by Kastro
Paper/Archive/CVE-2024-42479

CVE-2024-42479

ggerganov llama.cpp — llama.cpp allows write-what-where in rpc_server::set_tensor

EPSS 3%public exploitCVSS 10.0
critical.is priorityof 100
57
CVSS 10.0 (CNA)10.0 / 10 × 3030.0Not in CISA KEV+300.0EPSS percentile 85.4.854 × 2017.1Exploit: SSVC poc+1010.0SSVC exploitation: poc+100.0
Description · from the CVE record

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.

Affected

ProductAffectedFixed inSource
llama.cpp< b3561–CNA
llama.cpp< b3561b3561CISA-ADP

Scores by source

GitHub_M (CNA)10.0CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H
NVD9.8CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
CISA ADPSSVCExploitation: poc · Automatable: no · Technical impact: total

References

refgithub.com/ggerganov/llama.cpp/security/advisories/GHSA-wcr5-566p-9cwjrefgithub.com/ggerganov/llama.cpp/commit/b72942fac998672a79a1ae3c03b340f7e629980b