RFR: 8233159: Method::result_type should use calculated value in constMethod

coleen.phillimore at oracle.com coleen.phillimore at oracle.com
Wed Oct 30 14:25:36 UTC 2019



On 10/30/19 9:51 AM, Claes Redestad wrote:
> Hi Coleen,
>
> On 2019-10-30 14:09, coleen.phillimore at oracle.com wrote:
>>
>> This looks good. 
>
> thanks!
>
>> There are a couple other places where we iterate the signature with 
>> ResultTypeFinder and don't need to. I thought you were going to 
>> remove these also?
>
> Any particular place where you think we can trivially remove
> ResultTypeFinder use?
>
> I limited these changes to things that were both common between my
> experiments and JDK-8230199, as well as easily factored out to a
> separate RFE. Hopefully this makes JDK-8230199 easier to deal with.
>
> JDK-8230199 aims to consolidate so we only parse a method signature
> once, so doing intrusive changes in the interim didn't make immediate
> sense to me.

There's one in defaultMethods.cpp but it's a bit more work to change it.

There's one in jvmtiEnvBase, that could be:

   // Get information about method return type
   Symbol* signature = jvf->method()->signature();

   TosState fr_tos = as_TosState(jvf->method()->result_type());

I thought there were more at first glance.  Your change is fine. The 
jvmtiEnvBase code doesn't matter for performance.

Thanks,
Coleen
>
> Thanks!
>
> /Claes



More information about the hotspot-runtime-dev mailing list