RFR (M): 8238855: Move G1ConcurrentMark flag sanity checks to g1Arguments

[Kim Barrett]

> On Mar 20, 2020, at 4:32 AM, Thomas Schatzl <thomas.schatzl at oracle.com> wrote:
> 
> Hi all,
> 
>  can I have reviews for this change that moves (and deletes duplicate) flag checking from the G1ConcurrentMark class to the other G1 arguments processing?
> 
> Adds a test that checks whether the invariants before/after are still kept.
> 
> CR:
> https://bugs.openjdk.java.net/browse/JDK-8238855
> Webrev:
> http://cr.openjdk.java.net/~tschatzl/8238855/webrev/
> Testing:
> hs-tier1-5 with new test
> 
> Thanks,
>  Thomas

Looks good.

------------------------------------------------------------------------------
src/hotspot/share/gc/g1/g1Arguments.cpp
 107 void G1Arguments::initialize_mark_stack_size() {

Any particular reason for splitting this out into a separate function
that has one caller?  All the preceeding cases in the same caller are
just directly inlined, so this looks kind of out of place.

I don't object to this, but might actually prefer there to be more
like it.  That can wait for a followup.

------------------------------------------------------------------------------