RFR: 8371637: allocateNativeInternal sometimes return incorrectly aligned memory [v6]

[Almaz]

On Fri, 21 Nov 2025 14:48:38 GMT, Maurizio Cimadamore <mcimadamore at openjdk.org> wrote:

>>> Note that WG14 still defines a max alignment:
>>> > max_align_t
>> 
>> Can `MAX_MALLOC_ALIGN` be set to `alignof(max_align_t)`?
>> 
>>> It's in the fact that libc guarantees that all malloc calls will be aligned to that max alignment whereas WG14 does not (unless when you hold it at a certain angle ;-) )
>> 
>> I think this comes back to https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2293.htm.
>
>> > Note that WG14 still defines a max alignment:
>> > > max_align_t
>> 
>> Can `MAX_MALLOC_ALIGN` be set to `alignof(max_align_t)`?
>> 
>> > It's in the fact that libc guarantees that all malloc calls will be aligned to that max alignment whereas WG14 does not (unless when you hold it at a certain angle ;-) )
>> 
>> I think this comes back to https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2293.htm.
> 
> That would be best, I think, but can also be addressed in a separate PR.

@mcimadamore 

> but can also be addressed in a separate PR.

Do you want me to send a patch? Will that need an entry in JBS?

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

PR Comment: https://git.openjdk.org/jdk/pull/28235#issuecomment-3564244255