JDK-8221355 Performance regression after JDK-8155635 backport into 8u not in 8u212

[Severin Gehwolf]

On Wed, 2019-04-10 at 11:19 +0200, Aleksey Shipilev wrote:
> On 4/10/19 10:53 AM, Aleksey Shipilev wrote:
> > On 4/10/19 10:48 AM, Aleksey Shipilev wrote:
> > > On 4/10/19 10:39 AM, Langer, Christoph wrote:
> > > > As per the diff filter OpenJDK 8u212 vs. Oracle 8u212 [0], the issue JDK-8221355 "Performance
> > > > regression after JDK-8155635 backport into 8u" is part of Oracle's update but not in the OpenJDK
> > > > version.
> > > > 
> > > > This fix is not available in public but I guess it can be done by an experienced engineer from
> > > > the issue comments.
> > > 
> > > I'll take a look (or loop in Roland). Maybe Oracle folks do us a solid one and publish the fix?
> > 
> > Forgot to mention: this probably misses ongoing CPU, as the compiler fix there is risky and
> > invalidates lots of testing already done. Unless we collectively think the absence of it is the end
> > of the world.
> 
> Severin had successfully reproduced the regression, handing over to him. Plus, we think that the fix
> in the comment there is the final one:
>   https://bugs.openjdk.java.net/browse/JDK-8221355?focusedCommentId=14255045#comment-14255045

RFR is here:
http://mail.openjdk.java.net/pipermail/jdk8u-dev/2019-April/009079.html

Thanks,
Severin