On Tue, 21 Jul 2026, Arunpravin Paneer Selvam <[email protected]>
wrote:
> Add a new unit test case for buffer clearance issue during
> resume.
>
> Using a non-power-of-two mm size, allocate alternating blocks of
> 4KiB in an even sequence and free them as cleared. All alternate
> blocks should be marked as dirty and the split blocks should be
> merged back to their original size when the blocks clear reset
> function is called.
>
> Signed-off-by: Arunpravin Paneer Selvam <[email protected]>
> Reviewed-by: Matthew Auld <[email protected]>
> ---
> v6: Standalone repost of patch 3/3 from series 151686; the other two
> patches are already applied. Rebased onto the gpu_buddy rename
> (drm_buddy_test.c -> gpu_buddy_test.c, drm_buddy_* -> gpu_buddy_*)
> and updated the subject accordingly. No functional change; kept
> Matthew's Reviewed-by as the port is mechanical.
>
> drivers/gpu/tests/gpu_buddy_test.c | 41 ++++++++++++++++++++++++++++++
> 1 file changed, 41 insertions(+)
>
> diff --git a/drivers/gpu/tests/gpu_buddy_test.c
> b/drivers/gpu/tests/gpu_buddy_test.c
> index 7df5c2ae83bb..89698563c61b 100644
> --- a/drivers/gpu/tests/gpu_buddy_test.c
> +++ b/drivers/gpu/tests/gpu_buddy_test.c
> @@ -1002,6 +1002,47 @@ static void gpu_test_buddy_alloc_clear(struct kunit
> *test)
> "buddy_alloc hit an error size=%lu\n", ps);
> gpu_buddy_free_list(&mm, &allocated, GPU_BUDDY_CLEARED);
> gpu_buddy_fini(&mm);
> +
> + /*
> + * Using a non-power-of-two mm size, allocate alternating blocks of
> 4KiB in an
> + * even sequence and free them as cleared. All blocks should be marked
> as
> + * dirty and the split blocks should be merged back to their original
> + * size when the blocks clear reset function is called.
> + */
> + KUNIT_EXPECT_FALSE(test, gpu_buddy_init(&mm, mm_size, ps));
> + KUNIT_EXPECT_EQ(test, mm.max_order, max_order);
> +
> + i = 0;
> + n_pages = mm_size / ps;
> + do {
> + if (i % 2 == 0)
> + KUNIT_ASSERT_FALSE_MSG(test,
> gpu_buddy_alloc_blocks(&mm, 0, mm_size,
> + ps,
> ps, &allocated, 0),
> + "buddy_alloc hit an error size=%lu\n",
> ps);
> + } while (++i < n_pages);
Please prefer a for loop over do-while or while. This is so much more
obvious:
for (i = 0; i < n_pages; i++)
> +
> + gpu_buddy_free_list(&mm, &allocated, GPU_BUDDY_CLEARED);
> + gpu_buddy_reset_clear(&mm, false);
> + KUNIT_EXPECT_EQ(test, mm.clear_avail, 0);
> +
> + /*
> + * Using a non-power-of-two mm size, allocate alternating blocks of
> 4KiB in an
> + * odd sequence and free them as cleared. All blocks should be marked as
> + * cleared and the split blocks should be merged back to their original
> + * size when the blocks clear reset function is called.
> + */
> + i = 0;
> + do {
> + if (i % 2 != 0)
> + KUNIT_ASSERT_FALSE_MSG(test,
> gpu_buddy_alloc_blocks(&mm, 0, mm_size,
> + ps,
> ps, &allocated, 0),
> + "buddy_alloc hit an error size=%lu\n",
> ps);
> + } while (++i < n_pages);
> +
> + gpu_buddy_free_list(&mm, &allocated, GPU_BUDDY_CLEARED);
> + gpu_buddy_reset_clear(&mm, true);
> + KUNIT_EXPECT_EQ(test, mm.clear_avail, mm_size);
> + gpu_buddy_fini(&mm);
> }
>
> static void gpu_test_buddy_alloc_contiguous(struct kunit *test)
--
Jani Nikula, Intel
