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/SmallString.h"
16 #include "llvm/Config/config.h"
17 #include "llvm/Support/Errc.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 <cassert>
25 #include <cerrno>
26 #include <cstring>
27 #include <new>
28 #include <sys/types.h>
29 #include <system_error>
30 #if !defined(_MSC_VER) && !defined(__MINGW32__)
31 #include <unistd.h>
32 #else
33 #include <io.h>
34 #endif
35 using namespace llvm;
36 
37 //===----------------------------------------------------------------------===//
38 // MemoryBuffer implementation itself.
39 //===----------------------------------------------------------------------===//
40 
41 MemoryBuffer::~MemoryBuffer() { }
42 
43 /// init - Initialize this MemoryBuffer as a reference to externally allocated
44 /// memory, memory that we know is already null terminated.
45 void MemoryBuffer::init(const char *BufStart, const char *BufEnd,
46                         bool RequiresNullTerminator) {
47   assert((!RequiresNullTerminator || BufEnd[0] == 0) &&
48          "Buffer is not null terminated!");
49   BufferStart = BufStart;
50   BufferEnd = BufEnd;
51 }
52 
53 //===----------------------------------------------------------------------===//
54 // MemoryBufferMem implementation.
55 //===----------------------------------------------------------------------===//
56 
57 /// CopyStringRef - Copies contents of a StringRef into a block of memory and
58 /// null-terminates it.
59 static void CopyStringRef(char *Memory, StringRef Data) {
60   if (!Data.empty())
61     memcpy(Memory, Data.data(), Data.size());
62   Memory[Data.size()] = 0; // Null terminate string.
63 }
64 
65 namespace {
66 struct NamedBufferAlloc {
67   const Twine &Name;
68   NamedBufferAlloc(const Twine &Name) : Name(Name) {}
69 };
70 }
71 
72 void *operator new(size_t N, const NamedBufferAlloc &Alloc) {
73   SmallString<256> NameBuf;
74   StringRef NameRef = Alloc.Name.toStringRef(NameBuf);
75 
76   char *Mem = static_cast<char *>(operator new(N + NameRef.size() + 1));
77   CopyStringRef(Mem + N, NameRef);
78   return Mem;
79 }
80 
81 namespace {
82 /// MemoryBufferMem - Named MemoryBuffer pointing to a block of memory.
83 class MemoryBufferMem : public MemoryBuffer {
84 public:
85   MemoryBufferMem(StringRef InputData, bool RequiresNullTerminator) {
86     init(InputData.begin(), InputData.end(), RequiresNullTerminator);
87   }
88 
89   /// Disable sized deallocation for MemoryBufferMem, because it has
90   /// tail-allocated data.
91   void operator delete(void *p) { ::operator delete(p); }
92 
93   StringRef getBufferIdentifier() const override {
94     // The name is stored after the class itself.
95     return StringRef(reinterpret_cast<const char *>(this + 1));
96   }
97 
98   BufferKind getBufferKind() const override {
99     return MemoryBuffer_Malloc;
100   }
101 };
102 }
103 
104 static ErrorOr<std::unique_ptr<MemoryBuffer>>
105 getFileAux(const Twine &Filename, int64_t FileSize, uint64_t MapSize,
106            uint64_t Offset, bool RequiresNullTerminator, bool IsVolatile);
107 
108 std::unique_ptr<MemoryBuffer>
109 MemoryBuffer::getMemBuffer(StringRef InputData, StringRef BufferName,
110                            bool RequiresNullTerminator) {
111   auto *Ret = new (NamedBufferAlloc(BufferName))
112       MemoryBufferMem(InputData, RequiresNullTerminator);
113   return std::unique_ptr<MemoryBuffer>(Ret);
114 }
115 
116 std::unique_ptr<MemoryBuffer>
117 MemoryBuffer::getMemBuffer(MemoryBufferRef Ref, bool RequiresNullTerminator) {
118   return std::unique_ptr<MemoryBuffer>(getMemBuffer(
119       Ref.getBuffer(), Ref.getBufferIdentifier(), RequiresNullTerminator));
120 }
121 
122 std::unique_ptr<MemoryBuffer>
123 MemoryBuffer::getMemBufferCopy(StringRef InputData, const Twine &BufferName) {
124   std::unique_ptr<MemoryBuffer> Buf =
125       getNewUninitMemBuffer(InputData.size(), BufferName);
126   if (!Buf)
127     return nullptr;
128   memcpy(const_cast<char*>(Buf->getBufferStart()), InputData.data(),
129          InputData.size());
130   return Buf;
131 }
132 
133 std::unique_ptr<MemoryBuffer>
134 MemoryBuffer::getNewUninitMemBuffer(size_t Size, const Twine &BufferName) {
135   // Allocate space for the MemoryBuffer, the data and the name. It is important
136   // that MemoryBuffer and data are aligned so PointerIntPair works with them.
137   // TODO: Is 16-byte alignment enough?  We copy small object files with large
138   // alignment expectations into this buffer.
139   SmallString<256> NameBuf;
140   StringRef NameRef = BufferName.toStringRef(NameBuf);
141   size_t AlignedStringLen =
142       alignTo(sizeof(MemoryBufferMem) + NameRef.size() + 1, 16);
143   size_t RealLen = AlignedStringLen + Size + 1;
144   char *Mem = static_cast<char*>(operator new(RealLen, std::nothrow));
145   if (!Mem)
146     return nullptr;
147 
148   // The name is stored after the class itself.
149   CopyStringRef(Mem + sizeof(MemoryBufferMem), NameRef);
150 
151   // The buffer begins after the name and must be aligned.
152   char *Buf = Mem + AlignedStringLen;
153   Buf[Size] = 0; // Null terminate buffer.
154 
155   auto *Ret = new (Mem) MemoryBufferMem(StringRef(Buf, Size), true);
156   return std::unique_ptr<MemoryBuffer>(Ret);
157 }
158 
159 std::unique_ptr<MemoryBuffer>
160 MemoryBuffer::getNewMemBuffer(size_t Size, StringRef BufferName) {
161   std::unique_ptr<MemoryBuffer> SB = getNewUninitMemBuffer(Size, BufferName);
162   if (!SB)
163     return nullptr;
164   memset(const_cast<char*>(SB->getBufferStart()), 0, Size);
165   return SB;
166 }
167 
168 ErrorOr<std::unique_ptr<MemoryBuffer>>
169 MemoryBuffer::getFileOrSTDIN(const Twine &Filename, int64_t FileSize,
170                              bool RequiresNullTerminator) {
171   SmallString<256> NameBuf;
172   StringRef NameRef = Filename.toStringRef(NameBuf);
173 
174   if (NameRef == "-")
175     return getSTDIN();
176   return getFile(Filename, FileSize, RequiresNullTerminator);
177 }
178 
179 ErrorOr<std::unique_ptr<MemoryBuffer>>
180 MemoryBuffer::getFileSlice(const Twine &FilePath, uint64_t MapSize,
181                            uint64_t Offset, bool IsVolatile) {
182   return getFileAux(FilePath, -1, MapSize, Offset, false, IsVolatile);
183 }
184 
185 
186 //===----------------------------------------------------------------------===//
187 // MemoryBuffer::getFile implementation.
188 //===----------------------------------------------------------------------===//
189 
190 namespace {
191 /// \brief Memory maps a file descriptor using sys::fs::mapped_file_region.
192 ///
193 /// This handles converting the offset into a legal offset on the platform.
194 class MemoryBufferMMapFile : public MemoryBuffer {
195   sys::fs::mapped_file_region MFR;
196 
197   static uint64_t getLegalMapOffset(uint64_t Offset) {
198     return Offset & ~(sys::fs::mapped_file_region::alignment() - 1);
199   }
200 
201   static uint64_t getLegalMapSize(uint64_t Len, uint64_t Offset) {
202     return Len + (Offset - getLegalMapOffset(Offset));
203   }
204 
205   const char *getStart(uint64_t Len, uint64_t Offset) {
206     return MFR.const_data() + (Offset - getLegalMapOffset(Offset));
207   }
208 
209 public:
210   MemoryBufferMMapFile(bool RequiresNullTerminator, int FD, uint64_t Len,
211                        uint64_t Offset, std::error_code &EC)
212       : MFR(FD, sys::fs::mapped_file_region::readonly,
213             getLegalMapSize(Len, Offset), getLegalMapOffset(Offset), EC) {
214     if (!EC) {
215       const char *Start = getStart(Len, Offset);
216       init(Start, Start + Len, RequiresNullTerminator);
217     }
218   }
219 
220   /// Disable sized deallocation for MemoryBufferMMapFile, because it has
221   /// tail-allocated data.
222   void operator delete(void *p) { ::operator delete(p); }
223 
224   StringRef getBufferIdentifier() const override {
225     // The name is stored after the class itself.
226     return StringRef(reinterpret_cast<const char *>(this + 1));
227   }
228 
229   BufferKind getBufferKind() const override {
230     return MemoryBuffer_MMap;
231   }
232 };
233 }
234 
235 static ErrorOr<std::unique_ptr<MemoryBuffer>>
236 getMemoryBufferForStream(int FD, const Twine &BufferName) {
237   const ssize_t ChunkSize = 4096*4;
238   SmallString<ChunkSize> Buffer;
239   ssize_t ReadBytes;
240   // Read into Buffer until we hit EOF.
241   do {
242     Buffer.reserve(Buffer.size() + ChunkSize);
243     ReadBytes = sys::RetryAfterSignal(-1, read, FD, Buffer.end(), ChunkSize);
244     if (ReadBytes == -1)
245       return std::error_code(errno, std::generic_category());
246     Buffer.set_size(Buffer.size() + ReadBytes);
247   } while (ReadBytes != 0);
248 
249   return MemoryBuffer::getMemBufferCopy(Buffer, BufferName);
250 }
251 
252 
253 ErrorOr<std::unique_ptr<MemoryBuffer>>
254 MemoryBuffer::getFile(const Twine &Filename, int64_t FileSize,
255                       bool RequiresNullTerminator, bool IsVolatile) {
256   return getFileAux(Filename, FileSize, FileSize, 0,
257                     RequiresNullTerminator, IsVolatile);
258 }
259 
260 static ErrorOr<std::unique_ptr<MemoryBuffer>>
261 getOpenFileImpl(int FD, const Twine &Filename, uint64_t FileSize,
262                 uint64_t MapSize, int64_t Offset, bool RequiresNullTerminator,
263                 bool IsVolatile);
264 
265 static ErrorOr<std::unique_ptr<MemoryBuffer>>
266 getFileAux(const Twine &Filename, int64_t FileSize, uint64_t MapSize,
267            uint64_t Offset, bool RequiresNullTerminator, bool IsVolatile) {
268   int FD;
269   std::error_code EC = sys::fs::openFileForRead(Filename, FD);
270   if (EC)
271     return EC;
272 
273   ErrorOr<std::unique_ptr<MemoryBuffer>> Ret =
274       getOpenFileImpl(FD, Filename, FileSize, MapSize, Offset,
275                       RequiresNullTerminator, IsVolatile);
276   close(FD);
277   return Ret;
278 }
279 
280 static bool shouldUseMmap(int FD,
281                           size_t FileSize,
282                           size_t MapSize,
283                           off_t Offset,
284                           bool RequiresNullTerminator,
285                           int PageSize,
286                           bool IsVolatile) {
287   // mmap may leave the buffer without null terminator if the file size changed
288   // by the time the last page is mapped in, so avoid it if the file size is
289   // likely to change.
290   if (IsVolatile)
291     return false;
292 
293   // We don't use mmap for small files because this can severely fragment our
294   // address space.
295   if (MapSize < 4 * 4096 || MapSize < (unsigned)PageSize)
296     return false;
297 
298   if (!RequiresNullTerminator)
299     return true;
300 
301   // If we don't know the file size, use fstat to find out.  fstat on an open
302   // file descriptor is cheaper than stat on a random path.
303   // FIXME: this chunk of code is duplicated, but it avoids a fstat when
304   // RequiresNullTerminator = false and MapSize != -1.
305   if (FileSize == size_t(-1)) {
306     sys::fs::file_status Status;
307     if (sys::fs::status(FD, Status))
308       return false;
309     FileSize = Status.getSize();
310   }
311 
312   // If we need a null terminator and the end of the map is inside the file,
313   // we cannot use mmap.
314   size_t End = Offset + MapSize;
315   assert(End <= FileSize);
316   if (End != FileSize)
317     return false;
318 
319   // Don't try to map files that are exactly a multiple of the system page size
320   // if we need a null terminator.
321   if ((FileSize & (PageSize -1)) == 0)
322     return false;
323 
324 #if defined(__CYGWIN__)
325   // Don't try to map files that are exactly a multiple of the physical page size
326   // if we need a null terminator.
327   // FIXME: We should reorganize again getPageSize() on Win32.
328   if ((FileSize & (4096 - 1)) == 0)
329     return false;
330 #endif
331 
332   return true;
333 }
334 
335 static ErrorOr<std::unique_ptr<MemoryBuffer>>
336 getOpenFileImpl(int FD, const Twine &Filename, uint64_t FileSize,
337                 uint64_t MapSize, int64_t Offset, bool RequiresNullTerminator,
338                 bool IsVolatile) {
339   static int PageSize = sys::Process::getPageSize();
340 
341   // Default is to map the full file.
342   if (MapSize == uint64_t(-1)) {
343     // If we don't know the file size, use fstat to find out.  fstat on an open
344     // file descriptor is cheaper than stat on a random path.
345     if (FileSize == uint64_t(-1)) {
346       sys::fs::file_status Status;
347       std::error_code EC = sys::fs::status(FD, Status);
348       if (EC)
349         return EC;
350 
351       // If this not a file or a block device (e.g. it's a named pipe
352       // or character device), we can't trust the size. Create the memory
353       // buffer by copying off the stream.
354       sys::fs::file_type Type = Status.type();
355       if (Type != sys::fs::file_type::regular_file &&
356           Type != sys::fs::file_type::block_file)
357         return getMemoryBufferForStream(FD, Filename);
358 
359       FileSize = Status.getSize();
360     }
361     MapSize = FileSize;
362   }
363 
364   if (shouldUseMmap(FD, FileSize, MapSize, Offset, RequiresNullTerminator,
365                     PageSize, IsVolatile)) {
366     std::error_code EC;
367     std::unique_ptr<MemoryBuffer> Result(
368         new (NamedBufferAlloc(Filename))
369         MemoryBufferMMapFile(RequiresNullTerminator, FD, MapSize, Offset, EC));
370     if (!EC)
371       return std::move(Result);
372   }
373 
374   std::unique_ptr<MemoryBuffer> Buf =
375       MemoryBuffer::getNewUninitMemBuffer(MapSize, Filename);
376   if (!Buf) {
377     // Failed to create a buffer. The only way it can fail is if
378     // new(std::nothrow) returns 0.
379     return make_error_code(errc::not_enough_memory);
380   }
381 
382   char *BufPtr = const_cast<char *>(Buf->getBufferStart());
383 
384   size_t BytesLeft = MapSize;
385 #ifndef HAVE_PREAD
386   if (lseek(FD, Offset, SEEK_SET) == -1)
387     return std::error_code(errno, std::generic_category());
388 #endif
389 
390   while (BytesLeft) {
391 #ifdef HAVE_PREAD
392     ssize_t NumRead = sys::RetryAfterSignal(-1, ::pread, FD, BufPtr, BytesLeft,
393                                             MapSize - BytesLeft + Offset);
394 #else
395     ssize_t NumRead = sys::RetryAfterSignal(-1, ::read, FD, BufPtr, BytesLeft);
396 #endif
397     if (NumRead == -1) {
398       // Error while reading.
399       return std::error_code(errno, std::generic_category());
400     }
401     if (NumRead == 0) {
402       memset(BufPtr, 0, BytesLeft); // zero-initialize rest of the buffer.
403       break;
404     }
405     BytesLeft -= NumRead;
406     BufPtr += NumRead;
407   }
408 
409   return std::move(Buf);
410 }
411 
412 ErrorOr<std::unique_ptr<MemoryBuffer>>
413 MemoryBuffer::getOpenFile(int FD, const Twine &Filename, uint64_t FileSize,
414                           bool RequiresNullTerminator, bool IsVolatile) {
415   return getOpenFileImpl(FD, Filename, FileSize, FileSize, 0,
416                          RequiresNullTerminator, IsVolatile);
417 }
418 
419 ErrorOr<std::unique_ptr<MemoryBuffer>>
420 MemoryBuffer::getOpenFileSlice(int FD, const Twine &Filename, uint64_t MapSize,
421                                int64_t Offset, bool IsVolatile) {
422   assert(MapSize != uint64_t(-1));
423   return getOpenFileImpl(FD, Filename, -1, MapSize, Offset, false, IsVolatile);
424 }
425 
426 ErrorOr<std::unique_ptr<MemoryBuffer>> MemoryBuffer::getSTDIN() {
427   // Read in all of the data from stdin, we cannot mmap stdin.
428   //
429   // FIXME: That isn't necessarily true, we should try to mmap stdin and
430   // fallback if it fails.
431   sys::ChangeStdinToBinary();
432 
433   return getMemoryBufferForStream(0, "<stdin>");
434 }
435 
436 ErrorOr<std::unique_ptr<MemoryBuffer>>
437 MemoryBuffer::getFileAsStream(const Twine &Filename) {
438   int FD;
439   std::error_code EC = sys::fs::openFileForRead(Filename, FD);
440   if (EC)
441     return EC;
442   ErrorOr<std::unique_ptr<MemoryBuffer>> Ret =
443       getMemoryBufferForStream(FD, Filename);
444   close(FD);
445   return Ret;
446 }
447 
448 MemoryBufferRef MemoryBuffer::getMemBufferRef() const {
449   StringRef Data = getBuffer();
450   StringRef Identifier = getBufferIdentifier();
451   return MemoryBufferRef(Data, Identifier);
452 }
453