdev) { return container_of(ibdev, struct irdma_device, ibdev); } static inline struct irdma_ucontext *to_ucontext(struct ib_ucontext *ibucontext) { return container_of(ibucontext, struct irdma_ucontext, ibucontext); } static inline struct irdma_pd *to_iwpd(struct ib_pd *ibpd) { return container_of(ibpd, struct irdma_pd, ibpd); } static inline struct irdma_ah *to_iwah(struct ib_ah *ibah) { return container_of(ibah, struct irdma_ah, ibah); } static inline struct irdma_mr *to_iwmr(struct ib_mr *ibmr) { return container_of(ibmr, struct irdma_mr, ibmr); } static inline struct irdma_mr *to_iwmw(struct ib_mw *ibmw) { return container_of(ibmw, struct irdma_mr, ibmw); } static inline struct irdma_cq *to_iwcq(struct ib_cq *ibcq) { return container_of(ibcq, struct irdma_cq, ibcq); } static inline struct irdma_qp *to_iwqp(struct ib_qp *ibqp) { return container_of(ibqp, struct irdma_qp, ibqp); } static inline struct irdma_pci_f *dev_to_rf(struct irdma_sc_dev *dev) { return container_of(dev, struct irdma_pci_f, sc_dev); } /** * irdma_alloc_resource - allocate a resource * @iwdev: device pointer * @resource_array: resource bit array: * @max_resources: maximum resource number * @req_resources_num: Allocated resource number * @next: next free id **/ static inline int irdma_alloc_rsrc(struct irdma_pci_f *rf, unsigned long *rsrc_array, u32 max_rsrc, u32 *req_rsrc_num, u32 *next) { u32 rsrc_num; unsigned long flags; spin_lock_irqsave(&rf->rsrc_lock, flags); rsrc_num = find_next_zero_bit(rsrc_array, max_rsrc, *next); if (rsrc_num >= max_rsrc) { rsrc_num = find_first_zero_bit(rsrc_array, max_rsrc); if (rsrc_num >= max_rsrc) { spin_unlock_irqrestore(&rf->rsrc_lock, flags); irdma_debug(&rf->sc_dev, IRDMA_DEBUG_ERR, "resource [%d] allocation failed\n", rsrc_num); return -EOVERFLOW; } } __set_bit(rsrc_num, rsrc_array); *next = rsrc_num + 1; if (*next == max_rsrc) *next = 0; *req_rsrc_num = rsrc_num; spin_unlock_irqrestore(&rf->rsrc_lock, flags); return 0; } /** * irdma_free_resource - free a resource * @iwdev: device pointer * @resource_array: resource array for the resource_num * @resource_num: resource number to free **/ static inline void irdma_free_rsrc(struct irdma_pci_f *rf, unsigned long *rsrc_array, u32 rsrc_num) { unsigned long flags; spin_lock_irqsave(&rf->rsrc_lock, flags); __clear_bit(rsrc_num, rsrc_array); spin_unlock_irqrestore(&rf->rsrc_lock, flags); } int irdma_ctrl_init_hw(struct irdma_pci_f *rf); void irdma_ctrl_deinit_hw(struct irdma_pci_f *rf); int irdma_rt_init_hw(struct irdma_device *iwdev, struct irdma_l2params *l2params); void irdma_rt_deinit_hw(struct irdma_device *iwdev); void irdma_qp_add_ref(struct ib_qp *ibqp); void irdma_qp_rem_ref(struct ib_qp *ibqp); void irdma_free_lsmm_rsrc(struct irdma_qp *iwqp); struct ib_qp *irdma_get_qp(struct ib_device *ibdev, int qpn); void irdma_flush_wqes(struct irdma_qp *iwqp, u32 flush_mask); void irdma_manage_arp_cache(struct irdma_pci_f *rf, const unsigned char *mac_addr, u32 *ip_addr, u32 action); struct irdma_apbvt_entry *irdma_add_apbvt(struct irdma_device *iwdev, u16 port); void irdma_del_apbvt(struct irdma_device *iwdev, struct irdma_apbvt_entry *entry); struct irdma_cqp_request *irdma_alloc_and_get_cqp_request(struct irdma_cqp *cqp, bool wait); void irdma_free_cqp_request(struct irdma_cqp *cqp, struct irdma_cqp_request *cqp_request); void irdma_put_cqp_request(struct irdma_cqp *cqp, struct irdma_cqp_request *cqp_request); int irdma_alloc_local_mac_entry(struct irdma_pci_f *rf, u16 *mac_tbl_idx); int irdma_add_local_mac_entry(struct irdma_pci_f *rf, const u8 *mac_addr, u16 idx); void irdma_del_local_mac_entry(struct irdma_pci_f *rf, u16 idx); const char *irdma_get_ae_desc(u16 ae_id); u32 irdma_initialize_hw_rsrc(struct irdma_pci_f *rf); void irdma_port_ibevent(struct irdma_device *iwdev); void irdma_cm_disconn(struct irdma_qp *qp); bool irdma_cqp_crit_err(struct irdma_sc_dev *dev, u8 cqp_cmd, u16 maj_err_code, u16 min_err_code); int irdma_handle_cqp_op(struct irdma_pci_f *rf, struct irdma_cqp_request *cqp_request); int irdma_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask, struct ib_udata *udata); int irdma_modify_qp_roce(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask, struct ib_udata *udata); void irdma_cq_add_ref(struct ib_cq *ibcq); void irdma_cq_rem_ref(struct ib_cq *ibcq); void irdma_cq_wq_destroy(struct irdma_pci_f *rf, struct irdma_sc_cq *cq); void irdma_cleanup_pending_cqp_op(struct irdma_pci_f *rf); int irdma_hw_modify_qp(struct irdma_device *iwdev, struct irdma_qp *iwqp, struct irdma_modify_qp_info *info, bool wait); int irdma_qp_suspend_resume(struct irdma_sc_qp *qp, bool suspend); int irdma_manage_qhash(struct irdma_device *iwdev, struct irdma_cm_info *cminfo, enum irdma_quad_entry_type etype, enum irdma_quad_hash_manage_type mtype, void *cmnode, bool wait); void irdma_receive_ilq(struct irdma_sc_vsi *vsi, struct irdma_puda_buf *rbuf); void irdma_free_sqbuf(struct irdma_sc_vsi *vsi, void *bufp); void irdma_free_qp_rsrc(struct irdma_qp *iwqp); int irdma_setup_cm_core(struct irdma_device *iwdev, u8 ver); void irdma_cleanup_cm_core(struct irdma_cm_core *cm_core); void irdma_next_iw_state(struct irdma_qp *iwqp, u8 state, u8 del_hash, u8 term, u8 term_len); int irdma_send_syn(struct irdma_cm_node *cm_node, u32 sendack); int irdma_send_reset(struct irdma_cm_node *cm_node); struct irdma_cm_node *irdma_find_node(struct irdma_cm_core *cm_core, u16 rem_port, u32 *rem_addr, u16 loc_port, u32 *loc_addr, u16 vlan_id); int irdma_hw_flush_wqes(struct irdma_pci_f *rf, struct irdma_sc_qp *qp, struct irdma_qp_flush_info *info, bool wait); void irdma_gen_ae(struct irdma_pci_f *rf, struct irdma_sc_qp *qp, struct irdma_gen_ae_info *info, bool wait); void irdma_copy_ip_ntohl(u32 *dst, __be32 *src); void irdma_copy_ip_htonl(__be32 *dst, u32 *src); -u16 irdma_get_vlan_ipv4(u32 *addr); -if_t irdma_netdev_vlan_ipv6(u32 *addr, u16 *vlan_id, u8 *mac); +u16 irdma_get_vlan_ipv4(struct iw_cm_id *cm_id, u32 *addr); +if_t irdma_netdev_vlan_ipv6(struct iw_cm_id *cm_id, u32 *addr, u16 *vlan_id, + u8 *mac); struct ib_mr *irdma_reg_phys_mr(struct ib_pd *ib_pd, u64 addr, u64 size, int acc, u64 *iova_start); int irdma_upload_qp_context(struct irdma_qp *iwqp, bool freeze, bool raw); void irdma_del_hmc_objects(struct irdma_sc_dev *dev, struct irdma_hmc_info *hmc_info, bool privileged, bool reset, enum irdma_vers vers); void irdma_cqp_ce_handler(struct irdma_pci_f *rf, struct irdma_sc_cq *cq); int irdma_ah_cqp_op(struct irdma_pci_f *rf, struct irdma_sc_ah *sc_ah, u8 cmd, bool wait, void (*callback_fcn)(struct irdma_cqp_request *cqp_request), void *cb_param); void irdma_gsi_ud_qp_ah_cb(struct irdma_cqp_request *cqp_request); void irdma_udqp_qs_worker(struct work_struct *work); bool irdma_cq_empty(struct irdma_cq *iwcq); int irdma_netdevice_event(struct notifier_block *notifier, unsigned long event, void *ptr); void irdma_unregister_notifiers(struct irdma_device *iwdev); int irdma_register_notifiers(struct irdma_device *iwdev); void irdma_set_rf_user_cfg_params(struct irdma_pci_f *rf); void irdma_add_ip(struct irdma_device *iwdev); void irdma_add_handler(struct irdma_handler *hdl); void irdma_del_handler(struct irdma_handler *hdl); void cqp_compl_worker(struct work_struct *work); void irdma_cleanup_dead_qps(struct irdma_sc_vsi *vsi); #endif /* IRDMA_MAIN_H */ diff --git a/sys/dev/irdma/irdma_verbs.c b/sys/dev/irdma/irdma_verbs.c index 288b075ab79f..c0726d25b8ec 100644 --- a/sys/dev/irdma/irdma_verbs.c +++ b/sys/dev/irdma/irdma_verbs.c @@ -1,3663 +1,3663 @@ /*- * SPDX-License-Identifier: GPL-2.0 or Linux-OpenIB * * Copyright (c) 2015 - 2023 Intel Corporation * * 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 * OpenFabrics.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 "irdma_main.h" /** * irdma_query_device - get device attributes * @ibdev: device pointer from stack * @props: returning device attributes * @udata: user data */ static int irdma_query_device(struct ib_device *ibdev, struct ib_device_attr *props, struct ib_udata *udata) { struct irdma_device *iwdev = to_iwdev(ibdev); struct irdma_pci_f *rf = iwdev->rf; struct pci_dev *pcidev = iwdev->rf->pcidev; struct irdma_hw_attrs *hw_attrs = &rf->sc_dev.hw_attrs; if (udata->inlen || udata->outlen) return -EINVAL; memset(props, 0, sizeof(*props)); addrconf_addr_eui48((u8 *)&props->sys_image_guid, if_getlladdr(iwdev->netdev)); props->fw_ver = (u64)irdma_fw_major_ver(&rf->sc_dev) << 32 | irdma_fw_minor_ver(&rf->sc_dev); props->device_cap_flags = IB_DEVICE_MEM_WINDOW | IB_DEVICE_MEM_MGT_EXTENSIONS; props->device_cap_flags |= IB_DEVICE_LOCAL_DMA_LKEY; props->vendor_id = pcidev->vendor; props->vendor_part_id = pcidev->device; props->hw_ver = pcidev->revision; props->page_size_cap = hw_attrs->page_size_cap; props->max_mr_size = hw_attrs->max_mr_size; props->max_qp = rf->max_qp - rf->used_qps; props->max_qp_wr = hw_attrs->max_qp_wr; set_max_sge(props, rf); props->max_cq = rf->max_cq - rf->used_cqs; props->max_cqe = rf->max_cqe - 1; props->max_mr = rf->max_mr - rf->used_mrs; props->max_mw = props->max_mr; props->max_pd = rf->max_pd - rf->used_pds; props->max_sge_rd = hw_attrs->uk_attrs.max_hw_read_sges; props->max_qp_rd_atom = hw_attrs->max_hw_ird; props->max_qp_init_rd_atom = hw_attrs->max_hw_ord; if (rdma_protocol_roce(ibdev, 1)) { props->device_cap_flags |= IB_DEVICE_RC_RNR_NAK_GEN; props->max_pkeys = IRDMA_PKEY_TBL_SZ; props->max_ah = rf->max_ah; if (hw_attrs->uk_attrs.hw_rev == IRDMA_GEN_2) { props->max_mcast_grp = rf->max_mcg; props->max_mcast_qp_attach = IRDMA_MAX_MGS_PER_CTX; props->max_total_mcast_qp_attach = rf->max_qp * IRDMA_MAX_MGS_PER_CTX; } } props->max_fast_reg_page_list_len = IRDMA_MAX_PAGES_PER_FMR; if (hw_attrs->uk_attrs.hw_rev >= IRDMA_GEN_2) props->device_cap_flags |= IB_DEVICE_MEM_WINDOW_TYPE_2B; return 0; } static int irdma_mmap_legacy(struct irdma_ucontext *ucontext, struct vm_area_struct *vma) { u64 pfn; if (vma->vm_pgoff || vma->vm_end - vma->vm_start != PAGE_SIZE) return -EINVAL; vma->vm_private_data = ucontext; pfn = ((uintptr_t)ucontext->iwdev->rf->sc_dev.hw_regs[IRDMA_DB_ADDR_OFFSET] + pci_resource_start(ucontext->iwdev->rf->pcidev, 0)) >> PAGE_SHIFT; return rdma_user_mmap_io(&ucontext->ibucontext, vma, pfn, PAGE_SIZE, pgprot_noncached(vma->vm_page_prot)); } static inline bool find_key_in_mmap_tbl(struct irdma_ucontext *ucontext, u64 key) { struct irdma_user_mmap_entry *entry; HASH_FOR_EACH_POSSIBLE(ucontext->mmap_hash_tbl, entry, hlist, key) { if (entry->pgoff_key == key) return true; } return false; } struct irdma_user_mmap_entry * irdma_user_mmap_entry_add_hash(struct irdma_ucontext *ucontext, u64 bar_offset, enum irdma_mmap_flag mmap_flag, u64 *mmap_offset) { struct irdma_user_mmap_entry *entry = kzalloc(sizeof(*entry), GFP_KERNEL); unsigned long flags; int retry_cnt = 0; if (!entry) return NULL; entry->bar_offset = bar_offset; entry->mmap_flag = mmap_flag; entry->ucontext = ucontext; do { get_random_bytes(&entry->pgoff_key, sizeof(entry->pgoff_key)); /* The key is a page offset */ entry->pgoff_key >>= PAGE_SHIFT; /* In the event of a collision in the hash table, retry a new key */ spin_lock_irqsave(&ucontext->mmap_tbl_lock, flags); if (!find_key_in_mmap_tbl(ucontext, entry->pgoff_key)) { HASH_ADD(ucontext->mmap_hash_tbl, &entry->hlist, entry->pgoff_key); spin_unlock_irqrestore(&ucontext->mmap_tbl_lock, flags); goto hash_add_done; } spin_unlock_irqrestore(&ucontext->mmap_tbl_lock, flags); } while (retry_cnt++ < 10); irdma_debug(&ucontext->iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "mmap table add failed: Cannot find a unique key\n"); kfree(entry); return NULL; hash_add_done: /* libc mmap uses a byte offset */ *mmap_offset = entry->pgoff_key << PAGE_SHIFT; return entry; } static struct irdma_user_mmap_entry * irdma_find_user_mmap_entry(struct irdma_ucontext *ucontext, struct vm_area_struct *vma) { struct irdma_user_mmap_entry *entry; unsigned long flags; if (vma->vm_end - vma->vm_start != PAGE_SIZE) return NULL; spin_lock_irqsave(&ucontext->mmap_tbl_lock, flags); HASH_FOR_EACH_POSSIBLE(ucontext->mmap_hash_tbl, entry, hlist, vma->vm_pgoff) { if (entry->pgoff_key == vma->vm_pgoff) { spin_unlock_irqrestore(&ucontext->mmap_tbl_lock, flags); return entry; } } spin_unlock_irqrestore(&ucontext->mmap_tbl_lock, flags); return NULL; } void irdma_user_mmap_entry_del_hash(struct irdma_user_mmap_entry *entry) { struct irdma_ucontext *ucontext; unsigned long flags; if (!entry) return; ucontext = entry->ucontext; spin_lock_irqsave(&ucontext->mmap_tbl_lock, flags); HASH_DEL(ucontext->mmap_hash_tbl, &entry->hlist); spin_unlock_irqrestore(&ucontext->mmap_tbl_lock, flags); kfree(entry); } /** * irdma_mmap - user memory map * @context: context created during alloc * @vma: kernel info for user memory map */ static int irdma_mmap(struct ib_ucontext *context, struct vm_area_struct *vma) { struct irdma_user_mmap_entry *entry; struct irdma_ucontext *ucontext; u64 pfn; int ret; ucontext = to_ucontext(context); /* Legacy support for libi40iw with hard-coded mmap key */ if (ucontext->legacy_mode) return irdma_mmap_legacy(ucontext, vma); entry = irdma_find_user_mmap_entry(ucontext, vma); if (!entry) { irdma_debug(&ucontext->iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "pgoff[0x%lx] does not have valid entry\n", vma->vm_pgoff); return -EINVAL; } irdma_debug(&ucontext->iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "bar_offset [0x%lx] mmap_flag [%d]\n", entry->bar_offset, entry->mmap_flag); pfn = (entry->bar_offset + pci_resource_start(ucontext->iwdev->rf->pcidev, 0)) >> PAGE_SHIFT; switch (entry->mmap_flag) { case IRDMA_MMAP_IO_NC: ret = rdma_user_mmap_io(context, vma, pfn, PAGE_SIZE, pgprot_noncached(vma->vm_page_prot)); break; case IRDMA_MMAP_IO_WC: ret = rdma_user_mmap_io(context, vma, pfn, PAGE_SIZE, pgprot_writecombine(vma->vm_page_prot)); break; default: ret = -EINVAL; } if (ret) irdma_debug(&ucontext->iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "bar_offset [0x%lx] mmap_flag[%d] err[%d]\n", entry->bar_offset, entry->mmap_flag, ret); return ret; } /** * irdma_alloc_push_page - allocate a push page for qp * @iwqp: qp pointer */ static void irdma_alloc_push_page(struct irdma_qp *iwqp) { struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; struct irdma_device *iwdev = iwqp->iwdev; struct irdma_sc_qp *qp = &iwqp->sc_qp; int status; cqp_request = irdma_alloc_and_get_cqp_request(&iwdev->rf->cqp, true); if (!cqp_request) return; cqp_info = &cqp_request->info; cqp_info->cqp_cmd = IRDMA_OP_MANAGE_PUSH_PAGE; cqp_info->post_sq = 1; cqp_info->in.u.manage_push_page.info.push_idx = 0; cqp_info->in.u.manage_push_page.info.qs_handle = qp->vsi->qos[qp->user_pri].qs_handle; cqp_info->in.u.manage_push_page.info.free_page = 0; cqp_info->in.u.manage_push_page.info.push_page_type = 0; cqp_info->in.u.manage_push_page.cqp = &iwdev->rf->cqp.sc_cqp; cqp_info->in.u.manage_push_page.scratch = (uintptr_t)cqp_request; status = irdma_handle_cqp_op(iwdev->rf, cqp_request); if (!status && cqp_request->compl_info.op_ret_val < iwdev->rf->sc_dev.hw_attrs.max_hw_device_pages) { qp->push_idx = cqp_request->compl_info.op_ret_val; qp->push_offset = 0; } irdma_put_cqp_request(&iwdev->rf->cqp, cqp_request); } /** * irdma_get_pbl - Retrieve pbl from a list given a virtual * address * @va: user virtual address * @pbl_list: pbl list to search in (QP's or CQ's) */ struct irdma_pbl * irdma_get_pbl(unsigned long va, struct list_head *pbl_list) { struct irdma_pbl *iwpbl; list_for_each_entry(iwpbl, pbl_list, list) { if (iwpbl->user_base == va) { list_del(&iwpbl->list); iwpbl->on_list = false; return iwpbl; } } return NULL; } /** * irdma_clean_cqes - clean cq entries for qp * @iwqp: qp ptr (user or kernel) * @iwcq: cq ptr */ void irdma_clean_cqes(struct irdma_qp *iwqp, struct irdma_cq *iwcq) { struct irdma_cq_uk *ukcq = &iwcq->sc_cq.cq_uk; unsigned long flags; spin_lock_irqsave(&iwcq->lock, flags); irdma_uk_clean_cq(&iwqp->sc_qp.qp_uk, ukcq); spin_unlock_irqrestore(&iwcq->lock, flags); } static u64 irdma_compute_push_wqe_offset(struct irdma_device *iwdev, u32 page_idx){ u64 bar_off = (uintptr_t)iwdev->rf->sc_dev.hw_regs[IRDMA_DB_ADDR_OFFSET]; if (iwdev->rf->sc_dev.hw_attrs.uk_attrs.hw_rev == IRDMA_GEN_2) { /* skip over db page */ bar_off += IRDMA_HW_PAGE_SIZE; /* skip over reserved space */ bar_off += IRDMA_PF_BAR_RSVD; } /* push wqe page */ bar_off += (u64)page_idx * IRDMA_HW_PAGE_SIZE; return bar_off; } void irdma_remove_push_mmap_entries(struct irdma_qp *iwqp) { if (iwqp->push_db_mmap_entry) { irdma_user_mmap_entry_del_hash(iwqp->push_db_mmap_entry); iwqp->push_db_mmap_entry = NULL; } if (iwqp->push_wqe_mmap_entry) { irdma_user_mmap_entry_del_hash(iwqp->push_wqe_mmap_entry); iwqp->push_wqe_mmap_entry = NULL; } } static int irdma_setup_push_mmap_entries(struct irdma_ucontext *ucontext, struct irdma_qp *iwqp, u64 *push_wqe_mmap_key, u64 *push_db_mmap_key) { struct irdma_device *iwdev = ucontext->iwdev; u64 bar_off; WARN_ON_ONCE(iwdev->rf->sc_dev.hw_attrs.uk_attrs.hw_rev < IRDMA_GEN_2); bar_off = irdma_compute_push_wqe_offset(iwdev, iwqp->sc_qp.push_idx); iwqp->push_wqe_mmap_entry = irdma_user_mmap_entry_add_hash(ucontext, bar_off, IRDMA_MMAP_IO_WC, push_wqe_mmap_key); if (!iwqp->push_wqe_mmap_entry) return -ENOMEM; /* push doorbell page */ bar_off += IRDMA_HW_PAGE_SIZE; iwqp->push_db_mmap_entry = irdma_user_mmap_entry_add_hash(ucontext, bar_off, IRDMA_MMAP_IO_NC, push_db_mmap_key); if (!iwqp->push_db_mmap_entry) { irdma_user_mmap_entry_del_hash(iwqp->push_wqe_mmap_entry); return -ENOMEM; } return 0; } /** * irdma_setup_virt_qp - setup for allocation of virtual qp * @iwdev: irdma device * @iwqp: qp ptr * @init_info: initialize info to return */ void irdma_setup_virt_qp(struct irdma_device *iwdev, struct irdma_qp *iwqp, struct irdma_qp_init_info *init_info) { struct irdma_pbl *iwpbl = iwqp->iwpbl; struct irdma_qp_mr *qpmr = &iwpbl->qp_mr; iwqp->page = qpmr->sq_page; init_info->shadow_area_pa = qpmr->shadow; if (iwpbl->pbl_allocated) { init_info->virtual_map = true; init_info->sq_pa = qpmr->sq_pbl.idx; init_info->rq_pa = qpmr->rq_pbl.idx; } else { init_info->sq_pa = qpmr->sq_pbl.addr; init_info->rq_pa = qpmr->rq_pbl.addr; } } /** * irdma_setup_umode_qp - setup sq and rq size in user mode qp * @udata: user data * @iwdev: iwarp device * @iwqp: qp ptr (user or kernel) * @info: initialize info to return * @init_attr: Initial QP create attributes */ int irdma_setup_umode_qp(struct ib_udata *udata, struct irdma_device *iwdev, struct irdma_qp *iwqp, struct irdma_qp_init_info *info, struct ib_qp_init_attr *init_attr) { struct irdma_ucontext *ucontext = to_ucontext(iwqp->iwpd->ibpd.uobject->context); struct irdma_qp_uk_init_info *ukinfo = &info->qp_uk_init_info; struct irdma_create_qp_req req = {0}; unsigned long flags; int ret; ret = ib_copy_from_udata(&req, udata, min(sizeof(req), udata->inlen)); if (ret) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "ib_copy_from_data fail\n"); return ret; } iwqp->ctx_info.qp_compl_ctx = req.user_compl_ctx; iwqp->user_mode = 1; if (req.user_wqe_bufs) { info->qp_uk_init_info.legacy_mode = ucontext->legacy_mode; spin_lock_irqsave(&ucontext->qp_reg_mem_list_lock, flags); iwqp->iwpbl = irdma_get_pbl((unsigned long)req.user_wqe_bufs, &ucontext->qp_reg_mem_list); spin_unlock_irqrestore(&ucontext->qp_reg_mem_list_lock, flags); if (!iwqp->iwpbl) { ret = -ENODATA; irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "no pbl info\n"); return ret; } } if (!ucontext->use_raw_attrs) { /** * Maintain backward compat with older ABI which passes sq and * rq depth in quanta in cap.max_send_wr and cap.max_recv_wr. * There is no way to compute the correct value of * iwqp->max_send_wr/max_recv_wr in the kernel. */ iwqp->max_send_wr = init_attr->cap.max_send_wr; iwqp->max_recv_wr = init_attr->cap.max_recv_wr; ukinfo->sq_size = init_attr->cap.max_send_wr; ukinfo->rq_size = init_attr->cap.max_recv_wr; irdma_uk_calc_shift_wq(ukinfo, &ukinfo->sq_shift, &ukinfo->rq_shift); } else { ret = irdma_uk_calc_depth_shift_sq(ukinfo, &ukinfo->sq_depth, &ukinfo->sq_shift); if (ret) return ret; ret = irdma_uk_calc_depth_shift_rq(ukinfo, &ukinfo->rq_depth, &ukinfo->rq_shift); if (ret) return ret; iwqp->max_send_wr = (ukinfo->sq_depth - IRDMA_SQ_RSVD) >> ukinfo->sq_shift; iwqp->max_recv_wr = (ukinfo->rq_depth - IRDMA_RQ_RSVD) >> ukinfo->rq_shift; ukinfo->sq_size = ukinfo->sq_depth >> ukinfo->sq_shift; ukinfo->rq_size = ukinfo->rq_depth >> ukinfo->rq_shift; } irdma_setup_virt_qp(iwdev, iwqp, info); return 0; } /** * irdma_setup_kmode_qp - setup initialization for kernel mode qp * @iwdev: iwarp device * @iwqp: qp ptr (user or kernel) * @info: initialize info to return * @init_attr: Initial QP create attributes */ int irdma_setup_kmode_qp(struct irdma_device *iwdev, struct irdma_qp *iwqp, struct irdma_qp_init_info *info, struct ib_qp_init_attr *init_attr) { struct irdma_dma_mem *mem = &iwqp->kqp.dma_mem; u32 size; int status; struct irdma_qp_uk_init_info *ukinfo = &info->qp_uk_init_info; status = irdma_uk_calc_depth_shift_sq(ukinfo, &ukinfo->sq_depth, &ukinfo->sq_shift); if (status) return status; status = irdma_uk_calc_depth_shift_rq(ukinfo, &ukinfo->rq_depth, &ukinfo->rq_shift); if (status) return status; iwqp->kqp.sq_wrid_mem = kcalloc(ukinfo->sq_depth, sizeof(*iwqp->kqp.sq_wrid_mem), GFP_KERNEL); if (!iwqp->kqp.sq_wrid_mem) return -ENOMEM; iwqp->kqp.rq_wrid_mem = kcalloc(ukinfo->rq_depth, sizeof(*iwqp->kqp.rq_wrid_mem), GFP_KERNEL); if (!iwqp->kqp.rq_wrid_mem) { kfree(iwqp->kqp.sq_wrid_mem); iwqp->kqp.sq_wrid_mem = NULL; return -ENOMEM; } iwqp->kqp.sig_trk_mem = kcalloc(ukinfo->sq_depth, sizeof(u32), GFP_KERNEL); memset(iwqp->kqp.sig_trk_mem, 0, ukinfo->sq_depth * sizeof(u32)); if (!iwqp->kqp.sig_trk_mem) { kfree(iwqp->kqp.sq_wrid_mem); iwqp->kqp.sq_wrid_mem = NULL; kfree(iwqp->kqp.rq_wrid_mem); iwqp->kqp.rq_wrid_mem = NULL; return -ENOMEM; } ukinfo->sq_sigwrtrk_array = (void *)iwqp->kqp.sig_trk_mem; ukinfo->sq_wrtrk_array = iwqp->kqp.sq_wrid_mem; ukinfo->rq_wrid_array = iwqp->kqp.rq_wrid_mem; size = (ukinfo->sq_depth + ukinfo->rq_depth) * IRDMA_QP_WQE_MIN_SIZE; size += (IRDMA_SHADOW_AREA_SIZE << 3); mem->size = size; mem->va = irdma_allocate_dma_mem(&iwdev->rf->hw, mem, mem->size, 256); if (!mem->va) { kfree(iwqp->kqp.sq_wrid_mem); iwqp->kqp.sq_wrid_mem = NULL; kfree(iwqp->kqp.rq_wrid_mem); iwqp->kqp.rq_wrid_mem = NULL; return -ENOMEM; } ukinfo->sq = mem->va; info->sq_pa = mem->pa; ukinfo->rq = &ukinfo->sq[ukinfo->sq_depth]; info->rq_pa = info->sq_pa + (ukinfo->sq_depth * IRDMA_QP_WQE_MIN_SIZE); ukinfo->shadow_area = ukinfo->rq[ukinfo->rq_depth].elem; info->shadow_area_pa = info->rq_pa + (ukinfo->rq_depth * IRDMA_QP_WQE_MIN_SIZE); ukinfo->sq_size = ukinfo->sq_depth >> ukinfo->sq_shift; ukinfo->rq_size = ukinfo->rq_depth >> ukinfo->rq_shift; ukinfo->qp_id = iwqp->ibqp.qp_num; iwqp->max_send_wr = (ukinfo->sq_depth - IRDMA_SQ_RSVD) >> ukinfo->sq_shift; iwqp->max_recv_wr = (ukinfo->rq_depth - IRDMA_RQ_RSVD) >> ukinfo->rq_shift; init_attr->cap.max_send_wr = iwqp->max_send_wr; init_attr->cap.max_recv_wr = iwqp->max_recv_wr; return 0; } int irdma_cqp_create_qp_cmd(struct irdma_qp *iwqp) { struct irdma_pci_f *rf = iwqp->iwdev->rf; struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; struct irdma_create_qp_info *qp_info; int status; cqp_request = irdma_alloc_and_get_cqp_request(&rf->cqp, true); if (!cqp_request) return -ENOMEM; cqp_info = &cqp_request->info; qp_info = &cqp_request->info.in.u.qp_create.info; memset(qp_info, 0, sizeof(*qp_info)); qp_info->mac_valid = true; qp_info->cq_num_valid = true; qp_info->next_iwarp_state = IRDMA_QP_STATE_IDLE; cqp_info->cqp_cmd = IRDMA_OP_QP_CREATE; cqp_info->post_sq = 1; cqp_info->in.u.qp_create.qp = &iwqp->sc_qp; cqp_info->in.u.qp_create.scratch = (uintptr_t)cqp_request; status = irdma_handle_cqp_op(rf, cqp_request); irdma_put_cqp_request(&rf->cqp, cqp_request); return status; } void irdma_roce_fill_and_set_qpctx_info(struct irdma_qp *iwqp, struct irdma_qp_host_ctx_info *ctx_info) { struct irdma_device *iwdev = iwqp->iwdev; struct irdma_sc_dev *dev = &iwdev->rf->sc_dev; struct irdma_roce_offload_info *roce_info; struct irdma_udp_offload_info *udp_info; udp_info = &iwqp->udp_info; udp_info->snd_mss = ib_mtu_enum_to_int(ib_mtu_int_to_enum(iwdev->vsi.mtu)); udp_info->cwnd = iwdev->roce_cwnd; udp_info->rexmit_thresh = 2; udp_info->rnr_nak_thresh = 2; udp_info->src_port = 0xc000; udp_info->dst_port = ROCE_V2_UDP_DPORT; roce_info = &iwqp->roce_info; ether_addr_copy(roce_info->mac_addr, if_getlladdr(iwdev->netdev)); roce_info->rd_en = true; roce_info->wr_rdresp_en = true; roce_info->bind_en = true; roce_info->dcqcn_en = false; roce_info->rtomin = iwdev->roce_rtomin; roce_info->ack_credits = iwdev->roce_ackcreds; roce_info->ird_size = dev->hw_attrs.max_hw_ird; roce_info->ord_size = dev->hw_attrs.max_hw_ord; if (!iwqp->user_mode) { roce_info->priv_mode_en = true; roce_info->fast_reg_en = true; roce_info->udprivcq_en = true; } roce_info->roce_tver = 0; ctx_info->roce_info = &iwqp->roce_info; ctx_info->udp_info = &iwqp->udp_info; irdma_sc_qp_setctx_roce(&iwqp->sc_qp, iwqp->host_ctx.va, ctx_info); } void irdma_iw_fill_and_set_qpctx_info(struct irdma_qp *iwqp, struct irdma_qp_host_ctx_info *ctx_info) { struct irdma_device *iwdev = iwqp->iwdev; struct irdma_sc_dev *dev = &iwdev->rf->sc_dev; struct irdma_iwarp_offload_info *iwarp_info; iwarp_info = &iwqp->iwarp_info; ether_addr_copy(iwarp_info->mac_addr, if_getlladdr(iwdev->netdev)); iwarp_info->rd_en = true; iwarp_info->wr_rdresp_en = true; iwarp_info->bind_en = true; iwarp_info->ecn_en = true; iwarp_info->rtomin = 5; if (dev->hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) iwarp_info->ib_rd_en = true; if (!iwqp->user_mode) { iwarp_info->priv_mode_en = true; iwarp_info->fast_reg_en = true; } iwarp_info->ddp_ver = 1; iwarp_info->rdmap_ver = 1; ctx_info->iwarp_info = &iwqp->iwarp_info; ctx_info->iwarp_info_valid = true; irdma_sc_qp_setctx(&iwqp->sc_qp, iwqp->host_ctx.va, ctx_info); ctx_info->iwarp_info_valid = false; } int irdma_validate_qp_attrs(struct ib_qp_init_attr *init_attr, struct irdma_device *iwdev) { struct irdma_sc_dev *dev = &iwdev->rf->sc_dev; struct irdma_uk_attrs *uk_attrs = &dev->hw_attrs.uk_attrs; if (init_attr->create_flags) return -EOPNOTSUPP; if (init_attr->cap.max_inline_data > uk_attrs->max_hw_inline || init_attr->cap.max_send_sge > uk_attrs->max_hw_wq_frags || init_attr->cap.max_recv_sge > uk_attrs->max_hw_wq_frags) return -EINVAL; if (rdma_protocol_roce(&iwdev->ibdev, 1)) { if (init_attr->qp_type != IB_QPT_RC && init_attr->qp_type != IB_QPT_UD && init_attr->qp_type != IB_QPT_GSI) return -EOPNOTSUPP; } else { if (init_attr->qp_type != IB_QPT_RC) return -EOPNOTSUPP; } return 0; } void irdma_sched_qp_flush_work(struct irdma_qp *iwqp) { unsigned long flags; if (iwqp->sc_qp.qp_uk.destroy_pending) return; irdma_qp_add_ref(&iwqp->ibqp); spin_lock_irqsave(&iwqp->dwork_flush_lock, flags); if (mod_delayed_work(iwqp->iwdev->cleanup_wq, &iwqp->dwork_flush, msecs_to_jiffies(IRDMA_FLUSH_DELAY_MS))) irdma_qp_rem_ref(&iwqp->ibqp); spin_unlock_irqrestore(&iwqp->dwork_flush_lock, flags); } void irdma_flush_worker(struct work_struct *work) { struct delayed_work *dwork = to_delayed_work(work); struct irdma_qp *iwqp = container_of(dwork, struct irdma_qp, dwork_flush); irdma_generate_flush_completions(iwqp); /* For the add in irdma_sched_qp_flush_work */ irdma_qp_rem_ref(&iwqp->ibqp); } static int irdma_get_ib_acc_flags(struct irdma_qp *iwqp) { int acc_flags = 0; if (rdma_protocol_roce(iwqp->ibqp.device, 1)) { if (iwqp->roce_info.wr_rdresp_en) { acc_flags |= IB_ACCESS_LOCAL_WRITE; acc_flags |= IB_ACCESS_REMOTE_WRITE; } if (iwqp->roce_info.rd_en) acc_flags |= IB_ACCESS_REMOTE_READ; if (iwqp->roce_info.bind_en) acc_flags |= IB_ACCESS_MW_BIND; } else { if (iwqp->iwarp_info.wr_rdresp_en) { acc_flags |= IB_ACCESS_LOCAL_WRITE; acc_flags |= IB_ACCESS_REMOTE_WRITE; } if (iwqp->iwarp_info.rd_en) acc_flags |= IB_ACCESS_REMOTE_READ; if (iwqp->iwarp_info.bind_en) acc_flags |= IB_ACCESS_MW_BIND; } return acc_flags; } /** * irdma_query_qp - query qp attributes * @ibqp: qp pointer * @attr: attributes pointer * @attr_mask: Not used * @init_attr: qp attributes to return */ static int irdma_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask, struct ib_qp_init_attr *init_attr) { struct irdma_qp *iwqp = to_iwqp(ibqp); struct irdma_sc_qp *qp = &iwqp->sc_qp; memset(attr, 0, sizeof(*attr)); memset(init_attr, 0, sizeof(*init_attr)); attr->qp_state = iwqp->ibqp_state; attr->cur_qp_state = iwqp->ibqp_state; attr->cap.max_send_wr = iwqp->max_send_wr; attr->cap.max_recv_wr = iwqp->max_recv_wr; attr->cap.max_inline_data = qp->qp_uk.max_inline_data; attr->cap.max_send_sge = qp->qp_uk.max_sq_frag_cnt; attr->cap.max_recv_sge = qp->qp_uk.max_rq_frag_cnt; attr->qp_access_flags = irdma_get_ib_acc_flags(iwqp); attr->port_num = 1; if (rdma_protocol_roce(ibqp->device, 1)) { attr->path_mtu = ib_mtu_int_to_enum(iwqp->udp_info.snd_mss); attr->qkey = iwqp->roce_info.qkey; attr->rq_psn = iwqp->udp_info.epsn; attr->sq_psn = iwqp->udp_info.psn_nxt; attr->dest_qp_num = iwqp->roce_info.dest_qp; attr->pkey_index = iwqp->roce_info.p_key; attr->retry_cnt = iwqp->udp_info.rexmit_thresh; attr->rnr_retry = iwqp->udp_info.rnr_nak_thresh; attr->max_rd_atomic = iwqp->roce_info.ord_size; attr->max_dest_rd_atomic = iwqp->roce_info.ird_size; } init_attr->event_handler = iwqp->ibqp.event_handler; init_attr->qp_context = iwqp->ibqp.qp_context; init_attr->send_cq = iwqp->ibqp.send_cq; init_attr->recv_cq = iwqp->ibqp.recv_cq; init_attr->cap = attr->cap; return 0; } static int irdma_wait_for_suspend(struct irdma_qp *iwqp) { if (!wait_event_timeout(iwqp->iwdev->suspend_wq, !iwqp->suspend_pending, msecs_to_jiffies(IRDMA_EVENT_TIMEOUT_MS))) { iwqp->suspend_pending = false; irdma_dev_warn(&iwqp->iwdev->ibdev, "modify_qp timed out waiting for suspend. qp_id = %d, last_ae = 0x%x\n", iwqp->ibqp.qp_num, iwqp->last_aeq); return -EBUSY; } return 0; } /** * irdma_modify_qp_roce - modify qp request * @ibqp: qp's pointer for modify * @attr: access attributes * @attr_mask: state mask * @udata: user data */ int irdma_modify_qp_roce(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask, struct ib_udata *udata) { #define IRDMA_MODIFY_QP_MIN_REQ_LEN offsetofend(struct irdma_modify_qp_req, rq_flush) #define IRDMA_MODIFY_QP_MIN_RESP_LEN offsetofend(struct irdma_modify_qp_resp, push_valid) struct irdma_pd *iwpd = to_iwpd(ibqp->pd); struct irdma_qp *iwqp = to_iwqp(ibqp); struct irdma_device *iwdev = iwqp->iwdev; struct irdma_sc_dev *dev = &iwdev->rf->sc_dev; struct irdma_qp_host_ctx_info *ctx_info; struct irdma_roce_offload_info *roce_info; struct irdma_udp_offload_info *udp_info; struct irdma_modify_qp_info info = {0}; struct irdma_modify_qp_resp uresp = {}; struct irdma_modify_qp_req ureq; unsigned long flags; u8 issue_modify_qp = 0; int ret = 0; ctx_info = &iwqp->ctx_info; roce_info = &iwqp->roce_info; udp_info = &iwqp->udp_info; if (udata) { if ((udata->inlen && udata->inlen < IRDMA_MODIFY_QP_MIN_REQ_LEN) || (udata->outlen && udata->outlen < IRDMA_MODIFY_QP_MIN_RESP_LEN)) return -EINVAL; } if (attr_mask & ~IB_QP_ATTR_STANDARD_BITS) return -EOPNOTSUPP; if (attr_mask & IB_QP_DEST_QPN) roce_info->dest_qp = attr->dest_qp_num; if (attr_mask & IB_QP_PKEY_INDEX) { ret = irdma_query_pkey(ibqp->device, 0, attr->pkey_index, &roce_info->p_key); if (ret) return ret; } if (attr_mask & IB_QP_QKEY) roce_info->qkey = attr->qkey; if (attr_mask & IB_QP_PATH_MTU) udp_info->snd_mss = ib_mtu_enum_to_int(attr->path_mtu); if (attr_mask & IB_QP_SQ_PSN) { udp_info->psn_nxt = attr->sq_psn; udp_info->lsn = 0xffff; udp_info->psn_una = attr->sq_psn; udp_info->psn_max = attr->sq_psn; } if (attr_mask & IB_QP_RQ_PSN) udp_info->epsn = attr->rq_psn; if (attr_mask & IB_QP_RNR_RETRY) udp_info->rnr_nak_thresh = attr->rnr_retry; if (attr_mask & IB_QP_RETRY_CNT) udp_info->rexmit_thresh = attr->retry_cnt; ctx_info->roce_info->pd_id = iwpd->sc_pd.pd_id; if (attr_mask & IB_QP_AV) { struct irdma_av *av = &iwqp->roce_ah.av; u16 vlan_id = VLAN_N_VID; u32 local_ip[4] = {}; memset(&iwqp->roce_ah, 0, sizeof(iwqp->roce_ah)); if (attr->ah_attr.ah_flags & IB_AH_GRH) { udp_info->ttl = attr->ah_attr.grh.hop_limit; udp_info->flow_label = attr->ah_attr.grh.flow_label; udp_info->tos = attr->ah_attr.grh.traffic_class; udp_info->src_port = kc_rdma_get_udp_sport(udp_info->flow_label, ibqp->qp_num, roce_info->dest_qp); irdma_qp_rem_qos(&iwqp->sc_qp); dev->ws_remove(iwqp->sc_qp.vsi, ctx_info->user_pri); if (iwqp->sc_qp.vsi->dscp_mode) ctx_info->user_pri = iwqp->sc_qp.vsi->dscp_map[irdma_tos2dscp(udp_info->tos)]; else ctx_info->user_pri = rt_tos2priority(udp_info->tos); } ret = kc_irdma_set_roce_cm_info(iwqp, attr, &vlan_id); if (ret) return ret; if (dev->ws_add(iwqp->sc_qp.vsi, ctx_info->user_pri)) return -ENOMEM; iwqp->sc_qp.user_pri = ctx_info->user_pri; irdma_qp_add_qos(&iwqp->sc_qp); if (vlan_id >= VLAN_N_VID && iwdev->dcb_vlan_mode) vlan_id = 0; if (vlan_id < VLAN_N_VID) { udp_info->insert_vlan_tag = true; udp_info->vlan_tag = vlan_id | ctx_info->user_pri << VLAN_PRIO_SHIFT; } else { udp_info->insert_vlan_tag = false; } av->attrs = attr->ah_attr; rdma_gid2ip((struct sockaddr *)&av->dgid_addr, &attr->ah_attr.grh.dgid); if (av->net_type == RDMA_NETWORK_IPV6) { __be32 *daddr = av->dgid_addr.saddr_in6.sin6_addr.__u6_addr.__u6_addr32; __be32 *saddr = av->sgid_addr.saddr_in6.sin6_addr.__u6_addr.__u6_addr32; irdma_copy_ip_ntohl(&udp_info->dest_ip_addr[0], daddr); irdma_copy_ip_ntohl(&udp_info->local_ipaddr[0], saddr); udp_info->ipv4 = false; irdma_copy_ip_ntohl(local_ip, daddr); } else if (av->net_type == RDMA_NETWORK_IPV4) { __be32 saddr = av->sgid_addr.saddr_in.sin_addr.s_addr; __be32 daddr = av->dgid_addr.saddr_in.sin_addr.s_addr; local_ip[0] = ntohl(daddr); udp_info->ipv4 = true; udp_info->dest_ip_addr[0] = 0; udp_info->dest_ip_addr[1] = 0; udp_info->dest_ip_addr[2] = 0; udp_info->dest_ip_addr[3] = local_ip[0]; udp_info->local_ipaddr[0] = 0; udp_info->local_ipaddr[1] = 0; udp_info->local_ipaddr[2] = 0; udp_info->local_ipaddr[3] = ntohl(saddr); } else { return -EINVAL; } udp_info->arp_idx = irdma_add_arp(iwdev->rf, local_ip, ah_attr_to_dmac(attr->ah_attr)); } if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC) { if (attr->max_rd_atomic > dev->hw_attrs.max_hw_ord) { irdma_dev_err(&iwdev->ibdev, "rd_atomic = %d, above max_hw_ord=%d\n", attr->max_rd_atomic, dev->hw_attrs.max_hw_ord); return -EINVAL; } if (attr->max_rd_atomic) roce_info->ord_size = attr->max_rd_atomic; info.ord_valid = true; } if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC) { if (attr->max_dest_rd_atomic > dev->hw_attrs.max_hw_ird) { irdma_dev_err(&iwdev->ibdev, "rd_atomic = %d, above max_hw_ird=%d\n", attr->max_rd_atomic, dev->hw_attrs.max_hw_ird); return -EINVAL; } if (attr->max_dest_rd_atomic) roce_info->ird_size = attr->max_dest_rd_atomic; } if (attr_mask & IB_QP_ACCESS_FLAGS) { if (attr->qp_access_flags & IB_ACCESS_LOCAL_WRITE) roce_info->wr_rdresp_en = true; if (attr->qp_access_flags & IB_ACCESS_REMOTE_WRITE) roce_info->wr_rdresp_en = true; if (attr->qp_access_flags & IB_ACCESS_REMOTE_READ) roce_info->rd_en = true; } wait_event(iwqp->mod_qp_waitq, !atomic_read(&iwqp->hw_mod_qp_pend)); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "caller: %pS qp_id=%d to_ibqpstate=%d ibqpstate=%d irdma_qpstate=%d attr_mask=0x%x\n", __builtin_return_address(0), ibqp->qp_num, attr->qp_state, iwqp->ibqp_state, iwqp->iwarp_state, attr_mask); spin_lock_irqsave(&iwqp->lock, flags); if (attr_mask & IB_QP_STATE) { if (!ib_modify_qp_is_ok(iwqp->ibqp_state, attr->qp_state, iwqp->ibqp.qp_type, attr_mask)) { irdma_dev_warn(&iwdev->ibdev, "modify_qp invalid for qp_id=%d, old_state=0x%x, new_state=0x%x\n", iwqp->ibqp.qp_num, iwqp->ibqp_state, attr->qp_state); ret = -EINVAL; goto exit; } info.curr_iwarp_state = iwqp->iwarp_state; switch (attr->qp_state) { case IB_QPS_INIT: if (iwqp->iwarp_state > IRDMA_QP_STATE_IDLE) { ret = -EINVAL; goto exit; } if (iwqp->iwarp_state == IRDMA_QP_STATE_INVALID) { info.next_iwarp_state = IRDMA_QP_STATE_IDLE; issue_modify_qp = 1; } break; case IB_QPS_RTR: if (iwqp->iwarp_state > IRDMA_QP_STATE_IDLE) { ret = -EINVAL; goto exit; } info.arp_cache_idx_valid = true; info.cq_num_valid = true; info.next_iwarp_state = IRDMA_QP_STATE_RTR; issue_modify_qp = 1; break; case IB_QPS_RTS: if (iwqp->ibqp_state < IB_QPS_RTR || iwqp->ibqp_state == IB_QPS_ERR) { ret = -EINVAL; goto exit; } info.arp_cache_idx_valid = true; info.cq_num_valid = true; info.ord_valid = true; info.next_iwarp_state = IRDMA_QP_STATE_RTS; issue_modify_qp = 1; if (dev->hw_attrs.uk_attrs.hw_rev == IRDMA_GEN_2) iwdev->rf->check_fc(&iwdev->vsi, &iwqp->sc_qp); udp_info->cwnd = iwdev->roce_cwnd; roce_info->ack_credits = iwdev->roce_ackcreds; if (iwdev->push_mode && udata && iwqp->sc_qp.push_idx == IRDMA_INVALID_PUSH_PAGE_INDEX && dev->hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) { spin_unlock_irqrestore(&iwqp->lock, flags); irdma_alloc_push_page(iwqp); spin_lock_irqsave(&iwqp->lock, flags); } break; case IB_QPS_SQD: if (iwqp->iwarp_state == IRDMA_QP_STATE_SQD) goto exit; if (iwqp->iwarp_state != IRDMA_QP_STATE_RTS) { ret = -EINVAL; goto exit; } info.next_iwarp_state = IRDMA_QP_STATE_SQD; issue_modify_qp = 1; iwqp->suspend_pending = true; break; case IB_QPS_SQE: case IB_QPS_ERR: case IB_QPS_RESET: if (iwqp->iwarp_state == IRDMA_QP_STATE_ERROR) { spin_unlock_irqrestore(&iwqp->lock, flags); if (udata && udata->inlen) { if (ib_copy_from_udata(&ureq, udata, min(sizeof(ureq), udata->inlen))) return -EINVAL; irdma_flush_wqes(iwqp, (ureq.sq_flush ? IRDMA_FLUSH_SQ : 0) | (ureq.rq_flush ? IRDMA_FLUSH_RQ : 0) | IRDMA_REFLUSH); } return 0; } info.next_iwarp_state = IRDMA_QP_STATE_ERROR; issue_modify_qp = 1; break; default: ret = -EINVAL; goto exit; } iwqp->ibqp_state = attr->qp_state; } ctx_info->send_cq_num = iwqp->iwscq->sc_cq.cq_uk.cq_id; ctx_info->rcv_cq_num = iwqp->iwrcq->sc_cq.cq_uk.cq_id; irdma_sc_qp_setctx_roce(&iwqp->sc_qp, iwqp->host_ctx.va, ctx_info); spin_unlock_irqrestore(&iwqp->lock, flags); if (attr_mask & IB_QP_STATE) { if (issue_modify_qp) { ctx_info->rem_endpoint_idx = udp_info->arp_idx; if (irdma_hw_modify_qp(iwdev, iwqp, &info, true)) return -EINVAL; if (info.next_iwarp_state == IRDMA_QP_STATE_SQD) { ret = irdma_wait_for_suspend(iwqp); if (ret) return ret; } spin_lock_irqsave(&iwqp->lock, flags); if (iwqp->iwarp_state == info.curr_iwarp_state) { iwqp->iwarp_state = info.next_iwarp_state; iwqp->ibqp_state = attr->qp_state; } if (iwqp->ibqp_state > IB_QPS_RTS && !iwqp->flush_issued) { spin_unlock_irqrestore(&iwqp->lock, flags); irdma_flush_wqes(iwqp, IRDMA_FLUSH_SQ | IRDMA_FLUSH_RQ | IRDMA_FLUSH_WAIT); iwqp->flush_issued = 1; } else { spin_unlock_irqrestore(&iwqp->lock, flags); } } else { iwqp->ibqp_state = attr->qp_state; } if (udata && udata->outlen && dev->hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) { struct irdma_ucontext *ucontext; ucontext = to_ucontext(ibqp->uobject->context); if (iwqp->sc_qp.push_idx != IRDMA_INVALID_PUSH_PAGE_INDEX && !iwqp->push_wqe_mmap_entry && !irdma_setup_push_mmap_entries(ucontext, iwqp, &uresp.push_wqe_mmap_key, &uresp.push_db_mmap_key)) { uresp.push_valid = 1; uresp.push_offset = iwqp->sc_qp.push_offset; } uresp.rd_fence_rate = iwdev->rd_fence_rate; ret = ib_copy_to_udata(udata, &uresp, min(sizeof(uresp), udata->outlen)); if (ret) { irdma_remove_push_mmap_entries(iwqp); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "copy_to_udata failed\n"); return ret; } } } return 0; exit: spin_unlock_irqrestore(&iwqp->lock, flags); return ret; } /** * irdma_modify_qp - modify qp request * @ibqp: qp's pointer for modify * @attr: access attributes * @attr_mask: state mask * @udata: user data */ int irdma_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask, struct ib_udata *udata) { #define IRDMA_MODIFY_QP_MIN_REQ_LEN offsetofend(struct irdma_modify_qp_req, rq_flush) #define IRDMA_MODIFY_QP_MIN_RESP_LEN offsetofend(struct irdma_modify_qp_resp, push_valid) struct irdma_qp *iwqp = to_iwqp(ibqp); struct irdma_device *iwdev = iwqp->iwdev; struct irdma_sc_dev *dev = &iwdev->rf->sc_dev; struct irdma_qp_host_ctx_info *ctx_info; struct irdma_tcp_offload_info *tcp_info; struct irdma_iwarp_offload_info *offload_info; struct irdma_modify_qp_info info = {0}; struct irdma_modify_qp_resp uresp = {}; struct irdma_modify_qp_req ureq = {}; u8 issue_modify_qp = 0; u8 dont_wait = 0; int err; unsigned long flags; if (udata) { if ((udata->inlen && udata->inlen < IRDMA_MODIFY_QP_MIN_REQ_LEN) || (udata->outlen && udata->outlen < IRDMA_MODIFY_QP_MIN_RESP_LEN)) return -EINVAL; } if (attr_mask & ~IB_QP_ATTR_STANDARD_BITS) return -EOPNOTSUPP; ctx_info = &iwqp->ctx_info; offload_info = &iwqp->iwarp_info; tcp_info = &iwqp->tcp_info; wait_event(iwqp->mod_qp_waitq, !atomic_read(&iwqp->hw_mod_qp_pend)); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "caller: %pS qp_id=%d to_ibqpstate=%d ibqpstate=%d irdma_qpstate=%d last_aeq=%d hw_tcp_state=%d hw_iwarp_state=%d attr_mask=0x%x\n", __builtin_return_address(0), ibqp->qp_num, attr->qp_state, iwqp->ibqp_state, iwqp->iwarp_state, iwqp->last_aeq, iwqp->hw_tcp_state, iwqp->hw_iwarp_state, attr_mask); spin_lock_irqsave(&iwqp->lock, flags); if (attr_mask & IB_QP_STATE) { info.curr_iwarp_state = iwqp->iwarp_state; switch (attr->qp_state) { case IB_QPS_INIT: case IB_QPS_RTR: if (iwqp->iwarp_state > IRDMA_QP_STATE_IDLE) { err = -EINVAL; goto exit; } if (iwqp->iwarp_state == IRDMA_QP_STATE_INVALID) { info.next_iwarp_state = IRDMA_QP_STATE_IDLE; issue_modify_qp = 1; } if (iwdev->push_mode && udata && iwqp->sc_qp.push_idx == IRDMA_INVALID_PUSH_PAGE_INDEX && dev->hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) { spin_unlock_irqrestore(&iwqp->lock, flags); irdma_alloc_push_page(iwqp); spin_lock_irqsave(&iwqp->lock, flags); } break; case IB_QPS_RTS: if (iwqp->iwarp_state > IRDMA_QP_STATE_RTS || !iwqp->cm_id) { err = -EINVAL; goto exit; } issue_modify_qp = 1; iwqp->hw_tcp_state = IRDMA_TCP_STATE_ESTABLISHED; iwqp->hte_added = 1; info.next_iwarp_state = IRDMA_QP_STATE_RTS; info.tcp_ctx_valid = true; info.ord_valid = true; info.arp_cache_idx_valid = true; info.cq_num_valid = true; break; case IB_QPS_SQD: if (iwqp->hw_iwarp_state > IRDMA_QP_STATE_RTS) { err = 0; goto exit; } if (iwqp->iwarp_state == IRDMA_QP_STATE_CLOSING || iwqp->iwarp_state < IRDMA_QP_STATE_RTS) { err = 0; goto exit; } if (iwqp->iwarp_state > IRDMA_QP_STATE_CLOSING) { err = -EINVAL; goto exit; } info.next_iwarp_state = IRDMA_QP_STATE_CLOSING; issue_modify_qp = 1; break; case IB_QPS_SQE: if (iwqp->iwarp_state >= IRDMA_QP_STATE_TERMINATE) { err = -EINVAL; goto exit; } info.next_iwarp_state = IRDMA_QP_STATE_TERMINATE; issue_modify_qp = 1; break; case IB_QPS_ERR: case IB_QPS_RESET: if (iwqp->iwarp_state == IRDMA_QP_STATE_ERROR) { spin_unlock_irqrestore(&iwqp->lock, flags); if (udata && udata->inlen) { if (ib_copy_from_udata(&ureq, udata, min(sizeof(ureq), udata->inlen))) return -EINVAL; irdma_flush_wqes(iwqp, (ureq.sq_flush ? IRDMA_FLUSH_SQ : 0) | (ureq.rq_flush ? IRDMA_FLUSH_RQ : 0) | IRDMA_REFLUSH); } return 0; } if (iwqp->sc_qp.term_flags) { spin_unlock_irqrestore(&iwqp->lock, flags); irdma_terminate_del_timer(&iwqp->sc_qp); spin_lock_irqsave(&iwqp->lock, flags); } info.next_iwarp_state = IRDMA_QP_STATE_ERROR; if (iwqp->hw_tcp_state > IRDMA_TCP_STATE_CLOSED && iwdev->iw_status && iwqp->hw_tcp_state != IRDMA_TCP_STATE_TIME_WAIT) info.reset_tcp_conn = true; else dont_wait = 1; issue_modify_qp = 1; info.next_iwarp_state = IRDMA_QP_STATE_ERROR; break; default: err = -EINVAL; goto exit; } iwqp->ibqp_state = attr->qp_state; } if (attr_mask & IB_QP_ACCESS_FLAGS) { ctx_info->iwarp_info_valid = true; if (attr->qp_access_flags & IB_ACCESS_LOCAL_WRITE) offload_info->wr_rdresp_en = true; if (attr->qp_access_flags & IB_ACCESS_REMOTE_WRITE) offload_info->wr_rdresp_en = true; if (attr->qp_access_flags & IB_ACCESS_REMOTE_READ) offload_info->rd_en = true; } if (ctx_info->iwarp_info_valid) { ctx_info->send_cq_num = iwqp->iwscq->sc_cq.cq_uk.cq_id; ctx_info->rcv_cq_num = iwqp->iwrcq->sc_cq.cq_uk.cq_id; irdma_sc_qp_setctx(&iwqp->sc_qp, iwqp->host_ctx.va, ctx_info); } spin_unlock_irqrestore(&iwqp->lock, flags); if (attr_mask & IB_QP_STATE) { if (issue_modify_qp) { ctx_info->rem_endpoint_idx = tcp_info->arp_idx; if (irdma_hw_modify_qp(iwdev, iwqp, &info, true)) return -EINVAL; } spin_lock_irqsave(&iwqp->lock, flags); if (iwqp->iwarp_state == info.curr_iwarp_state) { iwqp->iwarp_state = info.next_iwarp_state; iwqp->ibqp_state = attr->qp_state; } spin_unlock_irqrestore(&iwqp->lock, flags); } if (issue_modify_qp && iwqp->ibqp_state > IB_QPS_RTS) { if (dont_wait) { if (iwqp->hw_tcp_state) { spin_lock_irqsave(&iwqp->lock, flags); iwqp->hw_tcp_state = IRDMA_TCP_STATE_CLOSED; iwqp->last_aeq = IRDMA_AE_RESET_SENT; spin_unlock_irqrestore(&iwqp->lock, flags); } irdma_cm_disconn(iwqp); } else { int close_timer_started; spin_lock_irqsave(&iwdev->cm_core.ht_lock, flags); if (iwqp->cm_node) { atomic_inc(&iwqp->cm_node->refcnt); spin_unlock_irqrestore(&iwdev->cm_core.ht_lock, flags); close_timer_started = atomic_inc_return(&iwqp->close_timer_started); if (iwqp->cm_id && close_timer_started == 1) irdma_schedule_cm_timer(iwqp->cm_node, (struct irdma_puda_buf *)iwqp, IRDMA_TIMER_TYPE_CLOSE, 1, 0); irdma_rem_ref_cm_node(iwqp->cm_node); } else { spin_unlock_irqrestore(&iwdev->cm_core.ht_lock, flags); } } } if (attr_mask & IB_QP_STATE && udata && udata->outlen && dev->hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) { struct irdma_ucontext *ucontext; ucontext = to_ucontext(ibqp->uobject->context); if (iwqp->sc_qp.push_idx != IRDMA_INVALID_PUSH_PAGE_INDEX && !iwqp->push_wqe_mmap_entry && !irdma_setup_push_mmap_entries(ucontext, iwqp, &uresp.push_wqe_mmap_key, &uresp.push_db_mmap_key)) { uresp.push_valid = 1; uresp.push_offset = iwqp->sc_qp.push_offset; } uresp.rd_fence_rate = iwdev->rd_fence_rate; err = ib_copy_to_udata(udata, &uresp, min(sizeof(uresp), udata->outlen)); if (err) { irdma_remove_push_mmap_entries(iwqp); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "copy_to_udata failed\n"); return err; } } return 0; exit: spin_unlock_irqrestore(&iwqp->lock, flags); return err; } /** * irdma_cq_free_rsrc - free up resources for cq * @rf: RDMA PCI function * @iwcq: cq ptr */ void irdma_cq_free_rsrc(struct irdma_pci_f *rf, struct irdma_cq *iwcq) { struct irdma_sc_cq *cq = &iwcq->sc_cq; if (!iwcq->user_mode) { irdma_free_dma_mem(rf->sc_dev.hw, &iwcq->kmem); irdma_free_dma_mem(rf->sc_dev.hw, &iwcq->kmem_shadow); } irdma_free_rsrc(rf, rf->allocated_cqs, cq->cq_uk.cq_id); } /** * irdma_free_cqbuf - worker to free a cq buffer * @work: provides access to the cq buffer to free */ static void irdma_free_cqbuf(struct work_struct *work) { struct irdma_cq_buf *cq_buf = container_of(work, struct irdma_cq_buf, work); irdma_free_dma_mem(cq_buf->hw, &cq_buf->kmem_buf); kfree(cq_buf); } /** * irdma_process_resize_list - remove resized cq buffers from the resize_list * @iwcq: cq which owns the resize_list * @iwdev: irdma device * @lcqe_buf: the buffer where the last cqe is received */ int irdma_process_resize_list(struct irdma_cq *iwcq, struct irdma_device *iwdev, struct irdma_cq_buf *lcqe_buf) { struct list_head *tmp_node, *list_node; struct irdma_cq_buf *cq_buf; int cnt = 0; list_for_each_safe(list_node, tmp_node, &iwcq->resize_list) { cq_buf = list_entry(list_node, struct irdma_cq_buf, list); if (cq_buf == lcqe_buf) return cnt; list_del(&cq_buf->list); queue_work(iwdev->cleanup_wq, &cq_buf->work); cnt++; } return cnt; } /** * irdma_resize_cq - resize cq * @ibcq: cq to be resized * @entries: desired cq size * @udata: user data */ static int irdma_resize_cq(struct ib_cq *ibcq, int entries, struct ib_udata *udata) { #define IRDMA_RESIZE_CQ_MIN_REQ_LEN offsetofend(struct irdma_resize_cq_req, user_cq_buffer) struct irdma_cq *iwcq = to_iwcq(ibcq); struct irdma_sc_dev *dev = iwcq->sc_cq.dev; struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; struct irdma_modify_cq_info *m_info; struct irdma_modify_cq_info info = {0}; struct irdma_dma_mem kmem_buf; struct irdma_cq_mr *cqmr_buf; struct irdma_pbl *iwpbl_buf; struct irdma_device *iwdev; struct irdma_pci_f *rf; struct irdma_cq_buf *cq_buf = NULL; unsigned long flags; int ret; iwdev = to_iwdev(ibcq->device); rf = iwdev->rf; if (!(rf->sc_dev.hw_attrs.uk_attrs.feature_flags & IRDMA_FEATURE_CQ_RESIZE)) return -EOPNOTSUPP; if (udata && udata->inlen < IRDMA_RESIZE_CQ_MIN_REQ_LEN) return -EINVAL; if (entries > rf->max_cqe) return -EINVAL; if (!iwcq->user_mode) { entries++; if (rf->sc_dev.hw_attrs.uk_attrs.hw_rev >= IRDMA_GEN_2) entries *= 2; } info.cq_size = max(entries, 4); if (info.cq_size == iwcq->sc_cq.cq_uk.cq_size - 1) return 0; if (udata) { struct irdma_resize_cq_req req = {}; struct irdma_ucontext *ucontext = to_ucontext(ibcq->uobject->context); /* CQ resize not supported with legacy GEN_1 libi40iw */ if (ucontext->legacy_mode) return -EOPNOTSUPP; if (ib_copy_from_udata(&req, udata, min(sizeof(req), udata->inlen))) return -EINVAL; spin_lock_irqsave(&ucontext->cq_reg_mem_list_lock, flags); iwpbl_buf = irdma_get_pbl((unsigned long)req.user_cq_buffer, &ucontext->cq_reg_mem_list); spin_unlock_irqrestore(&ucontext->cq_reg_mem_list_lock, flags); if (!iwpbl_buf) return -ENOMEM; cqmr_buf = &iwpbl_buf->cq_mr; if (iwpbl_buf->pbl_allocated) { info.virtual_map = true; info.pbl_chunk_size = 1; info.first_pm_pbl_idx = cqmr_buf->cq_pbl.idx; } else { info.cq_pa = cqmr_buf->cq_pbl.addr; } } else { /* Kmode CQ resize */ int rsize; rsize = info.cq_size * sizeof(struct irdma_cqe); kmem_buf.size = round_up(rsize, 256); kmem_buf.va = irdma_allocate_dma_mem(dev->hw, &kmem_buf, kmem_buf.size, 256); if (!kmem_buf.va) return -ENOMEM; info.cq_base = kmem_buf.va; info.cq_pa = kmem_buf.pa; cq_buf = kzalloc(sizeof(*cq_buf), GFP_KERNEL); if (!cq_buf) { ret = -ENOMEM; goto error; } } cqp_request = irdma_alloc_and_get_cqp_request(&rf->cqp, true); if (!cqp_request) { ret = -ENOMEM; goto error; } info.shadow_read_threshold = iwcq->sc_cq.shadow_read_threshold; info.cq_resize = true; cqp_info = &cqp_request->info; m_info = &cqp_info->in.u.cq_modify.info; memcpy(m_info, &info, sizeof(*m_info)); cqp_info->cqp_cmd = IRDMA_OP_CQ_MODIFY; cqp_info->in.u.cq_modify.cq = &iwcq->sc_cq; cqp_info->in.u.cq_modify.scratch = (uintptr_t)cqp_request; cqp_info->post_sq = 1; ret = irdma_handle_cqp_op(rf, cqp_request); irdma_put_cqp_request(&rf->cqp, cqp_request); if (ret) goto error; spin_lock_irqsave(&iwcq->lock, flags); if (cq_buf) { cq_buf->kmem_buf = iwcq->kmem; cq_buf->hw = dev->hw; memcpy(&cq_buf->cq_uk, &iwcq->sc_cq.cq_uk, sizeof(cq_buf->cq_uk)); INIT_WORK(&cq_buf->work, irdma_free_cqbuf); list_add_tail(&cq_buf->list, &iwcq->resize_list); iwcq->kmem = kmem_buf; } irdma_sc_cq_resize(&iwcq->sc_cq, &info); ibcq->cqe = info.cq_size - 1; spin_unlock_irqrestore(&iwcq->lock, flags); return 0; error: if (!udata) irdma_free_dma_mem(dev->hw, &kmem_buf); kfree(cq_buf); return ret; } /** * irdma_get_mr_access - get hw MR access permissions from IB access flags * @access: IB access flags */ static inline u16 irdma_get_mr_access(int access){ u16 hw_access = 0; hw_access |= (access & IB_ACCESS_LOCAL_WRITE) ? IRDMA_ACCESS_FLAGS_LOCALWRITE : 0; hw_access |= (access & IB_ACCESS_REMOTE_WRITE) ? IRDMA_ACCESS_FLAGS_REMOTEWRITE : 0; hw_access |= (access & IB_ACCESS_REMOTE_READ) ? IRDMA_ACCESS_FLAGS_REMOTEREAD : 0; hw_access |= (access & IB_ACCESS_MW_BIND) ? IRDMA_ACCESS_FLAGS_BIND_WINDOW : 0; hw_access |= (access & IB_ZERO_BASED) ? IRDMA_ACCESS_FLAGS_ZERO_BASED : 0; hw_access |= IRDMA_ACCESS_FLAGS_LOCALREAD; return hw_access; } /** * irdma_free_stag - free stag resource * @iwdev: irdma device * @stag: stag to free */ void irdma_free_stag(struct irdma_device *iwdev, u32 stag) { u32 stag_idx; stag_idx = (stag & iwdev->rf->mr_stagmask) >> IRDMA_CQPSQ_STAG_IDX_S; irdma_free_rsrc(iwdev->rf, iwdev->rf->allocated_mrs, stag_idx); } /** * irdma_create_stag - create random stag * @iwdev: irdma device */ u32 irdma_create_stag(struct irdma_device *iwdev) { u32 stag; u32 stag_index = 0; u32 next_stag_index; u32 driver_key; u32 random; u8 consumer_key; int ret; get_random_bytes(&random, sizeof(random)); consumer_key = (u8)random; driver_key = random & ~iwdev->rf->mr_stagmask; next_stag_index = (random & iwdev->rf->mr_stagmask) >> 8; next_stag_index %= iwdev->rf->max_mr; ret = irdma_alloc_rsrc(iwdev->rf, iwdev->rf->allocated_mrs, iwdev->rf->max_mr, &stag_index, &next_stag_index); if (ret) return 0; stag = stag_index << IRDMA_CQPSQ_STAG_IDX_S; stag |= driver_key; stag += (u32)consumer_key; return stag; } /** * irdma_check_mem_contiguous - check if pbls stored in arr are contiguous * @arr: lvl1 pbl array * @npages: page count * @pg_size: page size * */ static bool irdma_check_mem_contiguous(u64 *arr, u32 npages, u32 pg_size) { u32 pg_idx; for (pg_idx = 0; pg_idx < npages; pg_idx++) { if ((*arr + (pg_size * pg_idx)) != arr[pg_idx]) return false; } return true; } /** * irdma_check_mr_contiguous - check if MR is physically contiguous * @palloc: pbl allocation struct * @pg_size: page size */ static bool irdma_check_mr_contiguous(struct irdma_pble_alloc *palloc, u32 pg_size) { struct irdma_pble_level2 *lvl2 = &palloc->level2; struct irdma_pble_info *leaf = lvl2->leaf; u64 *arr = NULL; u64 *start_addr = NULL; int i; bool ret; if (palloc->level == PBLE_LEVEL_1) { arr = palloc->level1.addr; ret = irdma_check_mem_contiguous(arr, palloc->total_cnt, pg_size); return ret; } start_addr = leaf->addr; for (i = 0; i < lvl2->leaf_cnt; i++, leaf++) { arr = leaf->addr; if ((*start_addr + (i * pg_size * PBLE_PER_PAGE)) != *arr) return false; ret = irdma_check_mem_contiguous(arr, leaf->cnt, pg_size); if (!ret) return false; } return true; } /** * irdma_setup_pbles - copy user pg address to pble's * @rf: RDMA PCI function * @iwmr: mr pointer for this memory registration * @lvl: requested pble levels */ static int irdma_setup_pbles(struct irdma_pci_f *rf, struct irdma_mr *iwmr, u8 lvl) { struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_pble_alloc *palloc = &iwpbl->pble_alloc; struct irdma_pble_info *pinfo; u64 *pbl; int status; enum irdma_pble_level level = PBLE_LEVEL_1; if (lvl) { status = irdma_get_pble(rf->pble_rsrc, palloc, iwmr->page_cnt, lvl); if (status) return status; iwpbl->pbl_allocated = true; level = palloc->level; pinfo = (level == PBLE_LEVEL_1) ? &palloc->level1 : palloc->level2.leaf; pbl = pinfo->addr; } else { pbl = iwmr->pgaddrmem; } irdma_copy_user_pgaddrs(iwmr, pbl, level); if (lvl) iwmr->pgaddrmem[0] = *pbl; return 0; } /** * irdma_handle_q_mem - handle memory for qp and cq * @iwdev: irdma device * @req: information for q memory management * @iwpbl: pble struct * @lvl: pble level mask */ static int irdma_handle_q_mem(struct irdma_device *iwdev, struct irdma_mem_reg_req *req, struct irdma_pbl *iwpbl, u8 lvl) { struct irdma_pble_alloc *palloc = &iwpbl->pble_alloc; struct irdma_mr *iwmr = iwpbl->iwmr; struct irdma_qp_mr *qpmr = &iwpbl->qp_mr; struct irdma_cq_mr *cqmr = &iwpbl->cq_mr; struct irdma_hmc_pble *hmc_p; u64 *arr = iwmr->pgaddrmem; u32 pg_size, total; int err = 0; bool ret = true; pg_size = iwmr->page_size; err = irdma_setup_pbles(iwdev->rf, iwmr, lvl); if (err) return err; if (lvl) arr = palloc->level1.addr; switch (iwmr->type) { case IRDMA_MEMREG_TYPE_QP: total = req->sq_pages + req->rq_pages; hmc_p = &qpmr->sq_pbl; qpmr->shadow = (dma_addr_t) arr[total]; if (lvl) { ret = irdma_check_mem_contiguous(arr, req->sq_pages, pg_size); if (ret) ret = irdma_check_mem_contiguous(&arr[req->sq_pages], req->rq_pages, pg_size); } if (!ret) { hmc_p->idx = palloc->level1.idx; hmc_p = &qpmr->rq_pbl; hmc_p->idx = palloc->level1.idx + req->sq_pages; } else { hmc_p->addr = arr[0]; hmc_p = &qpmr->rq_pbl; hmc_p->addr = arr[req->sq_pages]; } break; case IRDMA_MEMREG_TYPE_CQ: hmc_p = &cqmr->cq_pbl; if (!cqmr->split) cqmr->shadow = (dma_addr_t) arr[req->cq_pages]; if (lvl) ret = irdma_check_mem_contiguous(arr, req->cq_pages, pg_size); if (!ret) hmc_p->idx = palloc->level1.idx; else hmc_p->addr = arr[0]; break; default: irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "MR type error\n"); err = -EINVAL; } if (lvl && ret) { irdma_free_pble(iwdev->rf->pble_rsrc, palloc); iwpbl->pbl_allocated = false; } return err; } /** * irdma_hw_alloc_mw - create the hw memory window * @iwdev: irdma device * @iwmr: pointer to memory window info */ int irdma_hw_alloc_mw(struct irdma_device *iwdev, struct irdma_mr *iwmr) { struct irdma_mw_alloc_info *info; struct irdma_pd *iwpd = to_iwpd(iwmr->ibmr.pd); struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; int status; cqp_request = irdma_alloc_and_get_cqp_request(&iwdev->rf->cqp, true); if (!cqp_request) return -ENOMEM; cqp_info = &cqp_request->info; info = &cqp_info->in.u.mw_alloc.info; memset(info, 0, sizeof(*info)); if (iwmr->ibmw.type == IB_MW_TYPE_1) info->mw_wide = true; info->page_size = PAGE_SIZE; info->mw_stag_index = iwmr->stag >> IRDMA_CQPSQ_STAG_IDX_S; info->pd_id = iwpd->sc_pd.pd_id; info->remote_access = true; cqp_info->cqp_cmd = IRDMA_OP_MW_ALLOC; cqp_info->post_sq = 1; cqp_info->in.u.mw_alloc.dev = &iwdev->rf->sc_dev; cqp_info->in.u.mw_alloc.scratch = (uintptr_t)cqp_request; status = irdma_handle_cqp_op(iwdev->rf, cqp_request); irdma_put_cqp_request(&iwdev->rf->cqp, cqp_request); return status; } /** * irdma_dealloc_mw - Dealloc memory window * @ibmw: memory window structure. */ static int irdma_dealloc_mw(struct ib_mw *ibmw) { struct ib_pd *ibpd = ibmw->pd; struct irdma_pd *iwpd = to_iwpd(ibpd); struct irdma_mr *iwmr = to_iwmr((struct ib_mr *)ibmw); struct irdma_device *iwdev = to_iwdev(ibmw->device); struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; struct irdma_dealloc_stag_info *info; cqp_request = irdma_alloc_and_get_cqp_request(&iwdev->rf->cqp, true); if (!cqp_request) return -ENOMEM; cqp_info = &cqp_request->info; info = &cqp_info->in.u.dealloc_stag.info; memset(info, 0, sizeof(*info)); info->pd_id = iwpd->sc_pd.pd_id; info->stag_idx = RS_64_1(ibmw->rkey, IRDMA_CQPSQ_STAG_IDX_S); info->mr = false; cqp_info->cqp_cmd = IRDMA_OP_DEALLOC_STAG; cqp_info->post_sq = 1; cqp_info->in.u.dealloc_stag.dev = &iwdev->rf->sc_dev; cqp_info->in.u.dealloc_stag.scratch = (uintptr_t)cqp_request; irdma_handle_cqp_op(iwdev->rf, cqp_request); irdma_put_cqp_request(&iwdev->rf->cqp, cqp_request); irdma_free_stag(iwdev, iwmr->stag); kfree(iwmr); return 0; } /** * irdma_hw_alloc_stag - cqp command to allocate stag * @iwdev: irdma device * @iwmr: irdma mr pointer */ int irdma_hw_alloc_stag(struct irdma_device *iwdev, struct irdma_mr *iwmr) { struct irdma_allocate_stag_info *info; struct ib_pd *pd = iwmr->ibmr.pd; struct irdma_pd *iwpd = to_iwpd(pd); struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; int status; cqp_request = irdma_alloc_and_get_cqp_request(&iwdev->rf->cqp, true); if (!cqp_request) return -ENOMEM; cqp_info = &cqp_request->info; info = &cqp_info->in.u.alloc_stag.info; memset(info, 0, sizeof(*info)); info->page_size = PAGE_SIZE; info->stag_idx = iwmr->stag >> IRDMA_CQPSQ_STAG_IDX_S; info->pd_id = iwpd->sc_pd.pd_id; info->total_len = iwmr->len; info->all_memory = (pd->flags & IB_PD_UNSAFE_GLOBAL_RKEY) ? true : false; info->remote_access = true; cqp_info->cqp_cmd = IRDMA_OP_ALLOC_STAG; cqp_info->post_sq = 1; cqp_info->in.u.alloc_stag.dev = &iwdev->rf->sc_dev; cqp_info->in.u.alloc_stag.scratch = (uintptr_t)cqp_request; status = irdma_handle_cqp_op(iwdev->rf, cqp_request); irdma_put_cqp_request(&iwdev->rf->cqp, cqp_request); if (!status) iwmr->is_hwreg = 1; return status; } /** * irdma_set_page - populate pbl list for fmr * @ibmr: ib mem to access iwarp mr pointer * @addr: page dma address fro pbl list */ static int irdma_set_page(struct ib_mr *ibmr, u64 addr) { struct irdma_mr *iwmr = to_iwmr(ibmr); struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_pble_alloc *palloc = &iwpbl->pble_alloc; u64 *pbl; if (unlikely(iwmr->npages == iwmr->page_cnt)) return -ENOMEM; if (palloc->level == PBLE_LEVEL_2) { struct irdma_pble_info *palloc_info = palloc->level2.leaf + (iwmr->npages >> PBLE_512_SHIFT); palloc_info->addr[iwmr->npages & (PBLE_PER_PAGE - 1)] = addr; } else { pbl = palloc->level1.addr; pbl[iwmr->npages] = addr; } iwmr->npages++; return 0; } /** * irdma_map_mr_sg - map of sg list for fmr * @ibmr: ib mem to access iwarp mr pointer * @sg: scatter gather list * @sg_nents: number of sg pages * @sg_offset: scatter gather list for fmr */ static int irdma_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents, unsigned int *sg_offset) { struct irdma_mr *iwmr = to_iwmr(ibmr); iwmr->npages = 0; return ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, irdma_set_page); } /** * irdma_hwreg_mr - send cqp command for memory registration * @iwdev: irdma device * @iwmr: irdma mr pointer * @access: access for MR */ int irdma_hwreg_mr(struct irdma_device *iwdev, struct irdma_mr *iwmr, u16 access) { struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_reg_ns_stag_info *stag_info; struct ib_pd *pd = iwmr->ibmr.pd; struct irdma_pd *iwpd = to_iwpd(pd); struct irdma_pble_alloc *palloc = &iwpbl->pble_alloc; struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; int ret; cqp_request = irdma_alloc_and_get_cqp_request(&iwdev->rf->cqp, true); if (!cqp_request) return -ENOMEM; cqp_info = &cqp_request->info; stag_info = &cqp_info->in.u.mr_reg_non_shared.info; memset(stag_info, 0, sizeof(*stag_info)); stag_info->va = iwpbl->user_base; stag_info->stag_idx = iwmr->stag >> IRDMA_CQPSQ_STAG_IDX_S; stag_info->stag_key = (u8)iwmr->stag; stag_info->total_len = iwmr->len; stag_info->all_memory = (pd->flags & IB_PD_UNSAFE_GLOBAL_RKEY) ? true : false; stag_info->access_rights = irdma_get_mr_access(access); stag_info->pd_id = iwpd->sc_pd.pd_id; if (stag_info->access_rights & IRDMA_ACCESS_FLAGS_ZERO_BASED) stag_info->addr_type = IRDMA_ADDR_TYPE_ZERO_BASED; else stag_info->addr_type = IRDMA_ADDR_TYPE_VA_BASED; stag_info->page_size = iwmr->page_size; if (iwpbl->pbl_allocated) { if (palloc->level == PBLE_LEVEL_1) { stag_info->first_pm_pbl_index = palloc->level1.idx; stag_info->chunk_size = 1; } else { stag_info->first_pm_pbl_index = palloc->level2.root.idx; stag_info->chunk_size = 3; } } else { stag_info->reg_addr_pa = iwmr->pgaddrmem[0]; } cqp_info->cqp_cmd = IRDMA_OP_MR_REG_NON_SHARED; cqp_info->post_sq = 1; cqp_info->in.u.mr_reg_non_shared.dev = &iwdev->rf->sc_dev; cqp_info->in.u.mr_reg_non_shared.scratch = (uintptr_t)cqp_request; ret = irdma_handle_cqp_op(iwdev->rf, cqp_request); irdma_put_cqp_request(&iwdev->rf->cqp, cqp_request); if (!ret) iwmr->is_hwreg = 1; return ret; } /* * irdma_alloc_iwmr - Allocate iwmr @region - memory region @pd - protection domain @virt - virtual address @reg_type - * registration type */ static struct irdma_mr * irdma_alloc_iwmr(struct ib_umem *region, struct ib_pd *pd, u64 virt, enum irdma_memreg_type reg_type) { struct irdma_pbl *iwpbl; struct irdma_mr *iwmr; iwmr = kzalloc(sizeof(*iwmr), GFP_KERNEL); if (!iwmr) return ERR_PTR(-ENOMEM); iwpbl = &iwmr->iwpbl; iwpbl->iwmr = iwmr; iwmr->region = region; iwmr->ibmr.pd = pd; iwmr->ibmr.device = pd->device; iwmr->ibmr.iova = virt; iwmr->type = reg_type; /* Some OOT versions of irdma_copy_user_pg_addr require the pg mask */ iwmr->page_msk = ~(IRDMA_HW_PAGE_SIZE - 1); iwmr->page_size = IRDMA_HW_PAGE_SIZE; iwmr->len = region->length; iwpbl->user_base = virt; iwmr->page_cnt = irdma_ib_umem_num_dma_blocks(region, iwmr->page_size, virt); return iwmr; } static void irdma_free_iwmr(struct irdma_mr *iwmr) { kfree(iwmr); } /* * irdma_reg_user_mr_type_mem - Handle memory registration @iwmr - irdma mr @access - access rights */ static int irdma_reg_user_mr_type_mem(struct irdma_mr *iwmr, int access) { struct irdma_device *iwdev = to_iwdev(iwmr->ibmr.device); struct irdma_pbl *iwpbl = &iwmr->iwpbl; u32 stag; int err; u8 lvl; lvl = iwmr->page_cnt != 1 ? PBLE_LEVEL_1 | PBLE_LEVEL_2 : PBLE_LEVEL_0; err = irdma_setup_pbles(iwdev->rf, iwmr, lvl); if (err) return err; if (lvl) { err = irdma_check_mr_contiguous(&iwpbl->pble_alloc, iwmr->page_size); if (err) { irdma_free_pble(iwdev->rf->pble_rsrc, &iwpbl->pble_alloc); iwpbl->pbl_allocated = false; } } stag = irdma_create_stag(iwdev); if (!stag) { err = -ENOMEM; goto free_pble; } iwmr->stag = stag; iwmr->ibmr.rkey = stag; iwmr->ibmr.lkey = stag; iwmr->access = access; err = irdma_hwreg_mr(iwdev, iwmr, access); if (err) goto err_hwreg; return 0; err_hwreg: irdma_free_stag(iwdev, stag); free_pble: if (iwpbl->pble_alloc.level != PBLE_LEVEL_0 && iwpbl->pbl_allocated) irdma_free_pble(iwdev->rf->pble_rsrc, &iwpbl->pble_alloc); return err; } /* * irdma_reg_user_mr_type_qp - Handle QP memory registration @req - memory reg req @udata - user info @iwmr - irdma mr */ static int irdma_reg_user_mr_type_qp(struct irdma_mem_reg_req req, struct ib_udata *udata, struct irdma_mr *iwmr) { struct irdma_device *iwdev = to_iwdev(iwmr->ibmr.device); struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_ucontext *ucontext; unsigned long flags; u32 total; int err; u8 lvl; total = req.sq_pages + req.rq_pages + IRDMA_SHADOW_PGCNT; if (total > iwmr->page_cnt) return -EINVAL; total = req.sq_pages + req.rq_pages; lvl = total > 2 ? PBLE_LEVEL_1 : PBLE_LEVEL_0; err = irdma_handle_q_mem(iwdev, &req, iwpbl, lvl); if (err) return err; ucontext = to_ucontext(iwmr->ibmr.pd->uobject->context); spin_lock_irqsave(&ucontext->qp_reg_mem_list_lock, flags); list_add_tail(&iwpbl->list, &ucontext->qp_reg_mem_list); iwpbl->on_list = true; spin_unlock_irqrestore(&ucontext->qp_reg_mem_list_lock, flags); return 0; } /* * irdma_reg_user_mr_type_cq - Handle CQ memory registration @req - memory reg req @udata - user info @iwmr - irdma mr */ static int irdma_reg_user_mr_type_cq(struct irdma_mem_reg_req req, struct ib_udata *udata, struct irdma_mr *iwmr) { struct irdma_device *iwdev = to_iwdev(iwmr->ibmr.device); struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_ucontext *ucontext; unsigned long flags; u32 total; int err; u8 lvl; total = req.cq_pages + ((iwdev->rf->sc_dev.hw_attrs.uk_attrs.feature_flags & IRDMA_FEATURE_CQ_RESIZE) ? 0 : IRDMA_SHADOW_PGCNT); if (total > iwmr->page_cnt) return -EINVAL; lvl = req.cq_pages > 1 ? PBLE_LEVEL_1 : PBLE_LEVEL_0; err = irdma_handle_q_mem(iwdev, &req, iwpbl, lvl); if (err) return err; ucontext = to_ucontext(iwmr->ibmr.pd->uobject->context); spin_lock_irqsave(&ucontext->cq_reg_mem_list_lock, flags); list_add_tail(&iwpbl->list, &ucontext->cq_reg_mem_list); iwpbl->on_list = true; spin_unlock_irqrestore(&ucontext->cq_reg_mem_list_lock, flags); return 0; } /** * irdma_reg_user_mr - Register a user memory region * @pd: ptr of pd * @start: virtual start address * @len: length of mr * @virt: virtual address * @access: access of mr * @udata: user data */ static struct ib_mr * irdma_reg_user_mr(struct ib_pd *pd, u64 start, u64 len, u64 virt, int access, struct ib_udata *udata) { #define IRDMA_MEM_REG_MIN_REQ_LEN offsetofend(struct irdma_mem_reg_req, sq_pages) struct irdma_device *iwdev = to_iwdev(pd->device); struct irdma_mem_reg_req req = {}; struct ib_umem *region; struct irdma_mr *iwmr; int err; if (len > iwdev->rf->sc_dev.hw_attrs.max_mr_size) return ERR_PTR(-EINVAL); if (udata->inlen < IRDMA_MEM_REG_MIN_REQ_LEN) return ERR_PTR(-EINVAL); region = ib_umem_get(pd->uobject->context, start, len, access, 0); if (IS_ERR(region)) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "Failed to create ib_umem region\n"); return (struct ib_mr *)region; } if (ib_copy_from_udata(&req, udata, min(sizeof(req), udata->inlen))) { ib_umem_release(region); return ERR_PTR(-EFAULT); } iwmr = irdma_alloc_iwmr(region, pd, virt, req.reg_type); if (IS_ERR(iwmr)) { ib_umem_release(region); return (struct ib_mr *)iwmr; } switch (req.reg_type) { case IRDMA_MEMREG_TYPE_QP: err = irdma_reg_user_mr_type_qp(req, udata, iwmr); if (err) goto error; break; case IRDMA_MEMREG_TYPE_CQ: err = irdma_reg_user_mr_type_cq(req, udata, iwmr); if (err) goto error; break; case IRDMA_MEMREG_TYPE_MEM: err = irdma_reg_user_mr_type_mem(iwmr, access); if (err) goto error; break; default: err = -EINVAL; goto error; } return &iwmr->ibmr; error: ib_umem_release(region); irdma_free_iwmr(iwmr); return ERR_PTR(err); } int irdma_hwdereg_mr(struct ib_mr *ib_mr) { struct irdma_device *iwdev = to_iwdev(ib_mr->device); struct irdma_mr *iwmr = to_iwmr(ib_mr); struct irdma_pd *iwpd = to_iwpd(ib_mr->pd); struct irdma_dealloc_stag_info *info; struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_cqp_request *cqp_request; struct cqp_cmds_info *cqp_info; int status; /* * Skip HW MR de-register when it is already de-registered during an MR re-reregister and the re-registration * fails */ if (!iwmr->is_hwreg) return 0; cqp_request = irdma_alloc_and_get_cqp_request(&iwdev->rf->cqp, true); if (!cqp_request) return -ENOMEM; cqp_info = &cqp_request->info; info = &cqp_info->in.u.dealloc_stag.info; memset(info, 0, sizeof(*info)); info->pd_id = iwpd->sc_pd.pd_id; info->stag_idx = RS_64_1(ib_mr->rkey, IRDMA_CQPSQ_STAG_IDX_S); info->mr = true; if (iwpbl->pbl_allocated) info->dealloc_pbl = true; cqp_info->cqp_cmd = IRDMA_OP_DEALLOC_STAG; cqp_info->post_sq = 1; cqp_info->in.u.dealloc_stag.dev = &iwdev->rf->sc_dev; cqp_info->in.u.dealloc_stag.scratch = (uintptr_t)cqp_request; status = irdma_handle_cqp_op(iwdev->rf, cqp_request); irdma_put_cqp_request(&iwdev->rf->cqp, cqp_request); if (!status) iwmr->is_hwreg = 0; return status; } /* * irdma_rereg_mr_trans - Re-register a user MR for a change translation. @iwmr: ptr of iwmr @start: virtual start * address @len: length of mr @virt: virtual address * * Re-register a user memory region when a change translation is requested. Re-register a new region while reusing the * stag from the original registration. */ struct ib_mr * irdma_rereg_mr_trans(struct irdma_mr *iwmr, u64 start, u64 len, u64 virt, struct ib_udata *udata) { struct irdma_device *iwdev = to_iwdev(iwmr->ibmr.device); struct irdma_pbl *iwpbl = &iwmr->iwpbl; struct irdma_pble_alloc *palloc = &iwpbl->pble_alloc; struct ib_pd *pd = iwmr->ibmr.pd; struct ib_umem *region; u8 lvl; int err; region = ib_umem_get(pd->uobject->context, start, len, iwmr->access, 0); if (IS_ERR(region)) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "Failed to create ib_umem region\n"); return (struct ib_mr *)region; } iwmr->region = region; iwmr->ibmr.iova = virt; iwmr->ibmr.pd = pd; iwmr->page_size = PAGE_SIZE; iwmr->len = region->length; iwpbl->user_base = virt; iwmr->page_cnt = irdma_ib_umem_num_dma_blocks(region, iwmr->page_size, virt); lvl = iwmr->page_cnt != 1 ? PBLE_LEVEL_1 | PBLE_LEVEL_2 : PBLE_LEVEL_0; err = irdma_setup_pbles(iwdev->rf, iwmr, lvl); if (err) goto error; if (lvl) { err = irdma_check_mr_contiguous(palloc, iwmr->page_size); if (err) { irdma_free_pble(iwdev->rf->pble_rsrc, palloc); iwpbl->pbl_allocated = false; } } err = irdma_hwreg_mr(iwdev, iwmr, iwmr->access); if (err) goto error; return &iwmr->ibmr; error: if (palloc->level != PBLE_LEVEL_0 && iwpbl->pbl_allocated) { irdma_free_pble(iwdev->rf->pble_rsrc, palloc); iwpbl->pbl_allocated = false; } ib_umem_release(region); iwmr->region = NULL; return ERR_PTR(err); } /** * irdma_reg_phys_mr - register kernel physical memory * @pd: ibpd pointer * @addr: physical address of memory to register * @size: size of memory to register * @access: Access rights * @iova_start: start of virtual address for physical buffers */ struct ib_mr * irdma_reg_phys_mr(struct ib_pd *pd, u64 addr, u64 size, int access, u64 *iova_start) { struct irdma_device *iwdev = to_iwdev(pd->device); struct irdma_pbl *iwpbl; struct irdma_mr *iwmr; u32 stag; int ret; iwmr = kzalloc(sizeof(*iwmr), GFP_KERNEL); if (!iwmr) return ERR_PTR(-ENOMEM); iwmr->ibmr.pd = pd; iwmr->ibmr.device = pd->device; iwpbl = &iwmr->iwpbl; iwpbl->iwmr = iwmr; iwmr->type = IRDMA_MEMREG_TYPE_MEM; iwpbl->user_base = *iova_start; stag = irdma_create_stag(iwdev); if (!stag) { ret = -ENOMEM; goto err; } iwmr->stag = stag; iwmr->ibmr.iova = *iova_start; iwmr->ibmr.rkey = stag; iwmr->ibmr.lkey = stag; iwmr->page_cnt = 1; iwmr->pgaddrmem[0] = addr; iwmr->len = size; iwmr->page_size = SZ_4K; ret = irdma_hwreg_mr(iwdev, iwmr, access); if (ret) { irdma_free_stag(iwdev, stag); goto err; } return &iwmr->ibmr; err: kfree(iwmr); return ERR_PTR(ret); } /** * irdma_get_dma_mr - register physical mem * @pd: ptr of pd * @acc: access for memory */ static struct ib_mr * irdma_get_dma_mr(struct ib_pd *pd, int acc) { u64 kva = 0; return irdma_reg_phys_mr(pd, 0, 0, acc, &kva); } /** * irdma_del_memlist - Deleting pbl list entries for CQ/QP * @iwmr: iwmr for IB's user page addresses * @ucontext: ptr to user context */ void irdma_del_memlist(struct irdma_mr *iwmr, struct irdma_ucontext *ucontext) { struct irdma_pbl *iwpbl = &iwmr->iwpbl; unsigned long flags; switch (iwmr->type) { case IRDMA_MEMREG_TYPE_CQ: spin_lock_irqsave(&ucontext->cq_reg_mem_list_lock, flags); if (iwpbl->on_list) { iwpbl->on_list = false; list_del(&iwpbl->list); } spin_unlock_irqrestore(&ucontext->cq_reg_mem_list_lock, flags); break; case IRDMA_MEMREG_TYPE_QP: spin_lock_irqsave(&ucontext->qp_reg_mem_list_lock, flags); if (iwpbl->on_list) { iwpbl->on_list = false; list_del(&iwpbl->list); } spin_unlock_irqrestore(&ucontext->qp_reg_mem_list_lock, flags); break; default: break; } } /** * irdma_copy_sg_list - copy sg list for qp * @sg_list: copied into sg_list * @sgl: copy from sgl * @num_sges: count of sg entries */ static void irdma_copy_sg_list(struct irdma_sge *sg_list, struct ib_sge *sgl, int num_sges) { unsigned int i; for (i = 0; i < num_sges; i++) { sg_list[i].tag_off = sgl[i].addr; sg_list[i].len = sgl[i].length; sg_list[i].stag = sgl[i].lkey; } } /** * irdma_post_send - kernel application wr * @ibqp: qp ptr for wr * @ib_wr: work request ptr * @bad_wr: return of bad wr if err */ static int irdma_post_send(struct ib_qp *ibqp, const struct ib_send_wr *ib_wr, const struct ib_send_wr **bad_wr) { struct irdma_qp *iwqp; struct irdma_qp_uk *ukqp; struct irdma_sc_dev *dev; struct irdma_post_sq_info info; int err = 0; unsigned long flags; bool inv_stag; struct irdma_ah *ah; iwqp = to_iwqp(ibqp); ukqp = &iwqp->sc_qp.qp_uk; dev = &iwqp->iwdev->rf->sc_dev; spin_lock_irqsave(&iwqp->lock, flags); while (ib_wr) { memset(&info, 0, sizeof(info)); inv_stag = false; info.wr_id = (ib_wr->wr_id); if ((ib_wr->send_flags & IB_SEND_SIGNALED) || iwqp->sig_all) info.signaled = true; if (ib_wr->send_flags & IB_SEND_FENCE) info.read_fence = true; switch (ib_wr->opcode) { case IB_WR_SEND_WITH_IMM: if (ukqp->qp_caps & IRDMA_SEND_WITH_IMM) { info.imm_data_valid = true; info.imm_data = ntohl(ib_wr->ex.imm_data); } else { err = -EINVAL; break; } /* fallthrough */ case IB_WR_SEND: case IB_WR_SEND_WITH_INV: if (ib_wr->opcode == IB_WR_SEND || ib_wr->opcode == IB_WR_SEND_WITH_IMM) { if (ib_wr->send_flags & IB_SEND_SOLICITED) info.op_type = IRDMA_OP_TYPE_SEND_SOL; else info.op_type = IRDMA_OP_TYPE_SEND; } else { if (ib_wr->send_flags & IB_SEND_SOLICITED) info.op_type = IRDMA_OP_TYPE_SEND_SOL_INV; else info.op_type = IRDMA_OP_TYPE_SEND_INV; info.stag_to_inv = ib_wr->ex.invalidate_rkey; } info.op.send.num_sges = ib_wr->num_sge; info.op.send.sg_list = (struct irdma_sge *)ib_wr->sg_list; if (iwqp->ibqp.qp_type == IB_QPT_UD || iwqp->ibqp.qp_type == IB_QPT_GSI) { ah = to_iwah(ud_wr(ib_wr)->ah); info.op.send.ah_id = ah->sc_ah.ah_info.ah_idx; info.op.send.qkey = ud_wr(ib_wr)->remote_qkey; info.op.send.dest_qp = ud_wr(ib_wr)->remote_qpn; } if (ib_wr->send_flags & IB_SEND_INLINE) err = irdma_uk_inline_send(ukqp, &info, false); else err = irdma_uk_send(ukqp, &info, false); break; case IB_WR_RDMA_WRITE_WITH_IMM: if (ukqp->qp_caps & IRDMA_WRITE_WITH_IMM) { info.imm_data_valid = true; info.imm_data = ntohl(ib_wr->ex.imm_data); } else { err = -EINVAL; break; } /* fallthrough */ case IB_WR_RDMA_WRITE: if (ib_wr->send_flags & IB_SEND_SOLICITED) info.op_type = IRDMA_OP_TYPE_RDMA_WRITE_SOL; else info.op_type = IRDMA_OP_TYPE_RDMA_WRITE; info.op.rdma_write.num_lo_sges = ib_wr->num_sge; info.op.rdma_write.lo_sg_list = (void *)ib_wr->sg_list; info.op.rdma_write.rem_addr.tag_off = rdma_wr(ib_wr)->remote_addr; info.op.rdma_write.rem_addr.stag = rdma_wr(ib_wr)->rkey; if (ib_wr->send_flags & IB_SEND_INLINE) err = irdma_uk_inline_rdma_write(ukqp, &info, false); else err = irdma_uk_rdma_write(ukqp, &info, false); break; case IB_WR_RDMA_READ_WITH_INV: inv_stag = true; /* fallthrough */ case IB_WR_RDMA_READ: if (ib_wr->num_sge > dev->hw_attrs.uk_attrs.max_hw_read_sges) { err = -EINVAL; break; } info.op_type = IRDMA_OP_TYPE_RDMA_READ; info.op.rdma_read.rem_addr.tag_off = rdma_wr(ib_wr)->remote_addr; info.op.rdma_read.rem_addr.stag = rdma_wr(ib_wr)->rkey; info.op.rdma_read.lo_sg_list = (void *)ib_wr->sg_list; info.op.rdma_read.num_lo_sges = ib_wr->num_sge; err = irdma_uk_rdma_read(ukqp, &info, inv_stag, false); break; case IB_WR_LOCAL_INV: info.op_type = IRDMA_OP_TYPE_INV_STAG; info.local_fence = info.read_fence; info.op.inv_local_stag.target_stag = ib_wr->ex.invalidate_rkey; err = irdma_uk_stag_local_invalidate(ukqp, &info, true); break; case IB_WR_REG_MR:{ struct irdma_mr *iwmr = to_iwmr(reg_wr(ib_wr)->mr); struct irdma_pble_alloc *palloc = &iwmr->iwpbl.pble_alloc; struct irdma_fast_reg_stag_info stag_info = {0}; stag_info.signaled = info.signaled; stag_info.read_fence = info.read_fence; stag_info.access_rights = irdma_get_mr_access(reg_wr(ib_wr)->access); stag_info.stag_key = reg_wr(ib_wr)->key & 0xff; stag_info.stag_idx = reg_wr(ib_wr)->key >> 8; stag_info.page_size = reg_wr(ib_wr)->mr->page_size; stag_info.wr_id = ib_wr->wr_id; stag_info.addr_type = IRDMA_ADDR_TYPE_VA_BASED; stag_info.va = (void *)(uintptr_t)iwmr->ibmr.iova; stag_info.total_len = iwmr->ibmr.length; if (palloc->level == PBLE_LEVEL_2) { stag_info.chunk_size = 3; stag_info.first_pm_pbl_index = palloc->level2.root.idx; } else { stag_info.chunk_size = 1; stag_info.first_pm_pbl_index = palloc->level1.idx; } stag_info.local_fence = ib_wr->send_flags & IB_SEND_FENCE; err = irdma_sc_mr_fast_register(&iwqp->sc_qp, &stag_info, true); break; } default: err = -EINVAL; irdma_debug(&iwqp->iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "upost_send bad opcode = 0x%x\n", ib_wr->opcode); break; } if (err) break; ib_wr = ib_wr->next; } if (!iwqp->flush_issued) { if (iwqp->hw_iwarp_state <= IRDMA_QP_STATE_RTS) irdma_uk_qp_post_wr(ukqp); spin_unlock_irqrestore(&iwqp->lock, flags); } else { spin_unlock_irqrestore(&iwqp->lock, flags); irdma_sched_qp_flush_work(iwqp); } if (err) *bad_wr = ib_wr; return err; } /** * irdma_post_recv - post receive wr for kernel application * @ibqp: ib qp pointer * @ib_wr: work request for receive * @bad_wr: bad wr caused an error */ static int irdma_post_recv(struct ib_qp *ibqp, const struct ib_recv_wr *ib_wr, const struct ib_recv_wr **bad_wr) { struct irdma_qp *iwqp = to_iwqp(ibqp); struct irdma_qp_uk *ukqp = &iwqp->sc_qp.qp_uk; struct irdma_post_rq_info post_recv = {0}; struct irdma_sge *sg_list = iwqp->sg_list; unsigned long flags; int err = 0; spin_lock_irqsave(&iwqp->lock, flags); while (ib_wr) { if (ib_wr->num_sge > ukqp->max_rq_frag_cnt) { err = -EINVAL; goto out; } post_recv.num_sges = ib_wr->num_sge; post_recv.wr_id = ib_wr->wr_id; irdma_copy_sg_list(sg_list, ib_wr->sg_list, ib_wr->num_sge); post_recv.sg_list = sg_list; err = irdma_uk_post_receive(ukqp, &post_recv); if (err) { irdma_debug(&iwqp->iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "post_recv err %d\n", err); goto out; } ib_wr = ib_wr->next; } out: spin_unlock_irqrestore(&iwqp->lock, flags); if (iwqp->flush_issued) irdma_sched_qp_flush_work(iwqp); if (err) *bad_wr = ib_wr; return err; } /** * irdma_flush_err_to_ib_wc_status - return change flush error code to IB status * @opcode: iwarp flush code */ static enum ib_wc_status irdma_flush_err_to_ib_wc_status(enum irdma_flush_opcode opcode) { switch (opcode) { case FLUSH_PROT_ERR: return IB_WC_LOC_PROT_ERR; case FLUSH_REM_ACCESS_ERR: return IB_WC_REM_ACCESS_ERR; case FLUSH_LOC_QP_OP_ERR: return IB_WC_LOC_QP_OP_ERR; case FLUSH_REM_OP_ERR: return IB_WC_REM_OP_ERR; case FLUSH_LOC_LEN_ERR: return IB_WC_LOC_LEN_ERR; case FLUSH_GENERAL_ERR: return IB_WC_WR_FLUSH_ERR; case FLUSH_MW_BIND_ERR: return IB_WC_MW_BIND_ERR; case FLUSH_REM_INV_REQ_ERR: return IB_WC_REM_INV_REQ_ERR; case FLUSH_RETRY_EXC_ERR: return IB_WC_RETRY_EXC_ERR; case FLUSH_FATAL_ERR: default: return IB_WC_FATAL_ERR; } } /** * irdma_process_cqe - process cqe info * @entry: processed cqe * @cq_poll_info: cqe info */ static void irdma_process_cqe(struct ib_wc *entry, struct irdma_cq_poll_info *cq_poll_info) { struct irdma_sc_qp *qp; entry->wc_flags = 0; entry->pkey_index = 0; entry->wr_id = cq_poll_info->wr_id; qp = cq_poll_info->qp_handle; entry->qp = qp->qp_uk.back_qp; if (cq_poll_info->error) { entry->status = (cq_poll_info->comp_status == IRDMA_COMPL_STATUS_FLUSHED) ? irdma_flush_err_to_ib_wc_status(cq_poll_info->minor_err) : IB_WC_GENERAL_ERR; entry->vendor_err = cq_poll_info->major_err << 16 | cq_poll_info->minor_err; } else { entry->status = IB_WC_SUCCESS; if (cq_poll_info->imm_valid) { entry->ex.imm_data = htonl(cq_poll_info->imm_data); entry->wc_flags |= IB_WC_WITH_IMM; } if (cq_poll_info->ud_smac_valid) { ether_addr_copy(entry->smac, cq_poll_info->ud_smac); entry->wc_flags |= IB_WC_WITH_SMAC; } if (cq_poll_info->ud_vlan_valid) { u16 vlan = cq_poll_info->ud_vlan & EVL_VLID_MASK; entry->sl = cq_poll_info->ud_vlan >> VLAN_PRIO_SHIFT; if (vlan) { entry->vlan_id = vlan; entry->wc_flags |= IB_WC_WITH_VLAN; } } else { entry->sl = 0; } } if (cq_poll_info->q_type == IRDMA_CQE_QTYPE_SQ) { set_ib_wc_op_sq(cq_poll_info, entry); } else { set_ib_wc_op_rq(cq_poll_info, entry, qp->qp_uk.qp_caps & IRDMA_SEND_WITH_IMM ? true : false); if (qp->qp_uk.qp_type != IRDMA_QP_TYPE_ROCE_UD && cq_poll_info->stag_invalid_set) { entry->ex.invalidate_rkey = cq_poll_info->inv_stag; entry->wc_flags |= IB_WC_WITH_INVALIDATE; } } if (qp->qp_uk.qp_type == IRDMA_QP_TYPE_ROCE_UD) { entry->src_qp = cq_poll_info->ud_src_qpn; entry->slid = 0; entry->wc_flags |= (IB_WC_GRH | IB_WC_WITH_NETWORK_HDR_TYPE); entry->network_hdr_type = cq_poll_info->ipv4 ? RDMA_NETWORK_IPV4 : RDMA_NETWORK_IPV6; } else { entry->src_qp = cq_poll_info->qp_id; } entry->byte_len = cq_poll_info->bytes_xfered; } /** * irdma_poll_one - poll one entry of the CQ * @ukcq: ukcq to poll * @cur_cqe: current CQE info to be filled in * @entry: ibv_wc object to be filled for non-extended CQ or NULL for extended CQ * * Returns the internal irdma device error code or 0 on success */ static inline int irdma_poll_one(struct irdma_cq_uk *ukcq, struct irdma_cq_poll_info *cur_cqe, struct ib_wc *entry) { int ret = irdma_uk_cq_poll_cmpl(ukcq, cur_cqe); if (ret) return ret; irdma_process_cqe(entry, cur_cqe); return 0; } /** * __irdma_poll_cq - poll cq for completion (kernel apps) * @iwcq: cq to poll * @num_entries: number of entries to poll * @entry: wr of a completed entry */ static int __irdma_poll_cq(struct irdma_cq *iwcq, int num_entries, struct ib_wc *entry) { struct list_head *tmp_node, *list_node; struct irdma_cq_buf *last_buf = NULL; struct irdma_cq_poll_info *cur_cqe = &iwcq->cur_cqe; struct irdma_cq_buf *cq_buf; int ret; struct irdma_device *iwdev; struct irdma_cq_uk *ukcq; bool cq_new_cqe = false; int resized_bufs = 0; int npolled = 0; iwdev = to_iwdev(iwcq->ibcq.device); ukcq = &iwcq->sc_cq.cq_uk; /* go through the list of previously resized CQ buffers */ list_for_each_safe(list_node, tmp_node, &iwcq->resize_list) { cq_buf = container_of(list_node, struct irdma_cq_buf, list); while (npolled < num_entries) { ret = irdma_poll_one(&cq_buf->cq_uk, cur_cqe, entry + npolled); if (!ret) { ++npolled; cq_new_cqe = true; continue; } if (ret == -ENOENT) break; /* QP using the CQ is destroyed. Skip reporting this CQE */ if (ret == -EFAULT) { cq_new_cqe = true; continue; } goto error; } /* save the resized CQ buffer which received the last cqe */ if (cq_new_cqe) last_buf = cq_buf; cq_new_cqe = false; } /* check the current CQ for new cqes */ while (npolled < num_entries) { ret = irdma_poll_one(ukcq, cur_cqe, entry + npolled); if (ret == -ENOENT) { ret = irdma_generated_cmpls(iwcq, cur_cqe); if (!ret) irdma_process_cqe(entry + npolled, cur_cqe); } if (!ret) { ++npolled; cq_new_cqe = true; continue; } if (ret == -ENOENT) break; /* QP using the CQ is destroyed. Skip reporting this CQE */ if (ret == -EFAULT) { cq_new_cqe = true; continue; } goto error; } if (cq_new_cqe) /* all previous CQ resizes are complete */ resized_bufs = irdma_process_resize_list(iwcq, iwdev, NULL); else if (last_buf) /* only CQ resizes up to the last_buf are complete */ resized_bufs = irdma_process_resize_list(iwcq, iwdev, last_buf); if (resized_bufs) /* report to the HW the number of complete CQ resizes */ irdma_uk_cq_set_resized_cnt(ukcq, resized_bufs); return npolled; error: irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "%s: Error polling CQ, irdma_err: %d\n", __func__, ret); return ret; } /** * irdma_poll_cq - poll cq for completion (kernel apps) * @ibcq: cq to poll * @num_entries: number of entries to poll * @entry: wr of a completed entry */ static int irdma_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry) { struct irdma_cq *iwcq; unsigned long flags; int ret; iwcq = to_iwcq(ibcq); spin_lock_irqsave(&iwcq->lock, flags); ret = __irdma_poll_cq(iwcq, num_entries, entry); spin_unlock_irqrestore(&iwcq->lock, flags); return ret; } /** * irdma_req_notify_cq - arm cq kernel application * @ibcq: cq to arm * @notify_flags: notofication flags */ static int irdma_req_notify_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags notify_flags) { struct irdma_cq *iwcq; struct irdma_cq_uk *ukcq; unsigned long flags; enum irdma_cmpl_notify cq_notify = IRDMA_CQ_COMPL_EVENT; bool promo_event = false; int ret = 0; iwcq = to_iwcq(ibcq); ukcq = &iwcq->sc_cq.cq_uk; spin_lock_irqsave(&iwcq->lock, flags); if (notify_flags == IB_CQ_SOLICITED) { cq_notify = IRDMA_CQ_COMPL_SOLICITED; } else { if (iwcq->last_notify == IRDMA_CQ_COMPL_SOLICITED) promo_event = true; } if (!atomic_cmpxchg(&iwcq->armed, 0, 1) || promo_event) { iwcq->last_notify = cq_notify; irdma_uk_cq_request_notification(ukcq, cq_notify); } if ((notify_flags & IB_CQ_REPORT_MISSED_EVENTS) && (!irdma_cq_empty(iwcq) || !list_empty(&iwcq->cmpl_generated))) ret = 1; spin_unlock_irqrestore(&iwcq->lock, flags); return ret; } /** * mcast_list_add - Add a new mcast item to list * @rf: RDMA PCI function * @new_elem: pointer to element to add */ static void mcast_list_add(struct irdma_pci_f *rf, struct mc_table_list *new_elem) { list_add(&new_elem->list, &rf->mc_qht_list.list); } /** * mcast_list_del - Remove an mcast item from list * @mc_qht_elem: pointer to mcast table list element */ static void mcast_list_del(struct mc_table_list *mc_qht_elem) { if (mc_qht_elem) list_del(&mc_qht_elem->list); } /** * mcast_list_lookup_ip - Search mcast list for address * @rf: RDMA PCI function * @ip_mcast: pointer to mcast IP address */ static struct mc_table_list * mcast_list_lookup_ip(struct irdma_pci_f *rf, u32 *ip_mcast) { struct mc_table_list *mc_qht_el; struct list_head *pos, *q; list_for_each_safe(pos, q, &rf->mc_qht_list.list) { mc_qht_el = list_entry(pos, struct mc_table_list, list); if (!memcmp(mc_qht_el->mc_info.dest_ip, ip_mcast, sizeof(mc_qht_el->mc_info.dest_ip))) return mc_qht_el; } return NULL; } /** * irdma_mcast_cqp_op - perform a mcast cqp operation * @iwdev: irdma device * @mc_grp_ctx: mcast group info * @op: operation * * returns error status */ static int irdma_mcast_cqp_op(struct irdma_device *iwdev, struct irdma_mcast_grp_info *mc_grp_ctx, u8 op) { struct cqp_cmds_info *cqp_info; struct irdma_cqp_request *cqp_request; int status; cqp_request = irdma_alloc_and_get_cqp_request(&iwdev->rf->cqp, true); if (!cqp_request) return -ENOMEM; cqp_request->info.in.u.mc_create.info = *mc_grp_ctx; cqp_info = &cqp_request->info; cqp_info->cqp_cmd = op; cqp_info->post_sq = 1; cqp_info->in.u.mc_create.scratch = (uintptr_t)cqp_request; cqp_info->in.u.mc_create.cqp = &iwdev->rf->cqp.sc_cqp; status = irdma_handle_cqp_op(iwdev->rf, cqp_request); irdma_put_cqp_request(&iwdev->rf->cqp, cqp_request); return status; } /** * irdma_attach_mcast - attach a qp to a multicast group * @ibqp: ptr to qp * @ibgid: pointer to global ID * @lid: local ID * * returns error status */ static int irdma_attach_mcast(struct ib_qp *ibqp, union ib_gid *ibgid, u16 lid) { struct irdma_qp *iwqp = to_iwqp(ibqp); struct irdma_device *iwdev = iwqp->iwdev; struct irdma_pci_f *rf = iwdev->rf; struct mc_table_list *mc_qht_elem; struct irdma_mcast_grp_ctx_entry_info mcg_info = {0}; unsigned long flags; u32 ip_addr[4] = {0}; u32 mgn; u32 no_mgs; int ret = 0; bool ipv4; u16 vlan_id; union irdma_sockaddr sgid_addr; unsigned char dmac[ETHER_ADDR_LEN]; rdma_gid2ip((struct sockaddr *)&sgid_addr, ibgid); if (!ipv6_addr_v4mapped((struct in6_addr *)ibgid)) { irdma_copy_ip_ntohl(ip_addr, sgid_addr.saddr_in6.sin6_addr.__u6_addr.__u6_addr32); - irdma_netdev_vlan_ipv6(ip_addr, &vlan_id, NULL); + irdma_netdev_vlan_ipv6(iwqp->cm_id, ip_addr, &vlan_id, NULL); ipv4 = false; irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "qp_id=%d, IP6address=%x:%x:%x:%x\n", ibqp->qp_num, IRDMA_PRINT_IP6(ip_addr)); irdma_mcast_mac_v6(ip_addr, dmac); } else { ip_addr[0] = ntohl(sgid_addr.saddr_in.sin_addr.s_addr); ipv4 = true; - vlan_id = irdma_get_vlan_ipv4(ip_addr); + vlan_id = irdma_get_vlan_ipv4(iwqp->cm_id, ip_addr); irdma_mcast_mac_v4(ip_addr, dmac); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "qp_id=%d, IP4address=%x, MAC=%x:%x:%x:%x:%x:%x\n", ibqp->qp_num, ip_addr[0], dmac[0], dmac[1], dmac[2], dmac[3], dmac[4], dmac[5]); } spin_lock_irqsave(&rf->qh_list_lock, flags); mc_qht_elem = mcast_list_lookup_ip(rf, ip_addr); if (!mc_qht_elem) { struct irdma_dma_mem *dma_mem_mc; spin_unlock_irqrestore(&rf->qh_list_lock, flags); mc_qht_elem = kzalloc(sizeof(*mc_qht_elem), GFP_KERNEL); if (!mc_qht_elem) return -ENOMEM; mc_qht_elem->mc_info.ipv4_valid = ipv4; memcpy(mc_qht_elem->mc_info.dest_ip, ip_addr, sizeof(mc_qht_elem->mc_info.dest_ip)); ret = irdma_alloc_rsrc(rf, rf->allocated_mcgs, rf->max_mcg, &mgn, &rf->next_mcg); if (ret) { kfree(mc_qht_elem); return -ENOMEM; } mc_qht_elem->mc_info.mgn = mgn; dma_mem_mc = &mc_qht_elem->mc_grp_ctx.dma_mem_mc; dma_mem_mc->size = sizeof(u64)* IRDMA_MAX_MGS_PER_CTX; dma_mem_mc->va = irdma_allocate_dma_mem(&rf->hw, dma_mem_mc, dma_mem_mc->size, IRDMA_HW_PAGE_SIZE); if (!dma_mem_mc->va) { irdma_free_rsrc(rf, rf->allocated_mcgs, mgn); kfree(mc_qht_elem); return -ENOMEM; } mc_qht_elem->mc_grp_ctx.mg_id = (u16)mgn; memcpy(mc_qht_elem->mc_grp_ctx.dest_ip_addr, ip_addr, sizeof(mc_qht_elem->mc_grp_ctx.dest_ip_addr)); mc_qht_elem->mc_grp_ctx.ipv4_valid = ipv4; mc_qht_elem->mc_grp_ctx.vlan_id = vlan_id; if (vlan_id < VLAN_N_VID) mc_qht_elem->mc_grp_ctx.vlan_valid = true; mc_qht_elem->mc_grp_ctx.hmc_fcn_id = iwdev->rf->sc_dev.hmc_fn_id; mc_qht_elem->mc_grp_ctx.qs_handle = iwqp->sc_qp.vsi->qos[iwqp->sc_qp.user_pri].qs_handle; ether_addr_copy(mc_qht_elem->mc_grp_ctx.dest_mac_addr, dmac); spin_lock_irqsave(&rf->qh_list_lock, flags); mcast_list_add(rf, mc_qht_elem); } else { if (mc_qht_elem->mc_grp_ctx.no_of_mgs == IRDMA_MAX_MGS_PER_CTX) { spin_unlock_irqrestore(&rf->qh_list_lock, flags); return -ENOMEM; } } mcg_info.qp_id = iwqp->ibqp.qp_num; no_mgs = mc_qht_elem->mc_grp_ctx.no_of_mgs; irdma_sc_add_mcast_grp(&mc_qht_elem->mc_grp_ctx, &mcg_info); spin_unlock_irqrestore(&rf->qh_list_lock, flags); /* Only if there is a change do we need to modify or create */ if (!no_mgs) { ret = irdma_mcast_cqp_op(iwdev, &mc_qht_elem->mc_grp_ctx, IRDMA_OP_MC_CREATE); } else if (no_mgs != mc_qht_elem->mc_grp_ctx.no_of_mgs) { ret = irdma_mcast_cqp_op(iwdev, &mc_qht_elem->mc_grp_ctx, IRDMA_OP_MC_MODIFY); } else { return 0; } if (ret) goto error; return 0; error: irdma_sc_del_mcast_grp(&mc_qht_elem->mc_grp_ctx, &mcg_info); if (!mc_qht_elem->mc_grp_ctx.no_of_mgs) { mcast_list_del(mc_qht_elem); irdma_free_dma_mem(&rf->hw, &mc_qht_elem->mc_grp_ctx.dma_mem_mc); irdma_free_rsrc(rf, rf->allocated_mcgs, mc_qht_elem->mc_grp_ctx.mg_id); kfree(mc_qht_elem); } return ret; } /** * irdma_detach_mcast - detach a qp from a multicast group * @ibqp: ptr to qp * @ibgid: pointer to global ID * @lid: local ID * * returns error status */ static int irdma_detach_mcast(struct ib_qp *ibqp, union ib_gid *ibgid, u16 lid) { struct irdma_qp *iwqp = to_iwqp(ibqp); struct irdma_device *iwdev = iwqp->iwdev; struct irdma_pci_f *rf = iwdev->rf; u32 ip_addr[4] = {0}; struct mc_table_list *mc_qht_elem; struct irdma_mcast_grp_ctx_entry_info mcg_info = {0}; int ret; unsigned long flags; union irdma_sockaddr sgid_addr; rdma_gid2ip((struct sockaddr *)&sgid_addr, ibgid); if (!ipv6_addr_v4mapped((struct in6_addr *)ibgid)) irdma_copy_ip_ntohl(ip_addr, sgid_addr.saddr_in6.sin6_addr.__u6_addr.__u6_addr32); else ip_addr[0] = ntohl(sgid_addr.saddr_in.sin_addr.s_addr); spin_lock_irqsave(&rf->qh_list_lock, flags); mc_qht_elem = mcast_list_lookup_ip(rf, ip_addr); if (!mc_qht_elem) { spin_unlock_irqrestore(&rf->qh_list_lock, flags); irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "address not found MCG\n"); return 0; } mcg_info.qp_id = iwqp->ibqp.qp_num; irdma_sc_del_mcast_grp(&mc_qht_elem->mc_grp_ctx, &mcg_info); if (!mc_qht_elem->mc_grp_ctx.no_of_mgs) { mcast_list_del(mc_qht_elem); spin_unlock_irqrestore(&rf->qh_list_lock, flags); ret = irdma_mcast_cqp_op(iwdev, &mc_qht_elem->mc_grp_ctx, IRDMA_OP_MC_DESTROY); if (ret) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "failed MC_DESTROY MCG\n"); spin_lock_irqsave(&rf->qh_list_lock, flags); mcast_list_add(rf, mc_qht_elem); spin_unlock_irqrestore(&rf->qh_list_lock, flags); return -EAGAIN; } irdma_free_dma_mem(&rf->hw, &mc_qht_elem->mc_grp_ctx.dma_mem_mc); irdma_free_rsrc(rf, rf->allocated_mcgs, mc_qht_elem->mc_grp_ctx.mg_id); kfree(mc_qht_elem); } else { spin_unlock_irqrestore(&rf->qh_list_lock, flags); ret = irdma_mcast_cqp_op(iwdev, &mc_qht_elem->mc_grp_ctx, IRDMA_OP_MC_MODIFY); if (ret) { irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "failed Modify MCG\n"); return ret; } } return 0; } /** * irdma_query_ah - Query address handle * @ibah: pointer to address handle * @ah_attr: address handle attributes */ static int irdma_query_ah(struct ib_ah *ibah, struct ib_ah_attr *ah_attr) { struct irdma_ah *ah = to_iwah(ibah); memset(ah_attr, 0, sizeof(*ah_attr)); if (ah->av.attrs.ah_flags & IB_AH_GRH) { ah_attr->ah_flags = IB_AH_GRH; ah_attr->grh.flow_label = ah->sc_ah.ah_info.flow_label; ah_attr->grh.traffic_class = ah->sc_ah.ah_info.tc_tos; ah_attr->grh.hop_limit = ah->sc_ah.ah_info.hop_ttl; ah_attr->grh.sgid_index = ah->sgid_index; ah_attr->grh.sgid_index = ah->sgid_index; memcpy(&ah_attr->grh.dgid, &ah->dgid, sizeof(ah_attr->grh.dgid)); } return 0; } static if_t irdma_get_netdev(struct ib_device *ibdev, u8 port_num) { struct irdma_device *iwdev = to_iwdev(ibdev); if (iwdev->netdev) { dev_hold(iwdev->netdev); return iwdev->netdev; } return NULL; } static void irdma_set_device_ops(struct ib_device *ibdev) { struct ib_device *dev_ops = ibdev; dev_ops->alloc_hw_stats = irdma_alloc_hw_stats; dev_ops->alloc_mr = irdma_alloc_mr; dev_ops->alloc_mw = irdma_alloc_mw; dev_ops->alloc_pd = irdma_alloc_pd; dev_ops->alloc_ucontext = irdma_alloc_ucontext; dev_ops->create_cq = irdma_create_cq; dev_ops->create_qp = irdma_create_qp; dev_ops->dealloc_mw = irdma_dealloc_mw; dev_ops->dealloc_pd = irdma_dealloc_pd; dev_ops->dealloc_ucontext = irdma_dealloc_ucontext; dev_ops->dereg_mr = irdma_dereg_mr; dev_ops->destroy_cq = irdma_destroy_cq; dev_ops->destroy_qp = irdma_destroy_qp; dev_ops->disassociate_ucontext = irdma_disassociate_ucontext; dev_ops->get_dev_fw_str = irdma_get_dev_fw_str; dev_ops->get_dma_mr = irdma_get_dma_mr; dev_ops->get_hw_stats = irdma_get_hw_stats; dev_ops->get_netdev = irdma_get_netdev; dev_ops->map_mr_sg = irdma_map_mr_sg; dev_ops->mmap = irdma_mmap; dev_ops->poll_cq = irdma_poll_cq; dev_ops->post_recv = irdma_post_recv; dev_ops->post_send = irdma_post_send; dev_ops->query_device = irdma_query_device; dev_ops->query_port = irdma_query_port; dev_ops->modify_port = irdma_modify_port; dev_ops->query_qp = irdma_query_qp; dev_ops->reg_user_mr = irdma_reg_user_mr; dev_ops->rereg_user_mr = irdma_rereg_user_mr; dev_ops->req_notify_cq = irdma_req_notify_cq; dev_ops->resize_cq = irdma_resize_cq; } static void irdma_set_device_mcast_ops(struct ib_device *ibdev) { struct ib_device *dev_ops = ibdev; dev_ops->attach_mcast = irdma_attach_mcast; dev_ops->detach_mcast = irdma_detach_mcast; } static void irdma_set_device_roce_ops(struct ib_device *ibdev) { struct ib_device *dev_ops = ibdev; dev_ops->create_ah = irdma_create_ah; dev_ops->destroy_ah = irdma_destroy_ah; dev_ops->get_link_layer = irdma_get_link_layer; dev_ops->get_port_immutable = irdma_roce_port_immutable; dev_ops->modify_qp = irdma_modify_qp_roce; dev_ops->query_ah = irdma_query_ah; dev_ops->query_gid = irdma_query_gid_roce; dev_ops->query_pkey = irdma_query_pkey; ibdev->add_gid = irdma_add_gid; ibdev->del_gid = irdma_del_gid; } static void irdma_set_device_iw_ops(struct ib_device *ibdev) { struct ib_device *dev_ops = ibdev; ibdev->uverbs_cmd_mask |= (1ull << IB_USER_VERBS_CMD_CREATE_AH) | (1ull << IB_USER_VERBS_CMD_DESTROY_AH); dev_ops->create_ah = irdma_create_ah_stub; dev_ops->destroy_ah = irdma_destroy_ah_stub; dev_ops->get_port_immutable = irdma_iw_port_immutable; dev_ops->modify_qp = irdma_modify_qp; dev_ops->query_gid = irdma_query_gid; dev_ops->query_pkey = irdma_iw_query_pkey; } static inline void irdma_set_device_gen1_ops(struct ib_device *ibdev) { } /** * irdma_init_roce_device - initialization of roce rdma device * @iwdev: irdma device */ static void irdma_init_roce_device(struct irdma_device *iwdev) { kc_set_roce_uverbs_cmd_mask(iwdev); iwdev->ibdev.node_type = RDMA_NODE_IB_CA; addrconf_addr_eui48((u8 *)&iwdev->ibdev.node_guid, if_getlladdr(iwdev->netdev)); irdma_set_device_roce_ops(&iwdev->ibdev); if (iwdev->rf->rdma_ver == IRDMA_GEN_2) irdma_set_device_mcast_ops(&iwdev->ibdev); } /** * irdma_init_iw_device - initialization of iwarp rdma device * @iwdev: irdma device */ static int irdma_init_iw_device(struct irdma_device *iwdev) { if_t netdev = iwdev->netdev; iwdev->ibdev.node_type = RDMA_NODE_RNIC; addrconf_addr_eui48((u8 *)&iwdev->ibdev.node_guid, if_getlladdr(netdev)); iwdev->ibdev.iwcm = kzalloc(sizeof(*iwdev->ibdev.iwcm), GFP_KERNEL); if (!iwdev->ibdev.iwcm) return -ENOMEM; iwdev->ibdev.iwcm->add_ref = irdma_qp_add_ref; iwdev->ibdev.iwcm->rem_ref = irdma_qp_rem_ref; iwdev->ibdev.iwcm->get_qp = irdma_get_qp; iwdev->ibdev.iwcm->connect = irdma_connect; iwdev->ibdev.iwcm->accept = irdma_accept; iwdev->ibdev.iwcm->reject = irdma_reject; iwdev->ibdev.iwcm->create_listen = irdma_create_listen; iwdev->ibdev.iwcm->destroy_listen = irdma_destroy_listen; memcpy(iwdev->ibdev.iwcm->ifname, if_name(netdev), sizeof(iwdev->ibdev.iwcm->ifname)); irdma_set_device_iw_ops(&iwdev->ibdev); return 0; } /** * irdma_init_rdma_device - initialization of rdma device * @iwdev: irdma device */ static int irdma_init_rdma_device(struct irdma_device *iwdev) { int ret; iwdev->ibdev.owner = THIS_MODULE; iwdev->ibdev.uverbs_abi_ver = IRDMA_ABI_VER; kc_set_rdma_uverbs_cmd_mask(iwdev); if (iwdev->roce_mode) { irdma_init_roce_device(iwdev); } else { ret = irdma_init_iw_device(iwdev); if (ret) return ret; } iwdev->ibdev.phys_port_cnt = 1; iwdev->ibdev.num_comp_vectors = iwdev->rf->ceqs_count; iwdev->ibdev.dev.parent = iwdev->rf->dev_ctx.dev; set_ibdev_dma_device(iwdev->ibdev, &iwdev->rf->pcidev->dev); irdma_set_device_ops(&iwdev->ibdev); if (iwdev->rf->rdma_ver == IRDMA_GEN_1) irdma_set_device_gen1_ops(&iwdev->ibdev); return 0; } /** * irdma_port_ibevent - indicate port event * @iwdev: irdma device */ void irdma_port_ibevent(struct irdma_device *iwdev) { struct ib_event event; event.device = &iwdev->ibdev; event.element.port_num = 1; event.event = iwdev->iw_status ? IB_EVENT_PORT_ACTIVE : IB_EVENT_PORT_ERR; ib_dispatch_event(&event); } /** * irdma_ib_unregister_device - unregister rdma device from IB * core * @iwdev: irdma device */ void irdma_ib_unregister_device(struct irdma_device *iwdev) { iwdev->iw_status = 0; irdma_port_ibevent(iwdev); ib_unregister_device(&iwdev->ibdev); dev_put(iwdev->netdev); kfree(iwdev->ibdev.iwcm); iwdev->ibdev.iwcm = NULL; } /** * irdma_ib_register_device - register irdma device to IB core * @iwdev: irdma device */ int irdma_ib_register_device(struct irdma_device *iwdev) { int ret; ret = irdma_init_rdma_device(iwdev); if (ret) return ret; dev_hold(iwdev->netdev); sprintf(iwdev->ibdev.name, "irdma-%s", if_name(iwdev->netdev)); ret = ib_register_device(&iwdev->ibdev, NULL); if (ret) goto error; iwdev->iw_status = 1; irdma_port_ibevent(iwdev); return 0; error: kfree(iwdev->ibdev.iwcm); iwdev->ibdev.iwcm = NULL; irdma_debug(&iwdev->rf->sc_dev, IRDMA_DEBUG_VERBS, "Register RDMA device fail\n"); return ret; }