Re: [PATCH V4 2/2] gpu/tests/gpu_buddy: Add KUnit test for gpu_buddy_allocated_addr_to_block

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Memory leak on KUnit assertion failure due to bypassed cleanup.
--

--- Patch [2]: [PATCH V4 2/2] gpu/tests/gpu_buddy: Add KUnit test for gpu_buddy_allocated_addr_to_block ---
commit 429e209c9177a128187845522d6b4613f6fbe30b
Author: Tejas Upadhyay <tejas.upadhyay@intel.com>

gpu/tests/gpu_buddy: Add KUnit test for gpu_buddy_allocated_addr_to_block

This patch adds a new KUnit test gpu_test_buddy_addr_to_block() to validate
the gpu_buddy_allocated_addr_to_block() helper function. The test verifies
exact address matching, unallocated addresses within the manager, and addresses
outside the manager bounds.

> diff --git a/drivers/gpu/tests/gpu_buddy_test.c b/drivers/gpu/tests/gpu_buddy_test.c
> index 89698563c61b2..5860bfc84a25a 100644
> --- a/drivers/gpu/tests/gpu_buddy_test.c
> +++ b/drivers/gpu/tests/gpu_buddy_test.c
> @@ -1422,6 +1422,50 @@ static void gpu_test_buddy_alloc_exceeds_max_order(struct kunit *test)
>  	gpu_buddy_fini(&mm);
>  }
>  
> +static void gpu_test_buddy_addr_to_block(struct kunit *test)
> +{
> +	struct gpu_buddy_block *allocated_block, *found_block;
> +	LIST_HEAD(allocated_list);
> +	const u64 test_size = SZ_4M + SZ_2M;
> +	const u64 alloc_start = SZ_4M;
> +	const u64 alloc_size = SZ_4K;
> +	const u64 chunk_size = SZ_4K;
> +	struct gpu_buddy mm;
> +
> +	KUNIT_ASSERT_FALSE_MSG(test, gpu_buddy_init(&mm, test_size, chunk_size),
> +			       "buddy_init failed\n");
> +
> +	KUNIT_ASSERT_FALSE_MSG(test, gpu_buddy_alloc_blocks(&mm, alloc_start,
> +							    alloc_start + alloc_size,
> +							    alloc_size, chunk_size,
> +							    &allocated_list, 0),
> +			       "buddy_alloc failed\n");

[Severity: Medium]
Will this leak memory if the allocation fails? KUNIT_ASSERT_FALSE_MSG() aborts
the test execution immediately on failure, bypassing the gpu_buddy_fini(&mm)
cleanup call at the end of the function. This would leak the memory allocated
by gpu_buddy_init().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260806053624.3215216-4-tejas.upadhyay@intel.com?part=2