[all-commits] [llvm/llvm-project] d15c45: [FPEnv][AMDGPU] Correct strictfp tests.

[Kevin P. Neal via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: d15c454bedc05775b5080e1d2130b0554d5e5a81
      https://github.com/llvm/llvm-project/commit/d15c454bedc05775b5080e1d2130b0554d5e5a81
  Author: Kevin P. Neal <kevin.neal at sas.com>
  Date:   2024-02-05 (Mon, 05 Feb 2024)

  Changed paths:
    M llvm/test/CodeGen/AMDGPU/fsub-as-fneg-src-modifier.ll
    M llvm/test/CodeGen/AMDGPU/global_atomics_scan_fadd.ll
    M llvm/test/CodeGen/AMDGPU/global_atomics_scan_fsub.ll
    M llvm/test/CodeGen/AMDGPU/llvm.get.fpmode.ll
    M llvm/test/CodeGen/AMDGPU/strict_fptrunc.ll
    M llvm/test/CodeGen/AMDGPU/strictfp_f16_abi_promote.ll

  Log Message:
  -----------
  [FPEnv][AMDGPU] Correct strictfp tests.

Correct AMDGPU strictfp tests to follow the rules documented in the
LangRef:
https://llvm.org/docs/LangRef.html#constrained-floating-point-intrinsics

These tests needed the strictfp attribute added to function calls and
some declarations.

Some of the tests now pass with D146845, others get farther along and
fail with D146845. The tests revealed that further work is required
in mostly AMDGPU atomics to get the tests passing.

Since I was here anyway I removed the strictfp attribute from some
constrained intrinsic declarations. They have this attribute by default.

Test changes verified with D146845.