CVE-2021-29521

TensorFlow is an end-to-end open source platform for machine learning. Specifying a negative dense shape in `tf.raw_ops.SparseCountSparseOutput` results in a segmentation fault being thrown out from the standard library as `std::vector` invariants are broken. This is because the implementation(https://github.com/tensorflow/tensorflow/blob/8f7b60ee8c0206a2c99802e3a4d1bb55d2bc0624/tensorflow/core/kernels/count_ops.cc#L199-L213) assumes the first element of the dense shape is always positive and uses it to initialize a `BatchedMap<T>` (i.e., `std::vector<absl::flat_hash_map<int64,T>>`(https://github.com/tensorflow/tensorflow/blob/8f7b60ee8c0206a2c99802e3a4d1bb55d2bc0624/tensorflow/core/kernels/count_ops.cc#L27)) data structure. If the `shape` tensor has more than one element, `num_batches` is the first value in `shape`. Ensuring that the `dense_shape` argument is a valid tensor shape (that is, all elements are non-negative) solves this issue. The fix will be included in TensorFlow 2.5.0. We will also cherrypick this commit on TensorFlow 2.4.2 and TensorFlow 2.3.3.

CVSS v3 5.5 MEDIUM
CVSS v2.0 2.1 LOW

5.5^/10

CVSS v3 : MEDIUM

Vector :

Exploitability : 1.8 / Impact : 3.6

Attack Vector LOCAL

Attack Complexity LOW

Privileges Required LOW

User Interaction NONE

Confidentiality Impact NONE

Integrity Impact NONE

Availability Impact HIGH

Scope UNCHANGED

2.1^/10

CVSS v2.0 : LOW

Vector : AV:L/AC:L/Au:N/C:N/I:N/A:P

Exploitability : 3.9 / Impact : 2.9

Access Vector LOCAL

Access Complexity LOW

Authentication NONE

Confidentiality Impact NONE

Integrity Impact NONE

Availability Impact PARTIAL

References

Link	Resource
https://github.com/tensorflow/tensorflow/commit/c57c0b9f3a4f8684f3489dd9a9ec627ad8b599f5	Patch Third Party Advisory
https://github.com/tensorflow/tensorflow/security/advisories/GHSA-hr84-fqvp-48mm	Exploit Patch Third Party Advisory

Configurations

Configuration 1 (hide)

OR	cpe:2.3:a:google:tensorflow::::::::
	cpe:2.3:a:google:tensorflow::::::::

History

No history.

Information

Published : 2021-05-14 20:15

Updated : 2024-02-28 18:28

NVD link : CVE-2021-29521

Mitre link : CVE-2021-29521

CVE.ORG link : CVE-2021-29521

JSON object : View

Products Affected

google

tensorflow

CWE

CWE-131

Incorrect Calculation of Buffer Size

{"id": "CVE-2021-29521", "metrics": {"cvssMetricV2": [{"type": "Primary", "source": "nvd@nist.gov", "cvssData": {"version": "2.0", "baseScore": 2.1, "accessVector": "LOCAL", "vectorString": "AV:L/AC:L/Au:N/C:N/I:N/A:P", "authentication": "NONE", "integrityImpact": "NONE", "accessComplexity": "LOW", "availabilityImpact": "PARTIAL", "confidentialityImpact": "NONE"}, "acInsufInfo": false, "impactScore": 2.9, "baseSeverity": "LOW", "obtainAllPrivilege": false, "exploitabilityScore": 3.9, "obtainUserPrivilege": false, "obtainOtherPrivilege": false, "userInteractionRequired": false}], "cvssMetricV31": [{"type": "Primary", "source": "nvd@nist.gov", "cvssData": {"scope": "UNCHANGED", "version": "3.1", "baseScore": 5.5, "attackVector": "LOCAL", "baseSeverity": "MEDIUM", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H", "integrityImpact": "NONE", "userInteraction": "NONE", "attackComplexity": "LOW", "availabilityImpact": "HIGH", "privilegesRequired": "LOW", "confidentialityImpact": "NONE"}, "impactScore": 3.6, "exploitabilityScore": 1.8}, {"type": "Secondary", "source": "security-advisories@github.com", "cvssData": {"scope": "UNCHANGED", "version": "3.1", "baseScore": 2.5, "attackVector": "LOCAL", "baseSeverity": "LOW", "vectorString": "CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L", "integrityImpact": "NONE", "userInteraction": "NONE", "attackComplexity": "HIGH", "availabilityImpact": "LOW", "privilegesRequired": "LOW", "confidentialityImpact": "NONE"}, "impactScore": 1.4, "exploitabilityScore": 1.0}]}, "published": "2021-05-14T20:15:11.567", "references": [{"url": "https://github.com/tensorflow/tensorflow/commit/c57c0b9f3a4f8684f3489dd9a9ec627ad8b599f5", "tags": ["Patch", "Third Party Advisory"], "source": "security-advisories@github.com"}, {"url": "https://github.com/tensorflow/tensorflow/security/advisories/GHSA-hr84-fqvp-48mm", "tags": ["Exploit", "Patch", "Third Party Advisory"], "source": "security-advisories@github.com"}], "vulnStatus": "Analyzed", "weaknesses": [{"type": "Primary", "source": "security-advisories@github.com", "description": [{"lang": "en", "value": "CWE-131"}]}], "descriptions": [{"lang": "en", "value": "TensorFlow is an end-to-end open source platform for machine learning. Specifying a negative dense shape in `tf.raw_ops.SparseCountSparseOutput` results in a segmentation fault being thrown out from the standard library as `std::vector` invariants are broken. This is because the implementation(https://github.com/tensorflow/tensorflow/blob/8f7b60ee8c0206a2c99802e3a4d1bb55d2bc0624/tensorflow/core/kernels/count_ops.cc#L199-L213) assumes the first element of the dense shape is always positive and uses it to initialize a `BatchedMap<T>` (i.e., `std::vector<absl::flat_hash_map<int64,T>>`(https://github.com/tensorflow/tensorflow/blob/8f7b60ee8c0206a2c99802e3a4d1bb55d2bc0624/tensorflow/core/kernels/count_ops.cc#L27)) data structure. If the `shape` tensor has more than one element, `num_batches` is the first value in `shape`. Ensuring that the `dense_shape` argument is a valid tensor shape (that is, all elements are non-negative) solves this issue. The fix will be included in TensorFlow 2.5.0. We will also cherrypick this commit on TensorFlow 2.4.2 and TensorFlow 2.3.3."}, {"lang": "es", "value": "TensorFlow es una plataforma de c\u00f3digo abierto de extremo a extremo para el aprendizaje autom\u00e1tico. Especificar una forma densa negativa en \"tf.raw_ops.SparseCountSparseOutput\" resulta en un error de segmentaci\u00f3n que es eliminado de la biblioteca est\u00e1ndar ya que los invariantes \"std::vector\" son rotos. Esto es debido a que la implementaci\u00f3n (https://github.com/tensorflow/tensorflow/blob/8f7b60ee8c0206a2c99802e3a4d1bb55d2bc0624/tensorflow/core/kernels/count_ops.cc#L199-L213) asume que el primer elemento de la forma densa es siempre positivo y lo usa para inicializar un \"BatchedMap(T)\" (es decir, \"std::vector (absl::flat_hash_map(int64, T))\" (https://github.com/tensorflow/tensorflow/blob/8f7b60ee8c0206a2c99802e3a4d1bb55d2bc0624/tensorflow/core/kernels/count_ops.cc#L27)) estructura de datos. Si el tensor de \"shape\" presenta m\u00e1s de un elemento,\"num_batches\" es el primer valor en \"shape\". Asegurarse de que el argumento \"dense_shape\" sea una forma de tensor v\u00e1lida (es decir, que todos los elementos no sean negativos) resuelve este problema. La correcci\u00f3n ser\u00e1 incluida en TensorFlow versi\u00f3n 2.5.0. Tambi\u00e9n seleccionaremos este commit en TensorFlow versi\u00f3n 2.4.2 y TensorFlow versi\u00f3n 2.3.3"}], "lastModified": "2021-05-20T17:19:30.450", "configurations": [{"nodes": [{"negate": false, "cpeMatch": [{"criteria": "cpe:2.3:a:google:tensorflow:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "0F83E0CF-CBF6-4C24-8683-3E7A5DC95BA9", "versionEndExcluding": "2.3.3", "versionStartIncluding": "2.3.0"}, {"criteria": "cpe:2.3:a:google:tensorflow:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "8259531B-A8AC-4F8B-B60F-B69DE4767C03", "versionEndExcluding": "2.4.2", "versionStartIncluding": "2.4.0"}], "operator": "OR"}]}], "sourceIdentifier": "security-advisories@github.com"}