On Fri, 27 Feb 2026 19:16:57 GMT, Frederic Parain <[email protected]> wrote:
> Small patch fixing a missing memory barrier in FlatArrayKlass::copy_array. > The unit test covers more scenario where a flat value has to be buffered > before to be written to another heap variable. So far, these tests haven't > shown any other issue (they were run repeatedly 50 times using Mach5). > > However, the unit test is quite heavy, testing 5 different scenarios, and > running each for 3 minutes, using all detected cores of the machine each > time. This might qualify this unit test as a "resourcehog" test, it then > should be moved to test/hotspot/jtreg/resourcehogs/ > > Fred lgtm. (Nice tests) I think that eventually we would like to get these semantics into jcstress, it has been a good tool for finding some breakage in the JMM in the past. ------------- Marked as reviewed by aboldtch (Committer). PR Review: https://git.openjdk.org/valhalla/pull/2183#pullrequestreview-3876724208
