RFR: 8265322: C2: Simplify control inputs for BarrierSetC2::obj_allocate

Yi Yang yyang at openjdk.java.net
Mon Apr 19 02:22:34 UTC 2021

On Fri, 16 Apr 2021 17:37:05 GMT, Vladimir Kozlov <kvn at openjdk.org> wrote:

> From compiler code POV the fix is reasonable and correct.

Thank you @vnkozlov!

Do you think it's reasonable to move PhaseMacroExpand::set_eden_pointers to BarrierSetC2? It seems that that's GC knowledge area about how to set eden_top/eden_end w or w/o turning UseTLAB.

> Note, the path when `initial_slow_test != NULL` is not rare. It is frequent for arrays allocation when `length` is not constant.

Yes, I missed the AllocateArrayNode when skimming the code. Thanks for pointing out this.

Let's wait for other hotspot gc/compiler folks for more reviews.


PR: https://git.openjdk.java.net/jdk/pull/3529

More information about the hotspot-compiler-dev mailing list