zstd_helpers.c 5.8 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147
  1. /*
  2. * Copyright (c) 2016-2021, Facebook, Inc.
  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. * You may select, at your option, one of the above-listed licenses.
  9. */
  10. #define ZSTD_STATIC_LINKING_ONLY
  11. #define ZDICT_STATIC_LINKING_ONLY
  12. #include <string.h>
  13. #include "zstd_helpers.h"
  14. #include "fuzz_helpers.h"
  15. #include "zstd.h"
  16. #include "zdict.h"
  17. const int kMinClevel = -3;
  18. const int kMaxClevel = 19;
  19. static void set(ZSTD_CCtx *cctx, ZSTD_cParameter param, int value)
  20. {
  21. FUZZ_ZASSERT(ZSTD_CCtx_setParameter(cctx, param, value));
  22. }
  23. static void setRand(ZSTD_CCtx *cctx, ZSTD_cParameter param, unsigned min,
  24. unsigned max, FUZZ_dataProducer_t *producer) {
  25. unsigned const value = FUZZ_dataProducer_uint32Range(producer, min, max);
  26. set(cctx, param, value);
  27. }
  28. ZSTD_compressionParameters FUZZ_randomCParams(size_t srcSize, FUZZ_dataProducer_t *producer)
  29. {
  30. /* Select compression parameters */
  31. ZSTD_compressionParameters cParams;
  32. cParams.windowLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_WINDOWLOG_MIN, 15);
  33. cParams.hashLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_HASHLOG_MIN, 15);
  34. cParams.chainLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_CHAINLOG_MIN, 16);
  35. cParams.searchLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_SEARCHLOG_MIN, 9);
  36. cParams.minMatch = FUZZ_dataProducer_uint32Range(producer, ZSTD_MINMATCH_MIN,
  37. ZSTD_MINMATCH_MAX);
  38. cParams.targetLength = FUZZ_dataProducer_uint32Range(producer, 0, 512);
  39. cParams.strategy = FUZZ_dataProducer_uint32Range(producer, ZSTD_STRATEGY_MIN, ZSTD_STRATEGY_MAX);
  40. return ZSTD_adjustCParams(cParams, srcSize, 0);
  41. }
  42. ZSTD_frameParameters FUZZ_randomFParams(FUZZ_dataProducer_t *producer)
  43. {
  44. /* Select frame parameters */
  45. ZSTD_frameParameters fParams;
  46. fParams.contentSizeFlag = FUZZ_dataProducer_uint32Range(producer, 0, 1);
  47. fParams.checksumFlag = FUZZ_dataProducer_uint32Range(producer, 0, 1);
  48. fParams.noDictIDFlag = FUZZ_dataProducer_uint32Range(producer, 0, 1);
  49. return fParams;
  50. }
  51. ZSTD_parameters FUZZ_randomParams(size_t srcSize, FUZZ_dataProducer_t *producer)
  52. {
  53. ZSTD_parameters params;
  54. params.cParams = FUZZ_randomCParams(srcSize, producer);
  55. params.fParams = FUZZ_randomFParams(producer);
  56. return params;
  57. }
  58. void FUZZ_setRandomParameters(ZSTD_CCtx *cctx, size_t srcSize, FUZZ_dataProducer_t *producer)
  59. {
  60. ZSTD_compressionParameters cParams = FUZZ_randomCParams(srcSize, producer);
  61. set(cctx, ZSTD_c_windowLog, cParams.windowLog);
  62. set(cctx, ZSTD_c_hashLog, cParams.hashLog);
  63. set(cctx, ZSTD_c_chainLog, cParams.chainLog);
  64. set(cctx, ZSTD_c_searchLog, cParams.searchLog);
  65. set(cctx, ZSTD_c_minMatch, cParams.minMatch);
  66. set(cctx, ZSTD_c_targetLength, cParams.targetLength);
  67. set(cctx, ZSTD_c_strategy, cParams.strategy);
  68. /* Select frame parameters */
  69. setRand(cctx, ZSTD_c_contentSizeFlag, 0, 1, producer);
  70. setRand(cctx, ZSTD_c_checksumFlag, 0, 1, producer);
  71. setRand(cctx, ZSTD_c_dictIDFlag, 0, 1, producer);
  72. /* Select long distance matching parameters */
  73. setRand(cctx, ZSTD_c_enableLongDistanceMatching, 0, 1, producer);
  74. setRand(cctx, ZSTD_c_ldmHashLog, ZSTD_HASHLOG_MIN, 16, producer);
  75. setRand(cctx, ZSTD_c_ldmMinMatch, ZSTD_LDM_MINMATCH_MIN,
  76. ZSTD_LDM_MINMATCH_MAX, producer);
  77. setRand(cctx, ZSTD_c_ldmBucketSizeLog, 0, ZSTD_LDM_BUCKETSIZELOG_MAX,
  78. producer);
  79. setRand(cctx, ZSTD_c_ldmHashRateLog, ZSTD_LDM_HASHRATELOG_MIN,
  80. ZSTD_LDM_HASHRATELOG_MAX, producer);
  81. /* Set misc parameters */
  82. setRand(cctx, ZSTD_c_nbWorkers, 0, 2, producer);
  83. setRand(cctx, ZSTD_c_rsyncable, 0, 1, producer);
  84. setRand(cctx, ZSTD_c_forceMaxWindow, 0, 1, producer);
  85. setRand(cctx, ZSTD_c_literalCompressionMode, 0, 2, producer);
  86. setRand(cctx, ZSTD_c_forceAttachDict, 0, 2, producer);
  87. setRand(cctx, ZSTD_c_splitBlocks, 0, 1, producer);
  88. if (FUZZ_dataProducer_uint32Range(producer, 0, 1) == 0) {
  89. setRand(cctx, ZSTD_c_srcSizeHint, ZSTD_SRCSIZEHINT_MIN, 2 * srcSize, producer);
  90. }
  91. if (FUZZ_dataProducer_uint32Range(producer, 0, 1) == 0) {
  92. setRand(cctx, ZSTD_c_targetCBlockSize, ZSTD_TARGETCBLOCKSIZE_MIN, ZSTD_TARGETCBLOCKSIZE_MAX, producer);
  93. }
  94. }
  95. FUZZ_dict_t FUZZ_train(void const* src, size_t srcSize, FUZZ_dataProducer_t *producer)
  96. {
  97. size_t const dictSize = MAX(srcSize / 8, 1024);
  98. size_t const totalSampleSize = dictSize * 11;
  99. FUZZ_dict_t dict = { FUZZ_malloc(dictSize), dictSize };
  100. char* const samples = (char*)FUZZ_malloc(totalSampleSize);
  101. unsigned nbSamples = 100;
  102. size_t* const samplesSizes = (size_t*)FUZZ_malloc(sizeof(size_t) * nbSamples);
  103. size_t pos = 0;
  104. size_t sample = 0;
  105. ZDICT_fastCover_params_t params;
  106. for (sample = 0; sample < nbSamples; ++sample) {
  107. size_t const remaining = totalSampleSize - pos;
  108. size_t const offset = FUZZ_dataProducer_uint32Range(producer, 0, MAX(srcSize, 1) - 1);
  109. size_t const limit = MIN(srcSize - offset, remaining);
  110. size_t const toCopy = MIN(limit, remaining / (nbSamples - sample));
  111. memcpy(samples + pos, src + offset, toCopy);
  112. pos += toCopy;
  113. samplesSizes[sample] = toCopy;
  114. }
  115. memset(samples + pos, 0, totalSampleSize - pos);
  116. memset(&params, 0, sizeof(params));
  117. params.accel = 5;
  118. params.k = 40;
  119. params.d = 8;
  120. params.f = 14;
  121. params.zParams.compressionLevel = 1;
  122. dict.size = ZDICT_trainFromBuffer_fastCover(dict.buff, dictSize,
  123. samples, samplesSizes, nbSamples, params);
  124. if (ZSTD_isError(dict.size)) {
  125. free(dict.buff);
  126. memset(&dict, 0, sizeof(dict));
  127. }
  128. free(samplesSizes);
  129. free(samples);
  130. return dict;
  131. }