GHSA-FRVG-QW88-HWW9 vllm p2p never connected zmq_exhaustion

Summary When the OffloadingConnector uses TieringOffloadingSpec with a P2P secondary tier, an API client can submit many unique non-listening peer addresses through kvtransferparams.prefill. vLLM keeps every never-established P2P session and its ZeroMQ sockets after the request finishes. On vLLM...

EUVD-2026-84218

vllm p2p never connected zmqexhaustion...

PT-2026-123436

Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.32.0 Description The Phi-4 multimodal audio encoder in phi4mm audio.py and phi4mm utils.py uses eager attention by default. This allows a remote client to submit a long audio clip that triggers the...

Allocation of Resources Without Limits or Throttling

Overview vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs Affected versions of this package are vulnerable to Allocation of Resources Without Limits or Throttling in the Rust gRPC converter rust/src/server/src/grpc/convert.rs via the CandidateTokens.topn fie...

Allocation of Resources Without Limits or Throttling

Overview vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs Affected versions of this package are vulnerable to Allocation of Resources Without Limits or Throttling via the Rust gRPC frontend's handling of outputcandidates.tokenids in Generate and GenerateStre...

CVE-2026-116180

vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.emptycpu in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sharing the engine can send promptlogprobs...

CVE-2026-116174

vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust gRPC Generate and GenerateStream service that forwards unchecked prompt token IDs to EngineCore. Attackers can submit requests with out-of-vocabulary prompt.tokenids values to trigger a fatal embedding...

CVE-2026-116173

vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to terminate EngineCore by bypassing logprobtokenids limits in the Rust gRPC frontend. Attackers can send Generate or GenerateStream requests with outputlogprobs enabled and more than 128...

CVE-2026-116180

The vLLM library (versions 0.8.5 before 0.31.0 ) is vulnerable to a Cross-Tenant Denial of Service (DoS) flaw. The root cause is located in vllm/v1/outputs.py, where LogprobsTensors.empty_cpu() allocates an uninitialized int32 tensor . Because the V1 model runner fails to write to prompt position...

CVE-2026-116180

vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.emptycpu in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sharing the engine can send promptlogprobs...

CVE-2026-116180 vLLM from 0.8.5 before 0.31.0 Cross-Tenant DoS via Uninitialized LogprobsTensors

vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.emptycpu in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sharing the engine can send promptlogprobs...

CVE-2026-116180 vLLM from 0.8.5 before 0.31.0 Cross-Tenant DoS via Uninitialized LogprobsTensors

vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.emptycpu in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sharing the engine can send promptlogprobs...

CVE-2026-116180 vLLM from 0.8.5 before 0.31.0 Cross-Tenant DoS via Uninitialized LogprobsTensors

vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.emptycpu in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sharing the engine can send promptlogprobs...

EUVD-2026-109441

vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.emptycpu in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sharing the engine can send promptlogprobs...

CVE-2026-116174

The vLLM library versions 0.22.0 before 0.24.0 is vulnerable to a Denial of Service (DoS) attack due to improper input validation in the Rust gRPC Generate and GenerateStream services. The vulnerability occurs because these services forward unchecked prompt token IDs to EngineCore . An attacker c...

EUVD-2026-109435

vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust gRPC Generate and GenerateStream service that forwards unchecked prompt token IDs to EngineCore. Attackers can submit requests with out-of-vocabulary prompt.tokenids values to trigger a fatal embedding...

CVE-2026-116173

The vLLM library versions 0.22.0 before 0.24.0 is vulnerable to a Denial of Service (DoS) attack due to improper input validation in the Rust gRPC frontend. A remote attacker can terminate the EngineCore by sending Generate or GenerateStream requests with output_logprobs enabled and more than 128...

CVE-2026-116173

vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to terminate EngineCore by bypassing logprobtokenids limits in the Rust gRPC frontend. Attackers can send Generate or GenerateStream requests with outputlogprobs enabled and more than 128...

EUVD-2026-109434

vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to terminate EngineCore by bypassing logprobtokenids limits in the Rust gRPC frontend. Attackers can send Generate or GenerateStream requests with outputlogprobs enabled and more than 128...

CVE-2026-116173 vLLM 0.22.0 before 0.24.0 Denial of Service via gRPC output_candidates.token_ids

vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to terminate EngineCore by bypassing logprobtokenids limits in the Rust gRPC frontend. Attackers can send Generate or GenerateStream requests with outputlogprobs enabled and more than 128...