[RFR] [8u] 8u222-b02 Upstream Sync

[Andrew John Hughes]

On 14/05/2019 16:10, Aleksey Shipilev wrote:
> On 5/14/19 4:55 PM, Aleksey Shipilev wrote:
>> hotspot:
>>
>> *) sharedRuntime_aarch64.cpp, templateInterpreter_aarch64.cpp, macroAssembler_x86.cpp,
>> jniHandles.cpp: new G1 blocks have to handle Shenandoah too. Luckily, we have this done in sh/jdk9
>> (archived), and it could be just picked up from there:
>>
>> http://hg.openjdk.java.net/shenandoah/jdk9/hotspot/file/37b802a7a71b/src/cpu/aarch64/vm/sharedRuntime_aarch64.cpp#l2067
>> http://hg.openjdk.java.net/shenandoah/jdk9/hotspot/file/37b802a7a71b/src/cpu/aarch64/vm/templateInterpreterGenerator_aarch64.cpp#l1419
>> http://hg.openjdk.java.net/shenandoah/jdk9/hotspot/file/37b802a7a71b/src/cpu/x86/vm/macroAssembler_x86.cpp#l5272
>> http://hg.openjdk.java.net/shenandoah/jdk9/hotspot/file/37b802a7a71b/src/share/vm/runtime/jniHandles.cpp#l118
>>
>> Otherwise looks okay.
> 
> I now remember another thing. The absence of Shenandoah additions in those blocks probably do not
> crash the VM yet, because we have the Shenandoah-specific workaround for it:
>   http://hg.openjdk.java.net/shenandoah/jdk8/hotspot/rev/119e9a5b24d5
>   http://hg.openjdk.java.net/shenandoah/jdk8/hotspot/rev/d305c31da9f5
> 
> Merge is not the proper place to remove that workaround. Maybe we just push this merge verbatim, and
> then we would do downstream work to remove the workaround and fix the issue properly with the
> changes above, then integrate it back wholesale.
> 
> -Aleksey
> 

I concur that a merge is not the proper place to fix this. Indeed, this
is why I included the AArch64 code for 8176100 as a separate changeset,
and not as part of the merge changeset.

I only tend to include additions to a merge changeset where it is
necessary to complete a product build on my development platform
(x86_64). I wouldn't expect that to break as x86_64 is supported in
upstream. It's expected that a merge may break for AArch64 & Shenandoah
as these are not considered upstream, and I think it adds clarity if the
necessary fixes are included as changesets in their own right.

I've pushed the merge. It'd be good if you could follow with the fixes
ASAP, so I can include them in testing of b03.

Thanks,
-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/0xCFDA0F9B35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222
https://keybase.io/gnu_andrew