HostInteropTask/interop-task-cuda-buffer-migrate.cpp fails flakily on unrelated changes

[againull]

Describe the bug

HostInteropTask/interop-task-cuda-buffer-migrate.cpp

2025-02-14T18:41:57.9751397Z [0/1] Running SYCL End-to-End tests
2025-02-14T18:41:57.9752363Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:64: note: run-only test mode enabled, only executing tests
2025-02-14T18:41:57.9753611Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:519: note: Targeted devices: cuda:gpu
2025-02-14T18:41:57.9754676Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:699: note: Found pre-installed AOT device compiler ocloc
2025-02-14T18:41:57.9755410Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:699: note: Found pre-installed AOT device compiler opencl-aot
2025-02-14T18:41:57.9766229Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:802: note: Aspects for cuda:gpu: ext_oneapi_mipmap_level_reference, usm_host_allocations, ext_oneapi_virtual_mem, ext_oneapi_bindless_sampled_image_fetch_3d, atomic64, online_compiler, usm_shared_allocations, ext_oneapi_bindless_images_shared_usm, online_linker, ext_oneapi_native_assert, ext_intel_device_info_uuid, ext_oneapi_external_semaphore_import, ext_oneapi_bindless_images_sample_2d_usm, ext_intel_free_memory, ext_intel_legacy_image, ext_oneapi_mipmap_anisotropy, ext_oneapi_cubemap, ext_oneapi_opportunistic_group, ext_intel_device_id, ext_oneapi_external_memory_import, usm_atomic_shared_allocations, ext_oneapi_ballot_group, queue_profiling, ext_oneapi_cuda_async_barrier, ext_oneapi_bindless_images, fp64, ext_oneapi_cubemap_seamless_filtering, ext_oneapi_image_array, usm_device_allocations, fp16, ext_oneapi_bindless_images_2d_usm, ext_oneapi_bindless_sampled_image_fetch_2d, ext_intel_pci_address, ext_oneapi_graph, ext_oneapi_mipmap, ext_oneapi_queue_profiling_tag, ext_oneapi_limited_graph, ext_oneapi_bindless_sampled_image_fetch_1d_usm, gpu, ext_oneapi_bindless_images_1d_usm, ext_oneapi_bindless_images_sample_1d_usm, ext_oneapi_unique_addressing_per_dim, ext_intel_memory_clock_rate, ext_intel_memory_bus_width, ext_oneapi_fixed_size_group, ext_oneapi_bindless_sampled_image_fetch_2d_usm
2025-02-14T18:41:57.9775121Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:814: note: SG sizes for cuda:gpu: 32
2025-02-14T18:41:57.9776503Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:823: note: Architectures for cuda:gpu: nvidia_gpu_sm_86
2025-02-14T18:41:57.9780685Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:901: note: Global features: any-device-is-cuda any-device-is-gpu any-target-is-nvidia build-and-run-mode cm-compiler cuda_dev_kit has_ndebug host=None level_zero_dev_kit linux llvm-link llvm-spirv non-root-user ocloc opencl-aot opencl_icd preview-breaking-changes-supported run-mode shell system-linux target=None test-mode-run-only xptifw zstd
2025-02-14T18:41:57.9784460Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:904: note: Per-device features:
2025-02-14T18:41:57.9793438Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:906: note: 	cuda:gpu: any-device-is-cuda any-device-is-gpu any-target-is-nvidia arch-nvidia_gpu_sm_86 aspect-atomic64 aspect-ext_intel_device_id aspect-ext_intel_device_info_uuid aspect-ext_intel_free_memory aspect-ext_intel_legacy_image aspect-ext_intel_memory_bus_width aspect-ext_intel_memory_clock_rate aspect-ext_intel_pci_address aspect-ext_oneapi_ballot_group aspect-ext_oneapi_bindless_images aspect-ext_oneapi_bindless_images_1d_usm aspect-ext_oneapi_bindless_images_2d_usm aspect-ext_oneapi_bindless_images_sample_1d_usm aspect-ext_oneapi_bindless_images_sample_2d_usm aspect-ext_oneapi_bindless_images_shared_usm aspect-ext_oneapi_bindless_sampled_image_fetch_1d_usm aspect-ext_oneapi_bindless_sampled_image_fetch_2d aspect-ext_oneapi_bindless_sampled_image_fetch_2d_usm aspect-ext_oneapi_bindless_sampled_image_fetch_3d aspect-ext_oneapi_cubemap aspect-ext_oneapi_cubemap_seamless_filtering aspect-ext_oneapi_cuda_async_barrier aspect-ext_oneapi_external_memory_import aspect-ext_oneapi_external_semaphore_import aspect-ext_oneapi_fixed_size_group aspect-ext_oneapi_graph aspect-ext_oneapi_image_array aspect-ext_oneapi_limited_graph aspect-ext_oneapi_mipmap aspect-ext_oneapi_mipmap_anisotropy aspect-ext_oneapi_mipmap_level_reference aspect-ext_oneapi_native_assert aspect-ext_oneapi_opportunistic_group aspect-ext_oneapi_queue_profiling_tag aspect-ext_oneapi_unique_addressing_per_dim aspect-ext_oneapi_virtual_mem aspect-fp16 aspect-fp64 aspect-gpu aspect-online_compiler aspect-online_linker aspect-queue_profiling aspect-usm_atomic_shared_allocations aspect-usm_device_allocations aspect-usm_host_allocations aspect-usm_shared_allocations build-and-run-mode cm-compiler cuda cuda_dev_kit gpu has_ndebug host=None level_zero_dev_kit linux llvm-link llvm-spirv non-root-user ocloc opencl-aot opencl_icd preview-breaking-changes-supported run-mode sg-32 shell system-linux target-nvidia target=None test-mode-run-only xptifw zstd
2025-02-14T18:41:57.9801077Z -- Testing: 2282 tests, 12 workers --
2025-02-14T18:41:57.9801510Z FAIL: SYCL :: HostInteropTask/interop-task-cuda-buffer-migrate.cpp (1421 of 2282)
2025-02-14T18:41:57.9802128Z ******************** TEST 'SYCL :: HostInteropTask/interop-task-cuda-buffer-migrate.cpp' FAILED ********************
2025-02-14T18:41:57.9802580Z Exit Code: 1
2025-02-14T18:41:57.9802687Z 
2025-02-14T18:41:57.9802771Z Command Output (stdout):
2025-02-14T18:41:57.9802971Z --
2025-02-14T18:41:57.9803125Z 1 devices found
2025-02-14T18:41:57.9803328Z Using on device 0: NVIDIA GeForce RTX 3090
2025-02-14T18:41:57.9803598Z Checking result on host: FAILED
2025-02-14T18:41:57.9803816Z 0 ?= 1
2025-02-14T18:41:57.9803908Z 
2025-02-14T18:41:57.9803971Z --
2025-02-14T18:41:57.9804133Z Command Output (stderr):
2025-02-14T18:41:57.9804326Z --
2025-02-14T18:41:57.9805088Z RUN: at line 4: env UR_CUDA_ENABLE_IMAGE_SUPPORT=1 ONEAPI_DEVICE_SELECTOR=cuda:gpu  /__w/llvm/llvm/build-e2e/HostInteropTask/Output/interop-task-cuda-buffer-migrate.cpp.tmp.out
2025-02-14T18:41:57.9806344Z + env UR_CUDA_ENABLE_IMAGE_SUPPORT=1 ONEAPI_DEVICE_SELECTOR=cuda:gpu /__w/llvm/llvm/build-e2e/HostInteropTask/Output/interop-task-cuda-buffer-migrate.cpp.tmp.out
2025-02-14T18:41:57.9806959Z 
2025-02-14T18:41:57.9807021Z --
2025-02-14T18:41:57.9807109Z 
2025-02-14T18:41:57.9807174Z ********************
2025-02-14T18:41:57.9807358Z Slowest Tests: