I was trying to make my young generation small and noticed that if I make the young generation too small, the Universe::print() mis-prints the percentage of the eden that is used.

    telling $ $Deployed/JDK-7/bin/java -showversion -Xmn512k -Xmx10m -XX:+PrintGCDetails SystemGC
    java version "1.7.0_25"
    Java(TM) SE Runtime Environment (build 1.7.0_25-b33)
    Java HotSpot(TM) Server VM (build 23.25-b02, mixed mode)

    [GC [PSYoungGen: 0K->0K(256K)] 326K->326K(9984K), 0.0005796 secs] [Times: user=0.00 sys=0.00, real=0.00 secs]
    [Full GC (System) [PSYoungGen: 0K->0K(256K)] [ParOldGen: 326K->230K(9728K)] 326K->230K(9984K) [PSPermGen: 1714K->1712K(16384K)], 0.0091519 secs] [Times: user=0.01 sys=0.00, real=0.01 secs]
    Heap
     PSYoungGen total 256K, used 0K [0xfa780000, 0xfa800000, 0xfa800000)
      eden space 0K, -2147483648% used [0xfa780000,0xfa780000,0xfa780000)
      from space 256K, 0% used [0xfa780000,0xfa780000,0xfa7c0000)
      to space 256K, 0% used [0xfa7c0000,0xfa7c0000,0xfa800000)
     ParOldGen total 9728K, used 230K [0xf9e00000, 0xfa780000, 0xfa780000)
      object space 9728K, 2% used [0xf9e00000,0xf9e39ba8,0xfa780000)
     PSPermGen total 16384K, used 1719K [0xf5e00000, 0xf6e00000, 0xf9e00000)
      object space 16384K, 10% used [0xf5e00000,0xf5fadea8,0xf6e00000)

Note the "-2147483648%" in there. SystemGC is a tiny program that just calls System.gc().

This is with

    telling $ $Deployed/JDK-7/bin/java -Xinternalversion
    Java HotSpot(TM) Server VM (23.25-b02) for solaris-x86 JRE (1.7.0_25-b33), built on Jun 17 2013 22:30:57 by "" with Sun Studio 12u1

I just tried it with

    java version "1.8.0-ea"
    Java(TM) SE Runtime Environment (build 1.8.0-ea-b100)
    Java HotSpot(TM) Server VM (build 25.0-b42, mixed mode)

and the bug is still there.

The story on my MacOSX box is worse

    carcharor $ $Deployed/JDK-7/bin/java -showversion -Xmn512k -Xmx10m -XX:+PrintGCDetails SystemGC
    java version "1.7.0_25"
    Java(TM) SE Runtime Environment (build 1.7.0_25-b33)
    Java HotSpot(TM) 64-Bit Server VM (build 23.25-b02, mixed mode)

    #
    # A fatal error has been detected by the Java Runtime Environment:
    #
    # SIGSEGV (0xb) at pc=0x000000010187dee8, pid=99365, tid=6403
    #
    # JRE version: 7.0_25-b33
    # Java VM: Java HotSpot(TM) 64-Bit Server VM (23.25-b02 mixed mode bsd-amd64 compressed oops)
    # Problematic frame:
    # V [libjvm.dylib+0x7dee8]
    #
    # Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
    #
    # An error report file with more information is saved as:
    # /Users/pbkessle/Play/Java/hs_err_pid99365.log
    #
    # If you would like to submit a bug report, please visit:
    # http://bugreport.sun.com/bugreport/crash.jsp
    #
    Abort trap

That's with

    carcharor $ $Deployed/JDK-7/bin/java -Xinternalversion
    Java HotSpot(TM) 64-Bit Server VM (23.25-b02) for bsd-amd64 JRE (1.7.0_25-b33), built on Jun 17 2013 21:35:53 by "java_re" with gcc 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00)

I have to go *all the way :-)* to -Xmn756k before it will run, and then it displays the same bug as my Solaris/i586 box

    carcharor $ $Deployed/JDK-7/bin/java -showversion -Xmn756k -Xmx10m -XX:+PrintGCDetails SystemGC
    java version "1.7.0_25"
    Java(TM) SE Runtime Environment (build 1.7.0_25-b33)
    Java HotSpot(TM) 64-Bit Server VM (build 23.25-b02, mixed mode)

    [GC [PSYoungGen: 0K->0K(512K)] 657K->657K(9728K), 0.0013460 secs] [Times: user=0.00 sys=0.00, real=0.00 secs]
    [Full GC (System) [PSYoungGen: 0K->0K(512K)] [ParOldGen: 657K->292K(9216K)] 657K->292K(9728K) [PSPermGen: 2666K->2664K(21504K)], 0.0224600 secs] [Times: user=0.02 sys=0.00, real=0.02 secs]
    Heap
     PSYoungGen total 512K, used 0K [0x00000007fff00000, 0x0000000800000000, 0x0000000800000000)
      eden space 0K, -2147483648% used [0x00000007fff00000,0x00000007fff00000,0x00000007fff00000)
      from space 512K, 0% used [0x00000007fff00000,0x00000007fff00000,0x00000007fff80000)
      to space 512K, 0% used [0x00000007fff80000,0x00000007fff80000,0x0000000800000000)
     ParOldGen total 9216K, used 293K [0x00000007ff600000, 0x00000007fff00000, 0x00000007fff00000)
      object space 9216K, 3% used [0x00000007ff600000,0x00000007ff649698,0x00000007fff00000)
     PSPermGen total 21504K, used 2675K [0x00000007fa400000, 0x00000007fb900000, 0x00000007ff600000)
      object space 21504K, 12% used [0x00000007fa400000,0x00000007fa69cca8,0x00000007fb900000)

I think the problem is in src/share/vm/gc_implementation/shared/mutableSpace.cpp

    237 void MutableSpace::print_short_on( outputStream* st) const {
    238 st->print(" space " SIZE_FORMAT "K, %d%% used", capacity_in_bytes() / K,
    239 (int) ((double) used_in_bytes() * 100 / capacity_in_bytes()));
    240 }

and the problem must be when capacity_in_bytes() is small. In fact, the bottom(), top() and end() of the eden are all the same, so its capacity_in_bytes() really is 0. Bleah, so now it's not a printing problem, it's a space-sizing problem. That code is too complicated for its own good.