[X86] Regenerate test using update_llc_test_checks.py.

llvm-svn: 321462

1 files changed, 128 insertions, 34 deletions

diff --git a/llvm/test/CodeGen/X86/gather-addresses.ll b/llvm/test/CodeGen/X86/gather-addresses.ll
index e09ad3e4e0b..6431847064f 100644
--- a/llvm/test/CodeGen/X86/gather-addresses.ll
+++ b/llvm/test/CodeGen/X86/gather-addresses.ll
@@ -1,3 +1,4 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=x86_64-linux -mcpu=nehalem < %s | FileCheck %s --check-prefix=LIN
 ; RUN: llc -mtriple=x86_64-win32 -mcpu=nehalem < %s | FileCheck %s --check-prefix=WIN
 ; RUN: llc -mtriple=i686-win32 -mcpu=nehalem < %s | FileCheck %s --check-prefix=LIN32
@@ -7,34 +8,59 @@
 ; use an efficient mov/shift sequence rather than shuffling each individual
 ; element out of the index vector.
 
-; CHECK-LABEL: foo:
-; LIN: movdqa	(%rsi), %xmm0
-; LIN: pand 	(%rdx), %xmm0
-; LIN: pextrq	$1, %xmm0, %r[[REG4:.+]]
-; LIN: movq 	%xmm0, %r[[REG2:.+]]
-; LIN: movslq	%e[[REG2]], %r[[REG1:.+]]
-; LIN: sarq    $32, %r[[REG2]]
-; LIN: movslq	%e[[REG4]], %r[[REG3:.+]]
-; LIN: sarq    $32, %r[[REG4]]
-; LIN: movsd    (%rdi,%r[[REG3]],8), %xmm1
-; LIN: movhpd   (%rdi,%r[[REG4]],8), %xmm1 
-; LIN: movq     %rdi, %xmm1 
-; LIN: movq     %r[[REG3]], %xmm0
-
-; WIN: movdqa	(%rdx), %xmm0
-; WIN: pand 	(%r8), %xmm0
-; WIN: pextrq	$1, %xmm0, %r[[REG4:.+]]
-; WIN: movq 	%xmm0, %r[[REG2:.+]]
-; WIN: movslq	%e[[REG2]], %r[[REG1:.+]]
-; WIN: sarq    $32, %r[[REG2]]
-; WIN: movslq	%e[[REG4]], %r[[REG3:.+]]
-; WIN: sarq    $32, %r[[REG4]]
-; WIN: movsd    (%rcx,%r[[REG3]],8), %xmm1
-; WIN: movhpd   (%rcx,%r[[REG4]],8), %xmm1
-; WIN: movdqa   (%r[[REG2]]), %xmm0
-; WIN: movq     %r[[REG2]], %xmm1
-
 define <4 x double> @foo(double* %p, <4 x i32>* %i, <4 x i32>* %h) nounwind {
+; LIN-LABEL: foo:
+; LIN:       # %bb.0:
+; LIN-NEXT:    movdqa (%rsi), %xmm0
+; LIN-NEXT:    pand (%rdx), %xmm0
+; LIN-NEXT:    pextrq $1, %xmm0, %rax
+; LIN-NEXT:    movq %xmm0, %rcx
+; LIN-NEXT:    movslq %ecx, %rdx
+; LIN-NEXT:    sarq $32, %rcx
+; LIN-NEXT:    movslq %eax, %rsi
+; LIN-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; LIN-NEXT:    movhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
+; LIN-NEXT:    sarq $32, %rax
+; LIN-NEXT:    movsd {{.*#+}} xmm1 = mem[0],zero
+; LIN-NEXT:    movhpd {{.*#+}} xmm1 = xmm1[0],mem[0]
+; LIN-NEXT:    retq
+;
+; WIN-LABEL: foo:
+; WIN:       # %bb.0:
+; WIN-NEXT:    movdqa (%rdx), %xmm0
+; WIN-NEXT:    pand (%r8), %xmm0
+; WIN-NEXT:    pextrq $1, %xmm0, %rax
+; WIN-NEXT:    movq %xmm0, %rdx
+; WIN-NEXT:    movslq %edx, %r8
+; WIN-NEXT:    sarq $32, %rdx
+; WIN-NEXT:    movslq %eax, %r9
+; WIN-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; WIN-NEXT:    movhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
+; WIN-NEXT:    sarq $32, %rax
+; WIN-NEXT:    movsd {{.*#+}} xmm1 = mem[0],zero
+; WIN-NEXT:    movhpd {{.*#+}} xmm1 = xmm1[0],mem[0]
+; WIN-NEXT:    retq
+;
+; LIN32-LABEL: foo:
+; LIN32:       # %bb.0:
+; LIN32-NEXT:    pushl %edi
+; LIN32-NEXT:    pushl %esi
+; LIN32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; LIN32-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; LIN32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; LIN32-NEXT:    movdqa (%edx), %xmm0
+; LIN32-NEXT:    pand (%ecx), %xmm0
+; LIN32-NEXT:    pextrd $1, %xmm0, %ecx
+; LIN32-NEXT:    pextrd $2, %xmm0, %edx
+; LIN32-NEXT:    pextrd $3, %xmm0, %esi
+; LIN32-NEXT:    movd %xmm0, %edi
+; LIN32-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; LIN32-NEXT:    movhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
+; LIN32-NEXT:    movsd {{.*#+}} xmm1 = mem[0],zero
+; LIN32-NEXT:    movhpd {{.*#+}} xmm1 = xmm1[0],mem[0]
+; LIN32-NEXT:    popl %esi
+; LIN32-NEXT:    popl %edi
+; LIN32-NEXT:    retl
   %a = load <4 x i32>, <4 x i32>* %i
   %b = load <4 x i32>, <4 x i32>* %h
   %j = and <4 x i32> %a, %b
@@ -60,13 +86,81 @@ define <4 x double> @foo(double* %p, <4 x i32>* %i, <4 x i32>* %h) nounwind {
 ; Check that the sequence previously used above, which bounces the vector off the
 ; cache works for x86-32. Note that in this case it will not be used for index
 ; calculation, since indexes are 32-bit, not 64.
-; CHECK-LABEL: old:
-; LIN32: movaps	%xmm0, (%esp)
-; LIN32-DAG: {{(mov|and)}}l	(%esp),
-; LIN32-DAG: {{(mov|and)}}l	4(%esp),
-; LIN32-DAG: {{(mov|and)}}l	8(%esp),
-; LIN32-DAG: {{(mov|and)}}l	12(%esp),
 define <4 x i64> @old(double* %p, <4 x i32>* %i, <4 x i32>* %h, i64 %f) nounwind {
+; LIN-LABEL: old:
+; LIN:       # %bb.0:
+; LIN-NEXT:    movdqa (%rsi), %xmm0
+; LIN-NEXT:    pand (%rdx), %xmm0
+; LIN-NEXT:    pextrq $1, %xmm0, %rax
+; LIN-NEXT:    movq %rax, %rdx
+; LIN-NEXT:    shrq $32, %rdx
+; LIN-NEXT:    movq %xmm0, %rsi
+; LIN-NEXT:    movq %rsi, %rdi
+; LIN-NEXT:    shrq $32, %rdi
+; LIN-NEXT:    andl %ecx, %esi
+; LIN-NEXT:    andl %ecx, %eax
+; LIN-NEXT:    andq %rcx, %rdi
+; LIN-NEXT:    andq %rcx, %rdx
+; LIN-NEXT:    movq %rdi, %xmm1
+; LIN-NEXT:    movq %rsi, %xmm0
+; LIN-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; LIN-NEXT:    movq %rdx, %xmm2
+; LIN-NEXT:    movq %rax, %xmm1
+; LIN-NEXT:    punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; LIN-NEXT:    retq
+;
+; WIN-LABEL: old:
+; WIN:       # %bb.0:
+; WIN-NEXT:    movdqa (%rdx), %xmm0
+; WIN-NEXT:    pand (%r8), %xmm0
+; WIN-NEXT:    pextrq $1, %xmm0, %r8
+; WIN-NEXT:    movq %r8, %rcx
+; WIN-NEXT:    shrq $32, %rcx
+; WIN-NEXT:    movq %xmm0, %rax
+; WIN-NEXT:    movq %rax, %rdx
+; WIN-NEXT:    shrq $32, %rdx
+; WIN-NEXT:    andl %r9d, %eax
+; WIN-NEXT:    andl %r9d, %r8d
+; WIN-NEXT:    andq %r9, %rdx
+; WIN-NEXT:    andq %r9, %rcx
+; WIN-NEXT:    movq %rdx, %xmm1
+; WIN-NEXT:    movq %rax, %xmm0
+; WIN-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; WIN-NEXT:    movq %rcx, %xmm2
+; WIN-NEXT:    movq %r8, %xmm1
+; WIN-NEXT:    punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; WIN-NEXT:    retq
+;
+; LIN32-LABEL: old:
+; LIN32:       # %bb.0:
+; LIN32-NEXT:    pushl %ebp
+; LIN32-NEXT:    movl %esp, %ebp
+; LIN32-NEXT:    pushl %esi
+; LIN32-NEXT:    andl $-16, %esp
+; LIN32-NEXT:    subl $32, %esp
+; LIN32-NEXT:    movl 20(%ebp), %eax
+; LIN32-NEXT:    movl 16(%ebp), %ecx
+; LIN32-NEXT:    movl 12(%ebp), %edx
+; LIN32-NEXT:    movaps (%edx), %xmm0
+; LIN32-NEXT:    andps (%ecx), %xmm0
+; LIN32-NEXT:    movaps %xmm0, (%esp)
+; LIN32-NEXT:    movl (%esp), %ecx
+; LIN32-NEXT:    andl %eax, %ecx
+; LIN32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; LIN32-NEXT:    andl %eax, %edx
+; LIN32-NEXT:    movl {{[0-9]+}}(%esp), %esi
+; LIN32-NEXT:    andl %eax, %esi
+; LIN32-NEXT:    andl {{[0-9]+}}(%esp), %eax
+; LIN32-NEXT:    movd %edx, %xmm1
+; LIN32-NEXT:    movd %ecx, %xmm0
+; LIN32-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; LIN32-NEXT:    movd %eax, %xmm2
+; LIN32-NEXT:    movd %esi, %xmm1
+; LIN32-NEXT:    punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; LIN32-NEXT:    leal -4(%ebp), %esp
+; LIN32-NEXT:    popl %esi
+; LIN32-NEXT:    popl %ebp
+; LIN32-NEXT:    retl
   %a = load <4 x i32>, <4 x i32>* %i
   %b = load <4 x i32>, <4 x i32>* %h
   %j = and <4 x i32> %a, %b
@@ -77,7 +171,7 @@ define <4 x i64> @old(double* %p, <4 x i32>* %i, <4 x i32>* %h, i64 %f) nounwind
   %q0 = zext i32 %d0 to i64
   %q1 = zext i32 %d1 to i64
   %q2 = zext i32 %d2 to i64
-  %q3 = zext i32 %d3 to i64  
+  %q3 = zext i32 %d3 to i64
   %r0 = and i64 %q0, %f
   %r1 = and i64 %q1, %f
   %r2 = and i64 %q2, %f