RFR: 8329088: Stack chunk thawing races with concurrent GC stack iteration [v2]

[Erik Österlund]

On Mon, 22 Apr 2024 09:39:44 GMT, Erik Österlund <eosterlund at openjdk.org> wrote:

>>> Unlike thawing, the freeze operation does not race with the GC by design.
>>>
>> Is this with the changes in the allocation code in this patch or even before those there was no race?
>
> Thanks for the detailed review @pchilano! I made the suggested updates.

> Hello 👋, I wonder @fisk if you were able to reproduce the situation where a GC thread sees inconsistent values for `sp` and `argsize` and whether it would make sense to write a test forcing that situation to happen to prove that this PR fixes the issue and to avoid regressions ? I'm not super experienced with the hotspot code, but I'd be happy to try to write a test for that if you think that it makes sense.

Writing a test for this would be rather tricky, as you have to put the GC in a rather particular state, to have it actually race with thawing. If you want to give it a shot, then that would be nice, but it won't be easy I think.

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

PR Comment: https://git.openjdk.org/jdk/pull/18643#issuecomment-2089897313