[ARM][MVE] Tail-predication: remove the BTC + 1 overflow checks

[SjoerdMeijer]

Summary

Adapt tail-predication to the new semantics of get.active.lane.mask as proposed in D86147. This means that:

• we can remove BTC + 1 overflow checks because now the loop tripcount is passed in to the intrinsic,
• we can immediately use that value to setup a counter for the number of elements processed by the loop and don't need to materialize BTC + 1.

[SjoerdMeijer]

Now with all tests fixed.

[samparker]

This looks much more simple :)

llvm/lib/Target/ARM/MVETailPredication.cpp

377

nit: maybe something more descriptive that highlights that this is the original scalar trip count or the number of elements.

394

same kind of nit, I guess having scalar and vector trip count names would be most clear?

399

Could you add a comment on why we're using SignedRange here? And below for the ranges and ceiling.

[SjoerdMeijer]

llvm/lib/Target/ARM/MVETailPredication.cpp

394

Yeah, it was a bit of an overloaded tripcount mess. I will update this and use these 2 terms:

• element count: the original scalar tripcount,
• tripcount = the (vector) tripcount.

[SjoerdMeijer]

llvm/lib/Target/ARM/MVETailPredication.cpp

399

Hm, good catch I think. The ranges are signed because they can just be negative, but extracting the ZExt value here for that probably doesn't make any sense. I will put a fixme in for that now, because after addressing the TODO above, the:

// 1) TODO: Check that the TripCount (TC) belongs ...

which is addressed in D86074, revising this part here is highest on my list.

[SjoerdMeijer]

Addressed comments

[samparker]

LGTM

llvm/lib/Target/ARM/MVETailPredication.cpp

 /// couple of final forms:
 /// - A tail-predicated loop, with implicit predication.
 /// - A loop containing multiple VCPT instructions, predicating multiple VPT
 ///   blocks of instructions operating on different vector types.
 ///
 /// This pass:
 /// 1) Checks if the predicates of the masked load/store instructions are
 ///    generated by intrinsic @llvm.get.active.lanes(). This intrinsic consumes
-///    the Backedge Taken Count (BTC) of the scalar loop as its second argument,
-///    which we extract to set up the number of elements processed by the loop.
+///    the the scalar loop tripcount as its second argument, which we extract
+///    to set up the number of elements processed by the loop.
 /// 2) Intrinsic @llvm.get.active.lanes() is then replaced by the MVE target
 ///    specific VCTP intrinsic to represent the effect of tail predication.
 ///    This will be picked up by the ARM Low-overhead loop pass, which performs
 ///    the final transformation to a DLSTP or WLSTP tail-predicated loop.
 
 #include "ARM.h"
 #include "ARMSubtarget.h"
 #include "ARMTargetTransformInfo.h"
static void Cleanup(SetVector<Instruction*> &MaybeDead, Loop *L) {
   }
 
   for (auto I : L->blocks())
     DeleteDeadPHIs(I);
 }
 
 // The active lane intrinsic has this form:
 //
-//    @llvm.get.active.lane.mask(IV, BTC)
+//    @llvm.get.active.lane.mask(IV, TC)
 //
 // Here we perform checks that this intrinsic behaves as expected,
 // which means:
 //
-// 1) The element count, which is calculated with BTC + 1, cannot overflow.
-// 2) The element count needs to be sufficiently large that the decrement of
-//    element counter doesn't overflow, which means that we need to prove:
+// 1) Check that the TripCount (TC) belongs to this loop (originally).
+// 2) The element count (TC) needs to be sufficiently large that the decrement
+//    of element counter doesn't overflow, which means that we need to prove:
 //        ceil(ElementCount / VectorWidth) >= TripCount
 //    by rounding up ElementCount up:
 //        ((ElementCount + (VectorWidth - 1)) / VectorWidth
 //    and evaluate if expression isKnownNonNegative:
 //        (((ElementCount + (VectorWidth - 1)) / VectorWidth) - TripCount
 // 3) The IV must be an induction phi with an increment equal to the
 //    vector width.
 bool MVETailPredication::IsSafeActiveMask(IntrinsicInst *ActiveLaneMask,
     Value *TripCount, FixedVectorType *VecTy) {
   bool ForceTailPredication =
     EnableTailPredication == TailPredication::ForceEnabledNoReductions ||
     EnableTailPredication == TailPredication::ForceEnabled;
 
-  // 1) Test whether entry to the loop is protected by a conditional
-  // BTC + 1 < 0. In other words, if the scalar trip count overflows,
-  // becomes negative, we shouldn't enter the loop and creating
-  // tripcount expression BTC + 1 is not safe. So, check that BTC
-  // isn't max. This is evaluated in unsigned, because the semantics
-  // of @get.active.lane.mask is a ULE comparison.
-  auto *BackedgeTakenCount = ActiveLaneMask->getOperand(1);
-  auto *BTC = SE->getSCEV(BackedgeTakenCount);
-  auto *MaxBTC = SE->getConstantMaxBackedgeTakenCount(L);
-
-  if (isa<SCEVCouldNotCompute>(MaxBTC)) {
-    LLVM_DEBUG(dbgs() << "ARM TP: Can't compute SCEV BTC expression: ";
-               BTC->dump());
-    return false;
-  }
-
-  APInt MaxInt = APInt(BTC->getType()->getScalarSizeInBits(), ~0);
-  if (cast<SCEVConstant>(MaxBTC)->getAPInt().eq(MaxInt) &&
-      !ForceTailPredication) {
-    LLVM_DEBUG(dbgs() << "ARM TP: Overflow possible, BTC can be int max: ";
-               BTC->dump());
-    return false;
-  }
+  // 1) TODO: Check that the TripCount (TC) belongs to this loop (originally).
+  auto *TCUse = ActiveLaneMask->getOperand(1);
 
   // 2) Prove that the sub expression is non-negative, i.e. it doesn't overflow:
   //
   //      (((ElementCount + (VectorWidth - 1)) / VectorWidth) - TripCount
   //
   // 2.1) First prove overflow can't happen in:
   //
   //      ElementCount + (VectorWidth - 1)
   //
   // Because of a lack of context, it is difficult to get a useful bounds on
   // this expression. But since ElementCount uses the same variables as the
   // TripCount (TC), for which we can find meaningful value ranges, we use that
   // instead and assert that:
   //
   //     upperbound(TC) <= UINT_MAX - VectorWidth
   //
-  auto *TC = SE->getSCEV(TripCount);
+  auto *TCDef = SE->getSCEV(TripCount);
   unsigned SizeInBits = TripCount->getType()->getScalarSizeInBits();
   int VectorWidth = VecTy->getNumElements();
   auto Diff =  APInt(SizeInBits, ~0) - APInt(SizeInBits, VectorWidth);
   uint64_t MaxMinusVW = Diff.getZExtValue();
-  uint64_t UpperboundTC = SE->getSignedRange(TC).getUpper().getZExtValue();
+  uint64_t UpperboundTC = SE->getSignedRange(TCDef).getUpper().getZExtValue();
 
   if (UpperboundTC > MaxMinusVW && !ForceTailPredication) {
     LLVM_DEBUG(dbgs() << "ARM TP: Overflow possible in tripcount rounding:\n";
                dbgs() << "upperbound(TC) <= UINT_MAX - VectorWidth\n";
                dbgs() << UpperboundTC << " <= " << MaxMinusVW << " == false\n";);
     return false;
   }
 
   // 2.2) Make sure overflow doesn't happen in final expression:
   //  (((ElementCount + (VectorWidth - 1)) / VectorWidth) - TripCount,
   // To do this, compare the full ranges of these subexpressions:
   //
   //     Range(Ceil) <= Range(TC)
   //
   // where Ceil = ElementCount + (VW-1) / VW. If Ceil and TC are runtime
   // values (and not constants), we have to compensate for the lowerbound value
-  // range to be off by 1. The reason is that BTC lives in the preheader in
+  // range to be off by 1. The reason is that the TC lives in the preheader in
   // this form:
   //
   //     %trip.count.minus = add nsw nuw i32 %N, -1
   //
   // For the loop to be executed, %N has to be >= 1 and as a result the value
   // range of %trip.count.minus has a lower bound of 0. Value %TC has this form:
   //
   //     %5 = add nuw nsw i32 %4, 1
   //     call void @llvm.set.loop.iterations.i32(i32 %5)
   //
   // where %5 is some expression using %N, which needs to have a lower bound of
   // 1. Thus, if the ranges of Ceil and TC are not a single constant but a set,
   // we first add 0 to TC such that we can do the <= comparison on both sets.
   //
-  auto *One = SE->getOne(TripCount->getType());
-  // ElementCount = BTC + 1
-  auto *ElementCount = SE->getAddExpr(BTC, One);
+  auto *ElementCount = SE->getSCEV(TCUse);
   // Tmp = ElementCount + (VW-1)
   auto *ECPlusVWMinus1 = SE->getAddExpr(ElementCount,
       SE->getSCEV(ConstantInt::get(TripCount->getType(), VectorWidth - 1)));
   // Ceil = ElementCount + (VW-1) / VW
   auto *Ceil = SE->getUDivExpr(ECPlusVWMinus1,
       SE->getSCEV(ConstantInt::get(TripCount->getType(), VectorWidth)));
 
   ConstantRange RangeCeil = SE->getSignedRange(Ceil) ;
-  ConstantRange RangeTC = SE->getSignedRange(TC) ;
+  ConstantRange RangeTC = SE->getSignedRange(TCDef) ;
   if (!RangeTC.isSingleElement()) {
     auto ZeroRange =
         ConstantRange(APInt(TripCount->getType()->getScalarSizeInBits(), 0));
     RangeTC = RangeTC.unionWith(ZeroRange);
   }
   if (!RangeTC.contains(RangeCeil) && !ForceTailPredication) {
     LLVM_DEBUG(dbgs() << "ARM TP: Overflow possible in sub\n");
     return false;
if (VectorWidth == StepValue)
     return true;
 
   LLVM_DEBUG(dbgs() << "ARM TP: Step value " << StepValue << " doesn't match "
              "vector width " << VectorWidth << "\n");
 
   return false;
 }
 
-// Materialize NumElements in the preheader block.
-static Value *getNumElements(BasicBlock *Preheader, Value *BTC) {
-  // First, check the preheader if it not already exist:
-  //
-  // preheader:
-  //    %BTC = add i32 %N, -1
-  //    ..
-  // vector.body:
-  //
-  // if %BTC already exists. We don't need to emit %NumElems = %BTC + 1,
-  // but instead can just return %N.
-  for (auto &I : *Preheader) {
-    if (I.getOpcode() != Instruction::Add || &I != BTC)
-      continue;
-    ConstantInt *MinusOne = nullptr;
-    if (!(MinusOne = dyn_cast<ConstantInt>(I.getOperand(1))))
-      continue;
-    if (MinusOne->getSExtValue() == -1) {
-      LLVM_DEBUG(dbgs() << "ARM TP: Found num elems: " << I << "\n");
-      return I.getOperand(0);
-    }
-  }
-
-  // But we do need to materialise BTC if it is not already there,
-  // e.g. if it is a constant.
-  IRBuilder<> Builder(Preheader->getTerminator());
-  Value *NumElements = Builder.CreateAdd(BTC,
-        ConstantInt::get(BTC->getType(), 1), "num.elements");
-  LLVM_DEBUG(dbgs() << "ARM TP: Created num elems: " << *NumElements << "\n");
-  return NumElements;
-}
-
 void MVETailPredication::InsertVCTPIntrinsic(IntrinsicInst *ActiveLaneMask,
     Value *TripCount, FixedVectorType *VecTy) {
   IRBuilder<> Builder(L->getLoopPreheader()->getTerminator());
   Module *M = L->getHeader()->getModule();
   Type *Ty = IntegerType::get(M->getContext(), 32);
   unsigned VectorWidth = VecTy->getNumElements();
 
-  // The backedge-taken count in @llvm.get.active.lane.mask, its 2nd operand,
-  // is one less than the trip count. So we need to find or create
-  // %num.elements = %BTC + 1 in the preheader.
-  Value *BTC = ActiveLaneMask->getOperand(1);
-  Builder.SetInsertPoint(L->getLoopPreheader()->getTerminator());
-  Value *NumElements = getNumElements(L->getLoopPreheader(), BTC);
-
   // Insert a phi to count the number of elements processed by the loop.
   Builder.SetInsertPoint(L->getHeader()->getFirstNonPHI()  );
   PHINode *Processed = Builder.CreatePHI(Ty, 2);
-  Processed->addIncoming(NumElements, L->getLoopPreheader());
+  Processed->addIncoming(ActiveLaneMask->getOperand(1), L->getLoopPreheader());
 
-  // Replace @llvm.get.active.mask() with the ARM specific VCTP intrinic, and thus
-  // represent the effect of tail predication.
+  // Replace @llvm.get.active.mask() with the ARM specific VCTP intrinic, and
+  // thus represent the effect of tail predication.
   Builder.SetInsertPoint(ActiveLaneMask);
-  ConstantInt *Factor =
-    ConstantInt::get(cast<IntegerType>(Ty), VectorWidth);
+  ConstantInt *Factor = ConstantInt::get(cast<IntegerType>(Ty), VectorWidth);
 
   Intrinsic::ID VCTPID;
   switch (VectorWidth) {
   default:
     llvm_unreachable("unexpected number of lanes");
   case 4:  VCTPID = Intrinsic::arm_mve_vctp32; break;
   case 8:  VCTPID = Intrinsic::arm_mve_vctp16; break;
   case 16: VCTPID = Intrinsic::arm_mve_vctp8; break;

llvm/test/CodeGen/Thumb2/LowOverheadLoops/basic-tail-pred.ll

vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <16 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <16 x i32> %broadcast.splatinsert, <16 x i32> undef, <16 x i32> zeroinitializer
   %induction = or <16 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
   %tmp = getelementptr inbounds i8, i8* %a, i32 %index
 
 ;  %tmp1 = icmp ule <16 x i32> %induction, %broadcast.splat11
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i8* %tmp to <16 x i8>*
   %wide.masked.load = tail call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %tmp2, i32 4, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %tmp3 = getelementptr inbounds i8, i8* %b, i32 %index
   %tmp4 = bitcast i8* %tmp3 to <16 x i8>*
   %wide.masked.load2 = tail call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %tmp4, i32 4, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %mul = mul nsw <16 x i8> %wide.masked.load2, %wide.masked.load
   %tmp6 = getelementptr inbounds i8, i8* %c, i32 %index
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %tmp = getelementptr inbounds i16, i16* %a, i32 %index
 
 ;  %tmp1 = icmp ule <8 x i32> %induction, %broadcast.splat11
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i16* %tmp to <8 x i16>*
   %wide.masked.load = tail call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp2, i32 4, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %tmp3 = getelementptr inbounds i16, i16* %b, i32 %index
   %tmp4 = bitcast i16* %tmp3 to <8 x i16>*
   %wide.masked.load2 = tail call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp4, i32 4, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %mul = mul nsw <8 x i16> %wide.masked.load2, %wide.masked.load
   %tmp6 = getelementptr inbounds i16, i16* %c, i32 %index
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
  ; %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat11
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
   %wide.masked.load = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load2 = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %mul = mul nsw <4 x i32> %wide.masked.load2, %wide.masked.load
   %tmp6 = getelementptr inbounds i32, i32* %c, i32 %index
   %tmp7 = bitcast i32* %tmp6 to <4 x i32>*
   tail call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %mul, <4 x i32>* %tmp7, i32 4, <4 x i1> %active.lane.mask)

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 ;  %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %extract.1.low = shufflevector <4 x i32> %wide.masked.load, <4 x i32> undef, < 2 x i32> < i32 0, i32 2>
   %extract.1.high = shufflevector <4 x i32> %wide.masked.load, <4 x i32> undef, < 2 x i32> < i32 1, i32 3>
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load2 = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %extract.2.low = shufflevector <4 x i32> %wide.masked.load2, <4 x i32> undef, < 2 x i32> < i32 0, i32 2>
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 
 ;  %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %wrong = icmp ult <4 x i32> %induction, %broadcast.splat11
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load12 = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %wrong, <4 x i32> undef)
   %tmp5 = mul nsw <4 x i32> %wide.masked.load12, %wide.masked.load
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 
 ;  %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %wrong = icmp ult <4 x i32> %induction, %broadcast.splat11
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load12 = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %tmp5 = mul nsw <4 x i32> %wide.masked.load12, %wide.masked.load

 ; TODO: Multiple intrinsics not yet supported.
 ; This is currently rejected, because if the vector body is unrolled, the step
 ; is not what we expect:
 ;
 ;   Step value 16 doesn't match vector width 4
 ;
 ; CHECK-LABEL: interleave4
 ; CHECK: vector.body:
-; CHECK:  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
-; CHECK:  %active.lane.mask{{.*}} = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v7, i32 %trip.count.minus.1)
-; CHECK:  %active.lane.mask{{.*}} = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v8, i32 %trip.count.minus.1)
-; CHECK:  %active.lane.mask{{.*}} = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v9, i32 %trip.count.minus.1)
+; CHECK:  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
+; CHECK:  %active.lane.mask{{.*}} = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v7, i32 %N)
+; CHECK:  %active.lane.mask{{.*}} = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v8, i32 %N)
+; CHECK:  %active.lane.mask{{.*}} = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v9, i32 %N)
 ;
 define dso_local void @interleave4(i32* noalias nocapture %A, i32* noalias nocapture readonly %B, i32* noalias nocapture readonly %C, i32 %N) local_unnamed_addr #0 {
 entry:
   %cmp8 = icmp sgt i32 %N, 0
   %v0 = add i32 %N, 15
   %v1 = lshr i32 %v0, 4
   %v2 = shl nuw i32 %v1, 4
   %v3 = add i32 %v2, -16
vector.body:
   %lsr.iv38 = phi i32* [ %scevgep39, %vector.body ], [ %scevgep37, %vector.ph ]
   %lsr.iv31 = phi i32* [ %scevgep32, %vector.body ], [ %scevgep30, %vector.ph ]
   %lsr.iv = phi i32* [ %scevgep25, %vector.body ], [ %scevgep, %vector.ph ]
   %index = phi i32 [ 0, %vector.ph ], [ %v14, %vector.body ]
   %v6 = phi i32 [ %v5, %vector.ph ], [ %v15, %vector.body ]
   %lsr.iv3840 = bitcast i32* %lsr.iv38 to <4 x i32>*
   %lsr.iv3133 = bitcast i32* %lsr.iv31 to <4 x i32>*
   %lsr.iv26 = bitcast i32* %lsr.iv to <4 x i32>*
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
   %v7 = add i32 %index, 4
-  %active.lane.mask15 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v7, i32 %trip.count.minus.1)
+  %active.lane.mask15 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v7, i32 %N)
   %v8 = add i32 %v7, 4
-  %active.lane.mask16 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v8, i32 %trip.count.minus.1)
+  %active.lane.mask16 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v8, i32 %N)
   %v9 = add i32 %v8, 4
-  %active.lane.mask17 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v9, i32 %trip.count.minus.1)
+  %active.lane.mask17 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %v9, i32 %N)
   %scevgep42 = getelementptr <4 x i32>, <4 x i32>* %lsr.iv3840, i32 -2
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %scevgep42, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %scevgep43 = getelementptr <4 x i32>, <4 x i32>* %lsr.iv3840, i32 -1
   %wide.masked.load18 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* nonnull %scevgep43, i32 4, <4 x i1> %active.lane.mask15, <4 x i32> undef)
   %wide.masked.load19 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* nonnull %lsr.iv3840, i32 4, <4 x i1> %active.lane.mask16, <4 x i32> undef)
   %scevgep41 = getelementptr <4 x i32>, <4 x i32>* %lsr.iv3840, i32 1
   %wide.masked.load20 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* nonnull %scevgep41, i32 4, <4 x i1> %active.lane.mask17, <4 x i32> undef)
   %scevgep34 = getelementptr <4 x i32>, <4 x i32>* %lsr.iv3133, i32 -2

llvm/test/CodeGen/Thumb2/LowOverheadLoops/clear-maskedinsts.ll

 ; CHECK-NEXT:    br i1 [[TOBOOL]], label [[VECTOR_BODY75_PREHEADER:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.body75.preheader:
 ; CHECK-NEXT:    call void @llvm.set.loop.iterations.i32(i32 [[TMP2]])
 ; CHECK-NEXT:    br label [[VECTOR_BODY75:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT71:%.*]] = insertelement <4 x i32> undef, i32 [[X]], i32 0
 ; CHECK-NEXT:    [[BROADCAST_SPLAT72:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT71]], <4 x i32> undef, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    call void @llvm.set.loop.iterations.i32(i32 [[TMP3]])
-; CHECK-NEXT:    [[NUM_ELEMENTS:%.*]] = add i32 [[TRIP_COUNT_MINUS_183]], 1
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[LSR_IV9:%.*]] = phi i32* [ [[SCEVGEP10:%.*]], [[VECTOR_BODY]] ], [ [[D:%.*]], [[VECTOR_PH]] ]
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP4:%.*]] = phi i32 [ [[TMP3]], [[VECTOR_PH]] ], [ [[TMP10:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[TMP5:%.*]] = phi i32 [ [[NUM_ELEMENTS]], [[VECTOR_PH]] ], [ [[TMP9:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP5:%.*]] = phi i32 [ [[N]], [[VECTOR_PH]] ], [ [[TMP9:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[LSR_IV911:%.*]] = bitcast i32* [[LSR_IV9]] to <4 x i32>*
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[INDEX]], i32 0
 ; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32> undef, i32 [[TRIP_COUNT_MINUS_183]], i32 0
 ; CHECK-NEXT:    [[TMP7:%.*]] = shufflevector <4 x i32> [[TMP6]], <4 x i32> undef, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP8:%.*]] = call <4 x i1> @llvm.arm.mve.vctp32(i32 [[TMP5]])
 ; CHECK-NEXT:    [[TMP9]] = sub i32 [[TMP5]], 4
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %4 = phi i32 [ %3, %vector.ph ], [ %8, %vector.body ]
   %lsr.iv911 = bitcast i32* %lsr.iv9 to <4 x i32>*
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %5 = insertelement <4 x i32> undef, i32 %trip.count.minus.183, i32 0
   %6 = shufflevector <4 x i32> %5, <4 x i32> undef, <4 x i32> zeroinitializer
-  %7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.183)
+  %7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %broadcast.splat72, <4 x i32>* %lsr.iv911, i32 4, <4 x i1> %7)
   %index.next = add i32 %index, 4
   %scevgep10 = getelementptr i32, i32* %lsr.iv9, i32 4
   %8 = call i32 @llvm.loop.decrement.reg.i32(i32 %4, i32 1)
   %9 = icmp ne i32 %8, 0
   br i1 %9, label %vector.body, label %for.cond.cleanup
 
 vector.body75:                                    ; preds = %vector.body75, %vector.body75.preheader

llvm/test/CodeGen/Thumb2/LowOverheadLoops/cond-vector-reduce-mve-codegen.ll

vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %add, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 
 ;  %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat12
-  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load.a = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load.b = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %tmp5 = getelementptr inbounds i32, i32* %c, i32 %index
   %tmp6 = bitcast i32* %tmp5 to <4 x i32>*
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %add, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 
 ;  %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat12
-  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load.a = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load.b = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %tmp5 = getelementptr inbounds i32, i32* %c, i32 %index
   %tmp6 = bitcast i32* %tmp5 to <4 x i32>*
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %add, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 
 ;  %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat12
-  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load.a = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load.b = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %sub = sub <4 x i32> %wide.masked.load.a, %wide.masked.load.b
   %cmp = icmp eq <4 x i32> %sub, <i32 0, i32 0, i32 0, i32 0>
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %add, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 
 ;  %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat12
-  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load.a = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load.b = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %sub = sub <4 x i32> %wide.masked.load.a, %wide.masked.load.b
   %cmp = icmp eq <4 x i32> %sub, <i32 0, i32 0, i32 0, i32 0>

 bb9:                                              ; preds = %bb9, %bb3
   %tmp10 = phi i32 [ 0, %bb3 ], [ %tmp25, %bb9 ]
   %tmp11 = insertelement <4 x i32> undef, i32 %tmp10, i32 0
   %tmp12 = shufflevector <4 x i32> %tmp11, <4 x i32> undef, <4 x i32> zeroinitializer
   %tmp13 = add <4 x i32> %tmp12, <i32 0, i32 1, i32 2, i32 3>
   %tmp14 = getelementptr inbounds i32, i32* %arg1, i32 %tmp10
 
   ; %tmp15 = icmp ule <4 x i32> %tmp13, %tmp8
-  %tmp15 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %tmp10, i32 %tmp6)
+  %tmp15 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %tmp10, i32 %arg2)
 
   %tmp16 = bitcast i32* %tmp14 to <4 x i32>*
   %tmp17 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp16, i32 4, <4 x i1> %tmp15, <4 x i32> undef)
   %tmp18 = icmp ne <4 x i32> %tmp17, zeroinitializer
   %tmp19 = getelementptr inbounds i32, i32* %arg, i32 %tmp10
   %tmp20 = and <4 x i1> %tmp18, %tmp15
   %tmp21 = bitcast i32* %tmp19 to <4 x i32>*
   %tmp22 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp21, i32 4, <4 x i1> %tmp20, <4 x i32> undef)

 bb12:                                             ; preds = %bb12, %bb4
   %tmp13 = phi i32 [ 0, %bb4 ], [ %tmp30, %bb12 ]
   %tmp14 = insertelement <4 x i32> undef, i32 %tmp13, i32 0
   %tmp15 = shufflevector <4 x i32> %tmp14, <4 x i32> undef, <4 x i32> zeroinitializer
   %tmp16 = add <4 x i32> %tmp15, <i32 0, i32 1, i32 2, i32 3>
   %tmp17 = getelementptr inbounds i32, i32* %arg, i32 %tmp13
 
   ; %tmp18 = icmp ule <4 x i32> %tmp16, %tmp9
-  %tmp18= call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %tmp13, i32 %tmp7)
+  %tmp18= call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %tmp13, i32 %arg3)
 
   %tmp19 = bitcast i32* %tmp17 to <4 x i32>*
   %tmp20 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp19, i32 4, <4 x i1> %tmp18, <4 x i32> undef)
   %tmp21 = icmp ne <4 x i32> %tmp20, zeroinitializer
   %tmp22 = icmp sle <4 x i32> %tmp20, %tmp11
   %tmp23 = getelementptr inbounds i32, i32* %arg1, i32 %tmp13
   %tmp24 = and <4 x i1> %tmp22, %tmp21
   %tmp25 = and <4 x i1> %tmp24, %tmp18

llvm/test/CodeGen/Thumb2/LowOverheadLoops/extending-loads.ll

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = or <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %0 = getelementptr inbounds i8, i8* %b, i32 %index
 
   ; %1 = icmp ule <8 x i32> %induction, %broadcast.splat11
-  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i8* %0 to <8 x i8>*
   %wide.masked.load = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %2, i32 1, <8 x i1> %1, <8 x i8> undef)
   %3 = sext <8 x i8> %wide.masked.load to <8 x i16>
   %4 = getelementptr inbounds i16, i16* %a, i32 %index
   %5 = bitcast i16* %4 to <8 x i16>*
   %wide.masked.load12 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %5, i32 2, <8 x i1> %1, <8 x i16> undef)
   %6 = add <8 x i16> %wide.masked.load12, %3

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = or <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %0 = getelementptr inbounds i8, i8* %b, i32 %index
 
   ; %1 = icmp ule <8 x i32> %induction, %broadcast.splat11
-  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i8* %0 to <8 x i8>*
   %wide.masked.load = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %2, i32 1, <8 x i1> %1, <8 x i8> undef)
   %3 = zext <8 x i8> %wide.masked.load to <8 x i16>
   %4 = getelementptr inbounds i16, i16* %a, i32 %index
   %5 = bitcast i16* %4 to <8 x i16>*
   %wide.masked.load12 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %5, i32 2, <8 x i1> %1, <8 x i16> undef)
   %6 = add <8 x i16> %wide.masked.load12, %3

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i16, i16* %b, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat9
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i16* %0 to <4 x i16>*
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %2, i32 2, <4 x i1> %1, <4 x i16> undef)
   %3 = sext <4 x i16> %wide.masked.load to <4 x i32>
   %4 = getelementptr inbounds i32, i32* %a, i32 %index
   %5 = bitcast i32* %4 to <4 x i32>*
   %wide.masked.load10 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %5, i32 4, <4 x i1> %1, <4 x i32> undef)
   %6 = add nsw <4 x i32> %wide.masked.load10, %3

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i16, i16* %b, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat9
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i16* %0 to <4 x i16>*
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %2, i32 2, <4 x i1> %1, <4 x i16> undef)
   %3 = zext <4 x i16> %wide.masked.load to <4 x i32>
   %4 = getelementptr inbounds i32, i32* %a, i32 %index
   %5 = bitcast i32* %4 to <4 x i32>*
   %wide.masked.load10 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %5, i32 4, <4 x i1> %1, <4 x i32> undef)
   %6 = add <4 x i32> %wide.masked.load10, %3

llvm/test/CodeGen/Thumb2/LowOverheadLoops/fast-fp-loops.ll

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %2 = getelementptr inbounds float, float* %b, i32 %index
 
   ; %3 = icmp ule <4 x i32> %induction, %broadcast.splat22
-  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %4 = bitcast float* %2 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %3, <4 x float> undef)
   %5 = getelementptr inbounds float, float* %c, i32 %index
   %6 = bitcast float* %5 to <4 x float>*
   %wide.masked.load23 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %6, i32 4, <4 x i1> %3, <4 x float> undef)
   %7 = fmul fast <4 x float> %wide.masked.load23, %wide.masked.load
   %8 = getelementptr inbounds float, float* %a, i32 %index
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x float> [ zeroinitializer, %vector.ph ], [ %6, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %b, i32 %index
 
 ;  %1 = icmp ule <4 x i32> %induction, %broadcast.splat12
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %c, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load13 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = fmul fast <4 x float> %wide.masked.load13, %wide.masked.load
   %6 = fadd fast <4 x float> %5, %vec.phi

llvm/test/CodeGen/Thumb2/LowOverheadLoops/mve-tail-data-types.ll

vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %5, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i8, i8* %b, i32 %index
 
 ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i8* %0 to <4 x i8>*
   %wide.masked.load = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %2, i32 1, <4 x i1> %1, <4 x i8> undef)
   %3 = zext <4 x i8> %wide.masked.load to <4 x i32>
   %4 = mul nuw nsw <4 x i32> %broadcast.splat13, %3
   %5 = add nuw nsw <4 x i32> %4, %vec.phi
   %index.next = add i32 %index, 4
   %6 = icmp eq i32 %index.next, %n.vec
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %5, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i16, i16* %b, i32 %index
 
 ;  %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i16* %0 to <4 x i16>*
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %2, i32 2, <4 x i1> %1, <4 x i16> undef)
   %3 = sext <4 x i16> %wide.masked.load to <4 x i32>
   %4 = mul nsw <4 x i32> %broadcast.splat13, %3
   %5 = add nsw <4 x i32> %4, %vec.phi
   %index.next = add i32 %index, 4
   %6 = icmp eq i32 %index.next, %n.vec
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %5, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i8, i8* %b, i32 %index
 
 ;  %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i8* %0 to <4 x i8>*
   %wide.masked.load = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %2, i32 1, <4 x i1> %1, <4 x i8> undef)
   %3 = zext <4 x i8> %wide.masked.load to <4 x i32>
   %4 = mul nuw nsw <4 x i32> %broadcast.splat13, %3
   %5 = add nuw nsw <4 x i32> %4, %vec.phi
   %index.next = add i32 %index, 4
   %6 = icmp eq i32 %index.next, %n.vec
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %5, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i16, i16* %b, i32 %index
 
 ;  %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i16* %0 to <4 x i16>*
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %2, i32 2, <4 x i1> %1, <4 x i16> undef)
   %3 = zext <4 x i16> %wide.masked.load to <4 x i32>
   %4 = mul nsw <4 x i32> %broadcast.splat13, %3
   %5 = add nsw <4 x i32> %4, %vec.phi
   %index.next = add i32 %index, 4
   %6 = icmp eq i32 %index.next, %n.vec
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %4, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i32, i32* %b, i32 %index
 
 ;  %1 = icmp ule <4 x i32> %induction, %broadcast.splat10
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %2, i32 4, <4 x i1> %1, <4 x i32> undef)
   %3 = mul nsw <4 x i32> %wide.masked.load, %broadcast.splat12
   %4 = add nsw <4 x i32> %3, %vec.phi
   %index.next = add i32 %index, 4
   %5 = icmp eq i32 %index.next, %n.vec
   br i1 %5, label %middle.block, label %vector.body

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %2 = getelementptr inbounds i8, i8* %a, i32 %index
 
   ; %3 = icmp ule <4 x i32> %induction, %broadcast.splat20
-  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %4 = bitcast i8* %2 to <4 x i8>*
   %wide.masked.load = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %4, i32 1, <4 x i1> %3, <4 x i8> undef)
   %5 = zext <4 x i8> %wide.masked.load to <4 x i32>
   %6 = getelementptr inbounds i8, i8* %b, i32 %index
   %7 = bitcast i8* %6 to <4 x i8>*
   %wide.masked.load21 = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %7, i32 1, <4 x i1> %3, <4 x i8> undef)
   %8 = zext <4 x i8> %wide.masked.load21 to <4 x i32>

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i16, i16* %a, i32 %index
 
 ;  %1 = icmp ule <4 x i32> %induction, %broadcast.splat13
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i16* %0 to <4 x i16>*
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %2, i32 2, <4 x i1> %1, <4 x i16> undef)
   %3 = sext <4 x i16> %wide.masked.load to <4 x i32>
   %4 = getelementptr inbounds i16, i16* %b, i32 %index
   %5 = bitcast i16* %4 to <4 x i16>*
   %wide.masked.load14 = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %5, i32 2, <4 x i1> %1, <4 x i16> undef)
   %6 = sext <4 x i16> %wide.masked.load14 to <4 x i32>

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %2 = getelementptr inbounds i8, i8* %a, i32 %index
 
 ;  %3 = icmp ule <4 x i32> %induction, %broadcast.splat20
-  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %4 = bitcast i8* %2 to <4 x i8>*
   %wide.masked.load = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %4, i32 1, <4 x i1> %3, <4 x i8> undef)
   %5 = zext <4 x i8> %wide.masked.load to <4 x i32>
   %6 = getelementptr inbounds i8, i8* %b, i32 %index
   %7 = bitcast i8* %6 to <4 x i8>*
   %wide.masked.load21 = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %7, i32 1, <4 x i1> %3, <4 x i8> undef)
   %8 = zext <4 x i8> %wide.masked.load21 to <4 x i32>

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i16, i16* %a, i32 %index
 
 ;  %1 = icmp ule <4 x i32> %induction, %broadcast.splat13
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i16* %0 to <4 x i16>*
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %2, i32 2, <4 x i1> %1, <4 x i16> undef)
   %3 = zext <4 x i16> %wide.masked.load to <4 x i32>
   %4 = getelementptr inbounds i16, i16* %b, i32 %index
   %5 = bitcast i16* %4 to <4 x i16>*
   %wide.masked.load14 = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %5, i32 2, <4 x i1> %1, <4 x i16> undef)
   %6 = zext <4 x i16> %wide.masked.load14 to <4 x i32>

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %2 = getelementptr inbounds i32, i32* %a, i32 %index
 
 ;  %3 = icmp ule <4 x i32> %induction, %broadcast.splat22
-  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %3 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %4 = bitcast i32* %2 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %4, i32 4, <4 x i1> %3, <4 x i32> undef)
   %5 = getelementptr inbounds i32, i32* %b, i32 %index
   %6 = bitcast i32* %5 to <4 x i32>*
   %wide.masked.load23 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %6, i32 4, <4 x i1> %3, <4 x i32> undef)
   %7 = mul nsw <4 x i32> %wide.masked.load23, %wide.masked.load
   %8 = add nsw <4 x i32> %7, %broadcast.splat25

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %0 = getelementptr inbounds i8, i8* %b, i32 %index
 
 ;  %1 = icmp ule <8 x i32> %induction, %broadcast.splat13
-  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i8* %0 to <8 x i8>*
   %wide.masked.load = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %2, i32 1, <8 x i1> %1, <8 x i8> undef)
   %3 = zext <8 x i8> %wide.masked.load to <8 x i16>
   %4 = getelementptr inbounds i8, i8* %c, i32 %index
   %5 = bitcast i8* %4 to <8 x i8>*
   %wide.masked.load14 = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %5, i32 1, <8 x i1> %1, <8 x i8> undef)
   %6 = zext <8 x i8> %wide.masked.load14 to <8 x i16>

llvm/test/CodeGen/Thumb2/LowOverheadLoops/nested.ll

 ; CHECK:       for.cond1.preheader.us:
 ; CHECK-NEXT:    [[I_025_US:%.*]] = phi i32 [ [[INC10_US:%.*]], [[MIDDLE_BLOCK:%.*]] ], [ 0, [[FOR_COND1_PREHEADER_US_PREHEADER]] ]
 ; CHECK-NEXT:    [[ARRAYIDX_US:%.*]] = getelementptr inbounds i16*, i16** [[A:%.*]], i32 [[I_025_US]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = load i16*, i16** [[ARRAYIDX_US]], align 4
 ; CHECK-NEXT:    [[ARRAYIDX8_US:%.*]] = getelementptr inbounds i32, i32* [[C:%.*]], i32 [[I_025_US]]
 ; CHECK-NEXT:    [[ARRAYIDX8_PROMOTED_US:%.*]] = load i32, i32* [[ARRAYIDX8_US]], align 4
 ; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> <i32 undef, i32 0, i32 0, i32 0>, i32 [[ARRAYIDX8_PROMOTED_US]], i32 0
 ; CHECK-NEXT:    call void @llvm.set.loop.iterations.i32(i32 [[TMP2]])
-; CHECK-NEXT:    [[NUM_ELEMENTS:%.*]] = add i32 [[TRIP_COUNT_MINUS_1]], 1
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[FOR_COND1_PREHEADER_US]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ [[TMP4]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP14:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP5:%.*]] = phi i32 [ [[TMP2]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP15:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[TMP0:%.*]] = phi i32 [ [[NUM_ELEMENTS]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP2:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP0:%.*]] = phi i32 [ [[N]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP2:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[INDEX]], i32 0
 ; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i16, i16* [[TMP3]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.vctp32(i32 [[TMP0]])
 ; CHECK-NEXT:    [[TMP2]] = sub i32 [[TMP0]], 4
 ; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i16* [[TMP6]] to <4 x i16>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* [[TMP8]], i32 2, <4 x i1> [[TMP1]], <4 x i16> undef)
vector.body:                                      ; preds = %vector.body, %for.cond1.preheader.us
   %vec.phi = phi <4 x i32> [ %tmp4, %for.cond1.preheader.us ], [ %tmp14, %vector.body ]
   %tmp5 = phi i32 [ %tmp2, %for.cond1.preheader.us ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp6 = getelementptr inbounds i16, i16* %tmp3, i32 %index
 
   ; %tmp7 = icmp ule <4 x i32> %induction, %broadcast.splat29
-  %tmp7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %tmp8 = bitcast i16* %tmp6 to <4 x i16>*
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %tmp8, i32 2, <4 x i1> %tmp7, <4 x i16> undef)
   %tmp9 = sext <4 x i16> %wide.masked.load to <4 x i32>
   %tmp10 = getelementptr inbounds i16, i16* %B, i32 %index
   %tmp11 = bitcast i16* %tmp10 to <4 x i16>*
   %wide.masked.load30 = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %tmp11, i32 2, <4 x i1> %tmp7, <4 x i16> undef)
   %tmp12 = sext <4 x i16> %wide.masked.load30 to <4 x i32>

 ; CHECK:       for.cond1.preheader.us:
 ; CHECK-NEXT:    [[I_024_US:%.*]] = phi i32 [ [[INC9_US:%.*]], [[MIDDLE_BLOCK:%.*]] ], [ 0, [[FOR_COND1_PREHEADER_US_PREHEADER]] ]
 ; CHECK-NEXT:    [[ARRAYIDX_US:%.*]] = getelementptr inbounds i32*, i32** [[A:%.*]], i32 [[I_024_US]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = load i32*, i32** [[ARRAYIDX_US]], align 4
 ; CHECK-NEXT:    [[ARRAYIDX7_US:%.*]] = getelementptr inbounds i32, i32* [[C:%.*]], i32 [[I_024_US]]
 ; CHECK-NEXT:    [[ARRAYIDX7_PROMOTED_US:%.*]] = load i32, i32* [[ARRAYIDX7_US]], align 4
 ; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> <i32 undef, i32 0, i32 0, i32 0>, i32 [[ARRAYIDX7_PROMOTED_US]], i32 0
 ; CHECK-NEXT:    call void @llvm.set.loop.iterations.i32(i32 [[TMP2]])
-; CHECK-NEXT:    [[NUM_ELEMENTS:%.*]] = add i32 [[TRIP_COUNT_MINUS_1]], 1
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[FOR_COND1_PREHEADER_US]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ [[TMP4]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP12:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP5:%.*]] = phi i32 [ [[TMP2]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP13:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[TMP0:%.*]] = phi i32 [ [[NUM_ELEMENTS]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP2:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP0:%.*]] = phi i32 [ [[N]], [[FOR_COND1_PREHEADER_US]] ], [ [[TMP2:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[INDEX]], i32 0
 ; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP3]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.vctp32(i32 [[TMP0]])
 ; CHECK-NEXT:    [[TMP2]] = sub i32 [[TMP0]], 4
 ; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i32* [[TMP6]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP8]], i32 4, <4 x i1> [[TMP1]], <4 x i32> undef)
vector.body:                                      ; preds = %vector.body, %for.cond1.preheader.us
   %vec.phi = phi <4 x i32> [ %tmp4, %for.cond1.preheader.us ], [ %tmp12, %vector.body ]
   %tmp5 = phi i32 [ %tmp2, %for.cond1.preheader.us ], [ %tmp13, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp6 = getelementptr inbounds i32, i32* %tmp3, i32 %index
 
   ; %tmp7 = icmp ule <4 x i32> %induction, %broadcast.splat28
-  %tmp7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %tmp8 = bitcast i32* %tmp6 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp8, i32 4, <4 x i1> %tmp7, <4 x i32> undef)
   %tmp9 = getelementptr inbounds i32, i32* %B, i32 %index
   %tmp10 = bitcast i32* %tmp9 to <4 x i32>*
   %wide.masked.load29 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp10, i32 4, <4 x i1> %tmp7, <4 x i32> undef)
   %tmp11 = mul nsw <4 x i32> %wide.masked.load29, %wide.masked.load
   %tmp12 = add nsw <4 x i32> %vec.phi, %tmp11

llvm/test/CodeGen/Thumb2/LowOverheadLoops/reductions.ll

vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <16 x i8> [ zeroinitializer, %vector.ph ], [ %i5, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %i1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %i2 = getelementptr inbounds i8, i8* %b, i32 %index
   %i3 = bitcast i8* %i2 to <16 x i8>*
   %wide.masked.load16 = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %i3, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %i4 = add <16 x i8> %wide.masked.load, %vec.phi
   %i5 = add <16 x i8> %i4, %wide.masked.load16
   %index.next = add i32 %index, 16
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <8 x i16> [ zeroinitializer, %vector.ph ], [ %i7, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <8 x i8>*
   %wide.masked.load = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i1, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i2 = zext <8 x i8> %wide.masked.load to <8 x i16>
   %i3 = getelementptr inbounds i8, i8* %b, i32 %index
   %i4 = bitcast i8* %i3 to <8 x i8>*
   %wide.masked.load17 = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i4, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i5 = zext <8 x i8> %wide.masked.load17 to <8 x i16>
   %i6 = add <8 x i16> %vec.phi, %i2
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <16 x i8> [ zeroinitializer, %vector.ph ], [ %i5, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %i1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %i2 = getelementptr inbounds i8, i8* %b, i32 %index
   %i3 = bitcast i8* %i2 to <16 x i8>*
   %wide.masked.load16 = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %i3, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %i4 = sub <16 x i8> %wide.masked.load, %wide.masked.load16
   %i5 = add <16 x i8> %i4, %vec.phi
   %index.next = add i32 %index, 16
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <8 x i16> [ zeroinitializer, %vector.ph ], [ %i7, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <8 x i8>*
   %wide.masked.load = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i1, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i2 = zext <8 x i8> %wide.masked.load to <8 x i16>
   %i3 = getelementptr inbounds i8, i8* %b, i32 %index
   %i4 = bitcast i8* %i3 to <8 x i8>*
   %wide.masked.load17 = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i4, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i5 = zext <8 x i8> %wide.masked.load17 to <8 x i16>
   %i6 = sub <8 x i16> %i5, %i2
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <16 x i8> [ zeroinitializer, %vector.ph ], [ %i5, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %i1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %i2 = getelementptr inbounds i8, i8* %b, i32 %index
   %i3 = bitcast i8* %i2 to <16 x i8>*
   %wide.masked.load15 = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %i3, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %i4 = mul <16 x i8> %wide.masked.load15, %wide.masked.load
   %i5 = add <16 x i8> %i4, %vec.phi
   %index.next = add i32 %index, 16
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <8 x i16> [ zeroinitializer, %vector.ph ], [ %i7, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <8 x i8>*
   %wide.masked.load = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i1, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i2 = zext <8 x i8> %wide.masked.load to <8 x i16>
   %i3 = getelementptr inbounds i8, i8* %b, i32 %index
   %i4 = bitcast i8* %i3 to <8 x i8>*
   %wide.masked.load17 = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i4, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i5 = zext <8 x i8> %wide.masked.load17 to <8 x i16>
   %i6 = mul <8 x i16> %i5, %i2
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %i7, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <4 x i8>*
   %wide.masked.load = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %i1, i32 1, <4 x i1> %active.lane.mask, <4 x i8> undef)
   %i2 = zext <4 x i8> %wide.masked.load to <4 x i32>
   %i3 = getelementptr inbounds i8, i8* %b, i32 %index
   %i4 = bitcast i8* %i3 to <4 x i8>*
   %wide.masked.load43 = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %i4, i32 1, <4 x i1> %active.lane.mask, <4 x i8> undef)
   %i5 = zext <4 x i8> %wide.masked.load43 to <4 x i32>
   %i6 = mul nuw nsw <4 x i32> %i5, %i2
vector.ph47:                                      ; preds = %middle.block
   %trip.count.minus.154 = add i32 %N, -1
   %i11 = insertelement <4 x i32> <i32 undef, i32 0, i32 0, i32 0>, i32 %i10, i32 0
   br label %vector.body46
 
 vector.body46:                                    ; preds = %vector.body46, %vector.ph47
   %index51 = phi i32 [ 0, %vector.ph47 ], [ %index.next52, %vector.body46 ]
   %vec.phi60 = phi <4 x i32> [ %i11, %vector.ph47 ], [ %i19, %vector.body46 ]
   %i12 = getelementptr inbounds i8, i8* %a, i32 %index51
-  %active.lane.mask61 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index51, i32 %trip.count.minus.154)
+  %active.lane.mask61 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index51, i32 %N)
   %i13 = bitcast i8* %i12 to <4 x i8>*
   %wide.masked.load62 = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %i13, i32 1, <4 x i1> %active.lane.mask61, <4 x i8> undef)
   %i14 = zext <4 x i8> %wide.masked.load62 to <4 x i32>
   %i15 = getelementptr inbounds i8, i8* %b, i32 %index51
   %i16 = bitcast i8* %i15 to <4 x i8>*
   %wide.masked.load63 = call <4 x i8> @llvm.masked.load.v4i8.p0v4i8(<4 x i8>* %i16, i32 1, <4 x i1> %active.lane.mask61, <4 x i8> undef)
   %i17 = zext <4 x i8> %wide.masked.load63 to <4 x i32>
   %i18 = mul nuw nsw <4 x i32> %i17, %i14
vector.ph:                                        ; preds = %entry
   %trip.count.minus.1 = add i32 %N, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <8 x i16> [ zeroinitializer, %vector.ph ], [ %i8, %vector.body ]
   %vec.phi.1 = phi <8 x i16> [ zeroinitializer, %vector.ph ], [ %i9, %vector.body ]
   %i = getelementptr inbounds i8, i8* %a, i32 %index
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
   %i1 = bitcast i8* %i to <8 x i8>*
   %wide.masked.load = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i1, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i2 = zext <8 x i8> %wide.masked.load to <8 x i16>
   %i3 = getelementptr inbounds i8, i8* %b, i32 %index
   %i4 = bitcast i8* %i3 to <8 x i8>*
   %wide.masked.load17 = call <8 x i8> @llvm.masked.load.v8i8.p0v8i8(<8 x i8>* %i4, i32 1, <8 x i1> %active.lane.mask, <8 x i8> undef)
   %i5 = zext <8 x i8> %wide.masked.load17 to <8 x i16>
   %i6 = mul <8 x i16> %i5, %i2

llvm/test/CodeGen/Thumb2/LowOverheadLoops/tail-pred-const.ll

vector.body:
   %lsr.iv1416 = bitcast i32* %lsr.iv14 to <4 x i32>*
   %lsr.iv1113 = bitcast i32* %lsr.iv11 to <4 x i32>*
   %lsr.iv10 = bitcast i32* %lsr.iv to <4 x i32>*
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
 
   ; %1 = icmp ult <4 x i32> %induction, <i32 32002, i32 32002, i32 32002, i32 32002>
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32002)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32003)
 
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv10, i32 4, <4 x i1> %1, <4 x i32> undef)
   %wide.masked.load9 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv1113, i32 4, <4 x i1> %1, <4 x i32> undef)
   %2 = add nsw <4 x i32> %wide.masked.load9, %wide.masked.load
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %2, <4 x i32>* %lsr.iv1416, i32 4, <4 x i1> %1)
   %index.next = add i32 %index, 4
   %scevgep = getelementptr i32, i32* %lsr.iv, i32 4
   %scevgep12 = getelementptr i32, i32* %lsr.iv11, i32 4
vector.body:
   %0 = phi i32 [ 8001, %entry ], [ %3, %vector.body ]
   %lsr.iv1416 = bitcast i32* %lsr.iv14 to <4 x i32>*
   %lsr.iv1113 = bitcast i32* %lsr.iv11 to <4 x i32>*
   %lsr.iv10 = bitcast i32* %lsr.iv to <4 x i32>*
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
 
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32002)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32003)
 
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv10, i32 4, <4 x i1> %1, <4 x i32> undef)
   %wide.masked.load9 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv1113, i32 4, <4 x i1> %1, <4 x i32> undef)
   %2 = add nsw <4 x i32> %wide.masked.load9, %wide.masked.load
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %2, <4 x i32>* %lsr.iv1416, i32 4, <4 x i1> %1)
   %index.next = add i32 %index, 4
   %scevgep = getelementptr i32, i32* %lsr.iv, i32 4
   %scevgep12 = getelementptr i32, i32* %lsr.iv11, i32 4
vector.body:
   %lsr.iv1416 = bitcast i32* %lsr.iv14 to <4 x i32>*
   %lsr.iv1113 = bitcast i32* %lsr.iv11 to <4 x i32>*
   %lsr.iv10 = bitcast i32* %lsr.iv to <4 x i32>*
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
 
 ; The induction variable %D is not an IV:
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %N, i32 32002)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %N, i32 32003)
 
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv10, i32 4, <4 x i1> %1, <4 x i32> undef)
   %wide.masked.load9 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv1113, i32 4, <4 x i1> %1, <4 x i32> undef)
   %2 = add nsw <4 x i32> %wide.masked.load9, %wide.masked.load
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %2, <4 x i32>* %lsr.iv1416, i32 4, <4 x i1> %1)
   %index.next = add i32 %index, 4
   %scevgep = getelementptr i32, i32* %lsr.iv, i32 4
   %scevgep12 = getelementptr i32, i32* %lsr.iv11, i32 4
vector.body:
   %0 = phi i32 [ 8001, %entry ], [ %3, %vector.body ]
   %lsr.iv1416 = bitcast i32* %lsr.iv14 to <4 x i32>*
   %lsr.iv1113 = bitcast i32* %lsr.iv11 to <4 x i32>*
   %lsr.iv10 = bitcast i32* %lsr.iv to <4 x i32>*
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
 
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32002)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32003)
 
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv10, i32 4, <4 x i1> %1, <4 x i32> undef)
   %wide.masked.load9 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv1113, i32 4, <4 x i1> %1, <4 x i32> undef)
   %2 = add nsw <4 x i32> %wide.masked.load9, %wide.masked.load
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %2, <4 x i32>* %lsr.iv1416, i32 4, <4 x i1> %1)
 
 ; %index is incremented with 3 and not 4, which is the vectorisation factor
 ; that we expect here:
vector.body:
   %index = phi i32 [ 0, %entry ], [ %index.next, %vector.body ]
   %0 = phi i32 [ 8001, %entry ], [ %3, %vector.body ]
   %lsr.iv1416 = bitcast i32* %lsr.iv14 to <4 x i32>*
   %lsr.iv1113 = bitcast i32* %lsr.iv11 to <4 x i32>*
   %lsr.iv10 = bitcast i32* %lsr.iv to <4 x i32>*
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32002)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 32003)
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv10, i32 4, <4 x i1> %1, <4 x i32> undef)
   %wide.masked.load9 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv1113, i32 4, <4 x i1> %1, <4 x i32> undef)
   %2 = add nsw <4 x i32> %wide.masked.load9, %wide.masked.load
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %2, <4 x i32>* %lsr.iv1416, i32 4, <4 x i1> %1)
 
 ; %index is incremented with some runtime value, i.e. not a constant:
   %index.next = add i32 %index, %N
 

llvm/test/CodeGen/Thumb2/LowOverheadLoops/tail-pred-intrinsic-add-sat.ll

vector.ph:                                        ; preds = %entry
   %trip.count.minus.1 = add i32 %blockSize, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr i16, i16* %pSrcA, i32 %index
   %next.gep20 = getelementptr i16, i16* %pDst, i32 %index
   %next.gep21 = getelementptr i16, i16* %pSrcB, i32 %index
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %blockSize)
   %0 = bitcast i16* %next.gep to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %0, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %1 = bitcast i16* %next.gep21 to <8 x i16>*
   %wide.masked.load24 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %1, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %2 = call <8 x i16> @llvm.uadd.sat.v8i16(<8 x i16> %wide.masked.load, <8 x i16> %wide.masked.load24)
   %3 = bitcast i16* %next.gep20 to <8 x i16>*
   call void @llvm.masked.store.v8i16.p0v8i16(<8 x i16> %2, <8 x i16>* %3, i32 2, <8 x i1> %active.lane.mask)
   %index.next = add i32 %index, 8
vector.ph:                                        ; preds = %entry
   %trip.count.minus.1 = add i32 %blockSize, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr i16, i16* %pSrcA, i32 %index
   %next.gep20 = getelementptr i16, i16* %pDst, i32 %index
   %next.gep21 = getelementptr i16, i16* %pSrcB, i32 %index
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %blockSize)
   %0 = bitcast i16* %next.gep to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %0, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %1 = bitcast i16* %next.gep21 to <8 x i16>*
   %wide.masked.load24 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %1, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %2 = call <8 x i16> @llvm.sadd.sat.v8i16(<8 x i16> %wide.masked.load, <8 x i16> %wide.masked.load24)
   %3 = bitcast i16* %next.gep20 to <8 x i16>*
   call void @llvm.masked.store.v8i16.p0v8i16(<8 x i16> %2, <8 x i16>* %3, i32 2, <8 x i1> %active.lane.mask)
   %index.next = add i32 %index, 8

llvm/test/CodeGen/Thumb2/LowOverheadLoops/tail-pred-intrinsic-fabs.ll

vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %blockSize, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr float, float* %pDst, i32 %index
   %next.gep13 = getelementptr float, float* %pSrcA, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %blockSize)
   %0 = bitcast float* %next.gep13 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %0, i32 4, <4 x i1> %active.lane.mask, <4 x float> undef)
   %1 = call fast <4 x float> @llvm.fabs.v4f32(<4 x float> %wide.masked.load)
   %2 = bitcast float* %next.gep to <4 x float>*
   call void @llvm.masked.store.v4f32.p0v4f32(<4 x float> %1, <4 x float>* %2, i32 4, <4 x i1> %active.lane.mask)
   %index.next = add i32 %index, 4
   %3 = icmp eq i32 %index.next, %n.vec
   br i1 %3, label %while.end, label %vector.body

llvm/test/CodeGen/Thumb2/LowOverheadLoops/tail-pred-intrinsic-round.ll

vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr float, float* %pSrcA, i32 %index
   %next.gep14 = getelementptr float, float* %pDst, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = bitcast float* %next.gep to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %0, i32 4, <4 x i1> %active.lane.mask, <4 x float> undef)
   %1 = call fast <4 x float> @llvm.round.v4f32(<4 x float> %wide.masked.load)
   %2 = bitcast float* %next.gep14 to <4 x float>*
   call void @llvm.masked.store.v4f32.p0v4f32(<4 x float> %1, <4 x float>* %2, i32 4, <4 x i1> %active.lane.mask)
   %index.next = add i32 %index, 4
   %3 = icmp eq i32 %index.next, %n.vec
   br i1 %3, label %for.cond.cleanup, label %vector.body
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr float, float* %pSrcA, i32 %index
   %next.gep14 = getelementptr float, float* %pDst, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = bitcast float* %next.gep to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %0, i32 4, <4 x i1> %active.lane.mask, <4 x float> undef)
   %1 = call fast <4 x float> @llvm.rint.v4f32(<4 x float> %wide.masked.load)
   %2 = bitcast float* %next.gep14 to <4 x float>*
   call void @llvm.masked.store.v4f32.p0v4f32(<4 x float> %1, <4 x float>* %2, i32 4, <4 x i1> %active.lane.mask)
   %index.next = add i32 %index, 4
   %3 = icmp eq i32 %index.next, %n.vec
   br i1 %3, label %for.cond.cleanup, label %vector.body
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr float, float* %pSrcA, i32 %index
   %next.gep14 = getelementptr float, float* %pDst, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = bitcast float* %next.gep to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %0, i32 4, <4 x i1> %active.lane.mask, <4 x float> undef)
   %1 = call fast <4 x float> @llvm.trunc.v4f32(<4 x float> %wide.masked.load)
   %2 = bitcast float* %next.gep14 to <4 x float>*
   call void @llvm.masked.store.v4f32.p0v4f32(<4 x float> %1, <4 x float>* %2, i32 4, <4 x i1> %active.lane.mask)
   %index.next = add i32 %index, 4
   %3 = icmp eq i32 %index.next, %n.vec
   br i1 %3, label %for.cond.cleanup, label %vector.body
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr float, float* %pSrcA, i32 %index
   %next.gep14 = getelementptr float, float* %pDst, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = bitcast float* %next.gep to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %0, i32 4, <4 x i1> %active.lane.mask, <4 x float> undef)
   %1 = call fast <4 x float> @llvm.ceil.v4f32(<4 x float> %wide.masked.load)
   %2 = bitcast float* %next.gep14 to <4 x float>*
   call void @llvm.masked.store.v4f32.p0v4f32(<4 x float> %1, <4 x float>* %2, i32 4, <4 x i1> %active.lane.mask)
   %index.next = add i32 %index, 4
   %3 = icmp eq i32 %index.next, %n.vec
   br i1 %3, label %for.cond.cleanup, label %vector.body
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr float, float* %pSrcA, i32 %index
   %next.gep14 = getelementptr float, float* %pDst, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = bitcast float* %next.gep to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %0, i32 4, <4 x i1> %active.lane.mask, <4 x float> undef)
   %1 = call fast <4 x float> @llvm.floor.v4f32(<4 x float> %wide.masked.load)
   %2 = bitcast float* %next.gep14 to <4 x float>*
   call void @llvm.masked.store.v4f32.p0v4f32(<4 x float> %1, <4 x float>* %2, i32 4, <4 x i1> %active.lane.mask)
   %index.next = add i32 %index, 4
   %3 = icmp eq i32 %index.next, %n.vec
   br i1 %3, label %for.cond.cleanup, label %vector.body
 
 for.cond.cleanup:                                 ; preds = %vector.body, %entry
   ret void
 }
 
 ; nearbyint shouldn't be tail predicated because it's lowered into multiple instructions
 define arm_aapcs_vfpcc void @nearbyint(float* noalias nocapture readonly %pSrcA, float* noalias nocapture %pDst, i32 %n) #0 {
 ; CHECK-LABEL: nearbyint:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    .save {r7, lr}
 ; CHECK-NEXT:    push {r7, lr}
 ; CHECK-NEXT:    cmp r2, #0
 ; CHECK-NEXT:    it eq
 ; CHECK-NEXT:    popeq {r7, pc}
 ; CHECK-NEXT:    adds r3, r2, #3
+; CHECK-NEXT:    vdup.32 q1, r2
 ; CHECK-NEXT:    bic r3, r3, #3
 ; CHECK-NEXT:    sub.w r12, r3, #4
 ; CHECK-NEXT:    movs r3, #1
 ; CHECK-NEXT:    add.w lr, r3, r12, lsr #2
 ; CHECK-NEXT:    adr r3, .LCPI5_0
-; CHECK-NEXT:    sub.w r12, r2, #1
 ; CHECK-NEXT:    vldrw.u32 q0, [r3]
-; CHECK-NEXT:    movs r2, #0
-; CHECK-NEXT:    vdup.32 q1, r12
+; CHECK-NEXT:    mov.w r12, #0
 ; CHECK-NEXT:    dls lr, lr
 ; CHECK-NEXT:  .LBB5_1: @ %vector.body
 ; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    vadd.i32 q2, q0, r2
-; CHECK-NEXT:    vdup.32 q3, r2
+; CHECK-NEXT:    vadd.i32 q2, q0, r12
+; CHECK-NEXT:    vdup.32 q3, r12
 ; CHECK-NEXT:    vcmp.u32 hi, q3, q2
-; CHECK-NEXT:    adds r2, #4
+; CHECK-NEXT:    add.w r12, r12, #4
 ; CHECK-NEXT:    vpnot
 ; CHECK-NEXT:    vpstt
 ; CHECK-NEXT:    vcmpt.u32 hi, q1, q2
 ; CHECK-NEXT:    vldrwt.u32 q2, [r0], #16
 ; CHECK-NEXT:    vrintr.f32 s15, s11
 ; CHECK-NEXT:    vrintr.f32 s14, s10
 ; CHECK-NEXT:    vrintr.f32 s13, s9
 ; CHECK-NEXT:    vrintr.f32 s12, s8
vector.ph:                                        ; preds = %entry
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %next.gep = getelementptr float, float* %pSrcA, i32 %index
   %next.gep14 = getelementptr float, float* %pDst, i32 %index
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = bitcast float* %next.gep to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %0, i32 4, <4 x i1> %active.lane.mask, <4 x float> undef)
   %1 = call fast <4 x float> @llvm.nearbyint.v4f32(<4 x float> %wide.masked.load)
   %2 = bitcast float* %next.gep14 to <4 x float>*
   call void @llvm.masked.store.v4f32.p0v4f32(<4 x float> %1, <4 x float>* %2, i32 4, <4 x i1> %active.lane.mask)
   %index.next = add i32 %index, 4
   %3 = icmp eq i32 %index.next, %n.vec
   br i1 %3, label %for.cond.cleanup, label %vector.body

llvm/test/CodeGen/Thumb2/LowOverheadLoops/tail-pred-intrinsic-sub-sat.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve -verify-machineinstrs -tail-predication=enabled -o - %s | FileCheck %s
 
 define arm_aapcs_vfpcc void @usub_sat(i16* noalias nocapture readonly %pSrcA, i16* noalias nocapture readonly %pSrcB, i16* noalias nocapture %pDst, i32 %blockSize) {
 ; CHECK-LABEL: usub_sat:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    .save {r7, lr}
 ; CHECK-NEXT:    push {r7, lr}
 ; CHECK-NEXT:    cmp r3, #0
 ; CHECK-NEXT:    it eq
 ; CHECK-NEXT:    popeq {r7, pc}
+; CHECK-NEXT:    subs r3, #1
 ; CHECK-NEXT:    dlstp.16 lr, r3
 ; CHECK-NEXT:  .LBB0_1: @ %vector.body
 ; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    vldrh.u16 q0, [r1], #16
 ; CHECK-NEXT:    vldrh.u16 q1, [r0], #16
 ; CHECK-NEXT:    vqsub.u16 q0, q1, q0
 ; CHECK-NEXT:    vstrh.16 q0, [r2], #16
 ; CHECK-NEXT:    letp lr, .LBB0_1

 define arm_aapcs_vfpcc void @ssub_sat(i16* noalias nocapture readonly %pSrcA, i16* noalias nocapture readonly %pSrcB, i16* noalias nocapture %pDst, i32 %blockSize) {
 ; CHECK-LABEL: ssub_sat:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    .save {r7, lr}
 ; CHECK-NEXT:    push {r7, lr}
 ; CHECK-NEXT:    cmp r3, #0
 ; CHECK-NEXT:    it eq
 ; CHECK-NEXT:    popeq {r7, pc}
+; CHECK-NEXT:    subs r3, #1
 ; CHECK-NEXT:    dlstp.16 lr, r3
 ; CHECK-NEXT:  .LBB1_1: @ %vector.body
 ; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    vldrh.u16 q0, [r1], #16
 ; CHECK-NEXT:    vldrh.u16 q1, [r0], #16
 ; CHECK-NEXT:    vqsub.s16 q0, q1, q0
 ; CHECK-NEXT:    vstrh.16 q0, [r2], #16
 ; CHECK-NEXT:    letp lr, .LBB1_1

llvm/test/CodeGen/Thumb2/LowOverheadLoops/tail-pred-widen.ll

vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %tmp = getelementptr inbounds i16, i16* %a, i32 %index
 
   ; %tmp1 = icmp ule <8 x i32> %induction, %broadcast.splat11
-   %tmp1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+   %tmp1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i16* %tmp to <8 x i16>*
   %wide.masked.load = tail call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp2, i32 4, <8 x i1> %tmp1, <8 x i16> undef)
   %tmp3 = getelementptr inbounds i16, i16* %b, i32 %index
   %tmp4 = bitcast i16* %tmp3 to <8 x i16>*
   %wide.masked.load2 = tail call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp4, i32 4, <8 x i1> %tmp1, <8 x i16> undef)
   %expand.1 = zext <8 x i16> %wide.masked.load to <8 x i32>
   %expand.2 = zext <8 x i16> %wide.masked.load2 to <8 x i32>
vector.body:                                      ; preds = %vector.body, %vector.ph
   %store.idx = phi i32 [ 0, %vector.ph ], [ %store.idx.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %tmp = getelementptr inbounds i16, i16* %a, i32 %index
 
   ; %tmp1 = icmp ule <8 x i32> %induction, %broadcast.splat11
-  %tmp1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i16* %tmp to <8 x i16>*
   %wide.masked.load = tail call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp2, i32 4, <8 x i1> %tmp1, <8 x i16> undef)
   %tmp3 = getelementptr inbounds i16, i16* %b, i32 %index
   %tmp4 = bitcast i16* %tmp3 to <8 x i16>*
   %wide.masked.load2 = tail call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp4, i32 4, <8 x i1> %tmp1, <8 x i16> undef)
   %extract.2.low = shufflevector <8 x i16> %wide.masked.load2, <8 x i16> undef, < 4 x i32> <i32 0, i32 1, i32 2, i32 3>
   %extract.2.high = shufflevector <8 x i16> %wide.masked.load2, <8 x i16> undef, < 4 x i32> <i32 4, i32 5, i32 6, i32 7>
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %tmp14 = phi i32 [ %tmp13, %vector.ph ], [ %tmp15, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %tmp = getelementptr inbounds i32, i32* %a, i32 %index
 
   ; %tmp1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %tmp1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %tmp2 = bitcast i32* %tmp to <4 x i32>*
   %wide.masked.load = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp2, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %tmp3 = getelementptr inbounds i32, i32* %b, i32 %index
   %tmp4 = bitcast i32* %tmp3 to <4 x i32>*
   %wide.masked.load2 = tail call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %tmp4, i32 4, <4 x i1> %tmp1, <4 x i32> undef)
   %expand.1 = zext <4 x i32> %wide.masked.load to <4 x i64>
   %expand.2 = zext <4 x i32> %wide.masked.load2 to <4 x i64>

llvm/test/CodeGen/Thumb2/LowOverheadLoops/tail-reduce.ll

vector.body:                                      ; preds = %vector.body, %vector.ph
   %vec.phi = phi <8 x i16> [ zeroinitializer, %vector.ph], [ %tmp8, %vector.body ]
   %3 = phi i32 [ %2, %vector.ph], [ %4, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %tmp2 = getelementptr inbounds i16, i16* %A, i32 %index
 
   ; %tmp3 = icmp ule <8 x i32> %induction, %broadcast.splat2
-  %tmp3 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %tmp)
+  %tmp3 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %tmp4 = bitcast i16* %tmp2 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp4, i32 4, <8 x i1> %tmp3, <8 x i16> undef)
   %tmp5 = getelementptr inbounds i16, i16* %B, i32 %index
   %tmp6 = bitcast i16* %tmp5 to <8 x i16>*
   %wide.masked.load3 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp6, i32 4, <8 x i1> %tmp3, <8 x i16> undef)
   %tmp7 = add <8 x i16> %wide.masked.load, %vec.phi
   %tmp8 = add <8 x i16> %tmp7, %wide.masked.load3
vector.body:                                      ; preds = %vector.body, %vector.ph
   %vec.phi = phi <8 x i16> [ zeroinitializer, %vector.ph], [ %tmp6, %vector.body ]
   %3 = phi i32 [ %2, %vector.ph], [ %4, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %tmp2 = getelementptr inbounds i16, i16* %A, i32 %index
 
   ; %tmp3 = icmp ule <8 x i32> %induction, %broadcast.splat2
-  %tmp3 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %tmp)
+  %tmp3 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %tmp4 = bitcast i16* %tmp2 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp4, i32 4, <8 x i1> %tmp3, <8 x i16> undef)
   %tmp5 = add <8 x i16> %vec.phi, %broadcast.splat4
   %tmp6 = add <8 x i16> %tmp5, %wide.masked.load
   %index.next = add nuw nsw i32 %index, 8
   %4 = call i32 @llvm.loop.decrement.reg.i32.i32.i32(i32 %3, i32 1)
   %5 = icmp ne i32 %4, 0
vector.body:                                      ; preds = %vector.body, %vector.ph
   %vec.phi = phi <8 x i16> [ zeroinitializer, %entry], [ %tmp6, %vector.body ]
   %3 = phi i32 [ %2, %entry ], [ %4, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %tmp2 = getelementptr inbounds i16, i16* %A, i32 %index
 
   ; %tmp3 = icmp ule <8 x i32> %induction, %broadcast.splat2
-  %tmp3 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %tmp)
+  %tmp3 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %tmp4 = bitcast i16* %tmp2 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %tmp4, i32 4, <8 x i1> %tmp3, <8 x i16> undef)
   %tmp5 = add <8 x i16> %vec.phi, %broadcast.splat4
   %tmp6 = add <8 x i16> %tmp5, %wide.masked.load
   %index.next = add nuw nsw i32 %index, 8
   %4 = call i32 @llvm.loop.decrement.reg.i32.i32.i32(i32 %3, i32 1)
   %5 = icmp ne i32 %4, 0

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %lsr.iv48 = phi i16* [ %scevgep49, %vector.body ], [ %lsr.iv46, %vector.ph ]
   %lsr.iv = phi i16* [ %scevgep, %vector.body ], [ %Input, %vector.ph ]
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %16, %vector.body ]
   %9 = phi i32 [ %7, %vector.ph ], [ %17, %vector.body ]
   %lsr.iv4850 = bitcast i16* %lsr.iv48 to <4 x i16>*
   %lsr.iv45 = bitcast i16* %lsr.iv to <4 x i16>*
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %8)
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %lsr.iv45, i32 2, <4 x i1> %active.lane.mask, <4 x i16> undef)
   %10 = sext <4 x i16> %wide.masked.load to <4 x i32>
   %wide.masked.load42 = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %lsr.iv4850, i32 2, <4 x i1> %active.lane.mask, <4 x i16> undef)
   %11 = sext <4 x i16> %wide.masked.load42 to <4 x i32>
   %12 = mul nsw <4 x i32> %11, %10
   %13 = insertelement <4 x i32> undef, i32 %conv1032, i32 0
   %14 = shufflevector <4 x i32> %13, <4 x i32> undef, <4 x i32> zeroinitializer
   %15 = ashr <4 x i32> %12, %14

llvm/test/CodeGen/Thumb2/LowOverheadLoops/varying-outer-2d-reduction.ll

 ; NOREDUCTIONS-NEXT:    letp lr, .LBB0_6
 ; NOREDUCTIONS-NEXT:  @ %bb.7: @ %middle.block
 ; NOREDUCTIONS-NEXT:    @ in Loop: Header=BB0_4 Depth=1
 ; NOREDUCTIONS-NEXT:    vaddv.u32 r0, q0
 ; NOREDUCTIONS-NEXT:    b .LBB0_3
 ; NOREDUCTIONS-NEXT:  .LBB0_8: @ %for.end17
 ; NOREDUCTIONS-NEXT:    add sp, #4
 ; NOREDUCTIONS-NEXT:    pop.w {r4, r5, r6, r7, r8, r9, r10, pc}
-;
 entry:
   %conv = sext i16 %N to i32
   %cmp36 = icmp sgt i16 %N, 0
   br i1 %cmp36, label %for.body.lr.ph, label %for.end17
 
 for.body.lr.ph:                                   ; preds = %entry
   %conv2 = sext i16 %Size to i32
   %conv1032 = zext i16 %Scale to i32

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %lsr.iv48 = phi i16* [ %scevgep49, %vector.body ], [ %lsr.iv46, %vector.ph ]
   %lsr.iv = phi i16* [ %scevgep, %vector.body ], [ %Input, %vector.ph ]
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %i16, %vector.body ]
   %i9 = phi i32 [ %i7, %vector.ph ], [ %i17, %vector.body ]
   %lsr.iv4850 = bitcast i16* %lsr.iv48 to <4 x i16>*
   %lsr.iv45 = bitcast i16* %lsr.iv to <4 x i16>*
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %i8)
   %wide.masked.load = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %lsr.iv45, i32 2, <4 x i1> %active.lane.mask, <4 x i16> undef)
   %i10 = sext <4 x i16> %wide.masked.load to <4 x i32>
   %wide.masked.load42 = call <4 x i16> @llvm.masked.load.v4i16.p0v4i16(<4 x i16>* %lsr.iv4850, i32 2, <4 x i1> %active.lane.mask, <4 x i16> undef)
   %i11 = sext <4 x i16> %wide.masked.load42 to <4 x i32>
   %i12 = mul nsw <4 x i32> %i11, %i10
   %i13 = insertelement <4 x i32> undef, i32 %conv1032, i32 0
   %i14 = shufflevector <4 x i32> %i13, <4 x i32> undef, <4 x i32> zeroinitializer
   %i15 = ashr <4 x i32> %i12, %i14

llvm/test/CodeGen/Thumb2/LowOverheadLoops/vector-arith-codegen.ll

vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %6, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i32, i32* %a, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat12
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %2, i32 4, <4 x i1> %1, <4 x i32> undef)
   %3 = getelementptr inbounds i32, i32* %b, i32 %index
   %4 = bitcast i32* %3 to <4 x i32>*
   %wide.masked.load13 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %4, i32 4, <4 x i1> %1, <4 x i32> undef)
   %5 = mul nsw <4 x i32> %wide.masked.load13, %wide.masked.load
   %6 = add nsw <4 x i32> %5, %vec.phi
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %3, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i32, i32* %a, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat10
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %2, i32 4, <4 x i1> %1, <4 x i32> undef)
   %3 = add nsw <4 x i32> %wide.masked.load, %vec.phi
   %index.next = add i32 %index, 4
   %4 = icmp eq i32 %index.next, %n.vec
   br i1 %4, label %middle.block, label %vector.body
 
vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.ph ], [ %3, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i32, i32* %a, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat10
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %2, i32 4, <4 x i1> %1, <4 x i32> undef)
   %3 = add nsw <4 x i32> %wide.masked.load, %vec.phi
   %index.next = add i32 %index, 4
   %4 = icmp eq i32 %index.next, %n.vec
   br i1 %4, label %middle.block, label %vector.body
 

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i32, i32* %b, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat9
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %2, i32 4, <4 x i1> %1, <4 x i32> undef)
   %3 = mul nsw <4 x i32> %wide.masked.load, %broadcast.splat11
   %4 = getelementptr inbounds i32, i32* %a, i32 %index
   %5 = bitcast i32* %4 to <4 x i32>*
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %3, <4 x i32>* %5, i32 4, <4 x i1> %1)
   %index.next = add i32 %index, 4

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds i32, i32* %b, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat9
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %2, i32 4, <4 x i1> %1, <4 x i32> undef)
   %3 = add nsw <4 x i32> %wide.masked.load, %broadcast.splat11
   %4 = getelementptr inbounds i32, i32* %a, i32 %index
   %5 = bitcast i32* %4 to <4 x i32>*
   call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %3, <4 x i32>* %5, i32 4, <4 x i1> %1)
   %index.next = add i32 %index, 4

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <16 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <16 x i32> %broadcast.splatinsert, <16 x i32> undef, <16 x i32> zeroinitializer
   %induction = add <16 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
   %0 = getelementptr inbounds i8, i8* %b, i32 %index
 
   ; %1 = icmp ule <16 x i32> %induction, %broadcast.splat13
-  %1 = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i8* %0 to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %2, i32 1, <16 x i1> %1, <16 x i8> undef)
   %3 = getelementptr inbounds i8, i8* %c, i32 %index
   %4 = bitcast i8* %3 to <16 x i8>*
   %wide.masked.load14 = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %4, i32 1, <16 x i1> %1, <16 x i8> undef)
   %5 = mul <16 x i8> %wide.masked.load14, %wide.masked.load
   %6 = getelementptr inbounds i8, i8* %a, i32 %index

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <8 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <8 x i32> %broadcast.splatinsert, <8 x i32> undef, <8 x i32> zeroinitializer
   %induction = add <8 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   %0 = getelementptr inbounds i16, i16* %b, i32 %index
 
   ; %1 = icmp ule <8 x i32> %induction, %broadcast.splat13
-  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %N)
 
   %2 = bitcast i16* %0 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %2, i32 2, <8 x i1> %1, <8 x i16> undef)
   %3 = getelementptr inbounds i16, i16* %c, i32 %index
   %4 = bitcast i16* %3 to <8 x i16>*
   %wide.masked.load14 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %4, i32 2, <8 x i1> %1, <8 x i16> undef)
   %5 = mul <8 x i16> %wide.masked.load14, %wide.masked.load
   %6 = getelementptr inbounds i16, i16* %a, i32 %index

llvm/test/CodeGen/Thumb2/LowOverheadLoops/vector-reduce-mve-tail.ll

vector.body:                                      ; preds = %vector.body, %vector.ph
   %6 = phi i32 [ %5, %vector.ph ], [ %10, %vector.body ]
   %lsr.iv24 = bitcast i32* %lsr.iv2 to <4 x i32>*
   %lsr.iv1 = bitcast i32* %lsr.iv to <4 x i32>*
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = add <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
 
   ; %7 = icmp ule <4 x i32> %induction, %broadcast.splat12
-  %7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %7 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %N)
 
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv24, i32 4, <4 x i1> %7, <4 x i32> undef)
   %wide.masked.load13 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %lsr.iv1, i32 4, <4 x i1> %7, <4 x i32> undef)
   %8 = mul nsw <4 x i32> %wide.masked.load13, %wide.masked.load
   %9 = add nsw <4 x i32> %8, %vec.phi
   %index.next = add i32 %index, 4
   %scevgep = getelementptr i32, i32* %lsr.iv, i32 4
   %scevgep3 = getelementptr i32, i32* %lsr.iv2, i32 4

llvm/test/CodeGen/Thumb2/mve-fma-loops.ll

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %wide.masked.load, <4 x float> %wide.masked.load12, <4 x float> %broadcast.splat14)
   %6 = getelementptr inbounds float, float* %z, i32 %index

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = fmul fast <4 x float> %wide.masked.load12, %wide.masked.load
   %6 = fadd fast <4 x float> %5, %broadcast.splat14

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %wide.masked.load, <4 x float> %broadcast.splat14, <4 x float> %wide.masked.load12)
   %6 = getelementptr inbounds float, float* %z, i32 %index

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = fmul fast <4 x float> %wide.masked.load, %broadcast.splat13
   %4 = getelementptr inbounds float, float* %y, i32 %index
   %5 = bitcast float* %4 to <4 x float>*
   %wide.masked.load14 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %5, i32 4, <4 x i1> %1, <4 x float> undef)
   %6 = fadd fast <4 x float> %3, %wide.masked.load14

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %wide.masked.load, <4 x float> %wide.masked.load12, <4 x float> %broadcast.splat14)
   %6 = getelementptr inbounds float, float* %z, i32 %index

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = fmul fast <4 x float> %wide.masked.load12, %wide.masked.load
   %6 = fsub fast <4 x float> %5, %broadcast.splat14

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = fneg fast <4 x float> %wide.masked.load12
   %6 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %wide.masked.load, <4 x float> %5, <4 x float> %broadcast.splat14)

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = fmul fast <4 x float> %wide.masked.load12, %wide.masked.load
   %6 = fsub fast <4 x float> %broadcast.splat14, %5

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %wide.masked.load, <4 x float> %broadcast.splat14, <4 x float> %wide.masked.load12)
   %6 = getelementptr inbounds float, float* %z, i32 %index

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = fmul fast <4 x float> %wide.masked.load, %broadcast.splat14
   %6 = fsub fast <4 x float> %wide.masked.load12, %5

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = getelementptr inbounds float, float* %y, i32 %index
   %4 = bitcast float* %3 to <4 x float>*
   %wide.masked.load12 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %4, i32 4, <4 x i1> %1, <4 x float> undef)
   %5 = fneg fast <4 x float> %wide.masked.load12
   %6 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %wide.masked.load, <4 x float> %broadcast.splat14, <4 x float> %5)

 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %broadcast.splatinsert = insertelement <4 x i32> undef, i32 %index, i32 0
   %broadcast.splat = shufflevector <4 x i32> %broadcast.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
   %induction = or <4 x i32> %broadcast.splat, <i32 0, i32 1, i32 2, i32 3>
   %0 = getelementptr inbounds float, float* %x, i32 %index
 
   ; %1 = icmp ule <4 x i32> %induction, %broadcast.splat11
-  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %1 = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
 
   %2 = bitcast float* %0 to <4 x float>*
   %wide.masked.load = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %2, i32 4, <4 x i1> %1, <4 x float> undef)
   %3 = fmul fast <4 x float> %wide.masked.load, %broadcast.splat13
   %4 = getelementptr inbounds float, float* %y, i32 %index
   %5 = bitcast float* %4 to <4 x float>*
   %wide.masked.load14 = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* %5, i32 4, <4 x i1> %1, <4 x float> undef)
   %6 = fsub fast <4 x float> %3, %wide.masked.load14

llvm/test/CodeGen/Thumb2/mve-gather-scatter-tailpred.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve.fp -enable-arm-maskedldst -enable-mem-access-versioning=false -enable-arm-maskedgatscat -tail-predication=force-enabled %s -o - | FileCheck %s
 
 define dso_local void @mve_gather_qi_wb(i32* noalias nocapture readonly %A, i32* noalias nocapture readonly %B, i32* noalias nocapture %C, i32 %n, i32 %m, i32 %l) {
 ; CHECK-LABEL: mve_gather_qi_wb:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    .save {r4, lr}
-; CHECK-NEXT:    push {r4, lr}
-; CHECK-NEXT:    adr r4, .LCPI0_0
+; CHECK-NEXT:    .save {r7, lr}
+; CHECK-NEXT:    push {r7, lr}
 ; CHECK-NEXT:    add.w r12, r0, r3, lsl #2
-; CHECK-NEXT:    vldrw.u32 q0, [r4]
-; CHECK-NEXT:    adds r0, r3, #1
+; CHECK-NEXT:    adr r0, .LCPI0_0
+; CHECK-NEXT:    vldrw.u32 q0, [r0]
 ; CHECK-NEXT:    vmov.i32 q2, #0x0
+; CHECK-NEXT:    dlstp.32 lr, r3
 ; CHECK-NEXT:    vadd.i32 q0, q0, r1
 ; CHECK-NEXT:    adds r1, r3, #4
-; CHECK-NEXT:    dlstp.32 lr, r0
 ; CHECK-NEXT:  .LBB0_1: @ %vector.body
 ; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    vldrw.u32 q1, [r12], #16
 ; CHECK-NEXT:    vldrw.u32 q3, [q0, #80]!
 ; CHECK-NEXT:    vmul.i32 q1, q3, q1
 ; CHECK-NEXT:    vadd.i32 q2, q2, q1
 ; CHECK-NEXT:    letp lr, .LBB0_1
 ; CHECK-NEXT:  @ %bb.2: @ %middle.block
 ; CHECK-NEXT:    vmov q0, q2
 ; CHECK-NEXT:    vaddv.u32 r0, q0
 ; CHECK-NEXT:    str.w r0, [r2, r1, lsl #2]
-; CHECK-NEXT:    pop {r4, pc}
+; CHECK-NEXT:    pop {r7, pc}
 ; CHECK-NEXT:    .p2align 4
 ; CHECK-NEXT:  @ %bb.3:
 ; CHECK-NEXT:  .LCPI0_0:
 ; CHECK-NEXT:    .long 4294967228 @ 0xffffffbc
 ; CHECK-NEXT:    .long 4294967248 @ 0xffffffd0
 ; CHECK-NEXT:    .long 4294967268 @ 0xffffffe4
 ; CHECK-NEXT:    .long 4294967288 @ 0xfffffff8
 entry:                                  ; preds = %middle.

 
 define dso_local void @mve_gatherscatter_offset(i32* noalias nocapture readonly %A, i32* noalias nocapture readonly %B, i32* noalias nocapture %C, i32 %n, i32 %m, i32 %l) {
 ; CHECK-LABEL: mve_gatherscatter_offset:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    .save {r4, lr}
 ; CHECK-NEXT:    push {r4, lr}
 ; CHECK-NEXT:    .vsave {d8, d9}
 ; CHECK-NEXT:    vpush {d8, d9}
-; CHECK-NEXT:    adr r4, .LCPI1_0
-; CHECK-NEXT:    add.w r12, r0, r3, lsl #2
-; CHECK-NEXT:    adds r0, r3, #1
-; CHECK-NEXT:    vldrw.u32 q1, [r4]
-; CHECK-NEXT:    adds r3, #4
+; CHECK-NEXT:    add.w r4, r0, r3, lsl #2
+; CHECK-NEXT:    adr r0, .LCPI1_0
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    add.w r12, r3, #4
 ; CHECK-NEXT:    vmov.i32 q3, #0x0
 ; CHECK-NEXT:    vmov.i32 q0, #0x14
-; CHECK-NEXT:    dlstp.32 lr, r0
+; CHECK-NEXT:    dlstp.32 lr, r3
 ; CHECK-NEXT:  .LBB1_1: @ %vector.body
 ; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    vldrw.u32 q2, [r1, q1, uxtw #2]
-; CHECK-NEXT:    vldrw.u32 q4, [r12], #16
+; CHECK-NEXT:    vldrw.u32 q4, [r4], #16
 ; CHECK-NEXT:    vmul.i32 q2, q2, q4
 ; CHECK-NEXT:    vstrw.32 q2, [r1, q1, uxtw #2]
 ; CHECK-NEXT:    vadd.i32 q1, q1, q0
 ; CHECK-NEXT:    vadd.i32 q3, q3, q2
 ; CHECK-NEXT:    letp lr, .LBB1_1
 ; CHECK-NEXT:  @ %bb.2: @ %middle.block
 ; CHECK-NEXT:    vmov q0, q3
 ; CHECK-NEXT:    vaddv.u32 r0, q0
-; CHECK-NEXT:    str.w r0, [r2, r3, lsl #2]
+; CHECK-NEXT:    str.w r0, [r2, r12, lsl #2]
 ; CHECK-NEXT:    vpop {d8, d9}
 ; CHECK-NEXT:    pop {r4, pc}
 ; CHECK-NEXT:    .p2align 4
 ; CHECK-NEXT:  @ %bb.3:
 ; CHECK-NEXT:  .LCPI1_0:
 ; CHECK-NEXT:    .long 3 @ 0x3
 ; CHECK-NEXT:    .long 8 @ 0x8
 ; CHECK-NEXT:    .long 13 @ 0xd
middle.block:                                     ; preds = %vector.body
   %exitcond81.not = icmp eq i32 %inc21.us.us, %n
   br label %end
 end:                                 ; preds = %middle.block
   ret void
 }
 define dso_local void @mve_scatter_qi(i32* noalias nocapture readonly %A, i32* noalias nocapture readonly %B, i32* noalias nocapture %C, i32 %n, i32 %m, i32 %l) {
 ; CHECK-LABEL: mve_scatter_qi:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    .save {r4, lr}
-; CHECK-NEXT:    push {r4, lr}
-; CHECK-NEXT:    adr r4, .LCPI2_0
+; CHECK-NEXT:    .save {r7, lr}
+; CHECK-NEXT:    push {r7, lr}
 ; CHECK-NEXT:    add.w r12, r0, r3, lsl #2
-; CHECK-NEXT:    vldrw.u32 q0, [r4]
-; CHECK-NEXT:    adds r0, r3, #1
+; CHECK-NEXT:    adr r0, .LCPI2_0
+; CHECK-NEXT:    vldrw.u32 q0, [r0]
 ; CHECK-NEXT:    vmov.i32 q3, #0x0
+; CHECK-NEXT:    vmov.i32 q2, #0x3
 ; CHECK-NEXT:    vadd.i32 q0, q0, r1
 ; CHECK-NEXT:    adds r1, r3, #4
-; CHECK-NEXT:    vmov.i32 q2, #0x3
-; CHECK-NEXT:    dlstp.32 lr, r0
+; CHECK-NEXT:    dlstp.32 lr, r3
 ; CHECK-NEXT:  .LBB2_1: @ %vector.body
 ; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    vldrw.u32 q1, [r12], #16
 ; CHECK-NEXT:    vmul.i32 q1, q1, q2
 ; CHECK-NEXT:    vstrw.32 q1, [q0, #80]!
 ; CHECK-NEXT:    vadd.i32 q3, q3, q1
 ; CHECK-NEXT:    letp lr, .LBB2_1
 ; CHECK-NEXT:  @ %bb.2: @ %middle.block
 ; CHECK-NEXT:    vmov q0, q3
 ; CHECK-NEXT:    vaddv.u32 r0, q0
 ; CHECK-NEXT:    str.w r0, [r2, r1, lsl #2]
-; CHECK-NEXT:    pop {r4, pc}
+; CHECK-NEXT:    pop {r7, pc}
 ; CHECK-NEXT:    .p2align 4
 ; CHECK-NEXT:  @ %bb.3:
 ; CHECK-NEXT:  .LCPI2_0:
 ; CHECK-NEXT:    .long 4294967228 @ 0xffffffbc
 ; CHECK-NEXT:    .long 4294967248 @ 0xffffffd0
 ; CHECK-NEXT:    .long 4294967268 @ 0xffffffe4
 ; CHECK-NEXT:    .long 4294967288 @ 0xfffffff8
 entry:                                  ; preds = %middle.

llvm/test/CodeGen/Thumb2/mve-vecreduce-loops.ll

vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 3
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i32 [ 0, %vector.ph ], [ %4, %vector.body ]
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i32, i32* %x, i32 %index
   %1 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %1, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %2 = select <4 x i1> %active.lane.mask, <4 x i32> %wide.masked.load, <4 x i32> zeroinitializer
   %3 = call i32 @llvm.experimental.vector.reduce.add.v4i32(<4 x i32> %2)
   %4 = add i32 %3, %vec.phi
   %index.next = add i32 %index, 4
   %5 = icmp eq i32 %index.next, %n.vec
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 3
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i32 [ 0, %vector.ph ], [ %7, %vector.body ]
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i32, i32* %x, i32 %index
   %1 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %1, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %2 = getelementptr inbounds i32, i32* %y, i32 %index
   %3 = bitcast i32* %2 to <4 x i32>*
   %wide.masked.load13 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %3, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %4 = mul nsw <4 x i32> %wide.masked.load13, %wide.masked.load
   %5 = select <4 x i1> %active.lane.mask, <4 x i32> %4, <4 x i32> zeroinitializer
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 7
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i32 [ 0, %vector.ph ], [ %5, %vector.body ]
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i16, i16* %x, i32 %index
   %1 = bitcast i16* %0 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %1, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %2 = sext <8 x i16> %wide.masked.load to <8 x i32>
   %3 = select <8 x i1> %active.lane.mask, <8 x i32> %2, <8 x i32> zeroinitializer
   %4 = call i32 @llvm.experimental.vector.reduce.add.v8i32(<8 x i32> %3)
   %5 = add i32 %4, %vec.phi
   %index.next = add i32 %index, 8
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 7
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i32 [ 0, %vector.ph ], [ %9, %vector.body ]
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i16, i16* %x, i32 %index
   %1 = bitcast i16* %0 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %1, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %2 = sext <8 x i16> %wide.masked.load to <8 x i32>
   %3 = getelementptr inbounds i16, i16* %y, i32 %index
   %4 = bitcast i16* %3 to <8 x i16>*
   %wide.masked.load14 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %4, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %5 = sext <8 x i16> %wide.masked.load14 to <8 x i32>
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 15
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i32 [ 0, %vector.ph ], [ %5, %vector.body ]
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i8, i8* %x, i32 %index
   %1 = bitcast i8* %0 to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %2 = zext <16 x i8> %wide.masked.load to <16 x i32>
   %3 = select <16 x i1> %active.lane.mask, <16 x i32> %2, <16 x i32> zeroinitializer
   %4 = call i32 @llvm.experimental.vector.reduce.add.v16i32(<16 x i32> %3)
   %5 = add i32 %4, %vec.phi
   %index.next = add i32 %index, 16
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 15
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i32 [ 0, %vector.ph ], [ %9, %vector.body ]
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i8, i8* %x, i32 %index
   %1 = bitcast i8* %0 to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %2 = zext <16 x i8> %wide.masked.load to <16 x i32>
   %3 = getelementptr inbounds i8, i8* %y, i32 %index
   %4 = bitcast i8* %3 to <16 x i8>*
   %wide.masked.load14 = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %4, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %5 = zext <16 x i8> %wide.masked.load14 to <16 x i32>
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 7
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i16 [ 0, %vector.ph ], [ %4, %vector.body ]
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i16, i16* %x, i32 %index
   %1 = bitcast i16* %0 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %1, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %2 = select <8 x i1> %active.lane.mask, <8 x i16> %wide.masked.load, <8 x i16> zeroinitializer
   %3 = call i16 @llvm.experimental.vector.reduce.add.v8i16(<8 x i16> %2)
   %4 = add i16 %3, %vec.phi
   %index.next = add i32 %index, 8
   %5 = icmp eq i32 %index.next, %n.vec
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 7
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i16 [ 0, %vector.ph ], [ %7, %vector.body ]
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i16, i16* %x, i32 %index
   %1 = bitcast i16* %0 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %1, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %2 = getelementptr inbounds i16, i16* %y, i32 %index
   %3 = bitcast i16* %2 to <8 x i16>*
   %wide.masked.load16 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %3, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %4 = mul <8 x i16> %wide.masked.load16, %wide.masked.load
   %5 = select <8 x i1> %active.lane.mask, <8 x i16> %4, <8 x i16> zeroinitializer
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 15
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i16 [ 0, %vector.ph ], [ %5, %vector.body ]
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i8, i8* %x, i32 %index
   %1 = bitcast i8* %0 to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %2 = zext <16 x i8> %wide.masked.load to <16 x i16>
   %3 = select <16 x i1> %active.lane.mask, <16 x i16> %2, <16 x i16> zeroinitializer
   %4 = call i16 @llvm.experimental.vector.reduce.add.v16i16(<16 x i16> %3)
   %5 = add i16 %4, %vec.phi
   %index.next = add i32 %index, 16
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 15
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i16 [ 0, %vector.ph ], [ %9, %vector.body ]
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i8, i8* %x, i32 %index
   %1 = bitcast i8* %0 to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %2 = zext <16 x i8> %wide.masked.load to <16 x i16>
   %3 = getelementptr inbounds i8, i8* %y, i32 %index
   %4 = bitcast i8* %3 to <16 x i8>*
   %wide.masked.load18 = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %4, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %5 = zext <16 x i8> %wide.masked.load18 to <16 x i16>
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 15
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i8 [ 0, %vector.ph ], [ %4, %vector.body ]
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i8, i8* %x, i32 %index
   %1 = bitcast i8* %0 to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %2 = select <16 x i1> %active.lane.mask, <16 x i8> %wide.masked.load, <16 x i8> zeroinitializer
   %3 = call i8 @llvm.experimental.vector.reduce.add.v16i8(<16 x i8> %2)
   %4 = add i8 %3, %vec.phi
   %index.next = add i32 %index, 16
   %5 = icmp eq i32 %index.next, %n.vec
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 15
   %n.vec = and i32 %n.rnd.up, -16
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i8 [ 0, %vector.ph ], [ %7, %vector.body ]
-  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <16 x i1> @llvm.get.active.lane.mask.v16i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i8, i8* %x, i32 %index
   %1 = bitcast i8* %0 to <16 x i8>*
   %wide.masked.load = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %1, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %2 = getelementptr inbounds i8, i8* %y, i32 %index
   %3 = bitcast i8* %2 to <16 x i8>*
   %wide.masked.load15 = call <16 x i8> @llvm.masked.load.v16i8.p0v16i8(<16 x i8>* %3, i32 1, <16 x i1> %active.lane.mask, <16 x i8> undef)
   %4 = mul <16 x i8> %wide.masked.load15, %wide.masked.load
   %5 = select <16 x i1> %active.lane.mask, <16 x i8> %4, <16 x i8> zeroinitializer
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 3
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i64 [ 0, %vector.ph ], [ %5, %vector.body ]
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i32, i32* %x, i32 %index
   %1 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %1, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %2 = sext <4 x i32> %wide.masked.load to <4 x i64>
   %3 = select <4 x i1> %active.lane.mask, <4 x i64> %2, <4 x i64> zeroinitializer
   %4 = call i64 @llvm.experimental.vector.reduce.add.v4i64(<4 x i64> %3)
   %5 = add i64 %4, %vec.phi
   %index.next = add i32 %index, 4
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 3
   %n.vec = and i32 %n.rnd.up, -4
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i64 [ 0, %vector.ph ], [ %9, %vector.body ]
-  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i32, i32* %x, i32 %index
   %1 = bitcast i32* %0 to <4 x i32>*
   %wide.masked.load = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %1, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %2 = sext <4 x i32> %wide.masked.load to <4 x i64>
   %3 = getelementptr inbounds i32, i32* %y, i32 %index
   %4 = bitcast i32* %3 to <4 x i32>*
   %wide.masked.load14 = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* %4, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
   %5 = sext <4 x i32> %wide.masked.load14 to <4 x i64>
vector.ph:                                        ; preds = %entry
   %n.rnd.up = add i32 %n, 7
   %n.vec = and i32 %n.rnd.up, -8
   %trip.count.minus.1 = add i32 %n, -1
   br label %vector.body
 
 vector.body:                                      ; preds = %vector.body, %vector.ph
   %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
   %vec.phi = phi i64 [ 0, %vector.ph ], [ %9, %vector.body ]
-  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %trip.count.minus.1)
+  %active.lane.mask = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32 %index, i32 %n)
   %0 = getelementptr inbounds i16, i16* %x, i32 %index
   %1 = bitcast i16* %0 to <8 x i16>*
   %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %1, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %2 = sext <8 x i16> %wide.masked.load to <8 x i64>
   %3 = getelementptr inbounds i16, i16* %y, i32 %index
   %4 = bitcast i16* %3 to <8 x i16>*
   %wide.masked.load14 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %4, i32 2, <8 x i1> %active.lane.mask, <8 x i16> undef)
   %5 = sext <8 x i16> %wide.masked.load14 to <8 x i64>