CVE-2026-44512Medium· 5.5▾ SunlitONNX has Null Pointer Dereference in Upsample Version Converter Adapter (Zero Inputs)
▾ Sunlit zone — Low / medium · no exploitation signal
impact 30.3 · likelihood 0 · exploitation 0
Need a working PoC? Pro members can cast a request and our team develops one — it lands right here.
Exploit-prediction probability, daily snapshots since Jul 9.
Disclosure to exploitation, from the record and what we observed since indexing it.
Disclosed via OSV
0.2%
0.2% → 0.2%
Last analysed / modified upstream
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.
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:
NDEBUG): bounds-check assertion compiled out → reads garbage pointer → SIGSEGVassert(Index < Length) at array_ref.h:159 → SIGABRTAn Upsample node with zero inputs passes graphProtoToGraph() because the import code only resolves input names present in the protobuf.
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).
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.
onnx >= 1.9.0, < 1.22.0Upgrade to a patched release:
onnx 1.22.0Connected by shared product, vendor, weakness, or advisory.
CVE-2024-27318High· 7.5Onnx Directory Traversal vulnerability
CVE-2026-34447Medium· 5.5ONNX: External Data Symlink Traversal
CVE-2026-28500High· 8.6ONNX Untrusted Model Repository Warnings Suppressed by silent=True in onnx.hub.load() — Silent Supply-Chain Attack
CVE-2024-27319Medium· 4.4Onnx Out-of-bounds Read vulnerability
CVE-2026-49114High· 7.1In ONNX before 1.21.0, the 'save_external_data' function builds the external-data file path from the model's external_data location field and opens it for writing without 'O_NOFOLLOW/O_EXCL', after a non-atomic 'os.path.isfile()' check
CVE-2026-34446Medium· 4.7ONNX: Arbitrary File Read via ExternalData Hardlink Bypass in ONNX load