1 //=-- InstrProf.cpp - Instrumented profiling format support -----------------=//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // This file contains support for clang's instrumentation based PGO and
11 // coverage.
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "llvm/IR/Constants.h"
16 #include "llvm/IR/Function.h"
17 #include "llvm/IR/Module.h"
18 #include "llvm/IR/GlobalVariable.h"
19 #include "llvm/ProfileData/InstrProf.h"
20 #include "llvm/Support/ErrorHandling.h"
21 #include "llvm/Support/ManagedStatic.h"
22 
23 using namespace llvm;
24 
25 namespace {
26 class InstrProfErrorCategoryType : public std::error_category {
27   const char *name() const LLVM_NOEXCEPT override { return "llvm.instrprof"; }
28   std::string message(int IE) const override {
29     instrprof_error E = static_cast<instrprof_error>(IE);
30     switch (E) {
31     case instrprof_error::success:
32       return "Success";
33     case instrprof_error::eof:
34       return "End of File";
35     case instrprof_error::unrecognized_format:
36       return "Unrecognized instrumentation profile encoding format";
37     case instrprof_error::bad_magic:
38       return "Invalid instrumentation profile data (bad magic)";
39     case instrprof_error::bad_header:
40       return "Invalid instrumentation profile data (file header is corrupt)";
41     case instrprof_error::unsupported_version:
42       return "Unsupported instrumentation profile format version";
43     case instrprof_error::unsupported_hash_type:
44       return "Unsupported instrumentation profile hash type";
45     case instrprof_error::too_large:
46       return "Too much profile data";
47     case instrprof_error::truncated:
48       return "Truncated profile data";
49     case instrprof_error::malformed:
50       return "Malformed instrumentation profile data";
51     case instrprof_error::unknown_function:
52       return "No profile data available for function";
53     case instrprof_error::hash_mismatch:
54       return "Function control flow change detected (hash mismatch)";
55     case instrprof_error::count_mismatch:
56       return "Function basic block count change detected (counter mismatch)";
57     case instrprof_error::counter_overflow:
58       return "Counter overflow";
59     case instrprof_error::value_site_count_mismatch:
60       return "Function value site count change detected (counter mismatch)";
61     }
62     llvm_unreachable("A value of instrprof_error has no message.");
63   }
64 };
65 }
66 
67 static ManagedStatic<InstrProfErrorCategoryType> ErrorCategory;
68 
69 const std::error_category &llvm::instrprof_category() {
70   return *ErrorCategory;
71 }
72 
73 namespace llvm {
74 
75 std::string getPGOFuncName(StringRef RawFuncName,
76                            GlobalValue::LinkageTypes Linkage,
77                            StringRef FileName) {
78 
79   // Function names may be prefixed with a binary '1' to indicate
80   // that the backend should not modify the symbols due to any platform
81   // naming convention. Do not include that '1' in the PGO profile name.
82   if (RawFuncName[0] == '\1')
83     RawFuncName = RawFuncName.substr(1);
84 
85   std::string FuncName = RawFuncName;
86   if (llvm::GlobalValue::isLocalLinkage(Linkage)) {
87     // For local symbols, prepend the main file name to distinguish them.
88     // Do not include the full path in the file name since there's no guarantee
89     // that it will stay the same, e.g., if the files are checked out from
90     // version control in different locations.
91     if (FileName.empty())
92       FuncName = FuncName.insert(0, "<unknown>:");
93     else
94       FuncName = FuncName.insert(0, FileName.str() + ":");
95   }
96   return FuncName;
97 }
98 
99 std::string getPGOFuncName(const Function &F) {
100   return getPGOFuncName(F.getName(), F.getLinkage(), F.getParent()->getName());
101 }
102 
103 GlobalVariable *createPGOFuncNameVar(Module &M,
104                                      GlobalValue::LinkageTypes Linkage,
105                                      StringRef FuncName) {
106 
107   // We generally want to match the function's linkage, but available_externally
108   // and extern_weak both have the wrong semantics, and anything that doesn't
109   // need to link across compilation units doesn't need to be visible at all.
110   if (Linkage == GlobalValue::ExternalWeakLinkage)
111     Linkage = GlobalValue::LinkOnceAnyLinkage;
112   else if (Linkage == GlobalValue::AvailableExternallyLinkage)
113     Linkage = GlobalValue::LinkOnceODRLinkage;
114   else if (Linkage == GlobalValue::InternalLinkage ||
115            Linkage == GlobalValue::ExternalLinkage)
116     Linkage = GlobalValue::PrivateLinkage;
117 
118   auto *Value = ConstantDataArray::getString(M.getContext(), FuncName, false);
119   auto FuncNameVar =
120       new GlobalVariable(M, Value->getType(), true, Linkage, Value,
121                          Twine(getInstrProfNameVarPrefix()) + FuncName);
122 
123   // Hide the symbol so that we correctly get a copy for each executable.
124   if (!GlobalValue::isLocalLinkage(FuncNameVar->getLinkage()))
125     FuncNameVar->setVisibility(GlobalValue::HiddenVisibility);
126 
127   return FuncNameVar;
128 }
129 
130 GlobalVariable *createPGOFuncNameVar(Function &F, StringRef FuncName) {
131   return createPGOFuncNameVar(*F.getParent(), F.getLinkage(), FuncName);
132 }
133 
134 uint32_t ValueProfRecord::getHeaderSize(uint32_t NumValueSites) {
135   uint32_t Size = offsetof(ValueProfRecord, SiteCountArray) +
136                   sizeof(uint8_t) * NumValueSites;
137   // Round the size to multiple of 8 bytes.
138   Size = (Size + 7) & ~7;
139   return Size;
140 }
141 
142 uint32_t ValueProfRecord::getSize(uint32_t NumValueSites,
143                                   uint32_t NumValueData) {
144   return getHeaderSize(NumValueSites) +
145          sizeof(InstrProfValueData) * NumValueData;
146 }
147 
148 void ValueProfRecord::deserializeTo(InstrProfRecord &Record,
149                                     InstrProfRecord::ValueMapType *VMap) {
150   Record.reserveSites(Kind, NumValueSites);
151 
152   InstrProfValueData *ValueData = this->getValueData();
153   for (uint64_t VSite = 0; VSite < NumValueSites; ++VSite) {
154     uint8_t ValueDataCount = this->SiteCountArray[VSite];
155     Record.addValueData(Kind, VSite, ValueData, ValueDataCount, VMap);
156     ValueData += ValueDataCount;
157   }
158 }
159 
160 void ValueProfRecord::serializeFrom(const InstrProfRecord &Record,
161                                     uint32_t ValueKind,
162                                     uint32_t NumValueSites) {
163   Kind = ValueKind;
164   this->NumValueSites = NumValueSites;
165   InstrProfValueData *DstVD = getValueData();
166   for (uint32_t S = 0; S < NumValueSites; S++) {
167     uint32_t ND = Record.getNumValueDataForSite(ValueKind, S);
168     SiteCountArray[S] = ND;
169     std::unique_ptr<InstrProfValueData[]> SrcVD =
170         Record.getValueForSite(ValueKind, S);
171     for (uint32_t I = 0; I < ND; I++) {
172       DstVD[I] = SrcVD[I];
173       switch (ValueKind) {
174       case IPVK_IndirectCallTarget:
175         DstVD[I].Value = IndexedInstrProf::ComputeHash(
176             IndexedInstrProf::HashType, (const char *)DstVD[I].Value);
177         break;
178       default:
179         llvm_unreachable("value kind not handled !");
180       }
181     }
182     DstVD += ND;
183   }
184 }
185 
186 template <class T>
187 static T swapToHostOrder(const unsigned char *&D, support::endianness Orig) {
188   using namespace support;
189   if (Orig == little)
190     return endian::readNext<T, little, unaligned>(D);
191   else
192     return endian::readNext<T, big, unaligned>(D);
193 }
194 
195 // For writing/serializing,  Old is the host endianness, and  New is
196 // byte order intended on disk. For Reading/deserialization, Old
197 // is the on-disk source endianness, and New is the host endianness.
198 void ValueProfRecord::swapBytes(support::endianness Old,
199                                 support::endianness New) {
200   using namespace support;
201   if (Old == New)
202     return;
203 
204   if (getHostEndianness() != Old) {
205     sys::swapByteOrder<uint32_t>(NumValueSites);
206     sys::swapByteOrder<uint32_t>(Kind);
207   }
208   uint32_t ND = getNumValueData();
209   InstrProfValueData *VD = getValueData();
210 
211   // No need to swap byte array: SiteCountArrray.
212   for (uint32_t I = 0; I < ND; I++) {
213     sys::swapByteOrder<uint64_t>(VD[I].Value);
214     sys::swapByteOrder<uint64_t>(VD[I].Count);
215   }
216   if (getHostEndianness() == Old) {
217     sys::swapByteOrder<uint32_t>(NumValueSites);
218     sys::swapByteOrder<uint32_t>(Kind);
219   }
220 }
221 
222 uint32_t ValueProfData::getSize(const InstrProfRecord &Record) {
223   uint32_t TotalSize = sizeof(ValueProfData);
224   uint32_t NumValueKinds = Record.getNumValueKinds();
225   if (NumValueKinds == 0)
226     return TotalSize;
227 
228   for (uint32_t Kind = IPVK_First; Kind <= IPVK_Last; Kind++) {
229     uint32_t NumValueSites = Record.getNumValueSites(Kind);
230     if (!NumValueSites)
231       continue;
232     TotalSize +=
233         ValueProfRecord::getSize(NumValueSites, Record.getNumValueData(Kind));
234   }
235   return TotalSize;
236 }
237 
238 void ValueProfData::deserializeTo(InstrProfRecord &Record,
239                                   InstrProfRecord::ValueMapType *VMap) {
240   if (NumValueKinds == 0)
241     return;
242 
243   ValueProfRecord *VR = getFirstValueProfRecord();
244   for (uint32_t K = 0; K < NumValueKinds; K++) {
245     VR->deserializeTo(Record, VMap);
246     VR = VR->getNext();
247   }
248 }
249 
250 static std::unique_ptr<ValueProfData> AllocValueProfData(uint32_t TotalSize) {
251   return std::unique_ptr<ValueProfData>(new (::operator new(TotalSize))
252                                             ValueProfData());
253 }
254 
255 std::unique_ptr<ValueProfData>
256 ValueProfData::serializeFrom(const InstrProfRecord &Record) {
257   uint32_t TotalSize = getSize(Record);
258 
259   std::unique_ptr<ValueProfData> VPD = AllocValueProfData(TotalSize);
260 
261   VPD->TotalSize = TotalSize;
262   VPD->NumValueKinds = Record.getNumValueKinds();
263   ValueProfRecord *VR = VPD->getFirstValueProfRecord();
264   for (uint32_t Kind = IPVK_First; Kind <= IPVK_Last; Kind++) {
265     uint32_t NumValueSites = Record.getNumValueSites(Kind);
266     if (!NumValueSites)
267       continue;
268     VR->serializeFrom(Record, Kind, NumValueSites);
269     VR = VR->getNext();
270   }
271   return VPD;
272 }
273 
274 ErrorOr<std::unique_ptr<ValueProfData>>
275 ValueProfData::getValueProfData(const unsigned char *D,
276                                 const unsigned char *const BufferEnd,
277                                 support::endianness Endianness) {
278   using namespace support;
279   if (D + sizeof(ValueProfData) > BufferEnd)
280     return instrprof_error::truncated;
281 
282   const unsigned char *Header = D;
283   uint32_t TotalSize = swapToHostOrder<uint32_t>(Header, Endianness);
284   uint32_t NumValueKinds = swapToHostOrder<uint32_t>(Header, Endianness);
285 
286   if (D + TotalSize > BufferEnd)
287     return instrprof_error::too_large;
288   if (NumValueKinds > IPVK_Last + 1)
289     return instrprof_error::malformed;
290   // Total size needs to be mulltiple of quadword size.
291   if (TotalSize % sizeof(uint64_t))
292     return instrprof_error::malformed;
293 
294   std::unique_ptr<ValueProfData> VPD = AllocValueProfData(TotalSize);
295 
296   memcpy(VPD.get(), D, TotalSize);
297   // Byte swap.
298   VPD->swapBytesToHost(Endianness);
299 
300   // Data integrety check:
301   ValueProfRecord *VR = VPD->getFirstValueProfRecord();
302   for (uint32_t K = 0; K < VPD->NumValueKinds; K++) {
303     if (VR->Kind > IPVK_Last)
304       return instrprof_error::malformed;
305     VR = VR->getNext();
306     if ((char *)VR - (char *)VPD.get() > (ptrdiff_t)TotalSize)
307       return instrprof_error::malformed;
308   }
309 
310   return std::move(VPD);
311 }
312 
313 void ValueProfData::swapBytesToHost(support::endianness Endianness) {
314   using namespace support;
315   if (Endianness == getHostEndianness())
316     return;
317 
318   sys::swapByteOrder<uint32_t>(TotalSize);
319   sys::swapByteOrder<uint32_t>(NumValueKinds);
320 
321   ValueProfRecord *VR = getFirstValueProfRecord();
322   for (uint32_t K = 0; K < NumValueKinds; K++) {
323     VR->swapBytes(Endianness, getHostEndianness());
324     VR = VR->getNext();
325   }
326 }
327 
328 void ValueProfData::swapBytesFromHost(support::endianness Endianness) {
329   using namespace support;
330   if (Endianness == getHostEndianness())
331     return;
332 
333   ValueProfRecord *VR = getFirstValueProfRecord();
334   for (uint32_t K = 0; K < NumValueKinds; K++) {
335     ValueProfRecord *NVR = VR->getNext();
336     VR->swapBytes(getHostEndianness(), Endianness);
337     VR = NVR;
338   }
339   sys::swapByteOrder<uint32_t>(TotalSize);
340   sys::swapByteOrder<uint32_t>(NumValueKinds);
341 }
342 
343 ValueProfRecord *ValueProfData::getFirstValueProfRecord() {
344   return reinterpret_cast<ValueProfRecord *>((char *)this +
345                                              sizeof(ValueProfData));
346 }
347 
348 uint32_t ValueProfRecord::getNumValueData() const {
349   uint32_t NumValueData = 0;
350   for (uint32_t I = 0; I < NumValueSites; I++)
351     NumValueData += SiteCountArray[I];
352   return NumValueData;
353 }
354 
355 ValueProfRecord *ValueProfRecord::getNext() {
356   return reinterpret_cast<ValueProfRecord *>((char *)this + getSize());
357 }
358 
359 InstrProfValueData *ValueProfRecord::getValueData() {
360   return reinterpret_cast<InstrProfValueData *>((char *)this +
361                                                 getHeaderSize(NumValueSites));
362 }
363 }
364