RFR: 8241232: -XX:+BootstrapJVMCI is not compatible with TieredStopAtLevel < CompLevel_full_optimization

Tobias Hartmann tobias.hartmann at oracle.com
Thu Mar 19 08:16:03 UTC 2020


Hi Jie,

looks good but I would prefer if the test just uses a @run main/othervm statement instead of
manually spawning a new VM process.

Best regards,
Tobias

On 19.03.20 01:57, jiefu(傅杰) wrote:
> Hi all,
> 
> JBS:    https://bugs.openjdk.java.net/browse/JDK-8241232
> Webrev: http://cr.openjdk.java.net/~jiefu/8241232/webrev.00/
> 
> An assert [1] was fired with -XX:+BootstrapJVMCI and TieredStopAtLevel<4.
> The reason is that for -XX:+BootstrapJVMCI the VM is hardcoded using CompLevel_full_optimization [2].
> So -XX:+BootstrapJVMCI seems not compatible with TieredStopAtLevel < 4.
> It might be better to fix it.
> 
> Testing:
>   - hotspot:tier1 on macosx with server-release/fastdebug
> 
> Please review it and give me some advice.
> 
> Thanks a lot.
> Best regards,
> Jie
> 
> [1] http://hg.openjdk.java.net/jdk/jdk/file/c1fdb04213e9/src/hotspot/share/compiler/compileBroker.cpp#l1246
> [2] http://hg.openjdk.java.net/jdk/jdk/file/743c9071c317/src/hotspot/share/jvmci/jvmciCompiler.cpp#l74
> 


More information about the hotspot-compiler-dev mailing list