[PATCH] D140260: [LoopVectorize] Fix crash on "vector->scalar" bitcast vectorization

[Florian Hahn via Phabricator via llvm-commits]

fhahn added inline comments.


================
Comment at: llvm/lib/Transforms/Vectorize/LoopVectorize.cpp:4736
+      if (isa<BitCastInst>(&I) && I.getOperand(0)->getType()->isVectorTy()) {
+        assert(TheLoop->hasLoopInvariantOperands(&I) &&
+               "Vector operand must be loop-invariant");
----------------
This should use isOutOfScope like the `ExtractValueInst` case?

It would also be good to add a test case where the vector bit cast is loop-dependent. At the moment, we bail out on any instruction with a vector return type, but it would be good to have a case if the restriction gets removed some day.


================
Comment at: llvm/lib/Transforms/Vectorize/LoopVectorize.cpp:4738
+               "Vector operand must be loop-invariant");
+        LLVM_DEBUG(dbgs() << "LV: Found uniform instruction: " << I << "\n");
+        addToWorklistIfAllowed(&I);
----------------
No need to dump there I think, the message will be printed once the entry in the worklist has been processed.


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:1
+; REQUIRES: assert
+; RUN: opt < %s -loop-vectorize -force-vector-width=2 -S 2>&1 | FileCheck %s
----------------
the test doesn't requires asserts.


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:2
+; REQUIRES: assert
+; RUN: opt < %s -loop-vectorize -force-vector-width=2 -S 2>&1 | FileCheck %s
+
----------------
This is missing CHECK lines? Should probably check at least the vector code generated.


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:4
+
+; Function Attrs: argmemonly nofree norecurse nosync nounwind uwtable
+define dso_local i64 @foo(ptr nocapture noundef %a, i32 noundef %n, <2 x float> %in2) {
----------------
not needed


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:5
+; Function Attrs: argmemonly nofree norecurse nosync nounwind uwtable
+define dso_local i64 @foo(ptr nocapture noundef %a, i32 noundef %n, <2 x float> %in2) {
+entry:
----------------
dso_local not needed.


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:8
+  %cmp3 = icmp sgt i32 %n, 0
+  br i1 %cmp3, label %for.body.preheader, label %for.cond.cleanup
+
----------------
those checks should not be needed


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:11
+for.body.preheader:                               ; preds = %entry
+  %wide.trip.count = zext i32 %n to i64
+  br label %for.body
----------------
no need for the zext here, could either use constant or pass `%n` as i64.


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:14
+
+for.cond.cleanup:                                 ; preds = %for.body, %entry
+  %result = phi i64 [ 0, %entry], [ %illegal, %for.body]
----------------
for easier readability, move the block to the end and call something like `exit:`?


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:19
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i64 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %arrayidx = getelementptr inbounds i32, ptr %a, i64 %indvars.iv
----------------
`indvars.*` prefix makes the names unnecessarily long, could be dropped.


================
Comment at: llvm/test/Transforms/LoopVectorize/vector2scalar-bitcast.ll:21
+  %arrayidx = getelementptr inbounds i32, ptr %a, i64 %indvars.iv
+  %illegal = bitcast <2 x float> %in2 to i64
+  %0 = load i32, ptr %arrayidx, align 4
----------------
Could you also add a variant where %illegal is used in the loop?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140260/new/

https://reviews.llvm.org/D140260