1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
|
/*
* html.c: a libFuzzer target to test several HTML parser interfaces.
*
* See Copyright for the status of this software.
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <libxml/HTMLparser.h>
#include <libxml/HTMLtree.h>
#include <libxml/catalog.h>
#include "fuzz.h"
int
LLVMFuzzerInitialize(int *argc ATTRIBUTE_UNUSED,
char ***argv ATTRIBUTE_UNUSED) {
xmlFuzzMemSetup();
xmlInitParser();
#ifdef LIBXML_CATALOG_ENABLED
xmlInitializeCatalog();
xmlCatalogSetDefaults(XML_CATA_ALLOW_NONE);
#endif
return 0;
}
int
LLVMFuzzerTestOneInput(const char *data, size_t size) {
xmlParserCtxtPtr ctxt;
htmlDocPtr doc;
const char *docBuffer;
size_t failurePos, docSize, maxChunkSize;
int opts, errorCode;
#ifdef LIBXML_OUTPUT_ENABLED
xmlOutputBufferPtr out = NULL;
#endif
xmlFuzzDataInit(data, size);
opts = (int) xmlFuzzReadInt(4);
failurePos = xmlFuzzReadInt(4) % (size + 100);
maxChunkSize = xmlFuzzReadInt(4) % (size + size / 8 + 1);
if (maxChunkSize == 0)
maxChunkSize = 1;
docBuffer = xmlFuzzReadRemaining(&docSize);
if (docBuffer == NULL) {
xmlFuzzDataCleanup();
return(0);
}
/* Pull parser */
xmlFuzzInjectFailure(failurePos);
ctxt = htmlNewParserCtxt();
if (ctxt == NULL) {
errorCode = XML_ERR_NO_MEMORY;
} else {
xmlCtxtSetErrorHandler(ctxt, xmlFuzzSErrorFunc, NULL);
doc = htmlCtxtReadMemory(ctxt, docBuffer, docSize, NULL, NULL, opts);
errorCode = ctxt->errNo;
xmlFuzzCheckFailureReport("htmlCtxtReadMemory",
errorCode == XML_ERR_NO_MEMORY,
errorCode == XML_IO_EIO);
if (doc != NULL) {
xmlDocPtr copy;
#ifdef LIBXML_OUTPUT_ENABLED
const xmlChar *content;
/*
* Also test the serializer. Call htmlDocContentDumpOutput with our
* own buffer to avoid encoding the output. The HTML encoding is
* excruciatingly slow (see htmlEntityValueLookup).
*/
out = xmlAllocOutputBuffer(NULL);
htmlDocContentDumpOutput(out, doc, NULL);
content = xmlOutputBufferGetContent(out);
xmlFuzzCheckFailureReport("htmlDocContentDumpOutput",
content == NULL, 0);
if (content == NULL) {
xmlOutputBufferClose(out);
out = NULL;
}
#endif
copy = xmlCopyDoc(doc, 1);
xmlFuzzCheckFailureReport("xmlCopyNode", copy == NULL, 0);
xmlFreeDoc(copy);
xmlFreeDoc(doc);
}
htmlFreeParserCtxt(ctxt);
}
/* Push parser */
#ifdef LIBXML_PUSH_ENABLED
xmlFuzzInjectFailure(failurePos);
ctxt = htmlCreatePushParserCtxt(NULL, NULL, NULL, 0, NULL,
XML_CHAR_ENCODING_NONE);
if (ctxt != NULL) {
size_t consumed;
int errorCodePush, numChunks, maxChunks;
xmlCtxtSetErrorHandler(ctxt, xmlFuzzSErrorFunc, NULL);
htmlCtxtUseOptions(ctxt, opts);
consumed = 0;
numChunks = 0;
maxChunks = 50 + docSize / 100;
while (numChunks == 0 ||
(consumed < docSize && numChunks < maxChunks)) {
size_t chunkSize;
int terminate;
numChunks += 1;
chunkSize = docSize - consumed;
if (numChunks < maxChunks && chunkSize > maxChunkSize) {
chunkSize = maxChunkSize;
terminate = 0;
} else {
terminate = 1;
}
htmlParseChunk(ctxt, docBuffer + consumed, chunkSize, terminate);
consumed += chunkSize;
}
errorCodePush = ctxt->errNo;
xmlFuzzCheckFailureReport("htmlParseChunk",
errorCodePush == XML_ERR_NO_MEMORY,
errorCodePush == XML_IO_EIO);
doc = ctxt->myDoc;
/*
* Push and pull parser differ in when exactly they
* stop parsing, and the error code is the *last* error
* reported, so we can't check whether the codes match.
*/
if (errorCode != XML_ERR_NO_MEMORY &&
errorCode != XML_IO_EIO &&
errorCodePush != XML_ERR_NO_MEMORY &&
errorCodePush != XML_IO_EIO &&
(errorCode == XML_ERR_OK) != (errorCodePush == XML_ERR_OK)) {
fprintf(stderr, "pull/push parser error mismatch: %d != %d\n",
errorCode, errorCodePush);
#if 0
FILE *f = fopen("c.html", "wb");
fwrite(docBuffer, docSize, 1, f);
fclose(f);
fprintf(stderr, "opts: %X\n", opts);
#endif
abort();
}
#ifdef LIBXML_OUTPUT_ENABLED
/*
* Verify that pull and push parser produce the same result.
*
* The NOBLANKS option doesn't work reliably in push mode.
*/
if ((opts & XML_PARSE_NOBLANKS) == 0 &&
errorCode == XML_ERR_OK &&
errorCodePush == XML_ERR_OK &&
out != NULL) {
xmlOutputBufferPtr outPush;
const xmlChar *content, *contentPush;
outPush = xmlAllocOutputBuffer(NULL);
htmlDocContentDumpOutput(outPush, doc, NULL);
content = xmlOutputBufferGetContent(out);
contentPush = xmlOutputBufferGetContent(outPush);
if (content != NULL && contentPush != NULL) {
size_t outSize = xmlOutputBufferGetSize(out);
if (outSize != xmlOutputBufferGetSize(outPush) ||
memcmp(content, contentPush, outSize) != 0) {
fprintf(stderr, "pull/push parser roundtrip "
"mismatch\n");
#if 0
FILE *f = fopen("c.html", "wb");
fwrite(docBuffer, docSize, 1, f);
fclose(f);
fprintf(stderr, "opts: %X\n", opts);
fprintf(stderr, "---\n%s\n---\n%s\n---\n",
xmlOutputBufferGetContent(out),
xmlOutputBufferGetContent(outPush));
#endif
abort();
}
}
xmlOutputBufferClose(outPush);
}
#endif
xmlFreeDoc(doc);
htmlFreeParserCtxt(ctxt);
}
#endif
/* Cleanup */
#ifdef LIBXML_OUTPUT_ENABLED
xmlOutputBufferClose(out);
#endif
xmlFuzzInjectFailure(0);
xmlFuzzDataCleanup();
xmlResetLastError();
return(0);
}
size_t
LLVMFuzzerCustomMutator(char *data, size_t size, size_t maxSize,
unsigned seed) {
static const xmlFuzzChunkDesc chunks[] = {
{ 4, XML_FUZZ_PROB_ONE / 10 }, /* opts */
{ 4, XML_FUZZ_PROB_ONE / 10 }, /* failurePos */
{ 0, 0 }
};
return xmlFuzzMutateChunks(chunks, data, size, maxSize, seed,
LLVMFuzzerMutate);
}
|
