[LoopUnroll] Separate peeling from unrolling

[nikic]

Summary

Loop peeling is currently performed as part of UnrollLoop(). The whole setup is somewhat odd, here is my current understanding of the situation:

• Outside test scenarios, peeling will always be performed as part of an unroll with Count=1: https://github.com/llvm/llvm-project/blob/ffb48d48e45c72ed81dda4983ccb06e800cdbbd0/llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp#L873-L879
• Of course, unrolling with count one does not do a great deal. My deduced reason for doing this is to run simplifyLoopAfterUnroll(), without which many tests indeed fail.
• When testing, it's possible to specify both -unroll-peel-count and -unroll-count, in which case both peeling and unrolling will be performed. However, if you actually try to do this, you apparently run into this miscompile: https://bugs.llvm.org/show_bug.cgi?id=45939 / D80080

Based on that reasoning (and suggested by @efriedma on the referenced review) I've moved the peeling code out of UnrollLoop() into tryToUnrollLoop(). An error is thrown if a test explicitly requests to perform both peeling and unrolling.

TBH the way peeling is implemented has left me very confused, so I'm not sure I'm doing the right thing here.

[lebedev.ri]

This doesn't make sense to me.
So now if we ask to unroll loop, and peel stuff first, it will only do the peeling?
The tests even show as much.

[nikic]

Yes, silently ignoring one option isn't good. I've changed the code to emit an error if both options are specified.

[lebedev.ri]

I was actually talking about not the explicit options, but how this works in the pipeline,
and i guess this somehow does not affect that use-case at all.
IMO that becomes even more confusing, but since this doesn't regress pipelines, i don't care.

[nikic]

In D103362#2788792, @lebedev.ri wrote:

I was actually talking about not the explicit options, but how this works in the pipeline,
and i guess this somehow does not affect that use-case at all.
IMO that becomes even more confusing, but since this doesn't regress pipelines, i don't care.

Yes, this does not affect the optimization pipeline, it only affects tests that specify explicit unroll counts.

I think it is worth noting that there are cases where performing both (non-PGO) peeling and unrolling could be sensible, namely when invariance peeling renders a previously non-analyzable loop analyzable. However, this requires that we first perform a peel and then re-analyze the loop for unrolling profitability from scratch, not that we perform a combined peel and unroll in a single step. As such, even if we want to support both peeling and unrolling a loop in the future, I believe this patch would still be the necessary first step in that direction.

[reames]

I'm not sure we can drop the behavior of both peeling and unrolling. The tests use the command lines, but I believe there's a pragma/metadata mechanism which can be used to achieve the same effect. (Vague memory of prior conversations w/ @Meinersbur)

Supporting the existing behavior with the new code structure doesn't seem too hard, would you mind updating to continue supporting both? (Well, while fixing the miscompile noted in D103620.)

[nikic]

In D103362#2796508, @reames wrote:

I'm not sure we can drop the behavior of both peeling and unrolling. The tests use the command lines, but I believe there's a pragma/metadata mechanism which can be used to achieve the same effect. (Vague memory of prior conversations w/ @Meinersbur)

While there is indeed metadata to set an unroll count, I don't believe there is metadata to set a peel count. If an unroll count is set via metadata, we will not attempt to compute a peel count.

Supporting the existing behavior with the new code structure doesn't seem too hard, would you mind updating to continue supporting both? (Well, while fixing the miscompile noted in D103620.)

I don't think the existing behavior makes sense, so I would prefer not to preserve it. It seems pointless to go out of our way to allow a scenario in testing that cannot appear as part of the optimization pipeline.

[reames]

@nikic I found your last comment convincing.

[reames]

LGTM

llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp

1177

In the future, we really should pull out a function with the other loop simplification, and use that here, but that's for the future.

[nikic]

If an explicit PeelCount has been provided for testing, use it directly and don't try to go through other unroll heuristics. Otherwise we might try to use the explicit test PeelCount together with a full unroll heuristic. (Once again, this doesn't affect the optimization pipeline, just testing.)

Also preserve existing tests by removing explicit unroll counts from tests that already have an explicit peel count. The tests are still reasonable apart from that.

[reames]

LGTM still holds.

llvm/include/llvm/Transforms/Utils/UnrollLoop.h

struct UnrollLoopOptions {
   unsigned Count;
   unsigned TripCount;
   bool Force;
   bool AllowRuntime;
   bool AllowExpensiveTripCount;
   bool PreserveCondBr;
   bool PreserveOnlyFirst;
   unsigned TripMultiple;
-  unsigned PeelCount;
   bool UnrollRemainder;
   bool ForgetAllSCEV;
 };
 
 LoopUnrollResult UnrollLoop(Loop *L, UnrollLoopOptions ULO, LoopInfo *LI,
                             ScalarEvolution *SE, DominatorTree *DT,
                             AssumptionCache *AC,
                             const llvm::TargetTransformInfo *TTI,

llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp

bool IsCountSetExplicitly = computeUnrollCount(
       L, TTI, DT, LI, SE, EphValues, &ORE, TripCount, MaxTripCount, MaxOrZero,
       TripMultiple, LoopSize, UP, PP, UseUpperBound);
   if (!UP.Count)
     return LoopUnrollResult::Unmodified;
   // Unroll factor (Count) must be less or equal to TripCount.
   if (TripCount && UP.Count > TripCount)
     UP.Count = TripCount;
 
+  if (PP.PeelCount) {
+    if (UP.Count != 1) {
+      report_fatal_error("Cannot specify both explicit peel count and "
+                         "explicit unroll count");
+    }
+
+    LLVM_DEBUG(dbgs() << "PEELING loop %" << L->getHeader()->getName()
+                      << " with iteration count " << PP.PeelCount << "!\n");
+    ORE.emit([&]() {
+      return OptimizationRemark(DEBUG_TYPE, "Peeled", L->getStartLoc(),
+                                L->getHeader())
+             << " peeled loop by " << ore::NV("PeelCount", PP.PeelCount)
+             << " iterations";
+    });
+
+    if (peelLoop(L, PP.PeelCount, LI, &SE, &DT, &AC, PreserveLCSSA)) {
+      simplifyLoopAfterUnroll(L, true, LI, &SE, &DT, &AC, &TTI);
+      // If the loop was peeled, we already "used up" the profile information
+      // we had, so we don't want to unroll or peel again.
+      if (PP.PeelProfiledIterations)
+        L->setLoopAlreadyUnrolled();
+      return LoopUnrollResult::PartiallyUnrolled;
+    }
+    return LoopUnrollResult::Unmodified;
+  }
+
   // Save loop properties before it is transformed.
   MDNode *OrigLoopID = L->getLoopID();
 
   // Unroll the loop.
   Loop *RemainderLoop = nullptr;
   LoopUnrollResult UnrollResult = UnrollLoop(
       L,
       {UP.Count, TripCount, UP.Force, UP.Runtime, UP.AllowExpensiveTripCount,
-       UseUpperBound, MaxOrZero, TripMultiple, PP.PeelCount, UP.UnrollRemainder,
+       UseUpperBound, MaxOrZero, TripMultiple, UP.UnrollRemainder,
        ForgetAllSCEV},
       LI, &SE, &DT, &AC, &TTI, &ORE, PreserveLCSSA, &RemainderLoop);
   if (UnrollResult == LoopUnrollResult::Unmodified)
     return LoopUnrollResult::Unmodified;
 
   if (RemainderLoop) {
     Optional<MDNode *> RemainderLoopID =
         makeFollowupLoopID(OrigLoopID, {LLVMLoopUnrollFollowupAll,
if (NewLoopID.hasValue()) {
       // Do not setLoopAlreadyUnrolled if loop attributes have been specified
       // explicitly.
       return UnrollResult;
     }
   }
 
   // If loop has an unroll count pragma or unrolled by explicitly set count
   // mark loop as unrolled to prevent unrolling beyond that requested.
-  // If the loop was peeled, we already "used up" the profile information
-  // we had, so we don't want to unroll or peel again.
-  if (UnrollResult != LoopUnrollResult::FullyUnrolled &&
-      (IsCountSetExplicitly || (PP.PeelProfiledIterations && PP.PeelCount)))
+  if (UnrollResult != LoopUnrollResult::FullyUnrolled && IsCountSetExplicitly)
     L->setLoopAlreadyUnrolled();
 
   return UnrollResult;
 }
 
 namespace {
 
 class LoopUnroll : public LoopPass {

llvm/lib/Transforms/Utils/LoopUnroll.cpp

 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Debug.h"
 #include "llvm/Support/GenericDomTree.h"
 #include "llvm/Support/MathExtras.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Transforms/Utils/BasicBlockUtils.h"
 #include "llvm/Transforms/Utils/Cloning.h"
 #include "llvm/Transforms/Utils/Local.h"
-#include "llvm/Transforms/Utils/LoopPeel.h"
 #include "llvm/Transforms/Utils/LoopSimplify.h"
 #include "llvm/Transforms/Utils/LoopUtils.h"
 #include "llvm/Transforms/Utils/SimplifyIndVar.h"
 #include "llvm/Transforms/Utils/UnrollLoop.h"
 #include "llvm/Transforms/Utils/ValueMapper.h"
 #include <algorithm>
 #include <assert.h>
 #include <type_traits>

 ///
 /// If AllowRuntime is true then UnrollLoop will consider unrolling loops that
 /// have a runtime (i.e. not compile time constant) trip count.  Unrolling these
 /// loops require a unroll "prologue" that runs "RuntimeTripCount % Count"
 /// iterations before branching into the unrolled loop.  UnrollLoop will not
 /// runtime-unroll the loop if computing RuntimeTripCount will be expensive and
 /// AllowExpensiveTripCount is false.
 ///
-/// If we want to perform PGO-based loop peeling, PeelCount is set to the
-/// number of iterations we want to peel off.
-///
 /// The LoopInfo Analysis that is passed will be kept consistent.
 ///
 /// This utility preserves LoopInfo. It will also preserve ScalarEvolution and
 /// DominatorTree if they are non-null.
 ///
 /// If RemainderLoop is non-null, it will receive the remainder loop (if
 /// required and not fully unrolled).
 LoopUnrollResult llvm::UnrollLoop(Loop *L, UnrollLoopOptions ULO, LoopInfo *LI,
if (ULO.TripMultiple != 1)
     LLVM_DEBUG(dbgs() << "  Trip Multiple = " << ULO.TripMultiple << "\n");
 
   // Effectively "DCE" unrolled iterations that are beyond the tripcount
   // and will never be executed.
   if (ULO.TripCount != 0 && ULO.Count > ULO.TripCount)
     ULO.Count = ULO.TripCount;
 
   // Don't enter the unroll code if there is nothing to do.
-  if (ULO.TripCount == 0 && ULO.Count < 2 && ULO.PeelCount == 0) {
+  if (ULO.TripCount == 0 && ULO.Count < 2) {
     LLVM_DEBUG(dbgs() << "Won't unroll; almost nothing to do\n");
     return LoopUnrollResult::Unmodified;
   }
 
   assert(ULO.Count > 0);
   assert(ULO.TripMultiple > 0);
   assert(ULO.TripCount == 0 || ULO.TripCount % ULO.TripMultiple == 0);
 
   // Are we eliminating the loop control altogether?
   bool CompletelyUnroll = ULO.Count == ULO.TripCount;
 
   // We assume a run-time trip count if the compiler cannot
   // figure out the loop trip count and the unroll-runtime
   // flag is specified.
   bool RuntimeTripCount =
       (ULO.TripCount == 0 && ULO.Count > 0 && ULO.AllowRuntime);
 
-  assert((!RuntimeTripCount || !ULO.PeelCount) &&
-         "Did not expect runtime trip-count unrolling "
-         "and peeling for the same loop");
-
-  bool Peeled = false;
-  if (ULO.PeelCount) {
-    Peeled = peelLoop(L, ULO.PeelCount, LI, SE, DT, AC, PreserveLCSSA);
-
-    // Successful peeling may result in a change in the loop preheader/trip
-    // counts. If we later unroll the loop, we want these to be updated.
-    if (Peeled) {
-      // According to our guards and profitability checks the only
-      // meaningful exit should be latch block. Other exits go to deopt,
-      // so we do not worry about them.
-      BasicBlock *ExitingBlock = L->getLoopLatch();
-      assert(ExitingBlock && "Loop without exiting block?");
-      assert(L->isLoopExiting(ExitingBlock) && "Latch is not exiting?");
-      ULO.TripCount = SE->getSmallConstantTripCount(L, ExitingBlock);
-      ULO.TripMultiple = SE->getSmallConstantTripMultiple(L, ExitingBlock);
-    }
-  }
-
   // All these values should be taken only after peeling because they might have
   // changed.
   BasicBlock *Preheader = L->getLoopPreheader();
   BasicBlock *Header = L->getHeader();
   BasicBlock *LatchBlock = L->getLoopLatch();
   SmallVector<BasicBlock *, 4> ExitBlocks;
   L->getExitBlocks(ExitBlocks);
   std::vector<BasicBlock *> OriginalLoopBlocks = L->getBlocks();
LoopUnrollResult llvm::UnrollLoop(Loop *L, UnrollLoopOptions ULO, LoopInfo *LI,
   // unconditional branch in the unrolled loop in some cases.
   BranchInst *ExitingBI = nullptr;
   bool LatchIsExiting = L->isLoopExiting(LatchBlock);
   if (LatchIsExiting)
     ExitingBI = LatchBI;
   else if (BasicBlock *ExitingBlock = L->getExitingBlock())
     ExitingBI = dyn_cast<BranchInst>(ExitingBlock->getTerminator());
   if (!LatchBI || (LatchBI->isConditional() && !LatchIsExiting)) {
-    // If the peeling guard is changed this assert may be relaxed or even
-    // deleted.
-    assert(!Peeled && "Peeling guard changed!");
     LLVM_DEBUG(
         dbgs() << "Can't unroll; a conditional latch must exit the loop");
     return LoopUnrollResult::Unmodified;
   }
   LLVM_DEBUG({
     if (ExitingBI)
       dbgs() << "  Exiting Block = " << ExitingBI->getParent()->getName()
              << "\n";
LLVM_DEBUG(dbgs() << "COMPLETELY UNROLLING loop %" << Header->getName()
                       << " with trip count " << ULO.TripCount << "!\n");
     if (ORE)
       ORE->emit([&]() {
         return OptimizationRemark(DEBUG_TYPE, "FullyUnrolled", L->getStartLoc(),
                                   L->getHeader())
                << "completely unrolled loop with "
                << NV("UnrollCount", ULO.TripCount) << " iterations";
       });
-  } else if (ULO.PeelCount) {
-    LLVM_DEBUG(dbgs() << "PEELING loop %" << Header->getName()
-                      << " with iteration count " << ULO.PeelCount << "!\n");
-    if (ORE)
-      ORE->emit([&]() {
-        return OptimizationRemark(DEBUG_TYPE, "Peeled", L->getStartLoc(),
-                                  L->getHeader())
-               << " peeled loop by " << NV("PeelCount", ULO.PeelCount)
-               << " iterations";
-      });
   } else {
     auto DiagBuilder = [&]() {
       OptimizationRemark Diag(DEBUG_TYPE, "PartialUnrolled", L->getStartLoc(),
                               L->getHeader());
       return Diag << "unrolled loop by a factor of "
                   << NV("UnrollCount", ULO.Count);
     };
 
if (Term && Term->isUnconditional()) {
       }
     }
   }
   // Apply updates to the DomTree.
   DT = &DTU.getDomTree();
 
   // At this point, the code is well formed.  We now simplify the unrolled loop,
   // doing constant propagation and dead code elimination as we go.
-  simplifyLoopAfterUnroll(L, !CompletelyUnroll && (ULO.Count > 1 || Peeled), LI,
-                          SE, DT, AC, TTI);
+  simplifyLoopAfterUnroll(L, !CompletelyUnroll && ULO.Count > 1, LI, SE, DT, AC,
+                          TTI);
 
   NumCompletelyUnrolled += CompletelyUnroll;
   ++NumUnrolled;
 
   Loop *OuterL = L->getParentLoop();
   // Update LoopInfo if the loop is completely removed.
   if (CompletelyUnroll)
     LI->erase(L);

llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp

#endif
   if (remainderLoop && UnrollRemainder) {
     LLVM_DEBUG(dbgs() << "Unrolling remainder loop\n");
     UnrollResult =
         UnrollLoop(remainderLoop,
                    {/*Count*/ Count - 1, /*TripCount*/ Count - 1,
                     /*Force*/ false, /*AllowRuntime*/ false,
                     /*AllowExpensiveTripCount*/ false, /*PreserveCondBr*/ true,
                     /*PreserveOnlyFirst*/ false, /*TripMultiple*/ 1,
-                    /*PeelCount*/ 0, /*UnrollRemainder*/ false, ForgetAllSCEV},
+                    /*UnrollRemainder*/ false, ForgetAllSCEV},
                    LI, SE, DT, AC, TTI, /*ORE*/ nullptr, PreserveLCSSA);
   }
 
   if (ResultLoop && UnrollResult != LoopUnrollResult::FullyUnrolled)
     *ResultLoop = remainderLoop;
   NumRuntimeUnrolled++;
   return true;
 }

llvm/test/Transforms/LoopUnroll/peel-loop-and-unroll.ll

+; RUN: not --crash opt -loop-unroll -unroll-peel-count=2 -unroll-count=2 -S < %s 2>&1 | FileCheck %s
+
+; CHECK: LLVM ERROR: Cannot specify both explicit peel count and explicit unroll count
+
+@a = global [8 x i32] zeroinitializer, align 16
+
+define void @test1() {
+entry:
+  br label %for.body
+
+for.body:                                         ; preds = %entry, %for.body
+  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
+  %arrayidx = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 %indvars.iv
+  %0 = trunc i64 %indvars.iv to i32
+  store i32 %0, i32* %arrayidx, align 4
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %exitcond = icmp ne i64 %indvars.iv.next, 8
+  br i1 %exitcond, label %for.body, label %for.exit
+
+for.exit:                        ; preds = %for.body
+  ret void
+}

llvm/test/Transforms/LoopUnroll/pr33437.ll

-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -S -loop-unroll -unroll-count=4 -unroll-peel-count=1 < %s | FileCheck %s
-
-declare zeroext i8 @patatino()
-
-define fastcc void @tinky() {
-; CHECK-LABEL: @tinky(
-; CHECK-NEXT:  entry:
-; CHECK-NEXT:    br label [[NEXT:%.*]]
-; CHECK:       loopexit:
-; CHECK-NEXT:    ret void
-; CHECK:       next:
-; CHECK-NEXT:    br label [[LOOP_PEEL_BEGIN:%.*]]
-; CHECK:       loop.peel.begin:
-; CHECK-NEXT:    br label [[LOOP_PEEL:%.*]]
-; CHECK:       loop.peel:
-; CHECK-NEXT:    [[CALL593_PEEL:%.*]] = tail call zeroext i8 @patatino()
-; CHECK-NEXT:    br i1 false, label [[LOOP_PEEL_NEXT:%.*]], label [[LOOPEXIT:%.*]]
-; CHECK:       loop.peel.next:
-; CHECK-NEXT:    br label [[LOOP_PEEL_NEXT1:%.*]]
-; CHECK:       loop.peel.next1:
-; CHECK-NEXT:    br label [[NEXT_PEEL_NEWPH:%.*]]
-; CHECK:       next.peel.newph:
-; CHECK-NEXT:    br label [[LOOP:%.*]]
-; CHECK:       loop:
-; CHECK-NEXT:    [[CALL593:%.*]] = tail call zeroext i8 @patatino()
-; CHECK-NEXT:    br label [[LOOPEXIT]]
-;
-entry:
-  br label %next
-
-loopexit:
-  ret void
-
-next:
-  br label %loop
-
-loop:
-  %a = phi i8 [ undef, %next ], [ %call593, %loop ]
-  %b = phi i32 [ 0, %next ], [ 1, %loop ]
-  %call593 = tail call zeroext i8 @patatino()
-  br i1 false, label %loop, label %loopexit
-}
-
-define void @tinky2() {
-; CHECK-LABEL: @tinky2(
-; CHECK-NEXT:  entry:
-; CHECK-NEXT:    br label [[NEXT:%.*]]
-; CHECK:       loopexit.loopexit:
-; CHECK-NEXT:    br label [[LOOPEXIT:%.*]]
-; CHECK:       loopexit:
-; CHECK-NEXT:    ret void
-; CHECK:       next:
-; CHECK-NEXT:    br label [[LOOP_PEEL_BEGIN:%.*]]
-; CHECK:       loop.peel.begin:
-; CHECK-NEXT:    br label [[LOOP_PEEL:%.*]]
-; CHECK:       loop.peel:
-; CHECK-NEXT:    [[CALL593_PEEL:%.*]] = tail call zeroext i8 @patatino()
-; CHECK-NEXT:    [[B_NEXT_PEEL:%.*]] = add i32 0, 1
-; CHECK-NEXT:    [[COND_PEEL:%.*]] = icmp ne i32 0, 30
-; CHECK-NEXT:    br i1 [[COND_PEEL]], label [[LOOP_PEEL_NEXT:%.*]], label [[LOOPEXIT]]
-; CHECK:       loop.peel.next:
-; CHECK-NEXT:    br label [[LOOP_PEEL_NEXT1:%.*]]
-; CHECK:       loop.peel.next1:
-; CHECK-NEXT:    br label [[NEXT_PEEL_NEWPH:%.*]]
-; CHECK:       next.peel.newph:
-; CHECK-NEXT:    br label [[LOOP:%.*]]
-; CHECK:       loop:
-; CHECK-NEXT:    [[B:%.*]] = phi i32 [ [[B_NEXT_PEEL]], [[NEXT_PEEL_NEWPH]] ], [ [[B_NEXT_3:%.*]], [[LOOP_2:%.*]] ]
-; CHECK-NEXT:    [[CALL593:%.*]] = tail call zeroext i8 @patatino()
-; CHECK-NEXT:    [[B_NEXT:%.*]] = add nuw nsw i32 [[B]], 1
-; CHECK-NEXT:    [[CALL593_1:%.*]] = tail call zeroext i8 @patatino()
-; CHECK-NEXT:    [[B_NEXT_1:%.*]] = add nuw nsw i32 [[B_NEXT]], 1
-; CHECK-NEXT:    [[COND_1:%.*]] = icmp ne i32 [[B_NEXT]], 30
-; CHECK-NEXT:    br i1 [[COND_1]], label [[LOOP_2]], label [[LOOPEXIT_LOOPEXIT:%.*]], !llvm.loop !0
-; CHECK:       loop.2:
-; CHECK-NEXT:    [[CALL593_2:%.*]] = tail call zeroext i8 @patatino()
-; CHECK-NEXT:    [[B_NEXT_2:%.*]] = add nuw nsw i32 [[B_NEXT_1]], 1
-; CHECK-NEXT:    [[CALL593_3:%.*]] = tail call zeroext i8 @patatino()
-; CHECK-NEXT:    [[B_NEXT_3]] = add nuw nsw i32 [[B_NEXT_2]], 1
-; CHECK-NEXT:    br label [[LOOP]], !llvm.loop !2
-;
-entry:
-  br label %next
-
-loopexit:
-  ret void
-
-next:
-  br label %loop
-
-loop:
-  %a = phi i8 [ undef, %next ], [ %call593, %loop ]
-  %b = phi i32 [ 0, %next ], [ %b.next, %loop ]
-  %call593 = tail call zeroext i8 @patatino()
-  %b.next = add i32 %b, 1
-  %cond = icmp ne i32 %b, 30
-  br i1 %cond, label %loop, label %loopexit
-}

llvm/test/Transforms/LoopUnroll/pr45939-peel-count-and-complete-unroll.ll

-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -loop-unroll -unroll-peel-count=2 -S %s | FileCheck --check-prefix=PEEL2 %s
-; RUN: opt -loop-unroll -unroll-peel-count=8 -S %s | FileCheck --check-prefix=PEEL8 %s
-; RUN: opt -loop-unroll -unroll-peel-count=2 -unroll-count=2 -S %s | FileCheck --check-prefix=PEEL2UNROLL2 %s
-
-; Test case for PR45939. Make sure unroll count is adjusted when loop is peeled and unrolled.
-
-@a = global [8 x i32] zeroinitializer, align 16
-
-define void @test1() {
-; PEEL2-LABEL: @test1(
-; PEEL2-NEXT:  entry:
-; PEEL2-NEXT:    br label [[FOR_BODY_PEEL_BEGIN:%.*]]
-; PEEL2:       for.body.peel.begin:
-; PEEL2-NEXT:    br label [[FOR_BODY_PEEL:%.*]]
-; PEEL2:       for.body.peel:
-; PEEL2-NEXT:    [[ARRAYIDX_PEEL:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 0
-; PEEL2-NEXT:    [[TMP0:%.*]] = trunc i64 0 to i32
-; PEEL2-NEXT:    store i32 [[TMP0]], i32* [[ARRAYIDX_PEEL]], align 4
-; PEEL2-NEXT:    [[INDVARS_IV_NEXT_PEEL:%.*]] = add nuw nsw i64 0, 1
-; PEEL2-NEXT:    [[EXITCOND_PEEL:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL]], 8
-; PEEL2-NEXT:    br i1 [[EXITCOND_PEEL]], label [[FOR_BODY_PEEL_NEXT:%.*]], label [[FOR_EXIT:%.*]]
-; PEEL2:       for.body.peel.next:
-; PEEL2-NEXT:    br label [[FOR_BODY_PEEL2:%.*]]
-; PEEL2:       for.body.peel2:
-; PEEL2-NEXT:    [[ARRAYIDX_PEEL3:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL]]
-; PEEL2-NEXT:    [[TMP1:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL]] to i32
-; PEEL2-NEXT:    store i32 [[TMP1]], i32* [[ARRAYIDX_PEEL3]], align 4
-; PEEL2-NEXT:    [[INDVARS_IV_NEXT_PEEL4:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL]], 1
-; PEEL2-NEXT:    [[EXITCOND_PEEL5:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL4]], 8
-; PEEL2-NEXT:    br i1 [[EXITCOND_PEEL5]], label [[FOR_BODY_PEEL_NEXT1:%.*]], label [[FOR_EXIT]]
-; PEEL2:       for.body.peel.next1:
-; PEEL2-NEXT:    br label [[FOR_BODY_PEEL_NEXT6:%.*]]
-; PEEL2:       for.body.peel.next6:
-; PEEL2-NEXT:    br label [[ENTRY_PEEL_NEWPH:%.*]]
-; PEEL2:       entry.peel.newph:
-; PEEL2-NEXT:    br label [[FOR_BODY:%.*]]
-; PEEL2:       for.body:
-; PEEL2-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL4]]
-; PEEL2-NEXT:    [[TMP2:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL4]] to i32
-; PEEL2-NEXT:    store i32 [[TMP2]], i32* [[ARRAYIDX]], align 4
-; PEEL2-NEXT:    store i32 3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @a, i64 0, i64 3), align 4
-; PEEL2-NEXT:    store i32 4, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @a, i64 0, i64 4), align 4
-; PEEL2-NEXT:    store i32 5, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @a, i64 0, i64 5), align 4
-; PEEL2-NEXT:    store i32 6, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @a, i64 0, i64 6), align 4
-; PEEL2-NEXT:    store i32 7, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @a, i64 0, i64 7), align 4
-; PEEL2-NEXT:    store i32 8, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @a, i64 1, i64 0), align 4
-; PEEL2-NEXT:    store i32 9, i32* getelementptr ([8 x i32], [8 x i32]* @a, i64 1, i64 1), align 4
-; PEEL2-NEXT:    br label [[FOR_EXIT]]
-; PEEL2:       for.exit:
-; PEEL2-NEXT:    ret void
-;
-; PEEL8-LABEL: @test1(
-; PEEL8-NEXT:  entry:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL_BEGIN:%.*]]
-; PEEL8:       for.body.peel.begin:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL:%.*]]
-; PEEL8:       for.body.peel:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 0
-; PEEL8-NEXT:    [[TMP0:%.*]] = trunc i64 0 to i32
-; PEEL8-NEXT:    store i32 [[TMP0]], i32* [[ARRAYIDX_PEEL]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL:%.*]] = add nuw nsw i64 0, 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL]], label [[FOR_BODY_PEEL_NEXT:%.*]], label [[FOR_EXIT:%.*]]
-; PEEL8:       for.body.peel.next:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL2:%.*]]
-; PEEL8:       for.body.peel2:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL3:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL]]
-; PEEL8-NEXT:    [[TMP1:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL]] to i32
-; PEEL8-NEXT:    store i32 [[TMP1]], i32* [[ARRAYIDX_PEEL3]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL4:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL]], 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL5:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL4]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL5]], label [[FOR_BODY_PEEL_NEXT1:%.*]], label [[FOR_EXIT]]
-; PEEL8:       for.body.peel.next1:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL7:%.*]]
-; PEEL8:       for.body.peel7:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL8:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL4]]
-; PEEL8-NEXT:    [[TMP2:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL4]] to i32
-; PEEL8-NEXT:    store i32 [[TMP2]], i32* [[ARRAYIDX_PEEL8]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL9:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL4]], 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL10:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL9]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL10]], label [[FOR_BODY_PEEL_NEXT6:%.*]], label [[FOR_EXIT]]
-; PEEL8:       for.body.peel.next6:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL12:%.*]]
-; PEEL8:       for.body.peel12:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL13:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL9]]
-; PEEL8-NEXT:    [[TMP3:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL9]] to i32
-; PEEL8-NEXT:    store i32 [[TMP3]], i32* [[ARRAYIDX_PEEL13]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL14:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL9]], 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL15:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL14]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL15]], label [[FOR_BODY_PEEL_NEXT11:%.*]], label [[FOR_EXIT]]
-; PEEL8:       for.body.peel.next11:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL17:%.*]]
-; PEEL8:       for.body.peel17:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL18:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL14]]
-; PEEL8-NEXT:    [[TMP4:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL14]] to i32
-; PEEL8-NEXT:    store i32 [[TMP4]], i32* [[ARRAYIDX_PEEL18]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL19:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL14]], 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL20:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL19]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL20]], label [[FOR_BODY_PEEL_NEXT16:%.*]], label [[FOR_EXIT]]
-; PEEL8:       for.body.peel.next16:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL22:%.*]]
-; PEEL8:       for.body.peel22:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL23:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL19]]
-; PEEL8-NEXT:    [[TMP5:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL19]] to i32
-; PEEL8-NEXT:    store i32 [[TMP5]], i32* [[ARRAYIDX_PEEL23]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL24:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL19]], 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL25:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL24]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL25]], label [[FOR_BODY_PEEL_NEXT21:%.*]], label [[FOR_EXIT]]
-; PEEL8:       for.body.peel.next21:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL27:%.*]]
-; PEEL8:       for.body.peel27:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL28:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL24]]
-; PEEL8-NEXT:    [[TMP6:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL24]] to i32
-; PEEL8-NEXT:    store i32 [[TMP6]], i32* [[ARRAYIDX_PEEL28]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL29:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL24]], 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL30:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL29]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL30]], label [[FOR_BODY_PEEL_NEXT26:%.*]], label [[FOR_EXIT]]
-; PEEL8:       for.body.peel.next26:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL32:%.*]]
-; PEEL8:       for.body.peel32:
-; PEEL8-NEXT:    [[ARRAYIDX_PEEL33:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL29]]
-; PEEL8-NEXT:    [[TMP7:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL29]] to i32
-; PEEL8-NEXT:    store i32 [[TMP7]], i32* [[ARRAYIDX_PEEL33]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_PEEL34:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL29]], 1
-; PEEL8-NEXT:    [[EXITCOND_PEEL35:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL34]], 8
-; PEEL8-NEXT:    br i1 [[EXITCOND_PEEL35]], label [[FOR_BODY_PEEL_NEXT31:%.*]], label [[FOR_EXIT]]
-; PEEL8:       for.body.peel.next31:
-; PEEL8-NEXT:    br label [[FOR_BODY_PEEL_NEXT36:%.*]]
-; PEEL8:       for.body.peel.next36:
-; PEEL8-NEXT:    br label [[ENTRY_PEEL_NEWPH:%.*]]
-; PEEL8:       entry.peel.newph:
-; PEEL8-NEXT:    br label [[FOR_BODY:%.*]]
-; PEEL8:       for.body:
-; PEEL8-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL34]]
-; PEEL8-NEXT:    [[TMP8:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL34]] to i32
-; PEEL8-NEXT:    store i32 [[TMP8]], i32* [[ARRAYIDX]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL34]], 1
-; PEEL8-NEXT:    [[ARRAYIDX_1:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT]]
-; PEEL8-NEXT:    [[TMP9:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
-; PEEL8-NEXT:    store i32 [[TMP9]], i32* [[ARRAYIDX_1]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_1:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT]], 1
-; PEEL8-NEXT:    [[ARRAYIDX_2:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_1]]
-; PEEL8-NEXT:    [[TMP10:%.*]] = trunc i64 [[INDVARS_IV_NEXT_1]] to i32
-; PEEL8-NEXT:    store i32 [[TMP10]], i32* [[ARRAYIDX_2]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_2:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_1]], 1
-; PEEL8-NEXT:    [[ARRAYIDX_3:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_2]]
-; PEEL8-NEXT:    [[TMP11:%.*]] = trunc i64 [[INDVARS_IV_NEXT_2]] to i32
-; PEEL8-NEXT:    store i32 [[TMP11]], i32* [[ARRAYIDX_3]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_3:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_2]], 1
-; PEEL8-NEXT:    [[ARRAYIDX_4:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_3]]
-; PEEL8-NEXT:    [[TMP12:%.*]] = trunc i64 [[INDVARS_IV_NEXT_3]] to i32
-; PEEL8-NEXT:    store i32 [[TMP12]], i32* [[ARRAYIDX_4]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_4:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_3]], 1
-; PEEL8-NEXT:    [[ARRAYIDX_5:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_4]]
-; PEEL8-NEXT:    [[TMP13:%.*]] = trunc i64 [[INDVARS_IV_NEXT_4]] to i32
-; PEEL8-NEXT:    store i32 [[TMP13]], i32* [[ARRAYIDX_5]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_5:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_4]], 1
-; PEEL8-NEXT:    [[ARRAYIDX_6:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_5]]
-; PEEL8-NEXT:    [[TMP14:%.*]] = trunc i64 [[INDVARS_IV_NEXT_5]] to i32
-; PEEL8-NEXT:    store i32 [[TMP14]], i32* [[ARRAYIDX_6]], align 4
-; PEEL8-NEXT:    [[INDVARS_IV_NEXT_6:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_5]], 1
-; PEEL8-NEXT:    [[ARRAYIDX_7:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_6]]
-; PEEL8-NEXT:    [[TMP15:%.*]] = trunc i64 [[INDVARS_IV_NEXT_6]] to i32
-; PEEL8-NEXT:    store i32 [[TMP15]], i32* [[ARRAYIDX_7]], align 4
-; PEEL8-NEXT:    br label [[FOR_EXIT]]
-; PEEL8:       for.exit:
-; PEEL8-NEXT:    ret void
-;
-; PEEL2UNROLL2-LABEL: @test1(
-; PEEL2UNROLL2-NEXT:  entry:
-; PEEL2UNROLL2-NEXT:    br label [[FOR_BODY_PEEL_BEGIN:%.*]]
-; PEEL2UNROLL2:       for.body.peel.begin:
-; PEEL2UNROLL2-NEXT:    br label [[FOR_BODY_PEEL:%.*]]
-; PEEL2UNROLL2:       for.body.peel:
-; PEEL2UNROLL2-NEXT:    [[ARRAYIDX_PEEL:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 0
-; PEEL2UNROLL2-NEXT:    [[TMP0:%.*]] = trunc i64 0 to i32
-; PEEL2UNROLL2-NEXT:    store i32 [[TMP0]], i32* [[ARRAYIDX_PEEL]], align 4
-; PEEL2UNROLL2-NEXT:    [[INDVARS_IV_NEXT_PEEL:%.*]] = add nuw nsw i64 0, 1
-; PEEL2UNROLL2-NEXT:    [[EXITCOND_PEEL:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL]], 8
-; PEEL2UNROLL2-NEXT:    br i1 [[EXITCOND_PEEL]], label [[FOR_BODY_PEEL_NEXT:%.*]], label [[FOR_EXIT:%.*]]
-; PEEL2UNROLL2:       for.body.peel.next:
-; PEEL2UNROLL2-NEXT:    br label [[FOR_BODY_PEEL2:%.*]]
-; PEEL2UNROLL2:       for.body.peel2:
-; PEEL2UNROLL2-NEXT:    [[ARRAYIDX_PEEL3:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT_PEEL]]
-; PEEL2UNROLL2-NEXT:    [[TMP1:%.*]] = trunc i64 [[INDVARS_IV_NEXT_PEEL]] to i32
-; PEEL2UNROLL2-NEXT:    store i32 [[TMP1]], i32* [[ARRAYIDX_PEEL3]], align 4
-; PEEL2UNROLL2-NEXT:    [[INDVARS_IV_NEXT_PEEL4:%.*]] = add nuw nsw i64 [[INDVARS_IV_NEXT_PEEL]], 1
-; PEEL2UNROLL2-NEXT:    [[EXITCOND_PEEL5:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_PEEL4]], 8
-; PEEL2UNROLL2-NEXT:    br i1 [[EXITCOND_PEEL5]], label [[FOR_BODY_PEEL_NEXT1:%.*]], label [[FOR_EXIT]]
-; PEEL2UNROLL2:       for.body.peel.next1:
-; PEEL2UNROLL2-NEXT:    br label [[FOR_BODY_PEEL_NEXT6:%.*]]
-; PEEL2UNROLL2:       for.body.peel.next6:
-; PEEL2UNROLL2-NEXT:    br label [[ENTRY_PEEL_NEWPH:%.*]]
-; PEEL2UNROLL2:       entry.peel.newph:
-; PEEL2UNROLL2-NEXT:    br label [[FOR_BODY:%.*]]
-; PEEL2UNROLL2:       for.body:
-; PEEL2UNROLL2-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT_PEEL4]], [[ENTRY_PEEL_NEWPH]] ], [ [[INDVARS_IV_NEXT_1:%.*]], [[FOR_BODY]] ]
-; PEEL2UNROLL2-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV]]
-; PEEL2UNROLL2-NEXT:    [[TMP2:%.*]] = trunc i64 [[INDVARS_IV]] to i32
-; PEEL2UNROLL2-NEXT:    store i32 [[TMP2]], i32* [[ARRAYIDX]], align 4
-; PEEL2UNROLL2-NEXT:    [[INDVARS_IV_NEXT:%.*]] = add nuw nsw i64 [[INDVARS_IV]], 1
-; PEEL2UNROLL2-NEXT:    [[ARRAYIDX_1:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 [[INDVARS_IV_NEXT]]
-; PEEL2UNROLL2-NEXT:    [[TMP3:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
-; PEEL2UNROLL2-NEXT:    store i32 [[TMP3]], i32* [[ARRAYIDX_1]], align 4
-; PEEL2UNROLL2-NEXT:    [[INDVARS_IV_NEXT_1]] = add nuw nsw i64 [[INDVARS_IV_NEXT]], 1
-; PEEL2UNROLL2-NEXT:    [[EXITCOND_1:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT_1]], 8
-; PEEL2UNROLL2-NEXT:    br i1 [[EXITCOND_1]], label [[FOR_BODY]], label [[FOR_EXIT_LOOPEXIT:%.*]], !llvm.loop [[LOOP0:![0-9]+]]
-; PEEL2UNROLL2:       for.exit.loopexit:
-; PEEL2UNROLL2-NEXT:    br label [[FOR_EXIT]]
-; PEEL2UNROLL2:       for.exit:
-; PEEL2UNROLL2-NEXT:    ret void
-;
-entry:
-  br label %for.body
-
-for.body:                                         ; preds = %entry, %for.body
-  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
-  %arrayidx = getelementptr inbounds [8 x i32], [8 x i32]* @a, i64 0, i64 %indvars.iv
-  %0 = trunc i64 %indvars.iv to i32
-  store i32 %0, i32* %arrayidx, align 4
-  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
-  %exitcond = icmp ne i64 %indvars.iv.next, 8
-  br i1 %exitcond, label %for.body, label %for.exit
-
-for.exit:                        ; preds = %for.body
-  ret void
-}

llvm/test/Transforms/LoopUnroll/wrong_assert_in_peeling.ll

 ; CHECK-NEXT:    br label [[BB1_PEEL_NEWPH:%.*]]
 ; CHECK:       bb1.peel.newph:
 ; CHECK-NEXT:    br label [[BB2:%.*]]
 ; CHECK:       bb2:
 ; CHECK-NEXT:    [[TMP3:%.*]] = phi i32 [ [[TMP4_PEEL]], [[BB1_PEEL_NEWPH]] ], [ [[TMP4:%.*]], [[BB12:%.*]] ]
 ; CHECK-NEXT:    [[TMP4]] = add nsw i32 [[TMP3]], [[TMP]]
 ; CHECK-NEXT:    br label [[BB5:%.*]]
 ; CHECK:       bb5:
-; CHECK-NEXT:    br i1 false, label [[BB7:%.*]], label [[BB15_LOOPEXIT:%.*]]
+; CHECK-NEXT:    br i1 undef, label [[BB7:%.*]], label [[BB15_LOOPEXIT:%.*]]
 ; CHECK:       bb7:
 ; CHECK-NEXT:    br i1 undef, label [[BB10:%.*]], label [[BB10]]
 ; CHECK:       bb10:
-; CHECK-NEXT:    br i1 false, label [[BB12]], label [[BB17_LOOPEXIT:%.*]]
+; CHECK-NEXT:    br i1 undef, label [[BB12]], label [[BB17_LOOPEXIT:%.*]]
 ; CHECK:       bb12:
-; CHECK-NEXT:    br i1 false, label [[BB13_LOOPEXIT:%.*]], label [[BB2]], !llvm.loop !0
+; CHECK-NEXT:    br i1 false, label [[BB13_LOOPEXIT:%.*]], label [[BB2]], !llvm.loop [[LOOP0:![0-9]+]]
 ; CHECK:       bb13.loopexit:
 ; CHECK-NEXT:    br label [[BB13]]
 ; CHECK:       bb13:
 ; CHECK-NEXT:    [[TMP14]] = add nsw i32 [[TMP]], -1
 ; CHECK-NEXT:    br label [[BB1]]
 ; CHECK:       bb15.loopexit:
 ; CHECK-NEXT:    br label [[BB15:%.*]]
 ; CHECK:       bb15.loopexit2: