mad.c

h内核

				break;
			}
		}
	} else {
		struct ib_mad_mgmt_class_table *class;
		struct ib_mad_mgmt_method_table *method;
		struct ib_mad_mgmt_vendor_class_table *vendor;
		struct ib_mad_mgmt_vendor_class *vendor_class;
		struct ib_vendor_mad *vendor_mad;
		int index;

		/*
		 * Routing is based on version, class, and method
		 * For "newer" vendor MADs, also based on OUI
		 */
		if (mad->mad_hdr.class_version >= MAX_MGMT_VERSION)
			goto out;
		if (!is_vendor_class(mad->mad_hdr.mgmt_class)) {
			class = port_priv->version[
					mad->mad_hdr.class_version].class;
			if (!class)
				goto out;
			method = class->method_table[convert_mgmt_class(
							mad->mad_hdr.mgmt_class)];
			if (method)
				mad_agent = method->agent[mad->mad_hdr.method &
							  ~IB_MGMT_METHOD_RESP];
		} else {
			vendor = port_priv->version[
					mad->mad_hdr.class_version].vendor;
			if (!vendor)
				goto out;
			vendor_class = vendor->vendor_class[vendor_class_index(
						mad->mad_hdr.mgmt_class)];
			if (!vendor_class)
				goto out;
			/* Find matching OUI */
			vendor_mad = (struct ib_vendor_mad *)mad;
			index = find_vendor_oui(vendor_class, vendor_mad->oui);
			if (index == -1)
				goto out;
			method = vendor_class->method_table[index];
			if (method) {
				mad_agent = method->agent[mad->mad_hdr.method &
							  ~IB_MGMT_METHOD_RESP];
			}
		}
	}

	if (mad_agent) {
		if (mad_agent->agent.recv_handler)
			atomic_inc(&mad_agent->refcount);
		else {
			printk(KERN_NOTICE PFX "No receive handler for client "
			       "%p on port %d\n",
			       &mad_agent->agent, port_priv->port_num);
			mad_agent = NULL;
		}
	}
out:
	spin_unlock_irqrestore(&port_priv->reg_lock, flags);

	return mad_agent;
}

static int validate_mad(struct ib_mad *mad, u32 qp_num)
{
	int valid = 0;

	/* Make sure MAD base version is understood */
	if (mad->mad_hdr.base_version != IB_MGMT_BASE_VERSION) {
		printk(KERN_ERR PFX "MAD received with unsupported base "
		       "version %d\n", mad->mad_hdr.base_version);
		goto out;
	}

	/* Filter SMI packets sent to other than QP0 */
	if ((mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED) ||
	    (mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)) {
		if (qp_num == 0)
			valid = 1;
	} else {
		/* Filter GSI packets sent to QP0 */
		if (qp_num != 0)
			valid = 1;
	}

out:
	return valid;
}

/*
 * Return start of fully reassembled MAD, or NULL, if MAD isn't assembled yet
 */
static struct ib_mad_private *
reassemble_recv(struct ib_mad_agent_private *mad_agent_priv,
		struct ib_mad_private *recv)
{
	/* Until we have RMPP, all receives are reassembled!... */
	INIT_LIST_HEAD(&recv->header.recv_wc.recv_buf.list);
	return recv;
}

static struct ib_mad_send_wr_private*
find_send_req(struct ib_mad_agent_private *mad_agent_priv,
	      u64 tid)
{
	struct ib_mad_send_wr_private *mad_send_wr;

	list_for_each_entry(mad_send_wr, &mad_agent_priv->wait_list,
			    agent_list) {
		if (mad_send_wr->tid == tid)
			return mad_send_wr;
	}

	/*
	 * 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 (mad_send_wr->tid == tid && mad_send_wr->timeout) {
			/* Verify request has not been canceled */
			return (mad_send_wr->status == IB_WC_SUCCESS) ?
				mad_send_wr : NULL;
		}
	}
	return NULL;
}

static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
				 struct ib_mad_private *recv,
				 int solicited)
{
	struct ib_mad_send_wr_private *mad_send_wr;
	struct ib_mad_send_wc mad_send_wc;
	unsigned long flags;

	/* Fully reassemble receive before processing */
	recv = reassemble_recv(mad_agent_priv, recv);
	if (!recv) {
		if (atomic_dec_and_test(&mad_agent_priv->refcount))
			wake_up(&mad_agent_priv->wait);
		return;
	}

	/* Complete corresponding request */
	if (solicited) {
		spin_lock_irqsave(&mad_agent_priv->lock, flags);
		mad_send_wr = find_send_req(mad_agent_priv,
					    recv->mad.mad.mad_hdr.tid);
		if (!mad_send_wr) {
			spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
			ib_free_recv_mad(&recv->header.recv_wc);
			if (atomic_dec_and_test(&mad_agent_priv->refcount))
				wake_up(&mad_agent_priv->wait);
			return;
		}
		/* Timeout = 0 means that we won't wait for a response */
		mad_send_wr->timeout = 0;
		spin_unlock_irqrestore(&mad_agent_priv->lock, flags);

		/* Defined behavior is to complete response before request */
		recv->header.recv_wc.wc->wr_id = mad_send_wr->wr_id;
		mad_agent_priv->agent.recv_handler(
						&mad_agent_priv->agent,
						&recv->header.recv_wc);
		atomic_dec(&mad_agent_priv->refcount);

		mad_send_wc.status = IB_WC_SUCCESS;
		mad_send_wc.vendor_err = 0;
		mad_send_wc.wr_id = mad_send_wr->wr_id;
		ib_mad_complete_send_wr(mad_send_wr, &mad_send_wc);
	} else {
		mad_agent_priv->agent.recv_handler(
						&mad_agent_priv->agent,
						&recv->header.recv_wc);
		if (atomic_dec_and_test(&mad_agent_priv->refcount))
			wake_up(&mad_agent_priv->wait);
	}
}

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_list_head *mad_list;
	struct ib_mad_agent_private *mad_agent;
	int solicited;

	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");

	mad_list = (struct ib_mad_list_head *)(unsigned long)wc->wr_id;
	qp_info = mad_list->mad_queue->qp_info;
	dequeue_mad(mad_list);

	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);

	/* Setup MAD receive work completion from "normal" work completion */
	recv->header.recv_wc.wc = wc;
	recv->header.recv_wc.mad_len = sizeof(struct ib_mad);
	recv->header.recv_wc.recv_buf.mad = &recv->mad.mad;
	recv->header.recv_wc.recv_buf.grh = &recv->grh;

	if (atomic_read(&qp_info->snoop_count))
		snoop_recv(qp_info, &recv->header.recv_wc, IB_MAD_SNOOP_RECVS);

	/* Validate MAD */
	if (!validate_mad(&recv->mad.mad, qp_info->qp->qp_num))
		goto out;

	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))
			goto out;
		if (!smi_check_forward_dr_smp(&recv->mad.smp))
			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,
					    port_priv->device,
					    port_priv->port_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,
						     &recv->mad.mad,
						     &response->mad.mad);
		if (ret & IB_MAD_RESULT_SUCCESS) {
			if (ret & IB_MAD_RESULT_CONSUMED)
				goto out;
			if (ret & IB_MAD_RESULT_REPLY) {
				/* Send response */
				if (!agent_send(response, &recv->grh, wc,
						port_priv->device,
						port_priv->port_num))
					response = NULL;
				goto out;
			}
		}
	}

	/* Determine corresponding MAD agent for incoming receive MAD */
	solicited = solicited_mad(&recv->mad.mad);
	mad_agent = find_mad_agent(port_priv, &recv->mad.mad, solicited);
	if (mad_agent) {
		ib_mad_complete_recv(mad_agent, recv, solicited);
		/*
		 * recv is freed up in error cases in ib_mad_complete_recv
		 * or via recv_handler in ib_mad_complete_recv()
		 */
		recv = NULL;
	}

out:
	/* Post another receive request for this QP */
	if (response) {
		ib_mad_post_receive_mads(qp_info, response);
		if (recv)
			kmem_cache_free(ib_mad_cache, recv);
	} else
		ib_mad_post_receive_mads(qp_info, recv);
}

static void adjust_timeout(struct ib_mad_agent_private *mad_agent_priv)
{
	struct ib_mad_send_wr_private *mad_send_wr;
	unsigned long delay;

	if (list_empty(&mad_agent_priv->wait_list)) {
		cancel_delayed_work(&mad_agent_priv->timed_work);
	} else {
		mad_send_wr = list_entry(mad_agent_priv->wait_list.next,
					 struct ib_mad_send_wr_private,
					 agent_list);

		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);
		}
	}
}

static void wait_for_response(struct ib_mad_agent_private *mad_agent_priv,
			      struct ib_mad_send_wr_private *mad_send_wr )
{
	struct ib_mad_send_wr_private *temp_mad_send_wr;
	struct list_head *list_item;
	unsigned long delay;

	list_del(&mad_send_wr->agent_list);

	delay = mad_send_wr->timeout;
	mad_send_wr->timeout += jiffies;

	list_for_each_prev(list_item, &mad_agent_priv->wait_list) {
		temp_mad_send_wr = list_entry(list_item,
					      struct ib_mad_send_wr_private,
					      agent_list);
		if (time_after(mad_send_wr->timeout,
			       temp_mad_send_wr->timeout))
			break;
	}
	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);
	}
}

/*
 * Process a send work completion
 */
static void ib_mad_complete_send_wr(struct ib_mad_send_wr_private *mad_send_wr,
				    struct ib_mad_send_wc *mad_send_wc)
{
	struct ib_mad_agent_private	*mad_agent_priv;
	unsigned long			flags;

	mad_agent_priv = container_of(mad_send_wr->agent,
				      struct ib_mad_agent_private, agent);

	spin_lock_irqsave(&mad_agent_priv->lock, flags);
	if (mad_send_wc->status != IB_WC_SUCCESS &&
	    mad_send_wr->status == IB_WC_SUCCESS) {
		mad_send_wr->status = mad_send_wc->status;
		mad_send_wr->refcount -= (mad_send_wr->timeout > 0);
	}

	if (--mad_send_wr->refcount > 0) {
		if (mad_send_wr->refcount == 1 && mad_send_wr->timeout &&
		    mad_send_wr->status == IB_WC_SUCCESS) {
			wait_for_response(mad_agent_priv, mad_send_wr);
		}
		spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
		return;
	}

	/* Remove send from MAD agent and notify client of completion */
	list_del(&mad_send_wr->agent_list);
	adjust_timeout(mad_agent_priv);
	spin_unlock_irqrestore(&mad_agent_priv->lock, flags);

	if (mad_send_wr->status != IB_WC_SUCCESS )
		mad_send_wc->status = mad_send_wr->status;
	mad_agent_priv->agent.send_handler(&mad_agent_priv->agent,
					    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);

	kfree(mad_send_wr);
}

static void ib_mad_send_done_handler(struct ib_mad_port_private *port_priv,
				     struct ib_wc *wc)
{
	struct ib_mad_send_wr_private	*mad_send_wr, *queued_send_wr;
	struct ib_mad_list_head		*mad_list;
	struct ib_mad_qp_info		*qp_info;
	struct ib_mad_queue		*send_queue;
	struct ib_send_wr		*bad_send_wr;
	unsigned long flags;
	int ret;

	mad_list = (struct ib_mad_list_head *)(unsigned long)wc->wr_id;
	mad_send_wr = container_of(mad_list, struct ib_mad_send_wr_private,
				   mad_list);
	send_queue = mad_list->mad_queue;
	qp_info = send_queue->qp_info;

retry:
	queued_send_wr = NULL;
	spin_lock_irqsave(&send_queue->lock, flags);
	list_del(&mad_list->list);

	/* Move queued send to the send queue */
	if (send_queue->count-- > send_queue->max_active) {
		mad_list = container_of(qp_info->overflow_list.next,
					struct ib_mad_list_head, list);
		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);
	}
	spin_unlock_irqrestore(&send_queue->lock, flags);

	/* Restore client wr_id in WC and complete send */
	wc->wr_id = mad_send_wr->wr_id;
	if (atomic_read(&qp_info->snoop_count))
		snoop_send(qp_info, &mad_send_wr->send_wr,
			   (struct ib_mad_send_wc *)wc,
			   IB_MAD_SNOOP_SEND_COMPLETIONS);
	ib_mad_complete_send_wr(mad_send_wr, (struct ib_mad_send_wc *)wc);

	if (queued_send_wr) {
		ret = ib_post_send(qp_info->qp, &queued_send_wr->send_wr,
				&bad_send_wr);
		if (ret) {
			printk(KERN_ERR PFX "ib_post_send failed: %d\n", ret);
			mad_send_wr = queued_send_wr;
			wc->status = IB_WC_LOC_QP_OP_ERR;
			goto retry;
		}
	}
}

static void mark_sends_for_retry(struct ib_mad_qp_info *qp_info)
{
	struct ib_mad_send_wr_private *mad_send_wr;
	struct ib_mad_list_head *mad_list;
	unsigned long flags;

	spin_lock_irqsave(&qp_info->send_queue.lock, flags);
	list_for_each_entry(mad_list, &qp_info->send_queue.list, list) {
		mad_send_wr = container_of(mad_list,
					   struct ib_mad_send_wr_private,
					   mad_list);