Vulnérabilités de llama.cpp

llama.cpp b5854 before b11275 Integer Overflow in gguf_init_from_reader

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 data blob and out-of-bounds reads by consumers.

• Mettez à jour vers 1:0.6.0+dfsg-1

llama.cpp b10398 before b11301 Integer Overflow in gguf_init_from_reader

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 overflow ggml_nbytes() and strides, causing out-of-bounds memory access when buffers are sized from wrapped byte counts.

• Mettez à jour vers 1:0.6.0+dfsg-1

llama.cpp b5899 before b10261 Heap Out-of-Bounds Write via plamo2 Tokenizer

llama.cpp b5899 before b10261 contains a heap out-of-bounds write vulnerability in llm_tokenizer_plamo2::build that allows attackers to corrupt heap memory via crafted GGUF models. Attackers can supply BYTE-typed token text like <0x-1> whose negative parsed value writes token ids before the bytes_ buffer during model loading, causing crashes or potential code execution.

• Mettez à jour vers 10271+dfsg-1

ggml-org llama.cpp Mirostat Sampler llama-sampler.cpp assertion

A security flaw has been discovered in ggml-org llama.cpp up to 0.7.0. Affected by this vulnerability is an unknown functionality of the file src/llama-sampler.cpp of the component Mirostat Sampler. The manipulation results in reachable assertion. It is possible to launch the attack remotely. The project was informed of the problem early through a pull request but has not reacted yet.

ggml-org llama.cpp Pocket-TTS pockettts-gen.cpp list_pockettts_state_slots out-of-bounds write

A weakness has been identified in ggml-org llama.cpp. Affected by this issue is the function list_pockettts_state_slots of the file tools/mtmd/models/pockettts-gen.cpp of the component Pocket-TTS. This manipulation of the argument mmproj causes out-of-bounds write. The attack requires local access. The exploit has been made available to the public and could be used for attacks. Upgrading to version b11796 can resolve this issue. Patch name: 0a1ef309cdbb9f2024f58a38fb4c42352c5db569. The affected component should be upgraded. The reported heap out‑of‑bounds write (gen‑audio / pocket‑tts state_out path) is no longer a silent exploitable heap wri

llama.cpp before b11393 Use-After-Free via common_chat_peg_mapper chat_parser

llama.cpp before b11393 contains a use-after-free and double free vulnerability in common_chat_peg_mapper::map that allows unauthenticated remote attackers to corrupt heap memory via a dangling current_tool pointer. Attackers can submit a chat_parser in a POST /completion request emitting a tool-id after a tool-close tag to crash llama-server and shape a heap write primitive.

ggml-org llama.cpp RPC Server ggml-rpc.cpp deserialize_tensor assertion

A vulnerability was detected in ggml-org llama.cpp up to 0.4.0. This impacts the function rpc_server::deserialize_tensor of the file ggml/src/ggml-rpc/ggml-rpc.cpp of the component RPC Server. Performing a manipulation of the argument ne results in reachable assertion. The attack is possible to be carried out remotely. The reported GitHub issue was closed automatically due to inactivity.

CVE-2026-52132

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-2026-52131

llama.cpp b5693 and before has a Reachable Assertion via the gguf_reader::read function.

CVE-2026-52130

llama.cpp b5693 and before is vulnerable to Uncontrolled Recursion in common/json-schema-to-grammar.cpp, resulting in a denial of service.

ggml-org llama.cpp ggml-RPC Server ggml-rpc.cpp graph_compute null pointer dereference

A vulnerability was determined in ggml-org llama.cpp bec4772f6. This affects the function rpc_server::graph_compute of the file ggml/src/ggml-rpc/ggml-rpc.cpp of the component ggml-RPC Server. Executing a manipulation can lead to null pointer dereference. The attack may be launched remotely. The pull request to fix this issue awaits acceptance.

ggml-org llama.cpp ggml-RPC Server ggml-rpc.cpp deserialize_tensor deserialization

A vulnerability was found in ggml-org llama.cpp bec4772f6. The impacted element is the function deserialize_tensor of the file ggml/src/ggml-rpc/ggml-rpc.cpp of the component ggml-RPC Server. Performing a manipulation of the argument op/op_params results in deserialization. The attack may be initiated remotely. This vulnerability is distinct from CVE-2026-34159 (GHSA-j8rj-fmpv-wcxw, PR #20908), which only added a buffer==nullptr rejection in create_node() and does not validate op or op_params. The reported GitHub issue was closed automatically due to inactivity.

DEBIAN-CVE-2026-39909

DEBIAN-CVE-2026-39909

• Mettez à jour vers 8611+dfsg-1

llama.cpp b7492–b9060 Use-After-Free in Tokenization Endpoints

llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in llama-server affecting six tokenization endpoints (/tokenize, /detokenize, /infill, /apply-template, /rerank, and /anthropic/count_tokens) that bypass the task queue and access ctx_server.vocab directly on HTTP worker threads. Attackers can exploit a time-of-check-time-of-use race condition where the main thread destroys and frees vocab after the synchronization lock is released but before the handler finishes using it, causing a crash or potential code execution when --sleep-idle-seconds is configured.

llama.cpp b7492–b9060 Use-After-Free RCE via llama-server

llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in the vocab pointer of llama-server when the --sleep-idle-seconds feature is enabled, allowing unauthenticated remote attackers to execute arbitrary code. Attackers can trigger the vulnerability by sending requests to affected endpoints while the server transitions to sleep mode, causing concurrent worker threads to dereference a freed vocab pointer that can be reclaimed with attacker-controlled data to achieve remote code execution.

llama.cpp b4882-b8487 Out-of-Bounds Read Information Disclosure

llama.cpp builds b4882 before b8487 contain an out-of-bounds read vulnerability in the recurrent memory state restore path that allows attackers with write access to the slot save directory to read memory past the end of the allocated cells array. Attackers can craft a malicious slot file with an oversized seq_id value to trigger an out-of-bounds read that leaks heap data including pointer values into server logs, defeating ASLR protections and facilitating further exploitation.

• Mettez à jour vers 8611+dfsg-1

llama.cpp b4882–b9058 Buffer Overflow in KV Cache State Restore

llama.cpp builds b4882 through b9058 contain a heap buffer overflow vulnerability in the KV cache state restore path where the state_read_data() function computes write size without overflow checking, allowing attackers with write access to the slot_save_path directory to corrupt heap memory. Attackers can craft malicious state files where cell_count multiplication overflows or exceeds tensor buffer allocation to write attacker-controlled bytes past buffer boundaries, potentially resulting in heap metadata corruption, model weight corruption, or arbitrary code execution via function pointer overwrite.

llama.cpp b3978–b9058 Integer Underflow via DRY Sampler

llama.cpp builds b3978 through b9058 contain an integer underflow and out-of-bounds read vulnerability in the DRY sampler that allows unauthenticated attackers to trigger a heap buffer underflow by sending a crafted HTTP request with dry_allowed_length set to INT32_MIN to the /v1/completions or /v1/chat/completions endpoints. Attackers can exploit this vulnerability to crash the server with SIGSEGV causing denial of service for all connected users, or corrupt token sampling probabilities by reading garbage values from memory before the allocated buffer.

llama.cpp b1283–b9058 Integer Overflow in llama_batch_init() Function

llama.cpp builds b1283 through b9058 contain an integer overflow vulnerability in the llama_batch_init() function where unchecked multiplications in malloc() calls can wrap past INT32_MAX when computing allocation sizes. Attackers can pass specially crafted parameters to trigger integer overflow, causing heap corruption and potentially achieving arbitrary code execution through subsequent batch operations that write past allocated buffer boundaries.

llama.cpp b1886–b7445 Race Condition Use-After-Free via llama-android.cpp

llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution.

• Mettez à jour vers 7593+dfsg-1