RFR (XXS) 8244667: Shenandoah: SBC2Support::test_gc_state takes loop for wrong control

Roman Kennke rkennke at redhat.com
Fri May 8 21:03:18 UTC 2020

Looks good, thank you!


> Bug:
>   https://bugs.openjdk.java.net/browse/JDK-8244667
> When doing the refactoring in JDK-8244509, I accidentally made a change that was not
> semantic-preserving: get_loop is called on new ctrl from IfTrueNode, not the ctrl we have been
> passed. It went correctly in test_null and test_in_cset, but not in test_gc_state.
> It did not show up with any test failures, but it is important to get right anyway.
> Fix (in webrev form, so you can take a look around):
>   https://cr.openjdk.java.net/~shade/8244667/webrev.01/
> Testing: hotspot_gc_shenandoah

More information about the hotspot-gc-dev mailing list