MemoryBuffer.cpp revision 263508 
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/Config/config.h"
18#include "llvm/Support/Errno.h"
19#include "llvm/Support/FileSystem.h"
20#include "llvm/Support/MathExtras.h"
21#include "llvm/Support/Path.h"
22#include "llvm/Support/Process.h"
23#include "llvm/Support/Program.h"
24#include "llvm/Support/system_error.h"
25#include <cassert>
26#include <cerrno>
27#include <cstdio>
28#include <cstring>
29#include <new>
30#include <sys/stat.h>
31#include <sys/types.h>
32#if !defined(_MSC_VER) && !defined(__MINGW32__)
33#include <unistd.h>
34#else
35#include <io.h>
36// Simplistic definitinos of these macros for use in getOpenFile.
37#ifndef S_ISREG
38#define S_ISREG(x) (1)
39#endif
40#ifndef S_ISBLK
41#define S_ISBLK(x) (0)
42#endif
43#endif
44using namespace llvm;
45
46//===----------------------------------------------------------------------===//
47// MemoryBuffer implementation itself.
48//===----------------------------------------------------------------------===//
49
50MemoryBuffer::~MemoryBuffer() { }
51
52/// init - Initialize this MemoryBuffer as a reference to externally allocated
53/// memory, memory that we know is already null terminated.
54void MemoryBuffer::init(const char *BufStart, const char *BufEnd,
55                        bool RequiresNullTerminator) {
56  assert((!RequiresNullTerminator || BufEnd[0] == 0) &&
57         "Buffer is not null terminated!");
58  BufferStart = BufStart;
59  BufferEnd = BufEnd;
60}
61
62//===----------------------------------------------------------------------===//
63// MemoryBufferMem implementation.
64//===----------------------------------------------------------------------===//
65
66/// CopyStringRef - Copies contents of a StringRef into a block of memory and
67/// null-terminates it.
68static void CopyStringRef(char *Memory, StringRef Data) {
69  memcpy(Memory, Data.data(), Data.size());
70  Memory[Data.size()] = 0; // Null terminate string.
71}
72
73namespace {
74struct NamedBufferAlloc {
75  StringRef Name;
76  NamedBufferAlloc(StringRef Name) : Name(Name) {}
77};
78}
79
80void *operator new(size_t N, const NamedBufferAlloc &Alloc) {
81  char *Mem = static_cast<char *>(operator new(N + Alloc.Name.size() + 1));
82  CopyStringRef(Mem + N, Alloc.Name);
83  return Mem;
84}
85
86namespace {
87/// MemoryBufferMem - Named MemoryBuffer pointing to a block of memory.
88class MemoryBufferMem : public MemoryBuffer {
89public:
90  MemoryBufferMem(StringRef InputData, bool RequiresNullTerminator) {
91    init(InputData.begin(), InputData.end(), RequiresNullTerminator);
92  }
93
94  virtual const char *getBufferIdentifier() const LLVM_OVERRIDE {
95     // The name is stored after the class itself.
96    return reinterpret_cast<const char*>(this + 1);
97  }
98
99  virtual BufferKind getBufferKind() const LLVM_OVERRIDE {
100    return MemoryBuffer_Malloc;
101  }
102};
103}
104
105/// getMemBuffer - Open the specified memory range as a MemoryBuffer.  Note
106/// that InputData must be a null terminated if RequiresNullTerminator is true!
107MemoryBuffer *MemoryBuffer::getMemBuffer(StringRef InputData,
108                                         StringRef BufferName,
109                                         bool RequiresNullTerminator) {
110  return new (NamedBufferAlloc(BufferName))
111      MemoryBufferMem(InputData, RequiresNullTerminator);
112}
113
114/// getMemBufferCopy - Open the specified memory range as a MemoryBuffer,
115/// copying the contents and taking ownership of it.  This has no requirements
116/// on EndPtr[0].
117MemoryBuffer *MemoryBuffer::getMemBufferCopy(StringRef InputData,
118                                             StringRef BufferName) {
119  MemoryBuffer *Buf = getNewUninitMemBuffer(InputData.size(), BufferName);
120  if (!Buf) return 0;
121  memcpy(const_cast<char*>(Buf->getBufferStart()), InputData.data(),
122         InputData.size());
123  return Buf;
124}
125
126/// getNewUninitMemBuffer - Allocate a new MemoryBuffer of the specified size
127/// that is not initialized.  Note that the caller should initialize the
128/// memory allocated by this method.  The memory is owned by the MemoryBuffer
129/// object.
130MemoryBuffer *MemoryBuffer::getNewUninitMemBuffer(size_t Size,
131                                                  StringRef BufferName) {
132  // Allocate space for the MemoryBuffer, the data and the name. It is important
133  // that MemoryBuffer and data are aligned so PointerIntPair works with them.
134  size_t AlignedStringLen =
135    RoundUpToAlignment(sizeof(MemoryBufferMem) + BufferName.size() + 1,
136                       sizeof(void*)); // TODO: Is sizeof(void*) enough?
137  size_t RealLen = AlignedStringLen + Size + 1;
138  char *Mem = static_cast<char*>(operator new(RealLen, std::nothrow));
139  if (!Mem) return 0;
140
141  // The name is stored after the class itself.
142  CopyStringRef(Mem + sizeof(MemoryBufferMem), BufferName);
143
144  // The buffer begins after the name and must be aligned.
145  char *Buf = Mem + AlignedStringLen;
146  Buf[Size] = 0; // Null terminate buffer.
147
148  return new (Mem) MemoryBufferMem(StringRef(Buf, Size), true);
149}
150
151/// getNewMemBuffer - Allocate a new MemoryBuffer of the specified size that
152/// is completely initialized to zeros.  Note that the caller should
153/// initialize the memory allocated by this method.  The memory is owned by
154/// the MemoryBuffer object.
155MemoryBuffer *MemoryBuffer::getNewMemBuffer(size_t Size, StringRef BufferName) {
156  MemoryBuffer *SB = getNewUninitMemBuffer(Size, BufferName);
157  if (!SB) return 0;
158  memset(const_cast<char*>(SB->getBufferStart()), 0, Size);
159  return SB;
160}
161
162
163/// getFileOrSTDIN - Open the specified file as a MemoryBuffer, or open stdin
164/// if the Filename is "-".  If an error occurs, this returns null and fills
165/// in *ErrStr with a reason.  If stdin is empty, this API (unlike getSTDIN)
166/// returns an empty buffer.
167error_code MemoryBuffer::getFileOrSTDIN(StringRef Filename,
168                                        OwningPtr<MemoryBuffer> &result,
169                                        int64_t FileSize) {
170  if (Filename == "-")
171    return getSTDIN(result);
172  return getFile(Filename, result, FileSize);
173}
174
175//===----------------------------------------------------------------------===//
176// MemoryBuffer::getFile implementation.
177//===----------------------------------------------------------------------===//
178
179namespace {
180/// \brief Memory maps a file descriptor using sys::fs::mapped_file_region.
181///
182/// This handles converting the offset into a legal offset on the platform.
183class MemoryBufferMMapFile : public MemoryBuffer {
184  sys::fs::mapped_file_region MFR;
185
186  static uint64_t getLegalMapOffset(uint64_t Offset) {
187    return Offset & ~(sys::fs::mapped_file_region::alignment() - 1);
188  }
189
190  static uint64_t getLegalMapSize(uint64_t Len, uint64_t Offset) {
191    return Len + (Offset - getLegalMapOffset(Offset));
192  }
193
194  const char *getStart(uint64_t Len, uint64_t Offset) {
195    return MFR.const_data() + (Offset - getLegalMapOffset(Offset));
196  }
197
198public:
199  MemoryBufferMMapFile(bool RequiresNullTerminator, int FD, uint64_t Len,
200                       uint64_t Offset, error_code EC)
201      : MFR(FD, false, sys::fs::mapped_file_region::readonly,
202            getLegalMapSize(Len, Offset), getLegalMapOffset(Offset), EC) {
203    if (!EC) {
204      const char *Start = getStart(Len, Offset);
205      init(Start, Start + Len, RequiresNullTerminator);
206    }
207  }
208
209  virtual const char *getBufferIdentifier() const LLVM_OVERRIDE {
210    // The name is stored after the class itself.
211    return reinterpret_cast<const char *>(this + 1);
212  }
213
214  virtual BufferKind getBufferKind() const LLVM_OVERRIDE {
215    return MemoryBuffer_MMap;
216  }
217};
218}
219
220static error_code getMemoryBufferForStream(int FD,
221                                           StringRef BufferName,
222                                           OwningPtr<MemoryBuffer> &result) {
223  const ssize_t ChunkSize = 4096*4;
224  SmallString<ChunkSize> Buffer;
225  ssize_t ReadBytes;
226  // Read into Buffer until we hit EOF.
227  do {
228    Buffer.reserve(Buffer.size() + ChunkSize);
229    ReadBytes = read(FD, Buffer.end(), ChunkSize);
230    if (ReadBytes == -1) {
231      if (errno == EINTR) continue;
232      return error_code(errno, posix_category());
233    }
234    Buffer.set_size(Buffer.size() + ReadBytes);
235  } while (ReadBytes != 0);
236
237  result.reset(MemoryBuffer::getMemBufferCopy(Buffer, BufferName));
238  return error_code::success();
239}
240
241static error_code getFileAux(const char *Filename,
242                             OwningPtr<MemoryBuffer> &result, int64_t FileSize,
243                             bool RequiresNullTerminator);
244
245error_code MemoryBuffer::getFile(Twine Filename,
246                                 OwningPtr<MemoryBuffer> &result,
247                                 int64_t FileSize,
248                                 bool RequiresNullTerminator) {
249  // Ensure the path is null terminated.
250  SmallString<256> PathBuf;
251  StringRef NullTerminatedName = Filename.toNullTerminatedStringRef(PathBuf);
252  return getFileAux(NullTerminatedName.data(), result, FileSize,
253                    RequiresNullTerminator);
254}
255
256static error_code getOpenFileImpl(int FD, const char *Filename,
257                                  OwningPtr<MemoryBuffer> &Result,
258                                  uint64_t FileSize, uint64_t MapSize,
259                                  int64_t Offset, bool RequiresNullTerminator);
260
261static error_code getFileAux(const char *Filename,
262                             OwningPtr<MemoryBuffer> &result, int64_t FileSize,
263                             bool RequiresNullTerminator) {
264  int FD;
265  error_code EC = sys::fs::openFileForRead(Filename, FD);
266  if (EC)
267    return EC;
268
269  error_code ret = getOpenFileImpl(FD, Filename, result, FileSize, FileSize, 0,
270                                   RequiresNullTerminator);
271  close(FD);
272  return ret;
273}
274
275static bool shouldUseMmap(int FD,
276                          size_t FileSize,
277                          size_t MapSize,
278                          off_t Offset,
279                          bool RequiresNullTerminator,
280                          int PageSize) {
281  // We don't use mmap for small files because this can severely fragment our
282  // address space.
283  if (MapSize < 4 * 4096 || MapSize < (unsigned)PageSize)
284    return false;
285
286  if (!RequiresNullTerminator)
287    return true;
288
289
290  // If we don't know the file size, use fstat to find out.  fstat on an open
291  // file descriptor is cheaper than stat on a random path.
292  // FIXME: this chunk of code is duplicated, but it avoids a fstat when
293  // RequiresNullTerminator = false and MapSize != -1.
294  if (FileSize == size_t(-1)) {
295    sys::fs::file_status Status;
296    error_code EC = sys::fs::status(FD, Status);
297    if (EC)
298      return EC;
299    FileSize = Status.getSize();
300  }
301
302  // If we need a null terminator and the end of the map is inside the file,
303  // we cannot use mmap.
304  size_t End = Offset + MapSize;
305  assert(End <= FileSize);
306  if (End != FileSize)
307    return false;
308
309#if defined(_WIN32) || defined(__CYGWIN__)
310  // Don't peek the next page if file is multiple of *physical* pagesize(4k)
311  // but is not multiple of AllocationGranularity(64k),
312  // when a null terminator is required.
313  // FIXME: It's not good to hardcode 4096 here. dwPageSize shows 4096.
314  if ((FileSize & (4096 - 1)) == 0)
315    return false;
316#endif
317
318  // Don't try to map files that are exactly a multiple of the system page size
319  // if we need a null terminator.
320  if ((FileSize & (PageSize -1)) == 0)
321    return false;
322
323  return true;
324}
325
326static error_code getOpenFileImpl(int FD, const char *Filename,
327                                  OwningPtr<MemoryBuffer> &result,
328                                  uint64_t FileSize, uint64_t MapSize,
329                                  int64_t Offset, bool RequiresNullTerminator) {
330  static int PageSize = sys::process::get_self()->page_size();
331
332  // Default is to map the full file.
333  if (MapSize == uint64_t(-1)) {
334    // If we don't know the file size, use fstat to find out.  fstat on an open
335    // file descriptor is cheaper than stat on a random path.
336    if (FileSize == uint64_t(-1)) {
337      sys::fs::file_status Status;
338      error_code EC = sys::fs::status(FD, Status);
339      if (EC)
340        return EC;
341
342      // If this not a file or a block device (e.g. it's a named pipe
343      // or character device), we can't trust the size. Create the memory
344      // buffer by copying off the stream.
345      sys::fs::file_type Type = Status.type();
346      if (Type != sys::fs::file_type::regular_file &&
347          Type != sys::fs::file_type::block_file)
348        return getMemoryBufferForStream(FD, Filename, result);
349
350      FileSize = Status.getSize();
351    }
352    MapSize = FileSize;
353  }
354
355  if (shouldUseMmap(FD, FileSize, MapSize, Offset, RequiresNullTerminator,
356                    PageSize)) {
357    error_code EC;
358    result.reset(new (NamedBufferAlloc(Filename)) MemoryBufferMMapFile(
359        RequiresNullTerminator, FD, MapSize, Offset, EC));
360    if (!EC)
361      return error_code::success();
362  }
363
364  MemoryBuffer *Buf = MemoryBuffer::getNewUninitMemBuffer(MapSize, Filename);
365  if (!Buf) {
366    // Failed to create a buffer. The only way it can fail is if
367    // new(std::nothrow) returns 0.
368    return make_error_code(errc::not_enough_memory);
369  }
370
371  OwningPtr<MemoryBuffer> SB(Buf);
372  char *BufPtr = const_cast<char*>(SB->getBufferStart());
373
374  size_t BytesLeft = MapSize;
375#ifndef HAVE_PREAD
376  if (lseek(FD, Offset, SEEK_SET) == -1)
377    return error_code(errno, posix_category());
378#endif
379
380  while (BytesLeft) {
381#ifdef HAVE_PREAD
382    ssize_t NumRead = ::pread(FD, BufPtr, BytesLeft, MapSize-BytesLeft+Offset);
383#else
384    ssize_t NumRead = ::read(FD, BufPtr, BytesLeft);
385#endif
386    if (NumRead == -1) {
387      if (errno == EINTR)
388        continue;
389      // Error while reading.
390      return error_code(errno, posix_category());
391    }
392    if (NumRead == 0) {
393      assert(0 && "We got inaccurate FileSize value or fstat reported an "
394                   "invalid file size.");
395      *BufPtr = '\0'; // null-terminate at the actual size.
396      break;
397    }
398    BytesLeft -= NumRead;
399    BufPtr += NumRead;
400  }
401
402  result.swap(SB);
403  return error_code::success();
404}
405
406error_code MemoryBuffer::getOpenFile(int FD, const char *Filename,
407                                     OwningPtr<MemoryBuffer> &Result,
408                                     uint64_t FileSize,
409                                     bool RequiresNullTerminator) {
410  return getOpenFileImpl(FD, Filename, Result, FileSize, FileSize, 0,
411                         RequiresNullTerminator);
412}
413
414error_code MemoryBuffer::getOpenFileSlice(int FD, const char *Filename,
415                                          OwningPtr<MemoryBuffer> &Result,
416                                          uint64_t MapSize, int64_t Offset) {
417  return getOpenFileImpl(FD, Filename, Result, -1, MapSize, Offset, false);
418}
419
420//===----------------------------------------------------------------------===//
421// MemoryBuffer::getSTDIN implementation.
422//===----------------------------------------------------------------------===//
423
424error_code MemoryBuffer::getSTDIN(OwningPtr<MemoryBuffer> &result) {
425  // Read in all of the data from stdin, we cannot mmap stdin.
426  //
427  // FIXME: That isn't necessarily true, we should try to mmap stdin and
428  // fallback if it fails.
429  sys::ChangeStdinToBinary();
430
431  return getMemoryBufferForStream(0, "<stdin>", result);
432}
433