| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| 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. |
| vLLM through 0.29.0 contains a denial of service vulnerability in the NIXL connector's prefix caching implementation that fails to properly validate block counts across multi-prompt completion requests in prefill/decode disaggregated deployments. Attackers can trigger an assertion failure in NixlBaseConnectorWorker._apply_prefix_caching by submitting completion requests with multiple prompts of varying lengths, causing the decode worker to terminate and become unavailable until restarted. |
| uutils coreutils versions before 0.10.0 apply setuid or setgid mode to install destinations before finalizing ownership changes, allowing privileged users to leave setuid executables owned by the privileged invoker when ownership changes fail. Attackers can execute leftover setuid files with elevated privileges when ownership change operations fail on capability-restricted systems. |
| vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. A single request with a negative token ID triggers a CUDA device-side assertion that poisons the GPU context, causing all subsequent requests to fail until the process restarts. |
| vLLM through 0.29.0 fails to properly clean up decode-side metadata for rejected inference requests in prefill/decode disaggregated deployments. Remote attackers can submit requests with max_tokens=0 to exhaust decode-worker memory without bound until the worker restarts. |
| vLLM through 0.29.0 fails to validate the tp_size parameter in kv_transfer_params on OpenAI-compatible completion endpoints, allowing attackers to allocate unbounded memory. Attackers can supply arbitrary tp_size values in prefill/decode disaggregated deployments to exhaust memory and trigger kernel OOM-kill of the decode worker process. |
| vLLM through 0.29.0 contains a denial of service vulnerability in P2P KV offloading when OffloadingConnector is configured with TieringOffloadingSpec and a peer-to-peer secondary tier. Attackers can supply arbitrary remote host and port values in kv_transfer_params to create unreachable peer sessions that retain ZeroMQ sockets until the context quota is exhausted, causing an uncaught ZMQError that crashes EngineCore and stops all inference. |
| Use of Externally-Controlled Input to Select Classes or Code ('Unsafe Reflection') vulnerability in Apache Calcite Avatica. Plugin instantiation (via AvaticaUtils#instantiatePlugin and other methods) initializes arbitrary classes via unrestricted calls to Class.forName(String) which by default triggers initialization. This may lead to the execution of static initializer blocks in arbitrary classes present in the classpath. The instantiation APIs should initialize and instantiate only classes implementing the specified plugin interface passed as input in conjunction with the desired classname. At the moment of writing, there are no well-known or widely used classes with dangerous static initializer blocks so the severity is low.
This issue affects Apache Calcite Avatica: before 1.29.0.
Users are recommended to upgrade to version 1.29.0, which fixes the issue. |
| The Okta Access Gateway improperly handles input sanitization and regular expression evaluation within its Protected Rule authorization check, resulting in an authorization bypass when an administrator has explicitly configured a Protected Rule policy on one or more application resources. |
| The Okta Hyperdrive Integration installer does not mask the OAuth client secret when passed as an MSI property. The credential is recorded in plaintext in the installer log, the Application Event Log, and the process command line, all of which are readable by an authenticated local user on the workstation. |
| Double free in Windows Hello allows an authorized attacker to elevate privileges locally. |
| Heap-based buffer overflow in Windows Credential Providers allows an authorized attacker to execute code over a network. |
| Heap-based buffer overflow in HID class driver allows an authorized attacker to elevate privileges locally. |
| Heap-based buffer overflow in Windows Link Layer Topology Discovery Protocol allows an unauthorized attacker to execute code over a network. |
| Use after free in Windows Broadcast DVR User Service allows an authorized attacker to elevate privileges locally. |
| Use after free in Windows Hello allows an authorized attacker to elevate privileges locally. |
| Use after free in Windows TCP/IP allows an authorized attacker to elevate privileges over a network. |
| Heap-based buffer overflow in Windows Universal Disk Format File System Driver (UDFS) allows an authorized attacker to elevate privileges locally. |
| Use after free in Windows TCP/IP allows an authorized attacker to elevate privileges over a network. |
| Stack-based buffer overflow in Windows Win32K allows an authorized attacker to elevate privileges over a network. |
