From nobody Wed Nov 27 10:30:44 2024 Received: from out30-110.freemail.mail.aliyun.com (out30-110.freemail.mail.aliyun.com [115.124.30.110]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id B0991C13C; Sat, 12 Oct 2024 01:29:25 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=115.124.30.110 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1728696569; cv=none; b=e+jVO1/ZVt1bfxEB8t5yMp7PmR01dZ8c3gIiVM8+4T/yhACFbT38YQ1/a7j1lEn9Z0zEdxW7npJbIS0H/iCNXPtagCK+GtlPVyIQcDPFza+qCAvYqKGEGNOVh3s/cEV7nIXmCwYdXRuzgVfnlagaaB6JfGWfFvlB8M4T+RzoTAc= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1728696569; c=relaxed/simple; bh=UyRRmdSiBSiSEtr7Ig4aUmF7hG5TYY9ph5NJyArBOaw=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=qzvu/xMukTRZW7dkuTG7QluzEYlok51j4zTx+kVorpA8w6ydxSQxI9KOFl16MK4q/caB/uOvzW1KAwuiIftDFRcAUnp9PX3dHWz1uBVr4PTPRzvlK0wSm9k5ybVYxm0UPr9ZZ+blsTz1X+5VITbKjUOqZ01s2mEKNyeBCVnvcD4= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com; spf=pass smtp.mailfrom=linux.alibaba.com; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b=aT1WEZZI; arc=none smtp.client-ip=115.124.30.110 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b="aT1WEZZI" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.alibaba.com; s=default; t=1728696563; h=From:To:Subject:Date:Message-Id:MIME-Version; bh=y1N2AVpnjdi1B9UGRgfVFOIURIhOEBNxoghNH/yKJVk=; b=aT1WEZZIgzCRLMdYtCd6tSTXZ3WtFd1gHw2jtlt2cVoq152DJrnyZ54bvrX8TEQBbNPFMbpFemIqDWrIPLoCPp40fbFN3sz36kn3POctOnHXiktBQWG3lBCtoDsON/SltY+i/cl5lpPoF2kmXqSn3iCePYDa23PIw6E1hjzgKhg= Received: from localhost(mailfrom:lulie@linux.alibaba.com fp:SMTPD_---0WGtAIG6_1728696561 cluster:ay36) by smtp.aliyun-inc.com; Sat, 12 Oct 2024 09:29:22 +0800 From: Philo Lu To: netdev@vger.kernel.org Cc: willemdebruijn.kernel@gmail.com, davem@davemloft.net, edumazet@google.com, kuba@kernel.org, pabeni@redhat.com, dsahern@kernel.org, antony.antony@secunet.com, steffen.klassert@secunet.com, linux-kernel@vger.kernel.org, dust.li@linux.alibaba.com, jakub@cloudflare.com, fred.cc@alibaba-inc.com, yubing.qiuyubing@alibaba-inc.com Subject: [PATCH v4 net-next 3/3] ipv4/udp: Add 4-tuple hash for connected socket Date: Sat, 12 Oct 2024 09:29:18 +0800 Message-Id: <20241012012918.70888-4-lulie@linux.alibaba.com> X-Mailer: git-send-email 2.32.0.3.g01195cf9f In-Reply-To: <20241012012918.70888-1-lulie@linux.alibaba.com> References: <20241012012918.70888-1-lulie@linux.alibaba.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Currently, the udp_table has two hash table, the port hash and portaddr hash. Usually for UDP servers, all sockets have the same local port and addr, so they are all on the same hash slot within a reuseport group. In some applications, UDP servers use connect() to manage clients. In particular, when firstly receiving from an unseen 4 tuple, a new socket is created and connect()ed to the remote addr:port, and then the fd is used exclusively by the client. Once there are connected sks in a reuseport group, udp has to score all sks in the same hash2 slot to find the best match. This could be inefficient with a large number of connections, resulting in high softirq overhead. To solve the problem, this patch implement 4-tuple hash for connected udp sockets. During connect(), hash4 slot is updated, as well as a corresponding counter, hash4_cnt, in hslot2. In __udp4_lib_lookup(), hslot4 will be searched firstly if the counter is non-zero. Otherwise, hslot2 is used like before. Note that only connected sockets enter this hash4 path, while un-connected ones are not affected. Signed-off-by: Philo Lu Signed-off-by: Cambda Zhu Signed-off-by: Fred Chen Signed-off-by: Yubing Qiu --- include/net/udp.h | 3 +- net/ipv4/udp.c | 142 ++++++++++++++++++++++++++++++++++++++++++++-- net/ipv6/udp.c | 2 +- 3 files changed, 141 insertions(+), 6 deletions(-) diff --git a/include/net/udp.h b/include/net/udp.h index 80f9622d0db3..5633b51cf8d4 100644 --- a/include/net/udp.h +++ b/include/net/udp.h @@ -226,7 +226,7 @@ static inline int udp_lib_hash(struct sock *sk) } =20 void udp_lib_unhash(struct sock *sk); -void udp_lib_rehash(struct sock *sk, u16 new_hash); +void udp_lib_rehash(struct sock *sk, u16 new_hash, u16 new_hash4); =20 static inline void udp_lib_close(struct sock *sk, long timeout) { @@ -319,6 +319,7 @@ int udp_rcv(struct sk_buff *skb); int udp_ioctl(struct sock *sk, int cmd, int *karg); int udp_init_sock(struct sock *sk); int udp_pre_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); +int udp_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); int __udp_disconnect(struct sock *sk, int flags); int udp_disconnect(struct sock *sk, int flags); __poll_t udp_poll(struct file *file, struct socket *sock, poll_table *wait= ); diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 1498ccb79c58..d7e3866617e0 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -478,6 +478,27 @@ static struct sock *udp4_lib_lookup2(const struct net = *net, return result; } =20 +static struct sock *udp4_lib_lookup4(const struct net *net, + __be32 saddr, __be16 sport, + __be32 daddr, unsigned int hnum, + int dif, int sdif, + struct udp_table *udptable) +{ + unsigned int hash4 =3D udp_ehashfn(net, daddr, hnum, saddr, sport); + const __portpair ports =3D INET_COMBINED_PORTS(sport, hnum); + struct udp_hslot *hslot4 =3D udp_hashslot4(udptable, hash4); + struct udp_sock *up; + struct sock *sk; + + INET_ADDR_COOKIE(acookie, saddr, daddr); + udp_lrpa_for_each_entry_rcu(up, &hslot4->head) { + sk =3D (struct sock *)up; + if (inet_match(net, sk, acookie, ports, dif, sdif)) + return sk; + } + return NULL; +} + /* UDP is nearly always wildcards out the wazoo, it makes no sense to try * harder than this. -DaveM */ @@ -493,6 +514,12 @@ struct sock *__udp4_lib_lookup(const struct net *net, = __be32 saddr, hash2 =3D ipv4_portaddr_hash(net, daddr, hnum); hslot2 =3D udp_hashslot2(udptable, hash2); =20 + if (UDP_HSLOT_MAIN(hslot2)->hash4_cnt) { + result =3D udp4_lib_lookup4(net, saddr, sport, daddr, hnum, dif, sdif, u= dptable); + if (result) /* udp4_lib_lookup4 return sk or NULL */ + return result; + } + /* Lookup connected or non-wildcard socket */ result =3D udp4_lib_lookup2(net, saddr, sport, daddr, hnum, dif, sdif, @@ -1931,6 +1958,85 @@ int udp_pre_connect(struct sock *sk, struct sockaddr= *uaddr, int addr_len) } EXPORT_SYMBOL(udp_pre_connect); =20 +/* In hash4, rehash can also happen in connect(), where hash4_cnt keeps un= changed. */ +static void udp4_rehash4(struct udp_table *udptable, struct sock *sk, u16 = newhash4) +{ + struct udp_hslot *hslot4, *nhslot4; + + hslot4 =3D udp_hashslot4(udptable, udp_sk(sk)->udp_lrpa_hash); + nhslot4 =3D udp_hashslot4(udptable, newhash4); + udp_sk(sk)->udp_lrpa_hash =3D newhash4; + + if (hslot4 !=3D nhslot4) { + spin_lock_bh(&hslot4->lock); + hlist_del_init_rcu(&udp_sk(sk)->udp_lrpa_node); + hslot4->count--; + spin_unlock_bh(&hslot4->lock); + + synchronize_rcu(); + + spin_lock_bh(&nhslot4->lock); + hlist_add_head_rcu(&udp_sk(sk)->udp_lrpa_node, &nhslot4->head); + nhslot4->count++; + spin_unlock_bh(&nhslot4->lock); + } +} + +/* call with sock lock */ +static void udp4_hash4(struct sock *sk) +{ + struct udp_hslot *hslot, *hslot4; + struct udp_hslot_main *hslotm2; + struct net *net =3D sock_net(sk); + struct udp_table *udptable; + unsigned int hash; + + if (sk_unhashed(sk) || inet_sk(sk)->inet_rcv_saddr =3D=3D htonl(INADDR_AN= Y)) + return; + + hash =3D udp_ehashfn(net, inet_sk(sk)->inet_rcv_saddr, inet_sk(sk)->inet_= num, + inet_sk(sk)->inet_daddr, inet_sk(sk)->inet_dport); + + udptable =3D net->ipv4.udp_table; + if (udp_hashed4(sk)) { + udp4_rehash4(udptable, sk, hash); + return; + } + + hslot =3D udp_hashslot(udptable, net, udp_sk(sk)->udp_port_hash); + hslotm2 =3D udp_hashslot2_main(udptable, udp_sk(sk)->udp_portaddr_hash); + hslot4 =3D udp_hashslot4(udptable, hash); + udp_sk(sk)->udp_lrpa_hash =3D hash; + + spin_lock_bh(&hslot->lock); + if (rcu_access_pointer(sk->sk_reuseport_cb)) + reuseport_detach_sock(sk); + + spin_lock(&hslot4->lock); + hlist_add_head_rcu(&udp_sk(sk)->udp_lrpa_node, &hslot4->head); + hslot4->count++; + spin_unlock(&hslot4->lock); + + spin_lock(&hslotm2->hslot.lock); + hslotm2->hash4_cnt++; + spin_unlock(&hslotm2->hslot.lock); + + spin_unlock_bh(&hslot->lock); +} + +int udp_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) +{ + int res; + + lock_sock(sk); + res =3D __ip4_datagram_connect(sk, uaddr, addr_len); + if (!res) + udp4_hash4(sk); + release_sock(sk); + return res; +} +EXPORT_SYMBOL(udp_connect); + int __udp_disconnect(struct sock *sk, int flags) { struct inet_sock *inet =3D inet_sk(sk); @@ -1972,7 +2078,7 @@ void udp_lib_unhash(struct sock *sk) { if (sk_hashed(sk)) { struct udp_table *udptable =3D udp_get_table_prot(sk); - struct udp_hslot *hslot, *hslot2; + struct udp_hslot *hslot, *hslot2, *hslot4; =20 hslot =3D udp_hashslot(udptable, sock_net(sk), udp_sk(sk)->udp_port_hash); @@ -1990,6 +2096,18 @@ void udp_lib_unhash(struct sock *sk) hlist_del_init_rcu(&udp_sk(sk)->udp_portaddr_node); hslot2->count--; spin_unlock(&hslot2->lock); + + if (udp_hashed4(sk)) { + hslot4 =3D udp_hashslot4(udptable, udp_sk(sk)->udp_lrpa_hash); + spin_lock(&hslot4->lock); + hlist_del_init_rcu(&udp_sk(sk)->udp_lrpa_node); + hslot4->count--; + spin_unlock(&hslot4->lock); + + spin_lock(&hslot2->lock); + UDP_HSLOT_MAIN(hslot2)->hash4_cnt--; + spin_unlock(&hslot2->lock); + } } spin_unlock_bh(&hslot->lock); } @@ -1999,7 +2117,7 @@ EXPORT_SYMBOL(udp_lib_unhash); /* * inet_rcv_saddr was changed, we must rehash secondary hash */ -void udp_lib_rehash(struct sock *sk, u16 newhash) +void udp_lib_rehash(struct sock *sk, u16 newhash, u16 newhash4) { if (sk_hashed(sk)) { struct udp_table *udptable =3D udp_get_table_prot(sk); @@ -2031,6 +2149,19 @@ void udp_lib_rehash(struct sock *sk, u16 newhash) spin_unlock(&nhslot2->lock); } =20 + if (udp_hashed4(sk)) { + udp4_rehash4(udptable, sk, newhash4); + + if (hslot2 !=3D nhslot2) { + spin_lock(&hslot2->lock); + UDP_HSLOT_MAIN(hslot2)->hash4_cnt--; + spin_unlock(&hslot2->lock); + + spin_lock(&nhslot2->lock); + UDP_HSLOT_MAIN(nhslot2)->hash4_cnt++; + spin_unlock(&nhslot2->lock); + } + } spin_unlock_bh(&hslot->lock); } } @@ -2042,7 +2173,10 @@ void udp_v4_rehash(struct sock *sk) u16 new_hash =3D ipv4_portaddr_hash(sock_net(sk), inet_sk(sk)->inet_rcv_saddr, inet_sk(sk)->inet_num); - udp_lib_rehash(sk, new_hash); + u16 new_hash4 =3D udp_ehashfn(sock_net(sk), + inet_sk(sk)->inet_rcv_saddr, inet_sk(sk)->inet_num, + inet_sk(sk)->inet_daddr, inet_sk(sk)->inet_dport); + udp_lib_rehash(sk, new_hash, new_hash4); } =20 static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb) @@ -2935,7 +3069,7 @@ struct proto udp_prot =3D { .owner =3D THIS_MODULE, .close =3D udp_lib_close, .pre_connect =3D udp_pre_connect, - .connect =3D ip4_datagram_connect, + .connect =3D udp_connect, .disconnect =3D udp_disconnect, .ioctl =3D udp_ioctl, .init =3D udp_init_sock, diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index bbf3352213c4..4d3dfcb48a39 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -111,7 +111,7 @@ void udp_v6_rehash(struct sock *sk) &sk->sk_v6_rcv_saddr, inet_sk(sk)->inet_num); =20 - udp_lib_rehash(sk, new_hash); + udp_lib_rehash(sk, new_hash, 0); /* 4-tuple hash not implemented */ } =20 static int compute_score(struct sock *sk, const struct net *net, --=20 2.32.0.3.g01195cf9f