[https://nvbugs/6337235][test] Fix MX/GMS model loader fixtures

[chienchunhung]

Description

Fix stale MX/GMS ModelLoader unit-test fixtures exposed by the expanded L0 unit-test coverage.

ModelLoader.load() now builds a SourceIdentity for GMS and MX paths. These branch-focused tests stub _load_and_validate_config() with a lightweight SimpleNamespace, so the real fingerprint builder can fail before the tests reach the GMS/MX behavior they are meant to cover. This PR stubs SourceIdentity.from_model_config() in those tests with a fixed real SourceIdentity, and makes the GMS RO backend return the same identity so the strict RO gate still executes successfully.

Reference: https://nvbugs/6337235

Test Coverage

• PYTHONPYCACHEPREFIX=/private/tmp/trtllm_pycache python3 -m py_compile tests/unittest/_torch/pyexecutor/test_model_loader_gms.py tests/unittest/_torch/pyexecutor/test_model_loader_mx.py
• git diff --check
• Commit hooks passed for the touched files. Skipped waive list check and validate-test-lists locally because they crash under this machine's Python 3.9 while parsing Python 3.10 union annotations in scripts/check_test_list.py.

Note: focused pytest was not runnable in this local environment because pytest is not installed for the system Python.

[chienchunhung]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #54892 [ run ] triggered by Bot. Commit: 9688833 Link to invocation

[tensorrt-cicd]

PR_Github #54892 [ run ] completed with state FAILURE. Commit: 9688833
/LLM/main/L0_MergeRequest_PR pipeline #43896 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