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 hash mismatch";
55     case instrprof_error::count_mismatch:
56       return "Function count mismatch";
57     case instrprof_error::counter_overflow:
58       return "Counter overflow";
59     case instrprof_error::value_site_count_mismatch:
60       return "Function's value site counts 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 namespace IndexedInstrProf {
135 
136 uint32_t ValueProfRecord::getHeaderSize(uint32_t NumValueSites) {
137   uint32_t Size = offsetof(ValueProfRecord, SiteCountArray) +
138                   sizeof(uint8_t) * NumValueSites;
139   // Round the size to multiple of 8 bytes.
140   Size = (Size + 7) & ~7;
141   return Size;
142 }
143 
144 uint32_t ValueProfRecord::getSize(uint32_t NumValueSites,
145                                   uint32_t NumValueData) {
146   return getHeaderSize(NumValueSites) +
147          sizeof(InstrProfValueData) * NumValueData;
148 }
149 
150 void ValueProfRecord::deserializeTo(InstrProfRecord &Record,
151                                     InstrProfRecord::ValueMapType *VMap) {
152   Record.reserveSites(Kind, NumValueSites);
153 
154   InstrProfValueData *ValueData = this->getValueData();
155   for (uint64_t VSite = 0; VSite < NumValueSites; ++VSite) {
156     uint8_t ValueDataCount = this->SiteCountArray[VSite];
157     Record.addValueData(Kind, VSite, ValueData, ValueDataCount, VMap);
158     ValueData += ValueDataCount;
159   }
160 }
161 
162 void ValueProfRecord::serializeFrom(const InstrProfRecord &Record,
163                                     uint32_t ValueKind,
164                                     uint32_t NumValueSites) {
165   Kind = ValueKind;
166   this->NumValueSites = NumValueSites;
167   InstrProfValueData *DstVD = getValueData();
168   for (uint32_t S = 0; S < NumValueSites; S++) {
169     uint32_t ND = Record.getNumValueDataForSite(ValueKind, S);
170     SiteCountArray[S] = ND;
171     std::unique_ptr<InstrProfValueData[]> SrcVD =
172         Record.getValueForSite(ValueKind, S);
173     for (uint32_t I = 0; I < ND; I++) {
174       DstVD[I] = SrcVD[I];
175       switch (ValueKind) {
176       case IPVK_IndirectCallTarget:
177         DstVD[I].Value = ComputeHash(HashType, (const char *)DstVD[I].Value);
178         break;
179       default:
180         llvm_unreachable("value kind not handled !");
181       }
182     }
183     DstVD += ND;
184   }
185 }
186 
187 template <class T>
188 static T swapToHostOrder(const unsigned char *&D, support::endianness Orig) {
189   using namespace support;
190   if (Orig == little)
191     return endian::readNext<T, little, unaligned>(D);
192   else
193     return endian::readNext<T, big, unaligned>(D);
194 }
195 
196 // For writing/serializing,  Old is the host endianness, and  New is
197 // byte order intended on disk. For Reading/deserialization, Old
198 // is the on-disk source endianness, and New is the host endianness.
199 void ValueProfRecord::swapBytes(support::endianness Old,
200                                 support::endianness New) {
201   using namespace support;
202   if (Old == New)
203     return;
204 
205   if (getHostEndianness() != Old) {
206     sys::swapByteOrder<uint32_t>(NumValueSites);
207     sys::swapByteOrder<uint32_t>(Kind);
208   }
209   uint32_t ND = getNumValueData();
210   InstrProfValueData *VD = getValueData();
211 
212   // No need to swap byte array: SiteCountArrray.
213   for (uint32_t I = 0; I < ND; I++) {
214     sys::swapByteOrder<uint64_t>(VD[I].Value);
215     sys::swapByteOrder<uint64_t>(VD[I].Count);
216   }
217   if (getHostEndianness() == Old) {
218     sys::swapByteOrder<uint32_t>(NumValueSites);
219     sys::swapByteOrder<uint32_t>(Kind);
220   }
221 }
222 
223 uint32_t ValueProfData::getSize(const InstrProfRecord &Record) {
224   uint32_t TotalSize = sizeof(ValueProfData);
225   uint32_t NumValueKinds = Record.getNumValueKinds();
226   if (NumValueKinds == 0)
227     return TotalSize;
228 
229   for (uint32_t Kind = IPVK_First; Kind <= IPVK_Last; Kind++) {
230     uint32_t NumValueSites = Record.getNumValueSites(Kind);
231     if (!NumValueSites)
232       continue;
233     TotalSize +=
234         ValueProfRecord::getSize(NumValueSites, Record.getNumValueData(Kind));
235   }
236   return TotalSize;
237 }
238 
239 void ValueProfData::deserializeTo(InstrProfRecord &Record,
240                                   InstrProfRecord::ValueMapType *VMap) {
241   if (NumValueKinds == 0)
242     return;
243 
244   ValueProfRecord *VR = getFirstValueProfRecord();
245   for (uint32_t K = 0; K < NumValueKinds; K++) {
246     VR->deserializeTo(Record, VMap);
247     VR = VR->getNext();
248   }
249 }
250 
251 static std::unique_ptr<ValueProfData> AllocValueProfData(uint32_t TotalSize) {
252   return std::unique_ptr<ValueProfData>(new (::operator new(TotalSize))
253                                             ValueProfData());
254 }
255 
256 std::unique_ptr<ValueProfData>
257 ValueProfData::serializeFrom(const InstrProfRecord &Record) {
258   uint32_t TotalSize = getSize(Record);
259 
260   std::unique_ptr<ValueProfData> VPD = AllocValueProfData(TotalSize);
261 
262   VPD->TotalSize = TotalSize;
263   VPD->NumValueKinds = Record.getNumValueKinds();
264   ValueProfRecord *VR = VPD->getFirstValueProfRecord();
265   for (uint32_t Kind = IPVK_First; Kind <= IPVK_Last; Kind++) {
266     uint32_t NumValueSites = Record.getNumValueSites(Kind);
267     if (!NumValueSites)
268       continue;
269     VR->serializeFrom(Record, Kind, NumValueSites);
270     VR = VR->getNext();
271   }
272   return VPD;
273 }
274 
275 ErrorOr<std::unique_ptr<ValueProfData>>
276 ValueProfData::getValueProfData(const unsigned char *D,
277                                 const unsigned char *const BufferEnd,
278                                 support::endianness Endianness) {
279   using namespace support;
280   if (D + sizeof(ValueProfData) > BufferEnd)
281     return instrprof_error::truncated;
282 
283   const unsigned char *Header = D;
284   uint32_t TotalSize = swapToHostOrder<uint32_t>(Header, Endianness);
285   uint32_t NumValueKinds = swapToHostOrder<uint32_t>(Header, Endianness);
286 
287   if (D + TotalSize > BufferEnd)
288     return instrprof_error::too_large;
289   if (NumValueKinds > IPVK_Last + 1)
290     return instrprof_error::malformed;
291   // Total size needs to be mulltiple of quadword size.
292   if (TotalSize % sizeof(uint64_t))
293     return instrprof_error::malformed;
294 
295   std::unique_ptr<ValueProfData> VPD = AllocValueProfData(TotalSize);
296 
297   memcpy(VPD.get(), D, TotalSize);
298   // Byte swap.
299   VPD->swapBytesToHost(Endianness);
300 
301   // Data integrety check:
302   ValueProfRecord *VR = VPD->getFirstValueProfRecord();
303   for (uint32_t K = 0; K < VPD->NumValueKinds; K++) {
304     if (VR->Kind > IPVK_Last)
305       return instrprof_error::malformed;
306     VR = VR->getNext();
307     if ((char *)VR - (char *)VPD.get() > (ptrdiff_t)TotalSize)
308       return instrprof_error::malformed;
309   }
310 
311   return std::move(VPD);
312 }
313 
314 void ValueProfData::swapBytesToHost(support::endianness Endianness) {
315   using namespace support;
316   if (Endianness == getHostEndianness())
317     return;
318 
319   sys::swapByteOrder<uint32_t>(TotalSize);
320   sys::swapByteOrder<uint32_t>(NumValueKinds);
321 
322   ValueProfRecord *VR = getFirstValueProfRecord();
323   for (uint32_t K = 0; K < NumValueKinds; K++) {
324     VR->swapBytes(Endianness, getHostEndianness());
325     VR = VR->getNext();
326   }
327 }
328 
329 void ValueProfData::swapBytesFromHost(support::endianness Endianness) {
330   using namespace support;
331   if (Endianness == getHostEndianness())
332     return;
333 
334   ValueProfRecord *VR = getFirstValueProfRecord();
335   for (uint32_t K = 0; K < NumValueKinds; K++) {
336     ValueProfRecord *NVR = VR->getNext();
337     VR->swapBytes(getHostEndianness(), Endianness);
338     VR = NVR;
339   }
340   sys::swapByteOrder<uint32_t>(TotalSize);
341   sys::swapByteOrder<uint32_t>(NumValueKinds);
342 }
343 
344 ValueProfRecord *ValueProfData::getFirstValueProfRecord() {
345   return reinterpret_cast<ValueProfRecord *>((char *)this +
346                                              sizeof(ValueProfData));
347 }
348 
349 uint32_t ValueProfRecord::getNumValueData() const {
350   uint32_t NumValueData = 0;
351   for (uint32_t I = 0; I < NumValueSites; I++)
352     NumValueData += SiteCountArray[I];
353   return NumValueData;
354 }
355 
356 ValueProfRecord *ValueProfRecord::getNext() {
357   return reinterpret_cast<ValueProfRecord *>((char *)this + getSize());
358 }
359 
360 InstrProfValueData *ValueProfRecord::getValueData() {
361   return reinterpret_cast<InstrProfValueData *>((char *)this +
362                                                 getHeaderSize(NumValueSites));
363 }
364 
365 } // End of IndexedInstrProf namespace.
366 }
367