[jdk11u-dev] RFR: 8272720: Fix the implementation of loop unrolling heuristic with LoopPercentProfileLimit
Jie Fu
jiefu at openjdk.java.net
Fri Nov 5 08:59:09 UTC 2021
On Fri, 5 Nov 2021 08:21:20 GMT, Jie Fu <jiefu at openjdk.org> wrote:
>> 8272720: Fix the implementation of loop unrolling heuristic with LoopPercentProfileLimit
>
>> /sponsor
>
> Thanks @johnshajiang for your sponsoring.
> > > The crash in GHA Windows build is a bit concerning... Maybe it's just infrastructure related, though. @DamonFool, could you check this?
> >
> >
> > Hi @RealCLanger ,
> > Thanks for your review.
> > The crash happened in `make` which shouldn't be caused by the hotspot code change.
> > ```
> > /usr/bin/bash: line 1: 1076 Segmentation fault (core dumped) /usr/bin/make -s -r -R -I /cygdrive/d/a/jdk11u-dev/jdk11u-dev/jdk/make/common SPEC=/cygdrive/d/a/jdk11u-dev/jdk11u-dev/jdk/build/windows-x64/spec.gmk MAKE_LOG_FLAGS="-s" -f ModuleWrapper.gmk -I/cygdrive/d/a/jdk11u-dev/jdk11u-dev/jdk/make -I/cygdrive/d/a/jdk11u-dev/jdk11u-dev/jdk/make/copy MODULE=java.base MAKEFILE_PREFIX=Copy
> > ```
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > In theory, there is no risk for this backport. Thanks.
>
> OK, yes, it must have been something in the environment. I just ran another GHA tests on a current branch which got past this step: https://github.com/RealCLanger/jdk11u-dev/actions/runs/1424808669
Thanks @RealCLanger for your experiment.
-------------
PR: https://git.openjdk.java.net/jdk11u-dev/pull/588
More information about the jdk-updates-dev
mailing list