همهٔ گزارش‌ها
متوسط

ONNX: Out-of-Bounds Read in RNN/GRU/LSTM Shape Inference (Old Opsets, Missing Rank Check)

متن اصلی

نسخه‌های متأثر

PyPIonnx
نسخه‌های متأثر
<= 1.21.0
نسخهٔ اصلاح‌شده
1.22.0
نسخه‌های متأثر
نامشخص
نسخهٔ اصلاح‌شده
اعلام نشده
نسخه‌های متأثر
نامشخص
نسخهٔ اصلاح‌شده
اعلام نشده
نسخه‌های متأثر
نامشخص
نسخهٔ اصلاح‌شده
اعلام نشده
نسخه‌های متأثر
نامشخص
نسخهٔ اصلاح‌شده
اعلام نشده
نسخه‌های متأثر
نامشخص
نسخهٔ اصلاح‌شده
اعلام نشده
نسخه‌های متأثر
نامشخص
نسخهٔ اصلاح‌شده
اعلام نشده

بهره‌برداری

CVE-2026-115919

احتمال بهره‌برداری در ۳۰ روز آینده: 0.18%

FIRST EPSS · ۱۴۰۵/۰۸/۲۷

شرح آسیب‌پذیری

Summary

ONNX shape inference performed an out-of-bounds read when processing a malformed RNN, GRU, or LSTM node from opsets 1–6 whose first input had a rank other than three.

RNNShapeInference_opset1_to_6() checked whether the input shape was present but accessed dim(0) and dim(1) without first checking its rank. A crafted model could therefore terminate a process that called onnx.shape_inference.infer_shapes(). The model passed onnx.checker.check_model() before reaching the vulnerable path.

Technical details

The vulnerable code read dimensions unconditionally:

if (hasInputShape(ctx, 0)) {
  const auto& first_input_shape = getInputShape(ctx, 0);
  seq_length = first_input_shape.dim(0);
  batch_size = first_input_shape.dim(1);
}

For an undersized shape, protobuf's repeated-field access read outside the intended array in release builds. The observed result was SIGSEGV; sanitizer and debug builds terminated on a protobuf bounds check.

The newer opset 7–13 and opset 14+ inference paths already required the first input to have rank three and were not affected.

Impact

An application that performs shape inference on an untrusted model can be terminated, causing denial of service. Exploitation does not require privileges or human interaction once the model is processed. No confidentiality or integrity impact was demonstrated.

Affected versions

  • Affected: ONNX through 1.21.0

  • Patched: ONNX 1.22.0

The vulnerable schemas are limited to RNN and LSTM opsets 1–6 and GRU opsets 3–6.

Remediation

Upgrade to ONNX 1.22.0 or later. The fix validates that the first input has rank three before accessing its dimensions.

The original malformed model was also re-tested against the published ONNX 1.23.2 wheel and correctly raises InferenceError.

Severity

The vulnerable ONNX component processes the malicious model through a local library API or file boundary and is not itself bound to a network stack. Applications that expose ONNX model processing over a network should calculate an Environmental score with Modified Attack Vector set to Network.

پیوندها

گزارش‌دهندگان

  • MinicoAI