On Wed, 7 Oct 2026 04:06:09 GMT, Dean Long <[email protected]> wrote:
> > I'm happy to organize all of this into a document on cr.openjdk.org if > > there's interest (so that you don't have to go over the PR comments so far). > > I think that would be useful. Dean, I documented everything I know so far (including the proposed implementation) [here](https://gist.github.com/raneashay/82a0e2b32b3c051e92c5686bc0905743). Feel free to skip Section 2, but I imagine that Section 3 (and perhaps Section 4) will be of most interest to you. I'm waiting to have my SSH key installed on cr.openjdk.org but once that's ready, I'll move the document there. > The idea would be the stack would have two modes, initially in Windows > "growth" mode, which ends when we hit the EXCEPTION_STACK_OVERFLOW, and then > after that HotSpot mode, where all stack pages are committed so we no longer > need the Windows guard page mechanism. I hadn't thought of that but after implementing it, I feel that the code becomes significantly more complex (and likely fragile) because of the additional state tracking between before-first-exception and after-first-exception. I _think_ I was able to get it to work, but I have very little confidence that I got it right. There are some related details that I have included in Section 5 of the linked document. ------------- PR Comment: https://git.openjdk.org/jdk/pull/32365#issuecomment-6133419126
