Now that we only run perf benchmarks on swarming, we should be able to clean up code that are needed for legacy perf waterfall.

Some examples:
* Benchmark's list json command ( issue 740090 )
* Recipe codes, e.g: DynamicPerfTests (https://cs.chromium.org/chromium/build/scripts/slave/recipe_modules/chromium_tests/steps.py?rcl=58d6e7a838e2ca20afbfd0e15e9c3dfe53b05d38&l=699)