cma.c

linux 内核源代码

	if (cma_get_net_info(ib_event->private_data, listen_id->ps,
			     &ip_ver, &port, &src, &dst))
		goto err;

	cma_save_net_info(&id->route.addr, &listen_id->route.addr,
			  ip_ver, port, src, dst);

	ret = rdma_translate_ip(&id->route.addr.src_addr,
				&id->route.addr.dev_addr);
	if (ret)
		goto err;

	id_priv = container_of(id, struct rdma_id_private, id);
	id_priv->state = CMA_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_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_disable_remove(listen_id, CMA_LISTEN))
		return -ECONNABORTED;

	memset(&event, 0, sizeof event);
	offset = cma_user_data_offset(listen_id->id.ps);
	event.event = RDMA_CM_EVENT_CONNECT_REQUEST;
	if (cma_is_ud_ps(listen_id->id.ps)) {
		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 {
		ib_send_cm_mra(cm_id, CMA_CM_MRA_SETTING, NULL, 0);
		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 out;
	}

	atomic_inc(&conn_id->dev_remove);
	mutex_lock(&lock);
	ret = cma_acquire_dev(conn_id);
	mutex_unlock(&lock);
	if (ret)
		goto release_conn_id;

	conn_id->cm_id.ib = cm_id;
	cm_id->context = conn_id;
	cm_id->cm_handler = cma_ib_handler;

	ret = conn_id->id.event_handler(&conn_id->id, &event);
	if (!ret)
		goto out;

	/* Destroy the CM ID by returning a non-zero value. */
	conn_id->cm_id.ib = NULL;

release_conn_id:
	cma_exch(conn_id, CMA_DESTROYING);
	cma_enable_remove(conn_id);
	rdma_destroy_id(&conn_id->id);

out:
	cma_enable_remove(listen_id);
	return ret;
}

static __be64 cma_get_service_id(enum rdma_port_space ps, struct sockaddr *addr)
{
	return cpu_to_be64(((u64)ps << 16) + be16_to_cpu(cma_port(addr)));
}

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;
	struct sdp_hh *sdp_data, *sdp_mask;
	__u32 ip4_addr;
	struct in6_addr ip6_addr;

	memset(compare, 0, sizeof *compare);
	cma_data = (void *) compare->data;
	cma_mask = (void *) compare->mask;
	sdp_data = (void *) compare->data;
	sdp_mask = (void *) compare->mask;

	switch (addr->sa_family) {
	case AF_INET:
		ip4_addr = ((struct sockaddr_in *) addr)->sin_addr.s_addr;
		if (ps == RDMA_PS_SDP) {
			sdp_set_ip_ver(sdp_data, 4);
			sdp_set_ip_ver(sdp_mask, 0xF);
			sdp_data->dst_addr.ip4.addr = ip4_addr;
			sdp_mask->dst_addr.ip4.addr = ~0;
		} else {
			cma_set_ip_ver(cma_data, 4);
			cma_set_ip_ver(cma_mask, 0xF);
			cma_data->dst_addr.ip4.addr = ip4_addr;
			cma_mask->dst_addr.ip4.addr = ~0;
		}
		break;
	case AF_INET6:
		ip6_addr = ((struct sockaddr_in6 *) addr)->sin6_addr;
		if (ps == RDMA_PS_SDP) {
			sdp_set_ip_ver(sdp_data, 6);
			sdp_set_ip_ver(sdp_mask, 0xF);
			sdp_data->dst_addr.ip6 = ip6_addr;
			memset(&sdp_mask->dst_addr.ip6, 0xFF,
			       sizeof sdp_mask->dst_addr.ip6);
		} else {
			cma_set_ip_ver(cma_data, 6);
			cma_set_ip_ver(cma_mask, 0xF);
			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;
	struct sockaddr_in *sin;
	int ret = 0;

	if (cma_disable_remove(id_priv, CMA_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:
		sin = (struct sockaddr_in *) &id_priv->id.route.addr.src_addr;
		*sin = iw_event->local_addr;
		sin = (struct sockaddr_in *) &id_priv->id.route.addr.dst_addr;
		*sin = iw_event->remote_addr;
		switch (iw_event->status) {
		case 0:
			event.event = RDMA_CM_EVENT_ESTABLISHED;
			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;
		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, CMA_DESTROYING);
		cma_enable_remove(id_priv);
		rdma_destroy_id(&id_priv->id);
		return ret;
	}

	cma_enable_remove(id_priv);
	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 sockaddr_in *sin;
	struct net_device *dev = NULL;
	struct rdma_cm_event event;
	int ret;

	listen_id = cm_id->context;
	if (cma_disable_remove(listen_id, CMA_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);
	if (!new_cm_id) {
		ret = -ENOMEM;
		goto out;
	}
	conn_id = container_of(new_cm_id, struct rdma_id_private, id);
	atomic_inc(&conn_id->dev_remove);
	conn_id->state = CMA_CONNECT;

	dev = ip_dev_find(iw_event->local_addr.sin_addr.s_addr);
	if (!dev) {
		ret = -EADDRNOTAVAIL;
		cma_enable_remove(conn_id);
		rdma_destroy_id(new_cm_id);
		goto out;
	}
	ret = rdma_copy_addr(&conn_id->id.route.addr.dev_addr, dev, NULL);
	if (ret) {
		cma_enable_remove(conn_id);
		rdma_destroy_id(new_cm_id);
		goto out;
	}

	mutex_lock(&lock);
	ret = cma_acquire_dev(conn_id);
	mutex_unlock(&lock);
	if (ret) {
		cma_enable_remove(conn_id);
		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;

	sin = (struct sockaddr_in *) &new_cm_id->route.addr.src_addr;
	*sin = iw_event->local_addr;
	sin = (struct sockaddr_in *) &new_cm_id->route.addr.dst_addr;
	*sin = iw_event->remote_addr;

	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;
	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, CMA_DESTROYING);
		cma_enable_remove(conn_id);
		rdma_destroy_id(&conn_id->id);
	}

out:
	if (dev)
		dev_put(dev);
	cma_enable_remove(listen_id);
	return ret;
}

static int cma_ib_listen(struct rdma_id_private *id_priv)
{
	struct ib_cm_compare_data compare_data;
	struct sockaddr *addr;
	__be64 svc_id;
	int ret;

	id_priv->cm_id.ib = ib_create_cm_id(id_priv->id.device, cma_req_handler,
					    id_priv);
	if (IS_ERR(id_priv->cm_id.ib))
		return PTR_ERR(id_priv->cm_id.ib);

	addr = &id_priv->id.route.addr.src_addr;
	svc_id = cma_get_service_id(id_priv->id.ps, addr);
	if (cma_any_addr(addr))
		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 sockaddr_in *sin;

	id_priv->cm_id.iw = iw_create_cm_id(id_priv->id.device,
					    iw_conn_req_handler,
					    id_priv);
	if (IS_ERR(id_priv->cm_id.iw))
		return PTR_ERR(id_priv->cm_id.iw);

	sin = (struct sockaddr_in *) &id_priv->id.route.addr.src_addr;
	id_priv->cm_id.iw->local_addr = *sin;

	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;

	id = rdma_create_id(cma_listen_handler, id_priv, id_priv->id.ps);
	if (IS_ERR(id))
		return;

	dev_id_priv = container_of(id, struct rdma_id_private, id);

	dev_id_priv->state = CMA_ADDR_BOUND;
	memcpy(&id->route.addr.src_addr, &id_priv->id.route.addr.src_addr,
	       ip_addr_size(&id_priv->id.route.addr.src_addr));

	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;

	ret = rdma_listen(id, id_priv->backlog);
	if (ret)
		printk(KERN_WARNING "RDMA CMA: cma_listen_on_dev, error %d, "
		       "listening on device %s", 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);
}

static int cma_bind_any(struct rdma_cm_id *id, sa_family_t af)
{
	struct sockaddr_in addr_in;

	memset(&addr_in, 0, sizeof addr_in);
	addr_in.sin_family = af;
	return rdma_bind_addr(id, (struct sockaddr *) &addr_in);
}

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 == CMA_IDLE) {
		ret = cma_bind_any(id, AF_INET);
		if (ret)
			return ret;
	}

	if (!cma_comp_exch(id_priv, CMA_ADDR_BOUND, CMA_LISTEN))
		return -EINVAL;

	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, CMA_LISTEN, CMA_ADDR_BOUND);
	return ret;
}
EXPORT_SYMBOL(rdma_listen);

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 = CMA_ROUTE_QUERY;
		work->new_state = CMA_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_addr *addr = &id_priv->id.route.addr;
	struct ib_sa_path_rec path_rec;
	ib_sa_comp_mask comp_mask;
	struct sockaddr_in6 *sin6;

	memset(&path_rec, 0, sizeof path_rec);
	ib_addr_get_sgid(&addr->dev_addr, &path_rec.sgid);
	ib_addr_get_dgid(&addr->dev_addr, &path_rec.dgid);
	path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(&addr->dev_addr));
	path_rec.numb_path = 1;
	path_rec.reversible = 1;
	path_rec.service_id = cma_get_service_id(id_priv->id.ps, &addr->dst_addr);

	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;

	if (addr->src_addr.sa_family == AF_INET) {
		path_rec.qos_class = cpu_to_be16((u16) id_priv->tos);
		comp_mask |= IB_SA_PATH_REC_QOS_CLASS;
	} else {
		sin6 = (struct sockaddr_in6 *) &addr->src_addr;
		path_rec.traffic_class = (u8) (be32_to_cpu(sin6->sin6_flowinfo) >> 20);
		comp_mask |= IB_SA_PATH_REC_TRAFFIC_CLASS;
	}

	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;

	atomic_inc(&id_priv->dev_remove);
	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, CMA_DESTROYING);
		destroy = 1;
	}