RFR: 8242901: Duplicate PSYoung/OldGen max size functions

Kim Barrett kim.barrett at oracle.com
Fri May 8 19:48:16 UTC 2020


Please review this cleanup of ParallelGC max generation size
accessors.  There are some lingering inconsistencies left over after
the removal of UseAdaptiveGCBoundary and associated cleanups.

We had the following:

PSYoungGen: max_size(), gen_size_limit()
Their values were calculated differently but (after removal of
UseAdaptiveGCBoundary) always have the same value.

PSOldGen: max_gen_size(), gen_size_limit()
They were virtual, but that's no longer needed or useful, as there are
no longer any derived classes.  They were both returning the value of
_max_gen_size.

Both classes have min_gen_size(), an accessor for _min_gen_size.

These have been made consistent for both classes:

min_gen_size() -- as before
max_gen_size() -- accessor for _max_gen_size

There are some assertions involving max_gen_size and the reserved size
for the generations that are perhaps redundant or could be tightend
to equality tests, but I've left them as is for now.

CR:
https://bugs.openjdk.java.net/browse/JDK-8242901

Webrev:
https://cr.openjdk.java.net/~kbarrett/8242901/open.00/

Testing:
mach5 tier1-5



More information about the hotspot-gc-dev mailing list