[CUDA] Add Validation of batch_indices in RoiAlign

[tianleiwu]

Description

This PR implements a device-side bounds check for batch_indices in the RoiAlign CUDA operator. This is a follow-up to #27543, which fixed the same vulnerability in the CPU implementation.

Previously, CheckROIAlignValidInput() only validated batch_indices when they were accessible on the host (CPU). For the CUDA EP, batch_indices reside in GPU memory, so host-side validation would require an expensive GPU-to-CPU copy, which could also break CUDA graph capture.

This change:

1. Passes batch_size from the host to the CUDA kernel.
2. Adds a check within the RoIAlignForward kernel to ensure 0 <= batch_index < batch_size.
3. If an invalid batch_index is encountered, the kernel sets the output value for that specific RoI element to 0 and returns early for that thread.

Impact

• Vulnerability fixed: Heap out-of-bounds read on GPU.
• Performance: Negligible impact as it's a simple range check within the existing kernel.
• Compatibility: No changes to ONNX models or public APIs.

Validation

• Existing RoiAlignTest suite.
• Added two new test cases: BatchIndicesOutOfRange_CUDA and BatchIndicesNegative_CUDA to verify that the CUDA provider correctly handles out-of-range batch indices.
• Verified that the CUDA provider handles opset 10 without falling back to the CPU EP for these tests.

[github-actions]

You can commit the suggested changes from lintrunner.

[Copilot]

Pull request overview

This PR addresses a security issue in the CUDA implementation of the RoiAlign operator by adding a device-side bounds check for batch_indices to prevent out-of-bounds reads when batch_indices contains invalid values.

Changes:

• Pass batch_size into the CUDA RoiAlign kernel and validate batch_indices on-device.
• Update CUDA RoiAlign implementation signatures (RoiAlignImpl / RoIAlignForward) to carry the new parameter.
• Add CUDA-specific unit tests that verify invalid batch_indices produce zero outputs instead of triggering unsafe reads.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
onnxruntime/core/providers/cuda/object_detection/roialign.cc	Passes batch_size (from X shape) into the CUDA implementation call.
onnxruntime/core/providers/cuda/object_detection/roialign_impl.h	Extends RoiAlignImpl API to accept batch_size.
onnxruntime/core/providers/cuda/object_detection/roialign_impl.cu	Adds device-side batch_indices range check inside the CUDA kernel.
onnxruntime/core/providers/cpu/object_detection/roialign.cc	Moves/retains explanatory comments around host-only validation logic.
onnxruntime/test/providers/cpu/object_detection/roialign_test.cc	Adds CUDA-focused tests for out-of-range and negative batch_indices.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated no new comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.