RFR: 8332362: Implement os::committed_in_range for MacOS and AIX [v2]

[Robert Toyonaga]

On Tue, 11 Jun 2024 13:51:13 GMT, Thomas Stuefe <stuefe at openjdk.org> wrote:

>> Robert Toyonaga has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Remove debug log. Enable tests for aix.
>
> test/hotspot/gtest/runtime/test_committed_virtualmemory.cpp line 217:
> 
>> 215: }
>> 216: 
>> 217: #ifndef _WINDOWS
> 
> Test should work on Windows, too

On Windows `os::committed_in_range` uses `VirtualQuery` which returns only whether the region is `MEM_COMMIT`. [A region that is set to `MEM_COMMIT` may not yet be faulted in.](https://learn.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-virtualalloc) So unlike mincore, it's not actually useful for getting a fine-grained measurement of physical memory used, although it's helpful to find the parts of the thread stacks that are committed. 

I did a quick check on my windows computer by committing a big amount of memory and checking the RSS. `os::committed_in_range` reports the memory as committed, but it's not reflected in the RSS. When I actually touch the chunk of committed memory, `os::committed_in_range` reports the same as before, but I can see it in the RSS this time.

> test/hotspot/jtreg/runtime/Thread/TestAlwaysPreTouchStacks.java line 43:
> 
>> 41:  * @test id=preTouch
>> 42:  * @summary Test AlwaysPreTouchThreadStacks
>> 43:  * @requires os.family != "aix"
> 
> Should work on aix now, no?

yes, that's my bad.  I neglected to take that out. Removed now.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/19455#discussion_r1635447323
PR Review Comment: https://git.openjdk.org/jdk/pull/19455#discussion_r1635450820