RFR: 8300657: Remove null filtering in CLD oop handle area
Erik Österlund
eosterlund at openjdk.org
Tue Jan 24 07:08:04 UTC 2023
On Tue, 24 Jan 2023 06:57:54 GMT, David Holmes <dholmes at openjdk.org> wrote:
> Seems fine.
>
>
>
> Aside: Not obvious to me whether closures should be responsible for doing their own null checks or whether the framework should guarantee no nulls get passed to them. You could go either way of course, just depends on the number of closure types versus the number of closure call sites, as to which approach minimises the code. At runtime hopefully there is only ever one null check per data item.
Thanks for the review! We have discussed this policy before, and I think the general consensus is that checking inside of the closure is what we want to do for global roots. That is typically the case already, so this change aligns this policy a bit.
-------------
PR: https://git.openjdk.org/jdk/pull/12145
More information about the hotspot-runtime-dev
mailing list