[VectorUtils] Don't try and truncate PHIs to a smaller bitwidth

We already try not to truncate PHIs in computeMinimalBitwidths. LoopVectorize can't handle it and we really don't need to, because both induction and reduction PHIs are truncated by other means.

However, we weren't bailing out in all the places we should have, and we ended up by returning a PHI to be truncated, which has caused PR27018.

This fixes PR17018.

llvm-svn: 264852

2 files changed, 73 insertions, 0 deletions

diff --git a/llvm/lib/Analysis/VectorUtils.cpp b/llvm/lib/Analysis/VectorUtils.cpp
index d9b1354daf7..2881645825a 100644
--- a/llvm/lib/Analysis/VectorUtils.cpp
+++ b/llvm/lib/Analysis/VectorUtils.cpp
@@ -499,6 +499,7 @@ llvm::computeMinimumValueSizes(ArrayRef<BasicBlock *> Blocks, DemandedBits &DB,
 
     uint64_t V = DB.getDemandedBits(I).getZExtValue();
     DBits[Leader] |= V;
+    DBits[I] = V;
 
     // Casts, loads and instructions outside of our range terminate a chain
     // successfully.
@@ -549,6 +550,20 @@ llvm::computeMinimumValueSizes(ArrayRef<BasicBlock *> Blocks, DemandedBits &DB,
     // Round up to a power of 2
     if (!isPowerOf2_64((uint64_t)MinBW))
       MinBW = NextPowerOf2(MinBW);
+
+    // We don't modify the types of PHIs. Reductions will already have been
+    // truncated if possible, and inductions' sizes will have been chosen by
+    // indvars.
+    // If we are required to shrink a PHI, abandon this entire equivalence class.
+    bool Abort = false;
+    for (auto MI = ECs.member_begin(I), ME = ECs.member_end(); MI != ME; ++MI)
+      if (isa<PHINode>(*MI) && MinBW < (*MI)->getType()->getScalarSizeInBits()) {
+        Abort = true;
+        break;
+      }
+    if (Abort)
+      continue;
+
     for (auto MI = ECs.member_begin(I), ME = ECs.member_end(); MI != ME; ++MI) {
       if (!isa<Instruction>(*MI))
         continue;
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/loop-vectorization-factors.ll b/llvm/test/Transforms/LoopVectorize/AArch64/loop-vectorization-factors.ll
index 51f899c2f64..c7ced757581 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/loop-vectorization-factors.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/loop-vectorization-factors.ll
@@ -205,5 +205,63 @@ for.body:                                         ; preds = %for.body, %for.body
   br i1 %exitcond, label %for.cond.cleanup, label %for.body
 }
 
+; CHECK-LABEL: @add_phifail(
+; CHECK: load <16 x i8>, <16 x i8>*
+; CHECK: add nuw nsw <16 x i32>
+; CHECK: store <16 x i8>
+; Function Attrs: nounwind
+define void @add_phifail(i8* noalias nocapture readonly %p, i8* noalias nocapture %q, i32 %len) #0 {
+entry:
+  %cmp8 = icmp sgt i32 %len, 0
+  br i1 %cmp8, label %for.body, label %for.cond.cleanup
+
+for.cond.cleanup:                                 ; preds = %for.body, %entry
+  ret void
+
+for.body:                                         ; preds = %entry, %for.body
+  %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
+  %a_phi = phi i32 [ %conv, %for.body ], [ 0, %entry ]
+  %arrayidx = getelementptr inbounds i8, i8* %p, i64 %indvars.iv
+  %0 = load i8, i8* %arrayidx
+  %conv = zext i8 %0 to i32
+  %add = add nuw nsw i32 %conv, 2
+  %conv1 = trunc i32 %add to i8
+  %arrayidx3 = getelementptr inbounds i8, i8* %q, i64 %indvars.iv
+  store i8 %conv1, i8* %arrayidx3
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
+  %exitcond = icmp eq i32 %lftr.wideiv, %len
+  br i1 %exitcond, label %for.cond.cleanup, label %for.body
+}
+
+; CHECK-LABEL: @add_phifail2(
+; CHECK: load <16 x i8>, <16 x i8>*
+; CHECK: add nuw nsw <16 x i32>
+; CHECK: store <16 x i8>
+; Function Attrs: nounwind
+define i8 @add_phifail2(i8* noalias nocapture readonly %p, i8* noalias nocapture %q, i32 %len) #0 {
+entry:
+  br label %for.body
+
+for.cond.cleanup:                                 ; preds = %for.body, %entry
+  %ret = trunc i32 %a_phi to i8
+  ret i8 %ret
+
+for.body:                                         ; preds = %entry, %for.body
+  %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
+  %a_phi = phi i32 [ %conv, %for.body ], [ 0, %entry ]
+  %arrayidx = getelementptr inbounds i8, i8* %p, i64 %indvars.iv
+  %0 = load i8, i8* %arrayidx
+  %conv = zext i8 %0 to i32
+  %add = add nuw nsw i32 %conv, 2
+  %conv1 = trunc i32 %add to i8
+  %arrayidx3 = getelementptr inbounds i8, i8* %q, i64 %indvars.iv
+  store i8 %conv1, i8* %arrayidx3
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
+  %exitcond = icmp eq i32 %lftr.wideiv, %len
+  br i1 %exitcond, label %for.cond.cleanup, label %for.body
+}
 
 attributes #0 = { nounwind }
+