[11u] Backport JEP 346 : Promptly Return Unused Committed Memory from G1

[Liang Mao]

Hi All,

The JEP 346 in JDK12 is a very attractive feature which could save
significant memory and is very useful to the customers of Alibaba
 cloud. Jdk11u is the latest LTS version and will exist as a mainstream
version for a long time. I'd like to backport the corresponding patches.

There're total 12 changesets:
3 changesets of features and fixes that JEP346 depends on
3 changesets of code refinements that JEP346 depends on
6 changesets of JEP346 feature implementation and bug fixes

8 changesets can be just applied and only 4 have very few conflicts
needed to be resolved(less than 20 lines). I have listed the changesets
below in a timeline order. 

I have an initial test with these changesets based on jdk-11.0.7+4 on
test/hotspot/jtreg/gc/ and specjbb2015. Everything looks fine except
gc/survivorAlignment/TestPromotionToSurvivor.java failed both with
my build and reference build of jdk-11.0.7+4.

1) https://bugs.openjdk.java.net/browse/JDK-8208669
 GC changes to allow enabling -Wreorder 
No conflict apply.

2) https://bugs.openjdk.java.net/browse/JDK-8071913
 Filter out entries to free/uncommitted regions during iteration 
Conflicts:
    src/hotspot/share/gc/g1/g1RemSet.cpp
Webrev：http://cr.openjdk.java.net/~ddong/liangmao/8071913/webrev/

3) https://bugs.openjdk.java.net/browse/JDK-6490394
 G1: Allow heap shrinking / memory unmapping after reclaiming regions during Remark 
No conflict apply

4) https://bugs.openjdk.java.net/browse/JDK-8213898
 CDS dumping of springboot asserts in G1ArchiveAllocator::alloc_new_region 
No conflict apply

5) https://bugs.openjdk.java.net/browse/JDK-8193312
 Rename VM_CGC_Operation to VM_G1Concurrent 
Conflicts: (vm_operations.hpp already renamed vmOperations.hpp)
src/hotspot/share/runtime/vm_operations.hpp 
Webrev：http://cr.openjdk.java.net/~ddong/liangmao/8193312/webrev/

6) https://bugs.openjdk.java.net/browse/JDK-8214791
 Consistently name gc files containing VM operations 
Conflicts:(vm_operations.hpp already renamed vmOperations.hpp)
    src/hotspot/share/gc/cms/cmsVMOperations.hpp
    src/hotspot/share/gc/g1/g1ConcurrentMark.cpp
Webrev：http://cr.openjdk.java.net/~ddong/liangmao/8214791/webrev/ 

7) https://bugs.openjdk.java.net/browse/JDK-8212657
 Implementation of JDK-8204089 Promptly Return Unused Committed Memory from G1 
No conflict apply

8) https://bugs.openjdk.java.net/browse/JDK-8215120
 32-bit build failures after JDK-8212657 (Promptly Return Unused Committed Memory from G1) 
No conflict apply

9) https://bugs.openjdk.java.net/browse/JDK-8215149
 TestOptionsWithRangesDynamic.java fails after JDK-8215120 
No conflict apply

10) https://bugs.openjdk.java.net/browse/JDK-8215548
 G1PeriodicGCSystemLoadThreshold needs to be a double 
No conflict apply

11) https://bugs.openjdk.java.net/browse/JDK-8216490
Spammy periodic GC log message contains random time stamp with periodic gc disabled
No conflict apply

12) https://bugs.openjdk.java.net/browse/JDK-8218880
 G1 crashes when issuing a periodic GC while the GCLocker is held 
Conflicts:
src/hotspot/share/gc/g1/g1YoungRemSetSamplingThread.cpp (copyright)
build failure:
src/hotspot/share/gc/g1/g1CollectedHeap.cpp
    } else if (GCLocker::should_discard(cause, gc_count_before)) {
      // Return false to be consistent with VMOp failure due to
      // another collection slipping in after our gc_count but before
      // our request is processed.  _gc_locker collections upgraded by
      // GCLockerInvokesConcurrent are handled above and never discarded.
-       return;
+       return false;Webrev：http://cr.openjdk.java.net/~ddong/liangmao/8218880/webrev/ 


PS:
7) JDK-8212657 might have a build error of uninitialized variable:
src/hotspot/share/gc/g1/g1YoungRemSetSamplingThread.cpp
bool G1YoungRemSetSamplingThread::should_start_periodic_gc() {
...
uintx time_since_last_gc;
11) JDK-8216490 will fix this.

Thanks,
Liang