Sign in Sign up
JA · EN · FR
SecPulse
Categories
CVE-2026-53923 high CVSS 7.5

Information Disclosure in vllm (CVE-2026-53923)

Summary

Vulnerability in vllm (CVE-2026-53923). Confidential information can be exposed externally. Exploitable via to_cuda_ggml_t. Mitigation: upgrade to 0.24.0 or later.

Original description

The description as published by the source.

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<<>>(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:…

Source: GitHub Security Advisories (Global) NVD GitHub Advisory

Auto-generated summary

Assembled by SecPulse from the source data (CVSS, CWE, affected versions) using templates — no AI involved. Check the original description and references below for exact details.

A vulnerability tracked as CVE-2026-53923 has been found in vllm.
Attackers can target a specific entry point like to_cuda_ggml_t over the network to misuse the product.
Confidential information can be exposed externally. CVSS score: 7.5/10.
What to do: upgrade vllm to 0.24.0 or later.
If unsure, ask your IT team or search "vllm CVE-2026-53923" on the vendor's site.

CVE-2026-53923 (vllm) — CWE-200 / CVSS v3 7.5
Attack vector: remote (network-reachable) / unauthenticated / no user interaction
Attack surface: to_cuda_ggml_t / int / dequantize_block_cuda / dequantize_row_q2_K_cuda
Patched: 0.24.0 — apply immediately
Plan: 1) Audit SBOM/dependencies, 2) Stage→prod upgrade, 3) Add WAF/proxy monitoring on affected endpoints, 4) Hunt IOCs in logs.
Refs: see the GHSA / vendor advisory / patched release linked on this page.

❓ What is the problem

Information disclosure (CWE-200) exists in vllm. Attackers reach the vulnerable code path via to_cuda_ggml_t without authentication.

📍 Affected scope

Check the vendor advisory for the affected versions of vllm. Attack surface: to_cuda_ggml_t / int / dequantize_block_cuda / dequantize_row_q2_K_cuda.

🔥 Severity

Severity: High (CVSS 7.5/10). Confidential information can be exposed externally.

🔧 How to fix

Update to 0.24.0.

🛡️ Workaround

Until the patch is applied: disable the affected feature, apply WAF rules, or restrict access via network ACLs.

🔍 Detection

Search webserver/proxy logs for unusual request patterns matching this CVE's known IOCs. Run grep -r 'vllm' . against your dependency files (package-lock.json, requirements.txt, go.sum) to find affected services.

Response actions

Concrete steps and command examples for SOC/SRE teams to execute in order

  1. 1
    Identify exposure identify 
    grep -r 'vllm' . | grep -v node_modules

    リポジトリと本番環境の依存ファイル (package-lock.json / requirements.txt / go.sum / Gemfile.lock 等) で vllm を grep し、稼働しているサービス・バージョンを把握する。

  2. 2
    Apply patch patch 
    Upgrade vllm to 0.24.0

    ステージング環境で 0.24.0 に上げて回帰テスト → 本番反映。回帰テストはアプリの主要ハッピーパスを含めること。

  3. 3
    Post-deployment verification verify 
    Confirm patched version is live in production

    パッチ適用後、ステージングで PoC または同等の悪用パターンを再現して脆弱性が閉じたことを確認。本番ではアクセスログ・WAF でアラート再発が無いか継続監視。

Affected packages

pip vllm
[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"0.24.0"}]}]

References

We use cookies for analytics and ad measurement. This setting applies to all Game co., Ltd. sites. Details