RFR: gc interface tweaks

[Aleksey Shipilev]

On 08/31/2018 11:55 AM, Roland Westrelin wrote:
>> After the brief look, this looks good. The usual plan (test in sh/jdk, and then propose upstream)
>> holds for this work, right?
> 
> Thanks for looking at that. Yes, the usual plan applies.

It seems that most of the tier3_gc_shenandoah tests are failing with this patch with:

#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007fc60b732eef, pid=29871, tid=29900
#
# JRE version: OpenJDK Runtime Environment (12.0) (fastdebug build
12-internal+0-adhoc.shade.shenandoah-jdk)
# Java VM: OpenJDK 64-Bit Server VM (fastdebug 12-internal+0-adhoc.shade.shenandoah-jdk, mixed mode,
tiered, compressed oops, shenandoah gc, linux-amd64)
# Problematic frame:
# V  [libjvm.so+0x17baeef]  ShenandoahBarrierSetC2::resolve_address(C2Access&) const+0x1f

Stack: [0x00007fc5e48a7000,0x00007fc5e49a8000],  sp=0x00007fc5e49a3770,  free space=1009k
Native frames: (J=compiled Java code, A=aot compiled Java code, j=interpreted, Vv=VM code, C=native
code)
V  [libjvm.so+0x17baeef]  ShenandoahBarrierSetC2::resolve_address(C2Access&) const+0x1f
V  [libjvm.so+0x6f0408]  BarrierSetC2::load_at(C2Access&, Type const*) const+0x608
V  [libjvm.so+0xe538ee]  GraphKit::access_load(Node*, Type const*, BasicType, unsigned long)+0xee
V  [libjvm.so+0x12a9a74]  LibraryCallKit::inline_native_getClass()+0x34
V  [libjvm.so+0x12c566e]  LibraryIntrinsic::generate(JVMState*)+0xbe
V  [libjvm.so+0xca6cc5]  Parse::do_call()+0x565
V  [libjvm.so+0x161b9aa]  Parse::do_one_bytecode()+0x2dea
V  [libjvm.so+0x1609b52]  Parse::do_one_block()+0x472



-Aleksey