[Analysis] Bail out for negative offsets in isDereferenceableAndAlignedInLoop

This patch is effectively NFC because it doesn't actually affect
any existing tests. It's not possible to even write a test for it,
because by accident negative offsets are caught by a subsequent
unsigned add overflow check. However, I think it's better to bail
out explicitly for negative offsets so that it's more consistent
with the unsigned remainder and add calculations.

Author: david-arm

llvm/lib/Analysis/Loads.cpp

@@ -313,6 +313,13 @@ bool llvm::isDereferenceableAndAlignedInLoop(LoadInst *LI, Loop *L,
     const auto *Offset = dyn_cast<SCEVConstant>(StartS->getOperand(0));
     const auto *NewBase = dyn_cast<SCEVUnknown>(StartS->getOperand(1));
     if (StartS->getNumOperands() == 2 && Offset && NewBase) {
+      // The following code below assumes the offset is unsigned, but GEP
+      // offsets are treated as signed so we can end up with a signed value
+      // here too. For example, suppose the initial PHI value is (i8 255),
+      // the offset will be treated as (i8 -1) and sign-extended to (i64 -1).
+      if (Offset->getAPInt().isNegative())
+        return false;
+
       // For the moment, restrict ourselves to the case where the offset is a
       // multiple of the requested alignment and the base is aligned.
       // TODO: generalize if a case found which warrants

llvm/test/Transforms/LoopVectorize/load-deref-pred-align.ll

@@ -211,8 +211,8 @@ define i8 @test_negative_off(i16 %len, ptr %test_base) {
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
-; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <2 x i8> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP10:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[PRED_LOAD_CONTINUE2:%.*]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <2 x i8> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP18:%.*]], [[PRED_LOAD_CONTINUE2]] ]
 ; CHECK-NEXT:    [[DOTCAST:%.*]] = trunc i32 [[INDEX]] to i16
 ; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = add i16 -1000, [[DOTCAST]]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i16 [[OFFSET_IDX]], 0
@@ -223,20 +223,35 @@ define i8 @test_negative_off(i16 %len, ptr %test_base) {
 ; CHECK-NEXT:    [[TMP5:%.*]] = load i1, ptr [[TMP3]], align 1
 ; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <2 x i1> poison, i1 [[TMP4]], i32 0
 ; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x i1> [[TMP6]], i1 [[TMP5]], i32 1
-; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr i8, ptr [[ALLOCA]], i16 [[TMP0]]
-; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr i8, ptr [[TMP8]], i32 0
-; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <2 x i8>, ptr [[TMP9]], align 1
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <2 x i1> [[TMP7]], <2 x i8> [[WIDE_LOAD]], <2 x i8> zeroinitializer
-; CHECK-NEXT:    [[TMP10]] = add <2 x i8> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <2 x i1> [[TMP7]], i32 0
+; CHECK-NEXT:    br i1 [[TMP8]], label [[PRED_LOAD_IF:%.*]], label [[PRED_LOAD_CONTINUE:%.*]]
+; CHECK:       pred.load.if:
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr i8, ptr [[ALLOCA]], i16 [[TMP0]]
+; CHECK-NEXT:    [[TMP10:%.*]] = load i8, ptr [[TMP9]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = insertelement <2 x i8> poison, i8 [[TMP10]], i32 0
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE]]
+; CHECK:       pred.load.continue:
+; CHECK-NEXT:    [[TMP12:%.*]] = phi <2 x i8> [ poison, [[VECTOR_BODY]] ], [ [[TMP11]], [[PRED_LOAD_IF]] ]
+; CHECK-NEXT:    [[TMP13:%.*]] = extractelement <2 x i1> [[TMP7]], i32 1
+; CHECK-NEXT:    br i1 [[TMP13]], label [[PRED_LOAD_IF1:%.*]], label [[PRED_LOAD_CONTINUE2]]
+; CHECK:       pred.load.if1:
+; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr i8, ptr [[ALLOCA]], i16 [[TMP1]]
+; CHECK-NEXT:    [[TMP15:%.*]] = load i8, ptr [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP16:%.*]] = insertelement <2 x i8> [[TMP12]], i8 [[TMP15]], i32 1
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE2]]
+; CHECK:       pred.load.continue2:
+; CHECK-NEXT:    [[TMP17:%.*]] = phi <2 x i8> [ [[TMP12]], [[PRED_LOAD_CONTINUE]] ], [ [[TMP16]], [[PRED_LOAD_IF1]] ]
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <2 x i1> [[TMP7]], <2 x i8> [[TMP17]], <2 x i8> zeroinitializer
+; CHECK-NEXT:    [[TMP18]] = add <2 x i8> [[VEC_PHI]], [[PREDPHI]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP11:%.*]] = icmp eq i32 [[INDEX_NEXT]], 12
-; CHECK-NEXT:    br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
+; CHECK-NEXT:    [[TMP19:%.*]] = icmp eq i32 [[INDEX_NEXT]], 12
+; CHECK-NEXT:    br i1 [[TMP19]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[TMP12:%.*]] = call i8 @llvm.vector.reduce.add.v2i8(<2 x i8> [[TMP10]])
+; CHECK-NEXT:    [[TMP20:%.*]] = call i8 @llvm.vector.reduce.add.v2i8(<2 x i8> [[TMP18]])
 ; CHECK-NEXT:    br i1 true, label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i16 [ -988, [[MIDDLE_BLOCK]] ], [ -1000, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i8 [ [[TMP12]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i8 [ [[TMP20]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i16 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
@@ -255,7 +270,7 @@ define i8 @test_negative_off(i16 %len, ptr %test_base) {
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i16 [[IV]], -990
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], !llvm.loop [[LOOP7:![0-9]+]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i8 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP12]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i8 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP20]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i8 [[ACCUM_NEXT_LCSSA]]
 ;
 entry: