RFR: 8193513: add support for printing a stack trace on class loading [v4]

David Holmes dholmes at openjdk.org
Mon Jul 3 22:56:53 UTC 2023


On Mon, 3 Jul 2023 21:28:43 GMT, Doug Simon <dnsimon at openjdk.org> wrote:

> The test uses -Xlog:all=trace:file=all.log so will indiscriminately log class+load+cause+native for every class loaded and the test seems to load a lot of classes.

Are you treating not setting `LogClassLoadingCauseFor` as being a wildcard for "all"? That'[s not what you had earlier and seems wrong - this needs to be opt-in.

> The test crashes the VM as VMError::print_native_stack is apparently not safe to call in all class loading contexts.

That is a distinct bug in itself. Probably print_native_stack needs additional safety/state checks. I will file a bug for that.

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

PR Comment: https://git.openjdk.org/jdk/pull/14553#issuecomment-1619232854


More information about the hotspot-dev mailing list