CVE-2024-42479: llama.cpp allows write-what-where in rpc_server::set_tensor
First published: Mon Aug 12 2024(Updated: )
llama.cpp provides LLM inference in C/C++. The unsafe `data` pointer member in the `rpc_tensor` structure can cause arbitrary address writing. This vulnerability is fixed in b3561.
Credit: security-advisories@github.com
| Affected Software | Affected Version | How to fix |
|---|---|---|
| Llama.cpp | <b3561 |
Never miss a vulnerability like this again
Sign up to SecAlerts for real-time vulnerability data matched to your software, aggregated from hundreds of sources.
Reference Links
Frequently Asked Questions
What is the severity of CVE-2024-42479?
CVE-2024-42479 has a high severity level due to its potential for arbitrary address writing.
How do I fix CVE-2024-42479?
To fix CVE-2024-42479, upgrade llama.cpp to version b3561 or later.
What type of vulnerability is CVE-2024-42479?
CVE-2024-42479 is a memory corruption vulnerability caused by unsafe pointer handling.
What software is affected by CVE-2024-42479?
CVE-2024-42479 affects versions of Ggerganov Llama.cpp prior to b3561.
Can CVE-2024-42479 lead to remote code execution?
Yes, the arbitrary address writing characteristic of CVE-2024-42479 may lead to remote code execution.
- agent/title
- agent/references
- agent/description
- agent/first-publish-date
- agent/type
- agent/author
- agent/event
- agent/severity
- agent/last-modified-date
- agent/weakness
- agent/remedy
- agent/softwarecombine
- collector/mitre-cve
- source/MITRE
- agent/source
- agent/tags
- collector/nvd-api
- source/NVD
- agent/software-canonical-lookup
- agent/software-canonical-lookup-request
- vendor/ggerganov
- canonical/llama.cpp