[RFC v3 0/3] Introduce KUnit tests for TTM subsystem

[RFC v3 0/3] Introduce KUnit tests for TTM subsystem

[Karolina Stolarek]

This series introduces KUnit[1] tests for TTM (Translation Table Manager)
subsystem, a memory manager used by graphics drivers to create and manage
memory buffers across different memory domains, such as system memory
or VRAM.

Unit tests implemented here cover two data structures:
  - ttm_device -- referred as a buffer object device, which stores
    resource managers and page pools
  - ttm_pool -- a struct of pools (ttm_pool_type) of different page
    orders and caching attributes, with pages that can be reused on
    the next buffer allocation

Use kunit_tool script to manually run the tests:

$ ./tools/testing/kunit/kunit.py run --kunitconfig=drivers/gpu/drm/ttm/tests

The kunit tool might not work with older python versions. To fix that,
apply [2] patch.

To build a kernel with TTM KUnit tests, enable CONFIG_DRM_TTM_KUNIT_TEST
symbol.

As for now, tests are architecture-agnostic (i.e. KUnit runner uses UML
kernel), which means that we have limited coverage in some places. For
example, we can't fully test the initialization of global page pools,
such as global_write_combined. It is to be decided if we want to stick
to UML or use CONFIG_X86 (at least to some extent).

These patches are just a beginning of the work to improve the test
coverage of TTM. Feel free to suggest changes, test cases or priorities.

Many thanks,
Karolina

v3:
  - Rename ttm_kunit_helper_alloc_device() to ttm_kunit_helper_init_device()
    (Christian)
  - Don't leak a full-blown drm_gem_object in ttm_kunit_helper_ttm_bo_init().
    (Christian). Create a small mock object just to get ttm_tt_init_fields()
    to init the right number of pages
  - As a follow up to the change above, delete ttm_kunit_helper_ttm_bo_fini()
    and just use ttm_bo_put()

v2:
  - Add missing symbol exports in ttm_kunit_helpers.c
  - Update helpers include to fix compilation issues (didn't catch it as
    KUnit tests weren't enabled in the kernel I tested, an oversight
    on my part)
  - Add checks for ttm_pool fields in ttm_pool_alloc_basic(), including the
    one for NUMA node id
  - Rebase the changes on the top of drm-tip

--------------------------------
[1] - https://www.kernel.org/doc/html/latest/dev-tools/kunit/index.html
[2] - https://lore.kernel.org/lkml/20230610175618.82271-1-sj@kernel.org/T/

Karolina Stolarek (3):
  drm/ttm: Introduce KUnit tests
  drm/ttm/tests: Add tests for ttm_device
  drm/ttm/tests: Add tests for ttm_pool

 drivers/gpu/drm/Kconfig                       |  15 +
 drivers/gpu/drm/ttm/Makefile                  |   1 +
 drivers/gpu/drm/ttm/tests/.kunitconfig        |   4 +
 drivers/gpu/drm/ttm/tests/Makefile            |   6 +
 drivers/gpu/drm/ttm/tests/ttm_device_test.c   | 213 +++++++++
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c |  74 ++++
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h |  33 ++
 drivers/gpu/drm/ttm/tests/ttm_pool_test.c     | 405 ++++++++++++++++++
 8 files changed, 751 insertions(+)
 create mode 100644 drivers/gpu/drm/ttm/tests/.kunitconfig
 create mode 100644 drivers/gpu/drm/ttm/tests/Makefile
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_device_test.c
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_pool_test.c

-- 
2.25.1

[RFC v3 1/3] drm/ttm: Introduce KUnit tests

[Karolina Stolarek]

Add the initial version of unit tests for ttm_device struct, together
with helper functions.

Signed-off-by: Karolina Stolarek <karolina.stolarek@intel.com>
---
 drivers/gpu/drm/Kconfig                       | 15 +++++
 drivers/gpu/drm/ttm/Makefile                  |  1 +
 drivers/gpu/drm/ttm/tests/.kunitconfig        |  4 ++
 drivers/gpu/drm/ttm/tests/Makefile            |  5 ++
 drivers/gpu/drm/ttm/tests/ttm_device_test.c   | 54 +++++++++++++++++
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c | 59 +++++++++++++++++++
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h | 29 +++++++++
 7 files changed, 167 insertions(+)
 create mode 100644 drivers/gpu/drm/ttm/tests/.kunitconfig
 create mode 100644 drivers/gpu/drm/ttm/tests/Makefile
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_device_test.c
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index afb3b2f5f425..53024e44a2d5 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -194,6 +194,21 @@ config DRM_TTM
 	  GPU memory types. Will be enabled automatically if a device driver
 	  uses it.
 
+config DRM_TTM_KUNIT_TEST
+        tristate "KUnit tests for TTM" if !KUNIT_ALL_TESTS
+        default n
+        depends on DRM && KUNIT
+        select DRM_TTM
+        select DRM_EXPORT_FOR_TESTS if m
+        select DRM_KUNIT_TEST_HELPERS
+        default KUNIT_ALL_TESTS
+        help
+          Enables unit tests for TTM, a GPU memory manager subsystem used
+          to manage memory buffers. This option is mostly useful for kernel
+          developers.
+
+          If in doubt, say "N".
+
 config DRM_BUDDY
 	tristate
 	depends on DRM
diff --git a/drivers/gpu/drm/ttm/Makefile b/drivers/gpu/drm/ttm/Makefile
index f906b22959cf..dad298127226 100644
--- a/drivers/gpu/drm/ttm/Makefile
+++ b/drivers/gpu/drm/ttm/Makefile
@@ -8,3 +8,4 @@ ttm-y := ttm_tt.o ttm_bo.o ttm_bo_util.o ttm_bo_vm.o ttm_module.o \
 ttm-$(CONFIG_AGP) += ttm_agp_backend.o
 
 obj-$(CONFIG_DRM_TTM) += ttm.o
+obj-$(CONFIG_DRM_TTM_KUNIT_TEST) += tests/
diff --git a/drivers/gpu/drm/ttm/tests/.kunitconfig b/drivers/gpu/drm/ttm/tests/.kunitconfig
new file mode 100644
index 000000000000..75fdce0cd98e
--- /dev/null
+++ b/drivers/gpu/drm/ttm/tests/.kunitconfig
@@ -0,0 +1,4 @@
+CONFIG_KUNIT=y
+CONFIG_DRM=y
+CONFIG_DRM_KUNIT_TEST_HELPERS=y
+CONFIG_DRM_TTM_KUNIT_TEST=y
diff --git a/drivers/gpu/drm/ttm/tests/Makefile b/drivers/gpu/drm/ttm/tests/Makefile
new file mode 100644
index 000000000000..7917805f37af
--- /dev/null
+++ b/drivers/gpu/drm/ttm/tests/Makefile
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0 AND MIT
+
+obj-$(CONFIG_DRM_TTM_KUNIT_TEST) += \
+        ttm_device_test.o \
+        ttm_kunit_helpers.o
diff --git a/drivers/gpu/drm/ttm/tests/ttm_device_test.c b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
new file mode 100644
index 000000000000..9cbfcf2911c7
--- /dev/null
+++ b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
@@ -0,0 +1,54 @@
+// SPDX-License-Identifier: GPL-2.0 AND MIT
+/*
+ * Copyright © 2023 Intel Corporation
+ */
+#include <drm/ttm/ttm_resource.h>
+#include <drm/ttm/ttm_device.h>
+#include <drm/ttm/ttm_placement.h>
+
+#include "ttm_kunit_helpers.h"
+
+static void ttm_device_init_basic(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	struct ttm_device *ttm_dev;
+	struct ttm_resource_manager *ttm_sys_man;
+	int err;
+
+	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
+
+	err = ttm_kunit_helper_init_device(test, ttm_dev, false, false);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	KUNIT_EXPECT_PTR_EQ(test, ttm_dev->funcs, &ttm_dev_funcs);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_dev->wq);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_dev->man_drv[TTM_PL_SYSTEM]);
+
+	ttm_sys_man = &ttm_dev->sysman;
+	KUNIT_ASSERT_NOT_NULL(test, ttm_sys_man);
+	KUNIT_EXPECT_TRUE(test, ttm_sys_man->use_tt);
+	KUNIT_EXPECT_TRUE(test, ttm_sys_man->use_type);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_sys_man->func);
+
+	KUNIT_EXPECT_PTR_EQ(test, ttm_dev->dev_mapping,
+			    priv->drm->anon_inode->i_mapping);
+
+	ttm_device_fini(ttm_dev);
+}
+
+static struct kunit_case ttm_device_test_cases[] = {
+	KUNIT_CASE(ttm_device_init_basic),
+	{}
+};
+
+static struct kunit_suite ttm_device_test_suite = {
+	.name = "ttm_device",
+	.init = ttm_test_devices_init,
+	.exit = ttm_test_devices_fini,
+	.test_cases = ttm_device_test_cases,
+};
+
+kunit_test_suites(&ttm_device_test_suite);
+
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
new file mode 100644
index 000000000000..428a8a4ac9dc
--- /dev/null
+++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
@@ -0,0 +1,59 @@
+// SPDX-License-Identifier: GPL-2.0 AND MIT
+/*
+ * Copyright © 2023 Intel Corporation
+ */
+#include "ttm_kunit_helpers.h"
+
+struct ttm_device_funcs ttm_dev_funcs = {
+};
+EXPORT_SYMBOL_GPL(ttm_dev_funcs);
+
+int ttm_kunit_helper_init_device(struct kunit *test,
+				  struct ttm_device *ttm,
+				  bool use_dma_alloc,
+				  bool use_dma32)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	struct drm_device *drm = priv->drm;
+	int err;
+
+	err = ttm_device_init(ttm, &ttm_dev_funcs, drm->dev,
+			      drm->anon_inode->i_mapping,
+			      drm->vma_offset_manager,
+			      use_dma_alloc, use_dma32);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(ttm_kunit_helper_init_device);
+
+int ttm_test_devices_init(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv;
+
+	priv = kunit_kzalloc(test, sizeof(*priv), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, priv);
+
+	test->priv = priv;
+
+	priv->dev = drm_kunit_helper_alloc_device(test);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, priv->dev);
+
+	priv->drm = __drm_kunit_helper_alloc_drm_device(test, priv->dev,
+							sizeof(*priv->drm), 0,
+							DRIVER_GEM);
+	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, priv->drm);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ttm_test_devices_init);
+
+void ttm_test_devices_fini(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+
+	drm_kunit_helper_free_device(test, priv->dev);
+	drm_dev_put(priv->drm);
+}
+EXPORT_SYMBOL_GPL(ttm_test_devices_fini);
+
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
new file mode 100644
index 000000000000..bf0b1000a8ee
--- /dev/null
+++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: GPL-2.0 AND MIT */
+/*
+ * Copyright © 2023 Intel Corporation
+ */
+#ifndef TTM_KUNIT_HELPERS_H
+#define TTM_KUNIT_HELPERS_H
+
+#include <drm/drm_drv.h>
+#include <drm/ttm/ttm_device.h>
+
+#include <drm/drm_kunit_helpers.h>
+#include <kunit/test.h>
+
+extern struct ttm_device_funcs ttm_dev_funcs;
+
+struct ttm_test_devices_priv {
+	struct drm_device *drm;
+	struct device *dev;
+};
+
+int ttm_kunit_helper_init_device(struct kunit *test,
+				  struct ttm_device *ttm,
+				  bool use_dma_alloc,
+				  bool use_dma32);
+
+int ttm_test_devices_init(struct kunit *test);
+void ttm_test_devices_fini(struct kunit *test);
+
+#endif // TTM_KUNIT_HELPERS_H
-- 
2.25.1

[RFC v3 2/3] drm/ttm/tests: Add tests for ttm_device

[Karolina Stolarek]

Test initialization and cleanup of the ttm_device struct, including
some error paths. Verify the creation of page pools if use_dma_alloc
param is true.

Signed-off-by: Karolina Stolarek <karolina.stolarek@intel.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/tests/ttm_device_test.c | 159 ++++++++++++++++++++
 1 file changed, 159 insertions(+)

diff --git a/drivers/gpu/drm/ttm/tests/ttm_device_test.c b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
index 9cbfcf2911c7..e3aa04441680 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_device_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
@@ -8,6 +8,13 @@
 
 #include "ttm_kunit_helpers.h"
 
+struct ttm_device_test_case {
+	const char *description;
+	bool use_dma_alloc;
+	bool use_dma32;
+	bool pools_init_expected;
+};
+
 static void ttm_device_init_basic(struct kunit *test)
 {
 	struct ttm_test_devices_priv *priv = test->priv;
@@ -37,8 +44,160 @@ static void ttm_device_init_basic(struct kunit *test)
 	ttm_device_fini(ttm_dev);
 }
 
+static void ttm_device_init_multiple(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	struct ttm_device *ttm_devs;
+	unsigned int i, num_dev = 3;
+	int err;
+
+	ttm_devs = kunit_kcalloc(test, num_dev, sizeof(*ttm_devs), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_devs);
+
+	for (i = 0; i < num_dev; i++) {
+		err = ttm_kunit_helper_init_device(test, &ttm_devs[i], false,
+						   false);
+		KUNIT_ASSERT_EQ(test, err, 0);
+
+		KUNIT_EXPECT_PTR_EQ(test, ttm_devs[i].dev_mapping,
+				    priv->drm->anon_inode->i_mapping);
+		KUNIT_ASSERT_NOT_NULL(test, ttm_devs[i].wq);
+		KUNIT_EXPECT_PTR_EQ(test, ttm_devs[i].funcs, &ttm_dev_funcs);
+		KUNIT_ASSERT_NOT_NULL(test, ttm_devs[i].man_drv[TTM_PL_SYSTEM]);
+	}
+
+	KUNIT_ASSERT_EQ(test, list_count_nodes(&ttm_devs[0].device_list), num_dev);
+
+	for (i = 0; i < num_dev; i++)
+		ttm_device_fini(&ttm_devs[i]);
+}
+
+static void ttm_device_fini_basic(struct kunit *test)
+{
+	struct ttm_device *ttm_dev;
+	struct ttm_resource_manager *man;
+	int err;
+
+	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
+
+	err = ttm_kunit_helper_init_device(test, ttm_dev, false, false);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	man = ttm_manager_type(ttm_dev, TTM_PL_SYSTEM);
+	KUNIT_ASSERT_NOT_NULL(test, man);
+
+	ttm_device_fini(ttm_dev);
+
+	KUNIT_ASSERT_FALSE(test, man->use_type);
+	KUNIT_ASSERT_TRUE(test, list_empty(&man->lru[0]));
+	KUNIT_ASSERT_NULL(test, ttm_dev->man_drv[TTM_PL_SYSTEM]);
+}
+
+static void ttm_device_init_no_vma_man(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	struct drm_device *drm = priv->drm;
+	struct ttm_device *ttm_dev;
+	struct drm_vma_offset_manager *vma_man;
+	int err;
+
+	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
+
+	/* Let's pretend there's no VMA manager allocated */
+	vma_man = drm->vma_offset_manager;
+	drm->vma_offset_manager = NULL;
+
+	err = ttm_kunit_helper_init_device(test, ttm_dev, false, false);
+	KUNIT_EXPECT_EQ(test, err, -EINVAL);
+
+	/* Bring the manager back for a graceful cleanup */
+	drm->vma_offset_manager = vma_man;
+}
+
+static const struct ttm_device_test_case ttm_device_cases[] = {
+	{
+		.description = "No DMA allocations, no DMA32 required",
+		.use_dma_alloc = false,
+		.use_dma32 = false,
+		.pools_init_expected = false,
+	},
+	{
+		.description = "DMA allocations, DMA32 required",
+		.use_dma_alloc = true,
+		.use_dma32 = true,
+		.pools_init_expected = true,
+	},
+	{
+		.description = "No DMA allocations, DMA32 required",
+		.use_dma_alloc = false,
+		.use_dma32 = true,
+		.pools_init_expected = false,
+	},
+	{
+		.description = "DMA allocations, no DMA32 required",
+		.use_dma_alloc = true,
+		.use_dma32 = false,
+		.pools_init_expected = true,
+	},
+};
+
+static void ttm_device_case_desc(const struct ttm_device_test_case *t, char *desc)
+{
+	strscpy(desc, t->description, KUNIT_PARAM_DESC_SIZE);
+}
+
+KUNIT_ARRAY_PARAM(ttm_device, ttm_device_cases, ttm_device_case_desc);
+
+static void ttm_device_init_pools(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	const struct ttm_device_test_case *params = test->param_value;
+	struct ttm_device *ttm_dev;
+	struct ttm_pool *pool;
+	struct ttm_pool_type pt;
+	int err;
+
+	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
+
+	err = ttm_kunit_helper_init_device(test, ttm_dev,
+					   params->use_dma_alloc,
+					   params->use_dma32);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	pool = &ttm_dev->pool;
+	KUNIT_ASSERT_NOT_NULL(test, pool);
+	KUNIT_EXPECT_PTR_EQ(test, pool->dev, priv->dev);
+	KUNIT_EXPECT_EQ(test, pool->use_dma_alloc, params->use_dma_alloc);
+	KUNIT_EXPECT_EQ(test, pool->use_dma32, params->use_dma32);
+
+	if (params->pools_init_expected) {
+		for (int i = 0; i < TTM_NUM_CACHING_TYPES; ++i) {
+			for (int j = 0; j <= MAX_ORDER; ++j) {
+				pt = pool->caching[i].orders[j];
+				KUNIT_EXPECT_PTR_EQ(test, pt.pool, pool);
+				KUNIT_EXPECT_EQ(test, pt.caching, i);
+				KUNIT_EXPECT_EQ(test, pt.order, j);
+
+				if (params->use_dma_alloc) {
+					KUNIT_ASSERT_FALSE(test,
+							   list_empty(&pt.pages));
+				}
+			}
+		}
+	}
+
+	ttm_device_fini(ttm_dev);
+}
+
 static struct kunit_case ttm_device_test_cases[] = {
 	KUNIT_CASE(ttm_device_init_basic),
+	KUNIT_CASE(ttm_device_init_multiple),
+	KUNIT_CASE(ttm_device_fini_basic),
+	KUNIT_CASE(ttm_device_init_no_vma_man),
+	KUNIT_CASE_PARAM(ttm_device_init_pools, ttm_device_gen_params),
 	{}
 };
 
-- 
2.25.1

[RFC v3 3/3] drm/ttm/tests: Add tests for ttm_pool

[Karolina Stolarek]

Add KUnit tests that exercise page allocation using page pools
and freeing pages, either by returning them to the pool or
freeing them. Add a basic test for ttm_pool cleanup. Introduce
helpers to create a dummy ttm_buffer_object.

Signed-off-by: Karolina Stolarek <karolina.stolarek@intel.com>
---
 drivers/gpu/drm/ttm/tests/Makefile            |   1 +
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c |  15 +
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h |   4 +
 drivers/gpu/drm/ttm/tests/ttm_pool_test.c     | 405 ++++++++++++++++++
 4 files changed, 425 insertions(+)
 create mode 100644 drivers/gpu/drm/ttm/tests/ttm_pool_test.c

diff --git a/drivers/gpu/drm/ttm/tests/Makefile b/drivers/gpu/drm/ttm/tests/Makefile
index 7917805f37af..ec87c4fc1ad5 100644
--- a/drivers/gpu/drm/ttm/tests/Makefile
+++ b/drivers/gpu/drm/ttm/tests/Makefile
@@ -2,4 +2,5 @@
 
 obj-$(CONFIG_DRM_TTM_KUNIT_TEST) += \
         ttm_device_test.o \
+        ttm_pool_test.o \
         ttm_kunit_helpers.o
diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
index 428a8a4ac9dc..9384679fccb0 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
@@ -26,6 +26,21 @@ int ttm_kunit_helper_init_device(struct kunit *test,
 }
 EXPORT_SYMBOL_GPL(ttm_kunit_helper_init_device);
 
+struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit *test,
+						       size_t size)
+{
+	struct drm_gem_object gem_obj = { .size = size };
+	struct ttm_buffer_object *bo;
+
+	bo = kunit_kzalloc(test, sizeof(*bo), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, bo);
+
+	bo->base = gem_obj;
+
+	return bo;
+}
+EXPORT_SYMBOL_GPL(ttm_kunit_helper_ttm_bo_init);
+
 int ttm_test_devices_init(struct kunit *test)
 {
 	struct ttm_test_devices_priv *priv;
diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
index bf0b1000a8ee..41ad38dd34e5 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
+++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
@@ -7,6 +7,7 @@
 
 #include <drm/drm_drv.h>
 #include <drm/ttm/ttm_device.h>
+#include <drm/ttm/ttm_bo.h>
 
 #include <drm/drm_kunit_helpers.h>
 #include <kunit/test.h>
@@ -23,6 +24,9 @@ int ttm_kunit_helper_init_device(struct kunit *test,
 				  bool use_dma_alloc,
 				  bool use_dma32);
 
+struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit *test,
+						       size_t size);
+
 int ttm_test_devices_init(struct kunit *test);
 void ttm_test_devices_fini(struct kunit *test);
 
diff --git a/drivers/gpu/drm/ttm/tests/ttm_pool_test.c b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
new file mode 100644
index 000000000000..85806f72df22
--- /dev/null
+++ b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
@@ -0,0 +1,405 @@
+// SPDX-License-Identifier: GPL-2.0 AND MIT
+/*
+ * Copyright © 2023 Intel Corporation
+ */
+#include <linux/mm.h>
+
+#include <drm/ttm/ttm_tt.h>
+#include <drm/ttm/ttm_pool.h>
+
+#include "ttm_kunit_helpers.h"
+
+struct ttm_pool_test_case {
+	const char *description;
+	unsigned int order;
+	bool use_dma_alloc;
+};
+
+static struct ttm_operation_ctx simple_ctx = {
+	.interruptible = true,
+	.no_wait_gpu = false,
+};
+
+static struct ttm_tt *mock_ttm_tt_init(struct kunit *test,
+				       uint32_t page_flags,
+				       enum ttm_caching caching,
+				       size_t size)
+{
+	struct ttm_tt *tt;
+	struct ttm_buffer_object *bo;
+	int err;
+
+	bo = ttm_kunit_helper_ttm_bo_init(test, size);
+	KUNIT_ASSERT_NOT_NULL(test, bo);
+
+	tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	err = ttm_tt_init(tt, bo, page_flags, caching, 0);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	/* We don't need this BO later, release it */
+	ttm_bo_put(bo);
+
+	return tt;
+}
+
+static struct ttm_pool *ttm_pool_pre_populated(struct kunit *test,
+					       size_t size,
+					       enum ttm_caching caching)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	struct ttm_pool *pool;
+	struct ttm_tt *tt;
+	int err;
+	unsigned long order = __fls(size / PAGE_SIZE);
+
+	tt = mock_ttm_tt_init(test, order, caching, size);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, pool);
+
+	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
+
+	err = ttm_pool_alloc(pool, tt, &simple_ctx);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+
+	return pool;
+}
+
+static const struct ttm_pool_test_case ttm_pool_basic_cases[] = {
+	{
+		.description = "One page",
+		.order = 0,
+	},
+	{
+		.description = "More than one page",
+		.order = 2,
+	},
+	{
+		.description = "Above the allocation limit",
+		.order = MAX_ORDER + 1,
+	},
+	{
+		.description = "One page, with coherent DMA mappings enabled",
+		.order = 0,
+		.use_dma_alloc = true,
+	},
+	{
+		.description = "Above the allocation limit, with coherent DMA mappings enabled",
+		.order = MAX_ORDER + 1,
+		.use_dma_alloc = true,
+	},
+};
+
+static void ttm_pool_alloc_case_desc(const struct ttm_pool_test_case *t,
+				     char *desc)
+{
+	strscpy(desc, t->description, KUNIT_PARAM_DESC_SIZE);
+}
+
+KUNIT_ARRAY_PARAM(ttm_pool_alloc_basic, ttm_pool_basic_cases,
+		  ttm_pool_alloc_case_desc);
+
+static void ttm_pool_alloc_basic(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	const struct ttm_pool_test_case *params = test->param_value;
+	struct ttm_tt *tt;
+	struct ttm_pool *pool;
+	struct page *fst_page, *last_page;
+	int err;
+	enum ttm_caching caching = ttm_uncached;
+	unsigned int expected_num_pages = 1 << params->order;
+	size_t size = expected_num_pages * PAGE_SIZE;
+
+	tt = mock_ttm_tt_init(test, 0, caching, size);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, pool);
+	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, params->use_dma_alloc,
+		      false);
+
+	KUNIT_ASSERT_PTR_EQ(test, pool->dev, priv->dev);
+	KUNIT_ASSERT_EQ(test, pool->nid, NUMA_NO_NODE);
+	KUNIT_ASSERT_EQ(test, pool->use_dma_alloc, params->use_dma_alloc);
+
+	err = ttm_pool_alloc(pool, tt, &simple_ctx);
+	KUNIT_ASSERT_EQ(test, err, 0);
+	KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
+
+	fst_page = tt->pages[0];
+	last_page = tt->pages[tt->num_pages - 1];
+
+	if (params->order <= MAX_ORDER) {
+		if (params->use_dma_alloc) {
+			KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
+			KUNIT_ASSERT_NOT_NULL(test, (void *)last_page->private);
+		} else {
+			KUNIT_ASSERT_EQ(test, fst_page->private, params->order);
+		}
+	} else {
+		if (params->use_dma_alloc) {
+			KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
+			KUNIT_ASSERT_NULL(test, (void *)last_page->private);
+		} else {
+			/*
+			 * We expect to alloc one big block, followed by
+			 * order 0 blocks
+			 */
+			KUNIT_ASSERT_EQ(test, fst_page->private,
+					min_t(unsigned int, MAX_ORDER,
+					      params->order));
+			KUNIT_ASSERT_EQ(test, last_page->private, 0);
+		}
+	}
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+	ttm_pool_fini(pool);
+}
+
+static void ttm_pool_alloc_basic_dma_addr(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	const struct ttm_pool_test_case *params = test->param_value;
+	struct ttm_tt *tt;
+	struct ttm_pool *pool;
+	struct ttm_buffer_object *bo;
+	dma_addr_t dma1, dma2;
+	int err;
+	enum ttm_caching caching = ttm_uncached;
+	unsigned int expected_num_pages = 1 << params->order;
+	size_t size = expected_num_pages * PAGE_SIZE;
+
+	tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	bo = ttm_kunit_helper_ttm_bo_init(test, size);
+	KUNIT_ASSERT_NOT_NULL(test, bo);
+
+	err = ttm_sg_tt_init(tt, bo, 0, caching);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, pool);
+
+	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
+
+	err = ttm_pool_alloc(pool, tt, &simple_ctx);
+	KUNIT_ASSERT_EQ(test, err, 0);
+	KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
+
+	dma1 = tt->dma_address[0];
+	dma2 = tt->dma_address[tt->num_pages - 1];
+
+	KUNIT_ASSERT_NOT_NULL(test, (void *)dma1);
+	KUNIT_ASSERT_NOT_NULL(test, (void *)dma2);
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+	ttm_pool_fini(pool);
+}
+
+static void ttm_pool_alloc_order_caching_match(struct kunit *test)
+{
+	struct ttm_tt *tt;
+	struct ttm_pool *pool;
+	struct ttm_pool_type *pt;
+	enum ttm_caching caching = ttm_uncached;
+	unsigned int order = 0;
+	size_t size = PAGE_SIZE;
+	int err;
+
+	pool = ttm_pool_pre_populated(test, size, caching);
+
+	pt = &pool->caching[caching].orders[order];
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
+
+	tt = mock_ttm_tt_init(test, 0, caching, size);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	err = ttm_pool_alloc(pool, tt, &simple_ctx);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+	ttm_pool_fini(pool);
+}
+
+static void ttm_pool_alloc_caching_mismatch(struct kunit *test)
+{
+	struct ttm_tt *tt;
+	struct ttm_pool *pool;
+	struct ttm_pool_type *pt_pool, *pt_tt;
+	int err;
+	enum ttm_caching tt_caching = ttm_uncached;
+	enum ttm_caching pool_caching = ttm_cached;
+	size_t size = PAGE_SIZE;
+	unsigned int order = 0;
+
+	pool = ttm_pool_pre_populated(test, size, pool_caching);
+
+	pt_pool = &pool->caching[pool_caching].orders[order];
+	pt_tt = &pool->caching[tt_caching].orders[order];
+
+	tt = mock_ttm_tt_init(test, 0, tt_caching, size);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
+	KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
+
+	err = ttm_pool_alloc(pool, tt, &simple_ctx);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
+
+	ttm_pool_fini(pool);
+}
+
+static void ttm_pool_alloc_order_mismatch(struct kunit *test)
+{
+	struct ttm_tt *tt;
+	struct ttm_pool *pool;
+	struct ttm_pool_type *pt_pool, *pt_tt;
+	int err;
+	enum ttm_caching caching = ttm_uncached;
+	unsigned int order = 2;
+	size_t fst_size = (1 << order) * PAGE_SIZE;
+	size_t snd_size = PAGE_SIZE;
+
+	pool = ttm_pool_pre_populated(test, fst_size, caching);
+
+	pt_pool = &pool->caching[caching].orders[order];
+	pt_tt = &pool->caching[caching].orders[0];
+
+	tt = mock_ttm_tt_init(test, 0, caching, snd_size);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
+	KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
+
+	err = ttm_pool_alloc(pool, tt, &simple_ctx);
+	KUNIT_ASSERT_EQ(test, err, 0);
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
+
+	ttm_pool_fini(pool);
+}
+
+static void ttm_pool_free_dma_alloc(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	struct ttm_tt *tt;
+	struct ttm_pool *pool;
+	struct ttm_pool_type *pt;
+	enum ttm_caching caching = ttm_uncached;
+	unsigned int order = 2;
+	size_t size = (1 << order) * PAGE_SIZE;
+
+	tt = mock_ttm_tt_init(test, 0, caching, size);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, pool);
+
+	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
+	ttm_pool_alloc(pool, tt, &simple_ctx);
+
+	pt = &pool->caching[caching].orders[order];
+	KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
+
+	ttm_pool_fini(pool);
+}
+
+static void ttm_pool_free_no_dma_alloc(struct kunit *test)
+{
+	struct ttm_test_devices_priv *priv = test->priv;
+	struct ttm_tt *tt;
+	struct ttm_pool *pool;
+	struct ttm_pool_type *pt;
+	enum ttm_caching caching = ttm_uncached;
+	unsigned int order = 2;
+	size_t size = (1 << order) * PAGE_SIZE;
+
+	tt = mock_ttm_tt_init(test, 0, caching, size);
+	KUNIT_ASSERT_NOT_NULL(test, tt);
+
+	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, pool);
+
+	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, false, false);
+	ttm_pool_alloc(pool, tt, &simple_ctx);
+
+	pt = &pool->caching[caching].orders[order];
+	KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
+
+	ttm_pool_free(pool, tt);
+	ttm_tt_fini(tt);
+
+	KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
+
+	ttm_pool_fini(pool);
+}
+
+static void ttm_pool_fini_basic(struct kunit *test)
+{
+	struct ttm_pool *pool;
+	struct ttm_pool_type *pt;
+	enum ttm_caching caching = ttm_uncached;
+	unsigned int order = 0;
+	size_t size = PAGE_SIZE;
+
+	pool = ttm_pool_pre_populated(test, size, caching);
+	pt = &pool->caching[caching].orders[order];
+
+	KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
+
+	ttm_pool_fini(pool);
+
+	KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
+}
+
+static struct kunit_case ttm_pool_test_cases[] = {
+	KUNIT_CASE_PARAM(ttm_pool_alloc_basic, ttm_pool_alloc_basic_gen_params),
+	KUNIT_CASE_PARAM(ttm_pool_alloc_basic_dma_addr,
+			 ttm_pool_alloc_basic_gen_params),
+	KUNIT_CASE(ttm_pool_alloc_order_caching_match),
+	KUNIT_CASE(ttm_pool_alloc_caching_mismatch),
+	KUNIT_CASE(ttm_pool_alloc_order_mismatch),
+	KUNIT_CASE(ttm_pool_free_dma_alloc),
+	KUNIT_CASE(ttm_pool_free_no_dma_alloc),
+	KUNIT_CASE(ttm_pool_fini_basic),
+	{}
+};
+
+static struct kunit_suite ttm_pool_test_suite = {
+	.name = "ttm_pool",
+	.init = ttm_test_devices_init,
+	.exit = ttm_test_devices_fini,
+	.test_cases = ttm_pool_test_cases,
+};
+
+kunit_test_suites(&ttm_pool_test_suite);
+
+MODULE_LICENSE("GPL");
-- 
2.25.1

Re: [RFC v3 3/3] drm/ttm/tests: Add tests for ttm_pool

[Christian König]

Am 03.07.23 um 09:58 schrieb Karolina Stolarek:
> Add KUnit tests that exercise page allocation using page pools
> and freeing pages, either by returning them to the pool or
> freeing them. Add a basic test for ttm_pool cleanup. Introduce
> helpers to create a dummy ttm_buffer_object.
>
> Signed-off-by: Karolina Stolarek <karolina.stolarek@intel.com>
> ---
>   drivers/gpu/drm/ttm/tests/Makefile            |   1 +
>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c |  15 +
>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h |   4 +
>   drivers/gpu/drm/ttm/tests/ttm_pool_test.c     | 405 ++++++++++++++++++
>   4 files changed, 425 insertions(+)
>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>
> diff --git a/drivers/gpu/drm/ttm/tests/Makefile b/drivers/gpu/drm/ttm/tests/Makefile
> index 7917805f37af..ec87c4fc1ad5 100644
> --- a/drivers/gpu/drm/ttm/tests/Makefile
> +++ b/drivers/gpu/drm/ttm/tests/Makefile
> @@ -2,4 +2,5 @@
>   
>   obj-$(CONFIG_DRM_TTM_KUNIT_TEST) += \
>           ttm_device_test.o \
> +        ttm_pool_test.o \
>           ttm_kunit_helpers.o
> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
> index 428a8a4ac9dc..9384679fccb0 100644
> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
> @@ -26,6 +26,21 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>   }
>   EXPORT_SYMBOL_GPL(ttm_kunit_helper_init_device);
>   
> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit *test,
> +						       size_t size)
> +{
> +	struct drm_gem_object gem_obj = { .size = size };
> +	struct ttm_buffer_object *bo;
> +
> +	bo = kunit_kzalloc(test, sizeof(*bo), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, bo);
> +
> +	bo->base = gem_obj;
> +
> +	return bo;
> +}
> +EXPORT_SYMBOL_GPL(ttm_kunit_helper_ttm_bo_init);
> +
>   int ttm_test_devices_init(struct kunit *test)
>   {
>   	struct ttm_test_devices_priv *priv;
> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
> index bf0b1000a8ee..41ad38dd34e5 100644
> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
> @@ -7,6 +7,7 @@
>   
>   #include <drm/drm_drv.h>
>   #include <drm/ttm/ttm_device.h>
> +#include <drm/ttm/ttm_bo.h>
>   
>   #include <drm/drm_kunit_helpers.h>
>   #include <kunit/test.h>
> @@ -23,6 +24,9 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>   				  bool use_dma_alloc,
>   				  bool use_dma32);
>   
> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit *test,
> +						       size_t size);
> +
>   int ttm_test_devices_init(struct kunit *test);
>   void ttm_test_devices_fini(struct kunit *test);
>   
> diff --git a/drivers/gpu/drm/ttm/tests/ttm_pool_test.c b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
> new file mode 100644
> index 000000000000..85806f72df22
> --- /dev/null
> +++ b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
> @@ -0,0 +1,405 @@
> +// SPDX-License-Identifier: GPL-2.0 AND MIT
> +/*
> + * Copyright © 2023 Intel Corporation
> + */
> +#include <linux/mm.h>
> +
> +#include <drm/ttm/ttm_tt.h>
> +#include <drm/ttm/ttm_pool.h>
> +
> +#include "ttm_kunit_helpers.h"
> +
> +struct ttm_pool_test_case {
> +	const char *description;
> +	unsigned int order;
> +	bool use_dma_alloc;
> +};
> +
> +static struct ttm_operation_ctx simple_ctx = {
> +	.interruptible = true,
> +	.no_wait_gpu = false,
> +};
> +
> +static struct ttm_tt *mock_ttm_tt_init(struct kunit *test,
> +				       uint32_t page_flags,
> +				       enum ttm_caching caching,
> +				       size_t size)
> +{
> +	struct ttm_tt *tt;
> +	struct ttm_buffer_object *bo;
> +	int err;
> +
> +	bo = ttm_kunit_helper_ttm_bo_init(test, size);
> +	KUNIT_ASSERT_NOT_NULL(test, bo);
> +
> +	tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	err = ttm_tt_init(tt, bo, page_flags, caching, 0);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +
> +	/* We don't need this BO later, release it */
> +	ttm_bo_put(bo);

That won't work. A BO must always outlive the TT object allocated for it.

Otherwise you can get crashes during TT release.

Regards,
Christian.

> +
> +	return tt;
> +}
> +
> +static struct ttm_pool *ttm_pool_pre_populated(struct kunit *test,
> +					       size_t size,
> +					       enum ttm_caching caching)
> +{
> +	struct ttm_test_devices_priv *priv = test->priv;
> +	struct ttm_pool *pool;
> +	struct ttm_tt *tt;
> +	int err;
> +	unsigned long order = __fls(size / PAGE_SIZE);
> +
> +	tt = mock_ttm_tt_init(test, order, caching, size);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, pool);
> +
> +	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
> +
> +	err = ttm_pool_alloc(pool, tt, &simple_ctx);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +
> +	return pool;
> +}
> +
> +static const struct ttm_pool_test_case ttm_pool_basic_cases[] = {
> +	{
> +		.description = "One page",
> +		.order = 0,
> +	},
> +	{
> +		.description = "More than one page",
> +		.order = 2,
> +	},
> +	{
> +		.description = "Above the allocation limit",
> +		.order = MAX_ORDER + 1,
> +	},
> +	{
> +		.description = "One page, with coherent DMA mappings enabled",
> +		.order = 0,
> +		.use_dma_alloc = true,
> +	},
> +	{
> +		.description = "Above the allocation limit, with coherent DMA mappings enabled",
> +		.order = MAX_ORDER + 1,
> +		.use_dma_alloc = true,
> +	},
> +};
> +
> +static void ttm_pool_alloc_case_desc(const struct ttm_pool_test_case *t,
> +				     char *desc)
> +{
> +	strscpy(desc, t->description, KUNIT_PARAM_DESC_SIZE);
> +}
> +
> +KUNIT_ARRAY_PARAM(ttm_pool_alloc_basic, ttm_pool_basic_cases,
> +		  ttm_pool_alloc_case_desc);
> +
> +static void ttm_pool_alloc_basic(struct kunit *test)
> +{
> +	struct ttm_test_devices_priv *priv = test->priv;
> +	const struct ttm_pool_test_case *params = test->param_value;
> +	struct ttm_tt *tt;
> +	struct ttm_pool *pool;
> +	struct page *fst_page, *last_page;
> +	int err;
> +	enum ttm_caching caching = ttm_uncached;
> +	unsigned int expected_num_pages = 1 << params->order;
> +	size_t size = expected_num_pages * PAGE_SIZE;
> +
> +	tt = mock_ttm_tt_init(test, 0, caching, size);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, pool);
> +	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, params->use_dma_alloc,
> +		      false);
> +
> +	KUNIT_ASSERT_PTR_EQ(test, pool->dev, priv->dev);
> +	KUNIT_ASSERT_EQ(test, pool->nid, NUMA_NO_NODE);
> +	KUNIT_ASSERT_EQ(test, pool->use_dma_alloc, params->use_dma_alloc);
> +
> +	err = ttm_pool_alloc(pool, tt, &simple_ctx);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +	KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
> +
> +	fst_page = tt->pages[0];
> +	last_page = tt->pages[tt->num_pages - 1];
> +
> +	if (params->order <= MAX_ORDER) {
> +		if (params->use_dma_alloc) {
> +			KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
> +			KUNIT_ASSERT_NOT_NULL(test, (void *)last_page->private);
> +		} else {
> +			KUNIT_ASSERT_EQ(test, fst_page->private, params->order);
> +		}
> +	} else {
> +		if (params->use_dma_alloc) {
> +			KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
> +			KUNIT_ASSERT_NULL(test, (void *)last_page->private);
> +		} else {
> +			/*
> +			 * We expect to alloc one big block, followed by
> +			 * order 0 blocks
> +			 */
> +			KUNIT_ASSERT_EQ(test, fst_page->private,
> +					min_t(unsigned int, MAX_ORDER,
> +					      params->order));
> +			KUNIT_ASSERT_EQ(test, last_page->private, 0);
> +		}
> +	}
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +	ttm_pool_fini(pool);
> +}
> +
> +static void ttm_pool_alloc_basic_dma_addr(struct kunit *test)
> +{
> +	struct ttm_test_devices_priv *priv = test->priv;
> +	const struct ttm_pool_test_case *params = test->param_value;
> +	struct ttm_tt *tt;
> +	struct ttm_pool *pool;
> +	struct ttm_buffer_object *bo;
> +	dma_addr_t dma1, dma2;
> +	int err;
> +	enum ttm_caching caching = ttm_uncached;
> +	unsigned int expected_num_pages = 1 << params->order;
> +	size_t size = expected_num_pages * PAGE_SIZE;
> +
> +	tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	bo = ttm_kunit_helper_ttm_bo_init(test, size);
> +	KUNIT_ASSERT_NOT_NULL(test, bo);
> +
> +	err = ttm_sg_tt_init(tt, bo, 0, caching);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +
> +	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, pool);
> +
> +	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
> +
> +	err = ttm_pool_alloc(pool, tt, &simple_ctx);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +	KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
> +
> +	dma1 = tt->dma_address[0];
> +	dma2 = tt->dma_address[tt->num_pages - 1];
> +
> +	KUNIT_ASSERT_NOT_NULL(test, (void *)dma1);
> +	KUNIT_ASSERT_NOT_NULL(test, (void *)dma2);
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +	ttm_pool_fini(pool);
> +}
> +
> +static void ttm_pool_alloc_order_caching_match(struct kunit *test)
> +{
> +	struct ttm_tt *tt;
> +	struct ttm_pool *pool;
> +	struct ttm_pool_type *pt;
> +	enum ttm_caching caching = ttm_uncached;
> +	unsigned int order = 0;
> +	size_t size = PAGE_SIZE;
> +	int err;
> +
> +	pool = ttm_pool_pre_populated(test, size, caching);
> +
> +	pt = &pool->caching[caching].orders[order];
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
> +
> +	tt = mock_ttm_tt_init(test, 0, caching, size);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	err = ttm_pool_alloc(pool, tt, &simple_ctx);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +
> +	KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +	ttm_pool_fini(pool);
> +}
> +
> +static void ttm_pool_alloc_caching_mismatch(struct kunit *test)
> +{
> +	struct ttm_tt *tt;
> +	struct ttm_pool *pool;
> +	struct ttm_pool_type *pt_pool, *pt_tt;
> +	int err;
> +	enum ttm_caching tt_caching = ttm_uncached;
> +	enum ttm_caching pool_caching = ttm_cached;
> +	size_t size = PAGE_SIZE;
> +	unsigned int order = 0;
> +
> +	pool = ttm_pool_pre_populated(test, size, pool_caching);
> +
> +	pt_pool = &pool->caching[pool_caching].orders[order];
> +	pt_tt = &pool->caching[tt_caching].orders[order];
> +
> +	tt = mock_ttm_tt_init(test, 0, tt_caching, size);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
> +	KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
> +
> +	err = ttm_pool_alloc(pool, tt, &simple_ctx);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
> +
> +	ttm_pool_fini(pool);
> +}
> +
> +static void ttm_pool_alloc_order_mismatch(struct kunit *test)
> +{
> +	struct ttm_tt *tt;
> +	struct ttm_pool *pool;
> +	struct ttm_pool_type *pt_pool, *pt_tt;
> +	int err;
> +	enum ttm_caching caching = ttm_uncached;
> +	unsigned int order = 2;
> +	size_t fst_size = (1 << order) * PAGE_SIZE;
> +	size_t snd_size = PAGE_SIZE;
> +
> +	pool = ttm_pool_pre_populated(test, fst_size, caching);
> +
> +	pt_pool = &pool->caching[caching].orders[order];
> +	pt_tt = &pool->caching[caching].orders[0];
> +
> +	tt = mock_ttm_tt_init(test, 0, caching, snd_size);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
> +	KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
> +
> +	err = ttm_pool_alloc(pool, tt, &simple_ctx);
> +	KUNIT_ASSERT_EQ(test, err, 0);
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
> +
> +	ttm_pool_fini(pool);
> +}
> +
> +static void ttm_pool_free_dma_alloc(struct kunit *test)
> +{
> +	struct ttm_test_devices_priv *priv = test->priv;
> +	struct ttm_tt *tt;
> +	struct ttm_pool *pool;
> +	struct ttm_pool_type *pt;
> +	enum ttm_caching caching = ttm_uncached;
> +	unsigned int order = 2;
> +	size_t size = (1 << order) * PAGE_SIZE;
> +
> +	tt = mock_ttm_tt_init(test, 0, caching, size);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, pool);
> +
> +	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
> +	ttm_pool_alloc(pool, tt, &simple_ctx);
> +
> +	pt = &pool->caching[caching].orders[order];
> +	KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
> +
> +	ttm_pool_fini(pool);
> +}
> +
> +static void ttm_pool_free_no_dma_alloc(struct kunit *test)
> +{
> +	struct ttm_test_devices_priv *priv = test->priv;
> +	struct ttm_tt *tt;
> +	struct ttm_pool *pool;
> +	struct ttm_pool_type *pt;
> +	enum ttm_caching caching = ttm_uncached;
> +	unsigned int order = 2;
> +	size_t size = (1 << order) * PAGE_SIZE;
> +
> +	tt = mock_ttm_tt_init(test, 0, caching, size);
> +	KUNIT_ASSERT_NOT_NULL(test, tt);
> +
> +	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
> +	KUNIT_ASSERT_NOT_NULL(test, pool);
> +
> +	ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, false, false);
> +	ttm_pool_alloc(pool, tt, &simple_ctx);
> +
> +	pt = &pool->caching[caching].orders[order];
> +	KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
> +
> +	ttm_pool_free(pool, tt);
> +	ttm_tt_fini(tt);
> +
> +	KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
> +
> +	ttm_pool_fini(pool);
> +}
> +
> +static void ttm_pool_fini_basic(struct kunit *test)
> +{
> +	struct ttm_pool *pool;
> +	struct ttm_pool_type *pt;
> +	enum ttm_caching caching = ttm_uncached;
> +	unsigned int order = 0;
> +	size_t size = PAGE_SIZE;
> +
> +	pool = ttm_pool_pre_populated(test, size, caching);
> +	pt = &pool->caching[caching].orders[order];
> +
> +	KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
> +
> +	ttm_pool_fini(pool);
> +
> +	KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
> +}
> +
> +static struct kunit_case ttm_pool_test_cases[] = {
> +	KUNIT_CASE_PARAM(ttm_pool_alloc_basic, ttm_pool_alloc_basic_gen_params),
> +	KUNIT_CASE_PARAM(ttm_pool_alloc_basic_dma_addr,
> +			 ttm_pool_alloc_basic_gen_params),
> +	KUNIT_CASE(ttm_pool_alloc_order_caching_match),
> +	KUNIT_CASE(ttm_pool_alloc_caching_mismatch),
> +	KUNIT_CASE(ttm_pool_alloc_order_mismatch),
> +	KUNIT_CASE(ttm_pool_free_dma_alloc),
> +	KUNIT_CASE(ttm_pool_free_no_dma_alloc),
> +	KUNIT_CASE(ttm_pool_fini_basic),
> +	{}
> +};
> +
> +static struct kunit_suite ttm_pool_test_suite = {
> +	.name = "ttm_pool",
> +	.init = ttm_test_devices_init,
> +	.exit = ttm_test_devices_fini,
> +	.test_cases = ttm_pool_test_cases,
> +};
> +
> +kunit_test_suites(&ttm_pool_test_suite);
> +
> +MODULE_LICENSE("GPL");

Re: [RFC v3 3/3] drm/ttm/tests: Add tests for ttm_pool

[Karolina Stolarek]

On 3.07.2023 13:21, Christian König wrote:
> 
> 
> Am 03.07.23 um 09:58 schrieb Karolina Stolarek:
>> Add KUnit tests that exercise page allocation using page pools
>> and freeing pages, either by returning them to the pool or
>> freeing them. Add a basic test for ttm_pool cleanup. Introduce
>> helpers to create a dummy ttm_buffer_object.
>>
>> Signed-off-by: Karolina Stolarek <karolina.stolarek@intel.com>
>> ---
>>   drivers/gpu/drm/ttm/tests/Makefile            |   1 +
>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c |  15 +
>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h |   4 +
>>   drivers/gpu/drm/ttm/tests/ttm_pool_test.c     | 405 ++++++++++++++++++
>>   4 files changed, 425 insertions(+)
>>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>
>> diff --git a/drivers/gpu/drm/ttm/tests/Makefile 
>> b/drivers/gpu/drm/ttm/tests/Makefile
>> index 7917805f37af..ec87c4fc1ad5 100644
>> --- a/drivers/gpu/drm/ttm/tests/Makefile
>> +++ b/drivers/gpu/drm/ttm/tests/Makefile
>> @@ -2,4 +2,5 @@
>>   obj-$(CONFIG_DRM_TTM_KUNIT_TEST) += \
>>           ttm_device_test.o \
>> +        ttm_pool_test.o \
>>           ttm_kunit_helpers.o
>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c 
>> b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>> index 428a8a4ac9dc..9384679fccb0 100644
>> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>> @@ -26,6 +26,21 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>>   }
>>   EXPORT_SYMBOL_GPL(ttm_kunit_helper_init_device);
>> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit 
>> *test,
>> +                               size_t size)
>> +{
>> +    struct drm_gem_object gem_obj = { .size = size };
>> +    struct ttm_buffer_object *bo;
>> +
>> +    bo = kunit_kzalloc(test, sizeof(*bo), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>> +
>> +    bo->base = gem_obj;
>> +
>> +    return bo;
>> +}
>> +EXPORT_SYMBOL_GPL(ttm_kunit_helper_ttm_bo_init);
>> +
>>   int ttm_test_devices_init(struct kunit *test)
>>   {
>>       struct ttm_test_devices_priv *priv;
>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h 
>> b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>> index bf0b1000a8ee..41ad38dd34e5 100644
>> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>> @@ -7,6 +7,7 @@
>>   #include <drm/drm_drv.h>
>>   #include <drm/ttm/ttm_device.h>
>> +#include <drm/ttm/ttm_bo.h>
>>   #include <drm/drm_kunit_helpers.h>
>>   #include <kunit/test.h>
>> @@ -23,6 +24,9 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>>                     bool use_dma_alloc,
>>                     bool use_dma32);
>> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit 
>> *test,
>> +                               size_t size);
>> +
>>   int ttm_test_devices_init(struct kunit *test);
>>   void ttm_test_devices_fini(struct kunit *test);
>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_pool_test.c 
>> b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>> new file mode 100644
>> index 000000000000..85806f72df22
>> --- /dev/null
>> +++ b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>> @@ -0,0 +1,405 @@
>> +// SPDX-License-Identifier: GPL-2.0 AND MIT
>> +/*
>> + * Copyright © 2023 Intel Corporation
>> + */
>> +#include <linux/mm.h>
>> +
>> +#include <drm/ttm/ttm_tt.h>
>> +#include <drm/ttm/ttm_pool.h>
>> +
>> +#include "ttm_kunit_helpers.h"
>> +
>> +struct ttm_pool_test_case {
>> +    const char *description;
>> +    unsigned int order;
>> +    bool use_dma_alloc;
>> +};
>> +
>> +static struct ttm_operation_ctx simple_ctx = {
>> +    .interruptible = true,
>> +    .no_wait_gpu = false,
>> +};
>> +
>> +static struct ttm_tt *mock_ttm_tt_init(struct kunit *test,
>> +                       uint32_t page_flags,
>> +                       enum ttm_caching caching,
>> +                       size_t size)
>> +{
>> +    struct ttm_tt *tt;
>> +    struct ttm_buffer_object *bo;
>> +    int err;
>> +
>> +    bo = ttm_kunit_helper_ttm_bo_init(test, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>> +
>> +    tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    err = ttm_tt_init(tt, bo, page_flags, caching, 0);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +
>> +    /* We don't need this BO later, release it */
>> +    ttm_bo_put(bo);
> 
> That won't work. A BO must always outlive the TT object allocated for it. >
> Otherwise you can get crashes during TT release.

OK, I can keep it around, probably as an another test->priv member.

I thought that for these mocks it should be fine -- ttm_tt_init_fields() 
only takes the scatter list from BO which is not in use (as for now, at 
least) and in ttm_tt_fini() we don't even touch that field.

All the best,
Karolina

> 
> Regards,
> Christian.
> 
>> +
>> +    return tt;
>> +}
>> +
>> +static struct ttm_pool *ttm_pool_pre_populated(struct kunit *test,
>> +                           size_t size,
>> +                           enum ttm_caching caching)
>> +{
>> +    struct ttm_test_devices_priv *priv = test->priv;
>> +    struct ttm_pool *pool;
>> +    struct ttm_tt *tt;
>> +    int err;
>> +    unsigned long order = __fls(size / PAGE_SIZE);
>> +
>> +    tt = mock_ttm_tt_init(test, order, caching, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>> +
>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>> +
>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +
>> +    return pool;
>> +}
>> +
>> +static const struct ttm_pool_test_case ttm_pool_basic_cases[] = {
>> +    {
>> +        .description = "One page",
>> +        .order = 0,
>> +    },
>> +    {
>> +        .description = "More than one page",
>> +        .order = 2,
>> +    },
>> +    {
>> +        .description = "Above the allocation limit",
>> +        .order = MAX_ORDER + 1,
>> +    },
>> +    {
>> +        .description = "One page, with coherent DMA mappings enabled",
>> +        .order = 0,
>> +        .use_dma_alloc = true,
>> +    },
>> +    {
>> +        .description = "Above the allocation limit, with coherent DMA 
>> mappings enabled",
>> +        .order = MAX_ORDER + 1,
>> +        .use_dma_alloc = true,
>> +    },
>> +};
>> +
>> +static void ttm_pool_alloc_case_desc(const struct ttm_pool_test_case *t,
>> +                     char *desc)
>> +{
>> +    strscpy(desc, t->description, KUNIT_PARAM_DESC_SIZE);
>> +}
>> +
>> +KUNIT_ARRAY_PARAM(ttm_pool_alloc_basic, ttm_pool_basic_cases,
>> +          ttm_pool_alloc_case_desc);
>> +
>> +static void ttm_pool_alloc_basic(struct kunit *test)
>> +{
>> +    struct ttm_test_devices_priv *priv = test->priv;
>> +    const struct ttm_pool_test_case *params = test->param_value;
>> +    struct ttm_tt *tt;
>> +    struct ttm_pool *pool;
>> +    struct page *fst_page, *last_page;
>> +    int err;
>> +    enum ttm_caching caching = ttm_uncached;
>> +    unsigned int expected_num_pages = 1 << params->order;
>> +    size_t size = expected_num_pages * PAGE_SIZE;
>> +
>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, params->use_dma_alloc,
>> +              false);
>> +
>> +    KUNIT_ASSERT_PTR_EQ(test, pool->dev, priv->dev);
>> +    KUNIT_ASSERT_EQ(test, pool->nid, NUMA_NO_NODE);
>> +    KUNIT_ASSERT_EQ(test, pool->use_dma_alloc, params->use_dma_alloc);
>> +
>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +    KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
>> +
>> +    fst_page = tt->pages[0];
>> +    last_page = tt->pages[tt->num_pages - 1];
>> +
>> +    if (params->order <= MAX_ORDER) {
>> +        if (params->use_dma_alloc) {
>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)last_page->private);
>> +        } else {
>> +            KUNIT_ASSERT_EQ(test, fst_page->private, params->order);
>> +        }
>> +    } else {
>> +        if (params->use_dma_alloc) {
>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
>> +            KUNIT_ASSERT_NULL(test, (void *)last_page->private);
>> +        } else {
>> +            /*
>> +             * We expect to alloc one big block, followed by
>> +             * order 0 blocks
>> +             */
>> +            KUNIT_ASSERT_EQ(test, fst_page->private,
>> +                    min_t(unsigned int, MAX_ORDER,
>> +                          params->order));
>> +            KUNIT_ASSERT_EQ(test, last_page->private, 0);
>> +        }
>> +    }
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +    ttm_pool_fini(pool);
>> +}
>> +
>> +static void ttm_pool_alloc_basic_dma_addr(struct kunit *test)
>> +{
>> +    struct ttm_test_devices_priv *priv = test->priv;
>> +    const struct ttm_pool_test_case *params = test->param_value;
>> +    struct ttm_tt *tt;
>> +    struct ttm_pool *pool;
>> +    struct ttm_buffer_object *bo;
>> +    dma_addr_t dma1, dma2;
>> +    int err;
>> +    enum ttm_caching caching = ttm_uncached;
>> +    unsigned int expected_num_pages = 1 << params->order;
>> +    size_t size = expected_num_pages * PAGE_SIZE;
>> +
>> +    tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    bo = ttm_kunit_helper_ttm_bo_init(test, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>> +
>> +    err = ttm_sg_tt_init(tt, bo, 0, caching);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +
>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>> +
>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>> +
>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +    KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
>> +
>> +    dma1 = tt->dma_address[0];
>> +    dma2 = tt->dma_address[tt->num_pages - 1];
>> +
>> +    KUNIT_ASSERT_NOT_NULL(test, (void *)dma1);
>> +    KUNIT_ASSERT_NOT_NULL(test, (void *)dma2);
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +    ttm_pool_fini(pool);
>> +}
>> +
>> +static void ttm_pool_alloc_order_caching_match(struct kunit *test)
>> +{
>> +    struct ttm_tt *tt;
>> +    struct ttm_pool *pool;
>> +    struct ttm_pool_type *pt;
>> +    enum ttm_caching caching = ttm_uncached;
>> +    unsigned int order = 0;
>> +    size_t size = PAGE_SIZE;
>> +    int err;
>> +
>> +    pool = ttm_pool_pre_populated(test, size, caching);
>> +
>> +    pt = &pool->caching[caching].orders[order];
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>> +
>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +
>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +    ttm_pool_fini(pool);
>> +}
>> +
>> +static void ttm_pool_alloc_caching_mismatch(struct kunit *test)
>> +{
>> +    struct ttm_tt *tt;
>> +    struct ttm_pool *pool;
>> +    struct ttm_pool_type *pt_pool, *pt_tt;
>> +    int err;
>> +    enum ttm_caching tt_caching = ttm_uncached;
>> +    enum ttm_caching pool_caching = ttm_cached;
>> +    size_t size = PAGE_SIZE;
>> +    unsigned int order = 0;
>> +
>> +    pool = ttm_pool_pre_populated(test, size, pool_caching);
>> +
>> +    pt_pool = &pool->caching[pool_caching].orders[order];
>> +    pt_tt = &pool->caching[tt_caching].orders[order];
>> +
>> +    tt = mock_ttm_tt_init(test, 0, tt_caching, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
>> +
>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
>> +
>> +    ttm_pool_fini(pool);
>> +}
>> +
>> +static void ttm_pool_alloc_order_mismatch(struct kunit *test)
>> +{
>> +    struct ttm_tt *tt;
>> +    struct ttm_pool *pool;
>> +    struct ttm_pool_type *pt_pool, *pt_tt;
>> +    int err;
>> +    enum ttm_caching caching = ttm_uncached;
>> +    unsigned int order = 2;
>> +    size_t fst_size = (1 << order) * PAGE_SIZE;
>> +    size_t snd_size = PAGE_SIZE;
>> +
>> +    pool = ttm_pool_pre_populated(test, fst_size, caching);
>> +
>> +    pt_pool = &pool->caching[caching].orders[order];
>> +    pt_tt = &pool->caching[caching].orders[0];
>> +
>> +    tt = mock_ttm_tt_init(test, 0, caching, snd_size);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
>> +
>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>> +    KUNIT_ASSERT_EQ(test, err, 0);
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
>> +
>> +    ttm_pool_fini(pool);
>> +}
>> +
>> +static void ttm_pool_free_dma_alloc(struct kunit *test)
>> +{
>> +    struct ttm_test_devices_priv *priv = test->priv;
>> +    struct ttm_tt *tt;
>> +    struct ttm_pool *pool;
>> +    struct ttm_pool_type *pt;
>> +    enum ttm_caching caching = ttm_uncached;
>> +    unsigned int order = 2;
>> +    size_t size = (1 << order) * PAGE_SIZE;
>> +
>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>> +
>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>> +    ttm_pool_alloc(pool, tt, &simple_ctx);
>> +
>> +    pt = &pool->caching[caching].orders[order];
>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>> +
>> +    ttm_pool_fini(pool);
>> +}
>> +
>> +static void ttm_pool_free_no_dma_alloc(struct kunit *test)
>> +{
>> +    struct ttm_test_devices_priv *priv = test->priv;
>> +    struct ttm_tt *tt;
>> +    struct ttm_pool *pool;
>> +    struct ttm_pool_type *pt;
>> +    enum ttm_caching caching = ttm_uncached;
>> +    unsigned int order = 2;
>> +    size_t size = (1 << order) * PAGE_SIZE;
>> +
>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>> +
>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>> +
>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, false, false);
>> +    ttm_pool_alloc(pool, tt, &simple_ctx);
>> +
>> +    pt = &pool->caching[caching].orders[order];
>> +    KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
>> +
>> +    ttm_pool_free(pool, tt);
>> +    ttm_tt_fini(tt);
>> +
>> +    KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
>> +
>> +    ttm_pool_fini(pool);
>> +}
>> +
>> +static void ttm_pool_fini_basic(struct kunit *test)
>> +{
>> +    struct ttm_pool *pool;
>> +    struct ttm_pool_type *pt;
>> +    enum ttm_caching caching = ttm_uncached;
>> +    unsigned int order = 0;
>> +    size_t size = PAGE_SIZE;
>> +
>> +    pool = ttm_pool_pre_populated(test, size, caching);
>> +    pt = &pool->caching[caching].orders[order];
>> +
>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>> +
>> +    ttm_pool_fini(pool);
>> +
>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>> +}
>> +
>> +static struct kunit_case ttm_pool_test_cases[] = {
>> +    KUNIT_CASE_PARAM(ttm_pool_alloc_basic, 
>> ttm_pool_alloc_basic_gen_params),
>> +    KUNIT_CASE_PARAM(ttm_pool_alloc_basic_dma_addr,
>> +             ttm_pool_alloc_basic_gen_params),
>> +    KUNIT_CASE(ttm_pool_alloc_order_caching_match),
>> +    KUNIT_CASE(ttm_pool_alloc_caching_mismatch),
>> +    KUNIT_CASE(ttm_pool_alloc_order_mismatch),
>> +    KUNIT_CASE(ttm_pool_free_dma_alloc),
>> +    KUNIT_CASE(ttm_pool_free_no_dma_alloc),
>> +    KUNIT_CASE(ttm_pool_fini_basic),
>> +    {}
>> +};
>> +
>> +static struct kunit_suite ttm_pool_test_suite = {
>> +    .name = "ttm_pool",
>> +    .init = ttm_test_devices_init,
>> +    .exit = ttm_test_devices_fini,
>> +    .test_cases = ttm_pool_test_cases,
>> +};
>> +
>> +kunit_test_suites(&ttm_pool_test_suite);
>> +
>> +MODULE_LICENSE("GPL");
> 

Re: [RFC v3 3/3] drm/ttm/tests: Add tests for ttm_pool

[Mauro Carvalho Chehab]

On Mon, 3 Jul 2023 13:21:43 +0200
Christian König <christian.koenig@amd.com> wrote:

> > +	/* We don't need this BO later, release it */
> > +	ttm_bo_put(bo);  
> 
> That won't work. A BO must always outlive the TT object allocated for it.
> 
> Otherwise you can get crashes during TT release.

Just my two cents: shouldn't the TT object be increasing Kobj refcount then,
in order to properly manage the BO object lifetime?

Regards,
Mauro

Re: [RFC v3 3/3] drm/ttm/tests: Add tests for ttm_pool

[Christian König]

Am 04.07.23 um 08:50 schrieb Mauro Carvalho Chehab:
> On Mon, 3 Jul 2023 13:21:43 +0200
> Christian König <christian.koenig@amd.com> wrote:
>
>>> +	/* We don't need this BO later, release it */
>>> +	ttm_bo_put(bo);
>> That won't work. A BO must always outlive the TT object allocated for it.
>>
>> Otherwise you can get crashes during TT release.
> Just my two cents: shouldn't the TT object be increasing Kobj refcount then,
> in order to properly manage the BO object lifetime?

No, because you then get a circle reference.

The TT object is usually destroyed when the BO which is referencing it 
is destroyed.

Regards,
Christian.

>
> Regards,
> Mauro

Re: [RFC v3 3/3] drm/ttm/tests: Add tests for ttm_pool

[Christian König]

Am 04.07.23 um 08:35 schrieb Karolina Stolarek:
> On 3.07.2023 13:21, Christian König wrote:
>>
>>
>> Am 03.07.23 um 09:58 schrieb Karolina Stolarek:
>>> Add KUnit tests that exercise page allocation using page pools
>>> and freeing pages, either by returning them to the pool or
>>> freeing them. Add a basic test for ttm_pool cleanup. Introduce
>>> helpers to create a dummy ttm_buffer_object.
>>>
>>> Signed-off-by: Karolina Stolarek <karolina.stolarek@intel.com>
>>> ---
>>>   drivers/gpu/drm/ttm/tests/Makefile            |   1 +
>>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c |  15 +
>>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h |   4 +
>>>   drivers/gpu/drm/ttm/tests/ttm_pool_test.c     | 405 
>>> ++++++++++++++++++
>>>   4 files changed, 425 insertions(+)
>>>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>>
>>> diff --git a/drivers/gpu/drm/ttm/tests/Makefile 
>>> b/drivers/gpu/drm/ttm/tests/Makefile
>>> index 7917805f37af..ec87c4fc1ad5 100644
>>> --- a/drivers/gpu/drm/ttm/tests/Makefile
>>> +++ b/drivers/gpu/drm/ttm/tests/Makefile
>>> @@ -2,4 +2,5 @@
>>>   obj-$(CONFIG_DRM_TTM_KUNIT_TEST) += \
>>>           ttm_device_test.o \
>>> +        ttm_pool_test.o \
>>>           ttm_kunit_helpers.o
>>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c 
>>> b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>>> index 428a8a4ac9dc..9384679fccb0 100644
>>> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>>> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>>> @@ -26,6 +26,21 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>>>   }
>>>   EXPORT_SYMBOL_GPL(ttm_kunit_helper_init_device);
>>> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit 
>>> *test,
>>> +                               size_t size)
>>> +{
>>> +    struct drm_gem_object gem_obj = { .size = size };
>>> +    struct ttm_buffer_object *bo;
>>> +
>>> +    bo = kunit_kzalloc(test, sizeof(*bo), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>>> +
>>> +    bo->base = gem_obj;
>>> +
>>> +    return bo;
>>> +}
>>> +EXPORT_SYMBOL_GPL(ttm_kunit_helper_ttm_bo_init);
>>> +
>>>   int ttm_test_devices_init(struct kunit *test)
>>>   {
>>>       struct ttm_test_devices_priv *priv;
>>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h 
>>> b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>>> index bf0b1000a8ee..41ad38dd34e5 100644
>>> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>>> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>>> @@ -7,6 +7,7 @@
>>>   #include <drm/drm_drv.h>
>>>   #include <drm/ttm/ttm_device.h>
>>> +#include <drm/ttm/ttm_bo.h>
>>>   #include <drm/drm_kunit_helpers.h>
>>>   #include <kunit/test.h>
>>> @@ -23,6 +24,9 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>>>                     bool use_dma_alloc,
>>>                     bool use_dma32);
>>> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit 
>>> *test,
>>> +                               size_t size);
>>> +
>>>   int ttm_test_devices_init(struct kunit *test);
>>>   void ttm_test_devices_fini(struct kunit *test);
>>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_pool_test.c 
>>> b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>> new file mode 100644
>>> index 000000000000..85806f72df22
>>> --- /dev/null
>>> +++ b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>> @@ -0,0 +1,405 @@
>>> +// SPDX-License-Identifier: GPL-2.0 AND MIT
>>> +/*
>>> + * Copyright © 2023 Intel Corporation
>>> + */
>>> +#include <linux/mm.h>
>>> +
>>> +#include <drm/ttm/ttm_tt.h>
>>> +#include <drm/ttm/ttm_pool.h>
>>> +
>>> +#include "ttm_kunit_helpers.h"
>>> +
>>> +struct ttm_pool_test_case {
>>> +    const char *description;
>>> +    unsigned int order;
>>> +    bool use_dma_alloc;
>>> +};
>>> +
>>> +static struct ttm_operation_ctx simple_ctx = {
>>> +    .interruptible = true,
>>> +    .no_wait_gpu = false,
>>> +};
>>> +
>>> +static struct ttm_tt *mock_ttm_tt_init(struct kunit *test,
>>> +                       uint32_t page_flags,
>>> +                       enum ttm_caching caching,
>>> +                       size_t size)
>>> +{
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_buffer_object *bo;
>>> +    int err;
>>> +
>>> +    bo = ttm_kunit_helper_ttm_bo_init(test, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>>> +
>>> +    tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    err = ttm_tt_init(tt, bo, page_flags, caching, 0);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +
>>> +    /* We don't need this BO later, release it */
>>> +    ttm_bo_put(bo);
>>
>> That won't work. A BO must always outlive the TT object allocated for 
>> it. >
>> Otherwise you can get crashes during TT release.
>
> OK, I can keep it around, probably as an another test->priv member.
>
> I thought that for these mocks it should be fine -- 
> ttm_tt_init_fields() only takes the scatter list from BO which is not 
> in use (as for now, at least) and in ttm_tt_fini() we don't even touch 
> that field.

Yeah, it might work for the mock test. But this is not how the interface 
is supposed to be used.

The lifetime of the TT object is tied to the one of the BO, e.g. when 
the BO is destroyed and has a TT object associated with it that object 
is destroyed as well.

Regards,
Christian.

>
> All the best,
> Karolina
>
>>
>> Regards,
>> Christian.
>>
>>> +
>>> +    return tt;
>>> +}
>>> +
>>> +static struct ttm_pool *ttm_pool_pre_populated(struct kunit *test,
>>> +                           size_t size,
>>> +                           enum ttm_caching caching)
>>> +{
>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_tt *tt;
>>> +    int err;
>>> +    unsigned long order = __fls(size / PAGE_SIZE);
>>> +
>>> +    tt = mock_ttm_tt_init(test, order, caching, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>> +
>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>>> +
>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +
>>> +    return pool;
>>> +}
>>> +
>>> +static const struct ttm_pool_test_case ttm_pool_basic_cases[] = {
>>> +    {
>>> +        .description = "One page",
>>> +        .order = 0,
>>> +    },
>>> +    {
>>> +        .description = "More than one page",
>>> +        .order = 2,
>>> +    },
>>> +    {
>>> +        .description = "Above the allocation limit",
>>> +        .order = MAX_ORDER + 1,
>>> +    },
>>> +    {
>>> +        .description = "One page, with coherent DMA mappings enabled",
>>> +        .order = 0,
>>> +        .use_dma_alloc = true,
>>> +    },
>>> +    {
>>> +        .description = "Above the allocation limit, with coherent 
>>> DMA mappings enabled",
>>> +        .order = MAX_ORDER + 1,
>>> +        .use_dma_alloc = true,
>>> +    },
>>> +};
>>> +
>>> +static void ttm_pool_alloc_case_desc(const struct 
>>> ttm_pool_test_case *t,
>>> +                     char *desc)
>>> +{
>>> +    strscpy(desc, t->description, KUNIT_PARAM_DESC_SIZE);
>>> +}
>>> +
>>> +KUNIT_ARRAY_PARAM(ttm_pool_alloc_basic, ttm_pool_basic_cases,
>>> +          ttm_pool_alloc_case_desc);
>>> +
>>> +static void ttm_pool_alloc_basic(struct kunit *test)
>>> +{
>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>> +    const struct ttm_pool_test_case *params = test->param_value;
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_pool *pool;
>>> +    struct page *fst_page, *last_page;
>>> +    int err;
>>> +    enum ttm_caching caching = ttm_uncached;
>>> +    unsigned int expected_num_pages = 1 << params->order;
>>> +    size_t size = expected_num_pages * PAGE_SIZE;
>>> +
>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, 
>>> params->use_dma_alloc,
>>> +              false);
>>> +
>>> +    KUNIT_ASSERT_PTR_EQ(test, pool->dev, priv->dev);
>>> +    KUNIT_ASSERT_EQ(test, pool->nid, NUMA_NO_NODE);
>>> +    KUNIT_ASSERT_EQ(test, pool->use_dma_alloc, params->use_dma_alloc);
>>> +
>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +    KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
>>> +
>>> +    fst_page = tt->pages[0];
>>> +    last_page = tt->pages[tt->num_pages - 1];
>>> +
>>> +    if (params->order <= MAX_ORDER) {
>>> +        if (params->use_dma_alloc) {
>>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
>>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)last_page->private);
>>> +        } else {
>>> +            KUNIT_ASSERT_EQ(test, fst_page->private, params->order);
>>> +        }
>>> +    } else {
>>> +        if (params->use_dma_alloc) {
>>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
>>> +            KUNIT_ASSERT_NULL(test, (void *)last_page->private);
>>> +        } else {
>>> +            /*
>>> +             * We expect to alloc one big block, followed by
>>> +             * order 0 blocks
>>> +             */
>>> +            KUNIT_ASSERT_EQ(test, fst_page->private,
>>> +                    min_t(unsigned int, MAX_ORDER,
>>> +                          params->order));
>>> +            KUNIT_ASSERT_EQ(test, last_page->private, 0);
>>> +        }
>>> +    }
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +    ttm_pool_fini(pool);
>>> +}
>>> +
>>> +static void ttm_pool_alloc_basic_dma_addr(struct kunit *test)
>>> +{
>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>> +    const struct ttm_pool_test_case *params = test->param_value;
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_buffer_object *bo;
>>> +    dma_addr_t dma1, dma2;
>>> +    int err;
>>> +    enum ttm_caching caching = ttm_uncached;
>>> +    unsigned int expected_num_pages = 1 << params->order;
>>> +    size_t size = expected_num_pages * PAGE_SIZE;
>>> +
>>> +    tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    bo = ttm_kunit_helper_ttm_bo_init(test, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>>> +
>>> +    err = ttm_sg_tt_init(tt, bo, 0, caching);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +
>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>> +
>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>>> +
>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +    KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
>>> +
>>> +    dma1 = tt->dma_address[0];
>>> +    dma2 = tt->dma_address[tt->num_pages - 1];
>>> +
>>> +    KUNIT_ASSERT_NOT_NULL(test, (void *)dma1);
>>> +    KUNIT_ASSERT_NOT_NULL(test, (void *)dma2);
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +    ttm_pool_fini(pool);
>>> +}
>>> +
>>> +static void ttm_pool_alloc_order_caching_match(struct kunit *test)
>>> +{
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_pool_type *pt;
>>> +    enum ttm_caching caching = ttm_uncached;
>>> +    unsigned int order = 0;
>>> +    size_t size = PAGE_SIZE;
>>> +    int err;
>>> +
>>> +    pool = ttm_pool_pre_populated(test, size, caching);
>>> +
>>> +    pt = &pool->caching[caching].orders[order];
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>>> +
>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +
>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +    ttm_pool_fini(pool);
>>> +}
>>> +
>>> +static void ttm_pool_alloc_caching_mismatch(struct kunit *test)
>>> +{
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_pool_type *pt_pool, *pt_tt;
>>> +    int err;
>>> +    enum ttm_caching tt_caching = ttm_uncached;
>>> +    enum ttm_caching pool_caching = ttm_cached;
>>> +    size_t size = PAGE_SIZE;
>>> +    unsigned int order = 0;
>>> +
>>> +    pool = ttm_pool_pre_populated(test, size, pool_caching);
>>> +
>>> +    pt_pool = &pool->caching[pool_caching].orders[order];
>>> +    pt_tt = &pool->caching[tt_caching].orders[order];
>>> +
>>> +    tt = mock_ttm_tt_init(test, 0, tt_caching, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
>>> +
>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
>>> +
>>> +    ttm_pool_fini(pool);
>>> +}
>>> +
>>> +static void ttm_pool_alloc_order_mismatch(struct kunit *test)
>>> +{
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_pool_type *pt_pool, *pt_tt;
>>> +    int err;
>>> +    enum ttm_caching caching = ttm_uncached;
>>> +    unsigned int order = 2;
>>> +    size_t fst_size = (1 << order) * PAGE_SIZE;
>>> +    size_t snd_size = PAGE_SIZE;
>>> +
>>> +    pool = ttm_pool_pre_populated(test, fst_size, caching);
>>> +
>>> +    pt_pool = &pool->caching[caching].orders[order];
>>> +    pt_tt = &pool->caching[caching].orders[0];
>>> +
>>> +    tt = mock_ttm_tt_init(test, 0, caching, snd_size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
>>> +
>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
>>> +
>>> +    ttm_pool_fini(pool);
>>> +}
>>> +
>>> +static void ttm_pool_free_dma_alloc(struct kunit *test)
>>> +{
>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_pool_type *pt;
>>> +    enum ttm_caching caching = ttm_uncached;
>>> +    unsigned int order = 2;
>>> +    size_t size = (1 << order) * PAGE_SIZE;
>>> +
>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>> +
>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>>> +    ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +
>>> +    pt = &pool->caching[caching].orders[order];
>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>>> +
>>> +    ttm_pool_fini(pool);
>>> +}
>>> +
>>> +static void ttm_pool_free_no_dma_alloc(struct kunit *test)
>>> +{
>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>> +    struct ttm_tt *tt;
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_pool_type *pt;
>>> +    enum ttm_caching caching = ttm_uncached;
>>> +    unsigned int order = 2;
>>> +    size_t size = (1 << order) * PAGE_SIZE;
>>> +
>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>> +
>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>> +
>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, false, false);
>>> +    ttm_pool_alloc(pool, tt, &simple_ctx);
>>> +
>>> +    pt = &pool->caching[caching].orders[order];
>>> +    KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
>>> +
>>> +    ttm_pool_free(pool, tt);
>>> +    ttm_tt_fini(tt);
>>> +
>>> +    KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
>>> +
>>> +    ttm_pool_fini(pool);
>>> +}
>>> +
>>> +static void ttm_pool_fini_basic(struct kunit *test)
>>> +{
>>> +    struct ttm_pool *pool;
>>> +    struct ttm_pool_type *pt;
>>> +    enum ttm_caching caching = ttm_uncached;
>>> +    unsigned int order = 0;
>>> +    size_t size = PAGE_SIZE;
>>> +
>>> +    pool = ttm_pool_pre_populated(test, size, caching);
>>> +    pt = &pool->caching[caching].orders[order];
>>> +
>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>>> +
>>> +    ttm_pool_fini(pool);
>>> +
>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>>> +}
>>> +
>>> +static struct kunit_case ttm_pool_test_cases[] = {
>>> +    KUNIT_CASE_PARAM(ttm_pool_alloc_basic, 
>>> ttm_pool_alloc_basic_gen_params),
>>> +    KUNIT_CASE_PARAM(ttm_pool_alloc_basic_dma_addr,
>>> +             ttm_pool_alloc_basic_gen_params),
>>> +    KUNIT_CASE(ttm_pool_alloc_order_caching_match),
>>> +    KUNIT_CASE(ttm_pool_alloc_caching_mismatch),
>>> +    KUNIT_CASE(ttm_pool_alloc_order_mismatch),
>>> +    KUNIT_CASE(ttm_pool_free_dma_alloc),
>>> +    KUNIT_CASE(ttm_pool_free_no_dma_alloc),
>>> +    KUNIT_CASE(ttm_pool_fini_basic),
>>> +    {}
>>> +};
>>> +
>>> +static struct kunit_suite ttm_pool_test_suite = {
>>> +    .name = "ttm_pool",
>>> +    .init = ttm_test_devices_init,
>>> +    .exit = ttm_test_devices_fini,
>>> +    .test_cases = ttm_pool_test_cases,
>>> +};
>>> +
>>> +kunit_test_suites(&ttm_pool_test_suite);
>>> +
>>> +MODULE_LICENSE("GPL");
>>

Re: [RFC v3 3/3] drm/ttm/tests: Add tests for ttm_pool

[Karolina Stolarek]

On 4.07.2023 09:01, Christian König wrote:
> Am 04.07.23 um 08:35 schrieb Karolina Stolarek:
>> On 3.07.2023 13:21, Christian König wrote:
>>>
>>>
>>> Am 03.07.23 um 09:58 schrieb Karolina Stolarek:
>>>> Add KUnit tests that exercise page allocation using page pools
>>>> and freeing pages, either by returning them to the pool or
>>>> freeing them. Add a basic test for ttm_pool cleanup. Introduce
>>>> helpers to create a dummy ttm_buffer_object.
>>>>
>>>> Signed-off-by: Karolina Stolarek <karolina.stolarek@intel.com>
>>>> ---
>>>>   drivers/gpu/drm/ttm/tests/Makefile            |   1 +
>>>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c |  15 +
>>>>   drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h |   4 +
>>>>   drivers/gpu/drm/ttm/tests/ttm_pool_test.c     | 405 
>>>> ++++++++++++++++++
>>>>   4 files changed, 425 insertions(+)
>>>>   create mode 100644 drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>>>
>>>> diff --git a/drivers/gpu/drm/ttm/tests/Makefile 
>>>> b/drivers/gpu/drm/ttm/tests/Makefile
>>>> index 7917805f37af..ec87c4fc1ad5 100644
>>>> --- a/drivers/gpu/drm/ttm/tests/Makefile
>>>> +++ b/drivers/gpu/drm/ttm/tests/Makefile
>>>> @@ -2,4 +2,5 @@
>>>>   obj-$(CONFIG_DRM_TTM_KUNIT_TEST) += \
>>>>           ttm_device_test.o \
>>>> +        ttm_pool_test.o \
>>>>           ttm_kunit_helpers.o
>>>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c 
>>>> b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>>>> index 428a8a4ac9dc..9384679fccb0 100644
>>>> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>>>> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
>>>> @@ -26,6 +26,21 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>>>>   }
>>>>   EXPORT_SYMBOL_GPL(ttm_kunit_helper_init_device);
>>>> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit 
>>>> *test,
>>>> +                               size_t size)
>>>> +{
>>>> +    struct drm_gem_object gem_obj = { .size = size };
>>>> +    struct ttm_buffer_object *bo;
>>>> +
>>>> +    bo = kunit_kzalloc(test, sizeof(*bo), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>>>> +
>>>> +    bo->base = gem_obj;
>>>> +
>>>> +    return bo;
>>>> +}
>>>> +EXPORT_SYMBOL_GPL(ttm_kunit_helper_ttm_bo_init);
>>>> +
>>>>   int ttm_test_devices_init(struct kunit *test)
>>>>   {
>>>>       struct ttm_test_devices_priv *priv;
>>>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h 
>>>> b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>>>> index bf0b1000a8ee..41ad38dd34e5 100644
>>>> --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>>>> +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
>>>> @@ -7,6 +7,7 @@
>>>>   #include <drm/drm_drv.h>
>>>>   #include <drm/ttm/ttm_device.h>
>>>> +#include <drm/ttm/ttm_bo.h>
>>>>   #include <drm/drm_kunit_helpers.h>
>>>>   #include <kunit/test.h>
>>>> @@ -23,6 +24,9 @@ int ttm_kunit_helper_init_device(struct kunit *test,
>>>>                     bool use_dma_alloc,
>>>>                     bool use_dma32);
>>>> +struct ttm_buffer_object *ttm_kunit_helper_ttm_bo_init(struct kunit 
>>>> *test,
>>>> +                               size_t size);
>>>> +
>>>>   int ttm_test_devices_init(struct kunit *test);
>>>>   void ttm_test_devices_fini(struct kunit *test);
>>>> diff --git a/drivers/gpu/drm/ttm/tests/ttm_pool_test.c 
>>>> b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>>> new file mode 100644
>>>> index 000000000000..85806f72df22
>>>> --- /dev/null
>>>> +++ b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
>>>> @@ -0,0 +1,405 @@
>>>> +// SPDX-License-Identifier: GPL-2.0 AND MIT
>>>> +/*
>>>> + * Copyright © 2023 Intel Corporation
>>>> + */
>>>> +#include <linux/mm.h>
>>>> +
>>>> +#include <drm/ttm/ttm_tt.h>
>>>> +#include <drm/ttm/ttm_pool.h>
>>>> +
>>>> +#include "ttm_kunit_helpers.h"
>>>> +
>>>> +struct ttm_pool_test_case {
>>>> +    const char *description;
>>>> +    unsigned int order;
>>>> +    bool use_dma_alloc;
>>>> +};
>>>> +
>>>> +static struct ttm_operation_ctx simple_ctx = {
>>>> +    .interruptible = true,
>>>> +    .no_wait_gpu = false,
>>>> +};
>>>> +
>>>> +static struct ttm_tt *mock_ttm_tt_init(struct kunit *test,
>>>> +                       uint32_t page_flags,
>>>> +                       enum ttm_caching caching,
>>>> +                       size_t size)
>>>> +{
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_buffer_object *bo;
>>>> +    int err;
>>>> +
>>>> +    bo = ttm_kunit_helper_ttm_bo_init(test, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>>>> +
>>>> +    tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    err = ttm_tt_init(tt, bo, page_flags, caching, 0);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +
>>>> +    /* We don't need this BO later, release it */
>>>> +    ttm_bo_put(bo);
>>>
>>> That won't work. A BO must always outlive the TT object allocated for 
>>> it. >
>>> Otherwise you can get crashes during TT release.
>>
>> OK, I can keep it around, probably as an another test->priv member.
>>
>> I thought that for these mocks it should be fine -- 
>> ttm_tt_init_fields() only takes the scatter list from BO which is not 
>> in use (as for now, at least) and in ttm_tt_fini() we don't even touch 
>> that field.
> 
> Yeah, it might work for the mock test. But this is not how the interface 
> is supposed to be used.

I'll correct it in the next version then, thanks for taking a look.

All the best,
Karolina

> 
> The lifetime of the TT object is tied to the one of the BO, e.g. when 
> the BO is destroyed and has a TT object associated with it that object 
> is destroyed as well.
> 
> Regards,
> Christian.
> 
>>
>> All the best,
>> Karolina
>>
>>>
>>> Regards,
>>> Christian.
>>>
>>>> +
>>>> +    return tt;
>>>> +}
>>>> +
>>>> +static struct ttm_pool *ttm_pool_pre_populated(struct kunit *test,
>>>> +                           size_t size,
>>>> +                           enum ttm_caching caching)
>>>> +{
>>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_tt *tt;
>>>> +    int err;
>>>> +    unsigned long order = __fls(size / PAGE_SIZE);
>>>> +
>>>> +    tt = mock_ttm_tt_init(test, order, caching, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>>> +
>>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>>>> +
>>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +
>>>> +    return pool;
>>>> +}
>>>> +
>>>> +static const struct ttm_pool_test_case ttm_pool_basic_cases[] = {
>>>> +    {
>>>> +        .description = "One page",
>>>> +        .order = 0,
>>>> +    },
>>>> +    {
>>>> +        .description = "More than one page",
>>>> +        .order = 2,
>>>> +    },
>>>> +    {
>>>> +        .description = "Above the allocation limit",
>>>> +        .order = MAX_ORDER + 1,
>>>> +    },
>>>> +    {
>>>> +        .description = "One page, with coherent DMA mappings enabled",
>>>> +        .order = 0,
>>>> +        .use_dma_alloc = true,
>>>> +    },
>>>> +    {
>>>> +        .description = "Above the allocation limit, with coherent 
>>>> DMA mappings enabled",
>>>> +        .order = MAX_ORDER + 1,
>>>> +        .use_dma_alloc = true,
>>>> +    },
>>>> +};
>>>> +
>>>> +static void ttm_pool_alloc_case_desc(const struct 
>>>> ttm_pool_test_case *t,
>>>> +                     char *desc)
>>>> +{
>>>> +    strscpy(desc, t->description, KUNIT_PARAM_DESC_SIZE);
>>>> +}
>>>> +
>>>> +KUNIT_ARRAY_PARAM(ttm_pool_alloc_basic, ttm_pool_basic_cases,
>>>> +          ttm_pool_alloc_case_desc);
>>>> +
>>>> +static void ttm_pool_alloc_basic(struct kunit *test)
>>>> +{
>>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>>> +    const struct ttm_pool_test_case *params = test->param_value;
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_pool *pool;
>>>> +    struct page *fst_page, *last_page;
>>>> +    int err;
>>>> +    enum ttm_caching caching = ttm_uncached;
>>>> +    unsigned int expected_num_pages = 1 << params->order;
>>>> +    size_t size = expected_num_pages * PAGE_SIZE;
>>>> +
>>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, 
>>>> params->use_dma_alloc,
>>>> +              false);
>>>> +
>>>> +    KUNIT_ASSERT_PTR_EQ(test, pool->dev, priv->dev);
>>>> +    KUNIT_ASSERT_EQ(test, pool->nid, NUMA_NO_NODE);
>>>> +    KUNIT_ASSERT_EQ(test, pool->use_dma_alloc, params->use_dma_alloc);
>>>> +
>>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +    KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
>>>> +
>>>> +    fst_page = tt->pages[0];
>>>> +    last_page = tt->pages[tt->num_pages - 1];
>>>> +
>>>> +    if (params->order <= MAX_ORDER) {
>>>> +        if (params->use_dma_alloc) {
>>>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
>>>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)last_page->private);
>>>> +        } else {
>>>> +            KUNIT_ASSERT_EQ(test, fst_page->private, params->order);
>>>> +        }
>>>> +    } else {
>>>> +        if (params->use_dma_alloc) {
>>>> +            KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
>>>> +            KUNIT_ASSERT_NULL(test, (void *)last_page->private);
>>>> +        } else {
>>>> +            /*
>>>> +             * We expect to alloc one big block, followed by
>>>> +             * order 0 blocks
>>>> +             */
>>>> +            KUNIT_ASSERT_EQ(test, fst_page->private,
>>>> +                    min_t(unsigned int, MAX_ORDER,
>>>> +                          params->order));
>>>> +            KUNIT_ASSERT_EQ(test, last_page->private, 0);
>>>> +        }
>>>> +    }
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +    ttm_pool_fini(pool);
>>>> +}
>>>> +
>>>> +static void ttm_pool_alloc_basic_dma_addr(struct kunit *test)
>>>> +{
>>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>>> +    const struct ttm_pool_test_case *params = test->param_value;
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_buffer_object *bo;
>>>> +    dma_addr_t dma1, dma2;
>>>> +    int err;
>>>> +    enum ttm_caching caching = ttm_uncached;
>>>> +    unsigned int expected_num_pages = 1 << params->order;
>>>> +    size_t size = expected_num_pages * PAGE_SIZE;
>>>> +
>>>> +    tt = kunit_kzalloc(test, sizeof(*tt), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    bo = ttm_kunit_helper_ttm_bo_init(test, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, bo);
>>>> +
>>>> +    err = ttm_sg_tt_init(tt, bo, 0, caching);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +
>>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>>> +
>>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>>>> +
>>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +    KUNIT_ASSERT_EQ(test, tt->num_pages, expected_num_pages);
>>>> +
>>>> +    dma1 = tt->dma_address[0];
>>>> +    dma2 = tt->dma_address[tt->num_pages - 1];
>>>> +
>>>> +    KUNIT_ASSERT_NOT_NULL(test, (void *)dma1);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, (void *)dma2);
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +    ttm_pool_fini(pool);
>>>> +}
>>>> +
>>>> +static void ttm_pool_alloc_order_caching_match(struct kunit *test)
>>>> +{
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_pool_type *pt;
>>>> +    enum ttm_caching caching = ttm_uncached;
>>>> +    unsigned int order = 0;
>>>> +    size_t size = PAGE_SIZE;
>>>> +    int err;
>>>> +
>>>> +    pool = ttm_pool_pre_populated(test, size, caching);
>>>> +
>>>> +    pt = &pool->caching[caching].orders[order];
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>>>> +
>>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +
>>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +    ttm_pool_fini(pool);
>>>> +}
>>>> +
>>>> +static void ttm_pool_alloc_caching_mismatch(struct kunit *test)
>>>> +{
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_pool_type *pt_pool, *pt_tt;
>>>> +    int err;
>>>> +    enum ttm_caching tt_caching = ttm_uncached;
>>>> +    enum ttm_caching pool_caching = ttm_cached;
>>>> +    size_t size = PAGE_SIZE;
>>>> +    unsigned int order = 0;
>>>> +
>>>> +    pool = ttm_pool_pre_populated(test, size, pool_caching);
>>>> +
>>>> +    pt_pool = &pool->caching[pool_caching].orders[order];
>>>> +    pt_tt = &pool->caching[tt_caching].orders[order];
>>>> +
>>>> +    tt = mock_ttm_tt_init(test, 0, tt_caching, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
>>>> +
>>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
>>>> +
>>>> +    ttm_pool_fini(pool);
>>>> +}
>>>> +
>>>> +static void ttm_pool_alloc_order_mismatch(struct kunit *test)
>>>> +{
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_pool_type *pt_pool, *pt_tt;
>>>> +    int err;
>>>> +    enum ttm_caching caching = ttm_uncached;
>>>> +    unsigned int order = 2;
>>>> +    size_t fst_size = (1 << order) * PAGE_SIZE;
>>>> +    size_t snd_size = PAGE_SIZE;
>>>> +
>>>> +    pool = ttm_pool_pre_populated(test, fst_size, caching);
>>>> +
>>>> +    pt_pool = &pool->caching[caching].orders[order];
>>>> +    pt_tt = &pool->caching[caching].orders[0];
>>>> +
>>>> +    tt = mock_ttm_tt_init(test, 0, caching, snd_size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt_tt->pages));
>>>> +
>>>> +    err = ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +    KUNIT_ASSERT_EQ(test, err, 0);
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_pool->pages));
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt_tt->pages));
>>>> +
>>>> +    ttm_pool_fini(pool);
>>>> +}
>>>> +
>>>> +static void ttm_pool_free_dma_alloc(struct kunit *test)
>>>> +{
>>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_pool_type *pt;
>>>> +    enum ttm_caching caching = ttm_uncached;
>>>> +    unsigned int order = 2;
>>>> +    size_t size = (1 << order) * PAGE_SIZE;
>>>> +
>>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>>> +
>>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, true, false);
>>>> +    ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +
>>>> +    pt = &pool->caching[caching].orders[order];
>>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>>>> +
>>>> +    ttm_pool_fini(pool);
>>>> +}
>>>> +
>>>> +static void ttm_pool_free_no_dma_alloc(struct kunit *test)
>>>> +{
>>>> +    struct ttm_test_devices_priv *priv = test->priv;
>>>> +    struct ttm_tt *tt;
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_pool_type *pt;
>>>> +    enum ttm_caching caching = ttm_uncached;
>>>> +    unsigned int order = 2;
>>>> +    size_t size = (1 << order) * PAGE_SIZE;
>>>> +
>>>> +    tt = mock_ttm_tt_init(test, 0, caching, size);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, tt);
>>>> +
>>>> +    pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
>>>> +    KUNIT_ASSERT_NOT_NULL(test, pool);
>>>> +
>>>> +    ttm_pool_init(pool, priv->dev, NUMA_NO_NODE, false, false);
>>>> +    ttm_pool_alloc(pool, tt, &simple_ctx);
>>>> +
>>>> +    pt = &pool->caching[caching].orders[order];
>>>> +    KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
>>>> +
>>>> +    ttm_pool_free(pool, tt);
>>>> +    ttm_tt_fini(tt);
>>>> +
>>>> +    KUNIT_ASSERT_TRUE(test, list_is_singular(&pt->pages));
>>>> +
>>>> +    ttm_pool_fini(pool);
>>>> +}
>>>> +
>>>> +static void ttm_pool_fini_basic(struct kunit *test)
>>>> +{
>>>> +    struct ttm_pool *pool;
>>>> +    struct ttm_pool_type *pt;
>>>> +    enum ttm_caching caching = ttm_uncached;
>>>> +    unsigned int order = 0;
>>>> +    size_t size = PAGE_SIZE;
>>>> +
>>>> +    pool = ttm_pool_pre_populated(test, size, caching);
>>>> +    pt = &pool->caching[caching].orders[order];
>>>> +
>>>> +    KUNIT_ASSERT_FALSE(test, list_empty(&pt->pages));
>>>> +
>>>> +    ttm_pool_fini(pool);
>>>> +
>>>> +    KUNIT_ASSERT_TRUE(test, list_empty(&pt->pages));
>>>> +}
>>>> +
>>>> +static struct kunit_case ttm_pool_test_cases[] = {
>>>> +    KUNIT_CASE_PARAM(ttm_pool_alloc_basic, 
>>>> ttm_pool_alloc_basic_gen_params),
>>>> +    KUNIT_CASE_PARAM(ttm_pool_alloc_basic_dma_addr,
>>>> +             ttm_pool_alloc_basic_gen_params),
>>>> +    KUNIT_CASE(ttm_pool_alloc_order_caching_match),
>>>> +    KUNIT_CASE(ttm_pool_alloc_caching_mismatch),
>>>> +    KUNIT_CASE(ttm_pool_alloc_order_mismatch),
>>>> +    KUNIT_CASE(ttm_pool_free_dma_alloc),
>>>> +    KUNIT_CASE(ttm_pool_free_no_dma_alloc),
>>>> +    KUNIT_CASE(ttm_pool_fini_basic),
>>>> +    {}
>>>> +};
>>>> +
>>>> +static struct kunit_suite ttm_pool_test_suite = {
>>>> +    .name = "ttm_pool",
>>>> +    .init = ttm_test_devices_init,
>>>> +    .exit = ttm_test_devices_fini,
>>>> +    .test_cases = ttm_pool_test_cases,
>>>> +};
>>>> +
>>>> +kunit_test_suites(&ttm_pool_test_suite);
>>>> +
>>>> +MODULE_LICENSE("GPL");
>>>
>