Vulnerability Name:

CVE-2021-37679 (CCN-207471)

Assigned:2021-08-11
Published:2021-08-11
Updated:2021-08-19
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.
CVSS v3 Severity:7.8 High (CVSS v3.1 Vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H)
6.8 Medium (Temporal CVSS v3.1 Vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H/E:U/RL:O/RC:C)
Exploitability Metrics:Attack Vector (AV): Local
Attack Complexity (AC): Low
Privileges Required (PR): Low
User Interaction (UI): None
Scope:Scope (S): Unchanged
Impact Metrics:Confidentiality (C): High
Integrity (I): High
Availibility (A): High
7.1 High (CCN CVSS v3.1 Vector: CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N)
6.2 Medium (CCN Temporal CVSS v3.1 Vector: CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N/E:U/RL:O/RC:C)
Exploitability Metrics:Attack Vector (AV): Local
Attack Complexity (AC): Low
Privileges Required (PR): None
User Interaction (UI): Required
Scope:Scope (S): Unchanged
Impact Metrics:Confidentiality (C): High
Integrity (I): High
Availibility (A): None
CVSS v2 Severity:4.6 Medium (CVSS v2 Vector: AV:L/AC:L/Au:N/C:P/I:P/A:P)
Exploitability Metrics:Access Vector (AV): Local
Access Complexity (AC): Low
Authentication (Au): None
Impact Metrics:Confidentiality (C): Partial
Integrity (I): Partial
Availibility (A): Partial
6.2 Medium (CCN CVSS v2 Vector: AV:L/AC:L/Au:S/C:C/I:C/A:N)
Exploitability Metrics:Access Vector (AV): Local
Access Complexity (AC): Low
Athentication (Au): Single_Instance
Impact Metrics:Confidentiality (C): Complete
Integrity (I): Complete
Availibility (A): None
Vulnerability Type:CWE-681
Vulnerability Consequences:Obtain Information
References:Source: MITRE
Type: CNA
CVE-2021-37679

Source: XF
Type: UNKNOWN
tensorflow-cve202137679-info-disc(207471)

Source: MISC
Type: Patch, Third Party Advisory
https://github.com/tensorflow/tensorflow/commit/4e2565483d0ffcadc719bd44893fb7f609bb5f12

Source: CCN
Type: TensorFlow GIT Repository
Heap OOB in nested tf.map_fn with RaggedTensor`s

Source: CONFIRM
Type: Third Party Advisory
https://github.com/tensorflow/tensorflow/security/advisories/GHSA-g8wg-cjwc-xhhp

Source: CCN
Type: IBM Security Bulletin 6523816 (Watson Discovery)
IBM Watson Discovery for IBM Cloud Pak for Data affected by vulnerability in TensorFlow

Vulnerable Configuration:Configuration 1:
  • cpe:/a:google:tensorflow:*:*:*:*:*:*:*:* (Version >= 2.3.0 and < 2.3.4)
  • OR cpe:/a:google:tensorflow:*:*:*:*:*:*:*:* (Version >= 2.4.0 and < 2.4.3)
  • OR cpe:/a:google:tensorflow:2.5.0:*:*:*:*:*:*:*
  • OR cpe:/a:google:tensorflow:2.6.0:rc0:*:*:*:*:*:*
  • OR cpe:/a:google:tensorflow:2.6.0:rc1:*:*:*:*:*:*
  • OR cpe:/a:google:tensorflow:2.6.0:rc2:*:*:*:*:*:*

  • Configuration CCN 1:
  • cpe:/a:google:tensorflow:-:*:*:*:-:*:*:*
  • AND
  • cpe:/a:ibm:watson_discovery:2.0.0:*:*:*:*:*:*:*
  • OR cpe:/a:ibm:watson_discovery:2.2.1:*:*:*:*:*:*:*

  • * Denotes that component is vulnerable
    BACK
    google tensorflow *
    google tensorflow *
    google tensorflow 2.5.0
    google tensorflow 2.6.0 rc0
    google tensorflow 2.6.0 rc1
    google tensorflow 2.6.0 rc2
    tensorflow tensorflow -
    ibm watson discovery 2.0.0
    ibm watson discovery 2.2.1