Foreign-Memory Access API

Maurizio Cimadamore maurizio.cimadamore at oracle.com
Tue Feb 4 13:20:29 UTC 2020


Hi Chris,
as you mention, supporting such a use case safely is a non starter. That 
of native interop tools that will come along once we start addressing 
the foreign function problem, there will be a tool to create an 
'unchecked' segment from a given address you might know by some other 
means. Of course this is an unsafe functionality as there would be no 
way for the API to know which spatial and temporal bounds should be 
attached to this segment, but there will be ways for the user to attach 
bounds him/herself.

Maurizio

On 04/02/2020 02:20, Chris T wrote:
> Hello,
>
>    I am looking into JEP 370 and I noted this particular line:
>
>     - Share memory across multiple processes, and
>
>    I didn't see it listed in the goals - hence the curiosity -:  is it
> really possible that I can get a non-heap memory area that can be used by
> two different processes (maybe JVMs)? If yes, how would that be achieved -
> through the API? Any considerations for security?
>    I am thinking a basic scenario of a common memory cache but we can
> imagine many more cases...
>
> Thanks!
>    Chris T


More information about the core-libs-dev mailing list