From e42e24c3cc072088756d84ef07b492ac2a3ae2e5 Mon Sep 17 00:00:00 2001 From: Matvejchikov Ilya Date: Mon, 24 Jul 2017 16:02:12 +0400 Subject: tcp: remove redundant argument from tcp_rcv_established() The last (4th) argument of tcp_rcv_established() is redundant as it always equals to skb->len and the skb itself is always passed as 2th agrument. There is no reason to have it. Signed-off-by: Ilya V. Matveychikov Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv6/tcp_ipv6.c') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 2521690d62d6..90a32576c3d0 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1296,7 +1296,7 @@ static int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb) } } - tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len); + tcp_rcv_established(sk, skb, tcp_hdr(skb)); if (opt_skb) goto ipv6_pktoptions; return 0; -- cgit From 3a3a4e3054137c5ff5d4d306ec834f6d25d7f95b Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Fri, 28 Jul 2017 22:18:57 +0200 Subject: ipv6: constify inet6_protocol structures The inet6_protocol structure is only passed as the first argument to inet6_add_protocol or inet6_del_protocol, both of which are declared as const. Thus the inet6_protocol structure itself can be const. Also drop __read_mostly where present on the newly const structures. Done with the help of Coccinelle. Signed-off-by: Julia Lawall Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv6/tcp_ipv6.c') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 90a32576c3d0..2968a33cca7d 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1945,7 +1945,7 @@ struct proto tcpv6_prot = { .diag_destroy = tcp_abort, }; -static struct inet6_protocol tcpv6_protocol = { +static const struct inet6_protocol tcpv6_protocol = { .early_demux = tcp_v6_early_demux, .early_demux_handler = tcp_v6_early_demux, .handler = tcp_v6_rcv, -- cgit From e7942d0633c47c791ece6afa038be9cf977226de Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Sun, 30 Jul 2017 03:57:18 +0200 Subject: tcp: remove prequeue support prequeue is a tcp receive optimization that moves part of rx processing from bh to process context. This only works if the socket being processed belongs to a process that is blocked in recv on that socket. In practice, this doesn't happen anymore that often because nowadays servers tend to use an event driven (epoll) model. Even normal client applications (web browsers) commonly use many tcp connections in parallel. This has measureable impact only in netperf (which uses plain recv and thus allows prequeue use) from host to locally running vm (~4%), however, there were no changes when using netperf between two physical hosts with ixgbe interfaces. Signed-off-by: Florian Westphal Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'net/ipv6/tcp_ipv6.c') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 2968a33cca7d..39ee8e7fc4bd 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1505,8 +1505,7 @@ process: tcp_segs_in(tcp_sk(sk), skb); ret = 0; if (!sock_owned_by_user(sk)) { - if (!tcp_prequeue(sk, skb)) - ret = tcp_v6_do_rcv(sk, skb); + ret = tcp_v6_do_rcv(sk, skb); } else if (tcp_add_backlog(sk, skb)) { goto discard_and_relse; } -- cgit From 39294c3df2a8dc5ec64513633446511f7e842acd Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Tue, 1 Aug 2017 18:27:28 +0200 Subject: Revert "ipv6: constify inet6_protocol structures" This reverts commit 3a3a4e3054137c5ff5d4d306ec834f6d25d7f95b. inet6_add_protocol and inet6_del_protocol include casts that remove the effect of the const annotation on their parameter, leading to possible runtime crashes. Reported-by: Eric Dumazet Signed-off-by: Julia Lawall Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv6/tcp_ipv6.c') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 39ee8e7fc4bd..ced5dcf37465 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1944,7 +1944,7 @@ struct proto tcpv6_prot = { .diag_destroy = tcp_abort, }; -static const struct inet6_protocol tcpv6_protocol = { +static struct inet6_protocol tcpv6_protocol = { .early_demux = tcp_v6_early_demux, .early_demux_handler = tcp_v6_early_demux, .handler = tcp_v6_rcv, -- cgit From 4297a0ef085729af98adab9131d128c576ed3044 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Mon, 7 Aug 2017 08:44:21 -0700 Subject: net: ipv6: add second dif to inet6 socket lookups Add a second device index, sdif, to inet6 socket lookups. sdif is the index for ingress devices enslaved to an l3mdev. It allows the lookups to consider the enslaved device as well as the L3 domain when searching for a socket. TCP moves the data in the cb. Prior to tcp_v4_rcv (e.g., early demux) the ingress index is obtained from IPCB using inet_sdif and after tcp_v4_rcv tcp_v4_sdif is used. Signed-off-by: David Ahern Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'net/ipv6/tcp_ipv6.c') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index ced5dcf37465..f776ec4ecf6d 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -350,7 +350,7 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, sk = __inet6_lookup_established(net, &tcp_hashinfo, &hdr->daddr, th->dest, &hdr->saddr, ntohs(th->source), - skb->dev->ifindex); + skb->dev->ifindex, inet6_sdif(skb)); if (!sk) { __ICMP6_INC_STATS(net, __in6_dev_get(skb->dev), @@ -918,7 +918,8 @@ static void tcp_v6_send_reset(const struct sock *sk, struct sk_buff *skb) &tcp_hashinfo, NULL, 0, &ipv6h->saddr, th->source, &ipv6h->daddr, - ntohs(th->source), tcp_v6_iif(skb)); + ntohs(th->source), tcp_v6_iif(skb), + tcp_v6_sdif(skb)); if (!sk1) goto out; @@ -1397,6 +1398,7 @@ static void tcp_v6_fill_cb(struct sk_buff *skb, const struct ipv6hdr *hdr, static int tcp_v6_rcv(struct sk_buff *skb) { + int sdif = inet6_sdif(skb); const struct tcphdr *th; const struct ipv6hdr *hdr; bool refcounted; @@ -1430,7 +1432,7 @@ static int tcp_v6_rcv(struct sk_buff *skb) lookup: sk = __inet6_lookup_skb(&tcp_hashinfo, skb, __tcp_hdrlen(th), - th->source, th->dest, inet6_iif(skb), + th->source, th->dest, inet6_iif(skb), sdif, &refcounted); if (!sk) goto no_tcp_socket; @@ -1563,7 +1565,8 @@ do_time_wait: skb, __tcp_hdrlen(th), &ipv6_hdr(skb)->saddr, th->source, &ipv6_hdr(skb)->daddr, - ntohs(th->dest), tcp_v6_iif(skb)); + ntohs(th->dest), tcp_v6_iif(skb), + sdif); if (sk2) { struct inet_timewait_sock *tw = inet_twsk(sk); inet_twsk_deschedule_put(tw); @@ -1610,7 +1613,7 @@ static void tcp_v6_early_demux(struct sk_buff *skb) sk = __inet6_lookup_established(dev_net(skb->dev), &tcp_hashinfo, &hdr->saddr, th->source, &hdr->daddr, ntohs(th->dest), - inet6_iif(skb)); + inet6_iif(skb), inet6_sdif(skb)); if (sk) { skb->sk = sk; skb->destructor = sock_edemux; -- cgit From 98aaa913b4ed250324429f0a9e6d5f77a3b5276c Mon Sep 17 00:00:00 2001 From: Mike Maloney Date: Tue, 22 Aug 2017 17:08:48 -0400 Subject: tcp: Extend SOF_TIMESTAMPING_RX_SOFTWARE to TCP recvmsg When SOF_TIMESTAMPING_RX_SOFTWARE is enabled for tcp sockets, return the timestamp corresponding to the highest sequence number data returned. Previously the skb->tstamp is overwritten when a TCP packet is placed in the out of order queue. While the packet is in the ooo queue, save the timestamp in the TCB_SKB_CB. This space is shared with the gso_* options which are only used on the tx path, and a previously unused 4 byte hole. When skbs are coalesced either in the sk_receive_queue or the out_of_order_queue always choose the timestamp of the appended skb to maintain the invariant of returning the timestamp of the last byte in the recvmsg buffer. Signed-off-by: Mike Maloney Acked-by: Willem de Bruijn Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net/ipv6/tcp_ipv6.c') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index d79a1af3252e..abba3bc2a3d9 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1394,6 +1394,8 @@ static void tcp_v6_fill_cb(struct sk_buff *skb, const struct ipv6hdr *hdr, TCP_SKB_CB(skb)->tcp_tw_isn = 0; TCP_SKB_CB(skb)->ip_dsfield = ipv6_get_dsfield(hdr); TCP_SKB_CB(skb)->sacked = 0; + TCP_SKB_CB(skb)->has_rxtstamp = + skb->tstamp || skb_hwtstamps(skb)->hwtstamp; } static int tcp_v6_rcv(struct sk_buff *skb) -- cgit From a8e3bb347d8c32c25830af55466e08979df4e2be Mon Sep 17 00:00:00 2001 From: David Ahern Date: Mon, 28 Aug 2017 15:14:20 -0700 Subject: net: Add comment that early_demux can change via sysctl Twice patches trying to constify inet{6}_protocol have been reverted: 39294c3df2a8 ("Revert "ipv6: constify inet6_protocol structures"") to revert 3a3a4e3054137 and then 03157937fe0b5 ("Revert "ipv4: make net_protocol const"") to revert aa8db499ea67. Add a comment that the structures can not be const because the early_demux field can change based on a sysctl. Signed-off-by: David Ahern Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'net/ipv6/tcp_ipv6.c') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index abba3bc2a3d9..38f76d8b231e 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1949,6 +1949,9 @@ struct proto tcpv6_prot = { .diag_destroy = tcp_abort, }; +/* thinking of making this const? Don't. + * early_demux can change based on sysctl. + */ static struct inet6_protocol tcpv6_protocol = { .early_demux = tcp_v6_early_demux, .early_demux_handler = tcp_v6_early_demux, -- cgit