diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index 0df180850a75c58740429e6289385eed9bcf3dcb..0d40a930c1922d486a027564caf20dca7ef1e194 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -1401,7 +1401,8 @@ static int bnxt_re_ib_reg(struct bnxt_re_dev *rdev)
 	if (!rdev->is_virtfn)
 		bnxt_re_set_resource_limits(rdev);
 
-	rc = bnxt_qplib_alloc_ctx(rdev->en_dev->pdev, &rdev->qplib_ctx, 0);
+	rc = bnxt_qplib_alloc_ctx(rdev->en_dev->pdev, &rdev->qplib_ctx, 0,
+				  bnxt_qplib_is_chip_gen_p5(&rdev->chip_ctx));
 	if (rc) {
 		pr_err("Failed to allocate QPLIB context: %#x\n", rc);
 		goto disable_rcfw;
diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
index 9c000245a6de96672d2a87040997c17931376f00..c6461e957078829e0c96cff7edb9ae3edd593c39 100644
--- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
+++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
@@ -482,11 +482,13 @@ int bnxt_qplib_init_rcfw(struct bnxt_qplib_rcfw *rcfw,
 	req.log2_dbr_pg_size = cpu_to_le16(PAGE_SHIFT -
 					   RCFW_DBR_BASE_PAGE_SHIFT);
 	/*
-	 * VFs need not setup the HW context area, PF
+	 * Gen P5 devices doesn't require this allocation
+	 * as the L2 driver does the same for RoCE also.
+	 * Also, VFs need not setup the HW context area, PF
 	 * shall setup this area for VF. Skipping the
 	 * HW programming
 	 */
-	if (is_virtfn)
+	if (is_virtfn || bnxt_qplib_is_chip_gen_p5(rcfw->res->cctx))
 		goto skip_ctx_setup;
 
 	level = ctx->qpc_tbl.level;
diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.c b/drivers/infiniband/hw/bnxt_re/qplib_res.c
index 57d4951679cb9da74fec015e431a381df7092915..c8502c2844a273a2ef1b5c44902c1561431779a4 100644
--- a/drivers/infiniband/hw/bnxt_re/qplib_res.c
+++ b/drivers/infiniband/hw/bnxt_re/qplib_res.c
@@ -330,13 +330,13 @@ void bnxt_qplib_free_ctx(struct pci_dev *pdev,
  */
 int bnxt_qplib_alloc_ctx(struct pci_dev *pdev,
 			 struct bnxt_qplib_ctx *ctx,
-			 bool virt_fn)
+			 bool virt_fn, bool is_p5)
 {
 	int i, j, k, rc = 0;
 	int fnz_idx = -1;
 	__le64 **pbl_ptr;
 
-	if (virt_fn)
+	if (virt_fn || is_p5)
 		goto stats_alloc;
 
 	/* QPC Tables */
@@ -762,7 +762,11 @@ static int bnxt_qplib_alloc_stats_ctx(struct pci_dev *pdev,
 {
 	memset(stats, 0, sizeof(*stats));
 	stats->fw_id = -1;
-	stats->size = sizeof(struct ctx_hw_stats);
+	/* 128 byte aligned context memory is required only for 57500.
+	 * However making this unconditional, it does not harm previous
+	 * generation.
+	 */
+	stats->size = ALIGN(sizeof(struct ctx_hw_stats), 128);
 	stats->dma = dma_alloc_coherent(&pdev->dev, stats->size,
 					&stats->dma_map, GFP_KERNEL);
 	if (!stats->dma) {
diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.h b/drivers/infiniband/hw/bnxt_re/qplib_res.h
index 35d862b449080a80643bf935bc128d694365bbe2..32cebd0f1436a6a41a224194c0aa03f3747e0398 100644
--- a/drivers/infiniband/hw/bnxt_re/qplib_res.h
+++ b/drivers/infiniband/hw/bnxt_re/qplib_res.h
@@ -252,5 +252,5 @@ void bnxt_qplib_free_ctx(struct pci_dev *pdev,
 			 struct bnxt_qplib_ctx *ctx);
 int bnxt_qplib_alloc_ctx(struct pci_dev *pdev,
 			 struct bnxt_qplib_ctx *ctx,
-			 bool virt_fn);
+			 bool virt_fn, bool is_p5);
 #endif /* __BNXT_QPLIB_RES_H__ */