CVE-2026-115916

01/תיאור (מקור, אנגלית)

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py computes max_token_id with an off-by-one upper bound, allowing a pre-tokenized prompt ID equal to the vocabulary size to pass validation. For a tensor-parallel size of one and an exact multiple-of-64 vocabulary, VocabParallelEmbedding.forward does not mask that value before F.embedding uses it, causing a CUDA device-side assertion and terminating the engine. This issue is fixed in version 0.32.0.

02/מוצרים מושפעים

vllm: vllm

03/קישורים4

• github.com/vllm-project/vllm/security/advisories/GHSA-q97c-p36‮… PatchVendor Advisory
• github.com/vllm-project/vllm/commit/37cecc8141e2dcba9dd534897a‮… Patch
• github.com/vllm-project/vllm/pull/59491 Patch
• github.com/vllm-project/vllm/releases/tag/v0.32.0 Release Notes