Raise inoperative error guards in ExportSession, XNNPACK node_visitor, and Samsung conv1d

[AxelNoun]

Summary

Error-guard sites that construct a failure which cannot actually fire. This PR inserts the missing raise (or replaces an always-true assert with the file's existing check_or_raise) so the documented failure mode is the one callers see.

Updated for main at 5e21c13: the two ExportSession guards were fixed upstream in the meantime by #21748, so the only code change left in export/export.py is the message typo. The regression tests for those two guards are kept, since nothing else covers them.

File	What happens on main today	Reachability
export/export.py ExportSession.print_delegation_info	Both guards already raise (#21748). Remaining here: the message says "atleast", and no test pins either guard.	Reachable. ExportSession is in executorch.export.__all__ and is @experimental. A pipeline with no lowering stage, or one that lists TO_EDGE_TRANSFORM_AND_LOWER / TO_BACKEND without having run it, reaches each guard. This PR fixes the wording and adds export/tests/test_print_delegation_info.py for both.
backends/xnnpack/operators/node_visitor.py per-channel axis else	assert f"Unsupported weight per channel quantization axis ...": the f-string is always truthy, so the assertion never fires. An assert would also disappear under python -O.	Not reachable on a normal export. The only call that passes swap_in_out_for_weights=True is backends/xnnpack/operators/op_conv2d.py:102, and that visitor already check_or_raises the axis to 0 (depthwise) or 1 (transpose) before define_tensor. This change restores a guard that can actually raise if that else is ever entered; it does not fix a mis-serialized .pte on the current conv2d path.
backends/samsung/_passes/conv1d_to_conv2d.py update_kernel else	RuntimeError("Weight of 1d conv should be constant tensor or Parameter obj") is built and discarded; weight_node.meta["val"] is still unsqueezed.	Not reachable on a normal Parameter / lifted-constant / buffer conv1d export. Those cases are handled by the preceding branches (the buffer branch was added on main by #22725). The else is the leftover for a weight that is none of those. This change only makes that documented RuntimeError actually raise if the branch is entered.

Test plan

Tests cover only the two reachable ExportSession guards. There is intentionally no test for the XNNPACK else and none for the Samsung else: both are unreachable on a normal export, and a test that injects that state would exercise a situation the production callers do not produce.

export/tests is already in pytest.ini testpaths:

pytest export/tests/test_print_delegation_info.py

Measured on 89cf698 (merge of main 5e21c13), CPython 3.12, torch 2.15.0.dev20261004+cpu:

• export/tests/test_print_delegation_info.py: 2 passed.
• Negative control, after git checkout origin/main -- export/export.py, same run: AssertionError: 'at least one of the lowering stages' not found in 'No delegation info available, atleast one of the lowering stages should be present', 1 failed / 1 passed.
• pytest export/tests --ignore=export/tests/test_target_recipes.py: 8 failed / 172 passed on this branch, and the same 8 failed / 170 passed on main alone. The 8 are a local-environment gap (exir/_serialize/program.fbs is copied in by the build, which this checkout has not run), not a regression. test_target_recipes.py does not collect locally (no coremltools).

CI

Merged main into the branch. The conflict in export/export.py was resolved in favour of main's _raise_if_released call, keeping the typo fix. The six reds triaged earlier on 12c3858 came from the merge base and their upstream fixes (#22802, #22754) are now in this branch.

Two checks are still red on 89cf698, neither of them from this change:

• pull / android / run-emulator: the emulator boots (sys.boot_completed returns 1), and reactivecircus/android-emulator-runner then fails on its own post-boot adb shell input keyevent 82 with cmd: Failure calling service input: Broken pipe (32), exit code 224. That step runs before the job script, and the log carries no output from scripts/run_android_emulator.sh. The same job also failed on main at efc7f2e when checked on 2026-10-06, and passed on the seven main runs before it.
• MLTEC_elastic - GitHubPOC (NXP Bamboo): outside this repository's workflows, and red on other PRs too, not just this one: of eight open PRs untouched since 2026-10-05, two carry this status and both are red. Here the status went pending at 2026-10-05T07:01:25Z and failure one second later, which reads as a trigger-side failure rather than anything built from this branch, but the build log is on an internal host so I cannot confirm that from here.

This change touches three Python files and adds one test.

cc @nil-is-all @digantdesai

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/22373

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

⚠️ 15 Awaiting Approval

As of commit aabf5c3 with merge base f2575be ():

AWAITING APPROVAL - The following workflows need approval before CI can run:

• Build documentation (gh)
• Build Presets (gh)
• Cadence Build & Test (gh)
• Lint (gh)
• periodic (gh)
• pull (gh)
• Test ARM Backend (gh)
• Test CoreML Backend (gh)
• Test Cortex-M Backend (gh)
• Test NXP Backend (gh)
• Test QNN Backend (gh)
• Test QNN Windows MSVC build (gh)
• Test Vulkan Backend (gh)
• Test WebGPU Backend (gh)
• Test XNNPACK Backend (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[AxelNoun]

@pytorchbot label "release notes: none"

@pytorchbot label "release notes: none"

[nil-is-all]

Thanks @AxelNoun, running CI again

[AxelNoun]

Thanks @AxelNoun, running CI again

Thanks ! @nil-is-all

[AxelNoun]

Thanks @nil-is-all. I triaged the 6 failures on 12c3858. All of them are already present on the merge base df6147a, none come from this PR:

• pull / unittest{,-editable} (linux, macos): 11 failures + 2 collection errors, all ModuleNotFoundError: No module named 'pyre_extensions' in backends/native/test and exir/native/test. Introduced by Add ReplaceCopyWithAliasPass (#22762) #22762, fixed on main by Remove pyre_extensions dependency from native preprocessing #22802. The same 13 test ids fail on df6147a, and the passing count goes 2800 → 2802: the two tests this PR adds run and pass.
• Test QNN Windows MSVC build (arm64, x64): MSVC errors in extension/llm/runner/model_metadata.h, fixed on main by Fix model metadata readers on MSVC #22754. This PR changes four error guards in three Python files and adds one Python test: no C++, no QNN, no build files.

The workflows check out the PR head rather than a merge with main, so re-running on 12c3858 would reproduce the same six reds, since the fixes only landed after df6147a. Happy to merge main in (500849b is green on all of these) if you'd prefer a green run; otherwise this should be reviewable as is. @digantdesai, could you take a look?

[nil-is-all]

Thanks @nil-is-all. I triaged the 6 failures on 12c3858. All of them are already present on the merge base df6147a, none come from this PR:

• pull / unittest{,-editable} (linux, macos): 11 failures + 2 collection errors, all ModuleNotFoundError: No module named 'pyre_extensions' in backends/native/test and exir/native/test. Introduced by Add ReplaceCopyWithAliasPass (#22762) #22762, fixed on main by Remove pyre_extensions dependency from native preprocessing #22802. The same 13 test ids fail on df6147a, and the passing count goes 2800 → 2802: the two tests this PR adds run and pass.
• Test QNN Windows MSVC build (arm64, x64): MSVC errors in extension/llm/runner/model_metadata.h, fixed on main by Fix model metadata readers on MSVC #22754. This PR changes four error guards in three Python files and adds one Python test: no C++, no QNN, no build files.

The workflows check out the PR head rather than a merge with main, so re-running on 12c3858 would reproduce the same six reds, since the fixes only landed after df6147a. Happy to merge main in (500849b is green on all of these) if you'd prefer a green run; otherwise this should be reviewable as is. @digantdesai, could you take a look?

Hi @AxelNoun, I agree the CI failures seem irrelevant. Would like to get a final OK from @digantdesai and then merge. Thanks for the ping

[digantdesai]

shoudln't we use 'why()'?

[AxelNoun]

why() here is WhyNoPartition (exir/backend/utils.py), and every call to it in this backend sits in partition/config/*.py, or in a pass that receives it as an argument: it logs a reason and the caller returns False, so the node is left on CPU. define_tensor runs after partitioning, at serialization time, so there is no why in scope and nothing left to decline. The options here are to raise or to serialize a wrong axis silently. I used the helper this file already uses (four other check_or_raise calls), and op_conv2d.py performs these same two axis checks one level up (lines 88 and 93), which is why this else should be unreachable on the conv2d path.

Your question did surface two things, though.

First, check_or_raise(False, ...) is the only literal-False call in the repo. Happy to give it a real condition, which also drops the else:

check_or_raise(
    quant_params.axis in (0, 1),
    f"Unsupported weight per channel quantization axis for depthwise conv2d / conv_transpose2d : {quant_params.axis}, expecting 0 / 1.",
)
if quant_params.axis == 0:
    quant_params.axis = len(dims) - 1
else:
    quant_params.axis = 0

Second, ConvolutionConfig.check_constraints uses why() for the transpose per-channel axis (gemm_configs.py:406), but there is no equivalent gate for the depthwise one. A per-channel depthwise conv with axis != 0 is therefore partitioned and then raises at op_conv2d.py:93, instead of staying on CPU. The axis is read verbatim from the dequant node's 4th arg (quant_params.py:272), so a quantizer other than XNNPACKQuantizer can produce it. Would you like that why() gate added here, or kept out of this PR (which only makes dead guards fire) and done as a follow-up?

[digantdesai]

Please rebase, and make sure CI is green. Thanks.

[AxelNoun]

@digantdesai I merged main (5e21c13) into the branch. The conflict in export/export.py was resolved in favour of main, where both guards already raise since #21748; this PR keeps the "at least" wording fix and the two regression tests. The branch still merges cleanly with current main, and 178 checks pass.

Two checks are red. Neither is related to the change, which touches three Python files and adds one test, with no Android or C++ code:

• pull / android / run-emulator: the emulator booted, then reactivecircus/android-emulator-runner failed on its own post-boot adb shell input keyevent 82 (Broken pipe (32), exit 224). The action runs that command before the job script, and the log has no output from scripts/run_android_emulator.sh. I don't have permission to re-run jobs here. @nil-is-all, could you re-run this one?
• MLTEC_elastic - GitHubPOC: an external NXP check, which also failed on the 8 other open PRs I checked on Oct 5.