Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Cross-merge networking fixes after downstream PR.

Conflicts:

net/sched/act_ct.c
  26488172b029 ("net/sched: Fix UAF when resolving a clash")
  3abbd7ed8b76 ("act_ct: prepare for stolen verdict coming from conntrack and nat engine")

No adjacent changes.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>

4 files changed, 242 insertions, 0 deletions

diff --git a/tools/testing/selftests/bpf/config b/tools/testing/selftests/bpf/config
index 5291e97df749..4ca84c8d9116 100644
--- a/tools/testing/selftests/bpf/config
+++ b/tools/testing/selftests/bpf/config
@@ -58,9 +58,12 @@ CONFIG_MPLS=y
 CONFIG_MPLS_IPTUNNEL=y
 CONFIG_MPLS_ROUTING=y
 CONFIG_MPTCP=y
+CONFIG_NET_ACT_SKBMOD=y
+CONFIG_NET_CLS=y
 CONFIG_NET_CLS_ACT=y
 CONFIG_NET_CLS_BPF=y
 CONFIG_NET_CLS_FLOWER=y
+CONFIG_NET_CLS_MATCHALL=y
 CONFIG_NET_FOU=y
 CONFIG_NET_FOU_IP_TUNNELS=y
 CONFIG_NET_IPGRE=y
diff --git a/tools/testing/selftests/bpf/prog_tests/tc_links.c b/tools/testing/selftests/bpf/prog_tests/tc_links.c
index bc9841144685..1af9ec1149aa 100644
--- a/tools/testing/selftests/bpf/prog_tests/tc_links.c
+++ b/tools/testing/selftests/bpf/prog_tests/tc_links.c
@@ -9,6 +9,8 @@
 #define ping_cmd "ping -q -c1 -w1 127.0.0.1 > /dev/null"
 
 #include "test_tc_link.skel.h"
+
+#include "netlink_helpers.h"
 #include "tc_helpers.h"
 
 void serial_test_tc_links_basic(void)
@@ -1787,6 +1789,65 @@ void serial_test_tc_links_ingress(void)
 	test_tc_links_ingress(BPF_TCX_INGRESS, false, false);
 }
 
+struct qdisc_req {
+	struct nlmsghdr  n;
+	struct tcmsg     t;
+	char             buf[1024];
+};
+
+static int qdisc_replace(int ifindex, const char *kind, bool block)
+{
+	struct rtnl_handle rth = { .fd = -1 };
+	struct qdisc_req req;
+	int err;
+
+	err = rtnl_open(&rth, 0);
+	if (!ASSERT_OK(err, "open_rtnetlink"))
+		return err;
+
+	memset(&req, 0, sizeof(req));
+	req.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct tcmsg));
+	req.n.nlmsg_flags = NLM_F_CREATE | NLM_F_REPLACE | NLM_F_REQUEST;
+	req.n.nlmsg_type = RTM_NEWQDISC;
+	req.t.tcm_family = AF_UNSPEC;
+	req.t.tcm_ifindex = ifindex;
+	req.t.tcm_parent = 0xfffffff1;
+
+	addattr_l(&req.n, sizeof(req), TCA_KIND, kind, strlen(kind) + 1);
+	if (block)
+		addattr32(&req.n, sizeof(req), TCA_INGRESS_BLOCK, 1);
+
+	err = rtnl_talk(&rth, &req.n, NULL);
+	ASSERT_OK(err, "talk_rtnetlink");
+	rtnl_close(&rth);
+	return err;
+}
+
+void serial_test_tc_links_dev_chain0(void)
+{
+	int err, ifindex;
+
+	ASSERT_OK(system("ip link add dev foo type veth peer name bar"), "add veth");
+	ifindex = if_nametoindex("foo");
+	ASSERT_NEQ(ifindex, 0, "non_zero_ifindex");
+	err = qdisc_replace(ifindex, "ingress", true);
+	if (!ASSERT_OK(err, "attaching ingress"))
+		goto cleanup;
+	ASSERT_OK(system("tc filter add block 1 matchall action skbmod swap mac"), "add block");
+	err = qdisc_replace(ifindex, "clsact", false);
+	if (!ASSERT_OK(err, "attaching clsact"))
+		goto cleanup;
+	/* Heuristic: kern_sync_rcu() alone does not work; a wait-time of ~5s
+	 * triggered the issue without the fix reliably 100% of the time.
+	 */
+	sleep(5);
+	ASSERT_OK(system("tc filter add dev foo ingress matchall action skbmod swap mac"), "add filter");
+cleanup:
+	ASSERT_OK(system("ip link del dev foo"), "del veth");
+	ASSERT_EQ(if_nametoindex("foo"), 0, "foo removed");
+	ASSERT_EQ(if_nametoindex("bar"), 0, "bar removed");
+}
+
 static void test_tc_links_dev_mixed(int target)
 {
 	LIBBPF_OPTS(bpf_tc_opts, tc_opts, .handle = 1, .priority = 1);
diff --git a/tools/testing/selftests/bpf/prog_tests/timer_lockup.c b/tools/testing/selftests/bpf/prog_tests/timer_lockup.c
new file mode 100644
index 000000000000..871d16cb95cf
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/timer_lockup.c
@@ -0,0 +1,91 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <sched.h>
+#include <test_progs.h>
+#include <pthread.h>
+#include <network_helpers.h>
+
+#include "timer_lockup.skel.h"
+
+static long cpu;
+static int *timer1_err;
+static int *timer2_err;
+static bool skip;
+
+volatile int k = 0;
+
+static void *timer_lockup_thread(void *arg)
+{
+	LIBBPF_OPTS(bpf_test_run_opts, opts,
+		.data_in	= &pkt_v4,
+		.data_size_in	= sizeof(pkt_v4),
+		.repeat		= 1000,
+	);
+	int i, prog_fd = *(int *)arg;
+	cpu_set_t cpuset;
+
+	CPU_ZERO(&cpuset);
+	CPU_SET(__sync_fetch_and_add(&cpu, 1), &cpuset);
+	ASSERT_OK(pthread_setaffinity_np(pthread_self(), sizeof(cpuset),
+				         &cpuset),
+		  "cpu affinity");
+
+	for (i = 0; !READ_ONCE(*timer1_err) && !READ_ONCE(*timer2_err); i++) {
+		bpf_prog_test_run_opts(prog_fd, &opts);
+		/* Skip the test if we can't reproduce the race in a reasonable
+		 * amount of time.
+		 */
+		if (i > 50) {
+			WRITE_ONCE(skip, true);
+			break;
+		}
+	}
+
+	return NULL;
+}
+
+void test_timer_lockup(void)
+{
+	int timer1_prog, timer2_prog;
+	struct timer_lockup *skel;
+	pthread_t thrds[2];
+	void *ret;
+
+	skel = timer_lockup__open_and_load();
+	if (!ASSERT_OK_PTR(skel, "timer_lockup__open_and_load"))
+		return;
+
+	timer1_prog = bpf_program__fd(skel->progs.timer1_prog);
+	timer2_prog = bpf_program__fd(skel->progs.timer2_prog);
+
+	timer1_err = &skel->bss->timer1_err;
+	timer2_err = &skel->bss->timer2_err;
+
+	if (!ASSERT_OK(pthread_create(&thrds[0], NULL, timer_lockup_thread,
+				      &timer1_prog),
+		       "pthread_create thread1"))
+		goto out;
+	if (!ASSERT_OK(pthread_create(&thrds[1], NULL, timer_lockup_thread,
+				      &timer2_prog),
+		       "pthread_create thread2")) {
+		pthread_exit(&thrds[0]);
+		goto out;
+	}
+
+	pthread_join(thrds[1], &ret);
+	pthread_join(thrds[0], &ret);
+
+	if (skip) {
+		test__skip();
+		goto out;
+	}
+
+	if (*timer1_err != -EDEADLK && *timer1_err != 0)
+		ASSERT_FAIL("timer1_err bad value");
+	if (*timer2_err != -EDEADLK && *timer2_err != 0)
+		ASSERT_FAIL("timer2_err bad value");
+out:
+	timer_lockup__destroy(skel);
+	return;
+}
diff --git a/tools/testing/selftests/bpf/progs/timer_lockup.c b/tools/testing/selftests/bpf/progs/timer_lockup.c
new file mode 100644
index 000000000000..3e520133281e
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/timer_lockup.c
@@ -0,0 +1,87 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/bpf.h>
+#include <time.h>
+#include <errno.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_tracing.h>
+#include "bpf_misc.h"
+
+char _license[] SEC("license") = "GPL";
+
+struct elem {
+	struct bpf_timer t;
+};
+
+struct {
+	__uint(type, BPF_MAP_TYPE_ARRAY);
+	__uint(max_entries, 1);
+	__type(key, int);
+	__type(value, struct elem);
+} timer1_map SEC(".maps");
+
+struct {
+	__uint(type, BPF_MAP_TYPE_ARRAY);
+	__uint(max_entries, 1);
+	__type(key, int);
+	__type(value, struct elem);
+} timer2_map SEC(".maps");
+
+int timer1_err;
+int timer2_err;
+
+static int timer_cb1(void *map, int *k, struct elem *v)
+{
+	struct bpf_timer *timer;
+	int key = 0;
+
+	timer = bpf_map_lookup_elem(&timer2_map, &key);
+	if (timer)
+		timer2_err = bpf_timer_cancel(timer);
+
+	return 0;
+}
+
+static int timer_cb2(void *map, int *k, struct elem *v)
+{
+	struct bpf_timer *timer;
+	int key = 0;
+
+	timer = bpf_map_lookup_elem(&timer1_map, &key);
+	if (timer)
+		timer1_err = bpf_timer_cancel(timer);
+
+	return 0;
+}
+
+SEC("tc")
+int timer1_prog(void *ctx)
+{
+	struct bpf_timer *timer;
+	int key = 0;
+
+	timer = bpf_map_lookup_elem(&timer1_map, &key);
+	if (timer) {
+		bpf_timer_init(timer, &timer1_map, CLOCK_BOOTTIME);
+		bpf_timer_set_callback(timer, timer_cb1);
+		bpf_timer_start(timer, 1, BPF_F_TIMER_CPU_PIN);
+	}
+
+	return 0;
+}
+
+SEC("tc")
+int timer2_prog(void *ctx)
+{
+	struct bpf_timer *timer;
+	int key = 0;
+
+	timer = bpf_map_lookup_elem(&timer2_map, &key);
+	if (timer) {
+		bpf_timer_init(timer, &timer2_map, CLOCK_BOOTTIME);
+		bpf_timer_set_callback(timer, timer_cb2);
+		bpf_timer_start(timer, 1, BPF_F_TIMER_CPU_PIN);
+	}
+
+	return 0;
+}