[diffusion] fix: handle disabled component residency in executor

[mickqian]

Motivation

Modifications

Accuracy Tests

Speed Tests and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.

Review and Merge Process

1. Ping Merge Oncalls to start the process. See the PR Merge Process.
2. Get approvals from CODEOWNERS and other reviewers.
3. Trigger CI tests with comments or contact authorized users to do so.
   • Common commands include /tag-and-rerun-ci, /tag-run-ci-label, /rerun-failed-ci
4. After green CI and required approvals, ask Merge Oncalls or people with Write permission to merge the PR.

[mickqian]

/tag-and-rerun-ci

[gemini-code-assist]

Code Review

This pull request introduces null checks for component_residency_manager in the PipelineExecutor class to prevent potential runtime errors. The review feedback identifies a potential AttributeError in begin_component_residency_request when the batch parameter is a list during grouped execution and suggests handling this case in both begin_component_residency_request and before_stage to ensure type consistency.

[gemini-code-assist]

While adding the null check is correct, the batch parameter can be a list of requests when called from execute_group. The component_residency_manager.begin_request method expects a single object satisfying the ResidencyBatch protocol (to access is_warmup), so passing a list will cause an AttributeError. Rebinding batch to its first element when it is a list ensures compatibility with grouped execution.

Suggested change

	if self.component_residency_manager is None:
	return
	if self.component_residency_manager is None:
	return
	if isinstance(batch, list):
	batch = batch[0]

[gemini-code-assist]

Similar to begin_component_residency_request, batch can be a list of requests during grouped execution. Although the current implementation of manager.before_stage does not access batch.is_warmup, it is better to pass the expected type to maintain consistency and avoid potential issues if the manager's implementation changes in the future.

        if self.component_residency_manager is None:
            return
        if isinstance(batch, list):
            batch = batch[0]

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!