Advisory

UBUNTU-CVE-2025-49847

Also known as: CVE-2025-49847

Summary

The record has no summary.

Details

llama.cpp is an inference of several LLM models in C/C++. Prior to version b5662, an attacker‐supplied GGUF model vocabulary can trigger a buffer overflow in llama.cpp’s vocabulary‐loading code. Specifically, the helper _try_copy in llama.cpp/src/vocab.cpp: llama_vocab::impl::token_to_piece() casts a very large size_t token length into an int32_t, causing the length check (if (length < (int32_t)size)) to be bypassed. As a result, memcpy is still called with that oversized size, letting a malicious model overwrite memory beyond the intended buffer. This can lead to arbitrary memory corruption and potential code execution. This issue has been patched in version b5662.

Severity

High Basis: CVSS v3 score 8.8, vector CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

Dates

Published
2025-06-17 20:15 UTC
Modified
2026-10-17 12:58 UTC

Affected packages and fixes

llama.cpp

Debian packages (deb) · Ubuntu 25.10

Affected

  • introduced 0 (ecosystem)

Fixed versions

None listed

Fix

No fixed version is listed.

llama.cpp

Debian packages (deb) · Ubuntu 26.10

Affected

  • introduced 0 (ecosystem)

Fixed versions

None listed

Fix

No fixed version is listed.

llama.cpp

Debian packages (deb) · Ubuntu 26.04

Affected

  • introduced 0 (ecosystem)

Fixed versions

None listed

Fix

No fixed version is listed.

Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.

Check your own computer: guard packages check

View this record on OSV.dev

Source

Source: Ubuntu Security Notices · CC BY-SA 4.0

Sources & licences