Files
packages/libs/glib2/patches/104-CVE-2024-34397-tests-Add-a-test-case-for-what-happens-if-a-unique-name-d.patch
Petr Štetiar 3932357ffb glib2: update to 2.74.7 to fix several CVEs
Bump glib2 to 2.74.7 which fixes CVE-2023-29499, CVE-2023-32611,
CVE-2023-32636, CVE-2023-32643, CVE-2023-32665 and on top of that
backport CVE-2024-34397 fix from Debian Bookworm glib2 package
2.74.6-2+deb12u2. While at it refresh the patches so they apply cleanly.

References: https://security-tracker.debian.org/tracker/source-package/glib2.0
Fixes: CVE-2023-29499, CVE-2023-32611, CVE-2023-32636, CVE-2023-32643, CVE-2023-32665, CVE-2024-34397
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2024-09-05 13:27:51 +02:00

80 lines
2.6 KiB
Diff

From: Simon McVittie <smcv@collabora.com>
Date: Fri, 8 Mar 2024 19:44:03 +0000
Subject: tests: Add a test-case for what happens if a unique name doesn't
exist
On GNOME/glib#3268 there was some concern about whether this would
allow an attacker to send signals and have them be matched to a
GDBusProxy in this situation, but it seems that was a false alarm.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Origin: upstream, https://gitlab.gnome.org/GNOME/glib/-/issues/3268
---
gio/tests/gdbus-subscribe.c | 48 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 48 insertions(+)
--- a/gio/tests/gdbus-subscribe.c
+++ b/gio/tests/gdbus-subscribe.c
@@ -358,6 +358,53 @@ static const TestPlan plan_limit_by_uniq
},
};
+static const TestPlan plan_nonexistent_unique_name =
+{
+ .description = "A subscription via a unique name that doesn't exist "
+ "accepts no messages",
+ .steps = {
+ {
+ /* Subscriber wants to receive signals from service */
+ .action = TEST_ACTION_SUBSCRIBE,
+ .u.subscribe = {
+ /* This relies on the implementation detail that the dbus-daemon
+ * (and presumably other bus implementations) never actually generates
+ * a unique name in this format */
+ .string_sender = ":0.this.had.better.not.exist",
+ .path = EXAMPLE_PATH,
+ .iface = EXAMPLE_INTERFACE,
+ },
+ },
+ {
+ /* Attacker wants to trick subscriber into thinking that service
+ * sent a signal */
+ .action = TEST_ACTION_EMIT_SIGNAL,
+ .u.signal = {
+ .sender = TEST_CONN_ATTACKER,
+ .path = EXAMPLE_PATH,
+ .iface = EXAMPLE_INTERFACE,
+ .member = FOO_SIGNAL,
+ .received_by_conn = 0,
+ .received_by_proxy = 0
+ },
+ },
+ {
+ /* Attacker tries harder, by sending a signal unicast directly to
+ * the subscriber */
+ .action = TEST_ACTION_EMIT_SIGNAL,
+ .u.signal = {
+ .sender = TEST_CONN_ATTACKER,
+ .unicast_to = TEST_CONN_SUBSCRIBER,
+ .path = EXAMPLE_PATH,
+ .iface = EXAMPLE_INTERFACE,
+ .member = FOO_SIGNAL,
+ .received_by_conn = 0,
+ .received_by_proxy = 0
+ },
+ },
+ },
+};
+
static const TestPlan plan_limit_by_well_known_name =
{
.description = "A subscription via a well-known name only accepts messages "
@@ -1051,6 +1098,7 @@ main (int argc,
ADD_SUBSCRIBE_TEST (broadcast_from_anyone);
ADD_SUBSCRIBE_TEST (match_twice);
ADD_SUBSCRIBE_TEST (limit_by_unique_name);
+ ADD_SUBSCRIBE_TEST (nonexistent_unique_name);
ADD_SUBSCRIBE_TEST (limit_by_well_known_name);
return g_test_run();