Uploaded image for project: 'JDK'
  1. JDK
  2. JDK-8153224 Monitor deflation prolong safepoints
  3. JDK-8217658

baseline_cleanups from Async Monitor Deflation project

    XMLWordPrintable

    Details

    • Type: Sub-task
    • Status: Resolved
    • Priority: P3
    • Resolution: Fixed
    • Affects Version/s: 13
    • Fix Version/s: 13
    • Component/s: hotspot
    • Labels:
      None
    • Subcomponent:
    • Resolved In Build:
      b06
    • CPU:
      generic
    • OS:
      generic

      Description

      This sub-task captures updates to the baseline code that are not
      directly related to Async Monitor Deflation.

      The cleanups include:

      - comment corrections and clarifications, fix typos
      - implied boolean expression corrections
      - delete/fix whitespace/blank lines
      - remove unecessary "ObjectSynchronizer::" prefix from calls
      - add additional info to existing asserts; this info was useful during
        my debugging of the Async Monitor Deflation code port.
      - use a consistent style for Thread::muxAcquire() id tags
      - ObjectSynchronizer::inflate_helper() return value is always
        ignore by callers so remove it.
      - add missing "else" to logical if-else-statement in
        ObjectSynchronizer::get_lock_owner()
      - guarantee that header() is NULL when
        ObjectSynchronizer::omRelease() is called
      - code cleanup/motion in ObjectSynchronizer::omFlush()
        - cleanup guarantee() and assert() calls
        - move some field resets to NULL to when they are needed
          and group them together
        - reorder global and per-thread code blocks to be the same
          instead of arbitrarily different
        - move some per-thread code out of the gListLock critical
          section since that lock protects the global monitor list
      - update existing 'monitorinflation' logging to:
        - be clear about the originating source of the log mesg
        - use 'Self' in ResourceMark ctrs which is more efficient
      - Replace a set_owner(NULL) call with a set_header(NULL)
        call in the "CASE: neutral" part of ObjectSynchronizer::inflate()
        when the object->cas_set_mark() fails. This is a rare race.
      - Update runtime/logging/MonitorInflationTest.java to match the
        modified output.

      In all of the above cleanups, there are only three real logic changes:

      - add missing "else" to logical if-else-statement in
        ObjectSynchronizer::get_lock_owner()
      - guarantee that header() is NULL when
        ObjectSynchronizer::omRelease() is called
      - Replace a set_owner(NULL) call with a set_header(NULL)
        call in the "CASE: neutral" part of ObjectSynchronizer::inflate()
        when the object->cas_set_mark() fails. This is a rare race.

      The first logic change is simply more efficient. The second and
      third logic changes are related to not leaving a bad or stale header
      value in a free ObjectMonitor. I have not been able to hypothesize
      any bad effects from this stale header value, but by cleaning it up,
      the upcoming ObjectMonitor audit code is more strict.

        Attachments

          Activity

            People

            Assignee:
            dcubed Daniel Daugherty
            Reporter:
            dcubed Daniel Daugherty
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Dates

              Created:
              Updated:
              Resolved: