[foreign-memaccess+abi] RFR: 8292034: Improve javadoc after memory segment/memory address unification [v2]

[Jorn Vernee]

On Fri, 23 Sep 2022 14:28:47 GMT, Maurizio Cimadamore <mcimadamore at openjdk.org> wrote:

>> src/java.base/share/classes/java/lang/foreign/MemorySegment.java line 91:
>> 
>>> 89:  * A heap segment obtained from one of the {@link #ofArray(int[])} factory methods has an address of zero.</li>
>>> 90:  * <li>The address of a native segment (including mapped segments) denotes the physical address of the region of
>>> 91:  * memory which backs the segment.</li>
>> 
>> I wonder if "physical address" could be confusing here. Since, a process in the operating system is typically assigned a virtual address space that maps to physical memory, and it's the virtual addresses we expose.
>> 
>> Maybe:
>> Suggestion:
>> 
>>  * <li>The address of a native segment (including mapped segments) denotes the address in native memory of the region of
>>  * memory which backs the segment.</li>
>
> I get your concern - however, by physical here we refer to the address that e.g. a C/C++ program would see. The fact that physical addresses are virtualized by the OS is a concern we gloss over here - honestly I don't think this is a big deal?

I get what is meant by it, but I don't think "physical" is a great term wrt intuitive meaning. It would maybe be nice to define the term somewhere up front if we want to stick with "physical".

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

PR: https://git.openjdk.org/panama-foreign/pull/730