Implement rdar://6480391, extending of equality icmp's to avoid a truncation.

I noticed this in the code compiled for a routine using std::map, which produced
this code:
	%25 = tail call i32 @memcmp(i8* %24, i8* %23, i32 6) nounwind readonly
	%.lobit.i = lshr i32 %25, 31		; <i32> [#uses=1]
	%tmp.i = trunc i32 %.lobit.i to i8		; <i8> [#uses=1]
	%toBool = icmp eq i8 %tmp.i, 0		; <i1> [#uses=1]
	br i1 %toBool, label %bb3, label %bb4
which compiled to:

	call	L_memcmp$stub
	shrl	$31, %eax
	testb	%al, %al
	jne	LBB1_11	## 

with this change, we compile it to:

	call	L_memcmp$stub
	testl	%eax, %eax
	js	LBB1_11

This triggers all the time in common code, with patters like this:

	%169 = and i32 %ply, 1		; <i32> [#uses=1]
	%170 = trunc i32 %169 to i8		; <i8> [#uses=1]
	%toBool = icmp ne i8 %170, 0		; <i1> [#uses=1]

 	%7 = lshr i32 %6, 24		; <i32> [#uses=1]
	%9 = trunc i32 %7 to i8		; <i8> [#uses=1]
	%10 = icmp ne i8 %9, 0		; <i1> [#uses=1]

etc



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@61985 91177308-0d34-0410-b5e6-96231b3b80d8

3 files changed, 41 insertions, 4 deletions

diff --git a/lib/Transforms/Scalar/InstructionCombining.cpp b/lib/Transforms/Scalar/InstructionCombining.cpp
index f59874b1a5..c1a8eab95b 100644
--- a/lib/Transforms/Scalar/InstructionCombining.cpp
+++ b/lib/Transforms/Scalar/InstructionCombining.cpp
@@ -6343,6 +6343,28 @@ Instruction *InstCombiner::visitICmpInstWithInstAndIntCst(ICmpInst &ICI,
   const APInt &RHSV = RHS->getValue();
   
   switch (LHSI->getOpcode()) {
+  case Instruction::Trunc:
+    if (ICI.isEquality() && LHSI->hasOneUse()) {
+      // Simplify icmp eq (trunc x to i8), 42 -> icmp eq x, 42|highbits if all
+      // of the high bits truncated out of x are known.
+      unsigned DstBits = LHSI->getType()->getPrimitiveSizeInBits(),
+             SrcBits = LHSI->getOperand(0)->getType()->getPrimitiveSizeInBits();
+      APInt Mask(APInt::getHighBitsSet(SrcBits, SrcBits-DstBits));
+      APInt KnownZero(SrcBits, 0), KnownOne(SrcBits, 0);
+      ComputeMaskedBits(LHSI->getOperand(0), Mask, KnownZero, KnownOne);
+      
+      // If all the high bits are known, we can do this xform.
+      if ((KnownZero|KnownOne).countLeadingOnes() >= SrcBits-DstBits) {
+        // Pull in the high bits from known-ones set.
+        APInt NewRHS(RHS->getValue());
+        NewRHS.zext(SrcBits);
+        NewRHS |= KnownOne;
+        return new ICmpInst(ICI.getPredicate(), LHSI->getOperand(0),
+                            ConstantInt::get(NewRHS));
+      }
+    }
+    break;
+      
   case Instruction::Xor:         // (icmp pred (xor X, XorCST), CI)
     if (ConstantInt *XorCST = dyn_cast<ConstantInt>(LHSI->getOperand(1))) {
       // If this is a comparison that tests the signbit (X < 0) or (x > -1),
diff --git a/test/Transforms/InstCombine/cast.ll b/test/Transforms/InstCombine/cast.ll
index dc5e92800f..9361ff24e9 100644
--- a/test/Transforms/InstCombine/cast.ll
+++ b/test/Transforms/InstCombine/cast.ll
@@ -1,7 +1,5 @@
 ; Tests to make sure elimination of casts is working correctly
-; RUN: llvm-as < %s | opt -instcombine | llvm-dis | \
-; RUN:    grep %c | notcast
-; END.
+; RUN: llvm-as < %s | opt -instcombine | llvm-dis | grep %c | notcast
 
 @inbuf = external global [32832 x i8]           ; <[32832 x i8]*> [#uses=1]
 
@@ -238,3 +236,21 @@ define i16 @test35(i16 %a) {
         ret i16 %c2
 }
 
+; icmp sgt i32 %a, -1
+; rdar://6480391
+define i1 @test36(i32 %a) {
+        %b = lshr i32 %a, 31
+        %c = trunc i32 %b to i8
+        %d = icmp eq i8 %c, 0
+        ret i1 %d
+}
+
+; ret i1 false
+define i1 @test37(i32 %a) {
+        %b = lshr i32 %a, 31
+        %c = or i32 %b, 512
+        %d = trunc i32 %c to i8
+        %e = icmp eq i8 %d, 11
+        ret i1 %e
+}
+
diff --git a/test/Transforms/InstCombine/icmp.ll b/test/Transforms/InstCombine/icmp.ll
index bb69ba3be4..27136d6943 100644
--- a/test/Transforms/InstCombine/icmp.ll
+++ b/test/Transforms/InstCombine/icmp.ll
@@ -1,5 +1,4 @@
 ; RUN: llvm-as < %s | opt -instcombine | llvm-dis | not grep icmp
-; END.
 
 define i32 @test1(i32 %X) {
 entry: