diff --git a/libhsakmt/tests/kfdtest/include/amdp2ptest.h b/libhsakmt/tests/kfdtest/include/amdp2ptest.h
index b465e24790..239e7e98d0 100644
--- a/libhsakmt/tests/kfdtest/include/amdp2ptest.h
+++ b/libhsakmt/tests/kfdtest/include/amdp2ptest.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2015-2018 Advanced Micro Devices, Inc. All Rights Reserved.
+ * Copyright 2015-2024 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -13,12 +13,11 @@
  *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
  * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
- *
  */
 #ifndef AMDP2PTEST_H_
 #define AMDP2PTEST_H_
@@ -44,6 +43,11 @@ struct AMDRDMA_IOCTL_GET_PAGES_PARAM {
 	/* Input parameters */
 	uint64_t addr;
 	uint64_t length;
+	uint64_t is_local;	/* 1 if this is the pointer to local
+				   allocation */
+
+	/* Output parameters */
+	uint64_t cpu_ptr;
 };
 
 
@@ -53,12 +57,6 @@ struct AMDRDMA_IOCTL_PUT_PAGES_PARAM {
 	uint64_t length;
 };
 
-struct AMDRDMA_IOCTL_IS_GPU_ADDRESS_PARAM {
-	/* Input parameters */
-	uint64_t addr;
-	uint64_t ret_value;
-};
-
 
 #define AMD2P2PTEST_IOCTL_GET_PAGE_SIZE	\
 _IOWR(AMDP2PTEST_IOCTL_MAGIC, 1, struct AMDRDMA_IOCTL_GET_PAGE_SIZE_PARAM *)
@@ -69,8 +67,5 @@ _IOWR(AMDP2PTEST_IOCTL_MAGIC, 2, struct AMDRDMA_IOCTL_GET_PAGES_PARAM *)
 #define AMD2P2PTEST_IOCTL_PUT_PAGES	\
 _IOW(AMDP2PTEST_IOCTL_MAGIC, 3, struct AMDRDMA_IOCTL_PUT_PAGES_PARAM *)
 
-#define AMD2P2PTEST_IOCTL_IS_GPU_ADDRESS	\
-_IOW(AMDP2PTEST_IOCTL_MAGIC, 4, struct AMDRDMA_IOCTL_IS_GPU_ADDRESS *)
-
 
 #endif  /* AMDP2PTEST_H */
diff --git a/libhsakmt/tests/kfdtest/src/RDMATest.cpp b/libhsakmt/tests/kfdtest/src/RDMATest.cpp
index d44d69895e..1931ea1741 100644
--- a/libhsakmt/tests/kfdtest/src/RDMATest.cpp
+++ b/libhsakmt/tests/kfdtest/src/RDMATest.cpp
@@ -112,3 +112,92 @@ TEST_F(RDMATest, GPUDirect) {
 
     TEST_END
 }
+
+TEST_F(RDMATest, ContiguousVRAMAllocation) {
+    TEST_REQUIRE_ENV_CAPABILITIES(ENVCAPS_64BITLINUX);
+    TEST_START(TESTPROFILE_RUNALL);
+
+    HSAuint64 AlternateVAGPU;
+
+    PM4Queue queue;
+    unsigned long BufferSize = 4UL << 30;
+
+    int defaultGPUNode = m_NodeInfo.HsaDefaultGPUNode();
+    ASSERT_GE(defaultGPUNode, 0) << "failed to get default GPU Node";
+
+    HsaMemoryBuffer isaBuffer(PAGE_SIZE, defaultGPUNode);
+    HsaMemoryBuffer srcSysBuffer(PAGE_SIZE, defaultGPUNode, false);
+    void *LocalBuffer;
+    HsaMemFlags memFlags = {0};
+    int ret;
+
+    memFlags.ui32.NonPaged = 1;
+    memFlags.ui32.Contiguous = 1;
+    ret = hsaKmtAllocMemory(defaultGPUNode, BufferSize, memFlags, &LocalBuffer);
+    if (ret == HSAKMT_STATUS_NOT_SUPPORTED) {
+        LOG() << "KFD does not support contiguous memory, skipping the test" << std::endl;
+        return;
+    }
+
+    ASSERT_SUCCESS(hsaKmtMapMemoryToGPU(srcSysBuffer.As<void*>(),
+                                        srcSysBuffer.Size(),
+                                        &AlternateVAGPU));
+    ASSERT_SUCCESS(hsaKmtMapMemoryToGPU(LocalBuffer, BufferSize, &AlternateVAGPU));
+
+    /* Fill up srcSysBuffer */
+    srcSysBuffer.Fill(0xfe);
+
+    /* Put 'copy dword' command to ISA buffer */
+    ASSERT_SUCCESS(m_pAsm->RunAssembleBuf(CopyDwordIsa, isaBuffer.As<char*>()));
+
+    ASSERT_SUCCESS(queue.Create(defaultGPUNode));
+    Dispatch dispatch(isaBuffer);
+
+    /* Submit the command to GPU so GPU will copy from system memory
+     * (srcSysBuffer) to local memory(LocalBuffer)
+     */
+    dispatch.SetArgs(srcSysBuffer.As<void*>(), LocalBuffer);
+    dispatch.Submit(queue);
+    dispatch.Sync(g_TestTimeOut);  // GPU executed the command
+
+    EXPECT_SUCCESS(queue.Destroy());
+
+    LocalMemoryAccess Rdma;
+    void *gpuAddr;
+
+    Rdma.Open();
+    if (Rdma.fd < 0) {
+        LOG() << "amdp2ptest.ko driver not loaded, skipping RDMA getpages" << std::endl;
+        goto exit;
+    }
+
+    /* GetPages asks the test driver to convert GPU virtual memory to DMA/
+     * Physical memory and save it in the list. rdma_mmap maps the memory to
+     * user space memory.
+     */
+    ret = Rdma.GetPages((uint64_t)LocalBuffer, BufferSize);
+    ASSERT_EQ(ret, 0) << "Failed to get pages";
+
+    gpuAddr = Rdma.MMap((uint64_t)LocalBuffer, BufferSize);
+    ASSERT_GE((int64_t)gpuAddr, 0) << "Failed to map RDMA address.";
+
+    printf("contiguous VRAM address %p size 0x%lx bytes\n", LocalBuffer, BufferSize);
+    printf("Pause to dump page table to check if allocation is contiguous\n");
+    printf("Press Enter key to continue\n");
+    getchar();
+
+    /* Read the memory to confirm that application can read the local memory
+     * correctly from the mapped address.
+     */
+    EXPECT_EQ(memcmp(gpuAddr, srcSysBuffer.As<void*>(), 4), 0);
+
+    Rdma.UnMap(gpuAddr, PAGE_SIZE);
+    Rdma.Close();
+
+exit:
+    EXPECT_SUCCESS(hsaKmtUnmapMemoryToGPU(srcSysBuffer.As<void*>()));
+    EXPECT_SUCCESS(hsaKmtUnmapMemoryToGPU(LocalBuffer));
+    EXPECT_SUCCESS(hsaKmtFreeMemory(LocalBuffer, BufferSize));
+
+    TEST_END
+}
diff --git a/libhsakmt/tests/kfdtest/src/RDMAUtil.cpp b/libhsakmt/tests/kfdtest/src/RDMAUtil.cpp
index 20246f9003..3a95f0e65c 100644
--- a/libhsakmt/tests/kfdtest/src/RDMAUtil.cpp
+++ b/libhsakmt/tests/kfdtest/src/RDMAUtil.cpp
@@ -60,7 +60,7 @@ void *LocalMemoryAccess::MMap(uint64_t offset, size_t size) {
     if (fd <= 0)
         return NULL;
 
-    gpuAddr = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, offset);
+    gpuAddr = mmap(NULL, size, PROT_READ, MAP_SHARED, fd, offset);
     return gpuAddr;
 }