[LoopUnroll] Make LoopPeeling respect the AllowPeeling preference.

[mcrosier]

Summary

AFAICT, the SimpleLoopUnrollPass isn't suppose to perform loop peeling.

Chad

[efriedma]

LGTM

It's possible that peeling which "simplifies" a loop should be done earlier, but completely ignoring AllowPeeling seems wrong.

[fhahn]

Thanks Chad, AllowPeeling and UserPeelCount taking precedence over other peel count computations looks good to me

[mkazantsev]

LGTM

[mcrosier]

Thanks everyone for the quick review. I'll commit shortly.

lib/Transforms/Utils/LoopUnrollPeel.cpp

void llvm::computePeelCount(Loop *L, unsigned LoopSize,
   UP.PeelCount = 0;
   if (!canPeel(L))
     return;
 
   // Only try to peel innermost loops.
   if (!L->empty())
     return;
 
+  // If the user provided a peel count, use that.
+  bool UserPeelCount = UnrollForcePeelCount.getNumOccurrences() > 0;
+  if (UserPeelCount) {
+    DEBUG(dbgs() << "Force-peeling first " << UnrollForcePeelCount
+                 << " iterations.\n");
+    UP.PeelCount = UnrollForcePeelCount;
+    return;
+  }
+
+  // Skip peeling if it's disabled.
+  if (!UP.AllowPeeling)
+    return;
+
   // Here we try to get rid of Phis which become invariants after 1, 2, ..., N
   // iterations of the loop. For this we compute the number for iterations after
   // which every Phi is guaranteed to become an invariant, and try to peel the
   // maximum number of iterations among these values, thus turning all those
   // Phis into invariants.
   // First, check that we can peel at least one iteration.
   if (2 * LoopSize <= UP.Threshold && UnrollPeelMaxCount > 0) {
     // Store the pre-calculated values here.
if (2 * LoopSize <= UP.Threshold && UnrollPeelMaxCount > 0) {
     }
   }
 
   // Bail if we know the statically calculated trip count.
   // In this case we rather prefer partial unrolling.
   if (TripCount)
     return;
 
-  // If the user provided a peel count, use that.
-  bool UserPeelCount = UnrollForcePeelCount.getNumOccurrences() > 0;
-  if (UserPeelCount) {
-    DEBUG(dbgs() << "Force-peeling first " << UnrollForcePeelCount
-                 << " iterations.\n");
-    UP.PeelCount = UnrollForcePeelCount;
-    return;
-  }
-
   // If we don't know the trip count, but have reason to believe the average
   // trip count is low, peeling should be beneficial, since we will usually
   // hit the peeled section.
   // We only do this in the presence of profile information, since otherwise
   // our estimates of the trip count are not reliable enough.
-  if (UP.AllowPeeling && L->getHeader()->getParent()->hasProfileData()) {
+  if (L->getHeader()->getParent()->hasProfileData()) {
     Optional<unsigned> PeelCount = getLoopEstimatedTripCount(L);
     if (!PeelCount)
       return;
 
     DEBUG(dbgs() << "Profile-based estimated trip count is " << *PeelCount
                  << "\n");
 
     if (*PeelCount) {

test/Transforms/LoopUnroll/peel-loop-not-forced.ll

 ; RUN: opt < %s -S -loop-unroll -unroll-threshold=30 | FileCheck %s
+; RUN: opt < %s -S -loop-unroll -unroll-threshold=30 -unroll-allow-peeling=false | FileCheck %s --check-prefix=DISABLE
 
 define i32 @invariant_backedge_1(i32 %a, i32 %b) {
 ; CHECK-LABEL: @invariant_backedge_1
 ; CHECK-NOT:     %plus = phi
 ; CHECK:       loop.peel:
 ; CHECK:       loop:
 ; CHECK:         %i = phi
 ; CHECK:         %sum = phi
+; DISABLE-LABEL: @invariant_backedge_1
+; DISABLE-NOT: loop.peel:
 entry:
   br label %loop
 
 loop:
   %i = phi i32 [ 0, %entry ], [ %inc, %loop ]
   %sum = phi i32 [ 0, %entry ], [ %incsum, %loop ]
   %plus = phi i32 [ %a, %entry ], [ %b, %loop ]