!fixup address comments, thanks

Author: fhahn

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

@@ -9409,11 +9409,11 @@ static bool processLoopInVPlanNativePath(
 
   {
     GeneratedRTChecks Checks(PSE, DT, LI, TTI, F->getDataLayout(), CM.CostKind);
-    InnerLoopVectorizer LB(L, PSE, LI, DT, TTI, AC, VF.Width, 1, &CM, BFI, PSI,
-                           Checks, BestPlan);
+    InnerLoopVectorizer LB(L, PSE, LI, DT, TTI, AC, VF.Width, /*UF=*/1, &CM,
+                           BFI, PSI, Checks, BestPlan);
     LLVM_DEBUG(dbgs() << "Vectorizing outer loop in \""
                       << L->getHeader()->getParent()->getName() << "\"\n");
-    LVP.addMinimumIterationCheck(BestPlan, VF.Width, 1,
+    LVP.addMinimumIterationCheck(BestPlan, VF.Width, /*UF=*/1,
                                  VF.MinProfitableTripCount);
 
     LVP.executePlan(VF.Width, 1, BestPlan, LB, DT, false);

llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp

@@ -740,6 +740,8 @@ void VPlanTransforms::addMinimumIterationCheck(
                              DebugLoc::getUnknown());
 
     // Don't execute the vector loop if (UMax - n) < (VF * UF).
+    // FIXME: Should only check VF * UF, but currently checks Step=max(VF*UF,
+    // minProfitableTripCount).
     TripCountCheck = Builder.createICmp(ICmpInst::ICMP_ULT, DistanceToMax,
                                         Builder.createExpandSCEV(Step), DL);
   }

llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp

@@ -3512,6 +3512,9 @@ VPlanTransforms::expandSCEVs(VPlan &Plan, ScalarEvolution &SE) {
     InsertPt = std::next(ExpSCEV->getIterator());
     ExpSCEV->eraseFromParent();
   }
+  assert(none_of(*Entry, IsaPred<VPExpandSCEVRecipe>) &&
+         "VPExpandSCEVRecipes must be be at the beginning of the entry block, "
+         "after any VPIRInstructions");
   for (Instruction &I : *EntryBB) {
     if (!Expander.isInsertedInstruction(&I) || isa<PHINode>(I))
       continue;

llvm/lib/Transforms/Vectorize/VPlanTransforms.h

@@ -62,9 +62,40 @@ struct VPlanTransforms {
   /// The created loop is wrapped in an initial skeleton to facilitate
   /// vectorization, consisting of a vector pre-header, an exit block for the
   /// main vector loop (middle.block) and a new block as preheader of the scalar
-  /// loop (scalar.ph). It also adds a canonical IV and its increment, using \p
-  /// InductionTy and \p IVDL, and creates a VPValue expression for the original
-  /// trip count.
+  /// loop (scalar.ph). See below for an illustration. It also adds a canonical
+  /// IV and its increment, using \p InductionTy and \p IVDL, and creates a
+  /// VPValue expression for the original trip count.
+  ///
+  ///    [ ] <-- Plan's entry VPIRBasicBlock, wrapping the original loop's
+  ///    / \       old preheader. Will contain iteration number check and SCEV
+  ///   |   |      expansions.
+  ///   |   |
+  ///   /   v
+  ///  |   [ ] <-- vector loop bypass (may consist of multiple blocks) will be
+  ///  |  / |      added later.
+  ///  | /  v
+  ///  ||  [ ]     <-- vector pre header.
+  ///  |/   |
+  ///  |    v
+  ///  |   [  ] \  <-- plain CFG loop wrapping original loop to be vectorized.
+  ///  |   [  ]_|
+  ///  |    |
+  ///  |    v
+  ///  |   [ ]   <--- middle-block with the branch to successors
+  ///  |   / |
+  ///  |  /  |
+  ///  | |   v
+  ///  \--->[ ]   <--- scalar preheader (initial a VPBasicBlock, which will be
+  ///    |   |        replaced later by a VPIRBasicBlock wrapping the scalar
+  ///    |   |         preheader basic block.
+  ///    |   |
+  ///        v      <-- edge from middle to exit iff epilogue is not required.
+  ///    |  [ ] \
+  ///    |  [ ]_|   <-- old scalar loop to handle remainder (scalar epilogue,
+  ///    |   |          header wrapped in VPIRBasicBlock).
+  ///    \   |
+  ///     \  v
+  ///      >[ ]     <-- original loop exit block(s), wrapped in VPIRBasicBlocks.
   LLVM_ABI_FOR_TEST static std::unique_ptr<VPlan>
   buildVPlan0(Loop *TheLoop, LoopInfo &LI, Type *InductionTy, DebugLoc IVDL,
               PredicatedScalarEvolution &PSE);