Coverage Report

Created: 2024-09-08 06:26

/src/zstd/tests/fuzz/zstd_helpers.c
Line
Count
Source (jump to first uncovered line)
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
 * You may select, at your option, one of the above-listed licenses.
9
 */
10
11
#define ZSTD_STATIC_LINKING_ONLY
12
#define ZDICT_STATIC_LINKING_ONLY
13
14
#include <string.h>
15
16
#include "zstd_helpers.h"
17
#include "fuzz_helpers.h"
18
#include "zstd.h"
19
#include "zdict.h"
20
#include "sequence_producer.h"
21
#include "fuzz_third_party_seq_prod.h"
22
23
const int kMinClevel = -3;
24
const int kMaxClevel = 19;
25
26
void* FUZZ_seqProdState = NULL;
27
28
static void set(ZSTD_CCtx *cctx, ZSTD_cParameter param, int value)
29
0
{
30
0
    FUZZ_ZASSERT(ZSTD_CCtx_setParameter(cctx, param, value));
31
0
}
32
33
static unsigned produceParamValue(unsigned min, unsigned max,
34
0
                                  FUZZ_dataProducer_t *producer) {
35
0
    return FUZZ_dataProducer_uint32Range(producer, min, max);
36
0
}
37
38
static void setRand(ZSTD_CCtx *cctx, ZSTD_cParameter param, unsigned min,
39
0
                    unsigned max, FUZZ_dataProducer_t *producer) {
40
0
    unsigned const value = produceParamValue(min, max, producer);
41
0
    set(cctx, param, value);
42
0
}
43
44
ZSTD_compressionParameters FUZZ_randomCParams(size_t srcSize, FUZZ_dataProducer_t *producer)
45
0
{
46
    /* Select compression parameters */
47
0
    ZSTD_compressionParameters cParams;
48
0
    cParams.windowLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_WINDOWLOG_MIN, 15);
49
0
    cParams.hashLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_HASHLOG_MIN, 15);
50
0
    cParams.chainLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_CHAINLOG_MIN, 16);
51
0
    cParams.searchLog = FUZZ_dataProducer_uint32Range(producer, ZSTD_SEARCHLOG_MIN, 9);
52
0
    cParams.minMatch = FUZZ_dataProducer_uint32Range(producer, ZSTD_MINMATCH_MIN,
53
0
                                          ZSTD_MINMATCH_MAX);
54
0
    cParams.targetLength = FUZZ_dataProducer_uint32Range(producer, 0, 512);
55
0
    cParams.strategy = FUZZ_dataProducer_uint32Range(producer, ZSTD_STRATEGY_MIN, ZSTD_STRATEGY_MAX);
56
0
    return ZSTD_adjustCParams(cParams, srcSize, 0);
57
0
}
58
59
ZSTD_frameParameters FUZZ_randomFParams(FUZZ_dataProducer_t *producer)
60
0
{
61
    /* Select frame parameters */
62
0
    ZSTD_frameParameters fParams;
63
0
    fParams.contentSizeFlag = FUZZ_dataProducer_uint32Range(producer, 0, 1);
64
0
    fParams.checksumFlag = FUZZ_dataProducer_uint32Range(producer, 0, 1);
65
0
    fParams.noDictIDFlag = FUZZ_dataProducer_uint32Range(producer, 0, 1);
66
0
    return fParams;
67
0
}
68
69
ZSTD_parameters FUZZ_randomParams(size_t srcSize, FUZZ_dataProducer_t *producer)
70
0
{
71
0
    ZSTD_parameters params;
72
0
    params.cParams = FUZZ_randomCParams(srcSize, producer);
73
0
    params.fParams = FUZZ_randomFParams(producer);
74
0
    return params;
75
0
}
76
77
0
static void setSequenceProducerParams(ZSTD_CCtx *cctx, FUZZ_dataProducer_t *producer) {
78
#ifdef FUZZ_THIRD_PARTY_SEQ_PROD
79
    ZSTD_registerSequenceProducer(
80
        cctx,
81
        FUZZ_seqProdState,
82
        FUZZ_thirdPartySeqProd
83
    );
84
#else
85
0
    ZSTD_registerSequenceProducer(
86
0
        cctx,
87
0
        NULL,
88
0
        simpleSequenceProducer
89
0
    );
90
0
#endif
91
92
#ifdef FUZZ_THIRD_PARTY_SEQ_PROD
93
    FUZZ_ZASSERT(ZSTD_CCtx_setParameter(cctx, ZSTD_c_enableSeqProducerFallback, 1));
94
#else
95
0
    setRand(cctx, ZSTD_c_enableSeqProducerFallback, 0, 1, producer);
96
0
#endif
97
0
    FUZZ_ZASSERT(ZSTD_CCtx_setParameter(cctx, ZSTD_c_nbWorkers, 0));
98
0
    FUZZ_ZASSERT(ZSTD_CCtx_setParameter(cctx, ZSTD_c_enableLongDistanceMatching, ZSTD_ps_disable));
99
0
}
100
101
void FUZZ_setRandomParameters(ZSTD_CCtx *cctx, size_t srcSize, FUZZ_dataProducer_t *producer)
102
0
{
103
0
    ZSTD_compressionParameters cParams = FUZZ_randomCParams(srcSize, producer);
104
0
    set(cctx, ZSTD_c_windowLog, cParams.windowLog);
105
0
    set(cctx, ZSTD_c_hashLog, cParams.hashLog);
106
0
    set(cctx, ZSTD_c_chainLog, cParams.chainLog);
107
0
    set(cctx, ZSTD_c_searchLog, cParams.searchLog);
108
0
    set(cctx, ZSTD_c_minMatch, cParams.minMatch);
109
0
    set(cctx, ZSTD_c_targetLength, cParams.targetLength);
110
0
    set(cctx, ZSTD_c_strategy, cParams.strategy);
111
    /* Select frame parameters */
112
0
    setRand(cctx, ZSTD_c_contentSizeFlag, 0, 1, producer);
113
0
    setRand(cctx, ZSTD_c_checksumFlag, 0, 1, producer);
114
0
    setRand(cctx, ZSTD_c_dictIDFlag, 0, 1, producer);
115
    /* Select long distance matching parameters */
116
0
    setRand(cctx, ZSTD_c_enableLongDistanceMatching, ZSTD_ps_auto, ZSTD_ps_disable, producer);
117
0
    setRand(cctx, ZSTD_c_ldmHashLog, ZSTD_HASHLOG_MIN, 16, producer);
118
0
    setRand(cctx, ZSTD_c_ldmMinMatch, ZSTD_LDM_MINMATCH_MIN,
119
0
            ZSTD_LDM_MINMATCH_MAX, producer);
120
0
    setRand(cctx, ZSTD_c_ldmBucketSizeLog, 0, ZSTD_LDM_BUCKETSIZELOG_MAX,
121
0
            producer);
122
0
    setRand(cctx, ZSTD_c_ldmHashRateLog, ZSTD_LDM_HASHRATELOG_MIN,
123
0
            ZSTD_LDM_HASHRATELOG_MAX, producer);
124
    /* Set misc parameters */
125
#ifndef ZSTD_MULTITHREAD
126
    // To reproduce with or without ZSTD_MULTITHREAD, we are going to use
127
    // the same amount of entropy.
128
    unsigned const nbWorkers_value = produceParamValue(0, 2, producer);
129
    unsigned const rsyncable_value = produceParamValue(0, 1, producer);
130
    (void)nbWorkers_value;
131
    (void)rsyncable_value;
132
    set(cctx, ZSTD_c_nbWorkers, 0);
133
    set(cctx, ZSTD_c_rsyncable, 0);
134
#else
135
0
    setRand(cctx, ZSTD_c_nbWorkers, 0, 2, producer);
136
0
    setRand(cctx, ZSTD_c_rsyncable, 0, 1, producer);
137
0
#endif
138
0
    setRand(cctx, ZSTD_c_useRowMatchFinder, 0, 2, producer);
139
0
    setRand(cctx, ZSTD_c_enableDedicatedDictSearch, 0, 1, producer);
140
0
    setRand(cctx, ZSTD_c_forceMaxWindow, 0, 1, producer);
141
0
    setRand(cctx, ZSTD_c_literalCompressionMode, 0, 2, producer);
142
0
    setRand(cctx, ZSTD_c_forceAttachDict, 0, 2, producer);
143
0
    setRand(cctx, ZSTD_c_useBlockSplitter, 0, 2, producer);
144
0
    setRand(cctx, ZSTD_c_deterministicRefPrefix, 0, 1, producer);
145
0
    setRand(cctx, ZSTD_c_prefetchCDictTables, 0, 2, producer);
146
0
    setRand(cctx, ZSTD_c_maxBlockSize, ZSTD_BLOCKSIZE_MAX_MIN, ZSTD_BLOCKSIZE_MAX, producer);
147
0
    setRand(cctx, ZSTD_c_validateSequences, 0, 1, producer);
148
0
    setRand(cctx, ZSTD_c_searchForExternalRepcodes, 0, 2, producer);
149
0
    if (FUZZ_dataProducer_uint32Range(producer, 0, 1) == 0) {
150
0
      setRand(cctx, ZSTD_c_srcSizeHint, ZSTD_SRCSIZEHINT_MIN, 2 * srcSize, producer);
151
0
    }
152
0
    if (FUZZ_dataProducer_uint32Range(producer, 0, 1) == 0) {
153
0
      setRand(cctx, ZSTD_c_targetCBlockSize, ZSTD_TARGETCBLOCKSIZE_MIN, ZSTD_TARGETCBLOCKSIZE_MAX, producer);
154
0
    }
155
156
#ifdef FUZZ_THIRD_PARTY_SEQ_PROD
157
    setSequenceProducerParams(cctx, producer);
158
#else
159
0
    if (FUZZ_dataProducer_uint32Range(producer, 0, 10) == 1) {
160
0
        setSequenceProducerParams(cctx, producer);
161
0
    } else {
162
0
        ZSTD_registerSequenceProducer(cctx, NULL, NULL);
163
0
    }
164
0
#endif
165
0
}
166
167
FUZZ_dict_t FUZZ_train(void const* src, size_t srcSize, FUZZ_dataProducer_t *producer)
168
0
{
169
0
    size_t const dictSize = MAX(srcSize / 8, 1024);
170
0
    size_t const totalSampleSize = dictSize * 11;
171
0
    FUZZ_dict_t dict = { FUZZ_malloc(dictSize), dictSize };
172
0
    char* const samples = (char*)FUZZ_malloc(totalSampleSize);
173
0
    unsigned nbSamples = 100;
174
0
    size_t* const samplesSizes = (size_t*)FUZZ_malloc(sizeof(size_t) * nbSamples);
175
0
    size_t pos = 0;
176
0
    size_t sample = 0;
177
0
    ZDICT_fastCover_params_t params;
178
179
0
    for (sample = 0; sample < nbSamples; ++sample) {
180
0
      size_t const remaining = totalSampleSize - pos;
181
0
      size_t const offset = FUZZ_dataProducer_uint32Range(producer, 0, MAX(srcSize, 1) - 1);
182
0
      size_t const limit = MIN(srcSize - offset, remaining);
183
0
      size_t const toCopy = MIN(limit, remaining / (nbSamples - sample));
184
0
      memcpy(samples + pos, (const char*)src + offset, toCopy);
185
0
      pos += toCopy;
186
0
      samplesSizes[sample] = toCopy;
187
0
    }
188
0
    memset(samples + pos, 0, totalSampleSize - pos);
189
190
0
    memset(&params, 0, sizeof(params));
191
0
    params.accel = 5;
192
0
    params.k = 40;
193
0
    params.d = 8;
194
0
    params.f = 14;
195
0
    params.zParams.compressionLevel = 1;
196
0
    dict.size = ZDICT_trainFromBuffer_fastCover(dict.buff, dictSize,
197
0
        samples, samplesSizes, nbSamples, params);
198
0
    if (ZSTD_isError(dict.size)) {
199
0
        free(dict.buff);
200
0
        memset(&dict, 0, sizeof(dict));
201
0
    }
202
203
0
    free(samplesSizes);
204
0
    free(samples);
205
206
0
    return dict;
207
0
}