RFR: 8280056: gtest/LargePageGtests.java#use-large-pages failed "os.release_one_mapping_multi_commits_vm" [v2]

[Afshin Zafari]

On Thu, 19 Oct 2023 01:37:28 GMT, David Holmes <dholmes at openjdk.org> wrote:

> That said, I really do not like polluting the VM code with this `fail_reason` just to allow a gtest to be more informative. Perhaps there should be logging at the lower-level instead?

In all the invocations of `attempt_reserve_merory_at` (except some WhiteBox methods) the return value is checked against `nullptr`. So `nullptr` is one of the possible and EXPECTED returns. Is it acceptable to change the gtest to check if  the return is not `nullptr`,  *then* it is EXPECTED to be the requested address?

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

PR Comment: https://git.openjdk.org/jdk/pull/16240#issuecomment-1770367120