ValueTracking: Make isBytewiseValue simpler and more powerful at the same time.

Turns out there is a simpler way of checking that all bytes in a word are equal
than binary decomposition.

llvm-svn: 228503

diff --git a/llvm/lib/Analysis/ValueTracking.cpp b/llvm/lib/Analysis/ValueTracking.cpp
index 91b6469..5ae8574 100644 (file)
--- a/llvm/lib/Analysis/ValueTracking.cpp
+++ b/llvm/lib/Analysis/ValueTracking.cpp
@@ -2121,26 +2121,16 @@ Value *llvm::isBytewiseValue(Value *V) {
     // Don't handle long double formats, which have strange constraints.
   }
 
-  // We can handle constant integers that are power of two in size and a
-  // multiple of 8 bits.
+  // We can handle constant integers that are multiple of 8 bits.
   if (ConstantInt *CI = dyn_cast<ConstantInt>(V)) {
-    unsigned Width = CI->getBitWidth();
-    if (isPowerOf2_32(Width) && Width > 8) {
-      // We can handle this value if the recursive binary decomposition is the
-      // same at all levels.
-      APInt Val = CI->getValue();
-      APInt Val2;
-      while (Val.getBitWidth() != 8) {
-        unsigned NextWidth = Val.getBitWidth()/2;
-        Val2  = Val.lshr(NextWidth);
-        Val2 = Val2.trunc(Val.getBitWidth()/2);
-        Val = Val.trunc(Val.getBitWidth()/2);
-
-        // If the top/bottom halves aren't the same, reject it.
-        if (Val != Val2)
-          return nullptr;
-      }
-      return ConstantInt::get(V->getContext(), Val);
+    if (CI->getBitWidth() % 8 == 0) {
+      assert(CI->getBitWidth() > 8 && "8 bits should be handled above!");
+
+      // We can check that all bytes of an integer are equal by making use of a
+      // little trick: rotate by 8 and check if it's still the same value.
+      if (CI->getValue() != CI->getValue().rotl(8))
+        return nullptr;
+      return ConstantInt::get(V->getContext(), CI->getValue().trunc(8));
     }
   }
 

diff --git a/llvm/test/Transforms/MemCpyOpt/form-memset.ll b/llvm/test/Transforms/MemCpyOpt/form-memset.ll
index d980b7f..0a40943 100644 (file)
--- a/llvm/test/Transforms/MemCpyOpt/form-memset.ll
+++ b/llvm/test/Transforms/MemCpyOpt/form-memset.ll
@@ -284,3 +284,18 @@ define void @test10(i8* nocapture %P) nounwind {
 ; CHECK-NOT: memset
 ; CHECK: ret void
 }
+
+; Memset followed by odd store.
+define void @test11(i32* nocapture %P) nounwind ssp {
+entry:
+  %add.ptr = getelementptr inbounds i32* %P, i64 3
+  %0 = bitcast i32* %add.ptr to i8*
+  tail call void @llvm.memset.p0i8.i64(i8* %0, i8 1, i64 11, i32 1, i1 false)
+  %arrayidx = getelementptr inbounds i32* %P, i64 0
+  %arrayidx.cast = bitcast i32* %arrayidx to i96*
+  store i96 310698676526526814092329217, i96* %arrayidx.cast, align 4
+  ret void
+; CHECK-LABEL: @test11(
+; CHECK-NOT: store
+; CHECK: call void @llvm.memset.p0i8.i64(i8* %1, i8 1, i64 23, i32 4, i1 false)
+}