Handle int overflow in rnn

[tianleiwu]

Description

Fixes two overflow/underflow bugs in the CPU RNN kernel (rnn.cc):

• SafeInt for GEMM M-dimension: seq_length * batch_size was computed as a raw int64_t multiply before narrow<int>(), meaning an overflow would be UB before the check could fire. Replaced with SafeInt<int64_t>(seq_length) * batch_size for a checked multiply.

• seq_length == 0 guard in Assign_Y_h: For the forward direction, last_time_step = seq_length - 1 underflows to -1 when seq_length == 0, producing a negative y_offset and out-of-bounds read. Added an early-exit that zero-fills Y_h for the direction and returns. Also handles sequence_lens[batch] == 0 (same underflow path), zeroing the affected batch slot and skipping via continue.

Motivation and Context

Silent UB from integer overflow/underflow in shape-derived index arithmetic can corrupt memory or produce incorrect results without any diagnostic signal. These cases are legal per the ONNX spec (empty sequences, per-batch zero-length sequences) and must be handled explicitly.

[Copilot]

Pull request overview

This PR updates the CPU RNN implementation to use onnxruntime::narrow<int>() instead of unchecked static_cast<int>() when passing shape-derived values into APIs that take int, aiming to prevent silent truncation/overflow.

Changes:

• Add an explicit include for core/common/narrow.h in rnn.cc.
• Replace several static_cast<int>(...) conversions with onnxruntime::narrow<int>(...) in CopyVector and Gemm calls.

---

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

[tianleiwu]

@copilot apply changes based on the comments in this thread

[Copilot]

@copilot apply changes based on the comments in this thread

Applied both changes from the review thread in commit 88cff57:

1. SafeInt for multiplication overflow: Changed onnxruntime::narrow<int>(seq_length * batch_size) to onnxruntime::narrow<int>(SafeInt<int64_t>(seq_length) * batch_size) so the multiplication itself is overflow-checked before narrowing.

2. seq_length == 0 handling in Assign_Y_h: Added an early-exit guard at the top of Assign_Y_h — if seq_length == 0, Y_h for the current direction is zeroed out and the function returns immediately. Also added handling for the case where sequence_lens[batch] == 0, which would also produce last_time_step = -1 — those batches are now zero-filled and skipped via continue.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

---

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

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

---

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