The following PRT job to c2_baseline has broken 5 MM_REGRESSION tests:

    java/lang/management/ThreadMXBean/FindDeadlocks.java
    java/lang/management/ThreadMXBean/GetAllLockedSynchronizers.java
    java/lang/management/ThreadMXBean/Locks.java
    java/lang/management/ThreadMXBean/ThreadBlockedCount.java
    java/lang/management/ThreadMXBean/ThreadStateTest.java

Original workspace:     lukasiewicz:/export/lukasiewicz1/jrose/mustang-intrinsics
Submitter:              jrose
Archived data:          /net/prt-archiver.sfbay/data/archived_workspaces/main/c2_baseline/2005/20050622152338.jrose.mustang-intrinsics/
Webrev:                 http://analemma.sfbay.sun.com/net/prt-archiver.sfbay/data/archived_workspaces/main/c2_baseline/2005/20050622152338.jrose.mustang-intrinsics/workspace/webrevs/webrev-2005.06.22/index.html

Partial 6281377: provide Unsafe lazy field set methods
This putback completes VM (but not JDK) changes necessary for 6281377.
There will be coordinated changes to JDK code in a few weeks.

Reviewed by: never, kbr

Fix verified (y/n): y

Verification testing:
 - Inspected PrintOptoAssembly for the expected intrinsic code.
 - Doug Lea has run library units for the new intrinsic.
[end jrose Wed Jun 22 20:08:01 2005 PDT]

Here is my analysis log entry:

New MM_REGRESSION failures (from 2005.06.24)
*   java/lang/management/ThreadMXBean/FindDeadlocks.java
*   java/lang/management/ThreadMXBean/Locks.java
*   java/lang/management/ThreadMXBean/ThreadBlockedCount.java
*   java/lang/management/ThreadMXBean/ThreadStateTest.java
        These tests failed due to an UnsatisfiedLinkError: park and/or
        UnsatisfiedLinkError: unpark on Linux IA32, on Solaris SPARC,
        Solaris X86 and Win32 Client VMs and Linux AMD64, Linux IA32,
        Solaris AMD64, Solaris SPARC, Solaris SPARC-64, Solaris X86
        and Win-AMD64 Server VMs. These failures have also been
        observed on Win32 Server VM.

        I verified that these tests passed with the previous JDK/VM:
        JDK-B41/20050623093904.dcubed.service_hs_merge; looks like we
        picked up the failures from either JDK-B42 or:
        20050623144856.kbr.c2_to_main_baseline

        Update: See update entry below.

*   java/lang/management/ThreadMXBean/GetAllLockedSynchronizers.java
        This test failed due to an IllegalMonitorStateException on
        Linux IA32, Solaris SPARC, Solaris X86 and Win32 Client VMs and
        Linux AMD64, Linux IA32, Solaris AMD64, Solaris SPARC, Solaris
        SPARC-64, Solaris X86 and Win-AMD64 Server VMs. This failure
        has also been observed on Win32 Server VM.

        I verified that these tests passed with the previous JDK/VM:
        JDK-B41/20050623093904.dcubed.service_hs_merge; looks like we
        picked up the failures from either JDK-B42 or:
        20050623144856.kbr.c2_to_main_baseline

        Update: I reran these 5 tests using a Mustang-B42 JDK with the
            Server VM:
                B42 VM                                 - no failures
                20050623063809.sbohne.rt_merge         - no failures
                20050623144856.kbr.c2_to_main_baseline - all 5 failed
            now from c2_baseline jobs:
                20050623143829.sgoldman.6288902         - all 5 failed
                20050623142908.never.c2                 - all 5 failed
                20050623135529.jrose.mustang-intrinsics - all 5 failed
                20050623095333.sgoldman.6281185         - all 5 failed
                20050622194805.kbr.c2_baseline          - all 5 failed
                20050622152338.jrose.mustang-intrinsics - all 5 failed
                20050622122040.kbr.main_to_c2_baseline  - all 5 passed

            It looks like the failure was introduced with John Rose's
            PRT job 20050622152338.jrose.mustang-intrinsics.


###@###.### 2005-06-28 18:07:43 GMT