Summary: After JDK-8035400 there were several opportunities to fix the visibility of several members of the G1ParScanThreadState class.
Reviewed-by: brutisso, mgerdin

Summary: Extract the G1ParScanThreadState class from G1CollectedHeap.?pp into its own files.
Reviewed-by: brutisso, mgerdin

Summary: Clean up usage of idx_t and uintptr_t when using it in conjunction with BitMap::set_map(), casting to the appropriate type. Fixes compilation on S390.
Reviewed-by: tschatzl
Contributed-by: NDan Horak <dhorak@redhat.com>

Summary: A few locations in the code use sizeof(this) which returns the size of the pointer instead of sizeof(classname) which returns the size of the sum of its members. This change fixes these errors and adds a few tests.
Reviewed-by: mgerdin, brutisso

Summary: Card re-dirtying code for verification and actual redirtying uses two different, almost completely identical card closures. Also the verification code still assumes a perm gen.
Reviewed-by: brutisso, jmasa

Summary: Parallelize card redirtying to decrease the time it takes.
Reviewed-by: brutisso

Summary: G1 does not retire allocation buffers after reference processing work when -XX:+ParallelRefProcEnabled is enabled. This causes wrong calculation of PLAB sizes, as the amount of space wasted is not updated correctly.
Reviewed-by: brutisso

Summary: The mentioned method is never used and out of date. So it is removed.
Reviewed-by: mgerdin, brutisso

Summary: After changes to the PerRegionTable where all these PRTs are linked together in an additional field, simplify iterating over all PRTs by using these links instead of walki
Reviewed-by: mgerdin, jwilhelm, brutisso

Summary: Instead of using a manually managed array for the in_cset_fast_test array, use a G1BiasedArray instance.
Reviewed-by: brutisso, mgerdin

8038930: G1CodeRootSet::test fails with assert(_num_chunks_handed_out == 0) failed: No elements must have been handed out yet
Summary: The test incorrectly assumed that it had been started with no other previous compilation activity. Fix this by allowing multiple code root free chunk lists, and use one separate from the global one to perform the test.
Reviewed-by: brutisso

Reviewed-by: brutisso, sjohanss

Summary: Removed all write barriers during reference processing and added explicit write barriers when iterating through the discovered list.
Reviewed-by: pliden, jmasa, tschatzl

Reviewed-by: tschatzl, mgerdin

Reviewed-by: brutisso, tschatzl

Reviewed-by: brutisso, ehelin

Reviewed-by: twisti, stefank

8043029: Change 8037816 breaks HS build with older GCC versions which don't support diagnostic pragmas
8043164: Format warning in traceStream.hpp
Summary: Backport of main fix + two corrections, enables clang compilation, turns on format attributes, corrects/mutes warnings
Reviewed-by: kvn, coleenp, iveresov, twisti

Summary: Only prune metaspace virtual spaces at safepoint so walking them is safe outside a safepoint.
Reviewed-by: mgerdin, mgronlun, hseigel, stefank

Reviewed-by: dholmes, kvn, stefank, twisti

Summary: temporarily disable the shutdown of the concurrent GC threads introduced in JDK-8037112
Reviewed-by: brutisso, tschatzl, jmasa

Reviewed-by: brutisso, tschatzl, mgerdin

Reviewed-by: tschatzl, ehelin

Reviewed-by: jwilhelm, ehelin, tschatzl

Reviewed-by: brutisso, mgerdin

Reviewed-by: tschatzl, pliden

Reviewed-by: tschatzl, pliden

Reviewed-by: tschatzl, jwilhelm

Summary: Replaced the verification call to a conditional one that only runs verification in debug builds.
Reviewed-by: brutisso, tschatzl

Reviewed-by: jmasa, coleenp

Summary: Revert JDK-6976350 because it does not improve PLAB fragmentation. To the contrary, it tends to increase the amount of wasted space with many threads.
Reviewed-by: brutisso

Summary: Move G1ParCopyClosure::copy_to_survivor_space to decrease code size.
Reviewed-by: stefank, jmasa

Summary: Remove the assumption that G1CollectedHeap::in_cset_fast_test needs to check for NULL references. Most of the time this is not required, making the code doing this check multiple times.
Reviewed-by: stefank, mgerdin, jmasa

Summary: Move methods that are not dependent on any of G1ParCopyClosure's template parameters into G1ParCopyHelper. Further remove unused methods and members of the class hierarchy.
Reviewed-by: mgerdin, stefank

Summary: When string and symbol table unlink are not performed in parallel, the claim index we check is not updated, and so a guarantee fails. Take this into account when checking the guarantee.
Reviewed-by: brutisso, jwilhelm

8033106: Wrong predicate for checking whether the correct amount of symbol table entries have been processed in G1
Summary: The change fixes the predicate check.
Reviewed-by: jmasa, tonyp, stefank

Summary: Clean up and slightly optimize reference handling from the GC reference task queue. Since we never push partial array chunks as narrowOop* we can manually specialize the code so that some code can be optimized away.
Reviewed-by: tonyp, brutisso, stefank

Summary: Made the flags Min- and MaxHeapFreeRatio manageable, and implemented support for these flags in ParallelGC.
Reviewed-by: sla, mgerdin, brutisso

Summary: Pad and cache-align BiasedMappedArray instances by default to avoid performance variability problems due to false sharing, as instances of this data structures are typically used for performance sensitive code.
Reviewed-by: brutisso, stefank

Summary: Mentioned closures are actually wrapped methods. This adds confusion to readers, and in this case also increases code size as G1ParScanHeapEvacClosure is part of the oop_oop_iterate() methods. Move them into G1ParScanThreadState as methods.
Reviewed-by: stefank