Critical severityNVD Advisory· Published Mar 24, 2023· Updated Feb 19, 2025
TensorFlow vulnerable to heap out-of-buffer read in the QuantizeAndDequantize operation
CVE-2023-25668
Description
TensorFlow is an open source platform for machine learning. Attackers using Tensorflow prior to 2.12.0 or 2.11.1 can access heap memory which is not in the control of user, leading to a crash or remote code execution. The fix will be included in TensorFlow version 2.12.0 and will also cherrypick this commit on TensorFlow version 2.11.1.
Affected packages
Versions sourced from the GitHub Security Advisory.
| Package | Affected versions | Patched versions |
|---|---|---|
tensorflowPyPI | < 2.11.1 | 2.11.1 |
tensorflow-cpuPyPI | < 2.11.1 | 2.11.1 |
tensorflow-gpuPyPI | < 2.11.1 | 2.11.1 |
Affected products
1- Range: < 2.11.1
Patches
17b174a0f2e40Fix asan issue with QuantizeAndDequantizeV2/V3/V4/V4Grad shape inference functions.
2 files changed · +82 −0
tensorflow/core/ops/array_ops.cc+16 −0 modified@@ -2879,6 +2879,10 @@ REGISTER_OP("QuantizeAndDequantizeV2") axis); } else if (axis != -1) { ShapeHandle input; + if (axis >= kint32max) { + return errors::InvalidArgument( + "Axis cannot be >= kint32max value, got ", axis); + } TF_RETURN_IF_ERROR(c->WithRankAtLeast(c->input(0), axis + 1, &input)); DimensionHandle depth; TF_RETURN_IF_ERROR( @@ -2914,6 +2918,10 @@ REGISTER_OP("QuantizeAndDequantizeV4") axis); } else if (axis != -1) { ShapeHandle input; + if (axis >= kint32max) { + return errors::InvalidArgument( + "Axis cannot be >= kint32max value, got ", axis); + } TF_RETURN_IF_ERROR(c->WithRankAtLeast(c->input(0), axis + 1, &input)); DimensionHandle depth; TF_RETURN_IF_ERROR( @@ -2945,6 +2953,10 @@ REGISTER_OP("QuantizeAndDequantizeV4Grad") axis); } else if (axis != -1) { ShapeHandle input; + if (axis >= kint32max) { + return errors::InvalidArgument( + "Axis cannot be >= kint32max value, got ", axis); + } TF_RETURN_IF_ERROR(c->WithRankAtLeast(c->input(0), axis + 1, &input)); DimensionHandle depth; TF_RETURN_IF_ERROR( @@ -2981,6 +2993,10 @@ REGISTER_OP("QuantizeAndDequantizeV3") axis); } else if (axis != -1) { ShapeHandle input; + if (axis >= kint32max) { + return errors::InvalidArgument( + "Axis cannot be >= kint32max value, got ", axis); + } TF_RETURN_IF_ERROR(c->WithRankAtLeast(c->input(0), axis + 1, &input)); DimensionHandle depth; TF_RETURN_IF_ERROR(
tensorflow/python/kernel_tests/array_ops/array_ops_test.py+66 −0 modified@@ -1856,6 +1856,72 @@ def testOutOfBoundAxis(self): max_range=input_max, axis=2**31 - 1)) + @test_util.run_v2_only + def testInvalidAxis(self): + + @def_function.function + def test_quantize_and_dequantize_v2(): + gen_array_ops.quantize_and_dequantize_v2( + input=[2.5], + input_min=[1.0], + input_max=[10.0], + signed_input=True, + num_bits=1, + range_given=True, + round_mode="HALF_TO_EVEN", + narrow_range=True, + axis=0x7fffffff) + + @def_function.function + def test_quantize_and_dequantize_v3(): + gen_array_ops.quantize_and_dequantize_v3( + input=[2.5], + input_min=[1.0], + input_max=[10.0], + num_bits=1, + signed_input=True, + range_given=True, + narrow_range=True, + axis=0x7fffffff) + + @def_function.function + def test_quantize_and_dequantize_v4(): + gen_array_ops.quantize_and_dequantize_v4( + input=[2.5], + input_min=[1.0], + input_max=[10.0], + signed_input=True, + num_bits=1, + range_given=True, + round_mode="HALF_TO_EVEN", + narrow_range=True, + axis=0x7fffffff) + + @def_function.function + def test_quantize_and_dequantize_v4_grad(): + gen_array_ops.quantize_and_dequantize_v4_grad( + gradients=[2.5], + input=[2.5], + input_min=[1.0], + input_max=[10.0], + axis=0x7fffffff) + + with self.assertRaisesRegex( + ValueError, "Axis cannot be >= kint32max value, got 2147483647"): + test_quantize_and_dequantize_v2() + + with self.assertRaisesRegex( + ValueError, "Axis cannot be >= kint32max value, got 2147483647"): + test_quantize_and_dequantize_v3() + + with self.assertRaisesRegex( + ValueError, "Axis cannot be >= kint32max value, got 2147483647"): + test_quantize_and_dequantize_v4() + + with self.assertRaisesRegex( + ValueError, "Axis cannot be >= kint32max value, got 2147483647"): + test_quantize_and_dequantize_v4_grad() + @test_util.run_all_in_graph_and_eager_modes class SortedSearchTest(test_util.TensorFlowTestCase):
Vulnerability mechanics
Generated by null/stub on May 9, 2026. Inputs: CWE entries + fix-commit diffs from this CVE's patches. Citations validated against bundle.
References
4- github.com/advisories/GHSA-gw97-ff7c-9v96ghsaADVISORY
- nvd.nist.gov/vuln/detail/CVE-2023-25668ghsaADVISORY
- github.com/tensorflow/tensorflow/commit/7b174a0f2e40ff3f3aa957aecddfd5aaae35eccbghsax_refsource_MISCWEB
- github.com/tensorflow/tensorflow/security/advisories/GHSA-gw97-ff7c-9v96ghsax_refsource_CONFIRMWEB
News mentions
0No linked articles in our index yet.