[dongfengy]

Description

Remove the stale waiver for unittest/_torch/modeling_v2/models/glm_moe_dsa/test_routing.py, restoring its CPU CI coverage. #20789 standardized the GLM route and its caller on the same interface, fixing the signature mismatch tracked by NVBug 6986044.

Test Coverage

• ComputeLab x86_64 validation at 69cab8a0a3c72bd81a80ddbe3b111935c82baee4: 21 passed, 0 failed, 0 errors, 0 skipped.
• Built a clean SM100 wheel (1.4.0rc3) from this commit in a fresh source-built image, installed it into a separate runtime, and verified native-library loading and imports from that runtime's site-packages.
• Ran the existing GLM routing suite with its normal pytest configuration and cpu_only marker. Slurm build/test job 5383196 completed with exit code 0.
• Repository pre-commit checks, test-list validation, and DCO: passed.
• Current-head CI: L0 #66920 passed. The GLM routing suite actually ran on both x86 CPU #10091 and ARM CPU #8643: 21 routing cases passed per architecture, 0 failures, 0 skips, plus the integration wrapper passed. No prior test results were reused.
• CI selected the affected CPU suite via CBTS waiveonly; both architecture builds, release checks, and test-list checks passed. This was a targeted CI pass, not the complete single-/multi-GPU L0 matrix. The suite remains registered in tests/integration/test_lists/test-db/l0_cpu.yml.
• The local validation covers Python routing behavior with mocked CUDA capabilities, not ARM wheel compatibility or model inference.

PR Checklist

• Please check this after reviewing the above items as appropriate for this PR.

This change restores existing CI coverage and adds no code paths, APIs, dependencies, or ownership changes.

Dev Engineer Review

The change removes the NVBug 6986044 waiver for unittest/_torch/modeling_v2/models/glm_moe_dsa/test_routing.py. It changes no source code or public APIs. No current review findings were supplied, so finding counts are unavailable.

QA Engineer Review

tests/integration/test_lists/waives.txt has one waiver removed. No test functions or selectors changed. The supplied context reports 21 passed, 0 failed, 0 errors, and 0 skipped in a CPU routing run, and reports that the related Slurm job exited with code 0. Current-head CI, including ARM CPU coverage, had not passed yet. Coverage verdict: needs follow-up for those unverified environments.

Per-File QA Perspective

• tests/integration/test_lists/waives.txt: Removes the waiver for the GLM MoE DSA routing test. tests/integration/test_lists/test-db/l0_cpu.yml lists the test for CPU CI.

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: NVIDIA/TensorRT-LLM
• Branch: main
• File: .coderabbit.yaml
• Commit: 1f9b707
• Layers: Repository YAML
• Review profile: CHILL
• Plan: Enterprise
• Run ID: 4675aea9-6ddf-4bea-bffe-de45f56c0457

Commits

Reviewing files that changed from the base of the PR and between b203764 and 69cab8a.

Files selected for processing: 1

• tests/integration/test_lists/waives.txt

Files with no reviewable changes: 1

• tests/integration/test_lists/waives.txt

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review.

Walkthrough

The integration test waiver list no longer contains the skip entry for test_routing.py associated with NVBUG 6986044.

Changes

Routing test waiver

Layer / File(s)	Summary
Remove routing test waiver tests/integration/test_lists/waives.txt	The skip-list entry for test_routing.py associated with NVBUG 6986044 was removed.

Priority: Normal

Estimated code review effort: 1 (Trivial) | ~2 minutes

Change: Bug fix

Merge Risk: Minimal · up to 69cab

No actionable merge-blocking risk is identified; complete the normal current-head CI checks before merging.

Pre-merge checks: 4

Check name	Status	Explanation
Title check		The title follows the required NVBugs and type format and clearly describes re-enabling the GLM modeling_v2 routing tests.
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.
Description check		The pull request description includes the required Description, Test Coverage, and PR Checklist sections. It clearly explains the stale waiver removal, the related fix, validation results, CI scope, a…

Full details: Description check

Explanation

The pull request description includes the required Description, Test Coverage, and PR Checklist sections. It clearly explains the stale waiver removal, the related fix, validation results, CI scope, and the absence of API, dependency, or ownership changes.

Finishing Touches

🧪 Generate unit tests (beta)

• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

_{Comment @coderabbitai help to get the list of available commands.}

[brnguyen2]

Approving — the comments below are optional touch-ups, not blockers.

GLM route() now has the same (model_config, llm_args, compute_capability) signature that modeling_v2_resolve calls with, and test_routing.py is still listed in l0_cpu.yml, so this waiver can come out.

The original failure was reported on the CPU-Generic-arm-1 stage, and the validation in the description is x86-only. Please let a full /bot run (not /bot skip) go green with that stage included before merging.

[github-actions]

Automatically added "ci: full pre-merge approved" because this PR has satisfied the required GitHub review approvals. Unresolved review conversations and other required checks remain independent merge requirements.

[dongfengy]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #81230 [ run ] triggered by Bot. Commit: 69cab8a Link to invocation

[tensorrt-cicd]

PR_Github #81230 [ run ] completed with state SUCCESS. Commit: 69cab8a
/LLM/main/L0_MergeRequest_PR pipeline #66920 completed with status: 'SUCCESS'

CI Report

Link to invocation