The patch fixes PR27392.

Summary:
 It is incorrect to compare TripCount (which is BECount + 1)
  with extraiters (or Count) to check if we should enter unrolled
  loop or not, because TripCount can potentially overflow
  (when BECount is max unsigned integer).
 While comparing BECount with (Count - 1) is overflow safe and
  therefore correct.

Reviewer: hfinkel

Differential Revision: http://reviews.llvm.org/D19256

From: Evgeny Stupachenko <evstupac@gmail.com>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267662 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Transforms/Utils/LoopUnrollRuntime.cpp

@@ -246,7 +246,7 @@ static void ConnectEpilog(Loop *L, Value *ModVal, BasicBlock *NewExit,
 
   Instruction *InsertPt = NewExit->getTerminator();
   IRBuilder<> B(InsertPt);
-  Value *BrLoopExit = B.CreateIsNotNull(ModVal);
+  Value *BrLoopExit = B.CreateIsNotNull(ModVal, "lcmp.mod");
   assert(Exit && "Loop must have a single exit block only");
   // Split the exit to maintain loop canonicalization guarantees
   SmallVector<BasicBlock*, 4> Preds(predecessors(Exit));
@@ -416,7 +416,7 @@ static void CloneLoopBlocks(Loop *L, Value *NewIter,
 ///
 /// ***Epilog case***
 ///        extraiters = tripcount % loopfactor
-///        if (extraiters == tripcount) jump LoopExit:
+///        if (tripcount < loopfactor) jump LoopExit:
 ///        unroll_iters = tripcount - extraiters
 /// Loop:  LoopBody; (executes unroll_iter times);
 ///        unroll_iter -= 1
@@ -575,14 +575,15 @@ bool llvm::UnrollRuntimeLoopRemainder(Loop *L, unsigned Count,
                           ConstantInt::get(BECount->getType(), Count),
                           "xtraiter");
   }
-  Value *CmpOperand =
+  Value *BranchVal =
-      UseEpilogRemainder ? TripCount :
+      UseEpilogRemainder ? B.CreateICmpULT(BECount,
-                           ConstantInt::get(TripCount->getType(), 0);
+                                           ConstantInt::get(BECount->getType(),
-  Value *BranchVal = B.CreateICmpNE(ModVal, CmpOperand, "lcmp.mod");
+                                                            Count - 1)) :
-  BasicBlock *FirstLoop = UseEpilogRemainder ? NewPreHeader : PrologPreHeader;
+                           B.CreateIsNotNull(ModVal, "lcmp.mod");
-  BasicBlock *SecondLoop = UseEpilogRemainder ? NewExit : PrologExit;
+  BasicBlock *RemainderLoop = UseEpilogRemainder ? NewExit : PrologPreHeader;
+  BasicBlock *UnrollingLoop = UseEpilogRemainder ? NewPreHeader : PrologExit;
   // Branch to either remainder (extra iterations) loop or unrolling loop.
-  B.CreateCondBr(BranchVal, FirstLoop, SecondLoop);
+  B.CreateCondBr(BranchVal, RemainderLoop, UnrollingLoop);
   PreHeaderBR->eraseFromParent();
   Function *F = Header->getParent();
   // Get an ordered list of blocks in the loop to help with the ordering of the

test/Transforms/LoopUnroll/AArch64/runtime-loop.ll

@@ -4,16 +4,14 @@
 ; Tests for unrolling loops with run-time trip counts
 
 ; EPILOG:  %xtraiter = and i32 %n
-; EPILOG:  %lcmp.mod = icmp ne i32 %xtraiter, %n
+; EPILOG:  for.body:
-; EPILOG:  br i1 %lcmp.mod, label %for.body.preheader.new, label %for.end.loopexit.unr-lcssa
+; EPILOG:  %lcmp.mod = icmp ne i32 %xtraiter, 0
+; EPILOG:  br i1 %lcmp.mod, label %for.body.epil.preheader, label %for.end.loopexit
+; EPILOG:  for.body.epil:
 
 ; PROLOG:  %xtraiter = and i32 %n
 ; PROLOG:  %lcmp.mod = icmp ne i32 %xtraiter, 0
 ; PROLOG:  br i1 %lcmp.mod, label %for.body.prol.preheader, label %for.body.prol.loopexit
-
-; EPILOG:  for.body:
-; EPILOG:  for.body.epil:
-
 ; PROLOG:  for.body.prol:
 ; PROLOG:  for.body:
 

test/Transforms/LoopUnroll/runtime-loop.ll

@@ -6,8 +6,8 @@ target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f3
 ; Tests for unrolling loops with run-time trip counts
 
 ; EPILOG: %xtraiter = and i32 %n
-; EPILOG:  %lcmp.mod = icmp ne i32 %xtraiter, %n
+; EPILOG:  %lcmp.mod = icmp ne i32 %xtraiter, 0
-; EPILOG:  br i1 %lcmp.mod, label %for.body.preheader.new, label %for.end.loopexit.unr-lcssa
+; EPILOG:  br i1 %lcmp.mod, label %for.body.epil.preheader, label %for.end.loopexit
 
 ; PROLOG: %xtraiter = and i32 %n
 ; PROLOG:  %lcmp.mod = icmp ne i32 %xtraiter, 0

test/Transforms/LoopUnroll/runtime-loop1.ll

@@ -5,7 +5,7 @@
 
 
 ; EPILOG: for.body.preheader:
-; EPILOG:   br i1 %lcmp.mod, label %for.body.preheader.new, label %for.end.loopexit.unr-lcssa, !dbg [[PH_LOC:![0-9]+]]
+; EPILOG:   br i1 %1, label %for.end.loopexit.unr-lcssa, label %for.body.preheader.new, !dbg [[PH_LOC:![0-9]+]]
 ; EPILOG: for.body:
 ; EPILOG:   br i1 %niter.ncmp.1, label %for.end.loopexit.unr-lcssa.loopexit, label %for.body, !dbg [[BODY_LOC:![0-9]+]]
 ; EPILOG-NOT: br i1 %niter.ncmp.2, label %for.end.loopexit{{.*}}, label %for.body

test/Transforms/LoopUnroll/tripcount-overflow.ll

@@ -1,4 +1,5 @@
-; RUN: opt < %s -S -unroll-runtime -unroll-count=2 -loop-unroll | FileCheck %s
+; RUN: opt < %s -S -unroll-runtime -unroll-count=2 -loop-unroll | FileCheck %s -check-prefix=EPILOG
+; RUN: opt < %s -S -unroll-runtime -unroll-count=2 -loop-unroll -unroll-runtime-epilog=false | FileCheck %s -check-prefix=PROLOG
 target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
 
 ; This test case documents how runtime loop unrolling handles the case
@@ -9,17 +10,28 @@ target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
 ; is divisible by 2.  The prologue then branches to the unrolled loop
 ; and executes the 2^32 iterations there, in groups of 2.
 
+; EPILOG: entry:
 
-; CHECK: entry:
+; EPILOG-NEXT: %0 = add i32 %N, 1
-; CHECK-NEXT: %0 = add i32 %N, 1
+; EPILOG-NEXT: %xtraiter = and i32 %0, 1
-; CHECK-NEXT: %xtraiter = and i32 %0, 1
+; EPILOG-NEXT: %1 = icmp ult i32 %N, 1
-; CHECK-NEXT: %lcmp.mod = icmp ne i32 %xtraiter, %0
+; EPILOG-NEXT: br i1 %1, label %while.end.unr-lcssa, label %entry.new
-; CHECK-NEXT: br i1 %lcmp.mod, label %entry.new, label %while.end.unr-lcssa
+; EPILOG: while.body:
 
-; CHECK: while.body.epil:
+; EPILOG: %lcmp.mod = icmp ne i32 %xtraiter, 0
-; CHECK: br label %while.end.epilog-lcssa
+; EPILOG-NEXT: br i1 %lcmp.mod, label %while.body.epil.preheader, label %while.end
+; EPILOG: while.body.epil:
 
-; CHECK: while.end.epilog-lcssa:
+; PROLOG: entry:
+; PROLOG-NEXT: %0 = add i32 %N, 1
+; PROLOG-NEXT: %xtraiter = and i32 %0, 1
+; PROLOG-NEXT: %lcmp.mod = icmp ne i32 %xtraiter, 0
+; PROLOG-NEXT: br i1 %lcmp.mod, label %while.body.prol.preheader, label %while.body.prol.loopexit
+; PROLOG: while.body.prol:
+
+; PROLOG: %1 = icmp ult i32 %N, 1
+; PROLOG-NEXT: br i1 %1, label %while.end, label %entry.new
+; PROLOG: while.body:
 
 ; Function Attrs: nounwind readnone ssp uwtable
 define i32 @foo(i32 %N) {