[None][test] DSv4 PR-6 coverage and import safety

[lfr-0531]

@coderabbitai summary

Description

PR-6 from the DSv4 a7728aa+ split. This PR keeps the remaining coverage, CI metadata, and import-safety polish separate from the functional DSv4/runtime/router/MoE changes.

Changes included:

• Add DSv4 Flash disaggregated serving smoke coverage and DSv4 Pro aggregate GSM8K / short-token-boundary coverage.
• Add related B200/GB200 CI entries that are still valid on latest main.
• Guard import-time CUDA probes in FlashInfer/Triton prefill/TileIR paths so import tensorrt_llm remains GPU-free when no CUDA device is visible.
• Add tests/unittest/others/test_import_gpu_free.py to pin that import contract.

Test Coverage

Passed:

• git commit -s pre-commit hook, including formatting, ruff, duplicate waive checks, and test-list AST validation.
• python3 -m py_compile on modified Python files.
• git diff --check github/main...HEAD.
• Strict changed-file conflict-marker search for <<<<<<< / >>>>>>>.
• PR-6 forbidden-pattern check: no DSv4 model/sparse backend/MoE/router/FMHA/pre-commit/pyproject files in the diff.

Not run / blocked:

• Runtime pytest and collect-only checks from the unbuilt worktree were blocked by missing tensorrt_llm.bindings; no C++/CUDA/header files changed, so build-install was not required for this PR.
• GPU-consuming DSv4 integration tests were not run. GPU pre-check showed all 8 B300 GPUs had high memory occupancy, so I avoided consuming GPU resources.

PR Checklist

Please review the following before submitting your PR:

• PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.

• PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.

• Test cases are provided for new code paths (see test instructions)

• If PR introduces API changes, an appropriate PR label is added - either api-compatible or api-breaking. For api-breaking, include BREAKING in the PR title.

• Any new dependencies have been scanned for license and vulnerabilities

• CODEOWNERS updated if ownership changes

• Documentation updated as needed

• Update tava architecture diagram if there is a significant design change in PR.

• The reviewers assigned automatically/manually are appropriate for the PR.

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

GitHub Bot Help

To see a list of available CI bot commands, please comment /bot help.

[lfr-0531]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #56362 [ run ] triggered by Bot. Commit: a64afb4 Link to invocation

[lfr-0531]

/bot kill

[tensorrt-cicd]

PR_Github #56363 [ kill ] triggered by Bot. Commit: 1f51d40 Link to invocation

[lfr-0531]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #56362 [ run ] completed with state ABORTED. Commit: a64afb4

Link to invocation

[tensorrt-cicd]

PR_Github #56363 [ kill ] completed with state SUCCESS. Commit: 1f51d40
Successfully killed previous jobs for commit 1f51d40

Link to invocation

[tensorrt-cicd]

PR_Github #56367 [ run ] triggered by Bot. Commit: 1f51d40 Link to invocation

[tensorrt-cicd]

PR_Github #56367 [ run ] completed with state SUCCESS. Commit: 1f51d40
/LLM/main/L0_MergeRequest_PR pipeline #45212 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

• Please check the failed tests and fix your PR
• If you cannot view the failures, ask the CI triggerer to share details
• Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

[lfr-0531]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #56601 [ run ] triggered by Bot. Commit: e93b8a9 Link to invocation

[tensorrt-cicd]

PR_Github #56601 [ run ] completed with state SUCCESS. Commit: e93b8a9
/LLM/main/L0_MergeRequest_PR pipeline #45428 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

• Please check the failed tests and fix your PR
• If you cannot view the failures, ask the CI triggerer to share details
• Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation