[8u] RFR: Backport of JDK-8222670: pathological case of JIT recompilation and code cache bloat

Liu, Xin xxinliu at amazon.com
Thu May 16 17:23:19 UTC 2019


Hi, Reviewers,

Could you review this backport patch?

BUG: https://bugs.openjdk.java.net/browse/JDK-8222670
Webrev: https://cr.openjdk.java.net/~xliu/8222670/webrev.8u-dev/

I can’t cleanly apply the original patch to jdk8u.  I made the following changes.

  1.  Include JDK-8223537.  It solved suspend problem in blocking mode. One line change.
  2.  Shorter wait period.  Otherwise, it’s easy to timeout for the  test.  It’s part of a JDK-8066770.
https://cr.openjdk.java.net/~xliu/8222670/webrev.8u-dev/test/compiler/whitebox/CompilerWhiteBoxTest.java.udiff.html

  1.  +1 for Tier4MinInvocationThreshold and Tier4CompileThreshold. Otherwise, we have to backport JDK-8059604, which looks like complicated.

https://cr.openjdk.java.net/~xliu/8222670/webrev.8u-dev/src/share/vm/prims/whitebox.cpp.udiff.html

Thanks,
--lx



More information about the jdk8u-dev mailing list