RFR: 8277531: Print actual default stacksize on Windows thread logging

[Thomas Stuefe]

On Tue, 23 Nov 2021 13:37:03 GMT, David Holmes <dholmes at openjdk.org> wrote:

> There are also issues if you combine -Xss with -XX:ThreadStackSize as the -Xss will be applied to the main thread by the launcher, but the ThreadStackSize will be applied to subsequent JavaThreads. But there is no way to tell that at present. This combination is treated as multiple occurrences of ThreadStackSize with last-option wins - except the launcher will already have only looked at -Xss.

I wondered about that too. We should fix this in the launcher. A bit of a mess :-/

Half-joking, we could spawn off a stack-size=0 thread, just to measure its stack size. But that would cost startup and feels a bit ridiculous tbh.

@tkiriyama : none of this is your fault, the patch would be fine if `os::default_stack_size()` would do what it advertises...

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

PR: https://git.openjdk.java.net/jdk/pull/6495