[LoopDeletion] Benefit from branches by undef conditions when symbolically executing 1st iteration

[mkazantsev]

Summary

We can exploit branches by undef condition. Frankly, the LangRef says that
such branches are UB, so we can assume that all outgoing edges of such blocks
are dead.

However, from practical perspective, we know that this is not supported correctly
in some other places. So we are being conservative about it.

Branch by undef is treated in the following way:

• If it is a loop-exiting branch, we always assume it exits the loop;
• If not, we arbitrarily assume it takes true value.

[mkazantsev]

Rebased on top of updated tests (w/o br i1 undef).

[aqjune]

llvm/lib/Transforms/Scalar/LoopDeletion.cpp

355

If BB -> IfFalse edge is already live, would it be beneficial if IfFalse is chosen instead?

[nikic]

LGTM. I don't like it, but it's in line with current approaches used elsewhere.

llvm/lib/Transforms/Scalar/LoopDeletion.cpp

355

Blocks are only visited once, so neither edge can be marked live at this point.

llvm/lib/Transforms/Scalar/LoopDeletion.cpp

for (auto *BB : RPOT) {
     if (!LHS->getType()->isIntegerTy()) {
       MarkAllSuccessorsLive(BB);
       continue;
     }
 
     // Can we prove constant true or false for this condition?
     LHS = getValueOnFirstIteration(LHS, FirstIterValue, SQ);
     RHS = getValueOnFirstIteration(RHS, FirstIterValue, SQ);
-    auto *KnownCondition =
-        dyn_cast_or_null<ConstantInt>(SimplifyICmpInst(Pred, LHS, RHS, SQ));
+    auto *KnownCondition = SimplifyICmpInst(Pred, LHS, RHS, SQ);
     if (!KnownCondition) {
+      // Failed to simplify.
       MarkAllSuccessorsLive(BB);
       continue;
     }
-    if (KnownCondition->isAllOnesValue())
+    if (isa<UndefValue>(KnownCondition)) {
+      // TODO: According to langref, branching by undef is undefined behavior.
+      // It means that, theoretically, we should be able to just continue
+      // without marking any successors as live. However, we are not certain
+      // how correct our compiler is at handling such cases. So we are being
+      // very conservative here.
+      //
+      // If there is a non-loop successor, always assume this branch leaves the
+      // loop. Otherwise, arbitrarily take IfTrue.
+      //
+      // Once we are certain that branching by undef is handled correctly by
+      // other transforms, we should not mark any successors live here.
+      if (L->contains(IfTrue) && L->contains(IfFalse))
+        MarkLiveEdge(BB, IfTrue);
+      continue;
+    }
+    auto *ConstCondition = dyn_cast<ConstantInt>(KnownCondition);
+    if (!ConstCondition) {
+      // Non-constant condition, cannot analyze any further.
+      MarkAllSuccessorsLive(BB);
+      continue;
+    }
+    if (ConstCondition->isAllOnesValue())
       MarkLiveEdge(BB, IfTrue);
     else
       MarkLiveEdge(BB, IfFalse);
   }
 
   // We can break the latch if it wasn't live.
   return !LiveEdges.count({ Latch, Header });
 }

llvm/test/Transforms/LoopDeletion/eval_first_iteration.ll

 done:                                             ; preds = %backedge
   %sum.next.lcssa = phi i32 [ %sum.next, %backedge ]
   ret i32 %sum.next.lcssa
 
 failure:
   unreachable
 }
 
-; TODO: We can break the backedge here by exploiting undef.
 define i32 @test_multiple_pred_undef_3(i1 %cond, i1 %cond2) {
 ; CHECK-LABEL: @test_multiple_pred_undef_3(
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
-; CHECK-NEXT:    [[SUM:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[SUM_NEXT:%.*]], [[BACKEDGE:%.*]] ]
+; CHECK-NEXT:    [[SUM:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ]
 ; CHECK-NEXT:    [[SUB:%.*]] = sub i32 4, [[SUM]]
 ; CHECK-NEXT:    [[IS_POSITIVE:%.*]] = icmp sgt i32 [[SUB]], 0
 ; CHECK-NEXT:    br i1 [[IS_POSITIVE]], label [[IF_TRUE:%.*]], label [[IF_FALSE:%.*]]
 ; CHECK:       if.true:
 ; CHECK-NEXT:    br i1 [[COND:%.*]], label [[IF_TRUE_1:%.*]], label [[IF_TRUE_2:%.*]]
 ; CHECK:       if.true.1:
-; CHECK-NEXT:    br label [[BACKEDGE]]
+; CHECK-NEXT:    br label [[BACKEDGE:%.*]]
 ; CHECK:       if.true.2:
 ; CHECK-NEXT:    br label [[BACKEDGE]]
 ; CHECK:       if.false:
 ; CHECK-NEXT:    br i1 [[COND2:%.*]], label [[IF_FALSE_1:%.*]], label [[IF_FALSE_2:%.*]]
 ; CHECK:       if.false.1:
 ; CHECK-NEXT:    br label [[BACKEDGE]]
 ; CHECK:       if.false.2:
 ; CHECK-NEXT:    br label [[BACKEDGE]]
 ; CHECK:       backedge:
 ; CHECK-NEXT:    [[MERGE_PHI:%.*]] = phi i32 [ 0, [[IF_FALSE_1]] ], [ 0, [[IF_FALSE_2]] ], [ undef, [[IF_TRUE_1]] ], [ undef, [[IF_TRUE_2]] ]
-; CHECK-NEXT:    [[SUM_NEXT]] = add i32 [[SUM]], [[MERGE_PHI]]
+; CHECK-NEXT:    [[SUM_NEXT:%.*]] = add i32 [[SUM]], [[MERGE_PHI]]
 ; CHECK-NEXT:    [[LOOP_COND:%.*]] = icmp ne i32 [[SUM_NEXT]], 4
-; CHECK-NEXT:    br i1 [[LOOP_COND]], label [[LOOP]], label [[DONE:%.*]]
+; CHECK-NEXT:    br i1 [[LOOP_COND]], label [[BACKEDGE_LOOP_CRIT_EDGE:%.*]], label [[DONE:%.*]]
+; CHECK:       backedge.loop_crit_edge:
+; CHECK-NEXT:    unreachable
 ; CHECK:       done:
 ; CHECK-NEXT:    [[SUM_NEXT_LCSSA:%.*]] = phi i32 [ [[SUM_NEXT]], [[BACKEDGE]] ]
 ; CHECK-NEXT:    ret i32 [[SUM_NEXT_LCSSA]]
 ; CHECK:       failure:
 ; CHECK-NEXT:    unreachable
 ;
 entry:
   br label %loop