Fix CVE-2026-58472: integer+buffer overflow in html_quote_string()

Backport fix for CVE-2026-58472 to wget-1.21.1. The patch
addresses an integer and buffer overflow in html_quote_string()
by using INT_ADD_OK for safe integer arithmetic when calculating
the output buffer size for HTML entity encoding. Also includes
the follow-up fix for an off-by-1 error in the size calculation
for each entity-encoded character, along with unit tests.

The combined fix is added as Patch6: wget-1.21-CVE-2026-58472.patch
based on upstream commits dd692d9 and f76978a.

CVE: CVE-2026-58472
Upstream patches:
 - dd692d9cea.patch
 - f76978a51b.patch
Resolves: RHEL-210636

This commit was backported by Ymir, a Red Hat Enterprise Linux software maintenance AI agent.

Assisted-by: Ymir
This commit is contained in:
RHEL Packaging Agent 2026-07-15 10:06:10 +00:00
parent 928040f599
commit a596752c69
2 changed files with 303 additions and 1 deletions

View File

@ -0,0 +1,296 @@
From 6e351fa40e7d9223b856002d62587c85cf382370 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Tim=20R=C3=BChsen?= <tim.ruehsen@gmx.de>
Date: Mon, 29 Jun 2026 19:13:15 +0200
Subject: [PATCH 1/2] * src/convert.c (html_quote_string): Fix integer+buffer
overflow
Reported-by: TristanInSec@gmail.com
---
src/convert.c | 31 ++++++++++++++++++++++++-------
1 file changed, 24 insertions(+), 7 deletions(-)
diff --git a/src/convert.c b/src/convert.c
index 683f719..741582f 100644
--- a/src/convert.c
+++ b/src/convert.c
@@ -36,6 +36,7 @@ as that of the covered work. */
#include <unistd.h>
#include <errno.h>
#include <assert.h>
+#include <intprops.h>
#include "convert.h"
#include "url.h"
#include "recur.h"
@@ -1169,21 +1170,37 @@ html_quote_string (const char *s)
{
const char *b = s;
char *p, *res;
- int i;
+ size_t i;
+ int ok;
/* Pass through the string, and count the new size. */
- for (i = 0; *s; s++, i++)
+ for (i = 0; *s; s++)
{
if (*s == '&')
- i += 4; /* `amp;' */
+ ok = INT_ADD_OK (i, 4, &i); /* `amp;' */
else if (*s == '<' || *s == '>')
- i += 3; /* `lt;' and `gt;' */
+ ok = INT_ADD_OK (i, 3, &i); /* `lt;' and `gt;' */
else if (*s == '\"')
- i += 5; /* `quot;' */
+ ok = INT_ADD_OK (i, 5, &i); /* `quot;' */
else if (*s == ' ')
- i += 4; /* #32; */
+ ok = INT_ADD_OK (i, 4, &i); /* #32; */
+ else
+ ok = INT_ADD_OK (i, 1, &i);
+
+ if (!ok)
+ {
+ DEBUGP (("Overflow detected in html_quote_string().\n"));
+ abort();
+ }
}
- res = xmalloc (i + 1);
+
+ if (!INT_ADD_OK (i, 1, &i))
+ {
+ DEBUGP (("Overflow detected in html_quote_string().\n"));
+ abort();
+ }
+
+ res = xmalloc (i);
s = b;
for (p = res; *s; s++)
{
From 8babd2417cee6f7433e6cee4982f7f4cd4787cf9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Tim=20R=C3=BChsen?= <tim.ruehsen@gmx.de>
Date: Thu, 2 Jul 2026 13:13:07 +0200
Subject: [PATCH 2/2] Regression: Fix buffer overflow in html_quote_string()
The regression has been introduced in commit dd692d9 and
is not part of any release.
The tests allow the address sanitizer to find the issue.
* src/convert.c: Fix string size calculation.
* tests/unit-tests.c: Added tests including tests for html_quote_string().
* tests/unit-tests.h: Add definitions for the test functions.
Reported-by: Trung Nguyen <trungnh@cystack.net>
---
src/convert.c | 148 +++++++++++++++++++++++++++++++++++++++++++--
tests/unit-tests.c | 4 ++
tests/unit-tests.h | 4 ++
3 files changed, 152 insertions(+), 4 deletions(-)
diff --git a/src/convert.c b/src/convert.c
index 741582f..24ab33b 100644
--- a/src/convert.c
+++ b/src/convert.c
@@ -48,6 +48,9 @@ as that of the covered work. */
#include "css-url.h"
#include "iri.h"
#include "xstrndup.h"
+#ifdef TESTING
+#include "../tests/unit-tests.h"
+#endif
static struct hash_table *dl_file_url_map;
struct hash_table *dl_url_file_map;
@@ -1177,13 +1180,13 @@ html_quote_string (const char *s)
for (i = 0; *s; s++)
{
if (*s == '&')
- ok = INT_ADD_OK (i, 4, &i); /* `amp;' */
+ ok = INT_ADD_OK (i, 4 + 1, &i); /* `amp;' */
else if (*s == '<' || *s == '>')
- ok = INT_ADD_OK (i, 3, &i); /* `lt;' and `gt;' */
+ ok = INT_ADD_OK (i, 3 + 1, &i); /* `lt;' and `gt;' */
else if (*s == '\"')
- ok = INT_ADD_OK (i, 5, &i); /* `quot;' */
+ ok = INT_ADD_OK (i, 5 + 1, &i); /* `quot;' */
else if (*s == ' ')
- ok = INT_ADD_OK (i, 4, &i); /* #32; */
+ ok = INT_ADD_OK (i, 4 + 1, &i); /* #32; */
else
ok = INT_ADD_OK (i, 1, &i);
@@ -1242,6 +1245,143 @@ html_quote_string (const char *s)
return res;
}
+#ifdef TESTING
+
+const char *
+test_construct_relative (void)
+{
+ static const struct {
+ const char *basefile;
+ const char *linkfile;
+ const char *expected;
+ } test_array[] = {
+ { "foo", "bar", "bar" },
+ { "A/foo", "A/bar", "bar" },
+ { "A/foo", "A/B/bar", "B/bar" },
+ { "A/X/foo", "A/Y/bar", "../Y/bar" },
+ { "X/", "Y/bar", "../Y/bar" },
+ { "/foo", "/bar", "bar" },
+ { "/a/b/c", "/a/b/d", "d" },
+ { "/a/b/c", "/a/b/c/d", "c/d" },
+ { "/a/b/c", "/a/b/c/d/e", "c/d/e" },
+ { "/a/b/c", "/x/y/z", "../../x/y/z" },
+ { "a/b", "c/d", "../c/d" },
+ { "./foo", "./bar", "bar" },
+ };
+
+ for (unsigned i = 0; i < countof (test_array); ++i)
+ {
+ char *result = construct_relative (test_array[i].basefile,
+ test_array[i].linkfile);
+ mu_assert ("test_construct_relative: wrong result",
+ strcmp (result, test_array[i].expected) == 0);
+ xfree (result);
+ }
+
+ return NULL;
+}
+
+const char *
+test_match_except_index (void)
+{
+ static const struct {
+ const char *s1;
+ const char *s2;
+ bool expected;
+ } test_array[] = {
+ { "foo/index.html", "foo/", true },
+ { "foo/", "foo/index.html", true },
+ { "foo", "foo/index.html", true },
+ { "foo", "foo/", true },
+ { "foo", "foo", true },
+ { "/foo/index.html", "/foo/", true },
+ { "/foo/", "/foo/index.html", true },
+ { "/foo", "/foo/index.html", true },
+ { "/foo", "/foo/", true },
+ { "foo/bar", "foo/qux", false },
+ { "foo/bar", "bar/foo", false },
+ };
+
+ for (unsigned i = 0; i < countof (test_array); ++i)
+ {
+ bool result = match_except_index (test_array[i].s1, test_array[i].s2);
+ mu_assert ("test_match_except_index: wrong result",
+ result == test_array[i].expected);
+ }
+
+ return NULL;
+}
+
+const char *
+test_find_fragment (void)
+{
+ static const struct {
+ const char *input;
+ int size;
+ bool has_fragment;
+ const char *fragment;
+ } test_array[] = {
+ { "http://example.com#section", 26, true, "#section" },
+ { "http://example.com", 18, false, NULL },
+ { "http://example.com?a=1#frag", 24, true, "#frag" },
+ { "http://example.com?a=1%26#frag", 28, true, "#frag" },
+ { "http://example.com?a=1&b=2#frag", 30, true, "#frag" },
+ { "a#b", 3, true, "#b" },
+ { "a", 1, false, NULL },
+ };
+ const char *bp, *ep;
+
+ for (unsigned i = 0; i < countof (test_array); ++i)
+ {
+ bool result = find_fragment (test_array[i].input,
+ test_array[i].size, &bp, &ep);
+ mu_assert ("test_find_fragment: wrong result",
+ result == test_array[i].has_fragment);
+ if (test_array[i].has_fragment)
+ {
+ mu_assert ("test_find_fragment: wrong fragment", bp != NULL);
+ mu_assert ("test_find_fragment: fragment mismatch",
+ strncmp (bp, test_array[i].fragment,
+ strlen (test_array[i].fragment)) == 0 &&
+ ep == test_array[i].input + test_array[i].size);
+ }
+ }
+
+ return NULL;
+}
+
+const char *
+test_html_quote_string (void)
+{
+ static const struct {
+ const char *input;
+ const char *expected;
+ } test_array[] = {
+ { "hello", "hello" },
+ { "a&b", "a&amp;b" },
+ { "<tag>", "&lt;tag&gt;" },
+ { "\"quote\"", "&quot;quote&quot;" },
+ { "space here", "space&#32;here" },
+ { "&<>\" ", "&amp;&lt;&gt;&quot;&#32;" },
+ { "no special", "no&#32;special" },
+ { "&&&&", "&amp;&amp;&amp;&amp;" },
+ { "<<>>", "&lt;&lt;&gt;&gt;" },
+ { "" , "" },
+ };
+
+ for (unsigned i = 0; i < countof (test_array); ++i)
+ {
+ char *result = html_quote_string (test_array[i].input);
+ mu_assert ("test_html_quote_string: wrong result",
+ strcmp (result, test_array[i].expected) == 0);
+ xfree (result);
+ }
+
+ return NULL;
+}
+
+#endif /* TESTING */
+
/*
* vim: et ts=2 sw=2
*/
diff --git a/tests/unit-tests.c b/tests/unit-tests.c
index 443e628..ac5ce69 100644
--- a/tests/unit-tests.c
+++ b/tests/unit-tests.c
@@ -65,6 +65,10 @@ all_tests(void)
mu_run_test (test_hsts_url_rewrite_congruent);
mu_run_test (test_hsts_read_database);
#endif
+ mu_run_test (test_construct_relative);
+ mu_run_test (test_match_except_index);
+ mu_run_test (test_find_fragment);
+ mu_run_test (test_html_quote_string);
return NULL;
}
diff --git a/tests/unit-tests.h b/tests/unit-tests.h
index dd10e54..8e43e96 100644
--- a/tests/unit-tests.h
+++ b/tests/unit-tests.h
@@ -61,6 +61,10 @@ const char *test_hsts_new_entry(void);
const char *test_hsts_url_rewrite_superdomain(void);
const char *test_hsts_url_rewrite_congruent(void);
const char *test_hsts_read_database(void);
+const char *test_construct_relative(void);
+const char *test_match_except_index(void);
+const char *test_find_fragment(void);
+const char *test_html_quote_string(void);
#endif /* TEST_H */

View File

@ -1,7 +1,7 @@
Summary: A utility for retrieving files using the HTTP or FTP protocols
Name: wget
Version: 1.21.1
Release: 8%{?dist}
Release: 9%{?dist}
License: GPLv3+
Url: http://www.gnu.org/software/wget/
Source: ftp://ftp.gnu.org/gnu/wget/wget-%{version}.tar.gz
@ -11,6 +11,9 @@ Patch2: wget-1.21-strtol.patch
Patch3: wget-1.21-metalink-man.patch
Patch4: wget-1.21-segfault.patch
Patch5: wget-1.21-CVE-2024-38428.patch
# https://gitlab.com/gnuwget/wget/-/commit/dd692d9cea5335b181d877ae917fe6e75587a812
# https://gitlab.com/gnuwget/wget/-/commit/f76978a51ba9365e7ecaed96c1cfb73197a38ca2
Patch6: wget-1.21-CVE-2026-58472.patch
Provides: webclient
Provides: bundled(gnulib)
@ -70,6 +73,9 @@ make check
%{_infodir}/*
%changelog
* Wed Jul 15 2026 RHEL Packaging Agent <redhat-ymir-agent@redhat.com> - 1.21.1-9
- Resolves: RHEL-210636 - Fix integer+buffer overflow in html_quote_string()
* Mon Jul 15 2024 Michal Ruprich <mruprich@redhat.com> - 1.21.1-8
- Resolves: RHEL-43226 - Misinterpretation of input may lead to improper behavior