RFR: 8344352: 32-bit builds crash after JDK-8305895 [v2]
Aleksey Shipilev
shade at openjdk.org
Mon Nov 18 13:50:43 UTC 2024
On Mon, 18 Nov 2024 13:41:40 GMT, Coleen Phillimore <coleenp at openjdk.org> wrote:
> Well, now that you say this, it does look like it should be called after command line processing. I think threads::create_vm might be a better place for this but there are initializations scattered all around, so you could probably make an argument for another place.
What I meant to say is that we currently call `ObjLayout::initialize()` after `Arguments::parse()` and `Arguments::apply_ergo()`, pretty much like what we do to initialize OS subsystem. I think that's the point where object layout-related arguments are definitely set. There is `Arguments::adjust_after_os()`, but I don't think it can reasonably change obj layout flags without breaking the rest of the setup.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/22196#issuecomment-2483101524
More information about the hotspot-runtime-dev
mailing list