[llvm] [X86][BF16] Do not combine FP_TRUNC + FP_EXTEND if they come from user (PR #91420)

[Phoebe Wang via llvm-commits]

================
@@ -2420,3 +2420,47 @@ define <16 x bfloat> @concat_dup_v8bf16(<8 x bfloat> %x, <8 x bfloat> %y) {
   %a = shufflevector <8 x bfloat> %x, <8 x bfloat> %y, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
   ret <16 x bfloat> %a
 }
+
+define float @trunc_ext(float %a) nounwind {
+; X86-LABEL: trunc_ext:
+; X86:       # %bb.0:
+; X86-NEXT:    pushl %eax
+; X86-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    vcvtneps2bf16 %xmm0, %xmm0
+; X86-NEXT:    vmovw %xmm0, %eax
+; X86-NEXT:    shll $16, %eax
+; X86-NEXT:    vmovd %eax, %xmm0
+; X86-NEXT:    vmovd %xmm0, (%esp)
+; X86-NEXT:    flds (%esp)
+; X86-NEXT:    popl %eax
+; X86-NEXT:    retl
+;
+; SSE2-LABEL: trunc_ext:
+; SSE2:       # %bb.0:
+; SSE2-NEXT:    pushq %rax
+; SSE2-NEXT:    callq __truncsfbf2 at PLT
----------------
phoebewang wrote:

What's the rounding conversion mean? A non-default rounding (RNE)? I don't think compiler guarantees rounding correntness except for the default rounding mode in default FP model. Only strict FP model guarantees it.

https://github.com/llvm/llvm-project/pull/91420