^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) * Copyright (c) 2006, 2019 Oracle and/or its affiliates. All rights reserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * This software is available to you under a choice of one of two
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * licenses. You may choose to be licensed under the terms of the GNU
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) * General Public License (GPL) Version 2, available from the file
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * COPYING in the main directory of this source tree, or the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) * OpenIB.org BSD license below:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) * Redistribution and use in source and binary forms, with or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) * without modification, are permitted provided that the following
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) * conditions are met:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) * - Redistributions of source code must retain the above
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) * copyright notice, this list of conditions and the following
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) * disclaimer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) * - Redistributions in binary form must reproduce the above
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) * copyright notice, this list of conditions and the following
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) * disclaimer in the documentation and/or other materials
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) * provided with the distribution.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) * SOFTWARE.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) #include <linux/kernel.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) #include <linux/in.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) #include <linux/if.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) #include <linux/netdevice.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) #include <linux/inetdevice.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) #include <linux/if_arp.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) #include <linux/delay.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) #include <linux/module.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) #include <net/addrconf.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) #include "rds_single_path.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) #include "rds.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) #include "ib.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) #include "ib_mr.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) static unsigned int rds_ib_mr_1m_pool_size = RDS_MR_1M_POOL_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) static unsigned int rds_ib_mr_8k_pool_size = RDS_MR_8K_POOL_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) unsigned int rds_ib_retry_count = RDS_IB_DEFAULT_RETRY_COUNT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) static atomic_t rds_ib_unloading;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) module_param(rds_ib_mr_1m_pool_size, int, 0444);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) MODULE_PARM_DESC(rds_ib_mr_1m_pool_size, " Max number of 1M mr per HCA");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) module_param(rds_ib_mr_8k_pool_size, int, 0444);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) MODULE_PARM_DESC(rds_ib_mr_8k_pool_size, " Max number of 8K mr per HCA");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) module_param(rds_ib_retry_count, int, 0444);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) MODULE_PARM_DESC(rds_ib_retry_count, " Number of hw retries before reporting an error");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) * we have a clumsy combination of RCU and a rwsem protecting this list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) * because it is used both in the get_mr fast path and while blocking in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) * the FMR flushing path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) DECLARE_RWSEM(rds_ib_devices_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) struct list_head rds_ib_devices;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) /* NOTE: if also grabbing ibdev lock, grab this first */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) DEFINE_SPINLOCK(ib_nodev_conns_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) LIST_HEAD(ib_nodev_conns);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) static void rds_ib_nodev_connect(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) struct rds_ib_connection *ic;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) spin_lock(&ib_nodev_conns_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) list_for_each_entry(ic, &ib_nodev_conns, ib_node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) rds_conn_connect_if_down(ic->conn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) spin_unlock(&ib_nodev_conns_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) static void rds_ib_dev_shutdown(struct rds_ib_device *rds_ibdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) struct rds_ib_connection *ic;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) unsigned long flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) spin_lock_irqsave(&rds_ibdev->spinlock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) list_for_each_entry(ic, &rds_ibdev->conn_list, ib_node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) rds_conn_path_drop(&ic->conn->c_path[0], true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) spin_unlock_irqrestore(&rds_ibdev->spinlock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) * rds_ib_destroy_mr_pool() blocks on a few things and mrs drop references
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) * from interrupt context so we push freing off into a work struct in krdsd.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) static void rds_ib_dev_free(struct work_struct *work)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) struct rds_ib_ipaddr *i_ipaddr, *i_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) struct rds_ib_device *rds_ibdev = container_of(work,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) struct rds_ib_device, free_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) if (rds_ibdev->mr_8k_pool)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) rds_ib_destroy_mr_pool(rds_ibdev->mr_8k_pool);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) if (rds_ibdev->mr_1m_pool)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) rds_ib_destroy_mr_pool(rds_ibdev->mr_1m_pool);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) if (rds_ibdev->pd)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) ib_dealloc_pd(rds_ibdev->pd);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) list_for_each_entry_safe(i_ipaddr, i_next, &rds_ibdev->ipaddr_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) list_del(&i_ipaddr->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) kfree(i_ipaddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) kfree(rds_ibdev->vector_load);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) kfree(rds_ibdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) void rds_ib_dev_put(struct rds_ib_device *rds_ibdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) BUG_ON(refcount_read(&rds_ibdev->refcount) == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) if (refcount_dec_and_test(&rds_ibdev->refcount))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) queue_work(rds_wq, &rds_ibdev->free_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) static int rds_ib_add_one(struct ib_device *device)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) struct rds_ib_device *rds_ibdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) /* Only handle IB (no iWARP) devices */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) if (device->node_type != RDMA_NODE_IB_CA)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) return -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) /* Device must support FRWR */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) if (!(device->attrs.device_cap_flags & IB_DEVICE_MEM_MGT_EXTENSIONS))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) return -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) rds_ibdev = kzalloc_node(sizeof(struct rds_ib_device), GFP_KERNEL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) ibdev_to_node(device));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) if (!rds_ibdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) spin_lock_init(&rds_ibdev->spinlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) refcount_set(&rds_ibdev->refcount, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) INIT_WORK(&rds_ibdev->free_work, rds_ib_dev_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) INIT_LIST_HEAD(&rds_ibdev->ipaddr_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) INIT_LIST_HEAD(&rds_ibdev->conn_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) rds_ibdev->max_wrs = device->attrs.max_qp_wr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) rds_ibdev->max_sge = min(device->attrs.max_send_sge, RDS_IB_MAX_SGE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) rds_ibdev->odp_capable =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) !!(device->attrs.device_cap_flags &
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) IB_DEVICE_ON_DEMAND_PAGING) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) !!(device->attrs.odp_caps.per_transport_caps.rc_odp_caps &
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) IB_ODP_SUPPORT_WRITE) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) !!(device->attrs.odp_caps.per_transport_caps.rc_odp_caps &
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) IB_ODP_SUPPORT_READ);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) rds_ibdev->max_1m_mrs = device->attrs.max_mr ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) min_t(unsigned int, (device->attrs.max_mr / 2),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) rds_ib_mr_1m_pool_size) : rds_ib_mr_1m_pool_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) rds_ibdev->max_8k_mrs = device->attrs.max_mr ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) min_t(unsigned int, ((device->attrs.max_mr / 2) * RDS_MR_8K_SCALE),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) rds_ib_mr_8k_pool_size) : rds_ib_mr_8k_pool_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) rds_ibdev->max_initiator_depth = device->attrs.max_qp_init_rd_atom;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) rds_ibdev->max_responder_resources = device->attrs.max_qp_rd_atom;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) rds_ibdev->vector_load = kcalloc(device->num_comp_vectors,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) sizeof(int),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) if (!rds_ibdev->vector_load) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) pr_err("RDS/IB: %s failed to allocate vector memory\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) __func__);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) ret = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) goto put_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) rds_ibdev->dev = device;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) rds_ibdev->pd = ib_alloc_pd(device, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) if (IS_ERR(rds_ibdev->pd)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) ret = PTR_ERR(rds_ibdev->pd);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) rds_ibdev->pd = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) goto put_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) rds_ibdev->mr_1m_pool =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) rds_ib_create_mr_pool(rds_ibdev, RDS_IB_MR_1M_POOL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) if (IS_ERR(rds_ibdev->mr_1m_pool)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) ret = PTR_ERR(rds_ibdev->mr_1m_pool);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) rds_ibdev->mr_1m_pool = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) goto put_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) rds_ibdev->mr_8k_pool =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) rds_ib_create_mr_pool(rds_ibdev, RDS_IB_MR_8K_POOL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) if (IS_ERR(rds_ibdev->mr_8k_pool)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) ret = PTR_ERR(rds_ibdev->mr_8k_pool);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) rds_ibdev->mr_8k_pool = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) goto put_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) rdsdebug("RDS/IB: max_mr = %d, max_wrs = %d, max_sge = %d, max_1m_mrs = %d, max_8k_mrs = %d\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) device->attrs.max_mr, rds_ibdev->max_wrs, rds_ibdev->max_sge,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) rds_ibdev->max_1m_mrs, rds_ibdev->max_8k_mrs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) pr_info("RDS/IB: %s: added\n", device->name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) down_write(&rds_ib_devices_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) list_add_tail_rcu(&rds_ibdev->list, &rds_ib_devices);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) up_write(&rds_ib_devices_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) refcount_inc(&rds_ibdev->refcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) ib_set_client_data(device, &rds_ib_client, rds_ibdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) rds_ib_nodev_connect();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) put_dev:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) rds_ib_dev_put(rds_ibdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) * New connections use this to find the device to associate with the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) * connection. It's not in the fast path so we're not concerned about the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) * performance of the IB call. (As of this writing, it uses an interrupt
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) * blocking spinlock to serialize walking a per-device list of all registered
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) * clients.)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) * RCU is used to handle incoming connections racing with device teardown.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) * Rather than use a lock to serialize removal from the client_data and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) * getting a new reference, we use an RCU grace period. The destruction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) * path removes the device from client_data and then waits for all RCU
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) * readers to finish.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) * A new connection can get NULL from this if its arriving on a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) * device that is in the process of being removed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) struct rds_ib_device *rds_ib_get_client_data(struct ib_device *device)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) struct rds_ib_device *rds_ibdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) rds_ibdev = ib_get_client_data(device, &rds_ib_client);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) if (rds_ibdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) refcount_inc(&rds_ibdev->refcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) return rds_ibdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) * The IB stack is letting us know that a device is going away. This can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) * happen if the underlying HCA driver is removed or if PCI hotplug is removing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) * the pci function, for example.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) * This can be called at any time and can be racing with any other RDS path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) static void rds_ib_remove_one(struct ib_device *device, void *client_data)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) struct rds_ib_device *rds_ibdev = client_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) rds_ib_dev_shutdown(rds_ibdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) /* stop connection attempts from getting a reference to this device. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) ib_set_client_data(device, &rds_ib_client, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) down_write(&rds_ib_devices_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) list_del_rcu(&rds_ibdev->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) up_write(&rds_ib_devices_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) * This synchronize rcu is waiting for readers of both the ib
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) * client data and the devices list to finish before we drop
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) * both of those references.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) synchronize_rcu();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) rds_ib_dev_put(rds_ibdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) rds_ib_dev_put(rds_ibdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) struct ib_client rds_ib_client = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) .name = "rds_ib",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) .add = rds_ib_add_one,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) .remove = rds_ib_remove_one
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) static int rds_ib_conn_info_visitor(struct rds_connection *conn,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) void *buffer)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) struct rds_info_rdma_connection *iinfo = buffer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) struct rds_ib_connection *ic = conn->c_transport_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) /* We will only ever look at IB transports */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) if (conn->c_trans != &rds_ib_transport)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) if (conn->c_isv6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) iinfo->src_addr = conn->c_laddr.s6_addr32[3];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) iinfo->dst_addr = conn->c_faddr.s6_addr32[3];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) if (ic) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) iinfo->tos = conn->c_tos;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) iinfo->sl = ic->i_sl;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) memset(&iinfo->src_gid, 0, sizeof(iinfo->src_gid));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) memset(&iinfo->dst_gid, 0, sizeof(iinfo->dst_gid));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) if (rds_conn_state(conn) == RDS_CONN_UP) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) struct rds_ib_device *rds_ibdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) rdma_read_gids(ic->i_cm_id, (union ib_gid *)&iinfo->src_gid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) (union ib_gid *)&iinfo->dst_gid);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) rds_ibdev = ic->rds_ibdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) iinfo->max_send_wr = ic->i_send_ring.w_nr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) iinfo->max_recv_wr = ic->i_recv_ring.w_nr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) iinfo->max_send_sge = rds_ibdev->max_sge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) rds_ib_get_mr_info(rds_ibdev, iinfo);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) iinfo->cache_allocs = atomic_read(&ic->i_cache_allocs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) /* IPv6 version of rds_ib_conn_info_visitor(). */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) static int rds6_ib_conn_info_visitor(struct rds_connection *conn,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) void *buffer)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) struct rds6_info_rdma_connection *iinfo6 = buffer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) struct rds_ib_connection *ic = conn->c_transport_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) /* We will only ever look at IB transports */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) if (conn->c_trans != &rds_ib_transport)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) iinfo6->src_addr = conn->c_laddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) iinfo6->dst_addr = conn->c_faddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) if (ic) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) iinfo6->tos = conn->c_tos;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) iinfo6->sl = ic->i_sl;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) memset(&iinfo6->src_gid, 0, sizeof(iinfo6->src_gid));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) memset(&iinfo6->dst_gid, 0, sizeof(iinfo6->dst_gid));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) if (rds_conn_state(conn) == RDS_CONN_UP) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) struct rds_ib_device *rds_ibdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) rdma_read_gids(ic->i_cm_id, (union ib_gid *)&iinfo6->src_gid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) (union ib_gid *)&iinfo6->dst_gid);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) rds_ibdev = ic->rds_ibdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) iinfo6->max_send_wr = ic->i_send_ring.w_nr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) iinfo6->max_recv_wr = ic->i_recv_ring.w_nr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) iinfo6->max_send_sge = rds_ibdev->max_sge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) rds6_ib_get_mr_info(rds_ibdev, iinfo6);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) iinfo6->cache_allocs = atomic_read(&ic->i_cache_allocs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) static void rds_ib_ic_info(struct socket *sock, unsigned int len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) struct rds_info_iterator *iter,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) struct rds_info_lengths *lens)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) u64 buffer[(sizeof(struct rds_info_rdma_connection) + 7) / 8];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) rds_for_each_conn_info(sock, len, iter, lens,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) rds_ib_conn_info_visitor,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) buffer,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) sizeof(struct rds_info_rdma_connection));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) /* IPv6 version of rds_ib_ic_info(). */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) static void rds6_ib_ic_info(struct socket *sock, unsigned int len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) struct rds_info_iterator *iter,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) struct rds_info_lengths *lens)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) u64 buffer[(sizeof(struct rds6_info_rdma_connection) + 7) / 8];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) rds_for_each_conn_info(sock, len, iter, lens,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) rds6_ib_conn_info_visitor,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) buffer,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) sizeof(struct rds6_info_rdma_connection));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) * Early RDS/IB was built to only bind to an address if there is an IPoIB
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) * device with that address set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) * If it were me, I'd advocate for something more flexible. Sending and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) * receiving should be device-agnostic. Transports would try and maintain
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) * connections between peers who have messages queued. Userspace would be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) * allowed to influence which paths have priority. We could call userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) * asserting this policy "routing".
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) static int rds_ib_laddr_check(struct net *net, const struct in6_addr *addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) __u32 scope_id)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) struct rdma_cm_id *cm_id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) struct sockaddr_in6 sin6;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) struct sockaddr_in sin;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) struct sockaddr *sa;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) bool isv4;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) isv4 = ipv6_addr_v4mapped(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) /* Create a CMA ID and try to bind it. This catches both
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) * IB and iWARP capable NICs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) cm_id = rdma_create_id(&init_net, rds_rdma_cm_event_handler,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) NULL, RDMA_PS_TCP, IB_QPT_RC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) if (IS_ERR(cm_id))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) return PTR_ERR(cm_id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) if (isv4) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) memset(&sin, 0, sizeof(sin));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) sin.sin_family = AF_INET;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) sin.sin_addr.s_addr = addr->s6_addr32[3];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) sa = (struct sockaddr *)&sin;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) memset(&sin6, 0, sizeof(sin6));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) sin6.sin6_family = AF_INET6;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) sin6.sin6_addr = *addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) sin6.sin6_scope_id = scope_id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) sa = (struct sockaddr *)&sin6;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) /* XXX Do a special IPv6 link local address check here. The
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) * reason is that rdma_bind_addr() always succeeds with IPv6
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) * link local address regardless it is indeed configured in a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) * system.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) if (ipv6_addr_type(addr) & IPV6_ADDR_LINKLOCAL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) struct net_device *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) if (scope_id == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) ret = -EADDRNOTAVAIL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) /* Use init_net for now as RDS is not network
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) * name space aware.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) dev = dev_get_by_index(&init_net, scope_id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) if (!dev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) ret = -EADDRNOTAVAIL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) if (!ipv6_chk_addr(&init_net, addr, dev, 1)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) dev_put(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) ret = -EADDRNOTAVAIL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) dev_put(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) ret = -EADDRNOTAVAIL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) /* rdma_bind_addr will only succeed for IB & iWARP devices */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) ret = rdma_bind_addr(cm_id, sa);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) /* due to this, we will claim to support iWARP devices unless we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) check node_type. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) if (ret || !cm_id->device ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) cm_id->device->node_type != RDMA_NODE_IB_CA)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) ret = -EADDRNOTAVAIL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) rdsdebug("addr %pI6c%%%u ret %d node type %d\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) addr, scope_id, ret,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) cm_id->device ? cm_id->device->node_type : -1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) rdma_destroy_id(cm_id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) static void rds_ib_unregister_client(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) ib_unregister_client(&rds_ib_client);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) /* wait for rds_ib_dev_free() to complete */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) flush_workqueue(rds_wq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) static void rds_ib_set_unloading(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) atomic_set(&rds_ib_unloading, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) static bool rds_ib_is_unloading(struct rds_connection *conn)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) struct rds_conn_path *cp = &conn->c_path[0];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) return (test_bit(RDS_DESTROY_PENDING, &cp->cp_flags) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) atomic_read(&rds_ib_unloading) != 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) void rds_ib_exit(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) rds_ib_set_unloading();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) synchronize_rcu();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) rds_info_deregister_func(RDS_INFO_IB_CONNECTIONS, rds_ib_ic_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) rds_info_deregister_func(RDS6_INFO_IB_CONNECTIONS, rds6_ib_ic_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) rds_ib_unregister_client();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) rds_ib_destroy_nodev_conns();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) rds_ib_sysctl_exit();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) rds_ib_recv_exit();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) rds_trans_unregister(&rds_ib_transport);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) rds_ib_mr_exit();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) static u8 rds_ib_get_tos_map(u8 tos)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) /* 1:1 user to transport map for RDMA transport.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) * In future, if custom map is desired, hook can export
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) * user configurable map.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) return tos;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) struct rds_transport rds_ib_transport = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) .laddr_check = rds_ib_laddr_check,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) .xmit_path_complete = rds_ib_xmit_path_complete,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) .xmit = rds_ib_xmit,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) .xmit_rdma = rds_ib_xmit_rdma,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) .xmit_atomic = rds_ib_xmit_atomic,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) .recv_path = rds_ib_recv_path,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) .conn_alloc = rds_ib_conn_alloc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) .conn_free = rds_ib_conn_free,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) .conn_path_connect = rds_ib_conn_path_connect,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) .conn_path_shutdown = rds_ib_conn_path_shutdown,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) .inc_copy_to_user = rds_ib_inc_copy_to_user,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) .inc_free = rds_ib_inc_free,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) .cm_initiate_connect = rds_ib_cm_initiate_connect,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) .cm_handle_connect = rds_ib_cm_handle_connect,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) .cm_connect_complete = rds_ib_cm_connect_complete,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) .stats_info_copy = rds_ib_stats_info_copy,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) .exit = rds_ib_exit,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) .get_mr = rds_ib_get_mr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) .sync_mr = rds_ib_sync_mr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) .free_mr = rds_ib_free_mr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) .flush_mrs = rds_ib_flush_mrs,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) .get_tos_map = rds_ib_get_tos_map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) .t_owner = THIS_MODULE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) .t_name = "infiniband",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) .t_unloading = rds_ib_is_unloading,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) .t_type = RDS_TRANS_IB
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) int rds_ib_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) INIT_LIST_HEAD(&rds_ib_devices);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) ret = rds_ib_mr_init();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) ret = ib_register_client(&rds_ib_client);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) goto out_mr_exit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) ret = rds_ib_sysctl_init();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) goto out_ibreg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) ret = rds_ib_recv_init();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) goto out_sysctl;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) rds_trans_register(&rds_ib_transport);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) rds_info_register_func(RDS_INFO_IB_CONNECTIONS, rds_ib_ic_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) rds_info_register_func(RDS6_INFO_IB_CONNECTIONS, rds6_ib_ic_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) out_sysctl:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) rds_ib_sysctl_exit();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) out_ibreg:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) rds_ib_unregister_client();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) out_mr_exit:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) rds_ib_mr_exit();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) MODULE_LICENSE("GPL");
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags