<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    Hi Jon,<br>
    <br>
    <div class="moz-cite-prefix">On 2016-06-08 04:06, Jon Masamitsu
      wrote:<br>
    </div>
    <blockquote
      cite="mid:da9677a6-4023-6ce8-4504-f15b984568c5@oracle.com"
      type="cite">
      <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
      Please find the new versions of the webrevs<br>
      <br>
      <a moz-do-not-send="true" class="moz-txt-link-freetext"
        href="http://cr.openjdk.java.net/%7Ejmasa/8026752/webrev.01/">http://cr.openjdk.java.net/~jmasa/8026752/webrev.01/</a><br>
      <a moz-do-not-send="true" class="moz-txt-link-freetext"
        href="http://cr.openjdk.java.net/%7Ejmasa/8026752/top_webrev.01/">http://cr.openjdk.java.net/~jmasa/8026752/top_webrev.01/</a><br>
      <br>
    </blockquote>
    Looks good. I think you could add
    Asserts.assertTrue(wb.metaspaceShouldConcurrentCollect()) before the
    call to System.gc() to make sure there was something to clear. No
    need to see a new webrev if you decide to do that. Reviewed.<br>
    <br>
    Thanks,<br>
    Stefan<br>
    <br>
    <blockquote
      cite="mid:da9677a6-4023-6ce8-4504-f15b984568c5@oracle.com"
      type="cite"> There is a delta (but I'd recommend just looking at
      the complete<br>
      webrevs).<br>
      <br>
      <a moz-do-not-send="true" class="moz-txt-link-freetext"
        href="http://cr.openjdk.java.net/%7Ejmasa/8026752/webrev_delta.00_01/">http://cr.openjdk.java.net/~jmasa/8026752/webrev_delta.00_01/</a><br>
      <br>
      Thanks.<br>
      <br>
      Jon<br>
      <br>
      <div class="moz-cite-prefix">On 5/27/2016 8:17 AM, Jon Masamitsu
        wrote:<br>
      </div>
      <blockquote
        cite="mid:1fa166df-b646-ead9-50ad-672ce7efc267@oracle.com"
        type="cite">
        <meta content="text/html; charset=utf-8"
          http-equiv="Content-Type">
        Stefan,<br>
        <br>
        Thanks for looking at this.  Please see in-line.<br>
        <br>
        <div class="moz-cite-prefix">On 5/27/2016 3:47 AM, Stefan
          Johansson wrote:<br>
        </div>
        <blockquote cite="mid:574825A8.80706@oracle.com" type="cite">
          <meta content="text/html; charset=utf-8"
            http-equiv="Content-Type">
          <tt>Hi Jon,</tt><br>
          <br>
          <div class="moz-cite-prefix">On 2016-05-27 07:46, Jon
            Masamitsu wrote:<br>
          </div>
          <blockquote
            cite="mid:cef5236e-31b9-0c05-f859-b58619f0ef88@oracle.com"
            type="cite">
            <meta http-equiv="content-type" content="text/html;
              charset=utf-8">
            <a moz-do-not-send="true" class="moz-txt-link-freetext"
              href="https://bugs.openjdk.java.net/browse/JDK-8026752">https://bugs.openjdk.java.net/browse/JDK-8026752</a><br>
            <br>
            If a concurrent CMS collection has been scheduled for
            Metaspace<br>
            needs, it should be cancelled if a full GC is done.<br>
            <br>
            <a moz-do-not-send="true" class="moz-txt-link-freetext"
              href="http://cr.openjdk.java.net/%7Ejmasa/8026752/webrev.00/">http://cr.openjdk.java.net/~jmasa/8026752/webrev.00/</a><br>
            <br>
            Fix was verified with the new test
            TestMetaspaceCMSCancel.java<br>
            Stability testing was done with gc_test_suite.<br>
            <br>
          </blockquote>
          <tt>I think the fix is good. Just one question regarding where
            to do the call. Is there a reason that you have put
            MetaspaceGC::set_should_concurrent_collect(false) in
            reset_after_compaction()? To me it would make more sense to
            have the call in CMSCollector::do_compaction_work(...).</tt><tt><br>
          </tt></blockquote>
        <br>
        <tt>No strong reason other than it is the place that occurred to
          me.<br>
          That reset_after_compaction() was added to take care of work<br>
          that needed to be done after a compaction so it seems a good<br>
          place to me.  If do_compaction_work() is a more obvious place
          to<br>
          you, I can move it there.<br>
          <br>
        </tt>
        <blockquote cite="mid:574825A8.80706@oracle.com" type="cite"><tt>
          </tt><tt><br>
          </tt><tt>Regarding the test, I think it is a bit unfortunate
            to have to sleep for 20s to verify this, especially since
            there have been efforts to shorten the test-time. What do
            you think about adding a WhiteBox-method for checking the
            value of MetaspaceGC::_should_concurrent_collect and have
            the test be something like:</tt><tt><br>
          </tt></blockquote>
        <blockquote cite="mid:574825A8.80706@oracle.com" type="cite"><tt>
          </tt><tt>assertTrue(WB.metaspaceShouldConcurrentCollect());</tt><tt><br>
          </tt><tt>System.gc();</tt><tt><br>
          </tt><tt>assertFalse(WB.metaspaceShouldConcurrentCollect());</tt><tt><br>
          </tt></blockquote>
        <br>
        <tt>Adding -XX:CMSWaitDuration=5000 to the command line makes it
          less likely that<br>
          a CMS concurrent collections starts and finishes before the
          call to System.gc().<br>
          The small MetaspaceSize threshold may have been exceeded
          during loading of<br>
          the systems classes (i.e., before the test program started
          running).  Adding the<br>
          pause() gives CMS time to start and complete a concurrent
          collection if the<br>
          concurrent collection is not short-circuited by the fix.  That
          is, I give the<br>
          test a better chance of failing.<br>
          <br>
          I don't think it is quite enough to check MetaspaceGC
          _should_concurrent_collect<br>
          because the concurrent collection could have already started.<br>
          <br>
          I could change the test to check the state of the CMS
          collector.  It should be<br>
          in "Resetting" or "Idling".  I'll look into that.<br>
          <br>
          Thanks again.<br>
          <br>
          Jon<br>
          <br>
        </tt>
        <blockquote cite="mid:574825A8.80706@oracle.com" type="cite"><tt>
          </tt><tt><br>
          </tt><tt>Doing this you should be able to avoid using the
            process builder as well.</tt><tt><br>
          </tt><tt><br>
          </tt><tt>Thanks,</tt><tt><br>
          </tt><tt>Stefan</tt><br>
          <blockquote
            cite="mid:cef5236e-31b9-0c05-f859-b58619f0ef88@oracle.com"
            type="cite"> Thanks.<br>
            <br>
            Jon<br>
          </blockquote>
          <br>
        </blockquote>
        <br>
      </blockquote>
      <br>
    </blockquote>
    <br>
  </body>
</html>