On 11/23/22 16:25, Maxime Ripard wrote:

The device managed resources are ran if the device has bus, which is not the case of a root_device.

Let's use a platform_device instead.

Signed-off-by: Maxime Ripard maxime@cerno.tech

drivers/gpu/drm/tests/drm_kunit_helpers.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/tests/drm_kunit_helpers.c b/drivers/gpu/drm/tests/drm_kunit_helpers.c index 2f974f45c5b4..651aa93a98a8 100644 --- a/drivers/gpu/drm/tests/drm_kunit_helpers.c +++ b/drivers/gpu/drm/tests/drm_kunit_helpers.c @@ -6,6 +6,7 @@ #include <kunit/resource.h> #include <linux/device.h> +#include <linux/platform_device.h> #include "drm_kunit_helpers.h" @@ -20,12 +21,26 @@ static const struct drm_mode_config_funcs drm_mode_config_funcs = { struct device *drm_kunit_helper_alloc_device(struct kunit *test) {

• return root_device_register(FAKE_DEVICE_NAME);

• struct platform_device *pdev;
• int ret;
• ret = platform_driver_register(&fake_platform_driver);

Maybe the platform_driver_register() is something that should be done in the drm_client_modeset_test_init() callback ?

I would also as a part of this series rename the kunit_suite from "drm_test_pick_cmdline_test_suite" to "drm_test_modeset_test_suite" or something like that. Since you are already setting the .init and .exit callbacks to allocate and release the devices. So it is more than just a test suite for cmdline params testing.

• KUNIT_ASSERT_EQ(test, ret, 0);
• pdev = platform_device_alloc(FAKE_DEVICE_NAME, PLATFORM_DEVID_NONE);
• KUNIT_ASSERT_NOT_ERR_OR_NULL(test, pdev);
• ret = platform_device_add(pdev);
• KUNIT_ASSERT_EQ(test, ret, 0);
• return &pdev->dev;

}

Probably won't do it as a part of this series to avoid making it cross subsystem, but I think there's room to further consolidation since this helper seems to be quite generic and something that would be useful for other subsystems. So it would make sense for kunit to have helpers to manage platform drivers and devices.

In fact, I see that drivers/base/test/test_async_driver_probe.c already has similar helpers that could be exposed to the rest of the kernel.

The patch looks good to me though and any of these changes could be done as a follow-up:

Reviewed-by: Javier Martinez Canillas javierm@redhat.com