[InstCombine] If Inst in unreachable refers to Inst change it to poison (llvm#65107)

Instructions in unreachable basic blocks are removed, but terminators
are not. In this case, even instructions that are only referenced by
a terminator, such as a return instruction, cannot be processed
properly.

This patch changes the operand of terminator instruction in an
unreachable basic block to poison if it refers to the instruction,
allowing the instruction to be properly processed.

Author: ParkHanbum

llvm/include/llvm/Transforms/Utils/Local.h

@@ -357,6 +357,10 @@ Value *salvageDebugInfoImpl(Instruction &I, uint64_t CurrentLocOps,
 bool replaceAllDbgUsesWith(Instruction &From, Value &To, Instruction &DomPoint,
                            DominatorTree &DT);
 
+/// If the terminator in an unreachable basic block refers to an instruction
+/// transform it to poison.
+bool handleUnreachableTerminator(Instruction *I);
+
 /// Remove all instructions from a basic block other than its terminator
 /// and any present EH pad instructions. Returns a pair where the first element
 /// is the number of instructions (excluding debug info intrinsics) that have

llvm/lib/Transforms/InstCombine/InstructionCombining.cpp

@@ -3231,9 +3231,7 @@ void InstCombinerImpl::handleUnreachableFrom(
     MadeIRChange = true;
   }
 
-  // RemoveDIs: to match behaviour in dbg.value mode, drop debug-info on
-  // terminator too.
-  BB->getTerminator()->dropDbgValues();
+  MadeIRChange |= handleUnreachableTerminator(BB->getTerminator());
 
   // Handle potentially dead successors.
   for (BasicBlock *Succ : successors(BB))

llvm/lib/Transforms/Utils/Local.cpp

@@ -2762,15 +2762,30 @@ bool llvm::replaceAllDbgUsesWith(Instruction &From, Value &To,
   return false;
 }
 
+bool llvm::handleUnreachableTerminator(Instruction *I) {
+  bool Changed = false;
+  // RemoveDIs: erase debug-info on this instruction manually.
+  I->dropDbgValues();
+  for (Use &U : I->operands()) {
+    Value *Op = U.get();
+    if (isa<Instruction>(Op) && !Op->getType()->isTokenTy()) {
+      U.set(PoisonValue::get(Op->getType()));
+      Changed = true;
+    }
+  }
+
+  return Changed;
+}
+
 std::pair<unsigned, unsigned>
 llvm::removeAllNonTerminatorAndEHPadInstructions(BasicBlock *BB) {
   unsigned NumDeadInst = 0;
   unsigned NumDeadDbgInst = 0;
   // Delete the instructions backwards, as it has a reduced likelihood of
   // having to update as many def-use and use-def chains.
   Instruction *EndInst = BB->getTerminator(); // Last not to be deleted.
-  // RemoveDIs: erasing debug-info must be done manually.
-  EndInst->dropDbgValues();
+  handleUnreachableTerminator(EndInst);
+
   while (EndInst != &BB->front()) {
     // Delete the next to last instruction.
     Instruction *Inst = &*--EndInst->getIterator();

llvm/test/Transforms/InstCombine/phi-select-constant.ll

@@ -140,12 +140,11 @@ end:
 define i16 @sink_to_unreachable_crash(i1 %a)  {
 ; CHECK-LABEL: @sink_to_unreachable_crash(
 ; CHECK-NEXT:  entry:
-; CHECK-NEXT:    [[S:%.*]] = select i1 [[A:%.*]], i16 0, i16 5
 ; CHECK-NEXT:    br label [[INF_LOOP:%.*]]
 ; CHECK:       inf_loop:
 ; CHECK-NEXT:    br label [[INF_LOOP]]
 ; CHECK:       unreachable:
-; CHECK-NEXT:    ret i16 [[S]]
+; CHECK-NEXT:    ret i16 poison
 ;
 entry:
   %s = select i1 %a, i16 0, i16 5

llvm/test/Transforms/InstCombine/pr63791.ll

@@ -17,7 +17,7 @@ define void @y() {
 ; CHECK:       for.cond5.preheader.i:
 ; CHECK-NEXT:    br i1 false, label [[FOR_INC19_I:%.*]], label [[FOR_COND1_LOOPEXIT_I:%.*]]
 ; CHECK:       for.inc19.i:
-; CHECK-NEXT:    br i1 false, label [[FOR_INC19_I]], label [[FOR_COND1_LOOPEXIT_I]]
+; CHECK-NEXT:    br i1 poison, label [[FOR_INC19_I]], label [[FOR_COND1_LOOPEXIT_I]]
 ;
 entry:
   br label %for.cond.i

llvm/test/Transforms/InstCombine/sink_to_unreachable.ll

@@ -157,3 +157,157 @@ bb3:
   %p = phi i32 [0, %bb1], [%a, %bb2]
   ret i32 %p
 }
+
+define i1 @sink_to_unreachable_ret(i16 %X)  {
+; CHECK-LABEL: @sink_to_unreachable_ret(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    br i1 true, label [[LOOP]], label [[UNREACH:%.*]]
+; CHECK:       unreach:
+; CHECK-NEXT:    ret i1 poison
+;
+entry:
+  br label %loop
+
+loop:
+  %p = icmp sgt i16 %X, 16
+  br i1 true, label %loop, label %unreach
+
+unreach:
+  ret i1 %p
+}
+
+define void @sink_to_unreachable_condbr(i16 %X)  {
+; CHECK-LABEL: @sink_to_unreachable_condbr(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    br i1 true, label [[LOOP]], label [[UNREACH:%.*]]
+; CHECK:       unreach:
+; CHECK-NEXT:    br i1 poison, label [[DUMMY:%.*]], label [[LOOP]]
+; CHECK:       dummy:
+; CHECK-NEXT:    unreachable
+;
+entry:
+  br label %loop
+
+loop:
+  %p = icmp sgt i16 %X, 16
+  br i1 true, label %loop, label %unreach
+
+unreach:
+  br i1 %p, label %dummy, label %loop
+
+dummy:
+  unreachable
+}
+
+define void @sink_to_unreachable_switch(i16 %X)  {
+; CHECK-LABEL: @sink_to_unreachable_switch(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    br i1 true, label [[LOOP]], label [[UNREACH:%.*]]
+; CHECK:       unreach:
+; CHECK-NEXT:    switch i16 poison, label [[UNREACH_RET:%.*]] [
+; CHECK-NEXT:    ]
+; CHECK:       unreach.ret:
+; CHECK-NEXT:    unreachable
+;
+entry:
+  br label %loop
+
+loop:
+  %quantum = srem i16 %X, 32
+  br i1 true, label %loop, label %unreach
+
+unreach:
+  switch i16 %quantum, label %unreach.ret []
+
+unreach.ret:
+  unreachable
+}
+
+define void @sink_to_unreachable_indirectbr(ptr %Ptr)  {
+; CHECK-LABEL: @sink_to_unreachable_indirectbr(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    br i1 true, label [[LOOP]], label [[UNREACH:%.*]]
+; CHECK:       unreach:
+; CHECK-NEXT:    indirectbr ptr poison, [label %loop]
+;
+entry:
+  br label %loop
+
+loop:
+  %gep = getelementptr inbounds ptr, ptr %Ptr, i16 1
+  br i1 true, label %loop, label %unreach
+
+unreach:
+  indirectbr ptr %gep, [label %loop]
+}
+
+define void @sink_to_unreachable_invoke(ptr %Ptr) personality ptr @__CxxFrameHandler3 {
+; CHECK-LABEL: @sink_to_unreachable_invoke(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    br i1 true, label [[LOOP]], label [[UNREACH:%.*]]
+; CHECK:       unreach:
+; CHECK-NEXT:    invoke void poison(i1 false)
+; CHECK-NEXT:            to label [[DUMMY:%.*]] unwind label [[ICATCH_DISPATCH:%.*]]
+; CHECK:       unreach2:
+; CHECK-NEXT:    invoke void @__CxxFrameHandler3(ptr poison)
+; CHECK-NEXT:            to label [[DUMMY]] unwind label [[ICATCH_DISPATCH]]
+; CHECK:       unreach3:
+; CHECK-NEXT:    [[CLEAN:%.*]] = cleanuppad within none []
+; CHECK-NEXT:    invoke void @__CxxFrameHandler3(ptr poison) [ "funclet"(token [[CLEAN]]) ]
+; CHECK-NEXT:            to label [[DUMMY]] unwind label [[ICATCH_DISPATCH]]
+; CHECK:       icatch.dispatch:
+; CHECK-NEXT:    [[TMP1:%.*]] = catchswitch within none [label %icatch] unwind to caller
+; CHECK:       icatch:
+; CHECK-NEXT:    [[TMP2:%.*]] = catchpad within [[TMP1]] [ptr null, i32 64, ptr null]
+; CHECK-NEXT:    catchret from [[TMP2]] to label [[DUMMY2:%.*]]
+; CHECK:       dummy:
+; CHECK-NEXT:    ret void
+; CHECK:       dummy2:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %loop
+
+loop:
+  %gep = getelementptr inbounds ptr, ptr %Ptr, i16 1
+  br i1 true, label %loop, label %unreach
+
+unreach:
+  invoke void %gep(i1 false)
+  to label %dummy unwind label %icatch.dispatch
+
+unreach2:
+  invoke void @__CxxFrameHandler3(ptr %gep)
+  to label %dummy unwind label %icatch.dispatch
+
+unreach3:
+  %clean = cleanuppad within none []
+  invoke void @__CxxFrameHandler3(ptr %gep) [ "funclet"(token %clean) ]
+  to label %dummy unwind label %icatch.dispatch
+
+icatch.dispatch:
+  %tmp1 = catchswitch within none [label %icatch] unwind to caller
+
+icatch:
+  %tmp2 = catchpad within %tmp1 [ptr null, i32 64, ptr null]
+  catchret from %tmp2 to label %dummy2
+
+dummy:
+  ret void
+
+dummy2:
+  ret void
+}
+
+declare void @may_throw()
+declare i32 @__CxxFrameHandler3(...)