RFR: 8191915: JCK tests produce incorrect results with C2

[Yangfei (Felix)]

Hi,

> -----Original Message-----
> From: Andrew Hughes [mailto:gnu.andrew at redhat.com]
> Sent: Thursday, March 18, 2021 3:47 AM
> To: Yangfei (Felix) <felix.yang at huawei.com>
> Cc: jdk8u-dev at openjdk.java.net; Guoge(JVM) <ge.guo at huawei.com>
> Subject: Re: RFR: 8191915: JCK tests produce incorrect results with C2
> 
> On 08:37 Mon 15 Mar     , Yangfei (Felix) wrote:
> > Hi,
> >
> >   Please review this 8u backport fixing a compatibility issue:
> >
> >   Original bug: https://bugs.openjdk.java.net/browse/JDK-8191915
> >   Original patch: http://hg.openjdk.java.net/jdk/jdk10/rev/4162f740b7bd
> >   Webrev for 8u:
> > http://cr.openjdk.java.net/~fyang/8191915-8u/webrev.00/
> >
> >   I am using gcc-7.5 and I find this issue can always be reproduced on both
> aarch64 and x84_64 linux with the provided jtreg test.
> >   Original patch does not apply cleanly after path shuffling, three
> adaptations for 8u:
> >   1. Adjusted copyright year update for both mathexactnode.cpp and
> mathexactnode.hpp;
> >   2. Adjusted the test command for the jtreg test in the original patch:
> LongMulOverflowTest.java;
> >   3. Removed test_mathexact.cpp in the original patch.
> >
> >   Performed full jtreg test on both aarch64 and x84_64 linux. OK?
> >
> > Thanks,
> > Felix
> 
> Looks good to me. Test fails without the patch and passes with here too.
> No GTest for the test_mathexact.cpp patch in 8u.
> 
> As this apparently affects a JCK test, I think it may be a candidate for inclusion
> in 8u292 during rampdown. Please flag with jdk8u-critical-request if you
> agree.

Yes, I agree we should have this for 8u292.  Flagged for approval.

Thanks,
Felix