1 //===- ObjCARCAnalysisUtils.h - ObjC ARC Analysis Utilities -----*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 /// \file
9 /// This file defines common analysis utilities used by the ObjC ARC Optimizer.
10 /// ARC stands for Automatic Reference Counting and is a system for managing
11 /// reference counts for objects in Objective C.
12 ///
13 /// WARNING: This file knows about certain library functions. It recognizes them
14 /// by name, and hardwires knowledge of their semantics.
15 ///
16 /// WARNING: This file knows about how certain Objective-C library functions are
17 /// used. Naive LLVM IR transformations which would otherwise be
18 /// behavior-preserving may break these assumptions.
19 ///
20 //===----------------------------------------------------------------------===//
21 
22 #ifndef LLVM_ANALYSIS_OBJCARCANALYSISUTILS_H
23 #define LLVM_ANALYSIS_OBJCARCANALYSISUTILS_H
24 
25 #include "llvm/Analysis/ObjCARCInstKind.h"
26 #include "llvm/Analysis/ValueTracking.h"
27 #include "llvm/IR/Constants.h"
28 #include "llvm/IR/Module.h"
29 #include "llvm/IR/ValueHandle.h"
30 #include <optional>
31 
32 namespace llvm {
33 
34 class AAResults;
35 
36 namespace objcarc {
37 
38 /// A handy option to enable/disable all ARC Optimizations.
39 extern bool EnableARCOpts;
40 
41 /// Test if the given module looks interesting to run ARC optimization
42 /// on.
ModuleHasARC(const Module & M)43 inline bool ModuleHasARC(const Module &M) {
44   return
45     M.getNamedValue("llvm.objc.retain") ||
46     M.getNamedValue("llvm.objc.release") ||
47     M.getNamedValue("llvm.objc.autorelease") ||
48     M.getNamedValue("llvm.objc.retainAutoreleasedReturnValue") ||
49     M.getNamedValue("llvm.objc.unsafeClaimAutoreleasedReturnValue") ||
50     M.getNamedValue("llvm.objc.retainBlock") ||
51     M.getNamedValue("llvm.objc.autoreleaseReturnValue") ||
52     M.getNamedValue("llvm.objc.autoreleasePoolPush") ||
53     M.getNamedValue("llvm.objc.loadWeakRetained") ||
54     M.getNamedValue("llvm.objc.loadWeak") ||
55     M.getNamedValue("llvm.objc.destroyWeak") ||
56     M.getNamedValue("llvm.objc.storeWeak") ||
57     M.getNamedValue("llvm.objc.initWeak") ||
58     M.getNamedValue("llvm.objc.moveWeak") ||
59     M.getNamedValue("llvm.objc.copyWeak") ||
60     M.getNamedValue("llvm.objc.retainedObject") ||
61     M.getNamedValue("llvm.objc.unretainedObject") ||
62     M.getNamedValue("llvm.objc.unretainedPointer") ||
63     M.getNamedValue("llvm.objc.clang.arc.use");
64 }
65 
66 /// This is a wrapper around getUnderlyingObject which also knows how to
67 /// look through objc_retain and objc_autorelease calls, which we know to return
68 /// their argument verbatim.
GetUnderlyingObjCPtr(const Value * V)69 inline const Value *GetUnderlyingObjCPtr(const Value *V) {
70   for (;;) {
71     V = getUnderlyingObject(V);
72     if (!IsForwarding(GetBasicARCInstKind(V)))
73       break;
74     V = cast<CallInst>(V)->getArgOperand(0);
75   }
76 
77   return V;
78 }
79 
80 /// A wrapper for GetUnderlyingObjCPtr used for results memoization.
GetUnderlyingObjCPtrCached(const Value * V,DenseMap<const Value *,std::pair<WeakVH,WeakTrackingVH>> & Cache)81 inline const Value *GetUnderlyingObjCPtrCached(
82     const Value *V,
83     DenseMap<const Value *, std::pair<WeakVH, WeakTrackingVH>> &Cache) {
84   // The entry is invalid if either value handle is null.
85   auto InCache = Cache.lookup(V);
86   if (InCache.first && InCache.second)
87     return InCache.second;
88 
89   const Value *Computed = GetUnderlyingObjCPtr(V);
90   Cache[V] =
91       std::make_pair(const_cast<Value *>(V), const_cast<Value *>(Computed));
92   return Computed;
93 }
94 
95 /// The RCIdentity root of a value \p V is a dominating value U for which
96 /// retaining or releasing U is equivalent to retaining or releasing V. In other
97 /// words, ARC operations on \p V are equivalent to ARC operations on \p U.
98 ///
99 /// We use this in the ARC optimizer to make it easier to match up ARC
100 /// operations by always mapping ARC operations to RCIdentityRoots instead of
101 /// pointers themselves.
102 ///
103 /// The two ways that we see RCIdentical values in ObjC are via:
104 ///
105 ///   1. PointerCasts
106 ///   2. Forwarding Calls that return their argument verbatim.
107 ///
108 /// Thus this function strips off pointer casts and forwarding calls. *NOTE*
109 /// This implies that two RCIdentical values must alias.
GetRCIdentityRoot(const Value * V)110 inline const Value *GetRCIdentityRoot(const Value *V) {
111   for (;;) {
112     V = V->stripPointerCasts();
113     if (!IsForwarding(GetBasicARCInstKind(V)))
114       break;
115     V = cast<CallInst>(V)->getArgOperand(0);
116   }
117   return V;
118 }
119 
120 /// Helper which calls const Value *GetRCIdentityRoot(const Value *V) and just
121 /// casts away the const of the result. For documentation about what an
122 /// RCIdentityRoot (and by extension GetRCIdentityRoot is) look at that
123 /// function.
GetRCIdentityRoot(Value * V)124 inline Value *GetRCIdentityRoot(Value *V) {
125   return const_cast<Value *>(GetRCIdentityRoot((const Value *)V));
126 }
127 
128 /// Assuming the given instruction is one of the special calls such as
129 /// objc_retain or objc_release, return the RCIdentity root of the argument of
130 /// the call.
GetArgRCIdentityRoot(Value * Inst)131 inline Value *GetArgRCIdentityRoot(Value *Inst) {
132   return GetRCIdentityRoot(cast<CallInst>(Inst)->getArgOperand(0));
133 }
134 
IsNullOrUndef(const Value * V)135 inline bool IsNullOrUndef(const Value *V) {
136   return isa<ConstantPointerNull>(V) || isa<UndefValue>(V);
137 }
138 
IsNoopInstruction(const Instruction * I)139 inline bool IsNoopInstruction(const Instruction *I) {
140   return isa<BitCastInst>(I) ||
141     (isa<GetElementPtrInst>(I) &&
142      cast<GetElementPtrInst>(I)->hasAllZeroIndices());
143 }
144 
145 /// Test whether the given value is possible a retainable object pointer.
IsPotentialRetainableObjPtr(const Value * Op)146 inline bool IsPotentialRetainableObjPtr(const Value *Op) {
147   // Pointers to static or stack storage are not valid retainable object
148   // pointers.
149   if (isa<Constant>(Op) || isa<AllocaInst>(Op))
150     return false;
151   // Special arguments can not be a valid retainable object pointer.
152   if (const Argument *Arg = dyn_cast<Argument>(Op))
153     if (Arg->hasPassPointeeByValueCopyAttr() || Arg->hasNestAttr() ||
154         Arg->hasStructRetAttr())
155       return false;
156   // Only consider values with pointer types.
157   //
158   // It seemes intuitive to exclude function pointer types as well, since
159   // functions are never retainable object pointers, however clang occasionally
160   // bitcasts retainable object pointers to function-pointer type temporarily.
161   PointerType *Ty = dyn_cast<PointerType>(Op->getType());
162   if (!Ty)
163     return false;
164   // Conservatively assume anything else is a potential retainable object
165   // pointer.
166   return true;
167 }
168 
169 bool IsPotentialRetainableObjPtr(const Value *Op, AAResults &AA);
170 
171 /// Helper for GetARCInstKind. Determines what kind of construct CS
172 /// is.
GetCallSiteClass(const CallBase & CB)173 inline ARCInstKind GetCallSiteClass(const CallBase &CB) {
174   for (const Use &U : CB.args())
175     if (IsPotentialRetainableObjPtr(U))
176       return CB.onlyReadsMemory() ? ARCInstKind::User : ARCInstKind::CallOrUser;
177 
178   return CB.onlyReadsMemory() ? ARCInstKind::None : ARCInstKind::Call;
179 }
180 
181 /// Return true if this value refers to a distinct and identifiable
182 /// object.
183 ///
184 /// This is similar to AliasAnalysis's isIdentifiedObject, except that it uses
185 /// special knowledge of ObjC conventions.
IsObjCIdentifiedObject(const Value * V)186 inline bool IsObjCIdentifiedObject(const Value *V) {
187   // Assume that call results and arguments have their own "provenance".
188   // Constants (including GlobalVariables) and Allocas are never
189   // reference-counted.
190   if (isa<CallInst>(V) || isa<InvokeInst>(V) ||
191       isa<Argument>(V) || isa<Constant>(V) ||
192       isa<AllocaInst>(V))
193     return true;
194 
195   if (const LoadInst *LI = dyn_cast<LoadInst>(V)) {
196     const Value *Pointer =
197       GetRCIdentityRoot(LI->getPointerOperand());
198     if (const GlobalVariable *GV = dyn_cast<GlobalVariable>(Pointer)) {
199       // A constant pointer can't be pointing to an object on the heap. It may
200       // be reference-counted, but it won't be deleted.
201       if (GV->isConstant())
202         return true;
203       StringRef Name = GV->getName();
204       // These special variables are known to hold values which are not
205       // reference-counted pointers.
206       if (Name.startswith("\01l_objc_msgSend_fixup_"))
207         return true;
208 
209       StringRef Section = GV->getSection();
210       if (Section.contains("__message_refs") ||
211           Section.contains("__objc_classrefs") ||
212           Section.contains("__objc_superrefs") ||
213           Section.contains("__objc_methname") || Section.contains("__cstring"))
214         return true;
215     }
216   }
217 
218   return false;
219 }
220 
221 enum class ARCMDKindID {
222   ImpreciseRelease,
223   CopyOnEscape,
224   NoObjCARCExceptions,
225 };
226 
227 /// A cache of MDKinds used by various ARC optimizations.
228 class ARCMDKindCache {
229   Module *M;
230 
231   /// The Metadata Kind for clang.imprecise_release metadata.
232   std::optional<unsigned> ImpreciseReleaseMDKind;
233 
234   /// The Metadata Kind for clang.arc.copy_on_escape metadata.
235   std::optional<unsigned> CopyOnEscapeMDKind;
236 
237   /// The Metadata Kind for clang.arc.no_objc_arc_exceptions metadata.
238   std::optional<unsigned> NoObjCARCExceptionsMDKind;
239 
240 public:
init(Module * Mod)241   void init(Module *Mod) {
242     M = Mod;
243     ImpreciseReleaseMDKind = std::nullopt;
244     CopyOnEscapeMDKind = std::nullopt;
245     NoObjCARCExceptionsMDKind = std::nullopt;
246   }
247 
get(ARCMDKindID ID)248   unsigned get(ARCMDKindID ID) {
249     switch (ID) {
250     case ARCMDKindID::ImpreciseRelease:
251       if (!ImpreciseReleaseMDKind)
252         ImpreciseReleaseMDKind =
253             M->getContext().getMDKindID("clang.imprecise_release");
254       return *ImpreciseReleaseMDKind;
255     case ARCMDKindID::CopyOnEscape:
256       if (!CopyOnEscapeMDKind)
257         CopyOnEscapeMDKind =
258             M->getContext().getMDKindID("clang.arc.copy_on_escape");
259       return *CopyOnEscapeMDKind;
260     case ARCMDKindID::NoObjCARCExceptions:
261       if (!NoObjCARCExceptionsMDKind)
262         NoObjCARCExceptionsMDKind =
263             M->getContext().getMDKindID("clang.arc.no_objc_arc_exceptions");
264       return *NoObjCARCExceptionsMDKind;
265     }
266     llvm_unreachable("Covered switch isn't covered?!");
267   }
268 };
269 
270 } // end namespace objcarc
271 } // end namespace llvm
272 
273 #endif
274