x64: Lower vany_true, vall_true, vhigh_bits, iconcat, and isplit in ISLE (#4787)

Lower vany_true, vall_true, vhigh_bits, iconcat, and isplit in ISLE.

cranelift/filetests/filetests/isa/x64/ishl.clif

@@ -17,20 +17,20 @@ block0(v0: i128, v1: i8):
 ;   pushq   %rbp
 ;   movq    %rsp, %rbp
 ; block0:
-;   movzbq  %dl, %rax
-;   movq    %rax, %rcx
+;   movzbq  %dl, %rcx
 ;   movq    %rdi, %rdx
 ;   shlq    %cl, %rdx, %rdx
 ;   shlq    %cl, %rsi, %rsi
-;   movq    %rcx, %r8
+;   movq    %rcx, %rax
 ;   movl    $64, %ecx
-;   subq    %rcx, %r8, %rcx
+;   movq    %rax, %r10
+;   subq    %rcx, %r10, %rcx
 ;   shrq    %cl, %rdi, %rdi
 ;   xorq    %rax, %rax, %rax
-;   testq   $127, %r8
+;   testq   $127, %r10
 ;   cmovzq  %rax, %rdi, %rdi
 ;   orq     %rdi, %rsi, %rdi
-;   testq   $64, %r8
+;   testq   $64, %r10
 ;   cmovzq  %rdx, %rax, %rax
 ;   cmovzq  %rdi, %rdx, %rdx
 ;   movq    %rbp, %rsp

cranelift/filetests/filetests/isa/x64/simd-logical-compile.clif

@@ -41,9 +41,9 @@ block0(v0: i64x2):
 ;   pushq   %rbp
 ;   movq    %rsp, %rbp
 ; block0:
-;   pxor    %xmm4, %xmm4, %xmm4
-;   pcmpeqq %xmm4, %xmm0, %xmm4
-;   ptest   %xmm4, %xmm4
+;   pxor    %xmm3, %xmm3, %xmm3
+;   pcmpeqq %xmm0, %xmm3, %xmm0
+;   ptest   %xmm0, %xmm0
 ;   setz    %al
 ;   movq    %rbp, %rsp
 ;   popq    %rbp

cranelift/filetests/filetests/isa/x64/sshr.clif

@@ -16,24 +16,25 @@ block0(v0: i128, v1: i8):
 ;   pushq   %rbp
 ;   movq    %rsp, %rbp
 ; block0:
-;   movzbq  %dl, %rdx
-;   movq    %rdx, %rcx
+;   movzbq  %dl, %rcx
 ;   shrq    %cl, %rdi, %rdi
-;   movq    %rsi, %r9
-;   sarq    %cl, %r9, %r9
+;   movq    %rsi, %rdx
+;   sarq    %cl, %rdx, %rdx
+;   movq    %rcx, %rax
 ;   movl    $64, %ecx
-;   subq    %rcx, %rdx, %rcx
-;   movq    %rsi, %r8
-;   shlq    %cl, %r8, %r8
-;   xorq    %r10, %r10, %r10
-;   testq   $127, %rdx
-;   cmovzq  %r10, %r8, %r8
-;   orq     %rdi, %r8, %rdi
+;   movq    %rax, %r11
+;   subq    %rcx, %r11, %rcx
+;   movq    %rsi, %rax
+;   shlq    %cl, %rax, %rax
+;   xorq    %r8, %r8, %r8
+;   testq   $127, %r11
+;   cmovzq  %r8, %rax, %rax
+;   orq     %rdi, %rax, %rdi
 ;   sarq    $63, %rsi, %rsi
-;   testq   $64, %rdx
-;   movq    %r9, %rax
+;   testq   $64, %r11
+;   movq    %rdx, %rax
 ;   cmovzq  %rdi, %rax, %rax
-;   cmovzq  %r9, %rsi, %rsi
+;   cmovzq  %rdx, %rsi, %rsi
 ;   movq    %rsi, %rdx
 ;   movq    %rbp, %rsp
 ;   popq    %rbp

cranelift/filetests/filetests/isa/x64/ushr.clif

@@ -15,24 +15,24 @@ block0(v0: i128, v1: i8):
 ;   pushq   %rbp
 ;   movq    %rsp, %rbp
 ; block0:
-;   movzbq  %dl, %rdx
-;   movq    %rdx, %rcx
+;   movzbq  %dl, %rcx
 ;   shrq    %cl, %rdi, %rdi
-;   movq    %rsi, %r9
-;   shrq    %cl, %r9, %r9
+;   movq    %rsi, %r8
+;   shrq    %cl, %r8, %r8
+;   movq    %rcx, %rax
 ;   movl    $64, %ecx
-;   movq    %rdx, %r10
-;   subq    %rcx, %r10, %rcx
+;   movq    %rax, %r11
+;   subq    %rcx, %r11, %rcx
 ;   shlq    %cl, %rsi, %rsi
-;   xorq    %r8, %r8, %r8
-;   testq   $127, %r10
-;   cmovzq  %r8, %rsi, %rsi
+;   xorq    %rax, %rax, %rax
+;   testq   $127, %r11
+;   cmovzq  %rax, %rsi, %rsi
 ;   orq     %rsi, %rdi, %rsi
 ;   xorq    %rdx, %rdx, %rdx
-;   testq   $64, %r10
-;   movq    %r9, %rax
+;   testq   $64, %r11
+;   movq    %r8, %rax
 ;   cmovzq  %rsi, %rax, %rax
-;   cmovzq  %r9, %rdx, %rdx
+;   cmovzq  %r8, %rdx, %rdx
 ;   movq    %rbp, %rsp
 ;   popq    %rbp
 ;   ret
@@ -188,7 +188,8 @@ block0(v0: i32, v1: i64, v2: i64):
 ;   pushq   %rbp
 ;   movq    %rsp, %rbp
 ; block0:
-;   movq    %rsi, %rcx
+;   movq    %rsi, %r9
+;   movq    %r9, %rcx
 ;   shrl    %cl, %edi, %edi
 ;   movq    %rdi, %rax
 ;   movq    %rbp, %rsp

cranelift/filetests/filetests/isa/x64/vhigh_bits.clif

@@ -0,0 +1,75 @@
+test compile precise-output
+target x86_64
+
+function %f1(i8x16) -> i8 {
+block0(v0: i8x16):
+  v1 = vhigh_bits.i8 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   pmovmskb %xmm0, %eax
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f2(i8x16) -> i16 {
+block0(v0: i8x16):
+  v1 = vhigh_bits.i16 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   pmovmskb %xmm0, %eax
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f3(i16x8) -> i8 {
+block0(v0: i16x8):
+  v1 = vhigh_bits.i8 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   packsswb %xmm0, %xmm0, %xmm0
+;   pmovmskb %xmm0, %eax
+;   shrq    $8, %rax, %rax
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f4(i32x4) -> i8 {
+block0(v0: i32x4):
+  v1 = vhigh_bits.i8 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   movmskps %xmm0, %eax
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+
+function %f5(i64x2) -> i8 {
+block0(v0: i64x2):
+  v1 = vhigh_bits.i8 v0
+  return v1
+}
+
+;   pushq   %rbp
+;   movq    %rsp, %rbp
+; block0:
+;   movmskpd %xmm0, %eax
+;   movq    %rbp, %rsp
+;   popq    %rbp
+;   ret
+