llamaCVEVulnerability Intelligence
/
Advisory · GitHub · PyPI · vllm

GHSA-5JV2-G5WQ-CMR4

vLLM: GGUF dequantize kernel int truncation exposes uninitialized GPU memory in multi-tenant serving

Parche disponible
Descargar PNG JSON
CVSS
—
medium
Fuente: —
ExplotaciónSin evidenciaNo figura en CISA KEV
Exploits públicos0Ninguno conocido
Parche1 versiónPrimer parche a 13 d de la divulgación
Exposición155 dDesde la divulgación, 17 jun 2026, 9:03 a. m.
Detección152 dDesde la divulgación hasta su ingreso al índice
Timeline

Ciclo de vida de la vulnerabilidad

Hoy
Divulgación
Publicada
Modificada
Explotación
Remediación
Parche 0.24.0
Monitoreo
Detectada
17 jun 26
30 jun 26
10 sept 26
17 nov 26
20 nov 26
17 jun 26
26 jul 26
3 sept 26
12 oct 26
20 nov 26
FechaFaseEventoDetalle
Divulgación Publicada GHSA-5JV2-G5WQ-CMR4 divulgada
Remediación Parche 0.24.0 vllm · 13 d desde divulgación
Divulgación Modificada Última actualización del registro
Monitoreo Detectada Ingresó al índice
Resumen
Descripción del advisory
Texto original · GitHub · en inglés

Summary

Integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels (`csrc/quantization/gguf/gguf_kernel.cu`) causes partial tensor processing. The output tensor is allocated at full size via `torch::empty` (uninitialized memory), but the dequantize CUDA kernel processes only a truncated number of elements. The unfilled portion of the output tensor retains whatever was previously in GPU memory. In multi-tenant inference deployments, this residual GPU memory may contain tensor data from other users' inference requests, constituting information disclosure.

Root Cause

The `to_cuda_ggml_t` function pointer type at `ggml-common.h:1067` declares its element count parameter as `int` (32-bit):

using to_cuda_ggml_t = void (*)(const void * __restrict__ x,
                                dst_t * __restrict__ y,
                                int k,              // 32-bit
                                cudaStream_t stream);

All dequantize kernel functions (`dequantize_block_cuda`, `dequantize_row_q2_K_cuda`, etc. in `dequantize.cuh`) inherit this `int k` parameter and use it as the kernel launch grid size:

static void dequantize_block_cuda(..., const int k, cudaStream_t stream) {
    const int num_blocks = (k + 2*CUDA_DEQUANTIZE_BLOCK_SIZE - 1) / (2*CUDA_DEQUANTIZE_BLOCK_SIZE);
    dequantize_block<<<num_blocks, CUDA_DEQUANTIZE_BLOCK_SIZE, 0, stream>>>(vx, y, k);
}

In `ggml_dequantize()` at `gguf_kernel.cu:85`, the caller passes `m * n` (an `int64_t` product) to this `int k` parameter:

at::Tensor DW = torch::empty({m, n}, options);    // line 80: full-size, UNINITIALIZED
// ...
to_cuda((void*)W.data_ptr(), (scalar_t*)DW.data_ptr(), m * n, stream);  // line 85: m*n truncated to int

When `m * n > INT_MAX`, the truncated `k` is smaller than the actual tensor size. The kernel processes `k` elements. The remaining `(m * n) - k` elements in `DW` are never written and contain stale GPU memory.

This is a single root cause -- the `int` type on the `k` parameter in `to_cuda_ggml_t` -- with a single fix: change `int k` to `int64_t k`. All dequantize functions inherit this type through the same typedef.

Affected Functions

All in `csrc/quantization/gguf/gguf_kernel.cu`:

| Function | Line | Allocation | Info Disclosure? |
|----------|------|-----------|-----------------|
| `ggml_dequantize` | 74 | `torch::empty({m, n})` at line 80 | Yes -- `m*n` truncated to `int k` at line 85 |
| `ggml_mul_mat_vec_a8` | 91 | `torch::empty({vecs, row})` at line 99 | Yes -- `int col = X.sizes()[1]` at line 94 |
| `ggml_mul_mat_a8` | 207 | `torch::empty({batch, row})` at line 215 | Yes -- `int col = X.sizes()[1]` at line 210 |
| `ggml_moe_a8` | 279 | `torch::empty({tokens*top_k, row})` at line 289 | Yes -- `int col = X.sizes()[1]` at line 285 |

All four functions allocate output tensors with `torch::empty` (uninitialized) and then run CUDA kernels that use truncated dimension values as loop bounds. The unfilled portion of each output tensor retains stale GPU memory.

`ggml_moe_a8_vec` (line 382) uses `torch::zeros` instead of `torch::empty`, so it is not affected by the info disclosure variant.

Impact: Information Disclosure in Multi-Tenant Serving

vLLM is designed for multi-tenant inference serving. GPU memory is reused across requests from different users. When the dequantize kernel partially fills an output tensor:

  1. The output tensor `DW` is allocated with `torch::empty` -- the buffer contains whatever was previously in that GPU memory region
  2. The dequantize kernel fills only a truncated portion of the buffer
  3. The unfilled portion retains residual data from prior GPU operations, which may include tensor data from other users' inference requests
  4. The contaminated tensor proceeds through the model computation
  5. No error or warning is generated -- the partial fill is silent

This is a confidentiality violation. In shared inference deployments (the primary vLLM use case), one user's inference data can leak into another user's model computation through residual GPU memory.

Attacker Control

The attacker crafts a GGUF model file with weight tensor dimensions whose product exceeds `INT_MAX` (e.g., a matrix with shape `[65536, 65536]` gives `m * n = 4,294,967,296`). The model is hosted on HuggingFace or any model hub. The victim loads the model with vLLM for inference serving. The truncation happens automatically during model weight dequantization.

Fix

A fix for this vulnerability was added here: https://github.com/vllm-project/vllm/pull/44971

Versiones

Rangos afectados

Afectada Corregida Última
vllm≥ 0.5.5 < 0.24.0
0.5.5
0.24.0
0.34.0
Remediación

Estado del parche por paquete

Paquete / productoEstadoCorregida enDesde divulgaciónÚltima versión
pypivllm
Corregida 0.24.0 13 d 0.34.0 no afectada
Versiones corregidas0.24.0
Explotación

Arsenal público

Sin exploits, módulos ni PoC públicos conocidos a la fecha de verificación.

Referencias
Advisoryhttps://nvd.nist.gov/vuln/detail/CVE-2026-53923
Advisoryhttps://github.com/advisories/GHSA-5jv2-g5wq-cmr4
Otrahttps://github.com/vllm-project/vllm/security/advisories/GHSA-5jv2-g5wq-cmr4
Otrahttps://github.com/vllm-project/vllm/pull/44971
Otrahttps://github.com/vllm-project/vllm/commit/f219788f91952827132fa4fdf916427cd20d225e
Otrahttps://github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2026-3403.yaml
Otrahttps://github.com/vllm-project/vllm
Otrahttps://pypi.org/project/vllm

Las referencias externas se muestran para copiar: cve.llama.ec no lo redirige a otros sitios.