2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* INET An implementation of the TCP/IP protocol suite for the LINUX
|
|
|
|
* operating system. INET is implemented using the BSD Socket
|
|
|
|
* interface as the means of communication with the user level.
|
|
|
|
*
|
|
|
|
* Definitions for the TCP module.
|
|
|
|
*
|
|
|
|
* Version: @(#)tcp.h 1.0.5 05/23/93
|
|
|
|
*
|
2005-05-06 06:16:16 +07:00
|
|
|
* Authors: Ross Biro
|
2005-04-17 05:20:36 +07:00
|
|
|
* Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public License
|
|
|
|
* as published by the Free Software Foundation; either version
|
|
|
|
* 2 of the License, or (at your option) any later version.
|
|
|
|
*/
|
|
|
|
#ifndef _TCP_H
|
|
|
|
#define _TCP_H
|
|
|
|
|
|
|
|
#define FASTRETRANS_DEBUG 1
|
|
|
|
|
|
|
|
#include <linux/list.h>
|
|
|
|
#include <linux/tcp.h>
|
2011-11-24 08:12:59 +07:00
|
|
|
#include <linux/bug.h>
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/cache.h>
|
|
|
|
#include <linux/percpu.h>
|
2005-11-11 04:01:24 +07:00
|
|
|
#include <linux/skbuff.h>
|
2006-05-24 07:55:33 +07:00
|
|
|
#include <linux/dmaengine.h>
|
2006-11-15 10:07:45 +07:00
|
|
|
#include <linux/crypto.h>
|
2008-02-08 12:49:26 +07:00
|
|
|
#include <linux/cryptohash.h>
|
TCPCT part 1d: define TCP cookie option, extend existing struct's
Data structures are carefully composed to require minimal additions.
For example, the struct tcp_options_received cookie_plus variable fits
between existing 16-bit and 8-bit variables, requiring no additional
space (taking alignment into consideration). There are no additions to
tcp_request_sock, and only 1 pointer in tcp_sock.
This is a significantly revised implementation of an earlier (year-old)
patch that no longer applies cleanly, with permission of the original
author (Adam Langley):
http://thread.gmane.org/gmane.linux.network/102586
The principle difference is using a TCP option to carry the cookie nonce,
instead of a user configured offset in the data. This is more flexible and
less subject to user configuration error. Such a cookie option has been
suggested for many years, and is also useful without SYN data, allowing
several related concepts to use the same extension option.
"Re: SYN floods (was: does history repeat itself?)", September 9, 1996.
http://www.merit.net/mail.archives/nanog/1996-09/msg00235.html
"Re: what a new TCP header might look like", May 12, 1998.
ftp://ftp.isi.edu/end2end/end2end-interest-1998.mail
These functions will also be used in subsequent patches that implement
additional features.
Requires:
TCPCT part 1a: add request_values parameter for sending SYNACK
TCPCT part 1b: generate Responder Cookie secret
TCPCT part 1c: sysctl_tcp_cookie_size, socket option TCP_COOKIE_TRANSACTIONS
Signed-off-by: William.Allen.Simpson@gmail.com
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-12-03 01:17:05 +07:00
|
|
|
#include <linux/kref.h>
|
2005-08-10 10:11:08 +07:00
|
|
|
|
|
|
|
#include <net/inet_connection_sock.h>
|
2005-08-10 10:44:40 +07:00
|
|
|
#include <net/inet_timewait_sock.h>
|
2005-08-10 10:00:51 +07:00
|
|
|
#include <net/inet_hashtables.h>
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <net/checksum.h>
|
[NET] Generalise TCP's struct open_request minisock infrastructure
Kept this first changeset minimal, without changing existing names to
ease peer review.
Basicaly tcp_openreq_alloc now receives the or_calltable, that in turn
has two new members:
->slab, that replaces tcp_openreq_cachep
->obj_size, to inform the size of the openreq descendant for
a specific protocol
The protocol specific fields in struct open_request were moved to a
class hierarchy, with the things that are common to all connection
oriented PF_INET protocols in struct inet_request_sock, the TCP ones
in tcp_request_sock, that is an inet_request_sock, that is an
open_request.
I.e. this uses the same approach used for the struct sock class
hierarchy, with sk_prot indicating if the protocol wants to use the
open_request infrastructure by filling in sk_prot->rsk_prot with an
or_calltable.
Results? Performance is improved and TCP v4 now uses only 64 bytes per
open request minisock, down from 96 without this patch :-)
Next changeset will rename some of the structs, fields and functions
mentioned above, struct or_calltable is way unclear, better name it
struct request_sock_ops, s/struct open_request/struct request_sock/g,
etc.
Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 12:46:52 +07:00
|
|
|
#include <net/request_sock.h>
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <net/sock.h>
|
|
|
|
#include <net/snmp.h>
|
|
|
|
#include <net/ip.h>
|
2005-08-10 10:08:28 +07:00
|
|
|
#include <net/tcp_states.h>
|
2007-05-27 16:04:16 +07:00
|
|
|
#include <net/inet_ecn.h>
|
2009-05-05 01:11:01 +07:00
|
|
|
#include <net/dst.h>
|
2005-08-10 10:08:28 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <linux/seq_file.h>
|
2011-12-12 04:47:02 +07:00
|
|
|
#include <linux/memcontrol.h>
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2005-08-10 10:07:35 +07:00
|
|
|
extern struct inet_hashinfo tcp_hashinfo;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2008-11-26 12:17:14 +07:00
|
|
|
extern struct percpu_counter tcp_orphan_count;
|
2005-04-17 05:20:36 +07:00
|
|
|
extern void tcp_time_wait(struct sock *sk, int state, int timeo);
|
|
|
|
|
|
|
|
#define MAX_TCP_HEADER (128 + MAX_HEADER)
|
2008-07-19 14:04:31 +07:00
|
|
|
#define MAX_TCP_OPTION_SPACE 40
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Never offer a window over 32767 without using window scaling. Some
|
|
|
|
* poor stacks do signed 16bit maths!
|
|
|
|
*/
|
|
|
|
#define MAX_TCP_WINDOW 32767U
|
|
|
|
|
|
|
|
/* Minimal accepted MSS. It is (60+60+8) - (20+20). */
|
|
|
|
#define TCP_MIN_MSS 88U
|
|
|
|
|
2006-03-21 08:53:41 +07:00
|
|
|
/* The least MTU to use for probing */
|
|
|
|
#define TCP_BASE_MSS 512
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* After receiving this amount of duplicate ACKs fast retransmit starts. */
|
|
|
|
#define TCP_FASTRETRANS_THRESH 3
|
|
|
|
|
|
|
|
/* Maximal reordering. */
|
|
|
|
#define TCP_MAX_REORDERING 127
|
|
|
|
|
|
|
|
/* Maximal number of ACKs sent quickly to accelerate slow-start. */
|
|
|
|
#define TCP_MAX_QUICKACKS 16U
|
|
|
|
|
|
|
|
/* urg_data states */
|
|
|
|
#define TCP_URG_VALID 0x0100
|
|
|
|
#define TCP_URG_NOTYET 0x0200
|
|
|
|
#define TCP_URG_READ 0x0400
|
|
|
|
|
|
|
|
#define TCP_RETR1 3 /*
|
|
|
|
* This is how many retries it does before it
|
|
|
|
* tries to figure out if the gateway is
|
|
|
|
* down. Minimal RFC value is 3; it corresponds
|
|
|
|
* to ~3sec-8min depending on RTO.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define TCP_RETR2 15 /*
|
|
|
|
* This should take at least
|
|
|
|
* 90 minutes to time out.
|
|
|
|
* RFC1122 says that the limit is 100 sec.
|
|
|
|
* 15 is ~13-30min depending on RTO.
|
|
|
|
*/
|
|
|
|
|
2012-08-31 09:48:31 +07:00
|
|
|
#define TCP_SYN_RETRIES 6 /* This is how many retries are done
|
|
|
|
* when active opening a connection.
|
|
|
|
* RFC1122 says the minimum retry MUST
|
|
|
|
* be at least 180secs. Nevertheless
|
|
|
|
* this value is corresponding to
|
|
|
|
* 63secs of retransmission with the
|
|
|
|
* current initial RTO.
|
|
|
|
*/
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2012-08-31 09:48:31 +07:00
|
|
|
#define TCP_SYNACK_RETRIES 5 /* This is how may retries are done
|
|
|
|
* when passive opening a connection.
|
|
|
|
* This is corresponding to 31secs of
|
|
|
|
* retransmission with the current
|
|
|
|
* initial RTO.
|
|
|
|
*/
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
#define TCP_TIMEWAIT_LEN (60*HZ) /* how long to wait to destroy TIME-WAIT
|
|
|
|
* state, about 60 seconds */
|
|
|
|
#define TCP_FIN_TIMEOUT TCP_TIMEWAIT_LEN
|
|
|
|
/* BSD style FIN_WAIT2 deadlock breaker.
|
|
|
|
* It used to be 3min, new value is 60sec,
|
|
|
|
* to combine FIN-WAIT-2 timeout with
|
|
|
|
* TIME-WAIT timer.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define TCP_DELACK_MAX ((unsigned)(HZ/5)) /* maximal time to delay before sending an ACK */
|
|
|
|
#if HZ >= 100
|
|
|
|
#define TCP_DELACK_MIN ((unsigned)(HZ/25)) /* minimal time to delay before sending an ACK */
|
|
|
|
#define TCP_ATO_MIN ((unsigned)(HZ/25))
|
|
|
|
#else
|
|
|
|
#define TCP_DELACK_MIN 4U
|
|
|
|
#define TCP_ATO_MIN 4U
|
|
|
|
#endif
|
|
|
|
#define TCP_RTO_MAX ((unsigned)(120*HZ))
|
|
|
|
#define TCP_RTO_MIN ((unsigned)(HZ/5))
|
2012-04-13 02:48:40 +07:00
|
|
|
#define TCP_TIMEOUT_INIT ((unsigned)(1*HZ)) /* RFC6298 2.1 initial RTO value */
|
2011-06-08 18:08:38 +07:00
|
|
|
#define TCP_TIMEOUT_FALLBACK ((unsigned)(3*HZ)) /* RFC 1122 initial RTO value, now
|
|
|
|
* used as a fallback RTO for the
|
|
|
|
* initial data transmission if no
|
|
|
|
* valid RTT sample has been acquired,
|
|
|
|
* most likely due to retrans in 3WHS.
|
|
|
|
*/
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
#define TCP_RESOURCE_PROBE_INTERVAL ((unsigned)(HZ/2U)) /* Maximal interval between probes
|
|
|
|
* for local resources.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define TCP_KEEPALIVE_TIME (120*60*HZ) /* two hours */
|
|
|
|
#define TCP_KEEPALIVE_PROBES 9 /* Max of 9 keepalive probes */
|
|
|
|
#define TCP_KEEPALIVE_INTVL (75*HZ)
|
|
|
|
|
|
|
|
#define MAX_TCP_KEEPIDLE 32767
|
|
|
|
#define MAX_TCP_KEEPINTVL 32767
|
|
|
|
#define MAX_TCP_KEEPCNT 127
|
|
|
|
#define MAX_TCP_SYNCNT 127
|
|
|
|
|
|
|
|
#define TCP_SYNQ_INTERVAL (HZ/5) /* Period of SYNACK timer */
|
|
|
|
|
|
|
|
#define TCP_PAWS_24DAYS (60 * 60 * 24 * 24)
|
|
|
|
#define TCP_PAWS_MSL 60 /* Per-host timestamps are invalidated
|
|
|
|
* after this time. It should be equal
|
|
|
|
* (or greater than) TCP_TIMEWAIT_LEN
|
|
|
|
* to provide reliability equal to one
|
|
|
|
* provided by timewait state.
|
|
|
|
*/
|
|
|
|
#define TCP_PAWS_WINDOW 1 /* Replay window for per-host
|
|
|
|
* timestamps. It must be less than
|
|
|
|
* minimal timewait lifetime.
|
|
|
|
*/
|
|
|
|
/*
|
|
|
|
* TCP option
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define TCPOPT_NOP 1 /* Padding */
|
|
|
|
#define TCPOPT_EOL 0 /* End of options */
|
|
|
|
#define TCPOPT_MSS 2 /* Segment size negotiating */
|
|
|
|
#define TCPOPT_WINDOW 3 /* Window scaling */
|
|
|
|
#define TCPOPT_SACK_PERM 4 /* SACK Permitted */
|
|
|
|
#define TCPOPT_SACK 5 /* SACK Block */
|
|
|
|
#define TCPOPT_TIMESTAMP 8 /* Better RTT estimations/PAWS */
|
2006-11-15 10:07:45 +07:00
|
|
|
#define TCPOPT_MD5SIG 19 /* MD5 Signature (RFC2385) */
|
2012-07-19 13:43:05 +07:00
|
|
|
#define TCPOPT_EXP 254 /* Experimental */
|
|
|
|
/* Magic number to be after the option value for sharing TCP
|
|
|
|
* experimental options. See draft-ietf-tcpm-experimental-options-00.txt
|
|
|
|
*/
|
|
|
|
#define TCPOPT_FASTOPEN_MAGIC 0xF989
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* TCP option lengths
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define TCPOLEN_MSS 4
|
|
|
|
#define TCPOLEN_WINDOW 3
|
|
|
|
#define TCPOLEN_SACK_PERM 2
|
|
|
|
#define TCPOLEN_TIMESTAMP 10
|
2006-11-15 10:07:45 +07:00
|
|
|
#define TCPOLEN_MD5SIG 18
|
2012-07-19 13:43:05 +07:00
|
|
|
#define TCPOLEN_EXP_FASTOPEN_BASE 4
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/* But this is what stacks really send out. */
|
|
|
|
#define TCPOLEN_TSTAMP_ALIGNED 12
|
|
|
|
#define TCPOLEN_WSCALE_ALIGNED 4
|
|
|
|
#define TCPOLEN_SACKPERM_ALIGNED 4
|
|
|
|
#define TCPOLEN_SACK_BASE 2
|
|
|
|
#define TCPOLEN_SACK_BASE_ALIGNED 4
|
|
|
|
#define TCPOLEN_SACK_PERBLOCK 8
|
2006-11-15 10:07:45 +07:00
|
|
|
#define TCPOLEN_MD5SIG_ALIGNED 20
|
2008-07-19 14:04:31 +07:00
|
|
|
#define TCPOLEN_MSS_ALIGNED 4
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/* Flags in tp->nonagle */
|
|
|
|
#define TCP_NAGLE_OFF 1 /* Nagle's algo is disabled */
|
|
|
|
#define TCP_NAGLE_CORK 2 /* Socket is corked */
|
2005-11-11 08:13:47 +07:00
|
|
|
#define TCP_NAGLE_PUSH 4 /* Cork is overridden for already queued data */
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2010-02-18 09:47:01 +07:00
|
|
|
/* TCP thin-stream limits */
|
|
|
|
#define TCP_THIN_LINEAR_RETRIES 6 /* After 6 linear retries, do exp. backoff */
|
|
|
|
|
2011-02-06 09:13:45 +07:00
|
|
|
/* TCP initial congestion window as per draft-hkchu-tcpm-initcwnd-01 */
|
2011-02-03 08:05:11 +07:00
|
|
|
#define TCP_INIT_CWND 10
|
|
|
|
|
2012-07-19 13:43:09 +07:00
|
|
|
/* Bit Flags for sysctl_tcp_fastopen */
|
|
|
|
#define TFO_CLIENT_ENABLE 1
|
2012-08-31 19:29:11 +07:00
|
|
|
#define TFO_SERVER_ENABLE 2
|
2012-07-19 13:43:11 +07:00
|
|
|
#define TFO_CLIENT_NO_COOKIE 4 /* Data in SYN w/o cookie option */
|
2012-07-19 13:43:09 +07:00
|
|
|
|
2012-08-31 19:29:11 +07:00
|
|
|
/* Process SYN data but skip cookie validation */
|
|
|
|
#define TFO_SERVER_COOKIE_NOT_CHKED 0x100
|
|
|
|
/* Accept SYN data w/o any cookie option */
|
|
|
|
#define TFO_SERVER_COOKIE_NOT_REQD 0x200
|
|
|
|
|
|
|
|
/* Force enable TFO on all listeners, i.e., not requiring the
|
|
|
|
* TCP_FASTOPEN socket option. SOCKOPT1/2 determine how to set max_qlen.
|
|
|
|
*/
|
|
|
|
#define TFO_SERVER_WO_SOCKOPT1 0x400
|
|
|
|
#define TFO_SERVER_WO_SOCKOPT2 0x800
|
|
|
|
/* Always create TFO child sockets on a TFO listener even when
|
|
|
|
* cookie/data not present. (For testing purpose!)
|
|
|
|
*/
|
|
|
|
#define TFO_SERVER_ALWAYS 0x1000
|
|
|
|
|
2005-08-10 10:44:40 +07:00
|
|
|
extern struct inet_timewait_death_row tcp_death_row;
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* sysctl variables for tcp */
|
|
|
|
extern int sysctl_tcp_timestamps;
|
|
|
|
extern int sysctl_tcp_window_scaling;
|
|
|
|
extern int sysctl_tcp_sack;
|
|
|
|
extern int sysctl_tcp_fin_timeout;
|
|
|
|
extern int sysctl_tcp_keepalive_time;
|
|
|
|
extern int sysctl_tcp_keepalive_probes;
|
|
|
|
extern int sysctl_tcp_keepalive_intvl;
|
|
|
|
extern int sysctl_tcp_syn_retries;
|
|
|
|
extern int sysctl_tcp_synack_retries;
|
|
|
|
extern int sysctl_tcp_retries1;
|
|
|
|
extern int sysctl_tcp_retries2;
|
|
|
|
extern int sysctl_tcp_orphan_retries;
|
|
|
|
extern int sysctl_tcp_syncookies;
|
2012-07-19 13:43:05 +07:00
|
|
|
extern int sysctl_tcp_fastopen;
|
2005-04-17 05:20:36 +07:00
|
|
|
extern int sysctl_tcp_retrans_collapse;
|
|
|
|
extern int sysctl_tcp_stdurg;
|
|
|
|
extern int sysctl_tcp_rfc1337;
|
|
|
|
extern int sysctl_tcp_abort_on_overflow;
|
|
|
|
extern int sysctl_tcp_max_orphans;
|
|
|
|
extern int sysctl_tcp_fack;
|
|
|
|
extern int sysctl_tcp_reordering;
|
|
|
|
extern int sysctl_tcp_dsack;
|
|
|
|
extern int sysctl_tcp_wmem[3];
|
|
|
|
extern int sysctl_tcp_rmem[3];
|
|
|
|
extern int sysctl_tcp_app_win;
|
|
|
|
extern int sysctl_tcp_adv_win_scale;
|
|
|
|
extern int sysctl_tcp_tw_reuse;
|
|
|
|
extern int sysctl_tcp_frto;
|
|
|
|
extern int sysctl_tcp_low_latency;
|
2006-05-24 08:02:55 +07:00
|
|
|
extern int sysctl_tcp_dma_copybreak;
|
2005-04-17 05:20:36 +07:00
|
|
|
extern int sysctl_tcp_nometrics_save;
|
|
|
|
extern int sysctl_tcp_moderate_rcvbuf;
|
|
|
|
extern int sysctl_tcp_tso_win_divisor;
|
2006-03-21 08:53:41 +07:00
|
|
|
extern int sysctl_tcp_mtu_probing;
|
|
|
|
extern int sysctl_tcp_base_mss;
|
2006-03-21 13:40:29 +07:00
|
|
|
extern int sysctl_tcp_workaround_signed_windows;
|
2006-06-14 12:33:04 +07:00
|
|
|
extern int sysctl_tcp_slow_start_after_idle;
|
2007-03-26 09:21:45 +07:00
|
|
|
extern int sysctl_tcp_max_ssthresh;
|
2010-02-18 09:47:01 +07:00
|
|
|
extern int sysctl_tcp_thin_linear_timeouts;
|
2010-02-18 11:48:19 +07:00
|
|
|
extern int sysctl_tcp_thin_dupack;
|
2012-05-02 20:30:03 +07:00
|
|
|
extern int sysctl_tcp_early_retrans;
|
tcp: TCP Small Queues
This introduce TSQ (TCP Small Queues)
TSQ goal is to reduce number of TCP packets in xmit queues (qdisc &
device queues), to reduce RTT and cwnd bias, part of the bufferbloat
problem.
sk->sk_wmem_alloc not allowed to grow above a given limit,
allowing no more than ~128KB [1] per tcp socket in qdisc/dev layers at a
given time.
TSO packets are sized/capped to half the limit, so that we have two
TSO packets in flight, allowing better bandwidth use.
As a side effect, setting the limit to 40000 automatically reduces the
standard gso max limit (65536) to 40000/2 : It can help to reduce
latencies of high prio packets, having smaller TSO packets.
This means we divert sock_wfree() to a tcp_wfree() handler, to
queue/send following frames when skb_orphan() [2] is called for the
already queued skbs.
Results on my dev machines (tg3/ixgbe nics) are really impressive,
using standard pfifo_fast, and with or without TSO/GSO.
Without reduction of nominal bandwidth, we have reduction of buffering
per bulk sender :
< 1ms on Gbit (instead of 50ms with TSO)
< 8ms on 100Mbit (instead of 132 ms)
I no longer have 4 MBytes backlogged in qdisc by a single netperf
session, and both side socket autotuning no longer use 4 Mbytes.
As skb destructor cannot restart xmit itself ( as qdisc lock might be
taken at this point ), we delegate the work to a tasklet. We use one
tasklest per cpu for performance reasons.
If tasklet finds a socket owned by the user, it sets TSQ_OWNED flag.
This flag is tested in a new protocol method called from release_sock(),
to eventually send new segments.
[1] New /proc/sys/net/ipv4/tcp_limit_output_bytes tunable
[2] skb_orphan() is usually called at TX completion time,
but some drivers call it in their start_xmit() handler.
These drivers should at least use BQL, or else a single TCP
session can still fill the whole NIC TX ring, since TSQ will
have no effect.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Dave Taht <dave.taht@bufferbloat.net>
Cc: Tom Herbert <therbert@google.com>
Cc: Matt Mathis <mattmathis@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-07-11 12:50:31 +07:00
|
|
|
extern int sysctl_tcp_limit_output_bytes;
|
2012-07-17 15:13:05 +07:00
|
|
|
extern int sysctl_tcp_challenge_ack_limit;
|
tcp: TCP_NOTSENT_LOWAT socket option
Idea of this patch is to add optional limitation of number of
unsent bytes in TCP sockets, to reduce usage of kernel memory.
TCP receiver might announce a big window, and TCP sender autotuning
might allow a large amount of bytes in write queue, but this has little
performance impact if a large part of this buffering is wasted :
Write queue needs to be large only to deal with large BDP, not
necessarily to cope with scheduling delays (incoming ACKS make room
for the application to queue more bytes)
For most workloads, using a value of 128 KB or less is OK to give
applications enough time to react to POLLOUT events in time
(or being awaken in a blocking sendmsg())
This patch adds two ways to set the limit :
1) Per socket option TCP_NOTSENT_LOWAT
2) A sysctl (/proc/sys/net/ipv4/tcp_notsent_lowat) for sockets
not using TCP_NOTSENT_LOWAT socket option (or setting a zero value)
Default value being UINT_MAX (0xFFFFFFFF), meaning this has no effect.
This changes poll()/select()/epoll() to report POLLOUT
only if number of unsent bytes is below tp->nosent_lowat
Note this might increase number of sendmsg()/sendfile() calls
when using non blocking sockets,
and increase number of context switches for blocking sockets.
Note this is not related to SO_SNDLOWAT (as SO_SNDLOWAT is
defined as :
Specify the minimum number of bytes in the buffer until
the socket layer will pass the data to the protocol)
Tested:
netperf sessions, and watching /proc/net/protocols "memory" column for TCP
With 200 concurrent netperf -t TCP_STREAM sessions, amount of kernel memory
used by TCP buffers shrinks by ~55 % (20567 pages instead of 45458)
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 45458 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 45458 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 20567 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 20567 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
Using 128KB has no bad effect on the throughput or cpu usage
of a single flow, although there is an increase of context switches.
A bonus is that we hold socket lock for a shorter amount
of time and should improve latencies of ACK processing.
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1651584 6291456 16384 20.00 17447.90 10^6bits/s 3.13 S -1.00 U 0.353 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
412,514 context-switches
200.034645535 seconds time elapsed
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1593240 6291456 16384 20.00 17321.16 10^6bits/s 3.35 S -1.00 U 0.381 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
2,675,818 context-switches
200.029651391 seconds time elapsed
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-By: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-07-23 10:27:07 +07:00
|
|
|
extern unsigned int sysctl_tcp_notsent_lowat;
|
tcp: TSO packets automatic sizing
After hearing many people over past years complaining against TSO being
bursty or even buggy, we are proud to present automatic sizing of TSO
packets.
One part of the problem is that tcp_tso_should_defer() uses an heuristic
relying on upcoming ACKS instead of a timer, but more generally, having
big TSO packets makes little sense for low rates, as it tends to create
micro bursts on the network, and general consensus is to reduce the
buffering amount.
This patch introduces a per socket sk_pacing_rate, that approximates
the current sending rate, and allows us to size the TSO packets so
that we try to send one packet every ms.
This field could be set by other transports.
Patch has no impact for high speed flows, where having large TSO packets
makes sense to reach line rate.
For other flows, this helps better packet scheduling and ACK clocking.
This patch increases performance of TCP flows in lossy environments.
A new sysctl (tcp_min_tso_segs) is added, to specify the
minimal size of a TSO packet (default being 2).
A follow-up patch will provide a new packet scheduler (FQ), using
sk_pacing_rate as an input to perform optional per flow pacing.
This explains why we chose to set sk_pacing_rate to twice the current
rate, allowing 'slow start' ramp up.
sk_pacing_rate = 2 * cwnd * mss / srtt
v2: Neal Cardwell reported a suspect deferring of last two segments on
initial write of 10 MSS, I had to change tcp_tso_should_defer() to take
into account tp->xmit_size_goal_segs
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Van Jacobson <vanj@google.com>
Cc: Tom Herbert <therbert@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-27 19:46:32 +07:00
|
|
|
extern int sysctl_tcp_min_tso_segs;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2010-11-10 06:24:26 +07:00
|
|
|
extern atomic_long_t tcp_memory_allocated;
|
2008-11-26 12:16:35 +07:00
|
|
|
extern struct percpu_counter tcp_sockets_allocated;
|
2005-04-17 05:20:36 +07:00
|
|
|
extern int tcp_memory_pressure;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The next routines deal with comparing 32 bit unsigned ints
|
|
|
|
* and worry about wraparound (automatic with unsigned arithmetic).
|
|
|
|
*/
|
|
|
|
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool before(__u32 seq1, __u32 seq2)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2007-01-05 03:25:16 +07:00
|
|
|
return (__s32)(seq1-seq2) < 0;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
[TCP]: Fix ambiguity in the `before' relation.
While looking at DCCP sequence numbers, I stumbled over a problem with
the following definition of before in tcp.h:
static inline int before(__u32 seq1, __u32 seq2)
{
return (__s32)(seq1-seq2) < 0;
}
Problem: This definition suffers from an an ambiguity, i.e. always
before(a, (a + 2^31) % 2^32)) = 1
before((a + 2^31) % 2^32), a) = 1
In text: when the difference between a and b amounts to 2^31,
a is always considered `before' b, the function can not decide.
The reason is that implicitly 0 is `before' 1 ... 2^31-1 ... 2^31
Solution: There is a simple fix, by defining before in such a way that
0 is no longer `before' 2^31, i.e. 0 `before' 1 ... 2^31-1
By not using the middle between 0 and 2^32, before can be made
unambiguous.
This is achieved by testing whether seq2-seq1 > 0 (using signed
32-bit arithmetic).
I attach a patch to codify this. Also the `after' relation is basically
a redefinition of `before', it is now defined as a macro after before.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-12-21 01:25:55 +07:00
|
|
|
#define after(seq2, seq1) before(seq1, seq2)
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/* is s2<=s1<=s3 ? */
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool between(__u32 seq1, __u32 seq2, __u32 seq3)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
return seq3 - seq2 >= seq1 - seq2;
|
|
|
|
}
|
|
|
|
|
2012-01-31 05:16:06 +07:00
|
|
|
static inline bool tcp_out_of_memory(struct sock *sk)
|
|
|
|
{
|
|
|
|
if (sk->sk_wmem_queued > SOCK_MIN_SNDBUF &&
|
|
|
|
sk_memory_allocated(sk) > sk_prot_mem_limits(sk, 2))
|
|
|
|
return true;
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2010-08-25 16:27:49 +07:00
|
|
|
static inline bool tcp_too_many_orphans(struct sock *sk, int shift)
|
2007-05-30 03:19:18 +07:00
|
|
|
{
|
2010-08-25 16:27:49 +07:00
|
|
|
struct percpu_counter *ocp = sk->sk_prot->orphan_count;
|
|
|
|
int orphans = percpu_counter_read_positive(ocp);
|
|
|
|
|
|
|
|
if (orphans << shift > sysctl_tcp_max_orphans) {
|
|
|
|
orphans = percpu_counter_sum_positive(ocp);
|
|
|
|
if (orphans << shift > sysctl_tcp_max_orphans)
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
return false;
|
2007-05-30 03:19:18 +07:00
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2012-01-31 05:16:06 +07:00
|
|
|
extern bool tcp_check_oom(struct sock *sk, int shift);
|
|
|
|
|
2009-04-19 16:43:48 +07:00
|
|
|
/* syncookies: remember time of last synqueue overflow */
|
|
|
|
static inline void tcp_synq_overflow(struct sock *sk)
|
|
|
|
{
|
|
|
|
tcp_sk(sk)->rx_opt.ts_recent_stamp = jiffies;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* syncookies: no recent synqueue overflow on this listening socket? */
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_synq_no_recent_overflow(const struct sock *sk)
|
2009-04-19 16:43:48 +07:00
|
|
|
{
|
|
|
|
unsigned long last_overflow = tcp_sk(sk)->rx_opt.ts_recent_stamp;
|
2011-06-08 18:08:38 +07:00
|
|
|
return time_after(jiffies, last_overflow + TCP_TIMEOUT_FALLBACK);
|
2009-04-19 16:43:48 +07:00
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
extern struct proto tcp_prot;
|
|
|
|
|
2008-07-18 18:02:08 +07:00
|
|
|
#define TCP_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.tcp_statistics, field)
|
|
|
|
#define TCP_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.tcp_statistics, field)
|
|
|
|
#define TCP_DEC_STATS(net, field) SNMP_DEC_STATS((net)->mib.tcp_statistics, field)
|
|
|
|
#define TCP_ADD_STATS_USER(net, field, val) SNMP_ADD_STATS_USER((net)->mib.tcp_statistics, field, val)
|
2010-04-22 14:00:24 +07:00
|
|
|
#define TCP_ADD_STATS(net, field, val) SNMP_ADD_STATS((net)->mib.tcp_statistics, field, val)
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2012-01-30 08:20:17 +07:00
|
|
|
extern void tcp_init_mem(struct net *net);
|
|
|
|
|
tcp: TCP Small Queues
This introduce TSQ (TCP Small Queues)
TSQ goal is to reduce number of TCP packets in xmit queues (qdisc &
device queues), to reduce RTT and cwnd bias, part of the bufferbloat
problem.
sk->sk_wmem_alloc not allowed to grow above a given limit,
allowing no more than ~128KB [1] per tcp socket in qdisc/dev layers at a
given time.
TSO packets are sized/capped to half the limit, so that we have two
TSO packets in flight, allowing better bandwidth use.
As a side effect, setting the limit to 40000 automatically reduces the
standard gso max limit (65536) to 40000/2 : It can help to reduce
latencies of high prio packets, having smaller TSO packets.
This means we divert sock_wfree() to a tcp_wfree() handler, to
queue/send following frames when skb_orphan() [2] is called for the
already queued skbs.
Results on my dev machines (tg3/ixgbe nics) are really impressive,
using standard pfifo_fast, and with or without TSO/GSO.
Without reduction of nominal bandwidth, we have reduction of buffering
per bulk sender :
< 1ms on Gbit (instead of 50ms with TSO)
< 8ms on 100Mbit (instead of 132 ms)
I no longer have 4 MBytes backlogged in qdisc by a single netperf
session, and both side socket autotuning no longer use 4 Mbytes.
As skb destructor cannot restart xmit itself ( as qdisc lock might be
taken at this point ), we delegate the work to a tasklet. We use one
tasklest per cpu for performance reasons.
If tasklet finds a socket owned by the user, it sets TSQ_OWNED flag.
This flag is tested in a new protocol method called from release_sock(),
to eventually send new segments.
[1] New /proc/sys/net/ipv4/tcp_limit_output_bytes tunable
[2] skb_orphan() is usually called at TX completion time,
but some drivers call it in their start_xmit() handler.
These drivers should at least use BQL, or else a single TCP
session can still fill the whole NIC TX ring, since TSQ will
have no effect.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Dave Taht <dave.taht@bufferbloat.net>
Cc: Tom Herbert <therbert@google.com>
Cc: Matt Mathis <mattmathis@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-07-11 12:50:31 +07:00
|
|
|
extern void tcp_tasklet_init(void);
|
|
|
|
|
2010-07-11 03:41:06 +07:00
|
|
|
extern void tcp_v4_err(struct sk_buff *skb, u32);
|
|
|
|
|
|
|
|
extern void tcp_shutdown (struct sock *sk, int how);
|
|
|
|
|
2012-06-28 12:01:22 +07:00
|
|
|
extern void tcp_v4_early_demux(struct sk_buff *skb);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp_v4_rcv(struct sk_buff *skb);
|
|
|
|
|
|
|
|
extern int tcp_v4_tw_remember_stamp(struct inet_timewait_sock *tw);
|
2010-07-11 03:41:55 +07:00
|
|
|
extern int tcp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
|
|
|
|
size_t size);
|
|
|
|
extern int tcp_sendpage(struct sock *sk, struct page *page, int offset,
|
|
|
|
size_t size, int flags);
|
tcp: TCP Small Queues
This introduce TSQ (TCP Small Queues)
TSQ goal is to reduce number of TCP packets in xmit queues (qdisc &
device queues), to reduce RTT and cwnd bias, part of the bufferbloat
problem.
sk->sk_wmem_alloc not allowed to grow above a given limit,
allowing no more than ~128KB [1] per tcp socket in qdisc/dev layers at a
given time.
TSO packets are sized/capped to half the limit, so that we have two
TSO packets in flight, allowing better bandwidth use.
As a side effect, setting the limit to 40000 automatically reduces the
standard gso max limit (65536) to 40000/2 : It can help to reduce
latencies of high prio packets, having smaller TSO packets.
This means we divert sock_wfree() to a tcp_wfree() handler, to
queue/send following frames when skb_orphan() [2] is called for the
already queued skbs.
Results on my dev machines (tg3/ixgbe nics) are really impressive,
using standard pfifo_fast, and with or without TSO/GSO.
Without reduction of nominal bandwidth, we have reduction of buffering
per bulk sender :
< 1ms on Gbit (instead of 50ms with TSO)
< 8ms on 100Mbit (instead of 132 ms)
I no longer have 4 MBytes backlogged in qdisc by a single netperf
session, and both side socket autotuning no longer use 4 Mbytes.
As skb destructor cannot restart xmit itself ( as qdisc lock might be
taken at this point ), we delegate the work to a tasklet. We use one
tasklest per cpu for performance reasons.
If tasklet finds a socket owned by the user, it sets TSQ_OWNED flag.
This flag is tested in a new protocol method called from release_sock(),
to eventually send new segments.
[1] New /proc/sys/net/ipv4/tcp_limit_output_bytes tunable
[2] skb_orphan() is usually called at TX completion time,
but some drivers call it in their start_xmit() handler.
These drivers should at least use BQL, or else a single TCP
session can still fill the whole NIC TX ring, since TSQ will
have no effect.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Dave Taht <dave.taht@bufferbloat.net>
Cc: Tom Herbert <therbert@google.com>
Cc: Matt Mathis <mattmathis@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-07-11 12:50:31 +07:00
|
|
|
extern void tcp_release_cb(struct sock *sk);
|
2013-04-12 18:31:52 +07:00
|
|
|
extern void tcp_wfree(struct sk_buff *skb);
|
2012-07-20 12:45:50 +07:00
|
|
|
extern void tcp_write_timer_handler(struct sock *sk);
|
|
|
|
extern void tcp_delack_timer_handler(struct sock *sk);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg);
|
|
|
|
extern int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
|
2011-10-21 16:22:42 +07:00
|
|
|
const struct tcphdr *th, unsigned int len);
|
2013-09-04 02:23:22 +07:00
|
|
|
extern void tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
|
|
|
|
const struct tcphdr *th, unsigned int len);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern void tcp_rcv_space_adjust(struct sock *sk);
|
|
|
|
extern void tcp_cleanup_rbuf(struct sock *sk, int copied);
|
|
|
|
extern int tcp_twsk_unique(struct sock *sk, struct sock *sktw, void *twp);
|
|
|
|
extern void tcp_twsk_destructor(struct sock *sk);
|
|
|
|
extern ssize_t tcp_splice_read(struct socket *sk, loff_t *ppos,
|
|
|
|
struct pipe_inode_info *pipe, size_t len,
|
|
|
|
unsigned int flags);
|
2007-11-07 14:30:13 +07:00
|
|
|
|
2005-08-10 10:10:42 +07:00
|
|
|
static inline void tcp_dec_quickack_mode(struct sock *sk,
|
|
|
|
const unsigned int pkts)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-08-10 10:10:42 +07:00
|
|
|
struct inet_connection_sock *icsk = inet_csk(sk);
|
2005-07-06 05:17:45 +07:00
|
|
|
|
2005-08-10 10:10:42 +07:00
|
|
|
if (icsk->icsk_ack.quick) {
|
|
|
|
if (pkts >= icsk->icsk_ack.quick) {
|
|
|
|
icsk->icsk_ack.quick = 0;
|
2005-07-06 05:17:45 +07:00
|
|
|
/* Leaving quickack mode we deflate ATO. */
|
2005-08-10 10:10:42 +07:00
|
|
|
icsk->icsk_ack.ato = TCP_ATO_MIN;
|
2005-07-06 05:17:45 +07:00
|
|
|
} else
|
2005-08-10 10:10:42 +07:00
|
|
|
icsk->icsk_ack.quick -= pkts;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-05-27 16:04:16 +07:00
|
|
|
#define TCP_ECN_OK 1
|
|
|
|
#define TCP_ECN_QUEUE_CWR 2
|
|
|
|
#define TCP_ECN_DEMAND_CWR 4
|
2011-09-23 03:02:19 +07:00
|
|
|
#define TCP_ECN_SEEN 8
|
2007-05-27 16:04:16 +07:00
|
|
|
|
2009-11-03 10:26:03 +07:00
|
|
|
enum tcp_tw_status {
|
2005-04-17 05:20:36 +07:00
|
|
|
TCP_TW_SUCCESS = 0,
|
|
|
|
TCP_TW_RST = 1,
|
|
|
|
TCP_TW_ACK = 2,
|
|
|
|
TCP_TW_SYN = 3
|
|
|
|
};
|
|
|
|
|
|
|
|
|
2010-07-11 03:41:06 +07:00
|
|
|
extern enum tcp_tw_status tcp_timewait_state_process(struct inet_timewait_sock *tw,
|
|
|
|
struct sk_buff *skb,
|
|
|
|
const struct tcphdr *th);
|
|
|
|
extern struct sock * tcp_check_req(struct sock *sk,struct sk_buff *skb,
|
|
|
|
struct request_sock *req,
|
2012-08-31 19:29:12 +07:00
|
|
|
struct request_sock **prev,
|
|
|
|
bool fastopen);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp_child_process(struct sock *parent, struct sock *child,
|
|
|
|
struct sk_buff *skb);
|
|
|
|
extern void tcp_enter_loss(struct sock *sk, int how);
|
|
|
|
extern void tcp_clear_retrans(struct tcp_sock *tp);
|
|
|
|
extern void tcp_update_metrics(struct sock *sk);
|
2012-07-10 06:07:30 +07:00
|
|
|
extern void tcp_init_metrics(struct sock *sk);
|
2012-07-10 14:49:14 +07:00
|
|
|
extern void tcp_metrics_init(void);
|
2012-07-10 17:14:24 +07:00
|
|
|
extern bool tcp_peer_is_proven(struct request_sock *req, struct dst_entry *dst, bool paws_check);
|
|
|
|
extern bool tcp_remember_stamp(struct sock *sk);
|
|
|
|
extern bool tcp_tw_remember_stamp(struct inet_timewait_sock *tw);
|
|
|
|
extern void tcp_fetch_timewait_stamp(struct sock *sk, struct dst_entry *dst);
|
2012-07-10 06:07:30 +07:00
|
|
|
extern void tcp_disable_fack(struct tcp_sock *tp);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern void tcp_close(struct sock *sk, long timeout);
|
2012-04-19 16:55:21 +07:00
|
|
|
extern void tcp_init_sock(struct sock *sk);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern unsigned int tcp_poll(struct file * file, struct socket *sock,
|
|
|
|
struct poll_table_struct *wait);
|
|
|
|
extern int tcp_getsockopt(struct sock *sk, int level, int optname,
|
|
|
|
char __user *optval, int __user *optlen);
|
|
|
|
extern int tcp_setsockopt(struct sock *sk, int level, int optname,
|
|
|
|
char __user *optval, unsigned int optlen);
|
|
|
|
extern int compat_tcp_getsockopt(struct sock *sk, int level, int optname,
|
|
|
|
char __user *optval, int __user *optlen);
|
|
|
|
extern int compat_tcp_setsockopt(struct sock *sk, int level, int optname,
|
|
|
|
char __user *optval, unsigned int optlen);
|
|
|
|
extern void tcp_set_keepalive(struct sock *sk, int val);
|
|
|
|
extern void tcp_syn_ack_timeout(struct sock *sk, struct request_sock *req);
|
|
|
|
extern int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
|
|
|
|
size_t len, int nonblock, int flags, int *addr_len);
|
2011-10-21 16:22:42 +07:00
|
|
|
extern void tcp_parse_options(const struct sk_buff *skb,
|
2013-03-17 15:23:34 +07:00
|
|
|
struct tcp_options_received *opt_rx,
|
2012-07-19 13:43:05 +07:00
|
|
|
int estab, struct tcp_fastopen_cookie *foc);
|
2011-10-21 16:22:42 +07:00
|
|
|
extern const u8 *tcp_parse_md5sig_option(const struct tcphdr *th);
|
2008-04-17 10:29:53 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* TCP v4 functions exported for the inet6 API
|
|
|
|
*/
|
|
|
|
|
2010-07-11 03:41:06 +07:00
|
|
|
extern void tcp_v4_send_check(struct sock *sk, struct sk_buff *skb);
|
|
|
|
extern int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb);
|
|
|
|
extern struct sock * tcp_create_openreq_child(struct sock *sk,
|
|
|
|
struct request_sock *req,
|
2005-04-17 05:20:36 +07:00
|
|
|
struct sk_buff *skb);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern struct sock * tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
|
|
|
|
struct request_sock *req,
|
|
|
|
struct dst_entry *dst);
|
|
|
|
extern int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb);
|
|
|
|
extern int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr,
|
|
|
|
int addr_len);
|
|
|
|
extern int tcp_connect(struct sock *sk);
|
|
|
|
extern struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
|
|
|
|
struct request_sock *req,
|
2012-08-31 19:29:12 +07:00
|
|
|
struct tcp_fastopen_cookie *foc);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp_disconnect(struct sock *sk, int flags);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2012-04-19 10:40:01 +07:00
|
|
|
void tcp_connect_init(struct sock *sk);
|
|
|
|
void tcp_finish_connect(struct sock *sk, struct sk_buff *skb);
|
2012-05-10 08:49:41 +07:00
|
|
|
int tcp_send_rcvq(struct sock *sk, struct msghdr *msg, size_t size);
|
2012-08-09 21:11:00 +07:00
|
|
|
void inet_sk_rx_dst_set(struct sock *sk, const struct sk_buff *skb);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/* From syncookies.c */
|
2008-03-24 12:21:28 +07:00
|
|
|
extern __u32 syncookie_secret[2][16-4+SHA_DIGEST_WORDS];
|
2013-08-27 13:50:13 +07:00
|
|
|
extern int __cookie_v4_check(const struct iphdr *iph, const struct tcphdr *th,
|
|
|
|
u32 cookie);
|
2005-04-17 05:20:36 +07:00
|
|
|
extern struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
|
|
|
|
struct ip_options *opt);
|
2011-09-19 08:02:55 +07:00
|
|
|
#ifdef CONFIG_SYN_COOKIES
|
2013-08-27 13:50:13 +07:00
|
|
|
extern u32 __cookie_v4_init_sequence(const struct iphdr *iph,
|
|
|
|
const struct tcphdr *th, u16 *mssp);
|
2005-04-17 05:20:36 +07:00
|
|
|
extern __u32 cookie_v4_init_sequence(struct sock *sk, struct sk_buff *skb,
|
|
|
|
__u16 *mss);
|
2011-09-19 08:02:55 +07:00
|
|
|
#else
|
|
|
|
static inline __u32 cookie_v4_init_sequence(struct sock *sk,
|
|
|
|
struct sk_buff *skb,
|
|
|
|
__u16 *mss)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2008-04-10 17:12:40 +07:00
|
|
|
extern __u32 cookie_init_timestamp(struct request_sock *req);
|
2013-01-05 23:10:48 +07:00
|
|
|
extern bool cookie_check_timestamp(struct tcp_options_received *opt,
|
|
|
|
struct net *net, bool *ecn_ok);
|
2008-04-10 17:12:40 +07:00
|
|
|
|
2008-02-08 12:49:26 +07:00
|
|
|
/* From net/ipv6/syncookies.c */
|
2013-08-27 13:50:15 +07:00
|
|
|
extern int __cookie_v6_check(const struct ipv6hdr *iph, const struct tcphdr *th,
|
|
|
|
u32 cookie);
|
2008-02-08 12:49:26 +07:00
|
|
|
extern struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb);
|
2011-09-19 08:02:55 +07:00
|
|
|
#ifdef CONFIG_SYN_COOKIES
|
2013-08-27 13:50:15 +07:00
|
|
|
extern u32 __cookie_v6_init_sequence(const struct ipv6hdr *iph,
|
|
|
|
const struct tcphdr *th, u16 *mssp);
|
2011-10-21 16:22:42 +07:00
|
|
|
extern __u32 cookie_v6_init_sequence(struct sock *sk, const struct sk_buff *skb,
|
2008-02-08 12:49:26 +07:00
|
|
|
__u16 *mss);
|
2011-09-19 08:02:55 +07:00
|
|
|
#else
|
|
|
|
static inline __u32 cookie_v6_init_sequence(struct sock *sk,
|
|
|
|
struct sk_buff *skb,
|
|
|
|
__u16 *mss)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif
|
2005-04-17 05:20:36 +07:00
|
|
|
/* tcp_output.c */
|
|
|
|
|
[TCP]: Sed magic converts func(sk, tp, ...) -> func(sk, ...)
This is (mostly) automated change using magic:
sed -e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e 's|struct sock \*sk,[\n\t ]*struct tcp_sock \*tp\([^{]*\n{\n\)|
struct sock \*sk\1\tstruct tcp_sock *tp = tcp_sk(sk);\n|g'
-e 's|struct sock \*sk, struct tcp_sock \*tp|
struct sock \*sk|g' -e 's|sk, tp\([^-]\)|sk\1|g'
Fixed four unused variable (tp) warnings that were introduced.
In addition, manually added newlines after local variables and
tweaked function arguments positioning.
$ gcc --version
gcc (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)
...
$ codiff -fV built-in.o.old built-in.o.new
net/ipv4/route.c:
rt_cache_flush | +14
1 function changed, 14 bytes added
net/ipv4/tcp.c:
tcp_setsockopt | -5
tcp_sendpage | -25
tcp_sendmsg | -16
3 functions changed, 46 bytes removed
net/ipv4/tcp_input.c:
tcp_try_undo_recovery | +3
tcp_try_undo_dsack | +2
tcp_mark_head_lost | -12
tcp_ack | -15
tcp_event_data_recv | -32
tcp_rcv_state_process | -10
tcp_rcv_established | +1
7 functions changed, 6 bytes added, 69 bytes removed, diff: -63
net/ipv4/tcp_output.c:
update_send_head | -9
tcp_transmit_skb | +19
tcp_cwnd_validate | +1
tcp_write_wakeup | -17
__tcp_push_pending_frames | -25
tcp_push_one | -8
tcp_send_fin | -4
7 functions changed, 20 bytes added, 63 bytes removed, diff: -43
built-in.o.new:
18 functions changed, 40 bytes added, 178 bytes removed, diff: -138
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2007-04-21 12:18:02 +07:00
|
|
|
extern void __tcp_push_pending_frames(struct sock *sk, unsigned int cur_mss,
|
|
|
|
int nonagle);
|
2012-05-17 06:15:34 +07:00
|
|
|
extern bool tcp_may_send_now(struct sock *sk);
|
2012-12-06 15:45:32 +07:00
|
|
|
extern int __tcp_retransmit_skb(struct sock *, struct sk_buff *);
|
2005-04-17 05:20:36 +07:00
|
|
|
extern int tcp_retransmit_skb(struct sock *, struct sk_buff *);
|
Revert Backoff [v3]: Revert RTO on ICMP destination unreachable
Here, an ICMP host/network unreachable message, whose payload fits to
TCP's SND.UNA, is taken as an indication that the RTO retransmission has
not been lost due to congestion, but because of a route failure
somewhere along the path.
With true congestion, a router won't trigger such a message and the
patched TCP will operate as standard TCP.
This patch reverts one RTO backoff, if an ICMP host/network unreachable
message, whose payload fits to TCP's SND.UNA, arrives.
Based on the new RTO, the retransmission timer is reset to reflect the
remaining time, or - if the revert clocked out the timer - a retransmission
is sent out immediately.
Backoffs are only reverted, if TCP is in RTO loss recovery, i.e. if
there have been retransmissions and reversible backoffs, already.
Changes from v2:
1) Renaming of skb in tcp_v4_err() moved to another patch.
2) Reintroduced tcp_bound_rto() and __tcp_set_rto().
3) Fixed code comments.
Signed-off-by: Damian Lukowski <damian@tvk.rwth-aachen.de>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-26 07:16:31 +07:00
|
|
|
extern void tcp_retransmit_timer(struct sock *sk);
|
2005-04-17 05:20:36 +07:00
|
|
|
extern void tcp_xmit_retransmit_queue(struct sock *);
|
|
|
|
extern void tcp_simple_retransmit(struct sock *);
|
|
|
|
extern int tcp_trim_head(struct sock *, struct sk_buff *, u32);
|
2005-09-02 12:47:01 +07:00
|
|
|
extern int tcp_fragment(struct sock *, struct sk_buff *, u32, unsigned int);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
extern void tcp_send_probe0(struct sock *);
|
|
|
|
extern void tcp_send_partial(struct sock *);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp_write_wakeup(struct sock *);
|
2005-04-17 05:20:36 +07:00
|
|
|
extern void tcp_send_fin(struct sock *sk);
|
2005-10-07 13:46:04 +07:00
|
|
|
extern void tcp_send_active_reset(struct sock *sk, gfp_t priority);
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp_send_synack(struct sock *);
|
2012-05-17 06:15:34 +07:00
|
|
|
extern bool tcp_syn_flood_action(struct sock *sk,
|
|
|
|
const struct sk_buff *skb,
|
|
|
|
const char *proto);
|
2005-07-06 05:24:38 +07:00
|
|
|
extern void tcp_push_one(struct sock *, unsigned int mss_now);
|
2005-04-17 05:20:36 +07:00
|
|
|
extern void tcp_send_ack(struct sock *sk);
|
|
|
|
extern void tcp_send_delayed_ack(struct sock *sk);
|
tcp: Tail loss probe (TLP)
This patch series implement the Tail loss probe (TLP) algorithm described
in http://tools.ietf.org/html/draft-dukkipati-tcpm-tcp-loss-probe-01. The
first patch implements the basic algorithm.
TLP's goal is to reduce tail latency of short transactions. It achieves
this by converting retransmission timeouts (RTOs) occuring due
to tail losses (losses at end of transactions) into fast recovery.
TLP transmits one packet in two round-trips when a connection is in
Open state and isn't receiving any ACKs. The transmitted packet, aka
loss probe, can be either new or a retransmission. When there is tail
loss, the ACK from a loss probe triggers FACK/early-retransmit based
fast recovery, thus avoiding a costly RTO. In the absence of loss,
there is no change in the connection state.
PTO stands for probe timeout. It is a timer event indicating
that an ACK is overdue and triggers a loss probe packet. The PTO value
is set to max(2*SRTT, 10ms) and is adjusted to account for delayed
ACK timer when there is only one oustanding packet.
TLP Algorithm
On transmission of new data in Open state:
-> packets_out > 1: schedule PTO in max(2*SRTT, 10ms).
-> packets_out == 1: schedule PTO in max(2*RTT, 1.5*RTT + 200ms)
-> PTO = min(PTO, RTO)
Conditions for scheduling PTO:
-> Connection is in Open state.
-> Connection is either cwnd limited or no new data to send.
-> Number of probes per tail loss episode is limited to one.
-> Connection is SACK enabled.
When PTO fires:
new_segment_exists:
-> transmit new segment.
-> packets_out++. cwnd remains same.
no_new_packet:
-> retransmit the last segment.
Its ACK triggers FACK or early retransmit based recovery.
ACK path:
-> rearm RTO at start of ACK processing.
-> reschedule PTO if need be.
In addition, the patch includes a small variation to the Early Retransmit
(ER) algorithm, such that ER and TLP together can in principle recover any
N-degree of tail loss through fast recovery. TLP is controlled by the same
sysctl as ER, tcp_early_retrans sysctl.
tcp_early_retrans==0; disables TLP and ER.
==1; enables RFC5827 ER.
==2; delayed ER.
==3; TLP and delayed ER. [DEFAULT]
==4; TLP only.
The TLP patch series have been extensively tested on Google Web servers.
It is most effective for short Web trasactions, where it reduced RTOs by 15%
and improved HTTP response time (average by 6%, 99th percentile by 10%).
The transmitted probes account for <0.5% of the overall transmissions.
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-03-11 17:00:43 +07:00
|
|
|
extern void tcp_send_loss_probe(struct sock *sk);
|
|
|
|
extern bool tcp_schedule_loss_probe(struct sock *sk);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2005-07-06 05:18:51 +07:00
|
|
|
/* tcp_input.c */
|
|
|
|
extern void tcp_cwnd_application_limited(struct sock *sk);
|
2012-05-02 20:30:04 +07:00
|
|
|
extern void tcp_resume_early_retransmit(struct sock *sk);
|
|
|
|
extern void tcp_rearm_rto(struct sock *sk);
|
2012-08-31 19:29:11 +07:00
|
|
|
extern void tcp_reset(struct sock *sk);
|
2005-07-06 05:18:51 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* tcp_timer.c */
|
|
|
|
extern void tcp_init_xmit_timers(struct sock *);
|
2005-08-10 10:10:42 +07:00
|
|
|
static inline void tcp_clear_xmit_timers(struct sock *sk)
|
|
|
|
{
|
|
|
|
inet_csk_clear_xmit_timers(sk);
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
extern unsigned int tcp_sync_mss(struct sock *sk, u32 pmtu);
|
2009-03-14 21:23:05 +07:00
|
|
|
extern unsigned int tcp_current_mss(struct sock *sk);
|
|
|
|
|
|
|
|
/* Bound MSS / TSO packet size with the half of the window */
|
|
|
|
static inline int tcp_bound_to_half_wnd(struct tcp_sock *tp, int pktsize)
|
|
|
|
{
|
2010-09-16 00:27:52 +07:00
|
|
|
int cutoff;
|
|
|
|
|
|
|
|
/* When peer uses tiny windows, there is no use in packetizing
|
|
|
|
* to sub-MSS pieces for the sake of SWS or making sure there
|
|
|
|
* are enough packets in the pipe for fast recovery.
|
|
|
|
*
|
|
|
|
* On the other hand, for extremely large MSS devices, handling
|
|
|
|
* smaller than MSS windows in this way does make sense.
|
|
|
|
*/
|
|
|
|
if (tp->max_window >= 512)
|
|
|
|
cutoff = (tp->max_window >> 1);
|
|
|
|
else
|
|
|
|
cutoff = tp->max_window;
|
|
|
|
|
|
|
|
if (cutoff && pktsize > cutoff)
|
|
|
|
return max_t(int, cutoff, 68U - tp->tcp_header_len);
|
2009-03-14 21:23:05 +07:00
|
|
|
else
|
|
|
|
return pktsize;
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
|
[INET_DIAG]: Move the tcp_diag interface to the proper place
With this the previous setup is back, i.e. tcp_diag can be built as a module,
as dccp_diag and both share the infrastructure available in inet_diag.
If one selects CONFIG_INET_DIAG as module CONFIG_INET_TCP_DIAG will also be
built as a module, as will CONFIG_INET_DCCP_DIAG, if CONFIG_IP_DCCP was
selected static or as a module, if CONFIG_INET_DIAG is y, being statically
linked CONFIG_INET_TCP_DIAG will follow suit and CONFIG_INET_DCCP_DIAG will be
built in the same manner as CONFIG_IP_DCCP.
Now to aim at UDP, converting it to use inet_hashinfo, so that we can use
iproute2 for UDP sockets as well.
Ah, just to show an example of this new infrastructure working for DCCP :-)
[root@qemu ~]# ./ss -dane
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 0 *:5001 *:* ino:942 sk:cfd503a0
ESTAB 0 0 127.0.0.1:5001 127.0.0.1:32770 ino:943 sk:cfd50a60
ESTAB 0 0 127.0.0.1:32770 127.0.0.1:5001 ino:947 sk:cfd50700
TIME-WAIT 0 0 127.0.0.1:32769 127.0.0.1:5001 timer:(timewait,3.430ms,0) ino:0 sk:cf209620
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-12 22:59:17 +07:00
|
|
|
/* tcp.c */
|
2011-10-21 16:22:42 +07:00
|
|
|
extern void tcp_get_info(const struct sock *, struct tcp_info *);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/* Read 'sendfile()'-style from a TCP socket */
|
|
|
|
typedef int (*sk_read_actor_t)(read_descriptor_t *, struct sk_buff *,
|
|
|
|
unsigned int, size_t);
|
|
|
|
extern int tcp_read_sock(struct sock *sk, read_descriptor_t *desc,
|
|
|
|
sk_read_actor_t recv_actor);
|
|
|
|
|
2006-01-04 07:03:49 +07:00
|
|
|
extern void tcp_initialize_rcv_mss(struct sock *sk);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
ipv6: RTAX_FEATURE_ALLFRAG causes inefficient TCP segment sizing
Quoting Tore Anderson from :
https://bugzilla.kernel.org/show_bug.cgi?id=42572
When RTAX_FEATURE_ALLFRAG is set on a route, the effective TCP segment
size does not take into account the size of the IPv6 Fragmentation
header that needs to be included in outbound packets, causing every
transmitted TCP segment to be fragmented across two IPv6 packets, the
latter of which will only contain 8 bytes of actual payload.
RTAX_FEATURE_ALLFRAG is typically set on a route in response to
receving a ICMPv6 Packet Too Big message indicating a Path MTU of less
than 1280 bytes. 1280 bytes is the minimum IPv6 MTU, however ICMPv6
PTBs with MTU < 1280 are still valid, in particular when an IPv6
packet is sent to an IPv4 destination through a stateless translator.
Any ICMPv4 Need To Fragment packets originated from the IPv4 part of
the path will be translated to ICMPv6 PTB which may then indicate an
MTU of less than 1280.
The Linux kernel refuses to reduce the effective MTU to anything below
1280 bytes, instead it sets it to exactly 1280 bytes, and
RTAX_FEATURE_ALLFRAG is also set. However, the TCP segment size appears
to be set to 1240 bytes (1280 Path MTU - 40 bytes of IPv6 header),
instead of 1232 (additionally taking into account the 8 bytes required
by the IPv6 Fragmentation extension header).
This in turn results in rather inefficient transmission, as every
transmitted TCP segment now is split in two fragments containing
1232+8 bytes of payload.
After this patch, all the outgoing packets that includes a
Fragmentation header all are "atomic" or "non-fragmented" fragments,
i.e., they both have Offset=0 and More Fragments=0.
With help from David S. Miller
Reported-by: Tore Anderson <tore@fud.no>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Cc: Tom Herbert <therbert@google.com>
Tested-by: Tore Anderson <tore@fud.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-04-24 14:37:38 +07:00
|
|
|
extern int tcp_mtu_to_mss(struct sock *sk, int pmtu);
|
|
|
|
extern int tcp_mss_to_mtu(struct sock *sk, int mss);
|
2006-03-21 08:53:41 +07:00
|
|
|
extern void tcp_mtup_init(struct sock *sk);
|
2012-08-31 19:29:11 +07:00
|
|
|
extern void tcp_init_buffer_space(struct sock *sk);
|
2006-03-21 08:53:41 +07:00
|
|
|
|
Revert Backoff [v3]: Revert RTO on ICMP destination unreachable
Here, an ICMP host/network unreachable message, whose payload fits to
TCP's SND.UNA, is taken as an indication that the RTO retransmission has
not been lost due to congestion, but because of a route failure
somewhere along the path.
With true congestion, a router won't trigger such a message and the
patched TCP will operate as standard TCP.
This patch reverts one RTO backoff, if an ICMP host/network unreachable
message, whose payload fits to TCP's SND.UNA, arrives.
Based on the new RTO, the retransmission timer is reset to reflect the
remaining time, or - if the revert clocked out the timer - a retransmission
is sent out immediately.
Backoffs are only reverted, if TCP is in RTO loss recovery, i.e. if
there have been retransmissions and reversible backoffs, already.
Changes from v2:
1) Renaming of skb in tcp_v4_err() moved to another patch.
2) Reintroduced tcp_bound_rto() and __tcp_set_rto().
3) Fixed code comments.
Signed-off-by: Damian Lukowski <damian@tvk.rwth-aachen.de>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-26 07:16:31 +07:00
|
|
|
static inline void tcp_bound_rto(const struct sock *sk)
|
|
|
|
{
|
|
|
|
if (inet_csk(sk)->icsk_rto > TCP_RTO_MAX)
|
|
|
|
inet_csk(sk)->icsk_rto = TCP_RTO_MAX;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline u32 __tcp_set_rto(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return (tp->srtt >> 3) + tp->rttvar;
|
|
|
|
}
|
|
|
|
|
2012-07-10 06:07:30 +07:00
|
|
|
extern void tcp_set_rto(struct sock *sk);
|
|
|
|
|
2006-01-04 07:03:49 +07:00
|
|
|
static inline void __tcp_fast_path_on(struct tcp_sock *tp, u32 snd_wnd)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
tp->pred_flags = htonl((tp->tcp_header_len << 26) |
|
|
|
|
ntohl(TCP_FLAG_ACK) |
|
|
|
|
snd_wnd);
|
|
|
|
}
|
|
|
|
|
2006-01-04 07:03:49 +07:00
|
|
|
static inline void tcp_fast_path_on(struct tcp_sock *tp)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
__tcp_fast_path_on(tp, tp->snd_wnd >> tp->rx_opt.snd_wscale);
|
|
|
|
}
|
|
|
|
|
[TCP]: Sed magic converts func(sk, tp, ...) -> func(sk, ...)
This is (mostly) automated change using magic:
sed -e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e 's|struct sock \*sk,[\n\t ]*struct tcp_sock \*tp\([^{]*\n{\n\)|
struct sock \*sk\1\tstruct tcp_sock *tp = tcp_sk(sk);\n|g'
-e 's|struct sock \*sk, struct tcp_sock \*tp|
struct sock \*sk|g' -e 's|sk, tp\([^-]\)|sk\1|g'
Fixed four unused variable (tp) warnings that were introduced.
In addition, manually added newlines after local variables and
tweaked function arguments positioning.
$ gcc --version
gcc (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)
...
$ codiff -fV built-in.o.old built-in.o.new
net/ipv4/route.c:
rt_cache_flush | +14
1 function changed, 14 bytes added
net/ipv4/tcp.c:
tcp_setsockopt | -5
tcp_sendpage | -25
tcp_sendmsg | -16
3 functions changed, 46 bytes removed
net/ipv4/tcp_input.c:
tcp_try_undo_recovery | +3
tcp_try_undo_dsack | +2
tcp_mark_head_lost | -12
tcp_ack | -15
tcp_event_data_recv | -32
tcp_rcv_state_process | -10
tcp_rcv_established | +1
7 functions changed, 6 bytes added, 69 bytes removed, diff: -63
net/ipv4/tcp_output.c:
update_send_head | -9
tcp_transmit_skb | +19
tcp_cwnd_validate | +1
tcp_write_wakeup | -17
__tcp_push_pending_frames | -25
tcp_push_one | -8
tcp_send_fin | -4
7 functions changed, 20 bytes added, 63 bytes removed, diff: -43
built-in.o.new:
18 functions changed, 40 bytes added, 178 bytes removed, diff: -138
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2007-04-21 12:18:02 +07:00
|
|
|
static inline void tcp_fast_path_check(struct sock *sk)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
[TCP]: Sed magic converts func(sk, tp, ...) -> func(sk, ...)
This is (mostly) automated change using magic:
sed -e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e 's|struct sock \*sk,[\n\t ]*struct tcp_sock \*tp\([^{]*\n{\n\)|
struct sock \*sk\1\tstruct tcp_sock *tp = tcp_sk(sk);\n|g'
-e 's|struct sock \*sk, struct tcp_sock \*tp|
struct sock \*sk|g' -e 's|sk, tp\([^-]\)|sk\1|g'
Fixed four unused variable (tp) warnings that were introduced.
In addition, manually added newlines after local variables and
tweaked function arguments positioning.
$ gcc --version
gcc (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)
...
$ codiff -fV built-in.o.old built-in.o.new
net/ipv4/route.c:
rt_cache_flush | +14
1 function changed, 14 bytes added
net/ipv4/tcp.c:
tcp_setsockopt | -5
tcp_sendpage | -25
tcp_sendmsg | -16
3 functions changed, 46 bytes removed
net/ipv4/tcp_input.c:
tcp_try_undo_recovery | +3
tcp_try_undo_dsack | +2
tcp_mark_head_lost | -12
tcp_ack | -15
tcp_event_data_recv | -32
tcp_rcv_state_process | -10
tcp_rcv_established | +1
7 functions changed, 6 bytes added, 69 bytes removed, diff: -63
net/ipv4/tcp_output.c:
update_send_head | -9
tcp_transmit_skb | +19
tcp_cwnd_validate | +1
tcp_write_wakeup | -17
__tcp_push_pending_frames | -25
tcp_push_one | -8
tcp_send_fin | -4
7 functions changed, 20 bytes added, 63 bytes removed, diff: -43
built-in.o.new:
18 functions changed, 40 bytes added, 178 bytes removed, diff: -138
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2007-04-21 12:18:02 +07:00
|
|
|
struct tcp_sock *tp = tcp_sk(sk);
|
|
|
|
|
2005-07-09 04:57:23 +07:00
|
|
|
if (skb_queue_empty(&tp->out_of_order_queue) &&
|
2005-04-17 05:20:36 +07:00
|
|
|
tp->rcv_wnd &&
|
|
|
|
atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf &&
|
|
|
|
!tp->urg_data)
|
|
|
|
tcp_fast_path_on(tp);
|
|
|
|
}
|
|
|
|
|
2009-05-05 01:11:01 +07:00
|
|
|
/* Compute the actual rto_min value */
|
|
|
|
static inline u32 tcp_rto_min(struct sock *sk)
|
|
|
|
{
|
2011-10-21 16:22:42 +07:00
|
|
|
const struct dst_entry *dst = __sk_dst_get(sk);
|
2009-05-05 01:11:01 +07:00
|
|
|
u32 rto_min = TCP_RTO_MIN;
|
|
|
|
|
|
|
|
if (dst && dst_metric_locked(dst, RTAX_RTO_MIN))
|
|
|
|
rto_min = dst_metric_rtt(dst, RTAX_RTO_MIN);
|
|
|
|
return rto_min;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Compute the actual receive window we are currently advertising.
|
|
|
|
* Rcv_nxt can be after the window if our peer push more data
|
|
|
|
* than the offered window.
|
|
|
|
*/
|
2006-01-04 07:03:49 +07:00
|
|
|
static inline u32 tcp_receive_window(const struct tcp_sock *tp)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
s32 win = tp->rcv_wup + tp->rcv_wnd - tp->rcv_nxt;
|
|
|
|
|
|
|
|
if (win < 0)
|
|
|
|
win = 0;
|
|
|
|
return (u32) win;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Choose a new window, without checks for shrinking, and without
|
|
|
|
* scaling applied to the result. The caller does these things
|
|
|
|
* if necessary. This is a "raw" window selection.
|
|
|
|
*/
|
2010-07-11 03:41:06 +07:00
|
|
|
extern u32 __tcp_select_window(struct sock *sk);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2012-04-19 10:40:39 +07:00
|
|
|
void tcp_send_window_probe(struct sock *sk);
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* TCP timestamps are only 32-bits, this causes a slight
|
|
|
|
* complication on 64-bit systems since we store a snapshot
|
2005-11-16 06:17:10 +07:00
|
|
|
* of jiffies in the buffer control blocks below. We decided
|
|
|
|
* to use only the low 32-bits of jiffies and hide the ugly
|
2005-04-17 05:20:36 +07:00
|
|
|
* casts with the following macro.
|
|
|
|
*/
|
|
|
|
#define tcp_time_stamp ((__u32)(jiffies))
|
|
|
|
|
2010-06-12 21:01:43 +07:00
|
|
|
#define tcp_flag_byte(th) (((u_int8_t *)th)[13])
|
|
|
|
|
|
|
|
#define TCPHDR_FIN 0x01
|
|
|
|
#define TCPHDR_SYN 0x02
|
|
|
|
#define TCPHDR_RST 0x04
|
|
|
|
#define TCPHDR_PSH 0x08
|
|
|
|
#define TCPHDR_ACK 0x10
|
|
|
|
#define TCPHDR_URG 0x20
|
|
|
|
#define TCPHDR_ECE 0x40
|
|
|
|
#define TCPHDR_CWR 0x80
|
|
|
|
|
2005-11-11 08:13:47 +07:00
|
|
|
/* This is what the send packet queuing engine uses to pass
|
2010-07-16 11:41:00 +07:00
|
|
|
* TCP per-packet control information to the transmission code.
|
|
|
|
* We also store the host-order sequence numbers in here too.
|
|
|
|
* This is 44 bytes if IPV6 is enabled.
|
|
|
|
* If this grows please adjust skbuff.h:skbuff->cb[xxx] size appropriately.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
struct tcp_skb_cb {
|
|
|
|
union {
|
|
|
|
struct inet_skb_parm h4;
|
2011-12-10 16:48:31 +07:00
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
2005-04-17 05:20:36 +07:00
|
|
|
struct inet6_skb_parm h6;
|
|
|
|
#endif
|
|
|
|
} header; /* For incoming frames */
|
|
|
|
__u32 seq; /* Starting sequence number */
|
|
|
|
__u32 end_seq; /* SEQ + FIN + SYN + datalen */
|
|
|
|
__u32 when; /* used to compute rtt's */
|
2011-09-28 00:25:05 +07:00
|
|
|
__u8 tcp_flags; /* TCP header flags. (tcp[13]) */
|
2012-04-16 14:08:06 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
__u8 sacked; /* State flags for SACK/FACK. */
|
|
|
|
#define TCPCB_SACKED_ACKED 0x01 /* SKB ACK'd by a SACK block */
|
|
|
|
#define TCPCB_SACKED_RETRANS 0x02 /* SKB retransmitted */
|
|
|
|
#define TCPCB_LOST 0x04 /* SKB is lost */
|
|
|
|
#define TCPCB_TAGBITS 0x07 /* All tag bits */
|
|
|
|
#define TCPCB_EVER_RETRANS 0x80 /* Ever retransmitted frame */
|
|
|
|
#define TCPCB_RETRANS (TCPCB_SACKED_RETRANS|TCPCB_EVER_RETRANS)
|
|
|
|
|
2012-04-16 14:08:06 +07:00
|
|
|
__u8 ip_dsfield; /* IPv4 tos or IPv6 dsfield */
|
|
|
|
/* 1 byte hole */
|
2005-04-17 05:20:36 +07:00
|
|
|
__u32 ack_seq; /* Sequence number ACK'd */
|
|
|
|
};
|
|
|
|
|
|
|
|
#define TCP_SKB_CB(__skb) ((struct tcp_skb_cb *)&((__skb)->cb[0]))
|
|
|
|
|
2012-05-04 12:14:02 +07:00
|
|
|
/* RFC3168 : 6.1.1 SYN packets must not have ECT/ECN bits set
|
|
|
|
*
|
|
|
|
* If we receive a SYN packet with these bits set, it means a network is
|
|
|
|
* playing bad games with TOS bits. In order to avoid possible false congestion
|
|
|
|
* notifications, we disable TCP ECN negociation.
|
|
|
|
*/
|
|
|
|
static inline void
|
2013-01-05 23:10:48 +07:00
|
|
|
TCP_ECN_create_request(struct request_sock *req, const struct sk_buff *skb,
|
|
|
|
struct net *net)
|
2012-05-04 12:14:02 +07:00
|
|
|
{
|
|
|
|
const struct tcphdr *th = tcp_hdr(skb);
|
|
|
|
|
2013-01-05 23:10:48 +07:00
|
|
|
if (net->ipv4.sysctl_tcp_ecn && th->ece && th->cwr &&
|
2012-05-04 12:14:02 +07:00
|
|
|
INET_ECN_is_not_ect(TCP_SKB_CB(skb)->ip_dsfield))
|
|
|
|
inet_rsk(req)->ecn_ok = 1;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Due to TSO, an SKB can be composed of multiple actual
|
|
|
|
* packets. To keep these tracked properly, we use this.
|
|
|
|
*/
|
|
|
|
static inline int tcp_skb_pcount(const struct sk_buff *skb)
|
|
|
|
{
|
2006-06-22 16:40:14 +07:00
|
|
|
return skb_shinfo(skb)->gso_segs;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* This is valid iff tcp_skb_pcount() > 1. */
|
|
|
|
static inline int tcp_skb_mss(const struct sk_buff *skb)
|
|
|
|
{
|
2006-06-22 16:40:14 +07:00
|
|
|
return skb_shinfo(skb)->gso_size;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2005-06-24 02:19:55 +07:00
|
|
|
/* Events passed to congestion control interface */
|
|
|
|
enum tcp_ca_event {
|
|
|
|
CA_EVENT_TX_START, /* first transmit when no packets in flight */
|
|
|
|
CA_EVENT_CWND_RESTART, /* congestion window restart */
|
|
|
|
CA_EVENT_COMPLETE_CWR, /* end of congestion recovery */
|
|
|
|
CA_EVENT_LOSS, /* loss timeout */
|
|
|
|
CA_EVENT_FAST_ACK, /* in sequence ack */
|
|
|
|
CA_EVENT_SLOW_ACK, /* other ack */
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Interface for adding new TCP congestion control handlers
|
|
|
|
*/
|
|
|
|
#define TCP_CA_NAME_MAX 16
|
2006-11-10 07:32:06 +07:00
|
|
|
#define TCP_CA_MAX 128
|
|
|
|
#define TCP_CA_BUF_MAX (TCP_CA_NAME_MAX*TCP_CA_MAX)
|
|
|
|
|
2007-04-24 12:26:16 +07:00
|
|
|
#define TCP_CONG_NON_RESTRICTED 0x1
|
|
|
|
#define TCP_CONG_RTT_STAMP 0x2
|
|
|
|
|
2005-06-24 02:19:55 +07:00
|
|
|
struct tcp_congestion_ops {
|
|
|
|
struct list_head list;
|
2007-04-24 12:26:16 +07:00
|
|
|
unsigned long flags;
|
2005-06-24 02:19:55 +07:00
|
|
|
|
|
|
|
/* initialize private data (optional) */
|
2005-08-10 14:03:31 +07:00
|
|
|
void (*init)(struct sock *sk);
|
2005-06-24 02:19:55 +07:00
|
|
|
/* cleanup private data (optional) */
|
2005-08-10 14:03:31 +07:00
|
|
|
void (*release)(struct sock *sk);
|
2005-06-24 02:19:55 +07:00
|
|
|
|
|
|
|
/* return slow start threshold (required) */
|
2005-08-10 14:03:31 +07:00
|
|
|
u32 (*ssthresh)(struct sock *sk);
|
2005-06-24 02:19:55 +07:00
|
|
|
/* lower bound for congestion window (optional) */
|
2006-06-06 07:30:08 +07:00
|
|
|
u32 (*min_cwnd)(const struct sock *sk);
|
2005-06-24 02:19:55 +07:00
|
|
|
/* do new cwnd calculation (required) */
|
2007-12-02 05:47:59 +07:00
|
|
|
void (*cong_avoid)(struct sock *sk, u32 ack, u32 in_flight);
|
2005-06-24 02:19:55 +07:00
|
|
|
/* call before changing ca_state (optional) */
|
2005-08-10 14:03:31 +07:00
|
|
|
void (*set_state)(struct sock *sk, u8 new_state);
|
2005-06-24 02:19:55 +07:00
|
|
|
/* call when cwnd event occurs (optional) */
|
2005-08-10 14:03:31 +07:00
|
|
|
void (*cwnd_event)(struct sock *sk, enum tcp_ca_event ev);
|
2005-06-24 02:19:55 +07:00
|
|
|
/* new value of cwnd after loss (optional) */
|
2005-08-10 14:03:31 +07:00
|
|
|
u32 (*undo_cwnd)(struct sock *sk);
|
2005-06-24 02:19:55 +07:00
|
|
|
/* hook for packet ack accounting (optional) */
|
2007-07-26 13:49:34 +07:00
|
|
|
void (*pkts_acked)(struct sock *sk, u32 num_acked, s32 rtt_us);
|
2005-08-12 22:51:49 +07:00
|
|
|
/* get info for inet_diag (optional) */
|
2005-08-10 14:03:31 +07:00
|
|
|
void (*get_info)(struct sock *sk, u32 ext, struct sk_buff *skb);
|
2005-06-24 02:19:55 +07:00
|
|
|
|
|
|
|
char name[TCP_CA_NAME_MAX];
|
|
|
|
struct module *owner;
|
|
|
|
};
|
|
|
|
|
|
|
|
extern int tcp_register_congestion_control(struct tcp_congestion_ops *type);
|
|
|
|
extern void tcp_unregister_congestion_control(struct tcp_congestion_ops *type);
|
|
|
|
|
2005-08-10 14:03:31 +07:00
|
|
|
extern void tcp_init_congestion_control(struct sock *sk);
|
|
|
|
extern void tcp_cleanup_congestion_control(struct sock *sk);
|
2005-06-24 02:19:55 +07:00
|
|
|
extern int tcp_set_default_congestion_control(const char *name);
|
|
|
|
extern void tcp_get_default_congestion_control(char *name);
|
2006-11-10 07:32:06 +07:00
|
|
|
extern void tcp_get_available_congestion_control(char *buf, size_t len);
|
2006-11-10 07:35:15 +07:00
|
|
|
extern void tcp_get_allowed_congestion_control(char *buf, size_t len);
|
|
|
|
extern int tcp_set_allowed_congestion_control(char *allowed);
|
2005-08-10 14:03:31 +07:00
|
|
|
extern int tcp_set_congestion_control(struct sock *sk, const char *name);
|
2006-01-04 07:03:49 +07:00
|
|
|
extern void tcp_slow_start(struct tcp_sock *tp);
|
2009-02-28 11:44:37 +07:00
|
|
|
extern void tcp_cong_avoid_ai(struct tcp_sock *tp, u32 w);
|
2005-06-24 02:19:55 +07:00
|
|
|
|
2005-06-24 10:37:36 +07:00
|
|
|
extern struct tcp_congestion_ops tcp_init_congestion_ops;
|
2005-08-10 14:03:31 +07:00
|
|
|
extern u32 tcp_reno_ssthresh(struct sock *sk);
|
2007-12-02 05:47:59 +07:00
|
|
|
extern void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 in_flight);
|
2006-06-06 07:30:08 +07:00
|
|
|
extern u32 tcp_reno_min_cwnd(const struct sock *sk);
|
2005-06-24 13:45:02 +07:00
|
|
|
extern struct tcp_congestion_ops tcp_reno;
|
2005-06-24 02:19:55 +07:00
|
|
|
|
2005-08-10 14:03:31 +07:00
|
|
|
static inline void tcp_set_ca_state(struct sock *sk, const u8 ca_state)
|
2005-06-24 02:19:55 +07:00
|
|
|
{
|
2005-08-10 14:03:31 +07:00
|
|
|
struct inet_connection_sock *icsk = inet_csk(sk);
|
|
|
|
|
|
|
|
if (icsk->icsk_ca_ops->set_state)
|
|
|
|
icsk->icsk_ca_ops->set_state(sk, ca_state);
|
|
|
|
icsk->icsk_ca_state = ca_state;
|
2005-06-24 02:19:55 +07:00
|
|
|
}
|
|
|
|
|
2005-08-10 14:03:31 +07:00
|
|
|
static inline void tcp_ca_event(struct sock *sk, const enum tcp_ca_event event)
|
2005-06-24 02:19:55 +07:00
|
|
|
{
|
2005-08-10 14:03:31 +07:00
|
|
|
const struct inet_connection_sock *icsk = inet_csk(sk);
|
|
|
|
|
|
|
|
if (icsk->icsk_ca_ops->cwnd_event)
|
|
|
|
icsk->icsk_ca_ops->cwnd_event(sk, event);
|
2005-06-24 02:19:55 +07:00
|
|
|
}
|
|
|
|
|
2007-08-09 19:14:46 +07:00
|
|
|
/* These functions determine how the current flow behaves in respect of SACK
|
|
|
|
* handling. SACK is negotiated with the peer, and therefore it can vary
|
|
|
|
* between different flows.
|
|
|
|
*
|
|
|
|
* tcp_is_sack - SACK enabled
|
|
|
|
* tcp_is_reno - No SACK
|
|
|
|
* tcp_is_fack - FACK enabled, implies SACK enabled
|
|
|
|
*/
|
|
|
|
static inline int tcp_is_sack(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->rx_opt.sack_ok;
|
|
|
|
}
|
|
|
|
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_is_reno(const struct tcp_sock *tp)
|
2007-08-09 19:14:46 +07:00
|
|
|
{
|
|
|
|
return !tcp_is_sack(tp);
|
|
|
|
}
|
|
|
|
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_is_fack(const struct tcp_sock *tp)
|
2007-08-09 19:14:46 +07:00
|
|
|
{
|
2011-12-20 20:23:24 +07:00
|
|
|
return tp->rx_opt.sack_ok & TCP_FACK_ENABLED;
|
2007-08-09 19:14:46 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_enable_fack(struct tcp_sock *tp)
|
|
|
|
{
|
2011-12-20 20:23:24 +07:00
|
|
|
tp->rx_opt.sack_ok |= TCP_FACK_ENABLED;
|
2007-08-09 19:14:46 +07:00
|
|
|
}
|
|
|
|
|
2012-05-02 20:30:03 +07:00
|
|
|
/* TCP early-retransmit (ER) is similar to but more conservative than
|
|
|
|
* the thin-dupack feature. Enable ER only if thin-dupack is disabled.
|
|
|
|
*/
|
|
|
|
static inline void tcp_enable_early_retrans(struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
tp->do_early_retrans = sysctl_tcp_early_retrans &&
|
tcp: Tail loss probe (TLP)
This patch series implement the Tail loss probe (TLP) algorithm described
in http://tools.ietf.org/html/draft-dukkipati-tcpm-tcp-loss-probe-01. The
first patch implements the basic algorithm.
TLP's goal is to reduce tail latency of short transactions. It achieves
this by converting retransmission timeouts (RTOs) occuring due
to tail losses (losses at end of transactions) into fast recovery.
TLP transmits one packet in two round-trips when a connection is in
Open state and isn't receiving any ACKs. The transmitted packet, aka
loss probe, can be either new or a retransmission. When there is tail
loss, the ACK from a loss probe triggers FACK/early-retransmit based
fast recovery, thus avoiding a costly RTO. In the absence of loss,
there is no change in the connection state.
PTO stands for probe timeout. It is a timer event indicating
that an ACK is overdue and triggers a loss probe packet. The PTO value
is set to max(2*SRTT, 10ms) and is adjusted to account for delayed
ACK timer when there is only one oustanding packet.
TLP Algorithm
On transmission of new data in Open state:
-> packets_out > 1: schedule PTO in max(2*SRTT, 10ms).
-> packets_out == 1: schedule PTO in max(2*RTT, 1.5*RTT + 200ms)
-> PTO = min(PTO, RTO)
Conditions for scheduling PTO:
-> Connection is in Open state.
-> Connection is either cwnd limited or no new data to send.
-> Number of probes per tail loss episode is limited to one.
-> Connection is SACK enabled.
When PTO fires:
new_segment_exists:
-> transmit new segment.
-> packets_out++. cwnd remains same.
no_new_packet:
-> retransmit the last segment.
Its ACK triggers FACK or early retransmit based recovery.
ACK path:
-> rearm RTO at start of ACK processing.
-> reschedule PTO if need be.
In addition, the patch includes a small variation to the Early Retransmit
(ER) algorithm, such that ER and TLP together can in principle recover any
N-degree of tail loss through fast recovery. TLP is controlled by the same
sysctl as ER, tcp_early_retrans sysctl.
tcp_early_retrans==0; disables TLP and ER.
==1; enables RFC5827 ER.
==2; delayed ER.
==3; TLP and delayed ER. [DEFAULT]
==4; TLP only.
The TLP patch series have been extensively tested on Google Web servers.
It is most effective for short Web trasactions, where it reduced RTOs by 15%
and improved HTTP response time (average by 6%, 99th percentile by 10%).
The transmitted probes account for <0.5% of the overall transmissions.
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-03-11 17:00:43 +07:00
|
|
|
sysctl_tcp_early_retrans < 4 && !sysctl_tcp_thin_dupack &&
|
|
|
|
sysctl_tcp_reordering == 3;
|
2012-05-02 20:30:03 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_disable_early_retrans(struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
tp->do_early_retrans = 0;
|
|
|
|
}
|
|
|
|
|
2007-08-09 18:37:30 +07:00
|
|
|
static inline unsigned int tcp_left_out(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->sacked_out + tp->lost_out;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* This determines how many packets are "in the network" to the best
|
|
|
|
* of our knowledge. In many cases it is conservative, but where
|
|
|
|
* detailed information is available from the receiver (via SACK
|
|
|
|
* blocks etc.) we can make more aggressive calculations.
|
|
|
|
*
|
|
|
|
* Use this for decisions involving congestion control, use just
|
|
|
|
* tp->packets_out to determine if the send queue is empty or not.
|
|
|
|
*
|
|
|
|
* Read this equation as:
|
|
|
|
*
|
|
|
|
* "Packets sent once on transmission queue" MINUS
|
|
|
|
* "Packets left network, but not honestly ACKed yet" PLUS
|
|
|
|
* "Packets fast retransmitted"
|
|
|
|
*/
|
2006-01-04 07:03:49 +07:00
|
|
|
static inline unsigned int tcp_packets_in_flight(const struct tcp_sock *tp)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2007-08-09 18:37:30 +07:00
|
|
|
return tp->packets_out - tcp_left_out(tp) + tp->retrans_out;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2009-09-15 15:30:10 +07:00
|
|
|
#define TCP_INFINITE_SSTHRESH 0x7fffffff
|
|
|
|
|
|
|
|
static inline bool tcp_in_initial_slowstart(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->snd_ssthresh >= TCP_INFINITE_SSTHRESH;
|
|
|
|
}
|
|
|
|
|
2012-09-03 00:38:04 +07:00
|
|
|
static inline bool tcp_in_cwnd_reduction(const struct sock *sk)
|
|
|
|
{
|
|
|
|
return (TCPF_CA_CWR | TCPF_CA_Recovery) &
|
|
|
|
(1 << inet_csk(sk)->icsk_ca_state);
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* If cwnd > ssthresh, we may raise ssthresh to be half-way to cwnd.
|
2012-09-03 00:38:04 +07:00
|
|
|
* The exception is cwnd reduction phase, when cwnd is decreasing towards
|
2005-04-17 05:20:36 +07:00
|
|
|
* ssthresh.
|
|
|
|
*/
|
2005-08-10 14:03:31 +07:00
|
|
|
static inline __u32 tcp_current_ssthresh(const struct sock *sk)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-08-10 14:03:31 +07:00
|
|
|
const struct tcp_sock *tp = tcp_sk(sk);
|
2011-10-21 16:22:42 +07:00
|
|
|
|
2012-09-03 00:38:04 +07:00
|
|
|
if (tcp_in_cwnd_reduction(sk))
|
2005-04-17 05:20:36 +07:00
|
|
|
return tp->snd_ssthresh;
|
|
|
|
else
|
|
|
|
return max(tp->snd_ssthresh,
|
|
|
|
((tp->snd_cwnd >> 1) +
|
|
|
|
(tp->snd_cwnd >> 2)));
|
|
|
|
}
|
|
|
|
|
2007-07-27 20:36:17 +07:00
|
|
|
/* Use define here intentionally to get WARN_ON location shown at the caller */
|
|
|
|
#define tcp_verify_left_out(tp) WARN_ON(tcp_left_out(tp) > tp->packets_out)
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2007-02-28 01:09:49 +07:00
|
|
|
extern void tcp_enter_cwr(struct sock *sk, const int set_ssthresh);
|
2011-10-21 16:22:42 +07:00
|
|
|
extern __u32 tcp_init_cwnd(const struct tcp_sock *tp, const struct dst_entry *dst);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2011-11-22 00:15:14 +07:00
|
|
|
/* The maximum number of MSS of available cwnd for which TSO defers
|
|
|
|
* sending if not using sysctl_tcp_tso_win_divisor.
|
|
|
|
*/
|
|
|
|
static inline __u32 tcp_max_tso_deferred_mss(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return 3;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Slow start with delack produces 3 packets of burst, so that
|
2008-04-16 05:26:39 +07:00
|
|
|
* it is safe "de facto". This will be the default - same as
|
|
|
|
* the default reordering threshold - but if reordering increases,
|
|
|
|
* we must be able to allow cwnd to burst at least this much in order
|
|
|
|
* to not pull it back when holes are filled.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
static __inline__ __u32 tcp_max_burst(const struct tcp_sock *tp)
|
|
|
|
{
|
2008-04-16 05:26:39 +07:00
|
|
|
return tp->reordering;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2007-12-31 19:48:41 +07:00
|
|
|
/* Returns end sequence number of the receiver's advertised window */
|
|
|
|
static inline u32 tcp_wnd_end(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->snd_una + tp->snd_wnd;
|
|
|
|
}
|
2012-05-17 06:15:34 +07:00
|
|
|
extern bool tcp_is_cwnd_limited(const struct sock *sk, u32 in_flight);
|
2005-11-11 07:53:30 +07:00
|
|
|
|
2007-10-24 11:08:54 +07:00
|
|
|
static inline void tcp_minshall_update(struct tcp_sock *tp, unsigned int mss,
|
2006-01-04 07:03:49 +07:00
|
|
|
const struct sk_buff *skb)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
if (skb->len < mss)
|
|
|
|
tp->snd_sml = TCP_SKB_CB(skb)->end_seq;
|
|
|
|
}
|
|
|
|
|
[TCP]: Sed magic converts func(sk, tp, ...) -> func(sk, ...)
This is (mostly) automated change using magic:
sed -e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e 's|struct sock \*sk,[\n\t ]*struct tcp_sock \*tp\([^{]*\n{\n\)|
struct sock \*sk\1\tstruct tcp_sock *tp = tcp_sk(sk);\n|g'
-e 's|struct sock \*sk, struct tcp_sock \*tp|
struct sock \*sk|g' -e 's|sk, tp\([^-]\)|sk\1|g'
Fixed four unused variable (tp) warnings that were introduced.
In addition, manually added newlines after local variables and
tweaked function arguments positioning.
$ gcc --version
gcc (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)
...
$ codiff -fV built-in.o.old built-in.o.new
net/ipv4/route.c:
rt_cache_flush | +14
1 function changed, 14 bytes added
net/ipv4/tcp.c:
tcp_setsockopt | -5
tcp_sendpage | -25
tcp_sendmsg | -16
3 functions changed, 46 bytes removed
net/ipv4/tcp_input.c:
tcp_try_undo_recovery | +3
tcp_try_undo_dsack | +2
tcp_mark_head_lost | -12
tcp_ack | -15
tcp_event_data_recv | -32
tcp_rcv_state_process | -10
tcp_rcv_established | +1
7 functions changed, 6 bytes added, 69 bytes removed, diff: -63
net/ipv4/tcp_output.c:
update_send_head | -9
tcp_transmit_skb | +19
tcp_cwnd_validate | +1
tcp_write_wakeup | -17
__tcp_push_pending_frames | -25
tcp_push_one | -8
tcp_send_fin | -4
7 functions changed, 20 bytes added, 63 bytes removed, diff: -43
built-in.o.new:
18 functions changed, 40 bytes added, 178 bytes removed, diff: -138
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2007-04-21 12:18:02 +07:00
|
|
|
static inline void tcp_check_probe_timer(struct sock *sk)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2011-10-21 16:22:42 +07:00
|
|
|
const struct tcp_sock *tp = tcp_sk(sk);
|
2005-08-10 10:10:42 +07:00
|
|
|
const struct inet_connection_sock *icsk = inet_csk(sk);
|
[TCP]: Sed magic converts func(sk, tp, ...) -> func(sk, ...)
This is (mostly) automated change using magic:
sed -e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e '/struct sock \*sk/ N' -e '/struct sock \*sk/ N'
-e 's|struct sock \*sk,[\n\t ]*struct tcp_sock \*tp\([^{]*\n{\n\)|
struct sock \*sk\1\tstruct tcp_sock *tp = tcp_sk(sk);\n|g'
-e 's|struct sock \*sk, struct tcp_sock \*tp|
struct sock \*sk|g' -e 's|sk, tp\([^-]\)|sk\1|g'
Fixed four unused variable (tp) warnings that were introduced.
In addition, manually added newlines after local variables and
tweaked function arguments positioning.
$ gcc --version
gcc (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)
...
$ codiff -fV built-in.o.old built-in.o.new
net/ipv4/route.c:
rt_cache_flush | +14
1 function changed, 14 bytes added
net/ipv4/tcp.c:
tcp_setsockopt | -5
tcp_sendpage | -25
tcp_sendmsg | -16
3 functions changed, 46 bytes removed
net/ipv4/tcp_input.c:
tcp_try_undo_recovery | +3
tcp_try_undo_dsack | +2
tcp_mark_head_lost | -12
tcp_ack | -15
tcp_event_data_recv | -32
tcp_rcv_state_process | -10
tcp_rcv_established | +1
7 functions changed, 6 bytes added, 69 bytes removed, diff: -63
net/ipv4/tcp_output.c:
update_send_head | -9
tcp_transmit_skb | +19
tcp_cwnd_validate | +1
tcp_write_wakeup | -17
__tcp_push_pending_frames | -25
tcp_push_one | -8
tcp_send_fin | -4
7 functions changed, 20 bytes added, 63 bytes removed, diff: -43
built-in.o.new:
18 functions changed, 40 bytes added, 178 bytes removed, diff: -138
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2007-04-21 12:18:02 +07:00
|
|
|
|
2005-08-10 10:10:42 +07:00
|
|
|
if (!tp->packets_out && !icsk->icsk_pending)
|
2005-08-10 10:11:08 +07:00
|
|
|
inet_csk_reset_xmit_timer(sk, ICSK_TIME_PROBE0,
|
|
|
|
icsk->icsk_rto, TCP_RTO_MAX);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2009-03-03 13:42:02 +07:00
|
|
|
static inline void tcp_init_wl(struct tcp_sock *tp, u32 seq)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
tp->snd_wl1 = seq;
|
|
|
|
}
|
|
|
|
|
2009-03-03 13:42:02 +07:00
|
|
|
static inline void tcp_update_wl(struct tcp_sock *tp, u32 seq)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
tp->snd_wl1 = seq;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Calculate(/check) TCP checksum
|
|
|
|
*/
|
2007-02-05 11:15:27 +07:00
|
|
|
static inline __sum16 tcp_v4_check(int len, __be32 saddr,
|
|
|
|
__be32 daddr, __wsum base)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
return csum_tcpudp_magic(saddr,daddr,len,IPPROTO_TCP,base);
|
|
|
|
}
|
|
|
|
|
2006-11-15 12:40:42 +07:00
|
|
|
static inline __sum16 __tcp_checksum_complete(struct sk_buff *skb)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-11-11 04:01:24 +07:00
|
|
|
return __skb_checksum_complete(skb);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_checksum_complete(struct sk_buff *skb)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2007-04-10 01:59:39 +07:00
|
|
|
return !skb_csum_unnecessary(skb) &&
|
2005-04-17 05:20:36 +07:00
|
|
|
__tcp_checksum_complete(skb);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Prequeue for VJ style copy to user, combined with checksumming. */
|
|
|
|
|
2006-01-04 07:03:49 +07:00
|
|
|
static inline void tcp_prequeue_init(struct tcp_sock *tp)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
tp->ucopy.task = NULL;
|
|
|
|
tp->ucopy.len = 0;
|
|
|
|
tp->ucopy.memory = 0;
|
|
|
|
skb_queue_head_init(&tp->ucopy.prequeue);
|
2006-05-24 07:55:33 +07:00
|
|
|
#ifdef CONFIG_NET_DMA
|
|
|
|
tp->ucopy.dma_chan = NULL;
|
|
|
|
tp->ucopy.wakeup = 0;
|
|
|
|
tp->ucopy.pinned_list = NULL;
|
|
|
|
tp->ucopy.dma_cookie = 0;
|
|
|
|
#endif
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2013-03-06 19:58:01 +07:00
|
|
|
extern bool tcp_prequeue(struct sock *sk, struct sk_buff *skb);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
#undef STATE_TRACE
|
|
|
|
|
|
|
|
#ifdef STATE_TRACE
|
|
|
|
static const char *statename[]={
|
|
|
|
"Unused","Established","Syn Sent","Syn Recv",
|
|
|
|
"Fin Wait 1","Fin Wait 2","Time Wait", "Close",
|
|
|
|
"Close Wait","Last ACK","Listen","Closing"
|
|
|
|
};
|
|
|
|
#endif
|
[TCP]: Uninline tcp_set_state
net/ipv4/tcp.c:
tcp_close_state | -226
tcp_done | -145
tcp_close | -564
tcp_disconnect | -141
4 functions changed, 1076 bytes removed, diff: -1076
net/ipv4/tcp_input.c:
tcp_fin | -86
tcp_rcv_state_process | -164
2 functions changed, 250 bytes removed, diff: -250
net/ipv4/tcp_ipv4.c:
tcp_v4_connect | -209
1 function changed, 209 bytes removed, diff: -209
net/ipv4/arp.c:
arp_ignore | +5
1 function changed, 5 bytes added, diff: +5
net/ipv6/tcp_ipv6.c:
tcp_v6_connect | -158
1 function changed, 158 bytes removed, diff: -158
net/sunrpc/xprtsock.c:
xs_sendpages | -2
1 function changed, 2 bytes removed, diff: -2
net/dccp/ccids/ccid3.c:
ccid3_update_send_interval | +7
1 function changed, 7 bytes added, diff: +7
net/ipv4/tcp.c:
tcp_set_state | +238
1 function changed, 238 bytes added, diff: +238
built-in.o:
12 functions changed, 250 bytes added, 1695 bytes removed, diff: -1445
I've no explanation why some unrelated changes seem to occur
consistently as well (arp_ignore, ccid3_update_send_interval;
I checked the arp_ignore asm and it seems to be due to some
reordered of operation order causing some extra opcodes to be
generated). Still, the benefits are pretty obvious from the
codiff's results.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-01-12 18:17:20 +07:00
|
|
|
extern void tcp_set_state(struct sock *sk, int state);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2007-04-21 07:11:46 +07:00
|
|
|
extern void tcp_done(struct sock *sk);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2006-01-04 07:03:49 +07:00
|
|
|
static inline void tcp_sack_reset(struct tcp_options_received *rx_opt)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
rx_opt->dsack = 0;
|
|
|
|
rx_opt->num_sacks = 0;
|
|
|
|
}
|
|
|
|
|
2013-06-12 05:35:32 +07:00
|
|
|
extern u32 tcp_default_init_rwnd(u32 mss);
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Determine a window scaling and initial window to offer. */
|
|
|
|
extern void tcp_select_initial_window(int __space, __u32 mss,
|
|
|
|
__u32 *rcv_wnd, __u32 *window_clamp,
|
2009-12-15 18:15:28 +07:00
|
|
|
int wscale_ok, __u8 *rcv_wscale,
|
|
|
|
__u32 init_rcv_wnd);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
static inline int tcp_win_from_space(int space)
|
|
|
|
{
|
|
|
|
return sysctl_tcp_adv_win_scale<=0 ?
|
|
|
|
(space>>(-sysctl_tcp_adv_win_scale)) :
|
|
|
|
space - (space>>sysctl_tcp_adv_win_scale);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Note: caller must be prepared to deal with negative returns */
|
|
|
|
static inline int tcp_space(const struct sock *sk)
|
|
|
|
{
|
|
|
|
return tcp_win_from_space(sk->sk_rcvbuf -
|
|
|
|
atomic_read(&sk->sk_rmem_alloc));
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline int tcp_full_space(const struct sock *sk)
|
|
|
|
{
|
|
|
|
return tcp_win_from_space(sk->sk_rcvbuf);
|
|
|
|
}
|
|
|
|
|
2006-01-04 07:03:49 +07:00
|
|
|
static inline void tcp_openreq_init(struct request_sock *req,
|
|
|
|
struct tcp_options_received *rx_opt,
|
|
|
|
struct sk_buff *skb)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
[NET] Generalise TCP's struct open_request minisock infrastructure
Kept this first changeset minimal, without changing existing names to
ease peer review.
Basicaly tcp_openreq_alloc now receives the or_calltable, that in turn
has two new members:
->slab, that replaces tcp_openreq_cachep
->obj_size, to inform the size of the openreq descendant for
a specific protocol
The protocol specific fields in struct open_request were moved to a
class hierarchy, with the things that are common to all connection
oriented PF_INET protocols in struct inet_request_sock, the TCP ones
in tcp_request_sock, that is an inet_request_sock, that is an
open_request.
I.e. this uses the same approach used for the struct sock class
hierarchy, with sk_prot indicating if the protocol wants to use the
open_request infrastructure by filling in sk_prot->rsk_prot with an
or_calltable.
Results? Performance is improved and TCP v4 now uses only 64 bytes per
open request minisock, down from 96 without this patch :-)
Next changeset will rename some of the structs, fields and functions
mentioned above, struct or_calltable is way unclear, better name it
struct request_sock_ops, s/struct open_request/struct request_sock/g,
etc.
Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 12:46:52 +07:00
|
|
|
struct inet_request_sock *ireq = inet_rsk(req);
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
req->rcv_wnd = 0; /* So that tcp_send_synack() knows! */
|
2008-04-10 17:12:40 +07:00
|
|
|
req->cookie_ts = 0;
|
[NET] Generalise TCP's struct open_request minisock infrastructure
Kept this first changeset minimal, without changing existing names to
ease peer review.
Basicaly tcp_openreq_alloc now receives the or_calltable, that in turn
has two new members:
->slab, that replaces tcp_openreq_cachep
->obj_size, to inform the size of the openreq descendant for
a specific protocol
The protocol specific fields in struct open_request were moved to a
class hierarchy, with the things that are common to all connection
oriented PF_INET protocols in struct inet_request_sock, the TCP ones
in tcp_request_sock, that is an inet_request_sock, that is an
open_request.
I.e. this uses the same approach used for the struct sock class
hierarchy, with sk_prot indicating if the protocol wants to use the
open_request infrastructure by filling in sk_prot->rsk_prot with an
or_calltable.
Results? Performance is improved and TCP v4 now uses only 64 bytes per
open request minisock, down from 96 without this patch :-)
Next changeset will rename some of the structs, fields and functions
mentioned above, struct or_calltable is way unclear, better name it
struct request_sock_ops, s/struct open_request/struct request_sock/g,
etc.
Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 12:46:52 +07:00
|
|
|
tcp_rsk(req)->rcv_isn = TCP_SKB_CB(skb)->seq;
|
2012-08-31 19:29:11 +07:00
|
|
|
tcp_rsk(req)->rcv_nxt = TCP_SKB_CB(skb)->seq + 1;
|
2012-09-22 11:18:55 +07:00
|
|
|
tcp_rsk(req)->snt_synack = 0;
|
2005-04-17 05:20:36 +07:00
|
|
|
req->mss = rx_opt->mss_clamp;
|
|
|
|
req->ts_recent = rx_opt->saw_tstamp ? rx_opt->rcv_tsval : 0;
|
[NET] Generalise TCP's struct open_request minisock infrastructure
Kept this first changeset minimal, without changing existing names to
ease peer review.
Basicaly tcp_openreq_alloc now receives the or_calltable, that in turn
has two new members:
->slab, that replaces tcp_openreq_cachep
->obj_size, to inform the size of the openreq descendant for
a specific protocol
The protocol specific fields in struct open_request were moved to a
class hierarchy, with the things that are common to all connection
oriented PF_INET protocols in struct inet_request_sock, the TCP ones
in tcp_request_sock, that is an inet_request_sock, that is an
open_request.
I.e. this uses the same approach used for the struct sock class
hierarchy, with sk_prot indicating if the protocol wants to use the
open_request infrastructure by filling in sk_prot->rsk_prot with an
or_calltable.
Results? Performance is improved and TCP v4 now uses only 64 bytes per
open request minisock, down from 96 without this patch :-)
Next changeset will rename some of the structs, fields and functions
mentioned above, struct or_calltable is way unclear, better name it
struct request_sock_ops, s/struct open_request/struct request_sock/g,
etc.
Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 12:46:52 +07:00
|
|
|
ireq->tstamp_ok = rx_opt->tstamp_ok;
|
|
|
|
ireq->sack_ok = rx_opt->sack_ok;
|
|
|
|
ireq->snd_wscale = rx_opt->snd_wscale;
|
|
|
|
ireq->wscale_ok = rx_opt->wscale_ok;
|
|
|
|
ireq->acked = 0;
|
|
|
|
ireq->ecn_ok = 0;
|
2007-04-11 11:04:22 +07:00
|
|
|
ireq->rmt_port = tcp_hdr(skb)->source;
|
2008-10-01 21:46:49 +07:00
|
|
|
ireq->loc_port = tcp_hdr(skb)->dest;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2008-07-17 10:28:10 +07:00
|
|
|
extern void tcp_enter_memory_pressure(struct sock *sk);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
static inline int keepalive_intvl_when(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->keepalive_intvl ? : sysctl_tcp_keepalive_intvl;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline int keepalive_time_when(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->keepalive_time ? : sysctl_tcp_keepalive_time;
|
|
|
|
}
|
|
|
|
|
2009-08-29 13:48:54 +07:00
|
|
|
static inline int keepalive_probes(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->keepalive_probes ? : sysctl_tcp_keepalive_probes;
|
|
|
|
}
|
|
|
|
|
2010-04-27 01:33:27 +07:00
|
|
|
static inline u32 keepalive_time_elapsed(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
const struct inet_connection_sock *icsk = &tp->inet_conn;
|
|
|
|
|
|
|
|
return min_t(u32, tcp_time_stamp - icsk->icsk_ack.lrcvtime,
|
|
|
|
tcp_time_stamp - tp->rcv_tstamp);
|
|
|
|
}
|
|
|
|
|
2005-08-10 10:10:42 +07:00
|
|
|
static inline int tcp_fin_time(const struct sock *sk)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-08-10 10:10:42 +07:00
|
|
|
int fin_timeout = tcp_sk(sk)->linger2 ? : sysctl_tcp_fin_timeout;
|
|
|
|
const int rto = inet_csk(sk)->icsk_rto;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2005-08-10 10:10:42 +07:00
|
|
|
if (fin_timeout < (rto << 2) - (rto >> 1))
|
|
|
|
fin_timeout = (rto << 2) - (rto >> 1);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
return fin_timeout;
|
|
|
|
}
|
|
|
|
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_paws_check(const struct tcp_options_received *rx_opt,
|
|
|
|
int paws_win)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2009-03-14 21:23:03 +07:00
|
|
|
if ((s32)(rx_opt->ts_recent - rx_opt->rcv_tsval) <= paws_win)
|
2012-05-17 06:15:34 +07:00
|
|
|
return true;
|
2009-03-14 21:23:03 +07:00
|
|
|
if (unlikely(get_seconds() >= rx_opt->ts_recent_stamp + TCP_PAWS_24DAYS))
|
2012-05-17 06:15:34 +07:00
|
|
|
return true;
|
2010-12-17 05:08:34 +07:00
|
|
|
/*
|
|
|
|
* Some OSes send SYN and SYNACK messages with tsval=0 tsecr=0,
|
|
|
|
* then following tcp messages have valid values. Ignore 0 value,
|
|
|
|
* or else 'negative' tsval might forbid us to accept their packets.
|
|
|
|
*/
|
|
|
|
if (!rx_opt->ts_recent)
|
2012-05-17 06:15:34 +07:00
|
|
|
return true;
|
|
|
|
return false;
|
2009-03-14 21:23:03 +07:00
|
|
|
}
|
|
|
|
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_paws_reject(const struct tcp_options_received *rx_opt,
|
|
|
|
int rst)
|
2009-03-14 21:23:03 +07:00
|
|
|
{
|
|
|
|
if (tcp_paws_check(rx_opt, 0))
|
2012-05-17 06:15:34 +07:00
|
|
|
return false;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/* RST segments are not recommended to carry timestamp,
|
|
|
|
and, if they do, it is recommended to ignore PAWS because
|
|
|
|
"their cleanup function should take precedence over timestamps."
|
|
|
|
Certainly, it is mistake. It is necessary to understand the reasons
|
|
|
|
of this constraint to relax it: if peer reboots, clock may go
|
|
|
|
out-of-sync and half-open connections will not be reset.
|
|
|
|
Actually, the problem would be not existing if all
|
|
|
|
the implementations followed draft about maintaining clock
|
|
|
|
via reboots. Linux-2.2 DOES NOT!
|
|
|
|
|
|
|
|
However, we can relax time bounds for RST segments to MSL.
|
|
|
|
*/
|
2007-03-05 07:12:44 +07:00
|
|
|
if (rst && get_seconds() >= rx_opt->ts_recent_stamp + TCP_PAWS_MSL)
|
2012-05-17 06:15:34 +07:00
|
|
|
return false;
|
|
|
|
return true;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2008-07-17 10:21:42 +07:00
|
|
|
static inline void tcp_mib_init(struct net *net)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
/* See RFC 2012 */
|
2008-07-17 10:27:38 +07:00
|
|
|
TCP_ADD_STATS_USER(net, TCP_MIB_RTOALGORITHM, 1);
|
|
|
|
TCP_ADD_STATS_USER(net, TCP_MIB_RTOMIN, TCP_RTO_MIN*1000/HZ);
|
|
|
|
TCP_ADD_STATS_USER(net, TCP_MIB_RTOMAX, TCP_RTO_MAX*1000/HZ);
|
|
|
|
TCP_ADD_STATS_USER(net, TCP_MIB_MAXCONN, -1);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2007-09-21 01:30:48 +07:00
|
|
|
/* from STCP */
|
2008-09-21 11:25:15 +07:00
|
|
|
static inline void tcp_clear_retrans_hints_partial(struct tcp_sock *tp)
|
2007-09-21 01:40:37 +07:00
|
|
|
{
|
2005-11-11 08:14:59 +07:00
|
|
|
tp->lost_skb_hint = NULL;
|
2008-09-21 11:25:15 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_clear_all_retrans_hints(struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
tcp_clear_retrans_hints_partial(tp);
|
2005-11-11 08:14:59 +07:00
|
|
|
tp->retransmit_skb_hint = NULL;
|
2007-09-21 01:37:19 +07:00
|
|
|
}
|
|
|
|
|
2006-11-15 10:07:45 +07:00
|
|
|
/* MD5 Signature */
|
|
|
|
struct crypto_hash;
|
|
|
|
|
2012-01-31 12:18:33 +07:00
|
|
|
union tcp_md5_addr {
|
|
|
|
struct in_addr a4;
|
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
|
|
|
struct in6_addr a6;
|
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
2006-11-15 10:07:45 +07:00
|
|
|
/* - key database */
|
|
|
|
struct tcp_md5sig_key {
|
2012-01-31 12:18:33 +07:00
|
|
|
struct hlist_node node;
|
2006-11-15 10:07:45 +07:00
|
|
|
u8 keylen;
|
2012-01-31 12:18:33 +07:00
|
|
|
u8 family; /* AF_INET or AF_INET6 */
|
|
|
|
union tcp_md5_addr addr;
|
|
|
|
u8 key[TCP_MD5SIG_MAXKEYLEN];
|
|
|
|
struct rcu_head rcu;
|
2006-11-15 10:07:45 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
/* - sock block */
|
|
|
|
struct tcp_md5sig_info {
|
2012-01-31 12:18:33 +07:00
|
|
|
struct hlist_head head;
|
2012-02-01 01:45:40 +07:00
|
|
|
struct rcu_head rcu;
|
2006-11-15 10:07:45 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
/* - pseudo header */
|
|
|
|
struct tcp4_pseudohdr {
|
|
|
|
__be32 saddr;
|
|
|
|
__be32 daddr;
|
|
|
|
__u8 pad;
|
|
|
|
__u8 protocol;
|
|
|
|
__be16 len;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct tcp6_pseudohdr {
|
|
|
|
struct in6_addr saddr;
|
|
|
|
struct in6_addr daddr;
|
|
|
|
__be32 len;
|
|
|
|
__be32 protocol; /* including padding */
|
|
|
|
};
|
|
|
|
|
|
|
|
union tcp_md5sum_block {
|
|
|
|
struct tcp4_pseudohdr ip4;
|
2011-12-10 16:48:31 +07:00
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
2006-11-15 10:07:45 +07:00
|
|
|
struct tcp6_pseudohdr ip6;
|
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
|
|
|
/* - pool: digest algorithm, hash description and scratch buffer */
|
|
|
|
struct tcp_md5sig_pool {
|
|
|
|
struct hash_desc md5_desc;
|
|
|
|
union tcp_md5sum_block md5_blk;
|
|
|
|
};
|
|
|
|
|
|
|
|
/* - functions */
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp_v4_md5_hash_skb(char *md5_hash, struct tcp_md5sig_key *key,
|
2011-10-24 13:46:04 +07:00
|
|
|
const struct sock *sk,
|
|
|
|
const struct request_sock *req,
|
|
|
|
const struct sk_buff *skb);
|
2012-01-31 12:18:33 +07:00
|
|
|
extern int tcp_md5_do_add(struct sock *sk, const union tcp_md5_addr *addr,
|
|
|
|
int family, const u8 *newkey,
|
|
|
|
u8 newkeylen, gfp_t gfp);
|
|
|
|
extern int tcp_md5_do_del(struct sock *sk, const union tcp_md5_addr *addr,
|
|
|
|
int family);
|
|
|
|
extern struct tcp_md5sig_key *tcp_v4_md5_lookup(struct sock *sk,
|
|
|
|
struct sock *addr_sk);
|
2006-11-15 10:07:45 +07:00
|
|
|
|
2008-04-18 10:45:16 +07:00
|
|
|
#ifdef CONFIG_TCP_MD5SIG
|
2012-01-31 12:18:33 +07:00
|
|
|
extern struct tcp_md5sig_key *tcp_md5_do_lookup(struct sock *sk,
|
|
|
|
const union tcp_md5_addr *addr, int family);
|
|
|
|
#define tcp_twsk_md5_key(twsk) ((twsk)->tw_md5_key)
|
2008-04-18 10:45:16 +07:00
|
|
|
#else
|
2012-01-31 12:18:33 +07:00
|
|
|
static inline struct tcp_md5sig_key *tcp_md5_do_lookup(struct sock *sk,
|
|
|
|
const union tcp_md5_addr *addr,
|
|
|
|
int family)
|
|
|
|
{
|
|
|
|
return NULL;
|
|
|
|
}
|
2008-04-18 10:45:16 +07:00
|
|
|
#define tcp_twsk_md5_key(twsk) NULL
|
|
|
|
#endif
|
|
|
|
|
2013-05-20 13:52:26 +07:00
|
|
|
extern bool tcp_alloc_md5sig_pool(void);
|
2006-11-15 10:07:45 +07:00
|
|
|
|
2010-05-16 14:34:04 +07:00
|
|
|
extern struct tcp_md5sig_pool *tcp_get_md5sig_pool(void);
|
2013-05-20 13:52:26 +07:00
|
|
|
static inline void tcp_put_md5sig_pool(void)
|
|
|
|
{
|
|
|
|
local_bh_enable();
|
|
|
|
}
|
2010-05-16 14:34:04 +07:00
|
|
|
|
2011-10-24 12:52:35 +07:00
|
|
|
extern int tcp_md5_hash_header(struct tcp_md5sig_pool *, const struct tcphdr *);
|
2011-10-21 16:22:42 +07:00
|
|
|
extern int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *, const struct sk_buff *,
|
2012-04-15 12:58:06 +07:00
|
|
|
unsigned int header_len);
|
2008-07-19 14:01:42 +07:00
|
|
|
extern int tcp_md5_hash_key(struct tcp_md5sig_pool *hp,
|
2011-10-21 16:22:42 +07:00
|
|
|
const struct tcp_md5sig_key *key);
|
2006-11-15 10:07:45 +07:00
|
|
|
|
2012-08-31 19:29:11 +07:00
|
|
|
/* From tcp_fastopen.c */
|
|
|
|
extern void tcp_fastopen_cache_get(struct sock *sk, u16 *mss,
|
|
|
|
struct tcp_fastopen_cookie *cookie,
|
|
|
|
int *syn_loss, unsigned long *last_syn_loss);
|
|
|
|
extern void tcp_fastopen_cache_set(struct sock *sk, u16 mss,
|
|
|
|
struct tcp_fastopen_cookie *cookie,
|
|
|
|
bool syn_lost);
|
2012-07-19 13:43:07 +07:00
|
|
|
struct tcp_fastopen_request {
|
|
|
|
/* Fast Open cookie. Size 0 means a cookie request */
|
|
|
|
struct tcp_fastopen_cookie cookie;
|
|
|
|
struct msghdr *data; /* data in MSG_FASTOPEN */
|
|
|
|
u16 copied; /* queued in tcp_connect() */
|
|
|
|
};
|
|
|
|
void tcp_free_fastopen_req(struct tcp_sock *tp);
|
|
|
|
|
2012-08-31 19:29:11 +07:00
|
|
|
extern struct tcp_fastopen_context __rcu *tcp_fastopen_ctx;
|
|
|
|
int tcp_fastopen_reset_cipher(void *key, unsigned int len);
|
2013-08-09 04:06:22 +07:00
|
|
|
extern void tcp_fastopen_cookie_gen(__be32 src, __be32 dst,
|
|
|
|
struct tcp_fastopen_cookie *foc);
|
2012-08-31 19:29:11 +07:00
|
|
|
|
|
|
|
#define TCP_FASTOPEN_KEY_LENGTH 16
|
|
|
|
|
|
|
|
/* Fastopen key context */
|
|
|
|
struct tcp_fastopen_context {
|
2013-06-25 15:21:06 +07:00
|
|
|
struct crypto_cipher *tfm;
|
|
|
|
__u8 key[TCP_FASTOPEN_KEY_LENGTH];
|
|
|
|
struct rcu_head rcu;
|
2012-08-31 19:29:11 +07:00
|
|
|
};
|
|
|
|
|
2007-03-08 03:12:44 +07:00
|
|
|
/* write queue abstraction */
|
|
|
|
static inline void tcp_write_queue_purge(struct sock *sk)
|
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
|
|
|
|
|
|
|
while ((skb = __skb_dequeue(&sk->sk_write_queue)) != NULL)
|
2007-12-31 15:11:19 +07:00
|
|
|
sk_wmem_free_skb(sk, skb);
|
|
|
|
sk_mem_reclaim(sk);
|
2009-12-03 13:24:02 +07:00
|
|
|
tcp_clear_all_retrans_hints(tcp_sk(sk));
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
2011-10-21 16:22:42 +07:00
|
|
|
static inline struct sk_buff *tcp_write_queue_head(const struct sock *sk)
|
2007-03-08 03:12:44 +07:00
|
|
|
{
|
2008-09-23 14:50:13 +07:00
|
|
|
return skb_peek(&sk->sk_write_queue);
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
2011-10-21 16:22:42 +07:00
|
|
|
static inline struct sk_buff *tcp_write_queue_tail(const struct sock *sk)
|
2007-03-08 03:12:44 +07:00
|
|
|
{
|
2008-09-23 14:50:13 +07:00
|
|
|
return skb_peek_tail(&sk->sk_write_queue);
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
2011-10-21 16:22:42 +07:00
|
|
|
static inline struct sk_buff *tcp_write_queue_next(const struct sock *sk,
|
|
|
|
const struct sk_buff *skb)
|
2007-03-08 03:12:44 +07:00
|
|
|
{
|
2008-09-23 14:50:13 +07:00
|
|
|
return skb_queue_next(&sk->sk_write_queue, skb);
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
2011-10-21 16:22:42 +07:00
|
|
|
static inline struct sk_buff *tcp_write_queue_prev(const struct sock *sk,
|
|
|
|
const struct sk_buff *skb)
|
tcp: Try to restore large SKBs while SACK processing
During SACK processing, most of the benefits of TSO are eaten by
the SACK blocks that one-by-one fragment SKBs to MSS sized chunks.
Then we're in problems when cleanup work for them has to be done
when a large cumulative ACK comes. Try to return back to pre-split
state already while more and more SACK info gets discovered by
combining newly discovered SACK areas with the previous skb if
that's SACKed as well.
This approach has a number of benefits:
1) The processing overhead is spread more equally over the RTT
2) Write queue has less skbs to process (affect everything
which has to walk in the queue past the sacked areas)
3) Write queue is consistent whole the time, so no other parts
of TCP has to be aware of this (this was not the case with
some other approach that was, well, quite intrusive all
around).
4) Clean_rtx_queue can release most of the pages using single
put_page instead of previous PAGE_SIZE/mss+1 calls
In case a hole is fully filled by the new SACK block, we attempt
to combine the next skb too which allows construction of skbs
that are even larger than what tso split them to and it handles
hole per on every nth patterns that often occur during slow start
overshoot pretty nicely. Though this to be really useful also
a retransmission would have to get lost since cumulative ACKs
advance one hole at a time in the most typical case.
TODO: handle upwards only merging. That should be rather easy
when segment is fully sacked but I'm leaving that as future
work item (it won't make very large difference anyway since
this current approach already covers quite a lot of normal
cases).
I was earlier thinking of some sophisticated way of tracking
timestamps of the first and the last segment but later on
realized that it won't be that necessary at all to store the
timestamp of the last segment. The cases that can occur are
basically either:
1) ambiguous => no sensible measurement can be taken anyway
2) non-ambiguous is due to reordering => having the timestamp
of the last segment there is just skewing things more off
than does some good since the ack got triggered by one of
the holes (besides some substle issues that would make
determining right hole/skb even harder problem). Anyway,
it has nothing to do with this change then.
I choose to route some abnormal looking cases with goto noop,
some could be handled differently (eg., by stopping the
walking at that skb but again). In general, they either
shouldn't happen at all or are rare enough to make no difference
in practice.
In theory this change (as whole) could cause some macroscale
regression (global) because of cache misses that are taken over
the round-trip time but it gets very likely better because of much
less (local) cache misses per other write queue walkers and the
big recovery clearing cumulative ack.
Worth to note that these benefits would be very easy to get also
without TSO/GSO being on as long as the data is in pages so that
we can merge them. Currently I won't let that happen because
DSACK splitting at fragment that would mess up pcounts due to
sk_can_gso in tcp_set_skb_tso_segs. Once DSACKs fragments gets
avoided, we have some conditions that can be made less strict.
TODO: I will probably have to convert the excessive pointer
passing to struct sacktag_state... :-)
My testing revealed that considerable amount of skbs couldn't
be shifted because they were cloned (most likely still awaiting
tx reclaim)...
[The rest is considering future work instead since I got
repeatably EFAULT to tcpdump's recvfrom when I added
pskb_expand_head to deal with clones, so I separated that
into another, later patch]
...To counter that, I gave up on the fifth advantage:
5) When growing previous SACK block, less allocs for new skbs
are done, basically a new alloc is needed only when new hole
is detected and when the previous skb runs out of frags space
...which now only happens of if reclaim is fast enough to dispose
the clone before the SACK block comes in (the window is RTT long),
otherwise we'll have to alloc some.
With clones being handled I got these numbers (will be somewhat
worse without that), taken with fine-grained mibs:
TCPSackShifted 398
TCPSackMerged 877
TCPSackShiftFallback 320
TCPSACKCOLLAPSEFALLBACKGSO 0
TCPSACKCOLLAPSEFALLBACKSKBBITS 0
TCPSACKCOLLAPSEFALLBACKSKBDATA 0
TCPSACKCOLLAPSEFALLBACKBELOW 0
TCPSACKCOLLAPSEFALLBACKFIRST 1
TCPSACKCOLLAPSEFALLBACKPREVBITS 318
TCPSACKCOLLAPSEFALLBACKMSS 1
TCPSACKCOLLAPSEFALLBACKNOHEAD 0
TCPSACKCOLLAPSEFALLBACKSHIFT 0
TCPSACKCOLLAPSENOOPSEQ 0
TCPSACKCOLLAPSENOOPSMALLPCOUNT 0
TCPSACKCOLLAPSENOOPSMALLLEN 0
TCPSACKCOLLAPSEHOLE 12
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-25 12:20:15 +07:00
|
|
|
{
|
|
|
|
return skb_queue_prev(&sk->sk_write_queue, skb);
|
|
|
|
}
|
|
|
|
|
2007-03-08 03:12:44 +07:00
|
|
|
#define tcp_for_write_queue(skb, sk) \
|
2008-09-23 14:50:13 +07:00
|
|
|
skb_queue_walk(&(sk)->sk_write_queue, skb)
|
2007-03-08 03:12:44 +07:00
|
|
|
|
|
|
|
#define tcp_for_write_queue_from(skb, sk) \
|
2008-09-23 14:50:13 +07:00
|
|
|
skb_queue_walk_from(&(sk)->sk_write_queue, skb)
|
2007-03-08 03:12:44 +07:00
|
|
|
|
2007-12-02 05:48:02 +07:00
|
|
|
#define tcp_for_write_queue_from_safe(skb, tmp, sk) \
|
2008-09-23 14:50:13 +07:00
|
|
|
skb_queue_walk_from_safe(&(sk)->sk_write_queue, skb, tmp)
|
2007-12-02 05:48:02 +07:00
|
|
|
|
2011-10-21 16:22:42 +07:00
|
|
|
static inline struct sk_buff *tcp_send_head(const struct sock *sk)
|
2007-03-08 03:12:44 +07:00
|
|
|
{
|
|
|
|
return sk->sk_send_head;
|
|
|
|
}
|
|
|
|
|
2008-09-23 14:50:13 +07:00
|
|
|
static inline bool tcp_skb_is_last(const struct sock *sk,
|
|
|
|
const struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
return skb_queue_is_last(&sk->sk_write_queue, skb);
|
|
|
|
}
|
|
|
|
|
2011-10-21 16:22:42 +07:00
|
|
|
static inline void tcp_advance_send_head(struct sock *sk, const struct sk_buff *skb)
|
2007-03-08 03:12:44 +07:00
|
|
|
{
|
2008-09-23 14:50:13 +07:00
|
|
|
if (tcp_skb_is_last(sk, skb))
|
2007-03-08 03:12:44 +07:00
|
|
|
sk->sk_send_head = NULL;
|
2008-09-23 14:50:13 +07:00
|
|
|
else
|
|
|
|
sk->sk_send_head = tcp_write_queue_next(sk, skb);
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_check_send_head(struct sock *sk, struct sk_buff *skb_unlinked)
|
|
|
|
{
|
|
|
|
if (sk->sk_send_head == skb_unlinked)
|
|
|
|
sk->sk_send_head = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_init_send_head(struct sock *sk)
|
|
|
|
{
|
|
|
|
sk->sk_send_head = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void __tcp_add_write_queue_tail(struct sock *sk, struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
__skb_queue_tail(&sk->sk_write_queue, skb);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_add_write_queue_tail(struct sock *sk, struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
__tcp_add_write_queue_tail(sk, skb);
|
|
|
|
|
|
|
|
/* Queue it, remembering where we must start sending. */
|
2007-12-02 05:48:06 +07:00
|
|
|
if (sk->sk_send_head == NULL) {
|
2007-03-08 03:12:44 +07:00
|
|
|
sk->sk_send_head = skb;
|
2007-12-02 05:48:06 +07:00
|
|
|
|
|
|
|
if (tcp_sk(sk)->highest_sack == NULL)
|
|
|
|
tcp_sk(sk)->highest_sack = skb;
|
|
|
|
}
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void __tcp_add_write_queue_head(struct sock *sk, struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
__skb_queue_head(&sk->sk_write_queue, skb);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Insert buff after skb on the write queue of sk. */
|
|
|
|
static inline void tcp_insert_write_queue_after(struct sk_buff *skb,
|
|
|
|
struct sk_buff *buff,
|
|
|
|
struct sock *sk)
|
|
|
|
{
|
2008-04-14 14:05:09 +07:00
|
|
|
__skb_queue_after(&sk->sk_write_queue, skb, buff);
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
2008-09-22 11:28:51 +07:00
|
|
|
/* Insert new before skb on the write queue of sk. */
|
2007-03-08 03:12:44 +07:00
|
|
|
static inline void tcp_insert_write_queue_before(struct sk_buff *new,
|
|
|
|
struct sk_buff *skb,
|
|
|
|
struct sock *sk)
|
|
|
|
{
|
2008-09-22 11:28:51 +07:00
|
|
|
__skb_queue_before(&sk->sk_write_queue, skb, new);
|
2007-11-20 14:24:09 +07:00
|
|
|
|
|
|
|
if (sk->sk_send_head == skb)
|
|
|
|
sk->sk_send_head = new;
|
2007-03-08 03:12:44 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_unlink_write_queue(struct sk_buff *skb, struct sock *sk)
|
|
|
|
{
|
|
|
|
__skb_unlink(skb, &sk->sk_write_queue);
|
|
|
|
}
|
|
|
|
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_write_queue_empty(struct sock *sk)
|
2007-03-08 03:12:44 +07:00
|
|
|
{
|
|
|
|
return skb_queue_empty(&sk->sk_write_queue);
|
|
|
|
}
|
|
|
|
|
2009-12-09 05:26:13 +07:00
|
|
|
static inline void tcp_push_pending_frames(struct sock *sk)
|
|
|
|
{
|
|
|
|
if (tcp_send_head(sk)) {
|
|
|
|
struct tcp_sock *tp = tcp_sk(sk);
|
|
|
|
|
|
|
|
__tcp_push_pending_frames(sk, tcp_current_mss(sk), tp->nonagle);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-02-28 05:52:52 +07:00
|
|
|
/* Start sequence of the skb just after the highest skb with SACKed
|
|
|
|
* bit, valid only if sacked_out > 0 or when the caller has ensured
|
|
|
|
* validity by itself.
|
2007-11-16 10:41:46 +07:00
|
|
|
*/
|
|
|
|
static inline u32 tcp_highest_sack_seq(struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
if (!tp->sacked_out)
|
|
|
|
return tp->snd_una;
|
2007-12-02 05:48:06 +07:00
|
|
|
|
|
|
|
if (tp->highest_sack == NULL)
|
|
|
|
return tp->snd_nxt;
|
|
|
|
|
2007-11-16 10:41:46 +07:00
|
|
|
return TCP_SKB_CB(tp->highest_sack)->seq;
|
|
|
|
}
|
|
|
|
|
2007-12-02 05:48:06 +07:00
|
|
|
static inline void tcp_advance_highest_sack(struct sock *sk, struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
tcp_sk(sk)->highest_sack = tcp_skb_is_last(sk, skb) ? NULL :
|
|
|
|
tcp_write_queue_next(sk, skb);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct sk_buff *tcp_highest_sack(struct sock *sk)
|
|
|
|
{
|
|
|
|
return tcp_sk(sk)->highest_sack;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void tcp_highest_sack_reset(struct sock *sk)
|
|
|
|
{
|
|
|
|
tcp_sk(sk)->highest_sack = tcp_write_queue_head(sk);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Called when old skb is about to be deleted (to be combined with new skb) */
|
|
|
|
static inline void tcp_highest_sack_combine(struct sock *sk,
|
|
|
|
struct sk_buff *old,
|
|
|
|
struct sk_buff *new)
|
|
|
|
{
|
|
|
|
if (tcp_sk(sk)->sacked_out && (old == tcp_sk(sk)->highest_sack))
|
|
|
|
tcp_sk(sk)->highest_sack = new;
|
|
|
|
}
|
|
|
|
|
2010-02-18 09:45:45 +07:00
|
|
|
/* Determines whether this is a thin stream (which may suffer from
|
|
|
|
* increased latency). Used to trigger latency-reducing mechanisms.
|
|
|
|
*/
|
2012-05-17 06:15:34 +07:00
|
|
|
static inline bool tcp_stream_is_thin(struct tcp_sock *tp)
|
2010-02-18 09:45:45 +07:00
|
|
|
{
|
|
|
|
return tp->packets_out < 4 && !tcp_in_initial_slowstart(tp);
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* /proc */
|
|
|
|
enum tcp_seq_states {
|
|
|
|
TCP_SEQ_STATE_LISTENING,
|
|
|
|
TCP_SEQ_STATE_OPENREQ,
|
|
|
|
TCP_SEQ_STATE_ESTABLISHED,
|
|
|
|
TCP_SEQ_STATE_TIME_WAIT,
|
|
|
|
};
|
|
|
|
|
2011-10-30 13:46:30 +07:00
|
|
|
int tcp_seq_open(struct inode *inode, struct file *file);
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
struct tcp_seq_afinfo {
|
2011-10-30 13:46:30 +07:00
|
|
|
char *name;
|
|
|
|
sa_family_t family;
|
|
|
|
const struct file_operations *seq_fops;
|
|
|
|
struct seq_operations seq_ops;
|
2005-04-17 05:20:36 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
struct tcp_iter_state {
|
2008-04-14 12:11:14 +07:00
|
|
|
struct seq_net_private p;
|
2005-04-17 05:20:36 +07:00
|
|
|
sa_family_t family;
|
|
|
|
enum tcp_seq_states state;
|
|
|
|
struct sock *syn_wait_sk;
|
2012-05-24 14:10:10 +07:00
|
|
|
int bucket, offset, sbucket, num;
|
|
|
|
kuid_t uid;
|
2010-06-07 14:43:42 +07:00
|
|
|
loff_t last_pos;
|
2005-04-17 05:20:36 +07:00
|
|
|
};
|
|
|
|
|
2008-03-21 18:14:45 +07:00
|
|
|
extern int tcp_proc_register(struct net *net, struct tcp_seq_afinfo *afinfo);
|
|
|
|
extern void tcp_proc_unregister(struct net *net, struct tcp_seq_afinfo *afinfo);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2005-08-16 12:18:02 +07:00
|
|
|
extern struct request_sock_ops tcp_request_sock_ops;
|
2008-02-08 12:49:26 +07:00
|
|
|
extern struct request_sock_ops tcp6_request_sock_ops;
|
2005-08-16 12:18:02 +07:00
|
|
|
|
2008-06-15 07:04:49 +07:00
|
|
|
extern void tcp_v4_destroy_sock(struct sock *sk);
|
2005-08-16 12:18:02 +07:00
|
|
|
|
2011-11-15 22:29:55 +07:00
|
|
|
extern struct sk_buff *tcp_tso_segment(struct sk_buff *skb,
|
|
|
|
netdev_features_t features);
|
2008-12-16 14:43:36 +07:00
|
|
|
extern struct sk_buff **tcp_gro_receive(struct sk_buff **head,
|
|
|
|
struct sk_buff *skb);
|
|
|
|
extern int tcp_gro_complete(struct sk_buff *skb);
|
2013-06-07 12:11:46 +07:00
|
|
|
|
|
|
|
extern void __tcp_v4_send_check(struct sk_buff *skb, __be32 saddr,
|
|
|
|
__be32 daddr);
|
2006-06-22 17:02:40 +07:00
|
|
|
|
tcp: TCP_NOTSENT_LOWAT socket option
Idea of this patch is to add optional limitation of number of
unsent bytes in TCP sockets, to reduce usage of kernel memory.
TCP receiver might announce a big window, and TCP sender autotuning
might allow a large amount of bytes in write queue, but this has little
performance impact if a large part of this buffering is wasted :
Write queue needs to be large only to deal with large BDP, not
necessarily to cope with scheduling delays (incoming ACKS make room
for the application to queue more bytes)
For most workloads, using a value of 128 KB or less is OK to give
applications enough time to react to POLLOUT events in time
(or being awaken in a blocking sendmsg())
This patch adds two ways to set the limit :
1) Per socket option TCP_NOTSENT_LOWAT
2) A sysctl (/proc/sys/net/ipv4/tcp_notsent_lowat) for sockets
not using TCP_NOTSENT_LOWAT socket option (or setting a zero value)
Default value being UINT_MAX (0xFFFFFFFF), meaning this has no effect.
This changes poll()/select()/epoll() to report POLLOUT
only if number of unsent bytes is below tp->nosent_lowat
Note this might increase number of sendmsg()/sendfile() calls
when using non blocking sockets,
and increase number of context switches for blocking sockets.
Note this is not related to SO_SNDLOWAT (as SO_SNDLOWAT is
defined as :
Specify the minimum number of bytes in the buffer until
the socket layer will pass the data to the protocol)
Tested:
netperf sessions, and watching /proc/net/protocols "memory" column for TCP
With 200 concurrent netperf -t TCP_STREAM sessions, amount of kernel memory
used by TCP buffers shrinks by ~55 % (20567 pages instead of 45458)
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 45458 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 45458 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 20567 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 20567 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
Using 128KB has no bad effect on the throughput or cpu usage
of a single flow, although there is an increase of context switches.
A bonus is that we hold socket lock for a shorter amount
of time and should improve latencies of ACK processing.
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1651584 6291456 16384 20.00 17447.90 10^6bits/s 3.13 S -1.00 U 0.353 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
412,514 context-switches
200.034645535 seconds time elapsed
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1593240 6291456 16384 20.00 17321.16 10^6bits/s 3.35 S -1.00 U 0.381 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
2,675,818 context-switches
200.029651391 seconds time elapsed
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-By: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-07-23 10:27:07 +07:00
|
|
|
static inline u32 tcp_notsent_lowat(const struct tcp_sock *tp)
|
|
|
|
{
|
|
|
|
return tp->notsent_lowat ?: sysctl_tcp_notsent_lowat;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool tcp_stream_memory_free(const struct sock *sk)
|
|
|
|
{
|
|
|
|
const struct tcp_sock *tp = tcp_sk(sk);
|
|
|
|
u32 notsent_bytes = tp->write_seq - tp->snd_nxt;
|
|
|
|
|
|
|
|
return notsent_bytes < tcp_notsent_lowat(tp);
|
|
|
|
}
|
|
|
|
|
2005-08-16 12:18:02 +07:00
|
|
|
#ifdef CONFIG_PROC_FS
|
2010-07-11 03:41:06 +07:00
|
|
|
extern int tcp4_proc_init(void);
|
2005-08-16 12:18:02 +07:00
|
|
|
extern void tcp4_proc_exit(void);
|
|
|
|
#endif
|
|
|
|
|
2006-11-15 10:07:45 +07:00
|
|
|
/* TCP af-specific functions */
|
|
|
|
struct tcp_sock_af_ops {
|
|
|
|
#ifdef CONFIG_TCP_MD5SIG
|
|
|
|
struct tcp_md5sig_key *(*md5_lookup) (struct sock *sk,
|
|
|
|
struct sock *addr_sk);
|
|
|
|
int (*calc_md5_hash) (char *location,
|
|
|
|
struct tcp_md5sig_key *md5,
|
2011-10-24 13:46:04 +07:00
|
|
|
const struct sock *sk,
|
|
|
|
const struct request_sock *req,
|
|
|
|
const struct sk_buff *skb);
|
2006-11-15 10:07:45 +07:00
|
|
|
int (*md5_parse) (struct sock *sk,
|
|
|
|
char __user *optval,
|
|
|
|
int optlen);
|
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
|
|
|
struct tcp_request_sock_ops {
|
|
|
|
#ifdef CONFIG_TCP_MD5SIG
|
|
|
|
struct tcp_md5sig_key *(*md5_lookup) (struct sock *sk,
|
|
|
|
struct request_sock *req);
|
2009-07-16 12:04:51 +07:00
|
|
|
int (*calc_md5_hash) (char *location,
|
|
|
|
struct tcp_md5sig_key *md5,
|
2011-10-24 13:46:04 +07:00
|
|
|
const struct sock *sk,
|
|
|
|
const struct request_sock *req,
|
|
|
|
const struct sk_buff *skb);
|
2006-11-15 10:07:45 +07:00
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
2013-06-07 12:11:46 +07:00
|
|
|
extern int tcpv4_offload_init(void);
|
|
|
|
|
2008-03-01 02:13:15 +07:00
|
|
|
extern void tcp_v4_init(void);
|
2005-08-16 12:18:02 +07:00
|
|
|
extern void tcp_init(void);
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
#endif /* _TCP_H */
|