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/System/Path.h"
  18 #include "llvm/System/Process.h"
  19 #include "llvm/System/Program.h"
  20 #include <cassert>
  21 #include <cstdio>
  22 #include <cstring>
  23 #include <cerrno>
  24 #include <sys/types.h>
  25 #include <sys/stat.h>
  26 #if !defined(_MSC_VER) && !defined(__MINGW32__)
  27 #include <unistd.h>
  28 #include <sys/uio.h>
  29 #else
  30 #include <io.h>
  31 #endif
  32 #include <fcntl.h>
  33 using namespace llvm;
  34
  35 //===----------------------------------------------------------------------===//
  36 // MemoryBuffer implementation itself.
  37 //===----------------------------------------------------------------------===//
  38
  39 MemoryBuffer::~MemoryBuffer() {
  40   if (MustDeleteBuffer)
  41     free((void*)BufferStart);
  42 }
  43
  44 /// initCopyOf - Initialize this source buffer with a copy of the specified
  45 /// memory range.  We make the copy so that we can null terminate it
  46 /// successfully.
  47 void MemoryBuffer::initCopyOf(const char *BufStart, const char *BufEnd) {
  48   size_t Size = BufEnd-BufStart;
  49   BufferStart = (char *)malloc((Size+1) * sizeof(char));
  50   BufferEnd = BufferStart+Size;
  51   memcpy(const_cast<char*>(BufferStart), BufStart, Size);
  52   *const_cast<char*>(BufferEnd) = 0;   // Null terminate buffer.
  53   MustDeleteBuffer = true;
  54 }
  55
  56 /// init - Initialize this MemoryBuffer as a reference to externally allocated
  57 /// memory, memory that we know is already null terminated.
  58 void MemoryBuffer::init(const char *BufStart, const char *BufEnd) {
  59   assert(BufEnd[0] == 0 && "Buffer is not null terminated!");
  60   BufferStart = BufStart;
  61   BufferEnd = BufEnd;
  62   MustDeleteBuffer = false;
  63 }
  64
  65 //===----------------------------------------------------------------------===//
  66 // MemoryBufferMem implementation.
  67 //===----------------------------------------------------------------------===//
  68
  69 namespace {
  70 class MemoryBufferMem : public MemoryBuffer {
  71   std::string FileID;
  72 public:
  73   MemoryBufferMem(const char *Start, const char *End, const char *FID,
  74                   bool Copy = false)
  75   : FileID(FID) {
  76     if (!Copy)
  77       init(Start, End);
  78     else
  79       initCopyOf(Start, End);
  80   }
  81
  82   virtual const char *getBufferIdentifier() const {
  83     return FileID.c_str();
  84   }
  85 };
  86 }
  87
  88 /// getMemBuffer - Open the specified memory range as a MemoryBuffer.  Note
  89 /// that EndPtr[0] must be a null byte and be accessible!
  90 MemoryBuffer *MemoryBuffer::getMemBuffer(const char *StartPtr,
  91                                          const char *EndPtr,
  92                                          const char *BufferName) {
  93   return new MemoryBufferMem(StartPtr, EndPtr, BufferName);
  94 }
  95
  96 /// getMemBufferCopy - Open the specified memory range as a MemoryBuffer,
  97 /// copying the contents and taking ownership of it.  This has no requirements
  98 /// on EndPtr[0].
  99 MemoryBuffer *MemoryBuffer::getMemBufferCopy(const char *StartPtr,
 100                                              const char *EndPtr,
 101                                              const char *BufferName) {
 102   return new MemoryBufferMem(StartPtr, EndPtr, BufferName, true);
 103 }
 104
 105 /// getNewUninitMemBuffer - Allocate a new MemoryBuffer of the specified size
 106 /// that is completely initialized to zeros.  Note that the caller should
 107 /// initialize the memory allocated by this method.  The memory is owned by
 108 /// the MemoryBuffer object.
 109 MemoryBuffer *MemoryBuffer::getNewUninitMemBuffer(size_t Size,
 110                                                   const char *BufferName) {
 111   char *Buf = (char *)malloc((Size+1) * sizeof(char));
 112   if (!Buf) return 0;
 113   Buf[Size] = 0;
 114   MemoryBufferMem *SB = new MemoryBufferMem(Buf, Buf+Size, BufferName);
 115   // The memory for this buffer is owned by the MemoryBuffer.
 116   SB->MustDeleteBuffer = true;
 117   return SB;
 118 }
 119
 120 /// getNewMemBuffer - Allocate a new MemoryBuffer of the specified size that
 121 /// is completely initialized to zeros.  Note that the caller should
 122 /// initialize the memory allocated by this method.  The memory is owned by
 123 /// the MemoryBuffer object.
 124 MemoryBuffer *MemoryBuffer::getNewMemBuffer(size_t Size,
 125                                             const char *BufferName) {
 126   MemoryBuffer *SB = getNewUninitMemBuffer(Size, BufferName);
 127   if (!SB) return 0;
 128   memset(const_cast<char*>(SB->getBufferStart()), 0, Size+1);
 129   return SB;
 130 }
 131
 132
 133 /// getFileOrSTDIN - Open the specified file as a MemoryBuffer, or open stdin
 134 /// if the Filename is "-".  If an error occurs, this returns null and fills
 135 /// in *ErrStr with a reason.  If stdin is empty, this API (unlike getSTDIN)
 136 /// returns an empty buffer.
 137 MemoryBuffer *MemoryBuffer::getFileOrSTDIN(const char *Filename,
 138                                            std::string *ErrStr,
 139                                            int64_t FileSize) {
 140   if (Filename[0] != '-' || Filename[1] != 0)
 141     return getFile(Filename, ErrStr, FileSize);
 142   MemoryBuffer *M = getSTDIN();
 143   if (M) return M;
 144
 145   // If stdin was empty, M is null.  Cons up an empty memory buffer now.
 146   const char *EmptyStr = "";
 147   return MemoryBuffer::getMemBuffer(EmptyStr, EmptyStr, "<stdin>");
 148 }
 149
 150 //===----------------------------------------------------------------------===//
 151 // MemoryBuffer::getFile implementation.
 152 //===----------------------------------------------------------------------===//
 153
 154 namespace {
 155 /// MemoryBufferMMapFile - This represents a file that was mapped in with the
 156 /// sys::Path::MapInFilePages method.  When destroyed, it calls the
 157 /// sys::Path::UnMapFilePages method.
 158 class MemoryBufferMMapFile : public MemoryBuffer {
 159   std::string Filename;
 160 public:
 161   MemoryBufferMMapFile(const char *filename, const char *Pages, uint64_t Size)
 162     : Filename(filename) {
 163     init(Pages, Pages+Size);
 164   }
 165
 166   virtual const char *getBufferIdentifier() const {
 167     return Filename.c_str();
 168   }
 169
 170   ~MemoryBufferMMapFile() {
 171     sys::Path::UnMapFilePages(getBufferStart(), getBufferSize());
 172   }
 173 };
 174 }
 175
 176 MemoryBuffer *MemoryBuffer::getFile(const char *Filename, std::string *ErrStr,
 177                                     int64_t FileSize) {
 178   int OpenFlags = 0;
 179 #ifdef O_BINARY
 180   OpenFlags |= O_BINARY;  // Open input file in binary mode on win32.
 181 #endif
 182   int FD = ::open(Filename, O_RDONLY|OpenFlags);
 183   if (FD == -1) {
 184     if (ErrStr) *ErrStr = "could not open file";
 185     return 0;
 186   }
 187
 188   // If we don't know the file size, use fstat to find out.  fstat on an open
 189   // file descriptor is cheaper than stat on a random path.
 190   if (FileSize == -1) {
 191     struct stat FileInfo;
 192     // TODO: This should use fstat64 when available.
 193     if (fstat(FD, &FileInfo) == -1) {
 194       if (ErrStr) *ErrStr = "could not get file length";
 195       ::close(FD);
 196       return 0;
 197     }
 198     FileSize = FileInfo.st_size;
 199   }
 200
 201
 202   // If the file is large, try to use mmap to read it in.  We don't use mmap
 203   // for small files, because this can severely fragment our address space. Also
 204   // don't try to map files that are exactly a multiple of the system page size,
 205   // as the file would not have the required null terminator.
 206   if (FileSize >= 4096*4 &&
 207       (FileSize & (sys::Process::GetPageSize()-1)) != 0) {
 208     if (const char *Pages = sys::Path::MapInFilePages(FD, FileSize)) {
 209       // Close the file descriptor, now that the whole file is in memory.
 210       ::close(FD);
 211       return new MemoryBufferMMapFile(Filename, Pages, FileSize);
 212     }
 213   }
 214
 215   MemoryBuffer *Buf = MemoryBuffer::getNewUninitMemBuffer(FileSize, Filename);
 216   if (!Buf) {
 217     // Failed to create a buffer.
 218     if (ErrStr) *ErrStr = "could not allocate buffer";
 219     ::close(FD);
 220     return 0;
 221   }
 222
 223   OwningPtr<MemoryBuffer> SB(Buf);
 224   char *BufPtr = const_cast<char*>(SB->getBufferStart());
 225
 226   size_t BytesLeft = FileSize;
 227   while (BytesLeft) {
 228     ssize_t NumRead = ::read(FD, BufPtr, BytesLeft);
 229     if (NumRead != -1) {
 230       BytesLeft -= NumRead;
 231       BufPtr += NumRead;
 232     } else if (errno == EINTR) {
 233       // try again
 234     } else {
 235       // error reading.
 236       close(FD);
 237       if (ErrStr) *ErrStr = "error reading file data";
 238       return 0;
 239     }
 240   }
 241   close(FD);
 242
 243   return SB.take();
 244 }
 245
 246 //===----------------------------------------------------------------------===//
 247 // MemoryBuffer::getSTDIN implementation.
 248 //===----------------------------------------------------------------------===//
 249
 250 namespace {
 251 class STDINBufferFile : public MemoryBuffer {
 252 public:
 253   virtual const char *getBufferIdentifier() const {
 254     return "<stdin>";
 255   }
 256 };
 257 }
 258
 259 MemoryBuffer *MemoryBuffer::getSTDIN() {
 260   char Buffer[4096*4];
 261
 262   std::vector<char> FileData;
 263
 264   // Read in all of the data from stdin, we cannot mmap stdin.
 265   sys::Program::ChangeStdinToBinary();
 266   size_t ReadBytes;
 267   do {
 268     ReadBytes = fread(Buffer, sizeof(char), sizeof(Buffer), stdin);
 269     FileData.insert(FileData.end(), Buffer, Buffer+ReadBytes);
 270   } while (ReadBytes == sizeof(Buffer));
 271
 272   FileData.push_back(0); // &FileData[Size] is invalid. So is &*FileData.end().
 273   size_t Size = FileData.size();
 274   if (Size <= 1)
 275     return 0;
 276   MemoryBuffer *B = new STDINBufferFile();
 277   B->initCopyOf(&FileData[0], &FileData[Size-1]);
 278   return B;
 279 }