RFR: several small changes including 8207040: [lworld] TestLWorld fails with "Meet Not Symmetric" with -Xcomp

[Tobias Hartmann]

On 13.07.2018 16:56, Tobias Hartmann wrote:
> I think we do reexecute the monitorenter because the uncommon trap

Sorry, I've pressed enter before finishing that sentence: I meant that it seems like we *are*
re-executing the monitorenter because the uncommon trap is created before the monitorenter is parsed
and with the obj still on the stack (see GraphKit::gen_value_type_guard()). Without that we hit:

#  Internal Error (/oracle/valhallaL/open/src/hotspot/share/opto/graphKit.cpp:2080), pid=3504, tid=3521
#  assert(sp() >= inputs) failed: must have enough JVMS stack to execute monitorenter: sp=0, inputs=1

Best regards,
Tobias