diff options
Diffstat (limited to 'drivers/infiniband/core')
39 files changed, 17310 insertions, 2968 deletions
diff --git a/drivers/infiniband/core/Makefile b/drivers/infiniband/core/Makefile index ec3353f24b2..ffd0af6734a 100644 --- a/drivers/infiniband/core/Makefile +++ b/drivers/infiniband/core/Makefile @@ -1,19 +1,33 @@ +infiniband-$(CONFIG_INFINIBAND_ADDR_TRANS) := rdma_cm.o +user_access-$(CONFIG_INFINIBAND_ADDR_TRANS) := rdma_ucm.o + obj-$(CONFIG_INFINIBAND) += ib_core.o ib_mad.o ib_sa.o \ - ib_cm.o + ib_cm.o iw_cm.o ib_addr.o \ + $(infiniband-y) obj-$(CONFIG_INFINIBAND_USER_MAD) += ib_umad.o -obj-$(CONFIG_INFINIBAND_USER_ACCESS) += ib_uverbs.o ib_ucm.o +obj-$(CONFIG_INFINIBAND_USER_ACCESS) += ib_uverbs.o ib_ucm.o \ + $(user_access-y) ib_core-y := packer.o ud_header.o verbs.o sysfs.o \ - device.o fmr_pool.o cache.o + device.o fmr_pool.o cache.o netlink.o +ib_core-$(CONFIG_INFINIBAND_USER_MEM) += umem.o ib_mad-y := mad.o smi.o agent.o mad_rmpp.o -ib_sa-y := sa_query.o +ib_sa-y := sa_query.o multicast.o ib_cm-y := cm.o +iw_cm-y := iwcm.o iwpm_util.o iwpm_msg.o + +rdma_cm-y := cma.o + +rdma_ucm-y := ucma.o + +ib_addr-y := addr.o + ib_umad-y := user_mad.o ib_ucm-y := ucm.o -ib_uverbs-y := uverbs_main.o uverbs_cmd.o uverbs_mem.o +ib_uverbs-y := uverbs_main.o uverbs_cmd.o uverbs_marshall.o diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c new file mode 100644 index 00000000000..8172d37f9ad --- /dev/null +++ b/drivers/infiniband/core/addr.c @@ -0,0 +1,565 @@ +/* + * Copyright (c) 2005 Voltaire Inc. All rights reserved. + * Copyright (c) 2002-2005, Network Appliance, Inc. All rights reserved. + * Copyright (c) 1999-2005, Mellanox Technologies, Inc. All rights reserved. + * Copyright (c) 2005 Intel Corporation. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/mutex.h> +#include <linux/inetdevice.h> +#include <linux/slab.h> +#include <linux/workqueue.h> +#include <linux/module.h> +#include <net/arp.h> +#include <net/neighbour.h> +#include <net/route.h> +#include <net/netevent.h> +#include <net/addrconf.h> +#include <net/ip6_route.h> +#include <rdma/ib_addr.h> +#include <rdma/ib.h> + +MODULE_AUTHOR("Sean Hefty"); +MODULE_DESCRIPTION("IB Address Translation"); +MODULE_LICENSE("Dual BSD/GPL"); + +struct addr_req { + struct list_head list; + struct sockaddr_storage src_addr; + struct sockaddr_storage dst_addr; + struct rdma_dev_addr *addr; + struct rdma_addr_client *client; + void *context; + void (*callback)(int status, struct sockaddr *src_addr, + struct rdma_dev_addr *addr, void *context); + unsigned long timeout; + int status; +}; + +static void process_req(struct work_struct *work); + +static DEFINE_MUTEX(lock); +static LIST_HEAD(req_list); +static DECLARE_DELAYED_WORK(work, process_req); +static struct workqueue_struct *addr_wq; + +int rdma_addr_size(struct sockaddr *addr) +{ + switch (addr->sa_family) { + case AF_INET: + return sizeof(struct sockaddr_in); + case AF_INET6: + return sizeof(struct sockaddr_in6); + case AF_IB: + return sizeof(struct sockaddr_ib); + default: + return 0; + } +} +EXPORT_SYMBOL(rdma_addr_size); + +static struct rdma_addr_client self; + +void rdma_addr_register_client(struct rdma_addr_client *client) +{ + atomic_set(&client->refcount, 1); + init_completion(&client->comp); +} +EXPORT_SYMBOL(rdma_addr_register_client); + +static inline void put_client(struct rdma_addr_client *client) +{ + if (atomic_dec_and_test(&client->refcount)) + complete(&client->comp); +} + +void rdma_addr_unregister_client(struct rdma_addr_client *client) +{ + put_client(client); + wait_for_completion(&client->comp); +} +EXPORT_SYMBOL(rdma_addr_unregister_client); + +int rdma_copy_addr(struct rdma_dev_addr *dev_addr, struct net_device *dev, + const unsigned char *dst_dev_addr) +{ + dev_addr->dev_type = dev->type; + memcpy(dev_addr->src_dev_addr, dev->dev_addr, MAX_ADDR_LEN); + memcpy(dev_addr->broadcast, dev->broadcast, MAX_ADDR_LEN); + if (dst_dev_addr) + memcpy(dev_addr->dst_dev_addr, dst_dev_addr, MAX_ADDR_LEN); + dev_addr->bound_dev_if = dev->ifindex; + return 0; +} +EXPORT_SYMBOL(rdma_copy_addr); + +int rdma_translate_ip(struct sockaddr *addr, struct rdma_dev_addr *dev_addr, + u16 *vlan_id) +{ + struct net_device *dev; + int ret = -EADDRNOTAVAIL; + + if (dev_addr->bound_dev_if) { + dev = dev_get_by_index(&init_net, dev_addr->bound_dev_if); + if (!dev) + return -ENODEV; + ret = rdma_copy_addr(dev_addr, dev, NULL); + dev_put(dev); + return ret; + } + + switch (addr->sa_family) { + case AF_INET: + dev = ip_dev_find(&init_net, + ((struct sockaddr_in *) addr)->sin_addr.s_addr); + + if (!dev) + return ret; + + ret = rdma_copy_addr(dev_addr, dev, NULL); + if (vlan_id) + *vlan_id = rdma_vlan_dev_vlan_id(dev); + dev_put(dev); + break; + +#if IS_ENABLED(CONFIG_IPV6) + case AF_INET6: + rcu_read_lock(); + for_each_netdev_rcu(&init_net, dev) { + if (ipv6_chk_addr(&init_net, + &((struct sockaddr_in6 *) addr)->sin6_addr, + dev, 1)) { + ret = rdma_copy_addr(dev_addr, dev, NULL); + if (vlan_id) + *vlan_id = rdma_vlan_dev_vlan_id(dev); + break; + } + } + rcu_read_unlock(); + break; +#endif + } + return ret; +} +EXPORT_SYMBOL(rdma_translate_ip); + +static void set_timeout(unsigned long time) +{ + unsigned long delay; + + delay = time - jiffies; + if ((long)delay <= 0) + delay = 1; + + mod_delayed_work(addr_wq, &work, delay); +} + +static void queue_req(struct addr_req *req) +{ + struct addr_req *temp_req; + + mutex_lock(&lock); + list_for_each_entry_reverse(temp_req, &req_list, list) { + if (time_after_eq(req->timeout, temp_req->timeout)) + break; + } + + list_add(&req->list, &temp_req->list); + + if (req_list.next == &req->list) + set_timeout(req->timeout); + mutex_unlock(&lock); +} + +static int dst_fetch_ha(struct dst_entry *dst, struct rdma_dev_addr *dev_addr, void *daddr) +{ + struct neighbour *n; + int ret; + + n = dst_neigh_lookup(dst, daddr); + + rcu_read_lock(); + if (!n || !(n->nud_state & NUD_VALID)) { + if (n) + neigh_event_send(n, NULL); + ret = -ENODATA; + } else { + ret = rdma_copy_addr(dev_addr, dst->dev, n->ha); + } + rcu_read_unlock(); + + if (n) + neigh_release(n); + + return ret; +} + +static int addr4_resolve(struct sockaddr_in *src_in, + struct sockaddr_in *dst_in, + struct rdma_dev_addr *addr) +{ + __be32 src_ip = src_in->sin_addr.s_addr; + __be32 dst_ip = dst_in->sin_addr.s_addr; + struct rtable *rt; + struct flowi4 fl4; + int ret; + + memset(&fl4, 0, sizeof(fl4)); + fl4.daddr = dst_ip; + fl4.saddr = src_ip; + fl4.flowi4_oif = addr->bound_dev_if; + rt = ip_route_output_key(&init_net, &fl4); + if (IS_ERR(rt)) { + ret = PTR_ERR(rt); + goto out; + } + src_in->sin_family = AF_INET; + src_in->sin_addr.s_addr = fl4.saddr; + + if (rt->dst.dev->flags & IFF_LOOPBACK) { + ret = rdma_translate_ip((struct sockaddr *)dst_in, addr, NULL); + if (!ret) + memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN); + goto put; + } + + /* If the device does ARP internally, return 'done' */ + if (rt->dst.dev->flags & IFF_NOARP) { + ret = rdma_copy_addr(addr, rt->dst.dev, NULL); + goto put; + } + + ret = dst_fetch_ha(&rt->dst, addr, &fl4.daddr); +put: + ip_rt_put(rt); +out: + return ret; +} + +#if IS_ENABLED(CONFIG_IPV6) +static int addr6_resolve(struct sockaddr_in6 *src_in, + struct sockaddr_in6 *dst_in, + struct rdma_dev_addr *addr) +{ + struct flowi6 fl6; + struct dst_entry *dst; + int ret; + + memset(&fl6, 0, sizeof fl6); + fl6.daddr = dst_in->sin6_addr; + fl6.saddr = src_in->sin6_addr; + fl6.flowi6_oif = addr->bound_dev_if; + + dst = ip6_route_output(&init_net, NULL, &fl6); + if ((ret = dst->error)) + goto put; + + if (ipv6_addr_any(&fl6.saddr)) { + ret = ipv6_dev_get_saddr(&init_net, ip6_dst_idev(dst)->dev, + &fl6.daddr, 0, &fl6.saddr); + if (ret) + goto put; + + src_in->sin6_family = AF_INET6; + src_in->sin6_addr = fl6.saddr; + } + + if (dst->dev->flags & IFF_LOOPBACK) { + ret = rdma_translate_ip((struct sockaddr *)dst_in, addr, NULL); + if (!ret) + memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN); + goto put; + } + + /* If the device does ARP internally, return 'done' */ + if (dst->dev->flags & IFF_NOARP) { + ret = rdma_copy_addr(addr, dst->dev, NULL); + goto put; + } + + ret = dst_fetch_ha(dst, addr, &fl6.daddr); +put: + dst_release(dst); + return ret; +} +#else +static int addr6_resolve(struct sockaddr_in6 *src_in, + struct sockaddr_in6 *dst_in, + struct rdma_dev_addr *addr) +{ + return -EADDRNOTAVAIL; +} +#endif + +static int addr_resolve(struct sockaddr *src_in, + struct sockaddr *dst_in, + struct rdma_dev_addr *addr) +{ + if (src_in->sa_family == AF_INET) { + return addr4_resolve((struct sockaddr_in *) src_in, + (struct sockaddr_in *) dst_in, addr); + } else + return addr6_resolve((struct sockaddr_in6 *) src_in, + (struct sockaddr_in6 *) dst_in, addr); +} + +static void process_req(struct work_struct *work) +{ + struct addr_req *req, *temp_req; + struct sockaddr *src_in, *dst_in; + struct list_head done_list; + + INIT_LIST_HEAD(&done_list); + + mutex_lock(&lock); + list_for_each_entry_safe(req, temp_req, &req_list, list) { + if (req->status == -ENODATA) { + src_in = (struct sockaddr *) &req->src_addr; + dst_in = (struct sockaddr *) &req->dst_addr; + req->status = addr_resolve(src_in, dst_in, req->addr); + if (req->status && time_after_eq(jiffies, req->timeout)) + req->status = -ETIMEDOUT; + else if (req->status == -ENODATA) + continue; + } + list_move_tail(&req->list, &done_list); + } + + if (!list_empty(&req_list)) { + req = list_entry(req_list.next, struct addr_req, list); + set_timeout(req->timeout); + } + mutex_unlock(&lock); + + list_for_each_entry_safe(req, temp_req, &done_list, list) { + list_del(&req->list); + req->callback(req->status, (struct sockaddr *) &req->src_addr, + req->addr, req->context); + put_client(req->client); + kfree(req); + } +} + +int rdma_resolve_ip(struct rdma_addr_client *client, + struct sockaddr *src_addr, struct sockaddr *dst_addr, + struct rdma_dev_addr *addr, int timeout_ms, + void (*callback)(int status, struct sockaddr *src_addr, + struct rdma_dev_addr *addr, void *context), + void *context) +{ + struct sockaddr *src_in, *dst_in; + struct addr_req *req; + int ret = 0; + + req = kzalloc(sizeof *req, GFP_KERNEL); + if (!req) + return -ENOMEM; + + src_in = (struct sockaddr *) &req->src_addr; + dst_in = (struct sockaddr *) &req->dst_addr; + + if (src_addr) { + if (src_addr->sa_family != dst_addr->sa_family) { + ret = -EINVAL; + goto err; + } + + memcpy(src_in, src_addr, rdma_addr_size(src_addr)); + } else { + src_in->sa_family = dst_addr->sa_family; + } + + memcpy(dst_in, dst_addr, rdma_addr_size(dst_addr)); + req->addr = addr; + req->callback = callback; + req->context = context; + req->client = client; + atomic_inc(&client->refcount); + + req->status = addr_resolve(src_in, dst_in, addr); + switch (req->status) { + case 0: + req->timeout = jiffies; + queue_req(req); + break; + case -ENODATA: + req->timeout = msecs_to_jiffies(timeout_ms) + jiffies; + queue_req(req); + break; + default: + ret = req->status; + atomic_dec(&client->refcount); + goto err; + } + return ret; +err: + kfree(req); + return ret; +} +EXPORT_SYMBOL(rdma_resolve_ip); + +void rdma_addr_cancel(struct rdma_dev_addr *addr) +{ + struct addr_req *req, *temp_req; + + mutex_lock(&lock); + list_for_each_entry_safe(req, temp_req, &req_list, list) { + if (req->addr == addr) { + req->status = -ECANCELED; + req->timeout = jiffies; + list_move(&req->list, &req_list); + set_timeout(req->timeout); + break; + } + } + mutex_unlock(&lock); +} +EXPORT_SYMBOL(rdma_addr_cancel); + +struct resolve_cb_context { + struct rdma_dev_addr *addr; + struct completion comp; +}; + +static void resolve_cb(int status, struct sockaddr *src_addr, + struct rdma_dev_addr *addr, void *context) +{ + memcpy(((struct resolve_cb_context *)context)->addr, addr, sizeof(struct + rdma_dev_addr)); + complete(&((struct resolve_cb_context *)context)->comp); +} + +int rdma_addr_find_dmac_by_grh(union ib_gid *sgid, union ib_gid *dgid, u8 *dmac, + u16 *vlan_id) +{ + int ret = 0; + struct rdma_dev_addr dev_addr; + struct resolve_cb_context ctx; + struct net_device *dev; + + union { + struct sockaddr _sockaddr; + struct sockaddr_in _sockaddr_in; + struct sockaddr_in6 _sockaddr_in6; + } sgid_addr, dgid_addr; + + + ret = rdma_gid2ip(&sgid_addr._sockaddr, sgid); + if (ret) + return ret; + + ret = rdma_gid2ip(&dgid_addr._sockaddr, dgid); + if (ret) + return ret; + + memset(&dev_addr, 0, sizeof(dev_addr)); + + ctx.addr = &dev_addr; + init_completion(&ctx.comp); + ret = rdma_resolve_ip(&self, &sgid_addr._sockaddr, &dgid_addr._sockaddr, + &dev_addr, 1000, resolve_cb, &ctx); + if (ret) + return ret; + + wait_for_completion(&ctx.comp); + + memcpy(dmac, dev_addr.dst_dev_addr, ETH_ALEN); + dev = dev_get_by_index(&init_net, dev_addr.bound_dev_if); + if (!dev) + return -ENODEV; + if (vlan_id) + *vlan_id = rdma_vlan_dev_vlan_id(dev); + dev_put(dev); + return ret; +} +EXPORT_SYMBOL(rdma_addr_find_dmac_by_grh); + +int rdma_addr_find_smac_by_sgid(union ib_gid *sgid, u8 *smac, u16 *vlan_id) +{ + int ret = 0; + struct rdma_dev_addr dev_addr; + union { + struct sockaddr _sockaddr; + struct sockaddr_in _sockaddr_in; + struct sockaddr_in6 _sockaddr_in6; + } gid_addr; + + ret = rdma_gid2ip(&gid_addr._sockaddr, sgid); + + if (ret) + return ret; + memset(&dev_addr, 0, sizeof(dev_addr)); + ret = rdma_translate_ip(&gid_addr._sockaddr, &dev_addr, vlan_id); + if (ret) + return ret; + + memcpy(smac, dev_addr.src_dev_addr, ETH_ALEN); + return ret; +} +EXPORT_SYMBOL(rdma_addr_find_smac_by_sgid); + +static int netevent_callback(struct notifier_block *self, unsigned long event, + void *ctx) +{ + if (event == NETEVENT_NEIGH_UPDATE) { + struct neighbour *neigh = ctx; + + if (neigh->nud_state & NUD_VALID) { + set_timeout(jiffies); + } + } + return 0; +} + +static struct notifier_block nb = { + .notifier_call = netevent_callback +}; + +static int __init addr_init(void) +{ + addr_wq = create_singlethread_workqueue("ib_addr"); + if (!addr_wq) + return -ENOMEM; + + register_netevent_notifier(&nb); + rdma_addr_register_client(&self); + return 0; +} + +static void __exit addr_cleanup(void) +{ + rdma_addr_unregister_client(&self); + unregister_netevent_notifier(&nb); + destroy_workqueue(addr_wq); +} + +module_init(addr_init); +module_exit(addr_cleanup); diff --git a/drivers/infiniband/core/agent.c b/drivers/infiniband/core/agent.c index 34b724afd28..2bc7f5af64f 100644 --- a/drivers/infiniband/core/agent.c +++ b/drivers/infiniband/core/agent.c @@ -3,7 +3,7 @@ * Copyright (c) 2004, 2005 Infinicon Corporation. All rights reserved. * Copyright (c) 2004, 2005 Intel Corporation. All rights reserved. * Copyright (c) 2004, 2005 Topspin Corporation. All rights reserved. - * Copyright (c) 2004, 2005 Voltaire Corporation. All rights reserved. + * Copyright (c) 2004-2007 Voltaire Corporation. All rights reserved. * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -34,7 +34,6 @@ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. * - * $Id: agent.c 1389 2004-12-27 22:56:47Z roland $ */ #include <linux/slab.h> @@ -42,6 +41,7 @@ #include "agent.h" #include "smi.h" +#include "mad_priv.h" #define SPFX "ib_agent: " @@ -59,8 +59,8 @@ __ib_get_agent_port(struct ib_device *device, int port_num) struct ib_agent_port_private *entry; list_for_each_entry(entry, &ib_agent_port_list, port_list) { - if (entry->agent[0]->device == device && - entry->agent[0]->port_num == port_num) + if (entry->agent[1]->device == device && + entry->agent[1]->port_num == port_num) return entry; } return NULL; @@ -78,70 +78,61 @@ ib_get_agent_port(struct ib_device *device, int port_num) return entry; } -int smi_check_local_dr_smp(struct ib_smp *smp, - struct ib_device *device, - int port_num) -{ - struct ib_agent_port_private *port_priv; - - if (smp->mgmt_class != IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) - return 1; - - port_priv = ib_get_agent_port(device, port_num); - if (!port_priv) { - printk(KERN_DEBUG SPFX "smi_check_local_dr_smp %s port %d " - "not open\n", device->name, port_num); - return 1; - } - - return smi_check_local_smp(port_priv->agent[0], smp); -} - -int agent_send_response(struct ib_mad *mad, struct ib_grh *grh, - struct ib_wc *wc, struct ib_device *device, - int port_num, int qpn) +void agent_send_response(struct ib_mad *mad, struct ib_grh *grh, + struct ib_wc *wc, struct ib_device *device, + int port_num, int qpn) { struct ib_agent_port_private *port_priv; struct ib_mad_agent *agent; struct ib_mad_send_buf *send_buf; struct ib_ah *ah; - int ret; + struct ib_mad_send_wr_private *mad_send_wr; + + if (device->node_type == RDMA_NODE_IB_SWITCH) + port_priv = ib_get_agent_port(device, 0); + else + port_priv = ib_get_agent_port(device, port_num); - port_priv = ib_get_agent_port(device, port_num); if (!port_priv) { printk(KERN_ERR SPFX "Unable to find port agent\n"); - return -ENODEV; + return; } agent = port_priv->agent[qpn]; ah = ib_create_ah_from_wc(agent->qp->pd, wc, grh, port_num); if (IS_ERR(ah)) { - ret = PTR_ERR(ah); - printk(KERN_ERR SPFX "ib_create_ah_from_wc error:%d\n", ret); - return ret; + printk(KERN_ERR SPFX "ib_create_ah_from_wc error %ld\n", + PTR_ERR(ah)); + return; } send_buf = ib_create_send_mad(agent, wc->src_qp, wc->pkey_index, 0, IB_MGMT_MAD_HDR, IB_MGMT_MAD_DATA, GFP_KERNEL); if (IS_ERR(send_buf)) { - ret = PTR_ERR(send_buf); - printk(KERN_ERR SPFX "ib_create_send_mad error:%d\n", ret); + printk(KERN_ERR SPFX "ib_create_send_mad error\n"); goto err1; } memcpy(send_buf->mad, mad, sizeof *mad); send_buf->ah = ah; - if ((ret = ib_post_send_mad(send_buf, NULL))) { - printk(KERN_ERR SPFX "ib_post_send_mad error:%d\n", ret); + + if (device->node_type == RDMA_NODE_IB_SWITCH) { + mad_send_wr = container_of(send_buf, + struct ib_mad_send_wr_private, + send_buf); + mad_send_wr->send_wr.wr.ud.port_num = port_num; + } + + if (ib_post_send_mad(send_buf, NULL)) { + printk(KERN_ERR SPFX "ib_post_send_mad error\n"); goto err2; } - return 0; + return; err2: ib_free_send_mad(send_buf); err1: ib_destroy_ah(ah); - return ret; } static void agent_send_handler(struct ib_mad_agent *mad_agent, @@ -165,14 +156,16 @@ int ib_agent_port_open(struct ib_device *device, int port_num) goto error1; } - /* Obtain send only MAD agent for SMI QP */ - port_priv->agent[0] = ib_register_mad_agent(device, port_num, - IB_QPT_SMI, NULL, 0, - &agent_send_handler, - NULL, NULL); - if (IS_ERR(port_priv->agent[0])) { - ret = PTR_ERR(port_priv->agent[0]); - goto error2; + if (rdma_port_get_link_layer(device, port_num) == IB_LINK_LAYER_INFINIBAND) { + /* Obtain send only MAD agent for SMI QP */ + port_priv->agent[0] = ib_register_mad_agent(device, port_num, + IB_QPT_SMI, NULL, 0, + &agent_send_handler, + NULL, NULL); + if (IS_ERR(port_priv->agent[0])) { + ret = PTR_ERR(port_priv->agent[0]); + goto error2; + } } /* Obtain send only MAD agent for GSI QP */ @@ -192,7 +185,8 @@ int ib_agent_port_open(struct ib_device *device, int port_num) return 0; error3: - ib_unregister_mad_agent(port_priv->agent[0]); + if (port_priv->agent[0]) + ib_unregister_mad_agent(port_priv->agent[0]); error2: kfree(port_priv); error1: @@ -215,7 +209,9 @@ int ib_agent_port_close(struct ib_device *device, int port_num) spin_unlock_irqrestore(&ib_agent_port_list_lock, flags); ib_unregister_mad_agent(port_priv->agent[1]); - ib_unregister_mad_agent(port_priv->agent[0]); + if (port_priv->agent[0]) + ib_unregister_mad_agent(port_priv->agent[0]); + kfree(port_priv); return 0; } diff --git a/drivers/infiniband/core/agent.h b/drivers/infiniband/core/agent.h index 86d72fab37b..6669287009c 100644 --- a/drivers/infiniband/core/agent.h +++ b/drivers/infiniband/core/agent.h @@ -32,8 +32,6 @@ * 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. - * - * $Id: agent.h 1389 2004-12-27 22:56:47Z roland $ */ #ifndef __AGENT_H_ @@ -46,8 +44,8 @@ extern int ib_agent_port_open(struct ib_device *device, int port_num); extern int ib_agent_port_close(struct ib_device *device, int port_num); -extern int agent_send_response(struct ib_mad *mad, struct ib_grh *grh, - struct ib_wc *wc, struct ib_device *device, - int port_num, int qpn); +extern void agent_send_response(struct ib_mad *mad, struct ib_grh *grh, + struct ib_wc *wc, struct ib_device *device, + int port_num, int qpn); #endif /* __AGENT_H_ */ diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c index c57a3871184..80f6cf2449f 100644 --- a/drivers/infiniband/core/cache.c +++ b/drivers/infiniband/core/cache.c @@ -31,14 +31,12 @@ * 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. - * - * $Id: cache.c 1349 2004-12-16 21:09:43Z roland $ */ #include <linux/module.h> #include <linux/errno.h> #include <linux/slab.h> -#include <linux/sched.h> /* INIT_WORK, schedule_work(), flush_scheduled_work() */ +#include <linux/workqueue.h> #include <rdma/ib_cache.h> @@ -62,12 +60,13 @@ struct ib_update_work { static inline int start_port(struct ib_device *device) { - return device->node_type == IB_NODE_SWITCH ? 0 : 1; + return (device->node_type == RDMA_NODE_IB_SWITCH) ? 0 : 1; } static inline int end_port(struct ib_device *device) { - return device->node_type == IB_NODE_SWITCH ? 0 : device->phys_port_cnt; + return (device->node_type == RDMA_NODE_IB_SWITCH) ? + 0 : device->phys_port_cnt; } int ib_get_cached_gid(struct ib_device *device, @@ -168,6 +167,7 @@ int ib_find_cached_pkey(struct ib_device *device, unsigned long flags; int i; int ret = -ENOENT; + int partial_ix = -1; if (port_num < start_port(device) || port_num > end_port(device)) return -EINVAL; @@ -180,6 +180,46 @@ int ib_find_cached_pkey(struct ib_device *device, for (i = 0; i < cache->table_len; ++i) if ((cache->table[i] & 0x7fff) == (pkey & 0x7fff)) { + if (cache->table[i] & 0x8000) { + *index = i; + ret = 0; + break; + } else + partial_ix = i; + } + + if (ret && partial_ix >= 0) { + *index = partial_ix; + ret = 0; + } + + read_unlock_irqrestore(&device->cache.lock, flags); + + return ret; +} +EXPORT_SYMBOL(ib_find_cached_pkey); + +int ib_find_exact_cached_pkey(struct ib_device *device, + u8 port_num, + u16 pkey, + u16 *index) +{ + struct ib_pkey_cache *cache; + unsigned long flags; + int i; + int ret = -ENOENT; + + if (port_num < start_port(device) || port_num > end_port(device)) + return -EINVAL; + + read_lock_irqsave(&device->cache.lock, flags); + + cache = device->cache.pkey_cache[port_num - start_port(device)]; + + *index = -1; + + for (i = 0; i < cache->table_len; ++i) + if (cache->table[i] == pkey) { *index = i; ret = 0; break; @@ -189,7 +229,25 @@ int ib_find_cached_pkey(struct ib_device *device, return ret; } -EXPORT_SYMBOL(ib_find_cached_pkey); +EXPORT_SYMBOL(ib_find_exact_cached_pkey); + +int ib_get_cached_lmc(struct ib_device *device, + u8 port_num, + u8 *lmc) +{ + unsigned long flags; + int ret = 0; + + if (port_num < start_port(device) || port_num > end_port(device)) + return -EINVAL; + + read_lock_irqsave(&device->cache.lock, flags); + *lmc = device->cache.lmc_cache[port_num - start_port(device)]; + read_unlock_irqrestore(&device->cache.lock, flags); + + return ret; +} +EXPORT_SYMBOL(ib_get_cached_lmc); static void ib_cache_update(struct ib_device *device, u8 port) @@ -251,6 +309,8 @@ static void ib_cache_update(struct ib_device *device, device->cache.pkey_cache[port - start_port(device)] = pkey_cache; device->cache.gid_cache [port - start_port(device)] = gid_cache; + device->cache.lmc_cache[port - start_port(device)] = tprops->lmc; + write_unlock_irq(&device->cache.lock); kfree(old_pkey_cache); @@ -264,9 +324,10 @@ err: kfree(tprops); } -static void ib_cache_task(void *work_ptr) +static void ib_cache_task(struct work_struct *_work) { - struct ib_update_work *work = work_ptr; + struct ib_update_work *work = + container_of(_work, struct ib_update_work, work); ib_cache_update(work->device, work->port_num); kfree(work); @@ -281,13 +342,15 @@ static void ib_cache_event(struct ib_event_handler *handler, event->event == IB_EVENT_PORT_ACTIVE || event->event == IB_EVENT_LID_CHANGE || event->event == IB_EVENT_PKEY_CHANGE || - event->event == IB_EVENT_SM_CHANGE) { + event->event == IB_EVENT_SM_CHANGE || + event->event == IB_EVENT_CLIENT_REREGISTER || + event->event == IB_EVENT_GID_CHANGE) { work = kmalloc(sizeof *work, GFP_ATOMIC); if (work) { - INIT_WORK(&work->work, ib_cache_task, work); + INIT_WORK(&work->work, ib_cache_task); work->device = event->device; work->port_num = event->element.port_num; - schedule_work(&work->work); + queue_work(ib_wq, &work->work); } } } @@ -302,10 +365,16 @@ static void ib_cache_setup_one(struct ib_device *device) kmalloc(sizeof *device->cache.pkey_cache * (end_port(device) - start_port(device) + 1), GFP_KERNEL); device->cache.gid_cache = - kmalloc(sizeof *device->cache.pkey_cache * + kmalloc(sizeof *device->cache.gid_cache * (end_port(device) - start_port(device) + 1), GFP_KERNEL); - if (!device->cache.pkey_cache || !device->cache.gid_cache) { + device->cache.lmc_cache = kmalloc(sizeof *device->cache.lmc_cache * + (end_port(device) - + start_port(device) + 1), + GFP_KERNEL); + + if (!device->cache.pkey_cache || !device->cache.gid_cache || + !device->cache.lmc_cache) { printk(KERN_WARNING "Couldn't allocate cache " "for %s\n", device->name); goto err; @@ -333,6 +402,7 @@ err_cache: err: kfree(device->cache.pkey_cache); kfree(device->cache.gid_cache); + kfree(device->cache.lmc_cache); } static void ib_cache_cleanup_one(struct ib_device *device) @@ -340,7 +410,7 @@ static void ib_cache_cleanup_one(struct ib_device *device) int p; ib_unregister_event_handler(&device->cache.event_handler); - flush_scheduled_work(); + flush_workqueue(ib_wq); for (p = 0; p <= end_port(device) - start_port(device); ++p) { kfree(device->cache.pkey_cache[p]); @@ -349,6 +419,7 @@ static void ib_cache_cleanup_one(struct ib_device *device) kfree(device->cache.pkey_cache); kfree(device->cache.gid_cache); + kfree(device->cache.lmc_cache); } static struct ib_client cache_client = { diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c index 2514de3480d..c3239170d8b 100644 --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2004, 2005 Intel Corporation. All rights reserved. + * Copyright (c) 2004-2007 Intel Corporation. All rights reserved. * Copyright (c) 2004 Topspin Corporation. All rights reserved. * Copyright (c) 2004, 2005 Voltaire Corporation. All rights reserved. * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. @@ -31,17 +31,23 @@ * 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. - * - * $Id: cm.c 2821 2005-07-08 17:07:28Z sean.hefty $ */ + +#include <linux/completion.h> #include <linux/dma-mapping.h> +#include <linux/device.h> +#include <linux/module.h> #include <linux/err.h> #include <linux/idr.h> #include <linux/interrupt.h> -#include <linux/pci.h> +#include <linux/random.h> #include <linux/rbtree.h> #include <linux/spinlock.h> +#include <linux/slab.h> +#include <linux/sysfs.h> #include <linux/workqueue.h> +#include <linux/kdev_t.h> +#include <linux/etherdevice.h> #include <rdma/ib_cache.h> #include <rdma/ib_cm.h> @@ -71,20 +77,99 @@ static struct ib_cm { struct rb_root remote_id_table; struct rb_root remote_sidr_table; struct idr local_id_table; + __be32 random_id_operand; + struct list_head timewait_list; struct workqueue_struct *wq; } cm; +/* Counter indexes ordered by attribute ID */ +enum { + CM_REQ_COUNTER, + CM_MRA_COUNTER, + CM_REJ_COUNTER, + CM_REP_COUNTER, + CM_RTU_COUNTER, + CM_DREQ_COUNTER, + CM_DREP_COUNTER, + CM_SIDR_REQ_COUNTER, + CM_SIDR_REP_COUNTER, + CM_LAP_COUNTER, + CM_APR_COUNTER, + CM_ATTR_COUNT, + CM_ATTR_ID_OFFSET = 0x0010, +}; + +enum { + CM_XMIT, + CM_XMIT_RETRIES, + CM_RECV, + CM_RECV_DUPLICATES, + CM_COUNTER_GROUPS +}; + +static char const counter_group_names[CM_COUNTER_GROUPS] + [sizeof("cm_rx_duplicates")] = { + "cm_tx_msgs", "cm_tx_retries", + "cm_rx_msgs", "cm_rx_duplicates" +}; + +struct cm_counter_group { + struct kobject obj; + atomic_long_t counter[CM_ATTR_COUNT]; +}; + +struct cm_counter_attribute { + struct attribute attr; + int index; +}; + +#define CM_COUNTER_ATTR(_name, _index) \ +struct cm_counter_attribute cm_##_name##_counter_attr = { \ + .attr = { .name = __stringify(_name), .mode = 0444 }, \ + .index = _index \ +} + +static CM_COUNTER_ATTR(req, CM_REQ_COUNTER); +static CM_COUNTER_ATTR(mra, CM_MRA_COUNTER); +static CM_COUNTER_ATTR(rej, CM_REJ_COUNTER); +static CM_COUNTER_ATTR(rep, CM_REP_COUNTER); +static CM_COUNTER_ATTR(rtu, CM_RTU_COUNTER); +static CM_COUNTER_ATTR(dreq, CM_DREQ_COUNTER); +static CM_COUNTER_ATTR(drep, CM_DREP_COUNTER); +static CM_COUNTER_ATTR(sidr_req, CM_SIDR_REQ_COUNTER); +static CM_COUNTER_ATTR(sidr_rep, CM_SIDR_REP_COUNTER); +static CM_COUNTER_ATTR(lap, CM_LAP_COUNTER); +static CM_COUNTER_ATTR(apr, CM_APR_COUNTER); + +static struct attribute *cm_counter_default_attrs[] = { + &cm_req_counter_attr.attr, + &cm_mra_counter_attr.attr, + &cm_rej_counter_attr.attr, + &cm_rep_counter_attr.attr, + &cm_rtu_counter_attr.attr, + &cm_dreq_counter_attr.attr, + &cm_drep_counter_attr.attr, + &cm_sidr_req_counter_attr.attr, + &cm_sidr_rep_counter_attr.attr, + &cm_lap_counter_attr.attr, + &cm_apr_counter_attr.attr, + NULL +}; + struct cm_port { struct cm_device *cm_dev; struct ib_mad_agent *mad_agent; + struct kobject port_obj; u8 port_num; + struct cm_counter_group counter_group[CM_COUNTER_GROUPS]; }; struct cm_device { struct list_head list; - struct ib_device *device; - __be64 ca_guid; - struct cm_port port[0]; + struct ib_device *ib_device; + struct device *device; + u8 ack_delay; + struct cm_port *port[0]; }; struct cm_av { @@ -92,11 +177,13 @@ struct cm_av { union ib_gid dgid; struct ib_ah_attr ah_attr; u16 pkey_index; - u8 packet_life_time; + u8 timeout; + u8 valid; + u8 smac[ETH_ALEN]; }; struct cm_work { - struct work_struct work; + struct delayed_work work; struct list_head list; struct cm_port *port; struct ib_mad_recv_wc *mad_recv_wc; /* Received MADs */ @@ -108,6 +195,7 @@ struct cm_work { struct cm_timewait_info { struct cm_work work; /* Must be first. */ + struct list_head list; struct rb_node remote_qp_node; struct rb_node remote_id_node; __be64 remote_ca_guid; @@ -121,8 +209,8 @@ struct cm_id_private { struct rb_node service_node; struct rb_node sidr_id_node; - spinlock_t lock; - wait_queue_head_t wait; + spinlock_t lock; /* Do not acquire inside cm.lock */ + struct completion comp; atomic_t refcount; struct ib_mad_send_buf *msg; @@ -130,6 +218,7 @@ struct cm_id_private { /* todo: use alternate port on send failure */ struct cm_av av; struct cm_av alt_av; + struct ib_cm_compare_data *compare_data; void *private_data; __be64 tid; @@ -140,26 +229,27 @@ struct cm_id_private { __be32 rq_psn; int timeout_ms; enum ib_mtu path_mtu; + __be16 pkey; u8 private_data_len; u8 max_cm_retries; u8 peer_to_peer; u8 responder_resources; u8 initiator_depth; - u8 local_ack_timeout; u8 retry_count; u8 rnr_retry_count; u8 service_timeout; + u8 target_ack_delay; struct list_head work_list; atomic_t work_count; }; -static void cm_work_handler(void *data); +static void cm_work_handler(struct work_struct *work); static inline void cm_deref_id(struct cm_id_private *cm_id_priv) { if (atomic_dec_and_test(&cm_id_priv->refcount)) - wake_up(&cm_id_priv->wait); + complete(&cm_id_priv->comp); } static int cm_alloc_msg(struct cm_id_private *cm_id_priv, @@ -174,7 +264,7 @@ static int cm_alloc_msg(struct cm_id_private *cm_id_priv, if (IS_ERR(ah)) return PTR_ERR(ah); - m = ib_create_send_mad(mad_agent, cm_id_priv->id.remote_cm_qpn, + m = ib_create_send_mad(mad_agent, cm_id_priv->id.remote_cm_qpn, cm_id_priv->av.pkey_index, 0, IB_MGMT_MAD_HDR, IB_MGMT_MAD_DATA, GFP_ATOMIC); @@ -233,11 +323,10 @@ static void * cm_copy_private_data(const void *private_data, if (!private_data || !private_data_len) return NULL; - data = kmalloc(private_data_len, GFP_KERNEL); + data = kmemdup(private_data, private_data_len, GFP_KERNEL); if (!data) return ERR_PTR(-ENOMEM); - memcpy(data, private_data, private_data_len); return data; } @@ -251,23 +340,13 @@ static void cm_set_private_data(struct cm_id_private *cm_id_priv, cm_id_priv->private_data_len = private_data_len; } -static void cm_set_ah_attr(struct ib_ah_attr *ah_attr, u8 port_num, - u16 dlid, u8 sl, u16 src_path_bits) -{ - memset(ah_attr, 0, sizeof ah_attr); - ah_attr->dlid = dlid; - ah_attr->sl = sl; - ah_attr->src_path_bits = src_path_bits; - ah_attr->port_num = port_num; -} - -static void cm_init_av_for_response(struct cm_port *port, - struct ib_wc *wc, struct cm_av *av) +static void cm_init_av_for_response(struct cm_port *port, struct ib_wc *wc, + struct ib_grh *grh, struct cm_av *av) { av->port = port; av->pkey_index = wc->pkey_index; - cm_set_ah_attr(&av->ah_attr, port->port_num, wc->slid, - wc->sl, wc->dlid_path_bits); + ib_init_ah_from_wc(port->cm_dev->ib_device, port->port_num, wc, + grh, &av->ah_attr); } static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av) @@ -280,9 +359,9 @@ static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av) read_lock_irqsave(&cm.device_lock, flags); list_for_each_entry(cm_dev, &cm.device_list, list) { - if (!ib_find_cached_gid(cm_dev->device, &path->sgid, + if (!ib_find_cached_gid(cm_dev->ib_device, &path->sgid, &p, NULL)) { - port = &cm_dev->port[p-1]; + port = cm_dev->port[p-1]; break; } } @@ -291,48 +370,52 @@ static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av) if (!port) return -EINVAL; - ret = ib_find_cached_pkey(cm_dev->device, port->port_num, + ret = ib_find_cached_pkey(cm_dev->ib_device, port->port_num, be16_to_cpu(path->pkey), &av->pkey_index); if (ret) return ret; av->port = port; - cm_set_ah_attr(&av->ah_attr, av->port->port_num, - be16_to_cpu(path->dlid), path->sl, - be16_to_cpu(path->slid) & 0x7F); - av->packet_life_time = path->packet_life_time; + ib_init_ah_from_path(cm_dev->ib_device, port->port_num, path, + &av->ah_attr); + av->timeout = path->packet_life_time + 1; + memcpy(av->smac, path->smac, sizeof(av->smac)); + + av->valid = 1; return 0; } static int cm_alloc_id(struct cm_id_private *cm_id_priv) { unsigned long flags; - int ret; - static int next_id; + int id; - do { - spin_lock_irqsave(&cm.lock, flags); - ret = idr_get_new_above(&cm.local_id_table, cm_id_priv, next_id++, - (__force int *) &cm_id_priv->id.local_id); - spin_unlock_irqrestore(&cm.lock, flags); - } while( (ret == -EAGAIN) && idr_pre_get(&cm.local_id_table, GFP_KERNEL) ); - return ret; + idr_preload(GFP_KERNEL); + spin_lock_irqsave(&cm.lock, flags); + + id = idr_alloc_cyclic(&cm.local_id_table, cm_id_priv, 0, 0, GFP_NOWAIT); + + spin_unlock_irqrestore(&cm.lock, flags); + idr_preload_end(); + + cm_id_priv->id.local_id = (__force __be32)id ^ cm.random_id_operand; + return id < 0 ? id : 0; } static void cm_free_id(__be32 local_id) { - unsigned long flags; - - spin_lock_irqsave(&cm.lock, flags); - idr_remove(&cm.local_id_table, (__force int) local_id); - spin_unlock_irqrestore(&cm.lock, flags); + spin_lock_irq(&cm.lock); + idr_remove(&cm.local_id_table, + (__force int) (local_id ^ cm.random_id_operand)); + spin_unlock_irq(&cm.lock); } static struct cm_id_private * cm_get_id(__be32 local_id, __be32 remote_id) { struct cm_id_private *cm_id_priv; - cm_id_priv = idr_find(&cm.local_id_table, (__force int) local_id); + cm_id_priv = idr_find(&cm.local_id_table, + (__force int) (local_id ^ cm.random_id_operand)); if (cm_id_priv) { if (cm_id_priv->id.remote_id == remote_id) atomic_inc(&cm_id_priv->refcount); @@ -346,15 +429,74 @@ static struct cm_id_private * cm_get_id(__be32 local_id, __be32 remote_id) static struct cm_id_private * cm_acquire_id(__be32 local_id, __be32 remote_id) { struct cm_id_private *cm_id_priv; - unsigned long flags; - spin_lock_irqsave(&cm.lock, flags); + spin_lock_irq(&cm.lock); cm_id_priv = cm_get_id(local_id, remote_id); - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock_irq(&cm.lock); return cm_id_priv; } +static void cm_mask_copy(u8 *dst, u8 *src, u8 *mask) +{ + int i; + + for (i = 0; i < IB_CM_COMPARE_SIZE / sizeof(unsigned long); i++) + ((unsigned long *) dst)[i] = ((unsigned long *) src)[i] & + ((unsigned long *) mask)[i]; +} + +static int cm_compare_data(struct ib_cm_compare_data *src_data, + struct ib_cm_compare_data *dst_data) +{ + u8 src[IB_CM_COMPARE_SIZE]; + u8 dst[IB_CM_COMPARE_SIZE]; + + if (!src_data || !dst_data) + return 0; + + cm_mask_copy(src, src_data->data, dst_data->mask); + cm_mask_copy(dst, dst_data->data, src_data->mask); + return memcmp(src, dst, IB_CM_COMPARE_SIZE); +} + +static int cm_compare_private_data(u8 *private_data, + struct ib_cm_compare_data *dst_data) +{ + u8 src[IB_CM_COMPARE_SIZE]; + + if (!dst_data) + return 0; + + cm_mask_copy(src, private_data, dst_data->mask); + return memcmp(src, dst_data->data, IB_CM_COMPARE_SIZE); +} + +/* + * Trivial helpers to strip endian annotation and compare; the + * endianness doesn't actually matter since we just need a stable + * order for the RB tree. + */ +static int be32_lt(__be32 a, __be32 b) +{ + return (__force u32) a < (__force u32) b; +} + +static int be32_gt(__be32 a, __be32 b) +{ + return (__force u32) a > (__force u32) b; +} + +static int be64_lt(__be64 a, __be64 b) +{ + return (__force u64) a < (__force u64) b; +} + +static int be64_gt(__be64 a, __be64 b) +{ + return (__force u64) a > (__force u64) b; +} + static struct cm_id_private * cm_insert_listen(struct cm_id_private *cm_id_priv) { struct rb_node **link = &cm.listen_service_table.rb_node; @@ -362,21 +504,29 @@ static struct cm_id_private * cm_insert_listen(struct cm_id_private *cm_id_priv) struct cm_id_private *cur_cm_id_priv; __be64 service_id = cm_id_priv->id.service_id; __be64 service_mask = cm_id_priv->id.service_mask; + int data_cmp; while (*link) { parent = *link; cur_cm_id_priv = rb_entry(parent, struct cm_id_private, service_node); + data_cmp = cm_compare_data(cm_id_priv->compare_data, + cur_cm_id_priv->compare_data); if ((cur_cm_id_priv->id.service_mask & service_id) == (service_mask & cur_cm_id_priv->id.service_id) && - (cm_id_priv->id.device == cur_cm_id_priv->id.device)) + (cm_id_priv->id.device == cur_cm_id_priv->id.device) && + !data_cmp) return cur_cm_id_priv; if (cm_id_priv->id.device < cur_cm_id_priv->id.device) link = &(*link)->rb_left; else if (cm_id_priv->id.device > cur_cm_id_priv->id.device) link = &(*link)->rb_right; - else if (service_id < cur_cm_id_priv->id.service_id) + else if (be64_lt(service_id, cur_cm_id_priv->id.service_id)) + link = &(*link)->rb_left; + else if (be64_gt(service_id, cur_cm_id_priv->id.service_id)) + link = &(*link)->rb_right; + else if (data_cmp < 0) link = &(*link)->rb_left; else link = &(*link)->rb_right; @@ -387,23 +537,31 @@ static struct cm_id_private * cm_insert_listen(struct cm_id_private *cm_id_priv) } static struct cm_id_private * cm_find_listen(struct ib_device *device, - __be64 service_id) + __be64 service_id, + u8 *private_data) { struct rb_node *node = cm.listen_service_table.rb_node; struct cm_id_private *cm_id_priv; + int data_cmp; while (node) { cm_id_priv = rb_entry(node, struct cm_id_private, service_node); + data_cmp = cm_compare_private_data(private_data, + cm_id_priv->compare_data); if ((cm_id_priv->id.service_mask & service_id) == cm_id_priv->id.service_id && - (cm_id_priv->id.device == device)) + (cm_id_priv->id.device == device) && !data_cmp) return cm_id_priv; if (device < cm_id_priv->id.device) node = node->rb_left; else if (device > cm_id_priv->id.device) node = node->rb_right; - else if (service_id < cm_id_priv->id.service_id) + else if (be64_lt(service_id, cm_id_priv->id.service_id)) + node = node->rb_left; + else if (be64_gt(service_id, cm_id_priv->id.service_id)) + node = node->rb_right; + else if (data_cmp < 0) node = node->rb_left; else node = node->rb_right; @@ -424,13 +582,13 @@ static struct cm_timewait_info * cm_insert_remote_id(struct cm_timewait_info parent = *link; cur_timewait_info = rb_entry(parent, struct cm_timewait_info, remote_id_node); - if (remote_id < cur_timewait_info->work.remote_id) + if (be32_lt(remote_id, cur_timewait_info->work.remote_id)) link = &(*link)->rb_left; - else if (remote_id > cur_timewait_info->work.remote_id) + else if (be32_gt(remote_id, cur_timewait_info->work.remote_id)) link = &(*link)->rb_right; - else if (remote_ca_guid < cur_timewait_info->remote_ca_guid) + else if (be64_lt(remote_ca_guid, cur_timewait_info->remote_ca_guid)) link = &(*link)->rb_left; - else if (remote_ca_guid > cur_timewait_info->remote_ca_guid) + else if (be64_gt(remote_ca_guid, cur_timewait_info->remote_ca_guid)) link = &(*link)->rb_right; else return cur_timewait_info; @@ -450,13 +608,13 @@ static struct cm_timewait_info * cm_find_remote_id(__be64 remote_ca_guid, while (node) { timewait_info = rb_entry(node, struct cm_timewait_info, remote_id_node); - if (remote_id < timewait_info->work.remote_id) + if (be32_lt(remote_id, timewait_info->work.remote_id)) node = node->rb_left; - else if (remote_id > timewait_info->work.remote_id) + else if (be32_gt(remote_id, timewait_info->work.remote_id)) node = node->rb_right; - else if (remote_ca_guid < timewait_info->remote_ca_guid) + else if (be64_lt(remote_ca_guid, timewait_info->remote_ca_guid)) node = node->rb_left; - else if (remote_ca_guid > timewait_info->remote_ca_guid) + else if (be64_gt(remote_ca_guid, timewait_info->remote_ca_guid)) node = node->rb_right; else return timewait_info; @@ -477,13 +635,13 @@ static struct cm_timewait_info * cm_insert_remote_qpn(struct cm_timewait_info parent = *link; cur_timewait_info = rb_entry(parent, struct cm_timewait_info, remote_qp_node); - if (remote_qpn < cur_timewait_info->remote_qpn) + if (be32_lt(remote_qpn, cur_timewait_info->remote_qpn)) link = &(*link)->rb_left; - else if (remote_qpn > cur_timewait_info->remote_qpn) + else if (be32_gt(remote_qpn, cur_timewait_info->remote_qpn)) link = &(*link)->rb_right; - else if (remote_ca_guid < cur_timewait_info->remote_ca_guid) + else if (be64_lt(remote_ca_guid, cur_timewait_info->remote_ca_guid)) link = &(*link)->rb_left; - else if (remote_ca_guid > cur_timewait_info->remote_ca_guid) + else if (be64_gt(remote_ca_guid, cur_timewait_info->remote_ca_guid)) link = &(*link)->rb_right; else return cur_timewait_info; @@ -507,9 +665,9 @@ static struct cm_id_private * cm_insert_remote_sidr(struct cm_id_private parent = *link; cur_cm_id_priv = rb_entry(parent, struct cm_id_private, sidr_id_node); - if (remote_id < cur_cm_id_priv->id.remote_id) + if (be32_lt(remote_id, cur_cm_id_priv->id.remote_id)) link = &(*link)->rb_left; - else if (remote_id > cur_cm_id_priv->id.remote_id) + else if (be32_gt(remote_id, cur_cm_id_priv->id.remote_id)) link = &(*link)->rb_right; else { int cmp; @@ -559,7 +717,7 @@ struct ib_cm_id *ib_create_cm_id(struct ib_device *device, goto error; spin_lock_init(&cm_id_priv->lock); - init_waitqueue_head(&cm_id_priv->wait); + init_completion(&cm_id_priv->comp); INIT_LIST_HEAD(&cm_id_priv->work_list); atomic_set(&cm_id_priv->work_count, -1); atomic_set(&cm_id_priv->refcount, 1); @@ -596,15 +754,27 @@ static inline int cm_convert_to_ms(int iba_time) return 1 << max(iba_time - 8, 0); } -static void cm_cleanup_timewait(struct cm_timewait_info *timewait_info) +/* + * calculate: 4.096x2^ack_timeout = 4.096x2^ack_delay + 2x4.096x2^life_time + * Because of how ack_timeout is stored, adding one doubles the timeout. + * To avoid large timeouts, select the max(ack_delay, life_time + 1), and + * increment it (round up) only if the other is within 50%. + */ +static u8 cm_ack_timeout(u8 ca_ack_delay, u8 packet_life_time) { - unsigned long flags; + int ack_timeout = packet_life_time + 1; - if (!timewait_info->inserted_remote_id && - !timewait_info->inserted_remote_qp) - return; + if (ack_timeout >= ca_ack_delay) + ack_timeout += (ca_ack_delay >= (ack_timeout - 1)); + else + ack_timeout = ca_ack_delay + + (ack_timeout >= (ca_ack_delay - 1)); - spin_lock_irqsave(&cm.lock, flags); + return min(31, ack_timeout); +} + +static void cm_cleanup_timewait(struct cm_timewait_info *timewait_info) +{ if (timewait_info->inserted_remote_id) { rb_erase(&timewait_info->remote_id_node, &cm.remote_id_table); timewait_info->inserted_remote_id = 0; @@ -614,7 +784,6 @@ static void cm_cleanup_timewait(struct cm_timewait_info *timewait_info) rb_erase(&timewait_info->remote_qp_node, &cm.remote_qp_table); timewait_info->inserted_remote_qp = 0; } - spin_unlock_irqrestore(&cm.lock, flags); } static struct cm_timewait_info * cm_create_timewait_info(__be32 local_id) @@ -626,8 +795,7 @@ static struct cm_timewait_info * cm_create_timewait_info(__be32 local_id) return ERR_PTR(-ENOMEM); timewait_info->work.local_id = local_id; - INIT_WORK(&timewait_info->work.work, cm_work_handler, - &timewait_info->work); + INIT_DELAYED_WORK(&timewait_info->work.work, cm_work_handler); timewait_info->work.cm_event.event = IB_CM_TIMEWAIT_EXIT; return timewait_info; } @@ -635,6 +803,12 @@ static struct cm_timewait_info * cm_create_timewait_info(__be32 local_id) static void cm_enter_timewait(struct cm_id_private *cm_id_priv) { int wait_time; + unsigned long flags; + + spin_lock_irqsave(&cm.lock, flags); + cm_cleanup_timewait(cm_id_priv->timewait_info); + list_add_tail(&cm_id_priv->timewait_info->list, &cm.timewait_list); + spin_unlock_irqrestore(&cm.lock, flags); /* * The cm_id could be destroyed by the user before we exit timewait. @@ -642,7 +816,7 @@ static void cm_enter_timewait(struct cm_id_private *cm_id_priv) * timewait before notifying the user that we've exited timewait. */ cm_id_priv->id.state = IB_CM_TIMEWAIT; - wait_time = cm_convert_to_ms(cm_id_priv->local_ack_timeout); + wait_time = cm_convert_to_ms(cm_id_priv->av.timeout); queue_delayed_work(cm.wq, &cm_id_priv->timewait_info->work.work, msecs_to_jiffies(wait_time)); cm_id_priv->timewait_info = NULL; @@ -650,114 +824,144 @@ static void cm_enter_timewait(struct cm_id_private *cm_id_priv) static void cm_reset_to_idle(struct cm_id_private *cm_id_priv) { + unsigned long flags; + cm_id_priv->id.state = IB_CM_IDLE; if (cm_id_priv->timewait_info) { + spin_lock_irqsave(&cm.lock, flags); cm_cleanup_timewait(cm_id_priv->timewait_info); + spin_unlock_irqrestore(&cm.lock, flags); kfree(cm_id_priv->timewait_info); cm_id_priv->timewait_info = NULL; } } -void ib_destroy_cm_id(struct ib_cm_id *cm_id) +static void cm_destroy_id(struct ib_cm_id *cm_id, int err) { struct cm_id_private *cm_id_priv; struct cm_work *work; - unsigned long flags; cm_id_priv = container_of(cm_id, struct cm_id_private, id); retest: - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); switch (cm_id->state) { case IB_CM_LISTEN: cm_id->state = IB_CM_IDLE; - spin_unlock_irqrestore(&cm_id_priv->lock, flags); - spin_lock_irqsave(&cm.lock, flags); + spin_unlock_irq(&cm_id_priv->lock); + spin_lock_irq(&cm.lock); rb_erase(&cm_id_priv->service_node, &cm.listen_service_table); - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock_irq(&cm.lock); break; case IB_CM_SIDR_REQ_SENT: cm_id->state = IB_CM_IDLE; ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); break; case IB_CM_SIDR_REQ_RCVD: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); cm_reject_sidr_req(cm_id_priv, IB_SIDR_REJECT); break; case IB_CM_REQ_SENT: ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); ib_send_cm_rej(cm_id, IB_CM_REJ_TIMEOUT, - &cm_id_priv->av.port->cm_dev->ca_guid, - sizeof cm_id_priv->av.port->cm_dev->ca_guid, + &cm_id_priv->id.device->node_guid, + sizeof cm_id_priv->id.device->node_guid, NULL, 0); break; + case IB_CM_REQ_RCVD: + if (err == -ENOMEM) { + /* Do not reject to allow future retries. */ + cm_reset_to_idle(cm_id_priv); + spin_unlock_irq(&cm_id_priv->lock); + } else { + spin_unlock_irq(&cm_id_priv->lock); + ib_send_cm_rej(cm_id, IB_CM_REJ_CONSUMER_DEFINED, + NULL, 0, NULL, 0); + } + break; case IB_CM_MRA_REQ_RCVD: case IB_CM_REP_SENT: case IB_CM_MRA_REP_RCVD: ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); /* Fall through */ - case IB_CM_REQ_RCVD: case IB_CM_MRA_REQ_SENT: case IB_CM_REP_RCVD: case IB_CM_MRA_REP_SENT: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); ib_send_cm_rej(cm_id, IB_CM_REJ_CONSUMER_DEFINED, NULL, 0, NULL, 0); break; case IB_CM_ESTABLISHED: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); + if (cm_id_priv->qp_type == IB_QPT_XRC_TGT) + break; ib_send_cm_dreq(cm_id, NULL, 0); goto retest; case IB_CM_DREQ_SENT: ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); cm_enter_timewait(cm_id_priv); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); break; case IB_CM_DREQ_RCVD: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); ib_send_cm_drep(cm_id, NULL, 0); break; default: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); break; } cm_free_id(cm_id->local_id); - atomic_dec(&cm_id_priv->refcount); - wait_event(cm_id_priv->wait, !atomic_read(&cm_id_priv->refcount)); + cm_deref_id(cm_id_priv); + wait_for_completion(&cm_id_priv->comp); while ((work = cm_dequeue_work(cm_id_priv)) != NULL) cm_free_work(work); - if (cm_id_priv->private_data && cm_id_priv->private_data_len) - kfree(cm_id_priv->private_data); + kfree(cm_id_priv->compare_data); + kfree(cm_id_priv->private_data); kfree(cm_id_priv); } + +void ib_destroy_cm_id(struct ib_cm_id *cm_id) +{ + cm_destroy_id(cm_id, 0); +} EXPORT_SYMBOL(ib_destroy_cm_id); -int ib_cm_listen(struct ib_cm_id *cm_id, - __be64 service_id, - __be64 service_mask) +int ib_cm_listen(struct ib_cm_id *cm_id, __be64 service_id, __be64 service_mask, + struct ib_cm_compare_data *compare_data) { struct cm_id_private *cm_id_priv, *cur_cm_id_priv; unsigned long flags; int ret = 0; - service_mask = service_mask ? service_mask : - __constant_cpu_to_be64(~0ULL); + service_mask = service_mask ? service_mask : ~cpu_to_be64(0); service_id &= service_mask; if ((service_id & IB_SERVICE_ID_AGN_MASK) == IB_CM_ASSIGN_SERVICE_ID && (service_id != IB_CM_ASSIGN_SERVICE_ID)) return -EINVAL; cm_id_priv = container_of(cm_id, struct cm_id_private, id); - BUG_ON(cm_id->state != IB_CM_IDLE); + if (cm_id->state != IB_CM_IDLE) + return -EINVAL; + + if (compare_data) { + cm_id_priv->compare_data = kzalloc(sizeof *compare_data, + GFP_KERNEL); + if (!cm_id_priv->compare_data) + return -ENOMEM; + cm_mask_copy(cm_id_priv->compare_data->data, + compare_data->data, compare_data->mask); + memcpy(cm_id_priv->compare_data->mask, compare_data->mask, + IB_CM_COMPARE_SIZE); + } cm_id->state = IB_CM_LISTEN; spin_lock_irqsave(&cm.lock, flags); if (service_id == IB_CM_ASSIGN_SERVICE_ID) { cm_id->service_id = cpu_to_be64(cm.listen_service_id++); - cm_id->service_mask = __constant_cpu_to_be64(~0ULL); + cm_id->service_mask = ~cpu_to_be64(0); } else { cm_id->service_id = service_id; cm_id->service_mask = service_mask; @@ -767,6 +971,8 @@ int ib_cm_listen(struct ib_cm_id *cm_id, if (cur_cm_id_priv) { cm_id->state = IB_CM_IDLE; + kfree(cm_id_priv->compare_data); + cm_id_priv->compare_data = NULL; ret = -EBUSY; } return ret; @@ -799,14 +1005,16 @@ static void cm_format_req(struct cm_req_msg *req_msg, struct cm_id_private *cm_id_priv, struct ib_cm_req_param *param) { + struct ib_sa_path_rec *pri_path = param->primary_path; + struct ib_sa_path_rec *alt_path = param->alternate_path; + cm_format_mad_hdr(&req_msg->hdr, CM_REQ_ATTR_ID, cm_form_tid(cm_id_priv, CM_MSG_SEQUENCE_REQ)); req_msg->local_comm_id = cm_id_priv->id.local_id; req_msg->service_id = param->service_id; - req_msg->local_ca_guid = cm_id_priv->av.port->cm_dev->ca_guid; + req_msg->local_ca_guid = cm_id_priv->id.device->node_guid; cm_req_set_local_qpn(req_msg, cpu_to_be32(param->qp_num)); - cm_req_set_resp_res(req_msg, param->responder_resources); cm_req_set_init_depth(req_msg, param->initiator_depth); cm_req_set_remote_resp_timeout(req_msg, param->remote_cm_response_timeout); @@ -815,40 +1023,57 @@ static void cm_format_req(struct cm_req_msg *req_msg, cm_req_set_starting_psn(req_msg, cpu_to_be32(param->starting_psn)); cm_req_set_local_resp_timeout(req_msg, param->local_cm_response_timeout); - cm_req_set_retry_count(req_msg, param->retry_count); req_msg->pkey = param->primary_path->pkey; cm_req_set_path_mtu(req_msg, param->primary_path->mtu); - cm_req_set_rnr_retry_count(req_msg, param->rnr_retry_count); cm_req_set_max_cm_retries(req_msg, param->max_cm_retries); - cm_req_set_srq(req_msg, param->srq); - - req_msg->primary_local_lid = param->primary_path->slid; - req_msg->primary_remote_lid = param->primary_path->dlid; - req_msg->primary_local_gid = param->primary_path->sgid; - req_msg->primary_remote_gid = param->primary_path->dgid; - cm_req_set_primary_flow_label(req_msg, param->primary_path->flow_label); - cm_req_set_primary_packet_rate(req_msg, param->primary_path->rate); - req_msg->primary_traffic_class = param->primary_path->traffic_class; - req_msg->primary_hop_limit = param->primary_path->hop_limit; - cm_req_set_primary_sl(req_msg, param->primary_path->sl); - cm_req_set_primary_subnet_local(req_msg, 1); /* local only... */ - cm_req_set_primary_local_ack_timeout(req_msg, - min(31, param->primary_path->packet_life_time + 1)); - if (param->alternate_path) { - req_msg->alt_local_lid = param->alternate_path->slid; - req_msg->alt_remote_lid = param->alternate_path->dlid; - req_msg->alt_local_gid = param->alternate_path->sgid; - req_msg->alt_remote_gid = param->alternate_path->dgid; + if (param->qp_type != IB_QPT_XRC_INI) { + cm_req_set_resp_res(req_msg, param->responder_resources); + cm_req_set_retry_count(req_msg, param->retry_count); + cm_req_set_rnr_retry_count(req_msg, param->rnr_retry_count); + cm_req_set_srq(req_msg, param->srq); + } + + if (pri_path->hop_limit <= 1) { + req_msg->primary_local_lid = pri_path->slid; + req_msg->primary_remote_lid = pri_path->dlid; + } else { + /* Work-around until there's a way to obtain remote LID info */ + req_msg->primary_local_lid = IB_LID_PERMISSIVE; + req_msg->primary_remote_lid = IB_LID_PERMISSIVE; + } + req_msg->primary_local_gid = pri_path->sgid; + req_msg->primary_remote_gid = pri_path->dgid; + cm_req_set_primary_flow_label(req_msg, pri_path->flow_label); + cm_req_set_primary_packet_rate(req_msg, pri_path->rate); + req_msg->primary_traffic_class = pri_path->traffic_class; + req_msg->primary_hop_limit = pri_path->hop_limit; + cm_req_set_primary_sl(req_msg, pri_path->sl); + cm_req_set_primary_subnet_local(req_msg, (pri_path->hop_limit <= 1)); + cm_req_set_primary_local_ack_timeout(req_msg, + cm_ack_timeout(cm_id_priv->av.port->cm_dev->ack_delay, + pri_path->packet_life_time)); + + if (alt_path) { + if (alt_path->hop_limit <= 1) { + req_msg->alt_local_lid = alt_path->slid; + req_msg->alt_remote_lid = alt_path->dlid; + } else { + req_msg->alt_local_lid = IB_LID_PERMISSIVE; + req_msg->alt_remote_lid = IB_LID_PERMISSIVE; + } + req_msg->alt_local_gid = alt_path->sgid; + req_msg->alt_remote_gid = alt_path->dgid; cm_req_set_alt_flow_label(req_msg, - param->alternate_path->flow_label); - cm_req_set_alt_packet_rate(req_msg, param->alternate_path->rate); - req_msg->alt_traffic_class = param->alternate_path->traffic_class; - req_msg->alt_hop_limit = param->alternate_path->hop_limit; - cm_req_set_alt_sl(req_msg, param->alternate_path->sl); - cm_req_set_alt_subnet_local(req_msg, 1); /* local only... */ + alt_path->flow_label); + cm_req_set_alt_packet_rate(req_msg, alt_path->rate); + req_msg->alt_traffic_class = alt_path->traffic_class; + req_msg->alt_hop_limit = alt_path->hop_limit; + cm_req_set_alt_sl(req_msg, alt_path->sl); + cm_req_set_alt_subnet_local(req_msg, (alt_path->hop_limit <= 1)); cm_req_set_alt_local_ack_timeout(req_msg, - min(31, param->alternate_path->packet_life_time + 1)); + cm_ack_timeout(cm_id_priv->av.port->cm_dev->ack_delay, + alt_path->packet_life_time)); } if (param->private_data && param->private_data_len) @@ -865,7 +1090,8 @@ static int cm_validate_req_param(struct ib_cm_req_param *param) if (!param->primary_path) return -EINVAL; - if (param->qp_type != IB_QPT_RC && param->qp_type != IB_QPT_UC) + if (param->qp_type != IB_QPT_RC && param->qp_type != IB_QPT_UC && + param->qp_type != IB_QPT_XRC_INI) return -EINVAL; if (param->private_data && @@ -904,8 +1130,10 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, cm_id_priv->timewait_info = cm_create_timewait_info(cm_id_priv-> id.local_id); - if (IS_ERR(cm_id_priv->timewait_info)) + if (IS_ERR(cm_id_priv->timewait_info)) { + ret = PTR_ERR(cm_id_priv->timewait_info); goto out; + } ret = cm_init_av_by_path(param->primary_path, &cm_id_priv->av); if (ret) @@ -917,7 +1145,7 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, goto error1; } cm_id->service_id = param->service_id; - cm_id->service_mask = __constant_cpu_to_be64(~0ULL); + cm_id->service_mask = ~cpu_to_be64(0); cm_id_priv->timeout_ms = cm_convert_to_ms( param->primary_path->packet_life_time) * 2 + cm_convert_to_ms( @@ -927,6 +1155,7 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, cm_id_priv->responder_resources = param->responder_resources; cm_id_priv->retry_count = param->retry_count; cm_id_priv->path_mtu = param->primary_path->mtu; + cm_id_priv->pkey = param->primary_path->pkey; cm_id_priv->qp_type = param->qp_type; ret = cm_alloc_msg(cm_id_priv, &cm_id_priv->msg); @@ -941,8 +1170,6 @@ int ib_send_cm_req(struct ib_cm_id *cm_id, cm_id_priv->local_qpn = cm_req_get_local_qpn(req_msg); cm_id_priv->rq_psn = cm_req_get_starting_psn(req_msg); - cm_id_priv->local_ack_timeout = - cm_req_get_primary_local_ack_timeout(req_msg); spin_lock_irqsave(&cm_id_priv->lock, flags); ret = ib_post_send_mad(cm_id_priv->msg, NULL); @@ -1089,7 +1316,6 @@ static void cm_format_req_event(struct cm_work *work, static void cm_process_work(struct cm_id_private *cm_id_priv, struct cm_work *work) { - unsigned long flags; int ret; /* We will typically only have the current event to report. */ @@ -1097,9 +1323,9 @@ static void cm_process_work(struct cm_id_private *cm_id_priv, cm_free_work(work); while (!ret && !atomic_add_negative(-1, &cm_id_priv->work_count)) { - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); work = cm_dequeue_work(cm_id_priv); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); BUG_ON(!work); ret = cm_id_priv->id.cm_handler(&cm_id_priv->id, &work->cm_event); @@ -1107,7 +1333,7 @@ static void cm_process_work(struct cm_id_private *cm_id_priv, } cm_deref_id(cm_id_priv); if (ret) - ib_destroy_cm_id(&cm_id_priv->id); + cm_destroy_id(&cm_id_priv->id, ret); } static void cm_format_mra(struct cm_mra_msg *mra_msg, @@ -1170,9 +1396,11 @@ static void cm_dup_req_handler(struct cm_work *work, struct cm_id_private *cm_id_priv) { struct ib_mad_send_buf *msg = NULL; - unsigned long flags; int ret; + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_REQ_COUNTER]); + /* Quick state check to discard duplicate REQs. */ if (cm_id_priv->id.state == IB_CM_REQ_RCVD) return; @@ -1181,7 +1409,7 @@ static void cm_dup_req_handler(struct cm_work *work, if (ret) return; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); switch (cm_id_priv->id.state) { case IB_CM_MRA_REQ_SENT: cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, @@ -1196,14 +1424,14 @@ static void cm_dup_req_handler(struct cm_work *work, default: goto unlock; } - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); ret = ib_post_send_mad(msg, NULL); if (ret) goto free; return; -unlock: spin_unlock_irqrestore(&cm_id_priv->lock, flags); +unlock: spin_unlock_irq(&cm_id_priv->lock); free: cm_free_msg(msg); } @@ -1213,49 +1441,81 @@ static struct cm_id_private * cm_match_req(struct cm_work *work, struct cm_id_private *listen_cm_id_priv, *cur_cm_id_priv; struct cm_timewait_info *timewait_info; struct cm_req_msg *req_msg; - unsigned long flags; req_msg = (struct cm_req_msg *)work->mad_recv_wc->recv_buf.mad; - /* Check for duplicate REQ and stale connections. */ - spin_lock_irqsave(&cm.lock, flags); + /* Check for possible duplicate REQ. */ + spin_lock_irq(&cm.lock); timewait_info = cm_insert_remote_id(cm_id_priv->timewait_info); - if (!timewait_info) - timewait_info = cm_insert_remote_qpn(cm_id_priv->timewait_info); - if (timewait_info) { cur_cm_id_priv = cm_get_id(timewait_info->work.local_id, timewait_info->work.remote_id); - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock_irq(&cm.lock); if (cur_cm_id_priv) { cm_dup_req_handler(work, cur_cm_id_priv); cm_deref_id(cur_cm_id_priv); - } else - cm_issue_rej(work->port, work->mad_recv_wc, - IB_CM_REJ_STALE_CONN, CM_MSG_RESPONSE_REQ, - NULL, 0); - goto error; + } + return NULL; + } + + /* Check for stale connections. */ + timewait_info = cm_insert_remote_qpn(cm_id_priv->timewait_info); + if (timewait_info) { + cm_cleanup_timewait(cm_id_priv->timewait_info); + spin_unlock_irq(&cm.lock); + cm_issue_rej(work->port, work->mad_recv_wc, + IB_CM_REJ_STALE_CONN, CM_MSG_RESPONSE_REQ, + NULL, 0); + return NULL; } /* Find matching listen request. */ listen_cm_id_priv = cm_find_listen(cm_id_priv->id.device, - req_msg->service_id); + req_msg->service_id, + req_msg->private_data); if (!listen_cm_id_priv) { - spin_unlock_irqrestore(&cm.lock, flags); + cm_cleanup_timewait(cm_id_priv->timewait_info); + spin_unlock_irq(&cm.lock); cm_issue_rej(work->port, work->mad_recv_wc, IB_CM_REJ_INVALID_SERVICE_ID, CM_MSG_RESPONSE_REQ, NULL, 0); - goto error; + goto out; } atomic_inc(&listen_cm_id_priv->refcount); atomic_inc(&cm_id_priv->refcount); cm_id_priv->id.state = IB_CM_REQ_RCVD; atomic_inc(&cm_id_priv->work_count); - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock_irq(&cm.lock); +out: return listen_cm_id_priv; +} -error: cm_cleanup_timewait(cm_id_priv->timewait_info); - return NULL; +/* + * Work-around for inter-subnet connections. If the LIDs are permissive, + * we need to override the LID/SL data in the REQ with the LID information + * in the work completion. + */ +static void cm_process_routed_req(struct cm_req_msg *req_msg, struct ib_wc *wc) +{ + if (!cm_req_get_primary_subnet_local(req_msg)) { + if (req_msg->primary_local_lid == IB_LID_PERMISSIVE) { + req_msg->primary_local_lid = cpu_to_be16(wc->slid); + cm_req_set_primary_sl(req_msg, wc->sl); + } + + if (req_msg->primary_remote_lid == IB_LID_PERMISSIVE) + req_msg->primary_remote_lid = cpu_to_be16(wc->dlid_path_bits); + } + + if (!cm_req_get_alt_subnet_local(req_msg)) { + if (req_msg->alt_local_lid == IB_LID_PERMISSIVE) { + req_msg->alt_local_lid = cpu_to_be16(wc->slid); + cm_req_set_alt_sl(req_msg, wc->sl); + } + + if (req_msg->alt_remote_lid == IB_LID_PERMISSIVE) + req_msg->alt_remote_lid = cpu_to_be16(wc->dlid_path_bits); + } } static int cm_req_handler(struct cm_work *work) @@ -1267,19 +1527,20 @@ static int cm_req_handler(struct cm_work *work) req_msg = (struct cm_req_msg *)work->mad_recv_wc->recv_buf.mad; - cm_id = ib_create_cm_id(work->port->cm_dev->device, NULL, NULL); + cm_id = ib_create_cm_id(work->port->cm_dev->ib_device, NULL, NULL); if (IS_ERR(cm_id)) return PTR_ERR(cm_id); cm_id_priv = container_of(cm_id, struct cm_id_private, id); cm_id_priv->id.remote_id = req_msg->local_comm_id; cm_init_av_for_response(work->port, work->mad_recv_wc->wc, + work->mad_recv_wc->recv_buf.grh, &cm_id_priv->av); cm_id_priv->timewait_info = cm_create_timewait_info(cm_id_priv-> id.local_id); if (IS_ERR(cm_id_priv->timewait_info)) { ret = PTR_ERR(cm_id_priv->timewait_info); - goto error1; + goto destroy; } cm_id_priv->timewait_info->work.remote_id = req_msg->local_comm_id; cm_id_priv->timewait_info->remote_ca_guid = req_msg->local_ca_guid; @@ -1288,22 +1549,37 @@ static int cm_req_handler(struct cm_work *work) listen_cm_id_priv = cm_match_req(work, cm_id_priv); if (!listen_cm_id_priv) { ret = -EINVAL; - goto error2; + kfree(cm_id_priv->timewait_info); + goto destroy; } cm_id_priv->id.cm_handler = listen_cm_id_priv->id.cm_handler; cm_id_priv->id.context = listen_cm_id_priv->id.context; cm_id_priv->id.service_id = req_msg->service_id; - cm_id_priv->id.service_mask = __constant_cpu_to_be64(~0ULL); + cm_id_priv->id.service_mask = ~cpu_to_be64(0); + cm_process_routed_req(req_msg, work->mad_recv_wc->wc); cm_format_paths_from_req(req_msg, &work->path[0], &work->path[1]); + + memcpy(work->path[0].dmac, cm_id_priv->av.ah_attr.dmac, ETH_ALEN); + work->path[0].vlan_id = cm_id_priv->av.ah_attr.vlan_id; ret = cm_init_av_by_path(&work->path[0], &cm_id_priv->av); - if (ret) - goto error3; + if (ret) { + ib_get_cached_gid(work->port->cm_dev->ib_device, + work->port->port_num, 0, &work->path[0].sgid); + ib_send_cm_rej(cm_id, IB_CM_REJ_INVALID_GID, + &work->path[0].sgid, sizeof work->path[0].sgid, + NULL, 0); + goto rejected; + } if (req_msg->alt_local_lid) { ret = cm_init_av_by_path(&work->path[1], &cm_id_priv->alt_av); - if (ret) - goto error3; + if (ret) { + ib_send_cm_rej(cm_id, IB_CM_REJ_INVALID_ALT_GID, + &work->path[0].sgid, + sizeof work->path[0].sgid, NULL, 0); + goto rejected; + } } cm_id_priv->tid = req_msg->hdr.tid; cm_id_priv->timeout_ms = cm_convert_to_ms( @@ -1313,9 +1589,8 @@ static int cm_req_handler(struct cm_work *work) cm_id_priv->initiator_depth = cm_req_get_resp_res(req_msg); cm_id_priv->responder_resources = cm_req_get_init_depth(req_msg); cm_id_priv->path_mtu = cm_req_get_path_mtu(req_msg); + cm_id_priv->pkey = req_msg->pkey; cm_id_priv->sq_psn = cm_req_get_starting_psn(req_msg); - cm_id_priv->local_ack_timeout = - cm_req_get_primary_local_ack_timeout(req_msg); cm_id_priv->retry_count = cm_req_get_retry_count(req_msg); cm_id_priv->rnr_retry_count = cm_req_get_rnr_retry_count(req_msg); cm_id_priv->qp_type = cm_req_get_qp_type(req_msg); @@ -1325,12 +1600,11 @@ static int cm_req_handler(struct cm_work *work) cm_deref_id(listen_cm_id_priv); return 0; -error3: atomic_dec(&cm_id_priv->refcount); +rejected: + atomic_dec(&cm_id_priv->refcount); cm_deref_id(listen_cm_id_priv); - cm_cleanup_timewait(cm_id_priv->timewait_info); -error2: kfree(cm_id_priv->timewait_info); - cm_id_priv->timewait_info = NULL; -error1: ib_destroy_cm_id(&cm_id_priv->id); +destroy: + ib_destroy_cm_id(cm_id); return ret; } @@ -1341,16 +1615,23 @@ static void cm_format_rep(struct cm_rep_msg *rep_msg, cm_format_mad_hdr(&rep_msg->hdr, CM_REP_ATTR_ID, cm_id_priv->tid); rep_msg->local_comm_id = cm_id_priv->id.local_id; rep_msg->remote_comm_id = cm_id_priv->id.remote_id; - cm_rep_set_local_qpn(rep_msg, cpu_to_be32(param->qp_num)); cm_rep_set_starting_psn(rep_msg, cpu_to_be32(param->starting_psn)); rep_msg->resp_resources = param->responder_resources; - rep_msg->initiator_depth = param->initiator_depth; - cm_rep_set_target_ack_delay(rep_msg, param->target_ack_delay); + cm_rep_set_target_ack_delay(rep_msg, + cm_id_priv->av.port->cm_dev->ack_delay); cm_rep_set_failover(rep_msg, param->failover_accepted); - cm_rep_set_flow_ctrl(rep_msg, param->flow_control); cm_rep_set_rnr_retry_count(rep_msg, param->rnr_retry_count); - cm_rep_set_srq(rep_msg, param->srq); - rep_msg->local_ca_guid = cm_id_priv->av.port->cm_dev->ca_guid; + rep_msg->local_ca_guid = cm_id_priv->id.device->node_guid; + + if (cm_id_priv->qp_type != IB_QPT_XRC_TGT) { + rep_msg->initiator_depth = param->initiator_depth; + cm_rep_set_flow_ctrl(rep_msg, param->flow_control); + cm_rep_set_srq(rep_msg, param->srq); + cm_rep_set_local_qpn(rep_msg, cpu_to_be32(param->qp_num)); + } else { + cm_rep_set_srq(rep_msg, 1); + cm_rep_set_local_eecn(rep_msg, cpu_to_be32(param->qp_num)); + } if (param->private_data && param->private_data_len) memcpy(rep_msg->private_data, param->private_data, @@ -1399,7 +1680,7 @@ int ib_send_cm_rep(struct ib_cm_id *cm_id, cm_id_priv->initiator_depth = param->initiator_depth; cm_id_priv->responder_resources = param->responder_resources; cm_id_priv->rq_psn = cm_rep_get_starting_psn(rep_msg); - cm_id_priv->local_qpn = cm_rep_get_local_qpn(rep_msg); + cm_id_priv->local_qpn = cpu_to_be32(param->qp_num & 0xFFFFFF); out: spin_unlock_irqrestore(&cm_id_priv->lock, flags); return ret; @@ -1470,7 +1751,7 @@ error: spin_unlock_irqrestore(&cm_id_priv->lock, flags); } EXPORT_SYMBOL(ib_send_cm_rtu); -static void cm_format_rep_event(struct cm_work *work) +static void cm_format_rep_event(struct cm_work *work, enum ib_qp_type qp_type) { struct cm_rep_msg *rep_msg; struct ib_cm_rep_event_param *param; @@ -1479,7 +1760,7 @@ static void cm_format_rep_event(struct cm_work *work) param = &work->cm_event.param.rep_rcvd; param->remote_ca_guid = rep_msg->local_ca_guid; param->remote_qkey = be32_to_cpu(rep_msg->local_qkey); - param->remote_qpn = be32_to_cpu(cm_rep_get_local_qpn(rep_msg)); + param->remote_qpn = be32_to_cpu(cm_rep_get_qpn(rep_msg, qp_type)); param->starting_psn = be32_to_cpu(cm_rep_get_starting_psn(rep_msg)); param->responder_resources = rep_msg->initiator_depth; param->initiator_depth = rep_msg->resp_resources; @@ -1496,7 +1777,6 @@ static void cm_dup_rep_handler(struct cm_work *work) struct cm_id_private *cm_id_priv; struct cm_rep_msg *rep_msg; struct ib_mad_send_buf *msg = NULL; - unsigned long flags; int ret; rep_msg = (struct cm_rep_msg *) work->mad_recv_wc->recv_buf.mad; @@ -1505,11 +1785,13 @@ static void cm_dup_rep_handler(struct cm_work *work) if (!cm_id_priv) return; + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_REP_COUNTER]); ret = cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg); if (ret) goto deref; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state == IB_CM_ESTABLISHED) cm_format_rtu((struct cm_rtu_msg *) msg->mad, cm_id_priv, cm_id_priv->private_data, @@ -1521,14 +1803,14 @@ static void cm_dup_rep_handler(struct cm_work *work) cm_id_priv->private_data_len); else goto unlock; - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); ret = ib_post_send_mad(msg, NULL); if (ret) goto free; goto deref; -unlock: spin_unlock_irqrestore(&cm_id_priv->lock, flags); +unlock: spin_unlock_irq(&cm_id_priv->lock); free: cm_free_msg(msg); deref: cm_deref_id(cm_id_priv); } @@ -1537,7 +1819,6 @@ static int cm_rep_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; struct cm_rep_msg *rep_msg; - unsigned long flags; int ret; rep_msg = (struct cm_rep_msg *)work->mad_recv_wc->recv_buf.mad; @@ -1547,47 +1828,60 @@ static int cm_rep_handler(struct cm_work *work) return -EINVAL; } + cm_format_rep_event(work, cm_id_priv->qp_type); + + spin_lock_irq(&cm_id_priv->lock); + switch (cm_id_priv->id.state) { + case IB_CM_REQ_SENT: + case IB_CM_MRA_REQ_RCVD: + break; + default: + spin_unlock_irq(&cm_id_priv->lock); + ret = -EINVAL; + goto error; + } + cm_id_priv->timewait_info->work.remote_id = rep_msg->local_comm_id; cm_id_priv->timewait_info->remote_ca_guid = rep_msg->local_ca_guid; - cm_id_priv->timewait_info->remote_qpn = cm_rep_get_local_qpn(rep_msg); + cm_id_priv->timewait_info->remote_qpn = cm_rep_get_qpn(rep_msg, cm_id_priv->qp_type); - spin_lock_irqsave(&cm.lock, flags); + spin_lock(&cm.lock); /* Check for duplicate REP. */ if (cm_insert_remote_id(cm_id_priv->timewait_info)) { - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock(&cm.lock); + spin_unlock_irq(&cm_id_priv->lock); ret = -EINVAL; goto error; } /* Check for a stale connection. */ if (cm_insert_remote_qpn(cm_id_priv->timewait_info)) { - spin_unlock_irqrestore(&cm.lock, flags); + rb_erase(&cm_id_priv->timewait_info->remote_id_node, + &cm.remote_id_table); + cm_id_priv->timewait_info->inserted_remote_id = 0; + spin_unlock(&cm.lock); + spin_unlock_irq(&cm_id_priv->lock); cm_issue_rej(work->port, work->mad_recv_wc, IB_CM_REJ_STALE_CONN, CM_MSG_RESPONSE_REP, NULL, 0); ret = -EINVAL; goto error; } - spin_unlock_irqrestore(&cm.lock, flags); - - cm_format_rep_event(work); + spin_unlock(&cm.lock); - spin_lock_irqsave(&cm_id_priv->lock, flags); - switch (cm_id_priv->id.state) { - case IB_CM_REQ_SENT: - case IB_CM_MRA_REQ_RCVD: - break; - default: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); - ret = -EINVAL; - goto error; - } cm_id_priv->id.state = IB_CM_REP_RCVD; cm_id_priv->id.remote_id = rep_msg->local_comm_id; - cm_id_priv->remote_qpn = cm_rep_get_local_qpn(rep_msg); + cm_id_priv->remote_qpn = cm_rep_get_qpn(rep_msg, cm_id_priv->qp_type); cm_id_priv->initiator_depth = rep_msg->resp_resources; cm_id_priv->responder_resources = rep_msg->initiator_depth; cm_id_priv->sq_psn = cm_rep_get_starting_psn(rep_msg); cm_id_priv->rnr_retry_count = cm_rep_get_rnr_retry_count(rep_msg); + cm_id_priv->target_ack_delay = cm_rep_get_target_ack_delay(rep_msg); + cm_id_priv->av.timeout = + cm_ack_timeout(cm_id_priv->target_ack_delay, + cm_id_priv->av.timeout - 1); + cm_id_priv->alt_av.timeout = + cm_ack_timeout(cm_id_priv->target_ack_delay, + cm_id_priv->alt_av.timeout - 1); /* todo: handle peer_to_peer */ @@ -1595,7 +1889,7 @@ static int cm_rep_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -1603,7 +1897,7 @@ static int cm_rep_handler(struct cm_work *work) cm_deref_id(cm_id_priv); return 0; -error: cm_cleanup_timewait(cm_id_priv->timewait_info); +error: cm_deref_id(cm_id_priv); return ret; } @@ -1611,17 +1905,16 @@ error: cm_cleanup_timewait(cm_id_priv->timewait_info); static int cm_establish_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; - unsigned long flags; int ret; - /* See comment in ib_cm_establish about lookup. */ + /* See comment in cm_establish about lookup. */ cm_id_priv = cm_acquire_id(work->local_id, work->remote_id); if (!cm_id_priv) return -EINVAL; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state != IB_CM_ESTABLISHED) { - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); goto out; } @@ -1629,7 +1922,7 @@ static int cm_establish_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -1645,7 +1938,6 @@ static int cm_rtu_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; struct cm_rtu_msg *rtu_msg; - unsigned long flags; int ret; rtu_msg = (struct cm_rtu_msg *)work->mad_recv_wc->recv_buf.mad; @@ -1656,10 +1948,12 @@ static int cm_rtu_handler(struct cm_work *work) work->cm_event.private_data = &rtu_msg->private_data; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state != IB_CM_REP_SENT && cm_id_priv->id.state != IB_CM_MRA_REP_RCVD) { - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_RTU_COUNTER]); goto out; } cm_id_priv->id.state = IB_CM_ESTABLISHED; @@ -1668,7 +1962,7 @@ static int cm_rtu_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -1714,6 +2008,10 @@ int ib_send_cm_dreq(struct ib_cm_id *cm_id, goto out; } + if (cm_id->lap_state == IB_CM_LAP_SENT || + cm_id->lap_state == IB_CM_MRA_LAP_RCVD) + ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); + ret = cm_alloc_msg(cm_id_priv, &msg); if (ret) { cm_enter_timewait(cm_id_priv); @@ -1800,23 +2098,52 @@ out: spin_unlock_irqrestore(&cm_id_priv->lock, flags); } EXPORT_SYMBOL(ib_send_cm_drep); +static int cm_issue_drep(struct cm_port *port, + struct ib_mad_recv_wc *mad_recv_wc) +{ + struct ib_mad_send_buf *msg = NULL; + struct cm_dreq_msg *dreq_msg; + struct cm_drep_msg *drep_msg; + int ret; + + ret = cm_alloc_response_msg(port, mad_recv_wc, &msg); + if (ret) + return ret; + + dreq_msg = (struct cm_dreq_msg *) mad_recv_wc->recv_buf.mad; + drep_msg = (struct cm_drep_msg *) msg->mad; + + cm_format_mad_hdr(&drep_msg->hdr, CM_DREP_ATTR_ID, dreq_msg->hdr.tid); + drep_msg->remote_comm_id = dreq_msg->local_comm_id; + drep_msg->local_comm_id = dreq_msg->remote_comm_id; + + ret = ib_post_send_mad(msg, NULL); + if (ret) + cm_free_msg(msg); + + return ret; +} + static int cm_dreq_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; struct cm_dreq_msg *dreq_msg; struct ib_mad_send_buf *msg = NULL; - unsigned long flags; int ret; dreq_msg = (struct cm_dreq_msg *)work->mad_recv_wc->recv_buf.mad; cm_id_priv = cm_acquire_id(dreq_msg->remote_comm_id, dreq_msg->local_comm_id); - if (!cm_id_priv) + if (!cm_id_priv) { + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_DREQ_COUNTER]); + cm_issue_drep(work->port, work->mad_recv_wc); return -EINVAL; + } work->cm_event.private_data = &dreq_msg->private_data; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->local_qpn != cm_dreq_get_remote_qpn(dreq_msg)) goto unlock; @@ -1826,20 +2153,30 @@ static int cm_dreq_handler(struct cm_work *work) ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); break; case IB_CM_ESTABLISHED: + if (cm_id_priv->id.lap_state == IB_CM_LAP_SENT || + cm_id_priv->id.lap_state == IB_CM_MRA_LAP_RCVD) + ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); + break; case IB_CM_MRA_REP_RCVD: break; case IB_CM_TIMEWAIT: + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_DREQ_COUNTER]); if (cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg)) goto unlock; cm_format_drep((struct cm_drep_msg *) msg->mad, cm_id_priv, cm_id_priv->private_data, cm_id_priv->private_data_len); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ib_post_send_mad(msg, NULL)) cm_free_msg(msg); goto deref; + case IB_CM_DREQ_RCVD: + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_DREQ_COUNTER]); + goto unlock; default: goto unlock; } @@ -1848,7 +2185,7 @@ static int cm_dreq_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -1856,7 +2193,7 @@ static int cm_dreq_handler(struct cm_work *work) cm_deref_id(cm_id_priv); return 0; -unlock: spin_unlock_irqrestore(&cm_id_priv->lock, flags); +unlock: spin_unlock_irq(&cm_id_priv->lock); deref: cm_deref_id(cm_id_priv); return -EINVAL; } @@ -1865,7 +2202,6 @@ static int cm_drep_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; struct cm_drep_msg *drep_msg; - unsigned long flags; int ret; drep_msg = (struct cm_drep_msg *)work->mad_recv_wc->recv_buf.mad; @@ -1876,10 +2212,10 @@ static int cm_drep_handler(struct cm_work *work) work->cm_event.private_data = &drep_msg->private_data; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state != IB_CM_DREQ_SENT && cm_id_priv->id.state != IB_CM_DREQ_RCVD) { - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); goto out; } cm_enter_timewait(cm_id_priv); @@ -1888,7 +2224,7 @@ static int cm_drep_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -1978,28 +2314,28 @@ static struct cm_id_private * cm_acquire_rejected_id(struct cm_rej_msg *rej_msg) { struct cm_timewait_info *timewait_info; struct cm_id_private *cm_id_priv; - unsigned long flags; __be32 remote_id; remote_id = rej_msg->local_comm_id; if (__be16_to_cpu(rej_msg->reason) == IB_CM_REJ_TIMEOUT) { - spin_lock_irqsave(&cm.lock, flags); + spin_lock_irq(&cm.lock); timewait_info = cm_find_remote_id( *((__be64 *) rej_msg->ari), remote_id); if (!timewait_info) { - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock_irq(&cm.lock); return NULL; } - cm_id_priv = idr_find(&cm.local_id_table, - (__force int) timewait_info->work.local_id); + cm_id_priv = idr_find(&cm.local_id_table, (__force int) + (timewait_info->work.local_id ^ + cm.random_id_operand)); if (cm_id_priv) { if (cm_id_priv->id.remote_id == remote_id) atomic_inc(&cm_id_priv->refcount); else cm_id_priv = NULL; } - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock_irq(&cm.lock); } else if (cm_rej_get_msg_rejected(rej_msg) == CM_MSG_RESPONSE_REQ) cm_id_priv = cm_acquire_id(rej_msg->remote_comm_id, 0); else @@ -2012,7 +2348,6 @@ static int cm_rej_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; struct cm_rej_msg *rej_msg; - unsigned long flags; int ret; rej_msg = (struct cm_rej_msg *)work->mad_recv_wc->recv_buf.mad; @@ -2022,7 +2357,7 @@ static int cm_rej_handler(struct cm_work *work) cm_format_rej_event(work); - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); switch (cm_id_priv->id.state) { case IB_CM_REQ_SENT: case IB_CM_MRA_REQ_RCVD: @@ -2042,11 +2377,20 @@ static int cm_rej_handler(struct cm_work *work) /* fall through */ case IB_CM_REP_RCVD: case IB_CM_MRA_REP_SENT: - case IB_CM_ESTABLISHED: cm_enter_timewait(cm_id_priv); break; + case IB_CM_ESTABLISHED: + if (cm_id_priv->id.lap_state == IB_CM_LAP_UNINIT || + cm_id_priv->id.lap_state == IB_CM_LAP_SENT) { + if (cm_id_priv->id.lap_state == IB_CM_LAP_SENT) + ib_cancel_mad(cm_id_priv->av.port->mad_agent, + cm_id_priv->msg); + cm_enter_timewait(cm_id_priv); + break; + } + /* fall through */ default: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); ret = -EINVAL; goto out; } @@ -2054,7 +2398,7 @@ static int cm_rej_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -2073,6 +2417,9 @@ int ib_send_cm_mra(struct ib_cm_id *cm_id, { struct cm_id_private *cm_id_priv; struct ib_mad_send_buf *msg; + enum ib_cm_state cm_state; + enum ib_cm_lap_state lap_state; + enum cm_msg_response msg_response; void *data; unsigned long flags; int ret; @@ -2089,48 +2436,42 @@ int ib_send_cm_mra(struct ib_cm_id *cm_id, spin_lock_irqsave(&cm_id_priv->lock, flags); switch(cm_id_priv->id.state) { case IB_CM_REQ_RCVD: - ret = cm_alloc_msg(cm_id_priv, &msg); - if (ret) - goto error1; - - cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, - CM_MSG_RESPONSE_REQ, service_timeout, - private_data, private_data_len); - ret = ib_post_send_mad(msg, NULL); - if (ret) - goto error2; - cm_id->state = IB_CM_MRA_REQ_SENT; + cm_state = IB_CM_MRA_REQ_SENT; + lap_state = cm_id->lap_state; + msg_response = CM_MSG_RESPONSE_REQ; break; case IB_CM_REP_RCVD: - ret = cm_alloc_msg(cm_id_priv, &msg); - if (ret) - goto error1; - - cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, - CM_MSG_RESPONSE_REP, service_timeout, - private_data, private_data_len); - ret = ib_post_send_mad(msg, NULL); - if (ret) - goto error2; - cm_id->state = IB_CM_MRA_REP_SENT; + cm_state = IB_CM_MRA_REP_SENT; + lap_state = cm_id->lap_state; + msg_response = CM_MSG_RESPONSE_REP; break; case IB_CM_ESTABLISHED: + if (cm_id->lap_state == IB_CM_LAP_RCVD) { + cm_state = cm_id->state; + lap_state = IB_CM_MRA_LAP_SENT; + msg_response = CM_MSG_RESPONSE_OTHER; + break; + } + default: + ret = -EINVAL; + goto error1; + } + + if (!(service_timeout & IB_CM_MRA_FLAG_DELAY)) { ret = cm_alloc_msg(cm_id_priv, &msg); if (ret) goto error1; cm_format_mra((struct cm_mra_msg *) msg->mad, cm_id_priv, - CM_MSG_RESPONSE_OTHER, service_timeout, + msg_response, service_timeout, private_data, private_data_len); ret = ib_post_send_mad(msg, NULL); if (ret) goto error2; - cm_id->lap_state = IB_CM_MRA_LAP_SENT; - break; - default: - ret = -EINVAL; - goto error1; } + + cm_id->state = cm_state; + cm_id->lap_state = lap_state; cm_id_priv->service_timeout = service_timeout; cm_set_private_data(cm_id_priv, data, private_data_len); spin_unlock_irqrestore(&cm_id_priv->lock, flags); @@ -2165,7 +2506,6 @@ static int cm_mra_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; struct cm_mra_msg *mra_msg; - unsigned long flags; int timeout, ret; mra_msg = (struct cm_mra_msg *)work->mad_recv_wc->recv_buf.mad; @@ -2177,9 +2517,9 @@ static int cm_mra_handler(struct cm_work *work) work->cm_event.param.mra_rcvd.service_timeout = cm_mra_get_service_timeout(mra_msg); timeout = cm_convert_to_ms(cm_mra_get_service_timeout(mra_msg)) + - cm_convert_to_ms(cm_id_priv->av.packet_life_time); + cm_convert_to_ms(cm_id_priv->av.timeout); - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); switch (cm_id_priv->id.state) { case IB_CM_REQ_SENT: if (cm_mra_get_msg_mraed(mra_msg) != CM_MSG_RESPONSE_REQ || @@ -2199,10 +2539,20 @@ static int cm_mra_handler(struct cm_work *work) if (cm_mra_get_msg_mraed(mra_msg) != CM_MSG_RESPONSE_OTHER || cm_id_priv->id.lap_state != IB_CM_LAP_SENT || ib_modify_mad(cm_id_priv->av.port->mad_agent, - cm_id_priv->msg, timeout)) + cm_id_priv->msg, timeout)) { + if (cm_id_priv->id.lap_state == IB_CM_MRA_LAP_RCVD) + atomic_long_inc(&work->port-> + counter_group[CM_RECV_DUPLICATES]. + counter[CM_MRA_COUNTER]); goto out; + } cm_id_priv->id.lap_state = IB_CM_MRA_LAP_RCVD; break; + case IB_CM_MRA_REQ_RCVD: + case IB_CM_MRA_REP_RCVD: + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_MRA_COUNTER]); + /* fall through */ default: goto out; } @@ -2212,7 +2562,7 @@ static int cm_mra_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -2220,7 +2570,7 @@ static int cm_mra_handler(struct cm_work *work) cm_deref_id(cm_id_priv); return 0; out: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); cm_deref_id(cm_id_priv); return -EINVAL; } @@ -2249,7 +2599,8 @@ static void cm_format_lap(struct cm_lap_msg *lap_msg, cm_lap_set_sl(lap_msg, alternate_path->sl); cm_lap_set_subnet_local(lap_msg, 1); /* local only... */ cm_lap_set_local_ack_timeout(lap_msg, - min(31, alternate_path->packet_life_time + 1)); + cm_ack_timeout(cm_id_priv->av.port->cm_dev->ack_delay, + alternate_path->packet_life_time)); if (private_data && private_data_len) memcpy(lap_msg->private_data, private_data, private_data_len); @@ -2271,11 +2622,19 @@ int ib_send_cm_lap(struct ib_cm_id *cm_id, cm_id_priv = container_of(cm_id, struct cm_id_private, id); spin_lock_irqsave(&cm_id_priv->lock, flags); if (cm_id->state != IB_CM_ESTABLISHED || - cm_id->lap_state != IB_CM_LAP_IDLE) { + (cm_id->lap_state != IB_CM_LAP_UNINIT && + cm_id->lap_state != IB_CM_LAP_IDLE)) { ret = -EINVAL; goto out; } + ret = cm_init_av_by_path(alternate_path, &cm_id_priv->alt_av); + if (ret) + goto out; + cm_id_priv->alt_av.timeout = + cm_ack_timeout(cm_id_priv->target_ack_delay, + cm_id_priv->alt_av.timeout - 1); + ret = cm_alloc_msg(cm_id_priv, &msg); if (ret) goto out; @@ -2300,7 +2659,8 @@ out: spin_unlock_irqrestore(&cm_id_priv->lock, flags); } EXPORT_SYMBOL(ib_send_cm_lap); -static void cm_format_path_from_lap(struct ib_sa_path_rec *path, +static void cm_format_path_from_lap(struct cm_id_private *cm_id_priv, + struct ib_sa_path_rec *path, struct cm_lap_msg *lap_msg) { memset(path, 0, sizeof *path); @@ -2312,10 +2672,10 @@ static void cm_format_path_from_lap(struct ib_sa_path_rec *path, path->hop_limit = lap_msg->alt_hop_limit; path->traffic_class = cm_lap_get_traffic_class(lap_msg); path->reversible = 1; - /* pkey is same as in REQ */ + path->pkey = cm_id_priv->pkey; path->sl = cm_lap_get_sl(lap_msg); path->mtu_selector = IB_SA_EQ; - /* mtu is same as in REQ */ + path->mtu = cm_id_priv->path_mtu; path->rate_selector = IB_SA_EQ; path->rate = cm_lap_get_packet_rate(lap_msg); path->packet_life_time_selector = IB_SA_EQ; @@ -2329,7 +2689,6 @@ static int cm_lap_handler(struct cm_work *work) struct cm_lap_msg *lap_msg; struct ib_cm_lap_event_param *param; struct ib_mad_send_buf *msg = NULL; - unsigned long flags; int ret; /* todo: verify LAP request and send reject APR if invalid. */ @@ -2341,17 +2700,20 @@ static int cm_lap_handler(struct cm_work *work) param = &work->cm_event.param.lap_rcvd; param->alternate_path = &work->path[0]; - cm_format_path_from_lap(param->alternate_path, lap_msg); + cm_format_path_from_lap(cm_id_priv, param->alternate_path, lap_msg); work->cm_event.private_data = &lap_msg->private_data; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state != IB_CM_ESTABLISHED) goto unlock; switch (cm_id_priv->id.lap_state) { + case IB_CM_LAP_UNINIT: case IB_CM_LAP_IDLE: break; case IB_CM_MRA_LAP_SENT: + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_LAP_COUNTER]); if (cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg)) goto unlock; @@ -2360,21 +2722,29 @@ static int cm_lap_handler(struct cm_work *work) cm_id_priv->service_timeout, cm_id_priv->private_data, cm_id_priv->private_data_len); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ib_post_send_mad(msg, NULL)) cm_free_msg(msg); goto deref; + case IB_CM_LAP_RCVD: + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_LAP_COUNTER]); + goto unlock; default: goto unlock; } cm_id_priv->id.lap_state = IB_CM_LAP_RCVD; cm_id_priv->tid = lap_msg->hdr.tid; + cm_init_av_for_response(work->port, work->mad_recv_wc->wc, + work->mad_recv_wc->recv_buf.grh, + &cm_id_priv->av); + cm_init_av_by_path(param->alternate_path, &cm_id_priv->alt_av); ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -2382,7 +2752,7 @@ static int cm_lap_handler(struct cm_work *work) cm_deref_id(cm_id_priv); return 0; -unlock: spin_unlock_irqrestore(&cm_id_priv->lock, flags); +unlock: spin_unlock_irq(&cm_id_priv->lock); deref: cm_deref_id(cm_id_priv); return -EINVAL; } @@ -2457,7 +2827,6 @@ static int cm_apr_handler(struct cm_work *work) { struct cm_id_private *cm_id_priv; struct cm_apr_msg *apr_msg; - unsigned long flags; int ret; apr_msg = (struct cm_apr_msg *)work->mad_recv_wc->recv_buf.mad; @@ -2471,11 +2840,11 @@ static int cm_apr_handler(struct cm_work *work) work->cm_event.param.apr_rcvd.info_len = apr_msg->info_length; work->cm_event.private_data = &apr_msg->private_data; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state != IB_CM_ESTABLISHED || (cm_id_priv->id.lap_state != IB_CM_LAP_SENT && cm_id_priv->id.lap_state != IB_CM_MRA_LAP_RCVD)) { - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); goto out; } cm_id_priv->id.lap_state = IB_CM_LAP_IDLE; @@ -2485,7 +2854,7 @@ static int cm_apr_handler(struct cm_work *work) ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -2501,28 +2870,29 @@ static int cm_timewait_handler(struct cm_work *work) { struct cm_timewait_info *timewait_info; struct cm_id_private *cm_id_priv; - unsigned long flags; int ret; timewait_info = (struct cm_timewait_info *)work; - cm_cleanup_timewait(timewait_info); + spin_lock_irq(&cm.lock); + list_del(&timewait_info->list); + spin_unlock_irq(&cm.lock); cm_id_priv = cm_acquire_id(timewait_info->work.local_id, timewait_info->work.remote_id); if (!cm_id_priv) return -EINVAL; - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state != IB_CM_TIMEWAIT || cm_id_priv->remote_qpn != timewait_info->remote_qpn) { - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); goto out; } cm_id_priv->id.state = IB_CM_IDLE; ret = atomic_inc_and_test(&cm_id_priv->work_count); if (!ret) list_add_tail(&work->list, &cm_id_priv->work_list); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); if (ret) cm_process_work(cm_id_priv, work); @@ -2541,7 +2911,7 @@ static void cm_format_sidr_req(struct cm_sidr_req_msg *sidr_req_msg, cm_format_mad_hdr(&sidr_req_msg->hdr, CM_SIDR_REQ_ATTR_ID, cm_form_tid(cm_id_priv, CM_MSG_SEQUENCE_SIDR)); sidr_req_msg->request_id = cm_id_priv->id.local_id; - sidr_req_msg->pkey = cpu_to_be16(param->pkey); + sidr_req_msg->pkey = param->path->pkey; sidr_req_msg->service_id = param->service_id; if (param->private_data && param->private_data_len) @@ -2567,7 +2937,7 @@ int ib_send_cm_sidr_req(struct ib_cm_id *cm_id, goto out; cm_id->service_id = param->service_id; - cm_id->service_mask = __constant_cpu_to_be64(~0ULL); + cm_id->service_mask = ~cpu_to_be64(0); cm_id_priv->timeout_ms = param->timeout_ms; cm_id_priv->max_cm_retries = param->max_cm_retries; ret = cm_alloc_msg(cm_id_priv, &msg); @@ -2619,9 +2989,8 @@ static int cm_sidr_req_handler(struct cm_work *work) struct cm_id_private *cm_id_priv, *cur_cm_id_priv; struct cm_sidr_req_msg *sidr_req_msg; struct ib_wc *wc; - unsigned long flags; - cm_id = ib_create_cm_id(work->port->cm_dev->device, NULL, NULL); + cm_id = ib_create_cm_id(work->port->cm_dev->ib_device, NULL, NULL); if (IS_ERR(cm_id)) return PTR_ERR(cm_id); cm_id_priv = container_of(cm_id, struct cm_id_private, id); @@ -2633,33 +3002,37 @@ static int cm_sidr_req_handler(struct cm_work *work) cm_id_priv->av.dgid.global.subnet_prefix = cpu_to_be64(wc->slid); cm_id_priv->av.dgid.global.interface_id = 0; cm_init_av_for_response(work->port, work->mad_recv_wc->wc, + work->mad_recv_wc->recv_buf.grh, &cm_id_priv->av); cm_id_priv->id.remote_id = sidr_req_msg->request_id; - cm_id_priv->id.state = IB_CM_SIDR_REQ_RCVD; cm_id_priv->tid = sidr_req_msg->hdr.tid; atomic_inc(&cm_id_priv->work_count); - spin_lock_irqsave(&cm.lock, flags); + spin_lock_irq(&cm.lock); cur_cm_id_priv = cm_insert_remote_sidr(cm_id_priv); if (cur_cm_id_priv) { - spin_unlock_irqrestore(&cm.lock, flags); + spin_unlock_irq(&cm.lock); + atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES]. + counter[CM_SIDR_REQ_COUNTER]); goto out; /* Duplicate message. */ } + cm_id_priv->id.state = IB_CM_SIDR_REQ_RCVD; cur_cm_id_priv = cm_find_listen(cm_id->device, - sidr_req_msg->service_id); + sidr_req_msg->service_id, + sidr_req_msg->private_data); if (!cur_cm_id_priv) { - rb_erase(&cm_id_priv->sidr_id_node, &cm.remote_sidr_table); - spin_unlock_irqrestore(&cm.lock, flags); - /* todo: reply with no match */ + spin_unlock_irq(&cm.lock); + cm_reject_sidr_req(cm_id_priv, IB_SIDR_UNSUPPORTED); goto out; /* No match. */ } atomic_inc(&cur_cm_id_priv->refcount); - spin_unlock_irqrestore(&cm.lock, flags); + atomic_inc(&cm_id_priv->refcount); + spin_unlock_irq(&cm.lock); cm_id_priv->id.cm_handler = cur_cm_id_priv->id.cm_handler; cm_id_priv->id.context = cur_cm_id_priv->id.context; cm_id_priv->id.service_id = sidr_req_msg->service_id; - cm_id_priv->id.service_mask = __constant_cpu_to_be64(~0ULL); + cm_id_priv->id.service_mask = ~cpu_to_be64(0); cm_format_sidr_req_event(work, &cur_cm_id_priv->id); cm_process_work(cm_id_priv, work); @@ -2755,7 +3128,6 @@ static int cm_sidr_rep_handler(struct cm_work *work) { struct cm_sidr_rep_msg *sidr_rep_msg; struct cm_id_private *cm_id_priv; - unsigned long flags; sidr_rep_msg = (struct cm_sidr_rep_msg *) work->mad_recv_wc->recv_buf.mad; @@ -2763,14 +3135,14 @@ static int cm_sidr_rep_handler(struct cm_work *work) if (!cm_id_priv) return -EINVAL; /* Unmatched reply. */ - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); if (cm_id_priv->id.state != IB_CM_SIDR_REQ_SENT) { - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); goto out; } cm_id_priv->id.state = IB_CM_IDLE; ib_cancel_mad(cm_id_priv->av.port->mad_agent, cm_id_priv->msg); - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); cm_format_sidr_rep_event(work); cm_process_work(cm_id_priv, work); @@ -2786,14 +3158,13 @@ static void cm_process_send_error(struct ib_mad_send_buf *msg, struct cm_id_private *cm_id_priv; struct ib_cm_event cm_event; enum ib_cm_state state; - unsigned long flags; int ret; memset(&cm_event, 0, sizeof cm_event); cm_id_priv = msg->context[0]; /* Discard old sends or ones without a response. */ - spin_lock_irqsave(&cm_id_priv->lock, flags); + spin_lock_irq(&cm_id_priv->lock); state = (enum ib_cm_state) (unsigned long) msg->context[1]; if (msg != cm_id_priv->msg || state != cm_id_priv->id.state) goto discard; @@ -2820,7 +3191,7 @@ static void cm_process_send_error(struct ib_mad_send_buf *msg, default: goto discard; } - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); cm_event.param.send_status = wc_status; /* No other events can occur on the cm_id at this point. */ @@ -2830,7 +3201,7 @@ static void cm_process_send_error(struct ib_mad_send_buf *msg, ib_destroy_cm_id(&cm_id_priv->id); return; discard: - spin_unlock_irqrestore(&cm_id_priv->lock, flags); + spin_unlock_irq(&cm_id_priv->lock); cm_free_msg(msg); } @@ -2838,6 +3209,27 @@ static void cm_send_handler(struct ib_mad_agent *mad_agent, struct ib_mad_send_wc *mad_send_wc) { struct ib_mad_send_buf *msg = mad_send_wc->send_buf; + struct cm_port *port; + u16 attr_index; + + port = mad_agent->context; + attr_index = be16_to_cpu(((struct ib_mad_hdr *) + msg->mad)->attr_id) - CM_ATTR_ID_OFFSET; + + /* + * If the send was in response to a received message (context[0] is not + * set to a cm_id), and is not a REJ, then it is a send that was + * manually retried. + */ + if (!msg->context[0] && (attr_index != CM_REJ_COUNTER)) + msg->retries = 1; + + atomic_long_add(1 + msg->retries, + &port->counter_group[CM_XMIT].counter[attr_index]); + if (msg->retries) + atomic_long_add(msg->retries, + &port->counter_group[CM_XMIT_RETRIES]. + counter[attr_index]); switch (mad_send_wc->status) { case IB_WC_SUCCESS: @@ -2853,9 +3245,9 @@ static void cm_send_handler(struct ib_mad_agent *mad_agent, } } -static void cm_work_handler(void *data) +static void cm_work_handler(struct work_struct *_work) { - struct cm_work *work = data; + struct cm_work *work = container_of(_work, struct cm_work, work.work); int ret; switch (work->cm_event.event) { @@ -2906,7 +3298,7 @@ static void cm_work_handler(void *data) cm_free_work(work); } -int ib_cm_establish(struct ib_cm_id *cm_id) +static int cm_establish(struct ib_cm_id *cm_id) { struct cm_id_private *cm_id_priv; struct cm_work *work; @@ -2945,22 +3337,61 @@ int ib_cm_establish(struct ib_cm_id *cm_id) * we need to find the cm_id once we're in the context of the * worker thread, rather than holding a reference on it. */ - INIT_WORK(&work->work, cm_work_handler, work); + INIT_DELAYED_WORK(&work->work, cm_work_handler); work->local_id = cm_id->local_id; work->remote_id = cm_id->remote_id; work->mad_recv_wc = NULL; work->cm_event.event = IB_CM_USER_ESTABLISHED; - queue_work(cm.wq, &work->work); + queue_delayed_work(cm.wq, &work->work, 0); out: return ret; } -EXPORT_SYMBOL(ib_cm_establish); + +static int cm_migrate(struct ib_cm_id *cm_id) +{ + struct cm_id_private *cm_id_priv; + unsigned long flags; + int ret = 0; + + cm_id_priv = container_of(cm_id, struct cm_id_private, id); + spin_lock_irqsave(&cm_id_priv->lock, flags); + if (cm_id->state == IB_CM_ESTABLISHED && + (cm_id->lap_state == IB_CM_LAP_UNINIT || + cm_id->lap_state == IB_CM_LAP_IDLE)) { + cm_id->lap_state = IB_CM_LAP_IDLE; + cm_id_priv->av = cm_id_priv->alt_av; + } else + ret = -EINVAL; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + return ret; +} + +int ib_cm_notify(struct ib_cm_id *cm_id, enum ib_event_type event) +{ + int ret; + + switch (event) { + case IB_EVENT_COMM_EST: + ret = cm_establish(cm_id); + break; + case IB_EVENT_PATH_MIG: + ret = cm_migrate(cm_id); + break; + default: + ret = -EINVAL; + } + return ret; +} +EXPORT_SYMBOL(ib_cm_notify); static void cm_recv_handler(struct ib_mad_agent *mad_agent, struct ib_mad_recv_wc *mad_recv_wc) { + struct cm_port *port = mad_agent->context; struct cm_work *work; enum ib_cm_event_type event; + u16 attr_id; int paths = 0; switch (mad_recv_wc->recv_buf.mad->mad_hdr.attr_id) { @@ -3005,6 +3436,10 @@ static void cm_recv_handler(struct ib_mad_agent *mad_agent, return; } + attr_id = be16_to_cpu(mad_recv_wc->recv_buf.mad->mad_hdr.attr_id); + atomic_long_inc(&port->counter_group[CM_RECV]. + counter[attr_id - CM_ATTR_ID_OFFSET]); + work = kmalloc(sizeof *work + sizeof(struct ib_sa_path_rec) * paths, GFP_KERNEL); if (!work) { @@ -3012,11 +3447,11 @@ static void cm_recv_handler(struct ib_mad_agent *mad_agent, return; } - INIT_WORK(&work->work, cm_work_handler, work); + INIT_DELAYED_WORK(&work->work, cm_work_handler); work->cm_event.event = event; work->mad_recv_wc = mad_recv_wc; - work->port = (struct cm_port *)mad_agent->context; - queue_work(cm.wq, &work->work); + work->port = port; + queue_delayed_work(cm.wq, &work->work, 0); } static int cm_init_qp_init_attr(struct cm_id_private *cm_id_priv, @@ -3039,10 +3474,10 @@ static int cm_init_qp_init_attr(struct cm_id_private *cm_id_priv, case IB_CM_ESTABLISHED: *qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS | IB_QP_PKEY_INDEX | IB_QP_PORT; - qp_attr->qp_access_flags = IB_ACCESS_LOCAL_WRITE | - IB_ACCESS_REMOTE_WRITE; + qp_attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE; if (cm_id_priv->responder_resources) - qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_READ; + qp_attr->qp_access_flags |= IB_ACCESS_REMOTE_READ | + IB_ACCESS_REMOTE_ATOMIC; qp_attr->pkey_index = cm_id_priv->av.pkey_index; qp_attr->port_num = cm_id_priv->av.port->port_num; ret = 0; @@ -3074,10 +3509,37 @@ static int cm_init_qp_rtr_attr(struct cm_id_private *cm_id_priv, *qp_attr_mask = IB_QP_STATE | IB_QP_AV | IB_QP_PATH_MTU | IB_QP_DEST_QPN | IB_QP_RQ_PSN; qp_attr->ah_attr = cm_id_priv->av.ah_attr; + if (!cm_id_priv->av.valid) { + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + return -EINVAL; + } + if (cm_id_priv->av.ah_attr.vlan_id != 0xffff) { + qp_attr->vlan_id = cm_id_priv->av.ah_attr.vlan_id; + *qp_attr_mask |= IB_QP_VID; + } + if (!is_zero_ether_addr(cm_id_priv->av.smac)) { + memcpy(qp_attr->smac, cm_id_priv->av.smac, + sizeof(qp_attr->smac)); + *qp_attr_mask |= IB_QP_SMAC; + } + if (cm_id_priv->alt_av.valid) { + if (cm_id_priv->alt_av.ah_attr.vlan_id != 0xffff) { + qp_attr->alt_vlan_id = + cm_id_priv->alt_av.ah_attr.vlan_id; + *qp_attr_mask |= IB_QP_ALT_VID; + } + if (!is_zero_ether_addr(cm_id_priv->alt_av.smac)) { + memcpy(qp_attr->alt_smac, + cm_id_priv->alt_av.smac, + sizeof(qp_attr->alt_smac)); + *qp_attr_mask |= IB_QP_ALT_SMAC; + } + } qp_attr->path_mtu = cm_id_priv->path_mtu; qp_attr->dest_qp_num = be32_to_cpu(cm_id_priv->remote_qpn); qp_attr->rq_psn = be32_to_cpu(cm_id_priv->rq_psn); - if (cm_id_priv->qp_type == IB_QPT_RC) { + if (cm_id_priv->qp_type == IB_QPT_RC || + cm_id_priv->qp_type == IB_QPT_XRC_TGT) { *qp_attr_mask |= IB_QP_MAX_DEST_RD_ATOMIC | IB_QP_MIN_RNR_TIMER; qp_attr->max_dest_rd_atomic = @@ -3086,6 +3548,9 @@ static int cm_init_qp_rtr_attr(struct cm_id_private *cm_id_priv, } if (cm_id_priv->alt_av.ah_attr.dlid) { *qp_attr_mask |= IB_QP_ALT_PATH; + qp_attr->alt_port_num = cm_id_priv->alt_av.port->port_num; + qp_attr->alt_pkey_index = cm_id_priv->alt_av.pkey_index; + qp_attr->alt_timeout = cm_id_priv->alt_av.timeout; qp_attr->alt_ah_attr = cm_id_priv->alt_av.ah_attr; } ret = 0; @@ -3107,24 +3572,44 @@ static int cm_init_qp_rts_attr(struct cm_id_private *cm_id_priv, spin_lock_irqsave(&cm_id_priv->lock, flags); switch (cm_id_priv->id.state) { + /* Allow transition to RTS before sending REP */ + case IB_CM_REQ_RCVD: + case IB_CM_MRA_REQ_SENT: + case IB_CM_REP_RCVD: case IB_CM_MRA_REP_SENT: case IB_CM_REP_SENT: case IB_CM_MRA_REP_RCVD: case IB_CM_ESTABLISHED: - *qp_attr_mask = IB_QP_STATE | IB_QP_SQ_PSN; - qp_attr->sq_psn = be32_to_cpu(cm_id_priv->sq_psn); - if (cm_id_priv->qp_type == IB_QPT_RC) { - *qp_attr_mask |= IB_QP_TIMEOUT | IB_QP_RETRY_CNT | - IB_QP_RNR_RETRY | - IB_QP_MAX_QP_RD_ATOMIC; - qp_attr->timeout = cm_id_priv->local_ack_timeout; - qp_attr->retry_cnt = cm_id_priv->retry_count; - qp_attr->rnr_retry = cm_id_priv->rnr_retry_count; - qp_attr->max_rd_atomic = cm_id_priv->initiator_depth; - } - if (cm_id_priv->alt_av.ah_attr.dlid) { - *qp_attr_mask |= IB_QP_PATH_MIG_STATE; + if (cm_id_priv->id.lap_state == IB_CM_LAP_UNINIT) { + *qp_attr_mask = IB_QP_STATE | IB_QP_SQ_PSN; + qp_attr->sq_psn = be32_to_cpu(cm_id_priv->sq_psn); + switch (cm_id_priv->qp_type) { + case IB_QPT_RC: + case IB_QPT_XRC_INI: + *qp_attr_mask |= IB_QP_RETRY_CNT | IB_QP_RNR_RETRY | + IB_QP_MAX_QP_RD_ATOMIC; + qp_attr->retry_cnt = cm_id_priv->retry_count; + qp_attr->rnr_retry = cm_id_priv->rnr_retry_count; + qp_attr->max_rd_atomic = cm_id_priv->initiator_depth; + /* fall through */ + case IB_QPT_XRC_TGT: + *qp_attr_mask |= IB_QP_TIMEOUT; + qp_attr->timeout = cm_id_priv->av.timeout; + break; + default: + break; + } + if (cm_id_priv->alt_av.ah_attr.dlid) { + *qp_attr_mask |= IB_QP_PATH_MIG_STATE; + qp_attr->path_mig_state = IB_MIG_REARM; + } + } else { + *qp_attr_mask = IB_QP_ALT_PATH | IB_QP_PATH_MIG_STATE; + qp_attr->alt_port_num = cm_id_priv->alt_av.port->port_num; + qp_attr->alt_pkey_index = cm_id_priv->alt_av.pkey_index; + qp_attr->alt_timeout = cm_id_priv->alt_av.timeout; + qp_attr->alt_ah_attr = cm_id_priv->alt_av.ah_attr; qp_attr->path_mig_state = IB_MIG_REARM; } ret = 0; @@ -3163,7 +3648,106 @@ int ib_cm_init_qp_attr(struct ib_cm_id *cm_id, } EXPORT_SYMBOL(ib_cm_init_qp_attr); -static void cm_add_one(struct ib_device *device) +static void cm_get_ack_delay(struct cm_device *cm_dev) +{ + struct ib_device_attr attr; + + if (ib_query_device(cm_dev->ib_device, &attr)) + cm_dev->ack_delay = 0; /* acks will rely on packet life time */ + else + cm_dev->ack_delay = attr.local_ca_ack_delay; +} + +static ssize_t cm_show_counter(struct kobject *obj, struct attribute *attr, + char *buf) +{ + struct cm_counter_group *group; + struct cm_counter_attribute *cm_attr; + + group = container_of(obj, struct cm_counter_group, obj); + cm_attr = container_of(attr, struct cm_counter_attribute, attr); + + return sprintf(buf, "%ld\n", + atomic_long_read(&group->counter[cm_attr->index])); +} + +static const struct sysfs_ops cm_counter_ops = { + .show = cm_show_counter +}; + +static struct kobj_type cm_counter_obj_type = { + .sysfs_ops = &cm_counter_ops, + .default_attrs = cm_counter_default_attrs +}; + +static void cm_release_port_obj(struct kobject *obj) +{ + struct cm_port *cm_port; + + cm_port = container_of(obj, struct cm_port, port_obj); + kfree(cm_port); +} + +static struct kobj_type cm_port_obj_type = { + .release = cm_release_port_obj +}; + +static char *cm_devnode(struct device *dev, umode_t *mode) +{ + if (mode) + *mode = 0666; + return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev)); +} + +struct class cm_class = { + .owner = THIS_MODULE, + .name = "infiniband_cm", + .devnode = cm_devnode, +}; +EXPORT_SYMBOL(cm_class); + +static int cm_create_port_fs(struct cm_port *port) +{ + int i, ret; + + ret = kobject_init_and_add(&port->port_obj, &cm_port_obj_type, + &port->cm_dev->device->kobj, + "%d", port->port_num); + if (ret) { + kfree(port); + return ret; + } + + for (i = 0; i < CM_COUNTER_GROUPS; i++) { + ret = kobject_init_and_add(&port->counter_group[i].obj, + &cm_counter_obj_type, + &port->port_obj, + "%s", counter_group_names[i]); + if (ret) + goto error; + } + + return 0; + +error: + while (i--) + kobject_put(&port->counter_group[i].obj); + kobject_put(&port->port_obj); + return ret; + +} + +static void cm_remove_port_fs(struct cm_port *port) +{ + int i; + + for (i = 0; i < CM_COUNTER_GROUPS; i++) + kobject_put(&port->counter_group[i].obj); + + kobject_put(&port->port_obj); +} + +static void cm_add_one(struct ib_device *ib_device) { struct cm_device *cm_dev; struct cm_port *port; @@ -3178,20 +3762,40 @@ static void cm_add_one(struct ib_device *device) int ret; u8 i; - cm_dev = kmalloc(sizeof(*cm_dev) + sizeof(*port) * - device->phys_port_cnt, GFP_KERNEL); + if (rdma_node_get_transport(ib_device->node_type) != RDMA_TRANSPORT_IB) + return; + + cm_dev = kzalloc(sizeof(*cm_dev) + sizeof(*port) * + ib_device->phys_port_cnt, GFP_KERNEL); if (!cm_dev) return; - cm_dev->device = device; - cm_dev->ca_guid = device->node_guid; + cm_dev->ib_device = ib_device; + cm_get_ack_delay(cm_dev); + + cm_dev->device = device_create(&cm_class, &ib_device->dev, + MKDEV(0, 0), NULL, + "%s", ib_device->name); + if (IS_ERR(cm_dev->device)) { + kfree(cm_dev); + return; + } set_bit(IB_MGMT_METHOD_SEND, reg_req.method_mask); - for (i = 1; i <= device->phys_port_cnt; i++) { - port = &cm_dev->port[i-1]; + for (i = 1; i <= ib_device->phys_port_cnt; i++) { + port = kzalloc(sizeof *port, GFP_KERNEL); + if (!port) + goto error1; + + cm_dev->port[i-1] = port; port->cm_dev = cm_dev; port->port_num = i; - port->mad_agent = ib_register_mad_agent(device, i, + + ret = cm_create_port_fs(port); + if (ret) + goto error1; + + port->mad_agent = ib_register_mad_agent(ib_device, i, IB_QPT_GSI, ®_req, 0, @@ -3199,33 +3803,37 @@ static void cm_add_one(struct ib_device *device) cm_recv_handler, port); if (IS_ERR(port->mad_agent)) - goto error1; + goto error2; - ret = ib_modify_port(device, i, 0, &port_modify); + ret = ib_modify_port(ib_device, i, 0, &port_modify); if (ret) - goto error2; + goto error3; } - ib_set_client_data(device, &cm_client, cm_dev); + ib_set_client_data(ib_device, &cm_client, cm_dev); write_lock_irqsave(&cm.device_lock, flags); list_add_tail(&cm_dev->list, &cm.device_list); write_unlock_irqrestore(&cm.device_lock, flags); return; -error2: +error3: ib_unregister_mad_agent(port->mad_agent); +error2: + cm_remove_port_fs(port); error1: port_modify.set_port_cap_mask = 0; port_modify.clr_port_cap_mask = IB_PORT_CM_SUP; while (--i) { - port = &cm_dev->port[i-1]; - ib_modify_port(device, port->port_num, 0, &port_modify); + port = cm_dev->port[i-1]; + ib_modify_port(ib_device, port->port_num, 0, &port_modify); ib_unregister_mad_agent(port->mad_agent); + cm_remove_port_fs(port); } + device_unregister(cm_dev->device); kfree(cm_dev); } -static void cm_remove_one(struct ib_device *device) +static void cm_remove_one(struct ib_device *ib_device) { struct cm_device *cm_dev; struct cm_port *port; @@ -3235,7 +3843,7 @@ static void cm_remove_one(struct ib_device *device) unsigned long flags; int i; - cm_dev = ib_get_client_data(device, &cm_client); + cm_dev = ib_get_client_data(ib_device, &cm_client); if (!cm_dev) return; @@ -3243,11 +3851,14 @@ static void cm_remove_one(struct ib_device *device) list_del(&cm_dev->list); write_unlock_irqrestore(&cm.device_lock, flags); - for (i = 1; i <= device->phys_port_cnt; i++) { - port = &cm_dev->port[i-1]; - ib_modify_port(device, port->port_num, 0, &port_modify); + for (i = 1; i <= ib_device->phys_port_cnt; i++) { + port = cm_dev->port[i-1]; + ib_modify_port(ib_device, port->port_num, 0, &port_modify); ib_unregister_mad_agent(port->mad_agent); + flush_workqueue(cm.wq); + cm_remove_port_fs(port); } + device_unregister(cm_dev->device); kfree(cm_dev); } @@ -3260,32 +3871,58 @@ static int __init ib_cm_init(void) rwlock_init(&cm.device_lock); spin_lock_init(&cm.lock); cm.listen_service_table = RB_ROOT; - cm.listen_service_id = __constant_be64_to_cpu(IB_CM_ASSIGN_SERVICE_ID); + cm.listen_service_id = be64_to_cpu(IB_CM_ASSIGN_SERVICE_ID); cm.remote_id_table = RB_ROOT; cm.remote_qp_table = RB_ROOT; cm.remote_sidr_table = RB_ROOT; idr_init(&cm.local_id_table); - idr_pre_get(&cm.local_id_table, GFP_KERNEL); + get_random_bytes(&cm.random_id_operand, sizeof cm.random_id_operand); + INIT_LIST_HEAD(&cm.timewait_list); + + ret = class_register(&cm_class); + if (ret) { + ret = -ENOMEM; + goto error1; + } cm.wq = create_workqueue("ib_cm"); - if (!cm.wq) - return -ENOMEM; + if (!cm.wq) { + ret = -ENOMEM; + goto error2; + } ret = ib_register_client(&cm_client); if (ret) - goto error; + goto error3; return 0; -error: +error3: destroy_workqueue(cm.wq); +error2: + class_unregister(&cm_class); +error1: + idr_destroy(&cm.local_id_table); return ret; } static void __exit ib_cm_cleanup(void) { - flush_workqueue(cm.wq); - destroy_workqueue(cm.wq); + struct cm_timewait_info *timewait_info, *tmp; + + spin_lock_irq(&cm.lock); + list_for_each_entry(timewait_info, &cm.timewait_list, list) + cancel_delayed_work(&timewait_info->work.work); + spin_unlock_irq(&cm.lock); + ib_unregister_client(&cm_client); + destroy_workqueue(cm.wq); + + list_for_each_entry_safe(timewait_info, tmp, &cm.timewait_list, list) { + list_del(&timewait_info->list); + kfree(timewait_info); + } + + class_unregister(&cm_class); idr_destroy(&cm.local_id_table); } diff --git a/drivers/infiniband/core/cm_msgs.h b/drivers/infiniband/core/cm_msgs.h index 4d3aee90c24..be068f47e47 100644 --- a/drivers/infiniband/core/cm_msgs.h +++ b/drivers/infiniband/core/cm_msgs.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2004 Intel Corporation. All rights reserved. + * Copyright (c) 2004, 2011 Intel Corporation. All rights reserved. * Copyright (c) 2004 Topspin Corporation. All rights reserved. * Copyright (c) 2004 Voltaire Corporation. All rights reserved. * @@ -35,6 +35,7 @@ #define CM_MSGS_H #include <rdma/ib_mad.h> +#include <rdma/ib_cm.h> /* * Parameters to routines below should be in network-byte order, and values @@ -43,18 +44,6 @@ #define IB_CM_CLASS_VERSION 2 /* IB specification 1.2 */ -#define CM_REQ_ATTR_ID __constant_htons(0x0010) -#define CM_MRA_ATTR_ID __constant_htons(0x0011) -#define CM_REJ_ATTR_ID __constant_htons(0x0012) -#define CM_REP_ATTR_ID __constant_htons(0x0013) -#define CM_RTU_ATTR_ID __constant_htons(0x0014) -#define CM_DREQ_ATTR_ID __constant_htons(0x0015) -#define CM_DREP_ATTR_ID __constant_htons(0x0016) -#define CM_SIDR_REQ_ATTR_ID __constant_htons(0x0017) -#define CM_SIDR_REP_ATTR_ID __constant_htons(0x0018) -#define CM_LAP_ATTR_ID __constant_htons(0x0019) -#define CM_APR_ATTR_ID __constant_htons(0x001A) - enum cm_msg_sequence { CM_MSG_SEQUENCE_REQ, CM_MSG_SEQUENCE_LAP, @@ -85,7 +74,7 @@ struct cm_req_msg { __be16 pkey; /* path MTU:4, RDC exists:1, RNR retry count:3. */ u8 offset50; - /* max CM Retries:4, SRQ:1, rsvd:3 */ + /* max CM Retries:4, SRQ:1, extended transport type:3 */ u8 offset51; __be16 primary_local_lid; @@ -174,6 +163,11 @@ static inline enum ib_qp_type cm_req_get_qp_type(struct cm_req_msg *req_msg) switch(transport_type) { case 0: return IB_QPT_RC; case 1: return IB_QPT_UC; + case 3: + switch (req_msg->offset51 & 0x7) { + case 1: return IB_QPT_XRC_TGT; + default: return 0; + } default: return 0; } } @@ -187,6 +181,12 @@ static inline void cm_req_set_qp_type(struct cm_req_msg *req_msg, req_msg->offset40) & 0xFFFFFFF9) | 0x2); break; + case IB_QPT_XRC_INI: + req_msg->offset40 = cpu_to_be32((be32_to_cpu( + req_msg->offset40) & + 0xFFFFFFF9) | 0x6); + req_msg->offset51 = (req_msg->offset51 & 0xF8) | 1; + break; default: req_msg->offset40 = cpu_to_be32(be32_to_cpu( req_msg->offset40) & @@ -526,6 +526,23 @@ static inline void cm_rep_set_local_qpn(struct cm_rep_msg *rep_msg, __be32 qpn) (be32_to_cpu(rep_msg->offset12) & 0x000000FF)); } +static inline __be32 cm_rep_get_local_eecn(struct cm_rep_msg *rep_msg) +{ + return cpu_to_be32(be32_to_cpu(rep_msg->offset16) >> 8); +} + +static inline void cm_rep_set_local_eecn(struct cm_rep_msg *rep_msg, __be32 eecn) +{ + rep_msg->offset16 = cpu_to_be32((be32_to_cpu(eecn) << 8) | + (be32_to_cpu(rep_msg->offset16) & 0x000000FF)); +} + +static inline __be32 cm_rep_get_qpn(struct cm_rep_msg *rep_msg, enum ib_qp_type qp_type) +{ + return (qp_type == IB_QPT_XRC_INI) ? + cm_rep_get_local_eecn(rep_msg) : cm_rep_get_local_qpn(rep_msg); +} + static inline __be32 cm_rep_get_starting_psn(struct cm_rep_msg *rep_msg) { return cpu_to_be32(be32_to_cpu(rep_msg->offset20) >> 8); @@ -770,6 +787,7 @@ struct cm_apr_msg { u8 info_length; u8 ap_status; + __be16 rsvd; u8 info[IB_CM_APR_INFO_LENGTH]; u8 private_data[IB_CM_APR_PRIVATE_DATA_SIZE]; diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c new file mode 100644 index 00000000000..d570030d899 --- /dev/null +++ b/drivers/infiniband/core/cma.c @@ -0,0 +1,3703 @@ +/* + * Copyright (c) 2005 Voltaire Inc. All rights reserved. + * Copyright (c) 2002-2005, Network Appliance, Inc. All rights reserved. + * Copyright (c) 1999-2005, Mellanox Technologies, Inc. All rights reserved. + * Copyright (c) 2005-2006 Intel Corporation. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/completion.h> +#include <linux/in.h> +#include <linux/in6.h> +#include <linux/mutex.h> +#include <linux/random.h> +#include <linux/idr.h> +#include <linux/inetdevice.h> +#include <linux/slab.h> +#include <linux/module.h> +#include <net/route.h> + +#include <net/tcp.h> +#include <net/ipv6.h> + +#include <rdma/rdma_cm.h> +#include <rdma/rdma_cm_ib.h> +#include <rdma/rdma_netlink.h> +#include <rdma/ib.h> +#include <rdma/ib_cache.h> +#include <rdma/ib_cm.h> +#include <rdma/ib_sa.h> +#include <rdma/iw_cm.h> + +MODULE_AUTHOR("Sean Hefty"); +MODULE_DESCRIPTION("Generic RDMA CM Agent"); +MODULE_LICENSE("Dual BSD/GPL"); + +#define CMA_CM_RESPONSE_TIMEOUT 20 +#define CMA_MAX_CM_RETRIES 15 +#define CMA_CM_MRA_SETTING (IB_CM_MRA_FLAG_DELAY | 24) +#define CMA_IBOE_PACKET_LIFETIME 18 + +static void cma_add_one(struct ib_device *device); +static void cma_remove_one(struct ib_device *device); + +static struct ib_client cma_client = { + .name = "cma", + .add = cma_add_one, + .remove = cma_remove_one +}; + +static struct ib_sa_client sa_client; +static struct rdma_addr_client addr_client; +static LIST_HEAD(dev_list); +static LIST_HEAD(listen_any_list); +static DEFINE_MUTEX(lock); +static struct workqueue_struct *cma_wq; +static DEFINE_IDR(tcp_ps); +static DEFINE_IDR(udp_ps); +static DEFINE_IDR(ipoib_ps); +static DEFINE_IDR(ib_ps); + +struct cma_device { + struct list_head list; + struct ib_device *device; + struct completion comp; + atomic_t refcount; + struct list_head id_list; +}; + +struct rdma_bind_list { + struct idr *ps; + struct hlist_head owners; + unsigned short port; +}; + +enum { + CMA_OPTION_AFONLY, +}; + +/* + * Device removal can occur at anytime, so we need extra handling to + * serialize notifying the user of device removal with other callbacks. + * We do this by disabling removal notification while a callback is in process, + * and reporting it after the callback completes. + */ +struct rdma_id_private { + struct rdma_cm_id id; + + struct rdma_bind_list *bind_list; + struct hlist_node node; + struct list_head list; /* listen_any_list or cma_device.list */ + struct list_head listen_list; /* per device listens */ + struct cma_device *cma_dev; + struct list_head mc_list; + + int internal_id; + enum rdma_cm_state state; + spinlock_t lock; + struct mutex qp_mutex; + + struct completion comp; + atomic_t refcount; + struct mutex handler_mutex; + + int backlog; + int timeout_ms; + struct ib_sa_query *query; + int query_id; + union { + struct ib_cm_id *ib; + struct iw_cm_id *iw; + } cm_id; + + u32 seq_num; + u32 qkey; + u32 qp_num; + pid_t owner; + u32 options; + u8 srq; + u8 tos; + u8 reuseaddr; + u8 afonly; +}; + +struct cma_multicast { + struct rdma_id_private *id_priv; + union { + struct ib_sa_multicast *ib; + } multicast; + struct list_head list; + void *context; + struct sockaddr_storage addr; + struct kref mcref; +}; + +struct cma_work { + struct work_struct work; + struct rdma_id_private *id; + enum rdma_cm_state old_state; + enum rdma_cm_state new_state; + struct rdma_cm_event event; +}; + +struct cma_ndev_work { + struct work_struct work; + struct rdma_id_private *id; + struct rdma_cm_event event; +}; + +struct iboe_mcast_work { + struct work_struct work; + struct rdma_id_private *id; + struct cma_multicast *mc; +}; + +union cma_ip_addr { + struct in6_addr ip6; + struct { + __be32 pad[3]; + __be32 addr; + } ip4; +}; + +struct cma_hdr { + u8 cma_version; + u8 ip_version; /* IP version: 7:4 */ + __be16 port; + union cma_ip_addr src_addr; + union cma_ip_addr dst_addr; +}; + +#define CMA_VERSION 0x00 + +static int cma_comp(struct rdma_id_private *id_priv, enum rdma_cm_state comp) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&id_priv->lock, flags); + ret = (id_priv->state == comp); + spin_unlock_irqrestore(&id_priv->lock, flags); + return ret; +} + +static int cma_comp_exch(struct rdma_id_private *id_priv, + enum rdma_cm_state comp, enum rdma_cm_state exch) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&id_priv->lock, flags); + if ((ret = (id_priv->state == comp))) + id_priv->state = exch; + spin_unlock_irqrestore(&id_priv->lock, flags); + return ret; +} + +static enum rdma_cm_state cma_exch(struct rdma_id_private *id_priv, + enum rdma_cm_state exch) +{ + unsigned long flags; + enum rdma_cm_state old; + + spin_lock_irqsave(&id_priv->lock, flags); + old = id_priv->state; + id_priv->state = exch; + spin_unlock_irqrestore(&id_priv->lock, flags); + return old; +} + +static inline u8 cma_get_ip_ver(struct cma_hdr *hdr) +{ + return hdr->ip_version >> 4; +} + +static inline void cma_set_ip_ver(struct cma_hdr *hdr, u8 ip_ver) +{ + hdr->ip_version = (ip_ver << 4) | (hdr->ip_version & 0xF); +} + +static void cma_attach_to_dev(struct rdma_id_private *id_priv, + struct cma_device *cma_dev) +{ + atomic_inc(&cma_dev->refcount); + id_priv->cma_dev = cma_dev; + id_priv->id.device = cma_dev->device; + id_priv->id.route.addr.dev_addr.transport = + rdma_node_get_transport(cma_dev->device->node_type); + list_add_tail(&id_priv->list, &cma_dev->id_list); +} + +static inline void cma_deref_dev(struct cma_device *cma_dev) +{ + if (atomic_dec_and_test(&cma_dev->refcount)) + complete(&cma_dev->comp); +} + +static inline void release_mc(struct kref *kref) +{ + struct cma_multicast *mc = container_of(kref, struct cma_multicast, mcref); + + kfree(mc->multicast.ib); + kfree(mc); +} + +static void cma_release_dev(struct rdma_id_private *id_priv) +{ + mutex_lock(&lock); + list_del(&id_priv->list); + cma_deref_dev(id_priv->cma_dev); + id_priv->cma_dev = NULL; + mutex_unlock(&lock); +} + +static inline struct sockaddr *cma_src_addr(struct rdma_id_private *id_priv) +{ + return (struct sockaddr *) &id_priv->id.route.addr.src_addr; +} + +static inline struct sockaddr *cma_dst_addr(struct rdma_id_private *id_priv) +{ + return (struct sockaddr *) &id_priv->id.route.addr.dst_addr; +} + +static inline unsigned short cma_family(struct rdma_id_private *id_priv) +{ + return id_priv->id.route.addr.src_addr.ss_family; +} + +static int cma_set_qkey(struct rdma_id_private *id_priv, u32 qkey) +{ + struct ib_sa_mcmember_rec rec; + int ret = 0; + + if (id_priv->qkey) { + if (qkey && id_priv->qkey != qkey) + return -EINVAL; + return 0; + } + + if (qkey) { + id_priv->qkey = qkey; + return 0; + } + + switch (id_priv->id.ps) { + case RDMA_PS_UDP: + case RDMA_PS_IB: + id_priv->qkey = RDMA_UDP_QKEY; + break; + case RDMA_PS_IPOIB: + ib_addr_get_mgid(&id_priv->id.route.addr.dev_addr, &rec.mgid); + ret = ib_sa_get_mcmember_rec(id_priv->id.device, + id_priv->id.port_num, &rec.mgid, + &rec); + if (!ret) + id_priv->qkey = be32_to_cpu(rec.qkey); + break; + default: + break; + } + return ret; +} + +static void cma_translate_ib(struct sockaddr_ib *sib, struct rdma_dev_addr *dev_addr) +{ + dev_addr->dev_type = ARPHRD_INFINIBAND; + rdma_addr_set_sgid(dev_addr, (union ib_gid *) &sib->sib_addr); + ib_addr_set_pkey(dev_addr, ntohs(sib->sib_pkey)); +} + +static int cma_translate_addr(struct sockaddr *addr, struct rdma_dev_addr *dev_addr) +{ + int ret; + + if (addr->sa_family != AF_IB) { + ret = rdma_translate_ip(addr, dev_addr, NULL); + } else { + cma_translate_ib((struct sockaddr_ib *) addr, dev_addr); + ret = 0; + } + + return ret; +} + +static int cma_acquire_dev(struct rdma_id_private *id_priv, + struct rdma_id_private *listen_id_priv) +{ + struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; + struct cma_device *cma_dev; + union ib_gid gid, iboe_gid; + int ret = -ENODEV; + u8 port, found_port; + enum rdma_link_layer dev_ll = dev_addr->dev_type == ARPHRD_INFINIBAND ? + IB_LINK_LAYER_INFINIBAND : IB_LINK_LAYER_ETHERNET; + + if (dev_ll != IB_LINK_LAYER_INFINIBAND && + id_priv->id.ps == RDMA_PS_IPOIB) + return -EINVAL; + + mutex_lock(&lock); + rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.src_addr, + &iboe_gid); + + memcpy(&gid, dev_addr->src_dev_addr + + rdma_addr_gid_offset(dev_addr), sizeof gid); + if (listen_id_priv && + rdma_port_get_link_layer(listen_id_priv->id.device, + listen_id_priv->id.port_num) == dev_ll) { + cma_dev = listen_id_priv->cma_dev; + port = listen_id_priv->id.port_num; + if (rdma_node_get_transport(cma_dev->device->node_type) == RDMA_TRANSPORT_IB && + rdma_port_get_link_layer(cma_dev->device, port) == IB_LINK_LAYER_ETHERNET) + ret = ib_find_cached_gid(cma_dev->device, &iboe_gid, + &found_port, NULL); + else + ret = ib_find_cached_gid(cma_dev->device, &gid, + &found_port, NULL); + + if (!ret && (port == found_port)) { + id_priv->id.port_num = found_port; + goto out; + } + } + list_for_each_entry(cma_dev, &dev_list, list) { + for (port = 1; port <= cma_dev->device->phys_port_cnt; ++port) { + if (listen_id_priv && + listen_id_priv->cma_dev == cma_dev && + listen_id_priv->id.port_num == port) + continue; + if (rdma_port_get_link_layer(cma_dev->device, port) == dev_ll) { + if (rdma_node_get_transport(cma_dev->device->node_type) == RDMA_TRANSPORT_IB && + rdma_port_get_link_layer(cma_dev->device, port) == IB_LINK_LAYER_ETHERNET) + ret = ib_find_cached_gid(cma_dev->device, &iboe_gid, &found_port, NULL); + else + ret = ib_find_cached_gid(cma_dev->device, &gid, &found_port, NULL); + + if (!ret && (port == found_port)) { + id_priv->id.port_num = found_port; + goto out; + } + } + } + } + +out: + if (!ret) + cma_attach_to_dev(id_priv, cma_dev); + + mutex_unlock(&lock); + return ret; +} + +/* + * Select the source IB device and address to reach the destination IB address. + */ +static int cma_resolve_ib_dev(struct rdma_id_private *id_priv) +{ + struct cma_device *cma_dev, *cur_dev; + struct sockaddr_ib *addr; + union ib_gid gid, sgid, *dgid; + u16 pkey, index; + u8 p; + int i; + + cma_dev = NULL; + addr = (struct sockaddr_ib *) cma_dst_addr(id_priv); + dgid = (union ib_gid *) &addr->sib_addr; + pkey = ntohs(addr->sib_pkey); + + list_for_each_entry(cur_dev, &dev_list, list) { + if (rdma_node_get_transport(cur_dev->device->node_type) != RDMA_TRANSPORT_IB) + continue; + + for (p = 1; p <= cur_dev->device->phys_port_cnt; ++p) { + if (ib_find_cached_pkey(cur_dev->device, p, pkey, &index)) + continue; + + for (i = 0; !ib_get_cached_gid(cur_dev->device, p, i, &gid); i++) { + if (!memcmp(&gid, dgid, sizeof(gid))) { + cma_dev = cur_dev; + sgid = gid; + id_priv->id.port_num = p; + goto found; + } + + if (!cma_dev && (gid.global.subnet_prefix == + dgid->global.subnet_prefix)) { + cma_dev = cur_dev; + sgid = gid; + id_priv->id.port_num = p; + } + } + } + } + + if (!cma_dev) + return -ENODEV; + +found: + cma_attach_to_dev(id_priv, cma_dev); + addr = (struct sockaddr_ib *) cma_src_addr(id_priv); + memcpy(&addr->sib_addr, &sgid, sizeof sgid); + cma_translate_ib(addr, &id_priv->id.route.addr.dev_addr); + return 0; +} + +static void cma_deref_id(struct rdma_id_private *id_priv) +{ + if (atomic_dec_and_test(&id_priv->refcount)) + complete(&id_priv->comp); +} + +static int cma_disable_callback(struct rdma_id_private *id_priv, + enum rdma_cm_state state) +{ + mutex_lock(&id_priv->handler_mutex); + if (id_priv->state != state) { + mutex_unlock(&id_priv->handler_mutex); + return -EINVAL; + } + return 0; +} + +struct rdma_cm_id *rdma_create_id(rdma_cm_event_handler event_handler, + void *context, enum rdma_port_space ps, + enum ib_qp_type qp_type) +{ + struct rdma_id_private *id_priv; + + id_priv = kzalloc(sizeof *id_priv, GFP_KERNEL); + if (!id_priv) + return ERR_PTR(-ENOMEM); + + id_priv->owner = task_pid_nr(current); + id_priv->state = RDMA_CM_IDLE; + id_priv->id.context = context; + id_priv->id.event_handler = event_handler; + id_priv->id.ps = ps; + id_priv->id.qp_type = qp_type; + spin_lock_init(&id_priv->lock); + mutex_init(&id_priv->qp_mutex); + init_completion(&id_priv->comp); + atomic_set(&id_priv->refcount, 1); + mutex_init(&id_priv->handler_mutex); + INIT_LIST_HEAD(&id_priv->listen_list); + INIT_LIST_HEAD(&id_priv->mc_list); + get_random_bytes(&id_priv->seq_num, sizeof id_priv->seq_num); + + return &id_priv->id; +} +EXPORT_SYMBOL(rdma_create_id); + +static int cma_init_ud_qp(struct rdma_id_private *id_priv, struct ib_qp *qp) +{ + struct ib_qp_attr qp_attr; + int qp_attr_mask, ret; + + qp_attr.qp_state = IB_QPS_INIT; + ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); + if (ret) + return ret; + + ret = ib_modify_qp(qp, &qp_attr, qp_attr_mask); + if (ret) + return ret; + + qp_attr.qp_state = IB_QPS_RTR; + ret = ib_modify_qp(qp, &qp_attr, IB_QP_STATE); + if (ret) + return ret; + + qp_attr.qp_state = IB_QPS_RTS; + qp_attr.sq_psn = 0; + ret = ib_modify_qp(qp, &qp_attr, IB_QP_STATE | IB_QP_SQ_PSN); + + return ret; +} + +static int cma_init_conn_qp(struct rdma_id_private *id_priv, struct ib_qp *qp) +{ + struct ib_qp_attr qp_attr; + int qp_attr_mask, ret; + + qp_attr.qp_state = IB_QPS_INIT; + ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); + if (ret) + return ret; + + return ib_modify_qp(qp, &qp_attr, qp_attr_mask); +} + +int rdma_create_qp(struct rdma_cm_id *id, struct ib_pd *pd, + struct ib_qp_init_attr *qp_init_attr) +{ + struct rdma_id_private *id_priv; + struct ib_qp *qp; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (id->device != pd->device) + return -EINVAL; + + qp = ib_create_qp(pd, qp_init_attr); + if (IS_ERR(qp)) + return PTR_ERR(qp); + + if (id->qp_type == IB_QPT_UD) + ret = cma_init_ud_qp(id_priv, qp); + else + ret = cma_init_conn_qp(id_priv, qp); + if (ret) + goto err; + + id->qp = qp; + id_priv->qp_num = qp->qp_num; + id_priv->srq = (qp->srq != NULL); + return 0; +err: + ib_destroy_qp(qp); + return ret; +} +EXPORT_SYMBOL(rdma_create_qp); + +void rdma_destroy_qp(struct rdma_cm_id *id) +{ + struct rdma_id_private *id_priv; + + id_priv = container_of(id, struct rdma_id_private, id); + mutex_lock(&id_priv->qp_mutex); + ib_destroy_qp(id_priv->id.qp); + id_priv->id.qp = NULL; + mutex_unlock(&id_priv->qp_mutex); +} +EXPORT_SYMBOL(rdma_destroy_qp); + +static int cma_modify_qp_rtr(struct rdma_id_private *id_priv, + struct rdma_conn_param *conn_param) +{ + struct ib_qp_attr qp_attr; + int qp_attr_mask, ret; + union ib_gid sgid; + + mutex_lock(&id_priv->qp_mutex); + if (!id_priv->id.qp) { + ret = 0; + goto out; + } + + /* Need to update QP attributes from default values. */ + qp_attr.qp_state = IB_QPS_INIT; + ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); + if (ret) + goto out; + + ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask); + if (ret) + goto out; + + qp_attr.qp_state = IB_QPS_RTR; + ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); + if (ret) + goto out; + + ret = ib_query_gid(id_priv->id.device, id_priv->id.port_num, + qp_attr.ah_attr.grh.sgid_index, &sgid); + if (ret) + goto out; + + if (rdma_node_get_transport(id_priv->cma_dev->device->node_type) + == RDMA_TRANSPORT_IB && + rdma_port_get_link_layer(id_priv->id.device, id_priv->id.port_num) + == IB_LINK_LAYER_ETHERNET) { + ret = rdma_addr_find_smac_by_sgid(&sgid, qp_attr.smac, NULL); + + if (ret) + goto out; + } + if (conn_param) + qp_attr.max_dest_rd_atomic = conn_param->responder_resources; + ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask); +out: + mutex_unlock(&id_priv->qp_mutex); + return ret; +} + +static int cma_modify_qp_rts(struct rdma_id_private *id_priv, + struct rdma_conn_param *conn_param) +{ + struct ib_qp_attr qp_attr; + int qp_attr_mask, ret; + + mutex_lock(&id_priv->qp_mutex); + if (!id_priv->id.qp) { + ret = 0; + goto out; + } + + qp_attr.qp_state = IB_QPS_RTS; + ret = rdma_init_qp_attr(&id_priv->id, &qp_attr, &qp_attr_mask); + if (ret) + goto out; + + if (conn_param) + qp_attr.max_rd_atomic = conn_param->initiator_depth; + ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask); +out: + mutex_unlock(&id_priv->qp_mutex); + return ret; +} + +static int cma_modify_qp_err(struct rdma_id_private *id_priv) +{ + struct ib_qp_attr qp_attr; + int ret; + + mutex_lock(&id_priv->qp_mutex); + if (!id_priv->id.qp) { + ret = 0; + goto out; + } + + qp_attr.qp_state = IB_QPS_ERR; + ret = ib_modify_qp(id_priv->id.qp, &qp_attr, IB_QP_STATE); +out: + mutex_unlock(&id_priv->qp_mutex); + return ret; +} + +static int cma_ib_init_qp_attr(struct rdma_id_private *id_priv, + struct ib_qp_attr *qp_attr, int *qp_attr_mask) +{ + struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; + int ret; + u16 pkey; + + if (rdma_port_get_link_layer(id_priv->id.device, id_priv->id.port_num) == + IB_LINK_LAYER_INFINIBAND) + pkey = ib_addr_get_pkey(dev_addr); + else + pkey = 0xffff; + + ret = ib_find_cached_pkey(id_priv->id.device, id_priv->id.port_num, + pkey, &qp_attr->pkey_index); + if (ret) + return ret; + + qp_attr->port_num = id_priv->id.port_num; + *qp_attr_mask = IB_QP_STATE | IB_QP_PKEY_INDEX | IB_QP_PORT; + + if (id_priv->id.qp_type == IB_QPT_UD) { + ret = cma_set_qkey(id_priv, 0); + if (ret) + return ret; + + qp_attr->qkey = id_priv->qkey; + *qp_attr_mask |= IB_QP_QKEY; + } else { + qp_attr->qp_access_flags = 0; + *qp_attr_mask |= IB_QP_ACCESS_FLAGS; + } + return 0; +} + +int rdma_init_qp_attr(struct rdma_cm_id *id, struct ib_qp_attr *qp_attr, + int *qp_attr_mask) +{ + struct rdma_id_private *id_priv; + int ret = 0; + + id_priv = container_of(id, struct rdma_id_private, id); + switch (rdma_node_get_transport(id_priv->id.device->node_type)) { + case RDMA_TRANSPORT_IB: + if (!id_priv->cm_id.ib || (id_priv->id.qp_type == IB_QPT_UD)) + ret = cma_ib_init_qp_attr(id_priv, qp_attr, qp_attr_mask); + else + ret = ib_cm_init_qp_attr(id_priv->cm_id.ib, qp_attr, + qp_attr_mask); + + if (qp_attr->qp_state == IB_QPS_RTR) + qp_attr->rq_psn = id_priv->seq_num; + break; + case RDMA_TRANSPORT_IWARP: + if (!id_priv->cm_id.iw) { + qp_attr->qp_access_flags = 0; + *qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS; + } else + ret = iw_cm_init_qp_attr(id_priv->cm_id.iw, qp_attr, + qp_attr_mask); + break; + default: + ret = -ENOSYS; + break; + } + + return ret; +} +EXPORT_SYMBOL(rdma_init_qp_attr); + +static inline int cma_zero_addr(struct sockaddr *addr) +{ + switch (addr->sa_family) { + case AF_INET: + return ipv4_is_zeronet(((struct sockaddr_in *)addr)->sin_addr.s_addr); + case AF_INET6: + return ipv6_addr_any(&((struct sockaddr_in6 *) addr)->sin6_addr); + case AF_IB: + return ib_addr_any(&((struct sockaddr_ib *) addr)->sib_addr); + default: + return 0; + } +} + +static inline int cma_loopback_addr(struct sockaddr *addr) +{ + switch (addr->sa_family) { + case AF_INET: + return ipv4_is_loopback(((struct sockaddr_in *) addr)->sin_addr.s_addr); + case AF_INET6: + return ipv6_addr_loopback(&((struct sockaddr_in6 *) addr)->sin6_addr); + case AF_IB: + return ib_addr_loopback(&((struct sockaddr_ib *) addr)->sib_addr); + default: + return 0; + } +} + +static inline int cma_any_addr(struct sockaddr *addr) +{ + return cma_zero_addr(addr) || cma_loopback_addr(addr); +} + +static int cma_addr_cmp(struct sockaddr *src, struct sockaddr *dst) +{ + if (src->sa_family != dst->sa_family) + return -1; + + switch (src->sa_family) { + case AF_INET: + return ((struct sockaddr_in *) src)->sin_addr.s_addr != + ((struct sockaddr_in *) dst)->sin_addr.s_addr; + case AF_INET6: + return ipv6_addr_cmp(&((struct sockaddr_in6 *) src)->sin6_addr, + &((struct sockaddr_in6 *) dst)->sin6_addr); + default: + return ib_addr_cmp(&((struct sockaddr_ib *) src)->sib_addr, + &((struct sockaddr_ib *) dst)->sib_addr); + } +} + +static __be16 cma_port(struct sockaddr *addr) +{ + struct sockaddr_ib *sib; + + switch (addr->sa_family) { + case AF_INET: + return ((struct sockaddr_in *) addr)->sin_port; + case AF_INET6: + return ((struct sockaddr_in6 *) addr)->sin6_port; + case AF_IB: + sib = (struct sockaddr_ib *) addr; + return htons((u16) (be64_to_cpu(sib->sib_sid) & + be64_to_cpu(sib->sib_sid_mask))); + default: + return 0; + } +} + +static inline int cma_any_port(struct sockaddr *addr) +{ + return !cma_port(addr); +} + +static void cma_save_ib_info(struct rdma_cm_id *id, struct rdma_cm_id *listen_id, + struct ib_sa_path_rec *path) +{ + struct sockaddr_ib *listen_ib, *ib; + + listen_ib = (struct sockaddr_ib *) &listen_id->route.addr.src_addr; + ib = (struct sockaddr_ib *) &id->route.addr.src_addr; + ib->sib_family = listen_ib->sib_family; + ib->sib_pkey = path->pkey; + ib->sib_flowinfo = path->flow_label; + memcpy(&ib->sib_addr, &path->sgid, 16); + ib->sib_sid = listen_ib->sib_sid; + ib->sib_sid_mask = cpu_to_be64(0xffffffffffffffffULL); + ib->sib_scope_id = listen_ib->sib_scope_id; + + ib = (struct sockaddr_ib *) &id->route.addr.dst_addr; + ib->sib_family = listen_ib->sib_family; + ib->sib_pkey = path->pkey; + ib->sib_flowinfo = path->flow_label; + memcpy(&ib->sib_addr, &path->dgid, 16); +} + +static void cma_save_ip4_info(struct rdma_cm_id *id, struct rdma_cm_id *listen_id, + struct cma_hdr *hdr) +{ + struct sockaddr_in *listen4, *ip4; + + listen4 = (struct sockaddr_in *) &listen_id->route.addr.src_addr; + ip4 = (struct sockaddr_in *) &id->route.addr.src_addr; + ip4->sin_family = listen4->sin_family; + ip4->sin_addr.s_addr = hdr->dst_addr.ip4.addr; + ip4->sin_port = listen4->sin_port; + + ip4 = (struct sockaddr_in *) &id->route.addr.dst_addr; + ip4->sin_family = listen4->sin_family; + ip4->sin_addr.s_addr = hdr->src_addr.ip4.addr; + ip4->sin_port = hdr->port; +} + +static void cma_save_ip6_info(struct rdma_cm_id *id, struct rdma_cm_id *listen_id, + struct cma_hdr *hdr) +{ + struct sockaddr_in6 *listen6, *ip6; + + listen6 = (struct sockaddr_in6 *) &listen_id->route.addr.src_addr; + ip6 = (struct sockaddr_in6 *) &id->route.addr.src_addr; + ip6->sin6_family = listen6->sin6_family; + ip6->sin6_addr = hdr->dst_addr.ip6; + ip6->sin6_port = listen6->sin6_port; + + ip6 = (struct sockaddr_in6 *) &id->route.addr.dst_addr; + ip6->sin6_family = listen6->sin6_family; + ip6->sin6_addr = hdr->src_addr.ip6; + ip6->sin6_port = hdr->port; +} + +static int cma_save_net_info(struct rdma_cm_id *id, struct rdma_cm_id *listen_id, + struct ib_cm_event *ib_event) +{ + struct cma_hdr *hdr; + + if ((listen_id->route.addr.src_addr.ss_family == AF_IB) && + (ib_event->event == IB_CM_REQ_RECEIVED)) { + cma_save_ib_info(id, listen_id, ib_event->param.req_rcvd.primary_path); + return 0; + } + + hdr = ib_event->private_data; + if (hdr->cma_version != CMA_VERSION) + return -EINVAL; + + switch (cma_get_ip_ver(hdr)) { + case 4: + cma_save_ip4_info(id, listen_id, hdr); + break; + case 6: + cma_save_ip6_info(id, listen_id, hdr); + break; + default: + return -EINVAL; + } + return 0; +} + +static inline int cma_user_data_offset(struct rdma_id_private *id_priv) +{ + return cma_family(id_priv) == AF_IB ? 0 : sizeof(struct cma_hdr); +} + +static void cma_cancel_route(struct rdma_id_private *id_priv) +{ + switch (rdma_port_get_link_layer(id_priv->id.device, id_priv->id.port_num)) { + case IB_LINK_LAYER_INFINIBAND: + if (id_priv->query) + ib_sa_cancel_query(id_priv->query_id, id_priv->query); + break; + default: + break; + } +} + +static void cma_cancel_listens(struct rdma_id_private *id_priv) +{ + struct rdma_id_private *dev_id_priv; + + /* + * Remove from listen_any_list to prevent added devices from spawning + * additional listen requests. + */ + mutex_lock(&lock); + list_del(&id_priv->list); + + while (!list_empty(&id_priv->listen_list)) { + dev_id_priv = list_entry(id_priv->listen_list.next, + struct rdma_id_private, listen_list); + /* sync with device removal to avoid duplicate destruction */ + list_del_init(&dev_id_priv->list); + list_del(&dev_id_priv->listen_list); + mutex_unlock(&lock); + + rdma_destroy_id(&dev_id_priv->id); + mutex_lock(&lock); + } + mutex_unlock(&lock); +} + +static void cma_cancel_operation(struct rdma_id_private *id_priv, + enum rdma_cm_state state) +{ + switch (state) { + case RDMA_CM_ADDR_QUERY: + rdma_addr_cancel(&id_priv->id.route.addr.dev_addr); + break; + case RDMA_CM_ROUTE_QUERY: + cma_cancel_route(id_priv); + break; + case RDMA_CM_LISTEN: + if (cma_any_addr(cma_src_addr(id_priv)) && !id_priv->cma_dev) + cma_cancel_listens(id_priv); + break; + default: + break; + } +} + +static void cma_release_port(struct rdma_id_private *id_priv) +{ + struct rdma_bind_list *bind_list = id_priv->bind_list; + + if (!bind_list) + return; + + mutex_lock(&lock); + hlist_del(&id_priv->node); + if (hlist_empty(&bind_list->owners)) { + idr_remove(bind_list->ps, bind_list->port); + kfree(bind_list); + } + mutex_unlock(&lock); +} + +static void cma_leave_mc_groups(struct rdma_id_private *id_priv) +{ + struct cma_multicast *mc; + + while (!list_empty(&id_priv->mc_list)) { + mc = container_of(id_priv->mc_list.next, + struct cma_multicast, list); + list_del(&mc->list); + switch (rdma_port_get_link_layer(id_priv->cma_dev->device, id_priv->id.port_num)) { + case IB_LINK_LAYER_INFINIBAND: + ib_sa_free_multicast(mc->multicast.ib); + kfree(mc); + break; + case IB_LINK_LAYER_ETHERNET: + kref_put(&mc->mcref, release_mc); + break; + default: + break; + } + } +} + +void rdma_destroy_id(struct rdma_cm_id *id) +{ + struct rdma_id_private *id_priv; + enum rdma_cm_state state; + + id_priv = container_of(id, struct rdma_id_private, id); + state = cma_exch(id_priv, RDMA_CM_DESTROYING); + cma_cancel_operation(id_priv, state); + + /* + * Wait for any active callback to finish. New callbacks will find + * the id_priv state set to destroying and abort. + */ + mutex_lock(&id_priv->handler_mutex); + mutex_unlock(&id_priv->handler_mutex); + + if (id_priv->cma_dev) { + switch (rdma_node_get_transport(id_priv->id.device->node_type)) { + case RDMA_TRANSPORT_IB: + if (id_priv->cm_id.ib) + ib_destroy_cm_id(id_priv->cm_id.ib); + break; + case RDMA_TRANSPORT_IWARP: + if (id_priv->cm_id.iw) + iw_destroy_cm_id(id_priv->cm_id.iw); + break; + default: + break; + } + cma_leave_mc_groups(id_priv); + cma_release_dev(id_priv); + } + + cma_release_port(id_priv); + cma_deref_id(id_priv); + wait_for_completion(&id_priv->comp); + + if (id_priv->internal_id) + cma_deref_id(id_priv->id.context); + + kfree(id_priv->id.route.path_rec); + kfree(id_priv); +} +EXPORT_SYMBOL(rdma_destroy_id); + +static int cma_rep_recv(struct rdma_id_private *id_priv) +{ + int ret; + + ret = cma_modify_qp_rtr(id_priv, NULL); + if (ret) + goto reject; + + ret = cma_modify_qp_rts(id_priv, NULL); + if (ret) + goto reject; + + ret = ib_send_cm_rtu(id_priv->cm_id.ib, NULL, 0); + if (ret) + goto reject; + + return 0; +reject: + cma_modify_qp_err(id_priv); + ib_send_cm_rej(id_priv->cm_id.ib, IB_CM_REJ_CONSUMER_DEFINED, + NULL, 0, NULL, 0); + return ret; +} + +static void cma_set_rep_event_data(struct rdma_cm_event *event, + struct ib_cm_rep_event_param *rep_data, + void *private_data) +{ + event->param.conn.private_data = private_data; + event->param.conn.private_data_len = IB_CM_REP_PRIVATE_DATA_SIZE; + event->param.conn.responder_resources = rep_data->responder_resources; + event->param.conn.initiator_depth = rep_data->initiator_depth; + event->param.conn.flow_control = rep_data->flow_control; + event->param.conn.rnr_retry_count = rep_data->rnr_retry_count; + event->param.conn.srq = rep_data->srq; + event->param.conn.qp_num = rep_data->remote_qpn; +} + +static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event) +{ + struct rdma_id_private *id_priv = cm_id->context; + struct rdma_cm_event event; + int ret = 0; + + if ((ib_event->event != IB_CM_TIMEWAIT_EXIT && + cma_disable_callback(id_priv, RDMA_CM_CONNECT)) || + (ib_event->event == IB_CM_TIMEWAIT_EXIT && + cma_disable_callback(id_priv, RDMA_CM_DISCONNECT))) + return 0; + + memset(&event, 0, sizeof event); + switch (ib_event->event) { + case IB_CM_REQ_ERROR: + case IB_CM_REP_ERROR: + event.event = RDMA_CM_EVENT_UNREACHABLE; + event.status = -ETIMEDOUT; + break; + case IB_CM_REP_RECEIVED: + if (id_priv->id.qp) { + event.status = cma_rep_recv(id_priv); + event.event = event.status ? RDMA_CM_EVENT_CONNECT_ERROR : + RDMA_CM_EVENT_ESTABLISHED; + } else { + event.event = RDMA_CM_EVENT_CONNECT_RESPONSE; + } + cma_set_rep_event_data(&event, &ib_event->param.rep_rcvd, + ib_event->private_data); + break; + case IB_CM_RTU_RECEIVED: + case IB_CM_USER_ESTABLISHED: + event.event = RDMA_CM_EVENT_ESTABLISHED; + break; + case IB_CM_DREQ_ERROR: + event.status = -ETIMEDOUT; /* fall through */ + case IB_CM_DREQ_RECEIVED: + case IB_CM_DREP_RECEIVED: + if (!cma_comp_exch(id_priv, RDMA_CM_CONNECT, + RDMA_CM_DISCONNECT)) + goto out; + event.event = RDMA_CM_EVENT_DISCONNECTED; + break; + case IB_CM_TIMEWAIT_EXIT: + event.event = RDMA_CM_EVENT_TIMEWAIT_EXIT; + break; + case IB_CM_MRA_RECEIVED: + /* ignore event */ + goto out; + case IB_CM_REJ_RECEIVED: + cma_modify_qp_err(id_priv); + event.status = ib_event->param.rej_rcvd.reason; + event.event = RDMA_CM_EVENT_REJECTED; + event.param.conn.private_data = ib_event->private_data; + event.param.conn.private_data_len = IB_CM_REJ_PRIVATE_DATA_SIZE; + break; + default: + printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d\n", + ib_event->event); + goto out; + } + + ret = id_priv->id.event_handler(&id_priv->id, &event); + if (ret) { + /* Destroy the CM ID by returning a non-zero value. */ + id_priv->cm_id.ib = NULL; + cma_exch(id_priv, RDMA_CM_DESTROYING); + mutex_unlock(&id_priv->handler_mutex); + rdma_destroy_id(&id_priv->id); + return ret; + } +out: + mutex_unlock(&id_priv->handler_mutex); + return ret; +} + +static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id, + struct ib_cm_event *ib_event) +{ + struct rdma_id_private *id_priv; + struct rdma_cm_id *id; + struct rdma_route *rt; + int ret; + + id = rdma_create_id(listen_id->event_handler, listen_id->context, + listen_id->ps, ib_event->param.req_rcvd.qp_type); + if (IS_ERR(id)) + return NULL; + + id_priv = container_of(id, struct rdma_id_private, id); + if (cma_save_net_info(id, listen_id, ib_event)) + goto err; + + rt = &id->route; + rt->num_paths = ib_event->param.req_rcvd.alternate_path ? 2 : 1; + rt->path_rec = kmalloc(sizeof *rt->path_rec * rt->num_paths, + GFP_KERNEL); + if (!rt->path_rec) + goto err; + + rt->path_rec[0] = *ib_event->param.req_rcvd.primary_path; + if (rt->num_paths == 2) + rt->path_rec[1] = *ib_event->param.req_rcvd.alternate_path; + + if (cma_any_addr(cma_src_addr(id_priv))) { + rt->addr.dev_addr.dev_type = ARPHRD_INFINIBAND; + rdma_addr_set_sgid(&rt->addr.dev_addr, &rt->path_rec[0].sgid); + ib_addr_set_pkey(&rt->addr.dev_addr, be16_to_cpu(rt->path_rec[0].pkey)); + } else { + ret = cma_translate_addr(cma_src_addr(id_priv), &rt->addr.dev_addr); + if (ret) + goto err; + } + rdma_addr_set_dgid(&rt->addr.dev_addr, &rt->path_rec[0].dgid); + + id_priv->state = RDMA_CM_CONNECT; + return id_priv; + +err: + rdma_destroy_id(id); + return NULL; +} + +static struct rdma_id_private *cma_new_udp_id(struct rdma_cm_id *listen_id, + struct ib_cm_event *ib_event) +{ + struct rdma_id_private *id_priv; + struct rdma_cm_id *id; + int ret; + + id = rdma_create_id(listen_id->event_handler, listen_id->context, + listen_id->ps, IB_QPT_UD); + if (IS_ERR(id)) + return NULL; + + id_priv = container_of(id, struct rdma_id_private, id); + if (cma_save_net_info(id, listen_id, ib_event)) + goto err; + + if (!cma_any_addr((struct sockaddr *) &id->route.addr.src_addr)) { + ret = cma_translate_addr(cma_src_addr(id_priv), &id->route.addr.dev_addr); + if (ret) + goto err; + } + + id_priv->state = RDMA_CM_CONNECT; + return id_priv; +err: + rdma_destroy_id(id); + return NULL; +} + +static void cma_set_req_event_data(struct rdma_cm_event *event, + struct ib_cm_req_event_param *req_data, + void *private_data, int offset) +{ + event->param.conn.private_data = private_data + offset; + event->param.conn.private_data_len = IB_CM_REQ_PRIVATE_DATA_SIZE - offset; + event->param.conn.responder_resources = req_data->responder_resources; + event->param.conn.initiator_depth = req_data->initiator_depth; + event->param.conn.flow_control = req_data->flow_control; + event->param.conn.retry_count = req_data->retry_count; + event->param.conn.rnr_retry_count = req_data->rnr_retry_count; + event->param.conn.srq = req_data->srq; + event->param.conn.qp_num = req_data->remote_qpn; +} + +static int cma_check_req_qp_type(struct rdma_cm_id *id, struct ib_cm_event *ib_event) +{ + return (((ib_event->event == IB_CM_REQ_RECEIVED) && + (ib_event->param.req_rcvd.qp_type == id->qp_type)) || + ((ib_event->event == IB_CM_SIDR_REQ_RECEIVED) && + (id->qp_type == IB_QPT_UD)) || + (!id->qp_type)); +} + +static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event) +{ + struct rdma_id_private *listen_id, *conn_id; + struct rdma_cm_event event; + int offset, ret; + + listen_id = cm_id->context; + if (!cma_check_req_qp_type(&listen_id->id, ib_event)) + return -EINVAL; + + if (cma_disable_callback(listen_id, RDMA_CM_LISTEN)) + return -ECONNABORTED; + + memset(&event, 0, sizeof event); + offset = cma_user_data_offset(listen_id); + event.event = RDMA_CM_EVENT_CONNECT_REQUEST; + if (ib_event->event == IB_CM_SIDR_REQ_RECEIVED) { + conn_id = cma_new_udp_id(&listen_id->id, ib_event); + event.param.ud.private_data = ib_event->private_data + offset; + event.param.ud.private_data_len = + IB_CM_SIDR_REQ_PRIVATE_DATA_SIZE - offset; + } else { + conn_id = cma_new_conn_id(&listen_id->id, ib_event); + cma_set_req_event_data(&event, &ib_event->param.req_rcvd, + ib_event->private_data, offset); + } + if (!conn_id) { + ret = -ENOMEM; + goto err1; + } + + mutex_lock_nested(&conn_id->handler_mutex, SINGLE_DEPTH_NESTING); + ret = cma_acquire_dev(conn_id, listen_id); + if (ret) + goto err2; + + conn_id->cm_id.ib = cm_id; + cm_id->context = conn_id; + cm_id->cm_handler = cma_ib_handler; + + /* + * Protect against the user destroying conn_id from another thread + * until we're done accessing it. + */ + atomic_inc(&conn_id->refcount); + ret = conn_id->id.event_handler(&conn_id->id, &event); + if (ret) + goto err3; + /* + * Acquire mutex to prevent user executing rdma_destroy_id() + * while we're accessing the cm_id. + */ + mutex_lock(&lock); + if (cma_comp(conn_id, RDMA_CM_CONNECT) && + (conn_id->id.qp_type != IB_QPT_UD)) + ib_send_cm_mra(cm_id, CMA_CM_MRA_SETTING, NULL, 0); + mutex_unlock(&lock); + mutex_unlock(&conn_id->handler_mutex); + mutex_unlock(&listen_id->handler_mutex); + cma_deref_id(conn_id); + return 0; + +err3: + cma_deref_id(conn_id); + /* Destroy the CM ID by returning a non-zero value. */ + conn_id->cm_id.ib = NULL; +err2: + cma_exch(conn_id, RDMA_CM_DESTROYING); + mutex_unlock(&conn_id->handler_mutex); +err1: + mutex_unlock(&listen_id->handler_mutex); + if (conn_id) + rdma_destroy_id(&conn_id->id); + return ret; +} + +__be64 rdma_get_service_id(struct rdma_cm_id *id, struct sockaddr *addr) +{ + if (addr->sa_family == AF_IB) + return ((struct sockaddr_ib *) addr)->sib_sid; + + return cpu_to_be64(((u64)id->ps << 16) + be16_to_cpu(cma_port(addr))); +} +EXPORT_SYMBOL(rdma_get_service_id); + +static void cma_set_compare_data(enum rdma_port_space ps, struct sockaddr *addr, + struct ib_cm_compare_data *compare) +{ + struct cma_hdr *cma_data, *cma_mask; + __be32 ip4_addr; + struct in6_addr ip6_addr; + + memset(compare, 0, sizeof *compare); + cma_data = (void *) compare->data; + cma_mask = (void *) compare->mask; + + switch (addr->sa_family) { + case AF_INET: + ip4_addr = ((struct sockaddr_in *) addr)->sin_addr.s_addr; + cma_set_ip_ver(cma_data, 4); + cma_set_ip_ver(cma_mask, 0xF); + if (!cma_any_addr(addr)) { + cma_data->dst_addr.ip4.addr = ip4_addr; + cma_mask->dst_addr.ip4.addr = htonl(~0); + } + break; + case AF_INET6: + ip6_addr = ((struct sockaddr_in6 *) addr)->sin6_addr; + cma_set_ip_ver(cma_data, 6); + cma_set_ip_ver(cma_mask, 0xF); + if (!cma_any_addr(addr)) { + cma_data->dst_addr.ip6 = ip6_addr; + memset(&cma_mask->dst_addr.ip6, 0xFF, + sizeof cma_mask->dst_addr.ip6); + } + break; + default: + break; + } +} + +static int cma_iw_handler(struct iw_cm_id *iw_id, struct iw_cm_event *iw_event) +{ + struct rdma_id_private *id_priv = iw_id->context; + struct rdma_cm_event event; + int ret = 0; + struct sockaddr *laddr = (struct sockaddr *)&iw_event->local_addr; + struct sockaddr *raddr = (struct sockaddr *)&iw_event->remote_addr; + + if (cma_disable_callback(id_priv, RDMA_CM_CONNECT)) + return 0; + + memset(&event, 0, sizeof event); + switch (iw_event->event) { + case IW_CM_EVENT_CLOSE: + event.event = RDMA_CM_EVENT_DISCONNECTED; + break; + case IW_CM_EVENT_CONNECT_REPLY: + memcpy(cma_src_addr(id_priv), laddr, + rdma_addr_size(laddr)); + memcpy(cma_dst_addr(id_priv), raddr, + rdma_addr_size(raddr)); + switch (iw_event->status) { + case 0: + event.event = RDMA_CM_EVENT_ESTABLISHED; + event.param.conn.initiator_depth = iw_event->ird; + event.param.conn.responder_resources = iw_event->ord; + break; + case -ECONNRESET: + case -ECONNREFUSED: + event.event = RDMA_CM_EVENT_REJECTED; + break; + case -ETIMEDOUT: + event.event = RDMA_CM_EVENT_UNREACHABLE; + break; + default: + event.event = RDMA_CM_EVENT_CONNECT_ERROR; + break; + } + break; + case IW_CM_EVENT_ESTABLISHED: + event.event = RDMA_CM_EVENT_ESTABLISHED; + event.param.conn.initiator_depth = iw_event->ird; + event.param.conn.responder_resources = iw_event->ord; + break; + default: + BUG_ON(1); + } + + event.status = iw_event->status; + event.param.conn.private_data = iw_event->private_data; + event.param.conn.private_data_len = iw_event->private_data_len; + ret = id_priv->id.event_handler(&id_priv->id, &event); + if (ret) { + /* Destroy the CM ID by returning a non-zero value. */ + id_priv->cm_id.iw = NULL; + cma_exch(id_priv, RDMA_CM_DESTROYING); + mutex_unlock(&id_priv->handler_mutex); + rdma_destroy_id(&id_priv->id); + return ret; + } + + mutex_unlock(&id_priv->handler_mutex); + return ret; +} + +static int iw_conn_req_handler(struct iw_cm_id *cm_id, + struct iw_cm_event *iw_event) +{ + struct rdma_cm_id *new_cm_id; + struct rdma_id_private *listen_id, *conn_id; + struct rdma_cm_event event; + int ret; + struct ib_device_attr attr; + struct sockaddr *laddr = (struct sockaddr *)&iw_event->local_addr; + struct sockaddr *raddr = (struct sockaddr *)&iw_event->remote_addr; + + listen_id = cm_id->context; + if (cma_disable_callback(listen_id, RDMA_CM_LISTEN)) + return -ECONNABORTED; + + /* Create a new RDMA id for the new IW CM ID */ + new_cm_id = rdma_create_id(listen_id->id.event_handler, + listen_id->id.context, + RDMA_PS_TCP, IB_QPT_RC); + if (IS_ERR(new_cm_id)) { + ret = -ENOMEM; + goto out; + } + conn_id = container_of(new_cm_id, struct rdma_id_private, id); + mutex_lock_nested(&conn_id->handler_mutex, SINGLE_DEPTH_NESTING); + conn_id->state = RDMA_CM_CONNECT; + + ret = rdma_translate_ip(laddr, &conn_id->id.route.addr.dev_addr, NULL); + if (ret) { + mutex_unlock(&conn_id->handler_mutex); + rdma_destroy_id(new_cm_id); + goto out; + } + + ret = cma_acquire_dev(conn_id, listen_id); + if (ret) { + mutex_unlock(&conn_id->handler_mutex); + rdma_destroy_id(new_cm_id); + goto out; + } + + conn_id->cm_id.iw = cm_id; + cm_id->context = conn_id; + cm_id->cm_handler = cma_iw_handler; + + memcpy(cma_src_addr(conn_id), laddr, rdma_addr_size(laddr)); + memcpy(cma_dst_addr(conn_id), raddr, rdma_addr_size(raddr)); + + ret = ib_query_device(conn_id->id.device, &attr); + if (ret) { + mutex_unlock(&conn_id->handler_mutex); + rdma_destroy_id(new_cm_id); + goto out; + } + + memset(&event, 0, sizeof event); + event.event = RDMA_CM_EVENT_CONNECT_REQUEST; + event.param.conn.private_data = iw_event->private_data; + event.param.conn.private_data_len = iw_event->private_data_len; + event.param.conn.initiator_depth = iw_event->ird; + event.param.conn.responder_resources = iw_event->ord; + + /* + * Protect against the user destroying conn_id from another thread + * until we're done accessing it. + */ + atomic_inc(&conn_id->refcount); + ret = conn_id->id.event_handler(&conn_id->id, &event); + if (ret) { + /* User wants to destroy the CM ID */ + conn_id->cm_id.iw = NULL; + cma_exch(conn_id, RDMA_CM_DESTROYING); + mutex_unlock(&conn_id->handler_mutex); + cma_deref_id(conn_id); + rdma_destroy_id(&conn_id->id); + goto out; + } + + mutex_unlock(&conn_id->handler_mutex); + cma_deref_id(conn_id); + +out: + mutex_unlock(&listen_id->handler_mutex); + return ret; +} + +static int cma_ib_listen(struct rdma_id_private *id_priv) +{ + struct ib_cm_compare_data compare_data; + struct sockaddr *addr; + struct ib_cm_id *id; + __be64 svc_id; + int ret; + + id = ib_create_cm_id(id_priv->id.device, cma_req_handler, id_priv); + if (IS_ERR(id)) + return PTR_ERR(id); + + id_priv->cm_id.ib = id; + + addr = cma_src_addr(id_priv); + svc_id = rdma_get_service_id(&id_priv->id, addr); + if (cma_any_addr(addr) && !id_priv->afonly) + ret = ib_cm_listen(id_priv->cm_id.ib, svc_id, 0, NULL); + else { + cma_set_compare_data(id_priv->id.ps, addr, &compare_data); + ret = ib_cm_listen(id_priv->cm_id.ib, svc_id, 0, &compare_data); + } + + if (ret) { + ib_destroy_cm_id(id_priv->cm_id.ib); + id_priv->cm_id.ib = NULL; + } + + return ret; +} + +static int cma_iw_listen(struct rdma_id_private *id_priv, int backlog) +{ + int ret; + struct iw_cm_id *id; + + id = iw_create_cm_id(id_priv->id.device, + iw_conn_req_handler, + id_priv); + if (IS_ERR(id)) + return PTR_ERR(id); + + id_priv->cm_id.iw = id; + + memcpy(&id_priv->cm_id.iw->local_addr, cma_src_addr(id_priv), + rdma_addr_size(cma_src_addr(id_priv))); + + ret = iw_cm_listen(id_priv->cm_id.iw, backlog); + + if (ret) { + iw_destroy_cm_id(id_priv->cm_id.iw); + id_priv->cm_id.iw = NULL; + } + + return ret; +} + +static int cma_listen_handler(struct rdma_cm_id *id, + struct rdma_cm_event *event) +{ + struct rdma_id_private *id_priv = id->context; + + id->context = id_priv->id.context; + id->event_handler = id_priv->id.event_handler; + return id_priv->id.event_handler(id, event); +} + +static void cma_listen_on_dev(struct rdma_id_private *id_priv, + struct cma_device *cma_dev) +{ + struct rdma_id_private *dev_id_priv; + struct rdma_cm_id *id; + int ret; + + if (cma_family(id_priv) == AF_IB && + rdma_node_get_transport(cma_dev->device->node_type) != RDMA_TRANSPORT_IB) + return; + + id = rdma_create_id(cma_listen_handler, id_priv, id_priv->id.ps, + id_priv->id.qp_type); + if (IS_ERR(id)) + return; + + dev_id_priv = container_of(id, struct rdma_id_private, id); + + dev_id_priv->state = RDMA_CM_ADDR_BOUND; + memcpy(cma_src_addr(dev_id_priv), cma_src_addr(id_priv), + rdma_addr_size(cma_src_addr(id_priv))); + + cma_attach_to_dev(dev_id_priv, cma_dev); + list_add_tail(&dev_id_priv->listen_list, &id_priv->listen_list); + atomic_inc(&id_priv->refcount); + dev_id_priv->internal_id = 1; + dev_id_priv->afonly = id_priv->afonly; + + ret = rdma_listen(id, id_priv->backlog); + if (ret) + printk(KERN_WARNING "RDMA CMA: cma_listen_on_dev, error %d, " + "listening on device %s\n", ret, cma_dev->device->name); +} + +static void cma_listen_on_all(struct rdma_id_private *id_priv) +{ + struct cma_device *cma_dev; + + mutex_lock(&lock); + list_add_tail(&id_priv->list, &listen_any_list); + list_for_each_entry(cma_dev, &dev_list, list) + cma_listen_on_dev(id_priv, cma_dev); + mutex_unlock(&lock); +} + +void rdma_set_service_type(struct rdma_cm_id *id, int tos) +{ + struct rdma_id_private *id_priv; + + id_priv = container_of(id, struct rdma_id_private, id); + id_priv->tos = (u8) tos; +} +EXPORT_SYMBOL(rdma_set_service_type); + +static void cma_query_handler(int status, struct ib_sa_path_rec *path_rec, + void *context) +{ + struct cma_work *work = context; + struct rdma_route *route; + + route = &work->id->id.route; + + if (!status) { + route->num_paths = 1; + *route->path_rec = *path_rec; + } else { + work->old_state = RDMA_CM_ROUTE_QUERY; + work->new_state = RDMA_CM_ADDR_RESOLVED; + work->event.event = RDMA_CM_EVENT_ROUTE_ERROR; + work->event.status = status; + } + + queue_work(cma_wq, &work->work); +} + +static int cma_query_ib_route(struct rdma_id_private *id_priv, int timeout_ms, + struct cma_work *work) +{ + struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; + struct ib_sa_path_rec path_rec; + ib_sa_comp_mask comp_mask; + struct sockaddr_in6 *sin6; + struct sockaddr_ib *sib; + + memset(&path_rec, 0, sizeof path_rec); + rdma_addr_get_sgid(dev_addr, &path_rec.sgid); + rdma_addr_get_dgid(dev_addr, &path_rec.dgid); + path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr)); + path_rec.numb_path = 1; + path_rec.reversible = 1; + path_rec.service_id = rdma_get_service_id(&id_priv->id, cma_dst_addr(id_priv)); + + comp_mask = IB_SA_PATH_REC_DGID | IB_SA_PATH_REC_SGID | + IB_SA_PATH_REC_PKEY | IB_SA_PATH_REC_NUMB_PATH | + IB_SA_PATH_REC_REVERSIBLE | IB_SA_PATH_REC_SERVICE_ID; + + switch (cma_family(id_priv)) { + case AF_INET: + path_rec.qos_class = cpu_to_be16((u16) id_priv->tos); + comp_mask |= IB_SA_PATH_REC_QOS_CLASS; + break; + case AF_INET6: + sin6 = (struct sockaddr_in6 *) cma_src_addr(id_priv); + path_rec.traffic_class = (u8) (be32_to_cpu(sin6->sin6_flowinfo) >> 20); + comp_mask |= IB_SA_PATH_REC_TRAFFIC_CLASS; + break; + case AF_IB: + sib = (struct sockaddr_ib *) cma_src_addr(id_priv); + path_rec.traffic_class = (u8) (be32_to_cpu(sib->sib_flowinfo) >> 20); + comp_mask |= IB_SA_PATH_REC_TRAFFIC_CLASS; + break; + } + + id_priv->query_id = ib_sa_path_rec_get(&sa_client, id_priv->id.device, + id_priv->id.port_num, &path_rec, + comp_mask, timeout_ms, + GFP_KERNEL, cma_query_handler, + work, &id_priv->query); + + return (id_priv->query_id < 0) ? id_priv->query_id : 0; +} + +static void cma_work_handler(struct work_struct *_work) +{ + struct cma_work *work = container_of(_work, struct cma_work, work); + struct rdma_id_private *id_priv = work->id; + int destroy = 0; + + mutex_lock(&id_priv->handler_mutex); + if (!cma_comp_exch(id_priv, work->old_state, work->new_state)) + goto out; + + if (id_priv->id.event_handler(&id_priv->id, &work->event)) { + cma_exch(id_priv, RDMA_CM_DESTROYING); + destroy = 1; + } +out: + mutex_unlock(&id_priv->handler_mutex); + cma_deref_id(id_priv); + if (destroy) + rdma_destroy_id(&id_priv->id); + kfree(work); +} + +static void cma_ndev_work_handler(struct work_struct *_work) +{ + struct cma_ndev_work *work = container_of(_work, struct cma_ndev_work, work); + struct rdma_id_private *id_priv = work->id; + int destroy = 0; + + mutex_lock(&id_priv->handler_mutex); + if (id_priv->state == RDMA_CM_DESTROYING || + id_priv->state == RDMA_CM_DEVICE_REMOVAL) + goto out; + + if (id_priv->id.event_handler(&id_priv->id, &work->event)) { + cma_exch(id_priv, RDMA_CM_DESTROYING); + destroy = 1; + } + +out: + mutex_unlock(&id_priv->handler_mutex); + cma_deref_id(id_priv); + if (destroy) + rdma_destroy_id(&id_priv->id); + kfree(work); +} + +static int cma_resolve_ib_route(struct rdma_id_private *id_priv, int timeout_ms) +{ + struct rdma_route *route = &id_priv->id.route; + struct cma_work *work; + int ret; + + work = kzalloc(sizeof *work, GFP_KERNEL); + if (!work) + return -ENOMEM; + + work->id = id_priv; + INIT_WORK(&work->work, cma_work_handler); + work->old_state = RDMA_CM_ROUTE_QUERY; + work->new_state = RDMA_CM_ROUTE_RESOLVED; + work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED; + + route->path_rec = kmalloc(sizeof *route->path_rec, GFP_KERNEL); + if (!route->path_rec) { + ret = -ENOMEM; + goto err1; + } + + ret = cma_query_ib_route(id_priv, timeout_ms, work); + if (ret) + goto err2; + + return 0; +err2: + kfree(route->path_rec); + route->path_rec = NULL; +err1: + kfree(work); + return ret; +} + +int rdma_set_ib_paths(struct rdma_cm_id *id, + struct ib_sa_path_rec *path_rec, int num_paths) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, + RDMA_CM_ROUTE_RESOLVED)) + return -EINVAL; + + id->route.path_rec = kmemdup(path_rec, sizeof *path_rec * num_paths, + GFP_KERNEL); + if (!id->route.path_rec) { + ret = -ENOMEM; + goto err; + } + + id->route.num_paths = num_paths; + return 0; +err: + cma_comp_exch(id_priv, RDMA_CM_ROUTE_RESOLVED, RDMA_CM_ADDR_RESOLVED); + return ret; +} +EXPORT_SYMBOL(rdma_set_ib_paths); + +static int cma_resolve_iw_route(struct rdma_id_private *id_priv, int timeout_ms) +{ + struct cma_work *work; + + work = kzalloc(sizeof *work, GFP_KERNEL); + if (!work) + return -ENOMEM; + + work->id = id_priv; + INIT_WORK(&work->work, cma_work_handler); + work->old_state = RDMA_CM_ROUTE_QUERY; + work->new_state = RDMA_CM_ROUTE_RESOLVED; + work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED; + queue_work(cma_wq, &work->work); + return 0; +} + +static int iboe_tos_to_sl(struct net_device *ndev, int tos) +{ + int prio; + struct net_device *dev; + + prio = rt_tos2priority(tos); + dev = ndev->priv_flags & IFF_802_1Q_VLAN ? + vlan_dev_real_dev(ndev) : ndev; + + if (dev->num_tc) + return netdev_get_prio_tc_map(dev, prio); + +#if IS_ENABLED(CONFIG_VLAN_8021Q) + if (ndev->priv_flags & IFF_802_1Q_VLAN) + return (vlan_dev_get_egress_qos_mask(ndev, prio) & + VLAN_PRIO_MASK) >> VLAN_PRIO_SHIFT; +#endif + return 0; +} + +static int cma_resolve_iboe_route(struct rdma_id_private *id_priv) +{ + struct rdma_route *route = &id_priv->id.route; + struct rdma_addr *addr = &route->addr; + struct cma_work *work; + int ret; + struct net_device *ndev = NULL; + + + work = kzalloc(sizeof *work, GFP_KERNEL); + if (!work) + return -ENOMEM; + + work->id = id_priv; + INIT_WORK(&work->work, cma_work_handler); + + route->path_rec = kzalloc(sizeof *route->path_rec, GFP_KERNEL); + if (!route->path_rec) { + ret = -ENOMEM; + goto err1; + } + + route->num_paths = 1; + + if (addr->dev_addr.bound_dev_if) + ndev = dev_get_by_index(&init_net, addr->dev_addr.bound_dev_if); + if (!ndev) { + ret = -ENODEV; + goto err2; + } + + route->path_rec->vlan_id = rdma_vlan_dev_vlan_id(ndev); + memcpy(route->path_rec->dmac, addr->dev_addr.dst_dev_addr, ETH_ALEN); + memcpy(route->path_rec->smac, ndev->dev_addr, ndev->addr_len); + + rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.src_addr, + &route->path_rec->sgid); + rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.dst_addr, + &route->path_rec->dgid); + + route->path_rec->hop_limit = 1; + route->path_rec->reversible = 1; + route->path_rec->pkey = cpu_to_be16(0xffff); + route->path_rec->mtu_selector = IB_SA_EQ; + route->path_rec->sl = iboe_tos_to_sl(ndev, id_priv->tos); + route->path_rec->mtu = iboe_get_mtu(ndev->mtu); + route->path_rec->rate_selector = IB_SA_EQ; + route->path_rec->rate = iboe_get_rate(ndev); + dev_put(ndev); + route->path_rec->packet_life_time_selector = IB_SA_EQ; + route->path_rec->packet_life_time = CMA_IBOE_PACKET_LIFETIME; + if (!route->path_rec->mtu) { + ret = -EINVAL; + goto err2; + } + + work->old_state = RDMA_CM_ROUTE_QUERY; + work->new_state = RDMA_CM_ROUTE_RESOLVED; + work->event.event = RDMA_CM_EVENT_ROUTE_RESOLVED; + work->event.status = 0; + + queue_work(cma_wq, &work->work); + + return 0; + +err2: + kfree(route->path_rec); + route->path_rec = NULL; +err1: + kfree(work); + return ret; +} + +int rdma_resolve_route(struct rdma_cm_id *id, int timeout_ms) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, RDMA_CM_ROUTE_QUERY)) + return -EINVAL; + + atomic_inc(&id_priv->refcount); + switch (rdma_node_get_transport(id->device->node_type)) { + case RDMA_TRANSPORT_IB: + switch (rdma_port_get_link_layer(id->device, id->port_num)) { + case IB_LINK_LAYER_INFINIBAND: + ret = cma_resolve_ib_route(id_priv, timeout_ms); + break; + case IB_LINK_LAYER_ETHERNET: + ret = cma_resolve_iboe_route(id_priv); + break; + default: + ret = -ENOSYS; + } + break; + case RDMA_TRANSPORT_IWARP: + ret = cma_resolve_iw_route(id_priv, timeout_ms); + break; + default: + ret = -ENOSYS; + break; + } + if (ret) + goto err; + + return 0; +err: + cma_comp_exch(id_priv, RDMA_CM_ROUTE_QUERY, RDMA_CM_ADDR_RESOLVED); + cma_deref_id(id_priv); + return ret; +} +EXPORT_SYMBOL(rdma_resolve_route); + +static void cma_set_loopback(struct sockaddr *addr) +{ + switch (addr->sa_family) { + case AF_INET: + ((struct sockaddr_in *) addr)->sin_addr.s_addr = htonl(INADDR_LOOPBACK); + break; + case AF_INET6: + ipv6_addr_set(&((struct sockaddr_in6 *) addr)->sin6_addr, + 0, 0, 0, htonl(1)); + break; + default: + ib_addr_set(&((struct sockaddr_ib *) addr)->sib_addr, + 0, 0, 0, htonl(1)); + break; + } +} + +static int cma_bind_loopback(struct rdma_id_private *id_priv) +{ + struct cma_device *cma_dev, *cur_dev; + struct ib_port_attr port_attr; + union ib_gid gid; + u16 pkey; + int ret; + u8 p; + + cma_dev = NULL; + mutex_lock(&lock); + list_for_each_entry(cur_dev, &dev_list, list) { + if (cma_family(id_priv) == AF_IB && + rdma_node_get_transport(cur_dev->device->node_type) != RDMA_TRANSPORT_IB) + continue; + + if (!cma_dev) + cma_dev = cur_dev; + + for (p = 1; p <= cur_dev->device->phys_port_cnt; ++p) { + if (!ib_query_port(cur_dev->device, p, &port_attr) && + port_attr.state == IB_PORT_ACTIVE) { + cma_dev = cur_dev; + goto port_found; + } + } + } + + if (!cma_dev) { + ret = -ENODEV; + goto out; + } + + p = 1; + +port_found: + ret = ib_get_cached_gid(cma_dev->device, p, 0, &gid); + if (ret) + goto out; + + ret = ib_get_cached_pkey(cma_dev->device, p, 0, &pkey); + if (ret) + goto out; + + id_priv->id.route.addr.dev_addr.dev_type = + (rdma_port_get_link_layer(cma_dev->device, p) == IB_LINK_LAYER_INFINIBAND) ? + ARPHRD_INFINIBAND : ARPHRD_ETHER; + + rdma_addr_set_sgid(&id_priv->id.route.addr.dev_addr, &gid); + ib_addr_set_pkey(&id_priv->id.route.addr.dev_addr, pkey); + id_priv->id.port_num = p; + cma_attach_to_dev(id_priv, cma_dev); + cma_set_loopback(cma_src_addr(id_priv)); +out: + mutex_unlock(&lock); + return ret; +} + +static void addr_handler(int status, struct sockaddr *src_addr, + struct rdma_dev_addr *dev_addr, void *context) +{ + struct rdma_id_private *id_priv = context; + struct rdma_cm_event event; + + memset(&event, 0, sizeof event); + mutex_lock(&id_priv->handler_mutex); + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, + RDMA_CM_ADDR_RESOLVED)) + goto out; + + memcpy(cma_src_addr(id_priv), src_addr, rdma_addr_size(src_addr)); + if (!status && !id_priv->cma_dev) + status = cma_acquire_dev(id_priv, NULL); + + if (status) { + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_RESOLVED, + RDMA_CM_ADDR_BOUND)) + goto out; + event.event = RDMA_CM_EVENT_ADDR_ERROR; + event.status = status; + } else + event.event = RDMA_CM_EVENT_ADDR_RESOLVED; + + if (id_priv->id.event_handler(&id_priv->id, &event)) { + cma_exch(id_priv, RDMA_CM_DESTROYING); + mutex_unlock(&id_priv->handler_mutex); + cma_deref_id(id_priv); + rdma_destroy_id(&id_priv->id); + return; + } +out: + mutex_unlock(&id_priv->handler_mutex); + cma_deref_id(id_priv); +} + +static int cma_resolve_loopback(struct rdma_id_private *id_priv) +{ + struct cma_work *work; + union ib_gid gid; + int ret; + + work = kzalloc(sizeof *work, GFP_KERNEL); + if (!work) + return -ENOMEM; + + if (!id_priv->cma_dev) { + ret = cma_bind_loopback(id_priv); + if (ret) + goto err; + } + + rdma_addr_get_sgid(&id_priv->id.route.addr.dev_addr, &gid); + rdma_addr_set_dgid(&id_priv->id.route.addr.dev_addr, &gid); + + work->id = id_priv; + INIT_WORK(&work->work, cma_work_handler); + work->old_state = RDMA_CM_ADDR_QUERY; + work->new_state = RDMA_CM_ADDR_RESOLVED; + work->event.event = RDMA_CM_EVENT_ADDR_RESOLVED; + queue_work(cma_wq, &work->work); + return 0; +err: + kfree(work); + return ret; +} + +static int cma_resolve_ib_addr(struct rdma_id_private *id_priv) +{ + struct cma_work *work; + int ret; + + work = kzalloc(sizeof *work, GFP_KERNEL); + if (!work) + return -ENOMEM; + + if (!id_priv->cma_dev) { + ret = cma_resolve_ib_dev(id_priv); + if (ret) + goto err; + } + + rdma_addr_set_dgid(&id_priv->id.route.addr.dev_addr, (union ib_gid *) + &(((struct sockaddr_ib *) &id_priv->id.route.addr.dst_addr)->sib_addr)); + + work->id = id_priv; + INIT_WORK(&work->work, cma_work_handler); + work->old_state = RDMA_CM_ADDR_QUERY; + work->new_state = RDMA_CM_ADDR_RESOLVED; + work->event.event = RDMA_CM_EVENT_ADDR_RESOLVED; + queue_work(cma_wq, &work->work); + return 0; +err: + kfree(work); + return ret; +} + +static int cma_bind_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, + struct sockaddr *dst_addr) +{ + if (!src_addr || !src_addr->sa_family) { + src_addr = (struct sockaddr *) &id->route.addr.src_addr; + src_addr->sa_family = dst_addr->sa_family; + if (dst_addr->sa_family == AF_INET6) { + ((struct sockaddr_in6 *) src_addr)->sin6_scope_id = + ((struct sockaddr_in6 *) dst_addr)->sin6_scope_id; + } else if (dst_addr->sa_family == AF_IB) { + ((struct sockaddr_ib *) src_addr)->sib_pkey = + ((struct sockaddr_ib *) dst_addr)->sib_pkey; + } + } + return rdma_bind_addr(id, src_addr); +} + +int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, + struct sockaddr *dst_addr, int timeout_ms) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (id_priv->state == RDMA_CM_IDLE) { + ret = cma_bind_addr(id, src_addr, dst_addr); + if (ret) + return ret; + } + + if (cma_family(id_priv) != dst_addr->sa_family) + return -EINVAL; + + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_ADDR_QUERY)) + return -EINVAL; + + atomic_inc(&id_priv->refcount); + memcpy(cma_dst_addr(id_priv), dst_addr, rdma_addr_size(dst_addr)); + if (cma_any_addr(dst_addr)) { + ret = cma_resolve_loopback(id_priv); + } else { + if (dst_addr->sa_family == AF_IB) { + ret = cma_resolve_ib_addr(id_priv); + } else { + ret = rdma_resolve_ip(&addr_client, cma_src_addr(id_priv), + dst_addr, &id->route.addr.dev_addr, + timeout_ms, addr_handler, id_priv); + } + } + if (ret) + goto err; + + return 0; +err: + cma_comp_exch(id_priv, RDMA_CM_ADDR_QUERY, RDMA_CM_ADDR_BOUND); + cma_deref_id(id_priv); + return ret; +} +EXPORT_SYMBOL(rdma_resolve_addr); + +int rdma_set_reuseaddr(struct rdma_cm_id *id, int reuse) +{ + struct rdma_id_private *id_priv; + unsigned long flags; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + spin_lock_irqsave(&id_priv->lock, flags); + if (reuse || id_priv->state == RDMA_CM_IDLE) { + id_priv->reuseaddr = reuse; + ret = 0; + } else { + ret = -EINVAL; + } + spin_unlock_irqrestore(&id_priv->lock, flags); + return ret; +} +EXPORT_SYMBOL(rdma_set_reuseaddr); + +int rdma_set_afonly(struct rdma_cm_id *id, int afonly) +{ + struct rdma_id_private *id_priv; + unsigned long flags; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + spin_lock_irqsave(&id_priv->lock, flags); + if (id_priv->state == RDMA_CM_IDLE || id_priv->state == RDMA_CM_ADDR_BOUND) { + id_priv->options |= (1 << CMA_OPTION_AFONLY); + id_priv->afonly = afonly; + ret = 0; + } else { + ret = -EINVAL; + } + spin_unlock_irqrestore(&id_priv->lock, flags); + return ret; +} +EXPORT_SYMBOL(rdma_set_afonly); + +static void cma_bind_port(struct rdma_bind_list *bind_list, + struct rdma_id_private *id_priv) +{ + struct sockaddr *addr; + struct sockaddr_ib *sib; + u64 sid, mask; + __be16 port; + + addr = cma_src_addr(id_priv); + port = htons(bind_list->port); + + switch (addr->sa_family) { + case AF_INET: + ((struct sockaddr_in *) addr)->sin_port = port; + break; + case AF_INET6: + ((struct sockaddr_in6 *) addr)->sin6_port = port; + break; + case AF_IB: + sib = (struct sockaddr_ib *) addr; + sid = be64_to_cpu(sib->sib_sid); + mask = be64_to_cpu(sib->sib_sid_mask); + sib->sib_sid = cpu_to_be64((sid & mask) | (u64) ntohs(port)); + sib->sib_sid_mask = cpu_to_be64(~0ULL); + break; + } + id_priv->bind_list = bind_list; + hlist_add_head(&id_priv->node, &bind_list->owners); +} + +static int cma_alloc_port(struct idr *ps, struct rdma_id_private *id_priv, + unsigned short snum) +{ + struct rdma_bind_list *bind_list; + int ret; + + bind_list = kzalloc(sizeof *bind_list, GFP_KERNEL); + if (!bind_list) + return -ENOMEM; + + ret = idr_alloc(ps, bind_list, snum, snum + 1, GFP_KERNEL); + if (ret < 0) + goto err; + + bind_list->ps = ps; + bind_list->port = (unsigned short)ret; + cma_bind_port(bind_list, id_priv); + return 0; +err: + kfree(bind_list); + return ret == -ENOSPC ? -EADDRNOTAVAIL : ret; +} + +static int cma_alloc_any_port(struct idr *ps, struct rdma_id_private *id_priv) +{ + static unsigned int last_used_port; + int low, high, remaining; + unsigned int rover; + + inet_get_local_port_range(&init_net, &low, &high); + remaining = (high - low) + 1; + rover = prandom_u32() % remaining + low; +retry: + if (last_used_port != rover && + !idr_find(ps, (unsigned short) rover)) { + int ret = cma_alloc_port(ps, id_priv, rover); + /* + * Remember previously used port number in order to avoid + * re-using same port immediately after it is closed. + */ + if (!ret) + last_used_port = rover; + if (ret != -EADDRNOTAVAIL) + return ret; + } + if (--remaining) { + rover++; + if ((rover < low) || (rover > high)) + rover = low; + goto retry; + } + return -EADDRNOTAVAIL; +} + +/* + * Check that the requested port is available. This is called when trying to + * bind to a specific port, or when trying to listen on a bound port. In + * the latter case, the provided id_priv may already be on the bind_list, but + * we still need to check that it's okay to start listening. + */ +static int cma_check_port(struct rdma_bind_list *bind_list, + struct rdma_id_private *id_priv, uint8_t reuseaddr) +{ + struct rdma_id_private *cur_id; + struct sockaddr *addr, *cur_addr; + + addr = cma_src_addr(id_priv); + hlist_for_each_entry(cur_id, &bind_list->owners, node) { + if (id_priv == cur_id) + continue; + + if ((cur_id->state != RDMA_CM_LISTEN) && reuseaddr && + cur_id->reuseaddr) + continue; + + cur_addr = cma_src_addr(cur_id); + if (id_priv->afonly && cur_id->afonly && + (addr->sa_family != cur_addr->sa_family)) + continue; + + if (cma_any_addr(addr) || cma_any_addr(cur_addr)) + return -EADDRNOTAVAIL; + + if (!cma_addr_cmp(addr, cur_addr)) + return -EADDRINUSE; + } + return 0; +} + +static int cma_use_port(struct idr *ps, struct rdma_id_private *id_priv) +{ + struct rdma_bind_list *bind_list; + unsigned short snum; + int ret; + + snum = ntohs(cma_port(cma_src_addr(id_priv))); + if (snum < PROT_SOCK && !capable(CAP_NET_BIND_SERVICE)) + return -EACCES; + + bind_list = idr_find(ps, snum); + if (!bind_list) { + ret = cma_alloc_port(ps, id_priv, snum); + } else { + ret = cma_check_port(bind_list, id_priv, id_priv->reuseaddr); + if (!ret) + cma_bind_port(bind_list, id_priv); + } + return ret; +} + +static int cma_bind_listen(struct rdma_id_private *id_priv) +{ + struct rdma_bind_list *bind_list = id_priv->bind_list; + int ret = 0; + + mutex_lock(&lock); + if (bind_list->owners.first->next) + ret = cma_check_port(bind_list, id_priv, 0); + mutex_unlock(&lock); + return ret; +} + +static struct idr *cma_select_inet_ps(struct rdma_id_private *id_priv) +{ + switch (id_priv->id.ps) { + case RDMA_PS_TCP: + return &tcp_ps; + case RDMA_PS_UDP: + return &udp_ps; + case RDMA_PS_IPOIB: + return &ipoib_ps; + case RDMA_PS_IB: + return &ib_ps; + default: + return NULL; + } +} + +static struct idr *cma_select_ib_ps(struct rdma_id_private *id_priv) +{ + struct idr *ps = NULL; + struct sockaddr_ib *sib; + u64 sid_ps, mask, sid; + + sib = (struct sockaddr_ib *) cma_src_addr(id_priv); + mask = be64_to_cpu(sib->sib_sid_mask) & RDMA_IB_IP_PS_MASK; + sid = be64_to_cpu(sib->sib_sid) & mask; + + if ((id_priv->id.ps == RDMA_PS_IB) && (sid == (RDMA_IB_IP_PS_IB & mask))) { + sid_ps = RDMA_IB_IP_PS_IB; + ps = &ib_ps; + } else if (((id_priv->id.ps == RDMA_PS_IB) || (id_priv->id.ps == RDMA_PS_TCP)) && + (sid == (RDMA_IB_IP_PS_TCP & mask))) { + sid_ps = RDMA_IB_IP_PS_TCP; + ps = &tcp_ps; + } else if (((id_priv->id.ps == RDMA_PS_IB) || (id_priv->id.ps == RDMA_PS_UDP)) && + (sid == (RDMA_IB_IP_PS_UDP & mask))) { + sid_ps = RDMA_IB_IP_PS_UDP; + ps = &udp_ps; + } + + if (ps) { + sib->sib_sid = cpu_to_be64(sid_ps | ntohs(cma_port((struct sockaddr *) sib))); + sib->sib_sid_mask = cpu_to_be64(RDMA_IB_IP_PS_MASK | + be64_to_cpu(sib->sib_sid_mask)); + } + return ps; +} + +static int cma_get_port(struct rdma_id_private *id_priv) +{ + struct idr *ps; + int ret; + + if (cma_family(id_priv) != AF_IB) + ps = cma_select_inet_ps(id_priv); + else + ps = cma_select_ib_ps(id_priv); + if (!ps) + return -EPROTONOSUPPORT; + + mutex_lock(&lock); + if (cma_any_port(cma_src_addr(id_priv))) + ret = cma_alloc_any_port(ps, id_priv); + else + ret = cma_use_port(ps, id_priv); + mutex_unlock(&lock); + + return ret; +} + +static int cma_check_linklocal(struct rdma_dev_addr *dev_addr, + struct sockaddr *addr) +{ +#if IS_ENABLED(CONFIG_IPV6) + struct sockaddr_in6 *sin6; + + if (addr->sa_family != AF_INET6) + return 0; + + sin6 = (struct sockaddr_in6 *) addr; + + if (!(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL)) + return 0; + + if (!sin6->sin6_scope_id) + return -EINVAL; + + dev_addr->bound_dev_if = sin6->sin6_scope_id; +#endif + return 0; +} + +int rdma_listen(struct rdma_cm_id *id, int backlog) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (id_priv->state == RDMA_CM_IDLE) { + id->route.addr.src_addr.ss_family = AF_INET; + ret = rdma_bind_addr(id, cma_src_addr(id_priv)); + if (ret) + return ret; + } + + if (!cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_LISTEN)) + return -EINVAL; + + if (id_priv->reuseaddr) { + ret = cma_bind_listen(id_priv); + if (ret) + goto err; + } + + id_priv->backlog = backlog; + if (id->device) { + switch (rdma_node_get_transport(id->device->node_type)) { + case RDMA_TRANSPORT_IB: + ret = cma_ib_listen(id_priv); + if (ret) + goto err; + break; + case RDMA_TRANSPORT_IWARP: + ret = cma_iw_listen(id_priv, backlog); + if (ret) + goto err; + break; + default: + ret = -ENOSYS; + goto err; + } + } else + cma_listen_on_all(id_priv); + + return 0; +err: + id_priv->backlog = 0; + cma_comp_exch(id_priv, RDMA_CM_LISTEN, RDMA_CM_ADDR_BOUND); + return ret; +} +EXPORT_SYMBOL(rdma_listen); + +int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr) +{ + struct rdma_id_private *id_priv; + int ret; + + if (addr->sa_family != AF_INET && addr->sa_family != AF_INET6 && + addr->sa_family != AF_IB) + return -EAFNOSUPPORT; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!cma_comp_exch(id_priv, RDMA_CM_IDLE, RDMA_CM_ADDR_BOUND)) + return -EINVAL; + + ret = cma_check_linklocal(&id->route.addr.dev_addr, addr); + if (ret) + goto err1; + + memcpy(cma_src_addr(id_priv), addr, rdma_addr_size(addr)); + if (!cma_any_addr(addr)) { + ret = cma_translate_addr(addr, &id->route.addr.dev_addr); + if (ret) + goto err1; + + ret = cma_acquire_dev(id_priv, NULL); + if (ret) + goto err1; + } + + if (!(id_priv->options & (1 << CMA_OPTION_AFONLY))) { + if (addr->sa_family == AF_INET) + id_priv->afonly = 1; +#if IS_ENABLED(CONFIG_IPV6) + else if (addr->sa_family == AF_INET6) + id_priv->afonly = init_net.ipv6.sysctl.bindv6only; +#endif + } + ret = cma_get_port(id_priv); + if (ret) + goto err2; + + return 0; +err2: + if (id_priv->cma_dev) + cma_release_dev(id_priv); +err1: + cma_comp_exch(id_priv, RDMA_CM_ADDR_BOUND, RDMA_CM_IDLE); + return ret; +} +EXPORT_SYMBOL(rdma_bind_addr); + +static int cma_format_hdr(void *hdr, struct rdma_id_private *id_priv) +{ + struct cma_hdr *cma_hdr; + + cma_hdr = hdr; + cma_hdr->cma_version = CMA_VERSION; + if (cma_family(id_priv) == AF_INET) { + struct sockaddr_in *src4, *dst4; + + src4 = (struct sockaddr_in *) cma_src_addr(id_priv); + dst4 = (struct sockaddr_in *) cma_dst_addr(id_priv); + + cma_set_ip_ver(cma_hdr, 4); + cma_hdr->src_addr.ip4.addr = src4->sin_addr.s_addr; + cma_hdr->dst_addr.ip4.addr = dst4->sin_addr.s_addr; + cma_hdr->port = src4->sin_port; + } else if (cma_family(id_priv) == AF_INET6) { + struct sockaddr_in6 *src6, *dst6; + + src6 = (struct sockaddr_in6 *) cma_src_addr(id_priv); + dst6 = (struct sockaddr_in6 *) cma_dst_addr(id_priv); + + cma_set_ip_ver(cma_hdr, 6); + cma_hdr->src_addr.ip6 = src6->sin6_addr; + cma_hdr->dst_addr.ip6 = dst6->sin6_addr; + cma_hdr->port = src6->sin6_port; + } + return 0; +} + +static int cma_sidr_rep_handler(struct ib_cm_id *cm_id, + struct ib_cm_event *ib_event) +{ + struct rdma_id_private *id_priv = cm_id->context; + struct rdma_cm_event event; + struct ib_cm_sidr_rep_event_param *rep = &ib_event->param.sidr_rep_rcvd; + int ret = 0; + + if (cma_disable_callback(id_priv, RDMA_CM_CONNECT)) + return 0; + + memset(&event, 0, sizeof event); + switch (ib_event->event) { + case IB_CM_SIDR_REQ_ERROR: + event.event = RDMA_CM_EVENT_UNREACHABLE; + event.status = -ETIMEDOUT; + break; + case IB_CM_SIDR_REP_RECEIVED: + event.param.ud.private_data = ib_event->private_data; + event.param.ud.private_data_len = IB_CM_SIDR_REP_PRIVATE_DATA_SIZE; + if (rep->status != IB_SIDR_SUCCESS) { + event.event = RDMA_CM_EVENT_UNREACHABLE; + event.status = ib_event->param.sidr_rep_rcvd.status; + break; + } + ret = cma_set_qkey(id_priv, rep->qkey); + if (ret) { + event.event = RDMA_CM_EVENT_ADDR_ERROR; + event.status = ret; + break; + } + ib_init_ah_from_path(id_priv->id.device, id_priv->id.port_num, + id_priv->id.route.path_rec, + &event.param.ud.ah_attr); + event.param.ud.qp_num = rep->qpn; + event.param.ud.qkey = rep->qkey; + event.event = RDMA_CM_EVENT_ESTABLISHED; + event.status = 0; + break; + default: + printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d\n", + ib_event->event); + goto out; + } + + ret = id_priv->id.event_handler(&id_priv->id, &event); + if (ret) { + /* Destroy the CM ID by returning a non-zero value. */ + id_priv->cm_id.ib = NULL; + cma_exch(id_priv, RDMA_CM_DESTROYING); + mutex_unlock(&id_priv->handler_mutex); + rdma_destroy_id(&id_priv->id); + return ret; + } +out: + mutex_unlock(&id_priv->handler_mutex); + return ret; +} + +static int cma_resolve_ib_udp(struct rdma_id_private *id_priv, + struct rdma_conn_param *conn_param) +{ + struct ib_cm_sidr_req_param req; + struct ib_cm_id *id; + void *private_data; + int offset, ret; + + memset(&req, 0, sizeof req); + offset = cma_user_data_offset(id_priv); + req.private_data_len = offset + conn_param->private_data_len; + if (req.private_data_len < conn_param->private_data_len) + return -EINVAL; + + if (req.private_data_len) { + private_data = kzalloc(req.private_data_len, GFP_ATOMIC); + if (!private_data) + return -ENOMEM; + } else { + private_data = NULL; + } + + if (conn_param->private_data && conn_param->private_data_len) + memcpy(private_data + offset, conn_param->private_data, + conn_param->private_data_len); + + if (private_data) { + ret = cma_format_hdr(private_data, id_priv); + if (ret) + goto out; + req.private_data = private_data; + } + + id = ib_create_cm_id(id_priv->id.device, cma_sidr_rep_handler, + id_priv); + if (IS_ERR(id)) { + ret = PTR_ERR(id); + goto out; + } + id_priv->cm_id.ib = id; + + req.path = id_priv->id.route.path_rec; + req.service_id = rdma_get_service_id(&id_priv->id, cma_dst_addr(id_priv)); + req.timeout_ms = 1 << (CMA_CM_RESPONSE_TIMEOUT - 8); + req.max_cm_retries = CMA_MAX_CM_RETRIES; + + ret = ib_send_cm_sidr_req(id_priv->cm_id.ib, &req); + if (ret) { + ib_destroy_cm_id(id_priv->cm_id.ib); + id_priv->cm_id.ib = NULL; + } +out: + kfree(private_data); + return ret; +} + +static int cma_connect_ib(struct rdma_id_private *id_priv, + struct rdma_conn_param *conn_param) +{ + struct ib_cm_req_param req; + struct rdma_route *route; + void *private_data; + struct ib_cm_id *id; + int offset, ret; + + memset(&req, 0, sizeof req); + offset = cma_user_data_offset(id_priv); + req.private_data_len = offset + conn_param->private_data_len; + if (req.private_data_len < conn_param->private_data_len) + return -EINVAL; + + if (req.private_data_len) { + private_data = kzalloc(req.private_data_len, GFP_ATOMIC); + if (!private_data) + return -ENOMEM; + } else { + private_data = NULL; + } + + if (conn_param->private_data && conn_param->private_data_len) + memcpy(private_data + offset, conn_param->private_data, + conn_param->private_data_len); + + id = ib_create_cm_id(id_priv->id.device, cma_ib_handler, id_priv); + if (IS_ERR(id)) { + ret = PTR_ERR(id); + goto out; + } + id_priv->cm_id.ib = id; + + route = &id_priv->id.route; + if (private_data) { + ret = cma_format_hdr(private_data, id_priv); + if (ret) + goto out; + req.private_data = private_data; + } + + req.primary_path = &route->path_rec[0]; + if (route->num_paths == 2) + req.alternate_path = &route->path_rec[1]; + + req.service_id = rdma_get_service_id(&id_priv->id, cma_dst_addr(id_priv)); + req.qp_num = id_priv->qp_num; + req.qp_type = id_priv->id.qp_type; + req.starting_psn = id_priv->seq_num; + req.responder_resources = conn_param->responder_resources; + req.initiator_depth = conn_param->initiator_depth; + req.flow_control = conn_param->flow_control; + req.retry_count = min_t(u8, 7, conn_param->retry_count); + req.rnr_retry_count = min_t(u8, 7, conn_param->rnr_retry_count); + req.remote_cm_response_timeout = CMA_CM_RESPONSE_TIMEOUT; + req.local_cm_response_timeout = CMA_CM_RESPONSE_TIMEOUT; + req.max_cm_retries = CMA_MAX_CM_RETRIES; + req.srq = id_priv->srq ? 1 : 0; + + ret = ib_send_cm_req(id_priv->cm_id.ib, &req); +out: + if (ret && !IS_ERR(id)) { + ib_destroy_cm_id(id); + id_priv->cm_id.ib = NULL; + } + + kfree(private_data); + return ret; +} + +static int cma_connect_iw(struct rdma_id_private *id_priv, + struct rdma_conn_param *conn_param) +{ + struct iw_cm_id *cm_id; + int ret; + struct iw_cm_conn_param iw_param; + + cm_id = iw_create_cm_id(id_priv->id.device, cma_iw_handler, id_priv); + if (IS_ERR(cm_id)) + return PTR_ERR(cm_id); + + id_priv->cm_id.iw = cm_id; + + memcpy(&cm_id->local_addr, cma_src_addr(id_priv), + rdma_addr_size(cma_src_addr(id_priv))); + memcpy(&cm_id->remote_addr, cma_dst_addr(id_priv), + rdma_addr_size(cma_dst_addr(id_priv))); + + ret = cma_modify_qp_rtr(id_priv, conn_param); + if (ret) + goto out; + + if (conn_param) { + iw_param.ord = conn_param->initiator_depth; + iw_param.ird = conn_param->responder_resources; + iw_param.private_data = conn_param->private_data; + iw_param.private_data_len = conn_param->private_data_len; + iw_param.qpn = id_priv->id.qp ? id_priv->qp_num : conn_param->qp_num; + } else { + memset(&iw_param, 0, sizeof iw_param); + iw_param.qpn = id_priv->qp_num; + } + ret = iw_cm_connect(cm_id, &iw_param); +out: + if (ret) { + iw_destroy_cm_id(cm_id); + id_priv->cm_id.iw = NULL; + } + return ret; +} + +int rdma_connect(struct rdma_cm_id *id, struct rdma_conn_param *conn_param) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!cma_comp_exch(id_priv, RDMA_CM_ROUTE_RESOLVED, RDMA_CM_CONNECT)) + return -EINVAL; + + if (!id->qp) { + id_priv->qp_num = conn_param->qp_num; + id_priv->srq = conn_param->srq; + } + + switch (rdma_node_get_transport(id->device->node_type)) { + case RDMA_TRANSPORT_IB: + if (id->qp_type == IB_QPT_UD) + ret = cma_resolve_ib_udp(id_priv, conn_param); + else + ret = cma_connect_ib(id_priv, conn_param); + break; + case RDMA_TRANSPORT_IWARP: + ret = cma_connect_iw(id_priv, conn_param); + break; + default: + ret = -ENOSYS; + break; + } + if (ret) + goto err; + + return 0; +err: + cma_comp_exch(id_priv, RDMA_CM_CONNECT, RDMA_CM_ROUTE_RESOLVED); + return ret; +} +EXPORT_SYMBOL(rdma_connect); + +static int cma_accept_ib(struct rdma_id_private *id_priv, + struct rdma_conn_param *conn_param) +{ + struct ib_cm_rep_param rep; + int ret; + + ret = cma_modify_qp_rtr(id_priv, conn_param); + if (ret) + goto out; + + ret = cma_modify_qp_rts(id_priv, conn_param); + if (ret) + goto out; + + memset(&rep, 0, sizeof rep); + rep.qp_num = id_priv->qp_num; + rep.starting_psn = id_priv->seq_num; + rep.private_data = conn_param->private_data; + rep.private_data_len = conn_param->private_data_len; + rep.responder_resources = conn_param->responder_resources; + rep.initiator_depth = conn_param->initiator_depth; + rep.failover_accepted = 0; + rep.flow_control = conn_param->flow_control; + rep.rnr_retry_count = min_t(u8, 7, conn_param->rnr_retry_count); + rep.srq = id_priv->srq ? 1 : 0; + + ret = ib_send_cm_rep(id_priv->cm_id.ib, &rep); +out: + return ret; +} + +static int cma_accept_iw(struct rdma_id_private *id_priv, + struct rdma_conn_param *conn_param) +{ + struct iw_cm_conn_param iw_param; + int ret; + + ret = cma_modify_qp_rtr(id_priv, conn_param); + if (ret) + return ret; + + iw_param.ord = conn_param->initiator_depth; + iw_param.ird = conn_param->responder_resources; + iw_param.private_data = conn_param->private_data; + iw_param.private_data_len = conn_param->private_data_len; + if (id_priv->id.qp) { + iw_param.qpn = id_priv->qp_num; + } else + iw_param.qpn = conn_param->qp_num; + + return iw_cm_accept(id_priv->cm_id.iw, &iw_param); +} + +static int cma_send_sidr_rep(struct rdma_id_private *id_priv, + enum ib_cm_sidr_status status, u32 qkey, + const void *private_data, int private_data_len) +{ + struct ib_cm_sidr_rep_param rep; + int ret; + + memset(&rep, 0, sizeof rep); + rep.status = status; + if (status == IB_SIDR_SUCCESS) { + ret = cma_set_qkey(id_priv, qkey); + if (ret) + return ret; + rep.qp_num = id_priv->qp_num; + rep.qkey = id_priv->qkey; + } + rep.private_data = private_data; + rep.private_data_len = private_data_len; + + return ib_send_cm_sidr_rep(id_priv->cm_id.ib, &rep); +} + +int rdma_accept(struct rdma_cm_id *id, struct rdma_conn_param *conn_param) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + + id_priv->owner = task_pid_nr(current); + + if (!cma_comp(id_priv, RDMA_CM_CONNECT)) + return -EINVAL; + + if (!id->qp && conn_param) { + id_priv->qp_num = conn_param->qp_num; + id_priv->srq = conn_param->srq; + } + + switch (rdma_node_get_transport(id->device->node_type)) { + case RDMA_TRANSPORT_IB: + if (id->qp_type == IB_QPT_UD) { + if (conn_param) + ret = cma_send_sidr_rep(id_priv, IB_SIDR_SUCCESS, + conn_param->qkey, + conn_param->private_data, + conn_param->private_data_len); + else + ret = cma_send_sidr_rep(id_priv, IB_SIDR_SUCCESS, + 0, NULL, 0); + } else { + if (conn_param) + ret = cma_accept_ib(id_priv, conn_param); + else + ret = cma_rep_recv(id_priv); + } + break; + case RDMA_TRANSPORT_IWARP: + ret = cma_accept_iw(id_priv, conn_param); + break; + default: + ret = -ENOSYS; + break; + } + + if (ret) + goto reject; + + return 0; +reject: + cma_modify_qp_err(id_priv); + rdma_reject(id, NULL, 0); + return ret; +} +EXPORT_SYMBOL(rdma_accept); + +int rdma_notify(struct rdma_cm_id *id, enum ib_event_type event) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!id_priv->cm_id.ib) + return -EINVAL; + + switch (id->device->node_type) { + case RDMA_NODE_IB_CA: + ret = ib_cm_notify(id_priv->cm_id.ib, event); + break; + default: + ret = 0; + break; + } + return ret; +} +EXPORT_SYMBOL(rdma_notify); + +int rdma_reject(struct rdma_cm_id *id, const void *private_data, + u8 private_data_len) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!id_priv->cm_id.ib) + return -EINVAL; + + switch (rdma_node_get_transport(id->device->node_type)) { + case RDMA_TRANSPORT_IB: + if (id->qp_type == IB_QPT_UD) + ret = cma_send_sidr_rep(id_priv, IB_SIDR_REJECT, 0, + private_data, private_data_len); + else + ret = ib_send_cm_rej(id_priv->cm_id.ib, + IB_CM_REJ_CONSUMER_DEFINED, NULL, + 0, private_data, private_data_len); + break; + case RDMA_TRANSPORT_IWARP: + ret = iw_cm_reject(id_priv->cm_id.iw, + private_data, private_data_len); + break; + default: + ret = -ENOSYS; + break; + } + return ret; +} +EXPORT_SYMBOL(rdma_reject); + +int rdma_disconnect(struct rdma_cm_id *id) +{ + struct rdma_id_private *id_priv; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!id_priv->cm_id.ib) + return -EINVAL; + + switch (rdma_node_get_transport(id->device->node_type)) { + case RDMA_TRANSPORT_IB: + ret = cma_modify_qp_err(id_priv); + if (ret) + goto out; + /* Initiate or respond to a disconnect. */ + if (ib_send_cm_dreq(id_priv->cm_id.ib, NULL, 0)) + ib_send_cm_drep(id_priv->cm_id.ib, NULL, 0); + break; + case RDMA_TRANSPORT_IWARP: + ret = iw_cm_disconnect(id_priv->cm_id.iw, 0); + break; + default: + ret = -EINVAL; + break; + } +out: + return ret; +} +EXPORT_SYMBOL(rdma_disconnect); + +static int cma_ib_mc_handler(int status, struct ib_sa_multicast *multicast) +{ + struct rdma_id_private *id_priv; + struct cma_multicast *mc = multicast->context; + struct rdma_cm_event event; + int ret; + + id_priv = mc->id_priv; + if (cma_disable_callback(id_priv, RDMA_CM_ADDR_BOUND) && + cma_disable_callback(id_priv, RDMA_CM_ADDR_RESOLVED)) + return 0; + + if (!status) + status = cma_set_qkey(id_priv, be32_to_cpu(multicast->rec.qkey)); + mutex_lock(&id_priv->qp_mutex); + if (!status && id_priv->id.qp) + status = ib_attach_mcast(id_priv->id.qp, &multicast->rec.mgid, + be16_to_cpu(multicast->rec.mlid)); + mutex_unlock(&id_priv->qp_mutex); + + memset(&event, 0, sizeof event); + event.status = status; + event.param.ud.private_data = mc->context; + if (!status) { + event.event = RDMA_CM_EVENT_MULTICAST_JOIN; + ib_init_ah_from_mcmember(id_priv->id.device, + id_priv->id.port_num, &multicast->rec, + &event.param.ud.ah_attr); + event.param.ud.qp_num = 0xFFFFFF; + event.param.ud.qkey = be32_to_cpu(multicast->rec.qkey); + } else + event.event = RDMA_CM_EVENT_MULTICAST_ERROR; + + ret = id_priv->id.event_handler(&id_priv->id, &event); + if (ret) { + cma_exch(id_priv, RDMA_CM_DESTROYING); + mutex_unlock(&id_priv->handler_mutex); + rdma_destroy_id(&id_priv->id); + return 0; + } + + mutex_unlock(&id_priv->handler_mutex); + return 0; +} + +static void cma_set_mgid(struct rdma_id_private *id_priv, + struct sockaddr *addr, union ib_gid *mgid) +{ + unsigned char mc_map[MAX_ADDR_LEN]; + struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; + struct sockaddr_in *sin = (struct sockaddr_in *) addr; + struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) addr; + + if (cma_any_addr(addr)) { + memset(mgid, 0, sizeof *mgid); + } else if ((addr->sa_family == AF_INET6) && + ((be32_to_cpu(sin6->sin6_addr.s6_addr32[0]) & 0xFFF0FFFF) == + 0xFF10A01B)) { + /* IPv6 address is an SA assigned MGID. */ + memcpy(mgid, &sin6->sin6_addr, sizeof *mgid); + } else if (addr->sa_family == AF_IB) { + memcpy(mgid, &((struct sockaddr_ib *) addr)->sib_addr, sizeof *mgid); + } else if ((addr->sa_family == AF_INET6)) { + ipv6_ib_mc_map(&sin6->sin6_addr, dev_addr->broadcast, mc_map); + if (id_priv->id.ps == RDMA_PS_UDP) + mc_map[7] = 0x01; /* Use RDMA CM signature */ + *mgid = *(union ib_gid *) (mc_map + 4); + } else { + ip_ib_mc_map(sin->sin_addr.s_addr, dev_addr->broadcast, mc_map); + if (id_priv->id.ps == RDMA_PS_UDP) + mc_map[7] = 0x01; /* Use RDMA CM signature */ + *mgid = *(union ib_gid *) (mc_map + 4); + } +} + +static int cma_join_ib_multicast(struct rdma_id_private *id_priv, + struct cma_multicast *mc) +{ + struct ib_sa_mcmember_rec rec; + struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; + ib_sa_comp_mask comp_mask; + int ret; + + ib_addr_get_mgid(dev_addr, &rec.mgid); + ret = ib_sa_get_mcmember_rec(id_priv->id.device, id_priv->id.port_num, + &rec.mgid, &rec); + if (ret) + return ret; + + ret = cma_set_qkey(id_priv, 0); + if (ret) + return ret; + + cma_set_mgid(id_priv, (struct sockaddr *) &mc->addr, &rec.mgid); + rec.qkey = cpu_to_be32(id_priv->qkey); + rdma_addr_get_sgid(dev_addr, &rec.port_gid); + rec.pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr)); + rec.join_state = 1; + + comp_mask = IB_SA_MCMEMBER_REC_MGID | IB_SA_MCMEMBER_REC_PORT_GID | + IB_SA_MCMEMBER_REC_PKEY | IB_SA_MCMEMBER_REC_JOIN_STATE | + IB_SA_MCMEMBER_REC_QKEY | IB_SA_MCMEMBER_REC_SL | + IB_SA_MCMEMBER_REC_FLOW_LABEL | + IB_SA_MCMEMBER_REC_TRAFFIC_CLASS; + + if (id_priv->id.ps == RDMA_PS_IPOIB) + comp_mask |= IB_SA_MCMEMBER_REC_RATE | + IB_SA_MCMEMBER_REC_RATE_SELECTOR | + IB_SA_MCMEMBER_REC_MTU_SELECTOR | + IB_SA_MCMEMBER_REC_MTU | + IB_SA_MCMEMBER_REC_HOP_LIMIT; + + mc->multicast.ib = ib_sa_join_multicast(&sa_client, id_priv->id.device, + id_priv->id.port_num, &rec, + comp_mask, GFP_KERNEL, + cma_ib_mc_handler, mc); + return PTR_ERR_OR_ZERO(mc->multicast.ib); +} + +static void iboe_mcast_work_handler(struct work_struct *work) +{ + struct iboe_mcast_work *mw = container_of(work, struct iboe_mcast_work, work); + struct cma_multicast *mc = mw->mc; + struct ib_sa_multicast *m = mc->multicast.ib; + + mc->multicast.ib->context = mc; + cma_ib_mc_handler(0, m); + kref_put(&mc->mcref, release_mc); + kfree(mw); +} + +static void cma_iboe_set_mgid(struct sockaddr *addr, union ib_gid *mgid) +{ + struct sockaddr_in *sin = (struct sockaddr_in *)addr; + struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)addr; + + if (cma_any_addr(addr)) { + memset(mgid, 0, sizeof *mgid); + } else if (addr->sa_family == AF_INET6) { + memcpy(mgid, &sin6->sin6_addr, sizeof *mgid); + } else { + mgid->raw[0] = 0xff; + mgid->raw[1] = 0x0e; + mgid->raw[2] = 0; + mgid->raw[3] = 0; + mgid->raw[4] = 0; + mgid->raw[5] = 0; + mgid->raw[6] = 0; + mgid->raw[7] = 0; + mgid->raw[8] = 0; + mgid->raw[9] = 0; + mgid->raw[10] = 0xff; + mgid->raw[11] = 0xff; + *(__be32 *)(&mgid->raw[12]) = sin->sin_addr.s_addr; + } +} + +static int cma_iboe_join_multicast(struct rdma_id_private *id_priv, + struct cma_multicast *mc) +{ + struct iboe_mcast_work *work; + struct rdma_dev_addr *dev_addr = &id_priv->id.route.addr.dev_addr; + int err; + struct sockaddr *addr = (struct sockaddr *)&mc->addr; + struct net_device *ndev = NULL; + + if (cma_zero_addr((struct sockaddr *)&mc->addr)) + return -EINVAL; + + work = kzalloc(sizeof *work, GFP_KERNEL); + if (!work) + return -ENOMEM; + + mc->multicast.ib = kzalloc(sizeof(struct ib_sa_multicast), GFP_KERNEL); + if (!mc->multicast.ib) { + err = -ENOMEM; + goto out1; + } + + cma_iboe_set_mgid(addr, &mc->multicast.ib->rec.mgid); + + mc->multicast.ib->rec.pkey = cpu_to_be16(0xffff); + if (id_priv->id.ps == RDMA_PS_UDP) + mc->multicast.ib->rec.qkey = cpu_to_be32(RDMA_UDP_QKEY); + + if (dev_addr->bound_dev_if) + ndev = dev_get_by_index(&init_net, dev_addr->bound_dev_if); + if (!ndev) { + err = -ENODEV; + goto out2; + } + mc->multicast.ib->rec.rate = iboe_get_rate(ndev); + mc->multicast.ib->rec.hop_limit = 1; + mc->multicast.ib->rec.mtu = iboe_get_mtu(ndev->mtu); + dev_put(ndev); + if (!mc->multicast.ib->rec.mtu) { + err = -EINVAL; + goto out2; + } + rdma_ip2gid((struct sockaddr *)&id_priv->id.route.addr.src_addr, + &mc->multicast.ib->rec.port_gid); + work->id = id_priv; + work->mc = mc; + INIT_WORK(&work->work, iboe_mcast_work_handler); + kref_get(&mc->mcref); + queue_work(cma_wq, &work->work); + + return 0; + +out2: + kfree(mc->multicast.ib); +out1: + kfree(work); + return err; +} + +int rdma_join_multicast(struct rdma_cm_id *id, struct sockaddr *addr, + void *context) +{ + struct rdma_id_private *id_priv; + struct cma_multicast *mc; + int ret; + + id_priv = container_of(id, struct rdma_id_private, id); + if (!cma_comp(id_priv, RDMA_CM_ADDR_BOUND) && + !cma_comp(id_priv, RDMA_CM_ADDR_RESOLVED)) + return -EINVAL; + + mc = kmalloc(sizeof *mc, GFP_KERNEL); + if (!mc) + return -ENOMEM; + + memcpy(&mc->addr, addr, rdma_addr_size(addr)); + mc->context = context; + mc->id_priv = id_priv; + + spin_lock(&id_priv->lock); + list_add(&mc->list, &id_priv->mc_list); + spin_unlock(&id_priv->lock); + + switch (rdma_node_get_transport(id->device->node_type)) { + case RDMA_TRANSPORT_IB: + switch (rdma_port_get_link_layer(id->device, id->port_num)) { + case IB_LINK_LAYER_INFINIBAND: + ret = cma_join_ib_multicast(id_priv, mc); + break; + case IB_LINK_LAYER_ETHERNET: + kref_init(&mc->mcref); + ret = cma_iboe_join_multicast(id_priv, mc); + break; + default: + ret = -EINVAL; + } + break; + default: + ret = -ENOSYS; + break; + } + + if (ret) { + spin_lock_irq(&id_priv->lock); + list_del(&mc->list); + spin_unlock_irq(&id_priv->lock); + kfree(mc); + } + return ret; +} +EXPORT_SYMBOL(rdma_join_multicast); + +void rdma_leave_multicast(struct rdma_cm_id *id, struct sockaddr *addr) +{ + struct rdma_id_private *id_priv; + struct cma_multicast *mc; + + id_priv = container_of(id, struct rdma_id_private, id); + spin_lock_irq(&id_priv->lock); + list_for_each_entry(mc, &id_priv->mc_list, list) { + if (!memcmp(&mc->addr, addr, rdma_addr_size(addr))) { + list_del(&mc->list); + spin_unlock_irq(&id_priv->lock); + + if (id->qp) + ib_detach_mcast(id->qp, + &mc->multicast.ib->rec.mgid, + be16_to_cpu(mc->multicast.ib->rec.mlid)); + if (rdma_node_get_transport(id_priv->cma_dev->device->node_type) == RDMA_TRANSPORT_IB) { + switch (rdma_port_get_link_layer(id->device, id->port_num)) { + case IB_LINK_LAYER_INFINIBAND: + ib_sa_free_multicast(mc->multicast.ib); + kfree(mc); + break; + case IB_LINK_LAYER_ETHERNET: + kref_put(&mc->mcref, release_mc); + break; + default: + break; + } + } + return; + } + } + spin_unlock_irq(&id_priv->lock); +} +EXPORT_SYMBOL(rdma_leave_multicast); + +static int cma_netdev_change(struct net_device *ndev, struct rdma_id_private *id_priv) +{ + struct rdma_dev_addr *dev_addr; + struct cma_ndev_work *work; + + dev_addr = &id_priv->id.route.addr.dev_addr; + + if ((dev_addr->bound_dev_if == ndev->ifindex) && + memcmp(dev_addr->src_dev_addr, ndev->dev_addr, ndev->addr_len)) { + printk(KERN_INFO "RDMA CM addr change for ndev %s used by id %p\n", + ndev->name, &id_priv->id); + work = kzalloc(sizeof *work, GFP_KERNEL); + if (!work) + return -ENOMEM; + + INIT_WORK(&work->work, cma_ndev_work_handler); + work->id = id_priv; + work->event.event = RDMA_CM_EVENT_ADDR_CHANGE; + atomic_inc(&id_priv->refcount); + queue_work(cma_wq, &work->work); + } + + return 0; +} + +static int cma_netdev_callback(struct notifier_block *self, unsigned long event, + void *ptr) +{ + struct net_device *ndev = netdev_notifier_info_to_dev(ptr); + struct cma_device *cma_dev; + struct rdma_id_private *id_priv; + int ret = NOTIFY_DONE; + + if (dev_net(ndev) != &init_net) + return NOTIFY_DONE; + + if (event != NETDEV_BONDING_FAILOVER) + return NOTIFY_DONE; + + if (!(ndev->flags & IFF_MASTER) || !(ndev->priv_flags & IFF_BONDING)) + return NOTIFY_DONE; + + mutex_lock(&lock); + list_for_each_entry(cma_dev, &dev_list, list) + list_for_each_entry(id_priv, &cma_dev->id_list, list) { + ret = cma_netdev_change(ndev, id_priv); + if (ret) + goto out; + } + +out: + mutex_unlock(&lock); + return ret; +} + +static struct notifier_block cma_nb = { + .notifier_call = cma_netdev_callback +}; + +static void cma_add_one(struct ib_device *device) +{ + struct cma_device *cma_dev; + struct rdma_id_private *id_priv; + + cma_dev = kmalloc(sizeof *cma_dev, GFP_KERNEL); + if (!cma_dev) + return; + + cma_dev->device = device; + + init_completion(&cma_dev->comp); + atomic_set(&cma_dev->refcount, 1); + INIT_LIST_HEAD(&cma_dev->id_list); + ib_set_client_data(device, &cma_client, cma_dev); + + mutex_lock(&lock); + list_add_tail(&cma_dev->list, &dev_list); + list_for_each_entry(id_priv, &listen_any_list, list) + cma_listen_on_dev(id_priv, cma_dev); + mutex_unlock(&lock); +} + +static int cma_remove_id_dev(struct rdma_id_private *id_priv) +{ + struct rdma_cm_event event; + enum rdma_cm_state state; + int ret = 0; + + /* Record that we want to remove the device */ + state = cma_exch(id_priv, RDMA_CM_DEVICE_REMOVAL); + if (state == RDMA_CM_DESTROYING) + return 0; + + cma_cancel_operation(id_priv, state); + mutex_lock(&id_priv->handler_mutex); + + /* Check for destruction from another callback. */ + if (!cma_comp(id_priv, RDMA_CM_DEVICE_REMOVAL)) + goto out; + + memset(&event, 0, sizeof event); + event.event = RDMA_CM_EVENT_DEVICE_REMOVAL; + ret = id_priv->id.event_handler(&id_priv->id, &event); +out: + mutex_unlock(&id_priv->handler_mutex); + return ret; +} + +static void cma_process_remove(struct cma_device *cma_dev) +{ + struct rdma_id_private *id_priv; + int ret; + + mutex_lock(&lock); + while (!list_empty(&cma_dev->id_list)) { + id_priv = list_entry(cma_dev->id_list.next, + struct rdma_id_private, list); + + list_del(&id_priv->listen_list); + list_del_init(&id_priv->list); + atomic_inc(&id_priv->refcount); + mutex_unlock(&lock); + + ret = id_priv->internal_id ? 1 : cma_remove_id_dev(id_priv); + cma_deref_id(id_priv); + if (ret) + rdma_destroy_id(&id_priv->id); + + mutex_lock(&lock); + } + mutex_unlock(&lock); + + cma_deref_dev(cma_dev); + wait_for_completion(&cma_dev->comp); +} + +static void cma_remove_one(struct ib_device *device) +{ + struct cma_device *cma_dev; + + cma_dev = ib_get_client_data(device, &cma_client); + if (!cma_dev) + return; + + mutex_lock(&lock); + list_del(&cma_dev->list); + mutex_unlock(&lock); + + cma_process_remove(cma_dev); + kfree(cma_dev); +} + +static int cma_get_id_stats(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct nlmsghdr *nlh; + struct rdma_cm_id_stats *id_stats; + struct rdma_id_private *id_priv; + struct rdma_cm_id *id = NULL; + struct cma_device *cma_dev; + int i_dev = 0, i_id = 0; + + /* + * We export all of the IDs as a sequence of messages. Each + * ID gets its own netlink message. + */ + mutex_lock(&lock); + + list_for_each_entry(cma_dev, &dev_list, list) { + if (i_dev < cb->args[0]) { + i_dev++; + continue; + } + + i_id = 0; + list_for_each_entry(id_priv, &cma_dev->id_list, list) { + if (i_id < cb->args[1]) { + i_id++; + continue; + } + + id_stats = ibnl_put_msg(skb, &nlh, cb->nlh->nlmsg_seq, + sizeof *id_stats, RDMA_NL_RDMA_CM, + RDMA_NL_RDMA_CM_ID_STATS, + NLM_F_MULTI); + if (!id_stats) + goto out; + + memset(id_stats, 0, sizeof *id_stats); + id = &id_priv->id; + id_stats->node_type = id->route.addr.dev_addr.dev_type; + id_stats->port_num = id->port_num; + id_stats->bound_dev_if = + id->route.addr.dev_addr.bound_dev_if; + + if (ibnl_put_attr(skb, nlh, + rdma_addr_size(cma_src_addr(id_priv)), + cma_src_addr(id_priv), + RDMA_NL_RDMA_CM_ATTR_SRC_ADDR)) + goto out; + if (ibnl_put_attr(skb, nlh, + rdma_addr_size(cma_src_addr(id_priv)), + cma_dst_addr(id_priv), + RDMA_NL_RDMA_CM_ATTR_DST_ADDR)) + goto out; + + id_stats->pid = id_priv->owner; + id_stats->port_space = id->ps; + id_stats->cm_state = id_priv->state; + id_stats->qp_num = id_priv->qp_num; + id_stats->qp_type = id->qp_type; + + i_id++; + } + + cb->args[1] = 0; + i_dev++; + } + +out: + mutex_unlock(&lock); + cb->args[0] = i_dev; + cb->args[1] = i_id; + + return skb->len; +} + +static const struct ibnl_client_cbs cma_cb_table[] = { + [RDMA_NL_RDMA_CM_ID_STATS] = { .dump = cma_get_id_stats, + .module = THIS_MODULE }, +}; + +static int __init cma_init(void) +{ + int ret; + + cma_wq = create_singlethread_workqueue("rdma_cm"); + if (!cma_wq) + return -ENOMEM; + + ib_sa_register_client(&sa_client); + rdma_addr_register_client(&addr_client); + register_netdevice_notifier(&cma_nb); + + ret = ib_register_client(&cma_client); + if (ret) + goto err; + + if (ibnl_add_client(RDMA_NL_RDMA_CM, RDMA_NL_RDMA_CM_NUM_OPS, cma_cb_table)) + printk(KERN_WARNING "RDMA CMA: failed to add netlink callback\n"); + + return 0; + +err: + unregister_netdevice_notifier(&cma_nb); + rdma_addr_unregister_client(&addr_client); + ib_sa_unregister_client(&sa_client); + destroy_workqueue(cma_wq); + return ret; +} + +static void __exit cma_cleanup(void) +{ + ibnl_remove_client(RDMA_NL_RDMA_CM); + ib_unregister_client(&cma_client); + unregister_netdevice_notifier(&cma_nb); + rdma_addr_unregister_client(&addr_client); + ib_sa_unregister_client(&sa_client); + destroy_workqueue(cma_wq); + idr_destroy(&tcp_ps); + idr_destroy(&udp_ps); + idr_destroy(&ipoib_ps); + idr_destroy(&ib_ps); +} + +module_init(cma_init); +module_exit(cma_cleanup); diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h index 7ad47a4b166..87d1936f5c1 100644 --- a/drivers/infiniband/core/core_priv.h +++ b/drivers/infiniband/core/core_priv.h @@ -28,8 +28,6 @@ * 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. - * - * $Id: core_priv.h 1349 2004-12-16 21:09:43Z roland $ */ #ifndef _CORE_PRIV_H @@ -40,7 +38,9 @@ #include <rdma/ib_verbs.h> -int ib_device_register_sysfs(struct ib_device *device); +int ib_device_register_sysfs(struct ib_device *device, + int (*port_callback)(struct ib_device *, + u8, struct kobject *)); void ib_device_unregister_sysfs(struct ib_device *device); int ib_sysfs_setup(void); @@ -49,4 +49,6 @@ void ib_sysfs_cleanup(void); int ib_cache_setup(void); void ib_cache_cleanup(void); +int ib_resolve_eth_l2_attrs(struct ib_qp *qp, + struct ib_qp_attr *qp_attr, int *qp_attr_mask); #endif /* _CORE_PRIV_H */ diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c index b2f3cb91d9b..18c1ece765f 100644 --- a/drivers/infiniband/core/device.c +++ b/drivers/infiniband/core/device.c @@ -29,16 +29,16 @@ * 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. - * - * $Id: device.c 1349 2004-12-16 21:09:43Z roland $ */ #include <linux/module.h> #include <linux/string.h> #include <linux/errno.h> +#include <linux/kernel.h> #include <linux/slab.h> #include <linux/init.h> #include <linux/mutex.h> +#include <rdma/rdma_netlink.h> #include "core_priv.h" @@ -52,6 +52,9 @@ struct ib_client_data { void * data; }; +struct workqueue_struct *ib_wq; +EXPORT_SYMBOL_GPL(ib_wq); + static LIST_HEAD(device_list); static LIST_HEAD(client_list); @@ -93,7 +96,7 @@ static int ib_device_check_mandatory(struct ib_device *device) }; int i; - for (i = 0; i < sizeof mandatory_table / sizeof mandatory_table[0]; ++i) { + for (i = 0; i < ARRAY_SIZE(mandatory_table); ++i) { if (!*(void **) ((void *) device + mandatory_table[i].offset)) { printk(KERN_WARNING "Device %s is missing mandatory function %s\n", device->name, mandatory_table[i].name); @@ -118,12 +121,12 @@ static struct ib_device *__ib_device_get_by_name(const char *name) static int alloc_name(char *name) { - long *inuse; + unsigned long *inuse; char buf[IB_DEVICE_NAME_MAX]; struct ib_device *device; int i; - inuse = (long *) get_zeroed_page(GFP_KERNEL); + inuse = (unsigned long *) get_zeroed_page(GFP_KERNEL); if (!inuse) return -ENOMEM; @@ -148,6 +151,18 @@ static int alloc_name(char *name) return 0; } +static int start_port(struct ib_device *device) +{ + return (device->node_type == RDMA_NODE_IB_SWITCH) ? 0 : 1; +} + + +static int end_port(struct ib_device *device) +{ + return (device->node_type == RDMA_NODE_IB_SWITCH) ? + 0 : device->phys_port_cnt; +} + /** * ib_alloc_device - allocate an IB device struct * @size:size of structure to allocate @@ -181,7 +196,7 @@ void ib_dealloc_device(struct ib_device *device) BUG_ON(device->reg_state != IB_DEV_UNREGISTERED); - ib_device_unregister_sysfs(device); + kobject_put(&device->dev.kobj); } EXPORT_SYMBOL(ib_dealloc_device); @@ -207,6 +222,45 @@ static int add_client_context(struct ib_device *device, struct ib_client *client return 0; } +static int read_port_table_lengths(struct ib_device *device) +{ + struct ib_port_attr *tprops = NULL; + int num_ports, ret = -ENOMEM; + u8 port_index; + + tprops = kmalloc(sizeof *tprops, GFP_KERNEL); + if (!tprops) + goto out; + + num_ports = end_port(device) - start_port(device) + 1; + + device->pkey_tbl_len = kmalloc(sizeof *device->pkey_tbl_len * num_ports, + GFP_KERNEL); + device->gid_tbl_len = kmalloc(sizeof *device->gid_tbl_len * num_ports, + GFP_KERNEL); + if (!device->pkey_tbl_len || !device->gid_tbl_len) + goto err; + + for (port_index = 0; port_index < num_ports; ++port_index) { + ret = ib_query_port(device, port_index + start_port(device), + tprops); + if (ret) + goto err; + device->pkey_tbl_len[port_index] = tprops->pkey_tbl_len; + device->gid_tbl_len[port_index] = tprops->gid_tbl_len; + } + + ret = 0; + goto out; + +err: + kfree(device->gid_tbl_len); + kfree(device->pkey_tbl_len); +out: + kfree(tprops); + return ret; +} + /** * ib_register_device - Register an IB device with IB core * @device:Device to register @@ -216,7 +270,9 @@ static int add_client_context(struct ib_device *device, struct ib_client *client * callback for each device that is added. @device must be allocated * with ib_alloc_device(). */ -int ib_register_device(struct ib_device *device) +int ib_register_device(struct ib_device *device, + int (*port_callback)(struct ib_device *, + u8, struct kobject *)) { int ret; @@ -238,10 +294,19 @@ int ib_register_device(struct ib_device *device) spin_lock_init(&device->event_handler_lock); spin_lock_init(&device->client_data_lock); - ret = ib_device_register_sysfs(device); + ret = read_port_table_lengths(device); + if (ret) { + printk(KERN_WARNING "Couldn't create table lengths cache for device %s\n", + device->name); + goto out; + } + + ret = ib_device_register_sysfs(device, port_callback); if (ret) { printk(KERN_WARNING "Couldn't register device %s with driver model\n", device->name); + kfree(device->gid_tbl_len); + kfree(device->pkey_tbl_len); goto out; } @@ -283,8 +348,13 @@ void ib_unregister_device(struct ib_device *device) list_del(&device->core_list); + kfree(device->gid_tbl_len); + kfree(device->pkey_tbl_len); + mutex_unlock(&device_mutex); + ib_device_unregister_sysfs(device); + spin_lock_irqsave(&device->client_data_lock, flags); list_for_each_entry_safe(context, tmp, &device->client_data_list, list) kfree(context); @@ -385,7 +455,7 @@ void *ib_get_client_data(struct ib_device *device, struct ib_client *client) EXPORT_SYMBOL(ib_get_client_data); /** - * ib_set_client_data - Get IB client context + * ib_set_client_data - Set IB client context * @device:Device to set context for * @client:Client to set context for * @data:Context to set @@ -505,10 +575,7 @@ int ib_query_port(struct ib_device *device, u8 port_num, struct ib_port_attr *port_attr) { - if (device->node_type == IB_NODE_SWITCH) { - if (port_num) - return -EINVAL; - } else if (port_num < 1 || port_num > device->phys_port_cnt) + if (port_num < start_port(device) || port_num > end_port(device)) return -EINVAL; return device->query_port(device, port_num, port_attr); @@ -560,6 +627,9 @@ int ib_modify_device(struct ib_device *device, int device_modify_mask, struct ib_device_modify *device_modify) { + if (!device->modify_device) + return -ENOSYS; + return device->modify_device(device, device_modify_mask, device_modify); } @@ -580,10 +650,10 @@ int ib_modify_port(struct ib_device *device, u8 port_num, int port_modify_mask, struct ib_port_modify *port_modify) { - if (device->node_type == IB_NODE_SWITCH) { - if (port_num) - return -EINVAL; - } else if (port_num < 1 || port_num > device->phys_port_cnt) + if (!device->modify_port) + return -ENOSYS; + + if (port_num < start_port(device) || port_num > end_port(device)) return -EINVAL; return device->modify_port(device, port_num, port_modify_mask, @@ -591,27 +661,124 @@ int ib_modify_port(struct ib_device *device, } EXPORT_SYMBOL(ib_modify_port); +/** + * ib_find_gid - Returns the port number and GID table index where + * a specified GID value occurs. + * @device: The device to query. + * @gid: The GID value to search for. + * @port_num: The port number of the device where the GID value was found. + * @index: The index into the GID table where the GID was found. This + * parameter may be NULL. + */ +int ib_find_gid(struct ib_device *device, union ib_gid *gid, + u8 *port_num, u16 *index) +{ + union ib_gid tmp_gid; + int ret, port, i; + + for (port = start_port(device); port <= end_port(device); ++port) { + for (i = 0; i < device->gid_tbl_len[port - start_port(device)]; ++i) { + ret = ib_query_gid(device, port, i, &tmp_gid); + if (ret) + return ret; + if (!memcmp(&tmp_gid, gid, sizeof *gid)) { + *port_num = port; + if (index) + *index = i; + return 0; + } + } + } + + return -ENOENT; +} +EXPORT_SYMBOL(ib_find_gid); + +/** + * ib_find_pkey - Returns the PKey table index where a specified + * PKey value occurs. + * @device: The device to query. + * @port_num: The port number of the device to search for the PKey. + * @pkey: The PKey value to search for. + * @index: The index into the PKey table where the PKey was found. + */ +int ib_find_pkey(struct ib_device *device, + u8 port_num, u16 pkey, u16 *index) +{ + int ret, i; + u16 tmp_pkey; + int partial_ix = -1; + + for (i = 0; i < device->pkey_tbl_len[port_num - start_port(device)]; ++i) { + ret = ib_query_pkey(device, port_num, i, &tmp_pkey); + if (ret) + return ret; + if ((pkey & 0x7fff) == (tmp_pkey & 0x7fff)) { + /* if there is full-member pkey take it.*/ + if (tmp_pkey & 0x8000) { + *index = i; + return 0; + } + if (partial_ix < 0) + partial_ix = i; + } + } + + /*no full-member, if exists take the limited*/ + if (partial_ix >= 0) { + *index = partial_ix; + return 0; + } + return -ENOENT; +} +EXPORT_SYMBOL(ib_find_pkey); + static int __init ib_core_init(void) { int ret; + ib_wq = alloc_workqueue("infiniband", 0, 0); + if (!ib_wq) + return -ENOMEM; + ret = ib_sysfs_setup(); - if (ret) + if (ret) { printk(KERN_WARNING "Couldn't create InfiniBand device class\n"); + goto err; + } + + ret = ibnl_init(); + if (ret) { + printk(KERN_WARNING "Couldn't init IB netlink interface\n"); + goto err_sysfs; + } ret = ib_cache_setup(); if (ret) { printk(KERN_WARNING "Couldn't set up InfiniBand P_Key/GID cache\n"); - ib_sysfs_cleanup(); + goto err_nl; } + return 0; + +err_nl: + ibnl_cleanup(); + +err_sysfs: + ib_sysfs_cleanup(); + +err: + destroy_workqueue(ib_wq); return ret; } static void __exit ib_core_cleanup(void) { ib_cache_cleanup(); + ibnl_cleanup(); ib_sysfs_cleanup(); + /* Make sure that any pending umem accounting work is done. */ + destroy_workqueue(ib_wq); } module_init(ib_core_init); diff --git a/drivers/infiniband/core/fmr_pool.c b/drivers/infiniband/core/fmr_pool.c index d34a6f1c4f4..9f5ad7cc33c 100644 --- a/drivers/infiniband/core/fmr_pool.c +++ b/drivers/infiniband/core/fmr_pool.c @@ -29,12 +29,11 @@ * 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. - * - * $Id: fmr_pool.c 2730 2005-06-28 16:43:03Z sean.hefty $ */ #include <linux/errno.h> #include <linux/spinlock.h> +#include <linux/export.h> #include <linux/slab.h> #include <linux/jhash.h> #include <linux/kthread.h> @@ -43,6 +42,8 @@ #include "core_priv.h" +#define PFX "fmr_pool: " + enum { IB_FMR_MAX_REMAPS = 32, @@ -54,7 +55,7 @@ enum { /* * If an FMR is not in use, then the list member will point to either * its pool's free_list (if the FMR can be mapped again; that is, - * remap_count < IB_FMR_MAX_REMAPS) or its pool's dirty_list (if the + * remap_count < pool->max_remaps) or its pool's dirty_list (if the * FMR needs to be unmapped before being remapped). In either of * these cases it is a bug if the ref_count is not 0. In other words, * if ref_count is > 0, then the list member must not be linked into @@ -84,6 +85,7 @@ struct ib_fmr_pool { int pool_size; int max_pages; + int max_remaps; int dirty_watermark; int dirty_len; struct list_head free_list; @@ -116,14 +118,13 @@ static inline struct ib_pool_fmr *ib_fmr_cache_lookup(struct ib_fmr_pool *pool, { struct hlist_head *bucket; struct ib_pool_fmr *fmr; - struct hlist_node *pos; if (!pool->cache_bucket) return NULL; bucket = pool->cache_bucket + ib_fmr_hash(*page_list); - hlist_for_each_entry(fmr, pos, bucket, cache_node) + hlist_for_each_entry(fmr, bucket, cache_node) if (io_virtual_address == fmr->io_virtual_address && page_list_len == fmr->page_list_len && !memcmp(page_list, fmr->page_list, @@ -149,14 +150,13 @@ static void ib_fmr_batch_release(struct ib_fmr_pool *pool) #ifdef DEBUG if (fmr->ref_count !=0) { - printk(KERN_WARNING "Unmapping FMR 0x%08x with ref count %d", + printk(KERN_WARNING PFX "Unmapping FMR 0x%08x with ref count %d\n", fmr, fmr->ref_count); } #endif } - list_splice(&pool->dirty_list, &unmap_list); - INIT_LIST_HEAD(&pool->dirty_list); + list_splice_init(&pool->dirty_list, &unmap_list); pool->dirty_len = 0; spin_unlock_irq(&pool->pool_lock); @@ -167,7 +167,7 @@ static void ib_fmr_batch_release(struct ib_fmr_pool *pool) ret = ib_unmap_fmr(&fmr_list); if (ret) - printk(KERN_WARNING "ib_unmap_fmr returned %d", ret); + printk(KERN_WARNING PFX "ib_unmap_fmr returned %d\n", ret); spin_lock_irq(&pool->pool_lock); list_splice(&unmap_list, &pool->free_list); @@ -179,8 +179,7 @@ static int ib_fmr_cleanup_thread(void *pool_ptr) struct ib_fmr_pool *pool = pool_ptr; do { - if (pool->dirty_len >= pool->dirty_watermark || - atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) < 0) { + if (atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) < 0) { ib_fmr_batch_release(pool); atomic_inc(&pool->flush_ser); @@ -191,8 +190,7 @@ static int ib_fmr_cleanup_thread(void *pool_ptr) } set_current_state(TASK_INTERRUPTIBLE); - if (pool->dirty_len < pool->dirty_watermark && - atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) >= 0 && + if (atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) >= 0 && !kthread_should_stop()) schedule(); __set_current_state(TASK_RUNNING); @@ -214,8 +212,10 @@ struct ib_fmr_pool *ib_create_fmr_pool(struct ib_pd *pd, { struct ib_device *device; struct ib_fmr_pool *pool; + struct ib_device_attr *attr; int i; int ret; + int max_remaps; if (!params) return ERR_PTR(-EINVAL); @@ -223,14 +223,34 @@ struct ib_fmr_pool *ib_create_fmr_pool(struct ib_pd *pd, device = pd->device; if (!device->alloc_fmr || !device->dealloc_fmr || !device->map_phys_fmr || !device->unmap_fmr) { - printk(KERN_WARNING "Device %s does not support fast memory regions", + printk(KERN_INFO PFX "Device %s does not support FMRs\n", device->name); return ERR_PTR(-ENOSYS); } + attr = kmalloc(sizeof *attr, GFP_KERNEL); + if (!attr) { + printk(KERN_WARNING PFX "couldn't allocate device attr struct\n"); + return ERR_PTR(-ENOMEM); + } + + ret = ib_query_device(device, attr); + if (ret) { + printk(KERN_WARNING PFX "couldn't query device: %d\n", ret); + kfree(attr); + return ERR_PTR(ret); + } + + if (!attr->max_map_per_fmr) + max_remaps = IB_FMR_MAX_REMAPS; + else + max_remaps = attr->max_map_per_fmr; + + kfree(attr); + pool = kmalloc(sizeof *pool, GFP_KERNEL); if (!pool) { - printk(KERN_WARNING "couldn't allocate pool struct"); + printk(KERN_WARNING PFX "couldn't allocate pool struct\n"); return ERR_PTR(-ENOMEM); } @@ -247,7 +267,7 @@ struct ib_fmr_pool *ib_create_fmr_pool(struct ib_pd *pd, kmalloc(IB_FMR_HASH_SIZE * sizeof *pool->cache_bucket, GFP_KERNEL); if (!pool->cache_bucket) { - printk(KERN_WARNING "Failed to allocate cache in pool"); + printk(KERN_WARNING PFX "Failed to allocate cache in pool\n"); ret = -ENOMEM; goto out_free_pool; } @@ -258,6 +278,7 @@ struct ib_fmr_pool *ib_create_fmr_pool(struct ib_pd *pd, pool->pool_size = 0; pool->max_pages = params->max_pages_per_fmr; + pool->max_remaps = max_remaps; pool->dirty_watermark = params->dirty_watermark; pool->dirty_len = 0; spin_lock_init(&pool->pool_lock); @@ -265,30 +286,33 @@ struct ib_fmr_pool *ib_create_fmr_pool(struct ib_pd *pd, atomic_set(&pool->flush_ser, 0); init_waitqueue_head(&pool->force_wait); - pool->thread = kthread_create(ib_fmr_cleanup_thread, - pool, - "ib_fmr(%s)", - device->name); + pool->thread = kthread_run(ib_fmr_cleanup_thread, + pool, + "ib_fmr(%s)", + device->name); if (IS_ERR(pool->thread)) { - printk(KERN_WARNING "couldn't start cleanup thread"); + printk(KERN_WARNING PFX "couldn't start cleanup thread\n"); ret = PTR_ERR(pool->thread); goto out_free_pool; } { struct ib_pool_fmr *fmr; - struct ib_fmr_attr attr = { - .max_pages = params->max_pages_per_fmr, - .max_maps = IB_FMR_MAX_REMAPS, - .page_size = PAGE_SHIFT + struct ib_fmr_attr fmr_attr = { + .max_pages = params->max_pages_per_fmr, + .max_maps = pool->max_remaps, + .page_shift = params->page_shift }; + int bytes_per_fmr = sizeof *fmr; + + if (pool->cache_bucket) + bytes_per_fmr += params->max_pages_per_fmr * sizeof (u64); for (i = 0; i < params->pool_size; ++i) { - fmr = kmalloc(sizeof *fmr + params->max_pages_per_fmr * sizeof (u64), - GFP_KERNEL); + fmr = kmalloc(bytes_per_fmr, GFP_KERNEL); if (!fmr) { - printk(KERN_WARNING "failed to allocate fmr struct " - "for FMR %d", i); + printk(KERN_WARNING PFX "failed to allocate fmr " + "struct for FMR %d\n", i); goto out_fail; } @@ -297,9 +321,10 @@ struct ib_fmr_pool *ib_create_fmr_pool(struct ib_pd *pd, fmr->ref_count = 0; INIT_HLIST_NODE(&fmr->cache_node); - fmr->fmr = ib_alloc_fmr(pd, params->access, &attr); + fmr->fmr = ib_alloc_fmr(pd, params->access, &fmr_attr); if (IS_ERR(fmr->fmr)) { - printk(KERN_WARNING "fmr_create failed for FMR %d", i); + printk(KERN_WARNING PFX "fmr_create failed " + "for FMR %d\n", i); kfree(fmr); goto out_fail; } @@ -354,7 +379,7 @@ void ib_destroy_fmr_pool(struct ib_fmr_pool *pool) } if (i < pool->pool_size) - printk(KERN_WARNING "pool still has %d regions registered", + printk(KERN_WARNING PFX "pool still has %d regions registered\n", pool->pool_size - i); kfree(pool->cache_bucket); @@ -371,19 +396,26 @@ EXPORT_SYMBOL(ib_destroy_fmr_pool); int ib_flush_fmr_pool(struct ib_fmr_pool *pool) { int serial; + struct ib_pool_fmr *fmr, *next; - atomic_inc(&pool->req_ser); /* - * It's OK if someone else bumps req_ser again here -- we'll - * just wait a little longer. + * The free_list holds FMRs that may have been used + * but have not been remapped enough times to be dirty. + * Put them on the dirty list now so that the cleanup + * thread will reap them too. */ - serial = atomic_read(&pool->req_ser); + spin_lock_irq(&pool->pool_lock); + list_for_each_entry_safe(fmr, next, &pool->free_list, list) { + if (fmr->remap_count > 0) + list_move(&fmr->list, &pool->dirty_list); + } + spin_unlock_irq(&pool->pool_lock); + serial = atomic_inc_return(&pool->req_ser); wake_up_process(pool->thread); if (wait_event_interruptible(pool->force_wait, - atomic_read(&pool->flush_ser) - - atomic_read(&pool->req_ser) >= 0)) + atomic_read(&pool->flush_ser) - serial >= 0)) return -EINTR; return 0; @@ -402,7 +434,7 @@ EXPORT_SYMBOL(ib_flush_fmr_pool); struct ib_pool_fmr *ib_fmr_pool_map_phys(struct ib_fmr_pool *pool_handle, u64 *page_list, int list_len, - u64 *io_virtual_address) + u64 io_virtual_address) { struct ib_fmr_pool *pool = pool_handle; struct ib_pool_fmr *fmr; @@ -416,7 +448,7 @@ struct ib_pool_fmr *ib_fmr_pool_map_phys(struct ib_fmr_pool *pool_handle, fmr = ib_fmr_cache_lookup(pool, page_list, list_len, - *io_virtual_address); + io_virtual_address); if (fmr) { /* found in cache */ ++fmr->ref_count; @@ -440,15 +472,14 @@ struct ib_pool_fmr *ib_fmr_pool_map_phys(struct ib_fmr_pool *pool_handle, spin_unlock_irqrestore(&pool->pool_lock, flags); result = ib_map_phys_fmr(fmr->fmr, page_list, list_len, - *io_virtual_address); + io_virtual_address); if (result) { spin_lock_irqsave(&pool->pool_lock, flags); list_add(&fmr->list, &pool->free_list); spin_unlock_irqrestore(&pool->pool_lock, flags); - printk(KERN_WARNING "fmr_map returns %d\n", - result); + printk(KERN_WARNING PFX "fmr_map returns %d\n", result); return ERR_PTR(result); } @@ -457,7 +488,7 @@ struct ib_pool_fmr *ib_fmr_pool_map_phys(struct ib_fmr_pool *pool_handle, fmr->ref_count = 1; if (pool->cache_bucket) { - fmr->io_virtual_address = *io_virtual_address; + fmr->io_virtual_address = io_virtual_address; fmr->page_list_len = list_len; memcpy(fmr->page_list, page_list, list_len * sizeof(*page_list)); @@ -489,18 +520,20 @@ int ib_fmr_pool_unmap(struct ib_pool_fmr *fmr) --fmr->ref_count; if (!fmr->ref_count) { - if (fmr->remap_count < IB_FMR_MAX_REMAPS) { + if (fmr->remap_count < pool->max_remaps) { list_add_tail(&fmr->list, &pool->free_list); } else { list_add_tail(&fmr->list, &pool->dirty_list); - ++pool->dirty_len; - wake_up_process(pool->thread); + if (++pool->dirty_len >= pool->dirty_watermark) { + atomic_inc(&pool->req_ser); + wake_up_process(pool->thread); + } } } #ifdef DEBUG if (fmr->ref_count < 0) - printk(KERN_WARNING "FMR %p has ref count %d < 0", + printk(KERN_WARNING PFX "FMR %p has ref count %d < 0\n", fmr, fmr->ref_count); #endif diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c new file mode 100644 index 00000000000..3d2e489ab73 --- /dev/null +++ b/drivers/infiniband/core/iwcm.c @@ -0,0 +1,1042 @@ +/* + * Copyright (c) 2004, 2005 Intel Corporation. All rights reserved. + * Copyright (c) 2004 Topspin Corporation. All rights reserved. + * Copyright (c) 2004, 2005 Voltaire Corporation. All rights reserved. + * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved. + * Copyright (c) 2005 Network Appliance, Inc. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ +#include <linux/dma-mapping.h> +#include <linux/err.h> +#include <linux/idr.h> +#include <linux/interrupt.h> +#include <linux/rbtree.h> +#include <linux/sched.h> +#include <linux/spinlock.h> +#include <linux/workqueue.h> +#include <linux/completion.h> +#include <linux/slab.h> +#include <linux/module.h> + +#include <rdma/iw_cm.h> +#include <rdma/ib_addr.h> + +#include "iwcm.h" + +MODULE_AUTHOR("Tom Tucker"); +MODULE_DESCRIPTION("iWARP CM"); +MODULE_LICENSE("Dual BSD/GPL"); + +static struct workqueue_struct *iwcm_wq; +struct iwcm_work { + struct work_struct work; + struct iwcm_id_private *cm_id; + struct list_head list; + struct iw_cm_event event; + struct list_head free_list; +}; + +/* + * The following services provide a mechanism for pre-allocating iwcm_work + * elements. The design pre-allocates them based on the cm_id type: + * LISTENING IDS: Get enough elements preallocated to handle the + * listen backlog. + * ACTIVE IDS: 4: CONNECT_REPLY, ESTABLISHED, DISCONNECT, CLOSE + * PASSIVE IDS: 3: ESTABLISHED, DISCONNECT, CLOSE + * + * Allocating them in connect and listen avoids having to deal + * with allocation failures on the event upcall from the provider (which + * is called in the interrupt context). + * + * One exception is when creating the cm_id for incoming connection requests. + * There are two cases: + * 1) in the event upcall, cm_event_handler(), for a listening cm_id. If + * the backlog is exceeded, then no more connection request events will + * be processed. cm_event_handler() returns -ENOMEM in this case. Its up + * to the provider to reject the connection request. + * 2) in the connection request workqueue handler, cm_conn_req_handler(). + * If work elements cannot be allocated for the new connect request cm_id, + * then IWCM will call the provider reject method. This is ok since + * cm_conn_req_handler() runs in the workqueue thread context. + */ + +static struct iwcm_work *get_work(struct iwcm_id_private *cm_id_priv) +{ + struct iwcm_work *work; + + if (list_empty(&cm_id_priv->work_free_list)) + return NULL; + work = list_entry(cm_id_priv->work_free_list.next, struct iwcm_work, + free_list); + list_del_init(&work->free_list); + return work; +} + +static void put_work(struct iwcm_work *work) +{ + list_add(&work->free_list, &work->cm_id->work_free_list); +} + +static void dealloc_work_entries(struct iwcm_id_private *cm_id_priv) +{ + struct list_head *e, *tmp; + + list_for_each_safe(e, tmp, &cm_id_priv->work_free_list) + kfree(list_entry(e, struct iwcm_work, free_list)); +} + +static int alloc_work_entries(struct iwcm_id_private *cm_id_priv, int count) +{ + struct iwcm_work *work; + + BUG_ON(!list_empty(&cm_id_priv->work_free_list)); + while (count--) { + work = kmalloc(sizeof(struct iwcm_work), GFP_KERNEL); + if (!work) { + dealloc_work_entries(cm_id_priv); + return -ENOMEM; + } + work->cm_id = cm_id_priv; + INIT_LIST_HEAD(&work->list); + put_work(work); + } + return 0; +} + +/* + * Save private data from incoming connection requests to + * iw_cm_event, so the low level driver doesn't have to. Adjust + * the event ptr to point to the local copy. + */ +static int copy_private_data(struct iw_cm_event *event) +{ + void *p; + + p = kmemdup(event->private_data, event->private_data_len, GFP_ATOMIC); + if (!p) + return -ENOMEM; + event->private_data = p; + return 0; +} + +static void free_cm_id(struct iwcm_id_private *cm_id_priv) +{ + dealloc_work_entries(cm_id_priv); + kfree(cm_id_priv); +} + +/* + * Release a reference on cm_id. If the last reference is being + * released, enable the waiting thread (in iw_destroy_cm_id) to + * get woken up, and return 1 if a thread is already waiting. + */ +static int iwcm_deref_id(struct iwcm_id_private *cm_id_priv) +{ + BUG_ON(atomic_read(&cm_id_priv->refcount)==0); + if (atomic_dec_and_test(&cm_id_priv->refcount)) { + BUG_ON(!list_empty(&cm_id_priv->work_list)); + complete(&cm_id_priv->destroy_comp); + return 1; + } + + return 0; +} + +static void add_ref(struct iw_cm_id *cm_id) +{ + struct iwcm_id_private *cm_id_priv; + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + atomic_inc(&cm_id_priv->refcount); +} + +static void rem_ref(struct iw_cm_id *cm_id) +{ + struct iwcm_id_private *cm_id_priv; + int cb_destroy; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + + /* + * Test bit before deref in case the cm_id gets freed on another + * thread. + */ + cb_destroy = test_bit(IWCM_F_CALLBACK_DESTROY, &cm_id_priv->flags); + if (iwcm_deref_id(cm_id_priv) && cb_destroy) { + BUG_ON(!list_empty(&cm_id_priv->work_list)); + free_cm_id(cm_id_priv); + } +} + +static int cm_event_handler(struct iw_cm_id *cm_id, struct iw_cm_event *event); + +struct iw_cm_id *iw_create_cm_id(struct ib_device *device, + iw_cm_handler cm_handler, + void *context) +{ + struct iwcm_id_private *cm_id_priv; + + cm_id_priv = kzalloc(sizeof(*cm_id_priv), GFP_KERNEL); + if (!cm_id_priv) + return ERR_PTR(-ENOMEM); + + cm_id_priv->state = IW_CM_STATE_IDLE; + cm_id_priv->id.device = device; + cm_id_priv->id.cm_handler = cm_handler; + cm_id_priv->id.context = context; + cm_id_priv->id.event_handler = cm_event_handler; + cm_id_priv->id.add_ref = add_ref; + cm_id_priv->id.rem_ref = rem_ref; + spin_lock_init(&cm_id_priv->lock); + atomic_set(&cm_id_priv->refcount, 1); + init_waitqueue_head(&cm_id_priv->connect_wait); + init_completion(&cm_id_priv->destroy_comp); + INIT_LIST_HEAD(&cm_id_priv->work_list); + INIT_LIST_HEAD(&cm_id_priv->work_free_list); + + return &cm_id_priv->id; +} +EXPORT_SYMBOL(iw_create_cm_id); + + +static int iwcm_modify_qp_err(struct ib_qp *qp) +{ + struct ib_qp_attr qp_attr; + + if (!qp) + return -EINVAL; + + qp_attr.qp_state = IB_QPS_ERR; + return ib_modify_qp(qp, &qp_attr, IB_QP_STATE); +} + +/* + * This is really the RDMAC CLOSING state. It is most similar to the + * IB SQD QP state. + */ +static int iwcm_modify_qp_sqd(struct ib_qp *qp) +{ + struct ib_qp_attr qp_attr; + + BUG_ON(qp == NULL); + qp_attr.qp_state = IB_QPS_SQD; + return ib_modify_qp(qp, &qp_attr, IB_QP_STATE); +} + +/* + * CM_ID <-- CLOSING + * + * Block if a passive or active connection is currently being processed. Then + * process the event as follows: + * - If we are ESTABLISHED, move to CLOSING and modify the QP state + * based on the abrupt flag + * - If the connection is already in the CLOSING or IDLE state, the peer is + * disconnecting concurrently with us and we've already seen the + * DISCONNECT event -- ignore the request and return 0 + * - Disconnect on a listening endpoint returns -EINVAL + */ +int iw_cm_disconnect(struct iw_cm_id *cm_id, int abrupt) +{ + struct iwcm_id_private *cm_id_priv; + unsigned long flags; + int ret = 0; + struct ib_qp *qp = NULL; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + /* Wait if we're currently in a connect or accept downcall */ + wait_event(cm_id_priv->connect_wait, + !test_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags)); + + spin_lock_irqsave(&cm_id_priv->lock, flags); + switch (cm_id_priv->state) { + case IW_CM_STATE_ESTABLISHED: + cm_id_priv->state = IW_CM_STATE_CLOSING; + + /* QP could be <nul> for user-mode client */ + if (cm_id_priv->qp) + qp = cm_id_priv->qp; + else + ret = -EINVAL; + break; + case IW_CM_STATE_LISTEN: + ret = -EINVAL; + break; + case IW_CM_STATE_CLOSING: + /* remote peer closed first */ + case IW_CM_STATE_IDLE: + /* accept or connect returned !0 */ + break; + case IW_CM_STATE_CONN_RECV: + /* + * App called disconnect before/without calling accept after + * connect_request event delivered. + */ + break; + case IW_CM_STATE_CONN_SENT: + /* Can only get here if wait above fails */ + default: + BUG(); + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + if (qp) { + if (abrupt) + ret = iwcm_modify_qp_err(qp); + else + ret = iwcm_modify_qp_sqd(qp); + + /* + * If both sides are disconnecting the QP could + * already be in ERR or SQD states + */ + ret = 0; + } + + return ret; +} +EXPORT_SYMBOL(iw_cm_disconnect); + +/* + * CM_ID <-- DESTROYING + * + * Clean up all resources associated with the connection and release + * the initial reference taken by iw_create_cm_id. + */ +static void destroy_cm_id(struct iw_cm_id *cm_id) +{ + struct iwcm_id_private *cm_id_priv; + unsigned long flags; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + /* + * Wait if we're currently in a connect or accept downcall. A + * listening endpoint should never block here. + */ + wait_event(cm_id_priv->connect_wait, + !test_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags)); + + spin_lock_irqsave(&cm_id_priv->lock, flags); + switch (cm_id_priv->state) { + case IW_CM_STATE_LISTEN: + cm_id_priv->state = IW_CM_STATE_DESTROYING; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + /* destroy the listening endpoint */ + cm_id->device->iwcm->destroy_listen(cm_id); + spin_lock_irqsave(&cm_id_priv->lock, flags); + break; + case IW_CM_STATE_ESTABLISHED: + cm_id_priv->state = IW_CM_STATE_DESTROYING; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + /* Abrupt close of the connection */ + (void)iwcm_modify_qp_err(cm_id_priv->qp); + spin_lock_irqsave(&cm_id_priv->lock, flags); + break; + case IW_CM_STATE_IDLE: + case IW_CM_STATE_CLOSING: + cm_id_priv->state = IW_CM_STATE_DESTROYING; + break; + case IW_CM_STATE_CONN_RECV: + /* + * App called destroy before/without calling accept after + * receiving connection request event notification or + * returned non zero from the event callback function. + * In either case, must tell the provider to reject. + */ + cm_id_priv->state = IW_CM_STATE_DESTROYING; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + cm_id->device->iwcm->reject(cm_id, NULL, 0); + spin_lock_irqsave(&cm_id_priv->lock, flags); + break; + case IW_CM_STATE_CONN_SENT: + case IW_CM_STATE_DESTROYING: + default: + BUG(); + break; + } + if (cm_id_priv->qp) { + cm_id_priv->id.device->iwcm->rem_ref(cm_id_priv->qp); + cm_id_priv->qp = NULL; + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + (void)iwcm_deref_id(cm_id_priv); +} + +/* + * This function is only called by the application thread and cannot + * be called by the event thread. The function will wait for all + * references to be released on the cm_id and then kfree the cm_id + * object. + */ +void iw_destroy_cm_id(struct iw_cm_id *cm_id) +{ + struct iwcm_id_private *cm_id_priv; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + BUG_ON(test_bit(IWCM_F_CALLBACK_DESTROY, &cm_id_priv->flags)); + + destroy_cm_id(cm_id); + + wait_for_completion(&cm_id_priv->destroy_comp); + + free_cm_id(cm_id_priv); +} +EXPORT_SYMBOL(iw_destroy_cm_id); + +/* + * CM_ID <-- LISTEN + * + * Start listening for connect requests. Generates one CONNECT_REQUEST + * event for each inbound connect request. + */ +int iw_cm_listen(struct iw_cm_id *cm_id, int backlog) +{ + struct iwcm_id_private *cm_id_priv; + unsigned long flags; + int ret; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + + ret = alloc_work_entries(cm_id_priv, backlog); + if (ret) + return ret; + + spin_lock_irqsave(&cm_id_priv->lock, flags); + switch (cm_id_priv->state) { + case IW_CM_STATE_IDLE: + cm_id_priv->state = IW_CM_STATE_LISTEN; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + ret = cm_id->device->iwcm->create_listen(cm_id, backlog); + if (ret) + cm_id_priv->state = IW_CM_STATE_IDLE; + spin_lock_irqsave(&cm_id_priv->lock, flags); + break; + default: + ret = -EINVAL; + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + return ret; +} +EXPORT_SYMBOL(iw_cm_listen); + +/* + * CM_ID <-- IDLE + * + * Rejects an inbound connection request. No events are generated. + */ +int iw_cm_reject(struct iw_cm_id *cm_id, + const void *private_data, + u8 private_data_len) +{ + struct iwcm_id_private *cm_id_priv; + unsigned long flags; + int ret; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + set_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + + spin_lock_irqsave(&cm_id_priv->lock, flags); + if (cm_id_priv->state != IW_CM_STATE_CONN_RECV) { + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + return -EINVAL; + } + cm_id_priv->state = IW_CM_STATE_IDLE; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + ret = cm_id->device->iwcm->reject(cm_id, private_data, + private_data_len); + + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + + return ret; +} +EXPORT_SYMBOL(iw_cm_reject); + +/* + * CM_ID <-- ESTABLISHED + * + * Accepts an inbound connection request and generates an ESTABLISHED + * event. Callers of iw_cm_disconnect and iw_destroy_cm_id will block + * until the ESTABLISHED event is received from the provider. + */ +int iw_cm_accept(struct iw_cm_id *cm_id, + struct iw_cm_conn_param *iw_param) +{ + struct iwcm_id_private *cm_id_priv; + struct ib_qp *qp; + unsigned long flags; + int ret; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + set_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + + spin_lock_irqsave(&cm_id_priv->lock, flags); + if (cm_id_priv->state != IW_CM_STATE_CONN_RECV) { + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + return -EINVAL; + } + /* Get the ib_qp given the QPN */ + qp = cm_id->device->iwcm->get_qp(cm_id->device, iw_param->qpn); + if (!qp) { + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + return -EINVAL; + } + cm_id->device->iwcm->add_ref(qp); + cm_id_priv->qp = qp; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + ret = cm_id->device->iwcm->accept(cm_id, iw_param); + if (ret) { + /* An error on accept precludes provider events */ + BUG_ON(cm_id_priv->state != IW_CM_STATE_CONN_RECV); + cm_id_priv->state = IW_CM_STATE_IDLE; + spin_lock_irqsave(&cm_id_priv->lock, flags); + if (cm_id_priv->qp) { + cm_id->device->iwcm->rem_ref(qp); + cm_id_priv->qp = NULL; + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + } + + return ret; +} +EXPORT_SYMBOL(iw_cm_accept); + +/* + * Active Side: CM_ID <-- CONN_SENT + * + * If successful, results in the generation of a CONNECT_REPLY + * event. iw_cm_disconnect and iw_cm_destroy will block until the + * CONNECT_REPLY event is received from the provider. + */ +int iw_cm_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *iw_param) +{ + struct iwcm_id_private *cm_id_priv; + int ret; + unsigned long flags; + struct ib_qp *qp; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + + ret = alloc_work_entries(cm_id_priv, 4); + if (ret) + return ret; + + set_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + spin_lock_irqsave(&cm_id_priv->lock, flags); + + if (cm_id_priv->state != IW_CM_STATE_IDLE) { + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + return -EINVAL; + } + + /* Get the ib_qp given the QPN */ + qp = cm_id->device->iwcm->get_qp(cm_id->device, iw_param->qpn); + if (!qp) { + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + return -EINVAL; + } + cm_id->device->iwcm->add_ref(qp); + cm_id_priv->qp = qp; + cm_id_priv->state = IW_CM_STATE_CONN_SENT; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + ret = cm_id->device->iwcm->connect(cm_id, iw_param); + if (ret) { + spin_lock_irqsave(&cm_id_priv->lock, flags); + if (cm_id_priv->qp) { + cm_id->device->iwcm->rem_ref(qp); + cm_id_priv->qp = NULL; + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + BUG_ON(cm_id_priv->state != IW_CM_STATE_CONN_SENT); + cm_id_priv->state = IW_CM_STATE_IDLE; + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + wake_up_all(&cm_id_priv->connect_wait); + } + + return ret; +} +EXPORT_SYMBOL(iw_cm_connect); + +/* + * Passive Side: new CM_ID <-- CONN_RECV + * + * Handles an inbound connect request. The function creates a new + * iw_cm_id to represent the new connection and inherits the client + * callback function and other attributes from the listening parent. + * + * The work item contains a pointer to the listen_cm_id and the event. The + * listen_cm_id contains the client cm_handler, context and + * device. These are copied when the device is cloned. The event + * contains the new four tuple. + * + * An error on the child should not affect the parent, so this + * function does not return a value. + */ +static void cm_conn_req_handler(struct iwcm_id_private *listen_id_priv, + struct iw_cm_event *iw_event) +{ + unsigned long flags; + struct iw_cm_id *cm_id; + struct iwcm_id_private *cm_id_priv; + int ret; + + /* + * The provider should never generate a connection request + * event with a bad status. + */ + BUG_ON(iw_event->status); + + cm_id = iw_create_cm_id(listen_id_priv->id.device, + listen_id_priv->id.cm_handler, + listen_id_priv->id.context); + /* If the cm_id could not be created, ignore the request */ + if (IS_ERR(cm_id)) + goto out; + + cm_id->provider_data = iw_event->provider_data; + cm_id->local_addr = iw_event->local_addr; + cm_id->remote_addr = iw_event->remote_addr; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + cm_id_priv->state = IW_CM_STATE_CONN_RECV; + + /* + * We could be destroying the listening id. If so, ignore this + * upcall. + */ + spin_lock_irqsave(&listen_id_priv->lock, flags); + if (listen_id_priv->state != IW_CM_STATE_LISTEN) { + spin_unlock_irqrestore(&listen_id_priv->lock, flags); + iw_cm_reject(cm_id, NULL, 0); + iw_destroy_cm_id(cm_id); + goto out; + } + spin_unlock_irqrestore(&listen_id_priv->lock, flags); + + ret = alloc_work_entries(cm_id_priv, 3); + if (ret) { + iw_cm_reject(cm_id, NULL, 0); + iw_destroy_cm_id(cm_id); + goto out; + } + + /* Call the client CM handler */ + ret = cm_id->cm_handler(cm_id, iw_event); + if (ret) { + iw_cm_reject(cm_id, NULL, 0); + set_bit(IWCM_F_CALLBACK_DESTROY, &cm_id_priv->flags); + destroy_cm_id(cm_id); + if (atomic_read(&cm_id_priv->refcount)==0) + free_cm_id(cm_id_priv); + } + +out: + if (iw_event->private_data_len) + kfree(iw_event->private_data); +} + +/* + * Passive Side: CM_ID <-- ESTABLISHED + * + * The provider generated an ESTABLISHED event which means that + * the MPA negotion has completed successfully and we are now in MPA + * FPDU mode. + * + * This event can only be received in the CONN_RECV state. If the + * remote peer closed, the ESTABLISHED event would be received followed + * by the CLOSE event. If the app closes, it will block until we wake + * it up after processing this event. + */ +static int cm_conn_est_handler(struct iwcm_id_private *cm_id_priv, + struct iw_cm_event *iw_event) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&cm_id_priv->lock, flags); + + /* + * We clear the CONNECT_WAIT bit here to allow the callback + * function to call iw_cm_disconnect. Calling iw_destroy_cm_id + * from a callback handler is not allowed. + */ + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + BUG_ON(cm_id_priv->state != IW_CM_STATE_CONN_RECV); + cm_id_priv->state = IW_CM_STATE_ESTABLISHED; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + ret = cm_id_priv->id.cm_handler(&cm_id_priv->id, iw_event); + wake_up_all(&cm_id_priv->connect_wait); + + return ret; +} + +/* + * Active Side: CM_ID <-- ESTABLISHED + * + * The app has called connect and is waiting for the established event to + * post it's requests to the server. This event will wake up anyone + * blocked in iw_cm_disconnect or iw_destroy_id. + */ +static int cm_conn_rep_handler(struct iwcm_id_private *cm_id_priv, + struct iw_cm_event *iw_event) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&cm_id_priv->lock, flags); + /* + * Clear the connect wait bit so a callback function calling + * iw_cm_disconnect will not wait and deadlock this thread + */ + clear_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags); + BUG_ON(cm_id_priv->state != IW_CM_STATE_CONN_SENT); + if (iw_event->status == 0) { + cm_id_priv->id.local_addr = iw_event->local_addr; + cm_id_priv->id.remote_addr = iw_event->remote_addr; + cm_id_priv->state = IW_CM_STATE_ESTABLISHED; + } else { + /* REJECTED or RESET */ + cm_id_priv->id.device->iwcm->rem_ref(cm_id_priv->qp); + cm_id_priv->qp = NULL; + cm_id_priv->state = IW_CM_STATE_IDLE; + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + ret = cm_id_priv->id.cm_handler(&cm_id_priv->id, iw_event); + + if (iw_event->private_data_len) + kfree(iw_event->private_data); + + /* Wake up waiters on connect complete */ + wake_up_all(&cm_id_priv->connect_wait); + + return ret; +} + +/* + * CM_ID <-- CLOSING + * + * If in the ESTABLISHED state, move to CLOSING. + */ +static void cm_disconnect_handler(struct iwcm_id_private *cm_id_priv, + struct iw_cm_event *iw_event) +{ + unsigned long flags; + + spin_lock_irqsave(&cm_id_priv->lock, flags); + if (cm_id_priv->state == IW_CM_STATE_ESTABLISHED) + cm_id_priv->state = IW_CM_STATE_CLOSING; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); +} + +/* + * CM_ID <-- IDLE + * + * If in the ESTBLISHED or CLOSING states, the QP will have have been + * moved by the provider to the ERR state. Disassociate the CM_ID from + * the QP, move to IDLE, and remove the 'connected' reference. + * + * If in some other state, the cm_id was destroyed asynchronously. + * This is the last reference that will result in waking up + * the app thread blocked in iw_destroy_cm_id. + */ +static int cm_close_handler(struct iwcm_id_private *cm_id_priv, + struct iw_cm_event *iw_event) +{ + unsigned long flags; + int ret = 0; + spin_lock_irqsave(&cm_id_priv->lock, flags); + + if (cm_id_priv->qp) { + cm_id_priv->id.device->iwcm->rem_ref(cm_id_priv->qp); + cm_id_priv->qp = NULL; + } + switch (cm_id_priv->state) { + case IW_CM_STATE_ESTABLISHED: + case IW_CM_STATE_CLOSING: + cm_id_priv->state = IW_CM_STATE_IDLE; + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + ret = cm_id_priv->id.cm_handler(&cm_id_priv->id, iw_event); + spin_lock_irqsave(&cm_id_priv->lock, flags); + break; + case IW_CM_STATE_DESTROYING: + break; + default: + BUG(); + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + return ret; +} + +static int process_event(struct iwcm_id_private *cm_id_priv, + struct iw_cm_event *iw_event) +{ + int ret = 0; + + switch (iw_event->event) { + case IW_CM_EVENT_CONNECT_REQUEST: + cm_conn_req_handler(cm_id_priv, iw_event); + break; + case IW_CM_EVENT_CONNECT_REPLY: + ret = cm_conn_rep_handler(cm_id_priv, iw_event); + break; + case IW_CM_EVENT_ESTABLISHED: + ret = cm_conn_est_handler(cm_id_priv, iw_event); + break; + case IW_CM_EVENT_DISCONNECT: + cm_disconnect_handler(cm_id_priv, iw_event); + break; + case IW_CM_EVENT_CLOSE: + ret = cm_close_handler(cm_id_priv, iw_event); + break; + default: + BUG(); + } + + return ret; +} + +/* + * Process events on the work_list for the cm_id. If the callback + * function requests that the cm_id be deleted, a flag is set in the + * cm_id flags to indicate that when the last reference is + * removed, the cm_id is to be destroyed. This is necessary to + * distinguish between an object that will be destroyed by the app + * thread asleep on the destroy_comp list vs. an object destroyed + * here synchronously when the last reference is removed. + */ +static void cm_work_handler(struct work_struct *_work) +{ + struct iwcm_work *work = container_of(_work, struct iwcm_work, work); + struct iw_cm_event levent; + struct iwcm_id_private *cm_id_priv = work->cm_id; + unsigned long flags; + int empty; + int ret = 0; + int destroy_id; + + spin_lock_irqsave(&cm_id_priv->lock, flags); + empty = list_empty(&cm_id_priv->work_list); + while (!empty) { + work = list_entry(cm_id_priv->work_list.next, + struct iwcm_work, list); + list_del_init(&work->list); + empty = list_empty(&cm_id_priv->work_list); + levent = work->event; + put_work(work); + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + + ret = process_event(cm_id_priv, &levent); + if (ret) { + set_bit(IWCM_F_CALLBACK_DESTROY, &cm_id_priv->flags); + destroy_cm_id(&cm_id_priv->id); + } + BUG_ON(atomic_read(&cm_id_priv->refcount)==0); + destroy_id = test_bit(IWCM_F_CALLBACK_DESTROY, &cm_id_priv->flags); + if (iwcm_deref_id(cm_id_priv)) { + if (destroy_id) { + BUG_ON(!list_empty(&cm_id_priv->work_list)); + free_cm_id(cm_id_priv); + } + return; + } + if (empty) + return; + spin_lock_irqsave(&cm_id_priv->lock, flags); + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); +} + +/* + * This function is called on interrupt context. Schedule events on + * the iwcm_wq thread to allow callback functions to downcall into + * the CM and/or block. Events are queued to a per-CM_ID + * work_list. If this is the first event on the work_list, the work + * element is also queued on the iwcm_wq thread. + * + * Each event holds a reference on the cm_id. Until the last posted + * event has been delivered and processed, the cm_id cannot be + * deleted. + * + * Returns: + * 0 - the event was handled. + * -ENOMEM - the event was not handled due to lack of resources. + */ +static int cm_event_handler(struct iw_cm_id *cm_id, + struct iw_cm_event *iw_event) +{ + struct iwcm_work *work; + struct iwcm_id_private *cm_id_priv; + unsigned long flags; + int ret = 0; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + + spin_lock_irqsave(&cm_id_priv->lock, flags); + work = get_work(cm_id_priv); + if (!work) { + ret = -ENOMEM; + goto out; + } + + INIT_WORK(&work->work, cm_work_handler); + work->cm_id = cm_id_priv; + work->event = *iw_event; + + if ((work->event.event == IW_CM_EVENT_CONNECT_REQUEST || + work->event.event == IW_CM_EVENT_CONNECT_REPLY) && + work->event.private_data_len) { + ret = copy_private_data(&work->event); + if (ret) { + put_work(work); + goto out; + } + } + + atomic_inc(&cm_id_priv->refcount); + if (list_empty(&cm_id_priv->work_list)) { + list_add_tail(&work->list, &cm_id_priv->work_list); + queue_work(iwcm_wq, &work->work); + } else + list_add_tail(&work->list, &cm_id_priv->work_list); +out: + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + return ret; +} + +static int iwcm_init_qp_init_attr(struct iwcm_id_private *cm_id_priv, + struct ib_qp_attr *qp_attr, + int *qp_attr_mask) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&cm_id_priv->lock, flags); + switch (cm_id_priv->state) { + case IW_CM_STATE_IDLE: + case IW_CM_STATE_CONN_SENT: + case IW_CM_STATE_CONN_RECV: + case IW_CM_STATE_ESTABLISHED: + *qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS; + qp_attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE| + IB_ACCESS_REMOTE_READ; + ret = 0; + break; + default: + ret = -EINVAL; + break; + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + return ret; +} + +static int iwcm_init_qp_rts_attr(struct iwcm_id_private *cm_id_priv, + struct ib_qp_attr *qp_attr, + int *qp_attr_mask) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&cm_id_priv->lock, flags); + switch (cm_id_priv->state) { + case IW_CM_STATE_IDLE: + case IW_CM_STATE_CONN_SENT: + case IW_CM_STATE_CONN_RECV: + case IW_CM_STATE_ESTABLISHED: + *qp_attr_mask = 0; + ret = 0; + break; + default: + ret = -EINVAL; + break; + } + spin_unlock_irqrestore(&cm_id_priv->lock, flags); + return ret; +} + +int iw_cm_init_qp_attr(struct iw_cm_id *cm_id, + struct ib_qp_attr *qp_attr, + int *qp_attr_mask) +{ + struct iwcm_id_private *cm_id_priv; + int ret; + + cm_id_priv = container_of(cm_id, struct iwcm_id_private, id); + switch (qp_attr->qp_state) { + case IB_QPS_INIT: + case IB_QPS_RTR: + ret = iwcm_init_qp_init_attr(cm_id_priv, + qp_attr, qp_attr_mask); + break; + case IB_QPS_RTS: + ret = iwcm_init_qp_rts_attr(cm_id_priv, + qp_attr, qp_attr_mask); + break; + default: + ret = -EINVAL; + break; + } + return ret; +} +EXPORT_SYMBOL(iw_cm_init_qp_attr); + +static int __init iw_cm_init(void) +{ + iwcm_wq = create_singlethread_workqueue("iw_cm_wq"); + if (!iwcm_wq) + return -ENOMEM; + + return 0; +} + +static void __exit iw_cm_cleanup(void) +{ + destroy_workqueue(iwcm_wq); +} + +module_init(iw_cm_init); +module_exit(iw_cm_cleanup); diff --git a/drivers/infiniband/core/iwcm.h b/drivers/infiniband/core/iwcm.h new file mode 100644 index 00000000000..3f6cc82564c --- /dev/null +++ b/drivers/infiniband/core/iwcm.h @@ -0,0 +1,62 @@ +/* + * Copyright (c) 2005 Network Appliance, Inc. All rights reserved. + * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef IWCM_H +#define IWCM_H + +enum iw_cm_state { + IW_CM_STATE_IDLE, /* unbound, inactive */ + IW_CM_STATE_LISTEN, /* listen waiting for connect */ + IW_CM_STATE_CONN_RECV, /* inbound waiting for user accept */ + IW_CM_STATE_CONN_SENT, /* outbound waiting for peer accept */ + IW_CM_STATE_ESTABLISHED, /* established */ + IW_CM_STATE_CLOSING, /* disconnect */ + IW_CM_STATE_DESTROYING /* object being deleted */ +}; + +struct iwcm_id_private { + struct iw_cm_id id; + enum iw_cm_state state; + unsigned long flags; + struct ib_qp *qp; + struct completion destroy_comp; + wait_queue_head_t connect_wait; + struct list_head work_list; + spinlock_t lock; + atomic_t refcount; + struct list_head work_free_list; +}; + +#define IWCM_F_CALLBACK_DESTROY 1 +#define IWCM_F_CONNECT_WAIT 2 + +#endif /* IWCM_H */ diff --git a/drivers/infiniband/core/iwpm_msg.c b/drivers/infiniband/core/iwpm_msg.c new file mode 100644 index 00000000000..b85ddbc979e --- /dev/null +++ b/drivers/infiniband/core/iwpm_msg.c @@ -0,0 +1,685 @@ +/* + * Copyright (c) 2014 Intel Corporation. All rights reserved. + * Copyright (c) 2014 Chelsio, Inc. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include "iwpm_util.h" + +static const char iwpm_ulib_name[] = "iWarpPortMapperUser"; +static int iwpm_ulib_version = 3; +static int iwpm_user_pid = IWPM_PID_UNDEFINED; +static atomic_t echo_nlmsg_seq; + +int iwpm_valid_pid(void) +{ + return iwpm_user_pid > 0; +} +EXPORT_SYMBOL(iwpm_valid_pid); + +/* + * iwpm_register_pid - Send a netlink query to user space + * for the iwarp port mapper pid + * + * nlmsg attributes: + * [IWPM_NLA_REG_PID_SEQ] + * [IWPM_NLA_REG_IF_NAME] + * [IWPM_NLA_REG_IBDEV_NAME] + * [IWPM_NLA_REG_ULIB_NAME] + */ +int iwpm_register_pid(struct iwpm_dev_data *pm_msg, u8 nl_client) +{ + struct sk_buff *skb = NULL; + struct iwpm_nlmsg_request *nlmsg_request = NULL; + struct nlmsghdr *nlh; + u32 msg_seq; + const char *err_str = ""; + int ret = -EINVAL; + + if (!iwpm_valid_client(nl_client)) { + err_str = "Invalid port mapper client"; + goto pid_query_error; + } + if (iwpm_registered_client(nl_client)) + return 0; + skb = iwpm_create_nlmsg(RDMA_NL_IWPM_REG_PID, &nlh, nl_client); + if (!skb) { + err_str = "Unable to create a nlmsg"; + goto pid_query_error; + } + nlh->nlmsg_seq = iwpm_get_nlmsg_seq(); + nlmsg_request = iwpm_get_nlmsg_request(nlh->nlmsg_seq, nl_client, GFP_KERNEL); + if (!nlmsg_request) { + err_str = "Unable to allocate netlink request"; + goto pid_query_error; + } + msg_seq = atomic_read(&echo_nlmsg_seq); + + /* fill in the pid request message */ + err_str = "Unable to put attribute of the nlmsg"; + ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq, IWPM_NLA_REG_PID_SEQ); + if (ret) + goto pid_query_error; + ret = ibnl_put_attr(skb, nlh, IWPM_IFNAME_SIZE, + pm_msg->if_name, IWPM_NLA_REG_IF_NAME); + if (ret) + goto pid_query_error; + ret = ibnl_put_attr(skb, nlh, IWPM_DEVNAME_SIZE, + pm_msg->dev_name, IWPM_NLA_REG_IBDEV_NAME); + if (ret) + goto pid_query_error; + ret = ibnl_put_attr(skb, nlh, IWPM_ULIBNAME_SIZE, + (char *)iwpm_ulib_name, IWPM_NLA_REG_ULIB_NAME); + if (ret) + goto pid_query_error; + + pr_debug("%s: Multicasting a nlmsg (dev = %s ifname = %s iwpm = %s)\n", + __func__, pm_msg->dev_name, pm_msg->if_name, iwpm_ulib_name); + + ret = ibnl_multicast(skb, nlh, RDMA_NL_GROUP_IWPM, GFP_KERNEL); + if (ret) { + skb = NULL; /* skb is freed in the netlink send-op handling */ + iwpm_set_registered(nl_client, 1); + iwpm_user_pid = IWPM_PID_UNAVAILABLE; + err_str = "Unable to send a nlmsg"; + goto pid_query_error; + } + nlmsg_request->req_buffer = pm_msg; + ret = iwpm_wait_complete_req(nlmsg_request); + return ret; +pid_query_error: + pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client); + if (skb) + dev_kfree_skb(skb); + if (nlmsg_request) + iwpm_free_nlmsg_request(&nlmsg_request->kref); + return ret; +} +EXPORT_SYMBOL(iwpm_register_pid); + +/* + * iwpm_add_mapping - Send a netlink add mapping message + * to the port mapper + * nlmsg attributes: + * [IWPM_NLA_MANAGE_MAPPING_SEQ] + * [IWPM_NLA_MANAGE_ADDR] + */ +int iwpm_add_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client) +{ + struct sk_buff *skb = NULL; + struct iwpm_nlmsg_request *nlmsg_request = NULL; + struct nlmsghdr *nlh; + u32 msg_seq; + const char *err_str = ""; + int ret = -EINVAL; + + if (!iwpm_valid_client(nl_client)) { + err_str = "Invalid port mapper client"; + goto add_mapping_error; + } + if (!iwpm_registered_client(nl_client)) { + err_str = "Unregistered port mapper client"; + goto add_mapping_error; + } + if (!iwpm_valid_pid()) + return 0; + skb = iwpm_create_nlmsg(RDMA_NL_IWPM_ADD_MAPPING, &nlh, nl_client); + if (!skb) { + err_str = "Unable to create a nlmsg"; + goto add_mapping_error; + } + nlh->nlmsg_seq = iwpm_get_nlmsg_seq(); + nlmsg_request = iwpm_get_nlmsg_request(nlh->nlmsg_seq, nl_client, GFP_KERNEL); + if (!nlmsg_request) { + err_str = "Unable to allocate netlink request"; + goto add_mapping_error; + } + msg_seq = atomic_read(&echo_nlmsg_seq); + /* fill in the add mapping message */ + err_str = "Unable to put attribute of the nlmsg"; + ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq, + IWPM_NLA_MANAGE_MAPPING_SEQ); + if (ret) + goto add_mapping_error; + ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage), + &pm_msg->loc_addr, IWPM_NLA_MANAGE_ADDR); + if (ret) + goto add_mapping_error; + nlmsg_request->req_buffer = pm_msg; + + ret = ibnl_unicast(skb, nlh, iwpm_user_pid); + if (ret) { + skb = NULL; /* skb is freed in the netlink send-op handling */ + iwpm_user_pid = IWPM_PID_UNDEFINED; + err_str = "Unable to send a nlmsg"; + goto add_mapping_error; + } + ret = iwpm_wait_complete_req(nlmsg_request); + return ret; +add_mapping_error: + pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client); + if (skb) + dev_kfree_skb(skb); + if (nlmsg_request) + iwpm_free_nlmsg_request(&nlmsg_request->kref); + return ret; +} +EXPORT_SYMBOL(iwpm_add_mapping); + +/* + * iwpm_add_and_query_mapping - Send a netlink add and query + * mapping message to the port mapper + * nlmsg attributes: + * [IWPM_NLA_QUERY_MAPPING_SEQ] + * [IWPM_NLA_QUERY_LOCAL_ADDR] + * [IWPM_NLA_QUERY_REMOTE_ADDR] + */ +int iwpm_add_and_query_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client) +{ + struct sk_buff *skb = NULL; + struct iwpm_nlmsg_request *nlmsg_request = NULL; + struct nlmsghdr *nlh; + u32 msg_seq; + const char *err_str = ""; + int ret = -EINVAL; + + if (!iwpm_valid_client(nl_client)) { + err_str = "Invalid port mapper client"; + goto query_mapping_error; + } + if (!iwpm_registered_client(nl_client)) { + err_str = "Unregistered port mapper client"; + goto query_mapping_error; + } + if (!iwpm_valid_pid()) + return 0; + ret = -ENOMEM; + skb = iwpm_create_nlmsg(RDMA_NL_IWPM_QUERY_MAPPING, &nlh, nl_client); + if (!skb) { + err_str = "Unable to create a nlmsg"; + goto query_mapping_error; + } + nlh->nlmsg_seq = iwpm_get_nlmsg_seq(); + nlmsg_request = iwpm_get_nlmsg_request(nlh->nlmsg_seq, + nl_client, GFP_KERNEL); + if (!nlmsg_request) { + err_str = "Unable to allocate netlink request"; + goto query_mapping_error; + } + msg_seq = atomic_read(&echo_nlmsg_seq); + + /* fill in the query message */ + err_str = "Unable to put attribute of the nlmsg"; + ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq, + IWPM_NLA_QUERY_MAPPING_SEQ); + if (ret) + goto query_mapping_error; + ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage), + &pm_msg->loc_addr, IWPM_NLA_QUERY_LOCAL_ADDR); + if (ret) + goto query_mapping_error; + ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage), + &pm_msg->rem_addr, IWPM_NLA_QUERY_REMOTE_ADDR); + if (ret) + goto query_mapping_error; + nlmsg_request->req_buffer = pm_msg; + + ret = ibnl_unicast(skb, nlh, iwpm_user_pid); + if (ret) { + skb = NULL; /* skb is freed in the netlink send-op handling */ + err_str = "Unable to send a nlmsg"; + goto query_mapping_error; + } + ret = iwpm_wait_complete_req(nlmsg_request); + return ret; +query_mapping_error: + pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client); + if (skb) + dev_kfree_skb(skb); + if (nlmsg_request) + iwpm_free_nlmsg_request(&nlmsg_request->kref); + return ret; +} +EXPORT_SYMBOL(iwpm_add_and_query_mapping); + +/* + * iwpm_remove_mapping - Send a netlink remove mapping message + * to the port mapper + * nlmsg attributes: + * [IWPM_NLA_MANAGE_MAPPING_SEQ] + * [IWPM_NLA_MANAGE_ADDR] + */ +int iwpm_remove_mapping(struct sockaddr_storage *local_addr, u8 nl_client) +{ + struct sk_buff *skb = NULL; + struct nlmsghdr *nlh; + u32 msg_seq; + const char *err_str = ""; + int ret = -EINVAL; + + if (!iwpm_valid_client(nl_client)) { + err_str = "Invalid port mapper client"; + goto remove_mapping_error; + } + if (!iwpm_registered_client(nl_client)) { + err_str = "Unregistered port mapper client"; + goto remove_mapping_error; + } + if (!iwpm_valid_pid()) + return 0; + skb = iwpm_create_nlmsg(RDMA_NL_IWPM_REMOVE_MAPPING, &nlh, nl_client); + if (!skb) { + ret = -ENOMEM; + err_str = "Unable to create a nlmsg"; + goto remove_mapping_error; + } + msg_seq = atomic_read(&echo_nlmsg_seq); + nlh->nlmsg_seq = iwpm_get_nlmsg_seq(); + err_str = "Unable to put attribute of the nlmsg"; + ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq, + IWPM_NLA_MANAGE_MAPPING_SEQ); + if (ret) + goto remove_mapping_error; + ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage), + local_addr, IWPM_NLA_MANAGE_ADDR); + if (ret) + goto remove_mapping_error; + + ret = ibnl_unicast(skb, nlh, iwpm_user_pid); + if (ret) { + skb = NULL; /* skb is freed in the netlink send-op handling */ + iwpm_user_pid = IWPM_PID_UNDEFINED; + err_str = "Unable to send a nlmsg"; + goto remove_mapping_error; + } + iwpm_print_sockaddr(local_addr, + "remove_mapping: Local sockaddr:"); + return 0; +remove_mapping_error: + pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client); + if (skb) + dev_kfree_skb_any(skb); + return ret; +} +EXPORT_SYMBOL(iwpm_remove_mapping); + +/* netlink attribute policy for the received response to register pid request */ +static const struct nla_policy resp_reg_policy[IWPM_NLA_RREG_PID_MAX] = { + [IWPM_NLA_RREG_PID_SEQ] = { .type = NLA_U32 }, + [IWPM_NLA_RREG_IBDEV_NAME] = { .type = NLA_STRING, + .len = IWPM_DEVNAME_SIZE - 1 }, + [IWPM_NLA_RREG_ULIB_NAME] = { .type = NLA_STRING, + .len = IWPM_ULIBNAME_SIZE - 1 }, + [IWPM_NLA_RREG_ULIB_VER] = { .type = NLA_U16 }, + [IWPM_NLA_RREG_PID_ERR] = { .type = NLA_U16 } +}; + +/* + * iwpm_register_pid_cb - Process a port mapper response to + * iwpm_register_pid() + */ +int iwpm_register_pid_cb(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct iwpm_nlmsg_request *nlmsg_request = NULL; + struct nlattr *nltb[IWPM_NLA_RREG_PID_MAX]; + struct iwpm_dev_data *pm_msg; + char *dev_name, *iwpm_name; + u32 msg_seq; + u8 nl_client; + u16 iwpm_version; + const char *msg_type = "Register Pid response"; + + if (iwpm_parse_nlmsg(cb, IWPM_NLA_RREG_PID_MAX, + resp_reg_policy, nltb, msg_type)) + return -EINVAL; + + msg_seq = nla_get_u32(nltb[IWPM_NLA_RREG_PID_SEQ]); + nlmsg_request = iwpm_find_nlmsg_request(msg_seq); + if (!nlmsg_request) { + pr_info("%s: Could not find a matching request (seq = %u)\n", + __func__, msg_seq); + return -EINVAL; + } + pm_msg = nlmsg_request->req_buffer; + nl_client = nlmsg_request->nl_client; + dev_name = (char *)nla_data(nltb[IWPM_NLA_RREG_IBDEV_NAME]); + iwpm_name = (char *)nla_data(nltb[IWPM_NLA_RREG_ULIB_NAME]); + iwpm_version = nla_get_u16(nltb[IWPM_NLA_RREG_ULIB_VER]); + + /* check device name, ulib name and version */ + if (strcmp(pm_msg->dev_name, dev_name) || + strcmp(iwpm_ulib_name, iwpm_name) || + iwpm_version != iwpm_ulib_version) { + + pr_info("%s: Incorrect info (dev = %s name = %s version = %d)\n", + __func__, dev_name, iwpm_name, iwpm_version); + nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR; + goto register_pid_response_exit; + } + iwpm_user_pid = cb->nlh->nlmsg_pid; + atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); + pr_debug("%s: iWarp Port Mapper (pid = %d) is available!\n", + __func__, iwpm_user_pid); + if (iwpm_valid_client(nl_client)) + iwpm_set_registered(nl_client, 1); +register_pid_response_exit: + nlmsg_request->request_done = 1; + /* always for found nlmsg_request */ + kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request); + barrier(); + wake_up(&nlmsg_request->waitq); + return 0; +} +EXPORT_SYMBOL(iwpm_register_pid_cb); + +/* netlink attribute policy for the received response to add mapping request */ +static const struct nla_policy resp_add_policy[IWPM_NLA_RMANAGE_MAPPING_MAX] = { + [IWPM_NLA_MANAGE_MAPPING_SEQ] = { .type = NLA_U32 }, + [IWPM_NLA_MANAGE_ADDR] = { .len = sizeof(struct sockaddr_storage) }, + [IWPM_NLA_MANAGE_MAPPED_LOC_ADDR] = { .len = sizeof(struct sockaddr_storage) }, + [IWPM_NLA_RMANAGE_MAPPING_ERR] = { .type = NLA_U16 } +}; + +/* + * iwpm_add_mapping_cb - Process a port mapper response to + * iwpm_add_mapping() + */ +int iwpm_add_mapping_cb(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct iwpm_sa_data *pm_msg; + struct iwpm_nlmsg_request *nlmsg_request = NULL; + struct nlattr *nltb[IWPM_NLA_RMANAGE_MAPPING_MAX]; + struct sockaddr_storage *local_sockaddr; + struct sockaddr_storage *mapped_sockaddr; + const char *msg_type; + u32 msg_seq; + + msg_type = "Add Mapping response"; + if (iwpm_parse_nlmsg(cb, IWPM_NLA_RMANAGE_MAPPING_MAX, + resp_add_policy, nltb, msg_type)) + return -EINVAL; + + atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); + + msg_seq = nla_get_u32(nltb[IWPM_NLA_MANAGE_MAPPING_SEQ]); + nlmsg_request = iwpm_find_nlmsg_request(msg_seq); + if (!nlmsg_request) { + pr_info("%s: Could not find a matching request (seq = %u)\n", + __func__, msg_seq); + return -EINVAL; + } + pm_msg = nlmsg_request->req_buffer; + local_sockaddr = (struct sockaddr_storage *) + nla_data(nltb[IWPM_NLA_MANAGE_ADDR]); + mapped_sockaddr = (struct sockaddr_storage *) + nla_data(nltb[IWPM_NLA_MANAGE_MAPPED_LOC_ADDR]); + + if (iwpm_compare_sockaddr(local_sockaddr, &pm_msg->loc_addr)) { + nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR; + goto add_mapping_response_exit; + } + if (mapped_sockaddr->ss_family != local_sockaddr->ss_family) { + pr_info("%s: Sockaddr family doesn't match the requested one\n", + __func__); + nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR; + goto add_mapping_response_exit; + } + memcpy(&pm_msg->mapped_loc_addr, mapped_sockaddr, + sizeof(*mapped_sockaddr)); + iwpm_print_sockaddr(&pm_msg->loc_addr, + "add_mapping: Local sockaddr:"); + iwpm_print_sockaddr(&pm_msg->mapped_loc_addr, + "add_mapping: Mapped local sockaddr:"); + +add_mapping_response_exit: + nlmsg_request->request_done = 1; + /* always for found request */ + kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request); + barrier(); + wake_up(&nlmsg_request->waitq); + return 0; +} +EXPORT_SYMBOL(iwpm_add_mapping_cb); + +/* netlink attribute policy for the response to add and query mapping request */ +static const struct nla_policy resp_query_policy[IWPM_NLA_RQUERY_MAPPING_MAX] = { + [IWPM_NLA_QUERY_MAPPING_SEQ] = { .type = NLA_U32 }, + [IWPM_NLA_QUERY_LOCAL_ADDR] = { .len = sizeof(struct sockaddr_storage) }, + [IWPM_NLA_QUERY_REMOTE_ADDR] = { .len = sizeof(struct sockaddr_storage) }, + [IWPM_NLA_RQUERY_MAPPED_LOC_ADDR] = { .len = sizeof(struct sockaddr_storage) }, + [IWPM_NLA_RQUERY_MAPPED_REM_ADDR] = { .len = sizeof(struct sockaddr_storage) }, + [IWPM_NLA_RQUERY_MAPPING_ERR] = { .type = NLA_U16 } +}; + +/* + * iwpm_add_and_query_mapping_cb - Process a port mapper response to + * iwpm_add_and_query_mapping() + */ +int iwpm_add_and_query_mapping_cb(struct sk_buff *skb, + struct netlink_callback *cb) +{ + struct iwpm_sa_data *pm_msg; + struct iwpm_nlmsg_request *nlmsg_request = NULL; + struct nlattr *nltb[IWPM_NLA_RQUERY_MAPPING_MAX]; + struct sockaddr_storage *local_sockaddr, *remote_sockaddr; + struct sockaddr_storage *mapped_loc_sockaddr, *mapped_rem_sockaddr; + const char *msg_type; + u32 msg_seq; + u16 err_code; + + msg_type = "Query Mapping response"; + if (iwpm_parse_nlmsg(cb, IWPM_NLA_RQUERY_MAPPING_MAX, + resp_query_policy, nltb, msg_type)) + return -EINVAL; + atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); + + msg_seq = nla_get_u32(nltb[IWPM_NLA_QUERY_MAPPING_SEQ]); + nlmsg_request = iwpm_find_nlmsg_request(msg_seq); + if (!nlmsg_request) { + pr_info("%s: Could not find a matching request (seq = %u)\n", + __func__, msg_seq); + return -EINVAL; + } + pm_msg = nlmsg_request->req_buffer; + local_sockaddr = (struct sockaddr_storage *) + nla_data(nltb[IWPM_NLA_QUERY_LOCAL_ADDR]); + remote_sockaddr = (struct sockaddr_storage *) + nla_data(nltb[IWPM_NLA_QUERY_REMOTE_ADDR]); + mapped_loc_sockaddr = (struct sockaddr_storage *) + nla_data(nltb[IWPM_NLA_RQUERY_MAPPED_LOC_ADDR]); + mapped_rem_sockaddr = (struct sockaddr_storage *) + nla_data(nltb[IWPM_NLA_RQUERY_MAPPED_REM_ADDR]); + + err_code = nla_get_u16(nltb[IWPM_NLA_RQUERY_MAPPING_ERR]); + if (err_code == IWPM_REMOTE_QUERY_REJECT) { + pr_info("%s: Received a Reject (pid = %u, echo seq = %u)\n", + __func__, cb->nlh->nlmsg_pid, msg_seq); + nlmsg_request->err_code = IWPM_REMOTE_QUERY_REJECT; + } + if (iwpm_compare_sockaddr(local_sockaddr, &pm_msg->loc_addr) || + iwpm_compare_sockaddr(remote_sockaddr, &pm_msg->rem_addr)) { + pr_info("%s: Incorrect local sockaddr\n", __func__); + nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR; + goto query_mapping_response_exit; + } + if (mapped_loc_sockaddr->ss_family != local_sockaddr->ss_family || + mapped_rem_sockaddr->ss_family != remote_sockaddr->ss_family) { + pr_info("%s: Sockaddr family doesn't match the requested one\n", + __func__); + nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR; + goto query_mapping_response_exit; + } + memcpy(&pm_msg->mapped_loc_addr, mapped_loc_sockaddr, + sizeof(*mapped_loc_sockaddr)); + memcpy(&pm_msg->mapped_rem_addr, mapped_rem_sockaddr, + sizeof(*mapped_rem_sockaddr)); + + iwpm_print_sockaddr(&pm_msg->loc_addr, + "query_mapping: Local sockaddr:"); + iwpm_print_sockaddr(&pm_msg->mapped_loc_addr, + "query_mapping: Mapped local sockaddr:"); + iwpm_print_sockaddr(&pm_msg->rem_addr, + "query_mapping: Remote sockaddr:"); + iwpm_print_sockaddr(&pm_msg->mapped_rem_addr, + "query_mapping: Mapped remote sockaddr:"); +query_mapping_response_exit: + nlmsg_request->request_done = 1; + /* always for found request */ + kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request); + barrier(); + wake_up(&nlmsg_request->waitq); + return 0; +} +EXPORT_SYMBOL(iwpm_add_and_query_mapping_cb); + +/* netlink attribute policy for the received request for mapping info */ +static const struct nla_policy resp_mapinfo_policy[IWPM_NLA_MAPINFO_REQ_MAX] = { + [IWPM_NLA_MAPINFO_ULIB_NAME] = { .type = NLA_STRING, + .len = IWPM_ULIBNAME_SIZE - 1 }, + [IWPM_NLA_MAPINFO_ULIB_VER] = { .type = NLA_U16 } +}; + +/* + * iwpm_mapping_info_cb - Process a port mapper request for mapping info + */ +int iwpm_mapping_info_cb(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct nlattr *nltb[IWPM_NLA_MAPINFO_REQ_MAX]; + const char *msg_type = "Mapping Info response"; + int iwpm_pid; + u8 nl_client; + char *iwpm_name; + u16 iwpm_version; + int ret = -EINVAL; + + if (iwpm_parse_nlmsg(cb, IWPM_NLA_MAPINFO_REQ_MAX, + resp_mapinfo_policy, nltb, msg_type)) { + pr_info("%s: Unable to parse nlmsg\n", __func__); + return ret; + } + iwpm_name = (char *)nla_data(nltb[IWPM_NLA_MAPINFO_ULIB_NAME]); + iwpm_version = nla_get_u16(nltb[IWPM_NLA_MAPINFO_ULIB_VER]); + if (strcmp(iwpm_ulib_name, iwpm_name) || + iwpm_version != iwpm_ulib_version) { + pr_info("%s: Invalid port mapper name = %s version = %d\n", + __func__, iwpm_name, iwpm_version); + return ret; + } + nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type); + if (!iwpm_valid_client(nl_client)) { + pr_info("%s: Invalid port mapper client = %d\n", + __func__, nl_client); + return ret; + } + iwpm_set_registered(nl_client, 0); + atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); + if (!iwpm_mapinfo_available()) + return 0; + iwpm_pid = cb->nlh->nlmsg_pid; + pr_debug("%s: iWarp Port Mapper (pid = %d) is available!\n", + __func__, iwpm_pid); + ret = iwpm_send_mapinfo(nl_client, iwpm_pid); + return ret; +} +EXPORT_SYMBOL(iwpm_mapping_info_cb); + +/* netlink attribute policy for the received mapping info ack */ +static const struct nla_policy ack_mapinfo_policy[IWPM_NLA_MAPINFO_NUM_MAX] = { + [IWPM_NLA_MAPINFO_SEQ] = { .type = NLA_U32 }, + [IWPM_NLA_MAPINFO_SEND_NUM] = { .type = NLA_U32 }, + [IWPM_NLA_MAPINFO_ACK_NUM] = { .type = NLA_U32 } +}; + +/* + * iwpm_ack_mapping_info_cb - Process a port mapper ack for + * the provided mapping info records + */ +int iwpm_ack_mapping_info_cb(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct nlattr *nltb[IWPM_NLA_MAPINFO_NUM_MAX]; + u32 mapinfo_send, mapinfo_ack; + const char *msg_type = "Mapping Info Ack"; + + if (iwpm_parse_nlmsg(cb, IWPM_NLA_MAPINFO_NUM_MAX, + ack_mapinfo_policy, nltb, msg_type)) + return -EINVAL; + mapinfo_send = nla_get_u32(nltb[IWPM_NLA_MAPINFO_SEND_NUM]); + mapinfo_ack = nla_get_u32(nltb[IWPM_NLA_MAPINFO_ACK_NUM]); + if (mapinfo_ack != mapinfo_send) + pr_info("%s: Invalid mapinfo number (sent = %u ack-ed = %u)\n", + __func__, mapinfo_send, mapinfo_ack); + atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); + return 0; +} +EXPORT_SYMBOL(iwpm_ack_mapping_info_cb); + +/* netlink attribute policy for the received port mapper error message */ +static const struct nla_policy map_error_policy[IWPM_NLA_ERR_MAX] = { + [IWPM_NLA_ERR_SEQ] = { .type = NLA_U32 }, + [IWPM_NLA_ERR_CODE] = { .type = NLA_U16 }, +}; + +/* + * iwpm_mapping_error_cb - Process a port mapper error message + */ +int iwpm_mapping_error_cb(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct iwpm_nlmsg_request *nlmsg_request = NULL; + int nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type); + struct nlattr *nltb[IWPM_NLA_ERR_MAX]; + u32 msg_seq; + u16 err_code; + const char *msg_type = "Mapping Error Msg"; + + if (iwpm_parse_nlmsg(cb, IWPM_NLA_ERR_MAX, + map_error_policy, nltb, msg_type)) + return -EINVAL; + + msg_seq = nla_get_u32(nltb[IWPM_NLA_ERR_SEQ]); + err_code = nla_get_u16(nltb[IWPM_NLA_ERR_CODE]); + pr_info("%s: Received msg seq = %u err code = %u client = %d\n", + __func__, msg_seq, err_code, nl_client); + /* look for nlmsg_request */ + nlmsg_request = iwpm_find_nlmsg_request(msg_seq); + if (!nlmsg_request) { + /* not all errors have associated requests */ + pr_debug("Could not find matching req (seq = %u)\n", msg_seq); + return 0; + } + atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq); + nlmsg_request->err_code = err_code; + nlmsg_request->request_done = 1; + /* always for found request */ + kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request); + barrier(); + wake_up(&nlmsg_request->waitq); + return 0; +} +EXPORT_SYMBOL(iwpm_mapping_error_cb); diff --git a/drivers/infiniband/core/iwpm_util.c b/drivers/infiniband/core/iwpm_util.c new file mode 100644 index 00000000000..69e9f84c160 --- /dev/null +++ b/drivers/infiniband/core/iwpm_util.c @@ -0,0 +1,607 @@ +/* + * Copyright (c) 2014 Chelsio, Inc. All rights reserved. + * Copyright (c) 2014 Intel Corporation. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include "iwpm_util.h" + +#define IWPM_HASH_BUCKET_SIZE 512 +#define IWPM_HASH_BUCKET_MASK (IWPM_HASH_BUCKET_SIZE - 1) + +static LIST_HEAD(iwpm_nlmsg_req_list); +static DEFINE_SPINLOCK(iwpm_nlmsg_req_lock); + +static struct hlist_head *iwpm_hash_bucket; +static DEFINE_SPINLOCK(iwpm_mapinfo_lock); + +static DEFINE_MUTEX(iwpm_admin_lock); +static struct iwpm_admin_data iwpm_admin; + +int iwpm_init(u8 nl_client) +{ + if (iwpm_valid_client(nl_client)) + return -EINVAL; + mutex_lock(&iwpm_admin_lock); + if (atomic_read(&iwpm_admin.refcount) == 0) { + iwpm_hash_bucket = kzalloc(IWPM_HASH_BUCKET_SIZE * + sizeof(struct hlist_head), GFP_KERNEL); + if (!iwpm_hash_bucket) { + mutex_unlock(&iwpm_admin_lock); + pr_err("%s Unable to create mapinfo hash table\n", __func__); + return -ENOMEM; + } + } + atomic_inc(&iwpm_admin.refcount); + mutex_unlock(&iwpm_admin_lock); + iwpm_set_valid(nl_client, 1); + return 0; +} +EXPORT_SYMBOL(iwpm_init); + +static void free_hash_bucket(void); + +int iwpm_exit(u8 nl_client) +{ + + if (!iwpm_valid_client(nl_client)) + return -EINVAL; + mutex_lock(&iwpm_admin_lock); + if (atomic_read(&iwpm_admin.refcount) == 0) { + mutex_unlock(&iwpm_admin_lock); + pr_err("%s Incorrect usage - negative refcount\n", __func__); + return -EINVAL; + } + if (atomic_dec_and_test(&iwpm_admin.refcount)) { + free_hash_bucket(); + pr_debug("%s: Mapinfo hash table is destroyed\n", __func__); + } + mutex_unlock(&iwpm_admin_lock); + iwpm_set_valid(nl_client, 0); + return 0; +} +EXPORT_SYMBOL(iwpm_exit); + +static struct hlist_head *get_hash_bucket_head(struct sockaddr_storage *, + struct sockaddr_storage *); + +int iwpm_create_mapinfo(struct sockaddr_storage *local_sockaddr, + struct sockaddr_storage *mapped_sockaddr, + u8 nl_client) +{ + struct hlist_head *hash_bucket_head; + struct iwpm_mapping_info *map_info; + unsigned long flags; + + if (!iwpm_valid_client(nl_client)) + return -EINVAL; + map_info = kzalloc(sizeof(struct iwpm_mapping_info), GFP_KERNEL); + if (!map_info) { + pr_err("%s: Unable to allocate a mapping info\n", __func__); + return -ENOMEM; + } + memcpy(&map_info->local_sockaddr, local_sockaddr, + sizeof(struct sockaddr_storage)); + memcpy(&map_info->mapped_sockaddr, mapped_sockaddr, + sizeof(struct sockaddr_storage)); + map_info->nl_client = nl_client; + + spin_lock_irqsave(&iwpm_mapinfo_lock, flags); + if (iwpm_hash_bucket) { + hash_bucket_head = get_hash_bucket_head( + &map_info->local_sockaddr, + &map_info->mapped_sockaddr); + hlist_add_head(&map_info->hlist_node, hash_bucket_head); + } + spin_unlock_irqrestore(&iwpm_mapinfo_lock, flags); + return 0; +} +EXPORT_SYMBOL(iwpm_create_mapinfo); + +int iwpm_remove_mapinfo(struct sockaddr_storage *local_sockaddr, + struct sockaddr_storage *mapped_local_addr) +{ + struct hlist_node *tmp_hlist_node; + struct hlist_head *hash_bucket_head; + struct iwpm_mapping_info *map_info = NULL; + unsigned long flags; + int ret = -EINVAL; + + spin_lock_irqsave(&iwpm_mapinfo_lock, flags); + if (iwpm_hash_bucket) { + hash_bucket_head = get_hash_bucket_head( + local_sockaddr, + mapped_local_addr); + hlist_for_each_entry_safe(map_info, tmp_hlist_node, + hash_bucket_head, hlist_node) { + + if (!iwpm_compare_sockaddr(&map_info->mapped_sockaddr, + mapped_local_addr)) { + + hlist_del_init(&map_info->hlist_node); + kfree(map_info); + ret = 0; + break; + } + } + } + spin_unlock_irqrestore(&iwpm_mapinfo_lock, flags); + return ret; +} +EXPORT_SYMBOL(iwpm_remove_mapinfo); + +static void free_hash_bucket(void) +{ + struct hlist_node *tmp_hlist_node; + struct iwpm_mapping_info *map_info; + unsigned long flags; + int i; + + /* remove all the mapinfo data from the list */ + spin_lock_irqsave(&iwpm_mapinfo_lock, flags); + for (i = 0; i < IWPM_HASH_BUCKET_SIZE; i++) { + hlist_for_each_entry_safe(map_info, tmp_hlist_node, + &iwpm_hash_bucket[i], hlist_node) { + + hlist_del_init(&map_info->hlist_node); + kfree(map_info); + } + } + /* free the hash list */ + kfree(iwpm_hash_bucket); + iwpm_hash_bucket = NULL; + spin_unlock_irqrestore(&iwpm_mapinfo_lock, flags); +} + +struct iwpm_nlmsg_request *iwpm_get_nlmsg_request(__u32 nlmsg_seq, + u8 nl_client, gfp_t gfp) +{ + struct iwpm_nlmsg_request *nlmsg_request = NULL; + unsigned long flags; + + nlmsg_request = kzalloc(sizeof(struct iwpm_nlmsg_request), gfp); + if (!nlmsg_request) { + pr_err("%s Unable to allocate a nlmsg_request\n", __func__); + return NULL; + } + spin_lock_irqsave(&iwpm_nlmsg_req_lock, flags); + list_add_tail(&nlmsg_request->inprocess_list, &iwpm_nlmsg_req_list); + spin_unlock_irqrestore(&iwpm_nlmsg_req_lock, flags); + + kref_init(&nlmsg_request->kref); + kref_get(&nlmsg_request->kref); + nlmsg_request->nlmsg_seq = nlmsg_seq; + nlmsg_request->nl_client = nl_client; + nlmsg_request->request_done = 0; + nlmsg_request->err_code = 0; + return nlmsg_request; +} + +void iwpm_free_nlmsg_request(struct kref *kref) +{ + struct iwpm_nlmsg_request *nlmsg_request; + unsigned long flags; + + nlmsg_request = container_of(kref, struct iwpm_nlmsg_request, kref); + + spin_lock_irqsave(&iwpm_nlmsg_req_lock, flags); + list_del_init(&nlmsg_request->inprocess_list); + spin_unlock_irqrestore(&iwpm_nlmsg_req_lock, flags); + + if (!nlmsg_request->request_done) + pr_debug("%s Freeing incomplete nlmsg request (seq = %u).\n", + __func__, nlmsg_request->nlmsg_seq); + kfree(nlmsg_request); +} + +struct iwpm_nlmsg_request *iwpm_find_nlmsg_request(__u32 echo_seq) +{ + struct iwpm_nlmsg_request *nlmsg_request; + struct iwpm_nlmsg_request *found_request = NULL; + unsigned long flags; + + spin_lock_irqsave(&iwpm_nlmsg_req_lock, flags); + list_for_each_entry(nlmsg_request, &iwpm_nlmsg_req_list, + inprocess_list) { + if (nlmsg_request->nlmsg_seq == echo_seq) { + found_request = nlmsg_request; + kref_get(&nlmsg_request->kref); + break; + } + } + spin_unlock_irqrestore(&iwpm_nlmsg_req_lock, flags); + return found_request; +} + +int iwpm_wait_complete_req(struct iwpm_nlmsg_request *nlmsg_request) +{ + int ret; + init_waitqueue_head(&nlmsg_request->waitq); + + ret = wait_event_timeout(nlmsg_request->waitq, + (nlmsg_request->request_done != 0), IWPM_NL_TIMEOUT); + if (!ret) { + ret = -EINVAL; + pr_info("%s: Timeout %d sec for netlink request (seq = %u)\n", + __func__, (IWPM_NL_TIMEOUT/HZ), nlmsg_request->nlmsg_seq); + } else { + ret = nlmsg_request->err_code; + } + kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request); + return ret; +} + +int iwpm_get_nlmsg_seq(void) +{ + return atomic_inc_return(&iwpm_admin.nlmsg_seq); +} + +int iwpm_valid_client(u8 nl_client) +{ + if (nl_client >= RDMA_NL_NUM_CLIENTS) + return 0; + return iwpm_admin.client_list[nl_client]; +} + +void iwpm_set_valid(u8 nl_client, int valid) +{ + if (nl_client >= RDMA_NL_NUM_CLIENTS) + return; + iwpm_admin.client_list[nl_client] = valid; +} + +/* valid client */ +int iwpm_registered_client(u8 nl_client) +{ + return iwpm_admin.reg_list[nl_client]; +} + +/* valid client */ +void iwpm_set_registered(u8 nl_client, int reg) +{ + iwpm_admin.reg_list[nl_client] = reg; +} + +int iwpm_compare_sockaddr(struct sockaddr_storage *a_sockaddr, + struct sockaddr_storage *b_sockaddr) +{ + if (a_sockaddr->ss_family != b_sockaddr->ss_family) + return 1; + if (a_sockaddr->ss_family == AF_INET) { + struct sockaddr_in *a4_sockaddr = + (struct sockaddr_in *)a_sockaddr; + struct sockaddr_in *b4_sockaddr = + (struct sockaddr_in *)b_sockaddr; + if (!memcmp(&a4_sockaddr->sin_addr, + &b4_sockaddr->sin_addr, sizeof(struct in_addr)) + && a4_sockaddr->sin_port == b4_sockaddr->sin_port) + return 0; + + } else if (a_sockaddr->ss_family == AF_INET6) { + struct sockaddr_in6 *a6_sockaddr = + (struct sockaddr_in6 *)a_sockaddr; + struct sockaddr_in6 *b6_sockaddr = + (struct sockaddr_in6 *)b_sockaddr; + if (!memcmp(&a6_sockaddr->sin6_addr, + &b6_sockaddr->sin6_addr, sizeof(struct in6_addr)) + && a6_sockaddr->sin6_port == b6_sockaddr->sin6_port) + return 0; + + } else { + pr_err("%s: Invalid sockaddr family\n", __func__); + } + return 1; +} + +struct sk_buff *iwpm_create_nlmsg(u32 nl_op, struct nlmsghdr **nlh, + int nl_client) +{ + struct sk_buff *skb = NULL; + + skb = dev_alloc_skb(NLMSG_GOODSIZE); + if (!skb) { + pr_err("%s Unable to allocate skb\n", __func__); + goto create_nlmsg_exit; + } + if (!(ibnl_put_msg(skb, nlh, 0, 0, nl_client, nl_op, + NLM_F_REQUEST))) { + pr_warn("%s: Unable to put the nlmsg header\n", __func__); + dev_kfree_skb(skb); + skb = NULL; + } +create_nlmsg_exit: + return skb; +} + +int iwpm_parse_nlmsg(struct netlink_callback *cb, int policy_max, + const struct nla_policy *nlmsg_policy, + struct nlattr *nltb[], const char *msg_type) +{ + int nlh_len = 0; + int ret; + const char *err_str = ""; + + ret = nlmsg_validate(cb->nlh, nlh_len, policy_max-1, nlmsg_policy); + if (ret) { + err_str = "Invalid attribute"; + goto parse_nlmsg_error; + } + ret = nlmsg_parse(cb->nlh, nlh_len, nltb, policy_max-1, nlmsg_policy); + if (ret) { + err_str = "Unable to parse the nlmsg"; + goto parse_nlmsg_error; + } + ret = iwpm_validate_nlmsg_attr(nltb, policy_max); + if (ret) { + err_str = "Invalid NULL attribute"; + goto parse_nlmsg_error; + } + return 0; +parse_nlmsg_error: + pr_warn("%s: %s (msg type %s ret = %d)\n", + __func__, err_str, msg_type, ret); + return ret; +} + +void iwpm_print_sockaddr(struct sockaddr_storage *sockaddr, char *msg) +{ + struct sockaddr_in6 *sockaddr_v6; + struct sockaddr_in *sockaddr_v4; + + switch (sockaddr->ss_family) { + case AF_INET: + sockaddr_v4 = (struct sockaddr_in *)sockaddr; + pr_debug("%s IPV4 %pI4: %u(0x%04X)\n", + msg, &sockaddr_v4->sin_addr, + ntohs(sockaddr_v4->sin_port), + ntohs(sockaddr_v4->sin_port)); + break; + case AF_INET6: + sockaddr_v6 = (struct sockaddr_in6 *)sockaddr; + pr_debug("%s IPV6 %pI6: %u(0x%04X)\n", + msg, &sockaddr_v6->sin6_addr, + ntohs(sockaddr_v6->sin6_port), + ntohs(sockaddr_v6->sin6_port)); + break; + default: + break; + } +} + +static u32 iwpm_ipv6_jhash(struct sockaddr_in6 *ipv6_sockaddr) +{ + u32 ipv6_hash = jhash(&ipv6_sockaddr->sin6_addr, sizeof(struct in6_addr), 0); + u32 hash = jhash_2words(ipv6_hash, (__force u32) ipv6_sockaddr->sin6_port, 0); + return hash; +} + +static u32 iwpm_ipv4_jhash(struct sockaddr_in *ipv4_sockaddr) +{ + u32 ipv4_hash = jhash(&ipv4_sockaddr->sin_addr, sizeof(struct in_addr), 0); + u32 hash = jhash_2words(ipv4_hash, (__force u32) ipv4_sockaddr->sin_port, 0); + return hash; +} + +static struct hlist_head *get_hash_bucket_head(struct sockaddr_storage + *local_sockaddr, + struct sockaddr_storage + *mapped_sockaddr) +{ + u32 local_hash, mapped_hash, hash; + + if (local_sockaddr->ss_family == AF_INET) { + local_hash = iwpm_ipv4_jhash((struct sockaddr_in *) local_sockaddr); + mapped_hash = iwpm_ipv4_jhash((struct sockaddr_in *) mapped_sockaddr); + + } else if (local_sockaddr->ss_family == AF_INET6) { + local_hash = iwpm_ipv6_jhash((struct sockaddr_in6 *) local_sockaddr); + mapped_hash = iwpm_ipv6_jhash((struct sockaddr_in6 *) mapped_sockaddr); + } else { + pr_err("%s: Invalid sockaddr family\n", __func__); + return NULL; + } + + if (local_hash == mapped_hash) /* if port mapper isn't available */ + hash = local_hash; + else + hash = jhash_2words(local_hash, mapped_hash, 0); + + return &iwpm_hash_bucket[hash & IWPM_HASH_BUCKET_MASK]; +} + +static int send_mapinfo_num(u32 mapping_num, u8 nl_client, int iwpm_pid) +{ + struct sk_buff *skb = NULL; + struct nlmsghdr *nlh; + u32 msg_seq; + const char *err_str = ""; + int ret = -EINVAL; + + skb = iwpm_create_nlmsg(RDMA_NL_IWPM_MAPINFO_NUM, &nlh, nl_client); + if (!skb) { + err_str = "Unable to create a nlmsg"; + goto mapinfo_num_error; + } + nlh->nlmsg_seq = iwpm_get_nlmsg_seq(); + msg_seq = 0; + err_str = "Unable to put attribute of mapinfo number nlmsg"; + ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq, IWPM_NLA_MAPINFO_SEQ); + if (ret) + goto mapinfo_num_error; + ret = ibnl_put_attr(skb, nlh, sizeof(u32), + &mapping_num, IWPM_NLA_MAPINFO_SEND_NUM); + if (ret) + goto mapinfo_num_error; + ret = ibnl_unicast(skb, nlh, iwpm_pid); + if (ret) { + skb = NULL; + err_str = "Unable to send a nlmsg"; + goto mapinfo_num_error; + } + pr_debug("%s: Sent mapping number = %d\n", __func__, mapping_num); + return 0; +mapinfo_num_error: + pr_info("%s: %s\n", __func__, err_str); + if (skb) + dev_kfree_skb(skb); + return ret; +} + +static int send_nlmsg_done(struct sk_buff *skb, u8 nl_client, int iwpm_pid) +{ + struct nlmsghdr *nlh = NULL; + int ret = 0; + + if (!skb) + return ret; + if (!(ibnl_put_msg(skb, &nlh, 0, 0, nl_client, + RDMA_NL_IWPM_MAPINFO, NLM_F_MULTI))) { + pr_warn("%s Unable to put NLMSG_DONE\n", __func__); + return -ENOMEM; + } + nlh->nlmsg_type = NLMSG_DONE; + ret = ibnl_unicast(skb, (struct nlmsghdr *)skb->data, iwpm_pid); + if (ret) + pr_warn("%s Unable to send a nlmsg\n", __func__); + return ret; +} + +int iwpm_send_mapinfo(u8 nl_client, int iwpm_pid) +{ + struct iwpm_mapping_info *map_info; + struct sk_buff *skb = NULL; + struct nlmsghdr *nlh; + int skb_num = 0, mapping_num = 0; + int i = 0, nlmsg_bytes = 0; + unsigned long flags; + const char *err_str = ""; + int ret; + + skb = dev_alloc_skb(NLMSG_GOODSIZE); + if (!skb) { + ret = -ENOMEM; + err_str = "Unable to allocate skb"; + goto send_mapping_info_exit; + } + skb_num++; + spin_lock_irqsave(&iwpm_mapinfo_lock, flags); + for (i = 0; i < IWPM_HASH_BUCKET_SIZE; i++) { + hlist_for_each_entry(map_info, &iwpm_hash_bucket[i], + hlist_node) { + if (map_info->nl_client != nl_client) + continue; + nlh = NULL; + if (!(ibnl_put_msg(skb, &nlh, 0, 0, nl_client, + RDMA_NL_IWPM_MAPINFO, NLM_F_MULTI))) { + ret = -ENOMEM; + err_str = "Unable to put the nlmsg header"; + goto send_mapping_info_unlock; + } + err_str = "Unable to put attribute of the nlmsg"; + ret = ibnl_put_attr(skb, nlh, + sizeof(struct sockaddr_storage), + &map_info->local_sockaddr, + IWPM_NLA_MAPINFO_LOCAL_ADDR); + if (ret) + goto send_mapping_info_unlock; + + ret = ibnl_put_attr(skb, nlh, + sizeof(struct sockaddr_storage), + &map_info->mapped_sockaddr, + IWPM_NLA_MAPINFO_MAPPED_ADDR); + if (ret) + goto send_mapping_info_unlock; + + iwpm_print_sockaddr(&map_info->local_sockaddr, + "send_mapping_info: Local sockaddr:"); + iwpm_print_sockaddr(&map_info->mapped_sockaddr, + "send_mapping_info: Mapped local sockaddr:"); + mapping_num++; + nlmsg_bytes += nlh->nlmsg_len; + + /* check if all mappings can fit in one skb */ + if (NLMSG_GOODSIZE - nlmsg_bytes < nlh->nlmsg_len * 2) { + /* and leave room for NLMSG_DONE */ + nlmsg_bytes = 0; + skb_num++; + spin_unlock_irqrestore(&iwpm_mapinfo_lock, + flags); + /* send the skb */ + ret = send_nlmsg_done(skb, nl_client, iwpm_pid); + skb = NULL; + if (ret) { + err_str = "Unable to send map info"; + goto send_mapping_info_exit; + } + if (skb_num == IWPM_MAPINFO_SKB_COUNT) { + ret = -ENOMEM; + err_str = "Insufficient skbs for map info"; + goto send_mapping_info_exit; + } + skb = dev_alloc_skb(NLMSG_GOODSIZE); + if (!skb) { + ret = -ENOMEM; + err_str = "Unable to allocate skb"; + goto send_mapping_info_exit; + } + spin_lock_irqsave(&iwpm_mapinfo_lock, flags); + } + } + } +send_mapping_info_unlock: + spin_unlock_irqrestore(&iwpm_mapinfo_lock, flags); +send_mapping_info_exit: + if (ret) { + pr_warn("%s: %s (ret = %d)\n", __func__, err_str, ret); + if (skb) + dev_kfree_skb(skb); + return ret; + } + send_nlmsg_done(skb, nl_client, iwpm_pid); + return send_mapinfo_num(mapping_num, nl_client, iwpm_pid); +} + +int iwpm_mapinfo_available(void) +{ + unsigned long flags; + int full_bucket = 0, i = 0; + + spin_lock_irqsave(&iwpm_mapinfo_lock, flags); + if (iwpm_hash_bucket) { + for (i = 0; i < IWPM_HASH_BUCKET_SIZE; i++) { + if (!hlist_empty(&iwpm_hash_bucket[i])) { + full_bucket = 1; + break; + } + } + } + spin_unlock_irqrestore(&iwpm_mapinfo_lock, flags); + return full_bucket; +} diff --git a/drivers/infiniband/core/iwpm_util.h b/drivers/infiniband/core/iwpm_util.h new file mode 100644 index 00000000000..9777c869a14 --- /dev/null +++ b/drivers/infiniband/core/iwpm_util.h @@ -0,0 +1,238 @@ +/* + * Copyright (c) 2014 Intel Corporation. All rights reserved. + * Copyright (c) 2014 Chelsio, Inc. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef _IWPM_UTIL_H +#define _IWPM_UTIL_H + +#include <linux/module.h> +#include <linux/io.h> +#include <linux/in.h> +#include <linux/in6.h> +#include <linux/spinlock.h> +#include <linux/kernel.h> +#include <linux/netdevice.h> +#include <linux/delay.h> +#include <linux/workqueue.h> +#include <linux/mutex.h> +#include <linux/jhash.h> +#include <linux/kref.h> +#include <net/netlink.h> +#include <linux/errno.h> +#include <rdma/iw_portmap.h> +#include <rdma/rdma_netlink.h> + + +#define IWPM_NL_RETRANS 3 +#define IWPM_NL_TIMEOUT (10*HZ) +#define IWPM_MAPINFO_SKB_COUNT 20 + +#define IWPM_PID_UNDEFINED -1 +#define IWPM_PID_UNAVAILABLE -2 + +struct iwpm_nlmsg_request { + struct list_head inprocess_list; + __u32 nlmsg_seq; + void *req_buffer; + u8 nl_client; + u8 request_done; + u16 err_code; + wait_queue_head_t waitq; + struct kref kref; +}; + +struct iwpm_mapping_info { + struct hlist_node hlist_node; + struct sockaddr_storage local_sockaddr; + struct sockaddr_storage mapped_sockaddr; + u8 nl_client; +}; + +struct iwpm_admin_data { + atomic_t refcount; + atomic_t nlmsg_seq; + int client_list[RDMA_NL_NUM_CLIENTS]; + int reg_list[RDMA_NL_NUM_CLIENTS]; +}; + +/** + * iwpm_get_nlmsg_request - Allocate and initialize netlink message request + * @nlmsg_seq: Sequence number of the netlink message + * @nl_client: The index of the netlink client + * @gfp: Indicates how the memory for the request should be allocated + * + * Returns the newly allocated netlink request object if successful, + * otherwise returns NULL + */ +struct iwpm_nlmsg_request *iwpm_get_nlmsg_request(__u32 nlmsg_seq, + u8 nl_client, gfp_t gfp); + +/** + * iwpm_free_nlmsg_request - Deallocate netlink message request + * @kref: Holds reference of netlink message request + */ +void iwpm_free_nlmsg_request(struct kref *kref); + +/** + * iwpm_find_nlmsg_request - Find netlink message request in the request list + * @echo_seq: Sequence number of the netlink request to find + * + * Returns the found netlink message request, + * if not found, returns NULL + */ +struct iwpm_nlmsg_request *iwpm_find_nlmsg_request(__u32 echo_seq); + +/** + * iwpm_wait_complete_req - Block while servicing the netlink request + * @nlmsg_request: Netlink message request to service + * + * Wakes up, after the request is completed or expired + * Returns 0 if the request is complete without error + */ +int iwpm_wait_complete_req(struct iwpm_nlmsg_request *nlmsg_request); + +/** + * iwpm_get_nlmsg_seq - Get the sequence number for a netlink + * message to send to the port mapper + * + * Returns the sequence number for the netlink message. + */ +int iwpm_get_nlmsg_seq(void); + +/** + * iwpm_valid_client - Check if the port mapper client is valid + * @nl_client: The index of the netlink client + * + * Valid clients need to call iwpm_init() before using + * the port mapper + */ +int iwpm_valid_client(u8 nl_client); + +/** + * iwpm_set_valid - Set the port mapper client to valid or not + * @nl_client: The index of the netlink client + * @valid: 1 if valid or 0 if invalid + */ +void iwpm_set_valid(u8 nl_client, int valid); + +/** + * iwpm_registered_client - Check if the port mapper client is registered + * @nl_client: The index of the netlink client + * + * Call iwpm_register_pid() to register a client + */ +int iwpm_registered_client(u8 nl_client); + +/** + * iwpm_set_registered - Set the port mapper client to registered or not + * @nl_client: The index of the netlink client + * @reg: 1 if registered or 0 if not + */ +void iwpm_set_registered(u8 nl_client, int reg); + +/** + * iwpm_send_mapinfo - Send local and mapped IPv4/IPv6 address info of + * a client to the user space port mapper + * @nl_client: The index of the netlink client + * @iwpm_pid: The pid of the user space port mapper + * + * If successful, returns the number of sent mapping info records + */ +int iwpm_send_mapinfo(u8 nl_client, int iwpm_pid); + +/** + * iwpm_mapinfo_available - Check if any mapping info records is available + * in the hash table + * + * Returns 1 if mapping information is available, otherwise returns 0 + */ +int iwpm_mapinfo_available(void); + +/** + * iwpm_compare_sockaddr - Compare two sockaddr storage structs + * + * Returns 0 if they are holding the same ip/tcp address info, + * otherwise returns 1 + */ +int iwpm_compare_sockaddr(struct sockaddr_storage *a_sockaddr, + struct sockaddr_storage *b_sockaddr); + +/** + * iwpm_validate_nlmsg_attr - Check for NULL netlink attributes + * @nltb: Holds address of each netlink message attributes + * @nla_count: Number of netlink message attributes + * + * Returns error if any of the nla_count attributes is NULL + */ +static inline int iwpm_validate_nlmsg_attr(struct nlattr *nltb[], + int nla_count) +{ + int i; + for (i = 1; i < nla_count; i++) { + if (!nltb[i]) + return -EINVAL; + } + return 0; +} + +/** + * iwpm_create_nlmsg - Allocate skb and form a netlink message + * @nl_op: Netlink message opcode + * @nlh: Holds address of the netlink message header in skb + * @nl_client: The index of the netlink client + * + * Returns the newly allcated skb, or NULL if the tailroom of the skb + * is insufficient to store the message header and payload + */ +struct sk_buff *iwpm_create_nlmsg(u32 nl_op, struct nlmsghdr **nlh, + int nl_client); + +/** + * iwpm_parse_nlmsg - Validate and parse the received netlink message + * @cb: Netlink callback structure + * @policy_max: Maximum attribute type to be expected + * @nlmsg_policy: Validation policy + * @nltb: Array to store policy_max parsed elements + * @msg_type: Type of netlink message + * + * Returns 0 on success or a negative error code + */ +int iwpm_parse_nlmsg(struct netlink_callback *cb, int policy_max, + const struct nla_policy *nlmsg_policy, + struct nlattr *nltb[], const char *msg_type); + +/** + * iwpm_print_sockaddr - Print IPv4/IPv6 address and TCP port + * @sockaddr: Socket address to print + * @msg: Message to print + */ +void iwpm_print_sockaddr(struct sockaddr_storage *sockaddr, char *msg); +#endif diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index d393b504bf2..ab31f136d04 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -1,7 +1,8 @@ /* - * Copyright (c) 2004, 2005 Voltaire, Inc. All rights reserved. + * Copyright (c) 2004-2007 Voltaire, Inc. All rights reserved. * Copyright (c) 2005 Intel Corporation. All rights reserved. * Copyright (c) 2005 Mellanox Technologies Ltd. All rights reserved. + * Copyright (c) 2009 HNR Consulting. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -31,9 +32,11 @@ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. * - * $Id: mad.c 2817 2005-07-07 11:29:26Z halr $ */ #include <linux/dma-mapping.h> +#include <linux/slab.h> +#include <linux/module.h> +#include <rdma/ib_cache.h> #include "mad_priv.h" #include "mad_rmpp.h" @@ -45,15 +48,21 @@ MODULE_DESCRIPTION("kernel IB MAD API"); MODULE_AUTHOR("Hal Rosenstock"); MODULE_AUTHOR("Sean Hefty"); +static int mad_sendq_size = IB_MAD_QP_SEND_SIZE; +static int mad_recvq_size = IB_MAD_QP_RECV_SIZE; -kmem_cache_t *ib_mad_cache; +module_param_named(send_queue_size, mad_sendq_size, int, 0444); +MODULE_PARM_DESC(send_queue_size, "Size of send queue in number of work requests"); +module_param_named(recv_queue_size, mad_recvq_size, int, 0444); +MODULE_PARM_DESC(recv_queue_size, "Size of receive queue in number of work requests"); + +static struct kmem_cache *ib_mad_cache; static struct list_head ib_mad_port_list; static u32 ib_mad_client_id = 0; /* Port list lock */ -static spinlock_t ib_mad_port_list_lock; - +static DEFINE_SPINLOCK(ib_mad_port_list_lock); /* Forward declarations */ static int method_in_use(struct ib_mad_mgmt_method_table **method, @@ -65,8 +74,8 @@ static struct ib_mad_agent_private *find_mad_agent( static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info, struct ib_mad_private *mad); static void cancel_mads(struct ib_mad_agent_private *mad_agent_priv); -static void timeout_sends(void *data); -static void local_completions(void *data); +static void timeout_sends(struct work_struct *work); +static void local_completions(struct work_struct *work); static int add_nonoui_reg_req(struct ib_mad_reg_req *mad_reg_req, struct ib_mad_agent_private *agent_priv, u8 mgmt_class); @@ -167,6 +176,15 @@ static int is_vendor_method_in_use( return 0; } +int ib_response_mad(struct ib_mad *mad) +{ + return ((mad->mad_hdr.method & IB_MGMT_METHOD_RESP) || + (mad->mad_hdr.method == IB_MGMT_METHOD_TRAP_REPRESS) || + ((mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_BM) && + (mad->mad_hdr.attr_mod & IB_BM_ATTR_MOD_RESP))); +} +EXPORT_SYMBOL(ib_response_mad); + /* * ib_register_mad_agent - Register to send/receive MADs */ @@ -227,6 +245,11 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, if (!is_vendor_oui(mad_reg_req->oui)) goto error1; } + /* Make sure class supplied is consistent with RMPP */ + if (!ib_is_mad_class_rmpp(mad_reg_req->mgmt_class)) { + if (rmpp_version) + goto error1; + } /* Make sure class supplied is consistent with QP type */ if (qp_type == IB_QPT_SMI) { if ((mad_reg_req->mgmt_class != @@ -254,6 +277,13 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, goto error1; } + /* Verify the QP requested is supported. For example, Ethernet devices + * will not have QP0 */ + if (!port_priv->qp_info[qpn].qp) { + ret = ERR_PTR(-EPROTONOSUPPORT); + goto error1; + } + /* Allocate structures */ mad_agent_priv = kzalloc(sizeof *mad_agent_priv, GFP_KERNEL); if (!mad_agent_priv) { @@ -269,13 +299,11 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, } if (mad_reg_req) { - reg_req = kmalloc(sizeof *reg_req, GFP_KERNEL); + reg_req = kmemdup(mad_reg_req, sizeof *reg_req, GFP_KERNEL); if (!reg_req) { ret = ERR_PTR(-ENOMEM); goto error3; } - /* Make a copy of the MAD registration request */ - memcpy(reg_req, mad_reg_req, sizeof *reg_req); } /* Now, fill in the various structures */ @@ -288,6 +316,16 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, mad_agent_priv->agent.context = context; mad_agent_priv->agent.qp = port_priv->qp_info[qpn].qp; mad_agent_priv->agent.port_num = port_num; + spin_lock_init(&mad_agent_priv->lock); + INIT_LIST_HEAD(&mad_agent_priv->send_list); + INIT_LIST_HEAD(&mad_agent_priv->wait_list); + INIT_LIST_HEAD(&mad_agent_priv->done_list); + INIT_LIST_HEAD(&mad_agent_priv->rmpp_list); + INIT_DELAYED_WORK(&mad_agent_priv->timed_work, timeout_sends); + INIT_LIST_HEAD(&mad_agent_priv->local_list); + INIT_WORK(&mad_agent_priv->local_work, local_completions); + atomic_set(&mad_agent_priv->refcount, 1); + init_completion(&mad_agent_priv->comp); spin_lock_irqsave(&port_priv->reg_lock, flags); mad_agent_priv->agent.hi_tid = ++ib_mad_client_id; @@ -337,18 +375,6 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, list_add_tail(&mad_agent_priv->agent_list, &port_priv->agent_list); spin_unlock_irqrestore(&port_priv->reg_lock, flags); - spin_lock_init(&mad_agent_priv->lock); - INIT_LIST_HEAD(&mad_agent_priv->send_list); - INIT_LIST_HEAD(&mad_agent_priv->wait_list); - INIT_LIST_HEAD(&mad_agent_priv->done_list); - INIT_LIST_HEAD(&mad_agent_priv->rmpp_list); - INIT_WORK(&mad_agent_priv->timed_work, timeout_sends, mad_agent_priv); - INIT_LIST_HEAD(&mad_agent_priv->local_list); - INIT_WORK(&mad_agent_priv->local_work, local_completions, - mad_agent_priv); - atomic_set(&mad_agent_priv->refcount, 1); - init_waitqueue_head(&mad_agent_priv->wait); - return &mad_agent_priv->agent; error4: @@ -394,19 +420,15 @@ static int register_snoop_agent(struct ib_mad_qp_info *qp_info, if (i == qp_info->snoop_table_size) { /* Grow table. */ - new_snoop_table = kmalloc(sizeof mad_snoop_priv * - qp_info->snoop_table_size + 1, - GFP_ATOMIC); + new_snoop_table = krealloc(qp_info->snoop_table, + sizeof mad_snoop_priv * + (qp_info->snoop_table_size + 1), + GFP_ATOMIC); if (!new_snoop_table) { i = -ENOMEM; goto out; } - if (qp_info->snoop_table) { - memcpy(new_snoop_table, qp_info->snoop_table, - sizeof mad_snoop_priv * - qp_info->snoop_table_size); - kfree(qp_info->snoop_table); - } + qp_info->snoop_table = new_snoop_table; qp_info->snoop_table_size++; } @@ -462,7 +484,7 @@ struct ib_mad_agent *ib_register_mad_snoop(struct ib_device *device, mad_snoop_priv->agent.qp = port_priv->qp_info[qpn].qp; mad_snoop_priv->agent.port_num = port_num; mad_snoop_priv->mad_snoop_flags = mad_snoop_flags; - init_waitqueue_head(&mad_snoop_priv->wait); + init_completion(&mad_snoop_priv->comp); mad_snoop_priv->snoop_index = register_snoop_agent( &port_priv->qp_info[qpn], mad_snoop_priv); @@ -481,6 +503,18 @@ error1: } EXPORT_SYMBOL(ib_register_mad_snoop); +static inline void deref_mad_agent(struct ib_mad_agent_private *mad_agent_priv) +{ + if (atomic_dec_and_test(&mad_agent_priv->refcount)) + complete(&mad_agent_priv->comp); +} + +static inline void deref_snoop_agent(struct ib_mad_snoop_private *mad_snoop_priv) +{ + if (atomic_dec_and_test(&mad_snoop_priv->refcount)) + complete(&mad_snoop_priv->comp); +} + static void unregister_mad_agent(struct ib_mad_agent_private *mad_agent_priv) { struct ib_mad_port_private *port_priv; @@ -504,9 +538,8 @@ static void unregister_mad_agent(struct ib_mad_agent_private *mad_agent_priv) flush_workqueue(port_priv->wq); ib_cancel_rmpp_recvs(mad_agent_priv); - atomic_dec(&mad_agent_priv->refcount); - wait_event(mad_agent_priv->wait, - !atomic_read(&mad_agent_priv->refcount)); + deref_mad_agent(mad_agent_priv); + wait_for_completion(&mad_agent_priv->comp); kfree(mad_agent_priv->reg_req); ib_dereg_mr(mad_agent_priv->agent.mr); @@ -524,9 +557,8 @@ static void unregister_mad_snoop(struct ib_mad_snoop_private *mad_snoop_priv) atomic_dec(&qp_info->snoop_count); spin_unlock_irqrestore(&qp_info->snoop_lock, flags); - atomic_dec(&mad_snoop_priv->refcount); - wait_event(mad_snoop_priv->wait, - !atomic_read(&mad_snoop_priv->refcount)); + deref_snoop_agent(mad_snoop_priv); + wait_for_completion(&mad_snoop_priv->comp); kfree(mad_snoop_priv); } @@ -555,13 +587,6 @@ int ib_unregister_mad_agent(struct ib_mad_agent *mad_agent) } EXPORT_SYMBOL(ib_unregister_mad_agent); -static inline int response_mad(struct ib_mad *mad) -{ - /* Trap represses are responses although response bit is reset */ - return ((mad->mad_hdr.method == IB_MGMT_METHOD_TRAP_REPRESS) || - (mad->mad_hdr.method & IB_MGMT_METHOD_RESP)); -} - static void dequeue_mad(struct ib_mad_list_head *mad_list) { struct ib_mad_queue *mad_queue; @@ -595,8 +620,7 @@ static void snoop_send(struct ib_mad_qp_info *qp_info, spin_unlock_irqrestore(&qp_info->snoop_lock, flags); mad_snoop_priv->agent.snoop_handler(&mad_snoop_priv->agent, send_buf, mad_send_wc); - if (atomic_dec_and_test(&mad_snoop_priv->refcount)) - wake_up(&mad_snoop_priv->wait); + deref_snoop_agent(mad_snoop_priv); spin_lock_irqsave(&qp_info->snoop_lock, flags); } spin_unlock_irqrestore(&qp_info->snoop_lock, flags); @@ -621,14 +645,14 @@ static void snoop_recv(struct ib_mad_qp_info *qp_info, spin_unlock_irqrestore(&qp_info->snoop_lock, flags); mad_snoop_priv->agent.recv_handler(&mad_snoop_priv->agent, mad_recv_wc); - if (atomic_dec_and_test(&mad_snoop_priv->refcount)) - wake_up(&mad_snoop_priv->wait); + deref_snoop_agent(mad_snoop_priv); spin_lock_irqsave(&qp_info->snoop_lock, flags); } spin_unlock_irqrestore(&qp_info->snoop_lock, flags); } -static void build_smp_wc(u64 wr_id, u16 slid, u16 pkey_index, u8 port_num, +static void build_smp_wc(struct ib_qp *qp, + u64 wr_id, u16 slid, u16 pkey_index, u8 port_num, struct ib_wc *wc) { memset(wc, 0, sizeof *wc); @@ -638,7 +662,7 @@ static void build_smp_wc(u64 wr_id, u16 slid, u16 pkey_index, u8 port_num, wc->pkey_index = pkey_index; wc->byte_len = sizeof(struct ib_mad) + sizeof(struct ib_grh); wc->src_qp = IB_QP0; - wc->qp_num = IB_QP0; + wc->qp = qp; wc->slid = slid; wc->sl = 0; wc->dlid_path_bits = 0; @@ -653,7 +677,7 @@ static void build_smp_wc(u64 wr_id, u16 slid, u16 pkey_index, u8 port_num, static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv, struct ib_mad_send_wr_private *mad_send_wr) { - int ret; + int ret = 0; struct ib_smp *smp = mad_send_wr->send_buf.mad; unsigned long flags; struct ib_mad_local_private *local; @@ -661,18 +685,34 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv, struct ib_mad_port_private *port_priv; struct ib_mad_agent_private *recv_mad_agent = NULL; struct ib_device *device = mad_agent_priv->agent.device; - u8 port_num = mad_agent_priv->agent.port_num; + u8 port_num; struct ib_wc mad_wc; struct ib_send_wr *send_wr = &mad_send_wr->send_wr; - if (!smi_handle_dr_smp_send(smp, device->node_type, port_num)) { + if (device->node_type == RDMA_NODE_IB_SWITCH && + smp->mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) + port_num = send_wr->wr.ud.port_num; + else + port_num = mad_agent_priv->agent.port_num; + + /* + * Directed route handling starts if the initial LID routed part of + * a request or the ending LID routed part of a response is empty. + * If we are at the start of the LID routed part, don't update the + * hop_ptr or hop_cnt. See section 14.2.2, Vol 1 IB spec. + */ + if ((ib_get_smp_direction(smp) ? smp->dr_dlid : smp->dr_slid) == + IB_LID_PERMISSIVE && + smi_handle_dr_smp_send(smp, device->node_type, port_num) == + IB_SMI_DISCARD) { ret = -EINVAL; printk(KERN_ERR PFX "Invalid directed route\n"); goto out; } + /* Check to post send on QP or process locally */ - ret = smi_check_local_dr_smp(smp, device, port_num); - if (!ret || !device->process_mad) + if (smi_check_local_smp(smp, device) == IB_SMI_DISCARD && + smi_check_local_returning_smp(smp, device) == IB_SMI_DISCARD) goto out; local = kmalloc(sizeof *local, GFP_ATOMIC); @@ -691,7 +731,8 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv, goto out; } - build_smp_wc(send_wr->wr_id, be16_to_cpu(smp->dr_slid), + build_smp_wc(mad_agent_priv->agent.qp, + send_wr->wr_id, be16_to_cpu(smp->dr_slid), send_wr->wr.ud.pkey_index, send_wr->wr.ud.port_num, &mad_wc); @@ -702,7 +743,7 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv, switch (ret) { case IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_REPLY: - if (response_mad(&mad_priv->mad.mad) && + if (ib_response_mad(&mad_priv->mad.mad) && mad_agent_priv->agent.recv_handler) { local->mad_priv = mad_priv; local->recv_mad_agent = mad_agent_priv; @@ -722,16 +763,17 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv, port_priv = ib_get_mad_port(mad_agent_priv->agent.device, mad_agent_priv->agent.port_num); if (port_priv) { - mad_priv->mad.mad.mad_hdr.tid = - ((struct ib_mad *)smp)->mad_hdr.tid; + memcpy(&mad_priv->mad.mad, smp, sizeof(struct ib_mad)); recv_mad_agent = find_mad_agent(port_priv, &mad_priv->mad.mad); } if (!port_priv || !recv_mad_agent) { + /* + * No receiving agent so drop packet and + * generate send completion. + */ kmem_cache_free(ib_mad_cache, mad_priv); - kfree(local); - ret = 0; - goto out; + break; } local->mad_priv = mad_priv; local->recv_mad_agent = recv_mad_agent; @@ -757,18 +799,67 @@ out: return ret; } -static int get_buf_length(int hdr_len, int data_len) +static int get_pad_size(int hdr_len, int data_len) { int seg_size, pad; seg_size = sizeof(struct ib_mad) - hdr_len; if (data_len && seg_size) { pad = seg_size - data_len % seg_size; - if (pad == seg_size) - pad = 0; + return pad == seg_size ? 0 : pad; } else - pad = seg_size; - return hdr_len + data_len + pad; + return seg_size; +} + +static void free_send_rmpp_list(struct ib_mad_send_wr_private *mad_send_wr) +{ + struct ib_rmpp_segment *s, *t; + + list_for_each_entry_safe(s, t, &mad_send_wr->rmpp_list, list) { + list_del(&s->list); + kfree(s); + } +} + +static int alloc_send_rmpp_list(struct ib_mad_send_wr_private *send_wr, + gfp_t gfp_mask) +{ + struct ib_mad_send_buf *send_buf = &send_wr->send_buf; + struct ib_rmpp_mad *rmpp_mad = send_buf->mad; + struct ib_rmpp_segment *seg = NULL; + int left, seg_size, pad; + + send_buf->seg_size = sizeof (struct ib_mad) - send_buf->hdr_len; + seg_size = send_buf->seg_size; + pad = send_wr->pad; + + /* Allocate data segments. */ + for (left = send_buf->data_len + pad; left > 0; left -= seg_size) { + seg = kmalloc(sizeof (*seg) + seg_size, gfp_mask); + if (!seg) { + printk(KERN_ERR "alloc_send_rmpp_segs: RMPP mem " + "alloc failed for len %zd, gfp %#x\n", + sizeof (*seg) + seg_size, gfp_mask); + free_send_rmpp_list(send_wr); + return -ENOMEM; + } + seg->num = ++send_buf->seg_count; + list_add_tail(&seg->list, &send_wr->rmpp_list); + } + + /* Zero any padding */ + if (pad) + memset(seg->data + seg_size - pad, 0, pad); + + rmpp_mad->rmpp_hdr.rmpp_version = send_wr->mad_agent_priv-> + agent.rmpp_version; + rmpp_mad->rmpp_hdr.rmpp_type = IB_MGMT_RMPP_TYPE_DATA; + ib_set_rmpp_flags(&rmpp_mad->rmpp_hdr, IB_MGMT_RMPP_FLAG_ACTIVE); + + send_wr->cur_seg = container_of(send_wr->rmpp_list.next, + struct ib_rmpp_segment, list); + send_wr->last_ack_seg = send_wr->cur_seg; + return 0; } struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent, @@ -779,32 +870,40 @@ struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent, { struct ib_mad_agent_private *mad_agent_priv; struct ib_mad_send_wr_private *mad_send_wr; - int buf_size; + int pad, message_size, ret, size; void *buf; mad_agent_priv = container_of(mad_agent, struct ib_mad_agent_private, agent); - buf_size = get_buf_length(hdr_len, data_len); + pad = get_pad_size(hdr_len, data_len); + message_size = hdr_len + data_len + pad; if ((!mad_agent->rmpp_version && - (rmpp_active || buf_size > sizeof(struct ib_mad))) || - (!rmpp_active && buf_size > sizeof(struct ib_mad))) + (rmpp_active || message_size > sizeof(struct ib_mad))) || + (!rmpp_active && message_size > sizeof(struct ib_mad))) return ERR_PTR(-EINVAL); - buf = kzalloc(sizeof *mad_send_wr + buf_size, gfp_mask); + size = rmpp_active ? hdr_len : sizeof(struct ib_mad); + buf = kzalloc(sizeof *mad_send_wr + size, gfp_mask); if (!buf) return ERR_PTR(-ENOMEM); - mad_send_wr = buf + buf_size; + mad_send_wr = buf + size; + INIT_LIST_HEAD(&mad_send_wr->rmpp_list); mad_send_wr->send_buf.mad = buf; + mad_send_wr->send_buf.hdr_len = hdr_len; + mad_send_wr->send_buf.data_len = data_len; + mad_send_wr->pad = pad; mad_send_wr->mad_agent_priv = mad_agent_priv; - mad_send_wr->sg_list[0].length = buf_size; + mad_send_wr->sg_list[0].length = hdr_len; mad_send_wr->sg_list[0].lkey = mad_agent->mr->lkey; + mad_send_wr->sg_list[1].length = sizeof(struct ib_mad) - hdr_len; + mad_send_wr->sg_list[1].lkey = mad_agent->mr->lkey; mad_send_wr->send_wr.wr_id = (unsigned long) mad_send_wr; mad_send_wr->send_wr.sg_list = mad_send_wr->sg_list; - mad_send_wr->send_wr.num_sge = 1; + mad_send_wr->send_wr.num_sge = 2; mad_send_wr->send_wr.opcode = IB_WR_SEND; mad_send_wr->send_wr.send_flags = IB_SEND_SIGNALED; mad_send_wr->send_wr.wr.ud.remote_qpn = remote_qpn; @@ -812,13 +911,11 @@ struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent, mad_send_wr->send_wr.wr.ud.pkey_index = pkey_index; if (rmpp_active) { - struct ib_rmpp_mad *rmpp_mad = mad_send_wr->send_buf.mad; - rmpp_mad->rmpp_hdr.paylen_newwin = cpu_to_be32(hdr_len - - IB_MGMT_RMPP_HDR + data_len); - rmpp_mad->rmpp_hdr.rmpp_version = mad_agent->rmpp_version; - rmpp_mad->rmpp_hdr.rmpp_type = IB_MGMT_RMPP_TYPE_DATA; - ib_set_rmpp_flags(&rmpp_mad->rmpp_hdr, - IB_MGMT_RMPP_FLAG_ACTIVE); + ret = alloc_send_rmpp_list(mad_send_wr, gfp_mask); + if (ret) { + kfree(buf); + return ERR_PTR(ret); + } } mad_send_wr->send_buf.mad_agent = mad_agent; @@ -827,16 +924,80 @@ struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent, } EXPORT_SYMBOL(ib_create_send_mad); +int ib_get_mad_data_offset(u8 mgmt_class) +{ + if (mgmt_class == IB_MGMT_CLASS_SUBN_ADM) + return IB_MGMT_SA_HDR; + else if ((mgmt_class == IB_MGMT_CLASS_DEVICE_MGMT) || + (mgmt_class == IB_MGMT_CLASS_DEVICE_ADM) || + (mgmt_class == IB_MGMT_CLASS_BIS)) + return IB_MGMT_DEVICE_HDR; + else if ((mgmt_class >= IB_MGMT_CLASS_VENDOR_RANGE2_START) && + (mgmt_class <= IB_MGMT_CLASS_VENDOR_RANGE2_END)) + return IB_MGMT_VENDOR_HDR; + else + return IB_MGMT_MAD_HDR; +} +EXPORT_SYMBOL(ib_get_mad_data_offset); + +int ib_is_mad_class_rmpp(u8 mgmt_class) +{ + if ((mgmt_class == IB_MGMT_CLASS_SUBN_ADM) || + (mgmt_class == IB_MGMT_CLASS_DEVICE_MGMT) || + (mgmt_class == IB_MGMT_CLASS_DEVICE_ADM) || + (mgmt_class == IB_MGMT_CLASS_BIS) || + ((mgmt_class >= IB_MGMT_CLASS_VENDOR_RANGE2_START) && + (mgmt_class <= IB_MGMT_CLASS_VENDOR_RANGE2_END))) + return 1; + return 0; +} +EXPORT_SYMBOL(ib_is_mad_class_rmpp); + +void *ib_get_rmpp_segment(struct ib_mad_send_buf *send_buf, int seg_num) +{ + struct ib_mad_send_wr_private *mad_send_wr; + struct list_head *list; + + mad_send_wr = container_of(send_buf, struct ib_mad_send_wr_private, + send_buf); + list = &mad_send_wr->cur_seg->list; + + if (mad_send_wr->cur_seg->num < seg_num) { + list_for_each_entry(mad_send_wr->cur_seg, list, list) + if (mad_send_wr->cur_seg->num == seg_num) + break; + } else if (mad_send_wr->cur_seg->num > seg_num) { + list_for_each_entry_reverse(mad_send_wr->cur_seg, list, list) + if (mad_send_wr->cur_seg->num == seg_num) + break; + } + return mad_send_wr->cur_seg->data; +} +EXPORT_SYMBOL(ib_get_rmpp_segment); + +static inline void *ib_get_payload(struct ib_mad_send_wr_private *mad_send_wr) +{ + if (mad_send_wr->send_buf.seg_count) + return ib_get_rmpp_segment(&mad_send_wr->send_buf, + mad_send_wr->seg_num); + else + return mad_send_wr->send_buf.mad + + mad_send_wr->send_buf.hdr_len; +} + void ib_free_send_mad(struct ib_mad_send_buf *send_buf) { struct ib_mad_agent_private *mad_agent_priv; + struct ib_mad_send_wr_private *mad_send_wr; mad_agent_priv = container_of(send_buf->mad_agent, struct ib_mad_agent_private, agent); - kfree(send_buf->mad); + mad_send_wr = container_of(send_buf, struct ib_mad_send_wr_private, + send_buf); - if (atomic_dec_and_test(&mad_agent_priv->refcount)) - wake_up(&mad_agent_priv->wait); + free_send_rmpp_list(mad_send_wr); + kfree(send_buf->mad); + deref_mad_agent(mad_agent_priv); } EXPORT_SYMBOL(ib_free_send_mad); @@ -857,10 +1018,26 @@ int ib_send_mad(struct ib_mad_send_wr_private *mad_send_wr) mad_agent = mad_send_wr->send_buf.mad_agent; sge = mad_send_wr->sg_list; - sge->addr = dma_map_single(mad_agent->device->dma_device, - mad_send_wr->send_buf.mad, sge->length, - DMA_TO_DEVICE); - pci_unmap_addr_set(mad_send_wr, mapping, sge->addr); + sge[0].addr = ib_dma_map_single(mad_agent->device, + mad_send_wr->send_buf.mad, + sge[0].length, + DMA_TO_DEVICE); + if (unlikely(ib_dma_mapping_error(mad_agent->device, sge[0].addr))) + return -ENOMEM; + + mad_send_wr->header_mapping = sge[0].addr; + + sge[1].addr = ib_dma_map_single(mad_agent->device, + ib_get_payload(mad_send_wr), + sge[1].length, + DMA_TO_DEVICE); + if (unlikely(ib_dma_mapping_error(mad_agent->device, sge[1].addr))) { + ib_dma_unmap_single(mad_agent->device, + mad_send_wr->header_mapping, + sge[0].length, DMA_TO_DEVICE); + return -ENOMEM; + } + mad_send_wr->payload_mapping = sge[1].addr; spin_lock_irqsave(&qp_info->send_queue.lock, flags); if (qp_info->send_queue.count < qp_info->send_queue.max_active) { @@ -877,11 +1054,14 @@ int ib_send_mad(struct ib_mad_send_wr_private *mad_send_wr) list_add_tail(&mad_send_wr->mad_list.list, list); } spin_unlock_irqrestore(&qp_info->send_queue.lock, flags); - if (ret) - dma_unmap_single(mad_agent->device->dma_device, - pci_unmap_addr(mad_send_wr, mapping), - sge->length, DMA_TO_DEVICE); - + if (ret) { + ib_dma_unmap_single(mad_agent->device, + mad_send_wr->header_mapping, + sge[0].length, DMA_TO_DEVICE); + ib_dma_unmap_single(mad_agent->device, + mad_send_wr->payload_mapping, + sge[1].length, DMA_TO_DEVICE); + } return ret; } @@ -913,6 +1093,13 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf, goto error; } + if (!ib_is_mad_class_rmpp(((struct ib_mad_hdr *) send_buf->mad)->mgmt_class)) { + if (mad_agent_priv->agent.rmpp_version) { + ret = -EINVAL; + goto error; + } + } + /* * Save pointer to next work request to post in case the * current one completes, and the user modifies the work @@ -934,7 +1121,9 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf, mad_send_wr->tid = ((struct ib_mad_hdr *) send_buf->mad)->tid; /* Timeout will be updated after send completes */ mad_send_wr->timeout = msecs_to_jiffies(send_buf->timeout_ms); - mad_send_wr->retries = send_buf->retries; + mad_send_wr->max_retries = send_buf->retries; + mad_send_wr->retries_left = send_buf->retries; + send_buf->retries = 0; /* Reference for work request to QP + response */ mad_send_wr->refcount = 1 + (mad_send_wr->timeout > 0); mad_send_wr->status = IB_WC_SUCCESS; @@ -1020,10 +1209,7 @@ static int method_in_use(struct ib_mad_mgmt_method_table **method, { int i; - for (i = find_first_bit(mad_reg_req->method_mask, IB_MGMT_MAX_METHODS); - i < IB_MGMT_MAX_METHODS; - i = find_next_bit(mad_reg_req->method_mask, IB_MGMT_MAX_METHODS, - 1+i)) { + for_each_set_bit(i, mad_reg_req->method_mask, IB_MGMT_MAX_METHODS) { if ((*method)->agent[i]) { printk(KERN_ERR PFX "Method %d already in use\n", i); return -EINVAL; @@ -1087,8 +1273,8 @@ static int find_vendor_oui(struct ib_mad_mgmt_vendor_class *vendor_class, int i; for (i = 0; i < MAX_MGMT_OUI; i++) - /* Is there matching OUI for this vendor class ? */ - if (!memcmp(vendor_class->oui[i], oui, 3)) + /* Is there matching OUI for this vendor class ? */ + if (!memcmp(vendor_class->oui[i], oui, 3)) return i; return -1; @@ -1157,13 +1343,9 @@ static int add_nonoui_reg_req(struct ib_mad_reg_req *mad_reg_req, goto error3; /* Finally, add in methods being registered */ - for (i = find_first_bit(mad_reg_req->method_mask, - IB_MGMT_MAX_METHODS); - i < IB_MGMT_MAX_METHODS; - i = find_next_bit(mad_reg_req->method_mask, IB_MGMT_MAX_METHODS, - 1+i)) { + for_each_set_bit(i, mad_reg_req->method_mask, IB_MGMT_MAX_METHODS) (*method)->agent[i] = agent_priv; - } + return 0; error3: @@ -1256,13 +1438,9 @@ check_in_use: goto error4; /* Finally, add in methods being registered */ - for (i = find_first_bit(mad_reg_req->method_mask, - IB_MGMT_MAX_METHODS); - i < IB_MGMT_MAX_METHODS; - i = find_next_bit(mad_reg_req->method_mask, IB_MGMT_MAX_METHODS, - 1+i)) { + for_each_set_bit(i, mad_reg_req->method_mask, IB_MGMT_MAX_METHODS) (*method)->agent[i] = agent_priv; - } + return 0; error4: @@ -1394,7 +1572,7 @@ find_mad_agent(struct ib_mad_port_private *port_priv, unsigned long flags; spin_lock_irqsave(&port_priv->reg_lock, flags); - if (response_mad(mad)) { + if (ib_response_mad(mad)) { u32 hi_tid; struct ib_mad_agent_private *entry; @@ -1428,6 +1606,9 @@ find_mad_agent(struct ib_mad_port_private *port_priv, mad->mad_hdr.class_version].class; if (!class) goto out; + if (convert_mgmt_class(mad->mad_hdr.mgmt_class) >= + IB_MGMT_MAX_METHODS) + goto out; method = class->method_table[convert_mgmt_class( mad->mad_hdr.mgmt_class)]; if (method) @@ -1509,29 +1690,107 @@ static int is_data_mad(struct ib_mad_agent_private *mad_agent_priv, (rmpp_mad->rmpp_hdr.rmpp_type == IB_MGMT_RMPP_TYPE_DATA); } +static inline int rcv_has_same_class(struct ib_mad_send_wr_private *wr, + struct ib_mad_recv_wc *rwc) +{ + return ((struct ib_mad *)(wr->send_buf.mad))->mad_hdr.mgmt_class == + rwc->recv_buf.mad->mad_hdr.mgmt_class; +} + +static inline int rcv_has_same_gid(struct ib_mad_agent_private *mad_agent_priv, + struct ib_mad_send_wr_private *wr, + struct ib_mad_recv_wc *rwc ) +{ + struct ib_ah_attr attr; + u8 send_resp, rcv_resp; + union ib_gid sgid; + struct ib_device *device = mad_agent_priv->agent.device; + u8 port_num = mad_agent_priv->agent.port_num; + u8 lmc; + + send_resp = ib_response_mad((struct ib_mad *)wr->send_buf.mad); + rcv_resp = ib_response_mad(rwc->recv_buf.mad); + + if (send_resp == rcv_resp) + /* both requests, or both responses. GIDs different */ + return 0; + + if (ib_query_ah(wr->send_buf.ah, &attr)) + /* Assume not equal, to avoid false positives. */ + return 0; + + if (!!(attr.ah_flags & IB_AH_GRH) != + !!(rwc->wc->wc_flags & IB_WC_GRH)) + /* one has GID, other does not. Assume different */ + return 0; + + if (!send_resp && rcv_resp) { + /* is request/response. */ + if (!(attr.ah_flags & IB_AH_GRH)) { + if (ib_get_cached_lmc(device, port_num, &lmc)) + return 0; + return (!lmc || !((attr.src_path_bits ^ + rwc->wc->dlid_path_bits) & + ((1 << lmc) - 1))); + } else { + if (ib_get_cached_gid(device, port_num, + attr.grh.sgid_index, &sgid)) + return 0; + return !memcmp(sgid.raw, rwc->recv_buf.grh->dgid.raw, + 16); + } + } + + if (!(attr.ah_flags & IB_AH_GRH)) + return attr.dlid == rwc->wc->slid; + else + return !memcmp(attr.grh.dgid.raw, rwc->recv_buf.grh->sgid.raw, + 16); +} + +static inline int is_direct(u8 class) +{ + return (class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE); +} + struct ib_mad_send_wr_private* -ib_find_send_mad(struct ib_mad_agent_private *mad_agent_priv, __be64 tid) +ib_find_send_mad(struct ib_mad_agent_private *mad_agent_priv, + struct ib_mad_recv_wc *wc) { - struct ib_mad_send_wr_private *mad_send_wr; + struct ib_mad_send_wr_private *wr; + struct ib_mad *mad; - list_for_each_entry(mad_send_wr, &mad_agent_priv->wait_list, - agent_list) { - if (mad_send_wr->tid == tid) - return mad_send_wr; + mad = (struct ib_mad *)wc->recv_buf.mad; + + list_for_each_entry(wr, &mad_agent_priv->wait_list, agent_list) { + if ((wr->tid == mad->mad_hdr.tid) && + rcv_has_same_class(wr, wc) && + /* + * Don't check GID for direct routed MADs. + * These might have permissive LIDs. + */ + (is_direct(wc->recv_buf.mad->mad_hdr.mgmt_class) || + rcv_has_same_gid(mad_agent_priv, wr, wc))) + return (wr->status == IB_WC_SUCCESS) ? wr : NULL; } /* * It's possible to receive the response before we've * been notified that the send has completed */ - list_for_each_entry(mad_send_wr, &mad_agent_priv->send_list, - agent_list) { - if (is_data_mad(mad_agent_priv, mad_send_wr->send_buf.mad) && - mad_send_wr->tid == tid && mad_send_wr->timeout) { + list_for_each_entry(wr, &mad_agent_priv->send_list, agent_list) { + if (is_data_mad(mad_agent_priv, wr->send_buf.mad) && + wr->tid == mad->mad_hdr.tid && + wr->timeout && + rcv_has_same_class(wr, wc) && + /* + * Don't check GID for direct routed MADs. + * These might have permissive LIDs. + */ + (is_direct(wc->recv_buf.mad->mad_hdr.mgmt_class) || + rcv_has_same_gid(mad_agent_priv, wr, wc))) /* Verify request has not been canceled */ - return (mad_send_wr->status == IB_WC_SUCCESS) ? - mad_send_wr : NULL; - } + return (wr->status == IB_WC_SUCCESS) ? wr : NULL; } return NULL; } @@ -1539,11 +1798,9 @@ ib_find_send_mad(struct ib_mad_agent_private *mad_agent_priv, __be64 tid) void ib_mark_mad_done(struct ib_mad_send_wr_private *mad_send_wr) { mad_send_wr->timeout = 0; - if (mad_send_wr->refcount == 1) { - list_del(&mad_send_wr->agent_list); - list_add_tail(&mad_send_wr->agent_list, + if (mad_send_wr->refcount == 1) + list_move_tail(&mad_send_wr->agent_list, &mad_send_wr->mad_agent_priv->done_list); - } } static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, @@ -1552,7 +1809,6 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, struct ib_mad_send_wr_private *mad_send_wr; struct ib_mad_send_wc mad_send_wc; unsigned long flags; - __be64 tid; INIT_LIST_HEAD(&mad_recv_wc->rmpp_list); list_add(&mad_recv_wc->recv_buf.list, &mad_recv_wc->rmpp_list); @@ -1560,22 +1816,19 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, mad_recv_wc = ib_process_rmpp_recv_wc(mad_agent_priv, mad_recv_wc); if (!mad_recv_wc) { - if (atomic_dec_and_test(&mad_agent_priv->refcount)) - wake_up(&mad_agent_priv->wait); + deref_mad_agent(mad_agent_priv); return; } } /* Complete corresponding request */ - if (response_mad(mad_recv_wc->recv_buf.mad)) { - tid = mad_recv_wc->recv_buf.mad->mad_hdr.tid; + if (ib_response_mad(mad_recv_wc->recv_buf.mad)) { spin_lock_irqsave(&mad_agent_priv->lock, flags); - mad_send_wr = ib_find_send_mad(mad_agent_priv, tid); + mad_send_wr = ib_find_send_mad(mad_agent_priv, mad_recv_wc); if (!mad_send_wr) { spin_unlock_irqrestore(&mad_agent_priv->lock, flags); ib_free_recv_mad(mad_recv_wc); - if (atomic_dec_and_test(&mad_agent_priv->refcount)) - wake_up(&mad_agent_priv->wait); + deref_mad_agent(mad_agent_priv); return; } ib_mark_mad_done(mad_send_wr); @@ -1594,24 +1847,40 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, } else { mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent, mad_recv_wc); - if (atomic_dec_and_test(&mad_agent_priv->refcount)) - wake_up(&mad_agent_priv->wait); + deref_mad_agent(mad_agent_priv); } } +static bool generate_unmatched_resp(struct ib_mad_private *recv, + struct ib_mad_private *response) +{ + if (recv->mad.mad.mad_hdr.method == IB_MGMT_METHOD_GET || + recv->mad.mad.mad_hdr.method == IB_MGMT_METHOD_SET) { + memcpy(response, recv, sizeof *response); + response->header.recv_wc.wc = &response->header.wc; + response->header.recv_wc.recv_buf.mad = &response->mad.mad; + response->header.recv_wc.recv_buf.grh = &response->grh; + response->mad.mad.mad_hdr.method = IB_MGMT_METHOD_GET_RESP; + response->mad.mad.mad_hdr.status = + cpu_to_be16(IB_MGMT_MAD_STATUS_UNSUPPORTED_METHOD_ATTRIB); + if (recv->mad.mad.mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) + response->mad.mad.mad_hdr.status |= IB_SMP_DIRECTION; + + return true; + } else { + return false; + } +} static void ib_mad_recv_done_handler(struct ib_mad_port_private *port_priv, struct ib_wc *wc) { struct ib_mad_qp_info *qp_info; struct ib_mad_private_header *mad_priv_hdr; - struct ib_mad_private *recv, *response; + struct ib_mad_private *recv, *response = NULL; struct ib_mad_list_head *mad_list; struct ib_mad_agent_private *mad_agent; - - response = kmem_cache_alloc(ib_mad_cache, GFP_KERNEL); - if (!response) - printk(KERN_ERR PFX "ib_mad_recv_done_handler no memory " - "for response buffer\n"); + int port_num; + int ret = IB_MAD_RESULT_SUCCESS; mad_list = (struct ib_mad_list_head *)(unsigned long)wc->wr_id; qp_info = mad_list->mad_queue->qp_info; @@ -1620,11 +1889,11 @@ static void ib_mad_recv_done_handler(struct ib_mad_port_private *port_priv, mad_priv_hdr = container_of(mad_list, struct ib_mad_private_header, mad_list); recv = container_of(mad_priv_hdr, struct ib_mad_private, header); - dma_unmap_single(port_priv->device->dma_device, - pci_unmap_addr(&recv->header, mapping), - sizeof(struct ib_mad_private) - - sizeof(struct ib_mad_private_header), - DMA_FROM_DEVICE); + ib_dma_unmap_single(port_priv->device, + recv->header.mapping, + sizeof(struct ib_mad_private) - + sizeof(struct ib_mad_private_header), + DMA_FROM_DEVICE); /* Setup MAD receive work completion from "normal" work completion */ recv->header.wc = *wc; @@ -1640,39 +1909,61 @@ static void ib_mad_recv_done_handler(struct ib_mad_port_private *port_priv, if (!validate_mad(&recv->mad.mad, qp_info->qp->qp_num)) goto out; + response = kmem_cache_alloc(ib_mad_cache, GFP_KERNEL); + if (!response) { + printk(KERN_ERR PFX "ib_mad_recv_done_handler no memory " + "for response buffer\n"); + goto out; + } + + if (port_priv->device->node_type == RDMA_NODE_IB_SWITCH) + port_num = wc->port_num; + else + port_num = port_priv->port_num; + if (recv->mad.mad.mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) { - if (!smi_handle_dr_smp_recv(&recv->mad.smp, - port_priv->device->node_type, - port_priv->port_num, - port_priv->device->phys_port_cnt)) + enum smi_forward_action retsmi; + + if (smi_handle_dr_smp_recv(&recv->mad.smp, + port_priv->device->node_type, + port_num, + port_priv->device->phys_port_cnt) == + IB_SMI_DISCARD) goto out; - if (!smi_check_forward_dr_smp(&recv->mad.smp)) + + retsmi = smi_check_forward_dr_smp(&recv->mad.smp); + if (retsmi == IB_SMI_LOCAL) goto local; - if (!smi_handle_dr_smp_send(&recv->mad.smp, - port_priv->device->node_type, - port_priv->port_num)) - goto out; - if (!smi_check_local_dr_smp(&recv->mad.smp, + + if (retsmi == IB_SMI_SEND) { /* don't forward */ + if (smi_handle_dr_smp_send(&recv->mad.smp, + port_priv->device->node_type, + port_num) == IB_SMI_DISCARD) + goto out; + + if (smi_check_local_smp(&recv->mad.smp, port_priv->device) == IB_SMI_DISCARD) + goto out; + } else if (port_priv->device->node_type == RDMA_NODE_IB_SWITCH) { + /* forward case for switches */ + memcpy(response, recv, sizeof(*response)); + response->header.recv_wc.wc = &response->header.wc; + response->header.recv_wc.recv_buf.mad = &response->mad.mad; + response->header.recv_wc.recv_buf.grh = &response->grh; + + agent_send_response(&response->mad.mad, + &response->grh, wc, port_priv->device, - port_priv->port_num)) + smi_get_fwd_port(&recv->mad.smp), + qp_info->qp->qp_num); + goto out; + } } local: /* Give driver "right of first refusal" on incoming MAD */ if (port_priv->device->process_mad) { - int ret; - - if (!response) { - printk(KERN_ERR PFX "No memory for response MAD\n"); - /* - * Is it better to assume that - * it wouldn't be processed ? - */ - goto out; - } - ret = port_priv->device->process_mad(port_priv->device, 0, port_priv->port_num, wc, &recv->grh, @@ -1685,7 +1976,7 @@ local: agent_send_response(&response->mad.mad, &recv->grh, wc, port_priv->device, - port_priv->port_num, + port_num, qp_info->qp->qp_num); goto out; } @@ -1700,6 +1991,10 @@ local: * or via recv_handler in ib_mad_complete_recv() */ recv = NULL; + } else if ((ret & IB_MAD_RESULT_SUCCESS) && + generate_unmatched_resp(recv, response)) { + agent_send_response(&response->mad.mad, &recv->grh, wc, + port_priv->device, port_num, qp_info->qp->qp_num); } out: @@ -1727,13 +2022,11 @@ static void adjust_timeout(struct ib_mad_agent_private *mad_agent_priv) if (time_after(mad_agent_priv->timeout, mad_send_wr->timeout)) { mad_agent_priv->timeout = mad_send_wr->timeout; - cancel_delayed_work(&mad_agent_priv->timed_work); delay = mad_send_wr->timeout - jiffies; if ((long)delay <= 0) delay = 1; - queue_delayed_work(mad_agent_priv->qp_info-> - port_priv->wq, - &mad_agent_priv->timed_work, delay); + mod_delayed_work(mad_agent_priv->qp_info->port_priv->wq, + &mad_agent_priv->timed_work, delay); } } } @@ -1766,11 +2059,9 @@ static void wait_for_response(struct ib_mad_send_wr_private *mad_send_wr) list_add(&mad_send_wr->agent_list, list_item); /* Reschedule a work item if we have a shorter timeout */ - if (mad_agent_priv->wait_list.next == &mad_send_wr->agent_list) { - cancel_delayed_work(&mad_agent_priv->timed_work); - queue_delayed_work(mad_agent_priv->qp_info->port_priv->wq, - &mad_agent_priv->timed_work, delay); - } + if (mad_agent_priv->wait_list.next == &mad_send_wr->agent_list) + mod_delayed_work(mad_agent_priv->qp_info->port_priv->wq, + &mad_agent_priv->timed_work, delay); } void ib_reset_mad_timeout(struct ib_mad_send_wr_private *mad_send_wr, @@ -1827,8 +2118,7 @@ void ib_mad_complete_send_wr(struct ib_mad_send_wr_private *mad_send_wr, mad_send_wc); /* Release reference on agent taken when sending */ - if (atomic_dec_and_test(&mad_agent_priv->refcount)) - wake_up(&mad_agent_priv->wait); + deref_mad_agent(mad_agent_priv); return; done: spin_unlock_irqrestore(&mad_agent_priv->lock, flags); @@ -1853,9 +2143,12 @@ static void ib_mad_send_done_handler(struct ib_mad_port_private *port_priv, qp_info = send_queue->qp_info; retry: - dma_unmap_single(mad_send_wr->send_buf.mad_agent->device->dma_device, - pci_unmap_addr(mad_send_wr, mapping), - mad_send_wr->sg_list[0].length, DMA_TO_DEVICE); + ib_dma_unmap_single(mad_send_wr->send_buf.mad_agent->device, + mad_send_wr->header_mapping, + mad_send_wr->sg_list[0].length, DMA_TO_DEVICE); + ib_dma_unmap_single(mad_send_wr->send_buf.mad_agent->device, + mad_send_wr->payload_mapping, + mad_send_wr->sg_list[1].length, DMA_TO_DEVICE); queued_send_wr = NULL; spin_lock_irqsave(&send_queue->lock, flags); list_del(&mad_list->list); @@ -1867,8 +2160,7 @@ retry: queued_send_wr = container_of(mad_list, struct ib_mad_send_wr_private, mad_list); - list_del(&mad_list->list); - list_add_tail(&mad_list->list, &send_queue->list); + list_move_tail(&mad_list->list, &send_queue->list); } spin_unlock_irqrestore(&send_queue->lock, flags); @@ -1968,12 +2260,12 @@ static void mad_error_handler(struct ib_mad_port_private *port_priv, /* * IB MAD completion callback */ -static void ib_mad_completion_handler(void *data) +static void ib_mad_completion_handler(struct work_struct *work) { struct ib_mad_port_private *port_priv; struct ib_wc wc; - port_priv = (struct ib_mad_port_private *)data; + port_priv = container_of(work, struct ib_mad_port_private, work); ib_req_notify_cq(port_priv->cq, IB_CQ_NEXT_COMP); while (ib_poll_cq(port_priv->cq, 1, &wc) == 1) { @@ -2007,15 +2299,13 @@ static void cancel_mads(struct ib_mad_agent_private *mad_agent_priv) list_for_each_entry_safe(mad_send_wr, temp_mad_send_wr, &mad_agent_priv->send_list, agent_list) { if (mad_send_wr->status == IB_WC_SUCCESS) { - mad_send_wr->status = IB_WC_WR_FLUSH_ERR; + mad_send_wr->status = IB_WC_WR_FLUSH_ERR; mad_send_wr->refcount -= (mad_send_wr->timeout > 0); } } /* Empty wait list to prevent receives from finding a request */ list_splice_init(&mad_agent_priv->wait_list, &cancel_list); - /* Empty local completion list as well */ - list_splice_init(&mad_agent_priv->local_list, &cancel_list); spin_unlock_irqrestore(&mad_agent_priv->lock, flags); /* Report all cancelled requests */ @@ -2094,37 +2384,41 @@ void ib_cancel_mad(struct ib_mad_agent *mad_agent, } EXPORT_SYMBOL(ib_cancel_mad); -static void local_completions(void *data) +static void local_completions(struct work_struct *work) { struct ib_mad_agent_private *mad_agent_priv; struct ib_mad_local_private *local; struct ib_mad_agent_private *recv_mad_agent; unsigned long flags; - int recv = 0; + int free_mad; struct ib_wc wc; struct ib_mad_send_wc mad_send_wc; - mad_agent_priv = (struct ib_mad_agent_private *)data; + mad_agent_priv = + container_of(work, struct ib_mad_agent_private, local_work); spin_lock_irqsave(&mad_agent_priv->lock, flags); while (!list_empty(&mad_agent_priv->local_list)) { local = list_entry(mad_agent_priv->local_list.next, struct ib_mad_local_private, completion_list); + list_del(&local->completion_list); spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + free_mad = 0; if (local->mad_priv) { recv_mad_agent = local->recv_mad_agent; if (!recv_mad_agent) { printk(KERN_ERR PFX "No receive MAD agent for local completion\n"); + free_mad = 1; goto local_send_completion; } - recv = 1; /* * Defined behavior is to complete response * before request */ - build_smp_wc((unsigned long) local->mad_send_wr, + build_smp_wc(recv_mad_agent->agent.qp, + (unsigned long) local->mad_send_wr, be16_to_cpu(IB_LID_PERMISSIVE), 0, recv_mad_agent->agent.port_num, &wc); @@ -2162,9 +2456,8 @@ local_send_completion: &mad_send_wc); spin_lock_irqsave(&mad_agent_priv->lock, flags); - list_del(&local->completion_list); atomic_dec(&mad_agent_priv->refcount); - if (!recv) + if (free_mad) kmem_cache_free(ib_mad_cache, local->mad_priv); kfree(local); } @@ -2175,9 +2468,12 @@ static int retry_send(struct ib_mad_send_wr_private *mad_send_wr) { int ret; - if (!mad_send_wr->retries--) + if (!mad_send_wr->retries_left) return -ETIMEDOUT; + mad_send_wr->retries_left--; + mad_send_wr->send_buf.retries++; + mad_send_wr->timeout = msecs_to_jiffies(mad_send_wr->send_buf.timeout_ms); if (mad_send_wr->mad_agent_priv->agent.rmpp_version) { @@ -2204,14 +2500,15 @@ static int retry_send(struct ib_mad_send_wr_private *mad_send_wr) return ret; } -static void timeout_sends(void *data) +static void timeout_sends(struct work_struct *work) { struct ib_mad_agent_private *mad_agent_priv; struct ib_mad_send_wr_private *mad_send_wr; struct ib_mad_send_wc mad_send_wc; unsigned long flags, delay; - mad_agent_priv = (struct ib_mad_agent_private *)data; + mad_agent_priv = container_of(work, struct ib_mad_agent_private, + timed_work.work); mad_send_wc.vendor_err = 0; spin_lock_irqsave(&mad_agent_priv->lock, flags); @@ -2254,8 +2551,12 @@ static void timeout_sends(void *data) static void ib_mad_thread_completion_handler(struct ib_cq *cq, void *arg) { struct ib_mad_port_private *port_priv = cq->cq_context; + unsigned long flags; - queue_work(port_priv->wq, &port_priv->work); + spin_lock_irqsave(&ib_mad_port_list_lock, flags); + if (!list_empty(&port_priv->port_list)) + queue_work(port_priv->wq, &port_priv->work); + spin_unlock_irqrestore(&ib_mad_port_list_lock, flags); } /* @@ -2293,13 +2594,17 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info, break; } } - sg_list.addr = dma_map_single(qp_info->port_priv-> - device->dma_device, - &mad_priv->grh, - sizeof *mad_priv - - sizeof mad_priv->header, - DMA_FROM_DEVICE); - pci_unmap_addr_set(&mad_priv->header, mapping, sg_list.addr); + sg_list.addr = ib_dma_map_single(qp_info->port_priv->device, + &mad_priv->grh, + sizeof *mad_priv - + sizeof mad_priv->header, + DMA_FROM_DEVICE); + if (unlikely(ib_dma_mapping_error(qp_info->port_priv->device, + sg_list.addr))) { + ret = -ENOMEM; + break; + } + mad_priv->header.mapping = sg_list.addr; recv_wr.wr_id = (unsigned long)&mad_priv->header.mad_list; mad_priv->header.mad_list.mad_queue = recv_queue; @@ -2314,12 +2619,11 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info, list_del(&mad_priv->header.mad_list.list); recv_queue->count--; spin_unlock_irqrestore(&recv_queue->lock, flags); - dma_unmap_single(qp_info->port_priv->device->dma_device, - pci_unmap_addr(&mad_priv->header, - mapping), - sizeof *mad_priv - - sizeof mad_priv->header, - DMA_FROM_DEVICE); + ib_dma_unmap_single(qp_info->port_priv->device, + mad_priv->header.mapping, + sizeof *mad_priv - + sizeof mad_priv->header, + DMA_FROM_DEVICE); kmem_cache_free(ib_mad_cache, mad_priv); printk(KERN_ERR PFX "ib_post_recv failed: %d\n", ret); break; @@ -2338,6 +2642,9 @@ static void cleanup_recv_queue(struct ib_mad_qp_info *qp_info) struct ib_mad_private *recv; struct ib_mad_list_head *mad_list; + if (!qp_info->qp) + return; + while (!list_empty(&qp_info->recv_queue.list)) { mad_list = list_entry(qp_info->recv_queue.list.next, @@ -2351,11 +2658,11 @@ static void cleanup_recv_queue(struct ib_mad_qp_info *qp_info) /* Remove from posted receive MAD list */ list_del(&mad_list->list); - dma_unmap_single(qp_info->port_priv->device->dma_device, - pci_unmap_addr(&recv->header, mapping), - sizeof(struct ib_mad_private) - - sizeof(struct ib_mad_private_header), - DMA_FROM_DEVICE); + ib_dma_unmap_single(qp_info->port_priv->device, + recv->header.mapping, + sizeof(struct ib_mad_private) - + sizeof(struct ib_mad_private_header), + DMA_FROM_DEVICE); kmem_cache_free(ib_mad_cache, recv); } @@ -2370,21 +2677,30 @@ static int ib_mad_port_start(struct ib_mad_port_private *port_priv) int ret, i; struct ib_qp_attr *attr; struct ib_qp *qp; + u16 pkey_index; attr = kmalloc(sizeof *attr, GFP_KERNEL); - if (!attr) { + if (!attr) { printk(KERN_ERR PFX "Couldn't kmalloc ib_qp_attr\n"); return -ENOMEM; } + ret = ib_find_pkey(port_priv->device, port_priv->port_num, + IB_DEFAULT_PKEY_FULL, &pkey_index); + if (ret) + pkey_index = 0; + for (i = 0; i < IB_MAD_QPS_CORE; i++) { qp = port_priv->qp_info[i].qp; + if (!qp) + continue; + /* * PKey index for QP1 is irrelevant but * one is needed for the Reset to Init transition */ attr->qp_state = IB_QPS_INIT; - attr->pkey_index = 0; + attr->pkey_index = pkey_index; attr->qkey = (qp->qp_num == 0) ? 0 : IB_QP1_QKEY; ret = ib_modify_qp(qp, attr, IB_QP_STATE | IB_QP_PKEY_INDEX | IB_QP_QKEY); @@ -2420,6 +2736,9 @@ static int ib_mad_port_start(struct ib_mad_port_private *port_priv) } for (i = 0; i < IB_MAD_QPS_CORE; i++) { + if (!port_priv->qp_info[i].qp) + continue; + ret = ib_mad_post_receive_mads(&port_priv->qp_info[i], NULL); if (ret) { printk(KERN_ERR PFX "Couldn't post receive WRs\n"); @@ -2472,8 +2791,8 @@ static int create_mad_qp(struct ib_mad_qp_info *qp_info, qp_init_attr.send_cq = qp_info->port_priv->cq; qp_init_attr.recv_cq = qp_info->port_priv->cq; qp_init_attr.sq_sig_type = IB_SIGNAL_ALL_WR; - qp_init_attr.cap.max_send_wr = IB_MAD_QP_SEND_SIZE; - qp_init_attr.cap.max_recv_wr = IB_MAD_QP_RECV_SIZE; + qp_init_attr.cap.max_send_wr = mad_sendq_size; + qp_init_attr.cap.max_recv_wr = mad_recvq_size; qp_init_attr.cap.max_send_sge = IB_MAD_SEND_REQ_MAX_SG; qp_init_attr.cap.max_recv_sge = IB_MAD_RECV_REQ_MAX_SG; qp_init_attr.qp_type = qp_type; @@ -2488,8 +2807,8 @@ static int create_mad_qp(struct ib_mad_qp_info *qp_info, goto error; } /* Use minimum queue sizes unless the CQ is resized */ - qp_info->send_queue.max_active = IB_MAD_QP_SEND_SIZE; - qp_info->recv_queue.max_active = IB_MAD_QP_RECV_SIZE; + qp_info->send_queue.max_active = mad_sendq_size; + qp_info->recv_queue.max_active = mad_recvq_size; return 0; error: @@ -2498,6 +2817,9 @@ error: static void destroy_mad_qp(struct ib_mad_qp_info *qp_info) { + if (!qp_info->qp) + return; + ib_destroy_qp(qp_info->qp); kfree(qp_info->snoop_table); } @@ -2513,6 +2835,7 @@ static int ib_mad_port_open(struct ib_device *device, struct ib_mad_port_private *port_priv; unsigned long flags; char name[sizeof "ib_mad123"]; + int has_smi; /* Create new device info */ port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL); @@ -2528,10 +2851,14 @@ static int ib_mad_port_open(struct ib_device *device, init_mad_qp(port_priv, &port_priv->qp_info[0]); init_mad_qp(port_priv, &port_priv->qp_info[1]); - cq_size = (IB_MAD_QP_SEND_SIZE + IB_MAD_QP_RECV_SIZE) * 2; + cq_size = mad_sendq_size + mad_recvq_size; + has_smi = rdma_port_get_link_layer(device, port_num) == IB_LINK_LAYER_INFINIBAND; + if (has_smi) + cq_size *= 2; + port_priv->cq = ib_create_cq(port_priv->device, ib_mad_thread_completion_handler, - NULL, port_priv, cq_size); + NULL, port_priv, cq_size, 0); if (IS_ERR(port_priv->cq)) { printk(KERN_ERR PFX "Couldn't create ib_mad CQ\n"); ret = PTR_ERR(port_priv->cq); @@ -2552,9 +2879,11 @@ static int ib_mad_port_open(struct ib_device *device, goto error5; } - ret = create_mad_qp(&port_priv->qp_info[0], IB_QPT_SMI); - if (ret) - goto error6; + if (has_smi) { + ret = create_mad_qp(&port_priv->qp_info[0], IB_QPT_SMI); + if (ret) + goto error6; + } ret = create_mad_qp(&port_priv->qp_info[1], IB_QPT_GSI); if (ret) goto error7; @@ -2565,7 +2894,11 @@ static int ib_mad_port_open(struct ib_device *device, ret = -ENOMEM; goto error8; } - INIT_WORK(&port_priv->work, ib_mad_completion_handler, port_priv); + INIT_WORK(&port_priv->work, ib_mad_completion_handler); + + spin_lock_irqsave(&ib_mad_port_list_lock, flags); + list_add_tail(&port_priv->port_list, &ib_mad_port_list); + spin_unlock_irqrestore(&ib_mad_port_list_lock, flags); ret = ib_mad_port_start(port_priv); if (ret) { @@ -2573,12 +2906,13 @@ static int ib_mad_port_open(struct ib_device *device, goto error9; } - spin_lock_irqsave(&ib_mad_port_list_lock, flags); - list_add_tail(&port_priv->port_list, &ib_mad_port_list); - spin_unlock_irqrestore(&ib_mad_port_list_lock, flags); return 0; error9: + spin_lock_irqsave(&ib_mad_port_list_lock, flags); + list_del_init(&port_priv->port_list); + spin_unlock_irqrestore(&ib_mad_port_list_lock, flags); + destroy_workqueue(port_priv->wq); error8: destroy_mad_qp(&port_priv->qp_info[1]); @@ -2615,11 +2949,9 @@ static int ib_mad_port_close(struct ib_device *device, int port_num) printk(KERN_ERR PFX "Port %d not found\n", port_num); return -ENODEV; } - list_del(&port_priv->port_list); + list_del_init(&port_priv->port_list); spin_unlock_irqrestore(&ib_mad_port_list_lock, flags); - /* Stop processing completions. */ - flush_workqueue(port_priv->wq); destroy_workqueue(port_priv->wq); destroy_mad_qp(&port_priv->qp_info[1]); destroy_mad_qp(&port_priv->qp_info[0]); @@ -2639,7 +2971,10 @@ static void ib_mad_init_device(struct ib_device *device) { int start, end, i; - if (device->node_type == IB_NODE_SWITCH) { + if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) + return; + + if (device->node_type == RDMA_NODE_IB_SWITCH) { start = 0; end = 0; } else { @@ -2686,7 +3021,10 @@ static void ib_mad_remove_device(struct ib_device *device) { int i, num_ports, cur_port; - if (device->node_type == IB_NODE_SWITCH) { + if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) + return; + + if (device->node_type == RDMA_NODE_IB_SWITCH) { num_ports = 1; cur_port = 0; } else { @@ -2714,13 +3052,16 @@ static int __init ib_mad_init_module(void) { int ret; - spin_lock_init(&ib_mad_port_list_lock); + mad_recvq_size = min(mad_recvq_size, IB_MAD_QP_MAX_SIZE); + mad_recvq_size = max(mad_recvq_size, IB_MAD_QP_MIN_SIZE); + + mad_sendq_size = min(mad_sendq_size, IB_MAD_QP_MAX_SIZE); + mad_sendq_size = max(mad_sendq_size, IB_MAD_QP_MIN_SIZE); ib_mad_cache = kmem_cache_create("ib_mad", sizeof(struct ib_mad_private), 0, SLAB_HWCACHE_ALIGN, - NULL, NULL); if (!ib_mad_cache) { printk(KERN_ERR PFX "Couldn't create ib_mad cache\n"); @@ -2747,12 +3088,8 @@ error1: static void __exit ib_mad_cleanup_module(void) { ib_unregister_client(&mad_client); - - if (kmem_cache_destroy(ib_mad_cache)) { - printk(KERN_DEBUG PFX "Failed to destroy ib_mad cache\n"); - } + kmem_cache_destroy(ib_mad_cache); } module_init(ib_mad_init_module); module_exit(ib_mad_cleanup_module); - diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h index 570f78682af..9430ab4969c 100644 --- a/drivers/infiniband/core/mad_priv.h +++ b/drivers/infiniband/core/mad_priv.h @@ -2,6 +2,7 @@ * Copyright (c) 2004, 2005, Voltaire, Inc. All rights reserved. * Copyright (c) 2005 Intel Corporation. All rights reserved. * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright (c) 2009 HNR Consulting. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -30,15 +31,13 @@ * 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. - * - * $Id: mad_priv.h 2730 2005-06-28 16:43:03Z sean.hefty $ */ #ifndef __IB_MAD_PRIV_H__ #define __IB_MAD_PRIV_H__ -#include <linux/pci.h> -#include <linux/kthread.h> +#include <linux/completion.h> +#include <linux/err.h> #include <linux/workqueue.h> #include <rdma/ib_mad.h> #include <rdma/ib_smi.h> @@ -51,6 +50,8 @@ /* QP and CQ parameters */ #define IB_MAD_QP_SEND_SIZE 128 #define IB_MAD_QP_RECV_SIZE 512 +#define IB_MAD_QP_MIN_SIZE 64 +#define IB_MAD_QP_MAX_SIZE 8192 #define IB_MAD_SEND_REQ_MAX_SG 2 #define IB_MAD_RECV_REQ_MAX_SG 1 @@ -72,7 +73,7 @@ struct ib_mad_private_header { struct ib_mad_list_head mad_list; struct ib_mad_recv_wc recv_wc; struct ib_wc wc; - DECLARE_PCI_UNMAP_ADDR(mapping) + u64 mapping; } __attribute__ ((packed)); struct ib_mad_private { @@ -85,6 +86,12 @@ struct ib_mad_private { } mad; } __attribute__ ((packed)); +struct ib_rmpp_segment { + struct list_head list; + u32 num; + u8 data[0]; +}; + struct ib_mad_agent_private { struct list_head agent_list; struct ib_mad_agent agent; @@ -95,14 +102,14 @@ struct ib_mad_agent_private { struct list_head send_list; struct list_head wait_list; struct list_head done_list; - struct work_struct timed_work; + struct delayed_work timed_work; unsigned long timeout; struct list_head local_list; struct work_struct local_work; struct list_head rmpp_list; atomic_t refcount; - wait_queue_head_t wait; + struct completion comp; }; struct ib_mad_snoop_private { @@ -111,7 +118,7 @@ struct ib_mad_snoop_private { int snoop_index; int mad_snoop_flags; atomic_t refcount; - wait_queue_head_t wait; + struct completion comp; }; struct ib_mad_send_wr_private { @@ -119,22 +126,25 @@ struct ib_mad_send_wr_private { struct list_head agent_list; struct ib_mad_agent_private *mad_agent_priv; struct ib_mad_send_buf send_buf; - DECLARE_PCI_UNMAP_ADDR(mapping) + u64 header_mapping; + u64 payload_mapping; struct ib_send_wr send_wr; struct ib_sge sg_list[IB_MAD_SEND_REQ_MAX_SG]; __be64 tid; unsigned long timeout; - int retries; + int max_retries; + int retries_left; int retry; int refcount; enum ib_wc_status status; /* RMPP control */ + struct list_head rmpp_list; + struct ib_rmpp_segment *last_ack_seg; + struct ib_rmpp_segment *cur_seg; int last_ack; int seg_num; int newwin; - int total_seg; - int data_offset; int pad; }; @@ -203,12 +213,11 @@ struct ib_mad_port_private { struct ib_mad_qp_info qp_info[IB_MAD_QPS_CORE]; }; -extern kmem_cache_t *ib_mad_cache; - int ib_send_mad(struct ib_mad_send_wr_private *mad_send_wr); struct ib_mad_send_wr_private * -ib_find_send_mad(struct ib_mad_agent_private *mad_agent_priv, __be64 tid); +ib_find_send_mad(struct ib_mad_agent_private *mad_agent_priv, + struct ib_mad_recv_wc *mad_recv_wc); void ib_mad_complete_send_wr(struct ib_mad_send_wr_private *mad_send_wr, struct ib_mad_send_wc *mad_send_wc); diff --git a/drivers/infiniband/core/mad_rmpp.c b/drivers/infiniband/core/mad_rmpp.c index 3249e1d8c07..f37878c9c06 100644 --- a/drivers/infiniband/core/mad_rmpp.c +++ b/drivers/infiniband/core/mad_rmpp.c @@ -1,6 +1,6 @@ /* * Copyright (c) 2005 Intel Inc. All rights reserved. - * Copyright (c) 2005 Voltaire, Inc. All rights reserved. + * Copyright (c) 2005-2006 Voltaire, Inc. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -29,11 +29,9 @@ * 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. - * - * $Id: mad_rmpp.c 1921 2005-03-02 22:58:44Z sean.hefty $ */ -#include <linux/dma-mapping.h> +#include <linux/slab.h> #include "mad_priv.h" #include "mad_rmpp.h" @@ -41,15 +39,16 @@ enum rmpp_state { RMPP_STATE_ACTIVE, RMPP_STATE_TIMEOUT, - RMPP_STATE_COMPLETE + RMPP_STATE_COMPLETE, + RMPP_STATE_CANCELING }; struct mad_rmpp_recv { struct ib_mad_agent_private *agent; struct list_head list; - struct work_struct timeout_work; - struct work_struct cleanup_work; - wait_queue_head_t wait; + struct delayed_work timeout_work; + struct delayed_work cleanup_work; + struct completion comp; enum rmpp_state state; spinlock_t lock; atomic_t refcount; @@ -60,6 +59,7 @@ struct mad_rmpp_recv { int last_ack; int seg_num; int newwin; + int repwin; __be64 tid; u32 src_qp; @@ -69,10 +69,16 @@ struct mad_rmpp_recv { u8 method; }; +static inline void deref_rmpp_recv(struct mad_rmpp_recv *rmpp_recv) +{ + if (atomic_dec_and_test(&rmpp_recv->refcount)) + complete(&rmpp_recv->comp); +} + static void destroy_rmpp_recv(struct mad_rmpp_recv *rmpp_recv) { - atomic_dec(&rmpp_recv->refcount); - wait_event(rmpp_recv->wait, !atomic_read(&rmpp_recv->refcount)); + deref_rmpp_recv(rmpp_recv); + wait_for_completion(&rmpp_recv->comp); ib_destroy_ah(rmpp_recv->ah); kfree(rmpp_recv); } @@ -84,41 +90,34 @@ void ib_cancel_rmpp_recvs(struct ib_mad_agent_private *agent) spin_lock_irqsave(&agent->lock, flags); list_for_each_entry(rmpp_recv, &agent->rmpp_list, list) { + if (rmpp_recv->state != RMPP_STATE_COMPLETE) + ib_free_recv_mad(rmpp_recv->rmpp_wc); + rmpp_recv->state = RMPP_STATE_CANCELING; + } + spin_unlock_irqrestore(&agent->lock, flags); + + list_for_each_entry(rmpp_recv, &agent->rmpp_list, list) { cancel_delayed_work(&rmpp_recv->timeout_work); cancel_delayed_work(&rmpp_recv->cleanup_work); } - spin_unlock_irqrestore(&agent->lock, flags); flush_workqueue(agent->qp_info->port_priv->wq); list_for_each_entry_safe(rmpp_recv, temp_rmpp_recv, &agent->rmpp_list, list) { list_del(&rmpp_recv->list); - if (rmpp_recv->state != RMPP_STATE_COMPLETE) - ib_free_recv_mad(rmpp_recv->rmpp_wc); destroy_rmpp_recv(rmpp_recv); } } -static int data_offset(u8 mgmt_class) -{ - if (mgmt_class == IB_MGMT_CLASS_SUBN_ADM) - return IB_MGMT_SA_HDR; - else if ((mgmt_class >= IB_MGMT_CLASS_VENDOR_RANGE2_START) && - (mgmt_class <= IB_MGMT_CLASS_VENDOR_RANGE2_END)) - return IB_MGMT_VENDOR_HDR; - else - return IB_MGMT_RMPP_HDR; -} - -static void format_ack(struct ib_rmpp_mad *ack, +static void format_ack(struct ib_mad_send_buf *msg, struct ib_rmpp_mad *data, struct mad_rmpp_recv *rmpp_recv) { + struct ib_rmpp_mad *ack = msg->mad; unsigned long flags; - memcpy(&ack->mad_hdr, &data->mad_hdr, - data_offset(data->mad_hdr.mgmt_class)); + memcpy(ack, &data->mad_hdr, msg->hdr_len); ack->mad_hdr.method ^= IB_MGMT_METHOD_RESP; ack->rmpp_hdr.rmpp_type = IB_MGMT_RMPP_TYPE_ACK; @@ -135,16 +134,16 @@ static void ack_recv(struct mad_rmpp_recv *rmpp_recv, struct ib_mad_recv_wc *recv_wc) { struct ib_mad_send_buf *msg; - int ret; + int ret, hdr_len; + hdr_len = ib_get_mad_data_offset(recv_wc->recv_buf.mad->mad_hdr.mgmt_class); msg = ib_create_send_mad(&rmpp_recv->agent->agent, recv_wc->wc->src_qp, - recv_wc->wc->pkey_index, 1, IB_MGMT_RMPP_HDR, - IB_MGMT_RMPP_DATA, GFP_KERNEL); - if (!msg) + recv_wc->wc->pkey_index, 1, hdr_len, + 0, GFP_KERNEL); + if (IS_ERR(msg)) return; - format_ack(msg->mad, (struct ib_rmpp_mad *) recv_wc->recv_buf.mad, - rmpp_recv); + format_ack(msg, (struct ib_rmpp_mad *) recv_wc->recv_buf.mad, rmpp_recv); msg->ah = rmpp_recv->ah; ret = ib_post_send_mad(msg, NULL); if (ret) @@ -156,29 +155,56 @@ static struct ib_mad_send_buf *alloc_response_msg(struct ib_mad_agent *agent, { struct ib_mad_send_buf *msg; struct ib_ah *ah; + int hdr_len; ah = ib_create_ah_from_wc(agent->qp->pd, recv_wc->wc, recv_wc->recv_buf.grh, agent->port_num); if (IS_ERR(ah)) return (void *) ah; + hdr_len = ib_get_mad_data_offset(recv_wc->recv_buf.mad->mad_hdr.mgmt_class); msg = ib_create_send_mad(agent, recv_wc->wc->src_qp, recv_wc->wc->pkey_index, 1, - IB_MGMT_RMPP_HDR, IB_MGMT_RMPP_DATA, - GFP_KERNEL); + hdr_len, 0, GFP_KERNEL); if (IS_ERR(msg)) ib_destroy_ah(ah); - else + else { msg->ah = ah; + msg->context[0] = ah; + } return msg; } -void ib_rmpp_send_handler(struct ib_mad_send_wc *mad_send_wc) +static void ack_ds_ack(struct ib_mad_agent_private *agent, + struct ib_mad_recv_wc *recv_wc) { - struct ib_rmpp_mad *rmpp_mad = mad_send_wc->send_buf->mad; + struct ib_mad_send_buf *msg; + struct ib_rmpp_mad *rmpp_mad; + int ret; + + msg = alloc_response_msg(&agent->agent, recv_wc); + if (IS_ERR(msg)) + return; + + rmpp_mad = msg->mad; + memcpy(rmpp_mad, recv_wc->recv_buf.mad, msg->hdr_len); + + rmpp_mad->mad_hdr.method ^= IB_MGMT_METHOD_RESP; + ib_set_rmpp_flags(&rmpp_mad->rmpp_hdr, IB_MGMT_RMPP_FLAG_ACTIVE); + rmpp_mad->rmpp_hdr.seg_num = 0; + rmpp_mad->rmpp_hdr.paylen_newwin = cpu_to_be32(1); - if (rmpp_mad->rmpp_hdr.rmpp_type != IB_MGMT_RMPP_TYPE_ACK) + ret = ib_post_send_mad(msg, NULL); + if (ret) { + ib_destroy_ah(msg->ah); + ib_free_send_mad(msg); + } +} + +void ib_rmpp_send_handler(struct ib_mad_send_wc *mad_send_wc) +{ + if (mad_send_wc->send_buf->context[0] == mad_send_wc->send_buf->ah) ib_destroy_ah(mad_send_wc->send_buf->ah); ib_free_send_mad(mad_send_wc->send_buf); } @@ -195,8 +221,7 @@ static void nack_recv(struct ib_mad_agent_private *agent, return; rmpp_mad = msg->mad; - memcpy(rmpp_mad, recv_wc->recv_buf.mad, - data_offset(recv_wc->recv_buf.mad->mad_hdr.mgmt_class)); + memcpy(rmpp_mad, recv_wc->recv_buf.mad, msg->hdr_len); rmpp_mad->mad_hdr.method ^= IB_MGMT_METHOD_RESP; rmpp_mad->rmpp_hdr.rmpp_version = IB_MGMT_RMPP_VERSION; @@ -213,9 +238,10 @@ static void nack_recv(struct ib_mad_agent_private *agent, } } -static void recv_timeout_handler(void *data) +static void recv_timeout_handler(struct work_struct *work) { - struct mad_rmpp_recv *rmpp_recv = data; + struct mad_rmpp_recv *rmpp_recv = + container_of(work, struct mad_rmpp_recv, timeout_work.work); struct ib_mad_recv_wc *rmpp_wc; unsigned long flags; @@ -234,12 +260,17 @@ static void recv_timeout_handler(void *data) ib_free_recv_mad(rmpp_wc); } -static void recv_cleanup_handler(void *data) +static void recv_cleanup_handler(struct work_struct *work) { - struct mad_rmpp_recv *rmpp_recv = data; + struct mad_rmpp_recv *rmpp_recv = + container_of(work, struct mad_rmpp_recv, cleanup_work.work); unsigned long flags; spin_lock_irqsave(&rmpp_recv->agent->lock, flags); + if (rmpp_recv->state == RMPP_STATE_CANCELING) { + spin_unlock_irqrestore(&rmpp_recv->agent->lock, flags); + return; + } list_del(&rmpp_recv->list); spin_unlock_irqrestore(&rmpp_recv->agent->lock, flags); destroy_rmpp_recv(rmpp_recv); @@ -264,9 +295,9 @@ create_rmpp_recv(struct ib_mad_agent_private *agent, goto error; rmpp_recv->agent = agent; - init_waitqueue_head(&rmpp_recv->wait); - INIT_WORK(&rmpp_recv->timeout_work, recv_timeout_handler, rmpp_recv); - INIT_WORK(&rmpp_recv->cleanup_work, recv_cleanup_handler, rmpp_recv); + init_completion(&rmpp_recv->comp); + INIT_DELAYED_WORK(&rmpp_recv->timeout_work, recv_timeout_handler); + INIT_DELAYED_WORK(&rmpp_recv->cleanup_work, recv_cleanup_handler); spin_lock_init(&rmpp_recv->lock); rmpp_recv->state = RMPP_STATE_ACTIVE; atomic_set(&rmpp_recv->refcount, 1); @@ -276,6 +307,7 @@ create_rmpp_recv(struct ib_mad_agent_private *agent, rmpp_recv->newwin = 1; rmpp_recv->seg_num = 1; rmpp_recv->last_ack = 0; + rmpp_recv->repwin = 1; mad_hdr = &mad_recv_wc->recv_buf.mad->mad_hdr; rmpp_recv->tid = mad_hdr->tid; @@ -290,12 +322,6 @@ error: kfree(rmpp_recv); return NULL; } -static inline void deref_rmpp_recv(struct mad_rmpp_recv *rmpp_recv) -{ - if (atomic_dec_and_test(&rmpp_recv->refcount)) - wake_up(&rmpp_recv->wait); -} - static struct mad_rmpp_recv * find_rmpp_recv(struct ib_mad_agent_private *agent, struct ib_mad_recv_wc *mad_recv_wc) @@ -376,7 +402,7 @@ static inline int window_size(struct ib_mad_agent_private *agent) static struct ib_mad_recv_buf * find_seg_location(struct list_head *rmpp_list, int seg_num) { - struct ib_mad_recv_buf *seg_buf; + struct ib_mad_recv_buf *seg_buf; int cur_seg_num; list_for_each_entry_reverse(seg_buf, rmpp_list, list) { @@ -408,7 +434,7 @@ static inline int get_mad_len(struct mad_rmpp_recv *rmpp_recv) rmpp_mad = (struct ib_rmpp_mad *)rmpp_recv->cur_seg_buf->mad; - hdr_size = data_offset(rmpp_mad->mad_hdr.mgmt_class); + hdr_size = ib_get_mad_data_offset(rmpp_mad->mad_hdr.mgmt_class); data_size = sizeof(struct ib_rmpp_mad) - hdr_size; pad = IB_MGMT_RMPP_DATA - be32_to_cpu(rmpp_mad->rmpp_hdr.paylen_newwin); if (pad > IB_MGMT_RMPP_DATA || pad < 0) @@ -433,44 +459,6 @@ static struct ib_mad_recv_wc * complete_rmpp(struct mad_rmpp_recv *rmpp_recv) return rmpp_wc; } -void ib_coalesce_recv_mad(struct ib_mad_recv_wc *mad_recv_wc, void *buf) -{ - struct ib_mad_recv_buf *seg_buf; - struct ib_rmpp_mad *rmpp_mad; - void *data; - int size, len, offset; - u8 flags; - - len = mad_recv_wc->mad_len; - if (len <= sizeof(struct ib_mad)) { - memcpy(buf, mad_recv_wc->recv_buf.mad, len); - return; - } - - offset = data_offset(mad_recv_wc->recv_buf.mad->mad_hdr.mgmt_class); - - list_for_each_entry(seg_buf, &mad_recv_wc->rmpp_list, list) { - rmpp_mad = (struct ib_rmpp_mad *)seg_buf->mad; - flags = ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr); - - if (flags & IB_MGMT_RMPP_FLAG_FIRST) { - data = rmpp_mad; - size = sizeof(*rmpp_mad); - } else { - data = (void *) rmpp_mad + offset; - if (flags & IB_MGMT_RMPP_FLAG_LAST) - size = len; - else - size = sizeof(*rmpp_mad) - offset; - } - - memcpy(buf, data, size); - len -= size; - buf += size; - } -} -EXPORT_SYMBOL(ib_coalesce_recv_mad); - static struct ib_mad_recv_wc * continue_rmpp(struct ib_mad_agent_private *agent, struct ib_mad_recv_wc *mad_recv_wc) @@ -570,66 +558,49 @@ start_rmpp(struct ib_mad_agent_private *agent, return mad_recv_wc; } -static inline u64 get_seg_addr(struct ib_mad_send_wr_private *mad_send_wr) -{ - return mad_send_wr->sg_list[0].addr + mad_send_wr->data_offset + - (sizeof(struct ib_rmpp_mad) - mad_send_wr->data_offset) * - (mad_send_wr->seg_num - 1); -} - static int send_next_seg(struct ib_mad_send_wr_private *mad_send_wr) { struct ib_rmpp_mad *rmpp_mad; int timeout; - u32 paylen; + u32 paylen = 0; rmpp_mad = mad_send_wr->send_buf.mad; ib_set_rmpp_flags(&rmpp_mad->rmpp_hdr, IB_MGMT_RMPP_FLAG_ACTIVE); - rmpp_mad->rmpp_hdr.seg_num = cpu_to_be32(mad_send_wr->seg_num); + rmpp_mad->rmpp_hdr.seg_num = cpu_to_be32(++mad_send_wr->seg_num); if (mad_send_wr->seg_num == 1) { rmpp_mad->rmpp_hdr.rmpp_rtime_flags |= IB_MGMT_RMPP_FLAG_FIRST; - paylen = mad_send_wr->total_seg * IB_MGMT_RMPP_DATA - + paylen = mad_send_wr->send_buf.seg_count * IB_MGMT_RMPP_DATA - mad_send_wr->pad; - rmpp_mad->rmpp_hdr.paylen_newwin = cpu_to_be32(paylen); - mad_send_wr->sg_list[0].length = sizeof(struct ib_rmpp_mad); - } else { - mad_send_wr->send_wr.num_sge = 2; - mad_send_wr->sg_list[0].length = mad_send_wr->data_offset; - mad_send_wr->sg_list[1].addr = get_seg_addr(mad_send_wr); - mad_send_wr->sg_list[1].length = sizeof(struct ib_rmpp_mad) - - mad_send_wr->data_offset; - mad_send_wr->sg_list[1].lkey = mad_send_wr->sg_list[0].lkey; - rmpp_mad->rmpp_hdr.paylen_newwin = 0; } - if (mad_send_wr->seg_num == mad_send_wr->total_seg) { + if (mad_send_wr->seg_num == mad_send_wr->send_buf.seg_count) { rmpp_mad->rmpp_hdr.rmpp_rtime_flags |= IB_MGMT_RMPP_FLAG_LAST; paylen = IB_MGMT_RMPP_DATA - mad_send_wr->pad; - rmpp_mad->rmpp_hdr.paylen_newwin = cpu_to_be32(paylen); } + rmpp_mad->rmpp_hdr.paylen_newwin = cpu_to_be32(paylen); /* 2 seconds for an ACK until we can find the packet lifetime */ timeout = mad_send_wr->send_buf.timeout_ms; if (!timeout || timeout > 2000) mad_send_wr->timeout = msecs_to_jiffies(2000); - mad_send_wr->seg_num++; + return ib_send_mad(mad_send_wr); } -static void abort_send(struct ib_mad_agent_private *agent, __be64 tid, - u8 rmpp_status) +static void abort_send(struct ib_mad_agent_private *agent, + struct ib_mad_recv_wc *mad_recv_wc, u8 rmpp_status) { struct ib_mad_send_wr_private *mad_send_wr; struct ib_mad_send_wc wc; unsigned long flags; spin_lock_irqsave(&agent->lock, flags); - mad_send_wr = ib_find_send_mad(agent, tid); + mad_send_wr = ib_find_send_mad(agent, mad_recv_wc); if (!mad_send_wr) goto out; /* Unmatched send */ - if ((mad_send_wr->last_ack == mad_send_wr->total_seg) || + if ((mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) || (!mad_send_wr->timeout) || (mad_send_wr->status != IB_WC_SUCCESS)) goto out; /* Send is already done */ @@ -645,6 +616,28 @@ out: spin_unlock_irqrestore(&agent->lock, flags); } +static inline void adjust_last_ack(struct ib_mad_send_wr_private *wr, + int seg_num) +{ + struct list_head *list; + + wr->last_ack = seg_num; + list = &wr->last_ack_seg->list; + list_for_each_entry(wr->last_ack_seg, list, list) + if (wr->last_ack_seg->num == seg_num) + break; +} + +static void process_ds_ack(struct ib_mad_agent_private *agent, + struct ib_mad_recv_wc *mad_recv_wc, int newwin) +{ + struct mad_rmpp_recv *rmpp_recv; + + rmpp_recv = find_rmpp_recv(agent, mad_recv_wc); + if (rmpp_recv && rmpp_recv->state == RMPP_STATE_COMPLETE) + rmpp_recv->repwin = newwin; +} + static void process_rmpp_ack(struct ib_mad_agent_private *agent, struct ib_mad_recv_wc *mad_recv_wc) { @@ -655,8 +648,7 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent, rmpp_mad = (struct ib_rmpp_mad *)mad_recv_wc->recv_buf.mad; if (rmpp_mad->rmpp_hdr.rmpp_status) { - abort_send(agent, rmpp_mad->mad_hdr.tid, - IB_MGMT_RMPP_STATUS_BAD_STATUS); + abort_send(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BAD_STATUS); nack_recv(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BAD_STATUS); return; } @@ -664,25 +656,34 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent, seg_num = be32_to_cpu(rmpp_mad->rmpp_hdr.seg_num); newwin = be32_to_cpu(rmpp_mad->rmpp_hdr.paylen_newwin); if (newwin < seg_num) { - abort_send(agent, rmpp_mad->mad_hdr.tid, - IB_MGMT_RMPP_STATUS_W2S); + abort_send(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_W2S); nack_recv(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_W2S); return; } spin_lock_irqsave(&agent->lock, flags); - mad_send_wr = ib_find_send_mad(agent, rmpp_mad->mad_hdr.tid); - if (!mad_send_wr) - goto out; /* Unmatched ACK */ + mad_send_wr = ib_find_send_mad(agent, mad_recv_wc); + if (!mad_send_wr) { + if (!seg_num) + process_ds_ack(agent, mad_recv_wc, newwin); + goto out; /* Unmatched or DS RMPP ACK */ + } - if ((mad_send_wr->last_ack == mad_send_wr->total_seg) || + if ((mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) && + (mad_send_wr->timeout)) { + spin_unlock_irqrestore(&agent->lock, flags); + ack_ds_ack(agent, mad_recv_wc); + return; /* Repeated ACK for DS RMPP transaction */ + } + + if ((mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) || (!mad_send_wr->timeout) || (mad_send_wr->status != IB_WC_SUCCESS)) goto out; /* Send is already done */ - if (seg_num > mad_send_wr->total_seg || seg_num > mad_send_wr->newwin) { + if (seg_num > mad_send_wr->send_buf.seg_count || + seg_num > mad_send_wr->newwin) { spin_unlock_irqrestore(&agent->lock, flags); - abort_send(agent, rmpp_mad->mad_hdr.tid, - IB_MGMT_RMPP_STATUS_S2B); + abort_send(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_S2B); nack_recv(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_S2B); return; } @@ -691,11 +692,11 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent, goto out; /* Old ACK */ if (seg_num > mad_send_wr->last_ack) { - mad_send_wr->last_ack = seg_num; - mad_send_wr->retries = mad_send_wr->send_buf.retries; + adjust_last_ack(mad_send_wr, seg_num); + mad_send_wr->retries_left = mad_send_wr->max_retries; } mad_send_wr->newwin = newwin; - if (mad_send_wr->last_ack == mad_send_wr->total_seg) { + if (mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) { /* If no response is expected, the ACK completes the send */ if (!mad_send_wr->send_buf.timeout_ms) { struct ib_mad_send_wc wc; @@ -712,17 +713,19 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent, if (mad_send_wr->refcount == 1) ib_reset_mad_timeout(mad_send_wr, mad_send_wr->send_buf.timeout_ms); + spin_unlock_irqrestore(&agent->lock, flags); + ack_ds_ack(agent, mad_recv_wc); + return; } else if (mad_send_wr->refcount == 1 && mad_send_wr->seg_num < mad_send_wr->newwin && - mad_send_wr->seg_num <= mad_send_wr->total_seg) { + mad_send_wr->seg_num < mad_send_wr->send_buf.seg_count) { /* Send failure will just result in a timeout/retry */ ret = send_next_seg(mad_send_wr); if (ret) goto out; mad_send_wr->refcount++; - list_del(&mad_send_wr->agent_list); - list_add_tail(&mad_send_wr->agent_list, + list_move_tail(&mad_send_wr->agent_list, &mad_send_wr->mad_agent_priv->send_list); } out: @@ -743,7 +746,7 @@ process_rmpp_data(struct ib_mad_agent_private *agent, goto bad; } - if (rmpp_hdr->seg_num == __constant_htonl(1)) { + if (rmpp_hdr->seg_num == cpu_to_be32(1)) { if (!(ib_get_rmpp_flags(rmpp_hdr) & IB_MGMT_RMPP_FLAG_FIRST)) { rmpp_status = IB_MGMT_RMPP_STATUS_BAD_SEG; goto bad; @@ -770,12 +773,10 @@ static void process_rmpp_stop(struct ib_mad_agent_private *agent, rmpp_mad = (struct ib_rmpp_mad *)mad_recv_wc->recv_buf.mad; if (rmpp_mad->rmpp_hdr.rmpp_status != IB_MGMT_RMPP_STATUS_RESX) { - abort_send(agent, rmpp_mad->mad_hdr.tid, - IB_MGMT_RMPP_STATUS_BAD_STATUS); + abort_send(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BAD_STATUS); nack_recv(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BAD_STATUS); } else - abort_send(agent, rmpp_mad->mad_hdr.tid, - rmpp_mad->rmpp_hdr.rmpp_status); + abort_send(agent, mad_recv_wc, rmpp_mad->rmpp_hdr.rmpp_status); } static void process_rmpp_abort(struct ib_mad_agent_private *agent, @@ -787,12 +788,10 @@ static void process_rmpp_abort(struct ib_mad_agent_private *agent, if (rmpp_mad->rmpp_hdr.rmpp_status < IB_MGMT_RMPP_STATUS_ABORT_MIN || rmpp_mad->rmpp_hdr.rmpp_status > IB_MGMT_RMPP_STATUS_ABORT_MAX) { - abort_send(agent, rmpp_mad->mad_hdr.tid, - IB_MGMT_RMPP_STATUS_BAD_STATUS); + abort_send(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BAD_STATUS); nack_recv(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BAD_STATUS); } else - abort_send(agent, rmpp_mad->mad_hdr.tid, - rmpp_mad->rmpp_hdr.rmpp_status); + abort_send(agent, mad_recv_wc, rmpp_mad->rmpp_hdr.rmpp_status); } struct ib_mad_recv_wc * @@ -806,8 +805,7 @@ ib_process_rmpp_recv_wc(struct ib_mad_agent_private *agent, return mad_recv_wc; if (rmpp_mad->rmpp_hdr.rmpp_version != IB_MGMT_RMPP_VERSION) { - abort_send(agent, rmpp_mad->mad_hdr.tid, - IB_MGMT_RMPP_STATUS_UNV); + abort_send(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_UNV); nack_recv(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_UNV); goto out; } @@ -825,8 +823,7 @@ ib_process_rmpp_recv_wc(struct ib_mad_agent_private *agent, process_rmpp_abort(agent, mad_recv_wc); break; default: - abort_send(agent, rmpp_mad->mad_hdr.tid, - IB_MGMT_RMPP_STATUS_BADT); + abort_send(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BADT); nack_recv(agent, mad_recv_wc, IB_MGMT_RMPP_STATUS_BADT); break; } @@ -835,34 +832,55 @@ out: return NULL; } +static int init_newwin(struct ib_mad_send_wr_private *mad_send_wr) +{ + struct ib_mad_agent_private *agent = mad_send_wr->mad_agent_priv; + struct ib_mad_hdr *mad_hdr = mad_send_wr->send_buf.mad; + struct mad_rmpp_recv *rmpp_recv; + struct ib_ah_attr ah_attr; + unsigned long flags; + int newwin = 1; + + if (!(mad_hdr->method & IB_MGMT_METHOD_RESP)) + goto out; + + spin_lock_irqsave(&agent->lock, flags); + list_for_each_entry(rmpp_recv, &agent->rmpp_list, list) { + if (rmpp_recv->tid != mad_hdr->tid || + rmpp_recv->mgmt_class != mad_hdr->mgmt_class || + rmpp_recv->class_version != mad_hdr->class_version || + (rmpp_recv->method & IB_MGMT_METHOD_RESP)) + continue; + + if (ib_query_ah(mad_send_wr->send_buf.ah, &ah_attr)) + continue; + + if (rmpp_recv->slid == ah_attr.dlid) { + newwin = rmpp_recv->repwin; + break; + } + } + spin_unlock_irqrestore(&agent->lock, flags); +out: + return newwin; +} + int ib_send_rmpp_mad(struct ib_mad_send_wr_private *mad_send_wr) { struct ib_rmpp_mad *rmpp_mad; - int i, total_len, ret; + int ret; rmpp_mad = mad_send_wr->send_buf.mad; if (!(ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE)) return IB_RMPP_RESULT_UNHANDLED; - if (rmpp_mad->rmpp_hdr.rmpp_type != IB_MGMT_RMPP_TYPE_DATA) + if (rmpp_mad->rmpp_hdr.rmpp_type != IB_MGMT_RMPP_TYPE_DATA) { + mad_send_wr->seg_num = 1; return IB_RMPP_RESULT_INTERNAL; + } - if (mad_send_wr->send_wr.num_sge > 1) - return -EINVAL; /* TODO: support num_sge > 1 */ - - mad_send_wr->seg_num = 1; - mad_send_wr->newwin = 1; - mad_send_wr->data_offset = data_offset(rmpp_mad->mad_hdr.mgmt_class); - - total_len = 0; - for (i = 0; i < mad_send_wr->send_wr.num_sge; i++) - total_len += mad_send_wr->send_wr.sg_list[i].length; - - mad_send_wr->total_seg = (total_len - mad_send_wr->data_offset) / - (sizeof(struct ib_rmpp_mad) - mad_send_wr->data_offset); - mad_send_wr->pad = total_len - IB_MGMT_RMPP_HDR - - be32_to_cpu(rmpp_mad->rmpp_hdr.paylen_newwin); + mad_send_wr->newwin = init_newwin(mad_send_wr); /* We need to wait for the final ACK even if there isn't a response */ mad_send_wr->refcount += (mad_send_wr->timeout == 0); @@ -893,14 +911,14 @@ int ib_process_rmpp_send_wc(struct ib_mad_send_wr_private *mad_send_wr, if (!mad_send_wr->timeout) return IB_RMPP_RESULT_PROCESSED; /* Response received */ - if (mad_send_wr->last_ack == mad_send_wr->total_seg) { + if (mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) { mad_send_wr->timeout = msecs_to_jiffies(mad_send_wr->send_buf.timeout_ms); return IB_RMPP_RESULT_PROCESSED; /* Send done */ } - if (mad_send_wr->seg_num > mad_send_wr->newwin || - mad_send_wr->seg_num > mad_send_wr->total_seg) + if (mad_send_wr->seg_num == mad_send_wr->newwin || + mad_send_wr->seg_num == mad_send_wr->send_buf.seg_count) return IB_RMPP_RESULT_PROCESSED; /* Wait for ACK */ ret = send_next_seg(mad_send_wr); @@ -921,10 +939,12 @@ int ib_retry_rmpp(struct ib_mad_send_wr_private *mad_send_wr) IB_MGMT_RMPP_FLAG_ACTIVE)) return IB_RMPP_RESULT_UNHANDLED; /* RMPP not active */ - if (mad_send_wr->last_ack == mad_send_wr->total_seg) + if (mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) return IB_RMPP_RESULT_PROCESSED; - mad_send_wr->seg_num = mad_send_wr->last_ack + 1; + mad_send_wr->seg_num = mad_send_wr->last_ack; + mad_send_wr->cur_seg = mad_send_wr->last_ack_seg; + ret = send_next_seg(mad_send_wr); if (ret) return IB_RMPP_RESULT_PROCESSED; diff --git a/drivers/infiniband/core/mad_rmpp.h b/drivers/infiniband/core/mad_rmpp.h index f0616fd2249..3d336bff114 100644 --- a/drivers/infiniband/core/mad_rmpp.h +++ b/drivers/infiniband/core/mad_rmpp.h @@ -28,8 +28,6 @@ * 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. - * - * $Id: mad_rmpp.h 1921 2005-02-25 22:58:44Z sean.hefty $ */ #ifndef __MAD_RMPP_H__ diff --git a/drivers/infiniband/core/multicast.c b/drivers/infiniband/core/multicast.c new file mode 100644 index 00000000000..d2360a8ef0b --- /dev/null +++ b/drivers/infiniband/core/multicast.c @@ -0,0 +1,898 @@ +/* + * Copyright (c) 2006 Intel Corporation. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/completion.h> +#include <linux/dma-mapping.h> +#include <linux/err.h> +#include <linux/interrupt.h> +#include <linux/export.h> +#include <linux/slab.h> +#include <linux/bitops.h> +#include <linux/random.h> + +#include <rdma/ib_cache.h> +#include "sa.h" + +static void mcast_add_one(struct ib_device *device); +static void mcast_remove_one(struct ib_device *device); + +static struct ib_client mcast_client = { + .name = "ib_multicast", + .add = mcast_add_one, + .remove = mcast_remove_one +}; + +static struct ib_sa_client sa_client; +static struct workqueue_struct *mcast_wq; +static union ib_gid mgid0; + +struct mcast_device; + +struct mcast_port { + struct mcast_device *dev; + spinlock_t lock; + struct rb_root table; + atomic_t refcount; + struct completion comp; + u8 port_num; +}; + +struct mcast_device { + struct ib_device *device; + struct ib_event_handler event_handler; + int start_port; + int end_port; + struct mcast_port port[0]; +}; + +enum mcast_state { + MCAST_JOINING, + MCAST_MEMBER, + MCAST_ERROR, +}; + +enum mcast_group_state { + MCAST_IDLE, + MCAST_BUSY, + MCAST_GROUP_ERROR, + MCAST_PKEY_EVENT +}; + +enum { + MCAST_INVALID_PKEY_INDEX = 0xFFFF +}; + +struct mcast_member; + +struct mcast_group { + struct ib_sa_mcmember_rec rec; + struct rb_node node; + struct mcast_port *port; + spinlock_t lock; + struct work_struct work; + struct list_head pending_list; + struct list_head active_list; + struct mcast_member *last_join; + int members[3]; + atomic_t refcount; + enum mcast_group_state state; + struct ib_sa_query *query; + int query_id; + u16 pkey_index; + u8 leave_state; + int retries; +}; + +struct mcast_member { + struct ib_sa_multicast multicast; + struct ib_sa_client *client; + struct mcast_group *group; + struct list_head list; + enum mcast_state state; + atomic_t refcount; + struct completion comp; +}; + +static void join_handler(int status, struct ib_sa_mcmember_rec *rec, + void *context); +static void leave_handler(int status, struct ib_sa_mcmember_rec *rec, + void *context); + +static struct mcast_group *mcast_find(struct mcast_port *port, + union ib_gid *mgid) +{ + struct rb_node *node = port->table.rb_node; + struct mcast_group *group; + int ret; + + while (node) { + group = rb_entry(node, struct mcast_group, node); + ret = memcmp(mgid->raw, group->rec.mgid.raw, sizeof *mgid); + if (!ret) + return group; + + if (ret < 0) + node = node->rb_left; + else + node = node->rb_right; + } + return NULL; +} + +static struct mcast_group *mcast_insert(struct mcast_port *port, + struct mcast_group *group, + int allow_duplicates) +{ + struct rb_node **link = &port->table.rb_node; + struct rb_node *parent = NULL; + struct mcast_group *cur_group; + int ret; + + while (*link) { + parent = *link; + cur_group = rb_entry(parent, struct mcast_group, node); + + ret = memcmp(group->rec.mgid.raw, cur_group->rec.mgid.raw, + sizeof group->rec.mgid); + if (ret < 0) + link = &(*link)->rb_left; + else if (ret > 0) + link = &(*link)->rb_right; + else if (allow_duplicates) + link = &(*link)->rb_left; + else + return cur_group; + } + rb_link_node(&group->node, parent, link); + rb_insert_color(&group->node, &port->table); + return NULL; +} + +static void deref_port(struct mcast_port *port) +{ + if (atomic_dec_and_test(&port->refcount)) + complete(&port->comp); +} + +static void release_group(struct mcast_group *group) +{ + struct mcast_port *port = group->port; + unsigned long flags; + + spin_lock_irqsave(&port->lock, flags); + if (atomic_dec_and_test(&group->refcount)) { + rb_erase(&group->node, &port->table); + spin_unlock_irqrestore(&port->lock, flags); + kfree(group); + deref_port(port); + } else + spin_unlock_irqrestore(&port->lock, flags); +} + +static void deref_member(struct mcast_member *member) +{ + if (atomic_dec_and_test(&member->refcount)) + complete(&member->comp); +} + +static void queue_join(struct mcast_member *member) +{ + struct mcast_group *group = member->group; + unsigned long flags; + + spin_lock_irqsave(&group->lock, flags); + list_add_tail(&member->list, &group->pending_list); + if (group->state == MCAST_IDLE) { + group->state = MCAST_BUSY; + atomic_inc(&group->refcount); + queue_work(mcast_wq, &group->work); + } + spin_unlock_irqrestore(&group->lock, flags); +} + +/* + * A multicast group has three types of members: full member, non member, and + * send only member. We need to keep track of the number of members of each + * type based on their join state. Adjust the number of members the belong to + * the specified join states. + */ +static void adjust_membership(struct mcast_group *group, u8 join_state, int inc) +{ + int i; + + for (i = 0; i < 3; i++, join_state >>= 1) + if (join_state & 0x1) + group->members[i] += inc; +} + +/* + * If a multicast group has zero members left for a particular join state, but + * the group is still a member with the SA, we need to leave that join state. + * Determine which join states we still belong to, but that do not have any + * active members. + */ +static u8 get_leave_state(struct mcast_group *group) +{ + u8 leave_state = 0; + int i; + + for (i = 0; i < 3; i++) + if (!group->members[i]) + leave_state |= (0x1 << i); + + return leave_state & group->rec.join_state; +} + +static int check_selector(ib_sa_comp_mask comp_mask, + ib_sa_comp_mask selector_mask, + ib_sa_comp_mask value_mask, + u8 selector, u8 src_value, u8 dst_value) +{ + int err; + + if (!(comp_mask & selector_mask) || !(comp_mask & value_mask)) + return 0; + + switch (selector) { + case IB_SA_GT: + err = (src_value <= dst_value); + break; + case IB_SA_LT: + err = (src_value >= dst_value); + break; + case IB_SA_EQ: + err = (src_value != dst_value); + break; + default: + err = 0; + break; + } + + return err; +} + +static int cmp_rec(struct ib_sa_mcmember_rec *src, + struct ib_sa_mcmember_rec *dst, ib_sa_comp_mask comp_mask) +{ + /* MGID must already match */ + + if (comp_mask & IB_SA_MCMEMBER_REC_PORT_GID && + memcmp(&src->port_gid, &dst->port_gid, sizeof src->port_gid)) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_QKEY && src->qkey != dst->qkey) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_MLID && src->mlid != dst->mlid) + return -EINVAL; + if (check_selector(comp_mask, IB_SA_MCMEMBER_REC_MTU_SELECTOR, + IB_SA_MCMEMBER_REC_MTU, dst->mtu_selector, + src->mtu, dst->mtu)) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_TRAFFIC_CLASS && + src->traffic_class != dst->traffic_class) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_PKEY && src->pkey != dst->pkey) + return -EINVAL; + if (check_selector(comp_mask, IB_SA_MCMEMBER_REC_RATE_SELECTOR, + IB_SA_MCMEMBER_REC_RATE, dst->rate_selector, + src->rate, dst->rate)) + return -EINVAL; + if (check_selector(comp_mask, + IB_SA_MCMEMBER_REC_PACKET_LIFE_TIME_SELECTOR, + IB_SA_MCMEMBER_REC_PACKET_LIFE_TIME, + dst->packet_life_time_selector, + src->packet_life_time, dst->packet_life_time)) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_SL && src->sl != dst->sl) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_FLOW_LABEL && + src->flow_label != dst->flow_label) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_HOP_LIMIT && + src->hop_limit != dst->hop_limit) + return -EINVAL; + if (comp_mask & IB_SA_MCMEMBER_REC_SCOPE && src->scope != dst->scope) + return -EINVAL; + + /* join_state checked separately, proxy_join ignored */ + + return 0; +} + +static int send_join(struct mcast_group *group, struct mcast_member *member) +{ + struct mcast_port *port = group->port; + int ret; + + group->last_join = member; + ret = ib_sa_mcmember_rec_query(&sa_client, port->dev->device, + port->port_num, IB_MGMT_METHOD_SET, + &member->multicast.rec, + member->multicast.comp_mask, + 3000, GFP_KERNEL, join_handler, group, + &group->query); + if (ret >= 0) { + group->query_id = ret; + ret = 0; + } + return ret; +} + +static int send_leave(struct mcast_group *group, u8 leave_state) +{ + struct mcast_port *port = group->port; + struct ib_sa_mcmember_rec rec; + int ret; + + rec = group->rec; + rec.join_state = leave_state; + group->leave_state = leave_state; + + ret = ib_sa_mcmember_rec_query(&sa_client, port->dev->device, + port->port_num, IB_SA_METHOD_DELETE, &rec, + IB_SA_MCMEMBER_REC_MGID | + IB_SA_MCMEMBER_REC_PORT_GID | + IB_SA_MCMEMBER_REC_JOIN_STATE, + 3000, GFP_KERNEL, leave_handler, + group, &group->query); + if (ret >= 0) { + group->query_id = ret; + ret = 0; + } + return ret; +} + +static void join_group(struct mcast_group *group, struct mcast_member *member, + u8 join_state) +{ + member->state = MCAST_MEMBER; + adjust_membership(group, join_state, 1); + group->rec.join_state |= join_state; + member->multicast.rec = group->rec; + member->multicast.rec.join_state = join_state; + list_move(&member->list, &group->active_list); +} + +static int fail_join(struct mcast_group *group, struct mcast_member *member, + int status) +{ + spin_lock_irq(&group->lock); + list_del_init(&member->list); + spin_unlock_irq(&group->lock); + return member->multicast.callback(status, &member->multicast); +} + +static void process_group_error(struct mcast_group *group) +{ + struct mcast_member *member; + int ret = 0; + u16 pkey_index; + + if (group->state == MCAST_PKEY_EVENT) + ret = ib_find_pkey(group->port->dev->device, + group->port->port_num, + be16_to_cpu(group->rec.pkey), &pkey_index); + + spin_lock_irq(&group->lock); + if (group->state == MCAST_PKEY_EVENT && !ret && + group->pkey_index == pkey_index) + goto out; + + while (!list_empty(&group->active_list)) { + member = list_entry(group->active_list.next, + struct mcast_member, list); + atomic_inc(&member->refcount); + list_del_init(&member->list); + adjust_membership(group, member->multicast.rec.join_state, -1); + member->state = MCAST_ERROR; + spin_unlock_irq(&group->lock); + + ret = member->multicast.callback(-ENETRESET, + &member->multicast); + deref_member(member); + if (ret) + ib_sa_free_multicast(&member->multicast); + spin_lock_irq(&group->lock); + } + + group->rec.join_state = 0; +out: + group->state = MCAST_BUSY; + spin_unlock_irq(&group->lock); +} + +static void mcast_work_handler(struct work_struct *work) +{ + struct mcast_group *group; + struct mcast_member *member; + struct ib_sa_multicast *multicast; + int status, ret; + u8 join_state; + + group = container_of(work, typeof(*group), work); +retest: + spin_lock_irq(&group->lock); + while (!list_empty(&group->pending_list) || + (group->state != MCAST_BUSY)) { + + if (group->state != MCAST_BUSY) { + spin_unlock_irq(&group->lock); + process_group_error(group); + goto retest; + } + + member = list_entry(group->pending_list.next, + struct mcast_member, list); + multicast = &member->multicast; + join_state = multicast->rec.join_state; + atomic_inc(&member->refcount); + + if (join_state == (group->rec.join_state & join_state)) { + status = cmp_rec(&group->rec, &multicast->rec, + multicast->comp_mask); + if (!status) + join_group(group, member, join_state); + else + list_del_init(&member->list); + spin_unlock_irq(&group->lock); + ret = multicast->callback(status, multicast); + } else { + spin_unlock_irq(&group->lock); + status = send_join(group, member); + if (!status) { + deref_member(member); + return; + } + ret = fail_join(group, member, status); + } + + deref_member(member); + if (ret) + ib_sa_free_multicast(&member->multicast); + spin_lock_irq(&group->lock); + } + + join_state = get_leave_state(group); + if (join_state) { + group->rec.join_state &= ~join_state; + spin_unlock_irq(&group->lock); + if (send_leave(group, join_state)) + goto retest; + } else { + group->state = MCAST_IDLE; + spin_unlock_irq(&group->lock); + release_group(group); + } +} + +/* + * Fail a join request if it is still active - at the head of the pending queue. + */ +static void process_join_error(struct mcast_group *group, int status) +{ + struct mcast_member *member; + int ret; + + spin_lock_irq(&group->lock); + member = list_entry(group->pending_list.next, + struct mcast_member, list); + if (group->last_join == member) { + atomic_inc(&member->refcount); + list_del_init(&member->list); + spin_unlock_irq(&group->lock); + ret = member->multicast.callback(status, &member->multicast); + deref_member(member); + if (ret) + ib_sa_free_multicast(&member->multicast); + } else + spin_unlock_irq(&group->lock); +} + +static void join_handler(int status, struct ib_sa_mcmember_rec *rec, + void *context) +{ + struct mcast_group *group = context; + u16 pkey_index = MCAST_INVALID_PKEY_INDEX; + + if (status) + process_join_error(group, status); + else { + ib_find_pkey(group->port->dev->device, group->port->port_num, + be16_to_cpu(rec->pkey), &pkey_index); + + spin_lock_irq(&group->port->lock); + group->rec = *rec; + if (group->state == MCAST_BUSY && + group->pkey_index == MCAST_INVALID_PKEY_INDEX) + group->pkey_index = pkey_index; + if (!memcmp(&mgid0, &group->rec.mgid, sizeof mgid0)) { + rb_erase(&group->node, &group->port->table); + mcast_insert(group->port, group, 1); + } + spin_unlock_irq(&group->port->lock); + } + mcast_work_handler(&group->work); +} + +static void leave_handler(int status, struct ib_sa_mcmember_rec *rec, + void *context) +{ + struct mcast_group *group = context; + + if (status && group->retries > 0 && + !send_leave(group, group->leave_state)) + group->retries--; + else + mcast_work_handler(&group->work); +} + +static struct mcast_group *acquire_group(struct mcast_port *port, + union ib_gid *mgid, gfp_t gfp_mask) +{ + struct mcast_group *group, *cur_group; + unsigned long flags; + int is_mgid0; + + is_mgid0 = !memcmp(&mgid0, mgid, sizeof mgid0); + if (!is_mgid0) { + spin_lock_irqsave(&port->lock, flags); + group = mcast_find(port, mgid); + if (group) + goto found; + spin_unlock_irqrestore(&port->lock, flags); + } + + group = kzalloc(sizeof *group, gfp_mask); + if (!group) + return NULL; + + group->retries = 3; + group->port = port; + group->rec.mgid = *mgid; + group->pkey_index = MCAST_INVALID_PKEY_INDEX; + INIT_LIST_HEAD(&group->pending_list); + INIT_LIST_HEAD(&group->active_list); + INIT_WORK(&group->work, mcast_work_handler); + spin_lock_init(&group->lock); + + spin_lock_irqsave(&port->lock, flags); + cur_group = mcast_insert(port, group, is_mgid0); + if (cur_group) { + kfree(group); + group = cur_group; + } else + atomic_inc(&port->refcount); +found: + atomic_inc(&group->refcount); + spin_unlock_irqrestore(&port->lock, flags); + return group; +} + +/* + * We serialize all join requests to a single group to make our lives much + * easier. Otherwise, two users could try to join the same group + * simultaneously, with different configurations, one could leave while the + * join is in progress, etc., which makes locking around error recovery + * difficult. + */ +struct ib_sa_multicast * +ib_sa_join_multicast(struct ib_sa_client *client, + struct ib_device *device, u8 port_num, + struct ib_sa_mcmember_rec *rec, + ib_sa_comp_mask comp_mask, gfp_t gfp_mask, + int (*callback)(int status, + struct ib_sa_multicast *multicast), + void *context) +{ + struct mcast_device *dev; + struct mcast_member *member; + struct ib_sa_multicast *multicast; + int ret; + + dev = ib_get_client_data(device, &mcast_client); + if (!dev) + return ERR_PTR(-ENODEV); + + member = kmalloc(sizeof *member, gfp_mask); + if (!member) + return ERR_PTR(-ENOMEM); + + ib_sa_client_get(client); + member->client = client; + member->multicast.rec = *rec; + member->multicast.comp_mask = comp_mask; + member->multicast.callback = callback; + member->multicast.context = context; + init_completion(&member->comp); + atomic_set(&member->refcount, 1); + member->state = MCAST_JOINING; + + member->group = acquire_group(&dev->port[port_num - dev->start_port], + &rec->mgid, gfp_mask); + if (!member->group) { + ret = -ENOMEM; + goto err; + } + + /* + * The user will get the multicast structure in their callback. They + * could then free the multicast structure before we can return from + * this routine. So we save the pointer to return before queuing + * any callback. + */ + multicast = &member->multicast; + queue_join(member); + return multicast; + +err: + ib_sa_client_put(client); + kfree(member); + return ERR_PTR(ret); +} +EXPORT_SYMBOL(ib_sa_join_multicast); + +void ib_sa_free_multicast(struct ib_sa_multicast *multicast) +{ + struct mcast_member *member; + struct mcast_group *group; + + member = container_of(multicast, struct mcast_member, multicast); + group = member->group; + + spin_lock_irq(&group->lock); + if (member->state == MCAST_MEMBER) + adjust_membership(group, multicast->rec.join_state, -1); + + list_del_init(&member->list); + + if (group->state == MCAST_IDLE) { + group->state = MCAST_BUSY; + spin_unlock_irq(&group->lock); + /* Continue to hold reference on group until callback */ + queue_work(mcast_wq, &group->work); + } else { + spin_unlock_irq(&group->lock); + release_group(group); + } + + deref_member(member); + wait_for_completion(&member->comp); + ib_sa_client_put(member->client); + kfree(member); +} +EXPORT_SYMBOL(ib_sa_free_multicast); + +int ib_sa_get_mcmember_rec(struct ib_device *device, u8 port_num, + union ib_gid *mgid, struct ib_sa_mcmember_rec *rec) +{ + struct mcast_device *dev; + struct mcast_port *port; + struct mcast_group *group; + unsigned long flags; + int ret = 0; + + dev = ib_get_client_data(device, &mcast_client); + if (!dev) + return -ENODEV; + + port = &dev->port[port_num - dev->start_port]; + spin_lock_irqsave(&port->lock, flags); + group = mcast_find(port, mgid); + if (group) + *rec = group->rec; + else + ret = -EADDRNOTAVAIL; + spin_unlock_irqrestore(&port->lock, flags); + + return ret; +} +EXPORT_SYMBOL(ib_sa_get_mcmember_rec); + +int ib_init_ah_from_mcmember(struct ib_device *device, u8 port_num, + struct ib_sa_mcmember_rec *rec, + struct ib_ah_attr *ah_attr) +{ + int ret; + u16 gid_index; + u8 p; + + ret = ib_find_cached_gid(device, &rec->port_gid, &p, &gid_index); + if (ret) + return ret; + + memset(ah_attr, 0, sizeof *ah_attr); + ah_attr->dlid = be16_to_cpu(rec->mlid); + ah_attr->sl = rec->sl; + ah_attr->port_num = port_num; + ah_attr->static_rate = rec->rate; + + ah_attr->ah_flags = IB_AH_GRH; + ah_attr->grh.dgid = rec->mgid; + + ah_attr->grh.sgid_index = (u8) gid_index; + ah_attr->grh.flow_label = be32_to_cpu(rec->flow_label); + ah_attr->grh.hop_limit = rec->hop_limit; + ah_attr->grh.traffic_class = rec->traffic_class; + + return 0; +} +EXPORT_SYMBOL(ib_init_ah_from_mcmember); + +static void mcast_groups_event(struct mcast_port *port, + enum mcast_group_state state) +{ + struct mcast_group *group; + struct rb_node *node; + unsigned long flags; + + spin_lock_irqsave(&port->lock, flags); + for (node = rb_first(&port->table); node; node = rb_next(node)) { + group = rb_entry(node, struct mcast_group, node); + spin_lock(&group->lock); + if (group->state == MCAST_IDLE) { + atomic_inc(&group->refcount); + queue_work(mcast_wq, &group->work); + } + if (group->state != MCAST_GROUP_ERROR) + group->state = state; + spin_unlock(&group->lock); + } + spin_unlock_irqrestore(&port->lock, flags); +} + +static void mcast_event_handler(struct ib_event_handler *handler, + struct ib_event *event) +{ + struct mcast_device *dev; + int index; + + dev = container_of(handler, struct mcast_device, event_handler); + if (rdma_port_get_link_layer(dev->device, event->element.port_num) != + IB_LINK_LAYER_INFINIBAND) + return; + + index = event->element.port_num - dev->start_port; + + switch (event->event) { + case IB_EVENT_PORT_ERR: + case IB_EVENT_LID_CHANGE: + case IB_EVENT_SM_CHANGE: + case IB_EVENT_CLIENT_REREGISTER: + mcast_groups_event(&dev->port[index], MCAST_GROUP_ERROR); + break; + case IB_EVENT_PKEY_CHANGE: + mcast_groups_event(&dev->port[index], MCAST_PKEY_EVENT); + break; + default: + break; + } +} + +static void mcast_add_one(struct ib_device *device) +{ + struct mcast_device *dev; + struct mcast_port *port; + int i; + int count = 0; + + if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) + return; + + dev = kmalloc(sizeof *dev + device->phys_port_cnt * sizeof *port, + GFP_KERNEL); + if (!dev) + return; + + if (device->node_type == RDMA_NODE_IB_SWITCH) + dev->start_port = dev->end_port = 0; + else { + dev->start_port = 1; + dev->end_port = device->phys_port_cnt; + } + + for (i = 0; i <= dev->end_port - dev->start_port; i++) { + if (rdma_port_get_link_layer(device, dev->start_port + i) != + IB_LINK_LAYER_INFINIBAND) + continue; + port = &dev->port[i]; + port->dev = dev; + port->port_num = dev->start_port + i; + spin_lock_init(&port->lock); + port->table = RB_ROOT; + init_completion(&port->comp); + atomic_set(&port->refcount, 1); + ++count; + } + + if (!count) { + kfree(dev); + return; + } + + dev->device = device; + ib_set_client_data(device, &mcast_client, dev); + + INIT_IB_EVENT_HANDLER(&dev->event_handler, device, mcast_event_handler); + ib_register_event_handler(&dev->event_handler); +} + +static void mcast_remove_one(struct ib_device *device) +{ + struct mcast_device *dev; + struct mcast_port *port; + int i; + + dev = ib_get_client_data(device, &mcast_client); + if (!dev) + return; + + ib_unregister_event_handler(&dev->event_handler); + flush_workqueue(mcast_wq); + + for (i = 0; i <= dev->end_port - dev->start_port; i++) { + if (rdma_port_get_link_layer(device, dev->start_port + i) == + IB_LINK_LAYER_INFINIBAND) { + port = &dev->port[i]; + deref_port(port); + wait_for_completion(&port->comp); + } + } + + kfree(dev); +} + +int mcast_init(void) +{ + int ret; + + mcast_wq = create_singlethread_workqueue("ib_mcast"); + if (!mcast_wq) + return -ENOMEM; + + ib_sa_register_client(&sa_client); + + ret = ib_register_client(&mcast_client); + if (ret) + goto err; + return 0; + +err: + ib_sa_unregister_client(&sa_client); + destroy_workqueue(mcast_wq); + return ret; +} + +void mcast_cleanup(void) +{ + ib_unregister_client(&mcast_client); + ib_sa_unregister_client(&sa_client); + destroy_workqueue(mcast_wq); +} diff --git a/drivers/infiniband/core/netlink.c b/drivers/infiniband/core/netlink.c new file mode 100644 index 00000000000..23dd5a5c759 --- /dev/null +++ b/drivers/infiniband/core/netlink.c @@ -0,0 +1,216 @@ +/* + * Copyright (c) 2010 Voltaire Inc. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#define pr_fmt(fmt) "%s:%s: " fmt, KBUILD_MODNAME, __func__ + +#include <linux/export.h> +#include <net/netlink.h> +#include <net/net_namespace.h> +#include <net/sock.h> +#include <rdma/rdma_netlink.h> + +struct ibnl_client { + struct list_head list; + int index; + int nops; + const struct ibnl_client_cbs *cb_table; +}; + +static DEFINE_MUTEX(ibnl_mutex); +static struct sock *nls; +static LIST_HEAD(client_list); + +int ibnl_add_client(int index, int nops, + const struct ibnl_client_cbs cb_table[]) +{ + struct ibnl_client *cur; + struct ibnl_client *nl_client; + + nl_client = kmalloc(sizeof *nl_client, GFP_KERNEL); + if (!nl_client) + return -ENOMEM; + + nl_client->index = index; + nl_client->nops = nops; + nl_client->cb_table = cb_table; + + mutex_lock(&ibnl_mutex); + + list_for_each_entry(cur, &client_list, list) { + if (cur->index == index) { + pr_warn("Client for %d already exists\n", index); + mutex_unlock(&ibnl_mutex); + kfree(nl_client); + return -EINVAL; + } + } + + list_add_tail(&nl_client->list, &client_list); + + mutex_unlock(&ibnl_mutex); + + return 0; +} +EXPORT_SYMBOL(ibnl_add_client); + +int ibnl_remove_client(int index) +{ + struct ibnl_client *cur, *next; + + mutex_lock(&ibnl_mutex); + list_for_each_entry_safe(cur, next, &client_list, list) { + if (cur->index == index) { + list_del(&(cur->list)); + mutex_unlock(&ibnl_mutex); + kfree(cur); + return 0; + } + } + pr_warn("Can't remove callback for client idx %d. Not found\n", index); + mutex_unlock(&ibnl_mutex); + + return -EINVAL; +} +EXPORT_SYMBOL(ibnl_remove_client); + +void *ibnl_put_msg(struct sk_buff *skb, struct nlmsghdr **nlh, int seq, + int len, int client, int op, int flags) +{ + unsigned char *prev_tail; + + prev_tail = skb_tail_pointer(skb); + *nlh = nlmsg_put(skb, 0, seq, RDMA_NL_GET_TYPE(client, op), + len, flags); + if (!*nlh) + goto out_nlmsg_trim; + (*nlh)->nlmsg_len = skb_tail_pointer(skb) - prev_tail; + return nlmsg_data(*nlh); + +out_nlmsg_trim: + nlmsg_trim(skb, prev_tail); + return NULL; +} +EXPORT_SYMBOL(ibnl_put_msg); + +int ibnl_put_attr(struct sk_buff *skb, struct nlmsghdr *nlh, + int len, void *data, int type) +{ + unsigned char *prev_tail; + + prev_tail = skb_tail_pointer(skb); + if (nla_put(skb, type, len, data)) + goto nla_put_failure; + nlh->nlmsg_len += skb_tail_pointer(skb) - prev_tail; + return 0; + +nla_put_failure: + nlmsg_trim(skb, prev_tail - nlh->nlmsg_len); + return -EMSGSIZE; +} +EXPORT_SYMBOL(ibnl_put_attr); + +static int ibnl_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh) +{ + struct ibnl_client *client; + int type = nlh->nlmsg_type; + int index = RDMA_NL_GET_CLIENT(type); + int op = RDMA_NL_GET_OP(type); + + list_for_each_entry(client, &client_list, list) { + if (client->index == index) { + if (op < 0 || op >= client->nops || + !client->cb_table[op].dump) + return -EINVAL; + + { + struct netlink_dump_control c = { + .dump = client->cb_table[op].dump, + .module = client->cb_table[op].module, + }; + return netlink_dump_start(nls, skb, nlh, &c); + } + } + } + + pr_info("Index %d wasn't found in client list\n", index); + return -EINVAL; +} + +static void ibnl_rcv(struct sk_buff *skb) +{ + mutex_lock(&ibnl_mutex); + netlink_rcv_skb(skb, &ibnl_rcv_msg); + mutex_unlock(&ibnl_mutex); +} + +int ibnl_unicast(struct sk_buff *skb, struct nlmsghdr *nlh, + __u32 pid) +{ + return nlmsg_unicast(nls, skb, pid); +} +EXPORT_SYMBOL(ibnl_unicast); + +int ibnl_multicast(struct sk_buff *skb, struct nlmsghdr *nlh, + unsigned int group, gfp_t flags) +{ + return nlmsg_multicast(nls, skb, 0, group, flags); +} +EXPORT_SYMBOL(ibnl_multicast); + +int __init ibnl_init(void) +{ + struct netlink_kernel_cfg cfg = { + .input = ibnl_rcv, + }; + + nls = netlink_kernel_create(&init_net, NETLINK_RDMA, &cfg); + if (!nls) { + pr_warn("Failed to create netlink socket\n"); + return -ENOMEM; + } + + return 0; +} + +void ibnl_cleanup(void) +{ + struct ibnl_client *cur, *next; + + mutex_lock(&ibnl_mutex); + list_for_each_entry_safe(cur, next, &client_list, list) { + list_del(&(cur->list)); + kfree(cur); + } + mutex_unlock(&ibnl_mutex); + + netlink_kernel_release(nls); +} diff --git a/drivers/infiniband/core/packer.c b/drivers/infiniband/core/packer.c index c972d723576..1b65986c0be 100644 --- a/drivers/infiniband/core/packer.c +++ b/drivers/infiniband/core/packer.c @@ -29,10 +29,9 @@ * 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. - * - * $Id: packer.c 1349 2004-12-16 21:09:43Z roland $ */ +#include <linux/export.h> #include <linux/string.h> #include <rdma/ib_pack.h> diff --git a/drivers/infiniband/core/sa.h b/drivers/infiniband/core/sa.h new file mode 100644 index 00000000000..b1d4bbf4ce5 --- /dev/null +++ b/drivers/infiniband/core/sa.h @@ -0,0 +1,66 @@ +/* + * Copyright (c) 2004 Topspin Communications. All rights reserved. + * Copyright (c) 2005 Voltaire, Inc. All rights reserved. + * Copyright (c) 2006 Intel Corporation. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#ifndef SA_H +#define SA_H + +#include <rdma/ib_sa.h> + +static inline void ib_sa_client_get(struct ib_sa_client *client) +{ + atomic_inc(&client->users); +} + +static inline void ib_sa_client_put(struct ib_sa_client *client) +{ + if (atomic_dec_and_test(&client->users)) + complete(&client->comp); +} + +int ib_sa_mcmember_rec_query(struct ib_sa_client *client, + struct ib_device *device, u8 port_num, + u8 method, + struct ib_sa_mcmember_rec *rec, + ib_sa_comp_mask comp_mask, + int timeout_ms, gfp_t gfp_mask, + void (*callback)(int status, + struct ib_sa_mcmember_rec *resp, + void *context), + void *context, + struct ib_sa_query **sa_query); + +int mcast_init(void); +void mcast_cleanup(void); + +#endif /* SA_H */ diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c index acda7d63d6f..233eaf541f5 100644 --- a/drivers/infiniband/core/sa_query.c +++ b/drivers/infiniband/core/sa_query.c @@ -1,6 +1,7 @@ /* * Copyright (c) 2004 Topspin Communications. All rights reserved. - * Copyright (c) 2005 Voltaire, Inc. All rights reserved. + * Copyright (c) 2005 Voltaire, Inc. All rights reserved. + * Copyright (c) 2006 Intel Corporation. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -29,8 +30,6 @@ * 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. - * - * $Id: sa_query.c 2811 2005-07-06 18:11:43Z halr $ */ #include <linux/module.h> @@ -39,14 +38,14 @@ #include <linux/random.h> #include <linux/spinlock.h> #include <linux/slab.h> -#include <linux/pci.h> #include <linux/dma-mapping.h> #include <linux/kref.h> #include <linux/idr.h> #include <linux/workqueue.h> - +#include <uapi/linux/if_ether.h> #include <rdma/ib_pack.h> -#include <rdma/ib_sa.h> +#include <rdma/ib_cache.h> +#include "sa.h" MODULE_AUTHOR("Roland Dreier"); MODULE_DESCRIPTION("InfiniBand subnet administration query support"); @@ -55,6 +54,8 @@ MODULE_LICENSE("Dual BSD/GPL"); struct ib_sa_sm_ah { struct ib_ah *ah; struct kref ref; + u16 pkey_index; + u8 src_path_mask; }; struct ib_sa_port { @@ -74,6 +75,7 @@ struct ib_sa_device { struct ib_sa_query { void (*callback)(struct ib_sa_query *, int, struct ib_sa_mad *); void (*release)(struct ib_sa_query *); + struct ib_sa_client *client; struct ib_sa_port *port; struct ib_mad_send_buf *mad_buf; struct ib_sa_sm_ah *sm_ah; @@ -92,6 +94,12 @@ struct ib_sa_path_query { struct ib_sa_query sa_query; }; +struct ib_sa_guidinfo_query { + void (*callback)(int, struct ib_sa_guidinfo_rec *, void *); + void *context; + struct ib_sa_query sa_query; +}; + struct ib_sa_mcmember_query { void (*callback)(int, struct ib_sa_mcmember_rec *, void *); void *context; @@ -107,10 +115,10 @@ static struct ib_client sa_client = { .remove = ib_sa_remove_one }; -static spinlock_t idr_lock; +static DEFINE_SPINLOCK(idr_lock); static DEFINE_IDR(query_idr); -static spinlock_t tid_lock; +static DEFINE_SPINLOCK(tid_lock); static u32 tid; #define PATH_REC_FIELD(field) \ @@ -119,14 +127,10 @@ static u32 tid; .field_name = "sa_path_rec:" #field static const struct ib_field path_rec_table[] = { - { RESERVED, + { PATH_REC_FIELD(service_id), .offset_words = 0, .offset_bits = 0, - .size_bits = 32 }, - { RESERVED, - .offset_words = 1, - .offset_bits = 0, - .size_bits = 32 }, + .size_bits = 64 }, { PATH_REC_FIELD(dgid), .offset_words = 2, .offset_bits = 0, @@ -175,7 +179,7 @@ static const struct ib_field path_rec_table[] = { .offset_words = 12, .offset_bits = 16, .size_bits = 16 }, - { RESERVED, + { PATH_REC_FIELD(qos_class), .offset_words = 13, .offset_bits = 0, .size_bits = 12 }, @@ -349,6 +353,34 @@ static const struct ib_field service_rec_table[] = { .size_bits = 2*64 }, }; +#define GUIDINFO_REC_FIELD(field) \ + .struct_offset_bytes = offsetof(struct ib_sa_guidinfo_rec, field), \ + .struct_size_bytes = sizeof((struct ib_sa_guidinfo_rec *) 0)->field, \ + .field_name = "sa_guidinfo_rec:" #field + +static const struct ib_field guidinfo_rec_table[] = { + { GUIDINFO_REC_FIELD(lid), + .offset_words = 0, + .offset_bits = 0, + .size_bits = 16 }, + { GUIDINFO_REC_FIELD(block_num), + .offset_words = 0, + .offset_bits = 16, + .size_bits = 8 }, + { GUIDINFO_REC_FIELD(res1), + .offset_words = 0, + .offset_bits = 24, + .size_bits = 8 }, + { GUIDINFO_REC_FIELD(res2), + .offset_words = 1, + .offset_bits = 0, + .size_bits = 32 }, + { GUIDINFO_REC_FIELD(guid_info_list), + .offset_words = 2, + .offset_bits = 0, + .size_bits = 512 }, +}; + static void free_sm_ah(struct kref *kref) { struct ib_sa_sm_ah *sm_ah = container_of(kref, struct ib_sa_sm_ah, ref); @@ -357,10 +389,11 @@ static void free_sm_ah(struct kref *kref) kfree(sm_ah); } -static void update_sm_ah(void *port_ptr) +static void update_sm_ah(struct work_struct *work) { - struct ib_sa_port *port = port_ptr; - struct ib_sa_sm_ah *new_ah, *old_ah; + struct ib_sa_port *port = + container_of(work, struct ib_sa_port, update_task); + struct ib_sa_sm_ah *new_ah; struct ib_port_attr port_attr; struct ib_ah_attr ah_attr; @@ -376,6 +409,12 @@ static void update_sm_ah(void *port_ptr) } kref_init(&new_ah->ref); + new_ah->src_path_mask = (1 << port_attr.lmc) - 1; + + new_ah->pkey_index = 0; + if (ib_find_pkey(port->agent->device, port->port_num, + IB_DEFAULT_PKEY_FULL, &new_ah->pkey_index)) + printk(KERN_ERR "Couldn't find index for default PKey\n"); memset(&ah_attr, 0, sizeof ah_attr); ah_attr.dlid = port_attr.sm_lid; @@ -390,12 +429,11 @@ static void update_sm_ah(void *port_ptr) } spin_lock_irq(&port->ah_lock); - old_ah = port->sm_ah; + if (port->sm_ah) + kref_put(&port->sm_ah->ref, free_sm_ah); port->sm_ah = new_ah; spin_unlock_irq(&port->ah_lock); - if (old_ah) - kref_put(&old_ah->ref, free_sm_ah); } static void ib_sa_event(struct ib_event_handler *handler, struct ib_event *event) @@ -404,15 +442,42 @@ static void ib_sa_event(struct ib_event_handler *handler, struct ib_event *event event->event == IB_EVENT_PORT_ACTIVE || event->event == IB_EVENT_LID_CHANGE || event->event == IB_EVENT_PKEY_CHANGE || - event->event == IB_EVENT_SM_CHANGE) { - struct ib_sa_device *sa_dev; - sa_dev = container_of(handler, typeof(*sa_dev), event_handler); - - schedule_work(&sa_dev->port[event->element.port_num - + event->event == IB_EVENT_SM_CHANGE || + event->event == IB_EVENT_CLIENT_REREGISTER) { + unsigned long flags; + struct ib_sa_device *sa_dev = + container_of(handler, typeof(*sa_dev), event_handler); + struct ib_sa_port *port = + &sa_dev->port[event->element.port_num - sa_dev->start_port]; + + if (rdma_port_get_link_layer(handler->device, port->port_num) != IB_LINK_LAYER_INFINIBAND) + return; + + spin_lock_irqsave(&port->ah_lock, flags); + if (port->sm_ah) + kref_put(&port->sm_ah->ref, free_sm_ah); + port->sm_ah = NULL; + spin_unlock_irqrestore(&port->ah_lock, flags); + + queue_work(ib_wq, &sa_dev->port[event->element.port_num - sa_dev->start_port].update_task); } } +void ib_sa_register_client(struct ib_sa_client *client) +{ + atomic_set(&client->users, 1); + init_completion(&client->comp); +} +EXPORT_SYMBOL(ib_sa_register_client); + +void ib_sa_unregister_client(struct ib_sa_client *client) +{ + ib_sa_client_put(client); + wait_for_completion(&client->comp); +} +EXPORT_SYMBOL(ib_sa_unregister_client); + /** * ib_sa_cancel_query - try to cancel an SA query * @id:ID of query to cancel @@ -441,6 +506,100 @@ void ib_sa_cancel_query(int id, struct ib_sa_query *query) } EXPORT_SYMBOL(ib_sa_cancel_query); +static u8 get_src_path_mask(struct ib_device *device, u8 port_num) +{ + struct ib_sa_device *sa_dev; + struct ib_sa_port *port; + unsigned long flags; + u8 src_path_mask; + + sa_dev = ib_get_client_data(device, &sa_client); + if (!sa_dev) + return 0x7f; + + port = &sa_dev->port[port_num - sa_dev->start_port]; + spin_lock_irqsave(&port->ah_lock, flags); + src_path_mask = port->sm_ah ? port->sm_ah->src_path_mask : 0x7f; + spin_unlock_irqrestore(&port->ah_lock, flags); + + return src_path_mask; +} + +int ib_init_ah_from_path(struct ib_device *device, u8 port_num, + struct ib_sa_path_rec *rec, struct ib_ah_attr *ah_attr) +{ + int ret; + u16 gid_index; + int force_grh; + + memset(ah_attr, 0, sizeof *ah_attr); + ah_attr->dlid = be16_to_cpu(rec->dlid); + ah_attr->sl = rec->sl; + ah_attr->src_path_bits = be16_to_cpu(rec->slid) & + get_src_path_mask(device, port_num); + ah_attr->port_num = port_num; + ah_attr->static_rate = rec->rate; + + force_grh = rdma_port_get_link_layer(device, port_num) == IB_LINK_LAYER_ETHERNET; + + if (rec->hop_limit > 1 || force_grh) { + ah_attr->ah_flags = IB_AH_GRH; + ah_attr->grh.dgid = rec->dgid; + + ret = ib_find_cached_gid(device, &rec->sgid, &port_num, + &gid_index); + if (ret) + return ret; + + ah_attr->grh.sgid_index = gid_index; + ah_attr->grh.flow_label = be32_to_cpu(rec->flow_label); + ah_attr->grh.hop_limit = rec->hop_limit; + ah_attr->grh.traffic_class = rec->traffic_class; + } + if (force_grh) { + memcpy(ah_attr->dmac, rec->dmac, ETH_ALEN); + ah_attr->vlan_id = rec->vlan_id; + } else { + ah_attr->vlan_id = 0xffff; + } + + return 0; +} +EXPORT_SYMBOL(ib_init_ah_from_path); + +static int alloc_mad(struct ib_sa_query *query, gfp_t gfp_mask) +{ + unsigned long flags; + + spin_lock_irqsave(&query->port->ah_lock, flags); + if (!query->port->sm_ah) { + spin_unlock_irqrestore(&query->port->ah_lock, flags); + return -EAGAIN; + } + kref_get(&query->port->sm_ah->ref); + query->sm_ah = query->port->sm_ah; + spin_unlock_irqrestore(&query->port->ah_lock, flags); + + query->mad_buf = ib_create_send_mad(query->port->agent, 1, + query->sm_ah->pkey_index, + 0, IB_MGMT_SA_HDR, IB_MGMT_SA_DATA, + gfp_mask); + if (IS_ERR(query->mad_buf)) { + kref_put(&query->sm_ah->ref, free_sm_ah); + return -ENOMEM; + } + + query->mad_buf->ah = query->sm_ah->ah; + + return 0; +} + +static void free_mad(struct ib_sa_query *query) +{ + ib_free_send_mad(query->mad_buf); + kref_put(&query->sm_ah->ref, free_sm_ah); +} + static void init_mad(struct ib_sa_mad *mad, struct ib_mad_agent *agent) { unsigned long flags; @@ -457,40 +616,33 @@ static void init_mad(struct ib_sa_mad *mad, struct ib_mad_agent *agent) spin_unlock_irqrestore(&tid_lock, flags); } -static int send_mad(struct ib_sa_query *query, int timeout_ms) +static int send_mad(struct ib_sa_query *query, int timeout_ms, gfp_t gfp_mask) { + bool preload = !!(gfp_mask & __GFP_WAIT); unsigned long flags; int ret, id; -retry: - if (!idr_pre_get(&query_idr, GFP_ATOMIC)) - return -ENOMEM; + if (preload) + idr_preload(gfp_mask); spin_lock_irqsave(&idr_lock, flags); - ret = idr_get_new(&query_idr, query, &id); + + id = idr_alloc(&query_idr, query, 0, 0, GFP_NOWAIT); + spin_unlock_irqrestore(&idr_lock, flags); - if (ret == -EAGAIN) - goto retry; - if (ret) - return ret; + if (preload) + idr_preload_end(); + if (id < 0) + return id; query->mad_buf->timeout_ms = timeout_ms; query->mad_buf->context[0] = query; query->id = id; - spin_lock_irqsave(&query->port->ah_lock, flags); - kref_get(&query->port->sm_ah->ref); - query->sm_ah = query->port->sm_ah; - spin_unlock_irqrestore(&query->port->ah_lock, flags); - - query->mad_buf->ah = query->sm_ah->ah; - ret = ib_post_send_mad(query->mad_buf, NULL); if (ret) { spin_lock_irqsave(&idr_lock, flags); idr_remove(&query_idr, id); spin_unlock_irqrestore(&idr_lock, flags); - - kref_put(&query->sm_ah->ref, free_sm_ah); } /* @@ -501,6 +653,18 @@ retry: return ret ? ret : id; } +void ib_sa_unpack_path(void *attribute, struct ib_sa_path_rec *rec) +{ + ib_unpack(path_rec_table, ARRAY_SIZE(path_rec_table), attribute, rec); +} +EXPORT_SYMBOL(ib_sa_unpack_path); + +void ib_sa_pack_path(struct ib_sa_path_rec *rec, void *attribute) +{ + ib_pack(path_rec_table, ARRAY_SIZE(path_rec_table), rec, attribute); +} +EXPORT_SYMBOL(ib_sa_pack_path); + static void ib_sa_path_rec_callback(struct ib_sa_query *sa_query, int status, struct ib_sa_mad *mad) @@ -513,6 +677,9 @@ static void ib_sa_path_rec_callback(struct ib_sa_query *sa_query, ib_unpack(path_rec_table, ARRAY_SIZE(path_rec_table), mad->data, &rec); + rec.vlan_id = 0xffff; + memset(rec.dmac, 0, ETH_ALEN); + memset(rec.smac, 0, ETH_ALEN); query->callback(status, &rec, query->context); } else query->callback(status, NULL, query->context); @@ -525,6 +692,7 @@ static void ib_sa_path_rec_release(struct ib_sa_query *sa_query) /** * ib_sa_path_rec_get - Start a Path get query + * @client:SA client * @device:device to send query on * @port_num: port number to send query on * @rec:Path Record to send in query @@ -547,7 +715,8 @@ static void ib_sa_path_rec_release(struct ib_sa_query *sa_query) * error code. Otherwise it is a query ID that can be used to cancel * the query. */ -int ib_sa_path_rec_get(struct ib_device *device, u8 port_num, +int ib_sa_path_rec_get(struct ib_sa_client *client, + struct ib_device *device, u8 port_num, struct ib_sa_path_rec *rec, ib_sa_comp_mask comp_mask, int timeout_ms, gfp_t gfp_mask, @@ -574,23 +743,21 @@ int ib_sa_path_rec_get(struct ib_device *device, u8 port_num, if (!query) return -ENOMEM; - query->sa_query.mad_buf = ib_create_send_mad(agent, 1, 0, - 0, IB_MGMT_SA_HDR, - IB_MGMT_SA_DATA, gfp_mask); - if (!query->sa_query.mad_buf) { - ret = -ENOMEM; + query->sa_query.port = port; + ret = alloc_mad(&query->sa_query, gfp_mask); + if (ret) goto err1; - } - query->callback = callback; - query->context = context; + ib_sa_client_get(client); + query->sa_query.client = client; + query->callback = callback; + query->context = context; mad = query->sa_query.mad_buf->mad; init_mad(mad, agent); query->sa_query.callback = callback ? ib_sa_path_rec_callback : NULL; query->sa_query.release = ib_sa_path_rec_release; - query->sa_query.port = port; mad->mad_hdr.method = IB_MGMT_METHOD_GET; mad->mad_hdr.attr_id = cpu_to_be16(IB_SA_ATTR_PATH_REC); mad->sa_hdr.comp_mask = comp_mask; @@ -599,7 +766,7 @@ int ib_sa_path_rec_get(struct ib_device *device, u8 port_num, *sa_query = &query->sa_query; - ret = send_mad(&query->sa_query, timeout_ms); + ret = send_mad(&query->sa_query, timeout_ms, gfp_mask); if (ret < 0) goto err2; @@ -607,7 +774,8 @@ int ib_sa_path_rec_get(struct ib_device *device, u8 port_num, err2: *sa_query = NULL; - ib_free_send_mad(query->sa_query.mad_buf); + ib_sa_client_put(query->sa_query.client); + free_mad(&query->sa_query); err1: kfree(query); @@ -639,6 +807,7 @@ static void ib_sa_service_rec_release(struct ib_sa_query *sa_query) /** * ib_sa_service_rec_query - Start Service Record operation + * @client:SA client * @device:device to send request on * @port_num: port number to send request on * @method:SA method - should be get, set, or delete @@ -663,7 +832,8 @@ static void ib_sa_service_rec_release(struct ib_sa_query *sa_query) * error code. Otherwise it is a request ID that can be used to cancel * the query. */ -int ib_sa_service_rec_query(struct ib_device *device, u8 port_num, u8 method, +int ib_sa_service_rec_query(struct ib_sa_client *client, + struct ib_device *device, u8 port_num, u8 method, struct ib_sa_service_rec *rec, ib_sa_comp_mask comp_mask, int timeout_ms, gfp_t gfp_mask, @@ -695,23 +865,21 @@ int ib_sa_service_rec_query(struct ib_device *device, u8 port_num, u8 method, if (!query) return -ENOMEM; - query->sa_query.mad_buf = ib_create_send_mad(agent, 1, 0, - 0, IB_MGMT_SA_HDR, - IB_MGMT_SA_DATA, gfp_mask); - if (!query->sa_query.mad_buf) { - ret = -ENOMEM; + query->sa_query.port = port; + ret = alloc_mad(&query->sa_query, gfp_mask); + if (ret) goto err1; - } - query->callback = callback; - query->context = context; + ib_sa_client_get(client); + query->sa_query.client = client; + query->callback = callback; + query->context = context; mad = query->sa_query.mad_buf->mad; init_mad(mad, agent); query->sa_query.callback = callback ? ib_sa_service_rec_callback : NULL; query->sa_query.release = ib_sa_service_rec_release; - query->sa_query.port = port; mad->mad_hdr.method = method; mad->mad_hdr.attr_id = cpu_to_be16(IB_SA_ATTR_SERVICE_REC); mad->sa_hdr.comp_mask = comp_mask; @@ -721,7 +889,7 @@ int ib_sa_service_rec_query(struct ib_device *device, u8 port_num, u8 method, *sa_query = &query->sa_query; - ret = send_mad(&query->sa_query, timeout_ms); + ret = send_mad(&query->sa_query, timeout_ms, gfp_mask); if (ret < 0) goto err2; @@ -729,7 +897,8 @@ int ib_sa_service_rec_query(struct ib_device *device, u8 port_num, u8 method, err2: *sa_query = NULL; - ib_free_send_mad(query->sa_query.mad_buf); + ib_sa_client_put(query->sa_query.client); + free_mad(&query->sa_query); err1: kfree(query); @@ -759,7 +928,8 @@ static void ib_sa_mcmember_rec_release(struct ib_sa_query *sa_query) kfree(container_of(sa_query, struct ib_sa_mcmember_query, sa_query)); } -int ib_sa_mcmember_rec_query(struct ib_device *device, u8 port_num, +int ib_sa_mcmember_rec_query(struct ib_sa_client *client, + struct ib_device *device, u8 port_num, u8 method, struct ib_sa_mcmember_rec *rec, ib_sa_comp_mask comp_mask, @@ -787,23 +957,21 @@ int ib_sa_mcmember_rec_query(struct ib_device *device, u8 port_num, if (!query) return -ENOMEM; - query->sa_query.mad_buf = ib_create_send_mad(agent, 1, 0, - 0, IB_MGMT_SA_HDR, - IB_MGMT_SA_DATA, gfp_mask); - if (!query->sa_query.mad_buf) { - ret = -ENOMEM; + query->sa_query.port = port; + ret = alloc_mad(&query->sa_query, gfp_mask); + if (ret) goto err1; - } - query->callback = callback; - query->context = context; + ib_sa_client_get(client); + query->sa_query.client = client; + query->callback = callback; + query->context = context; mad = query->sa_query.mad_buf->mad; init_mad(mad, agent); query->sa_query.callback = callback ? ib_sa_mcmember_rec_callback : NULL; query->sa_query.release = ib_sa_mcmember_rec_release; - query->sa_query.port = port; mad->mad_hdr.method = method; mad->mad_hdr.attr_id = cpu_to_be16(IB_SA_ATTR_MC_MEMBER_REC); mad->sa_hdr.comp_mask = comp_mask; @@ -813,7 +981,105 @@ int ib_sa_mcmember_rec_query(struct ib_device *device, u8 port_num, *sa_query = &query->sa_query; - ret = send_mad(&query->sa_query, timeout_ms); + ret = send_mad(&query->sa_query, timeout_ms, gfp_mask); + if (ret < 0) + goto err2; + + return ret; + +err2: + *sa_query = NULL; + ib_sa_client_put(query->sa_query.client); + free_mad(&query->sa_query); + +err1: + kfree(query); + return ret; +} + +/* Support GuidInfoRecord */ +static void ib_sa_guidinfo_rec_callback(struct ib_sa_query *sa_query, + int status, + struct ib_sa_mad *mad) +{ + struct ib_sa_guidinfo_query *query = + container_of(sa_query, struct ib_sa_guidinfo_query, sa_query); + + if (mad) { + struct ib_sa_guidinfo_rec rec; + + ib_unpack(guidinfo_rec_table, ARRAY_SIZE(guidinfo_rec_table), + mad->data, &rec); + query->callback(status, &rec, query->context); + } else + query->callback(status, NULL, query->context); +} + +static void ib_sa_guidinfo_rec_release(struct ib_sa_query *sa_query) +{ + kfree(container_of(sa_query, struct ib_sa_guidinfo_query, sa_query)); +} + +int ib_sa_guid_info_rec_query(struct ib_sa_client *client, + struct ib_device *device, u8 port_num, + struct ib_sa_guidinfo_rec *rec, + ib_sa_comp_mask comp_mask, u8 method, + int timeout_ms, gfp_t gfp_mask, + void (*callback)(int status, + struct ib_sa_guidinfo_rec *resp, + void *context), + void *context, + struct ib_sa_query **sa_query) +{ + struct ib_sa_guidinfo_query *query; + struct ib_sa_device *sa_dev = ib_get_client_data(device, &sa_client); + struct ib_sa_port *port; + struct ib_mad_agent *agent; + struct ib_sa_mad *mad; + int ret; + + if (!sa_dev) + return -ENODEV; + + if (method != IB_MGMT_METHOD_GET && + method != IB_MGMT_METHOD_SET && + method != IB_SA_METHOD_DELETE) { + return -EINVAL; + } + + port = &sa_dev->port[port_num - sa_dev->start_port]; + agent = port->agent; + + query = kmalloc(sizeof *query, gfp_mask); + if (!query) + return -ENOMEM; + + query->sa_query.port = port; + ret = alloc_mad(&query->sa_query, gfp_mask); + if (ret) + goto err1; + + ib_sa_client_get(client); + query->sa_query.client = client; + query->callback = callback; + query->context = context; + + mad = query->sa_query.mad_buf->mad; + init_mad(mad, agent); + + query->sa_query.callback = callback ? ib_sa_guidinfo_rec_callback : NULL; + query->sa_query.release = ib_sa_guidinfo_rec_release; + + mad->mad_hdr.method = method; + mad->mad_hdr.attr_id = cpu_to_be16(IB_SA_ATTR_GUID_INFO_REC); + mad->sa_hdr.comp_mask = comp_mask; + + ib_pack(guidinfo_rec_table, ARRAY_SIZE(guidinfo_rec_table), rec, + mad->data); + + *sa_query = &query->sa_query; + + ret = send_mad(&query->sa_query, timeout_ms, gfp_mask); if (ret < 0) goto err2; @@ -821,13 +1087,14 @@ int ib_sa_mcmember_rec_query(struct ib_device *device, u8 port_num, err2: *sa_query = NULL; - ib_free_send_mad(query->sa_query.mad_buf); + ib_sa_client_put(query->sa_query.client); + free_mad(&query->sa_query); err1: kfree(query); return ret; } -EXPORT_SYMBOL(ib_sa_mcmember_rec_query); +EXPORT_SYMBOL(ib_sa_guid_info_rec_query); static void send_handler(struct ib_mad_agent *agent, struct ib_mad_send_wc *mad_send_wc) @@ -855,8 +1122,8 @@ static void send_handler(struct ib_mad_agent *agent, idr_remove(&query_idr, query->id); spin_unlock_irqrestore(&idr_lock, flags); - ib_free_send_mad(mad_send_wc->send_buf); - kref_put(&query->sm_ah->ref, free_sm_ah); + free_mad(query); + ib_sa_client_put(query->client); query->release(query); } @@ -887,14 +1154,17 @@ static void ib_sa_add_one(struct ib_device *device) struct ib_sa_device *sa_dev; int s, e, i; - if (device->node_type == IB_NODE_SWITCH) + if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) + return; + + if (device->node_type == RDMA_NODE_IB_SWITCH) s = e = 0; else { s = 1; e = device->phys_port_cnt; } - sa_dev = kmalloc(sizeof *sa_dev + + sa_dev = kzalloc(sizeof *sa_dev + (e - s + 1) * sizeof (struct ib_sa_port), GFP_KERNEL); if (!sa_dev) @@ -904,9 +1174,12 @@ static void ib_sa_add_one(struct ib_device *device) sa_dev->end_port = e; for (i = 0; i <= e - s; ++i) { + spin_lock_init(&sa_dev->port[i].ah_lock); + if (rdma_port_get_link_layer(device, i + 1) != IB_LINK_LAYER_INFINIBAND) + continue; + sa_dev->port[i].sm_ah = NULL; sa_dev->port[i].port_num = i + s; - spin_lock_init(&sa_dev->port[i].ah_lock); sa_dev->port[i].agent = ib_register_mad_agent(device, i + s, IB_QPT_GSI, @@ -915,8 +1188,7 @@ static void ib_sa_add_one(struct ib_device *device) if (IS_ERR(sa_dev->port[i].agent)) goto err; - INIT_WORK(&sa_dev->port[i].update_task, - update_sm_ah, &sa_dev->port[i]); + INIT_WORK(&sa_dev->port[i].update_task, update_sm_ah); } ib_set_client_data(device, &sa_client, sa_dev); @@ -933,13 +1205,15 @@ static void ib_sa_add_one(struct ib_device *device) goto err; for (i = 0; i <= e - s; ++i) - update_sm_ah(&sa_dev->port[i]); + if (rdma_port_get_link_layer(device, i + 1) == IB_LINK_LAYER_INFINIBAND) + update_sm_ah(&sa_dev->port[i].update_task); return; err: while (--i >= 0) - ib_unregister_mad_agent(sa_dev->port[i].agent); + if (rdma_port_get_link_layer(device, i + 1) == IB_LINK_LAYER_INFINIBAND) + ib_unregister_mad_agent(sa_dev->port[i].agent); kfree(sa_dev); @@ -956,9 +1230,15 @@ static void ib_sa_remove_one(struct ib_device *device) ib_unregister_event_handler(&sa_dev->event_handler); + flush_workqueue(ib_wq); + for (i = 0; i <= sa_dev->end_port - sa_dev->start_port; ++i) { - ib_unregister_mad_agent(sa_dev->port[i].agent); - kref_put(&sa_dev->port[i].sm_ah->ref, free_sm_ah); + if (rdma_port_get_link_layer(device, i + 1) == IB_LINK_LAYER_INFINIBAND) { + ib_unregister_mad_agent(sa_dev->port[i].agent); + if (sa_dev->port[i].sm_ah) + kref_put(&sa_dev->port[i].sm_ah->ref, free_sm_ah); + } + } kfree(sa_dev); @@ -968,20 +1248,30 @@ static int __init ib_sa_init(void) { int ret; - spin_lock_init(&idr_lock); - spin_lock_init(&tid_lock); - get_random_bytes(&tid, sizeof tid); ret = ib_register_client(&sa_client); - if (ret) + if (ret) { printk(KERN_ERR "Couldn't register ib_sa client\n"); + goto err1; + } + + ret = mcast_init(); + if (ret) { + printk(KERN_ERR "Couldn't initialize multicast handling\n"); + goto err2; + } + return 0; +err2: + ib_unregister_client(&sa_client); +err1: return ret; } static void __exit ib_sa_cleanup(void) { + mcast_cleanup(); ib_unregister_client(&sa_client); idr_destroy(&query_idr); } diff --git a/drivers/infiniband/core/smi.c b/drivers/infiniband/core/smi.c index 35852e794e2..5855e4405d9 100644 --- a/drivers/infiniband/core/smi.c +++ b/drivers/infiniband/core/smi.c @@ -3,7 +3,7 @@ * Copyright (c) 2004, 2005 Infinicon Corporation. All rights reserved. * Copyright (c) 2004, 2005 Intel Corporation. All rights reserved. * Copyright (c) 2004, 2005 Topspin Corporation. All rights reserved. - * Copyright (c) 2004, 2005 Voltaire Corporation. All rights reserved. + * Copyright (c) 2004-2007 Voltaire Corporation. All rights reserved. * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -34,7 +34,6 @@ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. * - * $Id: smi.c 1389 2004-12-27 22:56:47Z roland $ */ #include <rdma/ib_smi.h> @@ -44,9 +43,8 @@ * Fixup a directed route SMP for sending * Return 0 if the SMP should be discarded */ -int smi_handle_dr_smp_send(struct ib_smp *smp, - u8 node_type, - int port_num) +enum smi_action smi_handle_dr_smp_send(struct ib_smp *smp, + u8 node_type, int port_num) { u8 hop_ptr, hop_cnt; @@ -54,69 +52,75 @@ int smi_handle_dr_smp_send(struct ib_smp *smp, hop_cnt = smp->hop_cnt; /* See section 14.2.2.2, Vol 1 IB spec */ + /* C14-6 -- valid hop_cnt values are from 0 to 63 */ + if (hop_cnt >= IB_SMP_MAX_PATH_HOPS) + return IB_SMI_DISCARD; + if (!ib_get_smp_direction(smp)) { /* C14-9:1 */ if (hop_cnt && hop_ptr == 0) { smp->hop_ptr++; return (smp->initial_path[smp->hop_ptr] == - port_num); + port_num ? IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-9:2 */ if (hop_ptr && hop_ptr < hop_cnt) { - if (node_type != IB_NODE_SWITCH) - return 0; + if (node_type != RDMA_NODE_IB_SWITCH) + return IB_SMI_DISCARD; /* smp->return_path set when received */ smp->hop_ptr++; return (smp->initial_path[smp->hop_ptr] == - port_num); + port_num ? IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-9:3 -- We're at the end of the DR segment of path */ if (hop_ptr == hop_cnt) { /* smp->return_path set when received */ smp->hop_ptr++; - return (node_type == IB_NODE_SWITCH || - smp->dr_dlid == IB_LID_PERMISSIVE); + return (node_type == RDMA_NODE_IB_SWITCH || + smp->dr_dlid == IB_LID_PERMISSIVE ? + IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-9:4 -- hop_ptr = hop_cnt + 1 -> give to SMA/SM */ /* C14-9:5 -- Fail unreasonable hop pointer */ - return (hop_ptr == hop_cnt + 1); + return (hop_ptr == hop_cnt + 1 ? IB_SMI_HANDLE : IB_SMI_DISCARD); } else { /* C14-13:1 */ if (hop_cnt && hop_ptr == hop_cnt + 1) { smp->hop_ptr--; return (smp->return_path[smp->hop_ptr] == - port_num); + port_num ? IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-13:2 */ if (2 <= hop_ptr && hop_ptr <= hop_cnt) { - if (node_type != IB_NODE_SWITCH) - return 0; + if (node_type != RDMA_NODE_IB_SWITCH) + return IB_SMI_DISCARD; smp->hop_ptr--; return (smp->return_path[smp->hop_ptr] == - port_num); + port_num ? IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-13:3 -- at the end of the DR segment of path */ if (hop_ptr == 1) { smp->hop_ptr--; /* C14-13:3 -- SMPs destined for SM shouldn't be here */ - return (node_type == IB_NODE_SWITCH || - smp->dr_slid == IB_LID_PERMISSIVE); + return (node_type == RDMA_NODE_IB_SWITCH || + smp->dr_slid == IB_LID_PERMISSIVE ? + IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-13:4 -- hop_ptr = 0 -> should have gone to SM */ if (hop_ptr == 0) - return 1; + return IB_SMI_HANDLE; /* C14-13:5 -- Check for unreasonable hop pointer */ - return 0; + return IB_SMI_DISCARD; } } @@ -124,10 +128,8 @@ int smi_handle_dr_smp_send(struct ib_smp *smp, * Adjust information for a received SMP * Return 0 if the SMP should be dropped */ -int smi_handle_dr_smp_recv(struct ib_smp *smp, - u8 node_type, - int port_num, - int phys_port_cnt) +enum smi_action smi_handle_dr_smp_recv(struct ib_smp *smp, u8 node_type, + int port_num, int phys_port_cnt) { u8 hop_ptr, hop_cnt; @@ -135,19 +137,24 @@ int smi_handle_dr_smp_recv(struct ib_smp *smp, hop_cnt = smp->hop_cnt; /* See section 14.2.2.2, Vol 1 IB spec */ + /* C14-6 -- valid hop_cnt values are from 0 to 63 */ + if (hop_cnt >= IB_SMP_MAX_PATH_HOPS) + return IB_SMI_DISCARD; + if (!ib_get_smp_direction(smp)) { /* C14-9:1 -- sender should have incremented hop_ptr */ if (hop_cnt && hop_ptr == 0) - return 0; + return IB_SMI_DISCARD; /* C14-9:2 -- intermediate hop */ if (hop_ptr && hop_ptr < hop_cnt) { - if (node_type != IB_NODE_SWITCH) - return 0; + if (node_type != RDMA_NODE_IB_SWITCH) + return IB_SMI_DISCARD; smp->return_path[hop_ptr] = port_num; /* smp->hop_ptr updated when sending */ - return (smp->initial_path[hop_ptr+1] <= phys_port_cnt); + return (smp->initial_path[hop_ptr+1] <= phys_port_cnt ? + IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-9:3 -- We're at the end of the DR segment of path */ @@ -156,13 +163,14 @@ int smi_handle_dr_smp_recv(struct ib_smp *smp, smp->return_path[hop_ptr] = port_num; /* smp->hop_ptr updated when sending */ - return (node_type == IB_NODE_SWITCH || - smp->dr_dlid == IB_LID_PERMISSIVE); + return (node_type == RDMA_NODE_IB_SWITCH || + smp->dr_dlid == IB_LID_PERMISSIVE ? + IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-9:4 -- hop_ptr = hop_cnt + 1 -> give to SMA/SM */ /* C14-9:5 -- fail unreasonable hop pointer */ - return (hop_ptr == hop_cnt + 1); + return (hop_ptr == hop_cnt + 1 ? IB_SMI_HANDLE : IB_SMI_DISCARD); } else { @@ -170,16 +178,17 @@ int smi_handle_dr_smp_recv(struct ib_smp *smp, if (hop_cnt && hop_ptr == hop_cnt + 1) { smp->hop_ptr--; return (smp->return_path[smp->hop_ptr] == - port_num); + port_num ? IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-13:2 */ if (2 <= hop_ptr && hop_ptr <= hop_cnt) { - if (node_type != IB_NODE_SWITCH) - return 0; + if (node_type != RDMA_NODE_IB_SWITCH) + return IB_SMI_DISCARD; /* smp->hop_ptr updated when sending */ - return (smp->return_path[hop_ptr-1] <= phys_port_cnt); + return (smp->return_path[hop_ptr-1] <= phys_port_cnt ? + IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-13:3 -- We're at the end of the DR segment of path */ @@ -187,23 +196,20 @@ int smi_handle_dr_smp_recv(struct ib_smp *smp, if (smp->dr_slid == IB_LID_PERMISSIVE) { /* giving SMP to SM - update hop_ptr */ smp->hop_ptr--; - return 1; + return IB_SMI_HANDLE; } /* smp->hop_ptr updated when sending */ - return (node_type == IB_NODE_SWITCH); + return (node_type == RDMA_NODE_IB_SWITCH ? + IB_SMI_HANDLE : IB_SMI_DISCARD); } /* C14-13:4 -- hop_ptr = 0 -> give to SM */ /* C14-13:5 -- Check for unreasonable hop pointer */ - return (hop_ptr == 0); + return (hop_ptr == 0 ? IB_SMI_HANDLE : IB_SMI_DISCARD); } } -/* - * Return 1 if the received DR SMP should be forwarded to the send queue - * Return 0 if the SMP should be completed up the stack - */ -int smi_check_forward_dr_smp(struct ib_smp *smp) +enum smi_forward_action smi_check_forward_dr_smp(struct ib_smp *smp) { u8 hop_ptr, hop_cnt; @@ -213,23 +219,35 @@ int smi_check_forward_dr_smp(struct ib_smp *smp) if (!ib_get_smp_direction(smp)) { /* C14-9:2 -- intermediate hop */ if (hop_ptr && hop_ptr < hop_cnt) - return 1; + return IB_SMI_FORWARD; /* C14-9:3 -- at the end of the DR segment of path */ if (hop_ptr == hop_cnt) - return (smp->dr_dlid == IB_LID_PERMISSIVE); + return (smp->dr_dlid == IB_LID_PERMISSIVE ? + IB_SMI_SEND : IB_SMI_LOCAL); /* C14-9:4 -- hop_ptr = hop_cnt + 1 -> give to SMA/SM */ if (hop_ptr == hop_cnt + 1) - return 1; + return IB_SMI_SEND; } else { - /* C14-13:2 */ + /* C14-13:2 -- intermediate hop */ if (2 <= hop_ptr && hop_ptr <= hop_cnt) - return 1; + return IB_SMI_FORWARD; /* C14-13:3 -- at the end of the DR segment of path */ if (hop_ptr == 1) - return (smp->dr_slid != IB_LID_PERMISSIVE); + return (smp->dr_slid != IB_LID_PERMISSIVE ? + IB_SMI_SEND : IB_SMI_LOCAL); } - return 0; + return IB_SMI_LOCAL; +} + +/* + * Return the forwarding port number from initial_path for outgoing SMP and + * from return_path for returning SMP + */ +int smi_get_fwd_port(struct ib_smp *smp) +{ + return (!ib_get_smp_direction(smp) ? smp->initial_path[smp->hop_ptr+1] : + smp->return_path[smp->hop_ptr-1]); } diff --git a/drivers/infiniband/core/smi.h b/drivers/infiniband/core/smi.h index 2b3c40198f8..aff96bac49b 100644 --- a/drivers/infiniband/core/smi.h +++ b/drivers/infiniband/core/smi.h @@ -3,7 +3,7 @@ * Copyright (c) 2004 Infinicon Corporation. All rights reserved. * Copyright (c) 2004 Intel Corporation. All rights reserved. * Copyright (c) 2004 Topspin Corporation. All rights reserved. - * Copyright (c) 2004 Voltaire Corporation. All rights reserved. + * Copyright (c) 2004-2007 Voltaire Corporation. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -33,7 +33,6 @@ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. * - * $Id: smi.h 1389 2004-12-27 22:56:47Z roland $ */ #ifndef __SMI_H_ @@ -41,29 +40,51 @@ #include <rdma/ib_smi.h> -int smi_handle_dr_smp_recv(struct ib_smp *smp, - u8 node_type, - int port_num, - int phys_port_cnt); -extern int smi_check_forward_dr_smp(struct ib_smp *smp); -extern int smi_handle_dr_smp_send(struct ib_smp *smp, - u8 node_type, - int port_num); -extern int smi_check_local_dr_smp(struct ib_smp *smp, - struct ib_device *device, - int port_num); +enum smi_action { + IB_SMI_DISCARD, + IB_SMI_HANDLE +}; + +enum smi_forward_action { + IB_SMI_LOCAL, /* SMP should be completed up the stack */ + IB_SMI_SEND, /* received DR SMP should be forwarded to the send queue */ + IB_SMI_FORWARD /* SMP should be forwarded (for switches only) */ +}; + +enum smi_action smi_handle_dr_smp_recv(struct ib_smp *smp, u8 node_type, + int port_num, int phys_port_cnt); +int smi_get_fwd_port(struct ib_smp *smp); +extern enum smi_forward_action smi_check_forward_dr_smp(struct ib_smp *smp); +extern enum smi_action smi_handle_dr_smp_send(struct ib_smp *smp, + u8 node_type, int port_num); /* - * Return 1 if the SMP should be handled by the local SMA/SM via process_mad + * Return IB_SMI_HANDLE if the SMP should be handled by the local SMA/SM + * via process_mad */ -static inline int smi_check_local_smp(struct ib_mad_agent *mad_agent, - struct ib_smp *smp) +static inline enum smi_action smi_check_local_smp(struct ib_smp *smp, + struct ib_device *device) { /* C14-9:3 -- We're at the end of the DR segment of path */ /* C14-9:4 -- Hop Pointer = Hop Count + 1 -> give to SMA/SM */ - return ((mad_agent->device->process_mad && + return ((device->process_mad && !ib_get_smp_direction(smp) && - (smp->hop_ptr == smp->hop_cnt + 1))); + (smp->hop_ptr == smp->hop_cnt + 1)) ? + IB_SMI_HANDLE : IB_SMI_DISCARD); +} + +/* + * Return IB_SMI_HANDLE if the SMP should be handled by the local SMA/SM + * via process_mad + */ +static inline enum smi_action smi_check_local_returning_smp(struct ib_smp *smp, + struct ib_device *device) +{ + /* C14-13:3 -- We're at the end of the DR segment of path */ + /* C14-13:4 -- Hop Pointer == 0 -> give to SM */ + return ((device->process_mad && + ib_get_smp_direction(smp) && + !smp->hop_ptr) ? IB_SMI_HANDLE : IB_SMI_DISCARD); } #endif /* __SMI_H_ */ diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c index 5982d687a00..cbd0383f622 100644 --- a/drivers/infiniband/core/sysfs.c +++ b/drivers/infiniband/core/sysfs.c @@ -30,13 +30,12 @@ * 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. - * - * $Id: sysfs.c 1349 2004-12-16 21:09:43Z roland $ */ #include "core_priv.h" #include <linux/slab.h> +#include <linux/stat.h> #include <linux/string.h> #include <rdma/ib_mad.h> @@ -68,11 +67,6 @@ struct port_table_attribute { int index; }; -static inline int ibdev_is_alive(const struct ib_device *dev) -{ - return dev->reg_state == IB_DEV_REGISTERED; -} - static ssize_t port_attr_show(struct kobject *kobj, struct attribute *attr, char *buf) { @@ -82,13 +76,11 @@ static ssize_t port_attr_show(struct kobject *kobj, if (!port_attr->show) return -EIO; - if (!ibdev_is_alive(p->ibdev)) - return -ENODEV; return port_attr->show(p, port_attr, buf); } -static struct sysfs_ops port_sysfs_ops = { +static const struct sysfs_ops port_sysfs_ops = { .show = port_attr_show }; @@ -112,7 +104,7 @@ static ssize_t state_show(struct ib_port *p, struct port_attribute *unused, return ret; return sprintf(buf, "%d: %s\n", attr.state, - attr.state >= 0 && attr.state <= ARRAY_SIZE(state_name) ? + attr.state >= 0 && attr.state < ARRAY_SIZE(state_name) ? state_name[attr.state] : "UNKNOWN"); } @@ -187,7 +179,7 @@ static ssize_t rate_show(struct ib_port *p, struct port_attribute *unused, { struct ib_port_attr attr; char *speed = ""; - int rate; + int rate; /* in deci-Gb/sec */ ssize_t ret; ret = ib_query_port(p->ibdev, p->port_num, &attr); @@ -195,11 +187,33 @@ static ssize_t rate_show(struct ib_port *p, struct port_attribute *unused, return ret; switch (attr.active_speed) { - case 2: speed = " DDR"; break; - case 4: speed = " QDR"; break; + case IB_SPEED_DDR: + speed = " DDR"; + rate = 50; + break; + case IB_SPEED_QDR: + speed = " QDR"; + rate = 100; + break; + case IB_SPEED_FDR10: + speed = " FDR10"; + rate = 100; + break; + case IB_SPEED_FDR: + speed = " FDR"; + rate = 140; + break; + case IB_SPEED_EDR: + speed = " EDR"; + rate = 250; + break; + case IB_SPEED_SDR: + default: /* default to SDR for invalid rates */ + rate = 25; + break; } - rate = 25 * ib_width_enum_to_int(attr.active_width) * attr.active_speed; + rate *= ib_width_enum_to_int(attr.active_width); if (rate < 0) return -EINVAL; @@ -231,6 +245,19 @@ static ssize_t phys_state_show(struct ib_port *p, struct port_attribute *unused, } } +static ssize_t link_layer_show(struct ib_port *p, struct port_attribute *unused, + char *buf) +{ + switch (rdma_port_get_link_layer(p->ibdev, p->port_num)) { + case IB_LINK_LAYER_INFINIBAND: + return sprintf(buf, "%s\n", "InfiniBand"); + case IB_LINK_LAYER_ETHERNET: + return sprintf(buf, "%s\n", "Ethernet"); + default: + return sprintf(buf, "%s\n", "Unknown"); + } +} + static PORT_ATTR_RO(state); static PORT_ATTR_RO(lid); static PORT_ATTR_RO(lid_mask_count); @@ -239,6 +266,7 @@ static PORT_ATTR_RO(sm_sl); static PORT_ATTR_RO(cap_mask); static PORT_ATTR_RO(rate); static PORT_ATTR_RO(phys_state); +static PORT_ATTR_RO(link_layer); static struct attribute *port_default_attrs[] = { &port_attr_state.attr, @@ -249,6 +277,7 @@ static struct attribute *port_default_attrs[] = { &port_attr_cap_mask.attr, &port_attr_rate.attr, &port_attr_phys_state.attr, + &port_attr_link_layer.attr, NULL }; @@ -264,15 +293,7 @@ static ssize_t show_port_gid(struct ib_port *p, struct port_attribute *attr, if (ret) return ret; - return sprintf(buf, "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n", - be16_to_cpu(((__be16 *) gid.raw)[0]), - be16_to_cpu(((__be16 *) gid.raw)[1]), - be16_to_cpu(((__be16 *) gid.raw)[2]), - be16_to_cpu(((__be16 *) gid.raw)[3]), - be16_to_cpu(((__be16 *) gid.raw)[4]), - be16_to_cpu(((__be16 *) gid.raw)[5]), - be16_to_cpu(((__be16 *) gid.raw)[6]), - be16_to_cpu(((__be16 *) gid.raw)[7])); + return sprintf(buf, "%pI6\n", gid.raw); } static ssize_t show_port_pkey(struct ib_port *p, struct port_attribute *attr, @@ -311,7 +332,7 @@ static ssize_t show_pma_counter(struct ib_port *p, struct port_attribute *attr, return sprintf(buf, "N/A (no PMA)\n"); in_mad = kzalloc(sizeof *in_mad, GFP_KERNEL); - out_mad = kmalloc(sizeof *in_mad, GFP_KERNEL); + out_mad = kmalloc(sizeof *out_mad, GFP_KERNEL); if (!in_mad || !out_mad) { ret = -ENOMEM; goto out; @@ -336,7 +357,7 @@ static ssize_t show_pma_counter(struct ib_port *p, struct port_attribute *attr, switch (width) { case 4: ret = sprintf(buf, "%u\n", (out_mad->data[40 + offset / 8] >> - (offset % 4)) & 0xf); + (4 - (offset % 8))) & 0xf); break; case 8: ret = sprintf(buf, "%u\n", out_mad->data[40 + offset / 8]); @@ -408,15 +429,19 @@ static void ib_port_release(struct kobject *kobj) struct attribute *a; int i; - for (i = 0; (a = p->gid_group.attrs[i]); ++i) - kfree(a); + if (p->gid_group.attrs) { + for (i = 0; (a = p->gid_group.attrs[i]); ++i) + kfree(a); - kfree(p->gid_group.attrs); + kfree(p->gid_group.attrs); + } - for (i = 0; (a = p->pkey_group.attrs[i]); ++i) - kfree(a); + if (p->pkey_group.attrs) { + for (i = 0; (a = p->pkey_group.attrs[i]); ++i) + kfree(a); - kfree(p->pkey_group.attrs); + kfree(p->pkey_group.attrs); + } kfree(p); } @@ -427,28 +452,25 @@ static struct kobj_type port_type = { .default_attrs = port_default_attrs }; -static void ib_device_release(struct class_device *cdev) +static void ib_device_release(struct device *device) { - struct ib_device *dev = container_of(cdev, struct ib_device, class_dev); + struct ib_device *dev = container_of(device, struct ib_device, dev); kfree(dev); } -static int ib_device_uevent(struct class_device *cdev, char **envp, - int num_envp, char *buf, int size) +static int ib_device_uevent(struct device *device, + struct kobj_uevent_env *env) { - struct ib_device *dev = container_of(cdev, struct ib_device, class_dev); - int i = 0, len = 0; + struct ib_device *dev = container_of(device, struct ib_device, dev); - if (add_uevent_var(envp, num_envp, &i, buf, size, &len, - "NAME=%s", dev->name)) + if (add_uevent_var(env, "NAME=%s", dev->name)) return -ENOMEM; /* * It would be nice to pass the node GUID with the event... */ - envp[i] = NULL; return 0; } @@ -472,14 +494,16 @@ alloc_group_attrs(ssize_t (*show)(struct ib_port *, goto err; if (snprintf(element->name, sizeof(element->name), - "%d", i) >= sizeof(element->name)) + "%d", i) >= sizeof(element->name)) { + kfree(element); goto err; + } element->attr.attr.name = element->name; element->attr.attr.mode = S_IRUGO; - element->attr.attr.owner = THIS_MODULE; element->attr.show = show; element->index = i; + sysfs_attr_init(&element->attr.attr); tab_attr[i] = &element->attr.attr; } @@ -493,7 +517,9 @@ err: return NULL; } -static int add_port(struct ib_device *device, int port_num) +static int add_port(struct ib_device *device, int port_num, + int (*port_callback)(struct ib_device *, + u8, struct kobject *)) { struct ib_port *p; struct ib_port_attr attr; @@ -510,30 +536,25 @@ static int add_port(struct ib_device *device, int port_num) p->ibdev = device; p->port_num = port_num; - p->kobj.ktype = &port_type; - p->kobj.parent = kobject_get(&device->ports_parent); - if (!p->kobj.parent) { - ret = -EBUSY; - goto err; + ret = kobject_init_and_add(&p->kobj, &port_type, + device->ports_parent, + "%d", port_num); + if (ret) { + kfree(p); + return ret; } - ret = kobject_set_name(&p->kobj, "%d", port_num); - if (ret) - goto err_put; - - ret = kobject_register(&p->kobj); - if (ret) - goto err_put; - ret = sysfs_create_group(&p->kobj, &pma_group); if (ret) goto err_put; p->gid_group.name = "gids"; p->gid_group.attrs = alloc_group_attrs(show_port_gid, attr.gid_tbl_len); - if (!p->gid_group.attrs) + if (!p->gid_group.attrs) { + ret = -ENOMEM; goto err_remove_pma; + } ret = sysfs_create_group(&p->kobj, &p->gid_group); if (ret) @@ -542,22 +563,35 @@ static int add_port(struct ib_device *device, int port_num) p->pkey_group.name = "pkeys"; p->pkey_group.attrs = alloc_group_attrs(show_port_pkey, attr.pkey_tbl_len); - if (!p->pkey_group.attrs) + if (!p->pkey_group.attrs) { + ret = -ENOMEM; goto err_remove_gid; + } ret = sysfs_create_group(&p->kobj, &p->pkey_group); if (ret) goto err_free_pkey; + if (port_callback) { + ret = port_callback(device, port_num, &p->kobj); + if (ret) + goto err_remove_pkey; + } + list_add_tail(&p->kobj.entry, &device->port_list); + kobject_uevent(&p->kobj, KOBJ_ADD); return 0; +err_remove_pkey: + sysfs_remove_group(&p->kobj, &p->pkey_group); + err_free_pkey: for (i = 0; i < attr.pkey_tbl_len; ++i) kfree(p->pkey_group.attrs[i]); kfree(p->pkey_group.attrs); + p->pkey_group.attrs = NULL; err_remove_gid: sysfs_remove_group(&p->kobj, &p->gid_group); @@ -567,42 +601,39 @@ err_free_gid: kfree(p->gid_group.attrs[i]); kfree(p->gid_group.attrs); + p->gid_group.attrs = NULL; err_remove_pma: sysfs_remove_group(&p->kobj, &pma_group); err_put: - kobject_put(&device->ports_parent); - -err: - kfree(p); + kobject_put(&p->kobj); return ret; } -static ssize_t show_node_type(struct class_device *cdev, char *buf) +static ssize_t show_node_type(struct device *device, + struct device_attribute *attr, char *buf) { - struct ib_device *dev = container_of(cdev, struct ib_device, class_dev); - - if (!ibdev_is_alive(dev)) - return -ENODEV; + struct ib_device *dev = container_of(device, struct ib_device, dev); switch (dev->node_type) { - case IB_NODE_CA: return sprintf(buf, "%d: CA\n", dev->node_type); - case IB_NODE_SWITCH: return sprintf(buf, "%d: switch\n", dev->node_type); - case IB_NODE_ROUTER: return sprintf(buf, "%d: router\n", dev->node_type); - default: return sprintf(buf, "%d: <unknown>\n", dev->node_type); + case RDMA_NODE_IB_CA: return sprintf(buf, "%d: CA\n", dev->node_type); + case RDMA_NODE_RNIC: return sprintf(buf, "%d: RNIC\n", dev->node_type); + case RDMA_NODE_USNIC: return sprintf(buf, "%d: usNIC\n", dev->node_type); + case RDMA_NODE_USNIC_UDP: return sprintf(buf, "%d: usNIC UDP\n", dev->node_type); + case RDMA_NODE_IB_SWITCH: return sprintf(buf, "%d: switch\n", dev->node_type); + case RDMA_NODE_IB_ROUTER: return sprintf(buf, "%d: router\n", dev->node_type); + default: return sprintf(buf, "%d: <unknown>\n", dev->node_type); } } -static ssize_t show_sys_image_guid(struct class_device *cdev, char *buf) +static ssize_t show_sys_image_guid(struct device *device, + struct device_attribute *dev_attr, char *buf) { - struct ib_device *dev = container_of(cdev, struct ib_device, class_dev); + struct ib_device *dev = container_of(device, struct ib_device, dev); struct ib_device_attr attr; ssize_t ret; - if (!ibdev_is_alive(dev)) - return -ENODEV; - ret = ib_query_device(dev, &attr); if (ret) return ret; @@ -614,12 +645,10 @@ static ssize_t show_sys_image_guid(struct class_device *cdev, char *buf) be16_to_cpu(((__be16 *) &attr.sys_image_guid)[3])); } -static ssize_t show_node_guid(struct class_device *cdev, char *buf) +static ssize_t show_node_guid(struct device *device, + struct device_attribute *attr, char *buf) { - struct ib_device *dev = container_of(cdev, struct ib_device, class_dev); - - if (!ibdev_is_alive(dev)) - return -ENODEV; + struct ib_device *dev = container_of(device, struct ib_device, dev); return sprintf(buf, "%04x:%04x:%04x:%04x\n", be16_to_cpu(((__be16 *) &dev->node_guid)[0]), @@ -628,91 +657,238 @@ static ssize_t show_node_guid(struct class_device *cdev, char *buf) be16_to_cpu(((__be16 *) &dev->node_guid)[3])); } -static CLASS_DEVICE_ATTR(node_type, S_IRUGO, show_node_type, NULL); -static CLASS_DEVICE_ATTR(sys_image_guid, S_IRUGO, show_sys_image_guid, NULL); -static CLASS_DEVICE_ATTR(node_guid, S_IRUGO, show_node_guid, NULL); +static ssize_t show_node_desc(struct device *device, + struct device_attribute *attr, char *buf) +{ + struct ib_device *dev = container_of(device, struct ib_device, dev); + + return sprintf(buf, "%.64s\n", dev->node_desc); +} + +static ssize_t set_node_desc(struct device *device, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct ib_device *dev = container_of(device, struct ib_device, dev); + struct ib_device_modify desc = {}; + int ret; + + if (!dev->modify_device) + return -EIO; + + memcpy(desc.node_desc, buf, min_t(int, count, 64)); + ret = ib_modify_device(dev, IB_DEVICE_MODIFY_NODE_DESC, &desc); + if (ret) + return ret; + + return count; +} + +static DEVICE_ATTR(node_type, S_IRUGO, show_node_type, NULL); +static DEVICE_ATTR(sys_image_guid, S_IRUGO, show_sys_image_guid, NULL); +static DEVICE_ATTR(node_guid, S_IRUGO, show_node_guid, NULL); +static DEVICE_ATTR(node_desc, S_IRUGO | S_IWUSR, show_node_desc, set_node_desc); -static struct class_device_attribute *ib_class_attributes[] = { - &class_device_attr_node_type, - &class_device_attr_sys_image_guid, - &class_device_attr_node_guid +static struct device_attribute *ib_class_attributes[] = { + &dev_attr_node_type, + &dev_attr_sys_image_guid, + &dev_attr_node_guid, + &dev_attr_node_desc }; static struct class ib_class = { .name = "infiniband", - .release = ib_device_release, - .uevent = ib_device_uevent, + .dev_release = ib_device_release, + .dev_uevent = ib_device_uevent, }; -int ib_device_register_sysfs(struct ib_device *device) +/* Show a given an attribute in the statistics group */ +static ssize_t show_protocol_stat(const struct device *device, + struct device_attribute *attr, char *buf, + unsigned offset) { - struct class_device *class_dev = &device->class_dev; + struct ib_device *dev = container_of(device, struct ib_device, dev); + union rdma_protocol_stats stats; + ssize_t ret; + + ret = dev->get_protocol_stats(dev, &stats); + if (ret) + return ret; + + return sprintf(buf, "%llu\n", + (unsigned long long) ((u64 *) &stats)[offset]); +} + +/* generate a read-only iwarp statistics attribute */ +#define IW_STATS_ENTRY(name) \ +static ssize_t show_##name(struct device *device, \ + struct device_attribute *attr, char *buf) \ +{ \ + return show_protocol_stat(device, attr, buf, \ + offsetof(struct iw_protocol_stats, name) / \ + sizeof (u64)); \ +} \ +static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL) + +IW_STATS_ENTRY(ipInReceives); +IW_STATS_ENTRY(ipInHdrErrors); +IW_STATS_ENTRY(ipInTooBigErrors); +IW_STATS_ENTRY(ipInNoRoutes); +IW_STATS_ENTRY(ipInAddrErrors); +IW_STATS_ENTRY(ipInUnknownProtos); +IW_STATS_ENTRY(ipInTruncatedPkts); +IW_STATS_ENTRY(ipInDiscards); +IW_STATS_ENTRY(ipInDelivers); +IW_STATS_ENTRY(ipOutForwDatagrams); +IW_STATS_ENTRY(ipOutRequests); +IW_STATS_ENTRY(ipOutDiscards); +IW_STATS_ENTRY(ipOutNoRoutes); +IW_STATS_ENTRY(ipReasmTimeout); +IW_STATS_ENTRY(ipReasmReqds); +IW_STATS_ENTRY(ipReasmOKs); +IW_STATS_ENTRY(ipReasmFails); +IW_STATS_ENTRY(ipFragOKs); +IW_STATS_ENTRY(ipFragFails); +IW_STATS_ENTRY(ipFragCreates); +IW_STATS_ENTRY(ipInMcastPkts); +IW_STATS_ENTRY(ipOutMcastPkts); +IW_STATS_ENTRY(ipInBcastPkts); +IW_STATS_ENTRY(ipOutBcastPkts); +IW_STATS_ENTRY(tcpRtoAlgorithm); +IW_STATS_ENTRY(tcpRtoMin); +IW_STATS_ENTRY(tcpRtoMax); +IW_STATS_ENTRY(tcpMaxConn); +IW_STATS_ENTRY(tcpActiveOpens); +IW_STATS_ENTRY(tcpPassiveOpens); +IW_STATS_ENTRY(tcpAttemptFails); +IW_STATS_ENTRY(tcpEstabResets); +IW_STATS_ENTRY(tcpCurrEstab); +IW_STATS_ENTRY(tcpInSegs); +IW_STATS_ENTRY(tcpOutSegs); +IW_STATS_ENTRY(tcpRetransSegs); +IW_STATS_ENTRY(tcpInErrs); +IW_STATS_ENTRY(tcpOutRsts); + +static struct attribute *iw_proto_stats_attrs[] = { + &dev_attr_ipInReceives.attr, + &dev_attr_ipInHdrErrors.attr, + &dev_attr_ipInTooBigErrors.attr, + &dev_attr_ipInNoRoutes.attr, + &dev_attr_ipInAddrErrors.attr, + &dev_attr_ipInUnknownProtos.attr, + &dev_attr_ipInTruncatedPkts.attr, + &dev_attr_ipInDiscards.attr, + &dev_attr_ipInDelivers.attr, + &dev_attr_ipOutForwDatagrams.attr, + &dev_attr_ipOutRequests.attr, + &dev_attr_ipOutDiscards.attr, + &dev_attr_ipOutNoRoutes.attr, + &dev_attr_ipReasmTimeout.attr, + &dev_attr_ipReasmReqds.attr, + &dev_attr_ipReasmOKs.attr, + &dev_attr_ipReasmFails.attr, + &dev_attr_ipFragOKs.attr, + &dev_attr_ipFragFails.attr, + &dev_attr_ipFragCreates.attr, + &dev_attr_ipInMcastPkts.attr, + &dev_attr_ipOutMcastPkts.attr, + &dev_attr_ipInBcastPkts.attr, + &dev_attr_ipOutBcastPkts.attr, + &dev_attr_tcpRtoAlgorithm.attr, + &dev_attr_tcpRtoMin.attr, + &dev_attr_tcpRtoMax.attr, + &dev_attr_tcpMaxConn.attr, + &dev_attr_tcpActiveOpens.attr, + &dev_attr_tcpPassiveOpens.attr, + &dev_attr_tcpAttemptFails.attr, + &dev_attr_tcpEstabResets.attr, + &dev_attr_tcpCurrEstab.attr, + &dev_attr_tcpInSegs.attr, + &dev_attr_tcpOutSegs.attr, + &dev_attr_tcpRetransSegs.attr, + &dev_attr_tcpInErrs.attr, + &dev_attr_tcpOutRsts.attr, + NULL +}; + +static struct attribute_group iw_stats_group = { + .name = "proto_stats", + .attrs = iw_proto_stats_attrs, +}; + +static void free_port_list_attributes(struct ib_device *device) +{ + struct kobject *p, *t; + + list_for_each_entry_safe(p, t, &device->port_list, entry) { + struct ib_port *port = container_of(p, struct ib_port, kobj); + list_del(&p->entry); + sysfs_remove_group(p, &pma_group); + sysfs_remove_group(p, &port->pkey_group); + sysfs_remove_group(p, &port->gid_group); + kobject_put(p); + } + + kobject_put(device->ports_parent); +} + +int ib_device_register_sysfs(struct ib_device *device, + int (*port_callback)(struct ib_device *, + u8, struct kobject *)) +{ + struct device *class_dev = &device->dev; int ret; int i; class_dev->class = &ib_class; - class_dev->class_data = device; - strlcpy(class_dev->class_id, device->name, BUS_ID_SIZE); + class_dev->parent = device->dma_device; + dev_set_name(class_dev, "%s", device->name); + dev_set_drvdata(class_dev, device); INIT_LIST_HEAD(&device->port_list); - ret = class_device_register(class_dev); + ret = device_register(class_dev); if (ret) goto err; for (i = 0; i < ARRAY_SIZE(ib_class_attributes); ++i) { - ret = class_device_create_file(class_dev, ib_class_attributes[i]); + ret = device_create_file(class_dev, ib_class_attributes[i]); if (ret) goto err_unregister; } - device->ports_parent.parent = kobject_get(&class_dev->kobj); - if (!device->ports_parent.parent) { - ret = -EBUSY; - goto err_unregister; - } - ret = kobject_set_name(&device->ports_parent, "ports"); - if (ret) - goto err_put; - ret = kobject_register(&device->ports_parent); - if (ret) + device->ports_parent = kobject_create_and_add("ports", + &class_dev->kobj); + if (!device->ports_parent) { + ret = -ENOMEM; goto err_put; + } - if (device->node_type == IB_NODE_SWITCH) { - ret = add_port(device, 0); + if (device->node_type == RDMA_NODE_IB_SWITCH) { + ret = add_port(device, 0, port_callback); if (ret) goto err_put; } else { - int i; - for (i = 1; i <= device->phys_port_cnt; ++i) { - ret = add_port(device, i); + ret = add_port(device, i, port_callback); if (ret) goto err_put; } } + if (device->node_type == RDMA_NODE_RNIC && device->get_protocol_stats) { + ret = sysfs_create_group(&class_dev->kobj, &iw_stats_group); + if (ret) + goto err_put; + } + return 0; err_put: - { - struct kobject *p, *t; - struct ib_port *port; - - list_for_each_entry_safe(p, t, &device->port_list, entry) { - list_del(&p->entry); - port = container_of(p, struct ib_port, kobj); - sysfs_remove_group(p, &pma_group); - sysfs_remove_group(p, &port->pkey_group); - sysfs_remove_group(p, &port->gid_group); - kobject_unregister(p); - } - } - - kobject_put(&class_dev->kobj); + free_port_list_attributes(device); err_unregister: - class_device_unregister(class_dev); + device_unregister(class_dev); err: return ret; @@ -720,20 +896,19 @@ err: void ib_device_unregister_sysfs(struct ib_device *device) { - struct kobject *p, *t; - struct ib_port *port; + /* Hold kobject until ib_dealloc_device() */ + struct kobject *kobj_dev = kobject_get(&device->dev.kobj); + int i; - list_for_each_entry_safe(p, t, &device->port_list, entry) { - list_del(&p->entry); - port = container_of(p, struct ib_port, kobj); - sysfs_remove_group(p, &pma_group); - sysfs_remove_group(p, &port->pkey_group); - sysfs_remove_group(p, &port->gid_group); - kobject_unregister(p); - } + if (device->node_type == RDMA_NODE_RNIC && device->get_protocol_stats) + sysfs_remove_group(kobj_dev, &iw_stats_group); + + free_port_list_attributes(device); + + for (i = 0; i < ARRAY_SIZE(ib_class_attributes); ++i) + device_remove_file(&device->dev, ib_class_attributes[i]); - kobject_unregister(&device->ports_parent); - class_device_unregister(&device->class_dev); + device_unregister(&device->dev); } int ib_sysfs_setup(void) diff --git a/drivers/infiniband/core/ucm.c b/drivers/infiniband/core/ucm.c index e95c4293a49..f2f63933e8a 100644 --- a/drivers/infiniband/core/ucm.c +++ b/drivers/infiniband/core/ucm.c @@ -29,25 +29,28 @@ * 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. - * - * $Id: ucm.c 2594 2005-06-13 19:46:02Z libor $ */ + +#include <linux/completion.h> #include <linux/init.h> #include <linux/fs.h> #include <linux/module.h> #include <linux/device.h> #include <linux/err.h> #include <linux/poll.h> +#include <linux/sched.h> #include <linux/file.h> #include <linux/mount.h> #include <linux/cdev.h> #include <linux/idr.h> #include <linux/mutex.h> +#include <linux/slab.h> #include <asm/uaccess.h> #include <rdma/ib_cm.h> #include <rdma/ib_user_cm.h> +#include <rdma/ib_marshall.h> MODULE_AUTHOR("Libor Michalek"); MODULE_DESCRIPTION("InfiniBand userspace Connection Manager access"); @@ -55,13 +58,13 @@ MODULE_LICENSE("Dual BSD/GPL"); struct ib_ucm_device { int devnum; - struct cdev dev; - struct class_device class_dev; + struct cdev cdev; + struct device dev; struct ib_device *ib_dev; }; struct ib_ucm_file { - struct semaphore mutex; + struct mutex file_mutex; struct file *filp; struct ib_ucm_device *device; @@ -72,7 +75,7 @@ struct ib_ucm_file { struct ib_ucm_context { int id; - wait_queue_head_t wait; + struct completion comp; atomic_t ref; int events_reported; @@ -138,7 +141,7 @@ static struct ib_ucm_context *ib_ucm_ctx_get(struct ib_ucm_file *file, int id) static void ib_ucm_ctx_put(struct ib_ucm_context *ctx) { if (atomic_dec_and_test(&ctx->ref)) - wake_up(&ctx->wait); + complete(&ctx->comp); } static inline int ib_ucm_new_cm_id(int event) @@ -150,7 +153,7 @@ static void ib_ucm_cleanup_events(struct ib_ucm_context *ctx) { struct ib_ucm_event *uevent; - down(&ctx->file->mutex); + mutex_lock(&ctx->file->file_mutex); list_del(&ctx->file_list); while (!list_empty(&ctx->events)) { @@ -158,41 +161,35 @@ static void ib_ucm_cleanup_events(struct ib_ucm_context *ctx) struct ib_ucm_event, ctx_list); list_del(&uevent->file_list); list_del(&uevent->ctx_list); + mutex_unlock(&ctx->file->file_mutex); /* clear incoming connections. */ if (ib_ucm_new_cm_id(uevent->resp.event)) ib_destroy_cm_id(uevent->cm_id); kfree(uevent); + mutex_lock(&ctx->file->file_mutex); } - up(&ctx->file->mutex); + mutex_unlock(&ctx->file->file_mutex); } static struct ib_ucm_context *ib_ucm_ctx_alloc(struct ib_ucm_file *file) { struct ib_ucm_context *ctx; - int result; ctx = kzalloc(sizeof *ctx, GFP_KERNEL); if (!ctx) return NULL; atomic_set(&ctx->ref, 1); - init_waitqueue_head(&ctx->wait); + init_completion(&ctx->comp); ctx->file = file; INIT_LIST_HEAD(&ctx->events); - do { - result = idr_pre_get(&ctx_id_table, GFP_KERNEL); - if (!result) - goto error; - - mutex_lock(&ctx_id_mutex); - result = idr_get_new(&ctx_id_table, ctx, &ctx->id); - mutex_unlock(&ctx_id_mutex); - } while (result == -EAGAIN); - - if (result) + mutex_lock(&ctx_id_mutex); + ctx->id = idr_alloc(&ctx_id_table, ctx, 0, 0, GFP_KERNEL); + mutex_unlock(&ctx_id_mutex); + if (ctx->id < 0) goto error; list_add_tail(&ctx->file_list, &file->ctxs); @@ -203,36 +200,6 @@ error: return NULL; } -static void ib_ucm_event_path_get(struct ib_ucm_path_rec *upath, - struct ib_sa_path_rec *kpath) -{ - if (!kpath || !upath) - return; - - memcpy(upath->dgid, kpath->dgid.raw, sizeof *upath->dgid); - memcpy(upath->sgid, kpath->sgid.raw, sizeof *upath->sgid); - - upath->dlid = kpath->dlid; - upath->slid = kpath->slid; - upath->raw_traffic = kpath->raw_traffic; - upath->flow_label = kpath->flow_label; - upath->hop_limit = kpath->hop_limit; - upath->traffic_class = kpath->traffic_class; - upath->reversible = kpath->reversible; - upath->numb_path = kpath->numb_path; - upath->pkey = kpath->pkey; - upath->sl = kpath->sl; - upath->mtu_selector = kpath->mtu_selector; - upath->mtu = kpath->mtu; - upath->rate_selector = kpath->rate_selector; - upath->rate = kpath->rate; - upath->packet_life_time = kpath->packet_life_time; - upath->preference = kpath->preference; - - upath->packet_life_time_selector = - kpath->packet_life_time_selector; -} - static void ib_ucm_event_req_get(struct ib_ucm_req_event_resp *ureq, struct ib_cm_req_event_param *kreq) { @@ -251,8 +218,10 @@ static void ib_ucm_event_req_get(struct ib_ucm_req_event_resp *ureq, ureq->srq = kreq->srq; ureq->port = kreq->port; - ib_ucm_event_path_get(&ureq->primary_path, kreq->primary_path); - ib_ucm_event_path_get(&ureq->alternate_path, kreq->alternate_path); + ib_copy_path_rec_to_user(&ureq->primary_path, kreq->primary_path); + if (kreq->alternate_path) + ib_copy_path_rec_to_user(&ureq->alternate_path, + kreq->alternate_path); } static void ib_ucm_event_rep_get(struct ib_ucm_rep_event_resp *urep, @@ -322,8 +291,8 @@ static int ib_ucm_event_process(struct ib_cm_event *evt, info = evt->param.rej_rcvd.ari; break; case IB_CM_LAP_RECEIVED: - ib_ucm_event_path_get(&uvt->resp.u.lap_resp.path, - evt->param.lap_rcvd.alternate_path); + ib_copy_path_rec_to_user(&uvt->resp.u.lap_resp.path, + evt->param.lap_rcvd.alternate_path); uvt->data_len = IB_CM_LAP_PRIVATE_DATA_SIZE; uvt->resp.present = IB_UCM_PRES_ALTERNATE; break; @@ -334,9 +303,9 @@ static int ib_ucm_event_process(struct ib_cm_event *evt, info = evt->param.apr_rcvd.apr_info; break; case IB_CM_SIDR_REQ_RECEIVED: - uvt->resp.u.sidr_req_resp.pkey = + uvt->resp.u.sidr_req_resp.pkey = evt->param.sidr_req_rcvd.pkey; - uvt->resp.u.sidr_req_resp.port = + uvt->resp.u.sidr_req_resp.port = evt->param.sidr_req_rcvd.port; uvt->data_len = IB_CM_SIDR_REQ_PRIVATE_DATA_SIZE; break; @@ -353,20 +322,18 @@ static int ib_ucm_event_process(struct ib_cm_event *evt, } if (uvt->data_len) { - uvt->data = kmalloc(uvt->data_len, GFP_KERNEL); + uvt->data = kmemdup(evt->private_data, uvt->data_len, GFP_KERNEL); if (!uvt->data) goto err1; - memcpy(uvt->data, evt->private_data, uvt->data_len); uvt->resp.present |= IB_UCM_PRES_DATA; } if (uvt->info_len) { - uvt->info = kmalloc(uvt->info_len, GFP_KERNEL); + uvt->info = kmemdup(info, uvt->info_len, GFP_KERNEL); if (!uvt->info) goto err2; - memcpy(uvt->info, info, uvt->info_len); uvt->resp.present |= IB_UCM_PRES_INFO; } return 0; @@ -400,11 +367,11 @@ static int ib_ucm_event_handler(struct ib_cm_id *cm_id, if (result) goto err2; - down(&ctx->file->mutex); + mutex_lock(&ctx->file->file_mutex); list_add_tail(&uevent->file_list, &ctx->file->events); list_add_tail(&uevent->ctx_list, &ctx->events); wake_up_interruptible(&ctx->file->poll_wait); - up(&ctx->file->mutex); + mutex_unlock(&ctx->file->file_mutex); return 0; err2: @@ -422,7 +389,6 @@ static ssize_t ib_ucm_event(struct ib_ucm_file *file, struct ib_ucm_event_get cmd; struct ib_ucm_event *uevent; int result = 0; - DEFINE_WAIT(wait); if (out_len < sizeof(struct ib_ucm_event_resp)) return -ENOSPC; @@ -430,31 +396,20 @@ static ssize_t ib_ucm_event(struct ib_ucm_file *file, if (copy_from_user(&cmd, inbuf, sizeof(cmd))) return -EFAULT; - down(&file->mutex); + mutex_lock(&file->file_mutex); while (list_empty(&file->events)) { + mutex_unlock(&file->file_mutex); - if (file->filp->f_flags & O_NONBLOCK) { - result = -EAGAIN; - break; - } - - if (signal_pending(current)) { - result = -ERESTARTSYS; - break; - } + if (file->filp->f_flags & O_NONBLOCK) + return -EAGAIN; - prepare_to_wait(&file->poll_wait, &wait, TASK_INTERRUPTIBLE); + if (wait_event_interruptible(file->poll_wait, + !list_empty(&file->events))) + return -ERESTARTSYS; - up(&file->mutex); - schedule(); - down(&file->mutex); - - finish_wait(&file->poll_wait, &wait); + mutex_lock(&file->file_mutex); } - if (result) - goto done; - uevent = list_entry(file->events.next, struct ib_ucm_event, file_list); if (ib_ucm_new_cm_id(uevent->resp.event)) { @@ -507,7 +462,7 @@ static ssize_t ib_ucm_event(struct ib_ucm_file *file, kfree(uevent->info); kfree(uevent); done: - up(&file->mutex); + mutex_unlock(&file->file_mutex); return result; } @@ -526,9 +481,9 @@ static ssize_t ib_ucm_create_id(struct ib_ucm_file *file, if (copy_from_user(&cmd, inbuf, sizeof(cmd))) return -EFAULT; - down(&file->mutex); + mutex_lock(&file->file_mutex); ctx = ib_ucm_ctx_alloc(file); - up(&file->mutex); + mutex_unlock(&file->file_mutex); if (!ctx) return -ENOMEM; @@ -586,8 +541,8 @@ static ssize_t ib_ucm_destroy_id(struct ib_ucm_file *file, if (IS_ERR(ctx)) return PTR_ERR(ctx); - atomic_dec(&ctx->ref); - wait_event(ctx->wait, !atomic_read(&ctx->ref)); + ib_ucm_ctx_put(ctx); + wait_for_completion(&ctx->comp); /* No new events will be generated after destroying the cm_id. */ ib_destroy_cm_id(ctx->cm_id); @@ -635,65 +590,11 @@ static ssize_t ib_ucm_attr_id(struct ib_ucm_file *file, return result; } -static void ib_ucm_copy_ah_attr(struct ib_ucm_ah_attr *dest_attr, - struct ib_ah_attr *src_attr) -{ - memcpy(dest_attr->grh_dgid, src_attr->grh.dgid.raw, - sizeof src_attr->grh.dgid); - dest_attr->grh_flow_label = src_attr->grh.flow_label; - dest_attr->grh_sgid_index = src_attr->grh.sgid_index; - dest_attr->grh_hop_limit = src_attr->grh.hop_limit; - dest_attr->grh_traffic_class = src_attr->grh.traffic_class; - - dest_attr->dlid = src_attr->dlid; - dest_attr->sl = src_attr->sl; - dest_attr->src_path_bits = src_attr->src_path_bits; - dest_attr->static_rate = src_attr->static_rate; - dest_attr->is_global = (src_attr->ah_flags & IB_AH_GRH); - dest_attr->port_num = src_attr->port_num; -} - -static void ib_ucm_copy_qp_attr(struct ib_ucm_init_qp_attr_resp *dest_attr, - struct ib_qp_attr *src_attr) -{ - dest_attr->cur_qp_state = src_attr->cur_qp_state; - dest_attr->path_mtu = src_attr->path_mtu; - dest_attr->path_mig_state = src_attr->path_mig_state; - dest_attr->qkey = src_attr->qkey; - dest_attr->rq_psn = src_attr->rq_psn; - dest_attr->sq_psn = src_attr->sq_psn; - dest_attr->dest_qp_num = src_attr->dest_qp_num; - dest_attr->qp_access_flags = src_attr->qp_access_flags; - - dest_attr->max_send_wr = src_attr->cap.max_send_wr; - dest_attr->max_recv_wr = src_attr->cap.max_recv_wr; - dest_attr->max_send_sge = src_attr->cap.max_send_sge; - dest_attr->max_recv_sge = src_attr->cap.max_recv_sge; - dest_attr->max_inline_data = src_attr->cap.max_inline_data; - - ib_ucm_copy_ah_attr(&dest_attr->ah_attr, &src_attr->ah_attr); - ib_ucm_copy_ah_attr(&dest_attr->alt_ah_attr, &src_attr->alt_ah_attr); - - dest_attr->pkey_index = src_attr->pkey_index; - dest_attr->alt_pkey_index = src_attr->alt_pkey_index; - dest_attr->en_sqd_async_notify = src_attr->en_sqd_async_notify; - dest_attr->sq_draining = src_attr->sq_draining; - dest_attr->max_rd_atomic = src_attr->max_rd_atomic; - dest_attr->max_dest_rd_atomic = src_attr->max_dest_rd_atomic; - dest_attr->min_rnr_timer = src_attr->min_rnr_timer; - dest_attr->port_num = src_attr->port_num; - dest_attr->timeout = src_attr->timeout; - dest_attr->retry_cnt = src_attr->retry_cnt; - dest_attr->rnr_retry = src_attr->rnr_retry; - dest_attr->alt_port_num = src_attr->alt_port_num; - dest_attr->alt_timeout = src_attr->alt_timeout; -} - static ssize_t ib_ucm_init_qp_attr(struct ib_ucm_file *file, const char __user *inbuf, int in_len, int out_len) { - struct ib_ucm_init_qp_attr_resp resp; + struct ib_uverbs_qp_attr resp; struct ib_ucm_init_qp_attr cmd; struct ib_ucm_context *ctx; struct ib_qp_attr qp_attr; @@ -716,7 +617,7 @@ static ssize_t ib_ucm_init_qp_attr(struct ib_ucm_file *file, if (result) goto out; - ib_ucm_copy_qp_attr(&resp, &qp_attr); + ib_copy_qp_attr_to_user(&resp, &qp_attr); if (copy_to_user((void __user *)(unsigned long)cmd.response, &resp, sizeof(resp))) @@ -727,6 +628,17 @@ out: return result; } +static int ucm_validate_listen(__be64 service_id, __be64 service_mask) +{ + service_id &= service_mask; + + if (((service_id & IB_CMA_SERVICE_ID_MASK) == IB_CMA_SERVICE_ID) || + ((service_id & IB_SDP_SERVICE_ID_MASK) == IB_SDP_SERVICE_ID)) + return -EINVAL; + + return 0; +} + static ssize_t ib_ucm_listen(struct ib_ucm_file *file, const char __user *inbuf, int in_len, int out_len) @@ -742,16 +654,22 @@ static ssize_t ib_ucm_listen(struct ib_ucm_file *file, if (IS_ERR(ctx)) return PTR_ERR(ctx); - result = ib_cm_listen(ctx->cm_id, cmd.service_id, cmd.service_mask); + result = ucm_validate_listen(cmd.service_id, cmd.service_mask); + if (result) + goto out; + + result = ib_cm_listen(ctx->cm_id, cmd.service_id, cmd.service_mask, + NULL); +out: ib_ucm_ctx_put(ctx); return result; } -static ssize_t ib_ucm_establish(struct ib_ucm_file *file, - const char __user *inbuf, - int in_len, int out_len) +static ssize_t ib_ucm_notify(struct ib_ucm_file *file, + const char __user *inbuf, + int in_len, int out_len) { - struct ib_ucm_establish cmd; + struct ib_ucm_notify cmd; struct ib_ucm_context *ctx; int result; @@ -762,7 +680,7 @@ static ssize_t ib_ucm_establish(struct ib_ucm_file *file, if (IS_ERR(ctx)) return PTR_ERR(ctx); - result = ib_cm_establish(ctx->cm_id); + result = ib_cm_notify(ctx->cm_id, (enum ib_event_type) cmd.event); ib_ucm_ctx_put(ctx); return result; } @@ -776,14 +694,9 @@ static int ib_ucm_alloc_data(const void **dest, u64 src, u32 len) if (!len) return 0; - data = kmalloc(len, GFP_KERNEL); - if (!data) - return -ENOMEM; - - if (copy_from_user(data, (void __user *)(unsigned long)src, len)) { - kfree(data); - return -EFAULT; - } + data = memdup_user((void __user *)(unsigned long)src, len); + if (IS_ERR(data)) + return PTR_ERR(data); *dest = data; return 0; @@ -791,7 +704,7 @@ static int ib_ucm_alloc_data(const void **dest, u64 src, u32 len) static int ib_ucm_path_get(struct ib_sa_path_rec **path, u64 src) { - struct ib_ucm_path_rec ucm_path; + struct ib_user_path_rec upath; struct ib_sa_path_rec *sa_path; *path = NULL; @@ -803,36 +716,14 @@ static int ib_ucm_path_get(struct ib_sa_path_rec **path, u64 src) if (!sa_path) return -ENOMEM; - if (copy_from_user(&ucm_path, (void __user *)(unsigned long)src, - sizeof(ucm_path))) { + if (copy_from_user(&upath, (void __user *)(unsigned long)src, + sizeof(upath))) { kfree(sa_path); return -EFAULT; } - memcpy(sa_path->dgid.raw, ucm_path.dgid, sizeof sa_path->dgid); - memcpy(sa_path->sgid.raw, ucm_path.sgid, sizeof sa_path->sgid); - - sa_path->dlid = ucm_path.dlid; - sa_path->slid = ucm_path.slid; - sa_path->raw_traffic = ucm_path.raw_traffic; - sa_path->flow_label = ucm_path.flow_label; - sa_path->hop_limit = ucm_path.hop_limit; - sa_path->traffic_class = ucm_path.traffic_class; - sa_path->reversible = ucm_path.reversible; - sa_path->numb_path = ucm_path.numb_path; - sa_path->pkey = ucm_path.pkey; - sa_path->sl = ucm_path.sl; - sa_path->mtu_selector = ucm_path.mtu_selector; - sa_path->mtu = ucm_path.mtu; - sa_path->rate_selector = ucm_path.rate_selector; - sa_path->rate = ucm_path.rate; - sa_path->packet_life_time = ucm_path.packet_life_time; - sa_path->preference = ucm_path.preference; - - sa_path->packet_life_time_selector = - ucm_path.packet_life_time_selector; - + ib_copy_path_rec_from_user(sa_path, &upath); *path = sa_path; return 0; } @@ -918,7 +809,6 @@ static ssize_t ib_ucm_send_rep(struct ib_ucm_file *file, param.private_data_len = cmd.len; param.responder_resources = cmd.responder_resources; param.initiator_depth = cmd.initiator_depth; - param.target_ack_delay = cmd.target_ack_delay; param.failover_accepted = cmd.failover_accepted; param.flow_control = cmd.flow_control; param.rnr_retry_count = cmd.rnr_retry_count; @@ -1128,7 +1018,6 @@ static ssize_t ib_ucm_send_sidr_req(struct ib_ucm_file *file, param.service_id = cmd.sid; param.timeout_ms = cmd.timeout; param.max_cm_retries = cmd.max_cm_retries; - param.pkey = cmd.pkey; ctx = ib_ucm_ctx_get(file, cmd.id); if (!IS_ERR(ctx)) { @@ -1192,7 +1081,7 @@ static ssize_t (*ucm_cmd_table[])(struct ib_ucm_file *file, [IB_USER_CM_CMD_DESTROY_ID] = ib_ucm_destroy_id, [IB_USER_CM_CMD_ATTR_ID] = ib_ucm_attr_id, [IB_USER_CM_CMD_LISTEN] = ib_ucm_listen, - [IB_USER_CM_CMD_ESTABLISH] = ib_ucm_establish, + [IB_USER_CM_CMD_NOTIFY] = ib_ucm_notify, [IB_USER_CM_CMD_SEND_REQ] = ib_ucm_send_req, [IB_USER_CM_CMD_SEND_REP] = ib_ucm_send_rep, [IB_USER_CM_CMD_SEND_RTU] = ib_ucm_send_rtu, @@ -1221,7 +1110,7 @@ static ssize_t ib_ucm_write(struct file *filp, const char __user *buf, if (copy_from_user(&hdr, buf, sizeof(hdr))) return -EFAULT; - if (hdr.cmd < 0 || hdr.cmd >= ARRAY_SIZE(ucm_cmd_table)) + if (hdr.cmd >= ARRAY_SIZE(ucm_cmd_table)) return -EINVAL; if (hdr.in + sizeof(hdr) > len) @@ -1249,6 +1138,14 @@ static unsigned int ib_ucm_poll(struct file *filp, return mask; } +/* + * ib_ucm_open() does not need the BKL: + * + * - no global state is referred to; + * - there is no ioctl method to race against; + * - no further module initialization is required for open to work + * after the device is registered. + */ static int ib_ucm_open(struct inode *inode, struct file *filp) { struct ib_ucm_file *file; @@ -1261,13 +1158,13 @@ static int ib_ucm_open(struct inode *inode, struct file *filp) INIT_LIST_HEAD(&file->ctxs); init_waitqueue_head(&file->poll_wait); - init_MUTEX(&file->mutex); + mutex_init(&file->file_mutex); filp->private_data = file; file->filp = filp; - file->device = container_of(inode->i_cdev, struct ib_ucm_device, dev); + file->device = container_of(inode->i_cdev, struct ib_ucm_device, cdev); - return 0; + return nonseekable_open(inode, filp); } static int ib_ucm_close(struct inode *inode, struct file *filp) @@ -1275,11 +1172,11 @@ static int ib_ucm_close(struct inode *inode, struct file *filp) struct ib_ucm_file *file = filp->private_data; struct ib_ucm_context *ctx; - down(&file->mutex); + mutex_lock(&file->file_mutex); while (!list_empty(&file->ctxs)) { ctx = list_entry(file->ctxs.next, struct ib_ucm_context, file_list); - up(&file->mutex); + mutex_unlock(&file->file_mutex); mutex_lock(&ctx_id_mutex); idr_remove(&ctx_id_table, ctx->id); @@ -1289,59 +1186,75 @@ static int ib_ucm_close(struct inode *inode, struct file *filp) ib_ucm_cleanup_events(ctx); kfree(ctx); - down(&file->mutex); + mutex_lock(&file->file_mutex); } - up(&file->mutex); + mutex_unlock(&file->file_mutex); kfree(file); return 0; } -static void ib_ucm_release_class_dev(struct class_device *class_dev) +static void ib_ucm_release_dev(struct device *dev) { - struct ib_ucm_device *dev; + struct ib_ucm_device *ucm_dev; - dev = container_of(class_dev, struct ib_ucm_device, class_dev); - cdev_del(&dev->dev); - clear_bit(dev->devnum, dev_map); - kfree(dev); + ucm_dev = container_of(dev, struct ib_ucm_device, dev); + cdev_del(&ucm_dev->cdev); + if (ucm_dev->devnum < IB_UCM_MAX_DEVICES) + clear_bit(ucm_dev->devnum, dev_map); + else + clear_bit(ucm_dev->devnum - IB_UCM_MAX_DEVICES, dev_map); + kfree(ucm_dev); } -static struct file_operations ucm_fops = { - .owner = THIS_MODULE, - .open = ib_ucm_open, +static const struct file_operations ucm_fops = { + .owner = THIS_MODULE, + .open = ib_ucm_open, .release = ib_ucm_close, - .write = ib_ucm_write, + .write = ib_ucm_write, .poll = ib_ucm_poll, + .llseek = no_llseek, }; -static struct class ucm_class = { - .name = "infiniband_cm", - .release = ib_ucm_release_class_dev -}; - -static ssize_t show_dev(struct class_device *class_dev, char *buf) +static ssize_t show_ibdev(struct device *dev, struct device_attribute *attr, + char *buf) { - struct ib_ucm_device *dev; - - dev = container_of(class_dev, struct ib_ucm_device, class_dev); - return print_dev_t(buf, dev->dev.dev); + struct ib_ucm_device *ucm_dev; + + ucm_dev = container_of(dev, struct ib_ucm_device, dev); + return sprintf(buf, "%s\n", ucm_dev->ib_dev->name); } -static CLASS_DEVICE_ATTR(dev, S_IRUGO, show_dev, NULL); +static DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL); -static ssize_t show_ibdev(struct class_device *class_dev, char *buf) +static dev_t overflow_maj; +static DECLARE_BITMAP(overflow_map, IB_UCM_MAX_DEVICES); +static int find_overflow_devnum(void) { - struct ib_ucm_device *dev; - - dev = container_of(class_dev, struct ib_ucm_device, class_dev); - return sprintf(buf, "%s\n", dev->ib_dev->name); + int ret; + + if (!overflow_maj) { + ret = alloc_chrdev_region(&overflow_maj, 0, IB_UCM_MAX_DEVICES, + "infiniband_cm"); + if (ret) { + printk(KERN_ERR "ucm: couldn't register dynamic device number\n"); + return ret; + } + } + + ret = find_first_zero_bit(overflow_map, IB_UCM_MAX_DEVICES); + if (ret >= IB_UCM_MAX_DEVICES) + return -1; + + return ret; } -static CLASS_DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL); static void ib_ucm_add_one(struct ib_device *device) { + int devnum; + dev_t base; struct ib_ucm_device *ucm_dev; - if (!device->alloc_ucontext) + if (!device->alloc_ucontext || + rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) return; ucm_dev = kzalloc(sizeof *ucm_dev, GFP_KERNEL); @@ -1350,40 +1263,49 @@ static void ib_ucm_add_one(struct ib_device *device) ucm_dev->ib_dev = device; - ucm_dev->devnum = find_first_zero_bit(dev_map, IB_UCM_MAX_DEVICES); - if (ucm_dev->devnum >= IB_UCM_MAX_DEVICES) - goto err; - - set_bit(ucm_dev->devnum, dev_map); + devnum = find_first_zero_bit(dev_map, IB_UCM_MAX_DEVICES); + if (devnum >= IB_UCM_MAX_DEVICES) { + devnum = find_overflow_devnum(); + if (devnum < 0) + goto err; + + ucm_dev->devnum = devnum + IB_UCM_MAX_DEVICES; + base = devnum + overflow_maj; + set_bit(devnum, overflow_map); + } else { + ucm_dev->devnum = devnum; + base = devnum + IB_UCM_BASE_DEV; + set_bit(devnum, dev_map); + } - cdev_init(&ucm_dev->dev, &ucm_fops); - ucm_dev->dev.owner = THIS_MODULE; - kobject_set_name(&ucm_dev->dev.kobj, "ucm%d", ucm_dev->devnum); - if (cdev_add(&ucm_dev->dev, IB_UCM_BASE_DEV + ucm_dev->devnum, 1)) + cdev_init(&ucm_dev->cdev, &ucm_fops); + ucm_dev->cdev.owner = THIS_MODULE; + kobject_set_name(&ucm_dev->cdev.kobj, "ucm%d", ucm_dev->devnum); + if (cdev_add(&ucm_dev->cdev, base, 1)) goto err; - ucm_dev->class_dev.class = &ucm_class; - ucm_dev->class_dev.dev = device->dma_device; - snprintf(ucm_dev->class_dev.class_id, BUS_ID_SIZE, "ucm%d", - ucm_dev->devnum); - if (class_device_register(&ucm_dev->class_dev)) + ucm_dev->dev.class = &cm_class; + ucm_dev->dev.parent = device->dma_device; + ucm_dev->dev.devt = ucm_dev->cdev.dev; + ucm_dev->dev.release = ib_ucm_release_dev; + dev_set_name(&ucm_dev->dev, "ucm%d", ucm_dev->devnum); + if (device_register(&ucm_dev->dev)) goto err_cdev; - if (class_device_create_file(&ucm_dev->class_dev, - &class_device_attr_dev)) - goto err_class; - if (class_device_create_file(&ucm_dev->class_dev, - &class_device_attr_ibdev)) - goto err_class; + if (device_create_file(&ucm_dev->dev, &dev_attr_ibdev)) + goto err_dev; ib_set_client_data(device, &ucm_client, ucm_dev); return; -err_class: - class_device_unregister(&ucm_dev->class_dev); +err_dev: + device_unregister(&ucm_dev->dev); err_cdev: - cdev_del(&ucm_dev->dev); - clear_bit(ucm_dev->devnum, dev_map); + cdev_del(&ucm_dev->cdev); + if (ucm_dev->devnum < IB_UCM_MAX_DEVICES) + clear_bit(devnum, dev_map); + else + clear_bit(devnum, overflow_map); err: kfree(ucm_dev); return; @@ -1396,14 +1318,11 @@ static void ib_ucm_remove_one(struct ib_device *device) if (!ucm_dev) return; - class_device_unregister(&ucm_dev->class_dev); + device_unregister(&ucm_dev->dev); } -static ssize_t show_abi_version(struct class *class, char *buf) -{ - return sprintf(buf, "%d\n", IB_USER_CM_ABI_VERSION); -} -static CLASS_ATTR(abi_version, S_IRUGO, show_abi_version, NULL); +static CLASS_ATTR_STRING(abi_version, S_IRUGO, + __stringify(IB_USER_CM_ABI_VERSION)); static int __init ib_ucm_init(void) { @@ -1413,41 +1332,37 @@ static int __init ib_ucm_init(void) "infiniband_cm"); if (ret) { printk(KERN_ERR "ucm: couldn't register device number\n"); - goto err; + goto error1; } - ret = class_register(&ucm_class); - if (ret) { - printk(KERN_ERR "ucm: couldn't create class infiniband_cm\n"); - goto err_chrdev; - } - - ret = class_create_file(&ucm_class, &class_attr_abi_version); + ret = class_create_file(&cm_class, &class_attr_abi_version.attr); if (ret) { printk(KERN_ERR "ucm: couldn't create abi_version attribute\n"); - goto err_class; + goto error2; } ret = ib_register_client(&ucm_client); if (ret) { printk(KERN_ERR "ucm: couldn't register client\n"); - goto err_class; + goto error3; } return 0; -err_class: - class_unregister(&ucm_class); -err_chrdev: +error3: + class_remove_file(&cm_class, &class_attr_abi_version.attr); +error2: unregister_chrdev_region(IB_UCM_BASE_DEV, IB_UCM_MAX_DEVICES); -err: +error1: return ret; } static void __exit ib_ucm_cleanup(void) { ib_unregister_client(&ucm_client); - class_unregister(&ucm_class); + class_remove_file(&cm_class, &class_attr_abi_version.attr); unregister_chrdev_region(IB_UCM_BASE_DEV, IB_UCM_MAX_DEVICES); + if (overflow_maj) + unregister_chrdev_region(overflow_maj, IB_UCM_MAX_DEVICES); idr_destroy(&ctx_id_table); } diff --git a/drivers/infiniband/core/ucma.c b/drivers/infiniband/core/ucma.c new file mode 100644 index 00000000000..56a4b7ca7ee --- /dev/null +++ b/drivers/infiniband/core/ucma.c @@ -0,0 +1,1632 @@ +/* + * Copyright (c) 2005-2006 Intel Corporation. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/completion.h> +#include <linux/file.h> +#include <linux/mutex.h> +#include <linux/poll.h> +#include <linux/sched.h> +#include <linux/idr.h> +#include <linux/in.h> +#include <linux/in6.h> +#include <linux/miscdevice.h> +#include <linux/slab.h> +#include <linux/sysctl.h> +#include <linux/module.h> + +#include <rdma/rdma_user_cm.h> +#include <rdma/ib_marshall.h> +#include <rdma/rdma_cm.h> +#include <rdma/rdma_cm_ib.h> +#include <rdma/ib_addr.h> +#include <rdma/ib.h> + +MODULE_AUTHOR("Sean Hefty"); +MODULE_DESCRIPTION("RDMA Userspace Connection Manager Access"); +MODULE_LICENSE("Dual BSD/GPL"); + +static unsigned int max_backlog = 1024; + +static struct ctl_table_header *ucma_ctl_table_hdr; +static struct ctl_table ucma_ctl_table[] = { + { + .procname = "max_backlog", + .data = &max_backlog, + .maxlen = sizeof max_backlog, + .mode = 0644, + .proc_handler = proc_dointvec, + }, + { } +}; + +struct ucma_file { + struct mutex mut; + struct file *filp; + struct list_head ctx_list; + struct list_head event_list; + wait_queue_head_t poll_wait; +}; + +struct ucma_context { + int id; + struct completion comp; + atomic_t ref; + int events_reported; + int backlog; + + struct ucma_file *file; + struct rdma_cm_id *cm_id; + u64 uid; + + struct list_head list; + struct list_head mc_list; +}; + +struct ucma_multicast { + struct ucma_context *ctx; + int id; + int events_reported; + + u64 uid; + struct list_head list; + struct sockaddr_storage addr; +}; + +struct ucma_event { + struct ucma_context *ctx; + struct ucma_multicast *mc; + struct list_head list; + struct rdma_cm_id *cm_id; + struct rdma_ucm_event_resp resp; +}; + +static DEFINE_MUTEX(mut); +static DEFINE_IDR(ctx_idr); +static DEFINE_IDR(multicast_idr); + +static inline struct ucma_context *_ucma_find_context(int id, + struct ucma_file *file) +{ + struct ucma_context *ctx; + + ctx = idr_find(&ctx_idr, id); + if (!ctx) + ctx = ERR_PTR(-ENOENT); + else if (ctx->file != file) + ctx = ERR_PTR(-EINVAL); + return ctx; +} + +static struct ucma_context *ucma_get_ctx(struct ucma_file *file, int id) +{ + struct ucma_context *ctx; + + mutex_lock(&mut); + ctx = _ucma_find_context(id, file); + if (!IS_ERR(ctx)) + atomic_inc(&ctx->ref); + mutex_unlock(&mut); + return ctx; +} + +static void ucma_put_ctx(struct ucma_context *ctx) +{ + if (atomic_dec_and_test(&ctx->ref)) + complete(&ctx->comp); +} + +static struct ucma_context *ucma_alloc_ctx(struct ucma_file *file) +{ + struct ucma_context *ctx; + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) + return NULL; + + atomic_set(&ctx->ref, 1); + init_completion(&ctx->comp); + INIT_LIST_HEAD(&ctx->mc_list); + ctx->file = file; + + mutex_lock(&mut); + ctx->id = idr_alloc(&ctx_idr, ctx, 0, 0, GFP_KERNEL); + mutex_unlock(&mut); + if (ctx->id < 0) + goto error; + + list_add_tail(&ctx->list, &file->ctx_list); + return ctx; + +error: + kfree(ctx); + return NULL; +} + +static struct ucma_multicast* ucma_alloc_multicast(struct ucma_context *ctx) +{ + struct ucma_multicast *mc; + + mc = kzalloc(sizeof(*mc), GFP_KERNEL); + if (!mc) + return NULL; + + mutex_lock(&mut); + mc->id = idr_alloc(&multicast_idr, mc, 0, 0, GFP_KERNEL); + mutex_unlock(&mut); + if (mc->id < 0) + goto error; + + mc->ctx = ctx; + list_add_tail(&mc->list, &ctx->mc_list); + return mc; + +error: + kfree(mc); + return NULL; +} + +static void ucma_copy_conn_event(struct rdma_ucm_conn_param *dst, + struct rdma_conn_param *src) +{ + if (src->private_data_len) + memcpy(dst->private_data, src->private_data, + src->private_data_len); + dst->private_data_len = src->private_data_len; + dst->responder_resources =src->responder_resources; + dst->initiator_depth = src->initiator_depth; + dst->flow_control = src->flow_control; + dst->retry_count = src->retry_count; + dst->rnr_retry_count = src->rnr_retry_count; + dst->srq = src->srq; + dst->qp_num = src->qp_num; +} + +static void ucma_copy_ud_event(struct rdma_ucm_ud_param *dst, + struct rdma_ud_param *src) +{ + if (src->private_data_len) + memcpy(dst->private_data, src->private_data, + src->private_data_len); + dst->private_data_len = src->private_data_len; + ib_copy_ah_attr_to_user(&dst->ah_attr, &src->ah_attr); + dst->qp_num = src->qp_num; + dst->qkey = src->qkey; +} + +static void ucma_set_event_context(struct ucma_context *ctx, + struct rdma_cm_event *event, + struct ucma_event *uevent) +{ + uevent->ctx = ctx; + switch (event->event) { + case RDMA_CM_EVENT_MULTICAST_JOIN: + case RDMA_CM_EVENT_MULTICAST_ERROR: + uevent->mc = (struct ucma_multicast *) + event->param.ud.private_data; + uevent->resp.uid = uevent->mc->uid; + uevent->resp.id = uevent->mc->id; + break; + default: + uevent->resp.uid = ctx->uid; + uevent->resp.id = ctx->id; + break; + } +} + +static int ucma_event_handler(struct rdma_cm_id *cm_id, + struct rdma_cm_event *event) +{ + struct ucma_event *uevent; + struct ucma_context *ctx = cm_id->context; + int ret = 0; + + uevent = kzalloc(sizeof(*uevent), GFP_KERNEL); + if (!uevent) + return event->event == RDMA_CM_EVENT_CONNECT_REQUEST; + + mutex_lock(&ctx->file->mut); + uevent->cm_id = cm_id; + ucma_set_event_context(ctx, event, uevent); + uevent->resp.event = event->event; + uevent->resp.status = event->status; + if (cm_id->qp_type == IB_QPT_UD) + ucma_copy_ud_event(&uevent->resp.param.ud, &event->param.ud); + else + ucma_copy_conn_event(&uevent->resp.param.conn, + &event->param.conn); + + if (event->event == RDMA_CM_EVENT_CONNECT_REQUEST) { + if (!ctx->backlog) { + ret = -ENOMEM; + kfree(uevent); + goto out; + } + ctx->backlog--; + } else if (!ctx->uid || ctx->cm_id != cm_id) { + /* + * We ignore events for new connections until userspace has set + * their context. This can only happen if an error occurs on a + * new connection before the user accepts it. This is okay, + * since the accept will just fail later. + */ + kfree(uevent); + goto out; + } + + list_add_tail(&uevent->list, &ctx->file->event_list); + wake_up_interruptible(&ctx->file->poll_wait); +out: + mutex_unlock(&ctx->file->mut); + return ret; +} + +static ssize_t ucma_get_event(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct ucma_context *ctx; + struct rdma_ucm_get_event cmd; + struct ucma_event *uevent; + int ret = 0; + + if (out_len < sizeof uevent->resp) + return -ENOSPC; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + mutex_lock(&file->mut); + while (list_empty(&file->event_list)) { + mutex_unlock(&file->mut); + + if (file->filp->f_flags & O_NONBLOCK) + return -EAGAIN; + + if (wait_event_interruptible(file->poll_wait, + !list_empty(&file->event_list))) + return -ERESTARTSYS; + + mutex_lock(&file->mut); + } + + uevent = list_entry(file->event_list.next, struct ucma_event, list); + + if (uevent->resp.event == RDMA_CM_EVENT_CONNECT_REQUEST) { + ctx = ucma_alloc_ctx(file); + if (!ctx) { + ret = -ENOMEM; + goto done; + } + uevent->ctx->backlog++; + ctx->cm_id = uevent->cm_id; + ctx->cm_id->context = ctx; + uevent->resp.id = ctx->id; + } + + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &uevent->resp, sizeof uevent->resp)) { + ret = -EFAULT; + goto done; + } + + list_del(&uevent->list); + uevent->ctx->events_reported++; + if (uevent->mc) + uevent->mc->events_reported++; + kfree(uevent); +done: + mutex_unlock(&file->mut); + return ret; +} + +static int ucma_get_qp_type(struct rdma_ucm_create_id *cmd, enum ib_qp_type *qp_type) +{ + switch (cmd->ps) { + case RDMA_PS_TCP: + *qp_type = IB_QPT_RC; + return 0; + case RDMA_PS_UDP: + case RDMA_PS_IPOIB: + *qp_type = IB_QPT_UD; + return 0; + case RDMA_PS_IB: + *qp_type = cmd->qp_type; + return 0; + default: + return -EINVAL; + } +} + +static ssize_t ucma_create_id(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_create_id cmd; + struct rdma_ucm_create_id_resp resp; + struct ucma_context *ctx; + enum ib_qp_type qp_type; + int ret; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ret = ucma_get_qp_type(&cmd, &qp_type); + if (ret) + return ret; + + mutex_lock(&file->mut); + ctx = ucma_alloc_ctx(file); + mutex_unlock(&file->mut); + if (!ctx) + return -ENOMEM; + + ctx->uid = cmd.uid; + ctx->cm_id = rdma_create_id(ucma_event_handler, ctx, cmd.ps, qp_type); + if (IS_ERR(ctx->cm_id)) { + ret = PTR_ERR(ctx->cm_id); + goto err1; + } + + resp.id = ctx->id; + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &resp, sizeof(resp))) { + ret = -EFAULT; + goto err2; + } + return 0; + +err2: + rdma_destroy_id(ctx->cm_id); +err1: + mutex_lock(&mut); + idr_remove(&ctx_idr, ctx->id); + mutex_unlock(&mut); + kfree(ctx); + return ret; +} + +static void ucma_cleanup_multicast(struct ucma_context *ctx) +{ + struct ucma_multicast *mc, *tmp; + + mutex_lock(&mut); + list_for_each_entry_safe(mc, tmp, &ctx->mc_list, list) { + list_del(&mc->list); + idr_remove(&multicast_idr, mc->id); + kfree(mc); + } + mutex_unlock(&mut); +} + +static void ucma_cleanup_mc_events(struct ucma_multicast *mc) +{ + struct ucma_event *uevent, *tmp; + + list_for_each_entry_safe(uevent, tmp, &mc->ctx->file->event_list, list) { + if (uevent->mc != mc) + continue; + + list_del(&uevent->list); + kfree(uevent); + } +} + +/* + * We cannot hold file->mut when calling rdma_destroy_id() or we can + * deadlock. We also acquire file->mut in ucma_event_handler(), and + * rdma_destroy_id() will wait until all callbacks have completed. + */ +static int ucma_free_ctx(struct ucma_context *ctx) +{ + int events_reported; + struct ucma_event *uevent, *tmp; + LIST_HEAD(list); + + /* No new events will be generated after destroying the id. */ + rdma_destroy_id(ctx->cm_id); + + ucma_cleanup_multicast(ctx); + + /* Cleanup events not yet reported to the user. */ + mutex_lock(&ctx->file->mut); + list_for_each_entry_safe(uevent, tmp, &ctx->file->event_list, list) { + if (uevent->ctx == ctx) + list_move_tail(&uevent->list, &list); + } + list_del(&ctx->list); + mutex_unlock(&ctx->file->mut); + + list_for_each_entry_safe(uevent, tmp, &list, list) { + list_del(&uevent->list); + if (uevent->resp.event == RDMA_CM_EVENT_CONNECT_REQUEST) + rdma_destroy_id(uevent->cm_id); + kfree(uevent); + } + + events_reported = ctx->events_reported; + kfree(ctx); + return events_reported; +} + +static ssize_t ucma_destroy_id(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_destroy_id cmd; + struct rdma_ucm_destroy_id_resp resp; + struct ucma_context *ctx; + int ret = 0; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + mutex_lock(&mut); + ctx = _ucma_find_context(cmd.id, file); + if (!IS_ERR(ctx)) + idr_remove(&ctx_idr, ctx->id); + mutex_unlock(&mut); + + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ucma_put_ctx(ctx); + wait_for_completion(&ctx->comp); + resp.events_reported = ucma_free_ctx(ctx); + + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &resp, sizeof(resp))) + ret = -EFAULT; + + return ret; +} + +static ssize_t ucma_bind_ip(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_bind_ip cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_bind_addr(ctx->cm_id, (struct sockaddr *) &cmd.addr); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_bind(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_bind cmd; + struct sockaddr *addr; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + addr = (struct sockaddr *) &cmd.addr; + if (cmd.reserved || !cmd.addr_size || (cmd.addr_size != rdma_addr_size(addr))) + return -EINVAL; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_bind_addr(ctx->cm_id, addr); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_resolve_ip(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_resolve_ip cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_resolve_addr(ctx->cm_id, (struct sockaddr *) &cmd.src_addr, + (struct sockaddr *) &cmd.dst_addr, + cmd.timeout_ms); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_resolve_addr(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_resolve_addr cmd; + struct sockaddr *src, *dst; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + src = (struct sockaddr *) &cmd.src_addr; + dst = (struct sockaddr *) &cmd.dst_addr; + if (cmd.reserved || (cmd.src_size && (cmd.src_size != rdma_addr_size(src))) || + !cmd.dst_size || (cmd.dst_size != rdma_addr_size(dst))) + return -EINVAL; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_resolve_addr(ctx->cm_id, src, dst, cmd.timeout_ms); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_resolve_route(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_resolve_route cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_resolve_route(ctx->cm_id, cmd.timeout_ms); + ucma_put_ctx(ctx); + return ret; +} + +static void ucma_copy_ib_route(struct rdma_ucm_query_route_resp *resp, + struct rdma_route *route) +{ + struct rdma_dev_addr *dev_addr; + + resp->num_paths = route->num_paths; + switch (route->num_paths) { + case 0: + dev_addr = &route->addr.dev_addr; + rdma_addr_get_dgid(dev_addr, + (union ib_gid *) &resp->ib_route[0].dgid); + rdma_addr_get_sgid(dev_addr, + (union ib_gid *) &resp->ib_route[0].sgid); + resp->ib_route[0].pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr)); + break; + case 2: + ib_copy_path_rec_to_user(&resp->ib_route[1], + &route->path_rec[1]); + /* fall through */ + case 1: + ib_copy_path_rec_to_user(&resp->ib_route[0], + &route->path_rec[0]); + break; + default: + break; + } +} + +static void ucma_copy_iboe_route(struct rdma_ucm_query_route_resp *resp, + struct rdma_route *route) +{ + + resp->num_paths = route->num_paths; + switch (route->num_paths) { + case 0: + rdma_ip2gid((struct sockaddr *)&route->addr.dst_addr, + (union ib_gid *)&resp->ib_route[0].dgid); + rdma_ip2gid((struct sockaddr *)&route->addr.src_addr, + (union ib_gid *)&resp->ib_route[0].sgid); + resp->ib_route[0].pkey = cpu_to_be16(0xffff); + break; + case 2: + ib_copy_path_rec_to_user(&resp->ib_route[1], + &route->path_rec[1]); + /* fall through */ + case 1: + ib_copy_path_rec_to_user(&resp->ib_route[0], + &route->path_rec[0]); + break; + default: + break; + } +} + +static void ucma_copy_iw_route(struct rdma_ucm_query_route_resp *resp, + struct rdma_route *route) +{ + struct rdma_dev_addr *dev_addr; + + dev_addr = &route->addr.dev_addr; + rdma_addr_get_dgid(dev_addr, (union ib_gid *) &resp->ib_route[0].dgid); + rdma_addr_get_sgid(dev_addr, (union ib_gid *) &resp->ib_route[0].sgid); +} + +static ssize_t ucma_query_route(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_query cmd; + struct rdma_ucm_query_route_resp resp; + struct ucma_context *ctx; + struct sockaddr *addr; + int ret = 0; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + memset(&resp, 0, sizeof resp); + addr = (struct sockaddr *) &ctx->cm_id->route.addr.src_addr; + memcpy(&resp.src_addr, addr, addr->sa_family == AF_INET ? + sizeof(struct sockaddr_in) : + sizeof(struct sockaddr_in6)); + addr = (struct sockaddr *) &ctx->cm_id->route.addr.dst_addr; + memcpy(&resp.dst_addr, addr, addr->sa_family == AF_INET ? + sizeof(struct sockaddr_in) : + sizeof(struct sockaddr_in6)); + if (!ctx->cm_id->device) + goto out; + + resp.node_guid = (__force __u64) ctx->cm_id->device->node_guid; + resp.port_num = ctx->cm_id->port_num; + switch (rdma_node_get_transport(ctx->cm_id->device->node_type)) { + case RDMA_TRANSPORT_IB: + switch (rdma_port_get_link_layer(ctx->cm_id->device, + ctx->cm_id->port_num)) { + case IB_LINK_LAYER_INFINIBAND: + ucma_copy_ib_route(&resp, &ctx->cm_id->route); + break; + case IB_LINK_LAYER_ETHERNET: + ucma_copy_iboe_route(&resp, &ctx->cm_id->route); + break; + default: + break; + } + break; + case RDMA_TRANSPORT_IWARP: + ucma_copy_iw_route(&resp, &ctx->cm_id->route); + break; + default: + break; + } + +out: + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &resp, sizeof(resp))) + ret = -EFAULT; + + ucma_put_ctx(ctx); + return ret; +} + +static void ucma_query_device_addr(struct rdma_cm_id *cm_id, + struct rdma_ucm_query_addr_resp *resp) +{ + if (!cm_id->device) + return; + + resp->node_guid = (__force __u64) cm_id->device->node_guid; + resp->port_num = cm_id->port_num; + resp->pkey = (__force __u16) cpu_to_be16( + ib_addr_get_pkey(&cm_id->route.addr.dev_addr)); +} + +static ssize_t ucma_query_addr(struct ucma_context *ctx, + void __user *response, int out_len) +{ + struct rdma_ucm_query_addr_resp resp; + struct sockaddr *addr; + int ret = 0; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + memset(&resp, 0, sizeof resp); + + addr = (struct sockaddr *) &ctx->cm_id->route.addr.src_addr; + resp.src_size = rdma_addr_size(addr); + memcpy(&resp.src_addr, addr, resp.src_size); + + addr = (struct sockaddr *) &ctx->cm_id->route.addr.dst_addr; + resp.dst_size = rdma_addr_size(addr); + memcpy(&resp.dst_addr, addr, resp.dst_size); + + ucma_query_device_addr(ctx->cm_id, &resp); + + if (copy_to_user(response, &resp, sizeof(resp))) + ret = -EFAULT; + + return ret; +} + +static ssize_t ucma_query_path(struct ucma_context *ctx, + void __user *response, int out_len) +{ + struct rdma_ucm_query_path_resp *resp; + int i, ret = 0; + + if (out_len < sizeof(*resp)) + return -ENOSPC; + + resp = kzalloc(out_len, GFP_KERNEL); + if (!resp) + return -ENOMEM; + + resp->num_paths = ctx->cm_id->route.num_paths; + for (i = 0, out_len -= sizeof(*resp); + i < resp->num_paths && out_len > sizeof(struct ib_path_rec_data); + i++, out_len -= sizeof(struct ib_path_rec_data)) { + + resp->path_data[i].flags = IB_PATH_GMP | IB_PATH_PRIMARY | + IB_PATH_BIDIRECTIONAL; + ib_sa_pack_path(&ctx->cm_id->route.path_rec[i], + &resp->path_data[i].path_rec); + } + + if (copy_to_user(response, resp, + sizeof(*resp) + (i * sizeof(struct ib_path_rec_data)))) + ret = -EFAULT; + + kfree(resp); + return ret; +} + +static ssize_t ucma_query_gid(struct ucma_context *ctx, + void __user *response, int out_len) +{ + struct rdma_ucm_query_addr_resp resp; + struct sockaddr_ib *addr; + int ret = 0; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + memset(&resp, 0, sizeof resp); + + ucma_query_device_addr(ctx->cm_id, &resp); + + addr = (struct sockaddr_ib *) &resp.src_addr; + resp.src_size = sizeof(*addr); + if (ctx->cm_id->route.addr.src_addr.ss_family == AF_IB) { + memcpy(addr, &ctx->cm_id->route.addr.src_addr, resp.src_size); + } else { + addr->sib_family = AF_IB; + addr->sib_pkey = (__force __be16) resp.pkey; + rdma_addr_get_sgid(&ctx->cm_id->route.addr.dev_addr, + (union ib_gid *) &addr->sib_addr); + addr->sib_sid = rdma_get_service_id(ctx->cm_id, (struct sockaddr *) + &ctx->cm_id->route.addr.src_addr); + } + + addr = (struct sockaddr_ib *) &resp.dst_addr; + resp.dst_size = sizeof(*addr); + if (ctx->cm_id->route.addr.dst_addr.ss_family == AF_IB) { + memcpy(addr, &ctx->cm_id->route.addr.dst_addr, resp.dst_size); + } else { + addr->sib_family = AF_IB; + addr->sib_pkey = (__force __be16) resp.pkey; + rdma_addr_get_dgid(&ctx->cm_id->route.addr.dev_addr, + (union ib_gid *) &addr->sib_addr); + addr->sib_sid = rdma_get_service_id(ctx->cm_id, (struct sockaddr *) + &ctx->cm_id->route.addr.dst_addr); + } + + if (copy_to_user(response, &resp, sizeof(resp))) + ret = -EFAULT; + + return ret; +} + +static ssize_t ucma_query(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_query cmd; + struct ucma_context *ctx; + void __user *response; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + response = (void __user *)(unsigned long) cmd.response; + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + switch (cmd.option) { + case RDMA_USER_CM_QUERY_ADDR: + ret = ucma_query_addr(ctx, response, out_len); + break; + case RDMA_USER_CM_QUERY_PATH: + ret = ucma_query_path(ctx, response, out_len); + break; + case RDMA_USER_CM_QUERY_GID: + ret = ucma_query_gid(ctx, response, out_len); + break; + default: + ret = -ENOSYS; + break; + } + + ucma_put_ctx(ctx); + return ret; +} + +static void ucma_copy_conn_param(struct rdma_cm_id *id, + struct rdma_conn_param *dst, + struct rdma_ucm_conn_param *src) +{ + dst->private_data = src->private_data; + dst->private_data_len = src->private_data_len; + dst->responder_resources =src->responder_resources; + dst->initiator_depth = src->initiator_depth; + dst->flow_control = src->flow_control; + dst->retry_count = src->retry_count; + dst->rnr_retry_count = src->rnr_retry_count; + dst->srq = src->srq; + dst->qp_num = src->qp_num; + dst->qkey = (id->route.addr.src_addr.ss_family == AF_IB) ? src->qkey : 0; +} + +static ssize_t ucma_connect(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_connect cmd; + struct rdma_conn_param conn_param; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + if (!cmd.conn_param.valid) + return -EINVAL; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ucma_copy_conn_param(ctx->cm_id, &conn_param, &cmd.conn_param); + ret = rdma_connect(ctx->cm_id, &conn_param); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_listen(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_listen cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ctx->backlog = cmd.backlog > 0 && cmd.backlog < max_backlog ? + cmd.backlog : max_backlog; + ret = rdma_listen(ctx->cm_id, ctx->backlog); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_accept(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_accept cmd; + struct rdma_conn_param conn_param; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + if (cmd.conn_param.valid) { + ucma_copy_conn_param(ctx->cm_id, &conn_param, &cmd.conn_param); + mutex_lock(&file->mut); + ret = rdma_accept(ctx->cm_id, &conn_param); + if (!ret) + ctx->uid = cmd.uid; + mutex_unlock(&file->mut); + } else + ret = rdma_accept(ctx->cm_id, NULL); + + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_reject(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_reject cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_reject(ctx->cm_id, cmd.private_data, cmd.private_data_len); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_disconnect(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_disconnect cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_disconnect(ctx->cm_id); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_init_qp_attr(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_init_qp_attr cmd; + struct ib_uverbs_qp_attr resp; + struct ucma_context *ctx; + struct ib_qp_attr qp_attr; + int ret; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + resp.qp_attr_mask = 0; + memset(&qp_attr, 0, sizeof qp_attr); + qp_attr.qp_state = cmd.qp_state; + ret = rdma_init_qp_attr(ctx->cm_id, &qp_attr, &resp.qp_attr_mask); + if (ret) + goto out; + + ib_copy_qp_attr_to_user(&resp, &qp_attr); + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &resp, sizeof(resp))) + ret = -EFAULT; + +out: + ucma_put_ctx(ctx); + return ret; +} + +static int ucma_set_option_id(struct ucma_context *ctx, int optname, + void *optval, size_t optlen) +{ + int ret = 0; + + switch (optname) { + case RDMA_OPTION_ID_TOS: + if (optlen != sizeof(u8)) { + ret = -EINVAL; + break; + } + rdma_set_service_type(ctx->cm_id, *((u8 *) optval)); + break; + case RDMA_OPTION_ID_REUSEADDR: + if (optlen != sizeof(int)) { + ret = -EINVAL; + break; + } + ret = rdma_set_reuseaddr(ctx->cm_id, *((int *) optval) ? 1 : 0); + break; + case RDMA_OPTION_ID_AFONLY: + if (optlen != sizeof(int)) { + ret = -EINVAL; + break; + } + ret = rdma_set_afonly(ctx->cm_id, *((int *) optval) ? 1 : 0); + break; + default: + ret = -ENOSYS; + } + + return ret; +} + +static int ucma_set_ib_path(struct ucma_context *ctx, + struct ib_path_rec_data *path_data, size_t optlen) +{ + struct ib_sa_path_rec sa_path; + struct rdma_cm_event event; + int ret; + + if (optlen % sizeof(*path_data)) + return -EINVAL; + + for (; optlen; optlen -= sizeof(*path_data), path_data++) { + if (path_data->flags == (IB_PATH_GMP | IB_PATH_PRIMARY | + IB_PATH_BIDIRECTIONAL)) + break; + } + + if (!optlen) + return -EINVAL; + + ib_sa_unpack_path(path_data->path_rec, &sa_path); + ret = rdma_set_ib_paths(ctx->cm_id, &sa_path, 1); + if (ret) + return ret; + + memset(&event, 0, sizeof event); + event.event = RDMA_CM_EVENT_ROUTE_RESOLVED; + return ucma_event_handler(ctx->cm_id, &event); +} + +static int ucma_set_option_ib(struct ucma_context *ctx, int optname, + void *optval, size_t optlen) +{ + int ret; + + switch (optname) { + case RDMA_OPTION_IB_PATH: + ret = ucma_set_ib_path(ctx, optval, optlen); + break; + default: + ret = -ENOSYS; + } + + return ret; +} + +static int ucma_set_option_level(struct ucma_context *ctx, int level, + int optname, void *optval, size_t optlen) +{ + int ret; + + switch (level) { + case RDMA_OPTION_ID: + ret = ucma_set_option_id(ctx, optname, optval, optlen); + break; + case RDMA_OPTION_IB: + ret = ucma_set_option_ib(ctx, optname, optval, optlen); + break; + default: + ret = -ENOSYS; + } + + return ret; +} + +static ssize_t ucma_set_option(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_set_option cmd; + struct ucma_context *ctx; + void *optval; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + optval = memdup_user((void __user *) (unsigned long) cmd.optval, + cmd.optlen); + if (IS_ERR(optval)) { + ret = PTR_ERR(optval); + goto out; + } + + ret = ucma_set_option_level(ctx, cmd.level, cmd.optname, optval, + cmd.optlen); + kfree(optval); + +out: + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_notify(struct ucma_file *file, const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_notify cmd; + struct ucma_context *ctx; + int ret; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + ctx = ucma_get_ctx(file, cmd.id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + ret = rdma_notify(ctx->cm_id, (enum ib_event_type) cmd.event); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_process_join(struct ucma_file *file, + struct rdma_ucm_join_mcast *cmd, int out_len) +{ + struct rdma_ucm_create_id_resp resp; + struct ucma_context *ctx; + struct ucma_multicast *mc; + struct sockaddr *addr; + int ret; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + addr = (struct sockaddr *) &cmd->addr; + if (cmd->reserved || !cmd->addr_size || (cmd->addr_size != rdma_addr_size(addr))) + return -EINVAL; + + ctx = ucma_get_ctx(file, cmd->id); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + mutex_lock(&file->mut); + mc = ucma_alloc_multicast(ctx); + if (!mc) { + ret = -ENOMEM; + goto err1; + } + + mc->uid = cmd->uid; + memcpy(&mc->addr, addr, cmd->addr_size); + ret = rdma_join_multicast(ctx->cm_id, (struct sockaddr *) &mc->addr, mc); + if (ret) + goto err2; + + resp.id = mc->id; + if (copy_to_user((void __user *)(unsigned long) cmd->response, + &resp, sizeof(resp))) { + ret = -EFAULT; + goto err3; + } + + mutex_unlock(&file->mut); + ucma_put_ctx(ctx); + return 0; + +err3: + rdma_leave_multicast(ctx->cm_id, (struct sockaddr *) &mc->addr); + ucma_cleanup_mc_events(mc); +err2: + mutex_lock(&mut); + idr_remove(&multicast_idr, mc->id); + mutex_unlock(&mut); + list_del(&mc->list); + kfree(mc); +err1: + mutex_unlock(&file->mut); + ucma_put_ctx(ctx); + return ret; +} + +static ssize_t ucma_join_ip_multicast(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_join_ip_mcast cmd; + struct rdma_ucm_join_mcast join_cmd; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + join_cmd.response = cmd.response; + join_cmd.uid = cmd.uid; + join_cmd.id = cmd.id; + join_cmd.addr_size = rdma_addr_size((struct sockaddr *) &cmd.addr); + join_cmd.reserved = 0; + memcpy(&join_cmd.addr, &cmd.addr, join_cmd.addr_size); + + return ucma_process_join(file, &join_cmd, out_len); +} + +static ssize_t ucma_join_multicast(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_join_mcast cmd; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + return ucma_process_join(file, &cmd, out_len); +} + +static ssize_t ucma_leave_multicast(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_destroy_id cmd; + struct rdma_ucm_destroy_id_resp resp; + struct ucma_multicast *mc; + int ret = 0; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + mutex_lock(&mut); + mc = idr_find(&multicast_idr, cmd.id); + if (!mc) + mc = ERR_PTR(-ENOENT); + else if (mc->ctx->file != file) + mc = ERR_PTR(-EINVAL); + else { + idr_remove(&multicast_idr, mc->id); + atomic_inc(&mc->ctx->ref); + } + mutex_unlock(&mut); + + if (IS_ERR(mc)) { + ret = PTR_ERR(mc); + goto out; + } + + rdma_leave_multicast(mc->ctx->cm_id, (struct sockaddr *) &mc->addr); + mutex_lock(&mc->ctx->file->mut); + ucma_cleanup_mc_events(mc); + list_del(&mc->list); + mutex_unlock(&mc->ctx->file->mut); + + ucma_put_ctx(mc->ctx); + resp.events_reported = mc->events_reported; + kfree(mc); + + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &resp, sizeof(resp))) + ret = -EFAULT; +out: + return ret; +} + +static void ucma_lock_files(struct ucma_file *file1, struct ucma_file *file2) +{ + /* Acquire mutex's based on pointer comparison to prevent deadlock. */ + if (file1 < file2) { + mutex_lock(&file1->mut); + mutex_lock(&file2->mut); + } else { + mutex_lock(&file2->mut); + mutex_lock(&file1->mut); + } +} + +static void ucma_unlock_files(struct ucma_file *file1, struct ucma_file *file2) +{ + if (file1 < file2) { + mutex_unlock(&file2->mut); + mutex_unlock(&file1->mut); + } else { + mutex_unlock(&file1->mut); + mutex_unlock(&file2->mut); + } +} + +static void ucma_move_events(struct ucma_context *ctx, struct ucma_file *file) +{ + struct ucma_event *uevent, *tmp; + + list_for_each_entry_safe(uevent, tmp, &ctx->file->event_list, list) + if (uevent->ctx == ctx) + list_move_tail(&uevent->list, &file->event_list); +} + +static ssize_t ucma_migrate_id(struct ucma_file *new_file, + const char __user *inbuf, + int in_len, int out_len) +{ + struct rdma_ucm_migrate_id cmd; + struct rdma_ucm_migrate_resp resp; + struct ucma_context *ctx; + struct fd f; + struct ucma_file *cur_file; + int ret = 0; + + if (copy_from_user(&cmd, inbuf, sizeof(cmd))) + return -EFAULT; + + /* Get current fd to protect against it being closed */ + f = fdget(cmd.fd); + if (!f.file) + return -ENOENT; + + /* Validate current fd and prevent destruction of id. */ + ctx = ucma_get_ctx(f.file->private_data, cmd.id); + if (IS_ERR(ctx)) { + ret = PTR_ERR(ctx); + goto file_put; + } + + cur_file = ctx->file; + if (cur_file == new_file) { + resp.events_reported = ctx->events_reported; + goto response; + } + + /* + * Migrate events between fd's, maintaining order, and avoiding new + * events being added before existing events. + */ + ucma_lock_files(cur_file, new_file); + mutex_lock(&mut); + + list_move_tail(&ctx->list, &new_file->ctx_list); + ucma_move_events(ctx, new_file); + ctx->file = new_file; + resp.events_reported = ctx->events_reported; + + mutex_unlock(&mut); + ucma_unlock_files(cur_file, new_file); + +response: + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &resp, sizeof(resp))) + ret = -EFAULT; + + ucma_put_ctx(ctx); +file_put: + fdput(f); + return ret; +} + +static ssize_t (*ucma_cmd_table[])(struct ucma_file *file, + const char __user *inbuf, + int in_len, int out_len) = { + [RDMA_USER_CM_CMD_CREATE_ID] = ucma_create_id, + [RDMA_USER_CM_CMD_DESTROY_ID] = ucma_destroy_id, + [RDMA_USER_CM_CMD_BIND_IP] = ucma_bind_ip, + [RDMA_USER_CM_CMD_RESOLVE_IP] = ucma_resolve_ip, + [RDMA_USER_CM_CMD_RESOLVE_ROUTE] = ucma_resolve_route, + [RDMA_USER_CM_CMD_QUERY_ROUTE] = ucma_query_route, + [RDMA_USER_CM_CMD_CONNECT] = ucma_connect, + [RDMA_USER_CM_CMD_LISTEN] = ucma_listen, + [RDMA_USER_CM_CMD_ACCEPT] = ucma_accept, + [RDMA_USER_CM_CMD_REJECT] = ucma_reject, + [RDMA_USER_CM_CMD_DISCONNECT] = ucma_disconnect, + [RDMA_USER_CM_CMD_INIT_QP_ATTR] = ucma_init_qp_attr, + [RDMA_USER_CM_CMD_GET_EVENT] = ucma_get_event, + [RDMA_USER_CM_CMD_GET_OPTION] = NULL, + [RDMA_USER_CM_CMD_SET_OPTION] = ucma_set_option, + [RDMA_USER_CM_CMD_NOTIFY] = ucma_notify, + [RDMA_USER_CM_CMD_JOIN_IP_MCAST] = ucma_join_ip_multicast, + [RDMA_USER_CM_CMD_LEAVE_MCAST] = ucma_leave_multicast, + [RDMA_USER_CM_CMD_MIGRATE_ID] = ucma_migrate_id, + [RDMA_USER_CM_CMD_QUERY] = ucma_query, + [RDMA_USER_CM_CMD_BIND] = ucma_bind, + [RDMA_USER_CM_CMD_RESOLVE_ADDR] = ucma_resolve_addr, + [RDMA_USER_CM_CMD_JOIN_MCAST] = ucma_join_multicast +}; + +static ssize_t ucma_write(struct file *filp, const char __user *buf, + size_t len, loff_t *pos) +{ + struct ucma_file *file = filp->private_data; + struct rdma_ucm_cmd_hdr hdr; + ssize_t ret; + + if (len < sizeof(hdr)) + return -EINVAL; + + if (copy_from_user(&hdr, buf, sizeof(hdr))) + return -EFAULT; + + if (hdr.cmd >= ARRAY_SIZE(ucma_cmd_table)) + return -EINVAL; + + if (hdr.in + sizeof(hdr) > len) + return -EINVAL; + + if (!ucma_cmd_table[hdr.cmd]) + return -ENOSYS; + + ret = ucma_cmd_table[hdr.cmd](file, buf + sizeof(hdr), hdr.in, hdr.out); + if (!ret) + ret = len; + + return ret; +} + +static unsigned int ucma_poll(struct file *filp, struct poll_table_struct *wait) +{ + struct ucma_file *file = filp->private_data; + unsigned int mask = 0; + + poll_wait(filp, &file->poll_wait, wait); + + if (!list_empty(&file->event_list)) + mask = POLLIN | POLLRDNORM; + + return mask; +} + +/* + * ucma_open() does not need the BKL: + * + * - no global state is referred to; + * - there is no ioctl method to race against; + * - no further module initialization is required for open to work + * after the device is registered. + */ +static int ucma_open(struct inode *inode, struct file *filp) +{ + struct ucma_file *file; + + file = kmalloc(sizeof *file, GFP_KERNEL); + if (!file) + return -ENOMEM; + + INIT_LIST_HEAD(&file->event_list); + INIT_LIST_HEAD(&file->ctx_list); + init_waitqueue_head(&file->poll_wait); + mutex_init(&file->mut); + + filp->private_data = file; + file->filp = filp; + + return nonseekable_open(inode, filp); +} + +static int ucma_close(struct inode *inode, struct file *filp) +{ + struct ucma_file *file = filp->private_data; + struct ucma_context *ctx, *tmp; + + mutex_lock(&file->mut); + list_for_each_entry_safe(ctx, tmp, &file->ctx_list, list) { + mutex_unlock(&file->mut); + + mutex_lock(&mut); + idr_remove(&ctx_idr, ctx->id); + mutex_unlock(&mut); + + ucma_free_ctx(ctx); + mutex_lock(&file->mut); + } + mutex_unlock(&file->mut); + kfree(file); + return 0; +} + +static const struct file_operations ucma_fops = { + .owner = THIS_MODULE, + .open = ucma_open, + .release = ucma_close, + .write = ucma_write, + .poll = ucma_poll, + .llseek = no_llseek, +}; + +static struct miscdevice ucma_misc = { + .minor = MISC_DYNAMIC_MINOR, + .name = "rdma_cm", + .nodename = "infiniband/rdma_cm", + .mode = 0666, + .fops = &ucma_fops, +}; + +static ssize_t show_abi_version(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + return sprintf(buf, "%d\n", RDMA_USER_CM_ABI_VERSION); +} +static DEVICE_ATTR(abi_version, S_IRUGO, show_abi_version, NULL); + +static int __init ucma_init(void) +{ + int ret; + + ret = misc_register(&ucma_misc); + if (ret) + return ret; + + ret = device_create_file(ucma_misc.this_device, &dev_attr_abi_version); + if (ret) { + printk(KERN_ERR "rdma_ucm: couldn't create abi_version attr\n"); + goto err1; + } + + ucma_ctl_table_hdr = register_net_sysctl(&init_net, "net/rdma_ucm", ucma_ctl_table); + if (!ucma_ctl_table_hdr) { + printk(KERN_ERR "rdma_ucm: couldn't register sysctl paths\n"); + ret = -ENOMEM; + goto err2; + } + return 0; +err2: + device_remove_file(ucma_misc.this_device, &dev_attr_abi_version); +err1: + misc_deregister(&ucma_misc); + return ret; +} + +static void __exit ucma_cleanup(void) +{ + unregister_net_sysctl_table(ucma_ctl_table_hdr); + device_remove_file(ucma_misc.this_device, &dev_attr_abi_version); + misc_deregister(&ucma_misc); + idr_destroy(&ctx_idr); +} + +module_init(ucma_init); +module_exit(ucma_cleanup); diff --git a/drivers/infiniband/core/ud_header.c b/drivers/infiniband/core/ud_header.c index 997c07db6d8..72feee620eb 100644 --- a/drivers/infiniband/core/ud_header.c +++ b/drivers/infiniband/core/ud_header.c @@ -29,12 +29,12 @@ * 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. - * - * $Id: ud_header.c 1349 2004-12-16 21:09:43Z roland $ */ #include <linux/errno.h> #include <linux/string.h> +#include <linux/export.h> +#include <linux/if_ether.h> #include <rdma/ib_pack.h> @@ -82,6 +82,40 @@ static const struct ib_field lrh_table[] = { .size_bits = 16 } }; +static const struct ib_field eth_table[] = { + { STRUCT_FIELD(eth, dmac_h), + .offset_words = 0, + .offset_bits = 0, + .size_bits = 32 }, + { STRUCT_FIELD(eth, dmac_l), + .offset_words = 1, + .offset_bits = 0, + .size_bits = 16 }, + { STRUCT_FIELD(eth, smac_h), + .offset_words = 1, + .offset_bits = 16, + .size_bits = 16 }, + { STRUCT_FIELD(eth, smac_l), + .offset_words = 2, + .offset_bits = 0, + .size_bits = 32 }, + { STRUCT_FIELD(eth, type), + .offset_words = 3, + .offset_bits = 0, + .size_bits = 16 } +}; + +static const struct ib_field vlan_table[] = { + { STRUCT_FIELD(vlan, tag), + .offset_words = 0, + .offset_bits = 0, + .size_bits = 16 }, + { STRUCT_FIELD(vlan, type), + .offset_words = 0, + .offset_bits = 16, + .size_bits = 16 } +}; + static const struct ib_field grh_table[] = { { STRUCT_FIELD(grh, ip_version), .offset_words = 0, @@ -182,45 +216,43 @@ static const struct ib_field deth_table[] = { /** * ib_ud_header_init - Initialize UD header structure * @payload_bytes:Length of packet payload + * @lrh_present: specify if LRH is present + * @eth_present: specify if Eth header is present + * @vlan_present: packet is tagged vlan * @grh_present:GRH flag (if non-zero, GRH will be included) + * @immediate_present: specify if immediate data is present * @header:Structure to initialize - * - * ib_ud_header_init() initializes the lrh.link_version, lrh.link_next_header, - * lrh.packet_length, grh.ip_version, grh.payload_length, - * grh.next_header, bth.opcode, bth.pad_count and - * bth.transport_header_version fields of a &struct ib_ud_header given - * the payload length and whether a GRH will be included. */ void ib_ud_header_init(int payload_bytes, + int lrh_present, + int eth_present, + int vlan_present, int grh_present, + int immediate_present, struct ib_ud_header *header) { - int header_len; - u16 packet_length; - memset(header, 0, sizeof *header); - header_len = - IB_LRH_BYTES + - IB_BTH_BYTES + - IB_DETH_BYTES; - if (grh_present) { - header_len += IB_GRH_BYTES; + if (lrh_present) { + u16 packet_length; + + header->lrh.link_version = 0; + header->lrh.link_next_header = + grh_present ? IB_LNH_IBA_GLOBAL : IB_LNH_IBA_LOCAL; + packet_length = (IB_LRH_BYTES + + IB_BTH_BYTES + + IB_DETH_BYTES + + (grh_present ? IB_GRH_BYTES : 0) + + payload_bytes + + 4 + /* ICRC */ + 3) / 4; /* round up */ + header->lrh.packet_length = cpu_to_be16(packet_length); } - header->lrh.link_version = 0; - header->lrh.link_next_header = - grh_present ? IB_LNH_IBA_GLOBAL : IB_LNH_IBA_LOCAL; - packet_length = (IB_LRH_BYTES + - IB_BTH_BYTES + - IB_DETH_BYTES + - payload_bytes + - 4 + /* ICRC */ - 3) / 4; /* round up */ - - header->grh_present = grh_present; + if (vlan_present) + header->eth.type = cpu_to_be16(ETH_P_8021Q); + if (grh_present) { - packet_length += IB_GRH_BYTES / 4; header->grh.ip_version = 6; header->grh.payload_length = cpu_to_be16((IB_BTH_BYTES + @@ -231,14 +263,18 @@ void ib_ud_header_init(int payload_bytes, header->grh.next_header = 0x1b; } - header->lrh.packet_length = cpu_to_be16(packet_length); - - if (header->immediate_present) + if (immediate_present) header->bth.opcode = IB_OPCODE_UD_SEND_ONLY_WITH_IMMEDIATE; else header->bth.opcode = IB_OPCODE_UD_SEND_ONLY; header->bth.pad_count = (4 - payload_bytes) & 3; header->bth.transport_header_version = 0; + + header->lrh_present = lrh_present; + header->eth_present = eth_present; + header->vlan_present = vlan_present; + header->grh_present = grh_present; + header->immediate_present = immediate_present; } EXPORT_SYMBOL(ib_ud_header_init); @@ -255,10 +291,21 @@ int ib_ud_header_pack(struct ib_ud_header *header, { int len = 0; - ib_pack(lrh_table, ARRAY_SIZE(lrh_table), - &header->lrh, buf); - len += IB_LRH_BYTES; - + if (header->lrh_present) { + ib_pack(lrh_table, ARRAY_SIZE(lrh_table), + &header->lrh, buf + len); + len += IB_LRH_BYTES; + } + if (header->eth_present) { + ib_pack(eth_table, ARRAY_SIZE(eth_table), + &header->eth, buf + len); + len += IB_ETH_BYTES; + } + if (header->vlan_present) { + ib_pack(vlan_table, ARRAY_SIZE(vlan_table), + &header->vlan, buf + len); + len += IB_VLAN_BYTES; + } if (header->grh_present) { ib_pack(grh_table, ARRAY_SIZE(grh_table), &header->grh, buf + len); diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c new file mode 100644 index 00000000000..a3a2e9c1639 --- /dev/null +++ b/drivers/infiniband/core/umem.c @@ -0,0 +1,287 @@ +/* + * Copyright (c) 2005 Topspin Communications. All rights reserved. + * Copyright (c) 2005 Cisco Systems. All rights reserved. + * Copyright (c) 2005 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/mm.h> +#include <linux/dma-mapping.h> +#include <linux/sched.h> +#include <linux/export.h> +#include <linux/hugetlb.h> +#include <linux/dma-attrs.h> +#include <linux/slab.h> + +#include "uverbs.h" + + +static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int dirty) +{ + struct scatterlist *sg; + struct page *page; + int i; + + if (umem->nmap > 0) + ib_dma_unmap_sg(dev, umem->sg_head.sgl, + umem->nmap, + DMA_BIDIRECTIONAL); + + for_each_sg(umem->sg_head.sgl, sg, umem->npages, i) { + + page = sg_page(sg); + if (umem->writable && dirty) + set_page_dirty_lock(page); + put_page(page); + } + + sg_free_table(&umem->sg_head); + return; + +} + +/** + * ib_umem_get - Pin and DMA map userspace memory. + * @context: userspace context to pin memory for + * @addr: userspace virtual address to start at + * @size: length of region to pin + * @access: IB_ACCESS_xxx flags for memory being pinned + * @dmasync: flush in-flight DMA when the memory region is written + */ +struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr, + size_t size, int access, int dmasync) +{ + struct ib_umem *umem; + struct page **page_list; + struct vm_area_struct **vma_list; + unsigned long locked; + unsigned long lock_limit; + unsigned long cur_base; + unsigned long npages; + int ret; + int i; + DEFINE_DMA_ATTRS(attrs); + struct scatterlist *sg, *sg_list_start; + int need_release = 0; + + if (dmasync) + dma_set_attr(DMA_ATTR_WRITE_BARRIER, &attrs); + + if (!can_do_mlock()) + return ERR_PTR(-EPERM); + + umem = kzalloc(sizeof *umem, GFP_KERNEL); + if (!umem) + return ERR_PTR(-ENOMEM); + + umem->context = context; + umem->length = size; + umem->offset = addr & ~PAGE_MASK; + umem->page_size = PAGE_SIZE; + /* + * We ask for writable memory if any access flags other than + * "remote read" are set. "Local write" and "remote write" + * obviously require write access. "Remote atomic" can do + * things like fetch and add, which will modify memory, and + * "MW bind" can change permissions by binding a window. + */ + umem->writable = !!(access & ~IB_ACCESS_REMOTE_READ); + + /* We assume the memory is from hugetlb until proved otherwise */ + umem->hugetlb = 1; + + page_list = (struct page **) __get_free_page(GFP_KERNEL); + if (!page_list) { + kfree(umem); + return ERR_PTR(-ENOMEM); + } + + /* + * if we can't alloc the vma_list, it's not so bad; + * just assume the memory is not hugetlb memory + */ + vma_list = (struct vm_area_struct **) __get_free_page(GFP_KERNEL); + if (!vma_list) + umem->hugetlb = 0; + + npages = PAGE_ALIGN(size + umem->offset) >> PAGE_SHIFT; + + down_write(¤t->mm->mmap_sem); + + locked = npages + current->mm->pinned_vm; + lock_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT; + + if ((locked > lock_limit) && !capable(CAP_IPC_LOCK)) { + ret = -ENOMEM; + goto out; + } + + cur_base = addr & PAGE_MASK; + + if (npages == 0) { + ret = -EINVAL; + goto out; + } + + ret = sg_alloc_table(&umem->sg_head, npages, GFP_KERNEL); + if (ret) + goto out; + + need_release = 1; + sg_list_start = umem->sg_head.sgl; + + while (npages) { + ret = get_user_pages(current, current->mm, cur_base, + min_t(unsigned long, npages, + PAGE_SIZE / sizeof (struct page *)), + 1, !umem->writable, page_list, vma_list); + + if (ret < 0) + goto out; + + umem->npages += ret; + cur_base += ret * PAGE_SIZE; + npages -= ret; + + for_each_sg(sg_list_start, sg, ret, i) { + if (vma_list && !is_vm_hugetlb_page(vma_list[i])) + umem->hugetlb = 0; + + sg_set_page(sg, page_list[i], PAGE_SIZE, 0); + } + + /* preparing for next loop */ + sg_list_start = sg; + } + + umem->nmap = ib_dma_map_sg_attrs(context->device, + umem->sg_head.sgl, + umem->npages, + DMA_BIDIRECTIONAL, + &attrs); + + if (umem->nmap <= 0) { + ret = -ENOMEM; + goto out; + } + + ret = 0; + +out: + if (ret < 0) { + if (need_release) + __ib_umem_release(context->device, umem, 0); + kfree(umem); + } else + current->mm->pinned_vm = locked; + + up_write(¤t->mm->mmap_sem); + if (vma_list) + free_page((unsigned long) vma_list); + free_page((unsigned long) page_list); + + return ret < 0 ? ERR_PTR(ret) : umem; +} +EXPORT_SYMBOL(ib_umem_get); + +static void ib_umem_account(struct work_struct *work) +{ + struct ib_umem *umem = container_of(work, struct ib_umem, work); + + down_write(&umem->mm->mmap_sem); + umem->mm->pinned_vm -= umem->diff; + up_write(&umem->mm->mmap_sem); + mmput(umem->mm); + kfree(umem); +} + +/** + * ib_umem_release - release memory pinned with ib_umem_get + * @umem: umem struct to release + */ +void ib_umem_release(struct ib_umem *umem) +{ + struct ib_ucontext *context = umem->context; + struct mm_struct *mm; + unsigned long diff; + + __ib_umem_release(umem->context->device, umem, 1); + + mm = get_task_mm(current); + if (!mm) { + kfree(umem); + return; + } + + diff = PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT; + + /* + * We may be called with the mm's mmap_sem already held. This + * can happen when a userspace munmap() is the call that drops + * the last reference to our file and calls our release + * method. If there are memory regions to destroy, we'll end + * up here and not be able to take the mmap_sem. In that case + * we defer the vm_locked accounting to the system workqueue. + */ + if (context->closing) { + if (!down_write_trylock(&mm->mmap_sem)) { + INIT_WORK(&umem->work, ib_umem_account); + umem->mm = mm; + umem->diff = diff; + + queue_work(ib_wq, &umem->work); + return; + } + } else + down_write(&mm->mmap_sem); + + current->mm->pinned_vm -= diff; + up_write(&mm->mmap_sem); + mmput(mm); + kfree(umem); +} +EXPORT_SYMBOL(ib_umem_release); + +int ib_umem_page_count(struct ib_umem *umem) +{ + int shift; + int i; + int n; + struct scatterlist *sg; + + shift = ilog2(umem->page_size); + + n = 0; + for_each_sg(umem->sg_head.sgl, sg, umem->nmap, i) + n += sg_dma_len(sg) >> shift; + + return n; +} +EXPORT_SYMBOL(ib_umem_page_count); diff --git a/drivers/infiniband/core/user_mad.c b/drivers/infiniband/core/user_mad.c index c908de8db5a..1acb9910055 100644 --- a/drivers/infiniband/core/user_mad.c +++ b/drivers/infiniband/core/user_mad.c @@ -1,7 +1,8 @@ /* * Copyright (c) 2004 Topspin Communications. All rights reserved. - * Copyright (c) 2005 Voltaire, Inc. All rights reserved. + * Copyright (c) 2005 Voltaire, Inc. All rights reserved. * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright (c) 2008 Cisco. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -30,8 +31,6 @@ * 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. - * - * $Id: user_mad.c 4010 2005-11-09 23:11:56Z roland $ */ #include <linux/module.h> @@ -40,14 +39,16 @@ #include <linux/err.h> #include <linux/fs.h> #include <linux/cdev.h> -#include <linux/pci.h> #include <linux/dma-mapping.h> #include <linux/poll.h> -#include <linux/rwsem.h> +#include <linux/mutex.h> #include <linux/kref.h> +#include <linux/compat.h> +#include <linux/sched.h> +#include <linux/semaphore.h> +#include <linux/slab.h> #include <asm/uaccess.h> -#include <asm/semaphore.h> #include <rdma/ib_mad.h> #include <rdma/ib_user_mad.h> @@ -65,12 +66,9 @@ enum { }; /* - * Our lifetime rules for these structs are the following: each time a - * device special file is opened, we look up the corresponding struct - * ib_umad_port by minor in the umad_port[] table while holding the - * port_lock. If this lookup succeeds, we take a reference on the - * ib_umad_port's struct ib_umad_device while still holding the - * port_lock; if the lookup fails, we fail the open(). We drop these + * Our lifetime rules for these structs are the following: + * device special file is opened, we take a reference on the + * ib_umad_port's struct ib_umad_device. We drop these * references in the corresponding close(). * * In addition to references coming from open character devices, there @@ -78,23 +76,18 @@ enum { * module's reference taken when allocating the ib_umad_device in * ib_umad_add_one(). * - * When destroying an ib_umad_device, we clear all of its - * ib_umad_ports from umad_port[] while holding port_lock before - * dropping the module's reference to the ib_umad_device. This is - * always safe because any open() calls will either succeed and obtain - * a reference before we clear the umad_port[] entries, or fail after - * we clear the umad_port[] entries. + * When destroying an ib_umad_device, we drop the module's reference. */ struct ib_umad_port { - struct cdev *dev; - struct class_device *class_dev; + struct cdev cdev; + struct device *dev; - struct cdev *sm_dev; - struct class_device *sm_class_dev; + struct cdev sm_cdev; + struct device *sm_dev; struct semaphore sm_sem; - struct rw_semaphore mutex; + struct mutex file_mutex; struct list_head file_list; struct ib_device *ib_dev; @@ -105,22 +98,27 @@ struct ib_umad_port { struct ib_umad_device { int start_port, end_port; - struct kref ref; + struct kobject kobj; struct ib_umad_port port[0]; }; struct ib_umad_file { + struct mutex mutex; struct ib_umad_port *port; struct list_head recv_list; + struct list_head send_list; struct list_head port_list; - spinlock_t recv_lock; + spinlock_t send_lock; wait_queue_head_t recv_wait; struct ib_mad_agent *agent[IB_UMAD_MAX_AGENTS]; int agents_dead; + u8 use_pkey_index; + u8 already_used; }; struct ib_umad_packet { struct ib_mad_send_buf *msg; + struct ib_mad_recv_wc *recv_wc; struct list_head list; int length; struct ib_user_mad mad; @@ -131,21 +129,30 @@ static struct class *umad_class; static const dev_t base_dev = MKDEV(IB_UMAD_MAJOR, IB_UMAD_MINOR_BASE); static DEFINE_SPINLOCK(port_lock); -static struct ib_umad_port *umad_port[IB_UMAD_MAX_PORTS]; -static DECLARE_BITMAP(dev_map, IB_UMAD_MAX_PORTS * 2); +static DECLARE_BITMAP(dev_map, IB_UMAD_MAX_PORTS); static void ib_umad_add_one(struct ib_device *device); static void ib_umad_remove_one(struct ib_device *device); -static void ib_umad_release_dev(struct kref *ref) +static void ib_umad_release_dev(struct kobject *kobj) { struct ib_umad_device *dev = - container_of(ref, struct ib_umad_device, ref); + container_of(kobj, struct ib_umad_device, kobj); kfree(dev); } -/* caller must hold port->mutex at least for reading */ +static struct kobj_type ib_umad_dev_ktype = { + .release = ib_umad_release_dev, +}; + +static int hdr_size(struct ib_umad_file *file) +{ + return file->use_pkey_index ? sizeof (struct ib_user_mad_hdr) : + sizeof (struct ib_user_mad_hdr_old); +} + +/* caller must hold file->mutex */ static struct ib_mad_agent *__get_agent(struct ib_umad_file *file, int id) { return file->agents_dead ? NULL : file->agent[id]; @@ -157,50 +164,47 @@ static int queue_packet(struct ib_umad_file *file, { int ret = 1; - down_read(&file->port->mutex); + mutex_lock(&file->mutex); for (packet->mad.hdr.id = 0; packet->mad.hdr.id < IB_UMAD_MAX_AGENTS; packet->mad.hdr.id++) if (agent == __get_agent(file, packet->mad.hdr.id)) { - spin_lock_irq(&file->recv_lock); list_add_tail(&packet->list, &file->recv_list); - spin_unlock_irq(&file->recv_lock); wake_up_interruptible(&file->recv_wait); ret = 0; break; } - up_read(&file->port->mutex); + mutex_unlock(&file->mutex); return ret; } +static void dequeue_send(struct ib_umad_file *file, + struct ib_umad_packet *packet) +{ + spin_lock_irq(&file->send_lock); + list_del(&packet->list); + spin_unlock_irq(&file->send_lock); +} + static void send_handler(struct ib_mad_agent *agent, struct ib_mad_send_wc *send_wc) { struct ib_umad_file *file = agent->context; - struct ib_umad_packet *timeout; struct ib_umad_packet *packet = send_wc->send_buf->context[0]; + dequeue_send(file, packet); ib_destroy_ah(packet->msg->ah); ib_free_send_mad(packet->msg); if (send_wc->status == IB_WC_RESP_TIMEOUT_ERR) { - timeout = kzalloc(sizeof *timeout + IB_MGMT_MAD_HDR, GFP_KERNEL); - if (!timeout) - goto out; - - timeout->length = IB_MGMT_MAD_HDR; - timeout->mad.hdr.id = packet->mad.hdr.id; - timeout->mad.hdr.status = ETIMEDOUT; - memcpy(timeout->mad.data, packet->mad.data, - sizeof (struct ib_mad_hdr)); - - if (queue_packet(file, agent, timeout)) - kfree(timeout); + packet->length = IB_MGMT_MAD_HDR; + packet->mad.hdr.status = ETIMEDOUT; + if (!queue_packet(file, agent, packet)) + return; } -out: kfree(packet); } @@ -209,43 +213,118 @@ static void recv_handler(struct ib_mad_agent *agent, { struct ib_umad_file *file = agent->context; struct ib_umad_packet *packet; - int length; if (mad_recv_wc->wc->status != IB_WC_SUCCESS) - goto out; + goto err1; - length = mad_recv_wc->mad_len; - packet = kzalloc(sizeof *packet + length, GFP_KERNEL); + packet = kzalloc(sizeof *packet, GFP_KERNEL); if (!packet) - goto out; - - packet->length = length; - - ib_coalesce_recv_mad(mad_recv_wc, packet->mad.data); - - packet->mad.hdr.status = 0; - packet->mad.hdr.length = length + sizeof (struct ib_user_mad); - packet->mad.hdr.qpn = cpu_to_be32(mad_recv_wc->wc->src_qp); - packet->mad.hdr.lid = cpu_to_be16(mad_recv_wc->wc->slid); - packet->mad.hdr.sl = mad_recv_wc->wc->sl; - packet->mad.hdr.path_bits = mad_recv_wc->wc->dlid_path_bits; + goto err1; + + packet->length = mad_recv_wc->mad_len; + packet->recv_wc = mad_recv_wc; + + packet->mad.hdr.status = 0; + packet->mad.hdr.length = hdr_size(file) + mad_recv_wc->mad_len; + packet->mad.hdr.qpn = cpu_to_be32(mad_recv_wc->wc->src_qp); + packet->mad.hdr.lid = cpu_to_be16(mad_recv_wc->wc->slid); + packet->mad.hdr.sl = mad_recv_wc->wc->sl; + packet->mad.hdr.path_bits = mad_recv_wc->wc->dlid_path_bits; + packet->mad.hdr.pkey_index = mad_recv_wc->wc->pkey_index; packet->mad.hdr.grh_present = !!(mad_recv_wc->wc->wc_flags & IB_WC_GRH); if (packet->mad.hdr.grh_present) { - /* XXX parse GRH */ - packet->mad.hdr.gid_index = 0; - packet->mad.hdr.hop_limit = 0; - packet->mad.hdr.traffic_class = 0; - memset(packet->mad.hdr.gid, 0, 16); - packet->mad.hdr.flow_label = 0; + struct ib_ah_attr ah_attr; + + ib_init_ah_from_wc(agent->device, agent->port_num, + mad_recv_wc->wc, mad_recv_wc->recv_buf.grh, + &ah_attr); + + packet->mad.hdr.gid_index = ah_attr.grh.sgid_index; + packet->mad.hdr.hop_limit = ah_attr.grh.hop_limit; + packet->mad.hdr.traffic_class = ah_attr.grh.traffic_class; + memcpy(packet->mad.hdr.gid, &ah_attr.grh.dgid, 16); + packet->mad.hdr.flow_label = cpu_to_be32(ah_attr.grh.flow_label); } if (queue_packet(file, agent, packet)) - kfree(packet); + goto err2; + return; -out: +err2: + kfree(packet); +err1: ib_free_recv_mad(mad_recv_wc); } +static ssize_t copy_recv_mad(struct ib_umad_file *file, char __user *buf, + struct ib_umad_packet *packet, size_t count) +{ + struct ib_mad_recv_buf *recv_buf; + int left, seg_payload, offset, max_seg_payload; + + /* We need enough room to copy the first (or only) MAD segment. */ + recv_buf = &packet->recv_wc->recv_buf; + if ((packet->length <= sizeof (*recv_buf->mad) && + count < hdr_size(file) + packet->length) || + (packet->length > sizeof (*recv_buf->mad) && + count < hdr_size(file) + sizeof (*recv_buf->mad))) + return -EINVAL; + + if (copy_to_user(buf, &packet->mad, hdr_size(file))) + return -EFAULT; + + buf += hdr_size(file); + seg_payload = min_t(int, packet->length, sizeof (*recv_buf->mad)); + if (copy_to_user(buf, recv_buf->mad, seg_payload)) + return -EFAULT; + + if (seg_payload < packet->length) { + /* + * Multipacket RMPP MAD message. Copy remainder of message. + * Note that last segment may have a shorter payload. + */ + if (count < hdr_size(file) + packet->length) { + /* + * The buffer is too small, return the first RMPP segment, + * which includes the RMPP message length. + */ + return -ENOSPC; + } + offset = ib_get_mad_data_offset(recv_buf->mad->mad_hdr.mgmt_class); + max_seg_payload = sizeof (struct ib_mad) - offset; + + for (left = packet->length - seg_payload, buf += seg_payload; + left; left -= seg_payload, buf += seg_payload) { + recv_buf = container_of(recv_buf->list.next, + struct ib_mad_recv_buf, list); + seg_payload = min(left, max_seg_payload); + if (copy_to_user(buf, ((void *) recv_buf->mad) + offset, + seg_payload)) + return -EFAULT; + } + } + return hdr_size(file) + packet->length; +} + +static ssize_t copy_send_mad(struct ib_umad_file *file, char __user *buf, + struct ib_umad_packet *packet, size_t count) +{ + ssize_t size = hdr_size(file) + packet->length; + + if (count < size) + return -EINVAL; + + if (copy_to_user(buf, &packet->mad, hdr_size(file))) + return -EFAULT; + + buf += hdr_size(file); + + if (copy_to_user(buf, packet->mad.data, packet->length)) + return -EFAULT; + + return size; +} + static ssize_t ib_umad_read(struct file *filp, char __user *buf, size_t count, loff_t *pos) { @@ -253,13 +332,13 @@ static ssize_t ib_umad_read(struct file *filp, char __user *buf, struct ib_umad_packet *packet; ssize_t ret; - if (count < sizeof (struct ib_user_mad) + sizeof (struct ib_mad)) + if (count < hdr_size(file)) return -EINVAL; - spin_lock_irq(&file->recv_lock); + mutex_lock(&file->mutex); while (list_empty(&file->recv_list)) { - spin_unlock_irq(&file->recv_lock); + mutex_unlock(&file->mutex); if (filp->f_flags & O_NONBLOCK) return -EAGAIN; @@ -268,36 +347,97 @@ static ssize_t ib_umad_read(struct file *filp, char __user *buf, !list_empty(&file->recv_list))) return -ERESTARTSYS; - spin_lock_irq(&file->recv_lock); + mutex_lock(&file->mutex); } packet = list_entry(file->recv_list.next, struct ib_umad_packet, list); list_del(&packet->list); - spin_unlock_irq(&file->recv_lock); + mutex_unlock(&file->mutex); - if (count < packet->length + sizeof (struct ib_user_mad)) { - /* Return length needed (and first RMPP segment) if too small */ - if (copy_to_user(buf, &packet->mad, - sizeof (struct ib_user_mad) + sizeof (struct ib_mad))) - ret = -EFAULT; - else - ret = -ENOSPC; - } else if (copy_to_user(buf, &packet->mad, - packet->length + sizeof (struct ib_user_mad))) - ret = -EFAULT; + if (packet->recv_wc) + ret = copy_recv_mad(file, buf, packet, count); else - ret = packet->length + sizeof (struct ib_user_mad); + ret = copy_send_mad(file, buf, packet, count); + if (ret < 0) { /* Requeue packet */ - spin_lock_irq(&file->recv_lock); + mutex_lock(&file->mutex); list_add(&packet->list, &file->recv_list); - spin_unlock_irq(&file->recv_lock); - } else + mutex_unlock(&file->mutex); + } else { + if (packet->recv_wc) + ib_free_recv_mad(packet->recv_wc); kfree(packet); + } return ret; } +static int copy_rmpp_mad(struct ib_mad_send_buf *msg, const char __user *buf) +{ + int left, seg; + + /* Copy class specific header */ + if ((msg->hdr_len > IB_MGMT_RMPP_HDR) && + copy_from_user(msg->mad + IB_MGMT_RMPP_HDR, buf + IB_MGMT_RMPP_HDR, + msg->hdr_len - IB_MGMT_RMPP_HDR)) + return -EFAULT; + + /* All headers are in place. Copy data segments. */ + for (seg = 1, left = msg->data_len, buf += msg->hdr_len; left > 0; + seg++, left -= msg->seg_size, buf += msg->seg_size) { + if (copy_from_user(ib_get_rmpp_segment(msg, seg), buf, + min(left, msg->seg_size))) + return -EFAULT; + } + return 0; +} + +static int same_destination(struct ib_user_mad_hdr *hdr1, + struct ib_user_mad_hdr *hdr2) +{ + if (!hdr1->grh_present && !hdr2->grh_present) + return (hdr1->lid == hdr2->lid); + + if (hdr1->grh_present && hdr2->grh_present) + return !memcmp(hdr1->gid, hdr2->gid, 16); + + return 0; +} + +static int is_duplicate(struct ib_umad_file *file, + struct ib_umad_packet *packet) +{ + struct ib_umad_packet *sent_packet; + struct ib_mad_hdr *sent_hdr, *hdr; + + hdr = (struct ib_mad_hdr *) packet->mad.data; + list_for_each_entry(sent_packet, &file->send_list, list) { + sent_hdr = (struct ib_mad_hdr *) sent_packet->mad.data; + + if ((hdr->tid != sent_hdr->tid) || + (hdr->mgmt_class != sent_hdr->mgmt_class)) + continue; + + /* + * No need to be overly clever here. If two new operations have + * the same TID, reject the second as a duplicate. This is more + * restrictive than required by the spec. + */ + if (!ib_response_mad((struct ib_mad *) hdr)) { + if (!ib_response_mad((struct ib_mad *) sent_hdr)) + return 1; + continue; + } else if (!ib_response_mad((struct ib_mad *) sent_hdr)) + continue; + + if (same_destination(&packet->mad.hdr, &sent_packet->mad.hdr)) + return 1; + } + + return 0; +} + static ssize_t ib_umad_write(struct file *filp, const char __user *buf, size_t count, loff_t *pos) { @@ -307,32 +447,34 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, struct ib_ah_attr ah_attr; struct ib_ah *ah; struct ib_rmpp_mad *rmpp_mad; - u8 method; __be64 *tid; - int ret, length, hdr_len, copy_offset; - int rmpp_active, has_rmpp_header; + int ret, data_len, hdr_len, copy_offset, rmpp_active; - if (count < sizeof (struct ib_user_mad) + IB_MGMT_RMPP_HDR) + if (count < hdr_size(file) + IB_MGMT_RMPP_HDR) return -EINVAL; - length = count - sizeof (struct ib_user_mad); - packet = kmalloc(sizeof *packet + IB_MGMT_RMPP_HDR, GFP_KERNEL); + packet = kzalloc(sizeof *packet + IB_MGMT_RMPP_HDR, GFP_KERNEL); if (!packet) return -ENOMEM; - if (copy_from_user(&packet->mad, buf, - sizeof (struct ib_user_mad) + IB_MGMT_RMPP_HDR)) { + if (copy_from_user(&packet->mad, buf, hdr_size(file))) { ret = -EFAULT; goto err; } - if (packet->mad.hdr.id < 0 || - packet->mad.hdr.id >= IB_UMAD_MAX_AGENTS) { + if (packet->mad.hdr.id >= IB_UMAD_MAX_AGENTS) { ret = -EINVAL; goto err; } - down_read(&file->port->mutex); + buf += hdr_size(file); + + if (copy_from_user(packet->mad.data, buf, IB_MGMT_RMPP_HDR)) { + ret = -EFAULT; + goto err; + } + + mutex_lock(&file->mutex); agent = __get_agent(file, packet->mad.hdr.id); if (!agent) { @@ -348,8 +490,9 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, if (packet->mad.hdr.grh_present) { ah_attr.ah_flags = IB_AH_GRH; memcpy(ah_attr.grh.dgid.raw, packet->mad.hdr.gid, 16); - ah_attr.grh.flow_label = be32_to_cpu(packet->mad.hdr.flow_label); - ah_attr.grh.hop_limit = packet->mad.hdr.hop_limit; + ah_attr.grh.sgid_index = packet->mad.hdr.gid_index; + ah_attr.grh.flow_label = be32_to_cpu(packet->mad.hdr.flow_label); + ah_attr.grh.hop_limit = packet->mad.hdr.hop_limit; ah_attr.grh.traffic_class = packet->mad.hdr.traffic_class; } @@ -360,91 +503,84 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, } rmpp_mad = (struct ib_rmpp_mad *) packet->mad.data; - if (rmpp_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_ADM) { - hdr_len = IB_MGMT_SA_HDR; - copy_offset = IB_MGMT_RMPP_HDR; - has_rmpp_header = 1; - } else if (rmpp_mad->mad_hdr.mgmt_class >= IB_MGMT_CLASS_VENDOR_RANGE2_START && - rmpp_mad->mad_hdr.mgmt_class <= IB_MGMT_CLASS_VENDOR_RANGE2_END) { - hdr_len = IB_MGMT_VENDOR_HDR; - copy_offset = IB_MGMT_RMPP_HDR; - has_rmpp_header = 1; - } else { - hdr_len = IB_MGMT_MAD_HDR; + hdr_len = ib_get_mad_data_offset(rmpp_mad->mad_hdr.mgmt_class); + if (!ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)) { copy_offset = IB_MGMT_MAD_HDR; - has_rmpp_header = 0; - } - - if (has_rmpp_header) + rmpp_active = 0; + } else { + copy_offset = IB_MGMT_RMPP_HDR; rmpp_active = ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE; - else - rmpp_active = 0; - - /* Validate that the management class can support RMPP */ - if (rmpp_active && !agent->rmpp_version) { - ret = -EINVAL; - goto err_ah; } + data_len = count - hdr_size(file) - hdr_len; packet->msg = ib_create_send_mad(agent, be32_to_cpu(packet->mad.hdr.qpn), - 0, rmpp_active, - hdr_len, length - hdr_len, - GFP_KERNEL); + packet->mad.hdr.pkey_index, rmpp_active, + hdr_len, data_len, GFP_KERNEL); if (IS_ERR(packet->msg)) { ret = PTR_ERR(packet->msg); goto err_ah; } - packet->msg->ah = ah; + packet->msg->ah = ah; packet->msg->timeout_ms = packet->mad.hdr.timeout_ms; - packet->msg->retries = packet->mad.hdr.retries; + packet->msg->retries = packet->mad.hdr.retries; packet->msg->context[0] = packet; - /* Copy MAD headers (RMPP header in place) */ + /* Copy MAD header. Any RMPP header is already in place. */ memcpy(packet->msg->mad, packet->mad.data, IB_MGMT_MAD_HDR); - /* Now, copy rest of message from user into send buffer */ - if (copy_from_user(packet->msg->mad + copy_offset, - buf + sizeof (struct ib_user_mad) + copy_offset, - length - copy_offset)) { - ret = -EFAULT; - goto err_msg; + + if (!rmpp_active) { + if (copy_from_user(packet->msg->mad + copy_offset, + buf + copy_offset, + hdr_len + data_len - copy_offset)) { + ret = -EFAULT; + goto err_msg; + } + } else { + ret = copy_rmpp_mad(packet->msg, buf); + if (ret) + goto err_msg; } /* - * If userspace is generating a request that will generate a - * response, we need to make sure the high-order part of the - * transaction ID matches the agent being used to send the - * MAD. + * Set the high-order part of the transaction ID to make MADs from + * different agents unique, and allow routing responses back to the + * original requestor. */ - method = ((struct ib_mad_hdr *) packet->msg->mad)->method; - - if (!(method & IB_MGMT_METHOD_RESP) && - method != IB_MGMT_METHOD_TRAP_REPRESS && - method != IB_MGMT_METHOD_SEND) { + if (!ib_response_mad(packet->msg->mad)) { tid = &((struct ib_mad_hdr *) packet->msg->mad)->tid; *tid = cpu_to_be64(((u64) agent->hi_tid) << 32 | (be64_to_cpup(tid) & 0xffffffff)); + rmpp_mad->mad_hdr.tid = *tid; } - ret = ib_post_send_mad(packet->msg, NULL); - if (ret) + spin_lock_irq(&file->send_lock); + ret = is_duplicate(file, packet); + if (!ret) + list_add_tail(&packet->list, &file->send_list); + spin_unlock_irq(&file->send_lock); + if (ret) { + ret = -EINVAL; goto err_msg; + } - up_read(&file->port->mutex); + ret = ib_post_send_mad(packet->msg, NULL); + if (ret) + goto err_send; + mutex_unlock(&file->mutex); return count; +err_send: + dequeue_send(file, packet); err_msg: ib_free_send_mad(packet->msg); - err_ah: ib_destroy_ah(ah); - err_up: - up_read(&file->port->mutex); - + mutex_unlock(&file->mutex); err: kfree(packet); return ret; @@ -465,22 +601,24 @@ static unsigned int ib_umad_poll(struct file *filp, struct poll_table_struct *wa return mask; } -static int ib_umad_reg_agent(struct ib_umad_file *file, unsigned long arg) +static int ib_umad_reg_agent(struct ib_umad_file *file, void __user *arg, + int compat_method_mask) { struct ib_user_mad_reg_req ureq; struct ib_mad_reg_req req; - struct ib_mad_agent *agent; + struct ib_mad_agent *agent = NULL; int agent_id; int ret; - down_write(&file->port->mutex); + mutex_lock(&file->port->file_mutex); + mutex_lock(&file->mutex); if (!file->port->ib_dev) { ret = -EPIPE; goto out; } - if (copy_from_user(&ureq, (void __user *) arg, sizeof ureq)) { + if (copy_from_user(&ureq, arg, sizeof ureq)) { ret = -EFAULT; goto out; } @@ -501,8 +639,18 @@ found: if (ureq.mgmt_class) { req.mgmt_class = ureq.mgmt_class; req.mgmt_class_version = ureq.mgmt_class_version; - memcpy(req.method_mask, ureq.method_mask, sizeof req.method_mask); - memcpy(req.oui, ureq.oui, sizeof req.oui); + memcpy(req.oui, ureq.oui, sizeof req.oui); + + if (compat_method_mask) { + u32 *umm = (u32 *) ureq.method_mask; + int i; + + for (i = 0; i < BITS_TO_LONGS(IB_MGMT_MAX_METHODS); ++i) + req.method_mask[i] = + umm[i * 2] | ((u64) umm[i * 2 + 1] << 32); + } else + memcpy(req.method_mask, ureq.method_mask, + sizeof req.method_mask); } agent = ib_register_mad_agent(file->port->ib_dev, file->port->port_num, @@ -512,36 +660,53 @@ found: send_handler, recv_handler, file); if (IS_ERR(agent)) { ret = PTR_ERR(agent); + agent = NULL; goto out; } if (put_user(agent_id, (u32 __user *) (arg + offsetof(struct ib_user_mad_reg_req, id)))) { ret = -EFAULT; - ib_unregister_mad_agent(agent); goto out; } + if (!file->already_used) { + file->already_used = 1; + if (!file->use_pkey_index) { + printk(KERN_WARNING "user_mad: process %s did not enable " + "P_Key index support.\n", current->comm); + printk(KERN_WARNING "user_mad: Documentation/infiniband/user_mad.txt " + "has info on the new ABI.\n"); + } + } + file->agent[agent_id] = agent; ret = 0; out: - up_write(&file->port->mutex); + mutex_unlock(&file->mutex); + + if (ret && agent) + ib_unregister_mad_agent(agent); + + mutex_unlock(&file->port->file_mutex); + return ret; } -static int ib_umad_unreg_agent(struct ib_umad_file *file, unsigned long arg) +static int ib_umad_unreg_agent(struct ib_umad_file *file, u32 __user *arg) { struct ib_mad_agent *agent = NULL; u32 id; int ret = 0; - if (get_user(id, (u32 __user *) arg)) + if (get_user(id, arg)) return -EFAULT; - down_write(&file->port->mutex); + mutex_lock(&file->port->file_mutex); + mutex_lock(&file->mutex); - if (id < 0 || id >= IB_UMAD_MAX_AGENTS || !__get_agent(file, id)) { + if (id >= IB_UMAD_MAX_AGENTS || !__get_agent(file, id)) { ret = -EINVAL; goto out; } @@ -550,11 +715,27 @@ static int ib_umad_unreg_agent(struct ib_umad_file *file, unsigned long arg) file->agent[id] = NULL; out: - up_write(&file->port->mutex); + mutex_unlock(&file->mutex); if (agent) ib_unregister_mad_agent(agent); + mutex_unlock(&file->port->file_mutex); + + return ret; +} + +static long ib_umad_enable_pkey(struct ib_umad_file *file) +{ + int ret = 0; + + mutex_lock(&file->mutex); + if (file->already_used) + ret = -EINVAL; + else + file->use_pkey_index = 1; + mutex_unlock(&file->mutex); + return ret; } @@ -563,45 +744,64 @@ static long ib_umad_ioctl(struct file *filp, unsigned int cmd, { switch (cmd) { case IB_USER_MAD_REGISTER_AGENT: - return ib_umad_reg_agent(filp->private_data, arg); + return ib_umad_reg_agent(filp->private_data, (void __user *) arg, 0); case IB_USER_MAD_UNREGISTER_AGENT: - return ib_umad_unreg_agent(filp->private_data, arg); + return ib_umad_unreg_agent(filp->private_data, (__u32 __user *) arg); + case IB_USER_MAD_ENABLE_PKEY: + return ib_umad_enable_pkey(filp->private_data); default: return -ENOIOCTLCMD; } } +#ifdef CONFIG_COMPAT +static long ib_umad_compat_ioctl(struct file *filp, unsigned int cmd, + unsigned long arg) +{ + switch (cmd) { + case IB_USER_MAD_REGISTER_AGENT: + return ib_umad_reg_agent(filp->private_data, compat_ptr(arg), 1); + case IB_USER_MAD_UNREGISTER_AGENT: + return ib_umad_unreg_agent(filp->private_data, compat_ptr(arg)); + case IB_USER_MAD_ENABLE_PKEY: + return ib_umad_enable_pkey(filp->private_data); + default: + return -ENOIOCTLCMD; + } +} +#endif + +/* + * ib_umad_open() does not need the BKL: + * + * - the ib_umad_port structures are properly reference counted, and + * everything else is purely local to the file being created, so + * races against other open calls are not a problem; + * - the ioctl method does not affect any global state outside of the + * file structure being operated on; + */ static int ib_umad_open(struct inode *inode, struct file *filp) { struct ib_umad_port *port; struct ib_umad_file *file; - int ret = 0; - - spin_lock(&port_lock); - port = umad_port[iminor(inode) - IB_UMAD_MINOR_BASE]; - if (port) - kref_get(&port->umad_dev->ref); - spin_unlock(&port_lock); + int ret = -ENXIO; - if (!port) - return -ENXIO; + port = container_of(inode->i_cdev, struct ib_umad_port, cdev); - down_write(&port->mutex); + mutex_lock(&port->file_mutex); - if (!port->ib_dev) { - ret = -ENXIO; + if (!port->ib_dev) goto out; - } + ret = -ENOMEM; file = kzalloc(sizeof *file, GFP_KERNEL); - if (!file) { - kref_put(&port->umad_dev->ref, ib_umad_release_dev); - ret = -ENOMEM; + if (!file) goto out; - } - spin_lock_init(&file->recv_lock); + mutex_init(&file->mutex); + spin_lock_init(&file->send_lock); INIT_LIST_HEAD(&file->recv_list); + INIT_LIST_HEAD(&file->send_list); init_waitqueue_head(&file->recv_wait); file->port = port; @@ -609,8 +809,17 @@ static int ib_umad_open(struct inode *inode, struct file *filp) list_add_tail(&file->port_list, &port->file_list); + ret = nonseekable_open(inode, filp); + if (ret) { + list_del(&file->port_list); + kfree(file); + goto out; + } + + kobject_get(&port->umad_dev->kobj); + out: - up_write(&port->mutex); + mutex_unlock(&port->file_mutex); return ret; } @@ -622,40 +831,47 @@ static int ib_umad_close(struct inode *inode, struct file *filp) int already_dead; int i; - down_write(&file->port->mutex); + mutex_lock(&file->port->file_mutex); + mutex_lock(&file->mutex); already_dead = file->agents_dead; file->agents_dead = 1; - list_for_each_entry_safe(packet, tmp, &file->recv_list, list) + list_for_each_entry_safe(packet, tmp, &file->recv_list, list) { + if (packet->recv_wc) + ib_free_recv_mad(packet->recv_wc); kfree(packet); + } list_del(&file->port_list); - downgrade_write(&file->port->mutex); + mutex_unlock(&file->mutex); if (!already_dead) for (i = 0; i < IB_UMAD_MAX_AGENTS; ++i) if (file->agent[i]) ib_unregister_mad_agent(file->agent[i]); - up_read(&file->port->mutex); + mutex_unlock(&file->port->file_mutex); kfree(file); - kref_put(&dev->ref, ib_umad_release_dev); + kobject_put(&dev->kobj); return 0; } -static struct file_operations umad_fops = { - .owner = THIS_MODULE, - .read = ib_umad_read, - .write = ib_umad_write, - .poll = ib_umad_poll, +static const struct file_operations umad_fops = { + .owner = THIS_MODULE, + .read = ib_umad_read, + .write = ib_umad_write, + .poll = ib_umad_poll, .unlocked_ioctl = ib_umad_ioctl, - .compat_ioctl = ib_umad_ioctl, - .open = ib_umad_open, - .release = ib_umad_close +#ifdef CONFIG_COMPAT + .compat_ioctl = ib_umad_compat_ioctl, +#endif + .open = ib_umad_open, + .release = ib_umad_close, + .llseek = no_llseek, }; static int ib_umad_sm_open(struct inode *inode, struct file *filp) @@ -666,14 +882,7 @@ static int ib_umad_sm_open(struct inode *inode, struct file *filp) }; int ret; - spin_lock(&port_lock); - port = umad_port[iminor(inode) - IB_UMAD_MINOR_BASE - IB_UMAD_MAX_PORTS]; - if (port) - kref_get(&port->umad_dev->ref); - spin_unlock(&port_lock); - - if (!port) - return -ENXIO; + port = container_of(inode->i_cdev, struct ib_umad_port, sm_cdev); if (filp->f_flags & O_NONBLOCK) { if (down_trylock(&port->sm_sem)) { @@ -688,17 +897,27 @@ static int ib_umad_sm_open(struct inode *inode, struct file *filp) } ret = ib_modify_port(port->ib_dev, port->port_num, 0, &props); - if (ret) { - up(&port->sm_sem); - goto fail; - } + if (ret) + goto err_up_sem; filp->private_data = port; + ret = nonseekable_open(inode, filp); + if (ret) + goto err_clr_sm_cap; + + kobject_get(&port->umad_dev->kobj); + return 0; +err_clr_sm_cap: + swap(props.set_port_cap_mask, props.clr_port_cap_mask); + ib_modify_port(port->ib_dev, port->port_num, 0, &props); + +err_up_sem: + up(&port->sm_sem); + fail: - kref_put(&port->umad_dev->ref, ib_umad_release_dev); return ret; } @@ -710,22 +929,23 @@ static int ib_umad_sm_close(struct inode *inode, struct file *filp) }; int ret = 0; - down_write(&port->mutex); + mutex_lock(&port->file_mutex); if (port->ib_dev) ret = ib_modify_port(port->ib_dev, port->port_num, 0, &props); - up_write(&port->mutex); + mutex_unlock(&port->file_mutex); up(&port->sm_sem); - kref_put(&port->umad_dev->ref, ib_umad_release_dev); + kobject_put(&port->umad_dev->kobj); return ret; } -static struct file_operations umad_sm_fops = { - .owner = THIS_MODULE, - .open = ib_umad_sm_open, - .release = ib_umad_sm_close +static const struct file_operations umad_sm_fops = { + .owner = THIS_MODULE, + .open = ib_umad_sm_open, + .release = ib_umad_sm_close, + .llseek = no_llseek, }; static struct ib_client umad_client = { @@ -734,113 +954,141 @@ static struct ib_client umad_client = { .remove = ib_umad_remove_one }; -static ssize_t show_ibdev(struct class_device *class_dev, char *buf) +static ssize_t show_ibdev(struct device *dev, struct device_attribute *attr, + char *buf) { - struct ib_umad_port *port = class_get_devdata(class_dev); + struct ib_umad_port *port = dev_get_drvdata(dev); if (!port) return -ENODEV; return sprintf(buf, "%s\n", port->ib_dev->name); } -static CLASS_DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL); +static DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL); -static ssize_t show_port(struct class_device *class_dev, char *buf) +static ssize_t show_port(struct device *dev, struct device_attribute *attr, + char *buf) { - struct ib_umad_port *port = class_get_devdata(class_dev); + struct ib_umad_port *port = dev_get_drvdata(dev); if (!port) return -ENODEV; return sprintf(buf, "%d\n", port->port_num); } -static CLASS_DEVICE_ATTR(port, S_IRUGO, show_port, NULL); +static DEVICE_ATTR(port, S_IRUGO, show_port, NULL); -static ssize_t show_abi_version(struct class *class, char *buf) +static CLASS_ATTR_STRING(abi_version, S_IRUGO, + __stringify(IB_USER_MAD_ABI_VERSION)); + +static dev_t overflow_maj; +static DECLARE_BITMAP(overflow_map, IB_UMAD_MAX_PORTS); +static int find_overflow_devnum(void) { - return sprintf(buf, "%d\n", IB_USER_MAD_ABI_VERSION); + int ret; + + if (!overflow_maj) { + ret = alloc_chrdev_region(&overflow_maj, 0, IB_UMAD_MAX_PORTS * 2, + "infiniband_mad"); + if (ret) { + printk(KERN_ERR "user_mad: couldn't register dynamic device number\n"); + return ret; + } + } + + ret = find_first_zero_bit(overflow_map, IB_UMAD_MAX_PORTS); + if (ret >= IB_UMAD_MAX_PORTS) + return -1; + + return ret; } -static CLASS_ATTR(abi_version, S_IRUGO, show_abi_version, NULL); static int ib_umad_init_port(struct ib_device *device, int port_num, + struct ib_umad_device *umad_dev, struct ib_umad_port *port) { + int devnum; + dev_t base; + spin_lock(&port_lock); - port->dev_num = find_first_zero_bit(dev_map, IB_UMAD_MAX_PORTS); - if (port->dev_num >= IB_UMAD_MAX_PORTS) { + devnum = find_first_zero_bit(dev_map, IB_UMAD_MAX_PORTS); + if (devnum >= IB_UMAD_MAX_PORTS) { spin_unlock(&port_lock); - return -1; + devnum = find_overflow_devnum(); + if (devnum < 0) + return -1; + + spin_lock(&port_lock); + port->dev_num = devnum + IB_UMAD_MAX_PORTS; + base = devnum + overflow_maj; + set_bit(devnum, overflow_map); + } else { + port->dev_num = devnum; + base = devnum + base_dev; + set_bit(devnum, dev_map); } - set_bit(port->dev_num, dev_map); spin_unlock(&port_lock); port->ib_dev = device; port->port_num = port_num; - init_MUTEX(&port->sm_sem); - init_rwsem(&port->mutex); + sema_init(&port->sm_sem, 1); + mutex_init(&port->file_mutex); INIT_LIST_HEAD(&port->file_list); - port->dev = cdev_alloc(); - if (!port->dev) - return -1; - port->dev->owner = THIS_MODULE; - port->dev->ops = &umad_fops; - kobject_set_name(&port->dev->kobj, "umad%d", port->dev_num); - if (cdev_add(port->dev, base_dev + port->dev_num, 1)) + cdev_init(&port->cdev, &umad_fops); + port->cdev.owner = THIS_MODULE; + port->cdev.kobj.parent = &umad_dev->kobj; + kobject_set_name(&port->cdev.kobj, "umad%d", port->dev_num); + if (cdev_add(&port->cdev, base, 1)) goto err_cdev; - port->class_dev = class_device_create(umad_class, NULL, port->dev->dev, - device->dma_device, - "umad%d", port->dev_num); - if (IS_ERR(port->class_dev)) + port->dev = device_create(umad_class, device->dma_device, + port->cdev.dev, port, + "umad%d", port->dev_num); + if (IS_ERR(port->dev)) goto err_cdev; - if (class_device_create_file(port->class_dev, &class_device_attr_ibdev)) - goto err_class; - if (class_device_create_file(port->class_dev, &class_device_attr_port)) - goto err_class; - - port->sm_dev = cdev_alloc(); - if (!port->sm_dev) - goto err_class; - port->sm_dev->owner = THIS_MODULE; - port->sm_dev->ops = &umad_sm_fops; - kobject_set_name(&port->sm_dev->kobj, "issm%d", port->dev_num); - if (cdev_add(port->sm_dev, base_dev + port->dev_num + IB_UMAD_MAX_PORTS, 1)) + if (device_create_file(port->dev, &dev_attr_ibdev)) + goto err_dev; + if (device_create_file(port->dev, &dev_attr_port)) + goto err_dev; + + base += IB_UMAD_MAX_PORTS; + cdev_init(&port->sm_cdev, &umad_sm_fops); + port->sm_cdev.owner = THIS_MODULE; + port->sm_cdev.kobj.parent = &umad_dev->kobj; + kobject_set_name(&port->sm_cdev.kobj, "issm%d", port->dev_num); + if (cdev_add(&port->sm_cdev, base, 1)) goto err_sm_cdev; - port->sm_class_dev = class_device_create(umad_class, NULL, port->sm_dev->dev, - device->dma_device, - "issm%d", port->dev_num); - if (IS_ERR(port->sm_class_dev)) + port->sm_dev = device_create(umad_class, device->dma_device, + port->sm_cdev.dev, port, + "issm%d", port->dev_num); + if (IS_ERR(port->sm_dev)) goto err_sm_cdev; - class_set_devdata(port->class_dev, port); - class_set_devdata(port->sm_class_dev, port); - - if (class_device_create_file(port->sm_class_dev, &class_device_attr_ibdev)) - goto err_sm_class; - if (class_device_create_file(port->sm_class_dev, &class_device_attr_port)) - goto err_sm_class; - - spin_lock(&port_lock); - umad_port[port->dev_num] = port; - spin_unlock(&port_lock); + if (device_create_file(port->sm_dev, &dev_attr_ibdev)) + goto err_sm_dev; + if (device_create_file(port->sm_dev, &dev_attr_port)) + goto err_sm_dev; return 0; -err_sm_class: - class_device_destroy(umad_class, port->sm_dev->dev); +err_sm_dev: + device_destroy(umad_class, port->sm_cdev.dev); err_sm_cdev: - cdev_del(port->sm_dev); + cdev_del(&port->sm_cdev); -err_class: - class_device_destroy(umad_class, port->dev->dev); +err_dev: + device_destroy(umad_class, port->cdev.dev); err_cdev: - cdev_del(port->dev); - clear_bit(port->dev_num, dev_map); + cdev_del(&port->cdev); + if (port->dev_num < IB_UMAD_MAX_PORTS) + clear_bit(devnum, dev_map); + else + clear_bit(devnum, overflow_map); return -1; } @@ -850,57 +1098,35 @@ static void ib_umad_kill_port(struct ib_umad_port *port) struct ib_umad_file *file; int id; - class_set_devdata(port->class_dev, NULL); - class_set_devdata(port->sm_class_dev, NULL); + dev_set_drvdata(port->dev, NULL); + dev_set_drvdata(port->sm_dev, NULL); - class_device_destroy(umad_class, port->dev->dev); - class_device_destroy(umad_class, port->sm_dev->dev); + device_destroy(umad_class, port->cdev.dev); + device_destroy(umad_class, port->sm_cdev.dev); - cdev_del(port->dev); - cdev_del(port->sm_dev); - - spin_lock(&port_lock); - umad_port[port->dev_num] = NULL; - spin_unlock(&port_lock); + cdev_del(&port->cdev); + cdev_del(&port->sm_cdev); - down_write(&port->mutex); + mutex_lock(&port->file_mutex); port->ib_dev = NULL; - /* - * Now go through the list of files attached to this port and - * unregister all of their MAD agents. We need to hold - * port->mutex while doing this to avoid racing with - * ib_umad_close(), but we can't hold the mutex for writing - * while calling ib_unregister_mad_agent(), since that might - * deadlock by calling back into queue_packet(). So we - * downgrade our lock to a read lock, and then drop and - * reacquire the write lock for the next iteration. - * - * We do list_del_init() on the file's list_head so that the - * list_del in ib_umad_close() is still OK, even after the - * file is removed from the list. - */ - while (!list_empty(&port->file_list)) { - file = list_entry(port->file_list.next, struct ib_umad_file, - port_list); - + list_for_each_entry(file, &port->file_list, port_list) { + mutex_lock(&file->mutex); file->agents_dead = 1; - list_del_init(&file->port_list); - - downgrade_write(&port->mutex); + mutex_unlock(&file->mutex); for (id = 0; id < IB_UMAD_MAX_AGENTS; ++id) if (file->agent[id]) ib_unregister_mad_agent(file->agent[id]); - - up_read(&port->mutex); - down_write(&port->mutex); } - up_write(&port->mutex); + mutex_unlock(&port->file_mutex); - clear_bit(port->dev_num, dev_map); + if (port->dev_num < IB_UMAD_MAX_PORTS) + clear_bit(port->dev_num, dev_map); + else + clear_bit(port->dev_num - IB_UMAD_MAX_PORTS, overflow_map); } static void ib_umad_add_one(struct ib_device *device) @@ -908,7 +1134,10 @@ static void ib_umad_add_one(struct ib_device *device) struct ib_umad_device *umad_dev; int s, e, i; - if (device->node_type == IB_NODE_SWITCH) + if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB) + return; + + if (device->node_type == RDMA_NODE_IB_SWITCH) s = e = 0; else { s = 1; @@ -921,7 +1150,7 @@ static void ib_umad_add_one(struct ib_device *device) if (!umad_dev) return; - kref_init(&umad_dev->ref); + kobject_init(&umad_dev->kobj, &ib_umad_dev_ktype); umad_dev->start_port = s; umad_dev->end_port = e; @@ -929,7 +1158,8 @@ static void ib_umad_add_one(struct ib_device *device) for (i = s; i <= e; ++i) { umad_dev->port[i - s].umad_dev = umad_dev; - if (ib_umad_init_port(device, i, &umad_dev->port[i - s])) + if (ib_umad_init_port(device, i, umad_dev, + &umad_dev->port[i - s])) goto err; } @@ -941,7 +1171,7 @@ err: while (--i >= s) ib_umad_kill_port(&umad_dev->port[i - s]); - kref_put(&umad_dev->ref, ib_umad_release_dev); + kobject_put(&umad_dev->kobj); } static void ib_umad_remove_one(struct ib_device *device) @@ -955,7 +1185,12 @@ static void ib_umad_remove_one(struct ib_device *device) for (i = 0; i <= umad_dev->end_port - umad_dev->start_port; ++i) ib_umad_kill_port(&umad_dev->port[i]); - kref_put(&umad_dev->ref, ib_umad_release_dev); + kobject_put(&umad_dev->kobj); +} + +static char *umad_devnode(struct device *dev, umode_t *mode) +{ + return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev)); } static int __init ib_umad_init(void) @@ -976,7 +1211,9 @@ static int __init ib_umad_init(void) goto out_chrdev; } - ret = class_create_file(umad_class, &class_attr_abi_version); + umad_class->devnode = umad_devnode; + + ret = class_create_file(umad_class, &class_attr_abi_version.attr); if (ret) { printk(KERN_ERR "user_mad: couldn't create abi_version attribute\n"); goto out_class; @@ -1005,6 +1242,8 @@ static void __exit ib_umad_cleanup(void) ib_unregister_client(&umad_client); class_destroy(umad_class); unregister_chrdev_region(base_dev, IB_UMAD_MAX_PORTS * 2); + if (overflow_maj) + unregister_chrdev_region(overflow_maj, IB_UMAD_MAX_PORTS * 2); } module_init(ib_umad_init); diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h index f7eecbc6af6..a283274a5a0 100644 --- a/drivers/infiniband/core/uverbs.h +++ b/drivers/infiniband/core/uverbs.h @@ -1,6 +1,6 @@ /* * Copyright (c) 2005 Topspin Communications. All rights reserved. - * Copyright (c) 2005 Cisco Systems. All rights reserved. + * Copyright (c) 2005, 2006 Cisco Systems. All rights reserved. * Copyright (c) 2005 Mellanox Technologies. All rights reserved. * Copyright (c) 2005 Voltaire, Inc. All rights reserved. * Copyright (c) 2005 PathScale, Inc. All rights reserved. @@ -32,8 +32,6 @@ * 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. - * - * $Id: uverbs.h 2559 2005-06-06 19:43:16Z roland $ */ #ifndef UVERBS_H @@ -42,10 +40,29 @@ #include <linux/kref.h> #include <linux/idr.h> #include <linux/mutex.h> +#include <linux/completion.h> +#include <linux/cdev.h> #include <rdma/ib_verbs.h> +#include <rdma/ib_umem.h> #include <rdma/ib_user_verbs.h> +#define INIT_UDATA(udata, ibuf, obuf, ilen, olen) \ + do { \ + (udata)->inbuf = (const void __user *) (ibuf); \ + (udata)->outbuf = (void __user *) (obuf); \ + (udata)->inlen = (ilen); \ + (udata)->outlen = (olen); \ + } while (0) + +#define INIT_UDATA_BUF_OR_NULL(udata, ibuf, obuf, ilen, olen) \ + do { \ + (udata)->inbuf = (ilen) ? (const void __user *) (ibuf) : NULL; \ + (udata)->outbuf = (olen) ? (void __user *) (obuf) : NULL; \ + (udata)->inlen = (ilen); \ + (udata)->outlen = (olen); \ + } while (0) + /* * Our lifetime rules for these structs are the following: * @@ -69,19 +86,22 @@ struct ib_uverbs_device { struct kref ref; - int devnum; - struct cdev *dev; - struct class_device *class_dev; - struct ib_device *ib_dev; int num_comp_vectors; + struct completion comp; + struct device *dev; + struct ib_device *ib_dev; + int devnum; + struct cdev cdev; + struct rb_root xrcd_tree; + struct mutex xrcd_tree_mutex; }; struct ib_uverbs_event_file { struct kref ref; - struct file *file; + int is_async; struct ib_uverbs_file *uverbs_file; spinlock_t lock; - int is_async; + int is_closed; wait_queue_head_t poll_wait; struct fasync_struct *async_queue; struct list_head event_list; @@ -118,9 +138,20 @@ struct ib_uevent_object { u32 events_reported; }; +struct ib_uxrcd_object { + struct ib_uobject uobject; + atomic_t refcnt; +}; + +struct ib_usrq_object { + struct ib_uevent_object uevent; + struct ib_uxrcd_object *uxrcd; +}; + struct ib_uqp_object { struct ib_uevent_object uevent; struct list_head mcast_list; + struct ib_uxrcd_object *uxrcd; }; struct ib_ucq_object { @@ -132,7 +163,7 @@ struct ib_ucq_object { u32 async_events_reported; }; -extern struct mutex ib_uverbs_idr_mutex; +extern spinlock_t ib_uverbs_idr_lock; extern struct idr ib_uverbs_pd_idr; extern struct idr ib_uverbs_mr_idr; extern struct idr ib_uverbs_mw_idr; @@ -140,10 +171,13 @@ extern struct idr ib_uverbs_ah_idr; extern struct idr ib_uverbs_cq_idr; extern struct idr ib_uverbs_qp_idr; extern struct idr ib_uverbs_srq_idr; +extern struct idr ib_uverbs_xrcd_idr; +extern struct idr ib_uverbs_rule_idr; + +void idr_remove_uobj(struct idr *idp, struct ib_uobject *uobj); struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file, - int is_async, int *fd); -void ib_uverbs_release_event_file(struct kref *ref); + int is_async); struct ib_uverbs_event_file *ib_uverbs_lookup_comp_file(int fd); void ib_uverbs_release_ucq(struct ib_uverbs_file *file, @@ -158,11 +192,23 @@ void ib_uverbs_qp_event_handler(struct ib_event *event, void *context_ptr); void ib_uverbs_srq_event_handler(struct ib_event *event, void *context_ptr); void ib_uverbs_event_handler(struct ib_event_handler *handler, struct ib_event *event); +void ib_uverbs_dealloc_xrcd(struct ib_uverbs_device *dev, struct ib_xrcd *xrcd); -int ib_umem_get(struct ib_device *dev, struct ib_umem *mem, - void *addr, size_t size, int write); -void ib_umem_release(struct ib_device *dev, struct ib_umem *umem); -void ib_umem_release_on_close(struct ib_device *dev, struct ib_umem *umem); +struct ib_uverbs_flow_spec { + union { + union { + struct ib_uverbs_flow_spec_hdr hdr; + struct { + __u32 type; + __u16 size; + __u16 reserved; + }; + }; + struct ib_uverbs_flow_spec_eth eth; + struct ib_uverbs_flow_spec_ipv4 ipv4; + struct ib_uverbs_flow_spec_tcp_udp tcp_udp; + }; +}; #define IB_UVERBS_DECLARE_CMD(name) \ ssize_t ib_uverbs_##name(struct ib_uverbs_file *file, \ @@ -176,12 +222,17 @@ IB_UVERBS_DECLARE_CMD(alloc_pd); IB_UVERBS_DECLARE_CMD(dealloc_pd); IB_UVERBS_DECLARE_CMD(reg_mr); IB_UVERBS_DECLARE_CMD(dereg_mr); +IB_UVERBS_DECLARE_CMD(alloc_mw); +IB_UVERBS_DECLARE_CMD(dealloc_mw); IB_UVERBS_DECLARE_CMD(create_comp_channel); IB_UVERBS_DECLARE_CMD(create_cq); +IB_UVERBS_DECLARE_CMD(resize_cq); IB_UVERBS_DECLARE_CMD(poll_cq); IB_UVERBS_DECLARE_CMD(req_notify_cq); IB_UVERBS_DECLARE_CMD(destroy_cq); IB_UVERBS_DECLARE_CMD(create_qp); +IB_UVERBS_DECLARE_CMD(open_qp); +IB_UVERBS_DECLARE_CMD(query_qp); IB_UVERBS_DECLARE_CMD(modify_qp); IB_UVERBS_DECLARE_CMD(destroy_qp); IB_UVERBS_DECLARE_CMD(post_send); @@ -193,6 +244,18 @@ IB_UVERBS_DECLARE_CMD(attach_mcast); IB_UVERBS_DECLARE_CMD(detach_mcast); IB_UVERBS_DECLARE_CMD(create_srq); IB_UVERBS_DECLARE_CMD(modify_srq); +IB_UVERBS_DECLARE_CMD(query_srq); IB_UVERBS_DECLARE_CMD(destroy_srq); +IB_UVERBS_DECLARE_CMD(create_xsrq); +IB_UVERBS_DECLARE_CMD(open_xrcd); +IB_UVERBS_DECLARE_CMD(close_xrcd); + +#define IB_UVERBS_DECLARE_EX_CMD(name) \ + int ib_uverbs_ex_##name(struct ib_uverbs_file *file, \ + struct ib_udata *ucore, \ + struct ib_udata *uhw) + +IB_UVERBS_DECLARE_EX_CMD(create_flow); +IB_UVERBS_DECLARE_EX_CMD(destroy_flow); #endif /* UVERBS_H */ diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c index 407b6284d7d..ea6203ee7bc 100644 --- a/drivers/infiniband/core/uverbs_cmd.c +++ b/drivers/infiniband/core/uverbs_cmd.c @@ -1,7 +1,8 @@ /* * Copyright (c) 2005 Topspin Communications. All rights reserved. - * Copyright (c) 2005 Cisco Systems. All rights reserved. + * Copyright (c) 2005, 2006, 2007 Cisco Systems. All rights reserved. * Copyright (c) 2005 PathScale, Inc. All rights reserved. + * Copyright (c) 2006 Mellanox Technologies. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -30,24 +31,254 @@ * 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. - * - * $Id: uverbs_cmd.c 2708 2005-06-24 17:27:21Z roland $ */ #include <linux/file.h> #include <linux/fs.h> +#include <linux/slab.h> #include <asm/uaccess.h> #include "uverbs.h" +#include "core_priv.h" + +struct uverbs_lock_class { + struct lock_class_key key; + char name[16]; +}; + +static struct uverbs_lock_class pd_lock_class = { .name = "PD-uobj" }; +static struct uverbs_lock_class mr_lock_class = { .name = "MR-uobj" }; +static struct uverbs_lock_class mw_lock_class = { .name = "MW-uobj" }; +static struct uverbs_lock_class cq_lock_class = { .name = "CQ-uobj" }; +static struct uverbs_lock_class qp_lock_class = { .name = "QP-uobj" }; +static struct uverbs_lock_class ah_lock_class = { .name = "AH-uobj" }; +static struct uverbs_lock_class srq_lock_class = { .name = "SRQ-uobj" }; +static struct uverbs_lock_class xrcd_lock_class = { .name = "XRCD-uobj" }; +static struct uverbs_lock_class rule_lock_class = { .name = "RULE-uobj" }; + +/* + * The ib_uobject locking scheme is as follows: + * + * - ib_uverbs_idr_lock protects the uverbs idrs themselves, so it + * needs to be held during all idr operations. When an object is + * looked up, a reference must be taken on the object's kref before + * dropping this lock. + * + * - Each object also has an rwsem. This rwsem must be held for + * reading while an operation that uses the object is performed. + * For example, while registering an MR, the associated PD's + * uobject.mutex must be held for reading. The rwsem must be held + * for writing while initializing or destroying an object. + * + * - In addition, each object has a "live" flag. If this flag is not + * set, then lookups of the object will fail even if it is found in + * the idr. This handles a reader that blocks and does not acquire + * the rwsem until after the object is destroyed. The destroy + * operation will set the live flag to 0 and then drop the rwsem; + * this will allow the reader to acquire the rwsem, see that the + * live flag is 0, and then drop the rwsem and its reference to + * object. The underlying storage will not be freed until the last + * reference to the object is dropped. + */ + +static void init_uobj(struct ib_uobject *uobj, u64 user_handle, + struct ib_ucontext *context, struct uverbs_lock_class *c) +{ + uobj->user_handle = user_handle; + uobj->context = context; + kref_init(&uobj->ref); + init_rwsem(&uobj->mutex); + lockdep_set_class_and_name(&uobj->mutex, &c->key, c->name); + uobj->live = 0; +} + +static void release_uobj(struct kref *kref) +{ + kfree(container_of(kref, struct ib_uobject, ref)); +} + +static void put_uobj(struct ib_uobject *uobj) +{ + kref_put(&uobj->ref, release_uobj); +} + +static void put_uobj_read(struct ib_uobject *uobj) +{ + up_read(&uobj->mutex); + put_uobj(uobj); +} + +static void put_uobj_write(struct ib_uobject *uobj) +{ + up_write(&uobj->mutex); + put_uobj(uobj); +} + +static int idr_add_uobj(struct idr *idr, struct ib_uobject *uobj) +{ + int ret; + + idr_preload(GFP_KERNEL); + spin_lock(&ib_uverbs_idr_lock); + + ret = idr_alloc(idr, uobj, 0, 0, GFP_NOWAIT); + if (ret >= 0) + uobj->id = ret; + + spin_unlock(&ib_uverbs_idr_lock); + idr_preload_end(); -#define INIT_UDATA(udata, ibuf, obuf, ilen, olen) \ - do { \ - (udata)->inbuf = (void __user *) (ibuf); \ - (udata)->outbuf = (void __user *) (obuf); \ - (udata)->inlen = (ilen); \ - (udata)->outlen = (olen); \ - } while (0) + return ret < 0 ? ret : 0; +} + +void idr_remove_uobj(struct idr *idr, struct ib_uobject *uobj) +{ + spin_lock(&ib_uverbs_idr_lock); + idr_remove(idr, uobj->id); + spin_unlock(&ib_uverbs_idr_lock); +} + +static struct ib_uobject *__idr_get_uobj(struct idr *idr, int id, + struct ib_ucontext *context) +{ + struct ib_uobject *uobj; + + spin_lock(&ib_uverbs_idr_lock); + uobj = idr_find(idr, id); + if (uobj) { + if (uobj->context == context) + kref_get(&uobj->ref); + else + uobj = NULL; + } + spin_unlock(&ib_uverbs_idr_lock); + + return uobj; +} + +static struct ib_uobject *idr_read_uobj(struct idr *idr, int id, + struct ib_ucontext *context, int nested) +{ + struct ib_uobject *uobj; + + uobj = __idr_get_uobj(idr, id, context); + if (!uobj) + return NULL; + + if (nested) + down_read_nested(&uobj->mutex, SINGLE_DEPTH_NESTING); + else + down_read(&uobj->mutex); + if (!uobj->live) { + put_uobj_read(uobj); + return NULL; + } + + return uobj; +} + +static struct ib_uobject *idr_write_uobj(struct idr *idr, int id, + struct ib_ucontext *context) +{ + struct ib_uobject *uobj; + + uobj = __idr_get_uobj(idr, id, context); + if (!uobj) + return NULL; + + down_write(&uobj->mutex); + if (!uobj->live) { + put_uobj_write(uobj); + return NULL; + } + + return uobj; +} + +static void *idr_read_obj(struct idr *idr, int id, struct ib_ucontext *context, + int nested) +{ + struct ib_uobject *uobj; + + uobj = idr_read_uobj(idr, id, context, nested); + return uobj ? uobj->object : NULL; +} + +static struct ib_pd *idr_read_pd(int pd_handle, struct ib_ucontext *context) +{ + return idr_read_obj(&ib_uverbs_pd_idr, pd_handle, context, 0); +} + +static void put_pd_read(struct ib_pd *pd) +{ + put_uobj_read(pd->uobject); +} + +static struct ib_cq *idr_read_cq(int cq_handle, struct ib_ucontext *context, int nested) +{ + return idr_read_obj(&ib_uverbs_cq_idr, cq_handle, context, nested); +} + +static void put_cq_read(struct ib_cq *cq) +{ + put_uobj_read(cq->uobject); +} + +static struct ib_ah *idr_read_ah(int ah_handle, struct ib_ucontext *context) +{ + return idr_read_obj(&ib_uverbs_ah_idr, ah_handle, context, 0); +} + +static void put_ah_read(struct ib_ah *ah) +{ + put_uobj_read(ah->uobject); +} + +static struct ib_qp *idr_read_qp(int qp_handle, struct ib_ucontext *context) +{ + return idr_read_obj(&ib_uverbs_qp_idr, qp_handle, context, 0); +} + +static struct ib_qp *idr_write_qp(int qp_handle, struct ib_ucontext *context) +{ + struct ib_uobject *uobj; + + uobj = idr_write_uobj(&ib_uverbs_qp_idr, qp_handle, context); + return uobj ? uobj->object : NULL; +} + +static void put_qp_read(struct ib_qp *qp) +{ + put_uobj_read(qp->uobject); +} + +static void put_qp_write(struct ib_qp *qp) +{ + put_uobj_write(qp->uobject); +} + +static struct ib_srq *idr_read_srq(int srq_handle, struct ib_ucontext *context) +{ + return idr_read_obj(&ib_uverbs_srq_idr, srq_handle, context, 0); +} + +static void put_srq_read(struct ib_srq *srq) +{ + put_uobj_read(srq->uobject); +} + +static struct ib_xrcd *idr_read_xrcd(int xrcd_handle, struct ib_ucontext *context, + struct ib_uobject **uobj) +{ + *uobj = idr_read_uobj(&ib_uverbs_xrcd_idr, xrcd_handle, context, 0); + return *uobj ? (*uobj)->object : NULL; +} + +static void put_xrcd_read(struct ib_uobject *uobj) +{ + put_uobj_read(uobj); +} ssize_t ib_uverbs_get_context(struct ib_uverbs_file *file, const char __user *buf, @@ -79,8 +310,10 @@ ssize_t ib_uverbs_get_context(struct ib_uverbs_file *file, in_len - sizeof cmd, out_len - sizeof resp); ucontext = ibdev->alloc_ucontext(ibdev, &udata); - if (IS_ERR(ucontext)) - return PTR_ERR(file->ucontext); + if (IS_ERR(ucontext)) { + ret = PTR_ERR(ucontext); + goto err; + } ucontext->device = ibdev; INIT_LIST_HEAD(&ucontext->pd_list); @@ -90,13 +323,21 @@ ssize_t ib_uverbs_get_context(struct ib_uverbs_file *file, INIT_LIST_HEAD(&ucontext->qp_list); INIT_LIST_HEAD(&ucontext->srq_list); INIT_LIST_HEAD(&ucontext->ah_list); + INIT_LIST_HEAD(&ucontext->xrcd_list); + INIT_LIST_HEAD(&ucontext->rule_list); + ucontext->closing = 0; resp.num_comp_vectors = file->device->num_comp_vectors; - filp = ib_uverbs_alloc_event_file(file, 1, &resp.async_fd); + ret = get_unused_fd_flags(O_CLOEXEC); + if (ret < 0) + goto err_free; + resp.async_fd = ret; + + filp = ib_uverbs_alloc_event_file(file, 1); if (IS_ERR(filp)) { ret = PTR_ERR(filp); - goto err_free; + goto err_fd; } if (copy_to_user((void __user *) (unsigned long) cmd.response, @@ -124,9 +365,11 @@ ssize_t ib_uverbs_get_context(struct ib_uverbs_file *file, return in_len; err_file: - put_unused_fd(resp.async_fd); fput(filp); +err_fd: + put_unused_fd(resp.async_fd); + err_free: ibdev->dealloc_ucontext(ucontext); @@ -244,6 +487,8 @@ ssize_t ib_uverbs_query_port(struct ib_uverbs_file *file, resp.active_width = attr.active_width; resp.active_speed = attr.active_speed; resp.phys_state = attr.phys_state; + resp.link_layer = rdma_port_get_link_layer(file->device->ib_dev, + cmd.port_num); if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) @@ -277,7 +522,8 @@ ssize_t ib_uverbs_alloc_pd(struct ib_uverbs_file *file, if (!uobj) return -ENOMEM; - uobj->context = file->ucontext; + init_uobj(uobj, 0, file->ucontext, &pd_lock_class); + down_write(&uobj->mutex); pd = file->device->ib_dev->alloc_pd(file->device->ib_dev, file->ucontext, &udata); @@ -290,20 +536,10 @@ ssize_t ib_uverbs_alloc_pd(struct ib_uverbs_file *file, pd->uobject = uobj; atomic_set(&pd->usecnt, 0); - mutex_lock(&ib_uverbs_idr_mutex); - -retry: - if (!idr_pre_get(&ib_uverbs_pd_idr, GFP_KERNEL)) { - ret = -ENOMEM; - goto err_up; - } - - ret = idr_get_new(&ib_uverbs_pd_idr, pd, &uobj->id); - - if (ret == -EAGAIN) - goto retry; + uobj->object = pd; + ret = idr_add_uobj(&ib_uverbs_pd_idr, uobj); if (ret) - goto err_up; + goto err_idr; memset(&resp, 0, sizeof resp); resp.pd_handle = uobj->id; @@ -311,26 +547,27 @@ retry: if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) { ret = -EFAULT; - goto err_idr; + goto err_copy; } mutex_lock(&file->mutex); list_add_tail(&uobj->list, &file->ucontext->pd_list); mutex_unlock(&file->mutex); - mutex_unlock(&ib_uverbs_idr_mutex); + uobj->live = 1; + + up_write(&uobj->mutex); return in_len; -err_idr: - idr_remove(&ib_uverbs_pd_idr, uobj->id); +err_copy: + idr_remove_uobj(&ib_uverbs_pd_idr, uobj); -err_up: - mutex_unlock(&ib_uverbs_idr_mutex); +err_idr: ib_dealloc_pd(pd); err: - kfree(uobj); + put_uobj_write(uobj); return ret; } @@ -339,37 +576,333 @@ ssize_t ib_uverbs_dealloc_pd(struct ib_uverbs_file *file, int in_len, int out_len) { struct ib_uverbs_dealloc_pd cmd; - struct ib_pd *pd; struct ib_uobject *uobj; - int ret = -EINVAL; + int ret; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + uobj = idr_write_uobj(&ib_uverbs_pd_idr, cmd.pd_handle, file->ucontext); + if (!uobj) + return -EINVAL; + + ret = ib_dealloc_pd(uobj->object); + if (!ret) + uobj->live = 0; + + put_uobj_write(uobj); + + if (ret) + return ret; + + idr_remove_uobj(&ib_uverbs_pd_idr, uobj); + + mutex_lock(&file->mutex); + list_del(&uobj->list); + mutex_unlock(&file->mutex); + + put_uobj(uobj); + + return in_len; +} + +struct xrcd_table_entry { + struct rb_node node; + struct ib_xrcd *xrcd; + struct inode *inode; +}; + +static int xrcd_table_insert(struct ib_uverbs_device *dev, + struct inode *inode, + struct ib_xrcd *xrcd) +{ + struct xrcd_table_entry *entry, *scan; + struct rb_node **p = &dev->xrcd_tree.rb_node; + struct rb_node *parent = NULL; + + entry = kmalloc(sizeof *entry, GFP_KERNEL); + if (!entry) + return -ENOMEM; + + entry->xrcd = xrcd; + entry->inode = inode; + + while (*p) { + parent = *p; + scan = rb_entry(parent, struct xrcd_table_entry, node); + + if (inode < scan->inode) { + p = &(*p)->rb_left; + } else if (inode > scan->inode) { + p = &(*p)->rb_right; + } else { + kfree(entry); + return -EEXIST; + } + } + + rb_link_node(&entry->node, parent, p); + rb_insert_color(&entry->node, &dev->xrcd_tree); + igrab(inode); + return 0; +} + +static struct xrcd_table_entry *xrcd_table_search(struct ib_uverbs_device *dev, + struct inode *inode) +{ + struct xrcd_table_entry *entry; + struct rb_node *p = dev->xrcd_tree.rb_node; + + while (p) { + entry = rb_entry(p, struct xrcd_table_entry, node); + + if (inode < entry->inode) + p = p->rb_left; + else if (inode > entry->inode) + p = p->rb_right; + else + return entry; + } + + return NULL; +} + +static struct ib_xrcd *find_xrcd(struct ib_uverbs_device *dev, struct inode *inode) +{ + struct xrcd_table_entry *entry; + + entry = xrcd_table_search(dev, inode); + if (!entry) + return NULL; + + return entry->xrcd; +} + +static void xrcd_table_delete(struct ib_uverbs_device *dev, + struct inode *inode) +{ + struct xrcd_table_entry *entry; + + entry = xrcd_table_search(dev, inode); + if (entry) { + iput(inode); + rb_erase(&entry->node, &dev->xrcd_tree); + kfree(entry); + } +} + +ssize_t ib_uverbs_open_xrcd(struct ib_uverbs_file *file, + const char __user *buf, int in_len, + int out_len) +{ + struct ib_uverbs_open_xrcd cmd; + struct ib_uverbs_open_xrcd_resp resp; + struct ib_udata udata; + struct ib_uxrcd_object *obj; + struct ib_xrcd *xrcd = NULL; + struct fd f = {NULL, 0}; + struct inode *inode = NULL; + int ret = 0; + int new_xrcd = 0; + + if (out_len < sizeof resp) + return -ENOSPC; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + INIT_UDATA(&udata, buf + sizeof cmd, + (unsigned long) cmd.response + sizeof resp, + in_len - sizeof cmd, out_len - sizeof resp); + + mutex_lock(&file->device->xrcd_tree_mutex); + + if (cmd.fd != -1) { + /* search for file descriptor */ + f = fdget(cmd.fd); + if (!f.file) { + ret = -EBADF; + goto err_tree_mutex_unlock; + } + + inode = file_inode(f.file); + xrcd = find_xrcd(file->device, inode); + if (!xrcd && !(cmd.oflags & O_CREAT)) { + /* no file descriptor. Need CREATE flag */ + ret = -EAGAIN; + goto err_tree_mutex_unlock; + } + + if (xrcd && cmd.oflags & O_EXCL) { + ret = -EINVAL; + goto err_tree_mutex_unlock; + } + } + + obj = kmalloc(sizeof *obj, GFP_KERNEL); + if (!obj) { + ret = -ENOMEM; + goto err_tree_mutex_unlock; + } + + init_uobj(&obj->uobject, 0, file->ucontext, &xrcd_lock_class); + + down_write(&obj->uobject.mutex); + + if (!xrcd) { + xrcd = file->device->ib_dev->alloc_xrcd(file->device->ib_dev, + file->ucontext, &udata); + if (IS_ERR(xrcd)) { + ret = PTR_ERR(xrcd); + goto err; + } + + xrcd->inode = inode; + xrcd->device = file->device->ib_dev; + atomic_set(&xrcd->usecnt, 0); + mutex_init(&xrcd->tgt_qp_mutex); + INIT_LIST_HEAD(&xrcd->tgt_qp_list); + new_xrcd = 1; + } + + atomic_set(&obj->refcnt, 0); + obj->uobject.object = xrcd; + ret = idr_add_uobj(&ib_uverbs_xrcd_idr, &obj->uobject); + if (ret) + goto err_idr; + + memset(&resp, 0, sizeof resp); + resp.xrcd_handle = obj->uobject.id; + + if (inode) { + if (new_xrcd) { + /* create new inode/xrcd table entry */ + ret = xrcd_table_insert(file->device, inode, xrcd); + if (ret) + goto err_insert_xrcd; + } + atomic_inc(&xrcd->usecnt); + } + + if (copy_to_user((void __user *) (unsigned long) cmd.response, + &resp, sizeof resp)) { + ret = -EFAULT; + goto err_copy; + } + + if (f.file) + fdput(f); + + mutex_lock(&file->mutex); + list_add_tail(&obj->uobject.list, &file->ucontext->xrcd_list); + mutex_unlock(&file->mutex); + + obj->uobject.live = 1; + up_write(&obj->uobject.mutex); + + mutex_unlock(&file->device->xrcd_tree_mutex); + return in_len; + +err_copy: + if (inode) { + if (new_xrcd) + xrcd_table_delete(file->device, inode); + atomic_dec(&xrcd->usecnt); + } + +err_insert_xrcd: + idr_remove_uobj(&ib_uverbs_xrcd_idr, &obj->uobject); + +err_idr: + ib_dealloc_xrcd(xrcd); + +err: + put_uobj_write(&obj->uobject); + +err_tree_mutex_unlock: + if (f.file) + fdput(f); + + mutex_unlock(&file->device->xrcd_tree_mutex); + + return ret; +} + +ssize_t ib_uverbs_close_xrcd(struct ib_uverbs_file *file, + const char __user *buf, int in_len, + int out_len) +{ + struct ib_uverbs_close_xrcd cmd; + struct ib_uobject *uobj; + struct ib_xrcd *xrcd = NULL; + struct inode *inode = NULL; + struct ib_uxrcd_object *obj; + int live; + int ret = 0; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); + mutex_lock(&file->device->xrcd_tree_mutex); + uobj = idr_write_uobj(&ib_uverbs_xrcd_idr, cmd.xrcd_handle, file->ucontext); + if (!uobj) { + ret = -EINVAL; + goto out; + } - pd = idr_find(&ib_uverbs_pd_idr, cmd.pd_handle); - if (!pd || pd->uobject->context != file->ucontext) + xrcd = uobj->object; + inode = xrcd->inode; + obj = container_of(uobj, struct ib_uxrcd_object, uobject); + if (atomic_read(&obj->refcnt)) { + put_uobj_write(uobj); + ret = -EBUSY; goto out; + } - uobj = pd->uobject; + if (!inode || atomic_dec_and_test(&xrcd->usecnt)) { + ret = ib_dealloc_xrcd(uobj->object); + if (!ret) + uobj->live = 0; + } + + live = uobj->live; + if (inode && ret) + atomic_inc(&xrcd->usecnt); + + put_uobj_write(uobj); - ret = ib_dealloc_pd(pd); if (ret) goto out; - idr_remove(&ib_uverbs_pd_idr, cmd.pd_handle); + if (inode && !live) + xrcd_table_delete(file->device, inode); + idr_remove_uobj(&ib_uverbs_xrcd_idr, uobj); mutex_lock(&file->mutex); list_del(&uobj->list); mutex_unlock(&file->mutex); - kfree(uobj); + put_uobj(uobj); + ret = in_len; out: - mutex_unlock(&ib_uverbs_idr_mutex); + mutex_unlock(&file->device->xrcd_tree_mutex); + return ret; +} - return ret ? ret : in_len; +void ib_uverbs_dealloc_xrcd(struct ib_uverbs_device *dev, + struct ib_xrcd *xrcd) +{ + struct inode *inode; + + inode = xrcd->inode; + if (inode && !atomic_dec_and_test(&xrcd->usecnt)) + return; + + ib_dealloc_xrcd(xrcd); + + if (inode) + xrcd_table_delete(dev, inode); } ssize_t ib_uverbs_reg_mr(struct ib_uverbs_file *file, @@ -379,7 +912,7 @@ ssize_t ib_uverbs_reg_mr(struct ib_uverbs_file *file, struct ib_uverbs_reg_mr cmd; struct ib_uverbs_reg_mr_resp resp; struct ib_udata udata; - struct ib_umem_object *obj; + struct ib_uobject *uobj; struct ib_pd *pd; struct ib_mr *mr; int ret; @@ -397,107 +930,75 @@ ssize_t ib_uverbs_reg_mr(struct ib_uverbs_file *file, if ((cmd.start & ~PAGE_MASK) != (cmd.hca_va & ~PAGE_MASK)) return -EINVAL; - /* - * Local write permission is required if remote write or - * remote atomic permission is also requested. - */ - if (cmd.access_flags & (IB_ACCESS_REMOTE_ATOMIC | IB_ACCESS_REMOTE_WRITE) && - !(cmd.access_flags & IB_ACCESS_LOCAL_WRITE)) - return -EINVAL; - - obj = kmalloc(sizeof *obj, GFP_KERNEL); - if (!obj) - return -ENOMEM; - - obj->uobject.context = file->ucontext; - - /* - * We ask for writable memory if any access flags other than - * "remote read" are set. "Local write" and "remote write" - * obviously require write access. "Remote atomic" can do - * things like fetch and add, which will modify memory, and - * "MW bind" can change permissions by binding a window. - */ - ret = ib_umem_get(file->device->ib_dev, &obj->umem, - (void *) (unsigned long) cmd.start, cmd.length, - !!(cmd.access_flags & ~IB_ACCESS_REMOTE_READ)); + ret = ib_check_mr_access(cmd.access_flags); if (ret) - goto err_free; + return ret; - obj->umem.virt_base = cmd.hca_va; + uobj = kmalloc(sizeof *uobj, GFP_KERNEL); + if (!uobj) + return -ENOMEM; - mutex_lock(&ib_uverbs_idr_mutex); + init_uobj(uobj, 0, file->ucontext, &mr_lock_class); + down_write(&uobj->mutex); - pd = idr_find(&ib_uverbs_pd_idr, cmd.pd_handle); - if (!pd || pd->uobject->context != file->ucontext) { + pd = idr_read_pd(cmd.pd_handle, file->ucontext); + if (!pd) { ret = -EINVAL; - goto err_up; - } - - if (!pd->device->reg_user_mr) { - ret = -ENOSYS; - goto err_up; + goto err_free; } - mr = pd->device->reg_user_mr(pd, &obj->umem, cmd.access_flags, &udata); + mr = pd->device->reg_user_mr(pd, cmd.start, cmd.length, cmd.hca_va, + cmd.access_flags, &udata); if (IS_ERR(mr)) { ret = PTR_ERR(mr); - goto err_up; + goto err_put; } mr->device = pd->device; mr->pd = pd; - mr->uobject = &obj->uobject; + mr->uobject = uobj; atomic_inc(&pd->usecnt); atomic_set(&mr->usecnt, 0); - memset(&resp, 0, sizeof resp); - resp.lkey = mr->lkey; - resp.rkey = mr->rkey; - -retry: - if (!idr_pre_get(&ib_uverbs_mr_idr, GFP_KERNEL)) { - ret = -ENOMEM; - goto err_unreg; - } - - ret = idr_get_new(&ib_uverbs_mr_idr, mr, &obj->uobject.id); - - if (ret == -EAGAIN) - goto retry; + uobj->object = mr; + ret = idr_add_uobj(&ib_uverbs_mr_idr, uobj); if (ret) goto err_unreg; - resp.mr_handle = obj->uobject.id; + memset(&resp, 0, sizeof resp); + resp.lkey = mr->lkey; + resp.rkey = mr->rkey; + resp.mr_handle = uobj->id; if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) { ret = -EFAULT; - goto err_idr; + goto err_copy; } + put_pd_read(pd); + mutex_lock(&file->mutex); - list_add_tail(&obj->uobject.list, &file->ucontext->mr_list); + list_add_tail(&uobj->list, &file->ucontext->mr_list); mutex_unlock(&file->mutex); - mutex_unlock(&ib_uverbs_idr_mutex); + uobj->live = 1; + + up_write(&uobj->mutex); return in_len; -err_idr: - idr_remove(&ib_uverbs_mr_idr, obj->uobject.id); +err_copy: + idr_remove_uobj(&ib_uverbs_mr_idr, uobj); err_unreg: ib_dereg_mr(mr); - atomic_dec(&pd->usecnt); - -err_up: - mutex_unlock(&ib_uverbs_idr_mutex); - ib_umem_release(file->device->ib_dev, &obj->umem); +err_put: + put_pd_read(pd); err_free: - kfree(obj); + put_uobj_write(uobj); return ret; } @@ -507,37 +1008,156 @@ ssize_t ib_uverbs_dereg_mr(struct ib_uverbs_file *file, { struct ib_uverbs_dereg_mr cmd; struct ib_mr *mr; - struct ib_umem_object *memobj; + struct ib_uobject *uobj; int ret = -EINVAL; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); - - mr = idr_find(&ib_uverbs_mr_idr, cmd.mr_handle); - if (!mr || mr->uobject->context != file->ucontext) - goto out; + uobj = idr_write_uobj(&ib_uverbs_mr_idr, cmd.mr_handle, file->ucontext); + if (!uobj) + return -EINVAL; - memobj = container_of(mr->uobject, struct ib_umem_object, uobject); + mr = uobj->object; ret = ib_dereg_mr(mr); + if (!ret) + uobj->live = 0; + + put_uobj_write(uobj); + if (ret) - goto out; + return ret; - idr_remove(&ib_uverbs_mr_idr, cmd.mr_handle); + idr_remove_uobj(&ib_uverbs_mr_idr, uobj); mutex_lock(&file->mutex); - list_del(&memobj->uobject.list); + list_del(&uobj->list); mutex_unlock(&file->mutex); - ib_umem_release(file->device->ib_dev, &memobj->umem); - kfree(memobj); + put_uobj(uobj); -out: - mutex_unlock(&ib_uverbs_idr_mutex); + return in_len; +} - return ret ? ret : in_len; +ssize_t ib_uverbs_alloc_mw(struct ib_uverbs_file *file, + const char __user *buf, int in_len, + int out_len) +{ + struct ib_uverbs_alloc_mw cmd; + struct ib_uverbs_alloc_mw_resp resp; + struct ib_uobject *uobj; + struct ib_pd *pd; + struct ib_mw *mw; + int ret; + + if (out_len < sizeof(resp)) + return -ENOSPC; + + if (copy_from_user(&cmd, buf, sizeof(cmd))) + return -EFAULT; + + uobj = kmalloc(sizeof(*uobj), GFP_KERNEL); + if (!uobj) + return -ENOMEM; + + init_uobj(uobj, 0, file->ucontext, &mw_lock_class); + down_write(&uobj->mutex); + + pd = idr_read_pd(cmd.pd_handle, file->ucontext); + if (!pd) { + ret = -EINVAL; + goto err_free; + } + + mw = pd->device->alloc_mw(pd, cmd.mw_type); + if (IS_ERR(mw)) { + ret = PTR_ERR(mw); + goto err_put; + } + + mw->device = pd->device; + mw->pd = pd; + mw->uobject = uobj; + atomic_inc(&pd->usecnt); + + uobj->object = mw; + ret = idr_add_uobj(&ib_uverbs_mw_idr, uobj); + if (ret) + goto err_unalloc; + + memset(&resp, 0, sizeof(resp)); + resp.rkey = mw->rkey; + resp.mw_handle = uobj->id; + + if (copy_to_user((void __user *)(unsigned long)cmd.response, + &resp, sizeof(resp))) { + ret = -EFAULT; + goto err_copy; + } + + put_pd_read(pd); + + mutex_lock(&file->mutex); + list_add_tail(&uobj->list, &file->ucontext->mw_list); + mutex_unlock(&file->mutex); + + uobj->live = 1; + + up_write(&uobj->mutex); + + return in_len; + +err_copy: + idr_remove_uobj(&ib_uverbs_mw_idr, uobj); + +err_unalloc: + ib_dealloc_mw(mw); + +err_put: + put_pd_read(pd); + +err_free: + put_uobj_write(uobj); + return ret; +} + +ssize_t ib_uverbs_dealloc_mw(struct ib_uverbs_file *file, + const char __user *buf, int in_len, + int out_len) +{ + struct ib_uverbs_dealloc_mw cmd; + struct ib_mw *mw; + struct ib_uobject *uobj; + int ret = -EINVAL; + + if (copy_from_user(&cmd, buf, sizeof(cmd))) + return -EFAULT; + + uobj = idr_write_uobj(&ib_uverbs_mw_idr, cmd.mw_handle, file->ucontext); + if (!uobj) + return -EINVAL; + + mw = uobj->object; + + ret = ib_dealloc_mw(mw); + if (!ret) + uobj->live = 0; + + put_uobj_write(uobj); + + if (ret) + return ret; + + idr_remove_uobj(&ib_uverbs_mw_idr, uobj); + + mutex_lock(&file->mutex); + list_del(&uobj->list); + mutex_unlock(&file->mutex); + + put_uobj(uobj); + + return in_len; } ssize_t ib_uverbs_create_comp_channel(struct ib_uverbs_file *file, @@ -547,6 +1167,7 @@ ssize_t ib_uverbs_create_comp_channel(struct ib_uverbs_file *file, struct ib_uverbs_create_comp_channel cmd; struct ib_uverbs_create_comp_channel_resp resp; struct file *filp; + int ret; if (out_len < sizeof resp) return -ENOSPC; @@ -554,9 +1175,16 @@ ssize_t ib_uverbs_create_comp_channel(struct ib_uverbs_file *file, if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - filp = ib_uverbs_alloc_event_file(file, 0, &resp.fd); - if (IS_ERR(filp)) + ret = get_unused_fd_flags(O_CLOEXEC); + if (ret < 0) + return ret; + resp.fd = ret; + + filp = ib_uverbs_alloc_event_file(file, 0); + if (IS_ERR(filp)) { + put_unused_fd(resp.fd); return PTR_ERR(filp); + } if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) { @@ -576,7 +1204,7 @@ ssize_t ib_uverbs_create_cq(struct ib_uverbs_file *file, struct ib_uverbs_create_cq cmd; struct ib_uverbs_create_cq_resp resp; struct ib_udata udata; - struct ib_ucq_object *uobj; + struct ib_ucq_object *obj; struct ib_uverbs_event_file *ev_file = NULL; struct ib_cq *cq; int ret; @@ -594,10 +1222,13 @@ ssize_t ib_uverbs_create_cq(struct ib_uverbs_file *file, if (cmd.comp_vector >= file->device->num_comp_vectors) return -EINVAL; - uobj = kmalloc(sizeof *uobj, GFP_KERNEL); - if (!uobj) + obj = kmalloc(sizeof *obj, GFP_KERNEL); + if (!obj) return -ENOMEM; + init_uobj(&obj->uobject, cmd.user_handle, file->ucontext, &cq_lock_class); + down_write(&obj->uobject.mutex); + if (cmd.comp_channel >= 0) { ev_file = ib_uverbs_lookup_comp_file(cmd.comp_channel); if (!ev_file) { @@ -606,137 +1237,179 @@ ssize_t ib_uverbs_create_cq(struct ib_uverbs_file *file, } } - uobj->uobject.user_handle = cmd.user_handle; - uobj->uobject.context = file->ucontext; - uobj->uverbs_file = file; - uobj->comp_events_reported = 0; - uobj->async_events_reported = 0; - INIT_LIST_HEAD(&uobj->comp_list); - INIT_LIST_HEAD(&uobj->async_list); + obj->uverbs_file = file; + obj->comp_events_reported = 0; + obj->async_events_reported = 0; + INIT_LIST_HEAD(&obj->comp_list); + INIT_LIST_HEAD(&obj->async_list); cq = file->device->ib_dev->create_cq(file->device->ib_dev, cmd.cqe, + cmd.comp_vector, file->ucontext, &udata); if (IS_ERR(cq)) { ret = PTR_ERR(cq); - goto err; + goto err_file; } cq->device = file->device->ib_dev; - cq->uobject = &uobj->uobject; + cq->uobject = &obj->uobject; cq->comp_handler = ib_uverbs_comp_handler; cq->event_handler = ib_uverbs_cq_event_handler; cq->cq_context = ev_file; atomic_set(&cq->usecnt, 0); - mutex_lock(&ib_uverbs_idr_mutex); - -retry: - if (!idr_pre_get(&ib_uverbs_cq_idr, GFP_KERNEL)) { - ret = -ENOMEM; - goto err_up; - } - - ret = idr_get_new(&ib_uverbs_cq_idr, cq, &uobj->uobject.id); - - if (ret == -EAGAIN) - goto retry; + obj->uobject.object = cq; + ret = idr_add_uobj(&ib_uverbs_cq_idr, &obj->uobject); if (ret) - goto err_up; + goto err_free; memset(&resp, 0, sizeof resp); - resp.cq_handle = uobj->uobject.id; + resp.cq_handle = obj->uobject.id; resp.cqe = cq->cqe; if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) { ret = -EFAULT; - goto err_idr; + goto err_copy; } mutex_lock(&file->mutex); - list_add_tail(&uobj->uobject.list, &file->ucontext->cq_list); + list_add_tail(&obj->uobject.list, &file->ucontext->cq_list); mutex_unlock(&file->mutex); - mutex_unlock(&ib_uverbs_idr_mutex); + obj->uobject.live = 1; + + up_write(&obj->uobject.mutex); return in_len; -err_idr: - idr_remove(&ib_uverbs_cq_idr, uobj->uobject.id); +err_copy: + idr_remove_uobj(&ib_uverbs_cq_idr, &obj->uobject); -err_up: - mutex_unlock(&ib_uverbs_idr_mutex); +err_free: ib_destroy_cq(cq); -err: +err_file: if (ev_file) - ib_uverbs_release_ucq(file, ev_file, uobj); - kfree(uobj); + ib_uverbs_release_ucq(file, ev_file, obj); + +err: + put_uobj_write(&obj->uobject); return ret; } +ssize_t ib_uverbs_resize_cq(struct ib_uverbs_file *file, + const char __user *buf, int in_len, + int out_len) +{ + struct ib_uverbs_resize_cq cmd; + struct ib_uverbs_resize_cq_resp resp; + struct ib_udata udata; + struct ib_cq *cq; + int ret = -EINVAL; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + INIT_UDATA(&udata, buf + sizeof cmd, + (unsigned long) cmd.response + sizeof resp, + in_len - sizeof cmd, out_len - sizeof resp); + + cq = idr_read_cq(cmd.cq_handle, file->ucontext, 0); + if (!cq) + return -EINVAL; + + ret = cq->device->resize_cq(cq, cmd.cqe, &udata); + if (ret) + goto out; + + resp.cqe = cq->cqe; + + if (copy_to_user((void __user *) (unsigned long) cmd.response, + &resp, sizeof resp.cqe)) + ret = -EFAULT; + +out: + put_cq_read(cq); + + return ret ? ret : in_len; +} + +static int copy_wc_to_user(void __user *dest, struct ib_wc *wc) +{ + struct ib_uverbs_wc tmp; + + tmp.wr_id = wc->wr_id; + tmp.status = wc->status; + tmp.opcode = wc->opcode; + tmp.vendor_err = wc->vendor_err; + tmp.byte_len = wc->byte_len; + tmp.ex.imm_data = (__u32 __force) wc->ex.imm_data; + tmp.qp_num = wc->qp->qp_num; + tmp.src_qp = wc->src_qp; + tmp.wc_flags = wc->wc_flags; + tmp.pkey_index = wc->pkey_index; + tmp.slid = wc->slid; + tmp.sl = wc->sl; + tmp.dlid_path_bits = wc->dlid_path_bits; + tmp.port_num = wc->port_num; + tmp.reserved = 0; + + if (copy_to_user(dest, &tmp, sizeof tmp)) + return -EFAULT; + + return 0; +} + ssize_t ib_uverbs_poll_cq(struct ib_uverbs_file *file, const char __user *buf, int in_len, int out_len) { struct ib_uverbs_poll_cq cmd; - struct ib_uverbs_poll_cq_resp *resp; + struct ib_uverbs_poll_cq_resp resp; + u8 __user *header_ptr; + u8 __user *data_ptr; struct ib_cq *cq; - struct ib_wc *wc; - int ret = 0; - int i; - int rsize; + struct ib_wc wc; + int ret; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - wc = kmalloc(cmd.ne * sizeof *wc, GFP_KERNEL); - if (!wc) - return -ENOMEM; + cq = idr_read_cq(cmd.cq_handle, file->ucontext, 0); + if (!cq) + return -EINVAL; - rsize = sizeof *resp + cmd.ne * sizeof(struct ib_uverbs_wc); - resp = kmalloc(rsize, GFP_KERNEL); - if (!resp) { - ret = -ENOMEM; - goto out_wc; - } + /* we copy a struct ib_uverbs_poll_cq_resp to user space */ + header_ptr = (void __user *)(unsigned long) cmd.response; + data_ptr = header_ptr + sizeof resp; - mutex_lock(&ib_uverbs_idr_mutex); - cq = idr_find(&ib_uverbs_cq_idr, cmd.cq_handle); - if (!cq || cq->uobject->context != file->ucontext) { - ret = -EINVAL; - goto out; - } + memset(&resp, 0, sizeof resp); + while (resp.count < cmd.ne) { + ret = ib_poll_cq(cq, 1, &wc); + if (ret < 0) + goto out_put; + if (!ret) + break; - resp->count = ib_poll_cq(cq, cmd.ne, wc); + ret = copy_wc_to_user(data_ptr, &wc); + if (ret) + goto out_put; - for (i = 0; i < resp->count; i++) { - resp->wc[i].wr_id = wc[i].wr_id; - resp->wc[i].status = wc[i].status; - resp->wc[i].opcode = wc[i].opcode; - resp->wc[i].vendor_err = wc[i].vendor_err; - resp->wc[i].byte_len = wc[i].byte_len; - resp->wc[i].imm_data = (__u32 __force) wc[i].imm_data; - resp->wc[i].qp_num = wc[i].qp_num; - resp->wc[i].src_qp = wc[i].src_qp; - resp->wc[i].wc_flags = wc[i].wc_flags; - resp->wc[i].pkey_index = wc[i].pkey_index; - resp->wc[i].slid = wc[i].slid; - resp->wc[i].sl = wc[i].sl; - resp->wc[i].dlid_path_bits = wc[i].dlid_path_bits; - resp->wc[i].port_num = wc[i].port_num; + data_ptr += sizeof(struct ib_uverbs_wc); + ++resp.count; } - if (copy_to_user((void __user *) (unsigned long) cmd.response, resp, rsize)) + if (copy_to_user(header_ptr, &resp, sizeof resp)) { ret = -EFAULT; + goto out_put; + } -out: - mutex_unlock(&ib_uverbs_idr_mutex); - kfree(resp); + ret = in_len; -out_wc: - kfree(wc); - return ret ? ret : in_len; +out_put: + put_cq_read(cq); + return ret; } ssize_t ib_uverbs_req_notify_cq(struct ib_uverbs_file *file, @@ -745,21 +1418,20 @@ ssize_t ib_uverbs_req_notify_cq(struct ib_uverbs_file *file, { struct ib_uverbs_req_notify_cq cmd; struct ib_cq *cq; - int ret = -EINVAL; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); - cq = idr_find(&ib_uverbs_cq_idr, cmd.cq_handle); - if (cq && cq->uobject->context == file->ucontext) { - ib_req_notify_cq(cq, cmd.solicited_only ? - IB_CQ_SOLICITED : IB_CQ_NEXT_COMP); - ret = in_len; - } - mutex_unlock(&ib_uverbs_idr_mutex); + cq = idr_read_cq(cmd.cq_handle, file->ucontext, 0); + if (!cq) + return -EINVAL; - return ret; + ib_req_notify_cq(cq, cmd.solicited_only ? + IB_CQ_SOLICITED : IB_CQ_NEXT_COMP); + + put_cq_read(cq); + + return in_len; } ssize_t ib_uverbs_destroy_cq(struct ib_uverbs_file *file, @@ -768,52 +1440,50 @@ ssize_t ib_uverbs_destroy_cq(struct ib_uverbs_file *file, { struct ib_uverbs_destroy_cq cmd; struct ib_uverbs_destroy_cq_resp resp; + struct ib_uobject *uobj; struct ib_cq *cq; - struct ib_ucq_object *uobj; + struct ib_ucq_object *obj; struct ib_uverbs_event_file *ev_file; - u64 user_handle; int ret = -EINVAL; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - memset(&resp, 0, sizeof resp); - - mutex_lock(&ib_uverbs_idr_mutex); + uobj = idr_write_uobj(&ib_uverbs_cq_idr, cmd.cq_handle, file->ucontext); + if (!uobj) + return -EINVAL; + cq = uobj->object; + ev_file = cq->cq_context; + obj = container_of(cq->uobject, struct ib_ucq_object, uobject); - cq = idr_find(&ib_uverbs_cq_idr, cmd.cq_handle); - if (!cq || cq->uobject->context != file->ucontext) - goto out; + ret = ib_destroy_cq(cq); + if (!ret) + uobj->live = 0; - user_handle = cq->uobject->user_handle; - uobj = container_of(cq->uobject, struct ib_ucq_object, uobject); - ev_file = cq->cq_context; + put_uobj_write(uobj); - ret = ib_destroy_cq(cq); if (ret) - goto out; + return ret; - idr_remove(&ib_uverbs_cq_idr, cmd.cq_handle); + idr_remove_uobj(&ib_uverbs_cq_idr, uobj); mutex_lock(&file->mutex); - list_del(&uobj->uobject.list); + list_del(&uobj->list); mutex_unlock(&file->mutex); - ib_uverbs_release_ucq(file, ev_file, uobj); + ib_uverbs_release_ucq(file, ev_file, obj); - resp.comp_events_reported = uobj->comp_events_reported; - resp.async_events_reported = uobj->async_events_reported; + memset(&resp, 0, sizeof resp); + resp.comp_events_reported = obj->comp_events_reported; + resp.async_events_reported = obj->async_events_reported; - kfree(uobj); + put_uobj(uobj); if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) - ret = -EFAULT; - -out: - mutex_unlock(&ib_uverbs_idr_mutex); + return -EFAULT; - return ret ? ret : in_len; + return in_len; } ssize_t ib_uverbs_create_qp(struct ib_uverbs_file *file, @@ -823,10 +1493,13 @@ ssize_t ib_uverbs_create_qp(struct ib_uverbs_file *file, struct ib_uverbs_create_qp cmd; struct ib_uverbs_create_qp_resp resp; struct ib_udata udata; - struct ib_uqp_object *uobj; - struct ib_pd *pd; - struct ib_cq *scq, *rcq; - struct ib_srq *srq; + struct ib_uqp_object *obj; + struct ib_device *device; + struct ib_pd *pd = NULL; + struct ib_xrcd *xrcd = NULL; + struct ib_uobject *uninitialized_var(xrcd_uobj); + struct ib_cq *scq = NULL, *rcq = NULL; + struct ib_srq *srq = NULL; struct ib_qp *qp; struct ib_qp_init_attr attr; int ret; @@ -837,27 +1510,57 @@ ssize_t ib_uverbs_create_qp(struct ib_uverbs_file *file, if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; + if (cmd.qp_type == IB_QPT_RAW_PACKET && !capable(CAP_NET_RAW)) + return -EPERM; + INIT_UDATA(&udata, buf + sizeof cmd, (unsigned long) cmd.response + sizeof resp, in_len - sizeof cmd, out_len - sizeof resp); - uobj = kmalloc(sizeof *uobj, GFP_KERNEL); - if (!uobj) + obj = kzalloc(sizeof *obj, GFP_KERNEL); + if (!obj) return -ENOMEM; - mutex_lock(&ib_uverbs_idr_mutex); + init_uobj(&obj->uevent.uobject, cmd.user_handle, file->ucontext, &qp_lock_class); + down_write(&obj->uevent.uobject.mutex); - pd = idr_find(&ib_uverbs_pd_idr, cmd.pd_handle); - scq = idr_find(&ib_uverbs_cq_idr, cmd.send_cq_handle); - rcq = idr_find(&ib_uverbs_cq_idr, cmd.recv_cq_handle); - srq = cmd.is_srq ? idr_find(&ib_uverbs_srq_idr, cmd.srq_handle) : NULL; + if (cmd.qp_type == IB_QPT_XRC_TGT) { + xrcd = idr_read_xrcd(cmd.pd_handle, file->ucontext, &xrcd_uobj); + if (!xrcd) { + ret = -EINVAL; + goto err_put; + } + device = xrcd->device; + } else { + if (cmd.qp_type == IB_QPT_XRC_INI) { + cmd.max_recv_wr = cmd.max_recv_sge = 0; + } else { + if (cmd.is_srq) { + srq = idr_read_srq(cmd.srq_handle, file->ucontext); + if (!srq || srq->srq_type != IB_SRQT_BASIC) { + ret = -EINVAL; + goto err_put; + } + } - if (!pd || pd->uobject->context != file->ucontext || - !scq || scq->uobject->context != file->ucontext || - !rcq || rcq->uobject->context != file->ucontext || - (cmd.is_srq && (!srq || srq->uobject->context != file->ucontext))) { - ret = -EINVAL; - goto err_up; + if (cmd.recv_cq_handle != cmd.send_cq_handle) { + rcq = idr_read_cq(cmd.recv_cq_handle, file->ucontext, 0); + if (!rcq) { + ret = -EINVAL; + goto err_put; + } + } + } + + scq = idr_read_cq(cmd.send_cq_handle, file->ucontext, !!rcq); + rcq = rcq ?: scq; + pd = idr_read_pd(cmd.pd_handle, file->ucontext); + if (!pd || !scq) { + ret = -EINVAL; + goto err_put; + } + + device = pd->device; } attr.event_handler = ib_uverbs_qp_event_handler; @@ -865,8 +1568,10 @@ ssize_t ib_uverbs_create_qp(struct ib_uverbs_file *file, attr.send_cq = scq; attr.recv_cq = rcq; attr.srq = srq; + attr.xrcd = xrcd; attr.sq_sig_type = cmd.sq_sig_all ? IB_SIGNAL_ALL_WR : IB_SIGNAL_REQ_WR; attr.qp_type = cmd.qp_type; + attr.create_flags = 0; attr.cap.max_send_wr = cmd.max_send_wr; attr.cap.max_recv_wr = cmd.max_recv_wr; @@ -874,50 +1579,48 @@ ssize_t ib_uverbs_create_qp(struct ib_uverbs_file *file, attr.cap.max_recv_sge = cmd.max_recv_sge; attr.cap.max_inline_data = cmd.max_inline_data; - uobj->uevent.uobject.user_handle = cmd.user_handle; - uobj->uevent.uobject.context = file->ucontext; - uobj->uevent.events_reported = 0; - INIT_LIST_HEAD(&uobj->uevent.event_list); - INIT_LIST_HEAD(&uobj->mcast_list); + obj->uevent.events_reported = 0; + INIT_LIST_HEAD(&obj->uevent.event_list); + INIT_LIST_HEAD(&obj->mcast_list); + + if (cmd.qp_type == IB_QPT_XRC_TGT) + qp = ib_create_qp(pd, &attr); + else + qp = device->create_qp(pd, &attr, &udata); - qp = pd->device->create_qp(pd, &attr, &udata); if (IS_ERR(qp)) { ret = PTR_ERR(qp); - goto err_up; - } - - qp->device = pd->device; - qp->pd = pd; - qp->send_cq = attr.send_cq; - qp->recv_cq = attr.recv_cq; - qp->srq = attr.srq; - qp->uobject = &uobj->uevent.uobject; - qp->event_handler = attr.event_handler; - qp->qp_context = attr.qp_context; - qp->qp_type = attr.qp_type; - atomic_inc(&pd->usecnt); - atomic_inc(&attr.send_cq->usecnt); - atomic_inc(&attr.recv_cq->usecnt); - if (attr.srq) - atomic_inc(&attr.srq->usecnt); - - memset(&resp, 0, sizeof resp); - resp.qpn = qp->qp_num; - -retry: - if (!idr_pre_get(&ib_uverbs_qp_idr, GFP_KERNEL)) { - ret = -ENOMEM; - goto err_destroy; + goto err_put; } - ret = idr_get_new(&ib_uverbs_qp_idr, qp, &uobj->uevent.uobject.id); + if (cmd.qp_type != IB_QPT_XRC_TGT) { + qp->real_qp = qp; + qp->device = device; + qp->pd = pd; + qp->send_cq = attr.send_cq; + qp->recv_cq = attr.recv_cq; + qp->srq = attr.srq; + qp->event_handler = attr.event_handler; + qp->qp_context = attr.qp_context; + qp->qp_type = attr.qp_type; + atomic_set(&qp->usecnt, 0); + atomic_inc(&pd->usecnt); + atomic_inc(&attr.send_cq->usecnt); + if (attr.recv_cq) + atomic_inc(&attr.recv_cq->usecnt); + if (attr.srq) + atomic_inc(&attr.srq->usecnt); + } + qp->uobject = &obj->uevent.uobject; - if (ret == -EAGAIN) - goto retry; + obj->uevent.uobject.object = qp; + ret = idr_add_uobj(&ib_uverbs_qp_idr, &obj->uevent.uobject); if (ret) goto err_destroy; - resp.qp_handle = uobj->uevent.uobject.id; + memset(&resp, 0, sizeof resp); + resp.qpn = qp->qp_num; + resp.qp_handle = obj->uevent.uobject.id; resp.max_recv_sge = attr.cap.max_recv_sge; resp.max_send_sge = attr.cap.max_send_sge; resp.max_recv_wr = attr.cap.max_recv_wr; @@ -927,40 +1630,271 @@ retry: if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) { ret = -EFAULT; - goto err_idr; + goto err_copy; + } + + if (xrcd) { + obj->uxrcd = container_of(xrcd_uobj, struct ib_uxrcd_object, + uobject); + atomic_inc(&obj->uxrcd->refcnt); + put_xrcd_read(xrcd_uobj); } + if (pd) + put_pd_read(pd); + if (scq) + put_cq_read(scq); + if (rcq && rcq != scq) + put_cq_read(rcq); + if (srq) + put_srq_read(srq); + mutex_lock(&file->mutex); - list_add_tail(&uobj->uevent.uobject.list, &file->ucontext->qp_list); + list_add_tail(&obj->uevent.uobject.list, &file->ucontext->qp_list); mutex_unlock(&file->mutex); - mutex_unlock(&ib_uverbs_idr_mutex); + obj->uevent.uobject.live = 1; + + up_write(&obj->uevent.uobject.mutex); return in_len; -err_idr: - idr_remove(&ib_uverbs_qp_idr, uobj->uevent.uobject.id); +err_copy: + idr_remove_uobj(&ib_uverbs_qp_idr, &obj->uevent.uobject); err_destroy: ib_destroy_qp(qp); - atomic_dec(&pd->usecnt); - atomic_dec(&attr.send_cq->usecnt); - atomic_dec(&attr.recv_cq->usecnt); - if (attr.srq) - atomic_dec(&attr.srq->usecnt); -err_up: - mutex_unlock(&ib_uverbs_idr_mutex); +err_put: + if (xrcd) + put_xrcd_read(xrcd_uobj); + if (pd) + put_pd_read(pd); + if (scq) + put_cq_read(scq); + if (rcq && rcq != scq) + put_cq_read(rcq); + if (srq) + put_srq_read(srq); + + put_uobj_write(&obj->uevent.uobject); + return ret; +} - kfree(uobj); +ssize_t ib_uverbs_open_qp(struct ib_uverbs_file *file, + const char __user *buf, int in_len, int out_len) +{ + struct ib_uverbs_open_qp cmd; + struct ib_uverbs_create_qp_resp resp; + struct ib_udata udata; + struct ib_uqp_object *obj; + struct ib_xrcd *xrcd; + struct ib_uobject *uninitialized_var(xrcd_uobj); + struct ib_qp *qp; + struct ib_qp_open_attr attr; + int ret; + + if (out_len < sizeof resp) + return -ENOSPC; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + INIT_UDATA(&udata, buf + sizeof cmd, + (unsigned long) cmd.response + sizeof resp, + in_len - sizeof cmd, out_len - sizeof resp); + + obj = kmalloc(sizeof *obj, GFP_KERNEL); + if (!obj) + return -ENOMEM; + + init_uobj(&obj->uevent.uobject, cmd.user_handle, file->ucontext, &qp_lock_class); + down_write(&obj->uevent.uobject.mutex); + + xrcd = idr_read_xrcd(cmd.pd_handle, file->ucontext, &xrcd_uobj); + if (!xrcd) { + ret = -EINVAL; + goto err_put; + } + + attr.event_handler = ib_uverbs_qp_event_handler; + attr.qp_context = file; + attr.qp_num = cmd.qpn; + attr.qp_type = cmd.qp_type; + + obj->uevent.events_reported = 0; + INIT_LIST_HEAD(&obj->uevent.event_list); + INIT_LIST_HEAD(&obj->mcast_list); + + qp = ib_open_qp(xrcd, &attr); + if (IS_ERR(qp)) { + ret = PTR_ERR(qp); + goto err_put; + } + + qp->uobject = &obj->uevent.uobject; + + obj->uevent.uobject.object = qp; + ret = idr_add_uobj(&ib_uverbs_qp_idr, &obj->uevent.uobject); + if (ret) + goto err_destroy; + + memset(&resp, 0, sizeof resp); + resp.qpn = qp->qp_num; + resp.qp_handle = obj->uevent.uobject.id; + + if (copy_to_user((void __user *) (unsigned long) cmd.response, + &resp, sizeof resp)) { + ret = -EFAULT; + goto err_remove; + } + + obj->uxrcd = container_of(xrcd_uobj, struct ib_uxrcd_object, uobject); + atomic_inc(&obj->uxrcd->refcnt); + put_xrcd_read(xrcd_uobj); + + mutex_lock(&file->mutex); + list_add_tail(&obj->uevent.uobject.list, &file->ucontext->qp_list); + mutex_unlock(&file->mutex); + + obj->uevent.uobject.live = 1; + + up_write(&obj->uevent.uobject.mutex); + + return in_len; + +err_remove: + idr_remove_uobj(&ib_uverbs_qp_idr, &obj->uevent.uobject); + +err_destroy: + ib_destroy_qp(qp); + +err_put: + put_xrcd_read(xrcd_uobj); + put_uobj_write(&obj->uevent.uobject); return ret; } +ssize_t ib_uverbs_query_qp(struct ib_uverbs_file *file, + const char __user *buf, int in_len, + int out_len) +{ + struct ib_uverbs_query_qp cmd; + struct ib_uverbs_query_qp_resp resp; + struct ib_qp *qp; + struct ib_qp_attr *attr; + struct ib_qp_init_attr *init_attr; + int ret; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + attr = kmalloc(sizeof *attr, GFP_KERNEL); + init_attr = kmalloc(sizeof *init_attr, GFP_KERNEL); + if (!attr || !init_attr) { + ret = -ENOMEM; + goto out; + } + + qp = idr_read_qp(cmd.qp_handle, file->ucontext); + if (!qp) { + ret = -EINVAL; + goto out; + } + + ret = ib_query_qp(qp, attr, cmd.attr_mask, init_attr); + + put_qp_read(qp); + + if (ret) + goto out; + + memset(&resp, 0, sizeof resp); + + resp.qp_state = attr->qp_state; + resp.cur_qp_state = attr->cur_qp_state; + resp.path_mtu = attr->path_mtu; + resp.path_mig_state = attr->path_mig_state; + resp.qkey = attr->qkey; + resp.rq_psn = attr->rq_psn; + resp.sq_psn = attr->sq_psn; + resp.dest_qp_num = attr->dest_qp_num; + resp.qp_access_flags = attr->qp_access_flags; + resp.pkey_index = attr->pkey_index; + resp.alt_pkey_index = attr->alt_pkey_index; + resp.sq_draining = attr->sq_draining; + resp.max_rd_atomic = attr->max_rd_atomic; + resp.max_dest_rd_atomic = attr->max_dest_rd_atomic; + resp.min_rnr_timer = attr->min_rnr_timer; + resp.port_num = attr->port_num; + resp.timeout = attr->timeout; + resp.retry_cnt = attr->retry_cnt; + resp.rnr_retry = attr->rnr_retry; + resp.alt_port_num = attr->alt_port_num; + resp.alt_timeout = attr->alt_timeout; + + memcpy(resp.dest.dgid, attr->ah_attr.grh.dgid.raw, 16); + resp.dest.flow_label = attr->ah_attr.grh.flow_label; + resp.dest.sgid_index = attr->ah_attr.grh.sgid_index; + resp.dest.hop_limit = attr->ah_attr.grh.hop_limit; + resp.dest.traffic_class = attr->ah_attr.grh.traffic_class; + resp.dest.dlid = attr->ah_attr.dlid; + resp.dest.sl = attr->ah_attr.sl; + resp.dest.src_path_bits = attr->ah_attr.src_path_bits; + resp.dest.static_rate = attr->ah_attr.static_rate; + resp.dest.is_global = !!(attr->ah_attr.ah_flags & IB_AH_GRH); + resp.dest.port_num = attr->ah_attr.port_num; + + memcpy(resp.alt_dest.dgid, attr->alt_ah_attr.grh.dgid.raw, 16); + resp.alt_dest.flow_label = attr->alt_ah_attr.grh.flow_label; + resp.alt_dest.sgid_index = attr->alt_ah_attr.grh.sgid_index; + resp.alt_dest.hop_limit = attr->alt_ah_attr.grh.hop_limit; + resp.alt_dest.traffic_class = attr->alt_ah_attr.grh.traffic_class; + resp.alt_dest.dlid = attr->alt_ah_attr.dlid; + resp.alt_dest.sl = attr->alt_ah_attr.sl; + resp.alt_dest.src_path_bits = attr->alt_ah_attr.src_path_bits; + resp.alt_dest.static_rate = attr->alt_ah_attr.static_rate; + resp.alt_dest.is_global = !!(attr->alt_ah_attr.ah_flags & IB_AH_GRH); + resp.alt_dest.port_num = attr->alt_ah_attr.port_num; + + resp.max_send_wr = init_attr->cap.max_send_wr; + resp.max_recv_wr = init_attr->cap.max_recv_wr; + resp.max_send_sge = init_attr->cap.max_send_sge; + resp.max_recv_sge = init_attr->cap.max_recv_sge; + resp.max_inline_data = init_attr->cap.max_inline_data; + resp.sq_sig_all = init_attr->sq_sig_type == IB_SIGNAL_ALL_WR; + + if (copy_to_user((void __user *) (unsigned long) cmd.response, + &resp, sizeof resp)) + ret = -EFAULT; + +out: + kfree(attr); + kfree(init_attr); + + return ret ? ret : in_len; +} + +/* Remove ignored fields set in the attribute mask */ +static int modify_qp_mask(enum ib_qp_type qp_type, int mask) +{ + switch (qp_type) { + case IB_QPT_XRC_INI: + return mask & ~(IB_QP_MAX_DEST_RD_ATOMIC | IB_QP_MIN_RNR_TIMER); + case IB_QPT_XRC_TGT: + return mask & ~(IB_QP_MAX_QP_RD_ATOMIC | IB_QP_RETRY_CNT | + IB_QP_RNR_RETRY); + default: + return mask; + } +} + ssize_t ib_uverbs_modify_qp(struct ib_uverbs_file *file, const char __user *buf, int in_len, int out_len) { struct ib_uverbs_modify_qp cmd; + struct ib_udata udata; struct ib_qp *qp; struct ib_qp_attr *attr; int ret; @@ -968,14 +1902,15 @@ ssize_t ib_uverbs_modify_qp(struct ib_uverbs_file *file, if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; + INIT_UDATA(&udata, buf + sizeof cmd, NULL, in_len - sizeof cmd, + out_len); + attr = kmalloc(sizeof *attr, GFP_KERNEL); if (!attr) return -ENOMEM; - mutex_lock(&ib_uverbs_idr_mutex); - - qp = idr_find(&ib_uverbs_qp_idr, cmd.qp_handle); - if (!qp || qp->uobject->context != file->ucontext) { + qp = idr_read_qp(cmd.qp_handle, file->ucontext); + if (!qp) { ret = -EINVAL; goto out; } @@ -990,7 +1925,7 @@ ssize_t ib_uverbs_modify_qp(struct ib_uverbs_file *file, attr->dest_qp_num = cmd.dest_qp_num; attr->qp_access_flags = cmd.qp_access_flags; attr->pkey_index = cmd.pkey_index; - attr->alt_pkey_index = cmd.pkey_index; + attr->alt_pkey_index = cmd.alt_pkey_index; attr->en_sqd_async_notify = cmd.en_sqd_async_notify; attr->max_rd_atomic = cmd.max_rd_atomic; attr->max_dest_rd_atomic = cmd.max_dest_rd_atomic; @@ -1026,14 +1961,24 @@ ssize_t ib_uverbs_modify_qp(struct ib_uverbs_file *file, attr->alt_ah_attr.ah_flags = cmd.alt_dest.is_global ? IB_AH_GRH : 0; attr->alt_ah_attr.port_num = cmd.alt_dest.port_num; - ret = ib_modify_qp(qp, attr, cmd.attr_mask); + if (qp->real_qp == qp) { + ret = ib_resolve_eth_l2_attrs(qp, attr, &cmd.attr_mask); + if (ret) + goto out; + ret = qp->device->modify_qp(qp, attr, + modify_qp_mask(qp->qp_type, cmd.attr_mask), &udata); + } else { + ret = ib_modify_qp(qp, attr, modify_qp_mask(qp->qp_type, cmd.attr_mask)); + } + + put_qp_read(qp); + if (ret) goto out; ret = in_len; out: - mutex_unlock(&ib_uverbs_idr_mutex); kfree(attr); return ret; @@ -1045,8 +1990,9 @@ ssize_t ib_uverbs_destroy_qp(struct ib_uverbs_file *file, { struct ib_uverbs_destroy_qp cmd; struct ib_uverbs_destroy_qp_resp resp; + struct ib_uobject *uobj; struct ib_qp *qp; - struct ib_uqp_object *uobj; + struct ib_uqp_object *obj; int ret = -EINVAL; if (copy_from_user(&cmd, buf, sizeof cmd)) @@ -1054,48 +2000,51 @@ ssize_t ib_uverbs_destroy_qp(struct ib_uverbs_file *file, memset(&resp, 0, sizeof resp); - mutex_lock(&ib_uverbs_idr_mutex); - - qp = idr_find(&ib_uverbs_qp_idr, cmd.qp_handle); - if (!qp || qp->uobject->context != file->ucontext) - goto out; - - uobj = container_of(qp->uobject, struct ib_uqp_object, uevent.uobject); + uobj = idr_write_uobj(&ib_uverbs_qp_idr, cmd.qp_handle, file->ucontext); + if (!uobj) + return -EINVAL; + qp = uobj->object; + obj = container_of(uobj, struct ib_uqp_object, uevent.uobject); - if (!list_empty(&uobj->mcast_list)) { - ret = -EBUSY; - goto out; + if (!list_empty(&obj->mcast_list)) { + put_uobj_write(uobj); + return -EBUSY; } ret = ib_destroy_qp(qp); + if (!ret) + uobj->live = 0; + + put_uobj_write(uobj); + if (ret) - goto out; + return ret; + + if (obj->uxrcd) + atomic_dec(&obj->uxrcd->refcnt); - idr_remove(&ib_uverbs_qp_idr, cmd.qp_handle); + idr_remove_uobj(&ib_uverbs_qp_idr, uobj); mutex_lock(&file->mutex); - list_del(&uobj->uevent.uobject.list); + list_del(&uobj->list); mutex_unlock(&file->mutex); - ib_uverbs_release_uevent(file, &uobj->uevent); + ib_uverbs_release_uevent(file, &obj->uevent); - resp.events_reported = uobj->uevent.events_reported; + resp.events_reported = obj->uevent.events_reported; - kfree(uobj); + put_uobj(uobj); if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) - ret = -EFAULT; - -out: - mutex_unlock(&ib_uverbs_idr_mutex); + return -EFAULT; - return ret ? ret : in_len; + return in_len; } ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, - const char __user *buf, int in_len, - int out_len) + const char __user *buf, int in_len, + int out_len) { struct ib_uverbs_post_send cmd; struct ib_uverbs_post_send_resp resp; @@ -1103,6 +2052,7 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, struct ib_send_wr *wr = NULL, *last, *next, *bad_wr; struct ib_qp *qp; int i, sg_ind; + int is_ud; ssize_t ret = -EINVAL; if (copy_from_user(&cmd, buf, sizeof cmd)) @@ -1119,12 +2069,11 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, if (!user_wr) return -ENOMEM; - mutex_lock(&ib_uverbs_idr_mutex); - - qp = idr_find(&ib_uverbs_qp_idr, cmd.qp_handle); - if (!qp || qp->uobject->context != file->ucontext) + qp = idr_read_qp(cmd.qp_handle, file->ucontext); + if (!qp) goto out; + is_ud = qp->qp_type == IB_QPT_UD; sg_ind = 0; last = NULL; for (i = 0; i < cmd.wr_count; ++i) { @@ -1132,12 +2081,12 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, buf + sizeof cmd + i * cmd.wqe_size, cmd.wqe_size)) { ret = -EFAULT; - goto out; + goto out_put; } if (user_wr->num_sge + sg_ind > cmd.sge_count) { ret = -EINVAL; - goto out; + goto out_put; } next = kmalloc(ALIGN(sizeof *next, sizeof (struct ib_sge)) + @@ -1145,7 +2094,7 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, GFP_KERNEL); if (!next) { ret = -ENOMEM; - goto out; + goto out_put; } if (!last) @@ -1159,27 +2108,39 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, next->num_sge = user_wr->num_sge; next->opcode = user_wr->opcode; next->send_flags = user_wr->send_flags; - next->imm_data = (__be32 __force) user_wr->imm_data; - if (qp->qp_type == IB_QPT_UD) { - next->wr.ud.ah = idr_find(&ib_uverbs_ah_idr, - user_wr->wr.ud.ah); + if (is_ud) { + next->wr.ud.ah = idr_read_ah(user_wr->wr.ud.ah, + file->ucontext); if (!next->wr.ud.ah) { ret = -EINVAL; - goto out; + goto out_put; } next->wr.ud.remote_qpn = user_wr->wr.ud.remote_qpn; next->wr.ud.remote_qkey = user_wr->wr.ud.remote_qkey; + if (next->opcode == IB_WR_SEND_WITH_IMM) + next->ex.imm_data = + (__be32 __force) user_wr->ex.imm_data; } else { switch (next->opcode) { - case IB_WR_RDMA_WRITE: case IB_WR_RDMA_WRITE_WITH_IMM: + next->ex.imm_data = + (__be32 __force) user_wr->ex.imm_data; + case IB_WR_RDMA_WRITE: case IB_WR_RDMA_READ: next->wr.rdma.remote_addr = user_wr->wr.rdma.remote_addr; next->wr.rdma.rkey = user_wr->wr.rdma.rkey; break; + case IB_WR_SEND_WITH_IMM: + next->ex.imm_data = + (__be32 __force) user_wr->ex.imm_data; + break; + case IB_WR_SEND_WITH_INV: + next->ex.invalidate_rkey = + user_wr->ex.invalidate_rkey; + break; case IB_WR_ATOMIC_CMP_AND_SWP: case IB_WR_ATOMIC_FETCH_AND_ADD: next->wr.atomic.remote_addr = @@ -1203,7 +2164,7 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, sg_ind * sizeof (struct ib_sge), next->num_sge * sizeof (struct ib_sge))) { ret = -EFAULT; - goto out; + goto out_put; } sg_ind += next->num_sge; } else @@ -1211,7 +2172,7 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, } resp.bad_wr = 0; - ret = qp->device->post_send(qp, wr, &bad_wr); + ret = qp->device->post_send(qp->real_qp, wr, &bad_wr); if (ret) for (next = wr; next; next = next->next) { ++resp.bad_wr; @@ -1223,15 +2184,18 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file, &resp, sizeof resp)) ret = -EFAULT; -out: - mutex_unlock(&ib_uverbs_idr_mutex); +out_put: + put_qp_read(qp); while (wr) { + if (is_ud && wr->wr.ud.ah) + put_ah_read(wr->wr.ud.ah); next = wr->next; kfree(wr); wr = next; } +out: kfree(user_wr); return ret ? ret : in_len; @@ -1323,8 +2287,8 @@ err: } ssize_t ib_uverbs_post_recv(struct ib_uverbs_file *file, - const char __user *buf, int in_len, - int out_len) + const char __user *buf, int in_len, + int out_len) { struct ib_uverbs_post_recv cmd; struct ib_uverbs_post_recv_resp resp; @@ -1341,14 +2305,15 @@ ssize_t ib_uverbs_post_recv(struct ib_uverbs_file *file, if (IS_ERR(wr)) return PTR_ERR(wr); - mutex_lock(&ib_uverbs_idr_mutex); - - qp = idr_find(&ib_uverbs_qp_idr, cmd.qp_handle); - if (!qp || qp->uobject->context != file->ucontext) + qp = idr_read_qp(cmd.qp_handle, file->ucontext); + if (!qp) goto out; resp.bad_wr = 0; - ret = qp->device->post_recv(qp, wr, &bad_wr); + ret = qp->device->post_recv(qp->real_qp, wr, &bad_wr); + + put_qp_read(qp); + if (ret) for (next = wr; next; next = next->next) { ++resp.bad_wr; @@ -1356,14 +2321,11 @@ ssize_t ib_uverbs_post_recv(struct ib_uverbs_file *file, break; } - if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) ret = -EFAULT; out: - mutex_unlock(&ib_uverbs_idr_mutex); - while (wr) { next = wr->next; kfree(wr); @@ -1374,8 +2336,8 @@ out: } ssize_t ib_uverbs_post_srq_recv(struct ib_uverbs_file *file, - const char __user *buf, int in_len, - int out_len) + const char __user *buf, int in_len, + int out_len) { struct ib_uverbs_post_srq_recv cmd; struct ib_uverbs_post_srq_recv_resp resp; @@ -1392,14 +2354,15 @@ ssize_t ib_uverbs_post_srq_recv(struct ib_uverbs_file *file, if (IS_ERR(wr)) return PTR_ERR(wr); - mutex_lock(&ib_uverbs_idr_mutex); - - srq = idr_find(&ib_uverbs_srq_idr, cmd.srq_handle); - if (!srq || srq->uobject->context != file->ucontext) + srq = idr_read_srq(cmd.srq_handle, file->ucontext); + if (!srq) goto out; resp.bad_wr = 0; ret = srq->device->post_srq_recv(srq, wr, &bad_wr); + + put_srq_read(srq); + if (ret) for (next = wr; next; next = next->next) { ++resp.bad_wr; @@ -1407,14 +2370,11 @@ ssize_t ib_uverbs_post_srq_recv(struct ib_uverbs_file *file, break; } - if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) ret = -EFAULT; out: - mutex_unlock(&ib_uverbs_idr_mutex); - while (wr) { next = wr->next; kfree(wr); @@ -1446,17 +2406,15 @@ ssize_t ib_uverbs_create_ah(struct ib_uverbs_file *file, if (!uobj) return -ENOMEM; - mutex_lock(&ib_uverbs_idr_mutex); + init_uobj(uobj, cmd.user_handle, file->ucontext, &ah_lock_class); + down_write(&uobj->mutex); - pd = idr_find(&ib_uverbs_pd_idr, cmd.pd_handle); - if (!pd || pd->uobject->context != file->ucontext) { + pd = idr_read_pd(cmd.pd_handle, file->ucontext); + if (!pd) { ret = -EINVAL; - goto err_up; + goto err; } - uobj->user_handle = cmd.user_handle; - uobj->context = file->ucontext; - attr.dlid = cmd.attr.dlid; attr.sl = cmd.attr.sl; attr.src_path_bits = cmd.attr.src_path_bits; @@ -1472,21 +2430,13 @@ ssize_t ib_uverbs_create_ah(struct ib_uverbs_file *file, ah = ib_create_ah(pd, &attr); if (IS_ERR(ah)) { ret = PTR_ERR(ah); - goto err_up; - } - - ah->uobject = uobj; - -retry: - if (!idr_pre_get(&ib_uverbs_ah_idr, GFP_KERNEL)) { - ret = -ENOMEM; - goto err_destroy; + goto err_put; } - ret = idr_get_new(&ib_uverbs_ah_idr, ah, &uobj->id); + ah->uobject = uobj; + uobj->object = ah; - if (ret == -EAGAIN) - goto retry; + ret = idr_add_uobj(&ib_uverbs_ah_idr, uobj); if (ret) goto err_destroy; @@ -1495,27 +2445,32 @@ retry: if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) { ret = -EFAULT; - goto err_idr; + goto err_copy; } + put_pd_read(pd); + mutex_lock(&file->mutex); list_add_tail(&uobj->list, &file->ucontext->ah_list); mutex_unlock(&file->mutex); - mutex_unlock(&ib_uverbs_idr_mutex); + uobj->live = 1; + + up_write(&uobj->mutex); return in_len; -err_idr: - idr_remove(&ib_uverbs_ah_idr, uobj->id); +err_copy: + idr_remove_uobj(&ib_uverbs_ah_idr, uobj); err_destroy: ib_destroy_ah(ah); -err_up: - mutex_unlock(&ib_uverbs_idr_mutex); +err_put: + put_pd_read(pd); - kfree(uobj); +err: + put_uobj_write(uobj); return ret; } @@ -1525,35 +2480,34 @@ ssize_t ib_uverbs_destroy_ah(struct ib_uverbs_file *file, struct ib_uverbs_destroy_ah cmd; struct ib_ah *ah; struct ib_uobject *uobj; - int ret = -EINVAL; + int ret; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); + uobj = idr_write_uobj(&ib_uverbs_ah_idr, cmd.ah_handle, file->ucontext); + if (!uobj) + return -EINVAL; + ah = uobj->object; - ah = idr_find(&ib_uverbs_ah_idr, cmd.ah_handle); - if (!ah || ah->uobject->context != file->ucontext) - goto out; + ret = ib_destroy_ah(ah); + if (!ret) + uobj->live = 0; - uobj = ah->uobject; + put_uobj_write(uobj); - ret = ib_destroy_ah(ah); if (ret) - goto out; + return ret; - idr_remove(&ib_uverbs_ah_idr, cmd.ah_handle); + idr_remove_uobj(&ib_uverbs_ah_idr, uobj); mutex_lock(&file->mutex); list_del(&uobj->list); mutex_unlock(&file->mutex); - kfree(uobj); - -out: - mutex_unlock(&ib_uverbs_idr_mutex); + put_uobj(uobj); - return ret ? ret : in_len; + return in_len; } ssize_t ib_uverbs_attach_mcast(struct ib_uverbs_file *file, @@ -1562,47 +2516,43 @@ ssize_t ib_uverbs_attach_mcast(struct ib_uverbs_file *file, { struct ib_uverbs_attach_mcast cmd; struct ib_qp *qp; - struct ib_uqp_object *uobj; + struct ib_uqp_object *obj; struct ib_uverbs_mcast_entry *mcast; - int ret = -EINVAL; + int ret; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); - - qp = idr_find(&ib_uverbs_qp_idr, cmd.qp_handle); - if (!qp || qp->uobject->context != file->ucontext) - goto out; + qp = idr_write_qp(cmd.qp_handle, file->ucontext); + if (!qp) + return -EINVAL; - uobj = container_of(qp->uobject, struct ib_uqp_object, uevent.uobject); + obj = container_of(qp->uobject, struct ib_uqp_object, uevent.uobject); - list_for_each_entry(mcast, &uobj->mcast_list, list) + list_for_each_entry(mcast, &obj->mcast_list, list) if (cmd.mlid == mcast->lid && !memcmp(cmd.gid, mcast->gid.raw, sizeof mcast->gid.raw)) { ret = 0; - goto out; + goto out_put; } mcast = kmalloc(sizeof *mcast, GFP_KERNEL); if (!mcast) { ret = -ENOMEM; - goto out; + goto out_put; } mcast->lid = cmd.mlid; memcpy(mcast->gid.raw, cmd.gid, sizeof mcast->gid.raw); ret = ib_attach_mcast(qp, &mcast->gid, cmd.mlid); - if (!ret) { - uobj = container_of(qp->uobject, struct ib_uqp_object, - uevent.uobject); - list_add_tail(&mcast->list, &uobj->mcast_list); - } else + if (!ret) + list_add_tail(&mcast->list, &obj->mcast_list); + else kfree(mcast); -out: - mutex_unlock(&ib_uverbs_idr_mutex); +out_put: + put_qp_write(qp); return ret ? ret : in_len; } @@ -1612,7 +2562,7 @@ ssize_t ib_uverbs_detach_mcast(struct ib_uverbs_file *file, int out_len) { struct ib_uverbs_detach_mcast cmd; - struct ib_uqp_object *uobj; + struct ib_uqp_object *obj; struct ib_qp *qp; struct ib_uverbs_mcast_entry *mcast; int ret = -EINVAL; @@ -1620,19 +2570,17 @@ ssize_t ib_uverbs_detach_mcast(struct ib_uverbs_file *file, if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); - - qp = idr_find(&ib_uverbs_qp_idr, cmd.qp_handle); - if (!qp || qp->uobject->context != file->ucontext) - goto out; + qp = idr_write_qp(cmd.qp_handle, file->ucontext); + if (!qp) + return -EINVAL; ret = ib_detach_mcast(qp, (union ib_gid *) cmd.gid, cmd.mlid); if (ret) - goto out; + goto out_put; - uobj = container_of(qp->uobject, struct ib_uqp_object, uevent.uobject); + obj = container_of(qp->uobject, struct ib_uqp_object, uevent.uobject); - list_for_each_entry(mcast, &uobj->mcast_list, list) + list_for_each_entry(mcast, &obj->mcast_list, list) if (cmd.mlid == mcast->lid && !memcmp(cmd.gid, mcast->gid.raw, sizeof mcast->gid.raw)) { list_del(&mcast->list); @@ -1640,123 +2588,456 @@ ssize_t ib_uverbs_detach_mcast(struct ib_uverbs_file *file, break; } -out: - mutex_unlock(&ib_uverbs_idr_mutex); +out_put: + put_qp_write(qp); return ret ? ret : in_len; } -ssize_t ib_uverbs_create_srq(struct ib_uverbs_file *file, - const char __user *buf, int in_len, - int out_len) +static int kern_spec_to_ib_spec(struct ib_uverbs_flow_spec *kern_spec, + union ib_flow_spec *ib_spec) +{ + if (kern_spec->reserved) + return -EINVAL; + + ib_spec->type = kern_spec->type; + + switch (ib_spec->type) { + case IB_FLOW_SPEC_ETH: + ib_spec->eth.size = sizeof(struct ib_flow_spec_eth); + if (ib_spec->eth.size != kern_spec->eth.size) + return -EINVAL; + memcpy(&ib_spec->eth.val, &kern_spec->eth.val, + sizeof(struct ib_flow_eth_filter)); + memcpy(&ib_spec->eth.mask, &kern_spec->eth.mask, + sizeof(struct ib_flow_eth_filter)); + break; + case IB_FLOW_SPEC_IPV4: + ib_spec->ipv4.size = sizeof(struct ib_flow_spec_ipv4); + if (ib_spec->ipv4.size != kern_spec->ipv4.size) + return -EINVAL; + memcpy(&ib_spec->ipv4.val, &kern_spec->ipv4.val, + sizeof(struct ib_flow_ipv4_filter)); + memcpy(&ib_spec->ipv4.mask, &kern_spec->ipv4.mask, + sizeof(struct ib_flow_ipv4_filter)); + break; + case IB_FLOW_SPEC_TCP: + case IB_FLOW_SPEC_UDP: + ib_spec->tcp_udp.size = sizeof(struct ib_flow_spec_tcp_udp); + if (ib_spec->tcp_udp.size != kern_spec->tcp_udp.size) + return -EINVAL; + memcpy(&ib_spec->tcp_udp.val, &kern_spec->tcp_udp.val, + sizeof(struct ib_flow_tcp_udp_filter)); + memcpy(&ib_spec->tcp_udp.mask, &kern_spec->tcp_udp.mask, + sizeof(struct ib_flow_tcp_udp_filter)); + break; + default: + return -EINVAL; + } + return 0; +} + +int ib_uverbs_ex_create_flow(struct ib_uverbs_file *file, + struct ib_udata *ucore, + struct ib_udata *uhw) +{ + struct ib_uverbs_create_flow cmd; + struct ib_uverbs_create_flow_resp resp; + struct ib_uobject *uobj; + struct ib_flow *flow_id; + struct ib_uverbs_flow_attr *kern_flow_attr; + struct ib_flow_attr *flow_attr; + struct ib_qp *qp; + int err = 0; + void *kern_spec; + void *ib_spec; + int i; + + if (ucore->inlen < sizeof(cmd)) + return -EINVAL; + + if (ucore->outlen < sizeof(resp)) + return -ENOSPC; + + err = ib_copy_from_udata(&cmd, ucore, sizeof(cmd)); + if (err) + return err; + + ucore->inbuf += sizeof(cmd); + ucore->inlen -= sizeof(cmd); + + if (cmd.comp_mask) + return -EINVAL; + + if ((cmd.flow_attr.type == IB_FLOW_ATTR_SNIFFER && + !capable(CAP_NET_ADMIN)) || !capable(CAP_NET_RAW)) + return -EPERM; + + if (cmd.flow_attr.num_of_specs > IB_FLOW_SPEC_SUPPORT_LAYERS) + return -EINVAL; + + if (cmd.flow_attr.size > ucore->inlen || + cmd.flow_attr.size > + (cmd.flow_attr.num_of_specs * sizeof(struct ib_uverbs_flow_spec))) + return -EINVAL; + + if (cmd.flow_attr.reserved[0] || + cmd.flow_attr.reserved[1]) + return -EINVAL; + + if (cmd.flow_attr.num_of_specs) { + kern_flow_attr = kmalloc(sizeof(*kern_flow_attr) + cmd.flow_attr.size, + GFP_KERNEL); + if (!kern_flow_attr) + return -ENOMEM; + + memcpy(kern_flow_attr, &cmd.flow_attr, sizeof(*kern_flow_attr)); + err = ib_copy_from_udata(kern_flow_attr + 1, ucore, + cmd.flow_attr.size); + if (err) + goto err_free_attr; + } else { + kern_flow_attr = &cmd.flow_attr; + } + + uobj = kmalloc(sizeof(*uobj), GFP_KERNEL); + if (!uobj) { + err = -ENOMEM; + goto err_free_attr; + } + init_uobj(uobj, 0, file->ucontext, &rule_lock_class); + down_write(&uobj->mutex); + + qp = idr_read_qp(cmd.qp_handle, file->ucontext); + if (!qp) { + err = -EINVAL; + goto err_uobj; + } + + flow_attr = kmalloc(sizeof(*flow_attr) + cmd.flow_attr.size, GFP_KERNEL); + if (!flow_attr) { + err = -ENOMEM; + goto err_put; + } + + flow_attr->type = kern_flow_attr->type; + flow_attr->priority = kern_flow_attr->priority; + flow_attr->num_of_specs = kern_flow_attr->num_of_specs; + flow_attr->port = kern_flow_attr->port; + flow_attr->flags = kern_flow_attr->flags; + flow_attr->size = sizeof(*flow_attr); + + kern_spec = kern_flow_attr + 1; + ib_spec = flow_attr + 1; + for (i = 0; i < flow_attr->num_of_specs && + cmd.flow_attr.size > offsetof(struct ib_uverbs_flow_spec, reserved) && + cmd.flow_attr.size >= + ((struct ib_uverbs_flow_spec *)kern_spec)->size; i++) { + err = kern_spec_to_ib_spec(kern_spec, ib_spec); + if (err) + goto err_free; + flow_attr->size += + ((union ib_flow_spec *) ib_spec)->size; + cmd.flow_attr.size -= ((struct ib_uverbs_flow_spec *)kern_spec)->size; + kern_spec += ((struct ib_uverbs_flow_spec *) kern_spec)->size; + ib_spec += ((union ib_flow_spec *) ib_spec)->size; + } + if (cmd.flow_attr.size || (i != flow_attr->num_of_specs)) { + pr_warn("create flow failed, flow %d: %d bytes left from uverb cmd\n", + i, cmd.flow_attr.size); + err = -EINVAL; + goto err_free; + } + flow_id = ib_create_flow(qp, flow_attr, IB_FLOW_DOMAIN_USER); + if (IS_ERR(flow_id)) { + err = PTR_ERR(flow_id); + goto err_free; + } + flow_id->qp = qp; + flow_id->uobject = uobj; + uobj->object = flow_id; + + err = idr_add_uobj(&ib_uverbs_rule_idr, uobj); + if (err) + goto destroy_flow; + + memset(&resp, 0, sizeof(resp)); + resp.flow_handle = uobj->id; + + err = ib_copy_to_udata(ucore, + &resp, sizeof(resp)); + if (err) + goto err_copy; + + put_qp_read(qp); + mutex_lock(&file->mutex); + list_add_tail(&uobj->list, &file->ucontext->rule_list); + mutex_unlock(&file->mutex); + + uobj->live = 1; + + up_write(&uobj->mutex); + kfree(flow_attr); + if (cmd.flow_attr.num_of_specs) + kfree(kern_flow_attr); + return 0; +err_copy: + idr_remove_uobj(&ib_uverbs_rule_idr, uobj); +destroy_flow: + ib_destroy_flow(flow_id); +err_free: + kfree(flow_attr); +err_put: + put_qp_read(qp); +err_uobj: + put_uobj_write(uobj); +err_free_attr: + if (cmd.flow_attr.num_of_specs) + kfree(kern_flow_attr); + return err; +} + +int ib_uverbs_ex_destroy_flow(struct ib_uverbs_file *file, + struct ib_udata *ucore, + struct ib_udata *uhw) +{ + struct ib_uverbs_destroy_flow cmd; + struct ib_flow *flow_id; + struct ib_uobject *uobj; + int ret; + + if (ucore->inlen < sizeof(cmd)) + return -EINVAL; + + ret = ib_copy_from_udata(&cmd, ucore, sizeof(cmd)); + if (ret) + return ret; + + if (cmd.comp_mask) + return -EINVAL; + + uobj = idr_write_uobj(&ib_uverbs_rule_idr, cmd.flow_handle, + file->ucontext); + if (!uobj) + return -EINVAL; + flow_id = uobj->object; + + ret = ib_destroy_flow(flow_id); + if (!ret) + uobj->live = 0; + + put_uobj_write(uobj); + + idr_remove_uobj(&ib_uverbs_rule_idr, uobj); + + mutex_lock(&file->mutex); + list_del(&uobj->list); + mutex_unlock(&file->mutex); + + put_uobj(uobj); + + return ret; +} + +static int __uverbs_create_xsrq(struct ib_uverbs_file *file, + struct ib_uverbs_create_xsrq *cmd, + struct ib_udata *udata) { - struct ib_uverbs_create_srq cmd; struct ib_uverbs_create_srq_resp resp; - struct ib_udata udata; - struct ib_uevent_object *uobj; + struct ib_usrq_object *obj; struct ib_pd *pd; struct ib_srq *srq; + struct ib_uobject *uninitialized_var(xrcd_uobj); struct ib_srq_init_attr attr; int ret; - if (out_len < sizeof resp) - return -ENOSPC; - - if (copy_from_user(&cmd, buf, sizeof cmd)) - return -EFAULT; + obj = kmalloc(sizeof *obj, GFP_KERNEL); + if (!obj) + return -ENOMEM; - INIT_UDATA(&udata, buf + sizeof cmd, - (unsigned long) cmd.response + sizeof resp, - in_len - sizeof cmd, out_len - sizeof resp); + init_uobj(&obj->uevent.uobject, cmd->user_handle, file->ucontext, &srq_lock_class); + down_write(&obj->uevent.uobject.mutex); - uobj = kmalloc(sizeof *uobj, GFP_KERNEL); - if (!uobj) - return -ENOMEM; + if (cmd->srq_type == IB_SRQT_XRC) { + attr.ext.xrc.xrcd = idr_read_xrcd(cmd->xrcd_handle, file->ucontext, &xrcd_uobj); + if (!attr.ext.xrc.xrcd) { + ret = -EINVAL; + goto err; + } - mutex_lock(&ib_uverbs_idr_mutex); + obj->uxrcd = container_of(xrcd_uobj, struct ib_uxrcd_object, uobject); + atomic_inc(&obj->uxrcd->refcnt); - pd = idr_find(&ib_uverbs_pd_idr, cmd.pd_handle); + attr.ext.xrc.cq = idr_read_cq(cmd->cq_handle, file->ucontext, 0); + if (!attr.ext.xrc.cq) { + ret = -EINVAL; + goto err_put_xrcd; + } + } - if (!pd || pd->uobject->context != file->ucontext) { + pd = idr_read_pd(cmd->pd_handle, file->ucontext); + if (!pd) { ret = -EINVAL; - goto err_up; + goto err_put_cq; } attr.event_handler = ib_uverbs_srq_event_handler; attr.srq_context = file; - attr.attr.max_wr = cmd.max_wr; - attr.attr.max_sge = cmd.max_sge; - attr.attr.srq_limit = cmd.srq_limit; + attr.srq_type = cmd->srq_type; + attr.attr.max_wr = cmd->max_wr; + attr.attr.max_sge = cmd->max_sge; + attr.attr.srq_limit = cmd->srq_limit; - uobj->uobject.user_handle = cmd.user_handle; - uobj->uobject.context = file->ucontext; - uobj->events_reported = 0; - INIT_LIST_HEAD(&uobj->event_list); + obj->uevent.events_reported = 0; + INIT_LIST_HEAD(&obj->uevent.event_list); - srq = pd->device->create_srq(pd, &attr, &udata); + srq = pd->device->create_srq(pd, &attr, udata); if (IS_ERR(srq)) { ret = PTR_ERR(srq); - goto err_up; + goto err_put; } - srq->device = pd->device; - srq->pd = pd; - srq->uobject = &uobj->uobject; + srq->device = pd->device; + srq->pd = pd; + srq->srq_type = cmd->srq_type; + srq->uobject = &obj->uevent.uobject; srq->event_handler = attr.event_handler; srq->srq_context = attr.srq_context; - atomic_inc(&pd->usecnt); - atomic_set(&srq->usecnt, 0); - - memset(&resp, 0, sizeof resp); -retry: - if (!idr_pre_get(&ib_uverbs_srq_idr, GFP_KERNEL)) { - ret = -ENOMEM; - goto err_destroy; + if (cmd->srq_type == IB_SRQT_XRC) { + srq->ext.xrc.cq = attr.ext.xrc.cq; + srq->ext.xrc.xrcd = attr.ext.xrc.xrcd; + atomic_inc(&attr.ext.xrc.cq->usecnt); + atomic_inc(&attr.ext.xrc.xrcd->usecnt); } - ret = idr_get_new(&ib_uverbs_srq_idr, srq, &uobj->uobject.id); + atomic_inc(&pd->usecnt); + atomic_set(&srq->usecnt, 0); - if (ret == -EAGAIN) - goto retry; + obj->uevent.uobject.object = srq; + ret = idr_add_uobj(&ib_uverbs_srq_idr, &obj->uevent.uobject); if (ret) goto err_destroy; - resp.srq_handle = uobj->uobject.id; + memset(&resp, 0, sizeof resp); + resp.srq_handle = obj->uevent.uobject.id; + resp.max_wr = attr.attr.max_wr; + resp.max_sge = attr.attr.max_sge; + if (cmd->srq_type == IB_SRQT_XRC) + resp.srqn = srq->ext.xrc.srq_num; - if (copy_to_user((void __user *) (unsigned long) cmd.response, + if (copy_to_user((void __user *) (unsigned long) cmd->response, &resp, sizeof resp)) { ret = -EFAULT; - goto err_idr; + goto err_copy; } + if (cmd->srq_type == IB_SRQT_XRC) { + put_uobj_read(xrcd_uobj); + put_cq_read(attr.ext.xrc.cq); + } + put_pd_read(pd); + mutex_lock(&file->mutex); - list_add_tail(&uobj->uobject.list, &file->ucontext->srq_list); + list_add_tail(&obj->uevent.uobject.list, &file->ucontext->srq_list); mutex_unlock(&file->mutex); - mutex_unlock(&ib_uverbs_idr_mutex); + obj->uevent.uobject.live = 1; - return in_len; + up_write(&obj->uevent.uobject.mutex); -err_idr: - idr_remove(&ib_uverbs_srq_idr, uobj->uobject.id); + return 0; + +err_copy: + idr_remove_uobj(&ib_uverbs_srq_idr, &obj->uevent.uobject); err_destroy: ib_destroy_srq(srq); - atomic_dec(&pd->usecnt); -err_up: - mutex_unlock(&ib_uverbs_idr_mutex); +err_put: + put_pd_read(pd); + +err_put_cq: + if (cmd->srq_type == IB_SRQT_XRC) + put_cq_read(attr.ext.xrc.cq); + +err_put_xrcd: + if (cmd->srq_type == IB_SRQT_XRC) { + atomic_dec(&obj->uxrcd->refcnt); + put_uobj_read(xrcd_uobj); + } - kfree(uobj); +err: + put_uobj_write(&obj->uevent.uobject); return ret; } +ssize_t ib_uverbs_create_srq(struct ib_uverbs_file *file, + const char __user *buf, int in_len, + int out_len) +{ + struct ib_uverbs_create_srq cmd; + struct ib_uverbs_create_xsrq xcmd; + struct ib_uverbs_create_srq_resp resp; + struct ib_udata udata; + int ret; + + if (out_len < sizeof resp) + return -ENOSPC; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + xcmd.response = cmd.response; + xcmd.user_handle = cmd.user_handle; + xcmd.srq_type = IB_SRQT_BASIC; + xcmd.pd_handle = cmd.pd_handle; + xcmd.max_wr = cmd.max_wr; + xcmd.max_sge = cmd.max_sge; + xcmd.srq_limit = cmd.srq_limit; + + INIT_UDATA(&udata, buf + sizeof cmd, + (unsigned long) cmd.response + sizeof resp, + in_len - sizeof cmd, out_len - sizeof resp); + + ret = __uverbs_create_xsrq(file, &xcmd, &udata); + if (ret) + return ret; + + return in_len; +} + +ssize_t ib_uverbs_create_xsrq(struct ib_uverbs_file *file, + const char __user *buf, int in_len, int out_len) +{ + struct ib_uverbs_create_xsrq cmd; + struct ib_uverbs_create_srq_resp resp; + struct ib_udata udata; + int ret; + + if (out_len < sizeof resp) + return -ENOSPC; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + INIT_UDATA(&udata, buf + sizeof cmd, + (unsigned long) cmd.response + sizeof resp, + in_len - sizeof cmd, out_len - sizeof resp); + + ret = __uverbs_create_xsrq(file, &cmd, &udata); + if (ret) + return ret; + + return in_len; +} + ssize_t ib_uverbs_modify_srq(struct ib_uverbs_file *file, const char __user *buf, int in_len, int out_len) { struct ib_uverbs_modify_srq cmd; + struct ib_udata udata; struct ib_srq *srq; struct ib_srq_attr attr; int ret; @@ -1764,70 +3045,116 @@ ssize_t ib_uverbs_modify_srq(struct ib_uverbs_file *file, if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); + INIT_UDATA(&udata, buf + sizeof cmd, NULL, in_len - sizeof cmd, + out_len); - srq = idr_find(&ib_uverbs_srq_idr, cmd.srq_handle); - if (!srq || srq->uobject->context != file->ucontext) { - ret = -EINVAL; - goto out; - } + srq = idr_read_srq(cmd.srq_handle, file->ucontext); + if (!srq) + return -EINVAL; attr.max_wr = cmd.max_wr; attr.srq_limit = cmd.srq_limit; - ret = ib_modify_srq(srq, &attr, cmd.attr_mask); + ret = srq->device->modify_srq(srq, &attr, cmd.attr_mask, &udata); -out: - mutex_unlock(&ib_uverbs_idr_mutex); + put_srq_read(srq); return ret ? ret : in_len; } +ssize_t ib_uverbs_query_srq(struct ib_uverbs_file *file, + const char __user *buf, + int in_len, int out_len) +{ + struct ib_uverbs_query_srq cmd; + struct ib_uverbs_query_srq_resp resp; + struct ib_srq_attr attr; + struct ib_srq *srq; + int ret; + + if (out_len < sizeof resp) + return -ENOSPC; + + if (copy_from_user(&cmd, buf, sizeof cmd)) + return -EFAULT; + + srq = idr_read_srq(cmd.srq_handle, file->ucontext); + if (!srq) + return -EINVAL; + + ret = ib_query_srq(srq, &attr); + + put_srq_read(srq); + + if (ret) + return ret; + + memset(&resp, 0, sizeof resp); + + resp.max_wr = attr.max_wr; + resp.max_sge = attr.max_sge; + resp.srq_limit = attr.srq_limit; + + if (copy_to_user((void __user *) (unsigned long) cmd.response, + &resp, sizeof resp)) + return -EFAULT; + + return in_len; +} + ssize_t ib_uverbs_destroy_srq(struct ib_uverbs_file *file, const char __user *buf, int in_len, int out_len) { struct ib_uverbs_destroy_srq cmd; struct ib_uverbs_destroy_srq_resp resp; + struct ib_uobject *uobj; struct ib_srq *srq; - struct ib_uevent_object *uobj; + struct ib_uevent_object *obj; int ret = -EINVAL; + struct ib_usrq_object *us; + enum ib_srq_type srq_type; if (copy_from_user(&cmd, buf, sizeof cmd)) return -EFAULT; - mutex_lock(&ib_uverbs_idr_mutex); - - memset(&resp, 0, sizeof resp); + uobj = idr_write_uobj(&ib_uverbs_srq_idr, cmd.srq_handle, file->ucontext); + if (!uobj) + return -EINVAL; + srq = uobj->object; + obj = container_of(uobj, struct ib_uevent_object, uobject); + srq_type = srq->srq_type; - srq = idr_find(&ib_uverbs_srq_idr, cmd.srq_handle); - if (!srq || srq->uobject->context != file->ucontext) - goto out; + ret = ib_destroy_srq(srq); + if (!ret) + uobj->live = 0; - uobj = container_of(srq->uobject, struct ib_uevent_object, uobject); + put_uobj_write(uobj); - ret = ib_destroy_srq(srq); if (ret) - goto out; + return ret; - idr_remove(&ib_uverbs_srq_idr, cmd.srq_handle); + if (srq_type == IB_SRQT_XRC) { + us = container_of(obj, struct ib_usrq_object, uevent); + atomic_dec(&us->uxrcd->refcnt); + } + + idr_remove_uobj(&ib_uverbs_srq_idr, uobj); mutex_lock(&file->mutex); - list_del(&uobj->uobject.list); + list_del(&uobj->list); mutex_unlock(&file->mutex); - ib_uverbs_release_uevent(file, uobj); + ib_uverbs_release_uevent(file, obj); - resp.events_reported = uobj->events_reported; + memset(&resp, 0, sizeof resp); + resp.events_reported = obj->events_reported; - kfree(uobj); + put_uobj(uobj); if (copy_to_user((void __user *) (unsigned long) cmd.response, &resp, sizeof resp)) ret = -EFAULT; -out: - mutex_unlock(&ib_uverbs_idr_mutex); - return ret ? ret : in_len; } diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c index 96ea79b63df..08219fb3338 100644 --- a/drivers/infiniband/core/uverbs_main.c +++ b/drivers/infiniband/core/uverbs_main.c @@ -1,6 +1,6 @@ /* * Copyright (c) 2005 Topspin Communications. All rights reserved. - * Copyright (c) 2005 Cisco Systems. All rights reserved. + * Copyright (c) 2005, 2006 Cisco Systems. All rights reserved. * Copyright (c) 2005 Mellanox Technologies. All rights reserved. * Copyright (c) 2005 Voltaire, Inc. All rights reserved. * Copyright (c) 2005 PathScale, Inc. All rights reserved. @@ -32,8 +32,6 @@ * 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. - * - * $Id: uverbs_main.c 2733 2005-06-28 19:14:34Z roland $ */ #include <linux/module.h> @@ -42,9 +40,11 @@ #include <linux/err.h> #include <linux/fs.h> #include <linux/poll.h> +#include <linux/sched.h> #include <linux/file.h> -#include <linux/mount.h> #include <linux/cdev.h> +#include <linux/anon_inodes.h> +#include <linux/slab.h> #include <asm/uaccess.h> @@ -54,8 +54,6 @@ MODULE_AUTHOR("Roland Dreier"); MODULE_DESCRIPTION("InfiniBand userspace verbs access"); MODULE_LICENSE("Dual BSD/GPL"); -#define INFINIBANDEVENTFS_MAGIC 0x49426576 /* "IBev" */ - enum { IB_UVERBS_MAJOR = 231, IB_UVERBS_BASE_MINOR = 192, @@ -66,7 +64,7 @@ enum { static struct class *uverbs_class; -DEFINE_MUTEX(ib_uverbs_idr_mutex); +DEFINE_SPINLOCK(ib_uverbs_idr_lock); DEFINE_IDR(ib_uverbs_pd_idr); DEFINE_IDR(ib_uverbs_mr_idr); DEFINE_IDR(ib_uverbs_mw_idr); @@ -74,42 +72,57 @@ DEFINE_IDR(ib_uverbs_ah_idr); DEFINE_IDR(ib_uverbs_cq_idr); DEFINE_IDR(ib_uverbs_qp_idr); DEFINE_IDR(ib_uverbs_srq_idr); +DEFINE_IDR(ib_uverbs_xrcd_idr); +DEFINE_IDR(ib_uverbs_rule_idr); -static spinlock_t map_lock; -static struct ib_uverbs_device *dev_table[IB_UVERBS_MAX_DEVICES]; +static DEFINE_SPINLOCK(map_lock); static DECLARE_BITMAP(dev_map, IB_UVERBS_MAX_DEVICES); static ssize_t (*uverbs_cmd_table[])(struct ib_uverbs_file *file, const char __user *buf, int in_len, int out_len) = { - [IB_USER_VERBS_CMD_GET_CONTEXT] = ib_uverbs_get_context, - [IB_USER_VERBS_CMD_QUERY_DEVICE] = ib_uverbs_query_device, - [IB_USER_VERBS_CMD_QUERY_PORT] = ib_uverbs_query_port, - [IB_USER_VERBS_CMD_ALLOC_PD] = ib_uverbs_alloc_pd, - [IB_USER_VERBS_CMD_DEALLOC_PD] = ib_uverbs_dealloc_pd, - [IB_USER_VERBS_CMD_REG_MR] = ib_uverbs_reg_mr, - [IB_USER_VERBS_CMD_DEREG_MR] = ib_uverbs_dereg_mr, + [IB_USER_VERBS_CMD_GET_CONTEXT] = ib_uverbs_get_context, + [IB_USER_VERBS_CMD_QUERY_DEVICE] = ib_uverbs_query_device, + [IB_USER_VERBS_CMD_QUERY_PORT] = ib_uverbs_query_port, + [IB_USER_VERBS_CMD_ALLOC_PD] = ib_uverbs_alloc_pd, + [IB_USER_VERBS_CMD_DEALLOC_PD] = ib_uverbs_dealloc_pd, + [IB_USER_VERBS_CMD_REG_MR] = ib_uverbs_reg_mr, + [IB_USER_VERBS_CMD_DEREG_MR] = ib_uverbs_dereg_mr, + [IB_USER_VERBS_CMD_ALLOC_MW] = ib_uverbs_alloc_mw, + [IB_USER_VERBS_CMD_DEALLOC_MW] = ib_uverbs_dealloc_mw, [IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL] = ib_uverbs_create_comp_channel, - [IB_USER_VERBS_CMD_CREATE_CQ] = ib_uverbs_create_cq, - [IB_USER_VERBS_CMD_POLL_CQ] = ib_uverbs_poll_cq, - [IB_USER_VERBS_CMD_REQ_NOTIFY_CQ] = ib_uverbs_req_notify_cq, - [IB_USER_VERBS_CMD_DESTROY_CQ] = ib_uverbs_destroy_cq, - [IB_USER_VERBS_CMD_CREATE_QP] = ib_uverbs_create_qp, - [IB_USER_VERBS_CMD_MODIFY_QP] = ib_uverbs_modify_qp, - [IB_USER_VERBS_CMD_DESTROY_QP] = ib_uverbs_destroy_qp, - [IB_USER_VERBS_CMD_POST_SEND] = ib_uverbs_post_send, - [IB_USER_VERBS_CMD_POST_RECV] = ib_uverbs_post_recv, - [IB_USER_VERBS_CMD_POST_SRQ_RECV] = ib_uverbs_post_srq_recv, - [IB_USER_VERBS_CMD_CREATE_AH] = ib_uverbs_create_ah, - [IB_USER_VERBS_CMD_DESTROY_AH] = ib_uverbs_destroy_ah, - [IB_USER_VERBS_CMD_ATTACH_MCAST] = ib_uverbs_attach_mcast, - [IB_USER_VERBS_CMD_DETACH_MCAST] = ib_uverbs_detach_mcast, - [IB_USER_VERBS_CMD_CREATE_SRQ] = ib_uverbs_create_srq, - [IB_USER_VERBS_CMD_MODIFY_SRQ] = ib_uverbs_modify_srq, - [IB_USER_VERBS_CMD_DESTROY_SRQ] = ib_uverbs_destroy_srq, + [IB_USER_VERBS_CMD_CREATE_CQ] = ib_uverbs_create_cq, + [IB_USER_VERBS_CMD_RESIZE_CQ] = ib_uverbs_resize_cq, + [IB_USER_VERBS_CMD_POLL_CQ] = ib_uverbs_poll_cq, + [IB_USER_VERBS_CMD_REQ_NOTIFY_CQ] = ib_uverbs_req_notify_cq, + [IB_USER_VERBS_CMD_DESTROY_CQ] = ib_uverbs_destroy_cq, + [IB_USER_VERBS_CMD_CREATE_QP] = ib_uverbs_create_qp, + [IB_USER_VERBS_CMD_QUERY_QP] = ib_uverbs_query_qp, + [IB_USER_VERBS_CMD_MODIFY_QP] = ib_uverbs_modify_qp, + [IB_USER_VERBS_CMD_DESTROY_QP] = ib_uverbs_destroy_qp, + [IB_USER_VERBS_CMD_POST_SEND] = ib_uverbs_post_send, + [IB_USER_VERBS_CMD_POST_RECV] = ib_uverbs_post_recv, + [IB_USER_VERBS_CMD_POST_SRQ_RECV] = ib_uverbs_post_srq_recv, + [IB_USER_VERBS_CMD_CREATE_AH] = ib_uverbs_create_ah, + [IB_USER_VERBS_CMD_DESTROY_AH] = ib_uverbs_destroy_ah, + [IB_USER_VERBS_CMD_ATTACH_MCAST] = ib_uverbs_attach_mcast, + [IB_USER_VERBS_CMD_DETACH_MCAST] = ib_uverbs_detach_mcast, + [IB_USER_VERBS_CMD_CREATE_SRQ] = ib_uverbs_create_srq, + [IB_USER_VERBS_CMD_MODIFY_SRQ] = ib_uverbs_modify_srq, + [IB_USER_VERBS_CMD_QUERY_SRQ] = ib_uverbs_query_srq, + [IB_USER_VERBS_CMD_DESTROY_SRQ] = ib_uverbs_destroy_srq, + [IB_USER_VERBS_CMD_OPEN_XRCD] = ib_uverbs_open_xrcd, + [IB_USER_VERBS_CMD_CLOSE_XRCD] = ib_uverbs_close_xrcd, + [IB_USER_VERBS_CMD_CREATE_XSRQ] = ib_uverbs_create_xsrq, + [IB_USER_VERBS_CMD_OPEN_QP] = ib_uverbs_open_qp, }; -static struct vfsmount *uverbs_event_mnt; +static int (*uverbs_ex_cmd_table[])(struct ib_uverbs_file *file, + struct ib_udata *ucore, + struct ib_udata *uhw) = { + [IB_USER_VERBS_EX_CMD_CREATE_FLOW] = ib_uverbs_ex_create_flow, + [IB_USER_VERBS_EX_CMD_DESTROY_FLOW] = ib_uverbs_ex_destroy_flow +}; static void ib_uverbs_add_one(struct ib_device *device); static void ib_uverbs_remove_one(struct ib_device *device); @@ -119,7 +132,15 @@ static void ib_uverbs_release_dev(struct kref *ref) struct ib_uverbs_device *dev = container_of(ref, struct ib_uverbs_device, ref); - kfree(dev); + complete(&dev->comp); +} + +static void ib_uverbs_release_event_file(struct kref *ref) +{ + struct ib_uverbs_event_file *file = + container_of(ref, struct ib_uverbs_event_file, ref); + + kfree(file); } void ib_uverbs_release_ucq(struct ib_uverbs_file *file, @@ -180,78 +201,100 @@ static int ib_uverbs_cleanup_ucontext(struct ib_uverbs_file *file, if (!context) return 0; - mutex_lock(&ib_uverbs_idr_mutex); + context->closing = 1; list_for_each_entry_safe(uobj, tmp, &context->ah_list, list) { - struct ib_ah *ah = idr_find(&ib_uverbs_ah_idr, uobj->id); - idr_remove(&ib_uverbs_ah_idr, uobj->id); + struct ib_ah *ah = uobj->object; + + idr_remove_uobj(&ib_uverbs_ah_idr, uobj); ib_destroy_ah(ah); - list_del(&uobj->list); + kfree(uobj); + } + + /* Remove MWs before QPs, in order to support type 2A MWs. */ + list_for_each_entry_safe(uobj, tmp, &context->mw_list, list) { + struct ib_mw *mw = uobj->object; + + idr_remove_uobj(&ib_uverbs_mw_idr, uobj); + ib_dealloc_mw(mw); + kfree(uobj); + } + + list_for_each_entry_safe(uobj, tmp, &context->rule_list, list) { + struct ib_flow *flow_id = uobj->object; + + idr_remove_uobj(&ib_uverbs_rule_idr, uobj); + ib_destroy_flow(flow_id); kfree(uobj); } list_for_each_entry_safe(uobj, tmp, &context->qp_list, list) { - struct ib_qp *qp = idr_find(&ib_uverbs_qp_idr, uobj->id); + struct ib_qp *qp = uobj->object; struct ib_uqp_object *uqp = container_of(uobj, struct ib_uqp_object, uevent.uobject); - idr_remove(&ib_uverbs_qp_idr, uobj->id); - ib_uverbs_detach_umcast(qp, uqp); - ib_destroy_qp(qp); - list_del(&uobj->list); + + idr_remove_uobj(&ib_uverbs_qp_idr, uobj); + if (qp != qp->real_qp) { + ib_close_qp(qp); + } else { + ib_uverbs_detach_umcast(qp, uqp); + ib_destroy_qp(qp); + } ib_uverbs_release_uevent(file, &uqp->uevent); kfree(uqp); } list_for_each_entry_safe(uobj, tmp, &context->cq_list, list) { - struct ib_cq *cq = idr_find(&ib_uverbs_cq_idr, uobj->id); + struct ib_cq *cq = uobj->object; struct ib_uverbs_event_file *ev_file = cq->cq_context; struct ib_ucq_object *ucq = container_of(uobj, struct ib_ucq_object, uobject); - idr_remove(&ib_uverbs_cq_idr, uobj->id); + + idr_remove_uobj(&ib_uverbs_cq_idr, uobj); ib_destroy_cq(cq); - list_del(&uobj->list); ib_uverbs_release_ucq(file, ev_file, ucq); kfree(ucq); } list_for_each_entry_safe(uobj, tmp, &context->srq_list, list) { - struct ib_srq *srq = idr_find(&ib_uverbs_srq_idr, uobj->id); + struct ib_srq *srq = uobj->object; struct ib_uevent_object *uevent = container_of(uobj, struct ib_uevent_object, uobject); - idr_remove(&ib_uverbs_srq_idr, uobj->id); + + idr_remove_uobj(&ib_uverbs_srq_idr, uobj); ib_destroy_srq(srq); - list_del(&uobj->list); ib_uverbs_release_uevent(file, uevent); kfree(uevent); } - /* XXX Free MWs */ - list_for_each_entry_safe(uobj, tmp, &context->mr_list, list) { - struct ib_mr *mr = idr_find(&ib_uverbs_mr_idr, uobj->id); - struct ib_device *mrdev = mr->device; - struct ib_umem_object *memobj; + struct ib_mr *mr = uobj->object; - idr_remove(&ib_uverbs_mr_idr, uobj->id); + idr_remove_uobj(&ib_uverbs_mr_idr, uobj); ib_dereg_mr(mr); + kfree(uobj); + } - memobj = container_of(uobj, struct ib_umem_object, uobject); - ib_umem_release_on_close(mrdev, &memobj->umem); + mutex_lock(&file->device->xrcd_tree_mutex); + list_for_each_entry_safe(uobj, tmp, &context->xrcd_list, list) { + struct ib_xrcd *xrcd = uobj->object; + struct ib_uxrcd_object *uxrcd = + container_of(uobj, struct ib_uxrcd_object, uobject); - list_del(&uobj->list); - kfree(memobj); + idr_remove_uobj(&ib_uverbs_xrcd_idr, uobj); + ib_uverbs_dealloc_xrcd(file->device, xrcd); + kfree(uxrcd); } + mutex_unlock(&file->device->xrcd_tree_mutex); list_for_each_entry_safe(uobj, tmp, &context->pd_list, list) { - struct ib_pd *pd = idr_find(&ib_uverbs_pd_idr, uobj->id); - idr_remove(&ib_uverbs_pd_idr, uobj->id); + struct ib_pd *pd = uobj->object; + + idr_remove_uobj(&ib_uverbs_pd_idr, uobj); ib_dealloc_pd(pd); - list_del(&uobj->list); kfree(uobj); } - mutex_unlock(&ib_uverbs_idr_mutex); - return context->device->dealloc_ucontext(context); } @@ -337,14 +380,6 @@ static unsigned int ib_uverbs_event_poll(struct file *filp, return pollflags; } -void ib_uverbs_release_event_file(struct kref *ref) -{ - struct ib_uverbs_event_file *file = - container_of(ref, struct ib_uverbs_event_file, ref); - - kfree(file); -} - static int ib_uverbs_event_fasync(int fd, struct file *filp, int on) { struct ib_uverbs_event_file *file = filp->private_data; @@ -358,7 +393,7 @@ static int ib_uverbs_event_close(struct inode *inode, struct file *filp) struct ib_uverbs_event *entry, *tmp; spin_lock_irq(&file->lock); - file->file = NULL; + file->is_closed = 1; list_for_each_entry_safe(entry, tmp, &file->event_list, list) { if (entry->counter) list_del(&entry->obj_list); @@ -366,8 +401,6 @@ static int ib_uverbs_event_close(struct inode *inode, struct file *filp) } spin_unlock_irq(&file->lock); - ib_uverbs_event_fasync(-1, filp, 0); - if (file->is_async) { ib_unregister_event_handler(&file->uverbs_file->event_handler); kref_put(&file->uverbs_file->ref, ib_uverbs_release_file); @@ -377,12 +410,13 @@ static int ib_uverbs_event_close(struct inode *inode, struct file *filp) return 0; } -static struct file_operations uverbs_event_fops = { +static const struct file_operations uverbs_event_fops = { .owner = THIS_MODULE, - .read = ib_uverbs_event_read, + .read = ib_uverbs_event_read, .poll = ib_uverbs_event_poll, .release = ib_uverbs_event_close, - .fasync = ib_uverbs_event_fasync + .fasync = ib_uverbs_event_fasync, + .llseek = no_llseek, }; void ib_uverbs_comp_handler(struct ib_cq *cq, void *cq_context) @@ -396,7 +430,7 @@ void ib_uverbs_comp_handler(struct ib_cq *cq, void *cq_context) return; spin_lock_irqsave(&file->lock, flags); - if (!file->file) { + if (file->is_closed) { spin_unlock_irqrestore(&file->lock, flags); return; } @@ -429,7 +463,7 @@ static void ib_uverbs_async_handler(struct ib_uverbs_file *file, unsigned long flags; spin_lock_irqsave(&file->async_file->lock, flags); - if (!file->async_file->file) { + if (file->async_file->is_closed) { spin_unlock_irqrestore(&file->async_file->lock, flags); return; } @@ -461,7 +495,6 @@ void ib_uverbs_cq_event_handler(struct ib_event *event, void *context_ptr) ib_uverbs_async_handler(uobj->uverbs_file, uobj->uobject.user_handle, event->event, &uobj->async_list, &uobj->async_events_reported); - } void ib_uverbs_qp_event_handler(struct ib_event *event, void *context_ptr) @@ -499,11 +532,10 @@ void ib_uverbs_event_handler(struct ib_event_handler *handler, } struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file, - int is_async, int *fd) + int is_async) { struct ib_uverbs_event_file *ev_file; struct file *filp; - int ret; ev_file = kmalloc(sizeof *ev_file, GFP_KERNEL); if (!ev_file) @@ -516,42 +548,14 @@ struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file, ev_file->uverbs_file = uverbs_file; ev_file->async_queue = NULL; ev_file->is_async = is_async; + ev_file->is_closed = 0; - *fd = get_unused_fd(); - if (*fd < 0) { - ret = *fd; - goto err; - } - - filp = get_empty_filp(); - if (!filp) { - ret = -ENFILE; - goto err_fd; - } - - ev_file->file = filp; - - /* - * fops_get() can't fail here, because we're coming from a - * system call on a uverbs file, which will already have a - * module reference. - */ - filp->f_op = fops_get(&uverbs_event_fops); - filp->f_vfsmnt = mntget(uverbs_event_mnt); - filp->f_dentry = dget(uverbs_event_mnt->mnt_root); - filp->f_mapping = filp->f_dentry->d_inode->i_mapping; - filp->f_flags = O_RDONLY; - filp->f_mode = FMODE_READ; - filp->private_data = ev_file; + filp = anon_inode_getfile("[infinibandevent]", &uverbs_event_fops, + ev_file, O_RDONLY); + if (IS_ERR(filp)) + kfree(ev_file); return filp; - -err_fd: - put_unused_fd(*fd); - -err: - kfree(ev_file); - return ERR_PTR(ret); } /* @@ -562,16 +566,15 @@ err: struct ib_uverbs_event_file *ib_uverbs_lookup_comp_file(int fd) { struct ib_uverbs_event_file *ev_file = NULL; - struct file *filp; + struct fd f = fdget(fd); - filp = fget(fd); - if (!filp) + if (!f.file) return NULL; - if (filp->f_op != &uverbs_event_fops) + if (f.file->f_op != &uverbs_event_fops) goto out; - ev_file = filp->private_data; + ev_file = f.file->private_data; if (ev_file->is_async) { ev_file = NULL; goto out; @@ -580,7 +583,7 @@ struct ib_uverbs_event_file *ib_uverbs_lookup_comp_file(int fd) kref_get(&ev_file->ref); out: - fput(filp); + fdput(f); return ev_file; } @@ -589,6 +592,7 @@ static ssize_t ib_uverbs_write(struct file *filp, const char __user *buf, { struct ib_uverbs_file *file = filp->private_data; struct ib_uverbs_cmd_hdr hdr; + __u32 flags; if (count < sizeof hdr) return -EINVAL; @@ -596,21 +600,110 @@ static ssize_t ib_uverbs_write(struct file *filp, const char __user *buf, if (copy_from_user(&hdr, buf, sizeof hdr)) return -EFAULT; - if (hdr.in_words * 4 != count) - return -EINVAL; + flags = (hdr.command & + IB_USER_VERBS_CMD_FLAGS_MASK) >> IB_USER_VERBS_CMD_FLAGS_SHIFT; - if (hdr.command < 0 || - hdr.command >= ARRAY_SIZE(uverbs_cmd_table) || - !uverbs_cmd_table[hdr.command] || - !(file->device->ib_dev->uverbs_cmd_mask & (1ull << hdr.command))) - return -EINVAL; + if (!flags) { + __u32 command; - if (!file->ucontext && - hdr.command != IB_USER_VERBS_CMD_GET_CONTEXT) - return -EINVAL; + if (hdr.command & ~(__u32)(IB_USER_VERBS_CMD_FLAGS_MASK | + IB_USER_VERBS_CMD_COMMAND_MASK)) + return -EINVAL; + + command = hdr.command & IB_USER_VERBS_CMD_COMMAND_MASK; + + if (command >= ARRAY_SIZE(uverbs_cmd_table) || + !uverbs_cmd_table[command]) + return -EINVAL; + + if (!file->ucontext && + command != IB_USER_VERBS_CMD_GET_CONTEXT) + return -EINVAL; + + if (!(file->device->ib_dev->uverbs_cmd_mask & (1ull << command))) + return -ENOSYS; + + if (hdr.in_words * 4 != count) + return -EINVAL; + + return uverbs_cmd_table[command](file, + buf + sizeof(hdr), + hdr.in_words * 4, + hdr.out_words * 4); + + } else if (flags == IB_USER_VERBS_CMD_FLAG_EXTENDED) { + __u32 command; + + struct ib_uverbs_ex_cmd_hdr ex_hdr; + struct ib_udata ucore; + struct ib_udata uhw; + int err; + size_t written_count = count; + + if (hdr.command & ~(__u32)(IB_USER_VERBS_CMD_FLAGS_MASK | + IB_USER_VERBS_CMD_COMMAND_MASK)) + return -EINVAL; + + command = hdr.command & IB_USER_VERBS_CMD_COMMAND_MASK; + + if (command >= ARRAY_SIZE(uverbs_ex_cmd_table) || + !uverbs_ex_cmd_table[command]) + return -ENOSYS; + + if (!file->ucontext) + return -EINVAL; - return uverbs_cmd_table[hdr.command](file, buf + sizeof hdr, - hdr.in_words * 4, hdr.out_words * 4); + if (!(file->device->ib_dev->uverbs_ex_cmd_mask & (1ull << command))) + return -ENOSYS; + + if (count < (sizeof(hdr) + sizeof(ex_hdr))) + return -EINVAL; + + if (copy_from_user(&ex_hdr, buf + sizeof(hdr), sizeof(ex_hdr))) + return -EFAULT; + + count -= sizeof(hdr) + sizeof(ex_hdr); + buf += sizeof(hdr) + sizeof(ex_hdr); + + if ((hdr.in_words + ex_hdr.provider_in_words) * 8 != count) + return -EINVAL; + + if (ex_hdr.cmd_hdr_reserved) + return -EINVAL; + + if (ex_hdr.response) { + if (!hdr.out_words && !ex_hdr.provider_out_words) + return -EINVAL; + + if (!access_ok(VERIFY_WRITE, + (void __user *) (unsigned long) ex_hdr.response, + (hdr.out_words + ex_hdr.provider_out_words) * 8)) + return -EFAULT; + } else { + if (hdr.out_words || ex_hdr.provider_out_words) + return -EINVAL; + } + + INIT_UDATA_BUF_OR_NULL(&ucore, buf, (unsigned long) ex_hdr.response, + hdr.in_words * 8, hdr.out_words * 8); + + INIT_UDATA_BUF_OR_NULL(&uhw, + buf + ucore.inlen, + (unsigned long) ex_hdr.response + ucore.outlen, + ex_hdr.provider_in_words * 8, + ex_hdr.provider_out_words * 8); + + err = uverbs_ex_cmd_table[command](file, + &ucore, + &uhw); + + if (err) + return err; + + return written_count; + } + + return -ENOSYS; } static int ib_uverbs_mmap(struct file *filp, struct vm_area_struct *vma) @@ -623,19 +716,26 @@ static int ib_uverbs_mmap(struct file *filp, struct vm_area_struct *vma) return file->device->ib_dev->mmap(file->ucontext, vma); } +/* + * ib_uverbs_open() does not need the BKL: + * + * - the ib_uverbs_device structures are properly reference counted and + * everything else is purely local to the file being created, so + * races against other open calls are not a problem; + * - there is no ioctl method to race against; + * - the open method will either immediately run -ENXIO, or all + * required initialization will be done. + */ static int ib_uverbs_open(struct inode *inode, struct file *filp) { struct ib_uverbs_device *dev; struct ib_uverbs_file *file; int ret; - spin_lock(&map_lock); - dev = dev_table[iminor(inode) - IB_UVERBS_BASE_MINOR]; + dev = container_of(inode->i_cdev, struct ib_uverbs_device, cdev); if (dev) kref_get(&dev->ref); - spin_unlock(&map_lock); - - if (!dev) + else return -ENXIO; if (!try_module_get(dev->ib_dev->owner)) { @@ -657,14 +757,13 @@ static int ib_uverbs_open(struct inode *inode, struct file *filp) filp->private_data = file; - return 0; + return nonseekable_open(inode, filp); err_module: module_put(dev->ib_dev->owner); err: kref_put(&dev->ref, ib_uverbs_release_dev); - return ret; } @@ -682,19 +781,21 @@ static int ib_uverbs_close(struct inode *inode, struct file *filp) return 0; } -static struct file_operations uverbs_fops = { - .owner = THIS_MODULE, - .write = ib_uverbs_write, - .open = ib_uverbs_open, - .release = ib_uverbs_close +static const struct file_operations uverbs_fops = { + .owner = THIS_MODULE, + .write = ib_uverbs_write, + .open = ib_uverbs_open, + .release = ib_uverbs_close, + .llseek = no_llseek, }; -static struct file_operations uverbs_mmap_fops = { - .owner = THIS_MODULE, - .write = ib_uverbs_write, +static const struct file_operations uverbs_mmap_fops = { + .owner = THIS_MODULE, + .write = ib_uverbs_write, .mmap = ib_uverbs_mmap, - .open = ib_uverbs_open, - .release = ib_uverbs_close + .open = ib_uverbs_open, + .release = ib_uverbs_close, + .llseek = no_llseek, }; static struct ib_client uverbs_client = { @@ -703,36 +804,65 @@ static struct ib_client uverbs_client = { .remove = ib_uverbs_remove_one }; -static ssize_t show_ibdev(struct class_device *class_dev, char *buf) +static ssize_t show_ibdev(struct device *device, struct device_attribute *attr, + char *buf) { - struct ib_uverbs_device *dev = class_get_devdata(class_dev); + struct ib_uverbs_device *dev = dev_get_drvdata(device); if (!dev) return -ENODEV; return sprintf(buf, "%s\n", dev->ib_dev->name); } -static CLASS_DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL); +static DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL); -static ssize_t show_dev_abi_version(struct class_device *class_dev, char *buf) +static ssize_t show_dev_abi_version(struct device *device, + struct device_attribute *attr, char *buf) { - struct ib_uverbs_device *dev = class_get_devdata(class_dev); + struct ib_uverbs_device *dev = dev_get_drvdata(device); if (!dev) return -ENODEV; return sprintf(buf, "%d\n", dev->ib_dev->uverbs_abi_ver); } -static CLASS_DEVICE_ATTR(abi_version, S_IRUGO, show_dev_abi_version, NULL); +static DEVICE_ATTR(abi_version, S_IRUGO, show_dev_abi_version, NULL); + +static CLASS_ATTR_STRING(abi_version, S_IRUGO, + __stringify(IB_USER_VERBS_ABI_VERSION)); -static ssize_t show_abi_version(struct class *class, char *buf) +static dev_t overflow_maj; +static DECLARE_BITMAP(overflow_map, IB_UVERBS_MAX_DEVICES); + +/* + * If we have more than IB_UVERBS_MAX_DEVICES, dynamically overflow by + * requesting a new major number and doubling the number of max devices we + * support. It's stupid, but simple. + */ +static int find_overflow_devnum(void) { - return sprintf(buf, "%d\n", IB_USER_VERBS_ABI_VERSION); + int ret; + + if (!overflow_maj) { + ret = alloc_chrdev_region(&overflow_maj, 0, IB_UVERBS_MAX_DEVICES, + "infiniband_verbs"); + if (ret) { + printk(KERN_ERR "user_verbs: couldn't register dynamic device number\n"); + return ret; + } + } + + ret = find_first_zero_bit(overflow_map, IB_UVERBS_MAX_DEVICES); + if (ret >= IB_UVERBS_MAX_DEVICES) + return -1; + + return ret; } -static CLASS_ATTR(abi_version, S_IRUGO, show_abi_version, NULL); static void ib_uverbs_add_one(struct ib_device *device) { + int devnum; + dev_t base; struct ib_uverbs_device *uverbs_dev; if (!device->alloc_ucontext) @@ -743,59 +873,68 @@ static void ib_uverbs_add_one(struct ib_device *device) return; kref_init(&uverbs_dev->ref); + init_completion(&uverbs_dev->comp); + uverbs_dev->xrcd_tree = RB_ROOT; + mutex_init(&uverbs_dev->xrcd_tree_mutex); spin_lock(&map_lock); - uverbs_dev->devnum = find_first_zero_bit(dev_map, IB_UVERBS_MAX_DEVICES); - if (uverbs_dev->devnum >= IB_UVERBS_MAX_DEVICES) { + devnum = find_first_zero_bit(dev_map, IB_UVERBS_MAX_DEVICES); + if (devnum >= IB_UVERBS_MAX_DEVICES) { spin_unlock(&map_lock); - goto err; + devnum = find_overflow_devnum(); + if (devnum < 0) + goto err; + + spin_lock(&map_lock); + uverbs_dev->devnum = devnum + IB_UVERBS_MAX_DEVICES; + base = devnum + overflow_maj; + set_bit(devnum, overflow_map); + } else { + uverbs_dev->devnum = devnum; + base = devnum + IB_UVERBS_BASE_DEV; + set_bit(devnum, dev_map); } - set_bit(uverbs_dev->devnum, dev_map); spin_unlock(&map_lock); uverbs_dev->ib_dev = device; - uverbs_dev->num_comp_vectors = 1; + uverbs_dev->num_comp_vectors = device->num_comp_vectors; - uverbs_dev->dev = cdev_alloc(); - if (!uverbs_dev->dev) - goto err; - uverbs_dev->dev->owner = THIS_MODULE; - uverbs_dev->dev->ops = device->mmap ? &uverbs_mmap_fops : &uverbs_fops; - kobject_set_name(&uverbs_dev->dev->kobj, "uverbs%d", uverbs_dev->devnum); - if (cdev_add(uverbs_dev->dev, IB_UVERBS_BASE_DEV + uverbs_dev->devnum, 1)) + cdev_init(&uverbs_dev->cdev, NULL); + uverbs_dev->cdev.owner = THIS_MODULE; + uverbs_dev->cdev.ops = device->mmap ? &uverbs_mmap_fops : &uverbs_fops; + kobject_set_name(&uverbs_dev->cdev.kobj, "uverbs%d", uverbs_dev->devnum); + if (cdev_add(&uverbs_dev->cdev, base, 1)) goto err_cdev; - uverbs_dev->class_dev = class_device_create(uverbs_class, NULL, - uverbs_dev->dev->dev, - device->dma_device, - "uverbs%d", uverbs_dev->devnum); - if (IS_ERR(uverbs_dev->class_dev)) + uverbs_dev->dev = device_create(uverbs_class, device->dma_device, + uverbs_dev->cdev.dev, uverbs_dev, + "uverbs%d", uverbs_dev->devnum); + if (IS_ERR(uverbs_dev->dev)) goto err_cdev; - class_set_devdata(uverbs_dev->class_dev, uverbs_dev); - - if (class_device_create_file(uverbs_dev->class_dev, &class_device_attr_ibdev)) + if (device_create_file(uverbs_dev->dev, &dev_attr_ibdev)) goto err_class; - if (class_device_create_file(uverbs_dev->class_dev, &class_device_attr_abi_version)) + if (device_create_file(uverbs_dev->dev, &dev_attr_abi_version)) goto err_class; - spin_lock(&map_lock); - dev_table[uverbs_dev->devnum] = uverbs_dev; - spin_unlock(&map_lock); - ib_set_client_data(device, &uverbs_client, uverbs_dev); return; err_class: - class_device_destroy(uverbs_class, uverbs_dev->dev->dev); + device_destroy(uverbs_class, uverbs_dev->cdev.dev); err_cdev: - cdev_del(uverbs_dev->dev); - clear_bit(uverbs_dev->devnum, dev_map); + cdev_del(&uverbs_dev->cdev); + if (uverbs_dev->devnum < IB_UVERBS_MAX_DEVICES) + clear_bit(devnum, dev_map); + else + clear_bit(devnum, overflow_map); err: kref_put(&uverbs_dev->ref, ib_uverbs_release_dev); + wait_for_completion(&uverbs_dev->comp); + kfree(uverbs_dev); return; } @@ -806,38 +945,31 @@ static void ib_uverbs_remove_one(struct ib_device *device) if (!uverbs_dev) return; - class_set_devdata(uverbs_dev->class_dev, NULL); - class_device_destroy(uverbs_class, uverbs_dev->dev->dev); - cdev_del(uverbs_dev->dev); + dev_set_drvdata(uverbs_dev->dev, NULL); + device_destroy(uverbs_class, uverbs_dev->cdev.dev); + cdev_del(&uverbs_dev->cdev); - spin_lock(&map_lock); - dev_table[uverbs_dev->devnum] = NULL; - spin_unlock(&map_lock); + if (uverbs_dev->devnum < IB_UVERBS_MAX_DEVICES) + clear_bit(uverbs_dev->devnum, dev_map); + else + clear_bit(uverbs_dev->devnum - IB_UVERBS_MAX_DEVICES, overflow_map); - clear_bit(uverbs_dev->devnum, dev_map); kref_put(&uverbs_dev->ref, ib_uverbs_release_dev); + wait_for_completion(&uverbs_dev->comp); + kfree(uverbs_dev); } -static struct super_block *uverbs_event_get_sb(struct file_system_type *fs_type, int flags, - const char *dev_name, void *data) +static char *uverbs_devnode(struct device *dev, umode_t *mode) { - return get_sb_pseudo(fs_type, "infinibandevent:", NULL, - INFINIBANDEVENTFS_MAGIC); + if (mode) + *mode = 0666; + return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev)); } -static struct file_system_type uverbs_event_fs = { - /* No owner field so module can be unloaded */ - .name = "infinibandeventfs", - .get_sb = uverbs_event_get_sb, - .kill_sb = kill_litter_super -}; - static int __init ib_uverbs_init(void) { int ret; - spin_lock_init(&map_lock); - ret = register_chrdev_region(IB_UVERBS_BASE_DEV, IB_UVERBS_MAX_DEVICES, "infiniband_verbs"); if (ret) { @@ -852,39 +984,22 @@ static int __init ib_uverbs_init(void) goto out_chrdev; } - ret = class_create_file(uverbs_class, &class_attr_abi_version); - if (ret) { - printk(KERN_ERR "user_verbs: couldn't create abi_version attribute\n"); - goto out_class; - } + uverbs_class->devnode = uverbs_devnode; - ret = register_filesystem(&uverbs_event_fs); + ret = class_create_file(uverbs_class, &class_attr_abi_version.attr); if (ret) { - printk(KERN_ERR "user_verbs: couldn't register infinibandeventfs\n"); + printk(KERN_ERR "user_verbs: couldn't create abi_version attribute\n"); goto out_class; } - uverbs_event_mnt = kern_mount(&uverbs_event_fs); - if (IS_ERR(uverbs_event_mnt)) { - ret = PTR_ERR(uverbs_event_mnt); - printk(KERN_ERR "user_verbs: couldn't mount infinibandeventfs\n"); - goto out_fs; - } - ret = ib_register_client(&uverbs_client); if (ret) { printk(KERN_ERR "user_verbs: couldn't register client\n"); - goto out_mnt; + goto out_class; } return 0; -out_mnt: - mntput(uverbs_event_mnt); - -out_fs: - unregister_filesystem(&uverbs_event_fs); - out_class: class_destroy(uverbs_class); @@ -898,10 +1013,10 @@ out: static void __exit ib_uverbs_cleanup(void) { ib_unregister_client(&uverbs_client); - mntput(uverbs_event_mnt); - unregister_filesystem(&uverbs_event_fs); class_destroy(uverbs_class); unregister_chrdev_region(IB_UVERBS_BASE_DEV, IB_UVERBS_MAX_DEVICES); + if (overflow_maj) + unregister_chrdev_region(overflow_maj, IB_UVERBS_MAX_DEVICES); idr_destroy(&ib_uverbs_pd_idr); idr_destroy(&ib_uverbs_mr_idr); idr_destroy(&ib_uverbs_mw_idr); diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c new file mode 100644 index 00000000000..e7bee46868d --- /dev/null +++ b/drivers/infiniband/core/uverbs_marshall.c @@ -0,0 +1,144 @@ +/* + * Copyright (c) 2005 Intel Corporation. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/export.h> +#include <rdma/ib_marshall.h> + +void ib_copy_ah_attr_to_user(struct ib_uverbs_ah_attr *dst, + struct ib_ah_attr *src) +{ + memcpy(dst->grh.dgid, src->grh.dgid.raw, sizeof src->grh.dgid); + dst->grh.flow_label = src->grh.flow_label; + dst->grh.sgid_index = src->grh.sgid_index; + dst->grh.hop_limit = src->grh.hop_limit; + dst->grh.traffic_class = src->grh.traffic_class; + memset(&dst->grh.reserved, 0, sizeof(dst->grh.reserved)); + dst->dlid = src->dlid; + dst->sl = src->sl; + dst->src_path_bits = src->src_path_bits; + dst->static_rate = src->static_rate; + dst->is_global = src->ah_flags & IB_AH_GRH ? 1 : 0; + dst->port_num = src->port_num; + dst->reserved = 0; +} +EXPORT_SYMBOL(ib_copy_ah_attr_to_user); + +void ib_copy_qp_attr_to_user(struct ib_uverbs_qp_attr *dst, + struct ib_qp_attr *src) +{ + dst->qp_state = src->qp_state; + dst->cur_qp_state = src->cur_qp_state; + dst->path_mtu = src->path_mtu; + dst->path_mig_state = src->path_mig_state; + dst->qkey = src->qkey; + dst->rq_psn = src->rq_psn; + dst->sq_psn = src->sq_psn; + dst->dest_qp_num = src->dest_qp_num; + dst->qp_access_flags = src->qp_access_flags; + + dst->max_send_wr = src->cap.max_send_wr; + dst->max_recv_wr = src->cap.max_recv_wr; + dst->max_send_sge = src->cap.max_send_sge; + dst->max_recv_sge = src->cap.max_recv_sge; + dst->max_inline_data = src->cap.max_inline_data; + + ib_copy_ah_attr_to_user(&dst->ah_attr, &src->ah_attr); + ib_copy_ah_attr_to_user(&dst->alt_ah_attr, &src->alt_ah_attr); + + dst->pkey_index = src->pkey_index; + dst->alt_pkey_index = src->alt_pkey_index; + dst->en_sqd_async_notify = src->en_sqd_async_notify; + dst->sq_draining = src->sq_draining; + dst->max_rd_atomic = src->max_rd_atomic; + dst->max_dest_rd_atomic = src->max_dest_rd_atomic; + dst->min_rnr_timer = src->min_rnr_timer; + dst->port_num = src->port_num; + dst->timeout = src->timeout; + dst->retry_cnt = src->retry_cnt; + dst->rnr_retry = src->rnr_retry; + dst->alt_port_num = src->alt_port_num; + dst->alt_timeout = src->alt_timeout; + memset(dst->reserved, 0, sizeof(dst->reserved)); +} +EXPORT_SYMBOL(ib_copy_qp_attr_to_user); + +void ib_copy_path_rec_to_user(struct ib_user_path_rec *dst, + struct ib_sa_path_rec *src) +{ + memcpy(dst->dgid, src->dgid.raw, sizeof src->dgid); + memcpy(dst->sgid, src->sgid.raw, sizeof src->sgid); + + dst->dlid = src->dlid; + dst->slid = src->slid; + dst->raw_traffic = src->raw_traffic; + dst->flow_label = src->flow_label; + dst->hop_limit = src->hop_limit; + dst->traffic_class = src->traffic_class; + dst->reversible = src->reversible; + dst->numb_path = src->numb_path; + dst->pkey = src->pkey; + dst->sl = src->sl; + dst->mtu_selector = src->mtu_selector; + dst->mtu = src->mtu; + dst->rate_selector = src->rate_selector; + dst->rate = src->rate; + dst->packet_life_time = src->packet_life_time; + dst->preference = src->preference; + dst->packet_life_time_selector = src->packet_life_time_selector; +} +EXPORT_SYMBOL(ib_copy_path_rec_to_user); + +void ib_copy_path_rec_from_user(struct ib_sa_path_rec *dst, + struct ib_user_path_rec *src) +{ + memcpy(dst->dgid.raw, src->dgid, sizeof dst->dgid); + memcpy(dst->sgid.raw, src->sgid, sizeof dst->sgid); + + dst->dlid = src->dlid; + dst->slid = src->slid; + dst->raw_traffic = src->raw_traffic; + dst->flow_label = src->flow_label; + dst->hop_limit = src->hop_limit; + dst->traffic_class = src->traffic_class; + dst->reversible = src->reversible; + dst->numb_path = src->numb_path; + dst->pkey = src->pkey; + dst->sl = src->sl; + dst->mtu_selector = src->mtu_selector; + dst->mtu = src->mtu; + dst->rate_selector = src->rate_selector; + dst->rate = src->rate; + dst->packet_life_time = src->packet_life_time; + dst->preference = src->preference; + dst->packet_life_time_selector = src->packet_life_time_selector; +} +EXPORT_SYMBOL(ib_copy_path_rec_from_user); diff --git a/drivers/infiniband/core/uverbs_mem.c b/drivers/infiniband/core/uverbs_mem.c deleted file mode 100644 index 36a32c31566..00000000000 --- a/drivers/infiniband/core/uverbs_mem.c +++ /dev/null @@ -1,222 +0,0 @@ -/* - * Copyright (c) 2005 Topspin Communications. All rights reserved. - * Copyright (c) 2005 Cisco Systems. All rights reserved. - * Copyright (c) 2005 Mellanox Technologies. All rights reserved. - * - * This software is available to you under a choice of one of two - * licenses. You may choose to be licensed under the terms of the GNU - * General Public License (GPL) Version 2, available from the file - * COPYING in the main directory of this source tree, or the - * OpenIB.org BSD license below: - * - * Redistribution and use in source and binary forms, with or - * without modification, are permitted provided that the following - * conditions are met: - * - * - Redistributions of source code must retain the above - * copyright notice, this list of conditions and the following - * disclaimer. - * - * - Redistributions in binary form must reproduce the above - * copyright notice, this list of conditions and the following - * disclaimer in the documentation and/or other materials - * provided with the distribution. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, - * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND - * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS - * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN - * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - * - * $Id: uverbs_mem.c 2743 2005-06-28 22:27:59Z roland $ - */ - -#include <linux/mm.h> -#include <linux/dma-mapping.h> - -#include "uverbs.h" - -struct ib_umem_account_work { - struct work_struct work; - struct mm_struct *mm; - unsigned long diff; -}; - - -static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int dirty) -{ - struct ib_umem_chunk *chunk, *tmp; - int i; - - list_for_each_entry_safe(chunk, tmp, &umem->chunk_list, list) { - dma_unmap_sg(dev->dma_device, chunk->page_list, - chunk->nents, DMA_BIDIRECTIONAL); - for (i = 0; i < chunk->nents; ++i) { - if (umem->writable && dirty) - set_page_dirty_lock(chunk->page_list[i].page); - put_page(chunk->page_list[i].page); - } - - kfree(chunk); - } -} - -int ib_umem_get(struct ib_device *dev, struct ib_umem *mem, - void *addr, size_t size, int write) -{ - struct page **page_list; - struct ib_umem_chunk *chunk; - unsigned long locked; - unsigned long lock_limit; - unsigned long cur_base; - unsigned long npages; - int ret = 0; - int off; - int i; - - if (!can_do_mlock()) - return -EPERM; - - page_list = (struct page **) __get_free_page(GFP_KERNEL); - if (!page_list) - return -ENOMEM; - - mem->user_base = (unsigned long) addr; - mem->length = size; - mem->offset = (unsigned long) addr & ~PAGE_MASK; - mem->page_size = PAGE_SIZE; - mem->writable = write; - - INIT_LIST_HEAD(&mem->chunk_list); - - npages = PAGE_ALIGN(size + mem->offset) >> PAGE_SHIFT; - - down_write(¤t->mm->mmap_sem); - - locked = npages + current->mm->locked_vm; - lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; - - if ((locked > lock_limit) && !capable(CAP_IPC_LOCK)) { - ret = -ENOMEM; - goto out; - } - - cur_base = (unsigned long) addr & PAGE_MASK; - - while (npages) { - ret = get_user_pages(current, current->mm, cur_base, - min_t(int, npages, - PAGE_SIZE / sizeof (struct page *)), - 1, !write, page_list, NULL); - - if (ret < 0) - goto out; - - cur_base += ret * PAGE_SIZE; - npages -= ret; - - off = 0; - - while (ret) { - chunk = kmalloc(sizeof *chunk + sizeof (struct scatterlist) * - min_t(int, ret, IB_UMEM_MAX_PAGE_CHUNK), - GFP_KERNEL); - if (!chunk) { - ret = -ENOMEM; - goto out; - } - - chunk->nents = min_t(int, ret, IB_UMEM_MAX_PAGE_CHUNK); - for (i = 0; i < chunk->nents; ++i) { - chunk->page_list[i].page = page_list[i + off]; - chunk->page_list[i].offset = 0; - chunk->page_list[i].length = PAGE_SIZE; - } - - chunk->nmap = dma_map_sg(dev->dma_device, - &chunk->page_list[0], - chunk->nents, - DMA_BIDIRECTIONAL); - if (chunk->nmap <= 0) { - for (i = 0; i < chunk->nents; ++i) - put_page(chunk->page_list[i].page); - kfree(chunk); - - ret = -ENOMEM; - goto out; - } - - ret -= chunk->nents; - off += chunk->nents; - list_add_tail(&chunk->list, &mem->chunk_list); - } - - ret = 0; - } - -out: - if (ret < 0) - __ib_umem_release(dev, mem, 0); - else - current->mm->locked_vm = locked; - - up_write(¤t->mm->mmap_sem); - free_page((unsigned long) page_list); - - return ret; -} - -void ib_umem_release(struct ib_device *dev, struct ib_umem *umem) -{ - __ib_umem_release(dev, umem, 1); - - down_write(¤t->mm->mmap_sem); - current->mm->locked_vm -= - PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT; - up_write(¤t->mm->mmap_sem); -} - -static void ib_umem_account(void *work_ptr) -{ - struct ib_umem_account_work *work = work_ptr; - - down_write(&work->mm->mmap_sem); - work->mm->locked_vm -= work->diff; - up_write(&work->mm->mmap_sem); - mmput(work->mm); - kfree(work); -} - -void ib_umem_release_on_close(struct ib_device *dev, struct ib_umem *umem) -{ - struct ib_umem_account_work *work; - struct mm_struct *mm; - - __ib_umem_release(dev, umem, 1); - - mm = get_task_mm(current); - if (!mm) - return; - - /* - * We may be called with the mm's mmap_sem already held. This - * can happen when a userspace munmap() is the call that drops - * the last reference to our file and calls our release - * method. If there are memory regions to destroy, we'll end - * up here and not be able to take the mmap_sem. Therefore we - * defer the vm_locked accounting to the system workqueue. - */ - - work = kmalloc(sizeof *work, GFP_KERNEL); - if (!work) - return; - - INIT_WORK(&work->work, ib_umem_account, work); - work->mm = mm; - work->diff = PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT; - - schedule_work(&work->work); -} diff --git a/drivers/infiniband/core/verbs.c b/drivers/infiniband/core/verbs.c index c857361be44..c2b89cc5dbc 100644 --- a/drivers/infiniband/core/verbs.c +++ b/drivers/infiniband/core/verbs.c @@ -5,7 +5,7 @@ * Copyright (c) 2004 Topspin Corporation. All rights reserved. * Copyright (c) 2004 Voltaire Corporation. All rights reserved. * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. - * Copyright (c) 2005 Cisco Systems. All rights reserved. + * Copyright (c) 2005, 2006 Cisco Systems. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -34,16 +34,117 @@ * 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. - * - * $Id: verbs.c 1349 2004-12-16 21:09:43Z roland $ */ #include <linux/errno.h> #include <linux/err.h> +#include <linux/export.h> #include <linux/string.h> +#include <linux/slab.h> #include <rdma/ib_verbs.h> #include <rdma/ib_cache.h> +#include <rdma/ib_addr.h> + +#include "core_priv.h" + +__attribute_const__ int ib_rate_to_mult(enum ib_rate rate) +{ + switch (rate) { + case IB_RATE_2_5_GBPS: return 1; + case IB_RATE_5_GBPS: return 2; + case IB_RATE_10_GBPS: return 4; + case IB_RATE_20_GBPS: return 8; + case IB_RATE_30_GBPS: return 12; + case IB_RATE_40_GBPS: return 16; + case IB_RATE_60_GBPS: return 24; + case IB_RATE_80_GBPS: return 32; + case IB_RATE_120_GBPS: return 48; + default: return -1; + } +} +EXPORT_SYMBOL(ib_rate_to_mult); + +__attribute_const__ enum ib_rate mult_to_ib_rate(int mult) +{ + switch (mult) { + case 1: return IB_RATE_2_5_GBPS; + case 2: return IB_RATE_5_GBPS; + case 4: return IB_RATE_10_GBPS; + case 8: return IB_RATE_20_GBPS; + case 12: return IB_RATE_30_GBPS; + case 16: return IB_RATE_40_GBPS; + case 24: return IB_RATE_60_GBPS; + case 32: return IB_RATE_80_GBPS; + case 48: return IB_RATE_120_GBPS; + default: return IB_RATE_PORT_CURRENT; + } +} +EXPORT_SYMBOL(mult_to_ib_rate); + +__attribute_const__ int ib_rate_to_mbps(enum ib_rate rate) +{ + switch (rate) { + case IB_RATE_2_5_GBPS: return 2500; + case IB_RATE_5_GBPS: return 5000; + case IB_RATE_10_GBPS: return 10000; + case IB_RATE_20_GBPS: return 20000; + case IB_RATE_30_GBPS: return 30000; + case IB_RATE_40_GBPS: return 40000; + case IB_RATE_60_GBPS: return 60000; + case IB_RATE_80_GBPS: return 80000; + case IB_RATE_120_GBPS: return 120000; + case IB_RATE_14_GBPS: return 14062; + case IB_RATE_56_GBPS: return 56250; + case IB_RATE_112_GBPS: return 112500; + case IB_RATE_168_GBPS: return 168750; + case IB_RATE_25_GBPS: return 25781; + case IB_RATE_100_GBPS: return 103125; + case IB_RATE_200_GBPS: return 206250; + case IB_RATE_300_GBPS: return 309375; + default: return -1; + } +} +EXPORT_SYMBOL(ib_rate_to_mbps); + +__attribute_const__ enum rdma_transport_type +rdma_node_get_transport(enum rdma_node_type node_type) +{ + switch (node_type) { + case RDMA_NODE_IB_CA: + case RDMA_NODE_IB_SWITCH: + case RDMA_NODE_IB_ROUTER: + return RDMA_TRANSPORT_IB; + case RDMA_NODE_RNIC: + return RDMA_TRANSPORT_IWARP; + case RDMA_NODE_USNIC: + return RDMA_TRANSPORT_USNIC; + case RDMA_NODE_USNIC_UDP: + return RDMA_TRANSPORT_USNIC_UDP; + default: + BUG(); + return 0; + } +} +EXPORT_SYMBOL(rdma_node_get_transport); + +enum rdma_link_layer rdma_port_get_link_layer(struct ib_device *device, u8 port_num) +{ + if (device->get_link_layer) + return device->get_link_layer(device, port_num); + + switch (rdma_node_get_transport(device->node_type)) { + case RDMA_TRANSPORT_IB: + return IB_LINK_LAYER_INFINIBAND; + case RDMA_TRANSPORT_IWARP: + case RDMA_TRANSPORT_USNIC: + case RDMA_TRANSPORT_USNIC_UDP: + return IB_LINK_LAYER_ETHERNET; + default: + return IB_LINK_LAYER_UNSPECIFIED; + } +} +EXPORT_SYMBOL(rdma_port_get_link_layer); /* Protection domains */ @@ -91,35 +192,67 @@ struct ib_ah *ib_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr) } EXPORT_SYMBOL(ib_create_ah); -struct ib_ah *ib_create_ah_from_wc(struct ib_pd *pd, struct ib_wc *wc, - struct ib_grh *grh, u8 port_num) +int ib_init_ah_from_wc(struct ib_device *device, u8 port_num, struct ib_wc *wc, + struct ib_grh *grh, struct ib_ah_attr *ah_attr) { - struct ib_ah_attr ah_attr; u32 flow_class; u16 gid_index; int ret; + int is_eth = (rdma_port_get_link_layer(device, port_num) == + IB_LINK_LAYER_ETHERNET); + + memset(ah_attr, 0, sizeof *ah_attr); + if (is_eth) { + if (!(wc->wc_flags & IB_WC_GRH)) + return -EPROTOTYPE; + + if (wc->wc_flags & IB_WC_WITH_SMAC && + wc->wc_flags & IB_WC_WITH_VLAN) { + memcpy(ah_attr->dmac, wc->smac, ETH_ALEN); + ah_attr->vlan_id = wc->vlan_id; + } else { + ret = rdma_addr_find_dmac_by_grh(&grh->dgid, &grh->sgid, + ah_attr->dmac, &ah_attr->vlan_id); + if (ret) + return ret; + } + } else { + ah_attr->vlan_id = 0xffff; + } - memset(&ah_attr, 0, sizeof ah_attr); - ah_attr.dlid = wc->slid; - ah_attr.sl = wc->sl; - ah_attr.src_path_bits = wc->dlid_path_bits; - ah_attr.port_num = port_num; + ah_attr->dlid = wc->slid; + ah_attr->sl = wc->sl; + ah_attr->src_path_bits = wc->dlid_path_bits; + ah_attr->port_num = port_num; if (wc->wc_flags & IB_WC_GRH) { - ah_attr.ah_flags = IB_AH_GRH; - ah_attr.grh.dgid = grh->sgid; + ah_attr->ah_flags = IB_AH_GRH; + ah_attr->grh.dgid = grh->sgid; - ret = ib_find_cached_gid(pd->device, &grh->dgid, &port_num, + ret = ib_find_cached_gid(device, &grh->dgid, &port_num, &gid_index); if (ret) - return ERR_PTR(ret); + return ret; - ah_attr.grh.sgid_index = (u8) gid_index; + ah_attr->grh.sgid_index = (u8) gid_index; flow_class = be32_to_cpu(grh->version_tclass_flow); - ah_attr.grh.flow_label = flow_class & 0xFFFFF; - ah_attr.grh.traffic_class = (flow_class >> 20) & 0xFF; - ah_attr.grh.hop_limit = grh->hop_limit; + ah_attr->grh.flow_label = flow_class & 0xFFFFF; + ah_attr->grh.hop_limit = 0xFF; + ah_attr->grh.traffic_class = (flow_class >> 20) & 0xFF; } + return 0; +} +EXPORT_SYMBOL(ib_init_ah_from_wc); + +struct ib_ah *ib_create_ah_from_wc(struct ib_pd *pd, struct ib_wc *wc, + struct ib_grh *grh, u8 port_num) +{ + struct ib_ah_attr ah_attr; + int ret; + + ret = ib_init_ah_from_wc(pd->device, port_num, wc, grh, &ah_attr); + if (ret) + return ERR_PTR(ret); return ib_create_ah(pd, &ah_attr); } @@ -173,6 +306,13 @@ struct ib_srq *ib_create_srq(struct ib_pd *pd, srq->uobject = NULL; srq->event_handler = srq_init_attr->event_handler; srq->srq_context = srq_init_attr->srq_context; + srq->srq_type = srq_init_attr->srq_type; + if (srq->srq_type == IB_SRQT_XRC) { + srq->ext.xrc.xrcd = srq_init_attr->ext.xrc.xrcd; + srq->ext.xrc.cq = srq_init_attr->ext.xrc.cq; + atomic_inc(&srq->ext.xrc.xrcd->usecnt); + atomic_inc(&srq->ext.xrc.cq->usecnt); + } atomic_inc(&pd->usecnt); atomic_set(&srq->usecnt, 0); } @@ -185,7 +325,9 @@ int ib_modify_srq(struct ib_srq *srq, struct ib_srq_attr *srq_attr, enum ib_srq_attr_mask srq_attr_mask) { - return srq->device->modify_srq(srq, srq_attr, srq_attr_mask); + return srq->device->modify_srq ? + srq->device->modify_srq(srq, srq_attr, srq_attr_mask, NULL) : + -ENOSYS; } EXPORT_SYMBOL(ib_modify_srq); @@ -200,16 +342,29 @@ EXPORT_SYMBOL(ib_query_srq); int ib_destroy_srq(struct ib_srq *srq) { struct ib_pd *pd; + enum ib_srq_type srq_type; + struct ib_xrcd *uninitialized_var(xrcd); + struct ib_cq *uninitialized_var(cq); int ret; if (atomic_read(&srq->usecnt)) return -EBUSY; pd = srq->pd; + srq_type = srq->srq_type; + if (srq_type == IB_SRQT_XRC) { + xrcd = srq->ext.xrc.xrcd; + cq = srq->ext.xrc.cq; + } ret = srq->device->destroy_srq(srq); - if (!ret) + if (!ret) { atomic_dec(&pd->usecnt); + if (srq_type == IB_SRQT_XRC) { + atomic_dec(&xrcd->usecnt); + atomic_dec(&cq->usecnt); + } + } return ret; } @@ -217,39 +372,544 @@ EXPORT_SYMBOL(ib_destroy_srq); /* Queue pairs */ +static void __ib_shared_qp_event_handler(struct ib_event *event, void *context) +{ + struct ib_qp *qp = context; + unsigned long flags; + + spin_lock_irqsave(&qp->device->event_handler_lock, flags); + list_for_each_entry(event->element.qp, &qp->open_list, open_list) + if (event->element.qp->event_handler) + event->element.qp->event_handler(event, event->element.qp->qp_context); + spin_unlock_irqrestore(&qp->device->event_handler_lock, flags); +} + +static void __ib_insert_xrcd_qp(struct ib_xrcd *xrcd, struct ib_qp *qp) +{ + mutex_lock(&xrcd->tgt_qp_mutex); + list_add(&qp->xrcd_list, &xrcd->tgt_qp_list); + mutex_unlock(&xrcd->tgt_qp_mutex); +} + +static struct ib_qp *__ib_open_qp(struct ib_qp *real_qp, + void (*event_handler)(struct ib_event *, void *), + void *qp_context) +{ + struct ib_qp *qp; + unsigned long flags; + + qp = kzalloc(sizeof *qp, GFP_KERNEL); + if (!qp) + return ERR_PTR(-ENOMEM); + + qp->real_qp = real_qp; + atomic_inc(&real_qp->usecnt); + qp->device = real_qp->device; + qp->event_handler = event_handler; + qp->qp_context = qp_context; + qp->qp_num = real_qp->qp_num; + qp->qp_type = real_qp->qp_type; + + spin_lock_irqsave(&real_qp->device->event_handler_lock, flags); + list_add(&qp->open_list, &real_qp->open_list); + spin_unlock_irqrestore(&real_qp->device->event_handler_lock, flags); + + return qp; +} + +struct ib_qp *ib_open_qp(struct ib_xrcd *xrcd, + struct ib_qp_open_attr *qp_open_attr) +{ + struct ib_qp *qp, *real_qp; + + if (qp_open_attr->qp_type != IB_QPT_XRC_TGT) + return ERR_PTR(-EINVAL); + + qp = ERR_PTR(-EINVAL); + mutex_lock(&xrcd->tgt_qp_mutex); + list_for_each_entry(real_qp, &xrcd->tgt_qp_list, xrcd_list) { + if (real_qp->qp_num == qp_open_attr->qp_num) { + qp = __ib_open_qp(real_qp, qp_open_attr->event_handler, + qp_open_attr->qp_context); + break; + } + } + mutex_unlock(&xrcd->tgt_qp_mutex); + return qp; +} +EXPORT_SYMBOL(ib_open_qp); + struct ib_qp *ib_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *qp_init_attr) { - struct ib_qp *qp; + struct ib_qp *qp, *real_qp; + struct ib_device *device; - qp = pd->device->create_qp(pd, qp_init_attr, NULL); + device = pd ? pd->device : qp_init_attr->xrcd->device; + qp = device->create_qp(pd, qp_init_attr, NULL); if (!IS_ERR(qp)) { - qp->device = pd->device; - qp->pd = pd; - qp->send_cq = qp_init_attr->send_cq; - qp->recv_cq = qp_init_attr->recv_cq; - qp->srq = qp_init_attr->srq; - qp->uobject = NULL; - qp->event_handler = qp_init_attr->event_handler; - qp->qp_context = qp_init_attr->qp_context; - qp->qp_type = qp_init_attr->qp_type; - atomic_inc(&pd->usecnt); - atomic_inc(&qp_init_attr->send_cq->usecnt); - atomic_inc(&qp_init_attr->recv_cq->usecnt); - if (qp_init_attr->srq) - atomic_inc(&qp_init_attr->srq->usecnt); + qp->device = device; + qp->real_qp = qp; + qp->uobject = NULL; + qp->qp_type = qp_init_attr->qp_type; + + atomic_set(&qp->usecnt, 0); + if (qp_init_attr->qp_type == IB_QPT_XRC_TGT) { + qp->event_handler = __ib_shared_qp_event_handler; + qp->qp_context = qp; + qp->pd = NULL; + qp->send_cq = qp->recv_cq = NULL; + qp->srq = NULL; + qp->xrcd = qp_init_attr->xrcd; + atomic_inc(&qp_init_attr->xrcd->usecnt); + INIT_LIST_HEAD(&qp->open_list); + + real_qp = qp; + qp = __ib_open_qp(real_qp, qp_init_attr->event_handler, + qp_init_attr->qp_context); + if (!IS_ERR(qp)) + __ib_insert_xrcd_qp(qp_init_attr->xrcd, real_qp); + else + real_qp->device->destroy_qp(real_qp); + } else { + qp->event_handler = qp_init_attr->event_handler; + qp->qp_context = qp_init_attr->qp_context; + if (qp_init_attr->qp_type == IB_QPT_XRC_INI) { + qp->recv_cq = NULL; + qp->srq = NULL; + } else { + qp->recv_cq = qp_init_attr->recv_cq; + atomic_inc(&qp_init_attr->recv_cq->usecnt); + qp->srq = qp_init_attr->srq; + if (qp->srq) + atomic_inc(&qp_init_attr->srq->usecnt); + } + + qp->pd = pd; + qp->send_cq = qp_init_attr->send_cq; + qp->xrcd = NULL; + + atomic_inc(&pd->usecnt); + atomic_inc(&qp_init_attr->send_cq->usecnt); + } } return qp; } EXPORT_SYMBOL(ib_create_qp); +static const struct { + int valid; + enum ib_qp_attr_mask req_param[IB_QPT_MAX]; + enum ib_qp_attr_mask req_param_add_eth[IB_QPT_MAX]; + enum ib_qp_attr_mask opt_param[IB_QPT_MAX]; + enum ib_qp_attr_mask opt_param_add_eth[IB_QPT_MAX]; +} qp_state_table[IB_QPS_ERR + 1][IB_QPS_ERR + 1] = { + [IB_QPS_RESET] = { + [IB_QPS_RESET] = { .valid = 1 }, + [IB_QPS_INIT] = { + .valid = 1, + .req_param = { + [IB_QPT_UD] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_QKEY), + [IB_QPT_RAW_PACKET] = IB_QP_PORT, + [IB_QPT_UC] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_RC] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_XRC_INI] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_XRC_TGT] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_SMI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + } + }, + }, + [IB_QPS_INIT] = { + [IB_QPS_RESET] = { .valid = 1 }, + [IB_QPS_ERR] = { .valid = 1 }, + [IB_QPS_INIT] = { + .valid = 1, + .opt_param = { + [IB_QPT_UD] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_QKEY), + [IB_QPT_UC] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_RC] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_XRC_INI] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_XRC_TGT] = (IB_QP_PKEY_INDEX | + IB_QP_PORT | + IB_QP_ACCESS_FLAGS), + [IB_QPT_SMI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + } + }, + [IB_QPS_RTR] = { + .valid = 1, + .req_param = { + [IB_QPT_UC] = (IB_QP_AV | + IB_QP_PATH_MTU | + IB_QP_DEST_QPN | + IB_QP_RQ_PSN), + [IB_QPT_RC] = (IB_QP_AV | + IB_QP_PATH_MTU | + IB_QP_DEST_QPN | + IB_QP_RQ_PSN | + IB_QP_MAX_DEST_RD_ATOMIC | + IB_QP_MIN_RNR_TIMER), + [IB_QPT_XRC_INI] = (IB_QP_AV | + IB_QP_PATH_MTU | + IB_QP_DEST_QPN | + IB_QP_RQ_PSN), + [IB_QPT_XRC_TGT] = (IB_QP_AV | + IB_QP_PATH_MTU | + IB_QP_DEST_QPN | + IB_QP_RQ_PSN | + IB_QP_MAX_DEST_RD_ATOMIC | + IB_QP_MIN_RNR_TIMER), + }, + .req_param_add_eth = { + [IB_QPT_RC] = (IB_QP_SMAC), + [IB_QPT_UC] = (IB_QP_SMAC), + [IB_QPT_XRC_INI] = (IB_QP_SMAC), + [IB_QPT_XRC_TGT] = (IB_QP_SMAC) + }, + .opt_param = { + [IB_QPT_UD] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + [IB_QPT_UC] = (IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX), + [IB_QPT_RC] = (IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX), + [IB_QPT_XRC_INI] = (IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX), + [IB_QPT_XRC_TGT] = (IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX), + [IB_QPT_SMI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + }, + .opt_param_add_eth = { + [IB_QPT_RC] = (IB_QP_ALT_SMAC | + IB_QP_VID | + IB_QP_ALT_VID), + [IB_QPT_UC] = (IB_QP_ALT_SMAC | + IB_QP_VID | + IB_QP_ALT_VID), + [IB_QPT_XRC_INI] = (IB_QP_ALT_SMAC | + IB_QP_VID | + IB_QP_ALT_VID), + [IB_QPT_XRC_TGT] = (IB_QP_ALT_SMAC | + IB_QP_VID | + IB_QP_ALT_VID) + } + } + }, + [IB_QPS_RTR] = { + [IB_QPS_RESET] = { .valid = 1 }, + [IB_QPS_ERR] = { .valid = 1 }, + [IB_QPS_RTS] = { + .valid = 1, + .req_param = { + [IB_QPT_UD] = IB_QP_SQ_PSN, + [IB_QPT_UC] = IB_QP_SQ_PSN, + [IB_QPT_RC] = (IB_QP_TIMEOUT | + IB_QP_RETRY_CNT | + IB_QP_RNR_RETRY | + IB_QP_SQ_PSN | + IB_QP_MAX_QP_RD_ATOMIC), + [IB_QPT_XRC_INI] = (IB_QP_TIMEOUT | + IB_QP_RETRY_CNT | + IB_QP_RNR_RETRY | + IB_QP_SQ_PSN | + IB_QP_MAX_QP_RD_ATOMIC), + [IB_QPT_XRC_TGT] = (IB_QP_TIMEOUT | + IB_QP_SQ_PSN), + [IB_QPT_SMI] = IB_QP_SQ_PSN, + [IB_QPT_GSI] = IB_QP_SQ_PSN, + }, + .opt_param = { + [IB_QPT_UD] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_UC] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PATH_MIG_STATE), + [IB_QPT_RC] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_MIN_RNR_TIMER | + IB_QP_PATH_MIG_STATE), + [IB_QPT_XRC_INI] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PATH_MIG_STATE), + [IB_QPT_XRC_TGT] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_MIN_RNR_TIMER | + IB_QP_PATH_MIG_STATE), + [IB_QPT_SMI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + } + } + }, + [IB_QPS_RTS] = { + [IB_QPS_RESET] = { .valid = 1 }, + [IB_QPS_ERR] = { .valid = 1 }, + [IB_QPS_RTS] = { + .valid = 1, + .opt_param = { + [IB_QPT_UD] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_UC] = (IB_QP_CUR_STATE | + IB_QP_ACCESS_FLAGS | + IB_QP_ALT_PATH | + IB_QP_PATH_MIG_STATE), + [IB_QPT_RC] = (IB_QP_CUR_STATE | + IB_QP_ACCESS_FLAGS | + IB_QP_ALT_PATH | + IB_QP_PATH_MIG_STATE | + IB_QP_MIN_RNR_TIMER), + [IB_QPT_XRC_INI] = (IB_QP_CUR_STATE | + IB_QP_ACCESS_FLAGS | + IB_QP_ALT_PATH | + IB_QP_PATH_MIG_STATE), + [IB_QPT_XRC_TGT] = (IB_QP_CUR_STATE | + IB_QP_ACCESS_FLAGS | + IB_QP_ALT_PATH | + IB_QP_PATH_MIG_STATE | + IB_QP_MIN_RNR_TIMER), + [IB_QPT_SMI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + } + }, + [IB_QPS_SQD] = { + .valid = 1, + .opt_param = { + [IB_QPT_UD] = IB_QP_EN_SQD_ASYNC_NOTIFY, + [IB_QPT_UC] = IB_QP_EN_SQD_ASYNC_NOTIFY, + [IB_QPT_RC] = IB_QP_EN_SQD_ASYNC_NOTIFY, + [IB_QPT_XRC_INI] = IB_QP_EN_SQD_ASYNC_NOTIFY, + [IB_QPT_XRC_TGT] = IB_QP_EN_SQD_ASYNC_NOTIFY, /* ??? */ + [IB_QPT_SMI] = IB_QP_EN_SQD_ASYNC_NOTIFY, + [IB_QPT_GSI] = IB_QP_EN_SQD_ASYNC_NOTIFY + } + }, + }, + [IB_QPS_SQD] = { + [IB_QPS_RESET] = { .valid = 1 }, + [IB_QPS_ERR] = { .valid = 1 }, + [IB_QPS_RTS] = { + .valid = 1, + .opt_param = { + [IB_QPT_UD] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_UC] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PATH_MIG_STATE), + [IB_QPT_RC] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_MIN_RNR_TIMER | + IB_QP_PATH_MIG_STATE), + [IB_QPT_XRC_INI] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PATH_MIG_STATE), + [IB_QPT_XRC_TGT] = (IB_QP_CUR_STATE | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_MIN_RNR_TIMER | + IB_QP_PATH_MIG_STATE), + [IB_QPT_SMI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + } + }, + [IB_QPS_SQD] = { + .valid = 1, + .opt_param = { + [IB_QPT_UD] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + [IB_QPT_UC] = (IB_QP_AV | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX | + IB_QP_PATH_MIG_STATE), + [IB_QPT_RC] = (IB_QP_PORT | + IB_QP_AV | + IB_QP_TIMEOUT | + IB_QP_RETRY_CNT | + IB_QP_RNR_RETRY | + IB_QP_MAX_QP_RD_ATOMIC | + IB_QP_MAX_DEST_RD_ATOMIC | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX | + IB_QP_MIN_RNR_TIMER | + IB_QP_PATH_MIG_STATE), + [IB_QPT_XRC_INI] = (IB_QP_PORT | + IB_QP_AV | + IB_QP_TIMEOUT | + IB_QP_RETRY_CNT | + IB_QP_RNR_RETRY | + IB_QP_MAX_QP_RD_ATOMIC | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX | + IB_QP_PATH_MIG_STATE), + [IB_QPT_XRC_TGT] = (IB_QP_PORT | + IB_QP_AV | + IB_QP_TIMEOUT | + IB_QP_MAX_DEST_RD_ATOMIC | + IB_QP_ALT_PATH | + IB_QP_ACCESS_FLAGS | + IB_QP_PKEY_INDEX | + IB_QP_MIN_RNR_TIMER | + IB_QP_PATH_MIG_STATE), + [IB_QPT_SMI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_PKEY_INDEX | + IB_QP_QKEY), + } + } + }, + [IB_QPS_SQE] = { + [IB_QPS_RESET] = { .valid = 1 }, + [IB_QPS_ERR] = { .valid = 1 }, + [IB_QPS_RTS] = { + .valid = 1, + .opt_param = { + [IB_QPT_UD] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_UC] = (IB_QP_CUR_STATE | + IB_QP_ACCESS_FLAGS), + [IB_QPT_SMI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + [IB_QPT_GSI] = (IB_QP_CUR_STATE | + IB_QP_QKEY), + } + } + }, + [IB_QPS_ERR] = { + [IB_QPS_RESET] = { .valid = 1 }, + [IB_QPS_ERR] = { .valid = 1 } + } +}; + +int ib_modify_qp_is_ok(enum ib_qp_state cur_state, enum ib_qp_state next_state, + enum ib_qp_type type, enum ib_qp_attr_mask mask, + enum rdma_link_layer ll) +{ + enum ib_qp_attr_mask req_param, opt_param; + + if (cur_state < 0 || cur_state > IB_QPS_ERR || + next_state < 0 || next_state > IB_QPS_ERR) + return 0; + + if (mask & IB_QP_CUR_STATE && + cur_state != IB_QPS_RTR && cur_state != IB_QPS_RTS && + cur_state != IB_QPS_SQD && cur_state != IB_QPS_SQE) + return 0; + + if (!qp_state_table[cur_state][next_state].valid) + return 0; + + req_param = qp_state_table[cur_state][next_state].req_param[type]; + opt_param = qp_state_table[cur_state][next_state].opt_param[type]; + + if (ll == IB_LINK_LAYER_ETHERNET) { + req_param |= qp_state_table[cur_state][next_state]. + req_param_add_eth[type]; + opt_param |= qp_state_table[cur_state][next_state]. + opt_param_add_eth[type]; + } + + if ((mask & req_param) != req_param) + return 0; + + if (mask & ~(req_param | opt_param | IB_QP_STATE)) + return 0; + + return 1; +} +EXPORT_SYMBOL(ib_modify_qp_is_ok); + +int ib_resolve_eth_l2_attrs(struct ib_qp *qp, + struct ib_qp_attr *qp_attr, int *qp_attr_mask) +{ + int ret = 0; + union ib_gid sgid; + + if ((*qp_attr_mask & IB_QP_AV) && + (rdma_port_get_link_layer(qp->device, qp_attr->ah_attr.port_num) == IB_LINK_LAYER_ETHERNET)) { + ret = ib_query_gid(qp->device, qp_attr->ah_attr.port_num, + qp_attr->ah_attr.grh.sgid_index, &sgid); + if (ret) + goto out; + if (rdma_link_local_addr((struct in6_addr *)qp_attr->ah_attr.grh.dgid.raw)) { + rdma_get_ll_mac((struct in6_addr *)qp_attr->ah_attr.grh.dgid.raw, qp_attr->ah_attr.dmac); + rdma_get_ll_mac((struct in6_addr *)sgid.raw, qp_attr->smac); + qp_attr->vlan_id = rdma_get_vlan_id(&sgid); + } else { + ret = rdma_addr_find_dmac_by_grh(&sgid, &qp_attr->ah_attr.grh.dgid, + qp_attr->ah_attr.dmac, &qp_attr->vlan_id); + if (ret) + goto out; + ret = rdma_addr_find_smac_by_sgid(&sgid, qp_attr->smac, NULL); + if (ret) + goto out; + } + *qp_attr_mask |= IB_QP_SMAC; + if (qp_attr->vlan_id < 0xFFFF) + *qp_attr_mask |= IB_QP_VID; + } +out: + return ret; +} +EXPORT_SYMBOL(ib_resolve_eth_l2_attrs); + + int ib_modify_qp(struct ib_qp *qp, struct ib_qp_attr *qp_attr, int qp_attr_mask) { - return qp->device->modify_qp(qp, qp_attr, qp_attr_mask); + int ret; + + ret = ib_resolve_eth_l2_attrs(qp, qp_attr, &qp_attr_mask); + if (ret) + return ret; + + return qp->device->modify_qp(qp->real_qp, qp_attr, qp_attr_mask, NULL); } EXPORT_SYMBOL(ib_modify_qp); @@ -259,11 +919,59 @@ int ib_query_qp(struct ib_qp *qp, struct ib_qp_init_attr *qp_init_attr) { return qp->device->query_qp ? - qp->device->query_qp(qp, qp_attr, qp_attr_mask, qp_init_attr) : + qp->device->query_qp(qp->real_qp, qp_attr, qp_attr_mask, qp_init_attr) : -ENOSYS; } EXPORT_SYMBOL(ib_query_qp); +int ib_close_qp(struct ib_qp *qp) +{ + struct ib_qp *real_qp; + unsigned long flags; + + real_qp = qp->real_qp; + if (real_qp == qp) + return -EINVAL; + + spin_lock_irqsave(&real_qp->device->event_handler_lock, flags); + list_del(&qp->open_list); + spin_unlock_irqrestore(&real_qp->device->event_handler_lock, flags); + + atomic_dec(&real_qp->usecnt); + kfree(qp); + + return 0; +} +EXPORT_SYMBOL(ib_close_qp); + +static int __ib_destroy_shared_qp(struct ib_qp *qp) +{ + struct ib_xrcd *xrcd; + struct ib_qp *real_qp; + int ret; + + real_qp = qp->real_qp; + xrcd = real_qp->xrcd; + + mutex_lock(&xrcd->tgt_qp_mutex); + ib_close_qp(qp); + if (atomic_read(&real_qp->usecnt) == 0) + list_del(&real_qp->xrcd_list); + else + real_qp = NULL; + mutex_unlock(&xrcd->tgt_qp_mutex); + + if (real_qp) { + ret = ib_destroy_qp(real_qp); + if (!ret) + atomic_dec(&xrcd->usecnt); + else + __ib_insert_xrcd_qp(xrcd, real_qp); + } + + return 0; +} + int ib_destroy_qp(struct ib_qp *qp) { struct ib_pd *pd; @@ -271,16 +979,25 @@ int ib_destroy_qp(struct ib_qp *qp) struct ib_srq *srq; int ret; - pd = qp->pd; - scq = qp->send_cq; - rcq = qp->recv_cq; - srq = qp->srq; + if (atomic_read(&qp->usecnt)) + return -EBUSY; + + if (qp->real_qp != qp) + return __ib_destroy_shared_qp(qp); + + pd = qp->pd; + scq = qp->send_cq; + rcq = qp->recv_cq; + srq = qp->srq; ret = qp->device->destroy_qp(qp); if (!ret) { - atomic_dec(&pd->usecnt); - atomic_dec(&scq->usecnt); - atomic_dec(&rcq->usecnt); + if (pd) + atomic_dec(&pd->usecnt); + if (scq) + atomic_dec(&scq->usecnt); + if (rcq) + atomic_dec(&rcq->usecnt); if (srq) atomic_dec(&srq->usecnt); } @@ -294,11 +1011,11 @@ EXPORT_SYMBOL(ib_destroy_qp); struct ib_cq *ib_create_cq(struct ib_device *device, ib_comp_handler comp_handler, void (*event_handler)(struct ib_event *, void *), - void *cq_context, int cqe) + void *cq_context, int cqe, int comp_vector) { struct ib_cq *cq; - cq = device->create_cq(device, cqe, NULL, NULL); + cq = device->create_cq(device, cqe, comp_vector, NULL, NULL); if (!IS_ERR(cq)) { cq->device = device; @@ -313,6 +1030,13 @@ struct ib_cq *ib_create_cq(struct ib_device *device, } EXPORT_SYMBOL(ib_create_cq); +int ib_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period) +{ + return cq->device->modify_cq ? + cq->device->modify_cq(cq, cq_count, cq_period) : -ENOSYS; +} +EXPORT_SYMBOL(ib_modify_cq); + int ib_destroy_cq(struct ib_cq *cq) { if (atomic_read(&cq->usecnt)) @@ -322,11 +1046,10 @@ int ib_destroy_cq(struct ib_cq *cq) } EXPORT_SYMBOL(ib_destroy_cq); -int ib_resize_cq(struct ib_cq *cq, - int cqe) +int ib_resize_cq(struct ib_cq *cq, int cqe) { return cq->device->resize_cq ? - cq->device->resize_cq(cq, cqe) : -ENOSYS; + cq->device->resize_cq(cq, cqe, NULL) : -ENOSYS; } EXPORT_SYMBOL(ib_resize_cq); @@ -335,6 +1058,11 @@ EXPORT_SYMBOL(ib_resize_cq); struct ib_mr *ib_get_dma_mr(struct ib_pd *pd, int mr_access_flags) { struct ib_mr *mr; + int err; + + err = ib_check_mr_access(mr_access_flags); + if (err) + return ERR_PTR(err); mr = pd->device->get_dma_mr(pd, mr_access_flags); @@ -357,6 +1085,14 @@ struct ib_mr *ib_reg_phys_mr(struct ib_pd *pd, u64 *iova_start) { struct ib_mr *mr; + int err; + + err = ib_check_mr_access(mr_access_flags); + if (err) + return ERR_PTR(err); + + if (!pd->device->reg_phys_mr) + return ERR_PTR(-ENOSYS); mr = pd->device->reg_phys_mr(pd, phys_buf_array, num_phys_buf, mr_access_flags, iova_start); @@ -384,6 +1120,10 @@ int ib_rereg_phys_mr(struct ib_mr *mr, struct ib_pd *old_pd; int ret; + ret = ib_check_mr_access(mr_access_flags); + if (ret) + return ret; + if (!mr->device->rereg_phys_mr) return -ENOSYS; @@ -429,20 +1169,106 @@ int ib_dereg_mr(struct ib_mr *mr) } EXPORT_SYMBOL(ib_dereg_mr); +struct ib_mr *ib_create_mr(struct ib_pd *pd, + struct ib_mr_init_attr *mr_init_attr) +{ + struct ib_mr *mr; + + if (!pd->device->create_mr) + return ERR_PTR(-ENOSYS); + + mr = pd->device->create_mr(pd, mr_init_attr); + + if (!IS_ERR(mr)) { + mr->device = pd->device; + mr->pd = pd; + mr->uobject = NULL; + atomic_inc(&pd->usecnt); + atomic_set(&mr->usecnt, 0); + } + + return mr; +} +EXPORT_SYMBOL(ib_create_mr); + +int ib_destroy_mr(struct ib_mr *mr) +{ + struct ib_pd *pd; + int ret; + + if (atomic_read(&mr->usecnt)) + return -EBUSY; + + pd = mr->pd; + ret = mr->device->destroy_mr(mr); + if (!ret) + atomic_dec(&pd->usecnt); + + return ret; +} +EXPORT_SYMBOL(ib_destroy_mr); + +struct ib_mr *ib_alloc_fast_reg_mr(struct ib_pd *pd, int max_page_list_len) +{ + struct ib_mr *mr; + + if (!pd->device->alloc_fast_reg_mr) + return ERR_PTR(-ENOSYS); + + mr = pd->device->alloc_fast_reg_mr(pd, max_page_list_len); + + if (!IS_ERR(mr)) { + mr->device = pd->device; + mr->pd = pd; + mr->uobject = NULL; + atomic_inc(&pd->usecnt); + atomic_set(&mr->usecnt, 0); + } + + return mr; +} +EXPORT_SYMBOL(ib_alloc_fast_reg_mr); + +struct ib_fast_reg_page_list *ib_alloc_fast_reg_page_list(struct ib_device *device, + int max_page_list_len) +{ + struct ib_fast_reg_page_list *page_list; + + if (!device->alloc_fast_reg_page_list) + return ERR_PTR(-ENOSYS); + + page_list = device->alloc_fast_reg_page_list(device, max_page_list_len); + + if (!IS_ERR(page_list)) { + page_list->device = device; + page_list->max_page_list_len = max_page_list_len; + } + + return page_list; +} +EXPORT_SYMBOL(ib_alloc_fast_reg_page_list); + +void ib_free_fast_reg_page_list(struct ib_fast_reg_page_list *page_list) +{ + page_list->device->free_fast_reg_page_list(page_list); +} +EXPORT_SYMBOL(ib_free_fast_reg_page_list); + /* Memory windows */ -struct ib_mw *ib_alloc_mw(struct ib_pd *pd) +struct ib_mw *ib_alloc_mw(struct ib_pd *pd, enum ib_mw_type type) { struct ib_mw *mw; if (!pd->device->alloc_mw) return ERR_PTR(-ENOSYS); - mw = pd->device->alloc_mw(pd); + mw = pd->device->alloc_mw(pd, type); if (!IS_ERR(mw)) { mw->device = pd->device; mw->pd = pd; mw->uobject = NULL; + mw->type = type; atomic_inc(&pd->usecnt); } @@ -516,22 +1342,106 @@ EXPORT_SYMBOL(ib_dealloc_fmr); int ib_attach_mcast(struct ib_qp *qp, union ib_gid *gid, u16 lid) { + int ret; + if (!qp->device->attach_mcast) return -ENOSYS; if (gid->raw[0] != 0xff || qp->qp_type != IB_QPT_UD) return -EINVAL; - return qp->device->attach_mcast(qp, gid, lid); + ret = qp->device->attach_mcast(qp, gid, lid); + if (!ret) + atomic_inc(&qp->usecnt); + return ret; } EXPORT_SYMBOL(ib_attach_mcast); int ib_detach_mcast(struct ib_qp *qp, union ib_gid *gid, u16 lid) { + int ret; + if (!qp->device->detach_mcast) return -ENOSYS; if (gid->raw[0] != 0xff || qp->qp_type != IB_QPT_UD) return -EINVAL; - return qp->device->detach_mcast(qp, gid, lid); + ret = qp->device->detach_mcast(qp, gid, lid); + if (!ret) + atomic_dec(&qp->usecnt); + return ret; } EXPORT_SYMBOL(ib_detach_mcast); + +struct ib_xrcd *ib_alloc_xrcd(struct ib_device *device) +{ + struct ib_xrcd *xrcd; + + if (!device->alloc_xrcd) + return ERR_PTR(-ENOSYS); + + xrcd = device->alloc_xrcd(device, NULL, NULL); + if (!IS_ERR(xrcd)) { + xrcd->device = device; + xrcd->inode = NULL; + atomic_set(&xrcd->usecnt, 0); + mutex_init(&xrcd->tgt_qp_mutex); + INIT_LIST_HEAD(&xrcd->tgt_qp_list); + } + + return xrcd; +} +EXPORT_SYMBOL(ib_alloc_xrcd); + +int ib_dealloc_xrcd(struct ib_xrcd *xrcd) +{ + struct ib_qp *qp; + int ret; + + if (atomic_read(&xrcd->usecnt)) + return -EBUSY; + + while (!list_empty(&xrcd->tgt_qp_list)) { + qp = list_entry(xrcd->tgt_qp_list.next, struct ib_qp, xrcd_list); + ret = ib_destroy_qp(qp); + if (ret) + return ret; + } + + return xrcd->device->dealloc_xrcd(xrcd); +} +EXPORT_SYMBOL(ib_dealloc_xrcd); + +struct ib_flow *ib_create_flow(struct ib_qp *qp, + struct ib_flow_attr *flow_attr, + int domain) +{ + struct ib_flow *flow_id; + if (!qp->device->create_flow) + return ERR_PTR(-ENOSYS); + + flow_id = qp->device->create_flow(qp, flow_attr, domain); + if (!IS_ERR(flow_id)) + atomic_inc(&qp->usecnt); + return flow_id; +} +EXPORT_SYMBOL(ib_create_flow); + +int ib_destroy_flow(struct ib_flow *flow_id) +{ + int err; + struct ib_qp *qp = flow_id->qp; + + err = qp->device->destroy_flow(flow_id); + if (!err) + atomic_dec(&qp->usecnt); + return err; +} +EXPORT_SYMBOL(ib_destroy_flow); + +int ib_check_mr_status(struct ib_mr *mr, u32 check_mask, + struct ib_mr_status *mr_status) +{ + return mr->device->check_mr_status ? + mr->device->check_mr_status(mr, check_mask, mr_status) : -ENOSYS; +} +EXPORT_SYMBOL(ib_check_mr_status); |