1 files changed, 672 insertions, 27 deletions

diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c
index c5104e0b2916..8388ae25640c 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.c
@@ -42,8 +42,162 @@
 #define DMA_ADDR_T_SHIFT		12
 #define BT_BA_SHIFT			32
 
-struct hns_roce_hem *hns_roce_alloc_hem(struct hns_roce_dev *hr_dev, int npages,
-					gfp_t gfp_mask)
+bool hns_roce_check_whether_mhop(struct hns_roce_dev *hr_dev, u32 type)
+{
+	if ((hr_dev->caps.qpc_hop_num && type == HEM_TYPE_QPC) ||
+	    (hr_dev->caps.mpt_hop_num && type == HEM_TYPE_MTPT) ||
+	    (hr_dev->caps.cqc_hop_num && type == HEM_TYPE_CQC) ||
+	    (hr_dev->caps.srqc_hop_num && type == HEM_TYPE_SRQC) ||
+	    (hr_dev->caps.cqe_hop_num && type == HEM_TYPE_CQE) ||
+	    (hr_dev->caps.mtt_hop_num && type == HEM_TYPE_MTT))
+		return true;
+
+	return false;
+}
+EXPORT_SYMBOL_GPL(hns_roce_check_whether_mhop);
+
+static bool hns_roce_check_hem_null(struct hns_roce_hem **hem, u64 start_idx,
+			    u32 bt_chunk_num)
+{
+	int i;
+
+	for (i = 0; i < bt_chunk_num; i++)
+		if (hem[start_idx + i])
+			return false;
+
+	return true;
+}
+
+static bool hns_roce_check_bt_null(u64 **bt, u64 start_idx, u32 bt_chunk_num)
+{
+	int i;
+
+	for (i = 0; i < bt_chunk_num; i++)
+		if (bt[start_idx + i])
+			return false;
+
+	return true;
+}
+
+static int hns_roce_get_bt_num(u32 table_type, u32 hop_num)
+{
+	if (check_whether_bt_num_3(table_type, hop_num))
+		return 3;
+	else if (check_whether_bt_num_2(table_type, hop_num))
+		return 2;
+	else if (check_whether_bt_num_1(table_type, hop_num))
+		return 1;
+	else
+		return 0;
+}
+
+int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
+			   struct hns_roce_hem_table *table, unsigned long *obj,
+			   struct hns_roce_hem_mhop *mhop)
+{
+	struct device *dev = hr_dev->dev;
+	u32 chunk_ba_num;
+	u32 table_idx;
+	u32 bt_num;
+	u32 chunk_size;
+
+	switch (table->type) {
+	case HEM_TYPE_QPC:
+		mhop->buf_chunk_size = 1 << (hr_dev->caps.qpc_buf_pg_sz
+					     + PAGE_SHIFT);
+		mhop->bt_chunk_size = 1 << (hr_dev->caps.qpc_ba_pg_sz
+					     + PAGE_SHIFT);
+		mhop->ba_l0_num = hr_dev->caps.qpc_bt_num;
+		mhop->hop_num = hr_dev->caps.qpc_hop_num;
+		break;
+	case HEM_TYPE_MTPT:
+		mhop->buf_chunk_size = 1 << (hr_dev->caps.mpt_buf_pg_sz
+					     + PAGE_SHIFT);
+		mhop->bt_chunk_size = 1 << (hr_dev->caps.mpt_ba_pg_sz
+					     + PAGE_SHIFT);
+		mhop->ba_l0_num = hr_dev->caps.mpt_bt_num;
+		mhop->hop_num = hr_dev->caps.mpt_hop_num;
+		break;
+	case HEM_TYPE_CQC:
+		mhop->buf_chunk_size = 1 << (hr_dev->caps.cqc_buf_pg_sz
+					     + PAGE_SHIFT);
+		mhop->bt_chunk_size = 1 << (hr_dev->caps.cqc_ba_pg_sz
+					    + PAGE_SHIFT);
+		mhop->ba_l0_num = hr_dev->caps.cqc_bt_num;
+		mhop->hop_num = hr_dev->caps.cqc_hop_num;
+		break;
+	case HEM_TYPE_SRQC:
+		mhop->buf_chunk_size = 1 << (hr_dev->caps.srqc_buf_pg_sz
+					     + PAGE_SHIFT);
+		mhop->bt_chunk_size = 1 << (hr_dev->caps.srqc_ba_pg_sz
+					     + PAGE_SHIFT);
+		mhop->ba_l0_num = hr_dev->caps.srqc_bt_num;
+		mhop->hop_num = hr_dev->caps.srqc_hop_num;
+		break;
+	case HEM_TYPE_MTT:
+		mhop->buf_chunk_size = 1 << (hr_dev->caps.mtt_buf_pg_sz
+					     + PAGE_SHIFT);
+		mhop->bt_chunk_size = 1 << (hr_dev->caps.mtt_ba_pg_sz
+					     + PAGE_SHIFT);
+		mhop->ba_l0_num = mhop->bt_chunk_size / 8;
+		mhop->hop_num = hr_dev->caps.mtt_hop_num;
+		break;
+	case HEM_TYPE_CQE:
+		mhop->buf_chunk_size = 1 << (hr_dev->caps.cqe_buf_pg_sz
+					     + PAGE_SHIFT);
+		mhop->bt_chunk_size = 1 << (hr_dev->caps.cqe_ba_pg_sz
+					     + PAGE_SHIFT);
+		mhop->ba_l0_num = mhop->bt_chunk_size / 8;
+		mhop->hop_num = hr_dev->caps.cqe_hop_num;
+		break;
+	default:
+		dev_err(dev, "Table %d not support multi-hop addressing!\n",
+			 table->type);
+		return -EINVAL;
+	}
+
+	if (!obj)
+		return 0;
+
+	/*
+	 * QPC/MTPT/CQC/SRQC alloc hem for buffer pages.
+	 * MTT/CQE alloc hem for bt pages.
+	 */
+	bt_num = hns_roce_get_bt_num(table->type, mhop->hop_num);
+	chunk_ba_num = mhop->bt_chunk_size / 8;
+	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);
+	switch (bt_num) {
+	case 3:
+		mhop->l2_idx = table_idx & (chunk_ba_num - 1);
+		mhop->l1_idx = table_idx / chunk_ba_num & (chunk_ba_num - 1);
+		mhop->l0_idx = table_idx / chunk_ba_num / chunk_ba_num;
+		break;
+	case 2:
+		mhop->l1_idx = table_idx & (chunk_ba_num - 1);
+		mhop->l0_idx = table_idx / chunk_ba_num;
+		break;
+	case 1:
+		mhop->l0_idx = table_idx;
+		break;
+	default:
+		dev_err(dev, "Table %d not support hop_num = %d!\n",
+			     table->type, mhop->hop_num);
+		return -EINVAL;
+	}
+	if (mhop->l0_idx >= mhop->ba_l0_num)
+		mhop->l0_idx %= mhop->ba_l0_num;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(hns_roce_calc_hem_mhop);
+
+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)
 {
 	struct hns_roce_hem_chunk *chunk = NULL;
 	struct hns_roce_hem *hem;
@@ -61,7 +215,7 @@ struct hns_roce_hem *hns_roce_alloc_hem(struct hns_roce_dev *hr_dev, int npages,
 	hem->refcount = 0;
 	INIT_LIST_HEAD(&hem->chunk_list);
 
-	order = get_order(HNS_ROCE_HEM_ALLOC_SIZE);
+	order = get_order(hem_alloc_size);
 
 	while (npages > 0) {
 		if (!chunk) {
@@ -84,7 +238,7 @@ struct hns_roce_hem *hns_roce_alloc_hem(struct hns_roce_dev *hr_dev, int npages,
 		 * memory, directly return fail.
 		 */
 		mem = &chunk->mem[chunk->npages];
-		buf = dma_alloc_coherent(&hr_dev->pdev->dev, PAGE_SIZE << order,
+		buf = dma_alloc_coherent(hr_dev->dev, PAGE_SIZE << order,
 				&sg_dma_address(mem), gfp_mask);
 		if (!buf)
 			goto fail;
@@ -115,7 +269,7 @@ void hns_roce_free_hem(struct hns_roce_dev *hr_dev, struct hns_roce_hem *hem)
 
 	list_for_each_entry_safe(chunk, tmp, &hem->chunk_list, list) {
 		for (i = 0; i < chunk->npages; ++i)
-			dma_free_coherent(&hr_dev->pdev->dev,
+			dma_free_coherent(hr_dev->dev,
 				   chunk->mem[i].length,
 				   lowmem_page_address(sg_page(&chunk->mem[i])),
 				   sg_dma_address(&chunk->mem[i]));
@@ -128,8 +282,8 @@ void hns_roce_free_hem(struct hns_roce_dev *hr_dev, struct hns_roce_hem *hem)
 static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
 			    struct hns_roce_hem_table *table, unsigned long obj)
 {
-	struct device *dev = &hr_dev->pdev->dev;
 	spinlock_t *lock = &hr_dev->bt_cmd_lock;
+	struct device *dev = hr_dev->dev;
 	unsigned long end = 0;
 	unsigned long flags;
 	struct hns_roce_hem_iter iter;
@@ -209,13 +363,184 @@ static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
 	return ret;
 }
 
+static int hns_roce_table_mhop_get(struct hns_roce_dev *hr_dev,
+				   struct hns_roce_hem_table *table,
+				   unsigned long obj)
+{
+	struct device *dev = hr_dev->dev;
+	struct hns_roce_hem_mhop mhop;
+	struct hns_roce_hem_iter iter;
+	u32 buf_chunk_size;
+	u32 bt_chunk_size;
+	u32 chunk_ba_num;
+	u32 hop_num;
+	u32 size;
+	u32 bt_num;
+	u64 hem_idx;
+	u64 bt_l1_idx = 0;
+	u64 bt_l0_idx = 0;
+	u64 bt_ba;
+	unsigned long mhop_obj = obj;
+	int bt_l1_allocated = 0;
+	int bt_l0_allocated = 0;
+	int step_idx;
+	int ret;
+
+	ret = hns_roce_calc_hem_mhop(hr_dev, table, &mhop_obj, &mhop);
+	if (ret)
+		return ret;
+
+	buf_chunk_size = mhop.buf_chunk_size;
+	bt_chunk_size = mhop.bt_chunk_size;
+	hop_num = mhop.hop_num;
+	chunk_ba_num = bt_chunk_size / 8;
+
+	bt_num = hns_roce_get_bt_num(table->type, hop_num);
+	switch (bt_num) {
+	case 3:
+		hem_idx = mhop.l0_idx * chunk_ba_num * chunk_ba_num +
+			  mhop.l1_idx * chunk_ba_num + mhop.l2_idx;
+		bt_l1_idx = mhop.l0_idx * chunk_ba_num + mhop.l1_idx;
+		bt_l0_idx = mhop.l0_idx;
+		break;
+	case 2:
+		hem_idx = mhop.l0_idx * chunk_ba_num + mhop.l1_idx;
+		bt_l0_idx = mhop.l0_idx;
+		break;
+	case 1:
+		hem_idx = mhop.l0_idx;
+		break;
+	default:
+		dev_err(dev, "Table %d not support hop_num = %d!\n",
+			     table->type, hop_num);
+		return -EINVAL;
+	}
+
+	mutex_lock(&table->mutex);
+
+	if (table->hem[hem_idx]) {
+		++table->hem[hem_idx]->refcount;
+		goto out;
+	}
+
+	/* alloc L1 BA's chunk */
+	if ((check_whether_bt_num_3(table->type, hop_num) ||
+		check_whether_bt_num_2(table->type, hop_num)) &&
+		!table->bt_l0[bt_l0_idx]) {
+		table->bt_l0[bt_l0_idx] = dma_alloc_coherent(dev, bt_chunk_size,
+					    &(table->bt_l0_dma_addr[bt_l0_idx]),
+					    GFP_KERNEL);
+		if (!table->bt_l0[bt_l0_idx]) {
+			ret = -ENOMEM;
+			goto out;
+		}
+		bt_l0_allocated = 1;
+
+		/* set base address to hardware */
+		if (table->type < HEM_TYPE_MTT) {
+			step_idx = 0;
+			if (hr_dev->hw->set_hem(hr_dev, table, obj, step_idx)) {
+				ret = -ENODEV;
+				dev_err(dev, "set HEM base address to HW failed!\n");
+				goto err_dma_alloc_l1;
+			}
+		}
+	}
+
+	/* alloc L2 BA's chunk */
+	if (check_whether_bt_num_3(table->type, hop_num) &&
+	    !table->bt_l1[bt_l1_idx])  {
+		table->bt_l1[bt_l1_idx] = dma_alloc_coherent(dev, bt_chunk_size,
+					    &(table->bt_l1_dma_addr[bt_l1_idx]),
+					    GFP_KERNEL);
+		if (!table->bt_l1[bt_l1_idx]) {
+			ret = -ENOMEM;
+			goto err_dma_alloc_l1;
+		}
+		bt_l1_allocated = 1;
+		*(table->bt_l0[bt_l0_idx] + mhop.l1_idx) =
+					       table->bt_l1_dma_addr[bt_l1_idx];
+
+		/* set base address to hardware */
+		step_idx = 1;
+		if (hr_dev->hw->set_hem(hr_dev, table, obj, step_idx)) {
+			ret = -ENODEV;
+			dev_err(dev, "set HEM base address to HW failed!\n");
+			goto err_alloc_hem_buf;
+		}
+	}
+
+	/*
+	 * alloc buffer space chunk for QPC/MTPT/CQC/SRQC.
+	 * alloc bt space chunk for MTT/CQE.
+	 */
+	size = table->type < HEM_TYPE_MTT ? buf_chunk_size : bt_chunk_size;
+	table->hem[hem_idx] = hns_roce_alloc_hem(hr_dev,
+						size >> PAGE_SHIFT,
+						size,
+						(table->lowmem ? GFP_KERNEL :
+						GFP_HIGHUSER) | __GFP_NOWARN);
+	if (!table->hem[hem_idx]) {
+		ret = -ENOMEM;
+		goto err_alloc_hem_buf;
+	}
+
+	hns_roce_hem_first(table->hem[hem_idx], &iter);
+	bt_ba = hns_roce_hem_addr(&iter);
+
+	if (table->type < HEM_TYPE_MTT) {
+		if (hop_num == 2) {
+			*(table->bt_l1[bt_l1_idx] + mhop.l2_idx) = bt_ba;
+			step_idx = 2;
+		} else if (hop_num == 1) {
+			*(table->bt_l0[bt_l0_idx] + mhop.l1_idx) = bt_ba;
+			step_idx = 1;
+		} else if (hop_num == HNS_ROCE_HOP_NUM_0) {
+			step_idx = 0;
+		}
+
+		/* set HEM base address to hardware */
+		if (hr_dev->hw->set_hem(hr_dev, table, obj, step_idx)) {
+			ret = -ENODEV;
+			dev_err(dev, "set HEM base address to HW failed!\n");
+			goto err_alloc_hem_buf;
+		}
+	} else if (hop_num == 2) {
+		*(table->bt_l0[bt_l0_idx] + mhop.l1_idx) = bt_ba;
+	}
+
+	++table->hem[hem_idx]->refcount;
+	goto out;
+
+err_alloc_hem_buf:
+	if (bt_l1_allocated) {
+		dma_free_coherent(dev, bt_chunk_size, table->bt_l1[bt_l1_idx],
+				  table->bt_l1_dma_addr[bt_l1_idx]);
+		table->bt_l1[bt_l1_idx] = NULL;
+	}
+
+err_dma_alloc_l1:
+	if (bt_l0_allocated) {
+		dma_free_coherent(dev, bt_chunk_size, table->bt_l0[bt_l0_idx],
+				  table->bt_l0_dma_addr[bt_l0_idx]);
+		table->bt_l0[bt_l0_idx] = NULL;
+	}
+
+out:
+	mutex_unlock(&table->mutex);
+	return ret;
+}
+
 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->pdev->dev;
+	struct device *dev = hr_dev->dev;
 	int ret = 0;
 	unsigned long i;
 
+	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)) / (HNS_ROCE_TABLE_CHUNK_SIZE /
 	     table->obj_size);
 
@@ -228,6 +553,7 @@ int hns_roce_table_get(struct hns_roce_dev *hr_dev,
 
 	table->hem[i] = hns_roce_alloc_hem(hr_dev,
 				       HNS_ROCE_TABLE_CHUNK_SIZE >> PAGE_SHIFT,
+				       HNS_ROCE_HEM_ALLOC_SIZE,
 				       (table->lowmem ? GFP_KERNEL :
 					GFP_HIGHUSER) | __GFP_NOWARN);
 	if (!table->hem[i]) {
@@ -237,6 +563,8 @@ int hns_roce_table_get(struct hns_roce_dev *hr_dev,
 
 	/* Set HEM base address(128K/page, pa) to Hardware */
 	if (hns_roce_set_hem(hr_dev, table, obj)) {
+		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");
 		goto out;
@@ -248,12 +576,131 @@ out:
 	return ret;
 }
 
+static void hns_roce_table_mhop_put(struct hns_roce_dev *hr_dev,
+				    struct hns_roce_hem_table *table,
+				    unsigned long obj,
+				    int check_refcount)
+{
+	struct device *dev = hr_dev->dev;
+	struct hns_roce_hem_mhop mhop;
+	unsigned long mhop_obj = obj;
+	u32 bt_chunk_size;
+	u32 chunk_ba_num;
+	u32 hop_num;
+	u32 start_idx;
+	u32 bt_num;
+	u64 hem_idx;
+	u64 bt_l1_idx = 0;
+	int ret;
+
+	ret = hns_roce_calc_hem_mhop(hr_dev, table, &mhop_obj, &mhop);
+	if (ret)
+		return;
+
+	bt_chunk_size = mhop.bt_chunk_size;
+	hop_num = mhop.hop_num;
+	chunk_ba_num = bt_chunk_size / 8;
+
+	bt_num = hns_roce_get_bt_num(table->type, hop_num);
+	switch (bt_num) {
+	case 3:
+		hem_idx = mhop.l0_idx * chunk_ba_num * chunk_ba_num +
+			  mhop.l1_idx * chunk_ba_num + mhop.l2_idx;
+		bt_l1_idx = mhop.l0_idx * chunk_ba_num + mhop.l1_idx;
+		break;
+	case 2:
+		hem_idx = mhop.l0_idx * chunk_ba_num + mhop.l1_idx;
+		break;
+	case 1:
+		hem_idx = mhop.l0_idx;
+		break;
+	default:
+		dev_err(dev, "Table %d not support hop_num = %d!\n",
+			     table->type, hop_num);
+		return;
+	}
+
+	mutex_lock(&table->mutex);
+
+	if (check_refcount && (--table->hem[hem_idx]->refcount > 0)) {
+		mutex_unlock(&table->mutex);
+		return;
+	}
+
+	if (table->type < HEM_TYPE_MTT && hop_num == 1) {
+		if (hr_dev->hw->clear_hem(hr_dev, table, obj, 1))
+			dev_warn(dev, "Clear HEM base address failed.\n");
+	} else if (table->type < HEM_TYPE_MTT && hop_num == 2) {
+		if (hr_dev->hw->clear_hem(hr_dev, table, obj, 2))
+			dev_warn(dev, "Clear HEM base address failed.\n");
+	} else if (table->type < HEM_TYPE_MTT &&
+		   hop_num == HNS_ROCE_HOP_NUM_0) {
+		if (hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
+			dev_warn(dev, "Clear HEM base address failed.\n");
+	}
+
+	/*
+	 * free buffer space chunk for QPC/MTPT/CQC/SRQC.
+	 * free bt space chunk for MTT/CQE.
+	 */
+	hns_roce_free_hem(hr_dev, table->hem[hem_idx]);
+	table->hem[hem_idx] = NULL;
+
+	if (check_whether_bt_num_2(table->type, hop_num)) {
+		start_idx = mhop.l0_idx * chunk_ba_num;
+		if (hns_roce_check_hem_null(table->hem, start_idx,
+					    chunk_ba_num)) {
+			if (table->type < HEM_TYPE_MTT &&
+			    hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
+				dev_warn(dev, "Clear HEM base address failed.\n");
+
+			dma_free_coherent(dev, bt_chunk_size,
+					  table->bt_l0[mhop.l0_idx],
+					  table->bt_l0_dma_addr[mhop.l0_idx]);
+			table->bt_l0[mhop.l0_idx] = NULL;
+		}
+	} else if (check_whether_bt_num_3(table->type, hop_num)) {
+		start_idx = mhop.l0_idx * chunk_ba_num * chunk_ba_num +
+			    mhop.l1_idx * chunk_ba_num;
+		if (hns_roce_check_hem_null(table->hem, start_idx,
+					    chunk_ba_num)) {
+			if (hr_dev->hw->clear_hem(hr_dev, table, obj, 1))
+				dev_warn(dev, "Clear HEM base address failed.\n");
+
+			dma_free_coherent(dev, bt_chunk_size,
+					  table->bt_l1[bt_l1_idx],
+					  table->bt_l1_dma_addr[bt_l1_idx]);
+			table->bt_l1[bt_l1_idx] = NULL;
+
+			start_idx = mhop.l0_idx * chunk_ba_num;
+			if (hns_roce_check_bt_null(table->bt_l1, start_idx,
+						   chunk_ba_num)) {
+				if (hr_dev->hw->clear_hem(hr_dev, table, obj,
+							  0))
+					dev_warn(dev, "Clear HEM base address failed.\n");
+
+				dma_free_coherent(dev, bt_chunk_size,
+					    table->bt_l0[mhop.l0_idx],
+					    table->bt_l0_dma_addr[mhop.l0_idx]);
+				table->bt_l0[mhop.l0_idx] = NULL;
+			}
+		}
+	}
+
+	mutex_unlock(&table->mutex);
+}
+
 void hns_roce_table_put(struct hns_roce_dev *hr_dev,
 			struct hns_roce_hem_table *table, unsigned long obj)
 {
-	struct device *dev = &hr_dev->pdev->dev;
+	struct device *dev = hr_dev->dev;
 	unsigned long i;
 
+	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)) /
 	    (HNS_ROCE_TABLE_CHUNK_SIZE / table->obj_size);
 
@@ -261,7 +708,7 @@ void hns_roce_table_put(struct hns_roce_dev *hr_dev,
 
 	if (--table->hem[i]->refcount == 0) {
 		/* Clear HEM base address */
-		if (hr_dev->hw->clear_hem(hr_dev, table, obj))
+		if (hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
 			dev_warn(dev, "Clear HEM base address failed.\n");
 
 		hns_roce_free_hem(hr_dev, table->hem[i]);
@@ -271,23 +718,46 @@ void hns_roce_table_put(struct hns_roce_dev *hr_dev,
 	mutex_unlock(&table->mutex);
 }
 
-void *hns_roce_table_find(struct hns_roce_hem_table *table, unsigned long obj,
-			  dma_addr_t *dma_handle)
+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;
-	unsigned long idx;
-	int i;
-	int offset, dma_offset;
+	struct hns_roce_hem_mhop mhop;
 	struct hns_roce_hem *hem;
 	struct page *page = NULL;
+	unsigned long mhop_obj = obj;
+	unsigned long idx;
+	int offset, dma_offset;
+	int i, j;
+	u32 hem_idx = 0;
 
 	if (!table->lowmem)
 		return NULL;
 
 	mutex_lock(&table->mutex);
-	idx = (obj & (table->num_obj - 1)) * table->obj_size;
-	hem = table->hem[idx / HNS_ROCE_TABLE_CHUNK_SIZE];
-	dma_offset = offset = idx % HNS_ROCE_TABLE_CHUNK_SIZE;
+
+	if (!hns_roce_check_whether_mhop(hr_dev, table->type)) {
+		idx = (obj & (table->num_obj - 1)) * table->obj_size;
+		hem = table->hem[idx / HNS_ROCE_TABLE_CHUNK_SIZE];
+		dma_offset = offset = idx % HNS_ROCE_TABLE_CHUNK_SIZE;
+	} else {
+		hns_roce_calc_hem_mhop(hr_dev, table, &mhop_obj, &mhop);
+		/* mtt mhop */
+		i = mhop.l0_idx;
+		j = mhop.l1_idx;
+		if (mhop.hop_num == 2)
+			hem_idx = i * (mhop.bt_chunk_size / 8) + j;
+		else if (mhop.hop_num == 1 ||
+			 mhop.hop_num == HNS_ROCE_HOP_NUM_0)
+			hem_idx = i;
+
+		hem = table->hem[hem_idx];
+		dma_offset = offset = (obj & (table->num_obj - 1)) *
+				       table->obj_size % mhop.bt_chunk_size;
+		if (mhop.hop_num == 2)
+			dma_offset = offset = 0;
+	}
 
 	if (!hem)
 		goto out;
@@ -314,14 +784,21 @@ out:
 	mutex_unlock(&table->mutex);
 	return page ? lowmem_page_address(page) + offset : NULL;
 }
+EXPORT_SYMBOL_GPL(hns_roce_table_find);
 
 int hns_roce_table_get_range(struct hns_roce_dev *hr_dev,
 			     struct hns_roce_hem_table *table,
 			     unsigned long start, unsigned long end)
 {
+	struct hns_roce_hem_mhop mhop;
 	unsigned long inc = HNS_ROCE_TABLE_CHUNK_SIZE / table->obj_size;
-	unsigned long i = 0;
-	int ret = 0;
+	unsigned long i;
+	int ret;
+
+	if (hns_roce_check_whether_mhop(hr_dev, table->type)) {
+		hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop);
+		inc = mhop.bt_chunk_size / table->obj_size;
+	}
 
 	/* Allocate MTT entry memory according to chunk(128K) */
 	for (i = start; i <= end; i += inc) {
@@ -344,10 +821,17 @@ void hns_roce_table_put_range(struct hns_roce_dev *hr_dev,
 			      struct hns_roce_hem_table *table,
 			      unsigned long start, unsigned long end)
 {
+	struct hns_roce_hem_mhop mhop;
+	unsigned long inc = HNS_ROCE_TABLE_CHUNK_SIZE / table->obj_size;
 	unsigned long i;
 
+	if (hns_roce_check_whether_mhop(hr_dev, table->type)) {
+		hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop);
+		inc = mhop.bt_chunk_size / table->obj_size;
+	}
+
 	for (i = start; i <= end;
-		i += HNS_ROCE_TABLE_CHUNK_SIZE / table->obj_size)
+		i += inc)
 		hns_roce_table_put(hr_dev, table, i);
 }
 
@@ -356,15 +840,119 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 			    unsigned long obj_size, unsigned long nobj,
 			    int use_lowmem)
 {
+	struct device *dev = hr_dev->dev;
 	unsigned long obj_per_chunk;
 	unsigned long num_hem;
 
-	obj_per_chunk = HNS_ROCE_TABLE_CHUNK_SIZE / obj_size;
-	num_hem = (nobj + obj_per_chunk - 1) / obj_per_chunk;
+	if (!hns_roce_check_whether_mhop(hr_dev, type)) {
+		obj_per_chunk = HNS_ROCE_TABLE_CHUNK_SIZE / obj_size;
+		num_hem = (nobj + obj_per_chunk - 1) / obj_per_chunk;
+
+		table->hem = kcalloc(num_hem, sizeof(*table->hem), GFP_KERNEL);
+		if (!table->hem)
+			return -ENOMEM;
+	} else {
+		unsigned long buf_chunk_size;
+		unsigned long bt_chunk_size;
+		unsigned long bt_chunk_num;
+		unsigned long num_bt_l0 = 0;
+		u32 hop_num;
+
+		switch (type) {
+		case HEM_TYPE_QPC:
+			buf_chunk_size = 1 << (hr_dev->caps.qpc_buf_pg_sz
+					+ PAGE_SHIFT);
+			bt_chunk_size = 1 << (hr_dev->caps.qpc_ba_pg_sz
+					+ PAGE_SHIFT);
+			num_bt_l0 = hr_dev->caps.qpc_bt_num;
+			hop_num = hr_dev->caps.qpc_hop_num;
+			break;
+		case HEM_TYPE_MTPT:
+			buf_chunk_size = 1 << (hr_dev->caps.mpt_buf_pg_sz
+					+ PAGE_SHIFT);
+			bt_chunk_size = 1 << (hr_dev->caps.mpt_ba_pg_sz
+					+ PAGE_SHIFT);
+			num_bt_l0 = hr_dev->caps.mpt_bt_num;
+			hop_num = hr_dev->caps.mpt_hop_num;
+			break;
+		case HEM_TYPE_CQC:
+			buf_chunk_size = 1 << (hr_dev->caps.cqc_buf_pg_sz
+					+ PAGE_SHIFT);
+			bt_chunk_size = 1 << (hr_dev->caps.cqc_ba_pg_sz
+					+ PAGE_SHIFT);
+			num_bt_l0 = hr_dev->caps.cqc_bt_num;
+			hop_num = hr_dev->caps.cqc_hop_num;
+			break;
+		case HEM_TYPE_SRQC:
+			buf_chunk_size = 1 << (hr_dev->caps.srqc_buf_pg_sz
+					+ PAGE_SHIFT);
+			bt_chunk_size = 1 << (hr_dev->caps.srqc_ba_pg_sz
+					+ PAGE_SHIFT);
+			num_bt_l0 = hr_dev->caps.srqc_bt_num;
+			hop_num = hr_dev->caps.srqc_hop_num;
+			break;
+		case HEM_TYPE_MTT:
+			buf_chunk_size = 1 << (hr_dev->caps.mtt_ba_pg_sz
+					+ PAGE_SHIFT);
+			bt_chunk_size = buf_chunk_size;
+			hop_num = hr_dev->caps.mtt_hop_num;
+			break;
+		case HEM_TYPE_CQE:
+			buf_chunk_size = 1 << (hr_dev->caps.cqe_ba_pg_sz
+					+ PAGE_SHIFT);
+			bt_chunk_size = buf_chunk_size;
+			hop_num = hr_dev->caps.cqe_hop_num;
+			break;
+		default:
+			dev_err(dev,
+			  "Table %d not support to init hem table here!\n",
+			  type);
+			return -EINVAL;
+		}
+		obj_per_chunk = buf_chunk_size / obj_size;
+		num_hem = (nobj + obj_per_chunk - 1) / obj_per_chunk;
+		bt_chunk_num = bt_chunk_size / 8;
+		if (table->type >= HEM_TYPE_MTT)
+			num_bt_l0 = bt_chunk_num;
+
+		table->hem = kcalloc(num_hem, sizeof(*table->hem),
+					 GFP_KERNEL);
+		if (!table->hem)
+			goto err_kcalloc_hem_buf;
+
+		if (check_whether_bt_num_3(table->type, hop_num)) {
+			unsigned long num_bt_l1;
+
+			num_bt_l1 = (num_hem + bt_chunk_num - 1) /
+					     bt_chunk_num;
+			table->bt_l1 = kcalloc(num_bt_l1,
+					       sizeof(*table->bt_l1),
+					       GFP_KERNEL);
+			if (!table->bt_l1)
+				goto err_kcalloc_bt_l1;
+
+			table->bt_l1_dma_addr = kcalloc(num_bt_l1,
+						 sizeof(*table->bt_l1_dma_addr),
+						 GFP_KERNEL);
+
+			if (!table->bt_l1_dma_addr)
+				goto err_kcalloc_l1_dma;
+		}
 
-	table->hem = kcalloc(num_hem, sizeof(*table->hem), GFP_KERNEL);
-	if (!table->hem)
-		return -ENOMEM;
+		if (check_whether_bt_num_2(table->type, hop_num) ||
+			check_whether_bt_num_3(table->type, hop_num)) {
+			table->bt_l0 = kcalloc(num_bt_l0, sizeof(*table->bt_l0),
+					       GFP_KERNEL);
+			if (!table->bt_l0)
+				goto err_kcalloc_bt_l0;
+
+			table->bt_l0_dma_addr = kcalloc(num_bt_l0,
+						 sizeof(*table->bt_l0_dma_addr),
+						 GFP_KERNEL);
+			if (!table->bt_l0_dma_addr)
+				goto err_kcalloc_l0_dma;
+		}
+	}
 
 	table->type = type;
 	table->num_hem = num_hem;
@@ -374,18 +962,72 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 	mutex_init(&table->mutex);
 
 	return 0;
+
+err_kcalloc_l0_dma:
+	kfree(table->bt_l0);
+	table->bt_l0 = NULL;
+
+err_kcalloc_bt_l0:
+	kfree(table->bt_l1_dma_addr);
+	table->bt_l1_dma_addr = NULL;
+
+err_kcalloc_l1_dma:
+	kfree(table->bt_l1);
+	table->bt_l1 = NULL;
+
+err_kcalloc_bt_l1:
+	kfree(table->hem);
+	table->hem = NULL;
+
+err_kcalloc_hem_buf:
+	return -ENOMEM;
+}
+
+static void hns_roce_cleanup_mhop_hem_table(struct hns_roce_dev *hr_dev,
+					    struct hns_roce_hem_table *table)
+{
+	struct hns_roce_hem_mhop mhop;
+	u32 buf_chunk_size;
+	int i;
+	u64 obj;
+
+	hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop);
+	buf_chunk_size = table->type < HEM_TYPE_MTT ? mhop.buf_chunk_size :
+					mhop.bt_chunk_size;
+
+	for (i = 0; i < table->num_hem; ++i) {
+		obj = i * buf_chunk_size / table->obj_size;
+		if (table->hem[i])
+			hns_roce_table_mhop_put(hr_dev, table, obj, 0);
+	}
+
+	kfree(table->hem);
+	table->hem = NULL;
+	kfree(table->bt_l1);
+	table->bt_l1 = NULL;
+	kfree(table->bt_l1_dma_addr);
+	table->bt_l1_dma_addr = NULL;
+	kfree(table->bt_l0);
+	table->bt_l0 = NULL;
+	kfree(table->bt_l0_dma_addr);
+	table->bt_l0_dma_addr = NULL;
 }
 
 void hns_roce_cleanup_hem_table(struct hns_roce_dev *hr_dev,
 				struct hns_roce_hem_table *table)
 {
-	struct device *dev = &hr_dev->pdev->dev;
+	struct device *dev = hr_dev->dev;
 	unsigned long i;
 
+	if (hns_roce_check_whether_mhop(hr_dev, table->type)) {
+		hns_roce_cleanup_mhop_hem_table(hr_dev, table);
+		return;
+	}
+
 	for (i = 0; i < table->num_hem; ++i)
 		if (table->hem[i]) {
 			if (hr_dev->hw->clear_hem(hr_dev, table,
-			    i * HNS_ROCE_TABLE_CHUNK_SIZE / table->obj_size))
+			    i * HNS_ROCE_TABLE_CHUNK_SIZE / table->obj_size, 0))
 				dev_err(dev, "Clear HEM base address failed.\n");
 
 			hns_roce_free_hem(hr_dev, table->hem[i]);
@@ -401,4 +1043,7 @@ void hns_roce_cleanup_hem(struct hns_roce_dev *hr_dev)
 	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);
 	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->mr_table.mtt_table);
+	if (hns_roce_check_whether_mhop(hr_dev, HEM_TYPE_CQE))
+		hns_roce_cleanup_hem_table(hr_dev,
+					   &hr_dev->mr_table.mtt_cqe_table);
 }