[jdk25u-dev] RFR: 8357570: [macOS] os::Bsd::available_memory() might return too low values

[Aleksey Shipilev]

On Fri, 28 Nov 2025 10:57:13 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

> This bugfix improves MacOS reporting, *and* makes further backports in container code cleaner. Notably, it is on the way to [JDK-8370572](https://bugs.openjdk.org/browse/JDK-8370572) backport. The full stack looks like this:
> 
> 
>     Clean JDK-8370966: Create regression test for the hierarchical memory limit fix in JDK-8370572
> (!) Clean JDK-8370572: Cgroups hierarchical memory limit is not honored after JDK-8322420
>     Clean JDK-8292984: Refactor internal container-related interfaces for clarity
>     Clean JDK-8356868: Not all cgroup parameters are made available
>     Clean JDK-8367485: os::physical_memory is broken in 32-bit JVMs when running on 64-bit OSes
>     Clean JDK-8357086: os::xxx functions returning memory size should return size_t
>>>> Clean JDK-8357570: [macOS] os::Bsd::available_memory() might return too low values
>     Clean JDK-8364248: Separate commit and reservation limit detection
> 
> 
> I looked where this code is used, and I think only management and reporting code realistically depends on it.
> 
> Additional testing:
>  - [x] GHA
>  - [x] MacOS AArch64 server fastdebug, `jdk_management`

Push approval granted.

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

PR Comment: https://git.openjdk.org/jdk25u-dev/pull/31#issuecomment-3596686595