RFR (S) 8140650: Method::is_accessor should cover getters and setters for all types

Aleksey Shipilev aleksey.shipilev at oracle.com
Tue Nov 3 15:20:50 UTC 2015


I would like to have a formal review for a minor nit in
Method::is_accessor. The definition for this method is inconsistent with
its intent: it should accept all accessors, but instead it only  accepts
the specific shapes of getters, and completely ignores setters. See:

This makes compilers to ignore many trivial methods that we might
otherwise inline when all other inline hints have failed. It seems to be
a lingering issue left from interpreters that had the "fast accessors".
While it is an open question should inlining policy treat accessors
differently or not (I stand by "yes, it should"), this is a fix that
makes is_accessors proper:

The only usage for the "old" style is_accessor is Zero, and they would
like to update them after we commit the change:

The patch passes JPRT, RBT (hotspot_all), and the new regression test.

It beats me whether this is a runtime, or compiler change -- JIRA bug
flip-flops on that -- so, sending to hotspot-dev at . I think it can be
pushed via hs-comp, given it impacts compilers mostly, and has the
compiler-specific test.


More information about the hotspot-dev mailing list