RFR (S) 8222893: markOopDesc::print_on() is a bit confused

[Daniel D. Daugherty]

Thanks for taking care of this bug.


On 5/3/19 11:31 AM, coleen.phillimore at oracle.com wrote:
> Summary: Add print_on for ObjectMonitor and make markOop printing 
> sensible and add test.
>
> Testing with mach5 ongoing, but tested locally.
>
> open webrev at http://cr.openjdk.java.net/~coleenp/2019/8222893.01/webrev

src/hotspot/share/oops/klass.cpp
     old L737:   ResourceMark rm;
         Why delete the ResourceMark?

     L744:      st->cr();
         As long as there are no tests that depend on this 'WizardMode'
         output style, this should be okay. (Please search the tests
         for use of WizardMode'.

src/hotspot/share/oops/markOop.cpp
     L47:     if (is_neutral()) {   // 001 biased bit in 3rd right bit
         Previously your comment was '// last bits = ???'
         I think that was a better style.

     L54:     } else if (has_bias_pattern()) {  // 101
         Previously your comment was '// last bits = ???'
         I think that was a better style.

src/hotspot/share/runtime/objectMonitor.cpp
     No comments.

src/hotspot/share/runtime/objectMonitor.hpp
     No comments.

test/hotspot/gtest/oops/test_markOop.cpp
     L67:     // Notify gets the lock inflated
         Actually it is wait() that gets the lock inflated.

     L88:   assert_test_pattern(h_obj, "is_biased");
         A comment about why a newly created object would be biased
         would help here. Something like:

         // Biased locking is enabled for java.lang.Object:

     There are some gtest pieces here that I'm not familiar with,
     but the test looks good to me. Obviously someone with more
     gtest experience should also review.

Thumbs up.

Dan

> bug link https://bugs.openjdk.java.net/browse/JDK-8222893
>
> Thanks,
> Coleen