RFR: 8363620: AArch64: reimplement emit_static_call_stub() [v6]

[Andrew Haley]

On Fri, 12 Dec 2025 10:13:27 GMT, Dean Long <dlong at openjdk.org> wrote:

> > I see your point. However, I don't see the comments to which you refer. Please, where are they?
> 
> Sorry, I meant the new comment in set_to_interpreted, in particular starting around line 145 "If a racing thread reaches the static call stub .... Initially we place an ISB at the start of the static call stub." The way I read that and the litmus test below it is that we need to go from ISB to B .+4 every time we update the MOVs.

Yes, we definitely do. The guarantee must be that the observability of the MOVs happens-before the observability of the B .+4.

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

PR Comment: https://git.openjdk.org/jdk/pull/26638#issuecomment-3650591237