RFR(S) : 8170227 : use vm.jvmci property in compiler/jvmci tests

Vladimir Kozlov vladimir.kozlov at oracle.com
Mon Nov 28 18:25:24 UTC 2016


Good.

Thanks,
Vladimir

On 11/28/16 8:18 AM, Igor Ignatyev wrote:
> http://cr.openjdk.java.net/~iignatyev/8170227/webrev.0/
>> 102 lines changed: 23 ins; 1 del; 78 mod;
>
> Hi all,
>
> could you please review the changset which update jvmci tests to use vm.jvmci @requiers property instead of cpu architecture bases expressions?
>
> there are two cases:
>  - jvmci tests which can be run on all platforms which support jvmci, in these tests '(vm.simpleArch == "x64" | vm.simpleArch == "sparcv9" | vm.simpleArch == "aarch64”)’ was replaced by vm.jvmci
>  - test/compiler/jvmci/jdk.vm.ci.code.test tests which require an arch specific subclass of TestAssembler class, for the time being, there are only amd64 and sparc versions. in such tests @requires expression was changed to 'vm.jvmci & (vm.simpleArch == "x64" | vm.simpleArch == "sparcv9”)'
>
> please be aware, this fix depends on 8170226 which adds vm.jvmci property.
>
> webrev: http://cr.openjdk.java.net/~iignatyev/8170227/webrev.0
> jbs: https://bugs.openjdk.java.net/browse/JDK-8170227
> testing: hotspot/compiler/jvmci on platforms/builds which support and do not support jvmci
>
> 8170226 RFR: http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2016-November/025005.html
>
> Thanks,
> — Igor
>
>
>


More information about the hotspot-compiler-dev mailing list