lib/Support/MemoryBuffer.cpp

   1 //===--- MemoryBuffer.cpp - Memory Buffer implementation ------------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 //  This file implements the MemoryBuffer interface.
  10 //
  11 //===----------------------------------------------------------------------===//
  12
  13 #include "llvm/Support/MemoryBuffer.h"
  14 #include "llvm/ADT/SmallString.h"
  15 #include "llvm/Config/config.h"
  16 #include "llvm/Support/Errc.h"
  17 #include "llvm/Support/Errno.h"
  18 #include "llvm/Support/FileSystem.h"
  19 #include "llvm/Support/MathExtras.h"
  20 #include "llvm/Support/Path.h"
  21 #include "llvm/Support/Process.h"
  22 #include "llvm/Support/Program.h"
  23 #include "llvm/Support/SmallVectorMemoryBuffer.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 template<typename MB>
  84 class MemoryBufferMem : public MB {
  85 public:
  86   MemoryBufferMem(StringRef InputData, bool RequiresNullTerminator) {
  87     MemoryBuffer::init(InputData.begin(), InputData.end(),
  88                        RequiresNullTerminator);
  89   }
  90
  91   /// Disable sized deallocation for MemoryBufferMem, because it has
  92   /// tail-allocated data.
  93   void operator delete(void *p) { ::operator delete(p); }
  94
  95   StringRef getBufferIdentifier() const override {
  96     // The name is stored after the class itself.
  97     return StringRef(reinterpret_cast<const char *>(this + 1));
  98   }
  99
 100   MemoryBuffer::BufferKind getBufferKind() const override {
 101     return MemoryBuffer::MemoryBuffer_Malloc;
 102   }
 103 };
 104 }
 105
 106 template <typename MB>
 107 static ErrorOr<std::unique_ptr<MB>>
 108 getFileAux(const Twine &Filename, int64_t FileSize, uint64_t MapSize,
 109            uint64_t Offset, bool RequiresNullTerminator, bool IsVolatile);
 110
 111 std::unique_ptr<MemoryBuffer>
 112 MemoryBuffer::getMemBuffer(StringRef InputData, StringRef BufferName,
 113                            bool RequiresNullTerminator) {
 114   auto *Ret = new (NamedBufferAlloc(BufferName))
 115       MemoryBufferMem<MemoryBuffer>(InputData, RequiresNullTerminator);
 116   return std::unique_ptr<MemoryBuffer>(Ret);
 117 }
 118
 119 std::unique_ptr<MemoryBuffer>
 120 MemoryBuffer::getMemBuffer(MemoryBufferRef Ref, bool RequiresNullTerminator) {
 121   return std::unique_ptr<MemoryBuffer>(getMemBuffer(
 122       Ref.getBuffer(), Ref.getBufferIdentifier(), RequiresNullTerminator));
 123 }
 124
 125 static ErrorOr<std::unique_ptr<WritableMemoryBuffer>>
 126 getMemBufferCopyImpl(StringRef InputData, const Twine &BufferName) {
 127   auto Buf = WritableMemoryBuffer::getNewUninitMemBuffer(InputData.size(), BufferName);
 128   if (!Buf)
 129     return make_error_code(errc::not_enough_memory);
 130   memcpy(Buf->getBufferStart(), InputData.data(), InputData.size());
 131   return std::move(Buf);
 132 }
 133
 134 std::unique_ptr<MemoryBuffer>
 135 MemoryBuffer::getMemBufferCopy(StringRef InputData, const Twine &BufferName) {
 136   auto Buf = getMemBufferCopyImpl(InputData, BufferName);
 137   if (Buf)
 138     return std::move(*Buf);
 139   return nullptr;
 140 }
 141
 142 ErrorOr<std::unique_ptr<MemoryBuffer>>
 143 MemoryBuffer::getFileOrSTDIN(const Twine &Filename, int64_t FileSize,
 144                              bool RequiresNullTerminator) {
 145   SmallString<256> NameBuf;
 146   StringRef NameRef = Filename.toStringRef(NameBuf);
 147
 148   if (NameRef == "-")
 149     return getSTDIN();
 150   return getFile(Filename, FileSize, RequiresNullTerminator);
 151 }
 152
 153 ErrorOr<std::unique_ptr<MemoryBuffer>>
 154 MemoryBuffer::getFileSlice(const Twine &FilePath, uint64_t MapSize,
 155                            uint64_t Offset, bool IsVolatile) {
 156   return getFileAux<MemoryBuffer>(FilePath, -1, MapSize, Offset, false,
 157                                   IsVolatile);
 158 }
 159
 160 //===----------------------------------------------------------------------===//
 161 // MemoryBuffer::getFile implementation.
 162 //===----------------------------------------------------------------------===//
 163
 164 namespace {
 165 /// Memory maps a file descriptor using sys::fs::mapped_file_region.
 166 ///
 167 /// This handles converting the offset into a legal offset on the platform.
 168 template<typename MB>
 169 class MemoryBufferMMapFile : public MB {
 170   sys::fs::mapped_file_region MFR;
 171
 172   static uint64_t getLegalMapOffset(uint64_t Offset) {
 173     return Offset & ~(sys::fs::mapped_file_region::alignment() - 1);
 174   }
 175
 176   static uint64_t getLegalMapSize(uint64_t Len, uint64_t Offset) {
 177     return Len + (Offset - getLegalMapOffset(Offset));
 178   }
 179
 180   const char *getStart(uint64_t Len, uint64_t Offset) {
 181     return MFR.const_data() + (Offset - getLegalMapOffset(Offset));
 182   }
 183
 184 public:
 185   MemoryBufferMMapFile(bool RequiresNullTerminator, int FD, uint64_t Len,
 186                        uint64_t Offset, std::error_code &EC)
 187       : MFR(FD, MB::Mapmode, getLegalMapSize(Len, Offset),
 188             getLegalMapOffset(Offset), EC) {
 189     if (!EC) {
 190       const char *Start = getStart(Len, Offset);
 191       MemoryBuffer::init(Start, Start + Len, RequiresNullTerminator);
 192     }
 193   }
 194
 195   /// Disable sized deallocation for MemoryBufferMMapFile, because it has
 196   /// tail-allocated data.
 197   void operator delete(void *p) { ::operator delete(p); }
 198
 199   StringRef getBufferIdentifier() const override {
 200     // The name is stored after the class itself.
 201     return StringRef(reinterpret_cast<const char *>(this + 1));
 202   }
 203
 204   MemoryBuffer::BufferKind getBufferKind() const override {
 205     return MemoryBuffer::MemoryBuffer_MMap;
 206   }
 207 };
 208 }
 209
 210 static ErrorOr<std::unique_ptr<WritableMemoryBuffer>>
 211 getMemoryBufferForStream(int FD, const Twine &BufferName) {
 212   const ssize_t ChunkSize = 4096*4;
 213   SmallString<ChunkSize> Buffer;
 214   ssize_t ReadBytes;
 215   // Read into Buffer until we hit EOF.
 216   do {
 217     Buffer.reserve(Buffer.size() + ChunkSize);
 218     ReadBytes = sys::RetryAfterSignal(-1, ::read, FD, Buffer.end(), ChunkSize);
 219     if (ReadBytes == -1)
 220       return std::error_code(errno, std::generic_category());
 221     Buffer.set_size(Buffer.size() + ReadBytes);
 222   } while (ReadBytes != 0);
 223
 224   return getMemBufferCopyImpl(Buffer, BufferName);
 225 }
 226
 227
 228 ErrorOr<std::unique_ptr<MemoryBuffer>>
 229 MemoryBuffer::getFile(const Twine &Filename, int64_t FileSize,
 230                       bool RequiresNullTerminator, bool IsVolatile) {
 231   return getFileAux<MemoryBuffer>(Filename, FileSize, FileSize, 0,
 232                                   RequiresNullTerminator, IsVolatile);
 233 }
 234
 235 template <typename MB>
 236 static ErrorOr<std::unique_ptr<MB>>
 237 getOpenFileImpl(int FD, const Twine &Filename, uint64_t FileSize,
 238                 uint64_t MapSize, int64_t Offset, bool RequiresNullTerminator,
 239                 bool IsVolatile);
 240
 241 template <typename MB>
 242 static ErrorOr<std::unique_ptr<MB>>
 243 getFileAux(const Twine &Filename, int64_t FileSize, uint64_t MapSize,
 244            uint64_t Offset, bool RequiresNullTerminator, bool IsVolatile) {
 245   int FD;
 246   std::error_code EC = sys::fs::openFileForRead(Filename, FD, sys::fs::OF_None);
 247
 248   if (EC)
 249     return EC;
 250
 251   auto Ret = getOpenFileImpl<MB>(FD, Filename, FileSize, MapSize, Offset,
 252                                  RequiresNullTerminator, IsVolatile);
 253   close(FD);
 254   return Ret;
 255 }
 256
 257 ErrorOr<std::unique_ptr<WritableMemoryBuffer>>
 258 WritableMemoryBuffer::getFile(const Twine &Filename, int64_t FileSize,
 259                               bool IsVolatile) {
 260   return getFileAux<WritableMemoryBuffer>(Filename, FileSize, FileSize, 0,
 261                                           /*RequiresNullTerminator*/ false,
 262                                           IsVolatile);
 263 }
 264
 265 ErrorOr<std::unique_ptr<WritableMemoryBuffer>>
 266 WritableMemoryBuffer::getFileSlice(const Twine &Filename, uint64_t MapSize,
 267                                    uint64_t Offset, bool IsVolatile) {
 268   return getFileAux<WritableMemoryBuffer>(Filename, -1, MapSize, Offset, false,
 269                                           IsVolatile);
 270 }
 271
 272 std::unique_ptr<WritableMemoryBuffer>
 273 WritableMemoryBuffer::getNewUninitMemBuffer(size_t Size, const Twine &BufferName) {
 274   using MemBuffer = MemoryBufferMem<WritableMemoryBuffer>;
 275   // Allocate space for the MemoryBuffer, the data and the name. It is important
 276   // that MemoryBuffer and data are aligned so PointerIntPair works with them.
 277   // TODO: Is 16-byte alignment enough?  We copy small object files with large
 278   // alignment expectations into this buffer.
 279   SmallString<256> NameBuf;
 280   StringRef NameRef = BufferName.toStringRef(NameBuf);
 281   size_t AlignedStringLen = alignTo(sizeof(MemBuffer) + NameRef.size() + 1, 16);
 282   size_t RealLen = AlignedStringLen + Size + 1;
 283   char *Mem = static_cast<char*>(operator new(RealLen, std::nothrow));
 284   if (!Mem)
 285     return nullptr;
 286
 287   // The name is stored after the class itself.
 288   CopyStringRef(Mem + sizeof(MemBuffer), NameRef);
 289
 290   // The buffer begins after the name and must be aligned.
 291   char *Buf = Mem + AlignedStringLen;
 292   Buf[Size] = 0; // Null terminate buffer.
 293
 294   auto *Ret = new (Mem) MemBuffer(StringRef(Buf, Size), true);
 295   return std::unique_ptr<WritableMemoryBuffer>(Ret);
 296 }
 297
 298 std::unique_ptr<WritableMemoryBuffer>
 299 WritableMemoryBuffer::getNewMemBuffer(size_t Size, const Twine &BufferName) {
 300   auto SB = WritableMemoryBuffer::getNewUninitMemBuffer(Size, BufferName);
 301   if (!SB)
 302     return nullptr;
 303   memset(SB->getBufferStart(), 0, Size);
 304   return SB;
 305 }
 306
 307 static bool shouldUseMmap(int FD,
 308                           size_t FileSize,
 309                           size_t MapSize,
 310                           off_t Offset,
 311                           bool RequiresNullTerminator,
 312                           int PageSize,
 313                           bool IsVolatile) {
 314   // mmap may leave the buffer without null terminator if the file size changed
 315   // by the time the last page is mapped in, so avoid it if the file size is
 316   // likely to change.
 317   if (IsVolatile)
 318     return false;
 319
 320   // We don't use mmap for small files because this can severely fragment our
 321   // address space.
 322   if (MapSize < 4 * 4096 || MapSize < (unsigned)PageSize)
 323     return false;
 324
 325   if (!RequiresNullTerminator)
 326     return true;
 327
 328   // If we don't know the file size, use fstat to find out.  fstat on an open
 329   // file descriptor is cheaper than stat on a random path.
 330   // FIXME: this chunk of code is duplicated, but it avoids a fstat when
 331   // RequiresNullTerminator = false and MapSize != -1.
 332   if (FileSize == size_t(-1)) {
 333     sys::fs::file_status Status;
 334     if (sys::fs::status(FD, Status))
 335       return false;
 336     FileSize = Status.getSize();
 337   }
 338
 339   // If we need a null terminator and the end of the map is inside the file,
 340   // we cannot use mmap.
 341   size_t End = Offset + MapSize;
 342   assert(End <= FileSize);
 343   if (End != FileSize)
 344     return false;
 345
 346   // Don't try to map files that are exactly a multiple of the system page size
 347   // if we need a null terminator.
 348   if ((FileSize & (PageSize -1)) == 0)
 349     return false;
 350
 351 #if defined(__CYGWIN__)
 352   // Don't try to map files that are exactly a multiple of the physical page size
 353   // if we need a null terminator.
 354   // FIXME: We should reorganize again getPageSize() on Win32.
 355   if ((FileSize & (4096 - 1)) == 0)
 356     return false;
 357 #endif
 358
 359   return true;
 360 }
 361
 362 static ErrorOr<std::unique_ptr<WriteThroughMemoryBuffer>>
 363 getReadWriteFile(const Twine &Filename, uint64_t FileSize, uint64_t MapSize,
 364                  uint64_t Offset) {
 365   int FD;
 366   std::error_code EC = sys::fs::openFileForReadWrite(
 367       Filename, FD, sys::fs::CD_OpenExisting, sys::fs::OF_None);
 368
 369   if (EC)
 370     return EC;
 371
 372   // Default is to map the full file.
 373   if (MapSize == uint64_t(-1)) {
 374     // If we don't know the file size, use fstat to find out.  fstat on an open
 375     // file descriptor is cheaper than stat on a random path.
 376     if (FileSize == uint64_t(-1)) {
 377       sys::fs::file_status Status;
 378       std::error_code EC = sys::fs::status(FD, Status);
 379       if (EC)
 380         return EC;
 381
 382       // If this not a file or a block device (e.g. it's a named pipe
 383       // or character device), we can't mmap it, so error out.
 384       sys::fs::file_type Type = Status.type();
 385       if (Type != sys::fs::file_type::regular_file &&
 386           Type != sys::fs::file_type::block_file)
 387         return make_error_code(errc::invalid_argument);
 388
 389       FileSize = Status.getSize();
 390     }
 391     MapSize = FileSize;
 392   }
 393
 394   std::unique_ptr<WriteThroughMemoryBuffer> Result(
 395       new (NamedBufferAlloc(Filename))
 396           MemoryBufferMMapFile<WriteThroughMemoryBuffer>(false, FD, MapSize,
 397                                                          Offset, EC));
 398   if (EC)
 399     return EC;
 400   return std::move(Result);
 401 }
 402
 403 ErrorOr<std::unique_ptr<WriteThroughMemoryBuffer>>
 404 WriteThroughMemoryBuffer::getFile(const Twine &Filename, int64_t FileSize) {
 405   return getReadWriteFile(Filename, FileSize, FileSize, 0);
 406 }
 407
 408 /// Map a subrange of the specified file as a WritableMemoryBuffer.
 409 ErrorOr<std::unique_ptr<WriteThroughMemoryBuffer>>
 410 WriteThroughMemoryBuffer::getFileSlice(const Twine &Filename, uint64_t MapSize,
 411                                        uint64_t Offset) {
 412   return getReadWriteFile(Filename, -1, MapSize, Offset);
 413 }
 414
 415 template <typename MB>
 416 static ErrorOr<std::unique_ptr<MB>>
 417 getOpenFileImpl(int FD, const Twine &Filename, uint64_t FileSize,
 418                 uint64_t MapSize, int64_t Offset, bool RequiresNullTerminator,
 419                 bool IsVolatile) {
 420   static int PageSize = sys::Process::getPageSize();
 421
 422   // Default is to map the full file.
 423   if (MapSize == uint64_t(-1)) {
 424     // If we don't know the file size, use fstat to find out.  fstat on an open
 425     // file descriptor is cheaper than stat on a random path.
 426     if (FileSize == uint64_t(-1)) {
 427       sys::fs::file_status Status;
 428       std::error_code EC = sys::fs::status(FD, Status);
 429       if (EC)
 430         return EC;
 431
 432       // If this not a file or a block device (e.g. it's a named pipe
 433       // or character device), we can't trust the size. Create the memory
 434       // buffer by copying off the stream.
 435       sys::fs::file_type Type = Status.type();
 436       if (Type != sys::fs::file_type::regular_file &&
 437           Type != sys::fs::file_type::block_file)
 438         return getMemoryBufferForStream(FD, Filename);
 439
 440       FileSize = Status.getSize();
 441     }
 442     MapSize = FileSize;
 443   }
 444
 445   if (shouldUseMmap(FD, FileSize, MapSize, Offset, RequiresNullTerminator,
 446                     PageSize, IsVolatile)) {
 447     std::error_code EC;
 448     std::unique_ptr<MB> Result(
 449         new (NamedBufferAlloc(Filename)) MemoryBufferMMapFile<MB>(
 450             RequiresNullTerminator, FD, MapSize, Offset, EC));
 451     if (!EC)
 452       return std::move(Result);
 453   }
 454
 455   auto Buf = WritableMemoryBuffer::getNewUninitMemBuffer(MapSize, Filename);
 456   if (!Buf) {
 457     // Failed to create a buffer. The only way it can fail is if
 458     // new(std::nothrow) returns 0.
 459     return make_error_code(errc::not_enough_memory);
 460   }
 461
 462   char *BufPtr = Buf.get()->getBufferStart();
 463
 464   size_t BytesLeft = MapSize;
 465 #ifndef HAVE_PREAD
 466   if (lseek(FD, Offset, SEEK_SET) == -1)
 467     return std::error_code(errno, std::generic_category());
 468 #endif
 469
 470   while (BytesLeft) {
 471 #ifdef HAVE_PREAD
 472     ssize_t NumRead = sys::RetryAfterSignal(-1, ::pread, FD, BufPtr, BytesLeft,
 473                                             MapSize - BytesLeft + Offset);
 474 #else
 475     ssize_t NumRead = sys::RetryAfterSignal(-1, ::read, FD, BufPtr, BytesLeft);
 476 #endif
 477     if (NumRead == -1) {
 478       // Error while reading.
 479       return std::error_code(errno, std::generic_category());
 480     }
 481     if (NumRead == 0) {
 482       memset(BufPtr, 0, BytesLeft); // zero-initialize rest of the buffer.
 483       break;
 484     }
 485     BytesLeft -= NumRead;
 486     BufPtr += NumRead;
 487   }
 488
 489   return std::move(Buf);
 490 }
 491
 492 ErrorOr<std::unique_ptr<MemoryBuffer>>
 493 MemoryBuffer::getOpenFile(int FD, const Twine &Filename, uint64_t FileSize,
 494                           bool RequiresNullTerminator, bool IsVolatile) {
 495   return getOpenFileImpl<MemoryBuffer>(FD, Filename, FileSize, FileSize, 0,
 496                          RequiresNullTerminator, IsVolatile);
 497 }
 498
 499 ErrorOr<std::unique_ptr<MemoryBuffer>>
 500 MemoryBuffer::getOpenFileSlice(int FD, const Twine &Filename, uint64_t MapSize,
 501                                int64_t Offset, bool IsVolatile) {
 502   assert(MapSize != uint64_t(-1));
 503   return getOpenFileImpl<MemoryBuffer>(FD, Filename, -1, MapSize, Offset, false,
 504                                        IsVolatile);
 505 }
 506
 507 ErrorOr<std::unique_ptr<MemoryBuffer>> MemoryBuffer::getSTDIN() {
 508   // Read in all of the data from stdin, we cannot mmap stdin.
 509   //
 510   // FIXME: That isn't necessarily true, we should try to mmap stdin and
 511   // fallback if it fails.
 512   sys::ChangeStdinToBinary();
 513
 514   return getMemoryBufferForStream(0, "<stdin>");
 515 }
 516
 517 ErrorOr<std::unique_ptr<MemoryBuffer>>
 518 MemoryBuffer::getFileAsStream(const Twine &Filename) {
 519   int FD;
 520   std::error_code EC = sys::fs::openFileForRead(Filename, FD, sys::fs::OF_None);
 521   if (EC)
 522     return EC;
 523   ErrorOr<std::unique_ptr<MemoryBuffer>> Ret =
 524       getMemoryBufferForStream(FD, Filename);
 525   close(FD);
 526   return Ret;
 527 }
 528
 529 MemoryBufferRef MemoryBuffer::getMemBufferRef() const {
 530   StringRef Data = getBuffer();
 531   StringRef Identifier = getBufferIdentifier();
 532   return MemoryBufferRef(Data, Identifier);
 533 }
 534
 535 SmallVectorMemoryBuffer::~SmallVectorMemoryBuffer() {}