CVE-2026-94627
vLLM through 0.29.0 GPU KV Cache Leak via Mooncake Transfer ID Collision
- State
- Published
- Severity
- 8.7 High
- Published
- Sep 21, 2026
- Updated
- Sep 22, 2026
- Reserved
- Sep 21, 2026
Description
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.
Scores
| CVSS | Score | Vector | Source |
|---|---|---|---|
| 4.0 | 8.7 High | CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N | VulnCheck |
| 3.1 | 7.5 High | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H | VulnCheck |
Weaknesses
- CWE-401 Missing Release of Memory after Effective Lifetime
Affected Products
| Product | Versions |
|---|---|
vllmvllm-project |
|
References
- Pull Request #49796issue-trackingpatch
- https://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py#L1978-L1989technical-description
- https://github.com/vllm-project/vllmproduct
- VulnCheck Advisory: vLLM through 0.29.0 GPU KV Cache Leak via Mooncake Transfer ID Collisionthird-party-advisory
Every record here comes from the CVE Program’s CVE List. © 1999–2026 The MITRE Corporation. CVE is a registered trademark of The MITRE Corporation. The records are reproduced under the CVE Program terms of use.