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