OSV-2024-1053 Heap-buffer-overflow in gguf_get_tensor_offset

Not scored Published

A heap-buffer-overflow in the GGUF tensor offset calculation allows attackers to cause crashes or potentially execute code via malicious model files.

Exploited in the wild
Not known
Login needed
Not needed
Attackable
Not stated
Victim action needed
Not needed
Fix
None listed

What is OSV-2024-1053?

OSV-2024-1053 is a memory corruption vulnerability in llamacpp, published on 13 Sep 2024.

Official description

OSS-Fuzz report: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=42538148

Crash type: Heap-buffer-overflow READ 8
Crash state:
gguf_get_tensor_offset
llama_model_loader::llama_tensor_weight::llama_tensor_weight
void std::__1::vector<llama_model_loader::llama_tensor_weight, std::__1::allocat

Which versions are affected by OSV-2024-1053?

1 open source package ship an affected version, including llamacpp.

PackageAffectedFixed in
llamacpp OSS-FuzzAll versionsNone

How do I fix OSV-2024-1053?

No fix has been published for OSV-2024-1053 yet. Limit exposure of the affected software until one is available.

Upgrade to the fixed version. Memory-safe languages, compiler hardening and fuzzing prevent most new cases. More about Memory corruption.

Is OSV-2024-1053 being exploited?

There is no known exploitation of OSV-2024-1053 in the wild and no public exploit code.

Sources

Is OSV-2024-1053 in your software?

Upload an SBOM and sbomify checks it against every vulnerability here, every day.

Check your SBOMs for free