dissecting memory sessions

Maurizio Cimadamore maurizio.cimadamore at oracle.com
Thu Nov 3 15:19:54 UTC 2022


On 03/11/2022 14:55, Gavin Ray wrote:
> I recall at one point there being an open PR to add some kind of 
> memory management/allocator interface (similar to Netty's 
> MemoryManager IIRC)
> Pull requests · openjdk/panama-foreign (github.com) 
> <https://urldefense.com/v3/__https://github.com/openjdk/panama-foreign/pull/509__;!!ACWV5N9M2RV99hQ!MGkwde_O1zcMAAqFsj5jOUAoqbw-JXW8ht4OYiP_Vx8YH3EScxEc5RJV4hMjkR5cRzQu1JfhGRhLJLrLqXlc5l4zvztN$>
> MemoryManager (Netty API Reference (5.0.0.Alpha4)) 
> <https://urldefense.com/v3/__https://netty.io/5.0/api/io.netty5.buffer/io/netty5/buffer/api/MemoryManager.html__;!!ACWV5N9M2RV99hQ!MGkwde_O1zcMAAqFsj5jOUAoqbw-JXW8ht4OYiP_Vx8YH3EScxEc5RJV4hMjkR5cRzQu1JfhGRhLJLrLqXlc5n1FAbMP$>
>
> Is this Arena interface filling the same sort of role that the 
> "MemorySegmentPool + Allocator" PR would have played?
> It seemed like a useful PR, but it doesn't look like it had the intent 
> to be merged, so I was just curious.
Hi Gavin,
Arena is a different thing. It can be used as a building block to 
implement custom allocators, such as the one you describe. But this is 
not about adding more custom allocators to the FFM API (that would be a 
separate discussion :-) ).
For now, the main thing is to make sure that the described proposal 
allows clients to cleanly build such custom allocators (see related 
email from Rado).

Thanks
Maurizio
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/panama-dev/attachments/20221103/b0525cf4/attachment.htm>


More information about the panama-dev mailing list