net: WireGuard secure network tunnel

WireGuard is a layer 3 secure networking tunnel made specifically for
the kernel, that aims to be much simpler and easier to audit than IPsec.
Extensive documentation and description of the protocol and
considerations, along with formal proofs of the cryptography, are
available at:

  * https://www.wireguard.com/
  * https://www.wireguard.com/papers/wireguard.pdf

This commit implements WireGuard as a simple network device driver,
accessible in the usual RTNL way used by virtual network drivers. It
makes use of the udp_tunnel APIs, GRO, GSO, NAPI, and the usual set of
networking subsystem APIs. It has a somewhat novel multicore queueing
system designed for maximum throughput and minimal latency of encryption
operations, but it is implemented modestly using workqueues and NAPI.
Configuration is done via generic Netlink, and following a review from
the Netlink maintainer a year ago, several high profile userspace tools
have already implemented the API.

This commit also comes with several different tests, both in-kernel
tests and out-of-kernel tests based on network namespaces, taking profit
of the fact that sockets used by WireGuard intentionally stay in the
namespace the WireGuard interface was originally created, exactly like
the semantics of userspace tun devices. See wireguard.com/netns/ for
pictures and examples.

The source code is fairly short, but rather than combining everything
into a single file, WireGuard is developed as cleanly separable files,
making auditing and comprehension easier. Things are laid out as
follows:

  * noise.[ch], cookie.[ch], messages.h: These implement the bulk of the
    cryptographic aspects of the protocol, and are mostly data-only in
    nature, taking in buffers of bytes and spitting out buffers of
    bytes. They also handle reference counting for their various shared
    pieces of data, like keys and key lists.

  * ratelimiter.[ch]: Used as an integral part of cookie.[ch] for
    ratelimiting certain types of cryptographic operations in accordance
    with particular WireGuard semantics.

  * allowedips.[ch], peerlookup.[ch]: The main lookup structures of
    WireGuard, the former being trie-like with particular semantics, an
    integral part of the design of the protocol, and the latter just
    being nice helper functions around the various hashtables we use.

  * device.[ch]: Implementation of functions for the netdevice and for
    rtnl, responsible for maintaining the life of a given interface and
    wiring it up to the rest of WireGuard.

  * peer.[ch]: Each interface has a list of peers, with helper functions
    available here for creation, destruction, and reference counting.

  * socket.[ch]: Implementation of functions related to udp_socket and
    the general set of kernel socket APIs, for sending and receiving
    ciphertext UDP packets, and taking care of WireGuard-specific sticky
    socket routing semantics for the automatic roaming.

  * netlink.[ch]: Userspace API entry point for configuring WireGuard
    peers and devices. The API has been implemented by several userspace
    tools and network management utility, and the WireGuard project
    distributes the basic wg(8) tool.

  * queueing.[ch]: Shared function on the rx and tx path for handling
    the various queues used in the multicore algorithms.

  * send.c: Handles encrypting outgoing packets in parallel on
    multiple cores, before sending them in order on a single core, via
    workqueues and ring buffers. Also handles sending handshake and cookie
    messages as part of the protocol, in parallel.

  * receive.c: Handles decrypting incoming packets in parallel on
    multiple cores, before passing them off in order to be ingested via
    the rest of the networking subsystem with GRO via the typical NAPI
    poll function. Also handles receiving handshake and cookie messages
    as part of the protocol, in parallel.

  * timers.[ch]: Uses the timer wheel to implement protocol particular
    event timeouts, and gives a set of very simple event-driven entry
    point functions for callers.

  * main.c, version.h: Initialization and deinitialization of the module.

  * selftest/*.h: Runtime unit tests for some of the most security
    sensitive functions.

  * tools/testing/selftests/wireguard/netns.sh: Aforementioned testing
    script using network namespaces.

This commit aims to be as self-contained as possible, implementing
WireGuard as a standalone module not needing much special handling or
coordination from the network subsystem. I expect for future
optimizations to the network stack to positively improve WireGuard, and
vice-versa, but for the time being, this exists as intentionally
standalone.

We introduce a menu option for CONFIG_WIREGUARD, as well as providing a
verbose debug log and self-tests via CONFIG_WIREGUARD_DEBUG.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Cc: David Miller <davem@davemloft.net>
Cc: Greg KH <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 221 insertions, 0 deletions

diff --git a/drivers/net/wireguard/peerlookup.c b/drivers/net/wireguard/peerlookup.c
new file mode 100644
index 000000000000..e4deb331476b
--- /dev/null
+++ b/drivers/net/wireguard/peerlookup.c
@@ -0,0 +1,221 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2015-2019 Jason A. Donenfeld <Jason@zx2c4.com>. All Rights Reserved.
+ */
+
+#include "peerlookup.h"
+#include "peer.h"
+#include "noise.h"
+
+static struct hlist_head *pubkey_bucket(struct pubkey_hashtable *table,
+					const u8 pubkey[NOISE_PUBLIC_KEY_LEN])
+{
+	/* siphash gives us a secure 64bit number based on a random key. Since
+	 * the bits are uniformly distributed, we can then mask off to get the
+	 * bits we need.
+	 */
+	const u64 hash = siphash(pubkey, NOISE_PUBLIC_KEY_LEN, &table->key);
+
+	return &table->hashtable[hash & (HASH_SIZE(table->hashtable) - 1)];
+}
+
+struct pubkey_hashtable *wg_pubkey_hashtable_alloc(void)
+{
+	struct pubkey_hashtable *table = kvmalloc(sizeof(*table), GFP_KERNEL);
+
+	if (!table)
+		return NULL;
+
+	get_random_bytes(&table->key, sizeof(table->key));
+	hash_init(table->hashtable);
+	mutex_init(&table->lock);
+	return table;
+}
+
+void wg_pubkey_hashtable_add(struct pubkey_hashtable *table,
+			     struct wg_peer *peer)
+{
+	mutex_lock(&table->lock);
+	hlist_add_head_rcu(&peer->pubkey_hash,
+			   pubkey_bucket(table, peer->handshake.remote_static));
+	mutex_unlock(&table->lock);
+}
+
+void wg_pubkey_hashtable_remove(struct pubkey_hashtable *table,
+				struct wg_peer *peer)
+{
+	mutex_lock(&table->lock);
+	hlist_del_init_rcu(&peer->pubkey_hash);
+	mutex_unlock(&table->lock);
+}
+
+/* Returns a strong reference to a peer */
+struct wg_peer *
+wg_pubkey_hashtable_lookup(struct pubkey_hashtable *table,
+			   const u8 pubkey[NOISE_PUBLIC_KEY_LEN])
+{
+	struct wg_peer *iter_peer, *peer = NULL;
+
+	rcu_read_lock_bh();
+	hlist_for_each_entry_rcu_bh(iter_peer, pubkey_bucket(table, pubkey),
+				    pubkey_hash) {
+		if (!memcmp(pubkey, iter_peer->handshake.remote_static,
+			    NOISE_PUBLIC_KEY_LEN)) {
+			peer = iter_peer;
+			break;
+		}
+	}
+	peer = wg_peer_get_maybe_zero(peer);
+	rcu_read_unlock_bh();
+	return peer;
+}
+
+static struct hlist_head *index_bucket(struct index_hashtable *table,
+				       const __le32 index)
+{
+	/* Since the indices are random and thus all bits are uniformly
+	 * distributed, we can find its bucket simply by masking.
+	 */
+	return &table->hashtable[(__force u32)index &
+				 (HASH_SIZE(table->hashtable) - 1)];
+}
+
+struct index_hashtable *wg_index_hashtable_alloc(void)
+{
+	struct index_hashtable *table = kvmalloc(sizeof(*table), GFP_KERNEL);
+
+	if (!table)
+		return NULL;
+
+	hash_init(table->hashtable);
+	spin_lock_init(&table->lock);
+	return table;
+}
+
+/* At the moment, we limit ourselves to 2^20 total peers, which generally might
+ * amount to 2^20*3 items in this hashtable. The algorithm below works by
+ * picking a random number and testing it. We can see that these limits mean we
+ * usually succeed pretty quickly:
+ *
+ * >>> def calculation(tries, size):
+ * ...     return (size / 2**32)**(tries - 1) *  (1 - (size / 2**32))
+ * ...
+ * >>> calculation(1, 2**20 * 3)
+ * 0.999267578125
+ * >>> calculation(2, 2**20 * 3)
+ * 0.0007318854331970215
+ * >>> calculation(3, 2**20 * 3)
+ * 5.360489012673497e-07
+ * >>> calculation(4, 2**20 * 3)
+ * 3.9261394135792216e-10
+ *
+ * At the moment, we don't do any masking, so this algorithm isn't exactly
+ * constant time in either the random guessing or in the hash list lookup. We
+ * could require a minimum of 3 tries, which would successfully mask the
+ * guessing. this would not, however, help with the growing hash lengths, which
+ * is another thing to consider moving forward.
+ */
+
+__le32 wg_index_hashtable_insert(struct index_hashtable *table,
+				 struct index_hashtable_entry *entry)
+{
+	struct index_hashtable_entry *existing_entry;
+
+	spin_lock_bh(&table->lock);
+	hlist_del_init_rcu(&entry->index_hash);
+	spin_unlock_bh(&table->lock);
+
+	rcu_read_lock_bh();
+
+search_unused_slot:
+	/* First we try to find an unused slot, randomly, while unlocked. */
+	entry->index = (__force __le32)get_random_u32();
+	hlist_for_each_entry_rcu_bh(existing_entry,
+				    index_bucket(table, entry->index),
+				    index_hash) {
+		if (existing_entry->index == entry->index)
+			/* If it's already in use, we continue searching. */
+			goto search_unused_slot;
+	}
+
+	/* Once we've found an unused slot, we lock it, and then double-check
+	 * that nobody else stole it from us.
+	 */
+	spin_lock_bh(&table->lock);
+	hlist_for_each_entry_rcu_bh(existing_entry,
+				    index_bucket(table, entry->index),
+				    index_hash) {
+		if (existing_entry->index == entry->index) {
+			spin_unlock_bh(&table->lock);
+			/* If it was stolen, we start over. */
+			goto search_unused_slot;
+		}
+	}
+	/* Otherwise, we know we have it exclusively (since we're locked),
+	 * so we insert.
+	 */
+	hlist_add_head_rcu(&entry->index_hash,
+			   index_bucket(table, entry->index));
+	spin_unlock_bh(&table->lock);
+
+	rcu_read_unlock_bh();
+
+	return entry->index;
+}
+
+bool wg_index_hashtable_replace(struct index_hashtable *table,
+				struct index_hashtable_entry *old,
+				struct index_hashtable_entry *new)
+{
+	if (unlikely(hlist_unhashed(&old->index_hash)))
+		return false;
+	spin_lock_bh(&table->lock);
+	new->index = old->index;
+	hlist_replace_rcu(&old->index_hash, &new->index_hash);
+
+	/* Calling init here NULLs out index_hash, and in fact after this
+	 * function returns, it's theoretically possible for this to get
+	 * reinserted elsewhere. That means the RCU lookup below might either
+	 * terminate early or jump between buckets, in which case the packet
+	 * simply gets dropped, which isn't terrible.
+	 */
+	INIT_HLIST_NODE(&old->index_hash);
+	spin_unlock_bh(&table->lock);
+	return true;
+}
+
+void wg_index_hashtable_remove(struct index_hashtable *table,
+			       struct index_hashtable_entry *entry)
+{
+	spin_lock_bh(&table->lock);
+	hlist_del_init_rcu(&entry->index_hash);
+	spin_unlock_bh(&table->lock);
+}
+
+/* Returns a strong reference to a entry->peer */
+struct index_hashtable_entry *
+wg_index_hashtable_lookup(struct index_hashtable *table,
+			  const enum index_hashtable_type type_mask,
+			  const __le32 index, struct wg_peer **peer)
+{
+	struct index_hashtable_entry *iter_entry, *entry = NULL;
+
+	rcu_read_lock_bh();
+	hlist_for_each_entry_rcu_bh(iter_entry, index_bucket(table, index),
+				    index_hash) {
+		if (iter_entry->index == index) {
+			if (likely(iter_entry->type & type_mask))
+				entry = iter_entry;
+			break;
+		}
+	}
+	if (likely(entry)) {
+		entry->peer = wg_peer_get_maybe_zero(entry->peer);
+		if (likely(entry->peer))
+			*peer = entry->peer;
+		else
+			entry = NULL;
+	}
+	rcu_read_unlock_bh();
+	return entry;
+}