lib/Support/MemoryBuffer.cpp

   1 //===--- MemoryBuffer.cpp - Memory Buffer implementation ------------------===//
   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 implements the MemoryBuffer interface.
  11 //
  12 //===----------------------------------------------------------------------===//
  13
  14 #include "llvm/Support/MemoryBuffer.h"
  15 #include "llvm/ADT/OwningPtr.h"
  16 #include "llvm/ADT/SmallString.h"
  17 #include "llvm/Support/MathExtras.h"
  18 #include "llvm/Support/Errno.h"
  19 #include "llvm/Support/Path.h"
  20 #include "llvm/Support/Process.h"
  21 #include "llvm/Support/Program.h"
  22 #include "llvm/Support/system_error.h"
  23 #include <cassert>
  24 #include <cstdio>
  25 #include <cstring>
  26 #include <cerrno>
  27 #include <new>
  28 #include <sys/types.h>
  29 #include <sys/stat.h>
  30 #if !defined(_MSC_VER) && !defined(__MINGW32__)
  31 #include <unistd.h>
  32 #include <sys/uio.h>
  33 #else
  34 #include <io.h>
  35 #endif
  36 #include <fcntl.h>
  37 using namespace llvm;
  38
  39 namespace { const llvm::error_code success; }
  40
  41 //===----------------------------------------------------------------------===//
  42 // MemoryBuffer implementation itself.
  43 //===----------------------------------------------------------------------===//
  44
  45 MemoryBuffer::~MemoryBuffer() { }
  46
  47 /// init - Initialize this MemoryBuffer as a reference to externally allocated
  48 /// memory, memory that we know is already null terminated.
  49 void MemoryBuffer::init(const char *BufStart, const char *BufEnd,
  50                         bool RequiresNullTerminator) {
  51   assert((!RequiresNullTerminator || BufEnd[0] == 0) &&
  52          "Buffer is not null terminated!");
  53   BufferStart = BufStart;
  54   BufferEnd = BufEnd;
  55 }
  56
  57 //===----------------------------------------------------------------------===//
  58 // MemoryBufferMem implementation.
  59 //===----------------------------------------------------------------------===//
  60
  61 /// CopyStringRef - Copies contents of a StringRef into a block of memory and
  62 /// null-terminates it.
  63 static void CopyStringRef(char *Memory, StringRef Data) {
  64   memcpy(Memory, Data.data(), Data.size());
  65   Memory[Data.size()] = 0; // Null terminate string.
  66 }
  67
  68 /// GetNamedBuffer - Allocates a new MemoryBuffer with Name copied after it.
  69 template <typename T>
  70 static T *GetNamedBuffer(StringRef Buffer, StringRef Name,
  71                          bool RequiresNullTerminator) {
  72   char *Mem = static_cast<char*>(operator new(sizeof(T) + Name.size() + 1));
  73   CopyStringRef(Mem + sizeof(T), Name);
  74   return new (Mem) T(Buffer, RequiresNullTerminator);
  75 }
  76
  77 namespace {
  78 /// MemoryBufferMem - Named MemoryBuffer pointing to a block of memory.
  79 class MemoryBufferMem : public MemoryBuffer {
  80 public:
  81   MemoryBufferMem(StringRef InputData, bool RequiresNullTerminator) {
  82     init(InputData.begin(), InputData.end(), RequiresNullTerminator);
  83   }
  84
  85   virtual const char *getBufferIdentifier() const {
  86      // The name is stored after the class itself.
  87     return reinterpret_cast<const char*>(this + 1);
  88   }
  89
  90   virtual BufferKind getBufferKind() const {
  91     return MemoryBuffer_Malloc;
  92   }
  93 };
  94 }
  95
  96 /// getMemBuffer - Open the specified memory range as a MemoryBuffer.  Note
  97 /// that InputData must be a null terminated if RequiresNullTerminator is true!
  98 MemoryBuffer *MemoryBuffer::getMemBuffer(StringRef InputData,
  99                                          StringRef BufferName,
 100                                          bool RequiresNullTerminator) {
 101   return GetNamedBuffer<MemoryBufferMem>(InputData, BufferName,
 102                                          RequiresNullTerminator);
 103 }
 104
 105 /// getMemBufferCopy - Open the specified memory range as a MemoryBuffer,
 106 /// copying the contents and taking ownership of it.  This has no requirements
 107 /// on EndPtr[0].
 108 MemoryBuffer *MemoryBuffer::getMemBufferCopy(StringRef InputData,
 109                                              StringRef BufferName) {
 110   MemoryBuffer *Buf = getNewUninitMemBuffer(InputData.size(), BufferName);
 111   if (!Buf) return 0;
 112   memcpy(const_cast<char*>(Buf->getBufferStart()), InputData.data(),
 113          InputData.size());
 114   return Buf;
 115 }
 116
 117 /// getNewUninitMemBuffer - Allocate a new MemoryBuffer of the specified size
 118 /// that is not initialized.  Note that the caller should initialize the
 119 /// memory allocated by this method.  The memory is owned by the MemoryBuffer
 120 /// object.
 121 MemoryBuffer *MemoryBuffer::getNewUninitMemBuffer(size_t Size,
 122                                                   StringRef BufferName) {
 123   // Allocate space for the MemoryBuffer, the data and the name. It is important
 124   // that MemoryBuffer and data are aligned so PointerIntPair works with them.
 125   size_t AlignedStringLen =
 126     RoundUpToAlignment(sizeof(MemoryBufferMem) + BufferName.size() + 1,
 127                        sizeof(void*)); // TODO: Is sizeof(void*) enough?
 128   size_t RealLen = AlignedStringLen + Size + 1;
 129   char *Mem = static_cast<char*>(operator new(RealLen, std::nothrow));
 130   if (!Mem) return 0;
 131
 132   // The name is stored after the class itself.
 133   CopyStringRef(Mem + sizeof(MemoryBufferMem), BufferName);
 134
 135   // The buffer begins after the name and must be aligned.
 136   char *Buf = Mem + AlignedStringLen;
 137   Buf[Size] = 0; // Null terminate buffer.
 138
 139   return new (Mem) MemoryBufferMem(StringRef(Buf, Size), true);
 140 }
 141
 142 /// getNewMemBuffer - Allocate a new MemoryBuffer of the specified size that
 143 /// is completely initialized to zeros.  Note that the caller should
 144 /// initialize the memory allocated by this method.  The memory is owned by
 145 /// the MemoryBuffer object.
 146 MemoryBuffer *MemoryBuffer::getNewMemBuffer(size_t Size, StringRef BufferName) {
 147   MemoryBuffer *SB = getNewUninitMemBuffer(Size, BufferName);
 148   if (!SB) return 0;
 149   memset(const_cast<char*>(SB->getBufferStart()), 0, Size);
 150   return SB;
 151 }
 152
 153
 154 /// getFileOrSTDIN - Open the specified file as a MemoryBuffer, or open stdin
 155 /// if the Filename is "-".  If an error occurs, this returns null and fills
 156 /// in *ErrStr with a reason.  If stdin is empty, this API (unlike getSTDIN)
 157 /// returns an empty buffer.
 158 error_code MemoryBuffer::getFileOrSTDIN(StringRef Filename,
 159                                         OwningPtr<MemoryBuffer> &result,
 160                                         int64_t FileSize) {
 161   if (Filename == "-")
 162     return getSTDIN(result);
 163   return getFile(Filename, result, FileSize);
 164 }
 165
 166 error_code MemoryBuffer::getFileOrSTDIN(const char *Filename,
 167                                         OwningPtr<MemoryBuffer> &result,
 168                                         int64_t FileSize) {
 169   if (strcmp(Filename, "-") == 0)
 170     return getSTDIN(result);
 171   return getFile(Filename, result, FileSize);
 172 }
 173
 174 //===----------------------------------------------------------------------===//
 175 // MemoryBuffer::getFile implementation.
 176 //===----------------------------------------------------------------------===//
 177
 178 namespace {
 179 /// MemoryBufferMMapFile - This represents a file that was mapped in with the
 180 /// sys::Path::MapInFilePages method.  When destroyed, it calls the
 181 /// sys::Path::UnMapFilePages method.
 182 class MemoryBufferMMapFile : public MemoryBufferMem {
 183 public:
 184   MemoryBufferMMapFile(StringRef Buffer, bool RequiresNullTerminator)
 185     : MemoryBufferMem(Buffer, RequiresNullTerminator) { }
 186
 187   ~MemoryBufferMMapFile() {
 188     static int PageSize = sys::Process::GetPageSize();
 189
 190     uintptr_t Start = reinterpret_cast<uintptr_t>(getBufferStart());
 191     size_t Size = getBufferSize();
 192     uintptr_t RealStart = Start & ~(PageSize - 1);
 193     size_t RealSize = Size + (Start - RealStart);
 194
 195     sys::Path::UnMapFilePages(reinterpret_cast<const char*>(RealStart),
 196                               RealSize);
 197   }
 198
 199   virtual BufferKind getBufferKind() const {
 200     return MemoryBuffer_MMap;
 201   }
 202 };
 203 }
 204
 205 error_code MemoryBuffer::getFile(StringRef Filename,
 206                                  OwningPtr<MemoryBuffer> &result,
 207                                  int64_t FileSize,
 208                                  bool RequiresNullTerminator) {
 209   // Ensure the path is null terminated.
 210   SmallString<256> PathBuf(Filename.begin(), Filename.end());
 211   return MemoryBuffer::getFile(PathBuf.c_str(), result, FileSize,
 212                                RequiresNullTerminator);
 213 }
 214
 215 error_code MemoryBuffer::getFile(const char *Filename,
 216                                  OwningPtr<MemoryBuffer> &result,
 217                                  int64_t FileSize,
 218                                  bool RequiresNullTerminator) {
 219   int OpenFlags = O_RDONLY;
 220 #ifdef O_BINARY
 221   OpenFlags |= O_BINARY;  // Open input file in binary mode on win32.
 222 #endif
 223   int FD = ::open(Filename, OpenFlags);
 224   if (FD == -1)
 225     return error_code(errno, posix_category());
 226
 227   error_code ret = getOpenFile(FD, Filename, result, FileSize, FileSize,
 228                                0, RequiresNullTerminator);
 229   close(FD);
 230   return ret;
 231 }
 232
 233 static bool shouldUseMmap(int FD,
 234                           size_t FileSize,
 235                           size_t MapSize,
 236                           off_t Offset,
 237                           bool RequiresNullTerminator,
 238                           int PageSize) {
 239   // We don't use mmap for small files because this can severely fragment our
 240   // address space.
 241   if (MapSize < 4096*4)
 242     return false;
 243
 244   if (!RequiresNullTerminator)
 245     return true;
 246
 247
 248   // If we don't know the file size, use fstat to find out.  fstat on an open
 249   // file descriptor is cheaper than stat on a random path.
 250   // FIXME: this chunk of code is duplicated, but it avoids a fstat when
 251   // RequiresNullTerminator = false and MapSize != -1.
 252   if (FileSize == size_t(-1)) {
 253     struct stat FileInfo;
 254     // TODO: This should use fstat64 when available.
 255     if (fstat(FD, &FileInfo) == -1) {
 256       return error_code(errno, posix_category());
 257     }
 258     FileSize = FileInfo.st_size;
 259   }
 260
 261   // If we need a null terminator and the end of the map is inside the file,
 262   // we cannot use mmap.
 263   size_t End = Offset + MapSize;
 264   assert(End <= FileSize);
 265   if (End != FileSize)
 266     return false;
 267
 268   // Don't try to map files that are exactly a multiple of the system page size
 269   // if we need a null terminator.
 270   if ((FileSize & (PageSize -1)) == 0)
 271     return false;
 272
 273   return true;
 274 }
 275
 276 error_code MemoryBuffer::getOpenFile(int FD, const char *Filename,
 277                                      OwningPtr<MemoryBuffer> &result,
 278                                      size_t FileSize, size_t MapSize,
 279                                      off_t Offset,
 280                                      bool RequiresNullTerminator) {
 281   static int PageSize = sys::Process::GetPageSize();
 282
 283   // Default is to map the full file.
 284   if (MapSize == size_t(-1)) {
 285     // If we don't know the file size, use fstat to find out.  fstat on an open
 286     // file descriptor is cheaper than stat on a random path.
 287     if (FileSize == size_t(-1)) {
 288       struct stat FileInfo;
 289       // TODO: This should use fstat64 when available.
 290       if (fstat(FD, &FileInfo) == -1) {
 291         return error_code(errno, posix_category());
 292       }
 293       FileSize = FileInfo.st_size;
 294     }
 295     MapSize = FileSize;
 296   }
 297
 298   if (shouldUseMmap(FD, FileSize, MapSize, Offset, RequiresNullTerminator,
 299                     PageSize)) {
 300     off_t RealMapOffset = Offset & ~(PageSize - 1);
 301     off_t Delta = Offset - RealMapOffset;
 302     size_t RealMapSize = MapSize + Delta;
 303
 304     if (const char *Pages = sys::Path::MapInFilePages(FD,
 305                                                       RealMapSize,
 306                                                       RealMapOffset)) {
 307       result.reset(GetNamedBuffer<MemoryBufferMMapFile>(
 308           StringRef(Pages + Delta, MapSize), Filename, RequiresNullTerminator));
 309       return success;
 310     }
 311   }
 312
 313   MemoryBuffer *Buf = MemoryBuffer::getNewUninitMemBuffer(MapSize, Filename);
 314   if (!Buf) {
 315     // Failed to create a buffer. The only way it can fail is if
 316     // new(std::nothrow) returns 0.
 317     return make_error_code(errc::not_enough_memory);
 318   }
 319
 320   OwningPtr<MemoryBuffer> SB(Buf);
 321   char *BufPtr = const_cast<char*>(SB->getBufferStart());
 322
 323   size_t BytesLeft = MapSize;
 324   if (lseek(FD, Offset, SEEK_SET) == -1)
 325     return error_code(errno, posix_category());
 326
 327   while (BytesLeft) {
 328     ssize_t NumRead = ::read(FD, BufPtr, BytesLeft);
 329     if (NumRead == -1) {
 330       if (errno == EINTR)
 331         continue;
 332       // Error while reading.
 333       return error_code(errno, posix_category());
 334     } else if (NumRead == 0) {
 335       // We hit EOF early, truncate and terminate buffer.
 336       Buf->BufferEnd = BufPtr;
 337       *BufPtr = 0;
 338       result.swap(SB);
 339       return success;
 340     }
 341     BytesLeft -= NumRead;
 342     BufPtr += NumRead;
 343   }
 344
 345   result.swap(SB);
 346   return success;
 347 }
 348
 349 //===----------------------------------------------------------------------===//
 350 // MemoryBuffer::getSTDIN implementation.
 351 //===----------------------------------------------------------------------===//
 352
 353 error_code MemoryBuffer::getSTDIN(OwningPtr<MemoryBuffer> &result) {
 354   // Read in all of the data from stdin, we cannot mmap stdin.
 355   //
 356   // FIXME: That isn't necessarily true, we should try to mmap stdin and
 357   // fallback if it fails.
 358   sys::Program::ChangeStdinToBinary();
 359
 360   const ssize_t ChunkSize = 4096*4;
 361   SmallString<ChunkSize> Buffer;
 362   ssize_t ReadBytes;
 363   // Read into Buffer until we hit EOF.
 364   do {
 365     Buffer.reserve(Buffer.size() + ChunkSize);
 366     ReadBytes = read(0, Buffer.end(), ChunkSize);
 367     if (ReadBytes == -1) {
 368       if (errno == EINTR) continue;
 369       return error_code(errno, posix_category());
 370     }
 371     Buffer.set_size(Buffer.size() + ReadBytes);
 372   } while (ReadBytes != 0);
 373
 374   result.reset(getMemBufferCopy(Buffer, "<stdin>"));
 375   return success;
 376 }