Remove null pointer from error message in mlir::odml

[apach301]

Issue type

Bug

Have you reproduced the bug with TensorFlow Nightly?

No

Source

source

TensorFlow version

2.16.1

Custom code

No

OS platform and distribution

No response

Mobile device

No response

Python version

No response

Bazel version

No response

GCC/compiler version

No response

CUDA/cuDNN version

No response

GPU model and memory

No response

Current behavior?

Possible null pointer dereference in mlir::odml::MatchInverseScalesOperand(): undefined variable inverse_scale_constant_op is printed in error message.

tensorflow/tensorflow/compiler/mlir/lite/stablehlo/transforms/compose_uniform_quantized_type_pass.cc

Lines 172 to 178 in 08a13d2

	auto inverse_scale_constant_op =
	dyn_cast_or_null<stablehlo::ConstantOp>(inverse_scales.getDefiningOp());
	if (!inverse_scale_constant_op) {
	llvm::dbgs()
	<< "Inverse scales should be a constant. Instead, it was defined by: "
	<< inverse_scale_constant_op << ".\n";
	return failure();

Standalone code to reproduce the issue

Bug was found with Svace static analyzer

Relevant log output

No response

[AsbestosLampshade]

Hi @apach301 is this issue still open?
If so, could you tell me how to replicate the bug so that I may have a go at it?

[apach301]

Hi @apach301 is this issue still open?

If so, could you tell me how to replicate the bug so that I may have a go at it?

This bug was detected by static analysis, therefore I have no input to reproduce it during execution. But I propose a fix in PR #67557

[AsbestosLampshade]

Hi @apach301 the PR looks great, if you would like any help with CI, do let me know.

Thanks and regards

[pkgoogle]

will wait for review.

[apach301]

will wait for review.

Hi @pkgoogle How to enable CI tests run in PR #67557? Is it wait any approval to run?

[pkgoogle]

Hi @apach301, we'll have to wait for review, we run CI/CD internally and review the PR.