Fix #1043: Fix memory leak in StridedMemoryView (#1048)

* Fix #1043: Fix memory leak in StridedMemoryView

* Move location of test

* Add relnote

Author: mdboom

cuda_core/cuda/core/experimental/_memoryview.pyx

@@ -105,6 +105,23 @@ cdef class StridedMemoryView:
         else:
             pass
 
+    def __dealloc__(self):
+        if self.dl_tensor == NULL:
+            return
+
+        if cpython.PyCapsule_IsValid(
+                self.metadata, DLPACK_VERSIONED_TENSOR_USED_NAME):
+            data = cpython.PyCapsule_GetPointer(
+                self.metadata, DLPACK_VERSIONED_TENSOR_USED_NAME)
+            dlm_tensor_ver = <DLManagedTensorVersioned*>data
+            dlm_tensor_ver.deleter(dlm_tensor_ver)
+        elif cpython.PyCapsule_IsValid(
+                self.metadata, DLPACK_TENSOR_USED_NAME):
+            data = cpython.PyCapsule_GetPointer(
+                self.metadata, DLPACK_TENSOR_USED_NAME)
+            dlm_tensor = <DLManagedTensor*>data
+            dlm_tensor.deleter(dlm_tensor)
+
     @property
     def shape(self) -> tuple[int]:
         if self._shape is None and self.exporting_obj is not None:

cuda_core/docs/source/release/0.X.Y-notes.rst

@@ -48,3 +48,4 @@ Fixes and enhancements
 - Make :class:`Buffer` creation more performant.
 - Enabled :class:`MemoryResource` subclasses to accept :class:`Device` objects, in addition to previously supported device ordinals.
 - Fixed a bug in :class:`Stream` and other classes where object cleanup would error during interpreter shutdown.
+- :class:`StridedMemoryView` of an underlying array using the DLPack protocol will no longer leak memory.

cuda_core/tests/test_memory.py

@@ -1,18 +1,24 @@
 # SPDX-FileCopyrightText: Copyright (c) 2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 # SPDX-License-Identifier: Apache-2.0
 
+import sys
+
 try:
     from cuda.bindings import driver
 except ImportError:
     from cuda import cuda as driver
-
+try:
+    import numpy as np
+except ImportError:
+    np = None
 import ctypes
 import platform
 
 import pytest
 from cuda.core.experimental import Buffer, Device, DeviceMemoryResource, MemoryResource
 from cuda.core.experimental._memory import DLDeviceType, IPCBufferDescriptor
 from cuda.core.experimental._utils.cuda_utils import handle_return
+from cuda.core.experimental.utils import StridedMemoryView
 
 POOL_SIZE = 2097152  # 2MB size
 
@@ -437,3 +443,14 @@ def test_mempool_attributes_ownership(mempool_device):
     with pytest.raises(RuntimeError, match="DeviceMemoryResource is expired"):
         _ = attributes.used_mem_high
     mr._mempool_handle = old_handle
+
+
+# Ensure that memory views dellocate their reference to dlpack tensors
+@pytest.mark.skipif(np is None, reason="numpy is not installed")
+def test_strided_memory_view_leak():
+    arr = np.zeros(1048576, dtype=np.uint8)
+    before = sys.getrefcount(arr)
+    for idx in range(10):
+        StridedMemoryView(arr, stream_ptr=-1)
+    after = sys.getrefcount(arr)
+    assert before == after