Skip to contentSkip to conversation

Add NETDEV_ADD_DEVICE() helper function

PATCH New Martin Doucha 3 participants 8 messages Aug 25, 2021 – Aug 26, 2021 Library code

lore
Share & tools
lore
#0 lore raw

PATCH 1/3 Add NETDEV_ADD_DEVICE() helper function

---
 doc/network-c-api.txt   |  4 ++++
 include/tst_netdevice.h |  5 +++++
 lib/tst_netdevice.c     | 44 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 53 insertions(+)
doc/network-c-api.txt +4 -0
index 145836a81..f24614fd8 100644
@@ -162,6 +162,10 @@ stage.
   on success, 0 on error. Add +"CONFIG_VETH"+ to +test.needs_kconfigs+ if your
   test calls this function.
 
+- +int NETDEV_ADD_DEVICE(const char *ifname, const char *devtype)+ - Creates
+  a new network device named +ifname+ of specified device type. Returns 1 on
+  success, 0 on error.
+
 - +int REMOVE_NETDEV(const char *ifname)+ ? Removes network device +ifname+.
   Returns 1 on success, 0 on error.
 
include/tst_netdevice.h +5 -0
index 3a6698731..0bd33765b 100644
@@ -23,6 +23,11 @@ int tst_create_veth_pair(const char *file, const int lineno,
 #define CREATE_VETH_PAIR(ifname1, ifname2) \
 	tst_create_veth_pair(__FILE__, __LINE__, (ifname1), (ifname2))
 
+int tst_netdev_add_device(const char *file, const int lineno,
+	const char *ifname, const char *devtype);
+#define NETDEV_ADD_DEVICE(ifname, devtype) \
+	tst_netdev_add_device(__FILE__, __LINE__, (ifname), (devtype))
+
 int tst_remove_netdev(const char *file, const int lineno, const char *ifname);
 #define REMOVE_NETDEV(ifname) tst_remove_netdev(__FILE__, __LINE__, (ifname))
 
lib/tst_netdevice.c +44 -0

50 lines not on this page · raw message

Signature · 1 line
2.32.0
#1 lore raw

PATCH 2/3 Rename REMOVE_NETDEV() to NETDEV_REMOVE_DEVICE()

Use the same naming style as in the rest of netdevice library.

---
 doc/network-c-api.txt   | 4 ++--
 include/tst_netdevice.h | 6 ++++--
 lib/tst_netdevice.c     | 3 ++-
 3 files changed, 8 insertions(+), 5 deletions(-)
doc/network-c-api.txt +2 -2
index f24614fd8..e910d159f 100644
@@ -166,8 +166,8 @@ stage.
   a new network device named +ifname+ of specified device type. Returns 1 on
   success, 0 on error.
 
-- +int REMOVE_NETDEV(const char *ifname)+ ? Removes network device +ifname+.
-  Returns 1 on success, 0 on error.
+- +int NETDEV_REMOVE_DEVICE(const char *ifname)+ ? Removes network device
+  +ifname+. Returns 1 on success, 0 on error.
 
 2.2 Network address management
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/tst_netdevice.h +4 -2
index 0bd33765b..f02661208 100644
@@ -28,8 +28,10 @@ int tst_netdev_add_device(const char *file, const int lineno,
 #define NETDEV_ADD_DEVICE(ifname, devtype) \
 	tst_netdev_add_device(__FILE__, __LINE__, (ifname), (devtype))
 
-int tst_remove_netdev(const char *file, const int lineno, const char *ifname);
-#define REMOVE_NETDEV(ifname) tst_remove_netdev(__FILE__, __LINE__, (ifname))
+int tst_netdev_remove_device(const char *file, const int lineno,
+	const char *ifname);
+#define NETDEV_REMOVE_DEVICE(ifname) \
+	tst_netdev_remove_device(__FILE__, __LINE__, (ifname))
 
 int tst_netdev_add_address(const char *file, const int lineno,
 	const char *ifname, unsigned int family, const void *address,
lib/tst_netdevice.c +2 -1
index 9a8c622e2..ba9a55b70 100644
@@ -201,7 +201,8 @@ int tst_netdev_add_device(const char *file, const int lineno,
 	return ret;
 }
 
-int tst_remove_netdev(const char *file, const int lineno, const char *ifname)
+int tst_netdev_remove_device(const char *file, const int lineno,
+	const char *ifname)
 {
 	struct ifinfomsg info = { .ifi_family = AF_UNSPEC };
 	struct tst_rtnl_context *ctx;
Signature · 1 line
2.32.0
#2 lore raw

PATCH 3/3 Add test for CVE 2021-3609

In reply to Martin Doucha

Fixes #863

---
 runtest/can                           |   1 +
 runtest/cve                           |   1 +
 testcases/network/can/cve/.gitignore  |   1 +
 testcases/network/can/cve/Makefile    |  10 ++
 testcases/network/can/cve/can_bcm01.c | 161 ++++++++++++++++++++++++++
 5 files changed, 174 insertions(+)
 create mode 100644 testcases/network/can/cve/.gitignore
 create mode 100644 testcases/network/can/cve/Makefile
 create mode 100644 testcases/network/can/cve/can_bcm01.c
runtest/can +1 -0
index b637183c6..23cbf9acd 100644
@@ -1,2 +1,3 @@
 can_filter can_filter
 can_rcv_own_msgs can_rcv_own_msgs
+can_bcm01 can_bcm01
runtest/cve +1 -0
index c27f58d8d..449f5ad7e 100644
@@ -66,6 +66,7 @@ cve-2020-14416 pty03
 cve-2020-25705 icmp_rate_limit01
 cve-2020-29373 io_uring02
 cve-2021-3444 bpf_prog05
+cve-2021-6309 can_bcm01
 cve-2021-22555 setsockopt08 -i 100
 cve-2021-26708 vsock01
 # Tests below may cause kernel memory leak
testcases/network/can/cve/.gitignore new file +1 -0
new file mode 100644
index 000000000..3d138b0b4
@@ -0,0 +1 @@
+/can_bcm01
testcases/network/can/cve/Makefile new file +10 -0
new file mode 100644
index 000000000..86f84e9f2
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: GPL-2.0-or-later
+
+top_srcdir		?= ../../../..
+
+include $(top_srcdir)/include/mk/testcases.mk
+
+can_bcm01: CFLAGS += -pthread
+can_bcm01: LDLIBS += -lrt
+
+include $(top_srcdir)/include/mk/generic_leaf_target.mk
testcases/network/can/cve/can_bcm01.c new file +161 -0

161 lines not on this page · raw message

Signature · 1 line
2.32.0
#3 lore raw

PATCH 3/3 Add test for CVE 2021-3609

Hello Martin,

Quoted 195 lines · Martin Doucha

Fixes #863

---
 runtest/can                           |   1 +
 runtest/cve                           |   1 +
 testcases/network/can/cve/.gitignore  |   1 +
 testcases/network/can/cve/Makefile    |  10 ++
 testcases/network/can/cve/can_bcm01.c | 161 ++++++++++++++++++++++++++
 5 files changed, 174 insertions(+)
 create mode 100644 testcases/network/can/cve/.gitignore
 create mode 100644 testcases/network/can/cve/Makefile
 create mode 100644 testcases/network/can/cve/can_bcm01.c
runtest/can +1 -0
index b637183c6..23cbf9acd 100644
@@ -1,2 +1,3 @@
 can_filter can_filter
 can_rcv_own_msgs can_rcv_own_msgs
+can_bcm01 can_bcm01
runtest/cve +1 -0
index c27f58d8d..449f5ad7e 100644
@@ -66,6 +66,7 @@ cve-2020-14416 pty03
 cve-2020-25705 icmp_rate_limit01
 cve-2020-29373 io_uring02
 cve-2021-3444 bpf_prog05
+cve-2021-6309 can_bcm01
 cve-2021-22555 setsockopt08 -i 100
 cve-2021-26708 vsock01
 # Tests below may cause kernel memory leak
testcases/network/can/cve/.gitignore new file +1 -0
new file mode 100644
index 000000000..3d138b0b4
@@ -0,0 +1 @@
+/can_bcm01
testcases/network/can/cve/Makefile new file +10 -0
new file mode 100644
index 000000000..86f84e9f2
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: GPL-2.0-or-later
+
+top_srcdir		?= ../../../..
+
+include $(top_srcdir)/include/mk/testcases.mk
+
+can_bcm01: CFLAGS += -pthread
+can_bcm01: LDLIBS += -lrt
+
+include $(top_srcdir)/include/mk/generic_leaf_target.mk
testcases/network/can/cve/can_bcm01.c new file +132 -0
new file mode 100644
index 000000000..b1816e6c2
@@ -0,0 +1,161 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2020 SUSE LLC <[email protected]>
+ *
+ * CVE-2021-3609
+ *
+ * Test for race condition vulnerability in CAN BCM. Fixed in:
+ *
+ *  commit d5f9023fa61ee8b94f37a93f08e94b136cf1e463
+ *  Author: Thadeu Lima de Souza Cascardo <[email protected]>
+ *  Date:   Sat Jun 19 13:18:13 2021 -0300
+ *
+ *  can: bcm: delay release of struct bcm_op after synchronize_rcu()
+ */
+
+#include "config.h"
+#include "tst_test.h"
+
+#ifdef HAVE_LINUX_CAN_H
+
+#include <linux/can.h>
+#include <linux/can/bcm.h>
+
+#include "tst_netdevice.h"
+#include "tst_fuzzy_sync.h"
+
+#define LTP_DEVICE "ltp_vcan0"
+
+struct test_payload {
+	struct bcm_msg_head head;
+	struct can_frame frame;
+};
+
+static int sock1 = -1, sock2 = -1;
+static struct tst_fzsync_pair fzsync_pair;
+
+static void setup(void)
+{
+	struct sockaddr_can addr = { .can_family = AF_CAN };
+
+	/*
+	 * Older kernels require explicit modprobe of vcan. Newer kernels
+	 * will load the modules automatically and support CAN in network
+	 * namespace which would eliminate the need for running the test
+	 * with root privileges.
+	 */
+	tst_cmd((const char*[]){"modprobe", "vcan", NULL}, NULL, NULL, 0);
+
+	NETDEV_ADD_DEVICE(LTP_DEVICE, "vcan");
+	NETDEV_SET_STATE(LTP_DEVICE, 1);
+	addr.can_ifindex = NETDEV_INDEX_BY_NAME(LTP_DEVICE);
+	addr.can_addr.tp.rx_id = 1;
+	sock1 = SAFE_SOCKET(AF_CAN, SOCK_DGRAM, CAN_BCM);
+	SAFE_CONNECT(sock1, (struct sockaddr *)&addr, sizeof(addr));
+
+	fzsync_pair.exec_loops = 100000;
+	tst_fzsync_pair_init(&fzsync_pair);
+}
+
+static void *thread_run(void *arg)
+{
+	struct test_payload data = {
+		{
+			.opcode = TX_SEND,
+			.flags = RX_NO_AUTOTIMER,
+			.count = -1,
+			.nframes = 1
+		},
+		{0}
+	};
+	struct iovec iov = {
+		.iov_base = &data,
+		.iov_len = sizeof(data)
+	};
+	struct msghdr msg = {
+		.msg_iov = &iov,
+		.msg_iovlen = 1
+	};
+
+	while (tst_fzsync_run_b(&fzsync_pair)) {
+		tst_fzsync_start_race_b(&fzsync_pair);
+		SAFE_SENDMSG(iov.iov_len, sock1, &msg, 0);
+		tst_fzsync_end_race_b(&fzsync_pair);
+	}
+
+	return arg;
+}
+
+static void run(void)
+{
+	struct sockaddr_can addr = { .can_family = AF_CAN };
+	struct bcm_msg_head data = {
+		.opcode = RX_SETUP,
+		.flags = RX_FILTER_ID | SETTIMER | STARTTIMER,
+		.ival1.tv_sec = 1,
+		.ival2.tv_sec = 1
+	};
+	struct iovec iov = {
+		.iov_base = &data,
+		.iov_len = sizeof(data)
+	};
+	struct msghdr msg = {
+		.msg_iov = &iov,
+		.msg_iovlen = 1,
+	};
+
+	tst_fzsync_pair_reset(&fzsync_pair, thread_run);
+
+	while (tst_fzsync_run_a(&fzsync_pair)) {
+		sock2 = SAFE_SOCKET(AF_CAN, SOCK_DGRAM, CAN_BCM);
+		SAFE_CONNECT(sock2, (struct sockaddr *)&addr, sizeof(addr));
+		SAFE_SENDMSG(iov.iov_len, sock2, &msg, 0);
+		tst_fzsync_start_race_a(&fzsync_pair);
+		SAFE_CLOSE(sock2);
+		tst_fzsync_end_race_a(&fzsync_pair);
+	}
+
+	tst_res(TPASS, "Nothing bad happened, probably");
+}
+
+static void cleanup(void)
+{
+	tst_fzsync_pair_cleanup(&fzsync_pair);
+
+	if (sock1 >= 0)
+		SAFE_CLOSE(sock1);
+
+	if (sock2 >= 0)
+		SAFE_CLOSE(sock2);
+
+	NETDEV_REMOVE_DEVICE(LTP_DEVICE);
+}

testcases/network/can/cve/can_bcm01.c line 142

Quoted 10 lines
testcases/network/can/cve/can_bcm01.c new file +10 -0
+
+static struct tst_test test = {
+	.test_all = run,
+	.setup = setup,
+	.cleanup = cleanup,
+	.taint_check = TST_TAINT_W | TST_TAINT_D,
+	.needs_root = 1,
+	.needs_kconfigs = (const char *[]) {
+		"CONFIG_USER_NS=y",
+		"CONFIG_NET_NS=y",

I think you dropped NS due to lack of CAN support on older kernels?

Otherwise LGTM!

Signature · 2 lines
Thank you,
Richard.
#4 lore raw

PATCH 3/3 Add test for CVE 2021-3609

Quoted 3 lines · Richard Palethorpe

Hello Martin,

I think you dropped NS due to lack of CAN support on older kernels?

Exactly. If I create a new network namespace, SAFE_SOCKET() will fail with EAFNOSUPPORT on kernel 4.4.

Signature · 7 lines
Martin Doucha   [email protected]
QA Engineer for Software Maintenance
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic
#5 lore raw

PATCH 3/3 Add test for CVE 2021-3609

Hi!

Quoted 4 lines
Quoted 1 line

I think you dropped NS due to lack of CAN support on older kernels?

Exactly. If I create a new network namespace, SAFE_SOCKET() will fail with EAFNOSUPPORT on kernel 4.4.

Then I guess that we have to remove the needs_kconfig before pushing, right?

Signature · 2 lines
Cyril Hrubis
[email protected]
#6 lore raw

PATCH 3/3 Add test for CVE 2021-3609

Quoted 8 lines · Cyril Hrubis

Hi!

Quoted 4 lines
Quoted 1 line

I think you dropped NS due to lack of CAN support on older kernels?

Exactly. If I create a new network namespace, SAFE_SOCKET() will fail with EAFNOSUPPORT on kernel 4.4.

Then I guess that we have to remove the needs_kconfig before pushing, right?

Yes, please. I forgot to remove needs_kconfig after rewriting setup() and cleanup() 5 times...

Signature · 7 lines
Martin Doucha   [email protected]
QA Engineer for Software Maintenance
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic
#7 lore raw

PATCH 3/3 Add test for CVE 2021-3609

Hi!

Quoted 5 lines
Quoted 2 lines

Then I guess that we have to remove the needs_kconfig before pushing, right?

Yes, please. I forgot to remove needs_kconfig after rewriting setup() and cleanup() 5 times...

Pushed with the namspace requirements removed, thanks.

Signature · 2 lines
Cyril Hrubis
[email protected]