RFR: 8340103: Add internal set_flag function to VMATree [v18]

[Thomas Stuefe]

On Thu, 7 Nov 2024 09:59:31 GMT, Johan Sjölen <jsjolen at openjdk.org> wrote:

>> You might be right about that. I thought that we simply ignored state and stack of released regions. Let me see what's going on.
>
> Ah, actually that is precisely what we see as an assert later on. `set_tag` must support address holes in some sense, otherwise we'll end up in a bad state.

I think register_mapping should always register released sections with mtNone/no stack. Maybe even assert if you specify something different. Then, set_tag should just ignore released sections, not change anything

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

PR Review Comment: https://git.openjdk.org/jdk/pull/20994#discussion_r1832457245