1 files changed, 335 insertions, 370 deletions

diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c
index c8db6f8ae018..3d479c63b117 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.c
@@ -31,14 +31,10 @@
  * SOFTWARE.
  */
 
-#include <linux/platform_device.h>
 #include "hns_roce_device.h"
 #include "hns_roce_hem.h"
 #include "hns_roce_common.h"
 
-#define DMA_ADDR_T_SHIFT		12
-#define BT_BA_SHIFT			32
-
 #define HEM_INDEX_BUF			BIT(0)
 #define HEM_INDEX_L0			BIT(1)
 #define HEM_INDEX_L1			BIT(2)
@@ -75,11 +71,14 @@ bool hns_roce_check_whether_mhop(struct hns_roce_dev *hr_dev, u32 type)
 	case HEM_TYPE_CQC_TIMER:
 		hop_num = hr_dev->caps.cqc_timer_hop_num;
 		break;
+	case HEM_TYPE_GMV:
+		hop_num = hr_dev->caps.gmv_hop_num;
+		break;
 	default:
 		return false;
 	}
 
-	return hop_num ? true : false;
+	return hop_num;
 }
 
 static bool hns_roce_check_hem_null(struct hns_roce_hem **hem, u64 hem_idx,
@@ -183,8 +182,16 @@ static int get_hem_table_config(struct hns_roce_dev *hr_dev,
 		mhop->ba_l0_num = hr_dev->caps.srqc_bt_num;
 		mhop->hop_num = hr_dev->caps.srqc_hop_num;
 		break;
+	case HEM_TYPE_GMV:
+		mhop->buf_chunk_size = 1 << (hr_dev->caps.gmv_buf_pg_sz +
+					     PAGE_SHIFT);
+		mhop->bt_chunk_size = 1 << (hr_dev->caps.gmv_ba_pg_sz +
+					    PAGE_SHIFT);
+		mhop->ba_l0_num = hr_dev->caps.gmv_bt_num;
+		mhop->hop_num = hr_dev->caps.gmv_hop_num;
+		break;
 	default:
-		dev_err(dev, "Table %d not support multi-hop addressing!\n",
+		dev_err(dev, "table %u not support multi-hop addressing!\n",
 			type);
 		return -EINVAL;
 	}
@@ -198,9 +205,9 @@ int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
 {
 	struct device *dev = hr_dev->dev;
 	u32 chunk_ba_num;
+	u32 chunk_size;
 	u32 table_idx;
 	u32 bt_num;
-	u32 chunk_size;
 
 	if (get_hem_table_config(hr_dev, mhop, table->type))
 		return -EINVAL;
@@ -216,8 +223,7 @@ int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
 	chunk_ba_num = mhop->bt_chunk_size / BA_BYTE_LEN;
 	chunk_size = table->type < HEM_TYPE_MTT ? mhop->buf_chunk_size :
 			      mhop->bt_chunk_size;
-	table_idx = (*obj & (table->num_obj - 1)) /
-		     (chunk_size / table->obj_size);
+	table_idx = *obj / (chunk_size / table->obj_size);
 	switch (bt_num) {
 	case 3:
 		mhop->l2_idx = table_idx & (chunk_ba_num - 1);
@@ -232,8 +238,8 @@ int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
 		mhop->l0_idx = table_idx;
 		break;
 	default:
-		dev_err(dev, "Table %d not support hop_num = %d!\n",
-			     table->type, mhop->hop_num);
+		dev_err(dev, "table %u not support hop_num = %u!\n",
+			table->type, mhop->hop_num);
 		return -EINVAL;
 	}
 	if (mhop->l0_idx >= mhop->ba_l0_num)
@@ -243,171 +249,54 @@ int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
 }
 
 static struct hns_roce_hem *hns_roce_alloc_hem(struct hns_roce_dev *hr_dev,
-					       int npages,
-					       unsigned long hem_alloc_size,
-					       gfp_t gfp_mask)
+					       unsigned long hem_alloc_size)
 {
-	struct hns_roce_hem_chunk *chunk = NULL;
 	struct hns_roce_hem *hem;
-	struct scatterlist *mem;
 	int order;
 	void *buf;
 
-	WARN_ON(gfp_mask & __GFP_HIGHMEM);
+	order = get_order(hem_alloc_size);
+	if (PAGE_SIZE << order != hem_alloc_size) {
+		dev_err(hr_dev->dev, "invalid hem_alloc_size: %lu!\n",
+			hem_alloc_size);
+		return NULL;
+	}
 
-	hem = kmalloc(sizeof(*hem),
-		      gfp_mask & ~(__GFP_HIGHMEM | __GFP_NOWARN));
+	hem = kmalloc(sizeof(*hem), GFP_KERNEL);
 	if (!hem)
 		return NULL;
 
-	hem->refcount = 0;
-	INIT_LIST_HEAD(&hem->chunk_list);
-
-	order = get_order(hem_alloc_size);
-
-	while (npages > 0) {
-		if (!chunk) {
-			chunk = kmalloc(sizeof(*chunk),
-				gfp_mask & ~(__GFP_HIGHMEM | __GFP_NOWARN));
-			if (!chunk)
-				goto fail;
-
-			sg_init_table(chunk->mem, HNS_ROCE_HEM_CHUNK_LEN);
-			chunk->npages = 0;
-			chunk->nsg = 0;
-			memset(chunk->buf, 0, sizeof(chunk->buf));
-			list_add_tail(&chunk->list, &hem->chunk_list);
-		}
+	buf = dma_alloc_coherent(hr_dev->dev, hem_alloc_size,
+				 &hem->dma, GFP_KERNEL);
+	if (!buf)
+		goto fail;
 
-		while (1 << order > npages)
-			--order;
-
-		/*
-		 * Alloc memory one time. If failed, don't alloc small block
-		 * memory, directly return fail.
-		 */
-		mem = &chunk->mem[chunk->npages];
-		buf = dma_alloc_coherent(hr_dev->dev, PAGE_SIZE << order,
-				&sg_dma_address(mem), gfp_mask);
-		if (!buf)
-			goto fail;
-
-		chunk->buf[chunk->npages] = buf;
-		sg_dma_len(mem) = PAGE_SIZE << order;
-
-		++chunk->npages;
-		++chunk->nsg;
-		npages -= 1 << order;
-	}
+	hem->buf = buf;
+	hem->size = hem_alloc_size;
 
 	return hem;
 
 fail:
-	hns_roce_free_hem(hr_dev, hem);
+	kfree(hem);
 	return NULL;
 }
 
 void hns_roce_free_hem(struct hns_roce_dev *hr_dev, struct hns_roce_hem *hem)
 {
-	struct hns_roce_hem_chunk *chunk, *tmp;
-	int i;
-
 	if (!hem)
 		return;
 
-	list_for_each_entry_safe(chunk, tmp, &hem->chunk_list, list) {
-		for (i = 0; i < chunk->npages; ++i)
-			dma_free_coherent(hr_dev->dev,
-				   sg_dma_len(&chunk->mem[i]),
-				   chunk->buf[i],
-				   sg_dma_address(&chunk->mem[i]));
-		kfree(chunk);
-	}
+	dma_free_coherent(hr_dev->dev, hem->size, hem->buf, hem->dma);
 
 	kfree(hem);
 }
 
-static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
-			    struct hns_roce_hem_table *table, unsigned long obj)
-{
-	spinlock_t *lock = &hr_dev->bt_cmd_lock;
-	struct device *dev = hr_dev->dev;
-	long end;
-	unsigned long flags;
-	struct hns_roce_hem_iter iter;
-	void __iomem *bt_cmd;
-	__le32 bt_cmd_val[2];
-	__le32 bt_cmd_h = 0;
-	__le32 bt_cmd_l = 0;
-	u64 bt_ba = 0;
-	int ret = 0;
-
-	/* Find the HEM(Hardware Entry Memory) entry */
-	unsigned long i = (obj & (table->num_obj - 1)) /
-			  (table->table_chunk_size / table->obj_size);
-
-	switch (table->type) {
-	case HEM_TYPE_QPC:
-	case HEM_TYPE_MTPT:
-	case HEM_TYPE_CQC:
-	case HEM_TYPE_SRQC:
-		roce_set_field(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, table->type);
-		break;
-	default:
-		return ret;
-	}
-
-	roce_set_field(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_IN_MDF_M,
-		       ROCEE_BT_CMD_H_ROCEE_BT_CMD_IN_MDF_S, obj);
-	roce_set_bit(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_S, 0);
-	roce_set_bit(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_HW_SYNS_S, 1);
-
-	/* Currently iter only a chunk */
-	for (hns_roce_hem_first(table->hem[i], &iter);
-	     !hns_roce_hem_last(&iter); hns_roce_hem_next(&iter)) {
-		bt_ba = hns_roce_hem_addr(&iter) >> DMA_ADDR_T_SHIFT;
-
-		spin_lock_irqsave(lock, flags);
-
-		bt_cmd = hr_dev->reg_base + ROCEE_BT_CMD_H_REG;
-
-		end = HW_SYNC_TIMEOUT_MSECS;
-		while (end > 0) {
-			if (!(readl(bt_cmd) >> BT_CMD_SYNC_SHIFT))
-				break;
-
-			mdelay(HW_SYNC_SLEEP_TIME_INTERVAL);
-			end -= HW_SYNC_SLEEP_TIME_INTERVAL;
-		}
-
-		if (end <= 0) {
-			dev_err(dev, "Write bt_cmd err,hw_sync is not zero.\n");
-			spin_unlock_irqrestore(lock, flags);
-			return -EBUSY;
-		}
-
-		bt_cmd_l = cpu_to_le32(bt_ba);
-		roce_set_field(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_BA_H_M,
-			       ROCEE_BT_CMD_H_ROCEE_BT_CMD_BA_H_S,
-			       bt_ba >> BT_BA_SHIFT);
-
-		bt_cmd_val[0] = bt_cmd_l;
-		bt_cmd_val[1] = bt_cmd_h;
-		hns_roce_write64_k(bt_cmd_val,
-				   hr_dev->reg_base + ROCEE_BT_CMD_L_REG);
-		spin_unlock_irqrestore(lock, flags);
-	}
-
-	return ret;
-}
-
 static int calc_hem_config(struct hns_roce_dev *hr_dev,
 			   struct hns_roce_hem_table *table, unsigned long obj,
 			   struct hns_roce_hem_mhop *mhop,
 			   struct hns_roce_hem_index *index)
 {
-	struct ib_device *ibdev = &hr_dev->ib_dev;
+	struct device *dev = hr_dev->dev;
 	unsigned long mhop_obj = obj;
 	u32 l0_idx, l1_idx, l2_idx;
 	u32 chunk_ba_num;
@@ -438,14 +327,14 @@ static int calc_hem_config(struct hns_roce_dev *hr_dev,
 		index->buf = l0_idx;
 		break;
 	default:
-		ibdev_err(ibdev, "Table %d not support mhop.hop_num = %d!\n",
-			  table->type, mhop->hop_num);
+		dev_err(dev, "table %u not support mhop.hop_num = %u!\n",
+			table->type, mhop->hop_num);
 		return -EINVAL;
 	}
 
 	if (unlikely(index->buf >= table->num_hem)) {
-		ibdev_err(ibdev, "Table %d exceed hem limt idx %llu,max %lu!\n",
-			  table->type, index->buf, table->num_hem);
+		dev_err(dev, "table %u exceed hem limt idx %llu, max %lu!\n",
+			table->type, index->buf, table->num_hem);
 		return -EINVAL;
 	}
 
@@ -485,8 +374,6 @@ static int alloc_mhop_hem(struct hns_roce_dev *hr_dev,
 {
 	u32 bt_size = mhop->bt_chunk_size;
 	struct device *dev = hr_dev->dev;
-	struct hns_roce_hem_iter iter;
-	gfp_t flag;
 	u64 bt_ba;
 	u32 size;
 	int ret;
@@ -525,17 +412,15 @@ static int alloc_mhop_hem(struct hns_roce_dev *hr_dev,
 	 * alloc bt space chunk for MTT/CQE.
 	 */
 	size = table->type < HEM_TYPE_MTT ? mhop->buf_chunk_size : bt_size;
-	flag = (table->lowmem ? GFP_KERNEL : GFP_HIGHUSER) | __GFP_NOWARN;
-	table->hem[index->buf] = hns_roce_alloc_hem(hr_dev, size >> PAGE_SHIFT,
-						    size, flag);
+	table->hem[index->buf] = hns_roce_alloc_hem(hr_dev, size);
 	if (!table->hem[index->buf]) {
 		ret = -ENOMEM;
 		goto err_alloc_hem;
 	}
 
 	index->inited |= HEM_INDEX_BUF;
-	hns_roce_hem_first(table->hem[index->buf], &iter);
-	bt_ba = hns_roce_hem_addr(&iter);
+	bt_ba = table->hem[index->buf]->dma;
+
 	if (table->type < HEM_TYPE_MTT) {
 		if (mhop->hop_num == 2)
 			*(table->bt_l1[index->l1] + mhop->l2_idx) = bt_ba;
@@ -557,14 +442,14 @@ static int set_mhop_hem(struct hns_roce_dev *hr_dev,
 			struct hns_roce_hem_mhop *mhop,
 			struct hns_roce_hem_index *index)
 {
-	struct ib_device *ibdev = &hr_dev->ib_dev;
-	int step_idx;
+	struct device *dev = hr_dev->dev;
+	u32 step_idx;
 	int ret = 0;
 
 	if (index->inited & HEM_INDEX_L0) {
 		ret = hr_dev->hw->set_hem(hr_dev, table, obj, 0);
 		if (ret) {
-			ibdev_err(ibdev, "set HEM step 0 failed!\n");
+			dev_err(dev, "set HEM step 0 failed!\n");
 			goto out;
 		}
 	}
@@ -572,7 +457,7 @@ static int set_mhop_hem(struct hns_roce_dev *hr_dev,
 	if (index->inited & HEM_INDEX_L1) {
 		ret = hr_dev->hw->set_hem(hr_dev, table, obj, 1);
 		if (ret) {
-			ibdev_err(ibdev, "set HEM step 1 failed!\n");
+			dev_err(dev, "set HEM step 1 failed!\n");
 			goto out;
 		}
 	}
@@ -584,7 +469,7 @@ static int set_mhop_hem(struct hns_roce_dev *hr_dev,
 			step_idx = mhop->hop_num;
 		ret = hr_dev->hw->set_hem(hr_dev, table, obj, step_idx);
 		if (ret)
-			ibdev_err(ibdev, "set HEM step last failed!\n");
+			dev_err(dev, "set HEM step last failed!\n");
 	}
 out:
 	return ret;
@@ -594,26 +479,26 @@ static int hns_roce_table_mhop_get(struct hns_roce_dev *hr_dev,
 				   struct hns_roce_hem_table *table,
 				   unsigned long obj)
 {
-	struct ib_device *ibdev = &hr_dev->ib_dev;
 	struct hns_roce_hem_index index = {};
 	struct hns_roce_hem_mhop mhop = {};
+	struct device *dev = hr_dev->dev;
 	int ret;
 
 	ret = calc_hem_config(hr_dev, table, obj, &mhop, &index);
 	if (ret) {
-		ibdev_err(ibdev, "calc hem config failed!\n");
+		dev_err(dev, "calc hem config failed!\n");
 		return ret;
 	}
 
 	mutex_lock(&table->mutex);
 	if (table->hem[index.buf]) {
-		++table->hem[index.buf]->refcount;
+		refcount_inc(&table->hem[index.buf]->refcount);
 		goto out;
 	}
 
 	ret = alloc_mhop_hem(hr_dev, table, &mhop, &index);
 	if (ret) {
-		ibdev_err(ibdev, "alloc mhop hem failed!\n");
+		dev_err(dev, "alloc mhop hem failed!\n");
 		goto out;
 	}
 
@@ -621,12 +506,12 @@ static int hns_roce_table_mhop_get(struct hns_roce_dev *hr_dev,
 	if (table->type < HEM_TYPE_MTT) {
 		ret = set_mhop_hem(hr_dev, table, obj, &mhop, &index);
 		if (ret) {
-			ibdev_err(ibdev, "set HEM address to HW failed!\n");
+			dev_err(dev, "set HEM address to HW failed!\n");
 			goto err_alloc;
 		}
 	}
 
-	++table->hem[index.buf]->refcount;
+	refcount_set(&table->hem[index.buf]->refcount, 1);
 	goto out;
 
 err_alloc:
@@ -640,42 +525,38 @@ int hns_roce_table_get(struct hns_roce_dev *hr_dev,
 		       struct hns_roce_hem_table *table, unsigned long obj)
 {
 	struct device *dev = hr_dev->dev;
-	int ret = 0;
 	unsigned long i;
+	int ret = 0;
 
 	if (hns_roce_check_whether_mhop(hr_dev, table->type))
 		return hns_roce_table_mhop_get(hr_dev, table, obj);
 
-	i = (obj & (table->num_obj - 1)) / (table->table_chunk_size /
-	     table->obj_size);
+	i = obj / (table->table_chunk_size / table->obj_size);
 
 	mutex_lock(&table->mutex);
 
 	if (table->hem[i]) {
-		++table->hem[i]->refcount;
+		refcount_inc(&table->hem[i]->refcount);
 		goto out;
 	}
 
-	table->hem[i] = hns_roce_alloc_hem(hr_dev,
-				       table->table_chunk_size >> PAGE_SHIFT,
-				       table->table_chunk_size,
-				       (table->lowmem ? GFP_KERNEL :
-					GFP_HIGHUSER) | __GFP_NOWARN);
+	table->hem[i] = hns_roce_alloc_hem(hr_dev, table->table_chunk_size);
 	if (!table->hem[i]) {
 		ret = -ENOMEM;
 		goto out;
 	}
 
 	/* Set HEM base address(128K/page, pa) to Hardware */
-	if (hns_roce_set_hem(hr_dev, table, obj)) {
+	ret = hr_dev->hw->set_hem(hr_dev, table, obj, HEM_HOP_STEP_DIRECT);
+	if (ret) {
 		hns_roce_free_hem(hr_dev, table->hem[i]);
 		table->hem[i] = NULL;
-		ret = -ENODEV;
-		dev_err(dev, "set HEM base address to HW failed.\n");
+		dev_err(dev, "set HEM base address to HW failed, ret = %d.\n",
+			ret);
 		goto out;
 	}
 
-	++table->hem[i]->refcount;
+	refcount_set(&table->hem[i]->refcount, 1);
 out:
 	mutex_unlock(&table->mutex);
 	return ret;
@@ -686,10 +567,11 @@ static void clear_mhop_hem(struct hns_roce_dev *hr_dev,
 			   struct hns_roce_hem_mhop *mhop,
 			   struct hns_roce_hem_index *index)
 {
-	struct ib_device *ibdev = &hr_dev->ib_dev;
+	struct device *dev = hr_dev->dev;
 	u32 hop_num = mhop->hop_num;
 	u32 chunk_ba_num;
-	int step_idx;
+	u32 step_idx;
+	int ret;
 
 	index->inited = HEM_INDEX_BUF;
 	chunk_ba_num = mhop->bt_chunk_size / BA_BYTE_LEN;
@@ -713,16 +595,24 @@ static void clear_mhop_hem(struct hns_roce_dev *hr_dev,
 		else
 			step_idx = hop_num;
 
-		if (hr_dev->hw->clear_hem(hr_dev, table, obj, step_idx))
-			ibdev_warn(ibdev, "Clear hop%d HEM failed.\n", hop_num);
-
-		if (index->inited & HEM_INDEX_L1)
-			if (hr_dev->hw->clear_hem(hr_dev, table, obj, 1))
-				ibdev_warn(ibdev, "Clear HEM step 1 failed.\n");
+		ret = hr_dev->hw->clear_hem(hr_dev, table, obj, step_idx);
+		if (ret)
+			dev_warn(dev, "failed to clear hop%u HEM, ret = %d.\n",
+				 hop_num, ret);
+
+		if (index->inited & HEM_INDEX_L1) {
+			ret = hr_dev->hw->clear_hem(hr_dev, table, obj, 1);
+			if (ret)
+				dev_warn(dev, "failed to clear HEM step 1, ret = %d.\n",
+					 ret);
+		}
 
-		if (index->inited & HEM_INDEX_L0)
-			if (hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
-				ibdev_warn(ibdev, "Clear HEM step 0 failed.\n");
+		if (index->inited & HEM_INDEX_L0) {
+			ret = hr_dev->hw->clear_hem(hr_dev, table, obj, 0);
+			if (ret)
+				dev_warn(dev, "failed to clear HEM step 0, ret = %d.\n",
+					 ret);
+		}
 	}
 }
 
@@ -731,22 +621,22 @@ static void hns_roce_table_mhop_put(struct hns_roce_dev *hr_dev,
 				    unsigned long obj,
 				    int check_refcount)
 {
-	struct ib_device *ibdev = &hr_dev->ib_dev;
 	struct hns_roce_hem_index index = {};
 	struct hns_roce_hem_mhop mhop = {};
+	struct device *dev = hr_dev->dev;
 	int ret;
 
 	ret = calc_hem_config(hr_dev, table, obj, &mhop, &index);
 	if (ret) {
-		ibdev_err(ibdev, "calc hem config failed!\n");
+		dev_err(dev, "calc hem config failed!\n");
 		return;
 	}
 
-	mutex_lock(&table->mutex);
-	if (check_refcount && (--table->hem[index.buf]->refcount > 0)) {
-		mutex_unlock(&table->mutex);
+	if (!check_refcount)
+		mutex_lock(&table->mutex);
+	else if (!refcount_dec_and_mutex_lock(&table->hem[index.buf]->refcount,
+					      &table->mutex))
 		return;
-	}
 
 	clear_mhop_hem(hr_dev, table, obj, &mhop, &index);
 	free_mhop_hem(hr_dev, table, &mhop, &index);
@@ -759,25 +649,26 @@ void hns_roce_table_put(struct hns_roce_dev *hr_dev,
 {
 	struct device *dev = hr_dev->dev;
 	unsigned long i;
+	int ret;
 
 	if (hns_roce_check_whether_mhop(hr_dev, table->type)) {
 		hns_roce_table_mhop_put(hr_dev, table, obj, 1);
 		return;
 	}
 
-	i = (obj & (table->num_obj - 1)) /
-	    (table->table_chunk_size / table->obj_size);
+	i = obj / (table->table_chunk_size / table->obj_size);
 
-	mutex_lock(&table->mutex);
+	if (!refcount_dec_and_mutex_lock(&table->hem[i]->refcount,
+					 &table->mutex))
+		return;
 
-	if (--table->hem[i]->refcount == 0) {
-		/* Clear HEM base address */
-		if (hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
-			dev_warn(dev, "Clear HEM base address failed.\n");
+	ret = hr_dev->hw->clear_hem(hr_dev, table, obj, HEM_HOP_STEP_DIRECT);
+	if (ret)
+		dev_warn_ratelimited(dev, "failed to clear HEM base address, ret = %d.\n",
+				     ret);
 
-		hns_roce_free_hem(hr_dev, table->hem[i]);
-		table->hem[i] = NULL;
-	}
+	hns_roce_free_hem(hr_dev, table->hem[i]);
+	table->hem[i] = NULL;
 
 	mutex_unlock(&table->mutex);
 }
@@ -786,27 +677,22 @@ void *hns_roce_table_find(struct hns_roce_dev *hr_dev,
 			  struct hns_roce_hem_table *table,
 			  unsigned long obj, dma_addr_t *dma_handle)
 {
-	struct hns_roce_hem_chunk *chunk;
 	struct hns_roce_hem_mhop mhop;
 	struct hns_roce_hem *hem;
-	void *addr = NULL;
 	unsigned long mhop_obj = obj;
 	unsigned long obj_per_chunk;
 	unsigned long idx_offset;
 	int offset, dma_offset;
-	int length;
-	int i, j;
+	void *addr = NULL;
 	u32 hem_idx = 0;
-
-	if (!table->lowmem)
-		return NULL;
+	int i, j;
 
 	mutex_lock(&table->mutex);
 
 	if (!hns_roce_check_whether_mhop(hr_dev, table->type)) {
 		obj_per_chunk = table->table_chunk_size / table->obj_size;
-		hem = table->hem[(obj & (table->num_obj - 1)) / obj_per_chunk];
-		idx_offset = (obj & (table->num_obj - 1)) % obj_per_chunk;
+		hem = table->hem[obj / obj_per_chunk];
+		idx_offset = obj % obj_per_chunk;
 		dma_offset = offset = idx_offset * table->obj_size;
 	} else {
 		u32 seg_size = 64; /* 8 bytes per BA and 8 BA per segment */
@@ -823,8 +709,7 @@ void *hns_roce_table_find(struct hns_roce_dev *hr_dev,
 			hem_idx = i;
 
 		hem = table->hem[hem_idx];
-		dma_offset = offset = (obj & (table->num_obj - 1)) * seg_size %
-				       mhop.bt_chunk_size;
+		dma_offset = offset = obj * seg_size % mhop.bt_chunk_size;
 		if (mhop.hop_num == 2)
 			dma_offset = offset = 0;
 	}
@@ -832,23 +717,8 @@ void *hns_roce_table_find(struct hns_roce_dev *hr_dev,
 	if (!hem)
 		goto out;
 
-	list_for_each_entry(chunk, &hem->chunk_list, list) {
-		for (i = 0; i < chunk->npages; ++i) {
-			length = sg_dma_len(&chunk->mem[i]);
-			if (dma_handle && dma_offset >= 0) {
-				if (length > (u32)dma_offset)
-					*dma_handle = sg_dma_address(
-						&chunk->mem[i]) + dma_offset;
-				dma_offset -= length;
-			}
-
-			if (length > (u32)offset) {
-				addr = chunk->buf[i] + offset;
-				goto out;
-			}
-			offset -= length;
-		}
-	}
+	*dma_handle = hem->dma + dma_offset;
+	addr = hem->buf + offset;
 
 out:
 	mutex_unlock(&table->mutex);
@@ -857,8 +727,7 @@ out:
 
 int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 			    struct hns_roce_hem_table *table, u32 type,
-			    unsigned long obj_size, unsigned long nobj,
-			    int use_lowmem)
+			    unsigned long obj_size, unsigned long nobj)
 {
 	unsigned long obj_per_chunk;
 	unsigned long num_hem;
@@ -866,7 +735,7 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 	if (!hns_roce_check_whether_mhop(hr_dev, type)) {
 		table->table_chunk_size = hr_dev->caps.chunk_sz;
 		obj_per_chunk = table->table_chunk_size / obj_size;
-		num_hem = (nobj + obj_per_chunk - 1) / obj_per_chunk;
+		num_hem = DIV_ROUND_UP(nobj, obj_per_chunk);
 
 		table->hem = kcalloc(num_hem, sizeof(*table->hem), GFP_KERNEL);
 		if (!table->hem)
@@ -876,7 +745,7 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 		unsigned long buf_chunk_size;
 		unsigned long bt_chunk_size;
 		unsigned long bt_chunk_num;
-		unsigned long num_bt_l0 = 0;
+		unsigned long num_bt_l0;
 		u32 hop_num;
 
 		if (get_hem_table_config(hr_dev, &mhop, type))
@@ -888,8 +757,9 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 		hop_num = mhop.hop_num;
 
 		obj_per_chunk = buf_chunk_size / obj_size;
-		num_hem = (nobj + obj_per_chunk - 1) / obj_per_chunk;
+		num_hem = DIV_ROUND_UP(nobj, obj_per_chunk);
 		bt_chunk_num = bt_chunk_size / BA_BYTE_LEN;
+
 		if (type >= HEM_TYPE_MTT)
 			num_bt_l0 = bt_chunk_num;
 
@@ -901,8 +771,7 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 		if (check_whether_bt_num_3(type, hop_num)) {
 			unsigned long num_bt_l1;
 
-			num_bt_l1 = (num_hem + bt_chunk_num - 1) /
-					     bt_chunk_num;
+			num_bt_l1 = DIV_ROUND_UP(num_hem, bt_chunk_num);
 			table->bt_l1 = kcalloc(num_bt_l1,
 					       sizeof(*table->bt_l1),
 					       GFP_KERNEL);
@@ -934,9 +803,7 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 
 	table->type = type;
 	table->num_hem = num_hem;
-	table->num_obj = nobj;
 	table->obj_size = obj_size;
-	table->lowmem = use_lowmem;
 	mutex_init(&table->mutex);
 
 	return 0;
@@ -966,8 +833,8 @@ static void hns_roce_cleanup_mhop_hem_table(struct hns_roce_dev *hr_dev,
 {
 	struct hns_roce_hem_mhop mhop;
 	u32 buf_chunk_size;
-	int i;
 	u64 obj;
+	int i;
 
 	if (hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop))
 		return;
@@ -997,27 +864,33 @@ void hns_roce_cleanup_hem_table(struct hns_roce_dev *hr_dev,
 {
 	struct device *dev = hr_dev->dev;
 	unsigned long i;
+	int obj;
+	int ret;
 
 	if (hns_roce_check_whether_mhop(hr_dev, table->type)) {
 		hns_roce_cleanup_mhop_hem_table(hr_dev, table);
+		mutex_destroy(&table->mutex);
 		return;
 	}
 
 	for (i = 0; i < table->num_hem; ++i)
 		if (table->hem[i]) {
-			if (hr_dev->hw->clear_hem(hr_dev, table,
-			    i * table->table_chunk_size / table->obj_size, 0))
-				dev_err(dev, "Clear HEM base address failed.\n");
+			obj = i * table->table_chunk_size / table->obj_size;
+			ret = hr_dev->hw->clear_hem(hr_dev, table, obj, 0);
+			if (ret)
+				dev_err(dev, "clear HEM base address failed, ret = %d.\n",
+					ret);
 
 			hns_roce_free_hem(hr_dev, table->hem[i]);
 		}
 
+	mutex_destroy(&table->mutex);
 	kfree(table->hem);
 }
 
 void hns_roce_cleanup_hem(struct hns_roce_dev *hr_dev)
 {
-	if (hr_dev->caps.srqc_entry_sz)
+	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_SRQ)
 		hns_roce_cleanup_hem_table(hr_dev,
 					   &hr_dev->srq_table.table);
 	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->cq_table.table);
@@ -1027,18 +900,22 @@ void hns_roce_cleanup_hem(struct hns_roce_dev *hr_dev)
 	if (hr_dev->caps.cqc_timer_entry_sz)
 		hns_roce_cleanup_hem_table(hr_dev,
 					   &hr_dev->cqc_timer_table);
-	if (hr_dev->caps.sccc_entry_sz)
+	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_QP_FLOW_CTRL)
 		hns_roce_cleanup_hem_table(hr_dev,
 					   &hr_dev->qp_table.sccc_table);
 	if (hr_dev->caps.trrl_entry_sz)
 		hns_roce_cleanup_hem_table(hr_dev,
 					   &hr_dev->qp_table.trrl_table);
+
+	if (hr_dev->caps.gmv_entry_sz)
+		hns_roce_cleanup_hem_table(hr_dev, &hr_dev->gmv_table);
+
 	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->qp_table.irrl_table);
 	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->qp_table.qp_table);
 	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->mr_table.mtpt_table);
 }
 
-struct roce_hem_item {
+struct hns_roce_hem_item {
 	struct list_head list; /* link all hems in the same bt level */
 	struct list_head sibling; /* link all hems in last hop for mtt */
 	void *addr;
@@ -1046,29 +923,36 @@ struct roce_hem_item {
 	size_t count; /* max ba numbers */
 	int start; /* start buf offset in this hem */
 	int end; /* end buf offset in this hem */
+	bool exist_bt;
+};
+
+/* All HEM items are linked in a tree structure */
+struct hns_roce_hem_head {
+	struct list_head branch[HNS_ROCE_MAX_BT_REGION];
+	struct list_head root;
+	struct list_head leaf;
 };
 
-static struct roce_hem_item *hem_list_alloc_item(struct hns_roce_dev *hr_dev,
-						   int start, int end,
-						   int count, bool exist_bt,
-						   int bt_level)
+static struct hns_roce_hem_item *
+hem_list_alloc_item(struct hns_roce_dev *hr_dev, int start, int end, int count,
+		    bool exist_bt)
 {
-	struct roce_hem_item *hem;
+	struct hns_roce_hem_item *hem;
 
 	hem = kzalloc(sizeof(*hem), GFP_KERNEL);
 	if (!hem)
 		return NULL;
 
 	if (exist_bt) {
-		hem->addr = dma_alloc_coherent(hr_dev->dev,
-						   count * BA_BYTE_LEN,
-						   &hem->dma_addr, GFP_KERNEL);
+		hem->addr = dma_alloc_coherent(hr_dev->dev, count * BA_BYTE_LEN,
+					       &hem->dma_addr, GFP_KERNEL);
 		if (!hem->addr) {
 			kfree(hem);
 			return NULL;
 		}
 	}
 
+	hem->exist_bt = exist_bt;
 	hem->count = count;
 	hem->start = start;
 	hem->end = end;
@@ -1079,51 +963,49 @@ static struct roce_hem_item *hem_list_alloc_item(struct hns_roce_dev *hr_dev,
 }
 
 static void hem_list_free_item(struct hns_roce_dev *hr_dev,
-			       struct roce_hem_item *hem, bool exist_bt)
+			       struct hns_roce_hem_item *hem)
 {
-	if (exist_bt)
+	if (hem->exist_bt)
 		dma_free_coherent(hr_dev->dev, hem->count * BA_BYTE_LEN,
 				  hem->addr, hem->dma_addr);
 	kfree(hem);
 }
 
 static void hem_list_free_all(struct hns_roce_dev *hr_dev,
-			      struct list_head *head, bool exist_bt)
+			      struct list_head *head)
 {
-	struct roce_hem_item *hem, *temp_hem;
+	struct hns_roce_hem_item *hem, *temp_hem;
 
 	list_for_each_entry_safe(hem, temp_hem, head, list) {
 		list_del(&hem->list);
-		hem_list_free_item(hr_dev, hem, exist_bt);
+		hem_list_free_item(hr_dev, hem);
 	}
 }
 
-static void hem_list_link_bt(struct hns_roce_dev *hr_dev, void *base_addr,
-			     u64 table_addr)
+static void hem_list_link_bt(void *base_addr, u64 table_addr)
 {
 	*(u64 *)(base_addr) = table_addr;
 }
 
 /* assign L0 table address to hem from root bt */
-static void hem_list_assign_bt(struct hns_roce_dev *hr_dev,
-			       struct roce_hem_item *hem, void *cpu_addr,
+static void hem_list_assign_bt(struct hns_roce_hem_item *hem, void *cpu_addr,
 			       u64 phy_addr)
 {
 	hem->addr = cpu_addr;
 	hem->dma_addr = (dma_addr_t)phy_addr;
 }
 
-static inline bool hem_list_page_is_in_range(struct roce_hem_item *hem,
+static inline bool hem_list_page_is_in_range(struct hns_roce_hem_item *hem,
 					     int offset)
 {
 	return (hem->start <= offset && offset <= hem->end);
 }
 
-static struct roce_hem_item *hem_list_search_item(struct list_head *ba_list,
-						    int page_offset)
+static struct hns_roce_hem_item *hem_list_search_item(struct list_head *ba_list,
+						      int page_offset)
 {
-	struct roce_hem_item *hem, *temp_hem;
-	struct roce_hem_item *found = NULL;
+	struct hns_roce_hem_item *hem, *temp_hem;
+	struct hns_roce_hem_item *found = NULL;
 
 	list_for_each_entry_safe(hem, temp_hem, ba_list, list) {
 		if (hem_list_page_is_in_range(hem, page_offset)) {
@@ -1147,15 +1029,15 @@ static bool hem_list_is_bottom_bt(int hopnum, int bt_level)
 	return bt_level >= (hopnum ? hopnum - 1 : hopnum);
 }
 
-/**
+/*
  * calc base address entries num
  * @hopnum: num of mutihop addressing
  * @bt_level: base address table level
  * @unit: ba entries per bt page
  */
-static u32 hem_list_calc_ba_range(int hopnum, int bt_level, int unit)
+static u64 hem_list_calc_ba_range(int hopnum, int bt_level, int unit)
 {
-	u32 step;
+	u64 step;
 	int max;
 	int i;
 
@@ -1180,7 +1062,7 @@ static u32 hem_list_calc_ba_range(int hopnum, int bt_level, int unit)
 	return step;
 }
 
-/**
+/*
  * calc the root ba entries which could cover all regions
  * @regions: buf region array
  * @region_cnt: array size of @regions
@@ -1191,11 +1073,15 @@ int hns_roce_hem_list_calc_root_ba(const struct hns_roce_buf_region *regions,
 {
 	struct hns_roce_buf_region *r;
 	int total = 0;
-	int step;
+	u64 step;
 	int i;
 
 	for (i = 0; i < region_cnt; i++) {
 		r = (struct hns_roce_buf_region *)&regions[i];
+		/* when r->hopnum = 0, the region should not occupy root_ba. */
+		if (!r->hopnum)
+			continue;
+
 		if (r->hopnum > 1) {
 			step = hem_list_calc_ba_range(r->hopnum, 1, unit);
 			if (step > 0)
@@ -1213,16 +1099,16 @@ static int hem_list_alloc_mid_bt(struct hns_roce_dev *hr_dev,
 				 int offset, struct list_head *mid_bt,
 				 struct list_head *btm_bt)
 {
-	struct roce_hem_item *hem_ptrs[HNS_ROCE_MAX_BT_LEVEL] = { NULL };
+	struct hns_roce_hem_item *hem_ptrs[HNS_ROCE_MAX_BT_LEVEL] = { NULL };
 	struct list_head temp_list[HNS_ROCE_MAX_BT_LEVEL];
-	struct roce_hem_item *cur, *pre;
+	struct hns_roce_hem_item *cur, *pre;
 	const int hopnum = r->hopnum;
 	int start_aligned;
 	int distance;
 	int ret = 0;
 	int max_ofs;
 	int level;
-	u32 step;
+	u64 step;
 	int end;
 
 	if (hopnum <= 1)
@@ -1234,7 +1120,7 @@ static int hem_list_alloc_mid_bt(struct hns_roce_dev *hr_dev,
 	}
 
 	if (offset < r->offset) {
-		dev_err(hr_dev->dev, "invalid offset %d,min %d!\n",
+		dev_err(hr_dev->dev, "invalid offset %d, min %u!\n",
 			offset, r->offset);
 		return -EINVAL;
 	}
@@ -1246,10 +1132,12 @@ static int hem_list_alloc_mid_bt(struct hns_roce_dev *hr_dev,
 
 	/* config L1 bt to last bt and link them to corresponding parent */
 	for (level = 1; level < hopnum; level++) {
-		cur = hem_list_search_item(&mid_bt[level], offset);
-		if (cur) {
-			hem_ptrs[level] = cur;
-			continue;
+		if (!hem_list_is_bottom_bt(hopnum, level)) {
+			cur = hem_list_search_item(&mid_bt[level], offset);
+			if (cur) {
+				hem_ptrs[level] = cur;
+				continue;
+			}
 		}
 
 		step = hem_list_calc_ba_range(hopnum, level, unit);
@@ -1259,9 +1147,9 @@ static int hem_list_alloc_mid_bt(struct hns_roce_dev *hr_dev,
 		}
 
 		start_aligned = (distance / step) * step + r->offset;
-		end = min_t(int, start_aligned + step - 1, max_ofs);
+		end = min_t(u64, start_aligned + step - 1, max_ofs);
 		cur = hem_list_alloc_item(hr_dev, start_aligned, end, unit,
-					  true, level);
+					  true);
 		if (!cur) {
 			ret = -ENOMEM;
 			goto err_exit;
@@ -1275,8 +1163,7 @@ static int hem_list_alloc_mid_bt(struct hns_roce_dev *hr_dev,
 		if (level > 1) {
 			pre = hem_ptrs[level - 1];
 			step = (cur->start - pre->start) / step * BA_BYTE_LEN;
-			hem_list_link_bt(hr_dev, pre->addr + step,
-					 cur->dma_addr);
+			hem_list_link_bt(pre->addr + step, cur->dma_addr);
 		}
 	}
 
@@ -1288,57 +1175,111 @@ static int hem_list_alloc_mid_bt(struct hns_roce_dev *hr_dev,
 
 err_exit:
 	for (level = 1; level < hopnum; level++)
-		hem_list_free_all(hr_dev, &temp_list[level], true);
+		hem_list_free_all(hr_dev, &temp_list[level]);
 
 	return ret;
 }
 
-static int hem_list_alloc_root_bt(struct hns_roce_dev *hr_dev,
-				  struct hns_roce_hem_list *hem_list, int unit,
-				  const struct hns_roce_buf_region *regions,
-				  int region_cnt)
+static struct hns_roce_hem_item *
+alloc_root_hem(struct hns_roce_dev *hr_dev, int unit, int *max_ba_num,
+	       const struct hns_roce_buf_region *regions, int region_cnt)
 {
-	struct roce_hem_item *hem, *temp_hem, *root_hem;
-	struct list_head temp_list[HNS_ROCE_MAX_BT_REGION];
 	const struct hns_roce_buf_region *r;
-	struct list_head temp_root;
-	struct list_head temp_btm;
-	void *cpu_base;
-	u64 phy_base;
-	int ret = 0;
+	struct hns_roce_hem_item *hem;
 	int ba_num;
 	int offset;
-	int total;
-	int step;
-	int i;
-
-	r = &regions[0];
-	root_hem = hem_list_search_item(&hem_list->root_bt, r->offset);
-	if (root_hem)
-		return 0;
 
 	ba_num = hns_roce_hem_list_calc_root_ba(regions, region_cnt, unit);
 	if (ba_num < 1)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
-	INIT_LIST_HEAD(&temp_root);
-	offset = r->offset;
+	if (ba_num > unit)
+		return ERR_PTR(-ENOBUFS);
+
+	offset = regions[0].offset;
 	/* indicate to last region */
 	r = &regions[region_cnt - 1];
-	root_hem = hem_list_alloc_item(hr_dev, offset, r->offset + r->count - 1,
-				       ba_num, true, 0);
-	if (!root_hem)
+	hem = hem_list_alloc_item(hr_dev, offset, r->offset + r->count - 1,
+				  ba_num, true);
+	if (!hem)
+		return ERR_PTR(-ENOMEM);
+
+	*max_ba_num = ba_num;
+
+	return hem;
+}
+
+static int alloc_fake_root_bt(struct hns_roce_dev *hr_dev, void *cpu_base,
+			      u64 phy_base, const struct hns_roce_buf_region *r,
+			      struct list_head *branch_head,
+			      struct list_head *leaf_head)
+{
+	struct hns_roce_hem_item *hem;
+
+	/* This is on the has_mtt branch, if r->hopnum
+	 * is 0, there is no root_ba to reuse for the
+	 * region's fake hem, so a dma_alloc request is
+	 * necessary here.
+	 */
+	hem = hem_list_alloc_item(hr_dev, r->offset, r->offset + r->count - 1,
+				  r->count, !r->hopnum);
+	if (!hem)
 		return -ENOMEM;
-	list_add(&root_hem->list, &temp_root);
 
-	hem_list->root_ba = root_hem->dma_addr;
+	/* The root_ba can be reused only when r->hopnum > 0. */
+	if (r->hopnum)
+		hem_list_assign_bt(hem, cpu_base, phy_base);
+	list_add(&hem->list, branch_head);
+	list_add(&hem->sibling, leaf_head);
 
-	INIT_LIST_HEAD(&temp_btm);
-	for (i = 0; i < region_cnt; i++)
-		INIT_LIST_HEAD(&temp_list[i]);
+	/* If r->hopnum == 0, 0 is returned,
+	 * so that the root_bt entry is not occupied.
+	 */
+	return r->hopnum ? r->count : 0;
+}
+
+static int setup_middle_bt(struct hns_roce_dev *hr_dev, void *cpu_base,
+			   int unit, const struct hns_roce_buf_region *r,
+			   const struct list_head *branch_head)
+{
+	struct hns_roce_hem_item *hem, *temp_hem;
+	int total = 0;
+	int offset;
+	u64 step;
+
+	step = hem_list_calc_ba_range(r->hopnum, 1, unit);
+	if (step < 1)
+		return -EINVAL;
+
+	/* if exist mid bt, link L1 to L0 */
+	list_for_each_entry_safe(hem, temp_hem, branch_head, list) {
+		offset = (hem->start - r->offset) / step * BA_BYTE_LEN;
+		hem_list_link_bt(cpu_base + offset, hem->dma_addr);
+		total++;
+	}
+
+	return total;
+}
+
+static int
+setup_root_hem(struct hns_roce_dev *hr_dev, struct hns_roce_hem_list *hem_list,
+	       int unit, int max_ba_num, struct hns_roce_hem_head *head,
+	       const struct hns_roce_buf_region *regions, int region_cnt)
+{
+	const struct hns_roce_buf_region *r;
+	struct hns_roce_hem_item *root_hem;
+	void *cpu_base;
+	u64 phy_base;
+	int i, total;
+	int ret;
+
+	root_hem = list_first_entry(&head->root,
+				    struct hns_roce_hem_item, list);
+	if (!root_hem)
+		return -ENOMEM;
 
 	total = 0;
-	for (i = 0; i < region_cnt && total < ba_num; i++) {
+	for (i = 0; i < region_cnt && total <= max_ba_num; i++) {
 		r = &regions[i];
 		if (!r->count)
 			continue;
@@ -1350,52 +1291,73 @@ static int hem_list_alloc_root_bt(struct hns_roce_dev *hr_dev,
 		/* if hopnum is 0 or 1, cut a new fake hem from the root bt
 		 * which's address share to all regions.
 		 */
-		if (hem_list_is_bottom_bt(r->hopnum, 0)) {
-			hem = hem_list_alloc_item(hr_dev, r->offset,
-						  r->offset + r->count - 1,
-						  r->count, false, 0);
-			if (!hem) {
-				ret = -ENOMEM;
-				goto err_exit;
-			}
-			hem_list_assign_bt(hr_dev, hem, cpu_base, phy_base);
-			list_add(&hem->list, &temp_list[i]);
-			list_add(&hem->sibling, &temp_btm);
-			total += r->count;
-		} else {
-			step = hem_list_calc_ba_range(r->hopnum, 1, unit);
-			if (step < 1) {
-				ret = -EINVAL;
-				goto err_exit;
-			}
-			/* if exist mid bt, link L1 to L0 */
-			list_for_each_entry_safe(hem, temp_hem,
-					  &hem_list->mid_bt[i][1], list) {
-				offset = (hem->start - r->offset) / step *
-					  BA_BYTE_LEN;
-				hem_list_link_bt(hr_dev, cpu_base + offset,
-						 hem->dma_addr);
-				total++;
-			}
-		}
+		if (hem_list_is_bottom_bt(r->hopnum, 0))
+			ret = alloc_fake_root_bt(hr_dev, cpu_base, phy_base, r,
+						 &head->branch[i], &head->leaf);
+		else
+			ret = setup_middle_bt(hr_dev, cpu_base, unit, r,
+					      &hem_list->mid_bt[i][1]);
+
+		if (ret < 0)
+			return ret;
+
+		total += ret;
 	}
 
-	list_splice(&temp_btm, &hem_list->btm_bt);
-	list_splice(&temp_root, &hem_list->root_bt);
+	list_splice(&head->leaf, &hem_list->btm_bt);
+	list_splice(&head->root, &hem_list->root_bt);
 	for (i = 0; i < region_cnt; i++)
-		list_splice(&temp_list[i], &hem_list->mid_bt[i][0]);
+		list_splice(&head->branch[i], &hem_list->mid_bt[i][0]);
 
 	return 0;
+}
 
-err_exit:
+static int hem_list_alloc_root_bt(struct hns_roce_dev *hr_dev,
+				  struct hns_roce_hem_list *hem_list, int unit,
+				  const struct hns_roce_buf_region *regions,
+				  int region_cnt)
+{
+	struct hns_roce_hem_item *root_hem;
+	struct hns_roce_hem_head head;
+	int max_ba_num;
+	int ret;
+	int i;
+
+	root_hem = hem_list_search_item(&hem_list->root_bt, regions[0].offset);
+	if (root_hem)
+		return 0;
+
+	max_ba_num = 0;
+	root_hem = alloc_root_hem(hr_dev, unit, &max_ba_num, regions,
+				  region_cnt);
+	if (IS_ERR(root_hem))
+		return PTR_ERR(root_hem);
+
+	/* List head for storing all allocated HEM items */
+	INIT_LIST_HEAD(&head.root);
+	INIT_LIST_HEAD(&head.leaf);
 	for (i = 0; i < region_cnt; i++)
-		hem_list_free_all(hr_dev, &temp_list[i], false);
+		INIT_LIST_HEAD(&head.branch[i]);
 
-	hem_list_free_all(hr_dev, &temp_root, true);
+	hem_list->root_ba = root_hem->dma_addr;
+	list_add(&root_hem->list, &head.root);
+	ret = setup_root_hem(hr_dev, hem_list, unit, max_ba_num, &head, regions,
+			     region_cnt);
+	if (ret) {
+		for (i = 0; i < region_cnt; i++)
+			hem_list_free_all(hr_dev, &head.branch[i]);
+
+		hem_list_free_all(hr_dev, &head.root);
+	}
 
 	return ret;
 }
 
+/* This is the bottom bt pages number of a 100G MR on 4K OS, assuming
+ * the bt page size is not expanded by cal_best_bt_pg_sz()
+ */
+#define RESCHED_LOOP_CNT_THRESHOLD_ON_4K 12800
+
 /* construct the base address table and link them by address hop config */
 int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev,
 			      struct hns_roce_hem_list *hem_list,
@@ -1404,8 +1366,9 @@ int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev,
 {
 	const struct hns_roce_buf_region *r;
 	int ofs, end;
-	int ret = 0;
+	int loop;
 	int unit;
+	int ret;
 	int i;
 
 	if (region_cnt > HNS_ROCE_MAX_BT_REGION) {
@@ -1421,13 +1384,16 @@ int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev,
 			continue;
 
 		end = r->offset + r->count;
-		for (ofs = r->offset; ofs < end; ofs += unit) {
+		for (ofs = r->offset, loop = 1; ofs < end; ofs += unit, loop++) {
+			if (!(loop % RESCHED_LOOP_CNT_THRESHOLD_ON_4K))
+				cond_resched();
+
 			ret = hem_list_alloc_mid_bt(hr_dev, r, unit, ofs,
 						    hem_list->mid_bt[i],
 						    &hem_list->btm_bt);
 			if (ret) {
 				dev_err(hr_dev->dev,
-					"alloc hem trunk fail ret=%d!\n", ret);
+					"alloc hem trunk fail ret = %d!\n", ret);
 				goto err_alloc;
 			}
 		}
@@ -1436,7 +1402,7 @@ int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev,
 	ret = hem_list_alloc_root_bt(hr_dev, hem_list, unit, regions,
 				     region_cnt);
 	if (ret)
-		dev_err(hr_dev->dev, "alloc hem root fail ret=%d!\n", ret);
+		dev_err(hr_dev->dev, "alloc hem root fail ret = %d!\n", ret);
 	else
 		return 0;
 
@@ -1453,10 +1419,9 @@ void hns_roce_hem_list_release(struct hns_roce_dev *hr_dev,
 
 	for (i = 0; i < HNS_ROCE_MAX_BT_REGION; i++)
 		for (j = 0; j < HNS_ROCE_MAX_BT_LEVEL; j++)
-			hem_list_free_all(hr_dev, &hem_list->mid_bt[i][j],
-					  j != 0);
+			hem_list_free_all(hr_dev, &hem_list->mid_bt[i][j]);
 
-	hem_list_free_all(hr_dev, &hem_list->root_bt, true);
+	hem_list_free_all(hr_dev, &hem_list->root_bt);
 	INIT_LIST_HEAD(&hem_list->btm_bt);
 	hem_list->root_ba = 0;
 }
@@ -1474,19 +1439,22 @@ void hns_roce_hem_list_init(struct hns_roce_hem_list *hem_list)
 
 void *hns_roce_hem_list_find_mtt(struct hns_roce_dev *hr_dev,
 				 struct hns_roce_hem_list *hem_list,
-				 int offset, int *mtt_cnt, u64 *phy_addr)
+				 int offset, int *mtt_cnt)
 {
 	struct list_head *head = &hem_list->btm_bt;
-	struct roce_hem_item *hem, *temp_hem;
+	struct hns_roce_hem_item *hem, *temp_hem;
 	void *cpu_base = NULL;
-	u64 phy_base = 0;
+	int loop = 1;
 	int nr = 0;
 
 	list_for_each_entry_safe(hem, temp_hem, head, sibling) {
+		if (!(loop % RESCHED_LOOP_CNT_THRESHOLD_ON_4K))
+			cond_resched();
+		loop++;
+
 		if (hem_list_page_is_in_range(hem, offset)) {
 			nr = offset - hem->start;
 			cpu_base = hem->addr + nr * BA_BYTE_LEN;
-			phy_base = hem->dma_addr + nr * BA_BYTE_LEN;
 			nr = hem->end + 1 - offset;
 			break;
 		}
@@ -1495,8 +1463,5 @@ void *hns_roce_hem_list_find_mtt(struct hns_roce_dev *hr_dev,
 	if (mtt_cnt)
 		*mtt_cnt = nr;
 
-	if (phy_addr)
-		*phy_addr = phy_base;
-
 	return cpu_base;
 }