diff options
| -rw-r--r-- | llvm/lib/Analysis/ScalarEvolutionExpander.cpp | 7 | ||||
| -rw-r--r-- | llvm/test/Analysis/ScalarEvolution/incorrect-offset-scaling.ll | 48 |
2 files changed, 55 insertions, 0 deletions
diff --git a/llvm/lib/Analysis/ScalarEvolutionExpander.cpp b/llvm/lib/Analysis/ScalarEvolutionExpander.cpp index 77164356d8e..77e4ec7ab40 100644 --- a/llvm/lib/Analysis/ScalarEvolutionExpander.cpp +++ b/llvm/lib/Analysis/ScalarEvolutionExpander.cpp @@ -1288,6 +1288,13 @@ Value *SCEVExpander::expandAddRecExprLiterally(const SCEVAddRecExpr *S) { if (!SE.dominates(Step, L->getHeader())) { PostLoopScale = Step; Step = SE.getConstant(Normalized->getType(), 1); + if (!Start->isZero()) { + // The normalization below assumes that Start is constant zero, so if + // it isn't re-associate Start to PostLoopOffset. + assert(!PostLoopOffset && "Start not-null but PostLoopOffset set?"); + PostLoopOffset = Start; + Start = SE.getConstant(Normalized->getType(), 0); + } Normalized = cast<SCEVAddRecExpr>(SE.getAddRecExpr( Start, Step, Normalized->getLoop(), diff --git a/llvm/test/Analysis/ScalarEvolution/incorrect-offset-scaling.ll b/llvm/test/Analysis/ScalarEvolution/incorrect-offset-scaling.ll new file mode 100644 index 00000000000..7ffb0936d10 --- /dev/null +++ b/llvm/test/Analysis/ScalarEvolution/incorrect-offset-scaling.ll @@ -0,0 +1,48 @@ +; RUN: opt -S -loop-reduce < %s | FileCheck %s + +target triple = "x86_64-unknown-unknown" +target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128" + +define void @incorrect_offset_scaling(i64, i64*) { +top: + br label %L + +L: ; preds = %idxend.10, %idxend, %L2, %top + br i1 undef, label %L, label %L1 + +L1: ; preds = %L1.preheader, %L2 + %r13 = phi i64 [ %r1, %L2 ], [ 1, %L ] +; CHECK: %lsr.iv = phi i64 [ 0, %L{{[^ ]+}} ], [ %lsr.iv.next, %L2 ] +; CHECK-NOT: %lsr.iv = phi i64 [ -1, %L{{[^ ]+}} ], [ %lsr.iv.next, %L2 ] +; CHECK: br + %r0 = add i64 %r13, -1 + br label %idxend.8 + +L2: ; preds = %idxend.8 + %r1 = add i64 %r13, 1 + br i1 undef, label %L, label %L1 + +if6: ; preds = %idxend.8 + %r2 = add i64 %0, -1 + %r3 = load i64, i64* %1, align 8 +; CHECK-NOT: %r2 +; CHECK: %r3 = load i64 + br label %ib + +idxend.8: ; preds = %L1 + br i1 undef, label %if6, label %L2 + +ib: ; preds = %if6 + %r4 = mul i64 %r3, %r0 + %r5 = add i64 %r2, %r4 + %r6 = icmp ult i64 %r5, undef +; CHECK %2 = mul i64 %lsr.iv, %r3 +; CHECK %3 = add i64 %1, -1 +; CHECK %4 = add i64 %0, %r3 +; CHECK %r6 + %r7 = getelementptr i64, i64* undef, i64 %r5 + store i64 1, i64* %r7, align 8 +; CHECK %5 = mul i64 %lsr.iv, %r3 +; CHECK %6 = add i64 %5, -1 + br label %L +} |
