On Wed, 30 Oct 2024 12:57:50 GMT, Maurizio Cimadamore <mcimadam...@openjdk.org> wrote:
> But then, somewhere else say: > > "The returned segment is backed by the same memory region as that of the > original segment. As such, the region of memory backing the returned segment > is truly deallocated only when the original segment's arena is closed." Bonus points if we can also add something re. the fact that this can cause two issues: 1. a client might access the reinterpreted segment _after_ its region of memory has been deallocated 2. the region of memory "outlives" the reinterpreted segment - e.g. even after the provided arena is closed ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/21761#discussion_r1822575744
