RFR: Fix Minimal VM build failures [v2]

Aleksey Shipilev shade at openjdk.java.net
Tue Sep 28 06:33:44 UTC 2021


On Mon, 27 Sep 2021 21:46:04 GMT, Serguei Spitsyn <sspitsyn at openjdk.org> wrote:

>> I think that makes sense. At first I was thinking why not just #ifdef out the whole API, but these APIs are mapped natives for `VirtualThread.notifyJvmtiMountBegin()` and `VirtualThread.notifyJvmtiMountEnd()`. Calls to them come from java, but should only happen if jvmti is enabled.
>
> I agree, it is a good idea. Of course, it makes sense to add for all 4 native functions:
>  JVM_VirtualThreadMountBegin, JVM_VirtualThreadMountEnd,
>  JVM_VirtualThreadUnmountBegin, JVM_VirtualThreadUnmountEnd
> 
> Additionally, asserts should be also under #if INCLUDE_JVMTI.

All right, see new commit.

-------------

PR: https://git.openjdk.java.net/loom/pull/67


More information about the loom-dev mailing list