x64: Migrate fcvt_from_sint and fcvt_low_from_sint to ISLE (#4650)

https://github.com/bytecodealliance/wasmtime/pull/4650

cranelift/filetests/filetests/isa/x64/fcvt.clif

@@ -0,0 +1,133 @@
+test compile precise-output
+target x86_64
+
+function %f1(i8) -> f32 {
+block0(v0: i8):
+  v1 = fcvt_from_sint.f32 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   movsbl  %dil, %ecx
+;   cvtsi2ss %ecx, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f2(i16) -> f32 {
+block0(v0: i16):
+  v1 = fcvt_from_sint.f32 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   movswl  %di, %ecx
+;   cvtsi2ss %ecx, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f3(i32) -> f32 {
+block0(v0: i32):
+  v1 = fcvt_from_sint.f32 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   cvtsi2ss %edi, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f4(i64) -> f32 {
+block0(v0: i64):
+  v1 = fcvt_from_sint.f32 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   cvtsi2ss %rdi, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f5(i8) -> f64 {
+block0(v0: i8):
+  v1 = fcvt_from_sint.f64 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   movsbl  %dil, %ecx
+;   cvtsi2sd %ecx, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f6(i16) -> f64 {
+block0(v0: i16):
+  v1 = fcvt_from_sint.f64 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   movswl  %di, %ecx
+;   cvtsi2sd %ecx, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f7(i32) -> f64 {
+block0(v0: i32):
+  v1 = fcvt_from_sint.f64 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   cvtsi2sd %edi, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f8(i64) -> f64 {
+block0(v0: i64):
+  v1 = fcvt_from_sint.f64 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   cvtsi2sd %rdi, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f9(i32x4) -> f64x2 {
+block0(v0: i32x4):
+  v1 = fcvt_low_from_sint.f64x2 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   cvtdq2pd %xmm0, %xmm0
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+