EAS Development for Mainline Linux

EAS development for mainline Linux takes place on LKML and Linaro eas-dev mailing lists where proposed patches are discussed with the community. Arm now provides a pre-integrated version of latest EAS patches for mainline, for upstream development use.

EAS mainline is a moving integration branch tracking tip:sched/core (or mainline release candidates) and adding the latest versions of EAS related patch sets targeted at the mainline kernel. The patches may already be under discussion on relevant open source mailing lists or waiting here while their dependencies get resolved and merged upstream.

The integration branch comes as is with synthetic test case testing only. The intention is to have a single branch containing the full EAS picture for mainline Linux kernel developers interested in tracking progress. This EAS mainline integration branch is not suitable for product development. For product, please use the AOSP common kernel.

Source code

EAS mainline integration branches are made available regularly on:

        git://linux-arm.org/linux-power.git

The releases are tagged with the date e.g. -20190310 so old integrations are available for comparison

Latest: integration-20190503

Tree: eas/next/integration

Base: eas/next/integration-base

New changes:

  1. CPU cooling device uses the Energy Model (3rd client next to CFS task scheduler class and CPUfreq governor schedutil)

Features:

  1. Util(ization) clamping, i.e. a mechanism which allows to "clamp" (i.e. filter) the utilization generated by RT and FAIR task within a range defined from user-space.
  2. EAS related tracepoints.

Features now in mainline:

  1. Frequency Invariant Engine (FIE) and CPU Invariant Engine (v4.13/v4.15).
  2. Idle CPU PELT (Per-Entity Load-Tracking) update (v4.17).
  3. Util(ization) Est(imated) (v4.17).
  4. Misfit task, i.e. forcing migration of running tasks that do not fit on the CPU they are currently running on (v4.20).
  5. Runtime scheduler domain flag detection (v4.20).
  6. Per-cpu Energy Model and Energy Aware Scheduling (v5.0).

Supported platforms:

  1. Arm Juno r0 (arm64)
  2. Arm Juno r2 (arm64)
  3. Hikey960 (arm64) (needs out-of-tree patches, provided by the hikey960_mainline topic branch which comes with this integration)
  4. Arm TC2 (arm32)

Next work items:

Work on getting a 100% pass-rate on all supported platforms.

  • Test Results (integration 20190503)
    • Platform Juno r0
    • Test case passed if failure rate < 10%

     

    • EAS behaviour: 

                                                                          pass rate
      EnergyModelWakeMigration: test_slack:               passed  534/534 (100.0%)
      EnergyModelWakeMigration: test_task_placement:      passed  534/534 (100.0%)
      OneSmallTask:             test_slack:               passed  533/533 (100.0%)
      OneSmallTask:             test_task_placement:      passed  533/533 (100.0%)
      RampDown:                 test_slack:               passed  534/534 (100.0%)
      RampDown:                 test_task_placement:      passed  534/534 (100.0%)
      RampUp:                   test_slack:               passed  534/534 (100.0%)
      RampUp:                   test_task_placement:      passed  534/534 (100.0%)
      ThreeSmallTasks:          test_slack:               passed  533/533 (100.0%)
      ThreeSmallTasks:          test_task_placement:      passed  533/533 (100.0%)
      TwoBigTasks:              test_slack:               passed  533/534 ( 99.8%)
      TwoBigTasks:              test_task_placement:      passed  534/534 (100.0%)
      TwoBigThreeSmall:         test_slack:               passed  534/534 (100.0%)
      TwoBigThreeSmall:         test_task_placement:      passed  513/534 ( 96.1%)

    • load tracking: 

                                                                          pass rate
      Invariance:               test_cpu_invariance:      passed  533/533 (100.0%)
      Invariance:               test_freq_invariance:     passed  533/533 (100.0%)
      Invariance:               test_task_load_avg:       passed  533/533 (100.0%)
      Invariance:               test_task_util_avg:       passed  533/533 (100.0%)
      OneTaskCPUMigration:      test_util_task_migration: passed  533/533 (100.0%)
PELTTasksTest:            test_load_avg_behaviour:  passed  533/533 (100.0%)
      PELTTasksTest:            test_load_avg_range:      passed  533/533 (100.0%)
      PELTTasksTest:            test_util_avg_behaviour:  passed  533/533 (100.0%)
      PELTTasksTest:            test_util_avg_range:      passed  533/533 (100.0%)
      TwoTasksCPUMigration:     test_util_task_migration: passed  532/532 (100.0%)
    • misfit: 

                                                                          pass rate
      StaggeredFinishes:        test_preempt_time:        passed  533/534 ( 99.8%)
      StaggeredFinishes:        test_throughput:          passed  527/534 ( 98.7%)
    • other: 

                                                                          pass rate
      CapacitySanity:           test_capacity_sanity:     passed  534/534 (100.0%)
      UserspaceSanity:          test_performance_sanity:  passed  534/534 (100.0%)
    • hotplug: 

                                                                          pass rate
      HotplugTorture:           test_cpus_alive:          passed  534/534 (100.0%)
      HotplugTorture:           test_target_alive:        passed  534/534 (100.0%)