IB/mad: include GID/class when matching receives

Received responses are currently matched against sent requests based
on TID only.  According to the spec, responses should match based on
the combination of TID, management class, and requester LID/GID.

Without the additional qualification, an agent that is responding to
two requests, both of which have the same TID, can match RMPP ACKs
with the incorrect transaction.  This problem can occur on the SM node
when responding to SA queries.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

drivers/infiniband/core/mad.c

@@ -1618,14 +1618,59 @@ 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_send_wr_private *wr,
+				   struct ib_mad_recv_wc *rwc )
+{
+	struct ib_ah_attr attr;
+	u8 send_resp, rcv_resp;
+
+	send_resp = ((struct ib_mad *)(wr->send_buf.mad))->
+		     mad_hdr.method & IB_MGMT_METHOD_RESP;
+	rcv_resp = rwc->recv_buf.mad->mad_hdr.method & IB_MGMT_METHOD_RESP;
+
+	if (!send_resp && rcv_resp)
+		/* is request/response. GID/LIDs are both local (same). */
+		return 1;
+
+	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))
+		return attr.dlid == rwc->wc->slid;
+	else if ((attr.ah_flags & IB_AH_GRH) &&
+		 (rwc->wc->wc_flags & IB_WC_GRH))
+		return memcmp(attr.grh.dgid.raw,
+			      rwc->recv_buf.grh->sgid.raw, 16) == 0;
+	else
+		/* one has GID, other does not.  Assume different */
+		return 0;
+}
 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)
 {
 	struct ib_mad_send_wr_private *mad_send_wr;
+	struct ib_mad *mad;
+
+	mad = (struct ib_mad *)mad_recv_wc->recv_buf.mad;
 
 	list_for_each_entry(mad_send_wr, &mad_agent_priv->wait_list,
 			    agent_list) {
-		if (mad_send_wr->tid == tid)
+		if ((mad_send_wr->tid == mad->mad_hdr.tid) &&
+		    rcv_has_same_class(mad_send_wr, mad_recv_wc) &&
+		    rcv_has_same_gid(mad_send_wr, mad_recv_wc))
 			return mad_send_wr;
 	}
 
@@ -1636,7 +1681,10 @@ ib_find_send_mad(struct ib_mad_agent_private *mad_agent_priv, __be64 tid)
 	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) {
+		    mad_send_wr->tid == mad->mad_hdr.tid &&
+		    mad_send_wr->timeout &&
+		    rcv_has_same_class(mad_send_wr, mad_recv_wc) &&
+		    rcv_has_same_gid(mad_send_wr, mad_recv_wc)) {
 			/* Verify request has not been canceled */
 			return (mad_send_wr->status == IB_WC_SUCCESS) ?
 				mad_send_wr : NULL;
@@ -1661,7 +1709,6 @@ 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);
@@ -1677,9 +1724,8 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
 
 	/* Complete corresponding request */
 	if (response_mad(mad_recv_wc->recv_buf.mad)) {
-		tid = mad_recv_wc->recv_buf.mad->mad_hdr.tid;
 		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);

drivers/infiniband/core/mad_priv.h

@@ -216,7 +216,8 @@ 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);

drivers/infiniband/core/mad_rmpp.c

@@ -562,15 +562,15 @@ static int send_next_seg(struct ib_mad_send_wr_private *mad_send_wr)
 	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 */
 
@@ -612,8 +612,7 @@ 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;
 	}
@@ -621,14 +620,13 @@ 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);
+	mad_send_wr = ib_find_send_mad(agent, mad_recv_wc);
 	if (!mad_send_wr)
 		goto out;	/* Unmatched ACK */
 
@@ -639,8 +637,7 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent,
 	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;
 	}
@@ -728,12 +725,10 @@ 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,
@@ -745,12 +740,10 @@ 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 *
@@ -764,8 +757,7 @@ 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;
 	}
@@ -783,8 +775,7 @@ 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;
 	}