CVE-2021-37679 (GCVE-0-2021-37679)

Vulnerability from cvelistv5 – Published: 2021-08-12 22:20 – Updated: 2024-08-04 01:23
VLAI
Title
Heap OOB in nested `tf.map_fn` with `RaggedTensor`s in TensorFlow
Summary
TensorFlow is an end-to-end open source platform for machine learning. In affected versions it is possible to nest a `tf.map_fn` within another `tf.map_fn` call. However, if the input tensor is a `RaggedTensor` and there is no function signature provided, code assumes the output is a fully specified tensor and fills output buffer with uninitialized contents from the heap. The `t` and `z` outputs should be identical, however this is not the case. The last row of `t` contains data from the heap which can be used to leak other memory information. The bug lies in the conversion from a `Variant` tensor to a `RaggedTensor`. The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/core/kernels/ragged_tensor_from_variant_op.cc#L177-L190) does not check that all inner shapes match and this results in the additional dimensions. The same implementation can result in data loss, if input tensor is tweaked. We have patched the issue in GitHub commit 4e2565483d0ffcadc719bd44893fb7f609bb5f12. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range.
Severity
7.1 (High)
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N
CWE
Assigner
GitHub_M CNA under the mitre root
CNA scorecard C 72/100 over 7958 records in the last 180 days details
References
Impacted products
Vendor Product Version CPE status
tensorflow tensorflow Affected: >= 2.5.0, < 2.5.1
Affected: >= 2.4.0, < 2.4.3
Affected: < 2.3.4
guessed Create a notification for this product.
Show details on NVD website
To clipboard 

{
  "containers": {
    "adp": [
      {
        "providerMetadata": {
          "dateUpdated": "2024-08-04T01:23:01.507Z",
          "orgId": "af854a3a-2127-422b-91ae-364da2661108",
          "shortName": "CVE"
        },
        "references": [
          {
            "tags": [
              "x_refsource_CONFIRM",
              "x_transferred"
            ],
            "url": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-g8wg-cjwc-xhhp"
          },
          {
            "tags": [
              "x_refsource_MISC",
              "x_transferred"
            ],
            "url": "https://github.com/tensorflow/tensorflow/commit/4e2565483d0ffcadc719bd44893fb7f609bb5f12"
          }
        ],
        "title": "CVE Program Container"
      }
    ],
    "cna": {
      "affected": [
        {
          "product": "tensorflow",
          "vendor": "tensorflow",
          "versions": [
            {
              "status": "affected",
              "version": "\u003e= 2.5.0, \u003c 2.5.1"
            },
            {
              "status": "affected",
              "version": "\u003e= 2.4.0, \u003c 2.4.3"
            },
            {
              "status": "affected",
              "version": "\u003c 2.3.4"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "TensorFlow is an end-to-end open source platform for machine learning. In affected versions it is possible to nest a `tf.map_fn` within another `tf.map_fn` call. However, if the input tensor is a `RaggedTensor` and there is no function signature provided, code assumes the output is a fully specified tensor and fills output buffer with uninitialized contents from the heap. The `t` and `z` outputs should be identical, however this is not the case. The last row of `t` contains data from the heap which can be used to leak other memory information. The bug lies in the conversion from a `Variant` tensor to a `RaggedTensor`. The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/core/kernels/ragged_tensor_from_variant_op.cc#L177-L190) does not check that all inner shapes match and this results in the additional dimensions. The same implementation can result in data loss, if input tensor is tweaked. We have patched the issue in GitHub commit 4e2565483d0ffcadc719bd44893fb7f609bb5f12. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range."
        }
      ],
      "metrics": [
        {
          "cvssV3_1": {
            "attackComplexity": "LOW",
            "attackVector": "LOCAL",
            "availabilityImpact": "NONE",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "confidentialityImpact": "HIGH",
            "integrityImpact": "HIGH",
            "privilegesRequired": "LOW",
            "scope": "UNCHANGED",
            "userInteraction": "NONE",
            "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N",
            "version": "3.1"
          }
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-125",
              "description": "CWE-125: Out-of-bounds Read",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2021-08-12T22:20:16.000Z",
        "orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
        "shortName": "GitHub_M"
      },
      "references": [
        {
          "tags": [
            "x_refsource_CONFIRM"
          ],
          "url": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-g8wg-cjwc-xhhp"
        },
        {
          "tags": [
            "x_refsource_MISC"
          ],
          "url": "https://github.com/tensorflow/tensorflow/commit/4e2565483d0ffcadc719bd44893fb7f609bb5f12"
        }
      ],
      "source": {
        "advisory": "GHSA-g8wg-cjwc-xhhp",
        "discovery": "UNKNOWN"
      },
      "title": "Heap OOB in nested `tf.map_fn` with `RaggedTensor`s in TensorFlow",
      "x_legacyV4Record": {
        "CVE_data_meta": {
          "ASSIGNER": "security-advisories@github.com",
          "ID": "CVE-2021-37679",
          "STATE": "PUBLIC",
          "TITLE": "Heap OOB in nested `tf.map_fn` with `RaggedTensor`s in TensorFlow"
        },
        "affects": {
          "vendor": {
            "vendor_data": [
              {
                "product": {
                  "product_data": [
                    {
                      "product_name": "tensorflow",
                      "version": {
                        "version_data": [
                          {
                            "version_value": "\u003e= 2.5.0, \u003c 2.5.1"
                          },
                          {
                            "version_value": "\u003e= 2.4.0, \u003c 2.4.3"
                          },
                          {
                            "version_value": "\u003c 2.3.4"
                          }
                        ]
                      }
                    }
                  ]
                },
                "vendor_name": "tensorflow"
              }
            ]
          }
        },
        "data_format": "MITRE",
        "data_type": "CVE",
        "data_version": "4.0",
        "description": {
          "description_data": [
            {
              "lang": "eng",
              "value": "TensorFlow is an end-to-end open source platform for machine learning. In affected versions it is possible to nest a `tf.map_fn` within another `tf.map_fn` call. However, if the input tensor is a `RaggedTensor` and there is no function signature provided, code assumes the output is a fully specified tensor and fills output buffer with uninitialized contents from the heap. The `t` and `z` outputs should be identical, however this is not the case. The last row of `t` contains data from the heap which can be used to leak other memory information. The bug lies in the conversion from a `Variant` tensor to a `RaggedTensor`. The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/core/kernels/ragged_tensor_from_variant_op.cc#L177-L190) does not check that all inner shapes match and this results in the additional dimensions. The same implementation can result in data loss, if input tensor is tweaked. We have patched the issue in GitHub commit 4e2565483d0ffcadc719bd44893fb7f609bb5f12. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range."
            }
          ]
        },
        "impact": {
          "cvss": {
            "attackComplexity": "LOW",
            "attackVector": "LOCAL",
            "availabilityImpact": "NONE",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "confidentialityImpact": "HIGH",
            "integrityImpact": "HIGH",
            "privilegesRequired": "LOW",
            "scope": "UNCHANGED",
            "userInteraction": "NONE",
            "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N",
            "version": "3.1"
          }
        },
        "problemtype": {
          "problemtype_data": [
            {
              "description": [
                {
                  "lang": "eng",
                  "value": "CWE-125: Out-of-bounds Read"
                }
              ]
            }
          ]
        },
        "references": {
          "reference_data": [
            {
              "name": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-g8wg-cjwc-xhhp",
              "refsource": "CONFIRM",
              "url": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-g8wg-cjwc-xhhp"
            },
            {
              "name": "https://github.com/tensorflow/tensorflow/commit/4e2565483d0ffcadc719bd44893fb7f609bb5f12",
              "refsource": "MISC",
              "url": "https://github.com/tensorflow/tensorflow/commit/4e2565483d0ffcadc719bd44893fb7f609bb5f12"
            }
          ]
        },
        "source": {
          "advisory": "GHSA-g8wg-cjwc-xhhp",
          "discovery": "UNKNOWN"
        }
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
    "assignerShortName": "GitHub_M",
    "cveId": "CVE-2021-37679",
    "datePublished": "2021-08-12T22:20:16.000Z",
    "dateReserved": "2021-07-29T00:00:00.000Z",
    "dateUpdated": "2024-08-04T01:23:01.507Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.1",
  "vulnerability-lookup:meta": {
    "epss": {
      "cve": "CVE-2021-37679",
      "date": "2026-11-08",
      "epss": "0.00181",
      "percentile": "0.07736"
    },
    "nvd": {
      "cve": {
        "configurations": [
          {
            "nodes": [
              {
                "cpeMatch": [
                  {
                    "criteria": "cpe:2.3:a:google:tensorflow:*:*:*:*:*:*:*:*",
                    "matchCriteriaId": "0F83C081-51CC-415F-A8C0-0A44C75E2CD6",
                    "versionEndExcluding": "2.3.4",
                    "versionStartIncluding": "2.3.0",
                    "vulnerable": true
                  },
                  {
                    "criteria": "cpe:2.3:a:google:tensorflow:*:*:*:*:*:*:*:*",
                    "matchCriteriaId": "BD3F2BF8-EBA9-42BF-8F9B-D918B880B15A",
                    "versionEndExcluding": "2.4.3",
                    "versionStartIncluding": "2.4.0",
                    "vulnerable": true
                  },
                  {
                    "criteria": "cpe:2.3:a:google:tensorflow:2.5.0:*:*:*:*:*:*:*",
                    "matchCriteriaId": "D03E99A7-4E3D-427D-A156-C0713E9FB02A",
                    "vulnerable": true
                  },
                  {
                    "criteria": "cpe:2.3:a:google:tensorflow:2.6.0:rc0:*:*:*:*:*:*",
                    "matchCriteriaId": "70FA6E48-6C57-40CA-809F-4E3D07CBF348",
                    "vulnerable": true
                  },
                  {
                    "criteria": "cpe:2.3:a:google:tensorflow:2.6.0:rc1:*:*:*:*:*:*",
                    "matchCriteriaId": "42187561-E491-434D-828C-F36701446634",
                    "vulnerable": true
                  },
                  {
                    "criteria": "cpe:2.3:a:google:tensorflow:2.6.0:rc2:*:*:*:*:*:*",
                    "matchCriteriaId": "C66B61C8-450A-4C5E-9174-F970D6DEE778",
                    "vulnerable": true
                  }
                ],
                "negate": false,
                "operator": "OR"
              }
            ]
          }
        ],
        "cveTags": [],
        "descriptions": [
          {
            "lang": "en",
            "value": "TensorFlow is an end-to-end open source platform for machine learning. In affected versions it is possible to nest a `tf.map_fn` within another `tf.map_fn` call. However, if the input tensor is a `RaggedTensor` and there is no function signature provided, code assumes the output is a fully specified tensor and fills output buffer with uninitialized contents from the heap. The `t` and `z` outputs should be identical, however this is not the case. The last row of `t` contains data from the heap which can be used to leak other memory information. The bug lies in the conversion from a `Variant` tensor to a `RaggedTensor`. The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/core/kernels/ragged_tensor_from_variant_op.cc#L177-L190) does not check that all inner shapes match and this results in the additional dimensions. The same implementation can result in data loss, if input tensor is tweaked. We have patched the issue in GitHub commit 4e2565483d0ffcadc719bd44893fb7f609bb5f12. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range."
          },
          {
            "lang": "es",
            "value": "TensorFlow es una plataforma de c\u00f3digo abierto de extremo a extremo para el aprendizaje autom\u00e1tico.\u0026#xa0;En las versiones afectadas, es posible anidar un \"tf.map_fn\" dentro de otra llamada \"tf.map_fn\".\u0026#xa0;Sin embargo, si el tensor de entrada es un \"RaggedTensor\" y no se proporciona una firma de funci\u00f3n, el c\u00f3digo asume que la salida es un tensor completamente especificado y llena el b\u00fafer de salida con contenido no inicializado de la pila.\u0026#xa0;Las salidas \"t\" y\" z\" deben ser id\u00e9nticas, sin embargo, este no es el caso.\u0026#xa0;La \u00faltima fila de \"t\" contiene datos de la pila que se pueden usar para filtrar otra informaci\u00f3n de la memoria.\u0026#xa0;El error radica en la conversi\u00f3n de un tensor \"Variant\" a un\" RaggedTensor\".\u0026#xa0;La [implementaci\u00f3n] (https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/core/kernels/ragged_tensor_from_variant_op.\u0026#xa0;cc # L177-L190) no comprueba que todas las formas internas coincidan y esto da como resultado las dimensiones adicionales.\u0026#xa0;La misma implementaci\u00f3n puede resultar en la p\u00e9rdida de datos, si se modifica el tensor de entrada.\u0026#xa0;Hemos solucionado el problema en GitHub commit 4e2565483d0ffcadc719bd44893fb7f609bb5f12.\u0026#xa0;La correcci\u00f3n ser\u00e1 incluida en TensorFlow versi\u00f3n 2.6.0.\u0026#xa0;Tambi\u00e9n seleccionaremos este commit en TensorFlow versi\u00f3n 2.5.1, TensorFlow versi\u00f3n 2.4.3 y TensorFlow versi\u00f3n 2.3.4, ya que estos tambi\u00e9n est\u00e1n afectados y a\u00fan se encuentran en el rango admitido."
          }
        ],
        "id": "CVE-2021-37679",
        "lastModified": "2024-11-21T06:15:41.070",
        "metrics": {
          "cvssMetricV2": [
            {
              "acInsufInfo": false,
              "baseSeverity": "MEDIUM",
              "cvssData": {
                "accessComplexity": "LOW",
                "accessVector": "LOCAL",
                "authentication": "NONE",
                "availabilityImpact": "PARTIAL",
                "baseScore": 4.6,
                "confidentialityImpact": "PARTIAL",
                "integrityImpact": "PARTIAL",
                "vectorString": "AV:L/AC:L/Au:N/C:P/I:P/A:P",
                "version": "2.0"
              },
              "exploitabilityScore": 3.9,
              "impactScore": 6.4,
              "obtainAllPrivilege": false,
              "obtainOtherPrivilege": false,
              "obtainUserPrivilege": false,
              "source": "nvd@nist.gov",
              "type": "Primary",
              "userInteractionRequired": false
            }
          ],
          "cvssMetricV31": [
            {
              "cvssData": {
                "attackComplexity": "LOW",
                "attackVector": "LOCAL",
                "availabilityImpact": "NONE",
                "baseScore": 7.1,
                "baseSeverity": "HIGH",
                "confidentialityImpact": "HIGH",
                "integrityImpact": "HIGH",
                "privilegesRequired": "LOW",
                "scope": "UNCHANGED",
                "userInteraction": "NONE",
                "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N",
                "version": "3.1"
              },
              "exploitabilityScore": 1.8,
              "impactScore": 5.2,
              "source": "security-advisories@github.com",
              "type": "Secondary"
            },
            {
              "cvssData": {
                "attackComplexity": "LOW",
                "attackVector": "LOCAL",
                "availabilityImpact": "HIGH",
                "baseScore": 7.8,
                "baseSeverity": "HIGH",
                "confidentialityImpact": "HIGH",
                "integrityImpact": "HIGH",
                "privilegesRequired": "LOW",
                "scope": "UNCHANGED",
                "userInteraction": "NONE",
                "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
                "version": "3.1"
              },
              "exploitabilityScore": 1.8,
              "impactScore": 5.9,
              "source": "nvd@nist.gov",
              "type": "Primary"
            }
          ]
        },
        "published": "2021-08-12T23:15:08.287",
        "references": [
          {
            "source": "security-advisories@github.com",
            "tags": [
              "Patch",
              "Third Party Advisory"
            ],
            "url": "https://github.com/tensorflow/tensorflow/commit/4e2565483d0ffcadc719bd44893fb7f609bb5f12"
          },
          {
            "source": "security-advisories@github.com",
            "tags": [
              "Third Party Advisory"
            ],
            "url": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-g8wg-cjwc-xhhp"
          },
          {
            "source": "af854a3a-2127-422b-91ae-364da2661108",
            "tags": [
              "Patch",
              "Third Party Advisory"
            ],
            "url": "https://github.com/tensorflow/tensorflow/commit/4e2565483d0ffcadc719bd44893fb7f609bb5f12"
          },
          {
            "source": "af854a3a-2127-422b-91ae-364da2661108",
            "tags": [
              "Third Party Advisory"
            ],
            "url": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-g8wg-cjwc-xhhp"
          }
        ],
        "sourceIdentifier": "security-advisories@github.com",
        "vulnStatus": "Modified",
        "weaknesses": [
          {
            "description": [
              {
                "lang": "en",
                "value": "CWE-125"
              }
            ],
            "source": "security-advisories@github.com",
            "type": "Secondary"
          },
          {
            "description": [
              {
                "lang": "en",
                "value": "CWE-681"
              }
            ],
            "source": "nvd@nist.gov",
            "type": "Primary"
          }
        ]
      }
    },
    "redhat_vex": {
      "current_release_date": "2026-03-27T13:44:52+00:00",
      "cve": "CVE-2021-37679",
      "id": "CVE-2021-37679",
      "initial_release_date": "2021-01-01T00:00:00+00:00",
      "product_status:known_not_affected": "1",
      "source": "Red Hat CSAF VEX",
      "status": "final",
      "title": "Heap OOB in nested `tf.map_fn` with `RaggedTensor`s in TensorFlow",
      "url": "https://security.access.redhat.com/data/csaf/v2/vex/2021/cve-2021-37679.json",
      "version": "3"
    }
  }
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.




Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Vulnerability timeline

Sightings and recorded dates, grouped by day · UTC

JSON
How durations are calculated

Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.

Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.

Loading timeline…




    Related by attack behaviour

    Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.