[chenfeiz0326]

Description

tests/unittest/scripts/test_subprocess_timeouts.py::test_subprocess_run_calls_have_timeouts (from #18812) requires every subprocess.run under tests/ to pass a timeout. #20024 added a gcc call in tests/unittest/scripts/test_bolt.py (_compile_so, which builds a scratch shared object) without one. Since then, the unittest/scripts stage has failed on pre-merge runs on CPU-Generic-x86-1/arm-1. In recent L0 builds it failed on 7 unrelated PRs (e.g. #20099, #20151, #19492, #19722, #19275, #20253, #20250):

FAILED unittest/scripts/test_subprocess_timeouts.py::test_subprocess_run_calls_have_timeouts
  AssertionError: subprocess.run calls without timeout:
    unittest/scripts/test_bolt.py:123

This PR adds timeout=60 to that call. It compiles a one-line C file, so 60 s is generous. A hang now fails the test instead of the stage.

Test Coverage

• I ran the policy test's own checker (_candidate_paths + _missing_timeout_lines) over the patched tree: 119 files scanned, 0 violations. On current main it reports tests/unittest/scripts/test_bolt.py:123.
• test_bolt.py behaviour is unchanged, apart from the bound on that gcc call.

PR Checklist

• PR description clearly explains what and why.
• Follows the TensorRT-LLM coding guidelines.
• Test cases are provided.

🤖 Generated with Claude Code

Dev Engineer Review

_compile_so now passes timeout=60 to subprocess.run when it invokes gcc. This satisfies the existing policy test, which checks that subprocess.run calls in the test tree specify a timeout. No other behavior change is reported. Current review finding counts are unavailable.

QA Engineer Review

tests/unittest/scripts/test_bolt.py changed. The change does not add or modify test functions; it bounds the compiler subprocess used by existing tests. The policy test checks for missing timeouts, but no test run result is provided. Coverage verdict: sufficient for this timeout-policy change. No CI or manual-QA test-list entry applies to this unit test file.

Per-File QA Perspective

• tests/unittest/scripts/test_bolt.py: Existing BOLT helper tests compile temporary shared objects through _compile_so; the compiler call now stops after 60 seconds. This unit test file is not listed in the integration CI or manual-QA lists.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Repository: NVIDIA/TensorRT-LLM/.coderabbit.yaml
• Review profile: CHILL
• Plan: Enterprise
• Run ID: 49c0fc6b-0ffd-4fee-aba3-bd009292f225

📥 Commits

Reviewing files that changed from the base of the PR and between 02f5556 and 23b8db8.

📒 Files selected for processing: 1

• tests/unittest/scripts/test_bolt.py

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 _compile_so helper now sets a 60-second timeout for the GCC subprocess. It continues to return whether compilation succeeded.

Changes

GCC compile timeout

Layer / File(s)	Summary
Set GCC subprocess timeout tests/unittest/scripts/test_bolt.py	The GCC subprocess call now uses a 60-second timeout. The helper still returns whether compilation succeeded.

Priority: ⬇️ Low

Estimated code review effort: 1 (Trivial) | ~3 minutes

Change: Bug fix

Merge Risk: ⚪ Minimal · up to 23b8d

The GCC test invocation is now bounded at 60 seconds, and no actionable merge-blocking risk is established.

Pre-merge checks: 5

Passed checks: 5 passed

Check name	Status	Explanation
Title check		The title clearly and concisely describes the main change: adding a timeout to the subprocess.run call in test_bolt. It follows the repository title format.
Description check		The description includes the required Description, Test Coverage, and PR Checklist sections. It explains the issue, the solution, the affected test, and the reported validation results.
Docstring Coverage		Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 1 functions across 1 files.
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.

✨ 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.}

[chenfeiz0326]

/bot run

[tensorrt-cicd]

PR_Github #77881 [ run ] triggered by Bot. Commit: 23b8db8 Link to invocation

[tensorrt-cicd]

PR_Github #77881 [ run ] completed with state SUCCESS. Commit: 23b8db8
/LLM/main/L0_MergeRequest_PR pipeline #64200 completed with status: 'SUCCESS'

CI Report

Link to invocation

[chenfeiz0326]

CI is green. @mlefeb01 @yihwang-nv, could one of you take a look? It's a one-line timeout=60 on the gcc call #20024 added to test_bolt.py. Without it, test_subprocess_timeouts (#18812) fails the unittest/scripts stage on every pre-merge run against current main, so other PRs (e.g. #20250) can't get a green CI until this lands.

[trtllm-agent]

[Repair Bot] NVBug 6929769 appears to have the same root cause as this PR.
No separate PR was opened.

• Failing test: pytest tests/unittest/scripts/test_subprocess_timeouts.py::test_subprocess_run_calls_have_timeouts -v
• Comparison: The existing PR's structured evidence fields are empty, but its description quotes the candidate's exact failure: test_subprocess_timeouts.py::test_subprocess_run_calls_have_timeouts fails with 'subprocess.run calls without timeout: unittest/scripts/test_bolt.py:123', on CPU-Generic-x86-1/arm-1. Both name the same cause: the _compile_so gcc helper added in [TRTLLMINF-336][fix] Stop emitting cortex-a53-843419 veneers in BOLT builds #20024 calls subprocess.run without a timeout. The test_bolt.py change is identical in both (adds timeout=60). The candidate also removes the

If this is a different root cause, an NVIDIA collaborator can request a separate attempt by commenting /repair-bot continue 6929769.

[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.