From kustos at gmx.net  Wed Sep  9 08:34:49 2020
From: kustos at gmx.net (Philippe Marschall)
Date: Wed, 9 Sep 2020 10:34:49 +0200
Subject: How to proceed with 8138732
Message-ID: <07104228-c122-f4f9-aaf6-4aad787585da@gmx.net>

Hello

I started working on 8138732 [2] as described in the issue [1]. However
the question about the impact and coordination with other projects came
up, eg.:

  - projects that implement their own intrinsics
  - Graal
  - somebody else?

How do we want to address this and coordinate with these projects?

Given all the possible impact do we still want the issue to be
implemented as described? Or do we want to have it implemented
differently or not at all?

  [1] https://bugs.openjdk.java.net/browse/JDK-8138732
  [2] https://github.com/openjdk/jdk/pull/45

Cheers
Philippe

From doug.simon at oracle.com  Wed Sep  9 09:06:45 2020
From: doug.simon at oracle.com (Doug Simon)
Date: Wed, 9 Sep 2020 11:06:45 +0200
Subject: How to proceed with 8138732
In-Reply-To: <07104228-c122-f4f9-aaf6-4aad787585da@gmx.net>
References: <07104228-c122-f4f9-aaf6-4aad787585da@gmx.net>
Message-ID: <CCFA4A51-B623-41DE-BAE4-194E67893EBA@oracle.com>

Hi Philippe,

From the Graal perspective, I don?t believe there?s anything directly impacted by this change. To the extent that Graal cares about HotSpot intrinsics, it simply uses <https://github.com/oracle/graal/blob/d9b4e452cc0ab52c3f1c2abdd08d239553948039/compiler/src/org.graalvm.compiler.hotspot.test/src/org/graalvm/compiler/hotspot/test/CheckGraalIntrinsics.java#L624> the list of intrinsics provided by JVMCI. This list is derived <https://github.com/openjdk/jdk/blob/ec5bd02186668e2cdda6352dd1ae3ed29434d60e/src/hotspot/share/jvmci/jvmciCompilerToVMInit.cpp#L181> from the VM_INTRINSICS_DO macro which is unrelated to @HotSpotIntrinsicCandidate.

-Doug

> On 9 Sep 2020, at 10:34, Philippe Marschall <kustos at gmx.net> wrote:
> 
> Hello
> 
> I started working on 8138732 [2] as described in the issue [1]. However
> the question about the impact and coordination with other projects came
> up, eg.:
> 
> - projects that implement their own intrinsics
> - Graal
> - somebody else?
> 
> How do we want to address this and coordinate with these projects?
> 
> Given all the possible impact do we still want the issue to be
> implemented as described? Or do we want to have it implemented
> differently or not at all?
> 
> [1] https://bugs.openjdk.java.net/browse/JDK-8138732
> [2] https://github.com/openjdk/jdk/pull/45
> 
> Cheers
> Philippe


From doug.simon at oracle.com  Tue Sep 15 12:24:41 2020
From: doug.simon at oracle.com (Doug Simon)
Date: Tue, 15 Sep 2020 14:24:41 +0200
Subject: JVMCI 20.3-b01 release
Message-ID: <7ED86374-3391-42C8-BB5E-DDBFD12E3056@oracle.com>

The 20.3-b01 release includes:

	? GR-25801: Add compilation ticks for mitigating against deadlock due to blocking JVMCI compilation (JDK-8250548).
	? GR-25133: Handle JVMCI counters expansion failure gracefully (JDK-8250556).
	? GR-25074: Exit VM with fatal error for libgraal internal problems(JDK-8249888).

It?s also the first JVMCI release with JDK 15 based binaries.

The OpenJDK based Windows, Linux and macOS binaries are available at:

https://github.com/graalvm/graal-jvmci-8/releases/tag/jvmci-20.3-b01 <https://github.com/graalvm/graal-jvmci-8/releases/tag/jvmci-20.3-b01>
https://github.com/graalvm/labs-openjdk-11/releases/tag/jvmci-20.3-b01 <https://github.com/graalvm/labs-openjdk-11/releases/tag/jvmci-20.3-b01>
https://github.com/graalvm/labs-openjdk-15/releases/tag/jvmci-20.3-b01 <https://github.com/graalvm/labs-openjdk-15/releases/tag/jvmci-20.3-b01>

The ?master? branches of the mx, graal and graal-enterprise repos have been updated to use jvmci-20.3-b01 in the CI system.

-Doug