RFR: Handle Op_Parm in get_load_addr()

Roman Kennke rkennke at redhat.com
Sat Jul 6 10:56:46 UTC 2019


One test hit a Parm in get_load_addr(). Let's handle it.

BTW: This (and others like Con, Calls, allocs) shouldn't be necessary
when we'd split LRBs through phis, they would get optimized out, and we
should get LRBs only where we really want them: after loads.

http://cr.openjdk.java.net/~rkennke/sfx-parm/webrev.00/

Testing: hotspot_gc_shenandoah

Ok?

Roman



More information about the shenandoah-dev mailing list