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
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
|
From 8973d58b7498fa5100a876815476b81fd1a2412a Mon Sep 17 00:00:00 2001
From: Daniel Veillard <veillard@redhat.com>
Date: Sat, 04 Feb 2012 11:07:44 +0000
Subject: Add hash randomization to hash and dict structures
Following http://www.ocert.org/advisories/ocert-2011-003.html
it seems that having hash randomization might be a good idea
when using XML with untrusted data
* configure.in: lookup for rand, srand and time
* dict.c: add randomization to dictionaries hash tables
* hash.c: add randomization to normal hash tables
---
diff --git a/configure.in b/configure.in
index fa80375..828b66a 100644
--- a/configure.in
+++ b/configure.in
@@ -512,6 +512,7 @@ AC_CHECK_FUNCS(strdup strndup strerror)
AC_CHECK_FUNCS(finite isnand fp_class class fpclass)
AC_CHECK_FUNCS(strftime localtime gettimeofday ftime)
AC_CHECK_FUNCS(stat _stat signal)
+AC_CHECK_FUNCS(rand srand time)
dnl Checking the standard string functions availability
AC_CHECK_FUNCS(printf sprintf fprintf snprintf vfprintf vsprintf vsnprintf sscanf,,
diff --git a/dict.c b/dict.c
index 3eff231..ae4966b 100644
--- a/dict.c
+++ b/dict.c
@@ -2,7 +2,7 @@
* dict.c: dictionary of reusable strings, just used to avoid allocation
* and freeing operations.
*
- * Copyright (C) 2003 Daniel Veillard.
+ * Copyright (C) 2003-2012 Daniel Veillard.
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
@@ -19,6 +19,28 @@
#define IN_LIBXML
#include "libxml.h"
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_TIME_H
+#include <time.h>
+#endif
+
+/*
+ * Following http://www.ocert.org/advisories/ocert-2011-003.html
+ * it seems that having hash randomization might be a good idea
+ * when using XML with untrusted data
+ * Note1: that it works correctly only if compiled with WITH_BIG_KEY
+ * which is the default.
+ * Note2: the fast function used for a small dict won't protect very
+ * well but since the attack is based on growing a very big hash
+ * list we will use the BigKey algo as soon as the hash size grows
+ * over MIN_DICT_SIZE so this actually works
+ */
+#if defined(HAVE_RAND) && defined(HAVE_SRAND) && defined(HAVE_TIME)
+#define DICT_RANDOMIZATION
+#endif
+
#include <string.h>
#ifdef HAVE_STDINT_H
#include <stdint.h>
@@ -44,23 +66,23 @@ typedef unsigned __int32 uint32_t;
#define WITH_BIG_KEY
#ifdef WITH_BIG_KEY
-#define xmlDictComputeKey(dict, name, len) \
- (((dict)->size == MIN_DICT_SIZE) ? \
- xmlDictComputeFastKey(name, len) : \
- xmlDictComputeBigKey(name, len))
-
-#define xmlDictComputeQKey(dict, prefix, plen, name, len) \
- (((prefix) == NULL) ? \
- (xmlDictComputeKey(dict, name, len)) : \
- (((dict)->size == MIN_DICT_SIZE) ? \
- xmlDictComputeFastQKey(prefix, plen, name, len) : \
- xmlDictComputeBigQKey(prefix, plen, name, len)))
+#define xmlDictComputeKey(dict, name, len) \
+ (((dict)->size == MIN_DICT_SIZE) ? \
+ xmlDictComputeFastKey(name, len, (dict)->seed) : \
+ xmlDictComputeBigKey(name, len, (dict)->seed))
+
+#define xmlDictComputeQKey(dict, prefix, plen, name, len) \
+ (((prefix) == NULL) ? \
+ (xmlDictComputeKey(dict, name, len)) : \
+ (((dict)->size == MIN_DICT_SIZE) ? \
+ xmlDictComputeFastQKey(prefix, plen, name, len, (dict)->seed) : \
+ xmlDictComputeBigQKey(prefix, plen, name, len, (dict)->seed)))
#else /* !WITH_BIG_KEY */
-#define xmlDictComputeKey(dict, name, len) \
- xmlDictComputeFastKey(name, len)
-#define xmlDictComputeQKey(dict, prefix, plen, name, len) \
- xmlDictComputeFastQKey(prefix, plen, name, len)
+#define xmlDictComputeKey(dict, name, len) \
+ xmlDictComputeFastKey(name, len, (dict)->seed)
+#define xmlDictComputeQKey(dict, prefix, plen, name, len) \
+ xmlDictComputeFastQKey(prefix, plen, name, len, (dict)->seed)
#endif /* WITH_BIG_KEY */
/*
@@ -98,6 +120,8 @@ struct _xmlDict {
xmlDictStringsPtr strings;
struct _xmlDict *subdict;
+ /* used for randomization */
+ int seed;
};
/*
@@ -125,6 +149,9 @@ static int xmlInitializeDict(void) {
if ((xmlDictMutex = xmlNewRMutex()) == NULL)
return(0);
+#ifdef DICT_RANDOMIZATION
+ srand(time(NULL));
+#endif
xmlDictInitialized = 1;
return(1);
}
@@ -277,13 +304,13 @@ found_pool:
*/
static uint32_t
-xmlDictComputeBigKey(const xmlChar* data, int namelen) {
+xmlDictComputeBigKey(const xmlChar* data, int namelen, int seed) {
uint32_t hash;
int i;
if (namelen <= 0 || data == NULL) return(0);
- hash = 0;
+ hash = seed;
for (i = 0;i < namelen; i++) {
hash += data[i];
@@ -310,12 +337,12 @@ xmlDictComputeBigKey(const xmlChar* data, int namelen) {
*/
static unsigned long
xmlDictComputeBigQKey(const xmlChar *prefix, int plen,
- const xmlChar *name, int len)
+ const xmlChar *name, int len, int seed)
{
uint32_t hash;
int i;
- hash = 0;
+ hash = seed;
for (i = 0;i < plen; i++) {
hash += prefix[i];
@@ -346,8 +373,8 @@ xmlDictComputeBigQKey(const xmlChar *prefix, int plen,
* for low hash table fill.
*/
static unsigned long
-xmlDictComputeFastKey(const xmlChar *name, int namelen) {
- unsigned long value = 0L;
+xmlDictComputeFastKey(const xmlChar *name, int namelen, int seed) {
+ unsigned long value = seed;
if (name == NULL) return(0);
value = *name;
@@ -381,9 +408,9 @@ xmlDictComputeFastKey(const xmlChar *name, int namelen) {
*/
static unsigned long
xmlDictComputeFastQKey(const xmlChar *prefix, int plen,
- const xmlChar *name, int len)
+ const xmlChar *name, int len, int seed)
{
- unsigned long value = 0L;
+ unsigned long value = (unsigned long) seed;
if (plen == 0)
value += 30 * (unsigned long) ':';
@@ -460,6 +487,11 @@ xmlDictCreate(void) {
dict->subdict = NULL;
if (dict->dict) {
memset(dict->dict, 0, MIN_DICT_SIZE * sizeof(xmlDictEntry));
+#ifdef DICT_RANDOMIZATION
+ dict->seed = rand();
+#else
+ dict->seed = 0;
+#endif
return(dict);
}
xmlFree(dict);
@@ -486,6 +518,7 @@ xmlDictCreateSub(xmlDictPtr sub) {
#ifdef DICT_DEBUG_PATTERNS
fprintf(stderr, "R");
#endif
+ dict->seed = sub->seed;
dict->subdict = sub;
xmlDictReference(dict->subdict);
}
diff --git a/hash.c b/hash.c
index b78bc2d..fe1424f 100644
--- a/hash.c
+++ b/hash.c
@@ -3,7 +3,7 @@
*
* Reference: Your favorite introductory book on algorithms
*
- * Copyright (C) 2000 Bjorn Reese and Daniel Veillard.
+ * Copyright (C) 2000,2012 Bjorn Reese and Daniel Veillard.
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
@@ -21,6 +21,22 @@
#include "libxml.h"
#include <string.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_TIME_H
+#include <time.h>
+#endif
+
+/*
+ * Following http://www.ocert.org/advisories/ocert-2011-003.html
+ * it seems that having hash randomization might be a good idea
+ * when using XML with untrusted data
+ */
+#if defined(HAVE_RAND) && defined(HAVE_SRAND) && defined(HAVE_TIME)
+#define HASH_RANDOMIZATION
+#endif
+
#include <libxml/parser.h>
#include <libxml/hash.h>
#include <libxml/xmlmemory.h>
@@ -31,6 +47,10 @@
/* #define DEBUG_GROW */
+#ifdef HASH_RANDOMIZATION
+static int hash_initialized = 0;
+#endif
+
/*
* A single entry in the hash table
*/
@@ -53,6 +73,9 @@ struct _xmlHashTable {
int size;
int nbElems;
xmlDictPtr dict;
+#ifdef HASH_RANDOMIZATION
+ int random_seed;
+#endif
};
/*
@@ -65,6 +88,9 @@ xmlHashComputeKey(xmlHashTablePtr table, const xmlChar *name,
unsigned long value = 0L;
char ch;
+#ifdef HASH_RANDOMIZATION
+ value = table->random_seed;
+#endif
if (name != NULL) {
value += 30 * (*name);
while ((ch = *name++) != 0) {
@@ -92,6 +118,9 @@ xmlHashComputeQKey(xmlHashTablePtr table,
unsigned long value = 0L;
char ch;
+#ifdef HASH_RANDOMIZATION
+ value = table->random_seed;
+#endif
if (prefix != NULL)
value += 30 * (*prefix);
else
@@ -156,6 +185,13 @@ xmlHashCreate(int size) {
table->table = xmlMalloc(size * sizeof(xmlHashEntry));
if (table->table) {
memset(table->table, 0, size * sizeof(xmlHashEntry));
+#ifdef HASH_RANDOMIZATION
+ if (!hash_initialized) {
+ srand(time(NULL));
+ hash_initialized = 1;
+ }
+ table->random_seed = rand();
+#endif
return(table);
}
xmlFree(table);
--
cgit v0.9.0.2
|
