Due to changes in allocator, the size of the allocation for contiguous region is not rounded up to a power-of-two and instead allocated as is. Thus, change the part of test that expected the allocation to fail.
Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/9311 Signed-off-by: Mikolaj Wasiak <mikolaj.was...@intel.com> --- .../gpu/drm/i915/selftests/intel_memory_region.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/i915/selftests/intel_memory_region.c b/drivers/gpu/drm/i915/selftests/intel_memory_region.c index f08f6674911e..ce848ae878c8 100644 --- a/drivers/gpu/drm/i915/selftests/intel_memory_region.c +++ b/drivers/gpu/drm/i915/selftests/intel_memory_region.c @@ -413,22 +413,17 @@ static int igt_mock_splintered_region(void *arg) close_objects(mem, &objects); - /* - * While we should be able allocate everything without any flag - * restrictions, if we consider I915_BO_ALLOC_CONTIGUOUS then we are - * actually limited to the largest power-of-two for the region size i.e - * max_order, due to the inner workings of the buddy allocator. So make - * sure that does indeed hold true. - */ - obj = igt_object_create(mem, &objects, size, I915_BO_ALLOC_CONTIGUOUS); - if (!IS_ERR(obj)) { - pr_err("%s too large contiguous allocation was not rejected\n", + if (!IS_ERR(obj) && !is_contiguous(obj)) { + pr_err("%s large allocation was not contiguous\n", __func__); err = -EINVAL; goto out_close; } + if (!IS_ERR(obj)) + close_objects(mem, &objects); + obj = igt_object_create(mem, &objects, rounddown_pow_of_two(size), I915_BO_ALLOC_CONTIGUOUS); if (IS_ERR(obj)) { -- 2.43.0
