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     delete [] 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 = new char[Size+1];
  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 = new char[Size+1];
 112   Buf[Size] = 0;
 113   MemoryBufferMem *SB = new MemoryBufferMem(Buf, Buf+Size, BufferName);
 114   // The memory for this buffer is owned by the MemoryBuffer.
 115   SB->MustDeleteBuffer = true;
 116   return SB;
 117 }
 118
 119 /// getNewMemBuffer - Allocate a new MemoryBuffer of the specified size that
 120 /// is completely initialized to zeros.  Note that the caller should
 121 /// initialize the memory allocated by this method.  The memory is owned by
 122 /// the MemoryBuffer object.
 123 MemoryBuffer *MemoryBuffer::getNewMemBuffer(size_t Size,
 124                                             const char *BufferName) {
 125   MemoryBuffer *SB = getNewUninitMemBuffer(Size, BufferName);
 126   memset(const_cast<char*>(SB->getBufferStart()), 0, Size+1);
 127   return SB;
 128 }
 129
 130
 131 /// getFileOrSTDIN - Open the specified file as a MemoryBuffer, or open stdin
 132 /// if the Filename is "-".  If an error occurs, this returns null and fills
 133 /// in *ErrStr with a reason.  If stdin is empty, this API (unlike getSTDIN)
 134 /// returns an empty buffer.
 135 MemoryBuffer *MemoryBuffer::getFileOrSTDIN(const char *Filename,
 136                                            std::string *ErrStr,
 137                                            int64_t FileSize) {
 138   if (Filename[0] != '-' || Filename[1] != 0)
 139     return getFile(Filename, ErrStr, FileSize);
 140   MemoryBuffer *M = getSTDIN();
 141   if (M) return M;
 142
 143   // If stdin was empty, M is null.  Cons up an empty memory buffer now.
 144   const char *EmptyStr = "";
 145   return MemoryBuffer::getMemBuffer(EmptyStr, EmptyStr, "<stdin>");
 146 }
 147
 148 //===----------------------------------------------------------------------===//
 149 // MemoryBuffer::getFile implementation.
 150 //===----------------------------------------------------------------------===//
 151
 152 namespace {
 153 /// MemoryBufferMMapFile - This represents a file that was mapped in with the
 154 /// sys::Path::MapInFilePages method.  When destroyed, it calls the
 155 /// sys::Path::UnMapFilePages method.
 156 class MemoryBufferMMapFile : public MemoryBuffer {
 157   std::string Filename;
 158 public:
 159   MemoryBufferMMapFile(const char *filename, const char *Pages, uint64_t Size)
 160     : Filename(filename) {
 161     init(Pages, Pages+Size);
 162   }
 163
 164   virtual const char *getBufferIdentifier() const {
 165     return Filename.c_str();
 166   }
 167
 168   ~MemoryBufferMMapFile() {
 169     sys::Path::UnMapFilePages(getBufferStart(), getBufferSize());
 170   }
 171 };
 172 }
 173
 174 MemoryBuffer *MemoryBuffer::getFile(const char *Filename, std::string *ErrStr,
 175                                     int64_t FileSize) {
 176   int OpenFlags = 0;
 177 #ifdef O_BINARY
 178   OpenFlags |= O_BINARY;  // Open input file in binary mode on win32.
 179 #endif
 180   int FD = ::open(Filename, O_RDONLY|OpenFlags);
 181   if (FD == -1) {
 182     if (ErrStr) *ErrStr = "could not open file";
 183     return 0;
 184   }
 185
 186   // If we don't know the file size, use fstat to find out.  fstat on an open
 187   // file descriptor is cheaper than stat on a random path.
 188   if (FileSize == -1) {
 189     struct stat FileInfo;
 190     // TODO: This should use fstat64 when available.
 191     if (fstat(FD, &FileInfo) == -1) {
 192       if (ErrStr) *ErrStr = "could not get file length";
 193       ::close(FD);
 194       return 0;
 195     }
 196     FileSize = FileInfo.st_size;
 197   }
 198
 199
 200   // If the file is large, try to use mmap to read it in.  We don't use mmap
 201   // for small files, because this can severely fragment our address space. Also
 202   // don't try to map files that are exactly a multiple of the system page size,
 203   // as the file would not have the required null terminator.
 204   if (FileSize >= 4096*4 &&
 205       (FileSize & (sys::Process::GetPageSize()-1)) != 0) {
 206     if (const char *Pages = sys::Path::MapInFilePages(FD, FileSize)) {
 207       // Close the file descriptor, now that the whole file is in memory.
 208       ::close(FD);
 209       return new MemoryBufferMMapFile(Filename, Pages, FileSize);
 210     }
 211   }
 212
 213   OwningPtr<MemoryBuffer> SB;
 214   SB.reset(MemoryBuffer::getNewUninitMemBuffer(FileSize, Filename));
 215   char *BufPtr = const_cast<char*>(SB->getBufferStart());
 216
 217   size_t BytesLeft = FileSize;
 218   while (BytesLeft) {
 219     ssize_t NumRead = ::read(FD, BufPtr, BytesLeft);
 220     if (NumRead != -1) {
 221       BytesLeft -= NumRead;
 222       BufPtr += NumRead;
 223     } else if (errno == EINTR) {
 224       // try again
 225     } else {
 226       // error reading.
 227       close(FD);
 228       if (ErrStr) *ErrStr = "error reading file data";
 229       return 0;
 230     }
 231   }
 232   close(FD);
 233
 234   return SB.take();
 235 }
 236
 237 //===----------------------------------------------------------------------===//
 238 // MemoryBuffer::getSTDIN implementation.
 239 //===----------------------------------------------------------------------===//
 240
 241 namespace {
 242 class STDINBufferFile : public MemoryBuffer {
 243 public:
 244   virtual const char *getBufferIdentifier() const {
 245     return "<stdin>";
 246   }
 247 };
 248 }
 249
 250 MemoryBuffer *MemoryBuffer::getSTDIN() {
 251   char Buffer[4096*4];
 252
 253   std::vector<char> FileData;
 254
 255   // Read in all of the data from stdin, we cannot mmap stdin.
 256   sys::Program::ChangeStdinToBinary();
 257   while (size_t ReadBytes = fread(Buffer, sizeof(char), 4096*4, stdin))
 258     FileData.insert(FileData.end(), Buffer, Buffer+ReadBytes);
 259
 260   FileData.push_back(0); // &FileData[Size] is invalid. So is &*FileData.end().
 261   size_t Size = FileData.size();
 262   if (Size <= 1)
 263     return 0;
 264   MemoryBuffer *B = new STDINBufferFile();
 265   B->initCopyOf(&FileData[0], &FileData[Size-1]);
 266   return B;
 267 }