RFR: JDK-8322854: Incorrect rematerialization of scalar replaced objects in C2

Cesar Soares Lucas cslucas at openjdk.org
Tue Feb 6 23:05:55 UTC 2024


On Fri, 2 Feb 2024 23:31:57 GMT, Vladimir Kozlov <kvn at openjdk.org> wrote:

>> @vnkozlov - Thank you for letting me know about those edge cases. I'll investigate what happens in those situations. 
>> I created this Gist to demonstrate the problem: https://gist.github.com/JohnTortugo/2e6f183b0bf1e465dc871246b410ef4c
>> 
>> @TobiHartmann - I'll try and reproduce these failures locally.
>
> @JohnTortugo are you still working on it? Please, let me know when you are ready for re-review and testing.

Hi @vnkozlov , @TobiHartmann - can you please take a look at the changes and run your internal tests on it? TIA!

@vnkozlov - I added an assert to check that there is a (non_io) memory projection in the Allocate. I ran `hotspot-all` in a few OS's and the assert didn't trigger. I kept it on the code just in case it may trigger with a different set of JVM options, OS/arch, etc.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/17562#issuecomment-1930915166


More information about the hotspot-compiler-dev mailing list