[kaffe] PowerPC problem

David.Cowan at apcc.com David.Cowan at apcc.com
Thu Oct 2 06:40:02 PDT 2003


I posted last week about a problem with a powerpc build I'm having.  I have
tried several things to resolve it and have hit a roadblock.

Thread.sleep doesn't work correctly.

When I compile with unix-jthreads a call to sleep never returns.  When I
compile with unix-pthreads the first call to sleep returns immediately
without sleeping and all subsequent calls never return.

I used the java -vmdebug SLOWLOCKS command to help debug the problem.
Following is output from this and the sample program I am running.
This output is with unix-jthreads

public class test {
     /**
     * The main method which gets executed...
     */
     public static void main(String args[])
     {
          System.out.println("sleeping " + Integer.parseInt(args[0]) + "
seconds");
          try
          {
               Thread.sleep(Integer.parseInt(args[0]));
          } catch (Exception e) {}
          System.out.println("done sleeping");
     }

}

_broadcastCond(**lkp=0x1025cae8, th=0x100e2018)
  getHeavyLock(**lkp=0x1025cae8, *lk=0x7fffe880, th=0x100e2018)
    got new lock
  putHeavyLock(**lkp=0x1025cae8, *lk=0x10269f58, th=0x100e2018)
slowUnlockMutex(**lkp=0x1025cae8, where=0x7fffe880, th=0x100e2018)
  getHeavyLock(**lkp=0x1025cae8, *lk=0x10269f59, th=0x100e2018)
    got cached lock
  putHeavyLock(**lkp=0x1025cae8, *lk=(nil), th=0x100e2018)
_waitCond(**lkp=0x1026c8d4, timeout=8, th=0x100e2018)
<-----------------   here is the native call to wait
  getHeavyLock(**lkp=0x1026c8d4, *lk=0x7fffeb70, th=0x100e2018)
    got new lock
  putHeavyLock(**lkp=0x1026c8d4, *lk=0x10269f58, th=0x100e2018)
slowUnlockMutex(**lkp=0x1026c8d4, where=0x7fffeb70, th=0x100e2018)
  getHeavyLock(**lkp=0x1026c8d4, *lk=0x10269f59, th=0x100e2018)
    got cached lock
  putHeavyLock(**lkp=0x1026c8d4, *lk=0x10269f58, th=0x100e2018)





More information about the kaffe mailing list