cm.c

linux 内核源代码

			node = node->rb_left;
		else if (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;
	}
	return NULL;
}

static struct cm_timewait_info * cm_insert_remote_id(struct cm_timewait_info
						     *timewait_info)
{
	struct rb_node **link = &cm.remote_id_table.rb_node;
	struct rb_node *parent = NULL;
	struct cm_timewait_info *cur_timewait_info;
	__be64 remote_ca_guid = timewait_info->remote_ca_guid;
	__be32 remote_id = timewait_info->work.remote_id;

	while (*link) {
		parent = *link;
		cur_timewait_info = rb_entry(parent, struct cm_timewait_info,
					     remote_id_node);
		if (remote_id < cur_timewait_info->work.remote_id)
			link = &(*link)->rb_left;
		else if (remote_id > cur_timewait_info->work.remote_id)
			link = &(*link)->rb_right;
		else if (remote_ca_guid < cur_timewait_info->remote_ca_guid)
			link = &(*link)->rb_left;
		else if (remote_ca_guid > cur_timewait_info->remote_ca_guid)
			link = &(*link)->rb_right;
		else
			return cur_timewait_info;
	}
	timewait_info->inserted_remote_id = 1;
	rb_link_node(&timewait_info->remote_id_node, parent, link);
	rb_insert_color(&timewait_info->remote_id_node, &cm.remote_id_table);
	return NULL;
}

static struct cm_timewait_info * cm_find_remote_id(__be64 remote_ca_guid,
						   __be32 remote_id)
{
	struct rb_node *node = cm.remote_id_table.rb_node;
	struct cm_timewait_info *timewait_info;

	while (node) {
		timewait_info = rb_entry(node, struct cm_timewait_info,
					 remote_id_node);
		if (remote_id < timewait_info->work.remote_id)
			node = node->rb_left;
		else if (remote_id > timewait_info->work.remote_id)
			node = node->rb_right;
		else if (remote_ca_guid < timewait_info->remote_ca_guid)
			node = node->rb_left;
		else if (remote_ca_guid > timewait_info->remote_ca_guid)
			node = node->rb_right;
		else
			return timewait_info;
	}
	return NULL;
}

static struct cm_timewait_info * cm_insert_remote_qpn(struct cm_timewait_info
						      *timewait_info)
{
	struct rb_node **link = &cm.remote_qp_table.rb_node;
	struct rb_node *parent = NULL;
	struct cm_timewait_info *cur_timewait_info;
	__be64 remote_ca_guid = timewait_info->remote_ca_guid;
	__be32 remote_qpn = timewait_info->remote_qpn;

	while (*link) {
		parent = *link;
		cur_timewait_info = rb_entry(parent, struct cm_timewait_info,
					     remote_qp_node);
		if (remote_qpn < cur_timewait_info->remote_qpn)
			link = &(*link)->rb_left;
		else if (remote_qpn > cur_timewait_info->remote_qpn)
			link = &(*link)->rb_right;
		else if (remote_ca_guid < cur_timewait_info->remote_ca_guid)
			link = &(*link)->rb_left;
		else if (remote_ca_guid > cur_timewait_info->remote_ca_guid)
			link = &(*link)->rb_right;
		else
			return cur_timewait_info;
	}
	timewait_info->inserted_remote_qp = 1;
	rb_link_node(&timewait_info->remote_qp_node, parent, link);
	rb_insert_color(&timewait_info->remote_qp_node, &cm.remote_qp_table);
	return NULL;
}

static struct cm_id_private * cm_insert_remote_sidr(struct cm_id_private
						    *cm_id_priv)
{
	struct rb_node **link = &cm.remote_sidr_table.rb_node;
	struct rb_node *parent = NULL;
	struct cm_id_private *cur_cm_id_priv;
	union ib_gid *port_gid = &cm_id_priv->av.dgid;
	__be32 remote_id = cm_id_priv->id.remote_id;

	while (*link) {
		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)
			link = &(*link)->rb_left;
		else if (remote_id > cur_cm_id_priv->id.remote_id)
			link = &(*link)->rb_right;
		else {
			int cmp;
			cmp = memcmp(port_gid, &cur_cm_id_priv->av.dgid,
				     sizeof *port_gid);
			if (cmp < 0)
				link = &(*link)->rb_left;
			else if (cmp > 0)
				link = &(*link)->rb_right;
			else
				return cur_cm_id_priv;
		}
	}
	rb_link_node(&cm_id_priv->sidr_id_node, parent, link);
	rb_insert_color(&cm_id_priv->sidr_id_node, &cm.remote_sidr_table);
	return NULL;
}

static void cm_reject_sidr_req(struct cm_id_private *cm_id_priv,
			       enum ib_cm_sidr_status status)
{
	struct ib_cm_sidr_rep_param param;

	memset(&param, 0, sizeof param);
	param.status = status;
	ib_send_cm_sidr_rep(&cm_id_priv->id, &param);
}

struct ib_cm_id *ib_create_cm_id(struct ib_device *device,
				 ib_cm_handler cm_handler,
				 void *context)
{
	struct cm_id_private *cm_id_priv;
	int ret;

	cm_id_priv = kzalloc(sizeof *cm_id_priv, GFP_KERNEL);
	if (!cm_id_priv)
		return ERR_PTR(-ENOMEM);

	cm_id_priv->id.state = IB_CM_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.remote_cm_qpn = 1;
	ret = cm_alloc_id(cm_id_priv);
	if (ret)
		goto error;

	spin_lock_init(&cm_id_priv->lock);
	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);
	return &cm_id_priv->id;

error:
	kfree(cm_id_priv);
	return ERR_PTR(-ENOMEM);
}
EXPORT_SYMBOL(ib_create_cm_id);

static struct cm_work * cm_dequeue_work(struct cm_id_private *cm_id_priv)
{
	struct cm_work *work;

	if (list_empty(&cm_id_priv->work_list))
		return NULL;

	work = list_entry(cm_id_priv->work_list.next, struct cm_work, list);
	list_del(&work->list);
	return work;
}

static void cm_free_work(struct cm_work *work)
{
	if (work->mad_recv_wc)
		ib_free_recv_mad(work->mad_recv_wc);
	kfree(work);
}

static inline int cm_convert_to_ms(int iba_time)
{
	/* approximate conversion to ms from 4.096us x 2^iba_time */
	return 1 << max(iba_time - 8, 0);
}

/*
 * 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)
{
	int ack_timeout = packet_life_time + 1;

	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));

	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;
	}

	if (timewait_info->inserted_remote_qp) {
		rb_erase(&timewait_info->remote_qp_node, &cm.remote_qp_table);
		timewait_info->inserted_remote_qp = 0;
	}
}

static struct cm_timewait_info * cm_create_timewait_info(__be32 local_id)
{
	struct cm_timewait_info *timewait_info;

	timewait_info = kzalloc(sizeof *timewait_info, GFP_KERNEL);
	if (!timewait_info)
		return ERR_PTR(-ENOMEM);

	timewait_info->work.local_id = local_id;
	INIT_DELAYED_WORK(&timewait_info->work.work, cm_work_handler);
	timewait_info->work.cm_event.event = IB_CM_TIMEWAIT_EXIT;
	return timewait_info;
}

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.
	 * To protect against this, we search for the cm_id after exiting
	 * 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->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;
}

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;
	}
}

static void cm_destroy_id(struct ib_cm_id *cm_id, int err)
{
	struct cm_id_private *cm_id_priv;
	struct cm_work *work;

	cm_id_priv = container_of(cm_id, struct cm_id_private, id);
retest:
	spin_lock_irq(&cm_id_priv->lock);
	switch (cm_id->state) {
	case IB_CM_LISTEN:
		cm_id->state = IB_CM_IDLE;
		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_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_irq(&cm_id_priv->lock);
		break;
	case IB_CM_SIDR_REQ_RCVD:
		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_irq(&cm_id_priv->lock);
		ib_send_cm_rej(cm_id, IB_CM_REJ_TIMEOUT,
			       &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_MRA_REQ_SENT:
	case IB_CM_REP_RCVD:
	case IB_CM_MRA_REP_SENT:
		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_irq(&cm_id_priv->lock);
		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_irq(&cm_id_priv->lock);
		break;
	case IB_CM_DREQ_RCVD:
		spin_unlock_irq(&cm_id_priv->lock);
		ib_send_cm_drep(cm_id, NULL, 0);
		break;
	default:
		spin_unlock_irq(&cm_id_priv->lock);
		break;
	}

	cm_free_id(cm_id->local_id);
	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);
	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,
		 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_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);
	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);
	} else {
		cm_id->service_id = service_id;
		cm_id->service_mask = service_mask;
	}
	cur_cm_id_priv = cm_insert_listen(cm_id_priv);
	spin_unlock_irqrestore(&cm.lock, flags);

	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;
}
EXPORT_SYMBOL(ib_cm_listen);

static __be64 cm_form_tid(struct cm_id_private *cm_id_priv,
			  enum cm_msg_sequence msg_seq)
{
	u64 hi_tid, low_tid;

	hi_tid   = ((u64) cm_id_priv->av.port->mad_agent->hi_tid) << 32;
	low_tid  = (u64) ((__force u32)cm_id_priv->id.local_id |
			  (msg_seq << 30));
	return cpu_to_be64(hi_tid | low_tid);
}

static void cm_format_mad_hdr(struct ib_mad_hdr *hdr,
			      __be16 attr_id, __be64 tid)
{
	hdr->base_version  = IB_MGMT_BASE_VERSION;
	hdr->mgmt_class	   = IB_MGMT_CLASS_CM;
	hdr->class_version = IB_CM_CLASS_VERSION;
	hdr->method	   = IB_MGMT_METHOD_SEND;
	hdr->attr_id	   = attr_id;
	hdr->tid	   = tid;
}

static void cm_format_req(struct cm_req_msg *req_msg,
			  struct cm_id_private *cm_id_priv,
			  struct ib_cm_req_param *param)
{
	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;