1 //===- GCMetadata.h - Garbage collector metadata ----------------*- 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 //
9 // This file declares the GCFunctionInfo and GCModuleInfo classes, which are
10 // used as a communication channel from the target code generator to the target
11 // garbage collectors. This interface allows code generators and garbage
12 // collectors to be developed independently.
13 //
14 // The GCFunctionInfo class logs the data necessary to build a type accurate
15 // stack map. The code generator outputs:
16 //
17 //   - Safe points as specified by the GCStrategy's NeededSafePoints.
18 //   - Stack offsets for GC roots, as specified by calls to llvm.gcroot
19 //
20 // As a refinement, liveness analysis calculates the set of live roots at each
21 // safe point. Liveness analysis is not presently performed by the code
22 // generator, so all roots are assumed live.
23 //
24 // GCModuleInfo simply collects GCFunctionInfo instances for each Function as
25 // they are compiled. This accretion is necessary for collectors which must emit
26 // a stack map for the compilation unit as a whole. Therefore, GCFunctionInfo
27 // outlives the MachineFunction from which it is derived and must not refer to
28 // any code generator data structures.
29 //
30 //===----------------------------------------------------------------------===//
31 
32 #ifndef LLVM_CODEGEN_GCMETADATA_H
33 #define LLVM_CODEGEN_GCMETADATA_H
34 
35 #include "llvm/ADT/DenseMap.h"
36 #include "llvm/ADT/SmallVector.h"
37 #include "llvm/ADT/StringMap.h"
38 #include "llvm/ADT/StringRef.h"
39 #include "llvm/IR/DebugLoc.h"
40 #include "llvm/IR/GCStrategy.h"
41 #include "llvm/IR/PassManager.h"
42 #include "llvm/Pass.h"
43 #include <algorithm>
44 #include <cstddef>
45 #include <cstdint>
46 #include <memory>
47 #include <vector>
48 
49 namespace llvm {
50 
51 class Constant;
52 class Function;
53 class MCSymbol;
54 
55 /// GCPoint - Metadata for a collector-safe point in machine code.
56 ///
57 struct GCPoint {
58   MCSymbol *Label;    ///< A label.
59   DebugLoc Loc;
60 
GCPointGCPoint61   GCPoint(MCSymbol *L, DebugLoc DL)
62       : Label(L), Loc(std::move(DL)) {}
63 };
64 
65 /// GCRoot - Metadata for a pointer to an object managed by the garbage
66 /// collector.
67 struct GCRoot {
68   int Num;                  ///< Usually a frame index.
69   int StackOffset = -1;     ///< Offset from the stack pointer.
70   const Constant *Metadata; ///< Metadata straight from the call
71                             ///< to llvm.gcroot.
72 
GCRootGCRoot73   GCRoot(int N, const Constant *MD) : Num(N), Metadata(MD) {}
74 };
75 
76 /// Garbage collection metadata for a single function.  Currently, this
77 /// information only applies to GCStrategies which use GCRoot.
78 class GCFunctionInfo {
79 public:
80   using iterator = std::vector<GCPoint>::iterator;
81   using roots_iterator = std::vector<GCRoot>::iterator;
82   using live_iterator = std::vector<GCRoot>::const_iterator;
83 
84 private:
85   const Function &F;
86   GCStrategy &S;
87   uint64_t FrameSize;
88   std::vector<GCRoot> Roots;
89   std::vector<GCPoint> SafePoints;
90 
91   // FIXME: Liveness. A 2D BitVector, perhaps?
92   //
93   //   BitVector Liveness;
94   //
95   //   bool islive(int point, int root) =
96   //     Liveness[point * SafePoints.size() + root]
97   //
98   // The bit vector is the more compact representation where >3.2% of roots
99   // are live per safe point (1.5% on 64-bit hosts).
100 
101 public:
102   GCFunctionInfo(const Function &F, GCStrategy &S);
103   ~GCFunctionInfo();
104 
105   /// Handle invalidation explicitly.
106   bool invalidate(Function &F, const PreservedAnalyses &PA,
107                   FunctionAnalysisManager::Invalidator &Inv);
108 
109   /// getFunction - Return the function to which this metadata applies.
getFunction()110   const Function &getFunction() const { return F; }
111 
112   /// getStrategy - Return the GC strategy for the function.
getStrategy()113   GCStrategy &getStrategy() { return S; }
114 
115   /// addStackRoot - Registers a root that lives on the stack. Num is the
116   ///                stack object ID for the alloca (if the code generator is
117   //                 using  MachineFrameInfo).
addStackRoot(int Num,const Constant * Metadata)118   void addStackRoot(int Num, const Constant *Metadata) {
119     Roots.push_back(GCRoot(Num, Metadata));
120   }
121 
122   /// removeStackRoot - Removes a root.
removeStackRoot(roots_iterator position)123   roots_iterator removeStackRoot(roots_iterator position) {
124     return Roots.erase(position);
125   }
126 
127   /// addSafePoint - Notes the existence of a safe point. Num is the ID of the
128   /// label just prior to the safe point (if the code generator is using
129   /// MachineModuleInfo).
addSafePoint(MCSymbol * Label,const DebugLoc & DL)130   void addSafePoint(MCSymbol *Label, const DebugLoc &DL) {
131     SafePoints.emplace_back(Label, DL);
132   }
133 
134   /// getFrameSize/setFrameSize - Records the function's frame size.
getFrameSize()135   uint64_t getFrameSize() const { return FrameSize; }
setFrameSize(uint64_t S)136   void setFrameSize(uint64_t S) { FrameSize = S; }
137 
138   /// begin/end - Iterators for safe points.
begin()139   iterator begin() { return SafePoints.begin(); }
end()140   iterator end() { return SafePoints.end(); }
size()141   size_t size() const { return SafePoints.size(); }
142 
143   /// roots_begin/roots_end - Iterators for all roots in the function.
roots_begin()144   roots_iterator roots_begin() { return Roots.begin(); }
roots_end()145   roots_iterator roots_end() { return Roots.end(); }
roots_size()146   size_t roots_size() const { return Roots.size(); }
147 
148   /// live_begin/live_end - Iterators for live roots at a given safe point.
live_begin(const iterator & p)149   live_iterator live_begin(const iterator &p) { return roots_begin(); }
live_end(const iterator & p)150   live_iterator live_end(const iterator &p) { return roots_end(); }
live_size(const iterator & p)151   size_t live_size(const iterator &p) const { return roots_size(); }
152 };
153 
154 struct GCStrategyMap {
155   StringMap<std::unique_ptr<GCStrategy>> StrategyMap;
156 
157   GCStrategyMap() = default;
158   GCStrategyMap(GCStrategyMap &&) = default;
159 
160   /// Handle invalidation explicitly.
161   bool invalidate(Module &M, const PreservedAnalyses &PA,
162                   ModuleAnalysisManager::Invalidator &Inv);
163 };
164 
165 /// An analysis pass which caches information about the entire Module.
166 /// Records a cache of the 'active' gc strategy objects for the current Module.
167 class CollectorMetadataAnalysis
168     : public AnalysisInfoMixin<CollectorMetadataAnalysis> {
169   friend struct AnalysisInfoMixin<CollectorMetadataAnalysis>;
170   static AnalysisKey Key;
171 
172 public:
173   using Result = GCStrategyMap;
174   Result run(Module &M, ModuleAnalysisManager &MAM);
175 };
176 
177 /// An analysis pass which caches information about the Function.
178 /// Records the function level information used by GCRoots.
179 /// This pass depends on `CollectorMetadataAnalysis`.
180 class GCFunctionAnalysis : public AnalysisInfoMixin<GCFunctionAnalysis> {
181   friend struct AnalysisInfoMixin<GCFunctionAnalysis>;
182   static AnalysisKey Key;
183 
184 public:
185   using Result = GCFunctionInfo;
186   Result run(Function &F, FunctionAnalysisManager &FAM);
187 };
188 
189 /// An analysis pass which caches information about the entire Module.
190 /// Records both the function level information used by GCRoots and a
191 /// cache of the 'active' gc strategy objects for the current Module.
192 class GCModuleInfo : public ImmutablePass {
193   /// An owning list of all GCStrategies which have been created
194   SmallVector<std::unique_ptr<GCStrategy>, 1> GCStrategyList;
195   /// A helper map to speedup lookups into the above list
196   StringMap<GCStrategy*> GCStrategyMap;
197 
198 public:
199   /// Lookup the GCStrategy object associated with the given gc name.
200   /// Objects are owned internally; No caller should attempt to delete the
201   /// returned objects.
202   GCStrategy *getGCStrategy(const StringRef Name);
203 
204   /// List of per function info objects.  In theory, Each of these
205   /// may be associated with a different GC.
206   using FuncInfoVec = std::vector<std::unique_ptr<GCFunctionInfo>>;
207 
208   FuncInfoVec::iterator funcinfo_begin() { return Functions.begin(); }
209   FuncInfoVec::iterator funcinfo_end() { return Functions.end(); }
210 
211 private:
212   /// Owning list of all GCFunctionInfos associated with this Module
213   FuncInfoVec Functions;
214 
215   /// Non-owning map to bypass linear search when finding the GCFunctionInfo
216   /// associated with a particular Function.
217   using finfo_map_type = DenseMap<const Function *, GCFunctionInfo *>;
218   finfo_map_type FInfoMap;
219 
220 public:
221   using iterator = SmallVector<std::unique_ptr<GCStrategy>, 1>::const_iterator;
222 
223   static char ID;
224 
225   GCModuleInfo();
226 
227   /// clear - Resets the pass. Any pass, which uses GCModuleInfo, should
228   /// call it in doFinalization().
229   ///
230   void clear();
231 
232   /// begin/end - Iterators for used strategies.
233   ///
234   iterator begin() const { return GCStrategyList.begin(); }
235   iterator end() const { return GCStrategyList.end(); }
236 
237   /// get - Look up function metadata.  This is currently assumed
238   /// have the side effect of initializing the associated GCStrategy.  That
239   /// will soon change.
240   GCFunctionInfo &getFunctionInfo(const Function &F);
241 };
242 
243 } // end namespace llvm
244 
245 #endif // LLVM_CODEGEN_GCMETADATA_H
246