RFR: 8164358: [JVMCI] expose Hotspot intrinsics and HotSpotIntrinsicCandidate info to JVMCI compilers

[Christian Thalinger]

src/jdk.vm.ci/share/classes/jdk.vm.ci.hotspot/src/jdk/vm/ci/hotspot/CompilerToVM.java

      *         [String name, Long value, ...] vmAddresses,
+     *         [String name, Long value, ...] vmAddresses,
      *         VMFlag[] vmFlags
+     *         VMIntrinsicMethod[] vmIntrinsics
Looks like a copy-paste problem.

src/share/vm/jvmci/jvmciCompilerToVM.cpp

Why do we have to do the prev_##var##_sid == sid dance?  I guess it’s some optimization but does it really matter?  If we don’t have to do that I would prefer is that macro would be a method (or two).

> On Aug 18, 2016, at 6:50 AM, Doug Simon <doug.simon at oracle.com> wrote:
> 
> Please review this change which allows JVMCI compilers to see what intrinsics are defined by the VM and which methods have been annotated by HotSpotIntrisicCandidate.
> 
> https://bugs.openjdk.java.net/browse/JDK-8164358
> http://cr.openjdk.java.net/~dnsimon/8164358/
> 
> Thanks,
> 
> -Doug

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20160818/6dc751f0/attachment-0001.html>