RFR: 8360651: Create OSContainer API for memory limit [v2]

[Stefan Johansson]

On Mon, 8 Sep 2025 17:54:10 GMT, Severin Gehwolf <sgehwolf at openjdk.org> wrote:

>> Please review this small addition to add a new `OSContainer::has_memory_limit()` API (Linux only - as with the entire OSContainer API) in preparation for [JDK-8350596](https://bugs.openjdk.org/browse/JDK-8350596) which proposes to increase the default `MaxRAMPercentage` when this new API returns true. The patch is pretty trivial. It's only the testing which amounts to the most lines in this patch.
>> 
>> Testing:
>> - [x] GHA
>> - [x] Hotspot container tests on x86_64 Linux on cgroup v1 and cgroup v2 (including the new tests).
>> 
>> Thoughts?
>
> Severin Gehwolf has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains four additional commits since the last revision:
> 
>  - Merge branch 'master' into jdk-8360651-mem-limit-api
>  - MemoryLimitTest whitespace fixes.
>  - TestContainerMemory whitespace fixes.
>  - 8360651: Create OSContainer API for memory limit

If you agree with my cleanup suggestion below feel free to incorporate it. Otherwise I'm ok with the change as is.

src/hotspot/os/linux/osContainer_linux.cpp line 81:

> 79:   bool any_mem_cpu_limit_present = false;
> 80:   bool controllers_read_only = cgroup_subsystem->is_containerized();
> 81:   _has_memory_limit = cgroup_subsystem->memory_limit_in_bytes() > 0;

I found those local `bool` a bit confusing and I don't really think they are needed. Played around a bit with also adding a _has_cpu_limit and ended up with the code below. To me that's a bit more readable, but a slightly larger change also adding has_cpu_limit. What do you think?

   * Step 1.) covers the basic in container use-cases. Step 2.) ensures
   * that limits enforced by other means (e.g. systemd slice) are properly
   * detected.
   */
  check_and_set_limits();

  const char *reason;
  if (cgroup_subsystem->is_containerized()) {
    // in-container case
    _is_containerized = true;
    reason = " because all controllers are mounted read-only (container case)";
  } else {
    // We can be in one of two cases:
    //  1.) On a physical Linux system without any limit
    //  2.) On a physical Linux system with a limit enforced by other means (like systemd slice)
    if (has_cpu_limit() || has_memory_limit()) {
      _is_containerized = true;
      reason = " because either a cpu or a memory limit is present";
    } else {
      // Not containerized
      reason = " because no cpu or memory limit is present";
    }
  }
  log_debug(os, container)("OSContainer::init: is_containerized() = %s%s",
                                                            _is_containerized ? "true" : "false",
                                                            reason);
}

void OSContainer::check_and_set_limits() {
  // Check for CPU limit
  if (os::Linux::active_processor_count() != cgroup_subsystem->active_processor_count()) {
    _has_cpu_limit = true;
  }

  // Check for memory limit
  if (cgroup_subsystem->memory_limit_in_bytes() > 0) {
    _has_memory_limit = true;
  }
}

-------------

Marked as reviewed by sjohanss (Reviewer).

PR Review: https://git.openjdk.org/jdk/pull/26020#pullrequestreview-3238630631
PR Review Comment: https://git.openjdk.org/jdk/pull/26020#discussion_r2358359868