Skip to main content
Redhat Developers  Logo
  • Products

    Featured

    • Red Hat Enterprise Linux
      Red Hat Enterprise Linux Icon
    • Red Hat OpenShift AI
      Red Hat OpenShift AI
    • Red Hat Enterprise Linux AI
      Linux icon inside of a brain
    • Image mode for Red Hat Enterprise Linux
      RHEL image mode
    • Red Hat OpenShift
      Openshift icon
    • Red Hat Ansible Automation Platform
      Ansible icon
    • Red Hat Developer Hub
      Developer Hub
    • View All Red Hat Products
    • Linux

      • Red Hat Enterprise Linux
      • Image mode for Red Hat Enterprise Linux
      • Red Hat Universal Base Images (UBI)
    • Java runtimes & frameworks

      • JBoss Enterprise Application Platform
      • Red Hat build of OpenJDK
    • Kubernetes

      • Red Hat OpenShift
      • Microsoft Azure Red Hat OpenShift
      • Red Hat OpenShift Virtualization
      • Red Hat OpenShift Lightspeed
    • Integration & App Connectivity

      • Red Hat Build of Apache Camel
      • Red Hat Service Interconnect
      • Red Hat Connectivity Link
    • AI/ML

      • Red Hat OpenShift AI
      • Red Hat Enterprise Linux AI
    • Automation

      • Red Hat Ansible Automation Platform
      • Red Hat Ansible Lightspeed
    • Developer tools

      • Red Hat Trusted Software Supply Chain
      • Podman Desktop
      • Red Hat OpenShift Dev Spaces
    • Developer Sandbox

      Developer Sandbox
      Try Red Hat products and technologies without setup or configuration fees for 30 days with this shared Openshift and Kubernetes cluster.
    • Try at no cost
  • Technologies

    Featured

    • AI/ML
      AI/ML Icon
    • Linux
      Linux Icon
    • Kubernetes
      Cloud icon
    • Automation
      Automation Icon showing arrows moving in a circle around a gear
    • View All Technologies
    • Programming Languages & Frameworks

      • Java
      • Python
      • JavaScript
    • System Design & Architecture

      • Red Hat architecture and design patterns
      • Microservices
      • Event-Driven Architecture
      • Databases
    • Developer Productivity

      • Developer productivity
      • Developer Tools
      • GitOps
    • Secure Development & Architectures

      • Security
      • Secure coding
    • Platform Engineering

      • DevOps
      • DevSecOps
      • Ansible automation for applications and services
    • Automated Data Processing

      • AI/ML
      • Data Science
      • Apache Kafka on Kubernetes
      • View All Technologies
    • Start exploring in the Developer Sandbox for free

      sandbox graphic
      Try Red Hat's products and technologies without setup or configuration.
    • Try at no cost
  • Learn

    Featured

    • Kubernetes & Cloud Native
      Openshift icon
    • Linux
      Rhel icon
    • Automation
      Ansible cloud icon
    • Java
      Java icon
    • AI/ML
      AI/ML Icon
    • View All Learning Resources

    E-Books

    • GitOps Cookbook
    • Podman in Action
    • Kubernetes Operators
    • The Path to GitOps
    • View All E-books

    Cheat Sheets

    • Linux Commands
    • Bash Commands
    • Git
    • systemd Commands
    • View All Cheat Sheets

    Documentation

    • API Catalog
    • Product Documentation
    • Legacy Documentation
    • Red Hat Learning

      Learning image
      Boost your technical skills to expert-level with the help of interactive lessons offered by various Red Hat Learning programs.
    • Explore Red Hat Learning
  • Developer Sandbox

    Developer Sandbox

    • Access Red Hat’s products and technologies without setup or configuration, and start developing quicker than ever before with our new, no-cost sandbox environments.
    • Explore Developer Sandbox

    Featured Developer Sandbox activities

    • Get started with your Developer Sandbox
    • OpenShift virtualization and application modernization using the Developer Sandbox
    • Explore all Developer Sandbox activities

    Ready to start developing apps?

    • Try at no cost
  • Blog
  • Events
  • Videos

Shenandoah GC in JDK 14, Part 1: Self-fixing barriers

March 4, 2020
Roman Kennke
Related topics:
JavaDeveloper Tools

Share:

    The development of the Shenandoah Garbage Collector (GC) in the upcoming JDK 14 has seen significant improvements. The first one covered here (self-fixing barriers) aims to reduce local latencies that are spent in barrier mid- and slow paths. The second will cover concurrent root processing and concurrent class unloading.

    Self-fixing barriers

    The self-fixing barriers improvement builds on the load reference barriers that went into JDK 13. A load reference barrier is employed after a load from a reference field or array element and before the loaded object is given to the rest of the application code. In pseudocode, the barrier looks like this:

    T load_reference_barrier(T* addr) {
      T obj = *addr;
    
      // Fast-path: accesses thread-local flag
      if (!is_gc_active()) return obj;
    
      // Mid-path 1: accesses small bitmap (byte per region, handful of KBs?)
      if (!is_in_collection_set(obj)) return obj;
    
      // Mid-path 2: accesses fwdptrs in object (entire heap?)
      T fwd = resolve_forwardee(obj);
      if (obj != fwd) return fwd;
    
      // Slow-path: call to runtime, once per location with non-forwarded object
      return load_reference_barrier_slowpath(obj);
    }
    

    The result is that whenever we load an object while the GC is active, and if that object is a reference into the collection set, we dive into resolving the object and possibly into the slow parts of the runtime to do the actual evacuation. Chances are that the object would be evacuated by the GC itself, and the barrier would discover this at the Mid-path 2 check, and then return from there. In the worst case, the barrier would call the runtime and do the whole thing.

    The performance-sensitive part of this story is that we would go to the Mid-path 2 check all of the time for relocated objects until the GC cycle is over and GC updates the interesting references. If we do the access in a hot loop, we always walk deep into the barrier while GC is running. And the Mid-path 2 check is expensive because it reaches far and wide.

    The idea behind self-fixing barriers is that when we have resolved the object and discovered the forwarded copy, we can just as well update the location right there. Since we are updating the reference to the object copy not in the collection set, on the next barrier invocation we would exit from Mid-path 1.

    This is what the barrier looks like with those changes:

    T load_reference_barrier(T* addr) {
      T obj = *addr;
    
      // Fast-path: accesses thread-local flag
      if (!is_gc_active()) return obj;
    
      // Mid-path 1: accesses small bitmap (byte per region, handful of KBs?)
      if (!is_in_collection_set(obj)) return obj;
    
      // Mid-path 2: accesses fwdptrs in objects (entire heap?)
      T fwd = resolve_forwardee(obj);
    
      if (obj != fwd) {
        // Can do the update here
        CAS(addr, fwd, obj);
        return fwd;
      }
    
      // Slow-path: call to runtime, once per location with non-forwarded object
      fwd = load_reference_barrier_slowpath(obj);
      // Can do the update here
      CAS(addr, fwd, obj);
      return fwd;
    }
    

    In other words, as soon as we get the forwardee, we dive into the slow path, where we can stamp the reference to the new copy back into the original address. We do this using a compare-and-set operation in order to avoid a potentially racing update of the same field by another Java thread that we must not override.

    Now, notice that we only fail the Mid-path 2 check once per non-updated location. When we fail this check, we can fix it and then never visit the darker parts of the barrier again. With that in mind, we can simplify the barrier by moving the entire update into the slow path itself:

    T load_reference_barrier(T* addr) {
      T obj = *addr;
    
      // Fast-path: accesses thread-local flag
      if (!is_gc_active()) return obj;
    
      // Mid-path 1: accesses small bitmap (byte per region, handful of KBs?)
      if (!is_in_collection_set(obj)) return obj;
    
      // Slow-path: call to runtime, once per non-updated location
      return load_reference_barrier_slowpath(obj, addr); // Update is actually here
    }
    

    Now we have simpler mutator-side barriers. The complication is in passing addr to the slow path, which requires fiddling with the interpreter, C1, and C2—this is why it is not done this way from the beginning. Also, notice the caveat. For non-updated locations, we used to exit earlier from the Mid-path 2 check. Now, we enter the runtime for them. While this behavior looks worse in the code, doing many Mid-path 2 checks for hot objects that are in the collection set is much more expensive than entering the runtime for fix-up once per location.

    These mechanics also leave less work to do for the GC workers during the update-references phase later.

    Conclusion

    Hopefully, you now have a better sense of how the self-fixing barriers in JDK 14's Shenandoah GC release can help to reduce local latencies in barrier mid- and slow paths. In the next article, I will cover concurrent root processing and concurrent class unloading. Together, these features reduce GC pause times, and thus global latencies, by moving GC work from the pause to the concurrent phase.

    Last updated: June 29, 2020

    Recent Posts

    • Container starting and termination order in a pod

    • More Essential AI tutorials for Node.js Developers

    • How to run a fraud detection AI model on RHEL CVMs

    • How we use software provenance at Red Hat

    • Alternatives to creating bootc images from scratch

    Red Hat Developers logo LinkedIn YouTube Twitter Facebook

    Products

    • Red Hat Enterprise Linux
    • Red Hat OpenShift
    • Red Hat Ansible Automation Platform

    Build

    • Developer Sandbox
    • Developer Tools
    • Interactive Tutorials
    • API Catalog

    Quicklinks

    • Learning Resources
    • E-books
    • Cheat Sheets
    • Blog
    • Events
    • Newsletter

    Communicate

    • About us
    • Contact sales
    • Find a partner
    • Report a website issue
    • Site Status Dashboard
    • Report a security problem

    RED HAT DEVELOPER

    Build here. Go anywhere.

    We serve the builders. The problem solvers who create careers with code.

    Join us if you’re a developer, software engineer, web designer, front-end designer, UX designer, computer scientist, architect, tester, product manager, project manager or team lead.

    Sign me up

    Red Hat legal and privacy links

    • About Red Hat
    • Jobs
    • Events
    • Locations
    • Contact Red Hat
    • Red Hat Blog
    • Inclusion at Red Hat
    • Cool Stuff Store
    • Red Hat Summit

    Red Hat legal and privacy links

    • Privacy statement
    • Terms of use
    • All policies and guidelines
    • Digital accessibility

    Report a website issue