Skip to content
Open
Show file tree
Hide file tree
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions onnxruntime/contrib_ops/cpu/bert/rotary_embedding_helper.h
Original file line number Diff line number Diff line change
Expand Up @@ -118,6 +118,16 @@ Status CheckInputs(const T* input,
if (head_size == 0) {
ORT_RETURN_IF_ERROR(detail::CheckedMulToInt32(cache_width, 2, "head_size", head_size));
}
// Validate that the rotary embedding dimension derived from cos_cache does not exceed hidden_size,
// which would cause an out-of-bounds read on the input tensor.
int effective_rotary_dim = cache_width * 2;
if (hidden_size > 0 && effective_rotary_dim > hidden_size) {
return ORT_MAKE_STATUS(ONNXRUNTIME, INVALID_ARGUMENT,
"RotaryEmbedding: cos_cache dimension (", cache_width,
" * 2 = ", effective_rotary_dim,
") exceeds input hidden_size (", hidden_size,
") when rotary_embedding_dim is 0");
}
} else {
if (!transposed) {
if (num_heads <= 0) {
Expand Down
28 changes: 28 additions & 0 deletions onnxruntime/test/contrib_ops/rotary_embedding_op_test.cc
Original file line number Diff line number Diff line change
Expand Up @@ -1172,5 +1172,33 @@ TEST(RotaryEmbeddingTest, ContribRotaryEmbedding_PositionIds_Negative_WebGPU_Pas
test.Run(OpTester::ExpectResult::kExpectSuccess, "", {}, nullptr, &execution_providers);
}

// Test that cos_cache dimension exceeding hidden_size is rejected when rotary_embedding_dim=0.
TEST(RotaryEmbeddingTest, ContribRotaryEmbedding_RejectsCosCacheExceedsHiddenSize) {
Comment thread
apsonawane marked this conversation as resolved.
Outdated
int batch_size = 1;
int sequence_length = 1;
int hidden_size = 64;
int half_rotary_dim = 64; // makes cos_cache_dims[1]*2 = 128 > hidden_size
int max_sequence_length = 2;

OpTester test("RotaryEmbedding", 1, onnxruntime::kMSDomain);
test.AddAttribute<int64_t>("interleaved", static_cast<int64_t>(0));
test.AddAttribute<int64_t>("num_heads", static_cast<int64_t>(1));
Comment thread
apsonawane marked this conversation as resolved.

test.AddInput<float>("input", {batch_size, sequence_length, hidden_size},
std::vector<float>(hidden_size, 42.0f));
test.AddInput<int64_t>("position_ids", {1}, {0});
test.AddInput<float>("cos_cache", {max_sequence_length, half_rotary_dim},
std::vector<float>(max_sequence_length * half_rotary_dim, 0.0f));
test.AddInput<float>("sin_cache", {max_sequence_length, half_rotary_dim},
std::vector<float>(max_sequence_length * half_rotary_dim, 1.0f));
test.AddOutput<float>("output", {batch_size, sequence_length, hidden_size},
std::vector<float>(hidden_size, 0.0f));

std::vector<std::unique_ptr<IExecutionProvider>> execution_providers;
execution_providers.push_back(DefaultCpuExecutionProvider());
test.Run(OpTester::ExpectResult::kExpectFailure,
"cos_cache dimension", {}, nullptr, &execution_providers);
}

} // namespace test
} // namespace onnxruntime
29 changes: 29 additions & 0 deletions onnxruntime/test/providers/cpu/llm/rotary_embedding_op_test.cc
Original file line number Diff line number Diff line change
Expand Up @@ -1412,5 +1412,34 @@ TEST(RotaryEmbeddingTest, RotaryEmbedding_PositionIds_OOB_InBatch_WebGPU_Passthr
test.Run(OpTester::ExpectResult::kExpectSuccess, "", {}, nullptr, &execution_providers);
}

// Test that cos_cache dimension exceeding hidden_size is rejected when rotary_embedding_dim=0.
TEST(RotaryEmbeddingTest, RotaryEmbedding_RejectsCosCacheExceedsHiddenSize) {

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nit: the name RejectsCosCacheExceedsHiddenSize overclaims here (same issue that was just fixed for the contrib test). This runs against the mainline ONNX-domain op, whose helper (core/providers/cpu/llm/rotary_embedding_helper.h) is not modified by this PR and has no hidden_size/OOB guard. With num_heads=1, head_size == hidden_size == 64, so the rejection actually comes from the pre-existing generic width check (head_size/2 = 32 != cos_cache_dims[1] = 64) — which is exactly the message this test asserts. So the test documents pre-existing dimension validation, not an OOB fix. Consider renaming (e.g. RejectsCosCacheWidthMismatch) and/or a one-line comment noting the mainline op needs no fix because it requires num_heads>0 for rank-3 and validates the exact cos_cache width.

// hidden_size = 64, cos_cache dim1 = 64 => effective rotary dim = 128 > 64
int batch_size = 1;
int sequence_length = 1;
int hidden_size = 64;
int half_rotary_dim = 64; // makes cos_cache_dims[1]*2 = 128 > hidden_size
int max_sequence_length = 2;

OpTester test("RotaryEmbedding", 23, onnxruntime::kOnnxDomain);
test.AddAttribute<int64_t>("interleaved", static_cast<int64_t>(0));
test.AddAttribute<int64_t>("num_heads", static_cast<int64_t>(1));

test.AddInput<float>("input", {batch_size, sequence_length, hidden_size},
std::vector<float>(hidden_size, 42.0f));
test.AddInput<float>("cos_cache", {max_sequence_length, half_rotary_dim},
std::vector<float>(max_sequence_length * half_rotary_dim, 0.0f));
test.AddInput<float>("sin_cache", {max_sequence_length, half_rotary_dim},
std::vector<float>(max_sequence_length * half_rotary_dim, 1.0f));
test.AddInput<int64_t>("position_ids", {1}, {0});
test.AddOutput<float>("output", {batch_size, sequence_length, hidden_size},
std::vector<float>(hidden_size, 0.0f));

std::vector<std::unique_ptr<IExecutionProvider>> execution_providers;
execution_providers.push_back(DefaultCpuExecutionProvider());
test.Run(OpTester::ExpectResult::kExpectFailure,
"cos_cache dimension", {}, nullptr, &execution_providers);
}

} // namespace test
} // namespace onnxruntime
Loading