CVE-2026-94627

vLLM Mooncake connector through 0.29.0 fails to properly manage GPU KV cache block ownership when concurrent child requests share a single transfer ID in prefill/decode disaggregated deployments. Attackers can trigger GPU memory exhaustion by submitting completion requests with multiple prompts, causing orphaned KV cache blocks to accumulate until process restart and eventually preventing legitimate requests from executing.
Configurations

Configuration 1 (hide)

cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*

History

29 Sep 2026, 12:46

Type Values Removed Values Added
CPE cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*
First Time Vllm vllm
Vllm
References () https://github.com/vllm-project/vllm - () https://github.com/vllm-project/vllm - Product
References () https://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py#L1978-L1989 - () https://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py#L1978-L1989 - Product
References () https://github.com/vllm-project/vllm/pull/49796 - () https://github.com/vllm-project/vllm/pull/49796 - Issue Tracking, Patch
References () https://www.vulncheck.com/advisories/vllm-through-0.29.0-gpu-kv-cache-leak-via-mooncake-transfer-id-collision - () https://www.vulncheck.com/advisories/vllm-through-0.29.0-gpu-kv-cache-leak-via-mooncake-transfer-id-collision - Third Party Advisory

21 Sep 2026, 22:17

Type Values Removed Values Added
New CVE

Information

Published : 2026-09-21 22:17

Updated : 2026-09-29 12:46


NVD link : CVE-2026-94627

Mitre link : CVE-2026-94627

CVE.ORG link : CVE-2026-94627


JSON object : View

Products Affected

vllm

  • vllm
CWE
CWE-401

Missing Release of Memory after Effective Lifetime