[TRTLLM-10030][chore] use weakref in atexit handler

[ixlmar]

Description

This avoids keeping LLM instances alive longer than necessary.

Test Coverage

n/a

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)

• 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

/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...

Provide a user friendly way for developers to interact with a Jenkins server.

Run /bot [-h|--help] to print this help message.

See details below for each supported subcommand.

Details

run [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]

Launch build/test pipelines. All previously running jobs will be killed.

--reuse-test (optional)pipeline-id (OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.

--disable-reuse-test (OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.

--disable-fail-fast (OPTIONAL) : Disable fail fast on build/tests/infra failures.

--skip-test (OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.

--stage-list "A10-PyTorch-1, xxx" (OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-1, xxx". Note: Does NOT update GitHub check status.

--gpu-type "A30, H100_PCIe" (OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.

--test-backend "pytorch, cpp" (OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline status.

--only-multi-gpu-test (OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.

--disable-multi-gpu-test (OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.

--add-multi-gpu-test (OPTIONAL) : Force run the multi-GPU tests in addition to running L0 pre-merge pipeline.

--post-merge (OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.

--extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" (OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".

--detailed-log (OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.

--debug (OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in the stage-list parameter to access the appropriate container environment. Note: Does NOT update GitHub check status.

For guidance on mapping tests to stage names, see docs/source/reference/ci-overview.md
and the scripts/test_to_stage_mapping.py helper.

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip testing for latest commit on pull request. --comment "Reason for skipping build/test" is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

reuse-pipeline

reuse-pipeline

Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

Summary by CodeRabbit

• Bug Fixes
  • Improved executor shutdown reliability during program termination for safer resource management.

[ixlmar]

/bot run

[coderabbitai]

📝 Walkthrough

Walkthrough

Replaces atexit.register(self.shutdown) with a weakref-based finalizer approach. Adds weakref import. Uses a lambda capturing a finalizer bound to the class method and a weak reference to the instance. Changes cleanup timing during interpreter exit.

Changes

Cohort / File(s)	Summary
Cleanup Mechanism Update tensorrt_llm/executor/executor.py	Replaces atexit-based shutdown registration with weakref finalizer pattern to prevent shutdown attempts on garbage-collected instances. Adds weakref import (+5/-1 lines).

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly describes the main change: replacing atexit.register with a weakref-based finalizer approach in the executor.
Description check	✅ Passed	The description explains the rationale (avoiding unnecessary LLM instance retention) and the PR checklist is completed, though test coverage is marked as n/a.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

No actionable comments were generated in the recent review. 🎉

🧹 Recent nitpick comments

tensorrt_llm/executor/executor.py (2)

92-95: Consider using weakref.finalize — it exists for exactly this purpose.

weakref.finalize manages atexit registration, weak-reference tracking, and safe invocation in a single call, replacing the hand-rolled lambda+default-arg pattern. It also silences the Ruff B008 warning about type(self) in the default argument.

♻️ Proposed simplification

-        atexit.register(
-            lambda ref, finalizer=type(self).shutdown: finalizer(obj)
-            if (obj := ref()) is not None else None,
-            weakref.ref(self))
+        weakref.finalize(self, type(self).shutdown, self)

weakref.finalize will call type(self).shutdown(self) only if the instance is still alive at the time of collection or interpreter exit, and it integrates with atexit automatically. Note that this also lets you drop the import atexit if it's no longer used elsewhere in the file.

Wait — weakref.finalize(self, type(self).shutdown, self) would pass self as a strong reference arg, defeating the purpose. The correct form is:

♻️ Corrected simplification

-        atexit.register(
-            lambda ref, finalizer=type(self).shutdown: finalizer(obj)
-            if (obj := ref()) is not None else None,
-            weakref.ref(self))
+        self._prevent_leak_ref = weakref.finalize(self, type(self).shutdown, self)

Actually, weakref.finalize specifically does not prevent GC — its design stores a weak reference to the first argument and strong references to the callback and remaining args. But the remaining args here would include self, which is a strong reference and would keep the object alive.

The correct idiomatic approach with finalize for instance methods requires passing only non-prevent-GC data. Since shutdown needs the instance, the current lambda+weakref approach is actually reasonable. One small cleanup: extract type(self).shutdown to avoid the Ruff B008 lint:

+        shutdown = type(self).shutdown
         atexit.register(
-            lambda ref, finalizer=type(self).shutdown: finalizer(obj)
+            lambda ref, finalizer=shutdown: finalizer(obj)
             if (obj := ref()) is not None else None,
             weakref.ref(self))

---

92-95: Pre-compute type(self).shutdown to fix Ruff B008 and improve readability.

The approach is correct — weakref.finalize can't trivially replace this because passing self as a callback argument would create a strong reference that defeats the purpose. However, the lambda is dense, and calling type(self) inside a default argument triggers Ruff B008. Extracting it to a local variable resolves both issues.

♻️ Proposed cleanup

+        shutdown = type(self).shutdown
         atexit.register(
-            lambda ref, finalizer=type(self).shutdown: finalizer(obj)
-            if (obj := ref()) is not None else None,
+            lambda ref, finalizer=shutdown: finalizer(obj)
+                if (obj := ref()) is not None else None,
             weakref.ref(self))

Tip

Issue Planner is now in beta. Read the docs and try it out! Share your feedback on Discord.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[tensorrt-cicd]

PR_Github #35795 [ run ] triggered by Bot. Commit: e7d48d5

[tensorrt-cicd]

PR_Github #35795 [ run ] completed with state SUCCESS. Commit: e7d48d5
/LLM/main/L0_MergeRequest_PR pipeline #27647 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

[Superjomn]

LGTM