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