[llvm] [PowerPC] Replace vspltisw+vadduwm instructions with xxleqv+vsubuwm for adding the vector {1, 1, 1, 1} (PR #160882)

[via llvm-commits]

================
@@ -16,17 +16,16 @@ define <2 x i64> @test_add(<2 x i64> %x, <2 x i64> %y) nounwind {
 define <2 x i64> @increment_by_one(<2 x i64> %x) nounwind {
 ; VSX-LABEL: increment_by_one:
 ; VSX:       # %bb.0:
-; VSX-NEXT:    vspltisw 3, 1
-; VSX-NEXT:    vupklsw 3, 3
-; VSX-NEXT:    vaddudm 2, 2, 3
+; VSX-NEXT:    xxleqv 35, 35, 35
+; VSX-NEXT:    vsubudm 2, 2, 3
 ; VSX-NEXT:    blr
 ;
 ; NOVSX-LABEL: increment_by_one:
 ; NOVSX:       # %bb.0:
 ; NOVSX-NEXT:    addis 3, 2, .LCPI1_0 at toc@ha
 ; NOVSX-NEXT:    addi 3, 3, .LCPI1_0 at toc@l
 ; NOVSX-NEXT:    lvx 3, 0, 3
-; NOVSX-NEXT:    vaddudm 2, 2, 3
+; NOVSX-NEXT:    vsubudm 2, 2, 3
----------------
Himadhith wrote:

This was because the code did not check for VSX attribute. The `hasVSX()` check fixed this.

https://github.com/llvm/llvm-project/pull/160882