GHSA-hwpq-hmq9-wj77

Опубликовано: 07 июл. 2026

Источник: github

Github: Прошло ревью

CVSS3: 5.5

Описание

ONNX has Null Pointer Dereference in Upsample Version Converter Adapter (Zero Inputs)

Summary

Null pointer dereference (SIGSEGV) in Upsample_6_7::adapt_upsample_6_7() (onnx/version_converter/adapters/upsample_6_7.h:31) when convert_version() processes a model with an Upsample node that has zero inputs. The adapter accesses node->inputs()[0]->sizes() without checking input count. 107-byte PoC crashes on Release build.

This is the same class of bug as the Cast adapter advisory (separate report) but in a different adapter, different file, and different operator.

Details

The Upsample 6→7 adapter validates attributes but not inputs:

// upsample_6_7.h:20-33 void adapt_upsample_6_7(..., Node* node) const { ONNX_ASSERTM( node->hasAttribute(width_scale_symbol) && node->hasAttribute(height_scale_symbol), "...") // Attribute check PASSES auto width_scale = node->f(width_scale_symbol); auto height_scale = node->f(height_scale_symbol); auto input_shape = node->inputs()[0]->sizes(); // ^^^^^^^^^^^^^^^^^^^^ // OOB when inputs().size() == 0 → SIGSEGV }

The PoC has an Upsample node at opset 6 with the required width_scale and height_scale attributes but zero inputs. The attribute assertions pass, then node->inputs()[0] on an empty ArrayRef:

Release builds (NDEBUG): bounds-check assertion compiled out → reads garbage pointer → SIGSEGV
Debug builds: assert(Index < Length) at array_ref.h:159 → SIGABRT

An Upsample node with zero inputs passes graphProtoToGraph() because the import code only resolves input names present in the protobuf.

PoC

import base64 import onnx from onnx import version_converter poc_b64 = "CAI6YQo8EgFZIghVcHNhbXBsZSoVCgt3aWR0aF9zY2FsZRUAAABAoAEBKhYKDGhlaWdodF9zY2FsZRUAAABAoAEBEgR0ZXN0YhsKAVkSFgoUCAESEAoCCAEKAggBCgIIBAoCCARCBAoAEAY=" model = onnx.load_from_string(base64.b64decode(poc_b64)) # CRASHES — Upsample_6_7 adapter dereferences empty inputs array version_converter.convert_version(model, 7) # SIGSEGV

107-byte PoC. Confirmed SIGSEGV on both onnx 1.21.0 (pip) and 1.22.0 (source build).

Impact

Any application that uses onnx.version_converter.convert_version() on untrusted models is vulnerable. This includes model conversion pipelines and tools that auto-upgrade opset versions for compatibility. The crash is unrecoverable (SIGSEGV).

This vulnerability is part of a systemic pattern across multiple version converter adapters. A full audit of all ~45 adapters was performed as part of the fix; eight adapters were found with the same class of unguarded indexed access (cast_9_8, softmax_12_13, softmax_13_12, upsample_6_7, upsample_9_10, group_normalization_20_21, broadcast_forward_compatibility, upsample_9_8) and all have been fixed in PR #7813.

Ссылки

Пакеты

Наименование

onnx

pip

Затронутые версииВерсия исправления

>= 1.9.0, < 1.22.0

1.22.0

EPSS

Процентиль: 8%

0.00183

Низкий

5.5 Medium

CVSS3

CVE ID

CVE-2026-44512

Дефекты

CWE-476

Связанные уязвимости

CVE-2026-44512

CVSS3: 5.5

ubuntu

около 1 месяца назад

Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. From 1.9.0 before 1.22.0, onnx.version_converter.convert_version() can dereference a null pointer in Upsample_6_7::adapt_upsample_6_7() in onnx/version_converter/adapters/upsample_6_7.h when processing an untrusted model with an Upsample node that has zero inputs, causing an unrecoverable denial of service. This issue is fixed in version 1.22.0.