From f82b681a511f4d61069e9586a9cf97bdef371ef3 Mon Sep 17 00:00:00 2001 From: Yuchung Cheng Date: Mon, 13 Jul 2015 12:10:20 -0700 Subject: tcp: don't use F-RTO on non-recurring timeouts Currently F-RTO may repeatedly send new data packets on non-recurring timeouts in CA_Loss mode. This is a bug because F-RTO (RFC5682) should only be used on either new recovery or recurring timeouts. This exacerbates the recovery progress during frequent timeout & repair, because we prioritize sending new data packets instead of repairing the holes when the bandwidth is already scarce. Fix it by correcting the test of a new recovery episode. Signed-off-by: Yuchung Cheng Signed-off-by: Neal Cardwell Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'net/ipv4') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 684f095d196e..728f5b3d3c64 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -1917,14 +1917,13 @@ void tcp_enter_loss(struct sock *sk) const struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; - bool new_recovery = false; + bool new_recovery = icsk->icsk_ca_state < TCP_CA_Recovery; bool is_reneg; /* is receiver reneging on SACKs? */ /* Reduce ssthresh if it has not yet been made inside this window. */ if (icsk->icsk_ca_state <= TCP_CA_Disorder || !after(tp->high_seq, tp->snd_una) || (icsk->icsk_ca_state == TCP_CA_Loss && !icsk->icsk_retransmits)) { - new_recovery = true; tp->prior_ssthresh = tcp_current_ssthresh(sk); tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk); tcp_ca_event(sk, CA_EVENT_LOSS); -- cgit v1.2.3-59-g8ed1b From 03645a11a570d52e70631838cb786eb4253eb463 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 14 Jul 2015 08:10:22 +0200 Subject: ipv6: lock socket in ip6_datagram_connect() ip6_datagram_connect() is doing a lot of socket changes without socket being locked. This looks wrong, at least for udp_lib_rehash() which could corrupt lists because of concurrent udp_sk(sk)->udp_portaddr_hash accesses. Signed-off-by: Eric Dumazet Acked-by: Herbert Xu Signed-off-by: David S. Miller --- include/net/ip.h | 1 + net/ipv4/datagram.c | 16 ++++++++++++---- net/ipv6/datagram.c | 20 +++++++++++++++----- 3 files changed, 28 insertions(+), 9 deletions(-) (limited to 'net/ipv4') diff --git a/include/net/ip.h b/include/net/ip.h index 0750a186ea63..d5fe9f2ab699 100644 --- a/include/net/ip.h +++ b/include/net/ip.h @@ -161,6 +161,7 @@ static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk) } /* datagram.c */ +int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); void ip4_datagram_release_cb(struct sock *sk); diff --git a/net/ipv4/datagram.c b/net/ipv4/datagram.c index 90c0e8386116..574fad9cca05 100644 --- a/net/ipv4/datagram.c +++ b/net/ipv4/datagram.c @@ -20,7 +20,7 @@ #include #include -int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) +int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) { struct inet_sock *inet = inet_sk(sk); struct sockaddr_in *usin = (struct sockaddr_in *) uaddr; @@ -39,8 +39,6 @@ int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) sk_dst_reset(sk); - lock_sock(sk); - oif = sk->sk_bound_dev_if; saddr = inet->inet_saddr; if (ipv4_is_multicast(usin->sin_addr.s_addr)) { @@ -82,9 +80,19 @@ int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) sk_dst_set(sk, &rt->dst); err = 0; out: - release_sock(sk); return err; } +EXPORT_SYMBOL(__ip4_datagram_connect); + +int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) +{ + int res; + + lock_sock(sk); + res = __ip4_datagram_connect(sk, uaddr, addr_len); + release_sock(sk); + return res; +} EXPORT_SYMBOL(ip4_datagram_connect); /* Because UDP xmit path can manipulate sk_dst_cache without holding diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c index 62d908e64eeb..b10a88986a98 100644 --- a/net/ipv6/datagram.c +++ b/net/ipv6/datagram.c @@ -40,7 +40,7 @@ static bool ipv6_mapped_addr_any(const struct in6_addr *a) return ipv6_addr_v4mapped(a) && (a->s6_addr32[3] == 0); } -int ip6_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) +static int __ip6_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) { struct sockaddr_in6 *usin = (struct sockaddr_in6 *) uaddr; struct inet_sock *inet = inet_sk(sk); @@ -56,7 +56,7 @@ int ip6_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) if (usin->sin6_family == AF_INET) { if (__ipv6_only_sock(sk)) return -EAFNOSUPPORT; - err = ip4_datagram_connect(sk, uaddr, addr_len); + err = __ip4_datagram_connect(sk, uaddr, addr_len); goto ipv4_connected; } @@ -98,9 +98,9 @@ int ip6_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) sin.sin_addr.s_addr = daddr->s6_addr32[3]; sin.sin_port = usin->sin6_port; - err = ip4_datagram_connect(sk, - (struct sockaddr *) &sin, - sizeof(sin)); + err = __ip4_datagram_connect(sk, + (struct sockaddr *) &sin, + sizeof(sin)); ipv4_connected: if (err) @@ -204,6 +204,16 @@ out: fl6_sock_release(flowlabel); return err; } + +int ip6_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) +{ + int res; + + lock_sock(sk); + res = __ip6_datagram_connect(sk, uaddr, addr_len); + release_sock(sk); + return res; +} EXPORT_SYMBOL_GPL(ip6_datagram_connect); int ip6_datagram_connect_v6_only(struct sock *sk, struct sockaddr *uaddr, -- cgit v1.2.3-59-g8ed1b From 0848f6428ba3a2e42db124d41ac6f548655735bf Mon Sep 17 00:00:00 2001 From: Edward Hyunkoo Jee Date: Tue, 21 Jul 2015 09:43:59 +0200 Subject: inet: frags: fix defragmented packet's IP header for af_packet When ip_frag_queue() computes positions, it assumes that the passed sk_buff does not contain L2 headers. However, when PACKET_FANOUT_FLAG_DEFRAG is used, IP reassembly functions can be called on outgoing packets that contain L2 headers. Also, IPv4 checksum is not corrected after reassembly. Fixes: 7736d33f4262 ("packet: Add pre-defragmentation support for ipv4 fanouts.") Signed-off-by: Edward Hyunkoo Jee Signed-off-by: Eric Dumazet Cc: Willem de Bruijn Cc: Jerry Chu Signed-off-by: David S. Miller --- net/ipv4/ip_fragment.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net/ipv4') diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index a50dc6d408d1..31f71b15cfba 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -351,7 +351,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb) ihl = ip_hdrlen(skb); /* Determine the position of this fragment. */ - end = offset + skb->len - ihl; + end = offset + skb->len - skb_network_offset(skb) - ihl; err = -EINVAL; /* Is this the final fragment? */ @@ -381,7 +381,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb) goto err; err = -ENOMEM; - if (!pskb_pull(skb, ihl)) + if (!pskb_pull(skb, skb_network_offset(skb) + ihl)) goto err; err = pskb_trim_rcsum(skb, end - offset); @@ -641,6 +641,8 @@ static int ip_frag_reasm(struct ipq *qp, struct sk_buff *prev, iph->frag_off = 0; } + ip_send_check(iph); + IP_INC_STATS_BH(net, IPSTATS_MIB_REASMOKS); qp->q.fragments = NULL; qp->q.fragments_tail = NULL; -- cgit v1.2.3-59-g8ed1b From 89e478a2aa58af2548b7f316e4d5b6bcc9eade5b Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 22 Jul 2015 07:02:00 +0200 Subject: tcp: suppress a division by zero warning Andrew Morton reported following warning on one ARM build with gcc-4.4 : net/ipv4/inet_hashtables.c: In function 'inet_ehash_locks_alloc': net/ipv4/inet_hashtables.c:617: warning: division by zero Even guarded with a test on sizeof(spinlock_t), compiler does not like current construct on a !CONFIG_SMP build. Remove the warning by using a temporary variable. Fixes: 095dc8e0c368 ("tcp: fix/cleanup inet_ehash_locks_alloc()") Reported-by: Andrew Morton Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/inet_hashtables.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'net/ipv4') diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index 5f9b063bbe8a..0cb9165421d4 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -624,22 +624,21 @@ EXPORT_SYMBOL_GPL(inet_hashinfo_init); int inet_ehash_locks_alloc(struct inet_hashinfo *hashinfo) { + unsigned int locksz = sizeof(spinlock_t); unsigned int i, nblocks = 1; - if (sizeof(spinlock_t) != 0) { + if (locksz != 0) { /* allocate 2 cache lines or at least one spinlock per cpu */ - nblocks = max_t(unsigned int, - 2 * L1_CACHE_BYTES / sizeof(spinlock_t), - 1); + nblocks = max(2U * L1_CACHE_BYTES / locksz, 1U); nblocks = roundup_pow_of_two(nblocks * num_possible_cpus()); /* no more locks than number of hash buckets */ nblocks = min(nblocks, hashinfo->ehash_mask + 1); - hashinfo->ehash_locks = kmalloc_array(nblocks, sizeof(spinlock_t), + hashinfo->ehash_locks = kmalloc_array(nblocks, locksz, GFP_KERNEL | __GFP_NOWARN); if (!hashinfo->ehash_locks) - hashinfo->ehash_locks = vmalloc(nblocks * sizeof(spinlock_t)); + hashinfo->ehash_locks = vmalloc(nblocks * locksz); if (!hashinfo->ehash_locks) return -ENOMEM; -- cgit v1.2.3-59-g8ed1b