[https://nvbugs/5813452][fix] Fix "Assertion failed: isLeaf() in kvCacheManager.cpp:465"

[Boreas618]

Summary by CodeRabbit

• Bug Fixes
  • Improved internal cache block management consistency to ensure proper resource cleanup during batch processing operations, enhancing stability and reliability of the system.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

Description

This PR fixes the Assertion failed: isLeaf() in kvCacheManager.cpp:465 issue observed when the MAX_UTILIZATION capacity scheduler is enabled.

In the updated code path, storeBlocks did not handle the case where a block already has a previous block (PB). If the block key is recomputed and the block is moved under a new parent (NB), PB can retain a stale reference to the child under the old blockKey. Later, when freeChildren() is called on PB, the stale entry (keyed by the old blockKey) is not removed. As a result, PB’s mNextBlocks is not empty and the isLeaf() assertion fails.

This is most likely triggered under MAX_UTILIZATION because pauses/resumes happen frequently. When a request is paused with its last KV block only partially filled, PB holds the child under the old block key. When the request resumes and the block becomes fully populated, the block key changes in the new code path, creating an inconsistency between PB’s mNextBlocks and the child’s updated key/parent linkage.

Based on my testing, the change in this PR addresses the stale reference and resolves the assertion failure.

Relevant issues / NV bugs

• [Bug]: Assertion failed: isLeaf() in kvCacheManager.cpp:465 during inference #10690
• [Bug]: Crash when enabling MAX_UTILIZATION #9931
• https://nvbugspro.nvidia.com/bug/5821094
• https://nvbugspro.nvidia.com/bug/5813452
• https://nvbugspro.nvidia.com/bug/5775544
• https://nvbugspro.nvidia.com/bug/5747903

Test Coverage

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.

[coderabbitai]

📝 Walkthrough

Walkthrough

The change modifies the storeBlocks method in WindowBlockManager to properly clean up block linkages when inserting non-matching blocks. When no existing matching block is found, the code now explicitly removes the previous block's connection to the current block via removeNextBlock before establishing new linkage, ensuring the radix-tree-like structure remains consistent.

Changes

Cohort / File(s)	Summary
Block linkage cleanup cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp	Modified WindowBlockManager::storeBlocks to call removeNextBlock on the previous block before creating new block linkages when no matching block exists, preventing orphaned successor references

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 follows the required template with NVBugs ID, type tag [fix], and a clear description of the assertion being fixed.
Description check	✅ Passed	PR description is comprehensive with clear problem statement, root cause analysis, and multiple linked issues, though Test Coverage section is empty.

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

✨ Finishing touches

• 📝 Generate docstrings

---

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.}