RFR: 8354231: x86: Purge FPU support from (Macro)Assembler after 32-bit x86 removal
Vladimir Kozlov
kvn at openjdk.org
Thu Apr 10 16:34:30 UTC 2025
On Wed, 9 Apr 2025 19:14:06 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:
> There are lots of cleanups possible in (Macro)Assembler after 32-bit x86 removal. This cleanup is focused on removing the large parts of x87 FPU support that are no longer needed. Note that _some_ math stubs are actually using x87 FPU in a very limited manner, but the implementations use separate definitions (under `LP64`). Those are left alone, and we clean up x86_32 code here.
>
> Additional testing:
> - [x] Linux x86_64 server fastdebug, `all`
Good.
src/hotspot/cpu/x86/assembler_x86.hpp line 1400:
> 1398: #ifndef _LP64
> 1399: void emms();
> 1400:
So you keeping this 32-bit code for now?
-------------
Marked as reviewed by kvn (Reviewer).
PR Review: https://git.openjdk.org/jdk/pull/24558#pullrequestreview-2757501758
PR Review Comment: https://git.openjdk.org/jdk/pull/24558#discussion_r2037817079
More information about the hotspot-dev
mailing list