rpm package
opensuse/llamacpp&distro=openSUSE Tumbleweed
pkg:rpm/opensuse/llamacpp&distro=openSUSE%20Tumbleweed
Vulnerabilities (9)
| CVE | Sev | CVSS | KEV | Affected versions | Fixed in | Published | Description |
|---|---|---|---|---|---|---|---|
| CVE-2026-119052 | Hig | 7.5 | < 0.8.0-1.1 | 0.8.0-1.1 | Oct 29, 2026 | llama.cpp b2877 before b11935 contains an uncaught exception vulnerability in the ggml-rpc server that allows unauthenticated remote attackers to crash the rpc-server by sending oversized message lengths. After the HELLO handshake, attackers can send a command such as SET_TENSOR | |
| CVE-2026-118375 | Hig | 8.1 | < 0.8.0-1.1 | 0.8.0-1.1 | Oct 28, 2026 | llama.cpp b8841 before b11865 contains an out-of-bounds write vulnerability in the ggml-rpc RDMA transport rdma_recv() function that allows unauthenticated peers to overflow the rpc-server stack. Attackers advertising RDMA capabilities in the HELLO message can send up to 256 KiB | |
| CVE-2026-116840 | Med | 6.1 | < 0.9.0-1.1 | 0.9.0-1.1 | Oct 27, 2026 | llama.cpp b5854 before b11275 contains an integer overflow vulnerability in gguf_init_from_reader that bypasses the CVE-2025-53630 size check when padding near-SIZE_MAX tensor sizes. Attackers can supply a crafted GGUF file whose tensor size pads to zero, causing an undersized da | |
| CVE-2026-116829 | Med | 6.6 | < 0.9.0-1.1 | 0.9.0-1.1 | Oct 27, 2026 | llama.cpp b10398 before b11301 contains an integer overflow vulnerability in gguf_init_from_reader that allows attackers to bypass tensor element-count validation by supplying shapes whose product wraps to zero. Attackers can craft GGUF files with enormous tensor extents that ove | |
| CVE-2026-52132 | Hig | 7.5 | < 0.4.0-1.1 | 0.4.0-1.1 | Sep 1, 2026 | llama.cpp through commit 97f06e9, when started with the --reranking flag, allows remote attackers to cause a denial of service (std::bad_alloc and HTTP 500) via a negative top_n value in a POST request to /rerank. | |
| CVE-2026-33298 | Hig | 7.8 | < 0.9.0-1.1 | 0.9.0-1.1 | Mar 24, 2026 | llama.cpp is an inference of several LLM models in C/C++. Prior to b7824, an integer overflow vulnerability in the `ggml_nbytes` function allows an attacker to bypass memory validation by crafting a GGUF file with specific tensor dimensions. This causes `ggml_nbytes` to return a | |
| CVE-2026-21869 | Hig | 8.8 | < 9500-1.1 | 9500-1.1 | Jan 8, 2026 | llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the n_discard parameter is parsed directly from JSON input in the llama.cpp server's completion endpoints without validation to ensure it's non-negative. When a negative value is supplied an | |
| CVE-2025-53630 | Hig | — | < 5889-1.1 | 5889-1.1 | Jul 10, 2025 | llama.cpp is an inference of several LLM models in C/C++. Integer Overflow in the gguf_init_from_file_impl function in ggml/src/gguf.cpp can lead to Heap Out-of-Bounds Read/Write. This vulnerability is fixed in commit 26a48ad699d50b6268900062661bd22f3e792579. | |
| CVE-2025-49847 | Hig | 8.8 | < 5699-1.1 | 5699-1.1 | Jun 17, 2025 | 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::tok |
- affected < 0.8.0-1.1fixed 0.8.0-1.1
llama.cpp b2877 before b11935 contains an uncaught exception vulnerability in the ggml-rpc server that allows unauthenticated remote attackers to crash the rpc-server by sending oversized message lengths. After the HELLO handshake, attackers can send a command such as SET_TENSOR
- affected < 0.8.0-1.1fixed 0.8.0-1.1
llama.cpp b8841 before b11865 contains an out-of-bounds write vulnerability in the ggml-rpc RDMA transport rdma_recv() function that allows unauthenticated peers to overflow the rpc-server stack. Attackers advertising RDMA capabilities in the HELLO message can send up to 256 KiB
- affected < 0.9.0-1.1fixed 0.9.0-1.1
llama.cpp b5854 before b11275 contains an integer overflow vulnerability in gguf_init_from_reader that bypasses the CVE-2025-53630 size check when padding near-SIZE_MAX tensor sizes. Attackers can supply a crafted GGUF file whose tensor size pads to zero, causing an undersized da
- affected < 0.9.0-1.1fixed 0.9.0-1.1
llama.cpp b10398 before b11301 contains an integer overflow vulnerability in gguf_init_from_reader that allows attackers to bypass tensor element-count validation by supplying shapes whose product wraps to zero. Attackers can craft GGUF files with enormous tensor extents that ove
- affected < 0.4.0-1.1fixed 0.4.0-1.1
llama.cpp through commit 97f06e9, when started with the --reranking flag, allows remote attackers to cause a denial of service (std::bad_alloc and HTTP 500) via a negative top_n value in a POST request to /rerank.
- affected < 0.9.0-1.1fixed 0.9.0-1.1
llama.cpp is an inference of several LLM models in C/C++. Prior to b7824, an integer overflow vulnerability in the `ggml_nbytes` function allows an attacker to bypass memory validation by crafting a GGUF file with specific tensor dimensions. This causes `ggml_nbytes` to return a
- affected < 9500-1.1fixed 9500-1.1
llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the n_discard parameter is parsed directly from JSON input in the llama.cpp server's completion endpoints without validation to ensure it's non-negative. When a negative value is supplied an
- affected < 5889-1.1fixed 5889-1.1
llama.cpp is an inference of several LLM models in C/C++. Integer Overflow in the gguf_init_from_file_impl function in ggml/src/gguf.cpp can lead to Heap Out-of-Bounds Read/Write. This vulnerability is fixed in commit 26a48ad699d50b6268900062661bd22f3e792579.
- affected < 5699-1.1fixed 5699-1.1
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::tok