RFR: 8293285: x86_64: Move libm stub implementations to StubGenerator

Fri Sep 2 00:28:02 UTC 2022

Move libm stub implementations (used for trigonometric math intrinsics) from MacroAssembler to StubGenerator.

x86_32-specific code is left intact, but all `macroAssembler_x86_*.cpp` files are renamed to `macroAssembler_x86_32_*.cpp` since there's no x86_64-specific code left there anymore.

Testing: hs-tier1 - hs-tier4

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

Depends on: https://git.openjdk.org/jdk/pull/10111

Commit messages:
 - Extract libm stub implementations into separate files

Changes: https://git.openjdk.org/jdk/pull/10133/files
 Webrev: https://webrevs.openjdk.org/?repo=jdk&pr=10133&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8293285
  Stats: 23104 lines in 26 files changed: 11818 ins; 11257 del; 29 mod
  Patch: https://git.openjdk.org/jdk/pull/10133.diff
  Fetch: git fetch https://git.openjdk.org/jdk pull/10133/head:pull/10133

PR: https://git.openjdk.org/jdk/pull/10133