From d779d350372cb1f684d739d5db013085cd2ac8a6 Mon Sep 17 00:00:00 2001
From: Simon Pilgrim <llvm-dev@redking.me.uk>
Date: Mon, 12 Oct 2015 23:07:06 +0000
Subject: [PATCH] [InstCombine] Tidied up SSE4A tests.

First stage of bugfix discussed in D13348

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@250121 91177308-0d34-0410-b5e6-96231b3b80d8
---
 test/Transforms/InstCombine/x86-sse4a.ll | 76 ++++++++++++------------
 1 file changed, 39 insertions(+), 37 deletions(-)

diff --git a/test/Transforms/InstCombine/x86-sse4a.ll b/test/Transforms/InstCombine/x86-sse4a.ll
index 4e2cd2f93c7..40404f39b97 100644
--- a/test/Transforms/InstCombine/x86-sse4a.ll
+++ b/test/Transforms/InstCombine/x86-sse4a.ll
@@ -1,10 +1,10 @@
 ; RUN: opt < %s -instcombine -S | FileCheck %s
 
 ; We should optimize these two redundant insertqi into one
-; CHECK: define <2 x i64> @testInsertTwice(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertTwice(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 32)
-; CHECK-NOT: insertqi
+; CHECK-LABEL: @testInsertTwice
+; CHECK-NEXT: %1 = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 32)
+; CHECK-NEXT: ret <2 x i64> %1
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 32)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 32, i8 32)
   ret <2 x i64> %2
@@ -13,110 +13,112 @@ define <2 x i64> @testInsertTwice(<2 x i64> %v, <2 x i64> %i) {
 ; The result of this insert is the second arg, since the top 64 bits of
 ; the result are undefined, and we copy the bottom 64 bits from the
 ; second arg
-; CHECK: define <2 x i64> @testInsert64Bits(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsert64Bits(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: ret <2 x i64> %i
+; CHECK-LABEL: @testInsert64Bits
+; CHECK-NEXT: ret <2 x i64> %i
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 64, i8 0)
   ret <2 x i64> %1
 }
 
 ; Test the several types of ranges and ordering that exist for two insertqi
-; CHECK: define <2 x i64> @testInsertContainedRange(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertContainedRange(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: %[[RES:.*]] = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 0)
-; CHECK: ret <2 x i64> %[[RES]]
+; CHECK-LABEL: @testInsertContainedRange
+; CHECK: %1 = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 0)
+; CHECK: ret <2 x i64> %1
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 0)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 16)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testInsertContainedRange_2(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertContainedRange_2(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: %[[RES:.*]] = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 0)
-; CHECK: ret <2 x i64> %[[RES]]
+; CHECK-LABEL: @testInsertContainedRange_2
+; CHECK-NEXT: %1 = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 0)
+; CHECK-NEXT: ret <2 x i64> %1
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 16)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 32, i8 0)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testInsertOverlappingRange(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertOverlappingRange(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: %[[RES:.*]] = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
-; CHECK: ret <2 x i64> %[[RES]]
+; CHECK-LABEL: @testInsertOverlappingRange
+; CHECK-NEXT: %1 = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
+; CHECK-NEXT: ret <2 x i64> %1
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 0)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 32, i8 16)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testInsertOverlappingRange_2(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertOverlappingRange_2(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: %[[RES:.*]] = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
-; CHECK: ret <2 x i64> %[[RES]]
+; CHECK-LABEL: @testInsertOverlappingRange_2
+; CHECK-NEXT: %1 = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
+; CHECK-NEXT: ret <2 x i64> %1
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 16)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 32, i8 0)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testInsertAdjacentRange(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertAdjacentRange(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: %[[RES:.*]] = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
-; CHECK: ret <2 x i64> %[[RES]]
+; CHECK-LABEL: @testInsertAdjacentRange
+; CHECK-NEXT: %1 = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
+; CHECK-NEXT: ret <2 x i64> %1
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 32, i8 0)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 32)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testInsertAdjacentRange_2(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertAdjacentRange_2(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: %[[RES:.*]] = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
-; CHECK: ret <2 x i64> %[[RES]]
+; CHECK-LABEL: @testInsertAdjacentRange_2
+; CHECK-NEXT: %1 = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 0)
+; CHECK-NEXT: ret <2 x i64> %1
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 32)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 32, i8 0)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testInsertDisjointRange(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertDisjointRange(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 0)
-; CHECK: tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 32)
+; CHECK-LABEL: @testInsertDisjointRange
+; CHECK-NEXT: %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 0)
+; CHECK-NEXT: %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 32)
+; CHECK-NEXT: ret <2 x i64> %2
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 0)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 32)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testInsertDisjointRange_2(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testInsertDisjointRange_2(<2 x i64> %v, <2 x i64> %i) {
-; CHECK:  tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 0)
-; CHECK:  tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 32)
+; CHECK-LABEL: @testInsertDisjointRange_2
+; CHECK-NEXT: %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 0)
+; CHECK-NEXT: %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 32)
+; CHECK-NEXT: ret <2 x i64> %2
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 16, i8 0)
   %2 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %1, <2 x i64> %i, i8 16, i8 32)
   ret <2 x i64> %2
 }
 
-; CHECK: define <2 x i64> @testZeroLength(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testZeroLength(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: ret <2 x i64> %i
+; CHECK-LABEL: @testZeroLength
+; CHECK-NEXT: ret <2 x i64> %i
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 0, i8 0)
   ret <2 x i64> %1
 }
 
-; CHECK: define <2 x i64> @testUndefinedInsertq_1(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testUndefinedInsertq_1(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: ret <2 x i64> undef
+; CHECK-LABEL: @testUndefinedInsertq_1
+; CHECK-NEXT: ret <2 x i64> undef
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 0, i8 16)
   ret <2 x i64> %1
 }
 
-; CHECK: define <2 x i64> @testUndefinedInsertq_2(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testUndefinedInsertq_2(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: ret <2 x i64> undef
+; CHECK-LABEL: @testUndefinedInsertq_2
+; CHECK-NEXT: ret <2 x i64> undef
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 48, i8 32)
   ret <2 x i64> %1
 }
 
-; CHECK: define <2 x i64> @testUndefinedInsertq_3(<2 x i64> %v, <2 x i64> %i)
 define <2 x i64> @testUndefinedInsertq_3(<2 x i64> %v, <2 x i64> %i) {
-; CHECK: ret <2 x i64> undef
+; CHECK-LABEL: @testUndefinedInsertq_3
+; CHECK-NEXT: ret <2 x i64> undef
   %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 64, i8 16)
   ret <2 x i64> %1
 }
-- 
2.34.1