{"items":[{"id":"NVI-00016294","kind":"vulnerability","title":"vLLM: GGUF dequantize kernel int truncation exposes uninitialized GPU memory in multi-tenant serving","revision":2,"unknowns":["Customer inventory, deployment, reachability and exploitation have not been inspected.","NVD lastModified does not establish freshness of every associated CPE record."],"lifecycle":"active","identifiers":["GHSA-5jv2-g5wq-cmr4","CVE-2026-53923"],"plain_summary":"GitHub lists this reviewed advisory for vllm. Published version ranges and source prerequisites are shown below; this is not an inspection of a visitor’s application.","review_status":"source_listed","control_topics":[],"public_actions":[{"summary":"Identify the exact component and deployed version, examine the linked publisher’s conditions, and retain evidence of any mitigation.","source_ids":["sr-03034ef0adcffa1734fbd30d070fbfd2"],"action_type":"investigate","limitations":"Confirm current vendor support, compatibility and deployment evidence. This report has not tested a fix in your system.","current_supported_target":null}],"schema_version":"1.1.0","source_metrics":[{"kind":"CVSS","as_of":"2026-07-17T16:20:59.000Z","value":"7.5 medium","meaning":"GitHub source severity; exploitability and local consequence require separate evidence.","source_id":"sr-6475672e8f62d780aff81e6a426cd839"},{"kind":"CVSS","as_of":"2026-10-25T23:10:00.167Z","value":"7.5 HIGH (3.1)","meaning":"nvd@nist.gov severity; no combined Null risk score.","source_id":"sr-03034ef0adcffa1734fbd30d070fbfd2"},{"kind":"CVSS","as_of":"2026-10-25T23:10:00.167Z","value":"5.3 MEDIUM (4.0)","meaning":"security-advisories@github.com severity; no combined Null risk score.","source_id":"sr-03034ef0adcffa1734fbd30d070fbfd2"}],"why_it_matters":"Published affected context helps an owner decide what evidence to gather. This report has not examined your application.","related_sources":[{"id":"sr-6475672e8f62d780aff81e6a426cd839","url":"https://github.com/advisories/GHSA-5jv2-g5wq-cmr4","basis":"vendor_advisory","notes":"GitHub Advisory Database; CC-BY-4.0. Each advisory links its original attribution. Null normalizes fields.","title":"GHSA-5jv2-g5wq-cmr4","publisher":"GitHub Advisory Database","observed_at":"2026-11-04T09:59:26.779Z"},{"id":"sr-03034ef0adcffa1734fbd30d070fbfd2","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-53923","basis":"official_catalog","notes":"This product uses data from the NVD API but is not endorsed or certified by the NVD.","title":"CVE-2026-53923","publisher":"NIST National Vulnerability Database","observed_at":"2026-11-08T01:48:05.462Z"}],"affected_context":[{"certainty":"source_listed","source_ids":["sr-6475672e8f62d780aff81e6a426cd839"],"technology":"vllm","known_exclusions":[],"conditions_summary":"## Summary\n\nInteger truncation of tensor dimensions in vLLM's GGUF dequantize kernels (`csrc/quantization/gguf/gguf_kernel.cu`) causes partial tensor processing. The output tensor is allocated at full size via `torch::empty` (uninitialized memory), but the dequantize CUDA kernel processes only a truncated number of elements. The unfilled portion of the output tensor retains whatever was previously in GPU memory. In multi-tenant inference deployments, this residual GPU memory may contain tensor data from other users' inference requests, constituting information disclosure.\n\n## Root Cause\n\nThe `to_cuda_ggml_t` function pointer type at `ggml-common.h:1067` declares its element count parameter as `int` (32-bit):\n\n```cpp\nusing to_cuda_ggml_t = void (*)(const void * __restrict__ x,\n                                dst_t * __restrict__ y,\n                                int k,              // 32-bit\n                                cudaStream_t stream);\n```\n\nAll dequantize kernel functions (`dequantize_block_cuda`, `dequantize_row_q2_K_cuda`, etc. in `dequantize.cuh`) inherit this `int k` parameter and use it as the kernel launch grid size:\n\n```cpp\nstatic void dequantize_block_cuda(..., con","public_affected_ranges":[">= 0.5.5, < 0.24.0"]}],"null_published_at":"2026-11-08T19:53:44.016Z","source_updated_at":"2026-07-17T16:20:59.000Z","source_disclosed_at":"2026-06-17T14:03:11.000Z","is_synthetic_fixture":false,"null_last_reviewed_at":null,"public_change_summary":"Source facts changed; compare the preserved prior revision.","null_first_observed_at":"2026-11-04T09:59:26.779Z"},{"id":"NVI-00016294","kind":"vulnerability","title":"vLLM: GGUF dequantize kernel int truncation exposes uninitialized GPU memory in multi-tenant serving","revision":1,"unknowns":["Customer inventory, deployment, reachability and exploitation have not been inspected."],"lifecycle":"active","identifiers":["GHSA-5jv2-g5wq-cmr4","CVE-2026-53923"],"plain_summary":"GitHub lists this reviewed advisory for vllm. Published version ranges and source prerequisites are shown below; this is not an inspection of a visitor’s application.","review_status":"source_listed","control_topics":[],"public_actions":[{"summary":"Identify the exact component and deployed version, examine the linked publisher’s conditions, and retain evidence of any mitigation.","source_ids":["sr-6475672e8f62d780aff81e6a426cd839"],"action_type":"investigate","limitations":"Confirm current vendor support, compatibility and deployment evidence. This report has not tested a fix in your system.","current_supported_target":null}],"schema_version":"1.1.0","source_metrics":[{"kind":"CVSS","as_of":"2026-07-17T16:20:59.000Z","value":"7.5 medium","meaning":"GitHub source severity; exploitability and local consequence require separate evidence.","source_id":"sr-6475672e8f62d780aff81e6a426cd839"}],"why_it_matters":"Published affected context helps an owner decide what evidence to gather. This report has not examined your application.","related_sources":[{"id":"sr-6475672e8f62d780aff81e6a426cd839","url":"https://github.com/advisories/GHSA-5jv2-g5wq-cmr4","basis":"vendor_advisory","notes":"GitHub Advisory Database; CC-BY-4.0. Each advisory links its original attribution. Null normalizes fields.","title":"GHSA-5jv2-g5wq-cmr4","publisher":"GitHub Advisory Database","observed_at":"2026-11-04T09:59:26.779Z"}],"affected_context":[{"certainty":"source_listed","source_ids":["sr-6475672e8f62d780aff81e6a426cd839"],"technology":"vllm","known_exclusions":[],"conditions_summary":"## Summary\n\nInteger truncation of tensor dimensions in vLLM's GGUF dequantize kernels (`csrc/quantization/gguf/gguf_kernel.cu`) causes partial tensor processing. The output tensor is allocated at full size via `torch::empty` (uninitialized memory), but the dequantize CUDA kernel processes only a truncated number of elements. The unfilled portion of the output tensor retains whatever was previously in GPU memory. In multi-tenant inference deployments, this residual GPU memory may contain tensor data from other users' inference requests, constituting information disclosure.\n\n## Root Cause\n\nThe `to_cuda_ggml_t` function pointer type at `ggml-common.h:1067` declares its element count parameter as `int` (32-bit):\n\n```cpp\nusing to_cuda_ggml_t = void (*)(const void * __restrict__ x,\n                                dst_t * __restrict__ y,\n                                int k,              // 32-bit\n                                cudaStream_t stream);\n```\n\nAll dequantize kernel functions (`dequantize_block_cuda`, `dequantize_row_q2_K_cuda`, etc. in `dequantize.cuh`) inherit this `int k` parameter and use it as the kernel launch grid size:\n\n```cpp\nstatic void dequantize_block_cuda(..., con","public_affected_ranges":[">= 0.5.5, < 0.24.0"]}],"null_published_at":"2026-11-04T10:04:27.928Z","source_updated_at":"2026-07-17T16:20:59.000Z","source_disclosed_at":"2026-06-17T14:03:11.000Z","is_synthetic_fixture":false,"null_last_reviewed_at":null,"public_change_summary":"First indexed by Null; source disclosure date is shown separately.","null_first_observed_at":"2026-11-04T09:59:26.779Z"}],"next_cursor":null,"snapshot_revision":"42594"}