All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Jason A. Donenfeld" <Jason@zx2c4.com>
To: netdev@vger.kernel.org, davem@davemloft.net
Cc: "Jason A. Donenfeld" <Jason@zx2c4.com>,
	Streun Fabio <fstreun@student.ethz.ch>,
	stable@vger.kernel.org
Subject: [PATCH net 08/10] wireguard: receive: drop handshakes if queue lock is contended
Date: Mon, 29 Nov 2021 10:39:27 -0500	[thread overview]
Message-ID: <20211129153929.3457-9-Jason@zx2c4.com> (raw)
In-Reply-To: <20211129153929.3457-1-Jason@zx2c4.com>

If we're being delivered packets from multiple CPUs so quickly that the
ring lock is contended for CPU tries, then it's safe to assume that the
queue is near capacity anyway, so just drop the packet rather than
spinning. This helps deal with multicore DoS that can interfere with
data path performance. It _still_ does not completely fix the issue, but
it again chips away at it.

Reported-by: Streun Fabio <fstreun@student.ethz.ch>
Cc: stable@vger.kernel.org
Fixes: e7096c131e51 ("net: WireGuard secure network tunnel")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
---
 drivers/net/wireguard/receive.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/drivers/net/wireguard/receive.c b/drivers/net/wireguard/receive.c
index f4e537e3e8ec..7b8df406c773 100644
--- a/drivers/net/wireguard/receive.c
+++ b/drivers/net/wireguard/receive.c
@@ -554,9 +554,19 @@ void wg_packet_receive(struct wg_device *wg, struct sk_buff *skb)
 	case cpu_to_le32(MESSAGE_HANDSHAKE_INITIATION):
 	case cpu_to_le32(MESSAGE_HANDSHAKE_RESPONSE):
 	case cpu_to_le32(MESSAGE_HANDSHAKE_COOKIE): {
-		int cpu;
-		if (unlikely(!rng_is_initialized() ||
-			     ptr_ring_produce_bh(&wg->handshake_queue.ring, skb))) {
+		int cpu, ret = -EBUSY;
+
+		if (unlikely(!rng_is_initialized()))
+			goto drop;
+		if (atomic_read(&wg->handshake_queue_len) > MAX_QUEUED_INCOMING_HANDSHAKES / 2) {
+			if (spin_trylock_bh(&wg->handshake_queue.ring.producer_lock)) {
+				ret = __ptr_ring_produce(&wg->handshake_queue.ring, skb);
+				spin_unlock_bh(&wg->handshake_queue.ring.producer_lock);
+			}
+		} else
+			ret = ptr_ring_produce_bh(&wg->handshake_queue.ring, skb);
+		if (ret) {
+	drop:
 			net_dbg_skb_ratelimited("%s: Dropping handshake packet from %pISpfsc\n",
 						wg->dev->name, skb);
 			goto err;
-- 
2.34.1


  parent reply	other threads:[~2021-11-29 17:50 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-29 15:39 [PATCH net 00/10] wireguard/siphash patches for 5.16-rc6 Jason A. Donenfeld
2021-11-29 15:39 ` [PATCH net 01/10] wireguard: allowedips: add missing __rcu annotation to satisfy sparse Jason A. Donenfeld
2021-11-29 15:39 ` [PATCH net 02/10] wireguard: selftests: increase default dmesg log size Jason A. Donenfeld
2021-11-29 15:39 ` [PATCH net 03/10] wireguard: selftests: actually test for routing loops Jason A. Donenfeld
2021-11-29 15:39 ` [PATCH net 04/10] wireguard: main: rename 'mod_init' & 'mod_exit' functions to be module-specific Jason A. Donenfeld
2021-11-29 15:39 ` [PATCH net 05/10] wireguard: selftests: rename DEBUG_PI_LIST to DEBUG_PLIST Jason A. Donenfeld
2021-11-29 15:39 ` [PATCH net 06/10] wireguard: device: reset peer src endpoint when netns exits Jason A. Donenfeld
2021-11-30 16:00   ` Toke Høiland-Jørgensen
2021-11-29 15:39 ` [PATCH net 07/10] wireguard: receive: use ring buffer for incoming handshakes Jason A. Donenfeld
2021-11-29 15:39 ` Jason A. Donenfeld [this message]
2021-11-29 15:39 ` [PATCH net 09/10] wireguard: ratelimiter: use kvcalloc() instead of kvzalloc() Jason A. Donenfeld
2021-11-29 15:39 ` [PATCH net 10/10] siphash: use _unaligned version by default Jason A. Donenfeld
2021-11-29 18:17 ` [PATCH net 00/10] wireguard/siphash patches for 5.16-rc6 Jakub Kicinski
2021-11-29 18:18   ` Jason A. Donenfeld

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211129153929.3457-9-Jason@zx2c4.com \
    --to=jason@zx2c4.com \
    --cc=davem@davemloft.net \
    --cc=fstreun@student.ethz.ch \
    --cc=netdev@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.