Uploaded image for project: 'JDK'
  1. JDK
  2. JDK-6365316

gamma launcher should create JVM from non-primordial thread

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Not an Issue
    • Icon: P4 P4
    • 7-pool
    • 6
    • hotspot
    • None
    • b59
    • 6
    • generic
    • generic

      After the fix for
           6316197 Java launcher should create JVM from non-primordial thread
      the gamma launcher has to be fixed correspondingly to be consistent.

      Please, see the email thread below:

      >> Do we have a bug ID for the gamma launcher issue?
      >> I would update it with my java launcher (SDK side) webrev.

      AFAIK, there is no bug, but perhaps you could file one ?
      Thanks

      >>
      >> Thanks,
      >> Serguei
      >>
      >> Kumar Srinivasan wrote:
      >>
      >>> Yes, we should fix the java launcher to be consistent with
      >>> real launcher, as Alan surmised its not high priority since its
      >>> used internally, I guess this can wait until dolphin (IMO).
      >>>
      >>>> <included Kumar since he putback the 6316197 fix> :-)
      >>>>
      >>>>
      >>>> Serguei Spitsyn writes:
      >>>> > Alan and Brian,
      >>>> > > I have just found that the gamma launcher still behaves the old way so it does not
      >>>> > run the JavaMain as a separated thread (as in the bug 6316197 fix).
      >>>> > This looks like abother side affect of the bug 6316197 fix Jim was talking about:
      >>>> > 6316197 Java launcher should create JVM from non-primordial thread
      >>>> > > Brian, do we have to submit a bug on this issue or it is Ok
      >>>> > that the gamma launcher does not follow the bug 6316197 fix?
      >>>> > > Thanks,
      >>>> > Serguei
      >>>> > > Alan Bateman wrote:
      >>>> > > > > > AFAIK the gamma launcher is only used internally in the hotspot team so > > I don't think it's a priority. If it makes life easier you just submit a > > bug to remind us to revisit it. Alternatively, you could update it and > > test it manually (by running gamma interactively). I'm not sure if > > automated tests are possible - certainly not for the regression suite in > > the SDK workspace as it doesn't have access to the gamma launcher.
      >>>> > > > > -Alan.
      >>>> > > > > > > Serguei Spitsyn wrote:
      >>>> > > > >> Hi Alan,
      >>>> > >>
      >>>> > >> It looks like Brian thinks that the gamma launcher needs
      >>>> > >> to be fixed correspondingly. Is it worth to do?
      >>>> > >> It is not a problem to fix it, but I have no idea how to test it.
      >>>> > >>
      >>>> > >> Thanks,
      >>>> > >> Serguei
      >>>> > >>
      >>>> > >> Brian Doherty wrote:
      >>>> > >>
      >>>> > >>> Hi Serguei,
      >>>> > >> ...
      >>>> > >>
      >>>> > >>> When I introduced the sun.java.command launcher property, I
      >>>> > >>> also modified the os/linux/launcher/java.c launcher in the
      >>>> > >>> hotspot workspace in order to keep the two launchers nearly
      >>>> > >>> the same. This launcher is known as the 'gamma' launcher,
      >>>> > >>> and people still use it.
      >>>> > >>>
      >>>> > >> ...
      >>>> > >>> Brian

      The fix for the gamma launcher should also pick up another java launcher's fix:
        6351349 On linux with the old thread lib, jps should return the same PID as $!

            zgu Zhengyu Gu
            sspitsyn Serguei Spitsyn
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved:
              Imported:
              Indexed: