<html>
  <head>
    <meta content="text/html; charset=windows-1252"
      http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    Hi all!<br>
    <br>
    I've changed tests mentioned: VM flavor and TieredCompilation
    options removed since these are rotated values in the testcycle.<br>
    Please see<br>
    <a class="moz-txt-link-freetext" href="http://cr.openjdk.java.net/~iignatyev/eistepan/8038098/webrev.01/">http://cr.openjdk.java.net/~iignatyev/eistepan/8038098/webrev.01/</a><br>
    <br>
    Thanks,<br>
    Jane<br>
    <br>
    <div class="moz-cite-prefix">On 16.09.2014 19:57, Vladimir Kozlov
      wrote:<br>
    </div>
    <blockquote cite="mid:54185DDD.60404@oracle.com" type="cite">On
      9/16/14 4:51 AM, Filipp Zhinkin wrote:
      <br>
      <blockquote type="cite">Igor, Vladimir,
        <br>
        <br>
        RTM tests will be skipped if the class of JVM used for a test
        execution is not 'server'.
        <br>
        Such check is done before the '-server' is passed in
        RTMTestBase, so '-server'
        <br>
        removal won't increase coverage.
        <br>
      </blockquote>
      <br>
      Okay, it make sense to have it here then.
      <br>
      I agree that it does not make sense to run RTM tests  with client
      VM.
      <br>
      <br>
      Vladimir
      <br>
      <br>
      <blockquote type="cite">
        <br>
        If we'll remove it, then I don't think that it will affect test
        execution:
        <br>
        if JVM class was not explicitly specified on the command line,
        then
        <br>
        a java launcher will choose it depending on available physical
        memory
        <br>
        and available processors count. So if JVM was 'server' when
        we're checking
        <br>
        its class, then it's unlikely that JVM started in RTMTestBase
        will become 'client'.
        <br>
        <br>
        But if we'll leave '-server' then tests continue to work as
        expected even
        <br>
        if a java launcher will start choosing JVM class using some
        other policy.
        <br>
        <br>
        Thanks,
        <br>
        Filipp.
        <br>
        <br>
        <br>
        On 09/16/2014 02:35 PM, Igor Ignatyev wrote:
        <br>
        <blockquote type="cite">Vladimir,
          <br>
          I doubt that we can remove -server from
          compiler/testlibrary/rtm/RTMTestBase.java. AFAIR, it requires
          -server, since
          <br>
          it uses server only flag 'UseRTMLocking'
          <br>
          <br>
          Filipp,
          <br>
          can you please verify it?
          <br>
          <br>
          Igor
          <br>
          <br>
          On 09/15/2014 09:42 PM, Vladimir Kozlov wrote:
          <br>
          <blockquote type="cite">Hi Evgeniya
            <br>
            <br>
            What about LoadWithMask.java test?
            <br>
            <br>
            Please, simple remove '-server' option. We always do testing
            with server
            <br>
            JVM so we can't miss the bug if it returns. We did go
            through such
            <br>
            cleanup in tests before and did not have any problems.
            <br>
            <br>
            Flag -XX:-TieredCompilation may affect bug reproduction.
            That is why we
            <br>
            specify it sometimes.
            <br>
            In general we can remove it too since we run Nightly in both
            <br>
            configurations with and without Tiered. But since it does
            not affect
            <br>
            what JVM is used we keep it.
            <br>
            <br>
            In compiler tests we should not specify neither -server or
            -client to
            <br>
            make sure the correct JVM is tested. This is one of our
            rules when we
            <br>
            add tests. Recently we added tests which have -server
            specified:
            <br>
            <br>
            compiler/IntegerArithmetic/TestIntegerComparison.java
            <br>
            compiler/testlibrary/rtm/RTMTestBase.java
            <br>
            compiler/uncommontrap/UncommonTrapStackBang.java
            <br>
            <br>
            compiler/stable/* test have  @run command with both -server
            and -client
            <br>
            configurations.
            <br>
            <br>
            All these tests should be cleaned up.
            <br>
            <br>
            Regards,
            <br>
            Vladimir
            <br>
            <br>
            On 9/15/14 7:11 AM, Evgeniya Stepanova wrote:
            <br>
            <blockquote type="cite">Hi Filipp,
              <br>
              <br>
              I want to keep first string "as is" because it is a
              regression test
              <br>
              and if we just remove -server option, it is possible
              <br>
              that bug have returned, but we won't find it before
              release because
              <br>
              options will not compose in the right way.
              <br>
            </blockquote>
            <br>
            <br>
            <br>
            <blockquote type="cite">The test is very fast, so the second
              run will not take significant
              <br>
              time, but we will be sure that there is no regression.
              <br>
              <br>
              As for C2-specific options, I agree. I've removed -client
              and
              <br>
              TieredCompilation options to use jtreg submitted.
              <br>
              <br>
              Please see updated diff:
              <br>
<a class="moz-txt-link-freetext" href="http://cr.openjdk.java.net/~iignatyev/eistepan/8038098/webrev.00/">http://cr.openjdk.java.net/~iignatyev/eistepan/8038098/webrev.00/</a>
              <br>
              <br>
              Thanks,
              <br>
              Jane
              <br>
              On 15.09.2014 16:01, Filipp Zhinkin wrote:
              <br>
              <blockquote type="cite">Hi Jane,
                <br>
                <br>
                mail's summary does not match the bug's name.
                <br>
                <br>
                As about your fix: I thought that such configs as
                client/server
                <br>
                and tired/non-tired should be rotated in nightly, aren't
                it?
                <br>
                <br>
                Maybe it does more sense to just remove '-server'
                option?
                <br>
                With '-client' the test may fail during test runs that
                were started
                <br>
                with some C2-specific options, like UseRTMLocking.
                <br>
                <br>
                Thanks,
                <br>
                Filipp.
                <br>
                <br>
                On 09/15/2014 03:42 PM, Evgeniya Stepanova wrote:
                <br>
                <blockquote type="cite">Hi all!
                  <br>
                  <br>
                  Please review fix for 8038098
                  <br>
                  <br>
                  bug: <a class="moz-txt-link-freetext" href="https://bugs.openjdk.java.net/browse/JDK-8038098">https://bugs.openjdk.java.net/browse/JDK-8038098</a>
                  <br>
                  <br>
                  Problem: Since we have inlining in C1 now, we should
                  check that
                  <br>
                  issue 8031743 will not appear in C1.
                  <br>
                  Solution: Added one more run in test for C1 inlining
                  checking
                  <br>
                  <br>
                  A am also need someone to push changes.
                  <br>
                  <br>
                  Diff is in the attachments.
                  <br>
                  I've also included diff to this letter since it is
                  very small.
                  <br>
                  <br>
                  <br>
                  Thanks,
                  <br>
                  Jane
                  <br>
                  <br>
                  diff -r 087443edb74a
                  test/compiler/codegen/LoadWithMask2.java
                  <br>
                  --- a/test/compiler/codegen/LoadWithMask2.java    Mon
                  Sep 15
                  <br>
                  08:08:22 2014 +0200
                  <br>
                  +++ b/test/compiler/codegen/LoadWithMask2.java    Mon
                  Sep 15
                  <br>
                  14:35:50 2014 +0400
                  <br>
                  @@ -26,6 +26,7 @@
                  <br>
                    * @bug 8031743
                  <br>
                    * @summary loadI2L_immI broken for negative memory
                  values
                  <br>
                    * @run main/othervm -server -Xbatch
                  -XX:-TieredCompilation
                  <br>
                  -XX:CompileCommand=compileonly,*.foo* LoadWithMask2
                  <br>
                  + * @run main/othervm -client -Xbatch
                  -XX:+TieredCompilation
                  <br>
                  -XX:CompileCommand=compileonly,*.foo* LoadWithMask2
                  <br>
                    *
                  <br>
                    */
                  <br>
                   public class LoadWithMask2 {
                  <br>
                </blockquote>
                <br>
              </blockquote>
              <br>
              --
              <br>
              /Evgeniya Stepanova/
              <br>
            </blockquote>
          </blockquote>
        </blockquote>
        <br>
      </blockquote>
    </blockquote>
    <br>
    <div class="moz-signature">-- <br>
      <i>Evgeniya Stepanova</i></div>
  </body>
</html>