import expat-2.2.5-9.el8

This commit is contained in:
CentOS Sources 2022-09-27 16:57:51 -04:00 committed by Stepan Oksanichenko
parent af98ea3183
commit 160a41e29f
6 changed files with 740 additions and 1 deletions

View File

@ -0,0 +1,200 @@
commit e8f285b522a907603501329e5b4212755f525fdf
Author: Tomas Korbar <tkorbar@redhat.com>
Date: Thu Mar 3 12:04:09 2022 +0100
CVE-2022-25235
diff --git a/lib/xmltok.c b/lib/xmltok.c
index 6b415d8..b55732a 100644
--- a/lib/xmltok.c
+++ b/lib/xmltok.c
@@ -103,13 +103,6 @@
+ ((((byte)[2]) >> 5) & 1)] \
& (1u << (((byte)[2]) & 0x1F)))
-#define UTF8_GET_NAMING(pages, p, n) \
- ((n) == 2 \
- ? UTF8_GET_NAMING2(pages, (const unsigned char *)(p)) \
- : ((n) == 3 \
- ? UTF8_GET_NAMING3(pages, (const unsigned char *)(p)) \
- : 0))
-
/* Detection of invalid UTF-8 sequences is based on Table 3.1B
of Unicode 3.2: http://www.unicode.org/unicode/reports/tr28/
with the additional restriction of not allowing the Unicode
diff --git a/lib/xmltok_impl.c b/lib/xmltok_impl.c
index 0403dd3..56d7a40 100644
--- a/lib/xmltok_impl.c
+++ b/lib/xmltok_impl.c
@@ -61,7 +61,7 @@
case BT_LEAD ## n: \
if (end - ptr < n) \
return XML_TOK_PARTIAL_CHAR; \
- if (!IS_NAME_CHAR(enc, ptr, n)) { \
+ if (IS_INVALID_CHAR(enc, ptr, n) || ! IS_NAME_CHAR(enc, ptr, n)) { \
*nextTokPtr = ptr; \
return XML_TOK_INVALID; \
} \
@@ -89,7 +89,7 @@
case BT_LEAD ## n: \
if (end - ptr < n) \
return XML_TOK_PARTIAL_CHAR; \
- if (!IS_NMSTRT_CHAR(enc, ptr, n)) { \
+ if (IS_INVALID_CHAR(enc, ptr, n) || ! IS_NMSTRT_CHAR(enc, ptr, n)) { \
*nextTokPtr = ptr; \
return XML_TOK_INVALID; \
} \
@@ -1117,6 +1117,10 @@ PREFIX(prologTok)(const ENCODING *enc, const char *ptr, const char *end,
case BT_LEAD ## n: \
if (end - ptr < n) \
return XML_TOK_PARTIAL_CHAR; \
+ if (IS_INVALID_CHAR(enc, ptr, n)) { \
+ *nextTokPtr = ptr; \
+ return XML_TOK_INVALID; \
+ } \
if (IS_NMSTRT_CHAR(enc, ptr, n)) { \
ptr += n; \
tok = XML_TOK_NAME; \
diff --git a/tests/runtests.c b/tests/runtests.c
index 278bfa1..0f3afde 100644
--- a/tests/runtests.c
+++ b/tests/runtests.c
@@ -6540,6 +6540,106 @@ START_TEST(test_utf8_in_cdata_section_2)
}
END_TEST
+START_TEST(test_utf8_in_start_tags) {
+ struct test_case {
+ bool goodName;
+ bool goodNameStart;
+ const char *tagName;
+ };
+
+ // The idea with the tests below is this:
+ // We want to cover 1-, 2- and 3-byte sequences, 4-byte sequences
+ // go to isNever and are hence not a concern.
+ //
+ // We start with a character that is a valid name character
+ // (or even name-start character, see XML 1.0r4 spec) and then we flip
+ // single bits at places where (1) the result leaves the UTF-8 encoding space
+ // and (2) we stay in the same n-byte sequence family.
+ //
+ // The flipped bits are highlighted in angle brackets in comments,
+ // e.g. "[<1>011 1001]" means we had [0011 1001] but we now flipped
+ // the most significant bit to 1 to leave UTF-8 encoding space.
+ struct test_case cases[] = {
+ // 1-byte UTF-8: [0xxx xxxx]
+ {true, true, "\x3A"}, // [0011 1010] = ASCII colon ':'
+ {false, false, "\xBA"}, // [<1>011 1010]
+ {true, false, "\x39"}, // [0011 1001] = ASCII nine '9'
+ {false, false, "\xB9"}, // [<1>011 1001]
+
+ // 2-byte UTF-8: [110x xxxx] [10xx xxxx]
+ {true, true, "\xDB\xA5"}, // [1101 1011] [1010 0101] =
+ // Arabic small waw U+06E5
+ {false, false, "\x9B\xA5"}, // [1<0>01 1011] [1010 0101]
+ {false, false, "\xDB\x25"}, // [1101 1011] [<0>010 0101]
+ {false, false, "\xDB\xE5"}, // [1101 1011] [1<1>10 0101]
+ {true, false, "\xCC\x81"}, // [1100 1100] [1000 0001] =
+ // combining char U+0301
+ {false, false, "\x8C\x81"}, // [1<0>00 1100] [1000 0001]
+ {false, false, "\xCC\x01"}, // [1100 1100] [<0>000 0001]
+ {false, false, "\xCC\xC1"}, // [1100 1100] [1<1>00 0001]
+
+ // 3-byte UTF-8: [1110 xxxx] [10xx xxxx] [10xxxxxx]
+ {true, true, "\xE0\xA4\x85"}, // [1110 0000] [1010 0100] [1000 0101] =
+ // Devanagari Letter A U+0905
+ {false, false, "\xA0\xA4\x85"}, // [1<0>10 0000] [1010 0100] [1000 0101]
+ {false, false, "\xE0\x24\x85"}, // [1110 0000] [<0>010 0100] [1000 0101]
+ {false, false, "\xE0\xE4\x85"}, // [1110 0000] [1<1>10 0100] [1000 0101]
+ {false, false, "\xE0\xA4\x05"}, // [1110 0000] [1010 0100] [<0>000 0101]
+ {false, false, "\xE0\xA4\xC5"}, // [1110 0000] [1010 0100] [1<1>00 0101]
+ {true, false, "\xE0\xA4\x81"}, // [1110 0000] [1010 0100] [1000 0001] =
+ // combining char U+0901
+ {false, false, "\xA0\xA4\x81"}, // [1<0>10 0000] [1010 0100] [1000 0001]
+ {false, false, "\xE0\x24\x81"}, // [1110 0000] [<0>010 0100] [1000 0001]
+ {false, false, "\xE0\xE4\x81"}, // [1110 0000] [1<1>10 0100] [1000 0001]
+ {false, false, "\xE0\xA4\x01"}, // [1110 0000] [1010 0100] [<0>000 0001]
+ {false, false, "\xE0\xA4\xC1"}, // [1110 0000] [1010 0100] [1<1>00 0001]
+ };
+ const bool atNameStart[] = {true, false};
+
+ size_t i = 0;
+ char doc[1024];
+ size_t failCount = 0;
+
+ for (; i < sizeof(cases) / sizeof(cases[0]); i++) {
+ size_t j = 0;
+ for (; j < sizeof(atNameStart) / sizeof(atNameStart[0]); j++) {
+ const bool expectedSuccess
+ = atNameStart[j] ? cases[i].goodNameStart : cases[i].goodName;
+ sprintf(doc, "<%s%s><!--", atNameStart[j] ? "" : "a", cases[i].tagName);
+ XML_Parser parser = XML_ParserCreate(NULL);
+
+ const enum XML_Status status
+ = XML_Parse(parser, doc, (int)strlen(doc), /*isFinal=*/XML_FALSE);
+
+ bool success = true;
+ if ((status == XML_STATUS_OK) != expectedSuccess) {
+ success = false;
+ }
+ if ((status == XML_STATUS_ERROR)
+ && (XML_GetErrorCode(parser) != XML_ERROR_INVALID_TOKEN)) {
+ success = false;
+ }
+
+ if (! success) {
+ fprintf(
+ stderr,
+ "FAIL case %2u (%sat name start, %u-byte sequence, error code %d)\n",
+ (unsigned)i + 1u, atNameStart[j] ? " " : "not ",
+ (unsigned)strlen(cases[i].tagName), XML_GetErrorCode(parser));
+ failCount++;
+ }
+
+ XML_ParserFree(parser);
+ }
+ }
+
+ if (failCount > 0) {
+ fail("UTF-8 regression detected");
+ }
+}
+END_TEST
+
+
/* Test trailing spaces in elements are accepted */
static void XMLCALL
record_element_end_handler(void *userData,
@@ -6734,6 +6834,15 @@ START_TEST(test_bad_doctype)
}
END_TEST
+START_TEST(test_bad_doctype_utf8) {
+ const char *text = "<!DOCTYPE \xDB\x25"
+ "doc><doc/>"; // [1101 1011] [<0>010 0101]
+ expect_failure(text, XML_ERROR_INVALID_TOKEN,
+ "Invalid UTF-8 in DOCTYPE not faulted");
+}
+END_TEST
+
+
START_TEST(test_bad_doctype_utf16)
{
const char text[] =
@@ -12256,6 +12365,7 @@ make_suite(void)
tcase_add_test(tc_basic, test_ext_entity_utf8_non_bom);
tcase_add_test(tc_basic, test_utf8_in_cdata_section);
tcase_add_test(tc_basic, test_utf8_in_cdata_section_2);
+ tcase_add_test(tc_basic, test_utf8_in_start_tags);
tcase_add_test(tc_basic, test_trailing_spaces_in_elements);
tcase_add_test(tc_basic, test_utf16_attribute);
tcase_add_test(tc_basic, test_utf16_second_attr);
@@ -12264,6 +12374,7 @@ make_suite(void)
tcase_add_test(tc_basic, test_bad_attr_desc_keyword);
tcase_add_test(tc_basic, test_bad_attr_desc_keyword_utf16);
tcase_add_test(tc_basic, test_bad_doctype);
+ tcase_add_test(tc_basic, test_bad_doctype_utf8);
tcase_add_test(tc_basic, test_bad_doctype_utf16);
tcase_add_test(tc_basic, test_bad_doctype_plus);
tcase_add_test(tc_basic, test_bad_doctype_star);

View File

@ -0,0 +1,19 @@
commit e5b609876e5a266725fba1c377b0ac95c737e6ed
Author: Tomas Korbar <tkorbar@redhat.com>
Date: Mon May 2 12:44:06 2022 +0200
Fix CVE-2022-25314
diff --git a/lib/xmlparse.c b/lib/xmlparse.c
index 1f1413f..ceeec26 100644
--- a/lib/xmlparse.c
+++ b/lib/xmlparse.c
@@ -7525,7 +7525,7 @@ static XML_Char *
copyString(const XML_Char *s,
const XML_Memory_Handling_Suite *memsuite)
{
- int charsRequired = 0;
+ size_t charsRequired = 0;
XML_Char *result;
/* First determine how long the string is */

View File

@ -0,0 +1,31 @@
commit 3a4141add108097fa548b196f5950c6663e1578e
Author: Tomas Korbar <tkorbar@redhat.com>
Date: Thu Mar 3 13:50:20 2022 +0100
CVE-2022-25315
diff --git a/lib/xmlparse.c b/lib/xmlparse.c
index f0061c8..45fda00 100644
--- a/lib/xmlparse.c
+++ b/lib/xmlparse.c
@@ -2508,6 +2508,7 @@ storeRawNames(XML_Parser parser)
while (tag) {
int bufSize;
int nameLen = sizeof(XML_Char) * (tag->name.strLen + 1);
+ size_t rawNameLen;
char *rawNameBuf = tag->buf + nameLen;
/* Stop if already stored. Since m_tagStack is a stack, we can stop
at the first entry that has already been copied; everything
@@ -2519,7 +2520,11 @@ storeRawNames(XML_Parser parser)
/* For re-use purposes we need to ensure that the
size of tag->buf is a multiple of sizeof(XML_Char).
*/
- bufSize = nameLen + ROUND_UP(tag->rawNameLength, sizeof(XML_Char));
+ rawNameLen = ROUND_UP(tag->rawNameLength, sizeof(XML_Char));
+ /* Detect and prevent integer overflow. */
+ if (rawNameLen > (size_t)INT_MAX - nameLen)
+ return XML_FALSE;
+ bufSize = nameLen + (int)rawNameLen;
if (bufSize > tag->bufEnd - tag->buf) {
char *temp = (char *)REALLOC(parser, tag->buf, bufSize);
if (temp == NULL)

View File

@ -0,0 +1,228 @@
commit f1b61e6fbaedbb2bbea736269a015d97d4df46ce
Author: Tomas Korbar <tkorbar@redhat.com>
Date: Tue May 3 13:42:54 2022 +0200
Fix CVE-2022-25313
diff --git a/lib/xmlparse.c b/lib/xmlparse.c
index ceeec26..d47e42c 100644
--- a/lib/xmlparse.c
+++ b/lib/xmlparse.c
@@ -7458,12 +7458,14 @@ build_node(XML_Parser parser,
}
static XML_Content *
-build_model (XML_Parser parser)
-{
- DTD * const dtd = parser->m_dtd; /* save one level of indirection */
+build_model(XML_Parser parser) {
+ /* Function build_model transforms the existing parser->m_dtd->scaffold
+ * array of CONTENT_SCAFFOLD tree nodes into a new array of
+ * XML_Content tree nodes followed by a gapless list of zero-terminated
+ * strings. */
+ DTD *const dtd = parser->m_dtd; /* save one level of indirection */
XML_Content *ret;
- XML_Content *cpos;
- XML_Char * str;
+ XML_Char *str; /* the current string writing location */
/* Detect and prevent integer overflow.
* The preprocessor guard addresses the "always false" warning
@@ -7486,13 +7488,99 @@ build_model (XML_Parser parser)
+ (dtd->contentStringLen * sizeof(XML_Char)));
ret = (XML_Content *)MALLOC(parser, allocsize);
- if (!ret)
+ if (! ret)
return NULL;
- str = (XML_Char *) (&ret[dtd->scaffCount]);
- cpos = &ret[1];
+ /* What follows is an iterative implementation (of what was previously done
+ * recursively in a dedicated function called "build_node". The old recursive
+ * build_node could be forced into stack exhaustion from input as small as a
+ * few megabyte, and so that was a security issue. Hence, a function call
+ * stack is avoided now by resolving recursion.)
+ *
+ * The iterative approach works as follows:
+ *
+ * - We have two writing pointers, both walking up the result array; one does
+ * the work, the other creates "jobs" for its colleague to do, and leads
+ * the way:
+ *
+ * - The faster one, pointer jobDest, always leads and writes "what job
+ * to do" by the other, once they reach that place in the
+ * array: leader "jobDest" stores the source node array index (relative
+ * to array dtd->scaffold) in field "numchildren".
+ *
+ * - The slower one, pointer dest, looks at the value stored in the
+ * "numchildren" field (which actually holds a source node array index
+ * at that time) and puts the real data from dtd->scaffold in.
+ *
+ * - Before the loop starts, jobDest writes source array index 0
+ * (where the root node is located) so that dest will have something to do
+ * when it starts operation.
+ *
+ * - Whenever nodes with children are encountered, jobDest appends
+ * them as new jobs, in order. As a result, tree node siblings are
+ * adjacent in the resulting array, for example:
+ *
+ * [0] root, has two children
+ * [1] first child of 0, has three children
+ * [3] first child of 1, does not have children
+ * [4] second child of 1, does not have children
+ * [5] third child of 1, does not have children
+ * [2] second child of 0, does not have children
+ *
+ * Or (the same data) presented in flat array view:
+ *
+ * [0] root, has two children
+ *
+ * [1] first child of 0, has three children
+ * [2] second child of 0, does not have children
+ *
+ * [3] first child of 1, does not have children
+ * [4] second child of 1, does not have children
+ * [5] third child of 1, does not have children
+ *
+ * - The algorithm repeats until all target array indices have been processed.
+ */
+ XML_Content *dest = ret; /* tree node writing location, moves upwards */
+ XML_Content *const destLimit = &ret[dtd->scaffCount];
+ XML_Content *jobDest = ret; /* next free writing location in target array */
+ str = (XML_Char *)&ret[dtd->scaffCount];
+
+ /* Add the starting job, the root node (index 0) of the source tree */
+ (jobDest++)->numchildren = 0;
+
+ for (; dest < destLimit; dest++) {
+ /* Retrieve source tree array index from job storage */
+ const int src_node = (int)dest->numchildren;
+
+ /* Convert item */
+ dest->type = dtd->scaffold[src_node].type;
+ dest->quant = dtd->scaffold[src_node].quant;
+ if (dest->type == XML_CTYPE_NAME) {
+ const XML_Char *src;
+ dest->name = str;
+ src = dtd->scaffold[src_node].name;
+ for (;;) {
+ *str++ = *src;
+ if (! *src)
+ break;
+ src++;
+ }
+ dest->numchildren = 0;
+ dest->children = NULL;
+ } else {
+ unsigned int i;
+ int cn;
+ dest->name = NULL;
+ dest->numchildren = dtd->scaffold[src_node].childcnt;
+ dest->children = jobDest;
+
+ /* Append scaffold indices of children to array */
+ for (i = 0, cn = dtd->scaffold[src_node].firstchild;
+ i < dest->numchildren; i++, cn = dtd->scaffold[cn].nextsib)
+ (jobDest++)->numchildren = (unsigned int)cn;
+ }
+ }
- build_node(parser, 0, ret, &cpos, &str);
return ret;
}
diff --git a/tests/runtests.c b/tests/runtests.c
index eacd163..569ad8c 100644
--- a/tests/runtests.c
+++ b/tests/runtests.c
@@ -2848,6 +2848,81 @@ START_TEST(test_dtd_elements)
}
END_TEST
+static void XMLCALL
+element_decl_check_model(void *UNUSED_P(userData), const XML_Char *name,
+ XML_Content *model) {
+ uint32_t errorFlags = 0;
+
+ /* Expected model array structure is this:
+ * [0] (type 6, quant 0)
+ * [1] (type 5, quant 0)
+ * [3] (type 4, quant 0, name "bar")
+ * [4] (type 4, quant 0, name "foo")
+ * [5] (type 4, quant 3, name "xyz")
+ * [2] (type 4, quant 2, name "zebra")
+ */
+ errorFlags |= ((xcstrcmp(name, XCS("junk")) == 0) ? 0 : (1u << 0));
+ errorFlags |= ((model != NULL) ? 0 : (1u << 1));
+
+ errorFlags |= ((model[0].type == XML_CTYPE_SEQ) ? 0 : (1u << 2));
+ errorFlags |= ((model[0].quant == XML_CQUANT_NONE) ? 0 : (1u << 3));
+ errorFlags |= ((model[0].numchildren == 2) ? 0 : (1u << 4));
+ errorFlags |= ((model[0].children == &model[1]) ? 0 : (1u << 5));
+ errorFlags |= ((model[0].name == NULL) ? 0 : (1u << 6));
+
+ errorFlags |= ((model[1].type == XML_CTYPE_CHOICE) ? 0 : (1u << 7));
+ errorFlags |= ((model[1].quant == XML_CQUANT_NONE) ? 0 : (1u << 8));
+ errorFlags |= ((model[1].numchildren == 3) ? 0 : (1u << 9));
+ errorFlags |= ((model[1].children == &model[3]) ? 0 : (1u << 10));
+ errorFlags |= ((model[1].name == NULL) ? 0 : (1u << 11));
+
+ errorFlags |= ((model[2].type == XML_CTYPE_NAME) ? 0 : (1u << 12));
+ errorFlags |= ((model[2].quant == XML_CQUANT_REP) ? 0 : (1u << 13));
+ errorFlags |= ((model[2].numchildren == 0) ? 0 : (1u << 14));
+ errorFlags |= ((model[2].children == NULL) ? 0 : (1u << 15));
+ errorFlags |= ((xcstrcmp(model[2].name, XCS("zebra")) == 0) ? 0 : (1u << 16));
+
+ errorFlags |= ((model[3].type == XML_CTYPE_NAME) ? 0 : (1u << 17));
+ errorFlags |= ((model[3].quant == XML_CQUANT_NONE) ? 0 : (1u << 18));
+ errorFlags |= ((model[3].numchildren == 0) ? 0 : (1u << 19));
+ errorFlags |= ((model[3].children == NULL) ? 0 : (1u << 20));
+ errorFlags |= ((xcstrcmp(model[3].name, XCS("bar")) == 0) ? 0 : (1u << 21));
+
+ errorFlags |= ((model[4].type == XML_CTYPE_NAME) ? 0 : (1u << 22));
+ errorFlags |= ((model[4].quant == XML_CQUANT_NONE) ? 0 : (1u << 23));
+ errorFlags |= ((model[4].numchildren == 0) ? 0 : (1u << 24));
+ errorFlags |= ((model[4].children == NULL) ? 0 : (1u << 25));
+ errorFlags |= ((xcstrcmp(model[4].name, XCS("foo")) == 0) ? 0 : (1u << 26));
+
+ errorFlags |= ((model[5].type == XML_CTYPE_NAME) ? 0 : (1u << 27));
+ errorFlags |= ((model[5].quant == XML_CQUANT_PLUS) ? 0 : (1u << 28));
+ errorFlags |= ((model[5].numchildren == 0) ? 0 : (1u << 29));
+ errorFlags |= ((model[5].children == NULL) ? 0 : (1u << 30));
+ errorFlags |= ((xcstrcmp(model[5].name, XCS("xyz")) == 0) ? 0 : (1u << 31));
+
+ XML_SetUserData(parser, (void *)(uintptr_t)errorFlags);
+ XML_FreeContentModel(parser, model);
+}
+
+START_TEST(test_dtd_elements_nesting) {
+ // Payload inspired by a test in Perl's XML::Parser
+ const char *text = "<!DOCTYPE foo [\n"
+ "<!ELEMENT junk ((bar|foo|xyz+), zebra*)>\n"
+ "]>\n"
+ "<foo/>";
+
+ XML_SetUserData(parser, (void *)(uintptr_t)-1);
+
+ XML_SetElementDeclHandler(parser, element_decl_check_model);
+ if (XML_Parse(parser, text, (int)strlen(text), XML_TRUE)
+ == XML_STATUS_ERROR)
+ xml_failure(parser);
+
+ if ((uint32_t)(uintptr_t)XML_GetUserData(parser) != 0)
+ fail("Element declaration model regression detected");
+}
+END_TEST
+
/* Test foreign DTD handling */
START_TEST(test_set_foreign_dtd)
{
@@ -12256,6 +12331,7 @@ make_suite(void)
tcase_add_test(tc_basic, test_memory_allocation);
tcase_add_test(tc_basic, test_default_current);
tcase_add_test(tc_basic, test_dtd_elements);
+ tcase_add_test(tc_basic, test_dtd_elements_nesting);
tcase_add_test(tc_basic, test_set_foreign_dtd);
tcase_add_test(tc_basic, test_foreign_dtd_not_standalone);
tcase_add_test(tc_basic, test_invalid_foreign_dtd);

View File

@ -0,0 +1,229 @@
commit fd5473ef5873048eadef344a1f16f71ad8eefe99
Author: Tomas Korbar <tkorbar@redhat.com>
Date: Mon Mar 14 12:17:41 2022 +0100
Protect against malicious namespace declarations
diff --git a/lib/xmlparse.c b/lib/xmlparse.c
index 581b9a4..6f3510b 100644
--- a/lib/xmlparse.c
+++ b/lib/xmlparse.c
@@ -661,8 +661,7 @@ XML_ParserCreate(const XML_Char *encodingName)
XML_Parser XMLCALL
XML_ParserCreateNS(const XML_Char *encodingName, XML_Char nsSep)
{
- XML_Char tmp[2];
- *tmp = nsSep;
+ XML_Char tmp[2] = {nsSep, 0};
return XML_ParserCreate_MM(encodingName, NULL, tmp);
}
@@ -1288,8 +1287,7 @@ XML_ExternalEntityParserCreate(XML_Parser oldParser,
would be otherwise.
*/
if (parser->m_ns) {
- XML_Char tmp[2];
- *tmp = parser->m_namespaceSeparator;
+ XML_Char tmp[2] = {parser->m_namespaceSeparator, 0};
parser = parserCreate(encodingName, &parser->m_mem, tmp, newDtd);
}
else {
@@ -3640,6 +3638,117 @@ storeAtts(XML_Parser parser, const ENCODING *enc,
return XML_ERROR_NONE;
}
+static XML_Bool
+is_rfc3986_uri_char(XML_Char candidate) {
+ // For the RFC 3986 ANBF grammar see
+ // https://datatracker.ietf.org/doc/html/rfc3986#appendix-A
+
+ switch (candidate) {
+ // From rule "ALPHA" (uppercase half)
+ case 'A':
+ case 'B':
+ case 'C':
+ case 'D':
+ case 'E':
+ case 'F':
+ case 'G':
+ case 'H':
+ case 'I':
+ case 'J':
+ case 'K':
+ case 'L':
+ case 'M':
+ case 'N':
+ case 'O':
+ case 'P':
+ case 'Q':
+ case 'R':
+ case 'S':
+ case 'T':
+ case 'U':
+ case 'V':
+ case 'W':
+ case 'X':
+ case 'Y':
+ case 'Z':
+
+ // From rule "ALPHA" (lowercase half)
+ case 'a':
+ case 'b':
+ case 'c':
+ case 'd':
+ case 'e':
+ case 'f':
+ case 'g':
+ case 'h':
+ case 'i':
+ case 'j':
+ case 'k':
+ case 'l':
+ case 'm':
+ case 'n':
+ case 'o':
+ case 'p':
+ case 'q':
+ case 'r':
+ case 's':
+ case 't':
+ case 'u':
+ case 'v':
+ case 'w':
+ case 'x':
+ case 'y':
+ case 'z':
+
+ // From rule "DIGIT"
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ case '8':
+ case '9':
+
+ // From rule "pct-encoded"
+ case '%':
+
+ // From rule "unreserved"
+ case '-':
+ case '.':
+ case '_':
+ case '~':
+
+ // From rule "gen-delims"
+ case ':':
+ case '/':
+ case '?':
+ case '#':
+ case '[':
+ case ']':
+ case '@':
+
+ // From rule "sub-delims"
+ case '!':
+ case '$':
+ case '&':
+ case '\'':
+ case '(':
+ case ')':
+ case '*':
+ case '+':
+ case ',':
+ case ';':
+ case '=':
+ return XML_TRUE;
+
+ default:
+ return XML_FALSE;
+ }
+}
+
/* addBinding() overwrites the value of prefix->binding without checking.
Therefore one must keep track of the old value outside of addBinding().
*/
@@ -3700,6 +3809,29 @@ addBinding(XML_Parser parser, PREFIX *prefix, const ATTRIBUTE_ID *attId,
if (!mustBeXML && isXMLNS
&& (len > xmlnsLen || uri[len] != xmlnsNamespace[len]))
isXMLNS = XML_FALSE;
+
+ // NOTE: While Expat does not validate namespace URIs against RFC 3986
+ // today (and is not REQUIRED to do so with regard to the XML 1.0
+ // namespaces specification) we have to at least make sure, that
+ // the application on top of Expat (that is likely splitting expanded
+ // element names ("qualified names") of form
+ // "[uri sep] local [sep prefix] '\0'" back into 1, 2 or 3 pieces
+ // in its element handler code) cannot be confused by an attacker
+ // putting additional namespace separator characters into namespace
+ // declarations. That would be ambiguous and not to be expected.
+ //
+ // While the HTML API docs of function XML_ParserCreateNS have been
+ // advising against use of a namespace separator character that can
+ // appear in a URI for >20 years now, some widespread applications
+ // are using URI characters (':' (colon) in particular) for a
+ // namespace separator, in practice. To keep these applications
+ // functional, we only reject namespaces URIs containing the
+ // application-chosen namespace separator if the chosen separator
+ // is a non-URI character with regard to RFC 3986.
+ if (parser->m_ns && (uri[len] == parser->m_namespaceSeparator)
+ && ! is_rfc3986_uri_char(uri[len])) {
+ return XML_ERROR_SYNTAX;
+ }
}
isXML = isXML && len == xmlLen;
isXMLNS = isXMLNS && len == xmlnsLen;
diff --git a/tests/runtests.c b/tests/runtests.c
index ecc6f47..eabd55d 100644
--- a/tests/runtests.c
+++ b/tests/runtests.c
@@ -7950,6 +7950,38 @@ START_TEST(test_ns_double_colon_doctype)
}
END_TEST
+START_TEST(test_ns_separator_in_uri) {
+ struct test_case {
+ enum XML_Status expectedStatus;
+ const char *doc;
+ XML_Char namesep;
+ };
+ struct test_case cases[] = {
+ {XML_STATUS_OK, "<doc xmlns='one_two' />", XCS('\n')},
+ {XML_STATUS_ERROR, "<doc xmlns='one&#x0A;two' />", XCS('\n')},
+ {XML_STATUS_OK, "<doc xmlns='one:two' />", XCS(':')},
+ };
+
+ size_t i = 0;
+ size_t failCount = 0;
+ for (; i < sizeof(cases) / sizeof(cases[0]); i++) {
+ XML_Parser parser = XML_ParserCreateNS(NULL, cases[i].namesep);
+ XML_SetElementHandler(parser, dummy_start_element, dummy_end_element);
+ if (XML_Parse(parser, cases[i].doc, (int)strlen(cases[i].doc),
+ /*isFinal*/ XML_TRUE)
+ != cases[i].expectedStatus) {
+ failCount++;
+ }
+ XML_ParserFree(parser);
+ }
+
+ if (failCount) {
+ fail("Namespace separator handling is broken");
+ }
+}
+END_TEST
+
+
/* Control variable; the number of times duff_allocator() will successfully allocate */
#define ALLOC_ALWAYS_SUCCEED (-1)
#define REALLOC_ALWAYS_SUCCEED (-1)
@@ -12290,6 +12322,7 @@ make_suite(void)
tcase_add_test(tc_namespace, test_ns_utf16_doctype);
tcase_add_test(tc_namespace, test_ns_invalid_doctype);
tcase_add_test(tc_namespace, test_ns_double_colon_doctype);
+ tcase_add_test(tc_namespace, test_ns_separator_in_uri);
suite_add_tcase(s, tc_misc);
tcase_add_checked_fixture(tc_misc, NULL, basic_teardown);

View File

@ -3,7 +3,7 @@
Summary: An XML parser library Summary: An XML parser library
Name: expat Name: expat
Version: %(echo %{unversion} | sed 's/_/./g') Version: %(echo %{unversion} | sed 's/_/./g')
Release: 5%{?dist} Release: 9%{?dist}
Source: https://github.com/libexpat/libexpat/archive/R_%{unversion}.tar.gz#/expat-%{version}.tar.gz Source: https://github.com/libexpat/libexpat/archive/R_%{unversion}.tar.gz#/expat-%{version}.tar.gz
URL: https://libexpat.github.io/ URL: https://libexpat.github.io/
License: MIT License: MIT
@ -15,6 +15,11 @@ Patch3: expat-2.2.5-Detect-and-prevent-integer-overflow-in-XML_GetBuffer.patch
Patch4: expat-2.2.5-Detect-and-prevent-troublesome-left-shifts.patch Patch4: expat-2.2.5-Detect-and-prevent-troublesome-left-shifts.patch
Patch5: expat-2.2.5-Prevent-integer-overflow-on-m_groupSize-in-function.patch Patch5: expat-2.2.5-Prevent-integer-overflow-on-m_groupSize-in-function.patch
Patch6: expat-2.2.5-Prevent-more-integer-overflows.patch Patch6: expat-2.2.5-Prevent-more-integer-overflows.patch
Patch7: expat-2.2.5-Protect-against-malicious-namespace-declarations.patch
Patch8: expat-2.2.5-Add-missing-validation-of-encoding.patch
Patch9: expat-2.2.5-Prevent-integer-overflow-in-storeRawNames.patch
Patch10: expat-2.2.5-Prevent-integer-overflow-in-copyString.patch
Patch11: expat-2.2.5-Prevent-stack-exhaustion-in-build_model.patch
%description %description
This is expat, the C library for parsing XML, written by James Clark. Expat This is expat, the C library for parsing XML, written by James Clark. Expat
@ -49,6 +54,11 @@ Install it if you need to link statically with expat.
%patch4 -p1 -b .CVE-2021-45960 %patch4 -p1 -b .CVE-2021-45960
%patch5 -p1 -b .CVE-2021-46143 %patch5 -p1 -b .CVE-2021-46143
%patch6 -p1 -b .CVE-2022-22822-CVE-2022-22827 %patch6 -p1 -b .CVE-2022-22822-CVE-2022-22827
%patch7 -p1 -b .CVE-2022-25236
%patch8 -p1 -b .CVE-2022-25235
%patch9 -p1 -b .CVE-2022-25315
%patch10 -p1 -b .CVE-2022-25314
%patch11 -p1 -b .CVE-2022-25313
sed -i 's/install-data-hook/do-nothing-please/' lib/Makefile.am sed -i 's/install-data-hook/do-nothing-please/' lib/Makefile.am
./buildconf.sh ./buildconf.sh
@ -87,6 +97,28 @@ make check
%{_libdir}/lib*.a %{_libdir}/lib*.a
%changelog %changelog
* Fri May 06 2022 Tomas Korbar <tkorbar@redhat.com> - 2.2.5-9
- Fix multiple CVEs
- Resolves: CVE-2022-25314
- Resolves: CVE-2022-25313
* Mon Mar 14 2022 Tomas Korbar <tkorbar@redhat.com> - 2.2.5-8
- Improve patch for CVE-2022-25236
- Related: CVE-2022-25236
* Fri Mar 04 2022 Tomas Korbar <tkorbar@redhat.com> - 2.2.5-7
- Fix patch for CVE-2022-25235
- Resolves: CVE-2022-25235
* Thu Mar 03 2022 Tomas Korbar <tkorbar@redhat.com> - 2.2.5-6
- Fix multiple CVEs
- CVE-2022-25236 expat: namespace-separator characters in "xmlns[:prefix]" attribute values can lead to arbitrary code execution
- CVE-2022-25235 expat: malformed 2- and 3-byte UTF-8 sequences can lead to arbitrary code execution
- CVE-2022-25315 expat: integer overflow in storeRawNames()
- Resolves: CVE-2022-25236
- Resolves: CVE-2022-25235
- Resolves: CVE-2022-25315
* Fri Feb 14 2022 Tomas Korbar <tkorbar@redhat.com> - 2.2.5-5 * Fri Feb 14 2022 Tomas Korbar <tkorbar@redhat.com> - 2.2.5-5
- Fix multiple CVEs - Fix multiple CVEs
- CVE-2022-23852 expat: integer overflow in function XML_GetBuffer - CVE-2022-23852 expat: integer overflow in function XML_GetBuffer