Heap out of bounds access in MakeEdge in TensorFlow
Description
In affected versions of TensorFlow under certain cases, loading a saved model can result in accessing uninitialized memory while building the computation graph. The MakeEdge function creates an edge between one output tensor of the src node (given by output_index) and the input slot of the dst node (given by input_index). This is only possible if the types of the tensors on both sides coincide, so the function begins by obtaining the corresponding DataType values and comparing these for equality. However, there is no check that the indices point to inside of the arrays they index into. Thus, this can result in accessing data out of bounds of the corresponding heap allocated arrays. In most scenarios, this can manifest as unitialized data access, but if the index points far away from the boundaries of the arrays this can be used to leak addresses from the library. This is fixed in versions 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2, and 2.4.0.
Affected packages
Versions sourced from the GitHub Security Advisory.
| Package | Affected versions | Patched versions |
|---|---|---|
tensorflowPyPI | < 1.15.5 | 1.15.5 |
tensorflowPyPI | >= 2.0.0, < 2.0.4 | 2.0.4 |
tensorflowPyPI | >= 2.1.0, < 2.1.3 | 2.1.3 |
tensorflowPyPI | >= 2.2.0, < 2.2.2 | 2.2.2 |
tensorflowPyPI | >= 2.3.0, < 2.3.2 | 2.3.2 |
tensorflow-cpuPyPI | < 1.15.5 | 1.15.5 |
tensorflow-cpuPyPI | >= 2.0.0, < 2.0.4 | 2.0.4 |
tensorflow-cpuPyPI | >= 2.1.0, < 2.1.3 | 2.1.3 |
tensorflow-cpuPyPI | >= 2.2.0, < 2.2.2 | 2.2.2 |
tensorflow-cpuPyPI | >= 2.3.0, < 2.3.2 | 2.3.2 |
tensorflow-gpuPyPI | < 1.15.5 | 1.15.5 |
tensorflow-gpuPyPI | >= 2.0.0, < 2.0.4 | 2.0.4 |
tensorflow-gpuPyPI | >= 2.1.0, < 2.1.3 | 2.1.3 |
tensorflow-gpuPyPI | >= 2.2.0, < 2.2.2 | 2.2.2 |
tensorflow-gpuPyPI | >= 2.3.0, < 2.3.2 | 2.3.2 |
Affected products
1- Range: < 1.15.5
Patches
10cc38aaa4064Prevent unitialized memory access in `GraphConstructor::MakeEdge`
1 file changed · +12 −0
tensorflow/core/common_runtime/graph_constructor.cc+12 −0 modified@@ -44,6 +44,7 @@ limitations under the License. #include "tensorflow/core/lib/gtl/inlined_vector.h" #include "tensorflow/core/lib/strings/scanner.h" #include "tensorflow/core/lib/strings/str_util.h" +#include "tensorflow/core/platform/errors.h" #include "tensorflow/core/platform/logging.h" #include "tensorflow/core/platform/macros.h" #include "tensorflow/core/public/version.h" @@ -1425,6 +1426,17 @@ void GraphConstructor::Undo() { Status GraphConstructor::MakeEdge(Node* src, int output_index, Node* dst, int input_index) { + if (output_index >= src->num_outputs()) { + return errors::InvalidArgument( + "Output ", output_index, " of node ", src->name(), + " does not exist. Node only has ", src->num_outputs(), " outputs."); + } + if (input_index >= dst->num_inputs()) { + return errors::InvalidArgument( + "Input ", input_index, " of node ", dst->name(), + " does not exist. Node only has ", dst->num_inputs(), " inputs."); + } + DataType src_out = src->output_type(output_index); DataType dst_in = dst->input_type(input_index); if (!TypesCompatible(dst_in, src_out)) {
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
7- github.com/advisories/GHSA-q263-fvxm-m5mwghsaADVISORY
- nvd.nist.gov/vuln/detail/CVE-2020-26271ghsaADVISORY
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2020-302.yamlghsaWEB
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2020-337.yamlghsaWEB
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow/PYSEC-2020-257.yamlghsaWEB
- github.com/tensorflow/tensorflow/commit/0cc38aaa4064fd9e79101994ce9872c6d91f816bghsax_refsource_MISCWEB
- github.com/tensorflow/tensorflow/security/advisories/GHSA-q263-fvxm-m5mwghsax_refsource_CONFIRMWEB
News mentions
0No linked articles in our index yet.