vLLM is an inference and serving engine for large language models. Prior to 0.32.0, the Phi-4 multimodal audio encoder in phi4mm_audio.py and phi4mm_utils.py uses eager attention by default, allowing a remote client to submit one legal long audio clip that causes MultiHeadedAttention.forward_attention to materialize and retain full attention matrices in self.attn. The scaled dot-product attention branch does not materialize these matrices. The resulting quadratic GPU-memory use can cause an out-of-memory condition and terminate EngineCore for all users. This issue is fixed in version 0.32.0.
{
"cna_assigner": "GitHub_M",
"cwe_ids": [
"CWE-770"
],
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/115xxx/CVE-2026-115913.json"
}{
"extracted_events": [
{
"cpe": "cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*",
"range": [
{
"introduced": "0.31.0"
},
{
"fixed": "0.32.0"
}
],
"source": "CPE_RANGE"
},
{
"range": [
{
"fixed": "3d52578e13454de0bb8352cac988347cdcc89928"
}
],
"source": "REFERENCES_COMMIT"
},
{
"original_tag": "v0.32.0",
"range": [
{
"fixed": "5df63a7203cdb9328d824a27b682e77c7b9928be"
}
],
"source": "REFERENCES_TAG"
}
]
}