Merge tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband

Pull InfiniBand/RDMA updates from Roland Dreier:

 - IPoIB fixes from Doug Ledford and Erez Shitrit

 - iSER updates from Sagi Grimberg

 - mlx4 GUID handling changes from Yishai Hadas

 - other misc fixes

* tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (51 commits)
  mlx5: wrong page mask if CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for 32Bit architectures
  IB/iser: Rewrite bounce buffer code path
  IB/iser: Bump version to 1.6
  IB/iser: Remove code duplication for a single DMA entry
  IB/iser: Pass struct iser_mem_reg to iser_fast_reg_mr and iser_reg_sig_mr
  IB/iser: Modify struct iser_mem_reg members
  IB/iser: Make fastreg pool cache friendly
  IB/iser: Move PI context alloc/free to routines
  IB/iser: Move fastreg descriptor pool get/put to helper functions
  IB/iser: Merge build page-vec into register page-vec
  IB/iser: Get rid of struct iser_rdma_regd
  IB/iser: Remove redundant assignments in iser_reg_page_vec
  IB/iser: Move memory reg/dereg routines to iser_memory.c
  IB/iser: Don't pass ib_device to fall_to_bounce_buff routine
  IB/iser: Remove a redundant struct iser_data_buf
  IB/iser: Remove redundant cmd_data_len calculation
  IB/iser: Fix wrong calculation of protection buffer length
  IB/iser: Handle fastreg/local_inv completion errors
  IB/iser: Fix unload during ep_poll wrong dereference
  ib_srpt: convert printk's to pr_* functions
  ...

6 files changed, 391 insertions, 166 deletions

diff --git a/drivers/infiniband/hw/mlx4/alias_GUID.c b/drivers/infiniband/hw/mlx4/alias_GUID.c
index a31e031afd87..0f00204d2ece 100644
--- a/drivers/infiniband/hw/mlx4/alias_GUID.c
+++ b/drivers/infiniband/hw/mlx4/alias_GUID.c
@@ -58,14 +58,19 @@ struct mlx4_alias_guid_work_context {
 	int			query_id;
 	struct list_head	list;
 	int			block_num;
+	ib_sa_comp_mask		guid_indexes;
+	u8			method;
 };
 
 struct mlx4_next_alias_guid_work {
 	u8 port;
 	u8 block_num;
+	u8 method;
 	struct mlx4_sriov_alias_guid_info_rec_det rec_det;
 };
 
+static int get_low_record_time_index(struct mlx4_ib_dev *dev, u8 port,
+				     int *resched_delay_sec);
 
 void mlx4_ib_update_cache_on_guid_change(struct mlx4_ib_dev *dev, int block_num,
 					 u8 port_num, u8 *p_data)
@@ -118,6 +123,57 @@ ib_sa_comp_mask mlx4_ib_get_aguid_comp_mask_from_ix(int index)
 	return IB_SA_COMP_MASK(4 + index);
 }
 
+void mlx4_ib_slave_alias_guid_event(struct mlx4_ib_dev *dev, int slave,
+				    int port,  int slave_init)
+{
+	__be64 curr_guid, required_guid;
+	int record_num = slave / 8;
+	int index = slave % 8;
+	int port_index = port - 1;
+	unsigned long flags;
+	int do_work = 0;
+
+	spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags);
+	if (dev->sriov.alias_guid.ports_guid[port_index].state_flags &
+	    GUID_STATE_NEED_PORT_INIT)
+		goto unlock;
+	if (!slave_init) {
+		curr_guid = *(__be64 *)&dev->sriov.
+			alias_guid.ports_guid[port_index].
+			all_rec_per_port[record_num].
+			all_recs[GUID_REC_SIZE * index];
+		if (curr_guid == cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL) ||
+		    !curr_guid)
+			goto unlock;
+		required_guid = cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL);
+	} else {
+		required_guid = mlx4_get_admin_guid(dev->dev, slave, port);
+		if (required_guid == cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL))
+			goto unlock;
+	}
+	*(__be64 *)&dev->sriov.alias_guid.ports_guid[port_index].
+		all_rec_per_port[record_num].
+		all_recs[GUID_REC_SIZE * index] = required_guid;
+	dev->sriov.alias_guid.ports_guid[port_index].
+		all_rec_per_port[record_num].guid_indexes
+		|= mlx4_ib_get_aguid_comp_mask_from_ix(index);
+	dev->sriov.alias_guid.ports_guid[port_index].
+		all_rec_per_port[record_num].status
+		= MLX4_GUID_INFO_STATUS_IDLE;
+	/* set to run immediately */
+	dev->sriov.alias_guid.ports_guid[port_index].
+		all_rec_per_port[record_num].time_to_run = 0;
+	dev->sriov.alias_guid.ports_guid[port_index].
+		all_rec_per_port[record_num].
+		guids_retry_schedule[index] = 0;
+	do_work = 1;
+unlock:
+	spin_unlock_irqrestore(&dev->sriov.alias_guid.ag_work_lock, flags);
+
+	if (do_work)
+		mlx4_ib_init_alias_guid_work(dev, port_index);
+}
+
 /*
  * Whenever new GUID is set/unset (guid table change) create event and
  * notify the relevant slave (master also should be notified).
@@ -138,10 +194,15 @@ void mlx4_ib_notify_slaves_on_guid_change(struct mlx4_ib_dev *dev,
 	enum slave_port_state prev_state;
 	__be64 tmp_cur_ag, form_cache_ag;
 	enum slave_port_gen_event gen_event;
+	struct mlx4_sriov_alias_guid_info_rec_det *rec;
+	unsigned long flags;
+	__be64 required_value;
 
 	if (!mlx4_is_master(dev->dev))
 		return;
 
+	rec = &dev->sriov.alias_guid.ports_guid[port_num - 1].
+			all_rec_per_port[block_num];
 	guid_indexes = be64_to_cpu((__force __be64) dev->sriov.alias_guid.
 				   ports_guid[port_num - 1].
 				   all_rec_per_port[block_num].guid_indexes);
@@ -166,8 +227,27 @@ void mlx4_ib_notify_slaves_on_guid_change(struct mlx4_ib_dev *dev,
 		 */
 		if (tmp_cur_ag != form_cache_ag)
 			continue;
-		mlx4_gen_guid_change_eqe(dev->dev, slave_id, port_num);
 
+		spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags);
+		required_value = *(__be64 *)&rec->all_recs[i * GUID_REC_SIZE];
+
+		if (required_value == cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL))
+			required_value = 0;
+
+		if (tmp_cur_ag == required_value) {
+			rec->guid_indexes = rec->guid_indexes &
+			       ~mlx4_ib_get_aguid_comp_mask_from_ix(i);
+		} else {
+			/* may notify port down if value is 0 */
+			if (tmp_cur_ag != MLX4_NOT_SET_GUID) {
+				spin_unlock_irqrestore(&dev->sriov.
+					alias_guid.ag_work_lock, flags);
+				continue;
+			}
+		}
+		spin_unlock_irqrestore(&dev->sriov.alias_guid.ag_work_lock,
+				       flags);
+		mlx4_gen_guid_change_eqe(dev->dev, slave_id, port_num);
 		/*2 cases: Valid GUID, and Invalid Guid*/
 
 		if (tmp_cur_ag != MLX4_NOT_SET_GUID) { /*valid GUID*/
@@ -188,10 +268,14 @@ void mlx4_ib_notify_slaves_on_guid_change(struct mlx4_ib_dev *dev,
 			set_and_calc_slave_port_state(dev->dev, slave_id, port_num,
 						      MLX4_PORT_STATE_IB_EVENT_GID_INVALID,
 						      &gen_event);
-			pr_debug("sending PORT DOWN event to slave: %d, port: %d\n",
-				 slave_id, port_num);
-			mlx4_gen_port_state_change_eqe(dev->dev, slave_id, port_num,
-						       MLX4_PORT_CHANGE_SUBTYPE_DOWN);
+			if (gen_event == SLAVE_PORT_GEN_EVENT_DOWN) {
+				pr_debug("sending PORT DOWN event to slave: %d, port: %d\n",
+					 slave_id, port_num);
+				mlx4_gen_port_state_change_eqe(dev->dev,
+							       slave_id,
+							       port_num,
+							       MLX4_PORT_CHANGE_SUBTYPE_DOWN);
+			}
 		}
 	}
 }
@@ -206,6 +290,9 @@ static void aliasguid_query_handler(int status,
 	int i;
 	struct mlx4_sriov_alias_guid_info_rec_det *rec;
 	unsigned long flags, flags1;
+	ib_sa_comp_mask declined_guid_indexes = 0;
+	ib_sa_comp_mask applied_guid_indexes = 0;
+	unsigned int resched_delay_sec = 0;
 
 	if (!context)
 		return;
@@ -216,9 +303,9 @@ static void aliasguid_query_handler(int status,
 		all_rec_per_port[cb_ctx->block_num];
 
 	if (status) {
-		rec->status = MLX4_GUID_INFO_STATUS_IDLE;
 		pr_debug("(port: %d) failed: status = %d\n",
 			 cb_ctx->port, status);
+		rec->time_to_run = ktime_get_real_ns() + 1 * NSEC_PER_SEC;
 		goto out;
 	}
 
@@ -235,57 +322,101 @@ static void aliasguid_query_handler(int status,
 	rec = &dev->sriov.alias_guid.ports_guid[port_index].
 		all_rec_per_port[guid_rec->block_num];
 
-	rec->status = MLX4_GUID_INFO_STATUS_SET;
-	rec->method = MLX4_GUID_INFO_RECORD_SET;
-
+	spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags);
 	for (i = 0 ; i < NUM_ALIAS_GUID_IN_REC; i++) {
-		__be64 tmp_cur_ag;
-		tmp_cur_ag = *(__be64 *)&guid_rec->guid_info_list[i * GUID_REC_SIZE];
+		__be64 sm_response, required_val;
+
+		if (!(cb_ctx->guid_indexes &
+			mlx4_ib_get_aguid_comp_mask_from_ix(i)))
+			continue;
+		sm_response = *(__be64 *)&guid_rec->guid_info_list
+				[i * GUID_REC_SIZE];
+		required_val = *(__be64 *)&rec->all_recs[i * GUID_REC_SIZE];
+		if (cb_ctx->method == MLX4_GUID_INFO_RECORD_DELETE) {
+			if (required_val ==
+			    cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL))
+				goto next_entry;
+
+			/* A new value was set till we got the response */
+			pr_debug("need to set new value %llx, record num %d, block_num:%d\n",
+				 be64_to_cpu(required_val),
+				 i, guid_rec->block_num);
+			goto entry_declined;
+		}
+
 		/* check if the SM didn't assign one of the records.
-		 * if it didn't, if it was not sysadmin request:
-		 * ask the SM to give a new GUID, (instead of the driver request).
+		 * if it didn't, re-ask for.
 		 */
-		if (tmp_cur_ag == MLX4_NOT_SET_GUID) {
-			mlx4_ib_warn(&dev->ib_dev, "%s:Record num %d in "
-				     "block_num: %d was declined by SM, "
-				     "ownership by %d (0 = driver, 1=sysAdmin,"
-				     " 2=None)\n", __func__, i,
-				     guid_rec->block_num, rec->ownership);
-			if (rec->ownership == MLX4_GUID_DRIVER_ASSIGN) {
-				/* if it is driver assign, asks for new GUID from SM*/
-				*(__be64 *)&rec->all_recs[i * GUID_REC_SIZE] =
-					MLX4_NOT_SET_GUID;
-
-				/* Mark the record as not assigned, and let it
-				 * be sent again in the next work sched.*/
-				rec->status = MLX4_GUID_INFO_STATUS_IDLE;
-				rec->guid_indexes |= mlx4_ib_get_aguid_comp_mask_from_ix(i);
-			}
+		if (sm_response == MLX4_NOT_SET_GUID) {
+			if (rec->guids_retry_schedule[i] == 0)
+				mlx4_ib_warn(&dev->ib_dev,
+					     "%s:Record num %d in  block_num: %d was declined by SM\n",
+					     __func__, i,
+					     guid_rec->block_num);
+			goto entry_declined;
 		} else {
 		       /* properly assigned record. */
 		       /* We save the GUID we just got from the SM in the
 			* admin_guid in order to be persistent, and in the
 			* request from the sm the process will ask for the same GUID */
-			if (rec->ownership == MLX4_GUID_SYSADMIN_ASSIGN &&
-			    tmp_cur_ag != *(__be64 *)&rec->all_recs[i * GUID_REC_SIZE]) {
-				/* the sysadmin assignment failed.*/
-				mlx4_ib_warn(&dev->ib_dev, "%s: Failed to set"
-					     " admin guid after SysAdmin "
-					     "configuration. "
-					     "Record num %d in block_num:%d "
-					     "was declined by SM, "
-					     "new val(0x%llx) was kept\n",
-					      __func__, i,
-					     guid_rec->block_num,
-					     be64_to_cpu(*(__be64 *) &
-							 rec->all_recs[i * GUID_REC_SIZE]));
+			if (required_val &&
+			    sm_response != required_val) {
+				/* Warn only on first retry */
+				if (rec->guids_retry_schedule[i] == 0)
+					mlx4_ib_warn(&dev->ib_dev, "%s: Failed to set"
+						     " admin guid after SysAdmin "
+						     "configuration. "
+						     "Record num %d in block_num:%d "
+						     "was declined by SM, "
+						     "new val(0x%llx) was kept, SM returned (0x%llx)\n",
+						      __func__, i,
+						     guid_rec->block_num,
+						     be64_to_cpu(required_val),
+						     be64_to_cpu(sm_response));
+				goto entry_declined;
 			} else {
-				memcpy(&rec->all_recs[i * GUID_REC_SIZE],
-				       &guid_rec->guid_info_list[i * GUID_REC_SIZE],
-				       GUID_REC_SIZE);
+				*(__be64 *)&rec->all_recs[i * GUID_REC_SIZE] =
+					sm_response;
+				if (required_val == 0)
+					mlx4_set_admin_guid(dev->dev,
+							    sm_response,
+							    (guid_rec->block_num
+							    * NUM_ALIAS_GUID_IN_REC) + i,
+							    cb_ctx->port);
+				goto next_entry;
 			}
 		}
+entry_declined:
+		declined_guid_indexes |= mlx4_ib_get_aguid_comp_mask_from_ix(i);
+		rec->guids_retry_schedule[i] =
+			(rec->guids_retry_schedule[i] == 0) ?  1 :
+			min((unsigned int)60,
+			    rec->guids_retry_schedule[i] * 2);
+		/* using the minimum value among all entries in that record */
+		resched_delay_sec = (resched_delay_sec == 0) ?
+				rec->guids_retry_schedule[i] :
+				min(resched_delay_sec,
+				    rec->guids_retry_schedule[i]);
+		continue;
+
+next_entry:
+		rec->guids_retry_schedule[i] = 0;
 	}
+
+	applied_guid_indexes =  cb_ctx->guid_indexes & ~declined_guid_indexes;
+	if (declined_guid_indexes ||
+	    rec->guid_indexes & ~(applied_guid_indexes)) {
+		pr_debug("record=%d wasn't fully set, guid_indexes=0x%llx applied_indexes=0x%llx, declined_indexes=0x%llx\n",
+			 guid_rec->block_num,
+			 be64_to_cpu((__force __be64)rec->guid_indexes),
+			 be64_to_cpu((__force __be64)applied_guid_indexes),
+			 be64_to_cpu((__force __be64)declined_guid_indexes));
+		rec->time_to_run = ktime_get_real_ns() +
+			resched_delay_sec * NSEC_PER_SEC;
+	} else {
+		rec->status = MLX4_GUID_INFO_STATUS_SET;
+	}
+	spin_unlock_irqrestore(&dev->sriov.alias_guid.ag_work_lock, flags);
 	/*
 	The func is call here to close the cases when the
 	sm doesn't send smp, so in the sa response the driver
@@ -297,10 +428,13 @@ static void aliasguid_query_handler(int status,
 out:
 	spin_lock_irqsave(&dev->sriov.going_down_lock, flags);
 	spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags1);
-	if (!dev->sriov.is_going_down)
+	if (!dev->sriov.is_going_down) {
+		get_low_record_time_index(dev, port_index, &resched_delay_sec);
 		queue_delayed_work(dev->sriov.alias_guid.ports_guid[port_index].wq,
 				   &dev->sriov.alias_guid.ports_guid[port_index].
-				   alias_guid_work, 0);
+				   alias_guid_work,
+				   msecs_to_jiffies(resched_delay_sec * 1000));
+	}
 	if (cb_ctx->sa_query) {
 		list_del(&cb_ctx->list);
 		kfree(cb_ctx);
@@ -317,9 +451,7 @@ static void invalidate_guid_record(struct mlx4_ib_dev *dev, u8 port, int index)
 	ib_sa_comp_mask comp_mask = 0;
 
 	dev->sriov.alias_guid.ports_guid[port - 1].all_rec_per_port[index].status
-		= MLX4_GUID_INFO_STATUS_IDLE;
-	dev->sriov.alias_guid.ports_guid[port - 1].all_rec_per_port[index].method
-		= MLX4_GUID_INFO_RECORD_SET;
+		= MLX4_GUID_INFO_STATUS_SET;
 
 	/* calculate the comp_mask for that record.*/
 	for (i = 0; i < NUM_ALIAS_GUID_IN_REC; i++) {
@@ -333,19 +465,21 @@ static void invalidate_guid_record(struct mlx4_ib_dev *dev, u8 port, int index)
 		need to assign GUIDs, then don't put it up for assignment.
 		*/
 		if (MLX4_GUID_FOR_DELETE_VAL == cur_admin_val ||
-		    (!index && !i) ||
-		    MLX4_GUID_NONE_ASSIGN == dev->sriov.alias_guid.
-		    ports_guid[port - 1].all_rec_per_port[index].ownership)
+		    (!index && !i))
 			continue;
 		comp_mask |= mlx4_ib_get_aguid_comp_mask_from_ix(i);
 	}
 	dev->sriov.alias_guid.ports_guid[port - 1].
-		all_rec_per_port[index].guid_indexes = comp_mask;
+		all_rec_per_port[index].guid_indexes |= comp_mask;
+	if (dev->sriov.alias_guid.ports_guid[port - 1].
+	    all_rec_per_port[index].guid_indexes)
+		dev->sriov.alias_guid.ports_guid[port - 1].
+		all_rec_per_port[index].status = MLX4_GUID_INFO_STATUS_IDLE;
+
 }
 
 static int set_guid_rec(struct ib_device *ibdev,
-			u8 port, int index,
-			struct mlx4_sriov_alias_guid_info_rec_det *rec_det)
+			struct mlx4_next_alias_guid_work *rec)
 {
 	int err;
 	struct mlx4_ib_dev *dev = to_mdev(ibdev);
@@ -354,6 +488,9 @@ static int set_guid_rec(struct ib_device *ibdev,
 	struct ib_port_attr attr;
 	struct mlx4_alias_guid_work_context *callback_context;
 	unsigned long resched_delay, flags, flags1;
+	u8 port = rec->port + 1;
+	int index = rec->block_num;
+	struct mlx4_sriov_alias_guid_info_rec_det *rec_det = &rec->rec_det;
 	struct list_head *head =
 		&dev->sriov.alias_guid.ports_guid[port - 1].cb_list;
 
@@ -380,6 +517,8 @@ static int set_guid_rec(struct ib_device *ibdev,
 	callback_context->port = port;
 	callback_context->dev = dev;
 	callback_context->block_num = index;
+	callback_context->guid_indexes = rec_det->guid_indexes;
+	callback_context->method = rec->method;
 
 	memset(&guid_info_rec, 0, sizeof (struct ib_sa_guidinfo_rec));
 
@@ -399,7 +538,7 @@ static int set_guid_rec(struct ib_device *ibdev,
 	callback_context->query_id =
 		ib_sa_guid_info_rec_query(dev->sriov.alias_guid.sa_client,
 					  ibdev, port, &guid_info_rec,
-					  comp_mask, rec_det->method, 1000,
+					  comp_mask, rec->method, 1000,
 					  GFP_KERNEL, aliasguid_query_handler,
 					  callback_context,
 					  &callback_context->sa_query);
@@ -434,6 +573,30 @@ out:
 	return err;
 }
 
+static void mlx4_ib_guid_port_init(struct mlx4_ib_dev *dev, int port)
+{
+	int j, k, entry;
+	__be64 guid;
+
+	/*Check if the SM doesn't need to assign the GUIDs*/
+	for (j = 0; j < NUM_ALIAS_GUID_REC_IN_PORT; j++) {
+		for (k = 0; k < NUM_ALIAS_GUID_IN_REC; k++) {
+			entry = j * NUM_ALIAS_GUID_IN_REC + k;
+			/* no request for the 0 entry (hw guid) */
+			if (!entry || entry > dev->dev->persist->num_vfs ||
+			    !mlx4_is_slave_active(dev->dev, entry))
+				continue;
+			guid = mlx4_get_admin_guid(dev->dev, entry, port);
+			*(__be64 *)&dev->sriov.alias_guid.ports_guid[port - 1].
+				all_rec_per_port[j].all_recs
+				[GUID_REC_SIZE * k] = guid;
+			pr_debug("guid was set, entry=%d, val=0x%llx, port=%d\n",
+				 entry,
+				 be64_to_cpu(guid),
+				 port);
+		}
+	}
+}
 void mlx4_ib_invalidate_all_guid_record(struct mlx4_ib_dev *dev, int port)
 {
 	int i;
@@ -443,6 +606,13 @@ void mlx4_ib_invalidate_all_guid_record(struct mlx4_ib_dev *dev, int port)
 
 	spin_lock_irqsave(&dev->sriov.going_down_lock, flags);
 	spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags1);
+
+	if (dev->sriov.alias_guid.ports_guid[port - 1].state_flags &
+		GUID_STATE_NEED_PORT_INIT) {
+		mlx4_ib_guid_port_init(dev, port);
+		dev->sriov.alias_guid.ports_guid[port - 1].state_flags &=
+			(~GUID_STATE_NEED_PORT_INIT);
+	}
 	for (i = 0; i < NUM_ALIAS_GUID_REC_IN_PORT; i++)
 		invalidate_guid_record(dev, port, i);
 
@@ -462,60 +632,107 @@ void mlx4_ib_invalidate_all_guid_record(struct mlx4_ib_dev *dev, int port)
 	spin_unlock_irqrestore(&dev->sriov.going_down_lock, flags);
 }
 
-/* The function returns the next record that was
- * not configured (or failed to be configured) */
-static int get_next_record_to_update(struct mlx4_ib_dev *dev, u8 port,
-				     struct mlx4_next_alias_guid_work *rec)
+static void set_required_record(struct mlx4_ib_dev *dev, u8 port,
+				struct mlx4_next_alias_guid_work *next_rec,
+				int record_index)
 {
-	int j;
-	unsigned long flags;
+	int i;
+	int lowset_time_entry = -1;
+	int lowest_time = 0;
+	ib_sa_comp_mask delete_guid_indexes = 0;
+	ib_sa_comp_mask set_guid_indexes = 0;
+	struct mlx4_sriov_alias_guid_info_rec_det *rec =
+			&dev->sriov.alias_guid.ports_guid[port].
+			all_rec_per_port[record_index];
 
-	for (j = 0; j < NUM_ALIAS_GUID_REC_IN_PORT; j++) {
-		spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags);
-		if (dev->sriov.alias_guid.ports_guid[port].all_rec_per_port[j].status ==
-		    MLX4_GUID_INFO_STATUS_IDLE) {
-			memcpy(&rec->rec_det,
-			       &dev->sriov.alias_guid.ports_guid[port].all_rec_per_port[j],
-			       sizeof (struct mlx4_sriov_alias_guid_info_rec_det));
-			rec->port = port;
-			rec->block_num = j;
-			dev->sriov.alias_guid.ports_guid[port].all_rec_per_port[j].status =
-				MLX4_GUID_INFO_STATUS_PENDING;
-			spin_unlock_irqrestore(&dev->sriov.alias_guid.ag_work_lock, flags);
-			return 0;
+	for (i = 0; i < NUM_ALIAS_GUID_IN_REC; i++) {
+		if (!(rec->guid_indexes &
+			mlx4_ib_get_aguid_comp_mask_from_ix(i)))
+			continue;
+
+		if (*(__be64 *)&rec->all_recs[i * GUID_REC_SIZE] ==
+				cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL))
+			delete_guid_indexes |=
+				mlx4_ib_get_aguid_comp_mask_from_ix(i);
+		else
+			set_guid_indexes |=
+				mlx4_ib_get_aguid_comp_mask_from_ix(i);
+
+		if (lowset_time_entry == -1 || rec->guids_retry_schedule[i] <=
+			lowest_time) {
+			lowset_time_entry = i;
+			lowest_time = rec->guids_retry_schedule[i];
 		}
-		spin_unlock_irqrestore(&dev->sriov.alias_guid.ag_work_lock, flags);
 	}
-	return -ENOENT;
+
+	memcpy(&next_rec->rec_det, rec, sizeof(*rec));
+	next_rec->port = port;
+	next_rec->block_num = record_index;
+
+	if (*(__be64 *)&rec->all_recs[lowset_time_entry * GUID_REC_SIZE] ==
+				cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL)) {
+		next_rec->rec_det.guid_indexes = delete_guid_indexes;
+		next_rec->method = MLX4_GUID_INFO_RECORD_DELETE;
+	} else {
+		next_rec->rec_det.guid_indexes = set_guid_indexes;
+		next_rec->method = MLX4_GUID_INFO_RECORD_SET;
+	}
 }
 
-static void set_administratively_guid_record(struct mlx4_ib_dev *dev, int port,
-					     int rec_index,
-					     struct mlx4_sriov_alias_guid_info_rec_det *rec_det)
+/* return index of record that should be updated based on lowest
+ * rescheduled time
+ */
+static int get_low_record_time_index(struct mlx4_ib_dev *dev, u8 port,
+				     int *resched_delay_sec)
 {
-	dev->sriov.alias_guid.ports_guid[port].all_rec_per_port[rec_index].guid_indexes =
-		rec_det->guid_indexes;
-	memcpy(dev->sriov.alias_guid.ports_guid[port].all_rec_per_port[rec_index].all_recs,
-	       rec_det->all_recs, NUM_ALIAS_GUID_IN_REC * GUID_REC_SIZE);
-	dev->sriov.alias_guid.ports_guid[port].all_rec_per_port[rec_index].status =
-		rec_det->status;
+	int record_index = -1;
+	u64 low_record_time = 0;
+	struct mlx4_sriov_alias_guid_info_rec_det rec;
+	int j;
+
+	for (j = 0; j < NUM_ALIAS_GUID_REC_IN_PORT; j++) {
+		rec = dev->sriov.alias_guid.ports_guid[port].
+			all_rec_per_port[j];
+		if (rec.status == MLX4_GUID_INFO_STATUS_IDLE &&
+		    rec.guid_indexes) {
+			if (record_index == -1 ||
+			    rec.time_to_run < low_record_time) {
+				record_index = j;
+				low_record_time = rec.time_to_run;
+			}
+		}
+	}
+	if (resched_delay_sec) {
+		u64 curr_time = ktime_get_real_ns();
+
+		*resched_delay_sec = (low_record_time < curr_time) ? 0 :
+			div_u64((low_record_time - curr_time), NSEC_PER_SEC);
+	}
+
+	return record_index;
 }
 
-static void set_all_slaves_guids(struct mlx4_ib_dev *dev, int port)
+/* The function returns the next record that was
+ * not configured (or failed to be configured) */
+static int get_next_record_to_update(struct mlx4_ib_dev *dev, u8 port,
+				     struct mlx4_next_alias_guid_work *rec)
 {
-	int j;
-	struct mlx4_sriov_alias_guid_info_rec_det rec_det ;
-
-	for (j = 0 ; j < NUM_ALIAS_GUID_REC_IN_PORT ; j++) {
-		memset(rec_det.all_recs, 0, NUM_ALIAS_GUID_IN_REC * GUID_REC_SIZE);
-		rec_det.guid_indexes = (!j ? 0 : IB_SA_GUIDINFO_REC_GID0) |
-			IB_SA_GUIDINFO_REC_GID1 | IB_SA_GUIDINFO_REC_GID2 |
-			IB_SA_GUIDINFO_REC_GID3 | IB_SA_GUIDINFO_REC_GID4 |
-			IB_SA_GUIDINFO_REC_GID5 | IB_SA_GUIDINFO_REC_GID6 |
-			IB_SA_GUIDINFO_REC_GID7;
-		rec_det.status = MLX4_GUID_INFO_STATUS_IDLE;
-		set_administratively_guid_record(dev, port, j, &rec_det);
+	unsigned long flags;
+	int record_index;
+	int ret = 0;
+
+	spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags);
+	record_index = get_low_record_time_index(dev, port, NULL);
+
+	if (record_index < 0) {
+		ret = -ENOENT;
+		goto out;
 	}
+
+	set_required_record(dev, port, rec, record_index);
+out:
+	spin_unlock_irqrestore(&dev->sriov.alias_guid.ag_work_lock, flags);
+	return ret;
 }
 
 static void alias_guid_work(struct work_struct *work)
@@ -545,9 +762,7 @@ static void alias_guid_work(struct work_struct *work)
 		goto out;
 	}
 
-	set_guid_rec(&dev->ib_dev, rec->port + 1, rec->block_num,
-		     &rec->rec_det);
-
+	set_guid_rec(&dev->ib_dev, rec);
 out:
 	kfree(rec);
 }
@@ -562,6 +777,12 @@ void mlx4_ib_init_alias_guid_work(struct mlx4_ib_dev *dev, int port)
 	spin_lock_irqsave(&dev->sriov.going_down_lock, flags);
 	spin_lock_irqsave(&dev->sriov.alias_guid.ag_work_lock, flags1);
 	if (!dev->sriov.is_going_down) {
+		/* If there is pending one should cancell then run, otherwise
+		  * won't run till previous one is ended as same work
+		  * struct is used.
+		  */
+		cancel_delayed_work(&dev->sriov.alias_guid.ports_guid[port].
+				    alias_guid_work);
 		queue_delayed_work(dev->sriov.alias_guid.ports_guid[port].wq,
 			   &dev->sriov.alias_guid.ports_guid[port].alias_guid_work, 0);
 	}
@@ -609,7 +830,7 @@ int mlx4_ib_init_alias_guid_service(struct mlx4_ib_dev *dev)
 {
 	char alias_wq_name[15];
 	int ret = 0;
-	int i, j, k;
+	int i, j;
 	union ib_gid gid;
 
 	if (!mlx4_is_master(dev->dev))
@@ -633,33 +854,25 @@ int mlx4_ib_init_alias_guid_service(struct mlx4_ib_dev *dev)
 	for (i = 0 ; i < dev->num_ports; i++) {
 		memset(&dev->sriov.alias_guid.ports_guid[i], 0,
 		       sizeof (struct mlx4_sriov_alias_guid_port_rec_det));
-		/*Check if the SM doesn't need to assign the GUIDs*/
+		dev->sriov.alias_guid.ports_guid[i].state_flags |=
+				GUID_STATE_NEED_PORT_INIT;
 		for (j = 0; j < NUM_ALIAS_GUID_REC_IN_PORT; j++) {
-			if (mlx4_ib_sm_guid_assign) {
-				dev->sriov.alias_guid.ports_guid[i].
-					all_rec_per_port[j].
-					ownership = MLX4_GUID_DRIVER_ASSIGN;
-				continue;
-			}
-			dev->sriov.alias_guid.ports_guid[i].all_rec_per_port[j].
-					ownership = MLX4_GUID_NONE_ASSIGN;
-			/*mark each val as it was deleted,
-			  till the sysAdmin will give it valid val*/
-			for (k = 0; k < NUM_ALIAS_GUID_IN_REC; k++) {
-				*(__be64 *)&dev->sriov.alias_guid.ports_guid[i].
-					all_rec_per_port[j].all_recs[GUID_REC_SIZE * k] =
-						cpu_to_be64(MLX4_GUID_FOR_DELETE_VAL);
-			}
+			/* mark each val as it was deleted */
+			memset(dev->sriov.alias_guid.ports_guid[i].
+				all_rec_per_port[j].all_recs, 0xFF,
+				sizeof(dev->sriov.alias_guid.ports_guid[i].
+				all_rec_per_port[j].all_recs));
 		}
 		INIT_LIST_HEAD(&dev->sriov.alias_guid.ports_guid[i].cb_list);
 		/*prepare the records, set them to be allocated by sm*/
+		if (mlx4_ib_sm_guid_assign)
+			for (j = 1; j < NUM_ALIAS_GUID_PER_PORT; j++)
+				mlx4_set_admin_guid(dev->dev, 0, j, i + 1);
 		for (j = 0 ; j < NUM_ALIAS_GUID_REC_IN_PORT; j++)
 			invalidate_guid_record(dev, i + 1, j);
 
 		dev->sriov.alias_guid.ports_guid[i].parent = &dev->sriov.alias_guid;
 		dev->sriov.alias_guid.ports_guid[i].port  = i;
-		if (mlx4_ib_sm_guid_assign)
-			set_all_slaves_guids(dev, i);
 
 		snprintf(alias_wq_name, sizeof alias_wq_name, "alias_guid%d", i);
 		dev->sriov.alias_guid.ports_guid[i].wq =
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index 59040265e361..9cd2b002d7ae 100644
--- a/drivers/infiniband/hw/mlx4/mad.c
+++ b/drivers/infiniband/hw/mlx4/mad.c
@@ -1430,6 +1430,10 @@ static int mlx4_ib_alloc_pv_bufs(struct mlx4_ib_demux_pv_ctx *ctx,
 							tun_qp->ring[i].addr,
 							rx_buf_size,
 							DMA_FROM_DEVICE);
+		if (ib_dma_mapping_error(ctx->ib_dev, tun_qp->ring[i].map)) {
+			kfree(tun_qp->ring[i].addr);
+			goto err;
+		}
 	}
 
 	for (i = 0; i < MLX4_NUM_TUNNEL_BUFS; i++) {
@@ -1442,6 +1446,11 @@ static int mlx4_ib_alloc_pv_bufs(struct mlx4_ib_demux_pv_ctx *ctx,
 					  tun_qp->tx_ring[i].buf.addr,
 					  tx_buf_size,
 					  DMA_TO_DEVICE);
+		if (ib_dma_mapping_error(ctx->ib_dev,
+					 tun_qp->tx_ring[i].buf.map)) {
+			kfree(tun_qp->tx_ring[i].buf.addr);
+			goto tx_err;
+		}
 		tun_qp->tx_ring[i].ah = NULL;
 	}
 	spin_lock_init(&tun_qp->tx_lock);
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 976bea794b5f..57070c529dfb 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -66,9 +66,9 @@ MODULE_DESCRIPTION("Mellanox ConnectX HCA InfiniBand driver");
 MODULE_LICENSE("Dual BSD/GPL");
 MODULE_VERSION(DRV_VERSION);
 
-int mlx4_ib_sm_guid_assign = 1;
+int mlx4_ib_sm_guid_assign = 0;
 module_param_named(sm_guid_assign, mlx4_ib_sm_guid_assign, int, 0444);
-MODULE_PARM_DESC(sm_guid_assign, "Enable SM alias_GUID assignment if sm_guid_assign > 0 (Default: 1)");
+MODULE_PARM_DESC(sm_guid_assign, "Enable SM alias_GUID assignment if sm_guid_assign > 0 (Default: 0)");
 
 static const char mlx4_ib_version[] =
 	DRV_NAME ": Mellanox ConnectX InfiniBand driver v"
@@ -2791,9 +2791,31 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr,
 	case MLX4_DEV_EVENT_SLAVE_INIT:
 		/* here, p is the slave id */
 		do_slave_init(ibdev, p, 1);
+		if (mlx4_is_master(dev)) {
+			int i;
+
+			for (i = 1; i <= ibdev->num_ports; i++) {
+				if (rdma_port_get_link_layer(&ibdev->ib_dev, i)
+					== IB_LINK_LAYER_INFINIBAND)
+					mlx4_ib_slave_alias_guid_event(ibdev,
+								       p, i,
+								       1);
+			}
+		}
 		return;
 
 	case MLX4_DEV_EVENT_SLAVE_SHUTDOWN:
+		if (mlx4_is_master(dev)) {
+			int i;
+
+			for (i = 1; i <= ibdev->num_ports; i++) {
+				if (rdma_port_get_link_layer(&ibdev->ib_dev, i)
+					== IB_LINK_LAYER_INFINIBAND)
+					mlx4_ib_slave_alias_guid_event(ibdev,
+								       p, i,
+								       0);
+			}
+		}
 		/* here, p is the slave id */
 		do_slave_init(ibdev, p, 0);
 		return;
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index f829fd935b79..fce3934372a1 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -342,14 +342,9 @@ struct mlx4_ib_ah {
 enum mlx4_guid_alias_rec_status {
 	MLX4_GUID_INFO_STATUS_IDLE,
 	MLX4_GUID_INFO_STATUS_SET,
-	MLX4_GUID_INFO_STATUS_PENDING,
 };
 
-enum mlx4_guid_alias_rec_ownership {
-	MLX4_GUID_DRIVER_ASSIGN,
-	MLX4_GUID_SYSADMIN_ASSIGN,
-	MLX4_GUID_NONE_ASSIGN, /*init state of each record*/
-};
+#define GUID_STATE_NEED_PORT_INIT 0x01
 
 enum mlx4_guid_alias_rec_method {
 	MLX4_GUID_INFO_RECORD_SET	= IB_MGMT_METHOD_SET,
@@ -360,8 +355,8 @@ struct mlx4_sriov_alias_guid_info_rec_det {
 	u8 all_recs[GUID_REC_SIZE * NUM_ALIAS_GUID_IN_REC];
 	ib_sa_comp_mask guid_indexes; /*indicates what from the 8 records are valid*/
 	enum mlx4_guid_alias_rec_status status; /*indicates the administraively status of the record.*/
-	u8 method; /*set or delete*/
-	enum mlx4_guid_alias_rec_ownership ownership; /*indicates who assign that alias_guid record*/
+	unsigned int guids_retry_schedule[NUM_ALIAS_GUID_IN_REC];
+	u64 time_to_run;
 };
 
 struct mlx4_sriov_alias_guid_port_rec_det {
@@ -369,6 +364,7 @@ struct mlx4_sriov_alias_guid_port_rec_det {
 	struct workqueue_struct *wq;
 	struct delayed_work alias_guid_work;
 	u8 port;
+	u32 state_flags;
 	struct mlx4_sriov_alias_guid *parent;
 	struct list_head cb_list;
 };
@@ -802,6 +798,8 @@ int add_sysfs_port_mcg_attr(struct mlx4_ib_dev *device, int port_num,
 void del_sysfs_port_mcg_attr(struct mlx4_ib_dev *device, int port_num,
 			     struct attribute *attr);
 ib_sa_comp_mask mlx4_ib_get_aguid_comp_mask_from_ix(int index);
+void mlx4_ib_slave_alias_guid_event(struct mlx4_ib_dev *dev, int slave,
+				    int port, int slave_init);
 
 int mlx4_ib_device_register_sysfs(struct mlx4_ib_dev *device) ;
 
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index ed2bd6701f9b..02fc91c68027 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -566,6 +566,10 @@ static int alloc_proxy_bufs(struct ib_device *dev, struct mlx4_ib_qp *qp)
 			ib_dma_map_single(dev, qp->sqp_proxy_rcv[i].addr,
 					  sizeof (struct mlx4_ib_proxy_sqp_hdr),
 					  DMA_FROM_DEVICE);
+		if (ib_dma_mapping_error(dev, qp->sqp_proxy_rcv[i].map)) {
+			kfree(qp->sqp_proxy_rcv[i].addr);
+			goto err;
+		}
 	}
 	return 0;
 
@@ -2605,8 +2609,7 @@ static int build_lso_seg(struct mlx4_wqe_lso_seg *wqe, struct ib_send_wr *wr,
 
 	memcpy(wqe->header, wr->wr.ud.header, wr->wr.ud.hlen);
 
-	*lso_hdr_sz  = cpu_to_be32((wr->wr.ud.mss - wr->wr.ud.hlen) << 16 |
-				   wr->wr.ud.hlen);
+	*lso_hdr_sz  = cpu_to_be32(wr->wr.ud.mss << 16 | wr->wr.ud.hlen);
 	*lso_seg_len = halign;
 	return 0;
 }
diff --git a/drivers/infiniband/hw/mlx4/sysfs.c b/drivers/infiniband/hw/mlx4/sysfs.c
index d10c2b8a5dad..6797108ce873 100644
--- a/drivers/infiniband/hw/mlx4/sysfs.c
+++ b/drivers/infiniband/hw/mlx4/sysfs.c
@@ -46,21 +46,17 @@
 static ssize_t show_admin_alias_guid(struct device *dev,
 			      struct device_attribute *attr, char *buf)
 {
-	int record_num;/*0-15*/
-	int guid_index_in_rec; /*0 - 7*/
 	struct mlx4_ib_iov_sysfs_attr *mlx4_ib_iov_dentry =
 		container_of(attr, struct mlx4_ib_iov_sysfs_attr, dentry);
 	struct mlx4_ib_iov_port *port = mlx4_ib_iov_dentry->ctx;
 	struct mlx4_ib_dev *mdev = port->dev;
+	__be64 sysadmin_ag_val;
 
-	record_num = mlx4_ib_iov_dentry->entry_num / 8 ;
-	guid_index_in_rec = mlx4_ib_iov_dentry->entry_num % 8 ;
+	sysadmin_ag_val = mlx4_get_admin_guid(mdev->dev,
+					      mlx4_ib_iov_dentry->entry_num,
+					      port->num);
 
-	return sprintf(buf, "%llx\n",
-		       be64_to_cpu(*(__be64 *)&mdev->sriov.alias_guid.
-				   ports_guid[port->num - 1].
-				   all_rec_per_port[record_num].
-				   all_recs[8 * guid_index_in_rec]));
+	return sprintf(buf, "%llx\n", be64_to_cpu(sysadmin_ag_val));
 }
 
 /* store_admin_alias_guid stores the (new) administratively assigned value of that GUID.
@@ -80,6 +76,7 @@ static ssize_t store_admin_alias_guid(struct device *dev,
 	struct mlx4_ib_iov_port *port = mlx4_ib_iov_dentry->ctx;
 	struct mlx4_ib_dev *mdev = port->dev;
 	u64 sysadmin_ag_val;
+	unsigned long flags;
 
 	record_num = mlx4_ib_iov_dentry->entry_num / 8;
 	guid_index_in_rec = mlx4_ib_iov_dentry->entry_num % 8;
@@ -87,6 +84,7 @@ static ssize_t store_admin_alias_guid(struct device *dev,
 		pr_err("GUID 0 block 0 is RO\n");
 		return count;
 	}
+	spin_lock_irqsave(&mdev->sriov.alias_guid.ag_work_lock, flags);
 	sscanf(buf, "%llx", &sysadmin_ag_val);
 	*(__be64 *)&mdev->sriov.alias_guid.ports_guid[port->num - 1].
 		all_rec_per_port[record_num].
@@ -96,33 +94,15 @@ static ssize_t store_admin_alias_guid(struct device *dev,
 	/* Change the state to be pending for update */
 	mdev->sriov.alias_guid.ports_guid[port->num - 1].all_rec_per_port[record_num].status
 		= MLX4_GUID_INFO_STATUS_IDLE ;
-
-	mdev->sriov.alias_guid.ports_guid[port->num - 1].all_rec_per_port[record_num].method
-		= MLX4_GUID_INFO_RECORD_SET;
-
-	switch (sysadmin_ag_val) {
-	case MLX4_GUID_FOR_DELETE_VAL:
-		mdev->sriov.alias_guid.ports_guid[port->num - 1].all_rec_per_port[record_num].method
-			= MLX4_GUID_INFO_RECORD_DELETE;
-		mdev->sriov.alias_guid.ports_guid[port->num - 1].all_rec_per_port[record_num].ownership
-			= MLX4_GUID_SYSADMIN_ASSIGN;
-		break;
-	/* The sysadmin requests the SM to re-assign */
-	case MLX4_NOT_SET_GUID:
-		mdev->sriov.alias_guid.ports_guid[port->num - 1].all_rec_per_port[record_num].ownership
-			= MLX4_GUID_DRIVER_ASSIGN;
-		break;
-	/* The sysadmin requests a specific value.*/
-	default:
-		mdev->sriov.alias_guid.ports_guid[port->num - 1].all_rec_per_port[record_num].ownership
-			= MLX4_GUID_SYSADMIN_ASSIGN;
-		break;
-	}
+	mlx4_set_admin_guid(mdev->dev, cpu_to_be64(sysadmin_ag_val),
+			    mlx4_ib_iov_dentry->entry_num,
+			    port->num);
 
 	/* set the record index */
 	mdev->sriov.alias_guid.ports_guid[port->num - 1].all_rec_per_port[record_num].guid_indexes
-		= mlx4_ib_get_aguid_comp_mask_from_ix(guid_index_in_rec);
+		|= mlx4_ib_get_aguid_comp_mask_from_ix(guid_index_in_rec);
 
+	spin_unlock_irqrestore(&mdev->sriov.alias_guid.ag_work_lock, flags);
 	mlx4_ib_init_alias_guid_work(mdev, port->num - 1);
 
 	return count;