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