[llvm] r331799 - [LV] Fix for PR37248, Broadcast codegen incorrectly assumed vector loop body is single basic block

[Hideki Saito via llvm-commits]

Author: hsaito
Date: Tue May  8 11:57:34 2018
New Revision: 331799

URL: http://llvm.org/viewvc/llvm-project?rev=331799&view=rev
Log:
[LV] Fix for PR37248, Broadcast codegen incorrectly assumed vector loop body is single basic block

Summary:
Broadcast code generation emitted instructions in pre-header, while the instruction they are dependent on in the vector loop body.
This resulted in an IL verification error ---- value used before defined.


Reviewers: rengolin, fhahn, hfinkel

Reviewed By: rengolin, fhahn

Subscribers: dcaballe, Ka-Ka, llvm-commits

Differential Revision: https://reviews.llvm.org/D46302

Added:
    llvm/trunk/test/Transforms/LoopVectorize/pr37248.ll
Modified:
    llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp

Modified: llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp?rev=331799&r1=331798&r2=331799&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp (original)
+++ llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp Tue May  8 11:57:34 2018
@@ -1737,14 +1737,16 @@ struct LoopVectorize : public FunctionPa
 //===----------------------------------------------------------------------===//
 
 Value *InnerLoopVectorizer::getBroadcastInstrs(Value *V) {
-  // We need to place the broadcast of invariant variables outside the loop.
+  // We need to place the broadcast of invariant variables outside the loop,
+  // but only if it's proven safe to do so. Else, broadcast will be inside
+  // vector loop body.
   Instruction *Instr = dyn_cast<Instruction>(V);
-  bool NewInstr = (Instr && Instr->getParent() == LoopVectorBody);
-  bool Invariant = OrigLoop->isLoopInvariant(V) && !NewInstr;
-
+  bool SafeToHoist = OrigLoop->isLoopInvariant(V) &&
+                     (!Instr ||
+                      DT->dominates(Instr->getParent(), LoopVectorPreHeader));
   // Place the code for broadcasting invariant variables in the new preheader.
   IRBuilder<>::InsertPointGuard Guard(Builder);
-  if (Invariant)
+  if (SafeToHoist)
     Builder.SetInsertPoint(LoopVectorPreHeader->getTerminator());
 
   // Broadcast the scalar into all locations in the vector.

Added: llvm/trunk/test/Transforms/LoopVectorize/pr37248.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/LoopVectorize/pr37248.ll?rev=331799&view=auto
==============================================================================
--- llvm/trunk/test/Transforms/LoopVectorize/pr37248.ll (added)
+++ llvm/trunk/test/Transforms/LoopVectorize/pr37248.ll Tue May  8 11:57:34 2018
@@ -0,0 +1,42 @@
+; RUN: opt -passes='loop-vectorize' -force-vector-width=2 -S < %s | FileCheck %s
+;
+; Forcing VF=2 to trigger vector code gen
+;
+; This is a test case that let's vectorizer's code gen to generate
+; more than one BasicBlocks in the loop body (emulated masked scatter)
+; for those targets that do not support masked scatter. Broadcast
+; code generation was previously dependent on loop body being
+; a single basic block and this test case exposed incorrect code gen
+; resulting in an assert in IL verification. Test passes if IL verification
+; does not fail.
+;
+; Performing minimal check in the output to ensure the loop is actually
+; vectorized.
+;
+; CHECK: vector.body
+
+ at a = external global [2 x i16], align 1
+
+define void @f1() {
+entry:
+  br label %for.body
+
+for.body:                                         ; preds = %land.end, %entry
+  %0 = phi i32 [ undef, %entry ], [ %dec, %land.end ]
+  br i1 undef, label %land.end, label %land.rhs
+
+land.rhs:                                         ; preds = %for.body
+  %1 = load i32, i32* undef, align 1
+  br label %land.end
+
+land.end:                                         ; preds = %land.rhs, %for.body
+  %2 = trunc i32 %0 to i16
+  %arrayidx = getelementptr inbounds [2 x i16], [2 x i16]* @a, i16 0, i16 %2
+  store i16 undef, i16* %arrayidx, align 1
+  %dec = add nsw i32 %0, -1
+  %cmp = icmp sgt i32 %0, 1
+  br i1 %cmp, label %for.body, label %for.cond.for.end_crit_edge
+
+for.cond.for.end_crit_edge:                       ; preds = %land.end
+  unreachable
+}