[Attributor] Liveness for internal functions.

For an internal function, if all its call sites are dead, the body of the function is considered dead.

Reviewers: jdoerfert, uenoku

Subscribers: hiraditya, llvm-commits

Differential Revision: https://reviews.llvm.org/D66155

llvm-svn: 369470

llvm/include/llvm/Transforms/IPO/Attributor.h

@@ -297,6 +297,9 @@ struct IRPosition {
     if (auto *Arg = dyn_cast<Argument>(&V))
       if (!Arg->getParent()->isDeclaration())
         return &Arg->getParent()->getEntryBlock().front();
+    if (auto *F = dyn_cast<Function>(&V))
+      if (!F->isDeclaration())
+        return &(F->getEntryBlock().front());
     return nullptr;
   }
   const Instruction *getCtxI() const {

llvm/lib/Transforms/IPO/Attributor.cpp

@@ -1570,13 +1570,22 @@ struct AAIsDeadImpl : public AAIsDead {
 
   void initialize(Attributor &A) override {
     const Function *F = getAssociatedFunction();
+
+    if (F->hasInternalLinkage())
+      return;
+
     if (!F || !F->hasExactDefinition()) {
       indicatePessimisticFixpoint();
       return;
     }
 
+    exploreFromEntry(A, F);
+  }
+
+  void exploreFromEntry(Attributor &A, const Function *F) {
     ToBeExploredPaths.insert(&(F->getEntryBlock().front()));
     AssumedLiveBlocks.insert(&(F->getEntryBlock()));
+
     for (size_t i = 0; i < ToBeExploredPaths.size(); ++i)
       if (const Instruction *NextNoReturnI =
               findNextNoReturn(A, ToBeExploredPaths[i]))
@@ -1606,7 +1615,12 @@ struct AAIsDeadImpl : public AAIsDead {
            "Attempted to manifest an invalid state!");
 
     ChangeStatus HasChanged = ChangeStatus::UNCHANGED;
-    const Function &F = *getAssociatedFunction();
+    Function &F = *getAssociatedFunction();
+
+    if (AssumedLiveBlocks.empty()) {
+      F.replaceAllUsesWith(UndefValue::get(F.getType()));
+      return ChangeStatus::CHANGED;
+    }
 
     // Flag to determine if we can change an invoke to a call assuming the
     // callee is nounwind. This is not possible if the personality of the
@@ -1725,6 +1739,13 @@ struct AAIsDeadFunction final : public AAIsDeadImpl {
 
   /// See AbstractAttribute::trackStatistics()
   void trackStatistics() const override {
+    STATS_DECL(DeadInternalFunction, Function,
+               "Number of internal functions classified as dead (no live callsite)");
+    BUILD_STAT_NAME(DeadInternalFunction, Function) +=
+        (getAssociatedFunction()->hasInternalLinkage() &&
+         AssumedLiveBlocks.empty())
+            ? 1
+            : 0;
     STATS_DECL(DeadBlocks, Function,
                "Number of basic blocks classified as dead");
     BUILD_STAT_NAME(DeadBlocks, Function) +=
@@ -1796,10 +1817,25 @@ const Instruction *AAIsDeadImpl::findNextNoReturn(Attributor &A,
 }
 
 ChangeStatus AAIsDeadImpl::updateImpl(Attributor &A) {
+  const Function *F = getAssociatedFunction();
+  ChangeStatus Status = ChangeStatus::UNCHANGED;
+
+  if (F->hasInternalLinkage() && AssumedLiveBlocks.empty()) {
+    auto CallSiteCheck = [&](CallSite) { return false; };
+
+    // All callsites of F are dead.
+    if (A.checkForAllCallSites(CallSiteCheck, *this, true))
+      return ChangeStatus::UNCHANGED;
+
+    // There exists at least one live call site, so we explore the function.
+    Status = ChangeStatus::CHANGED;
+
+    exploreFromEntry(A, F);
+  }
+
   // Temporary collection to iterate over existing noreturn instructions. This
   // will alow easier modification of NoReturnCalls collection
   SmallVector<const Instruction *, 8> NoReturnChanged;
-  ChangeStatus Status = ChangeStatus::UNCHANGED;
 
   for (const Instruction *I : NoReturnCalls)
     NoReturnChanged.push_back(I);
@@ -2249,6 +2285,11 @@ bool Attributor::isAssumedDead(const AbstractAttribute &AA,
   if (!LivenessAA)
     LivenessAA =
         &getAAFor<AAIsDead>(AA, IRPosition::function(*CtxI->getFunction()));
+
+  // Don't check liveness for AAIsDead.
+  if (&AA == LivenessAA)
+    return false;
+
   if (!LivenessAA->isAssumedDead(CtxI))
     return false;
 

llvm/test/Transforms/FunctionAttrs/align.ll

@@ -139,6 +139,7 @@ define internal i8* @f3(i8* readnone %0) local_unnamed_addr #0 {
 ; Better than IR information
 ; ATTRIBUTOR: define align 32 i32* @test7(i32* returned align 32 %p)
 define align 4 i32* @test7(i32* align 32 %p) #0 {
+  tail call i8* @f1(i8* align 8 dereferenceable(1) @a1)
   ret i32* %p
 }
 

llvm/test/Transforms/FunctionAttrs/liveness.ll

@@ -12,6 +12,27 @@ declare i32 @foo_noreturn() noreturn
 
 declare i32 @bar() nosync readnone
 
+; This internal function has no live call sites, so all its BBs are considered dead,
+; and nothing should be deduced for it.
+
+; CHECK: define internal i32 @dead_internal_func(i32 %0)
+define internal i32 @dead_internal_func(i32 %0) {
+  %2 = icmp slt i32 %0, 1
+  br i1 %2, label %3, label %5
+
+; <label>:3:                                      ; preds = %5, %1
+  %4 = phi i32 [ 1, %1 ], [ %8, %5 ]
+  ret i32 %4
+
+; <label>:5:                                      ; preds = %1, %5
+  %6 = phi i32 [ %9, %5 ], [ 1, %1 ]
+  %7 = phi i32 [ %8, %5 ], [ 1, %1 ]
+  %8 = mul nsw i32 %6, %7
+  %9 = add nuw nsw i32 %6, 1
+  %10 = icmp eq i32 %6, %0
+  br i1 %10, label %3, label %5
+}
+
 ; CHECK: Function Attrs: nofree norecurse nounwind uwtable willreturn
 define i32 @volatile_load(i32*) norecurse nounwind uwtable {
   %2 = load volatile i32, i32* %0, align 4
@@ -35,6 +56,8 @@ entry:
   call void @no_return_call()
   ; CHECK: call void @no_return_call()
   ; CHECK-NEXT: unreachable
+  call i32 @dead_internal_func(i32 10)
+  ; CHECK call i32 undef(i32 10)
   %cmp = icmp eq i32 %a, 0
   br i1 %cmp, label %cond.true, label %cond.false
 
@@ -96,13 +119,17 @@ cond.true:                                        ; preds = %entry
   call void @no_return_call()
   ; CHECK: call void @no_return_call()
   ; CHECK-NEXT: unreachable
+  call i32 @dead_internal_func(i32 10)
+  ; CHECK call i32 undef(i32 10)
   %call = call i32 @foo()
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
   call void @no_return_call()
   ; CHECK: call void @no_return_call()
   ; CHECK-NEXT: unreachable
+  call i32 @dead_internal_func(i32 10)
+  ; CHECK call i32 undef(i32 10)
   %call1 = call i32 @bar()
   br label %cond.end
 
@@ -311,3 +338,73 @@ cond.end:                                               ; preds = %cond.if, %con
   %8 = phi i32 [ %1, %cond.elseif ], [ 0, %cond.else ], [ 0, %cond.if ]
   ret i32 %8
 }
+
+; SCC test
+;
+; char a1 __attribute__((aligned(8)));
+; char a2 __attribute__((aligned(16)));
+;
+; char* f1(char* a ){
+;     return a?a:f2(&a1);
+; }
+; char* f2(char* a){
+;     return a?f1(a):f3(&a2);
+; }
+;
+; char* f3(char* a){
+;     return a?&a1: f1(&a2);
+; }
+
+@a1 = common global i8 0, align 8
+@a2 = common global i8 0, align 16
+
+define internal i8* @f1(i8* readnone %0) local_unnamed_addr #0 {
+; ATTRIBUTOR: define internal i8* @f1(i8* readnone %0)
+  %2 = icmp eq i8* %0, null
+  br i1 %2, label %3, label %5
+
+; <label>:3:                                      ; preds = %1
+; ATTRIBUTOR: %4 = tail call i8* undef(i8* nonnull align 8 @a1)
+  %4 = tail call i8* @f2(i8* nonnull @a1)
+  br label %5
+
+; <label>:5:                                      ; preds = %1, %3
+  %6 = phi i8* [ %4, %3 ], [ %0, %1 ]
+  ret i8* %6
+}
+
+define internal i8* @f2(i8* readnone %0) local_unnamed_addr #0 {
+; ATTRIBUTOR: define internal i8* @f2(i8* readnone %0)
+  %2 = icmp eq i8* %0, null
+  br i1 %2, label %5, label %3
+
+; <label>:3:                                      ; preds = %1
+
+; ATTRIBUTOR: %4 = tail call i8* undef(i8* nonnull align 8 %0)
+  %4 = tail call i8* @f1(i8* nonnull %0)
+  br label %7
+
+; <label>:5:                                      ; preds = %1
+; ATTRIBUTOR: %6 = tail call i8* undef(i8* nonnull align 16 @a2)
+  %6 = tail call i8* @f3(i8* nonnull @a2)
+  br label %7
+
+; <label>:7:                                      ; preds = %5, %3
+  %8 = phi i8* [ %4, %3 ], [ %6, %5 ]
+  ret i8* %8
+}
+
+define internal i8* @f3(i8* readnone %0) local_unnamed_addr #0 {
+; ATTRIBUTOR: define internal i8* @f3(i8* readnone %0)
+  %2 = icmp eq i8* %0, null
+  br i1 %2, label %3, label %5
+
+; <label>:3:                                      ; preds = %1
+; ATTRIBUTOR: %4 = tail call i8* undef(i8* nonnull align 16 @a2)
+  %4 = tail call i8* @f1(i8* nonnull @a2)
+  br label %5
+
+; <label>:5:                                      ; preds = %1, %3
+  %6 = phi i8* [ %4, %3 ], [ @a1, %1 ]
+  ret i8* %6
+}

llvm/test/Transforms/FunctionAttrs/read_write_returned_arguments_scc.ll

@@ -1,4 +1,4 @@
-; RUN: opt -functionattrs -enable-nonnull-arg-prop -attributor -attributor-disable=false -S < %s | FileCheck %s
+; RUN: opt -functionattrs -enable-nonnull-arg-prop -attributor -attributor-disable=false -attributor-max-iterations=33 -S < %s | FileCheck %s
 ;
 ; This is an evolved example to stress test SCC parameter attribute propagation.
 ; The SCC in this test is made up of the following six function, three of which