[8u] RFR (S) 8218201: Failures when vmIntrinsics::_getClass is not inlined

[Andrew Hughes]

On Thu, 8 Aug 2019 at 16:35, Aleksey Shipilev <shade at redhat.com> wrote:
>
> Bug:
>   https://bugs.openjdk.java.net/browse/JDK-8218201
>   https://hg.openjdk.java.net/jdk/jdk/rev/a2d3ca8062b9
>
> 8u version needs a few adjustments, because context is a bit different. The new change should be
> functionally equivalent to what we have in jdk/jdk. 8u webrev:
>   https://cr.openjdk.java.net/~shade/8218201/webrev.8u.01/
>
> Testing: affected test (fails without the patch, passes with it); tier1
>
> --
> Thanks,
> -Aleksey
>

Patch looks good.

FWIW, the fillInStackTrace case is removed in OpenJDK 9 by JDK-8076112
[0] (which adds HotSpotIntrinsicCandidate), but I can't see anything
in that change which warrants its removal. I suspect it may be more
related to the stack walking changes in 9.

Elements of JDK-8076112 were backported for the recent security fix,
JDK-8223511. I wonder if backporting the rest may be a worthwhile
debugging aid for intrinsics on 8u, especially with regard to the
varying support on different architectures.

Thanks,
-- 
Andrew :)

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

Web Site: http://fuseyism.com
Twitter: https://twitter.com/gnu_andrew_java
PGP Key: ed25519/0xCFDA0F9B35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222