RFR [11] (XS): 8195064 : LogCompilation: Chicken and egg startup problem with types

Tobias Hartmann tobias.hartmann at oracle.com
Mon Jan 15 08:50:41 UTC 2018


Hi Eric,

this looks good to me.

I just wanted to ask if we should add some tests but I've seen that you've filed JDK-8195069 for that.

Best regards,
Tobias

On 12.01.2018 23:40, Eric Caspole wrote:
> Hi everybody,
> 
> There is a startup condition problem in LogCompilation where it asserting a result from the type table is not null but
> no types have yet been placed in the table.
> 
> I have been using LogCompilation with this workaround manually applied for about a month and I haven't seen any problems
> from removing the check.
> 
> 
> JBS: https://bugs.openjdk.java.net/browse/JDK-8195064
> 
> webrev: http://cr.openjdk.java.net/~ecaspole/JDK-8195064/webrev/
> 
> Thanks,
> 
> Eric
> 


More information about the hotspot-compiler-dev mailing list