From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757546Ab0BLDwY (ORCPT ); Thu, 11 Feb 2010 22:52:24 -0500 Received: from mail-bw0-f219.google.com ([209.85.218.219]:51343 "EHLO mail-bw0-f219.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753340Ab0BLDwW (ORCPT ); Thu, 11 Feb 2010 22:52:22 -0500 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=subject:from:to:cc:in-reply-to:references:content-type:date :message-id:mime-version:x-mailer:content-transfer-encoding; b=lBNnzfYYhCViV6DUDMHaS6pLxAm+yMS8VGeZF+pNMNdhf+kL0c5jl5ZbtN5y4XzBmK dIrNVOfsZfszIGxfmIROi2l3rfiePh/Ms1nI4LFfGyB7SNFLRVkW15t8ESFmeOFoHF5T i/vdmPqtTJ/DUo96B6OVwgkIwczqzXewRJggY= Subject: Re: [net-next PATCH v3 2/3] net: TCP thin linear timeouts From: Eric Dumazet To: Andreas Petlund Cc: "netdev@vger.kernel.org" , Ilpo =?ISO-8859-1?Q?J=E4rvinen?= , Arnd Hannemann , LKML , shemminger@vyatta.com, David Miller , william.allen.simpson@gmail.com, damian@tvk.rwth-aachen.de In-Reply-To: <4B73F30D.6040205@simula.no> References: <4B73F30D.6040205@simula.no> Content-Type: text/plain; charset="UTF-8" Date: Fri, 12 Feb 2010 04:52:13 +0100 Message-ID: <1265946733.2891.10.camel@edumazet-laptop> Mime-Version: 1.0 X-Mailer: Evolution 2.28.1 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Le jeudi 11 février 2010 à 13:07 +0100, Andreas Petlund a écrit : > Major changes: > -Possible to disable mechanisms by socket option > -Socket option value boundary check > > > Signed-off-by: Andreas Petlund > --- > include/linux/sysctl.h | 1 + > include/linux/tcp.h | 3 +++ > include/net/tcp.h | 4 ++++ > net/ipv4/sysctl_net_ipv4.c | 7 +++++++ > net/ipv4/tcp.c | 7 +++++++ > net/ipv4/tcp_timer.c | 19 ++++++++++++++++++- > 6 files changed, 40 insertions(+), 1 deletions(-) > > diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h > index 9f236cd..d840d75 100644 > --- a/include/linux/sysctl.h > +++ b/include/linux/sysctl.h > @@ -425,6 +425,7 @@ enum > NET_TCP_ALLOWED_CONG_CONTROL=123, > NET_TCP_MAX_SSTHRESH=124, > NET_TCP_FRTO_RESPONSE=125, > + NET_TCP_FORCE_THIN_LINEAR_TIMEOUTS=126, > }; > > enum { > diff --git a/include/linux/tcp.h b/include/linux/tcp.h > index 7fee8a4..67da706 100644 > --- a/include/linux/tcp.h > +++ b/include/linux/tcp.h > @@ -103,6 +103,7 @@ enum { > #define TCP_CONGESTION 13 /* Congestion control algorithm */ > #define TCP_MD5SIG 14 /* TCP MD5 Signature (RFC2385) */ > #define TCP_COOKIE_TRANSACTIONS 15 /* TCP Cookie Transactions */ > +#define TCP_THIN_LT 16 /* Use linear timeouts for thin streams*/ > > /* for TCP_INFO socket option */ > #define TCPI_OPT_TIMESTAMPS 1 > @@ -341,6 +342,8 @@ struct tcp_sock { > u16 advmss; /* Advertised MSS */ > u8 frto_counter; /* Number of new acks after RTO */ > u8 nonagle; /* Disable Nagle algorithm? */ > + u8 thin_lt : 1,/* Use linear timeouts for thin streams */ > + thin_undef : 7; > > /* RTT measurement */ > u32 srtt; /* smoothed round trip time << 3 */ > diff --git a/include/net/tcp.h b/include/net/tcp.h > index e5e2056..bc5856a 100644 > --- a/include/net/tcp.h > +++ b/include/net/tcp.h > @@ -196,6 +196,9 @@ extern void tcp_time_wait(struct sock *sk, int state, int timeo); > #define TCP_NAGLE_CORK 2 /* Socket is corked */ > #define TCP_NAGLE_PUSH 4 /* Cork is overridden for already queued data */ > > +/* TCP thin-stream limits */ > +#define TCP_THIN_LT_RETRIES 6 /* After 6 linear retries, do exp. backoff */ > + > extern struct inet_timewait_death_row tcp_death_row; > > /* sysctl variables for tcp */ > @@ -241,6 +244,7 @@ extern int sysctl_tcp_workaround_signed_windows; > extern int sysctl_tcp_slow_start_after_idle; > extern int sysctl_tcp_max_ssthresh; > extern int sysctl_tcp_cookie_size; > +extern int sysctl_tcp_force_thin_linear_timeouts; > > extern atomic_t tcp_memory_allocated; > extern struct percpu_counter tcp_sockets_allocated; > diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c > index 7e3712c..cb2ed35 100644 > --- a/net/ipv4/sysctl_net_ipv4.c > +++ b/net/ipv4/sysctl_net_ipv4.c > @@ -576,6 +576,13 @@ static struct ctl_table ipv4_table[] = { > .proc_handler = proc_dointvec > }, > { > + .procname = "tcp_force_thin_linear_timeouts", > + .data = &sysctl_tcp_force_thin_linear_timeouts, > + .maxlen = sizeof(int), > + .mode = 0644, > + .proc_handler = proc_dointvec > + }, > + { > .procname = "udp_mem", > .data = &sysctl_udp_mem, > .maxlen = sizeof(sysctl_udp_mem), > diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c > index d5d69ea..ce9aeb0 100644 > --- a/net/ipv4/tcp.c > +++ b/net/ipv4/tcp.c > @@ -2229,6 +2229,13 @@ static int do_tcp_setsockopt(struct sock *sk, int level, > } > break; > > + case TCP_THIN_LT: > + if (val < 0 || val > 1) > + err = -EINVAL; > + else > + tp->thin_lt = val; > + break; > + > case TCP_CORK: > /* When set indicates to always queue non-full frames. > * Later the user clears this option and we transmit > diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c > index de7d1bf..a682479 100644 > --- a/net/ipv4/tcp_timer.c > +++ b/net/ipv4/tcp_timer.c > @@ -29,6 +29,7 @@ int sysctl_tcp_keepalive_intvl __read_mostly = TCP_KEEPALIVE_INTVL; > int sysctl_tcp_retries1 __read_mostly = TCP_RETR1; > int sysctl_tcp_retries2 __read_mostly = TCP_RETR2; > int sysctl_tcp_orphan_retries __read_mostly; > +int sysctl_tcp_force_thin_linear_timeouts __read_mostly; > > static void tcp_write_timer(unsigned long); > static void tcp_delack_timer(unsigned long); > @@ -415,7 +416,23 @@ void tcp_retransmit_timer(struct sock *sk) > icsk->icsk_retransmits++; > > out_reset_timer: > - icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX); > + /* If stream is thin, use linear timeouts. Since 'icsk_backoff' is > + * used to reset timer, set to 0. Recalculate 'icsk_rto' as this > + * might be increased if the stream oscillates between thin and thick, > + * thus the old value might already be too high compared to the value > + * set by 'tcp_set_rto' in tcp_input.c which resets the rto without > + * backoff. Limit to TCP_THIN_LT_RETRIES before initiating exponential > + * backoff behaviour to avoid continue hammering linear-timeout > + * retransmissions into a black hole*/ > + if ((tp->thin_lt || sysctl_tcp_force_thin_linear_timeouts) && > + tcp_stream_is_thin(sk) && sk->sk_state == TCP_ESTABLISHED && > + icsk->icsk_retransmits <= TCP_THIN_LT_RETRIES) { > + icsk->icsk_backoff = 0; > + icsk->icsk_rto = min(__tcp_set_rto(tp), TCP_RTO_MAX); > + } else { > + /* Use normal (exponential) backoff */ > + icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX); > + } > inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, icsk->icsk_rto, TCP_RTO_MAX); > if (retransmits_timed_out(sk, sysctl_tcp_retries1 + 1)) > __sk_dst_reset(sk); Hi Anreas Could you include a section in Documentation/networking/ip-sysctl.txt about tcp_force_thin_linear_timeouts setting ? Also, you should provide some documentation to be included in man pages (man 7 tcp), to Michael Kerrisk ( ), both for tcp_force_thin_linear_timeouts and TCP_THIN_LT (Same applies for patch 3/3 and tcp_force_thin_dupack / TCP_THIN_DUPACK )