RFR: 8282475: SafeFetch should not rely on existence of Thread::current

[David Holmes]

On Mon, 7 Mar 2022 14:24:01 GMT, Thomas Stuefe <stuefe at openjdk.org> wrote:

> We need to change the wx state of the current pthread in order to be able to execute stub routines. Otherwise, we would crash right away when trying to execute the SafeFetch stub.

Oh I see - that is unfortunate. I don't like messing with other people's threads.

> May I ask how that would affect the code that uses the methods (includes, ...)?

@parttimenerd there would be no change - they continue to include os.hpp, which will include the os/cpu specific header files.

> We could have generic wrappers like: ...

@tstuefe  I think this is going a little too far in this fix. I'm looking for simplicity. All the WX related code should be buried in the os/cpu file for BSD/Aarch64 and the callsites all using MACOS_AARCH64_ONLY.

Splitting BSD from macOS would also be a future RFE.

Thanks.

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

PR: https://git.openjdk.java.net/jdk/pull/7727