VYPR
← All advisories
High severityNVD Advisory· Published Feb 4, 2022· Updated Apr 22, 2025

Segfault in `simplifyBroadcast` in Tensorflow

CVE-2022-23593

Description

Tensorflow is an Open Source Machine Learning Framework. The simplifyBroadcast function in the MLIR-TFRT infrastructure in TensorFlow is vulnerable to a segfault (hence, denial of service), if called with scalar shapes. If all shapes are scalar, then maxRank is 0, so we build an empty SmallVector. The fix will be included in TensorFlow 2.8.0. This is the only affected version.

Affected packages

Versions sourced from the GitHub Security Advisory.

PackageAffected versionsPatched versions
tensorflowPyPI
>= 2.8.0-rc0, < 2.8.02.8.0
tensorflow-cpuPyPI
>= 2.8.0-rc0, < 2.8.02.8.0
tensorflow-gpuPyPI
>= 2.8.0-rc0, < 2.8.02.8.0

Affected products

1

Patches

1
35f0fabb4c17

Avoid Segfault for scalar shapes.

https://github.com/tensorflow/tensorflowAdrian KuegelDec 21, 2021via ghsa
commit
2 files changed · +15 −0
  • tensorflow/compiler/mlir/tfrt/jit/transforms/tf_cpurt_symbolic_shape_optimization.cc+4 −0 modified
    @@ -157,6 +157,10 @@ llvm::Optional<Value> simplifyBroadcast(ShapeComponentAnalysis& analysis,
     shapes_found.push_back(*found_shape);
     maxRank = std::max(maxRank, found_shape->size());
   }
+  if (maxRank == 0) {
+    return Value(builder->create<tensor::FromElementsOp>(
+        loc, shapes[0].getType(), SmallVector<Value>()));
+  }
 
   SmallVector<const ShapeComponentAnalysis::SymbolicExpr*> joined_dimensions(
       maxRank);
  • tensorflow/compiler/mlir/tfrt/python_tests/regression_tests/scalar_broadcast.mlir+11 −0 added
    @@ -0,0 +1,11 @@
+builtin.func @test(%V__0 : tensor<i1> { python_test_attrs.static_type = tensor<i1> }, %V__1 : tensor<f32> { python_test_attrs.static_type = tensor<f32> }, %V__2 : tensor<f32> { python_test_attrs.static_type = tensor<f32> }) -> tensor<f32> {
+  %0 = "tf.Cast"(%V__0) : (tensor<i1>) -> tensor<i1>
+  %1 = "tf.Selu"(%V__2) : (tensor<f32>) -> tensor<f32>
+  %2 = "tf.NextAfter"(%1, %V__2) : (tensor<f32>, tensor<f32>) -> tensor<f32>
+  %3 = "tf.Elu"(%2) : (tensor<f32>) -> tensor<f32>
+  %4 = "tf.Cosh"(%3) : (tensor<f32>) -> tensor<f32>
+  %5 = "tf.Elu"(%4) : (tensor<f32>) -> tensor<f32>
+  %6 = "tf.Div"(%V__1, %5) : (tensor<f32>, tensor<f32>) -> tensor<f32>
+  %7 = "tf.Select"(%0, %6, %V__1) : (tensor<i1>, tensor<f32>, tensor<f32>) -> tensor<f32>
+  return %7 : tensor<f32>
+}
\ No newline at end of file

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

News mentions

0

No linked articles in our index yet.