CVE-2022-21731: Tensorflow Type Confusion Leading To Segfault Vulnerability

HIGH
Advisory Identifier: GHSA-m4hf-j54p-p353
Share
CVSS 6.5
CWE-843 CWE-754
Published
Advisory Summary

CVE-2022-21731 is a HIGH (CVSS 6.5) type confusion leading to segfault vulnerability (CWE-843: Access of Resource Using Incompatible Type ('Type Confusion')) affecting Tensorflow libraries in the PyPI ecosystem. The issue involves type confusion leading to segfault in tensorflow. Remediation is available by upgrading affected packages to version 2.5.3, 2.6.3, 2.7.1.

CVE-2022-21731 at a Glance

CVE Identifier CVE-2022-21731
GHSA Identifier GHSA-m4hf-j54p-p353
Severity HIGH
CVSS v3.1 Base Score 6.5
CWE Classification CWE-843: Access of Resource Using Incompatible Type ('Type Confusion')
Affected Product Tensorflow
Package Ecosystem PyPI
Published Date
Last Updated

Security Impact

CVSS: 6.5

The security impact metrics below are parsed directly from the CVSS v3.1 vector string (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H).

Attack Vector Network (AV:N)
Attack Complexity Low (AC:L)
Privileges Required Low (PR:L)
User Interaction None (UI:N)
Scope Unchanged (S:U)
Confidentiality None (C:N)
Integrity None (I:N)
Availability High (A:H)

How to Fix CVE-2022-21731

To remediate this vulnerability, upgrade all affected Tensorflow dependencies to a patched release. Both direct and transitive dependencies should be checked, as frameworks or security libraries may pull in vulnerable versions transitively.

tensorflow fixed: 2.5.3 
pip install tensorflow>=2.5.3
tensorflow fixed: 2.6.3 
pip install tensorflow>=2.6.3
tensorflow fixed: 2.7.1 
pip install tensorflow>=2.7.1
tensorflow-cpu fixed: 2.5.3 
pip install tensorflow-cpu>=2.5.3
tensorflow-cpu fixed: 2.6.3 
pip install tensorflow-cpu>=2.6.3
tensorflow-cpu fixed: 2.7.1 
pip install tensorflow-cpu>=2.7.1
tensorflow-gpu fixed: 2.5.3 
pip install tensorflow-gpu>=2.5.3
tensorflow-gpu fixed: 2.6.3 
pip install tensorflow-gpu>=2.6.3
tensorflow-gpu fixed: 2.7.1 
pip install tensorflow-gpu>=2.7.1
Verify Direct and Transitive Dependencies

Run the following command in your project directory to verify that all resolved dependencies use the patched release: 

pip show tensorflow

How to Check if You Are Affected

To determine whether your Python environment is affected by CVE-2022-21731, inspect your installed packages using pip:

pip show tensorflow

Analyze the terminal output to locate any version of Tensorflow in your dependency tree. If any resolved version matches the vulnerable ranges listed in the table below (for example, versions earlier than 2.5.3), your application is affected.

Note: The pip show tensorflow command checks currently installed packages in your active Python environment.

Technical Details

Vulnerability Mechanism & Technical Analysis

This vulnerability is classified under CWE-843: Access of Resource Using Incompatible Type ('Type Confusion'), where the product allocates or initializes a resource such as a pointer, object, or variable using one type, but it later accesses that resource using a type that is incompatible with the original type. In Tensorflow, software state validation or input boundaries are circumvented during processing. Review the official source advisory statement below for specific affected components, classes, and reproduction details.

Source Advisory Statement

### Impact The [implementation of shape inference for `ConcatV2`](https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/common_shape_fns.cc#L1961-L2059) can be used to trigger a denial of service attack via a segfault caused by a type confusion: ```python import tensorflow as tf @tf.function def test(): y = tf.raw_ops.ConcatV2( values=[[1,2,3],[4,5,6]], axis = 0xb500005b) return y test() ``` The `axis` argument is translated into `concat_dim` in the `ConcatShapeHelper` helper function. Then, a value for `min_rank` is computed based on `concat_dim`. This is then used to validate that the `values` tensor has at least the required rank: ```cc int64_t concat_dim; if (concat_dim_t->dtype() == DT_INT32) { concat_dim = static_cast<int64_t>(concat_dim_t->flat<int32>()(0)); } else { concat_dim = concat_dim_t->flat<int64_t>()(0); } // Minimum required number of dimensions. const int min_rank = concat_dim < 0 ? -concat_dim : concat_dim + 1; // ... ShapeHandle input = c->input(end_value_index - 1); TF_RETURN_IF_ERROR(c->WithRankAtLeast(input, min_rank, &input)); ``` However, [`WithRankAtLeast`](https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/shape_inference.cc#L345-L358) receives the lower bound as a 64-bits value and then compares it against the maximum 32-bits integer value that could be represented: ```cc Status InferenceContext::WithRankAtLeast(ShapeHandle shape, int64_t rank, ShapeHandle* out) { if (rank > kint32max) { return errors::InvalidArgument("Rank cannot exceed kint32max"); } // ... } ``` Due to the fact that `min_rank` is a 32-bits value and the value of `axis`, the `rank` argument is a [negative value](https://godbolt.org/z/Gcr5haMob), so the error check is bypassed. ### Patches We have patched the issue in GitHub commit [08d7b00c0a5a20926363849f611729f53f3ec022](https://github.com/tensorflow/tensorflow/commit/08d7b00c0a5a20926363849f611729f53f3ec022). The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range. ### For more information Please 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. ### Attribution This vulnerability has been reported by Yu Tian of Qihoo 360 AIVul Team.

CVE-2022-21731 Affected Tensorflow Versions

9 component(s)
Fixed Versions:
2.5.32.6.32.7.1
Product / Package Ecosystem Affected Versions Fixed Version
tensorflow
pypi/tensorflow
PyPI
< 2.5.3
2.5.3
tensorflow
pypi/tensorflow
PyPI
≥ 2.6.0
< 2.6.3
2.6.3
tensorflow
pypi/tensorflow
PyPI
≥ 2.7.0
< 2.7.1
All 1 versions
2.7.0
2.7.1
tensorflow-cpu
pypi/tensorflow-cpu
PyPI
< 2.5.3
2.5.3
tensorflow-cpu
pypi/tensorflow-cpu
PyPI
≥ 2.6.0
< 2.6.3
2.6.3
tensorflow-cpu
pypi/tensorflow-cpu
PyPI
≥ 2.7.0
< 2.7.1
All 1 versions
2.7.0
2.7.1
tensorflow-gpu
pypi/tensorflow-gpu
PyPI
< 2.5.3
2.5.3
tensorflow-gpu
pypi/tensorflow-gpu
PyPI
≥ 2.6.0
< 2.6.3
2.6.3
tensorflow-gpu
pypi/tensorflow-gpu
PyPI
≥ 2.7.0
< 2.7.1
All 1 versions
2.7.0
2.7.1

References

Tensorflow Source Repository REPOSITORY
https://github.com/tensorflow/tensorflow
GitHub Advisory / Wiki ADVISORY
https://github.com/tensorflow/tensorflow/security/advisories/GHSA-m4hf-j54p-p353
NVD — CVE-2022-21731 NVD
https://nvd.nist.gov/vuln/detail/CVE-2022-21731
Security Fix Commit FIX COMMIT
https://github.com/tensorflow/tensorflow/commit/08d7b00c0a5a20926363849f611729f53f3ec022
Tensorflow Source Repository REPOSITORY
https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2022-55.yaml
Tensorflow Source Repository REPOSITORY
https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2022-110.yaml
Tensorflow Source Repository REPOSITORY
https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/common_shape_fns.cc#L1961-L2059
Tensorflow Source Repository REPOSITORY
https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/shape_inference.cc#L345-L358

Frequently Asked Questions

What is CVE-2022-21731?
CVE-2022-21731 is a security vulnerability in Tensorflow involving Type confusion leading to segfault in Tensorflow under CWE-843: Access of Resource Using Incompatible Type ('Type Confusion').
Which Tensorflow versions are affected?
Affected versions include releases prior to 2.5.3, 2.6.3, 2.7.1 across affected packages: tensorflow, tensorflow-cpu, tensorflow-gpu.
What version fixes CVE-2022-21731?
Upgrading to 2.5.3, 2.6.3, 2.7.1 (or the latest patched release for your respective branch) resolves the vulnerability.
Is tensorflow affected?
Yes, tensorflow is affected in releases prior to 2.5.3, 2.6.3, 2.7.1. Upgrade your dependency to 2.5.3, 2.6.3, 2.7.1 or later.
Is tensorflow-cpu affected?
Yes, tensorflow-cpu is affected in releases prior to 2.5.3, 2.6.3, 2.7.1. Upgrade your dependency to 2.5.3, 2.6.3, 2.7.1 or later.
What is the CVSS score of CVE-2022-21731?
CVE-2022-21731 has a CVSS v3.1 base score of 6.5 (HIGH) with vector string CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H.
What CWE is associated with CVE-2022-21731?
CVE-2022-21731 is classified under CWE-843: Access of Resource Using Incompatible Type ('Type Confusion') (The product allocates or initializes a resource such as a pointer, object, or variable using one type, but it later accesses that resource using a type that is incompatible with the original type.).
How can I check PyPI dependencies for CVE-2022-21731?
Run the dependency tree command: 'pip show tensorflow'. This command lists direct and transitive dependencies in your build configuration so you can verify if a vulnerable version is resolved.
Machine-readable OSV JSON OSV v1.5 Schema
Normalized schema output
View Raw JSON 
{
  "schema_version" : "1.4.0",
  "id" : "GHSA-m4hf-j54p-p353",
  "modified" : "2024-11-13T22:12:12Z",
  "published" : "2022-02-10T00:19:50Z",
  "aliases" : [ "CVE-2022-21731" ],
  "summary" : "Type confusion leading to segfault in Tensorflow",
  "details" : "### Impact \nThe [implementation of shape inference for `ConcatV2`](https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/common_shape_fns.cc#L1961-L2059) can be used to trigger a denial of service attack via a segfault caused by a type confusion:\n\n```python\nimport tensorflow as tf\n\n@tf.function\ndef test():\n  y = tf.raw_ops.ConcatV2(\n    values=[[1,2,3],[4,5,6]],\n    axis = 0xb500005b)\n  return y\n\ntest()\n```\n\nThe `axis` argument is translated into `concat_dim` in the `ConcatShapeHelper` helper function. Then, a value for `min_rank` is computed based on `concat_dim`. This is then used to validate that the `values` tensor has at least the required rank:\n\n```cc\n  int64_t concat_dim;\n  if (concat_dim_t->dtype() == DT_INT32) {\n    concat_dim = static_cast<int64_t>(concat_dim_t->flat<int32>()(0));\n  } else {\n    concat_dim = concat_dim_t->flat<int64_t>()(0);\n  }\n\n  // Minimum required number of dimensions.\n  const int min_rank = concat_dim < 0 ? -concat_dim : concat_dim + 1;\n\n  // ...\n  ShapeHandle input = c->input(end_value_index - 1);\n  TF_RETURN_IF_ERROR(c->WithRankAtLeast(input, min_rank, &input));\n```\n\nHowever, [`WithRankAtLeast`](https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/shape_inference.cc#L345-L358) receives the lower bound as a 64-bits value and then compares it against the maximum 32-bits integer value that could be represented:\n\n```cc\nStatus InferenceContext::WithRankAtLeast(ShapeHandle shape, int64_t rank,\n                                         ShapeHandle* out) {\n  if (rank > kint32max) {\n    return errors::InvalidArgument(\"Rank cannot exceed kint32max\");\n  }\n  // ...\n}\n```\n\nDue to the fact that `min_rank` is a 32-bits value and the value of `axis`, the `rank` argument is a [negative value](https://godbolt.org/z/Gcr5haMob), so the error check is bypassed.\n\n### Patches\nWe have patched the issue in GitHub commit [08d7b00c0a5a20926363849f611729f53f3ec022](https://github.com/tensorflow/tensorflow/commit/08d7b00c0a5a20926363849f611729f53f3ec022).\n\nThe fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.\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 Yu Tian of Qihoo 360 AIVul Team.",
  "severity" : [ {
    "type" : "CVSS_V3",
    "score" : "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H"
  }, {
    "type" : "CVSS_V4",
    "score" : "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N"
  } ],
  "affected" : [ {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "0"
      }, {
        "fixed" : "2.5.3"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "2.6.0"
      }, {
        "fixed" : "2.6.3"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "2.7.0"
      }, {
        "fixed" : "2.7.1"
      } ]
    } ],
    "versions" : [ "2.7.0" ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow-cpu"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "0"
      }, {
        "fixed" : "2.5.3"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow-cpu"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "2.6.0"
      }, {
        "fixed" : "2.6.3"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow-cpu"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "2.7.0"
      }, {
        "fixed" : "2.7.1"
      } ]
    } ],
    "versions" : [ "2.7.0" ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow-gpu"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "0"
      }, {
        "fixed" : "2.5.3"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow-gpu"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "2.6.0"
      }, {
        "fixed" : "2.6.3"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "tensorflow-gpu"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "2.7.0"
      }, {
        "fixed" : "2.7.1"
      } ]
    } ],
    "versions" : [ "2.7.0" ]
  } ],
  "references" : [ {
    "type" : "WEB",
    "url" : "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-m4hf-j54p-p353"
  }, {
    "type" : "ADVISORY",
    "url" : "https://nvd.nist.gov/vuln/detail/CVE-2022-21731"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/tensorflow/tensorflow/commit/08d7b00c0a5a20926363849f611729f53f3ec022"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2022-55.yaml"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2022-110.yaml"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/tensorflow/tensorflow"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/common_shape_fns.cc#L1961-L2059"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/tensorflow/tensorflow/blob/5100e359aef5c8021f2e71c7b986420b85ce7b3d/tensorflow/core/framework/shape_inference.cc#L345-L358"
  } ],
  "database_specific" : {
    "cwe_ids" : [ "CWE-754", "CWE-843" ],
    "severity" : "HIGH",
    "github_reviewed" : true,
    "github_reviewed_at" : "2022-02-03T19:01:09Z",
    "nvd_published_at" : "2022-02-03T12:15:00Z"
  }
}