| 1 |
/* |
| 2 |
* Copyright (c) Meta Platforms, Inc. and affiliates. |
| 3 |
* All rights reserved. |
| 4 |
* |
| 5 |
* This source code is licensed under both the BSD-style license (found in the |
| 6 |
* LICENSE file in the root directory of this source tree) and the GPLv2 (found |
| 7 |
* in the COPYING file in the root directory of this source tree). |
| 8 |
*/ |
| 9 |
#include "platform.h" /* Large Files support, SET_BINARY_MODE */ |
| 10 |
#include "Pzstd.h" |
| 11 |
#include "SkippableFrame.h" |
| 12 |
#include "utils/FileSystem.h" |
| 13 |
#include "utils/Portability.h" |
| 14 |
#include "utils/Range.h" |
| 15 |
#include "utils/ScopeGuard.h" |
| 16 |
#include "utils/ThreadPool.h" |
| 17 |
#include "utils/WorkQueue.h" |
| 18 |
|
| 19 |
#include <algorithm> |
| 20 |
#include <chrono> |
| 21 |
#include <cinttypes> |
| 22 |
#include <cstddef> |
| 23 |
#include <cstdio> |
| 24 |
#include <memory> |
| 25 |
#include <string> |
| 26 |
|
| 27 |
|
| 28 |
namespace pzstd { |
| 29 |
|
| 30 |
namespace { |
| 31 |
#ifdef _WIN32 |
| 32 |
const std::string nullOutput = "nul"; |
| 33 |
#else |
| 34 |
const std::string nullOutput = "/dev/null"; |
| 35 |
#endif |
| 36 |
} |
| 37 |
|
| 38 |
using std::size_t; |
| 39 |
|
| 40 |
static std::uintmax_t fileSizeOrZero(const std::string &file) { |
| 41 |
if (file == "-") { |
| 42 |
return 0; |
| 43 |
} |
| 44 |
std::error_code ec; |
| 45 |
auto size = file_size(file, ec); |
| 46 |
if (ec) { |
| 47 |
size = 0; |
| 48 |
} |
| 49 |
return size; |
| 50 |
} |
| 51 |
|
| 52 |
static std::uint64_t handleOneInput(const Options &options, |
| 53 |
const std::string &inputFile, |
| 54 |
FILE* inputFd, |
| 55 |
const std::string &outputFile, |
| 56 |
FILE* outputFd, |
| 57 |
SharedState& state) { |
| 58 |
auto inputSize = fileSizeOrZero(inputFile); |
| 59 |
// WorkQueue outlives ThreadPool so in the case of error we are certain |
| 60 |
// we don't accidentally try to call push() on it after it is destroyed |
| 61 |
WorkQueue<std::shared_ptr<BufferWorkQueue>> outs{options.numThreads + 1}; |
| 62 |
std::uint64_t bytesRead; |
| 63 |
std::uint64_t bytesWritten; |
| 64 |
{ |
| 65 |
// Initialize the (de)compression thread pool with numThreads |
| 66 |
ThreadPool executor(options.numThreads); |
| 67 |
// Run the reader thread on an extra thread |
| 68 |
ThreadPool readExecutor(1); |
| 69 |
if (!options.decompress) { |
| 70 |
// Add a job that reads the input and starts all the compression jobs |
| 71 |
readExecutor.add( |
| 72 |
[&state, &outs, &executor, inputFd, inputSize, &options, &bytesRead] { |
| 73 |
bytesRead = asyncCompressChunks( |
| 74 |
state, |
| 75 |
outs, |
| 76 |
executor, |
| 77 |
inputFd, |
| 78 |
inputSize, |
| 79 |
options.numThreads, |
| 80 |
options.determineParameters()); |
| 81 |
}); |
| 82 |
// Start writing |
| 83 |
bytesWritten = writeFile(state, outs, outputFd, options.decompress); |
| 84 |
} else { |
| 85 |
// Add a job that reads the input and starts all the decompression jobs |
| 86 |
readExecutor.add([&state, &outs, &executor, inputFd, &bytesRead] { |
| 87 |
bytesRead = asyncDecompressFrames(state, outs, executor, inputFd); |
| 88 |
}); |
| 89 |
// Start writing |
| 90 |
bytesWritten = writeFile(state, outs, outputFd, options.decompress); |
| 91 |
} |
| 92 |
} |
| 93 |
if (!state.errorHolder.hasError()) { |
| 94 |
std::string inputFileName = inputFile == "-" ? "stdin" : inputFile; |
| 95 |
std::string outputFileName = outputFile == "-" ? "stdout" : outputFile; |
| 96 |
if (!options.decompress) { |
| 97 |
double ratio = static_cast<double>(bytesWritten) / |
| 98 |
static_cast<double>(bytesRead + !bytesRead); |
| 99 |
state.log(kLogInfo, "%-20s :%6.2f%% (%6" PRIu64 " => %6" PRIu64 |
| 100 |
" bytes, %s)\n", |
| 101 |
inputFileName.c_str(), ratio * 100, bytesRead, bytesWritten, |
| 102 |
outputFileName.c_str()); |
| 103 |
} else { |
| 104 |
state.log(kLogInfo, "%-20s: %" PRIu64 " bytes \n", |
| 105 |
inputFileName.c_str(),bytesWritten); |
| 106 |
} |
| 107 |
} |
| 108 |
return bytesWritten; |
| 109 |
} |
| 110 |
|
| 111 |
static FILE *openInputFile(const std::string &inputFile, |
| 112 |
ErrorHolder &errorHolder) { |
| 113 |
if (inputFile == "-") { |
| 114 |
SET_BINARY_MODE(stdin); |
| 115 |
return stdin; |
| 116 |
} |
| 117 |
// Check if input file is a directory |
| 118 |
{ |
| 119 |
std::error_code ec; |
| 120 |
if (is_directory(inputFile, ec)) { |
| 121 |
errorHolder.setError("Output file is a directory -- ignored"); |
| 122 |
return nullptr; |
| 123 |
} |
| 124 |
} |
| 125 |
auto inputFd = std::fopen(inputFile.c_str(), "rb"); |
| 126 |
if (!errorHolder.check(inputFd != nullptr, "Failed to open input file")) { |
| 127 |
return nullptr; |
| 128 |
} |
| 129 |
return inputFd; |
| 130 |
} |
| 131 |
|
| 132 |
static FILE *openOutputFile(const Options &options, |
| 133 |
const std::string &outputFile, |
| 134 |
SharedState& state) { |
| 135 |
if (outputFile == "-") { |
| 136 |
SET_BINARY_MODE(stdout); |
| 137 |
return stdout; |
| 138 |
} |
| 139 |
// Check if the output file exists and then open it |
| 140 |
if (!options.overwrite && outputFile != nullOutput) { |
| 141 |
auto outputFd = std::fopen(outputFile.c_str(), "rb"); |
| 142 |
if (outputFd != nullptr) { |
| 143 |
std::fclose(outputFd); |
| 144 |
if (!state.log.logsAt(kLogInfo)) { |
| 145 |
state.errorHolder.setError("Output file exists"); |
| 146 |
return nullptr; |
| 147 |
} |
| 148 |
state.log( |
| 149 |
kLogInfo, |
| 150 |
"pzstd: %s already exists; do you wish to overwrite (y/n) ? ", |
| 151 |
outputFile.c_str()); |
| 152 |
int c = getchar(); |
| 153 |
if (c != 'y' && c != 'Y') { |
| 154 |
state.errorHolder.setError("Not overwritten"); |
| 155 |
return nullptr; |
| 156 |
} |
| 157 |
} |
| 158 |
} |
| 159 |
auto outputFd = std::fopen(outputFile.c_str(), "wb"); |
| 160 |
if (!state.errorHolder.check( |
| 161 |
outputFd != nullptr, "Failed to open output file")) { |
| 162 |
return nullptr; |
| 163 |
} |
| 164 |
return outputFd; |
| 165 |
} |
| 166 |
|
| 167 |
int pzstdMain(const Options &options) { |
| 168 |
int returnCode = 0; |
| 169 |
SharedState state(options); |
| 170 |
for (const auto& input : options.inputFiles) { |
| 171 |
// Setup the shared state |
| 172 |
auto printErrorGuard = makeScopeGuard([&] { |
| 173 |
if (state.errorHolder.hasError()) { |
| 174 |
returnCode = 1; |
| 175 |
state.log(kLogError, "pzstd: %s: %s.\n", input.c_str(), |
| 176 |
state.errorHolder.getError().c_str()); |
| 177 |
} |
| 178 |
}); |
| 179 |
// Open the input file |
| 180 |
auto inputFd = openInputFile(input, state.errorHolder); |
| 181 |
if (inputFd == nullptr) { |
| 182 |
continue; |
| 183 |
} |
| 184 |
auto closeInputGuard = makeScopeGuard([&] { std::fclose(inputFd); }); |
| 185 |
// Open the output file |
| 186 |
auto outputFile = options.getOutputFile(input); |
| 187 |
if (!state.errorHolder.check(outputFile != "", |
| 188 |
"Input file does not have extension .zst")) { |
| 189 |
continue; |
| 190 |
} |
| 191 |
auto outputFd = openOutputFile(options, outputFile, state); |
| 192 |
if (outputFd == nullptr) { |
| 193 |
continue; |
| 194 |
} |
| 195 |
auto closeOutputGuard = makeScopeGuard([&] { std::fclose(outputFd); }); |
| 196 |
// (de)compress the file |
| 197 |
handleOneInput(options, input, inputFd, outputFile, outputFd, state); |
| 198 |
if (state.errorHolder.hasError()) { |
| 199 |
continue; |
| 200 |
} |
| 201 |
// Delete the input file if necessary |
| 202 |
if (!options.keepSource) { |
| 203 |
// Be sure that we are done and have written everything before we delete |
| 204 |
if (!state.errorHolder.check(std::fclose(inputFd) == 0, |
| 205 |
"Failed to close input file")) { |
| 206 |
continue; |
| 207 |
} |
| 208 |
closeInputGuard.dismiss(); |
| 209 |
if (!state.errorHolder.check(std::fclose(outputFd) == 0, |
| 210 |
"Failed to close output file")) { |
| 211 |
continue; |
| 212 |
} |
| 213 |
closeOutputGuard.dismiss(); |
| 214 |
if (std::remove(input.c_str()) != 0) { |
| 215 |
state.errorHolder.setError("Failed to remove input file"); |
| 216 |
continue; |
| 217 |
} |
| 218 |
} |
| 219 |
} |
| 220 |
// Returns 1 if any of the files failed to (de)compress. |
| 221 |
return returnCode; |
| 222 |
} |
| 223 |
|
| 224 |
/// Construct a `ZSTD_inBuffer` that points to the data in `buffer`. |
| 225 |
static ZSTD_inBuffer makeZstdInBuffer(const Buffer& buffer) { |
| 226 |
return ZSTD_inBuffer{buffer.data(), buffer.size(), 0}; |
| 227 |
} |
| 228 |
|
| 229 |
/** |
| 230 |
* Advance `buffer` and `inBuffer` by the amount of data read, as indicated by |
| 231 |
* `inBuffer.pos`. |
| 232 |
*/ |
| 233 |
void advance(Buffer& buffer, ZSTD_inBuffer& inBuffer) { |
| 234 |
auto pos = inBuffer.pos; |
| 235 |
inBuffer.src = static_cast<const unsigned char*>(inBuffer.src) + pos; |
| 236 |
inBuffer.size -= pos; |
| 237 |
inBuffer.pos = 0; |
| 238 |
return buffer.advance(pos); |
| 239 |
} |
| 240 |
|
| 241 |
/// Construct a `ZSTD_outBuffer` that points to the data in `buffer`. |
| 242 |
static ZSTD_outBuffer makeZstdOutBuffer(Buffer& buffer) { |
| 243 |
return ZSTD_outBuffer{buffer.data(), buffer.size(), 0}; |
| 244 |
} |
| 245 |
|
| 246 |
/** |
| 247 |
* Split `buffer` and advance `outBuffer` by the amount of data written, as |
| 248 |
* indicated by `outBuffer.pos`. |
| 249 |
*/ |
| 250 |
Buffer split(Buffer& buffer, ZSTD_outBuffer& outBuffer) { |
| 251 |
auto pos = outBuffer.pos; |
| 252 |
outBuffer.dst = static_cast<unsigned char*>(outBuffer.dst) + pos; |
| 253 |
outBuffer.size -= pos; |
| 254 |
outBuffer.pos = 0; |
| 255 |
return buffer.splitAt(pos); |
| 256 |
} |
| 257 |
|
| 258 |
/** |
| 259 |
* Stream chunks of input from `in`, compress it, and stream it out to `out`. |
| 260 |
* |
| 261 |
* @param state The shared state |
| 262 |
* @param in Queue that we `pop()` input buffers from |
| 263 |
* @param out Queue that we `push()` compressed output buffers to |
| 264 |
* @param maxInputSize An upper bound on the size of the input |
| 265 |
*/ |
| 266 |
static void compress( |
| 267 |
SharedState& state, |
| 268 |
std::shared_ptr<BufferWorkQueue> in, |
| 269 |
std::shared_ptr<BufferWorkQueue> out, |
| 270 |
size_t maxInputSize) { |
| 271 |
auto& errorHolder = state.errorHolder; |
| 272 |
auto guard = makeScopeGuard([&] { |
| 273 |
in->finish(); |
| 274 |
out->finish(); |
| 275 |
}); |
| 276 |
// Initialize the CCtx |
| 277 |
auto ctx = state.cStreamPool->get(); |
| 278 |
if (!errorHolder.check(ctx != nullptr, "Failed to allocate ZSTD_CStream")) { |
| 279 |
return; |
| 280 |
} |
| 281 |
{ |
| 282 |
auto err = ZSTD_CCtx_reset(ctx.get(), ZSTD_reset_session_only); |
| 283 |
if (!errorHolder.check(!ZSTD_isError(err), ZSTD_getErrorName(err))) { |
| 284 |
return; |
| 285 |
} |
| 286 |
} |
| 287 |
|
| 288 |
// Allocate space for the result |
| 289 |
auto outBuffer = Buffer(ZSTD_compressBound(maxInputSize)); |
| 290 |
auto zstdOutBuffer = makeZstdOutBuffer(outBuffer); |
| 291 |
{ |
| 292 |
Buffer inBuffer; |
| 293 |
// Read a buffer in from the input queue |
| 294 |
while (in->pop(inBuffer) && !errorHolder.hasError()) { |
| 295 |
auto zstdInBuffer = makeZstdInBuffer(inBuffer); |
| 296 |
// Compress the whole buffer and send it to the output queue |
| 297 |
while (!inBuffer.empty() && !errorHolder.hasError()) { |
| 298 |
if (!errorHolder.check( |
| 299 |
!outBuffer.empty(), "ZSTD_compressBound() was too small")) { |
| 300 |
return; |
| 301 |
} |
| 302 |
// Compress |
| 303 |
auto err = |
| 304 |
ZSTD_compressStream(ctx.get(), &zstdOutBuffer, &zstdInBuffer); |
| 305 |
if (!errorHolder.check(!ZSTD_isError(err), ZSTD_getErrorName(err))) { |
| 306 |
return; |
| 307 |
} |
| 308 |
// Split the compressed data off outBuffer and pass to the output queue |
| 309 |
out->push(split(outBuffer, zstdOutBuffer)); |
| 310 |
// Forget about the data we already compressed |
| 311 |
advance(inBuffer, zstdInBuffer); |
| 312 |
} |
| 313 |
} |
| 314 |
} |
| 315 |
// Write the epilog |
| 316 |
size_t bytesLeft; |
| 317 |
do { |
| 318 |
if (!errorHolder.check( |
| 319 |
!outBuffer.empty(), "ZSTD_compressBound() was too small")) { |
| 320 |
return; |
| 321 |
} |
| 322 |
bytesLeft = ZSTD_endStream(ctx.get(), &zstdOutBuffer); |
| 323 |
if (!errorHolder.check( |
| 324 |
!ZSTD_isError(bytesLeft), ZSTD_getErrorName(bytesLeft))) { |
| 325 |
return; |
| 326 |
} |
| 327 |
out->push(split(outBuffer, zstdOutBuffer)); |
| 328 |
} while (bytesLeft != 0 && !errorHolder.hasError()); |
| 329 |
} |
| 330 |
|
| 331 |
/** |
| 332 |
* Calculates how large each independently compressed frame should be. |
| 333 |
* |
| 334 |
* @param size The size of the source if known, 0 otherwise |
| 335 |
* @param numThreads The number of threads available to run compression jobs on |
| 336 |
* @param params The zstd parameters to be used for compression |
| 337 |
*/ |
| 338 |
static size_t calculateStep( |
| 339 |
std::uintmax_t size, |
| 340 |
size_t numThreads, |
| 341 |
const ZSTD_parameters ¶ms) { |
| 342 |
(void)size; |
| 343 |
(void)numThreads; |
| 344 |
// Not validated to work correctly for window logs > 23. |
| 345 |
// It will definitely fail if windowLog + 2 is >= 4GB because |
| 346 |
// the skippable frame can only store sizes up to 4GB. |
| 347 |
assert(params.cParams.windowLog <= 23); |
| 348 |
return size_t{1} << (params.cParams.windowLog + 2); |
| 349 |
} |
| 350 |
|
| 351 |
namespace { |
| 352 |
enum class FileStatus { Continue, Done, Error }; |
| 353 |
/// Determines the status of the file descriptor `fd`. |
| 354 |
FileStatus fileStatus(FILE* fd) { |
| 355 |
if (std::feof(fd)) { |
| 356 |
return FileStatus::Done; |
| 357 |
} else if (std::ferror(fd)) { |
| 358 |
return FileStatus::Error; |
| 359 |
} |
| 360 |
return FileStatus::Continue; |
| 361 |
} |
| 362 |
} // anonymous namespace |
| 363 |
|
| 364 |
/** |
| 365 |
* Reads `size` data in chunks of `chunkSize` and puts it into `queue`. |
| 366 |
* Will read less if an error or EOF occurs. |
| 367 |
* Returns the status of the file after all of the reads have occurred. |
| 368 |
*/ |
| 369 |
static FileStatus |
| 370 |
readData(BufferWorkQueue& queue, size_t chunkSize, size_t size, FILE* fd, |
| 371 |
std::uint64_t *totalBytesRead) { |
| 372 |
Buffer buffer(size); |
| 373 |
while (!buffer.empty()) { |
| 374 |
auto bytesRead = |
| 375 |
std::fread(buffer.data(), 1, std::min(chunkSize, buffer.size()), fd); |
| 376 |
*totalBytesRead += bytesRead; |
| 377 |
queue.push(buffer.splitAt(bytesRead)); |
| 378 |
auto status = fileStatus(fd); |
| 379 |
if (status != FileStatus::Continue) { |
| 380 |
return status; |
| 381 |
} |
| 382 |
} |
| 383 |
return FileStatus::Continue; |
| 384 |
} |
| 385 |
|
| 386 |
std::uint64_t asyncCompressChunks( |
| 387 |
SharedState& state, |
| 388 |
WorkQueue<std::shared_ptr<BufferWorkQueue>>& chunks, |
| 389 |
ThreadPool& executor, |
| 390 |
FILE* fd, |
| 391 |
std::uintmax_t size, |
| 392 |
size_t numThreads, |
| 393 |
ZSTD_parameters params) { |
| 394 |
auto chunksGuard = makeScopeGuard([&] { chunks.finish(); }); |
| 395 |
std::uint64_t bytesRead = 0; |
| 396 |
|
| 397 |
// Break the input up into chunks of size `step` and compress each chunk |
| 398 |
// independently. |
| 399 |
size_t step = calculateStep(size, numThreads, params); |
| 400 |
state.log(kLogDebug, "Chosen frame size: %zu\n", step); |
| 401 |
auto status = FileStatus::Continue; |
| 402 |
while (status == FileStatus::Continue && !state.errorHolder.hasError()) { |
| 403 |
// Make a new input queue that we will put the chunk's input data into. |
| 404 |
auto in = std::make_shared<BufferWorkQueue>(); |
| 405 |
auto inGuard = makeScopeGuard([&] { in->finish(); }); |
| 406 |
// Make a new output queue that compress will put the compressed data into. |
| 407 |
auto out = std::make_shared<BufferWorkQueue>(); |
| 408 |
// Start compression in the thread pool |
| 409 |
executor.add([&state, in, out, step] { |
| 410 |
return compress( |
| 411 |
state, std::move(in), std::move(out), step); |
| 412 |
}); |
| 413 |
// Pass the output queue to the writer thread. |
| 414 |
chunks.push(std::move(out)); |
| 415 |
state.log(kLogVerbose, "%s\n", "Starting a new frame"); |
| 416 |
// Fill the input queue for the compression job we just started |
| 417 |
status = readData(*in, ZSTD_CStreamInSize(), step, fd, &bytesRead); |
| 418 |
} |
| 419 |
state.errorHolder.check(status != FileStatus::Error, "Error reading input"); |
| 420 |
return bytesRead; |
| 421 |
} |
| 422 |
|
| 423 |
/** |
| 424 |
* Decompress a frame, whose data is streamed into `in`, and stream the output |
| 425 |
* to `out`. |
| 426 |
* |
| 427 |
* @param state The shared state |
| 428 |
* @param in Queue that we `pop()` input buffers from. It contains |
| 429 |
* exactly one compressed frame. |
| 430 |
* @param out Queue that we `push()` decompressed output buffers to |
| 431 |
*/ |
| 432 |
static void decompress( |
| 433 |
SharedState& state, |
| 434 |
std::shared_ptr<BufferWorkQueue> in, |
| 435 |
std::shared_ptr<BufferWorkQueue> out) { |
| 436 |
auto& errorHolder = state.errorHolder; |
| 437 |
auto guard = makeScopeGuard([&] { |
| 438 |
in->finish(); |
| 439 |
out->finish(); |
| 440 |
}); |
| 441 |
// Initialize the DCtx |
| 442 |
auto ctx = state.dStreamPool->get(); |
| 443 |
if (!errorHolder.check(ctx != nullptr, "Failed to allocate ZSTD_DStream")) { |
| 444 |
return; |
| 445 |
} |
| 446 |
{ |
| 447 |
auto err = ZSTD_DCtx_reset(ctx.get(), ZSTD_reset_session_only); |
| 448 |
if (!errorHolder.check(!ZSTD_isError(err), ZSTD_getErrorName(err))) { |
| 449 |
return; |
| 450 |
} |
| 451 |
} |
| 452 |
|
| 453 |
const size_t outSize = ZSTD_DStreamOutSize(); |
| 454 |
Buffer inBuffer; |
| 455 |
size_t returnCode = 0; |
| 456 |
// Read a buffer in from the input queue |
| 457 |
while (in->pop(inBuffer) && !errorHolder.hasError()) { |
| 458 |
auto zstdInBuffer = makeZstdInBuffer(inBuffer); |
| 459 |
// Decompress the whole buffer and send it to the output queue |
| 460 |
while (!inBuffer.empty() && !errorHolder.hasError()) { |
| 461 |
// Allocate a buffer with at least outSize bytes. |
| 462 |
Buffer outBuffer(outSize); |
| 463 |
auto zstdOutBuffer = makeZstdOutBuffer(outBuffer); |
| 464 |
// Decompress |
| 465 |
returnCode = |
| 466 |
ZSTD_decompressStream(ctx.get(), &zstdOutBuffer, &zstdInBuffer); |
| 467 |
if (!errorHolder.check( |
| 468 |
!ZSTD_isError(returnCode), ZSTD_getErrorName(returnCode))) { |
| 469 |
return; |
| 470 |
} |
| 471 |
// Pass the buffer with the decompressed data to the output queue |
| 472 |
out->push(split(outBuffer, zstdOutBuffer)); |
| 473 |
// Advance past the input we already read |
| 474 |
advance(inBuffer, zstdInBuffer); |
| 475 |
if (returnCode == 0) { |
| 476 |
// The frame is over, prepare to (maybe) start a new frame |
| 477 |
ZSTD_initDStream(ctx.get()); |
| 478 |
} |
| 479 |
} |
| 480 |
} |
| 481 |
if (!errorHolder.check(returnCode <= 1, "Incomplete block")) { |
| 482 |
return; |
| 483 |
} |
| 484 |
// We've given ZSTD_decompressStream all of our data, but there may still |
| 485 |
// be data to read. |
| 486 |
while (returnCode == 1) { |
| 487 |
// Allocate a buffer with at least outSize bytes. |
| 488 |
Buffer outBuffer(outSize); |
| 489 |
auto zstdOutBuffer = makeZstdOutBuffer(outBuffer); |
| 490 |
// Pass in no input. |
| 491 |
ZSTD_inBuffer zstdInBuffer{nullptr, 0, 0}; |
| 492 |
// Decompress |
| 493 |
returnCode = |
| 494 |
ZSTD_decompressStream(ctx.get(), &zstdOutBuffer, &zstdInBuffer); |
| 495 |
if (!errorHolder.check( |
| 496 |
!ZSTD_isError(returnCode), ZSTD_getErrorName(returnCode))) { |
| 497 |
return; |
| 498 |
} |
| 499 |
// Pass the buffer with the decompressed data to the output queue |
| 500 |
out->push(split(outBuffer, zstdOutBuffer)); |
| 501 |
} |
| 502 |
} |
| 503 |
|
| 504 |
std::uint64_t asyncDecompressFrames( |
| 505 |
SharedState& state, |
| 506 |
WorkQueue<std::shared_ptr<BufferWorkQueue>>& frames, |
| 507 |
ThreadPool& executor, |
| 508 |
FILE* fd) { |
| 509 |
auto framesGuard = makeScopeGuard([&] { frames.finish(); }); |
| 510 |
std::uint64_t totalBytesRead = 0; |
| 511 |
|
| 512 |
// Split the source up into its component frames. |
| 513 |
// If we find our recognized skippable frame we know the next frames size |
| 514 |
// which means that we can decompress each standard frame in independently. |
| 515 |
// Otherwise, we will decompress using only one decompression task. |
| 516 |
const size_t chunkSize = ZSTD_DStreamInSize(); |
| 517 |
auto status = FileStatus::Continue; |
| 518 |
while (status == FileStatus::Continue && !state.errorHolder.hasError()) { |
| 519 |
// Make a new input queue that we will put the frames's bytes into. |
| 520 |
auto in = std::make_shared<BufferWorkQueue>(); |
| 521 |
auto inGuard = makeScopeGuard([&] { in->finish(); }); |
| 522 |
// Make a output queue that decompress will put the decompressed data into |
| 523 |
auto out = std::make_shared<BufferWorkQueue>(); |
| 524 |
|
| 525 |
size_t frameSize; |
| 526 |
{ |
| 527 |
// Calculate the size of the next frame. |
| 528 |
// frameSize is 0 if the frame info can't be decoded. |
| 529 |
Buffer buffer(SkippableFrame::kSize); |
| 530 |
auto bytesRead = std::fread(buffer.data(), 1, buffer.size(), fd); |
| 531 |
totalBytesRead += bytesRead; |
| 532 |
status = fileStatus(fd); |
| 533 |
if (bytesRead == 0 && status != FileStatus::Continue) { |
| 534 |
break; |
| 535 |
} |
| 536 |
buffer.subtract(buffer.size() - bytesRead); |
| 537 |
frameSize = SkippableFrame::tryRead(buffer.range()); |
| 538 |
in->push(std::move(buffer)); |
| 539 |
} |
| 540 |
if (frameSize == 0) { |
| 541 |
// We hit a non SkippableFrame, so this will be the last job. |
| 542 |
// Make sure that we don't use too much memory |
| 543 |
in->setMaxSize(64); |
| 544 |
out->setMaxSize(64); |
| 545 |
} |
| 546 |
// Start decompression in the thread pool |
| 547 |
executor.add([&state, in, out] { |
| 548 |
return decompress(state, std::move(in), std::move(out)); |
| 549 |
}); |
| 550 |
// Pass the output queue to the writer thread |
| 551 |
frames.push(std::move(out)); |
| 552 |
if (frameSize == 0) { |
| 553 |
// We hit a non SkippableFrame ==> not compressed by pzstd or corrupted |
| 554 |
// Pass the rest of the source to this decompression task |
| 555 |
state.log(kLogVerbose, "%s\n", |
| 556 |
"Input not in pzstd format, falling back to serial decompression"); |
| 557 |
while (status == FileStatus::Continue && !state.errorHolder.hasError()) { |
| 558 |
status = readData(*in, chunkSize, chunkSize, fd, &totalBytesRead); |
| 559 |
} |
| 560 |
break; |
| 561 |
} |
| 562 |
state.log(kLogVerbose, "Decompressing a frame of size %zu", frameSize); |
| 563 |
// Fill the input queue for the decompression job we just started |
| 564 |
status = readData(*in, chunkSize, frameSize, fd, &totalBytesRead); |
| 565 |
} |
| 566 |
state.errorHolder.check(status != FileStatus::Error, "Error reading input"); |
| 567 |
return totalBytesRead; |
| 568 |
} |
| 569 |
|
| 570 |
/// Write `data` to `fd`, returns true iff success. |
| 571 |
static bool writeData(ByteRange data, FILE* fd) { |
| 572 |
while (!data.empty()) { |
| 573 |
data.advance(std::fwrite(data.begin(), 1, data.size(), fd)); |
| 574 |
if (std::ferror(fd)) { |
| 575 |
return false; |
| 576 |
} |
| 577 |
} |
| 578 |
return true; |
| 579 |
} |
| 580 |
|
| 581 |
std::uint64_t writeFile( |
| 582 |
SharedState& state, |
| 583 |
WorkQueue<std::shared_ptr<BufferWorkQueue>>& outs, |
| 584 |
FILE* outputFd, |
| 585 |
bool decompress) { |
| 586 |
auto& errorHolder = state.errorHolder; |
| 587 |
auto outsFinishGuard = makeScopeGuard([&outs] { outs.finish(); }); |
| 588 |
auto lineClearGuard = makeScopeGuard([&state] { |
| 589 |
state.log.clear(kLogInfo); |
| 590 |
}); |
| 591 |
std::uint64_t bytesWritten = 0; |
| 592 |
std::shared_ptr<BufferWorkQueue> out; |
| 593 |
// Grab the output queue for each decompression job (in order). |
| 594 |
while (outs.pop(out)) { |
| 595 |
auto outFinishGuard = makeScopeGuard([&out] { out->finish(); }); |
| 596 |
if (errorHolder.hasError()) { |
| 597 |
continue; |
| 598 |
} |
| 599 |
if (!decompress) { |
| 600 |
// If we are compressing and want to write skippable frames we can't |
| 601 |
// start writing before compression is done because we need to know the |
| 602 |
// compressed size. |
| 603 |
// Wait for the compressed size to be available and write skippable frame |
| 604 |
assert(uint64_t(out->size()) < uint64_t(1) << 32); |
| 605 |
SkippableFrame frame(uint32_t(out->size())); |
| 606 |
if (!writeData(frame.data(), outputFd)) { |
| 607 |
errorHolder.setError("Failed to write output"); |
| 608 |
return bytesWritten; |
| 609 |
} |
| 610 |
bytesWritten += frame.kSize; |
| 611 |
} |
| 612 |
// For each chunk of the frame: Pop it from the queue and write it |
| 613 |
Buffer buffer; |
| 614 |
while (out->pop(buffer) && !errorHolder.hasError()) { |
| 615 |
if (!writeData(buffer.range(), outputFd)) { |
| 616 |
errorHolder.setError("Failed to write output"); |
| 617 |
return bytesWritten; |
| 618 |
} |
| 619 |
bytesWritten += buffer.size(); |
| 620 |
state.log.update(kLogInfo, "Written: %u MB ", |
| 621 |
static_cast<std::uint32_t>(bytesWritten >> 20)); |
| 622 |
} |
| 623 |
} |
| 624 |
return bytesWritten; |
| 625 |
} |
| 626 |
} |