GHSA-q97c-p368-75c4 CVE-2026-115916 PYSEC-2026-4347
vLLM: Out-of-vocabulary prompt token id passes input validation and triggers a CUDA device-side assert (engine crash / DoS)
- Severity
- 6.5 / 10
- Medium · CVSS 3.1
- Exploit chance (EPSS)
- 0.31%
- Chance of exploitation in the next 30 days. Higher than 24.10% of scored vulnerabilities.
- Known exploited
- No
- Not in CISA's catalog
- Fix
0.32.0- Upgrade vllm to this version or later
Description
For a model whose vocabulary size is an exact multiple of the embedding padding size (64) — e.g. any Llama-2-family model, vocab 32000 — a single request carrying a pre-tokenized prompt with token id equal to the vocab size (32000) passes vLLM's input-id validation but is one past the last embedding row. At tensor-parallel size 1 the embedding lookup does not mask indices, so the id reaches F.embedding out of range and raises a CUDA device-side assert, which poisons the engine's CUDA context and takes the engine down. Any client that can send pre-tokenized input (/v1/completions with prompt as token ids, /tokenize-fed flows, or the kv_transfer_params["prompt_token_ids"] reuse path) can trigger it.
Root cause
vllm/v1/engine/input_processor.py accepts a token id when id <= max(tokenizer.max_token_id, model_config.get_vocab_size() - 1).
vllm/tokenizers/hf.py computes max_token_id = max(max(vocab.values()), tokenizer.vocab_size). For Llama-2 tokenizers max(vocab.values()) = 31999 but tokenizer.vocab_size = 32000, so max_token_id = 32000 — one greater than the largest valid id.
The engine bound therefore becomes max(32000, 31999) = 32000, and id 32000 is accepted.
The model's embedding has pad_vocab_size(32000, 64) = 32000 rows (no padding), valid indices 0..31999. At VocabParallelEmbedding.forward with tp_size == 1 there is no masking, so index 32000 hits F.embedding out of range.
Affected
- tp=1 (single GPU per model — common per-copy layout). tp>1 masks via
get_masked_input_and_mask, so the id is clamped, not an OOB — likely wrong output rather than a crash. - Models whose vocab is a multiple of 64 with
tokenizer.vocab_sizeone past the max real id (Llama-2 family, vocab 32000, is the clear case). Qwen2.5 is NOT affected (padded headroom).
Suggested fix
max_token_id in vllm/tokenizers/hf.py now uses tokenizer.vocab_size - 1, so an id equal to the vocabulary size fails input validation instead of reaching the embedding. Fixed in MR59491, released in v0.32.0.
Reporter
Shijie Lyu (GitHub shijie-lyu lshjhf@gmail.com). Reported privately per SECURITY.
This advisory was fixed in MR59491.
Affected versions
PyPI
How to fix
Move to the fixed version or later with your package manager, update the lockfile, and run your tests. If the package only arrives through another dependency, upgrade that dependency first. If none of its releases pulls a fixed version yet, force the version with your package manager's override.
vllm PyPI → 0.32.0
- pip
pip install --upgrade "vllm>=0.32.0"- poetry
poetry add "vllm>=0.32.0"- uv
uv add "vllm>=0.32.0"
Severity breakdown
CVSS 3.1 6.5
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
- Attack vector
- Network
- Attack complexity
- Low
- Privileges required
- Low
- User interaction
- None
- Scope
- Unchanged
- Confidentiality
- None
- Integrity
- None
- Availability
- High
Weakness
CWE-129: Improper Validation of Array Index
The product uses untrusted input when calculating or using an array index, but the product does not validate or incorrectly validates the index to ensure the index references a valid position within the array.
How to prevent it
- Architecture and Design: Use an input validation framework such as Struts or the OWASP ESAPI Validation API. Note that using a framework does not automatically address all input validation problems; be mindful of weaknesses that could arise from misusing the framework itself (CWE-1173).
- Architecture and Design: For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server. Even though client-side checks provide minimal benefits with respect to server-side security, they are still useful. First, they can support intrusion detection. If the server receives input that should have been rejected by the client, then it may be an indication of an attack. Second, client-side error-checking can provide helpful feedback to the user about the expectations for valid input. Third, there may be a reduction in server-side processing time for accidental input errors, although this is typically a small savings.
- Requirements: Use a language that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid. For example, Ada allows the programmer to constrain the values of a variable and languages such as Java and Ruby will allow the programmer to handle exceptions when an out-of-bounds index is accessed.
- Operation, Build and Compilation: Run or compile the software using features or extensions that randomly arrange the positions of a program's executable and libraries in memory. Because this makes the addresses unpredictable, it can prevent an attacker from reliably jumping to exploitable code. Examples include Address Space Layout Randomization (ASLR) [REF-58] [REF-60] and Position-Independent Executables (PIE) [REF-64]. Imported modules may be similarly realigned if their default memory addresses conflict with other modules, in a process known as "rebasing" (for Windows) and "prelinking" (for Linux) [REF-1332] using randomly generated addresses. ASLR for libraries cannot be used in conjunction with prelink since it would require relocating the libraries at run-time, defeating the whole purpose of prelinking. For more information on these techniques see D3-SAOR (Segment Address Offset Randomization) from D3FEND [REF-1335].
- Operation: Use a CPU and operating system that offers Data Execution Protection (using hardware NX or XD bits) or the equivalent techniques that simulate this feature in software, such as PaX [REF-60] [REF-61]. These techniques ensure that any instruction executed is exclusively at a memory address that is part of the code segment. For more information on these techniques see D3-PSEP (Process Segment Execution Prevention) from D3FEND [REF-1336].
- Implementation: Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright. When accessing a user-controlled array index, use a stringent range of values that are within the target array. Make sure that you do not allow negative values to be used. That is, verify the minimum as well as the maximum of the range of acceptable values.
- Implementation: Be especially careful to validate all input when invoking code that crosses language boundaries, such as from an interpreted language to native code. This could create an unexpected interaction between the language boundaries. Ensure that you are not violating any of the expectations of the language with which you are interfacing. For example, even though Java may not be susceptible to buffer overflows, providing a large argument in a call to native code might trigger an overflow.
- Architecture and Design, Operation: Run your code using the lowest privileges that are required to accomplish the necessary tasks [REF-76]. If possible, create isolated accounts with limited privileges that are only used for a single task. That way, a successful attack will not immediately give the attacker access to the rest of the software or its environment. For example, database applications rarely need to run as the database administrator, especially in day-to-day operations.
- Architecture and Design, Operation: Run the code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which files can be accessed in a particular directory or which commands can be executed by the software. OS-level examples include the Unix chroot jail, AppArmor, and SELinux. In general, managed code may provide some protection. For example, java.io.FilePermission in the Java SecurityManager allows the software to specify restrictions on file operations. This may not be a feasible solution, and it only limits the impact to the operating system; the rest of the application may still be subject to compromise. Be careful to avoid CWE-243 and other weaknesses related to jails.
CWE-193: Off-by-one Error
A product calculates or uses an incorrect maximum or minimum value that is 1 more, or 1 less, than the correct value.
How to prevent it
- Implementation: When copying character arrays or using character manipulation methods, the correct size parameter must be used to account for the null terminator that needs to be added at the end of the array. Some examples of functions susceptible to this weakness in C include strcpy(), strncpy(), strcat(), strncat(), printf(), sprintf(), scanf() and sscanf().
References
Advisories
Fixes and patches
Packages
Other links
Timeline
- Published
- Last updated by the source
- Latest EPSS score
Related advisories
| Advisory | Severity | Why it's related |
|---|---|---|
| GHSA-8397-8qf6-f2gm vLLM: Phi-4 multimodal audio encoder materializes and retains full attention matrices: a single max-length audio OOM-crashes the vLLM engine | Medium | Also affects vllm |
| GHSA-x567-49g5-p4cf vLLM: Responses API bypasses trust_request_chat_template for chat_template_kwargs | Medium | Also affects vllm |
| GHSA-vvpg-rrww-wpmh vLLM: Permissive default CORS origins allow cross-origin requests to the vLLM HTTP server | Medium | Also affects vllm |
| GHSA-h3rc-j82p-2c3h vLLM: Untrusted EPS inputs can trigger denial of service or Ghostscript RCE | High | Also affects vllm |
| GHSA-gq5q-89c7-vhvx vLLM DeepSeek-OCR image processing vulnerable to denial of service | Medium | Also affects vllm |
| GHSA-6757-hggm-hp2w vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled | High | Also affects vllm |
| GHSA-542m-9234-8gvf vLLM: DoS via unbounded chat_template Jinja2 rendering exhausts the renderer thread pool | Medium | Also affects vllm |
| PYSEC-2026-4353 vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote attackers to crash the engine by sending inconsistent multimodal… | High | Also affects vllm |
Credits: jperezdealgaba, shijie-lyu
MergeHarbor checks every repository's dependencies against this database, ranks what matters by exploitability and reachability, routes it to the owning team and opens the fix pull request.
Check your repos with MergeHarbor