From a0fd7cc87a018df1a17f9d3f0bd994c1f22c6b34 Mon Sep 17 00:00:00 2001 From: "Jason A. Donenfeld" Date: Wed, 6 May 2020 15:33:02 -0600 Subject: wireguard: selftests: use normal kernel stack size on ppc64 While at some point it might have made sense to be running these tests on ppc64 with 4k stacks, the kernel hasn't actually used 4k stacks on 64-bit powerpc in a long time, and more interesting things that we test don't really work when we deviate from the default (16k). So, we stop pushing our luck in this commit, and return to the default instead of the minimum. Signed-off-by: Jason A. Donenfeld Signed-off-by: David S. Miller --- tools/testing/selftests/wireguard/qemu/arch/powerpc64le.config | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/testing/selftests/wireguard/qemu/arch/powerpc64le.config b/tools/testing/selftests/wireguard/qemu/arch/powerpc64le.config index 990c510a9cfa..f52f1e2bc7f6 100644 --- a/tools/testing/selftests/wireguard/qemu/arch/powerpc64le.config +++ b/tools/testing/selftests/wireguard/qemu/arch/powerpc64le.config @@ -10,3 +10,4 @@ CONFIG_CMDLINE_BOOL=y CONFIG_CMDLINE="console=hvc0 wg.success=hvc1" CONFIG_SECTION_MISMATCH_WARN_ONLY=y CONFIG_FRAME_WARN=1280 +CONFIG_THREAD_SHIFT=14 -- cgit v1.2.3 From b673e24aad36981f327a6570412ffa7754de8911 Mon Sep 17 00:00:00 2001 From: "Jason A. Donenfeld" Date: Wed, 6 May 2020 15:33:03 -0600 Subject: wireguard: socket: remove errant restriction on looping to self It's already possible to create two different interfaces and loop packets between them. This has always been possible with tunnels in the kernel, and isn't specific to wireguard. Therefore, the networking stack already needs to deal with that. At the very least, the packet winds up exceeding the MTU and is discarded at that point. So, since this is already something that happens, there's no need to forbid the not very exceptional case of routing a packet back to the same interface; this loop is no different than others, and we shouldn't special case it, but rather rely on generic handling of loops in general. This also makes it easier to do interesting things with wireguard such as onion routing. At the same time, we add a selftest for this, ensuring that both onion routing works and infinite routing loops do not crash the kernel. We also add a test case for wireguard interfaces nesting packets and sending traffic between each other, as well as the loop in this case too. We make sure to send some throughput-heavy traffic for this use case, to stress out any possible recursion issues with the locks around workqueues. Fixes: e7096c131e51 ("net: WireGuard secure network tunnel") Signed-off-by: Jason A. Donenfeld Signed-off-by: David S. Miller --- drivers/net/wireguard/socket.c | 12 ------- tools/testing/selftests/wireguard/netns.sh | 54 ++++++++++++++++++++++++++++-- 2 files changed, 51 insertions(+), 15 deletions(-) diff --git a/drivers/net/wireguard/socket.c b/drivers/net/wireguard/socket.c index b0d6541582d3..f9018027fc13 100644 --- a/drivers/net/wireguard/socket.c +++ b/drivers/net/wireguard/socket.c @@ -76,12 +76,6 @@ static int send4(struct wg_device *wg, struct sk_buff *skb, net_dbg_ratelimited("%s: No route to %pISpfsc, error %d\n", wg->dev->name, &endpoint->addr, ret); goto err; - } else if (unlikely(rt->dst.dev == skb->dev)) { - ip_rt_put(rt); - ret = -ELOOP; - net_dbg_ratelimited("%s: Avoiding routing loop to %pISpfsc\n", - wg->dev->name, &endpoint->addr); - goto err; } if (cache) dst_cache_set_ip4(cache, &rt->dst, fl.saddr); @@ -149,12 +143,6 @@ static int send6(struct wg_device *wg, struct sk_buff *skb, net_dbg_ratelimited("%s: No route to %pISpfsc, error %d\n", wg->dev->name, &endpoint->addr, ret); goto err; - } else if (unlikely(dst->dev == skb->dev)) { - dst_release(dst); - ret = -ELOOP; - net_dbg_ratelimited("%s: Avoiding routing loop to %pISpfsc\n", - wg->dev->name, &endpoint->addr); - goto err; } if (cache) dst_cache_set_ip6(cache, dst, &fl.saddr); diff --git a/tools/testing/selftests/wireguard/netns.sh b/tools/testing/selftests/wireguard/netns.sh index 936e1ca9410e..17a1f53ceba0 100755 --- a/tools/testing/selftests/wireguard/netns.sh +++ b/tools/testing/selftests/wireguard/netns.sh @@ -48,8 +48,11 @@ cleanup() { exec 2>/dev/null printf "$orig_message_cost" > /proc/sys/net/core/message_cost ip0 link del dev wg0 + ip0 link del dev wg1 ip1 link del dev wg0 + ip1 link del dev wg1 ip2 link del dev wg0 + ip2 link del dev wg1 local to_kill="$(ip netns pids $netns0) $(ip netns pids $netns1) $(ip netns pids $netns2)" [[ -n $to_kill ]] && kill $to_kill pp ip netns del $netns1 @@ -77,18 +80,20 @@ ip0 link set wg0 netns $netns2 key1="$(pp wg genkey)" key2="$(pp wg genkey)" key3="$(pp wg genkey)" +key4="$(pp wg genkey)" pub1="$(pp wg pubkey <<<"$key1")" pub2="$(pp wg pubkey <<<"$key2")" pub3="$(pp wg pubkey <<<"$key3")" +pub4="$(pp wg pubkey <<<"$key4")" psk="$(pp wg genpsk)" [[ -n $key1 && -n $key2 && -n $psk ]] configure_peers() { ip1 addr add 192.168.241.1/24 dev wg0 - ip1 addr add fd00::1/24 dev wg0 + ip1 addr add fd00::1/112 dev wg0 ip2 addr add 192.168.241.2/24 dev wg0 - ip2 addr add fd00::2/24 dev wg0 + ip2 addr add fd00::2/112 dev wg0 n1 wg set wg0 \ private-key <(echo "$key1") \ @@ -230,9 +235,38 @@ n1 ping -W 1 -c 1 192.168.241.2 n1 wg set wg0 private-key <(echo "$key3") n2 wg set wg0 peer "$pub3" preshared-key <(echo "$psk") allowed-ips 192.168.241.1/32 peer "$pub1" remove n1 ping -W 1 -c 1 192.168.241.2 +n2 wg set wg0 peer "$pub3" remove + +# Test that we can route wg through wg +ip1 addr flush dev wg0 +ip2 addr flush dev wg0 +ip1 addr add fd00::5:1/112 dev wg0 +ip2 addr add fd00::5:2/112 dev wg0 +n1 wg set wg0 private-key <(echo "$key1") peer "$pub2" preshared-key <(echo "$psk") allowed-ips fd00::5:2/128 endpoint 127.0.0.1:2 +n2 wg set wg0 private-key <(echo "$key2") listen-port 2 peer "$pub1" preshared-key <(echo "$psk") allowed-ips fd00::5:1/128 endpoint 127.212.121.99:9998 +ip1 link add wg1 type wireguard +ip2 link add wg1 type wireguard +ip1 addr add 192.168.241.1/24 dev wg1 +ip1 addr add fd00::1/112 dev wg1 +ip2 addr add 192.168.241.2/24 dev wg1 +ip2 addr add fd00::2/112 dev wg1 +ip1 link set mtu 1340 up dev wg1 +ip2 link set mtu 1340 up dev wg1 +n1 wg set wg1 listen-port 5 private-key <(echo "$key3") peer "$pub4" allowed-ips 192.168.241.2/32,fd00::2/128 endpoint [fd00::5:2]:5 +n2 wg set wg1 listen-port 5 private-key <(echo "$key4") peer "$pub3" allowed-ips 192.168.241.1/32,fd00::1/128 endpoint [fd00::5:1]:5 +tests +# Try to set up a routing loop between the two namespaces +ip1 link set netns $netns0 dev wg1 +ip0 addr add 192.168.241.1/24 dev wg1 +ip0 link set up dev wg1 +n0 ping -W 1 -c 1 192.168.241.2 +n1 wg set wg0 peer "$pub2" endpoint 192.168.241.2:7 +ip2 link del wg0 +ip2 link del wg1 +! n0 ping -W 1 -c 10 -f 192.168.241.2 || false # Should not crash kernel +ip0 link del wg1 ip1 link del wg0 -ip2 link del wg0 # Test using NAT. We now change the topology to this: # ┌────────────────────────────────────────┐ ┌────────────────────────────────────────────────┐ ┌────────────────────────────────────────┐ @@ -282,6 +316,20 @@ pp sleep 3 n2 ping -W 1 -c 1 192.168.241.1 n1 wg set wg0 peer "$pub2" persistent-keepalive 0 +# Test that onion routing works, even when it loops +n1 wg set wg0 peer "$pub3" allowed-ips 192.168.242.2/32 endpoint 192.168.241.2:5 +ip1 addr add 192.168.242.1/24 dev wg0 +ip2 link add wg1 type wireguard +ip2 addr add 192.168.242.2/24 dev wg1 +n2 wg set wg1 private-key <(echo "$key3") listen-port 5 peer "$pub1" allowed-ips 192.168.242.1/32 +ip2 link set wg1 up +n1 ping -W 1 -c 1 192.168.242.2 +ip2 link del wg1 +n1 wg set wg0 peer "$pub3" endpoint 192.168.242.2:5 +! n1 ping -W 1 -c 1 192.168.242.2 || false # Should not crash kernel +n1 wg set wg0 peer "$pub3" remove +ip1 addr del 192.168.242.1/24 dev wg0 + # Do a wg-quick(8)-style policy routing for the default route, making sure vethc has a v6 address to tease out bugs. ip1 -6 addr add fc00::9/96 dev vethc ip1 -6 route add default via fc00::1 -- cgit v1.2.3 From 4005f5c3c9d006157ba716594e0d70c88a235c5e Mon Sep 17 00:00:00 2001 From: "Jason A. Donenfeld" Date: Wed, 6 May 2020 15:33:04 -0600 Subject: wireguard: send/receive: cond_resched() when processing worker ringbuffers Users with pathological hardware reported CPU stalls on CONFIG_ PREEMPT_VOLUNTARY=y, because the ringbuffers would stay full, meaning these workers would never terminate. That turned out not to be okay on systems without forced preemption, which Sultan observed. This commit adds a cond_resched() to the bottom of each loop iteration, so that these workers don't hog the core. Note that we don't need this on the napi poll worker, since that terminates after its budget is expended. Suggested-by: Sultan Alsawaf Reported-by: Wang Jian Fixes: e7096c131e51 ("net: WireGuard secure network tunnel") Signed-off-by: Jason A. Donenfeld Signed-off-by: David S. Miller --- drivers/net/wireguard/receive.c | 2 ++ drivers/net/wireguard/send.c | 4 ++++ 2 files changed, 6 insertions(+) diff --git a/drivers/net/wireguard/receive.c b/drivers/net/wireguard/receive.c index 267f202f1931..2566e13a292d 100644 --- a/drivers/net/wireguard/receive.c +++ b/drivers/net/wireguard/receive.c @@ -516,6 +516,8 @@ void wg_packet_decrypt_worker(struct work_struct *work) &PACKET_CB(skb)->keypair->receiving)) ? PACKET_STATE_CRYPTED : PACKET_STATE_DEAD; wg_queue_enqueue_per_peer_napi(skb, state); + if (need_resched()) + cond_resched(); } } diff --git a/drivers/net/wireguard/send.c b/drivers/net/wireguard/send.c index 3e030d614df5..dc3079e17c7f 100644 --- a/drivers/net/wireguard/send.c +++ b/drivers/net/wireguard/send.c @@ -281,6 +281,8 @@ void wg_packet_tx_worker(struct work_struct *work) wg_noise_keypair_put(keypair, false); wg_peer_put(peer); + if (need_resched()) + cond_resched(); } } @@ -304,6 +306,8 @@ void wg_packet_encrypt_worker(struct work_struct *work) } wg_queue_enqueue_per_peer(&PACKET_PEER(first)->tx_queue, first, state); + if (need_resched()) + cond_resched(); } } -- cgit v1.2.3 From 4fed818ef54b08d4b29200e416cce65546ad5312 Mon Sep 17 00:00:00 2001 From: "Jason A. Donenfeld" Date: Wed, 6 May 2020 15:33:05 -0600 Subject: wireguard: selftests: initalize ipv6 members to NULL to squelch clang warning Without setting these to NULL, clang complains in certain configurations that have CONFIG_IPV6=n: In file included from drivers/net/wireguard/ratelimiter.c:223: drivers/net/wireguard/selftest/ratelimiter.c:173:34: error: variable 'skb6' is uninitialized when used here [-Werror,-Wuninitialized] ret = timings_test(skb4, hdr4, skb6, hdr6, &test_count); ^~~~ drivers/net/wireguard/selftest/ratelimiter.c:123:29: note: initialize the variable 'skb6' to silence this warning struct sk_buff *skb4, *skb6; ^ = NULL drivers/net/wireguard/selftest/ratelimiter.c:173:40: error: variable 'hdr6' is uninitialized when used here [-Werror,-Wuninitialized] ret = timings_test(skb4, hdr4, skb6, hdr6, &test_count); ^~~~ drivers/net/wireguard/selftest/ratelimiter.c:125:22: note: initialize the variable 'hdr6' to silence this warning struct ipv6hdr *hdr6; ^ We silence this warning by setting the variables to NULL as the warning suggests. Reported-by: Arnd Bergmann Signed-off-by: Jason A. Donenfeld Signed-off-by: David S. Miller --- drivers/net/wireguard/selftest/ratelimiter.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/wireguard/selftest/ratelimiter.c b/drivers/net/wireguard/selftest/ratelimiter.c index bcd6462e4540..007cd4457c5f 100644 --- a/drivers/net/wireguard/selftest/ratelimiter.c +++ b/drivers/net/wireguard/selftest/ratelimiter.c @@ -120,9 +120,9 @@ bool __init wg_ratelimiter_selftest(void) enum { TRIALS_BEFORE_GIVING_UP = 5000 }; bool success = false; int test = 0, trials; - struct sk_buff *skb4, *skb6; + struct sk_buff *skb4, *skb6 = NULL; struct iphdr *hdr4; - struct ipv6hdr *hdr6; + struct ipv6hdr *hdr6 = NULL; if (IS_ENABLED(CONFIG_KASAN) || IS_ENABLED(CONFIG_UBSAN)) return true; -- cgit v1.2.3 From 243f2148937adc72bcaaa590d482d599c936efde Mon Sep 17 00:00:00 2001 From: "Jason A. Donenfeld" Date: Wed, 6 May 2020 15:33:06 -0600 Subject: wireguard: send/receive: use explicit unlikely branch instead of implicit coalescing It's very unlikely that send will become true. It's nearly always false between 0 and 120 seconds of a session, and in most cases becomes true only between 120 and 121 seconds before becoming false again. So, unlikely(send) is clearly the right option here. What happened before was that we had this complex boolean expression with multiple likely and unlikely clauses nested. Since this is evaluated left-to-right anyway, the whole thing got converted to unlikely. So, we can clean this up to better represent what's going on. The generated code is the same. Suggested-by: Sultan Alsawaf Signed-off-by: Jason A. Donenfeld Signed-off-by: David S. Miller --- drivers/net/wireguard/receive.c | 13 ++++++------- drivers/net/wireguard/send.c | 15 ++++++--------- 2 files changed, 12 insertions(+), 16 deletions(-) diff --git a/drivers/net/wireguard/receive.c b/drivers/net/wireguard/receive.c index 2566e13a292d..3bb5b9ae7cd1 100644 --- a/drivers/net/wireguard/receive.c +++ b/drivers/net/wireguard/receive.c @@ -226,21 +226,20 @@ void wg_packet_handshake_receive_worker(struct work_struct *work) static void keep_key_fresh(struct wg_peer *peer) { struct noise_keypair *keypair; - bool send = false; + bool send; if (peer->sent_lastminute_handshake) return; rcu_read_lock_bh(); keypair = rcu_dereference_bh(peer->keypairs.current_keypair); - if (likely(keypair && READ_ONCE(keypair->sending.is_valid)) && - keypair->i_am_the_initiator && - unlikely(wg_birthdate_has_expired(keypair->sending.birthdate, - REJECT_AFTER_TIME - KEEPALIVE_TIMEOUT - REKEY_TIMEOUT))) - send = true; + send = keypair && READ_ONCE(keypair->sending.is_valid) && + keypair->i_am_the_initiator && + wg_birthdate_has_expired(keypair->sending.birthdate, + REJECT_AFTER_TIME - KEEPALIVE_TIMEOUT - REKEY_TIMEOUT); rcu_read_unlock_bh(); - if (send) { + if (unlikely(send)) { peer->sent_lastminute_handshake = true; wg_packet_send_queued_handshake_initiation(peer, false); } diff --git a/drivers/net/wireguard/send.c b/drivers/net/wireguard/send.c index dc3079e17c7f..6687db699803 100644 --- a/drivers/net/wireguard/send.c +++ b/drivers/net/wireguard/send.c @@ -124,20 +124,17 @@ void wg_packet_send_handshake_cookie(struct wg_device *wg, static void keep_key_fresh(struct wg_peer *peer) { struct noise_keypair *keypair; - bool send = false; + bool send; rcu_read_lock_bh(); keypair = rcu_dereference_bh(peer->keypairs.current_keypair); - if (likely(keypair && READ_ONCE(keypair->sending.is_valid)) && - (unlikely(atomic64_read(&keypair->sending.counter.counter) > - REKEY_AFTER_MESSAGES) || - (keypair->i_am_the_initiator && - unlikely(wg_birthdate_has_expired(keypair->sending.birthdate, - REKEY_AFTER_TIME))))) - send = true; + send = keypair && READ_ONCE(keypair->sending.is_valid) && + (atomic64_read(&keypair->sending.counter.counter) > REKEY_AFTER_MESSAGES || + (keypair->i_am_the_initiator && + wg_birthdate_has_expired(keypair->sending.birthdate, REKEY_AFTER_TIME))); rcu_read_unlock_bh(); - if (send) + if (unlikely(send)) wg_packet_send_queued_handshake_initiation(peer, false); } -- cgit v1.2.3