d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
aebcd3034e
freebsd-dev/sys/netinet/tcp_input.c

3867 lines
110 KiB
C
Raw Normal View History

/* -> /*- for license, minor formatting changes
2005-01-07 01:45:51 +00:00
/*-
Finish 4.4-Lite-2 merge: randomize TCP initial sequence numbers to make ISS-guessing spoofing attacks harder.
1995-10-03 16:54:17 +00:00
* Copyright (c) 1982, 1986, 1988, 1990, 1993, 1994, 1995
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* The Regents of the University of California. All rights reserved.
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
* Copyright (c) 2007-2008,2010
* Swinburne University of Technology, Melbourne, Australia.
* Copyright (c) 2009-2010 Lawrence Stewart <lstewart@freebsd.org>
* Copyright (c) 2010 The FreeBSD Foundation
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
* Copyright (c) 2010-2011 Juniper Networks, Inc.
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
* All rights reserved.
*
* Portions of this software were developed at the Centre for Advanced Internet
Use the full and proper company name for Swinburne University of Technology throughout the source tree. Requested by: Grenville Armitage, Director of CAIA at Swinburne University of Technology MFC after: 3 days
2011-04-12 08:13:18 +00:00
* Architectures, Swinburne University of Technology, by Lawrence Stewart,
* James Healy and David Hayes, made possible in part by a grant from the Cisco
* University Research Program Fund at Community Foundation Silicon Valley.
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
*
* Portions of this software were developed at the Centre for Advanced
* Internet Architectures, Swinburne University of Technology, Melbourne,
* Australia by David Hayes under sponsorship from the FreeBSD Foundation.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
* Portions of this software were developed by Robert N. M. Watson under
* contract to Juniper Networks, Inc.
*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
Finish 4.4-Lite-2 merge: randomize TCP initial sequence numbers to make ISS-guessing spoofing attacks harder.
1995-10-03 16:54:17 +00:00
* @(#)tcp_input.c 8.12 (Berkeley) 5/24/95
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Add FBSDID to all files in netinet so that people can more easily include file version information in bug reports. Approved by: re (kensmith)
2007-10-07 20:44:24 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Initial import of RFC 2385 (TCP-MD5) digest support. This is the first of two commits; bringing in the kernel support first. This can be enabled by compiling a kernel with options TCP_SIGNATURE and FAST_IPSEC. For the uninitiated, this is a TCP option which provides for a means of authenticating TCP sessions which came into being before IPSEC. It is still relevant today, however, as it is used by many commercial router vendors, particularly with BGP, and as such has become a requirement for interconnect at many major Internet points of presence. Several parts of the TCP and IP headers, including the segment payload, are digested with MD5, including a shared secret. The PF_KEY interface is used to manage the secrets using security associations in the SADB. There is a limitation here in that as there is no way to map a TCP flow per-port back to an SPI without polluting tcpcb or using the SPD; the code to do the latter is unstable at this time. Therefore this code only supports per-host keying granularity. Whilst FAST_IPSEC is mutually exclusive with KAME IPSEC (and thus IPv6), TCP_SIGNATURE applies only to IPv4. For the vast majority of prospective users of this feature, this will not pose any problem. This implementation is output-only; that is, the option is honoured when responding to a host initiating a TCP session, but no effort is made [yet] to authenticate inbound traffic. This is, however, sufficient to interwork with Cisco equipment. Tested with a Cisco 2501 running IOS 12.0(27), and Quagga 0.96.4 with local patches. Patches for tcpdump to validate TCP-MD5 sessions are also available from me upon request. Sponsored by: sentex.net
2004-02-11 04:26:04 +00:00
#include "opt_inet.h"
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#include "opt_inet6.h"
Fixed the problem that IPsec connection hangs when bigger data is sent. -opt_ipsec.h was missing on some tcp files (sorry for basic mistake) -made buildable as above fix -also added some missing IPv4 mapped IPv6 addr consideration into ipsec4_getpolicybysock
2000-01-15 14:56:38 +00:00
#include "opt_ipsec.h"
Make TCPDEBUG a new-style option.
1997-09-16 18:36:06 +00:00
#include "opt_tcpdebug.h"
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
Start adding new style sysctl here too.
1995-11-09 20:23:09 +00:00
#include <sys/kernel.h>
- Add some helper hook points to the TCP stack. The hooks allow Khelp modules to access inbound/outbound events and associated data for established TCP connections. The hooks only run if at least one hook function is registered for the hook point, ensuring the impact on the stack is effectively nil when no TCP Khelp modules are loaded. struct tcp_hhook_data is passed as contextual data to any registered Khelp module hook functions. - Add an OSD (Object Specific Data) pointer to struct tcpcb to allow Khelp modules to associate per-connection data with the TCP control block. - Bump __FreeBSD_version and add a note to UPDATING regarding to ABI changes introduced by this commit and r216753. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: FreeBSD Foundation Reviewed by: bz, others along the way MFC after: 3 months
2010-12-28 12:13:30 +00:00
#include <sys/hhook.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/malloc.h>
#include <sys/mbuf.h>
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
#include <sys/proc.h> /* for proc0 declaration */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/protosw.h>
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
#include <sys/sdt.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <sys/signalvar.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/socket.h>
#include <sys/socketvar.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <sys/sysctl.h>
Log TCP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.tcp.log_in_vain: 1 Log UDP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.udp.log_in_vain: 1 Suggested by: Warren Toomey <wkt@cs.adfa.oz.au>
1996-04-04 10:46:44 +00:00
#include <sys/syslog.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <sys/systm.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Finish 4.4-Lite-2 merge: randomize TCP initial sequence numbers to make ISS-guessing spoofing attacks harder.
1995-10-03 16:54:17 +00:00
#include <machine/cpu.h> /* before tcp_seq.h, for tcp_random18() */
Convert the tcp segment reassembly queue to UMA and limit the maximum amount of segments it will hold. The following tuneables and sysctls control the behaviour of the tcp segment reassembly queue: net.inet.tcp.reass.maxsegments (loader tuneable) specifies the maximum number of segments all tcp reassemly queues can hold (defaults to 1/16 of nmbclusters). net.inet.tcp.reass.maxqlen specifies the maximum number of segments any individual tcp session queue can hold (defaults to 48). net.inet.tcp.reass.cursegments (readonly) counts the number of segments currently in all reassembly queues. net.inet.tcp.reass.overflows (readonly) counts how often either the global or local queue limit has been reached. Tested by: bms, silby Reviewed by: bms, silby
2004-02-24 15:27:41 +00:00
#include <vm/uma.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <net/if.h>
The r48589 promised to remove implicit inclusion of if_var.h soon. Prepare to this event, adding if_var.h to files that do need it. Also, include all includes that now are included due to implicit pollution via if_var.h Sponsored by: Netflix Sponsored by: Nginx, Inc.
2013-10-26 17:58:36 +00:00
#include <net/if_var.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <net/route.h>
Merge the remainder of kern_vimage.c and vimage.h into vnet.c and vnet.h, we now use jails (rather than vimages) as the abstraction for virtualization management, and what remained was specific to virtual network stacks. Minor cleanups are done in the process, and comments updated to reflect these changes. Reviewed by: bz Approved by: re (vimage blanket)
2009-08-01 19:26:27 +00:00
#include <net/vnet.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
#define TCPSTATES /* for logging */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/in.h>
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
#include <netinet/in_kdtrace.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <netinet/in_pcb.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/in_systm.h>
#include <netinet/ip.h>
Remove unneeded includes and provide more accurate description to others. Submitted by: garys PR: kern/86437
2006-02-18 17:05:00 +00:00
#include <netinet/ip_icmp.h> /* required for icmp_var.h */
#include <netinet/icmp_var.h> /* for ICMP_BANDLIM */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/ip_var.h>
Consolidate all IP Options handling functions into ip_options.[ch] and include ip_options.h into all files making use of IP Options functions. From ip_input.c rev 1.306: ip_dooptions(struct mbuf *m, int pass) save_rte(m, option, dst) ip_srcroute(m0) ip_stripoptions(m, mopt) From ip_output.c rev 1.249: ip_insertoptions(m, opt, phlen) ip_optcopy(ip, jp) ip_pcbopts(struct inpcb *inp, int optname, struct mbuf *m) No functional changes in this commit. Discussed with: rwatson Sponsored by: TCP/IP Optimization Fundraise 2005
2005-11-18 20:12:40 +00:00
#include <netinet/ip_options.h>
sync with kame tree as of july00. tons of bug fixes/improvements. API changes: - additional IPv6 ioctls - IPsec PF_KEY API was changed, it is mandatory to upgrade setkey(8). (also syntax change)
2000-07-04 16:35:15 +00:00
#include <netinet/ip6.h>
#include <netinet/icmp6.h>
#include <netinet6/in6_pcb.h>
Cleanup TCP files from unnecessary interface related includes.
2016-01-21 22:24:20 +00:00
#include <netinet6/in6_var.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <netinet6/ip6_var.h>
#include <netinet6/nd6.h>
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
#ifdef TCP_RFC7413
#include <netinet/tcp_fastopen.h>
#endif
- Rename cc.h to more meaningful tcp_cc.h. - Declare it a kernel only include, which it already is. - Don't include tcp.h implicitly from tcp_cc.h
2016-01-21 22:34:51 +00:00
#include <netinet/tcp.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#include <netinet6/tcp6_var.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/tcpip.h>
Rename netinet/tcp_cc.h to netinet/cc/cc.h. Discussed with: lstewart
2016-01-27 17:59:39 +00:00
#include <netinet/cc/cc.h>
There are times when it would be really nice to have a record of the last few packets and/or state transitions from each TCP socket. That would help with narrowing down certain problems we see in the field that are hard to reproduce without understanding the history of how we got into a certain state. This change provides just that. It saves copies of the last N packets in a list in the tcpcb. When the tcpcb is destroyed, the list is freed. I thought this was likely to be more performance-friendly than saving copies of the tcpcb. Plus, with the packets, you should be able to reverse-engineer what happened to the tcpcb. To enable the feature, you will need to compile a kernel with the TCPPCAP option. Even then, the feature defaults to being deactivated. You can activate it by setting a positive value for the number of captured packets. You can do that on either a global basis or on a per-socket basis (via a setsockopt call). There is no way to get the packets out of the kernel other than using kmem or getting a coredump. I thought that would help some of the legal/privacy concerns regarding such a feature. However, it should be possible to add a future effort to export them in PCAP format. I tested this at low scale, and found that there were no mbuf leaks and the peak mbuf usage appeared to be unchanged with and without the feature. The main performance concern I can envision is the number of mbufs that would be used on systems with a large number of sockets. If you save five packets per direction per socket and have 3,000 sockets, that will consume at least 30,000 mbufs just to keep these packets. I tried to reduce the concerns associated with this by limiting the number of clusters (not mbufs) that could be used for this feature. Again, in my testing, that appears to work correctly. Differential Revision: D3100 Submitted by: Jonathan Looney <jlooney at juniper dot net> Reviewed by: gnn, hiren
2015-10-14 00:35:37 +00:00
#ifdef TCPPCAP
#include <netinet/tcp_pcap.h>
#endif
Export the contents of the syncache to netstat. Approved by: re (kensmith) MFC after: 2 weeks
2007-07-27 00:57:06 +00:00
#include <netinet/tcp_syncache.h>
Made TCPDEBUG truely optional. Based on changes I made in FreeBSD 1.1.5. Fixed somebody's idea of a joke - about the first half of the lines in in_proto.c were spaced over by one space.
1994-09-15 10:36:56 +00:00
#ifdef TCPDEBUG
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/tcp_debug.h>
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#endif /* TCPDEBUG */
- Updated TOE support in the kernel. - Stateful TCP offload drivers for Terminator 3 and 4 (T3 and T4) ASICs. These are available as t3_tom and t4_tom modules that augment cxgb(4) and cxgbe(4) respectively. The cxgb/cxgbe drivers continue to work as usual with or without these extra features. - iWARP driver for Terminator 3 ASIC (kernel verbs). T4 iWARP in the works and will follow soon. Build-tested with make universe. 30s overview ============ What interfaces support TCP offload? Look for TOE4 and/or TOE6 in the capabilities of an interface: # ifconfig -m | grep TOE Enable/disable TCP offload on an interface (just like any other ifnet capability): # ifconfig cxgbe0 toe # ifconfig cxgbe0 -toe Which connections are offloaded? Look for toe4 and/or toe6 in the output of netstat and sockstat: # netstat -np tcp | grep toe # sockstat -46c | grep toe Reviewed by: bz, gnn Sponsored by: Chelsio communications. MFC after: ~3 months (after 9.1, and after ensuring MFC is feasible)
2012-06-19 07:34:13 +00:00
#ifdef TCP_OFFLOAD
#include <netinet/tcp_offload.h>
#endif
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
Commit the change from FAST_IPSEC to IPSEC. The FAST_IPSEC option is now deprecated, as well as the KAME IPsec code. What was FAST_IPSEC is now IPSEC. Approved by: re Sponsored by: Secure Computing
2007-07-03 12:13:45 +00:00
#ifdef IPSEC
Tie new "Fast IPsec" code into the build. This involves the usual configuration stuff as well as conditional code in the IPv4 and IPv6 areas. Everything is conditional on FAST_IPSEC which is mutually exclusive with IPSEC (KAME IPsec implmentation). As noted previously, don't use FAST_IPSEC with INET6 at the moment. Reviewed by: KAME, rwatson Approved by: silence Supported by: Vernier Networks
2002-10-16 02:25:05 +00:00
#include <netipsec/ipsec.h>
#include <netipsec/ipsec6.h>
Commit the change from FAST_IPSEC to IPSEC. The FAST_IPSEC option is now deprecated, as well as the KAME IPsec code. What was FAST_IPSEC is now IPSEC. Approved by: re Sponsored by: Secure Computing
2007-07-03 12:13:45 +00:00
#endif /*IPSEC*/
Tie new "Fast IPsec" code into the build. This involves the usual configuration stuff as well as conditional code in the IPv4 and IPv6 areas. Everything is conditional on FAST_IPSEC which is mutually exclusive with IPSEC (KAME IPsec implmentation). As noted previously, don't use FAST_IPSEC with INET6 at the moment. Reviewed by: KAME, rwatson Approved by: silence Supported by: Vernier Networks
2002-10-16 02:25:05 +00:00
Add support for offloading IP/TCP/UDP checksums to NIC hardware which supports them.
2000-03-27 19:14:27 +00:00
#include <machine/in_cksum.h>
Complete break-out of sys/sys/mac.h into sys/security/mac/mac_framework.h begun with a repo-copy of mac.h to mac_framework.h. sys/mac.h now contains the userspace and user<->kernel API and definitions, with all in-kernel interfaces moved to mac_framework.h, which is now included across most of the kernel instead. This change is the first step in a larger cleanup and sweep of MAC Framework interfaces in the kernel, and will not be MFC'd. Obtained from: TrustedBSD Project Sponsored by: SPARTA
2006-10-22 11:52:19 +00:00
#include <security/mac/mac_framework.h>
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
const int tcprexmtthresh = 3;
New style sysctl & staticize alot of stuff.
1995-11-14 20:34:56 +00:00
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
int tcp_log_in_vain = 0;
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, log_in_vain, CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&tcp_log_in_vain, 0,
"Log all incoming TCP segments to closed ports");
Log TCP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.tcp.log_in_vain: 1 Log UDP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.udp.log_in_vain: 1 Suggested by: Warren Toomey <wkt@cs.adfa.oz.au>
1996-04-04 10:46:44 +00:00
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, blackhole) = 0;
#define V_blackhole VNET(blackhole)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, blackhole, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(blackhole), 0,
"Do not send RST on segments to closed ports");
Add net.inet.tcp.blackhole and net.inet.udp.blackhole sysctl knobs. With these knobs on, refused connection attempts are dropped without sending a RST, or Port unreachable in the UDP case. In the TCP case, sending of RST is inhibited iff the incoming segment was a SYN. Docs and rc.conf settings to follow.
1999-08-17 12:17:53 +00:00
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_delack_enabled) = 1;
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, delayed_ack, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_delack_enabled), 0,
Add sysctl descriptions to many SYSCTL_XXXs PR: kern/11197 Submitted by: Adrian Chadd <adrian@FreeBSD.org> Reviewed by: billf(spelling/style/minor nits) Looked at by: bde(style)
1999-05-03 23:57:32 +00:00
"Delay ACK to try and piggyback it onto a data packet");
Changes to support the addition of a new sysctl variable: net.inet.tcp.delack_enabled Which defaults to 1 and can be set to 0 to disable TCP delayed-ack processing (i.e. all acks are immediate).
1998-02-26 05:25:39 +00:00
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, drop_synfin) = 0;
#define V_drop_synfin VNET(drop_synfin)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(drop_synfin), 0,
"Drop TCP packets with SYN+FIN set");
Fix some more disordering, as well as the description string for the net.inet.tcp.drop_synfin sysctl, which for some mysterious reason said "Drop TCP packets with FIN+ACK set" (instead of "...with SYN+FIN set")
1999-09-14 16:14:05 +00:00
Calculate the correct amount of bytes that are in-flight for a connection as suggested by RFC 6675. Currently differnt places in the stack tries to guess this in suboptimal ways. The main problem is that current calculations don't take sacked bytes into account. Sacked bytes are the bytes receiver acked via SACK option. This is suboptimal because it assumes that network has more outstanding (unacked) bytes than the actual value and thus sends less data by setting congestion window lower than what's possible which in turn may cause slower recovery from losses. As an example, one of the current calculations looks something like this: snd_nxt - snd_fack + sackhint.sack_bytes_rexmit New proposal from RFC 6675 is: snd_max - snd_una - sackhint.sacked_bytes + sackhint.sack_bytes_rexmit which takes sacked bytes into account which is a new addition to the sackhint struct. Only thing we are missing from RFC 6675 is isLost() i.e. segment being considered lost and thus adjusting pipe based on that which makes this calculation a bit on conservative side. The approach is very simple. We already process each ack with sack info in tcp_sack_doack() and extract sack blocks/holes out of it. We'd now also track this new variable sacked_bytes which keeps track of total sacked bytes reported. One downside to this approach is that we may get incorrect count of sacked_bytes if the other end decides to drop sack info in the ack because of memory pressure or some other reasons. But in this (not very likely) case also the pipe calculation would be conservative which is okay as opposed to being aggressive in sending packets into the network. Next step is to use this more accurate pipe estimation to drive congestion window adjustments. In collaboration with: rrs Reviewed by: jason_eggnet dot com, rrs MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D3971
2015-10-28 22:57:51 +00:00
VNET_DEFINE(int, tcp_do_rfc6675_pipe) = 0;
Improve the sysctl node name. X-MFC with: r290122 Sponsored by: Limelight Networks
2015-11-05 02:09:48 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, rfc6675_pipe, CTLFLAG_VNET | CTLFLAG_RW,
Calculate the correct amount of bytes that are in-flight for a connection as suggested by RFC 6675. Currently differnt places in the stack tries to guess this in suboptimal ways. The main problem is that current calculations don't take sacked bytes into account. Sacked bytes are the bytes receiver acked via SACK option. This is suboptimal because it assumes that network has more outstanding (unacked) bytes than the actual value and thus sends less data by setting congestion window lower than what's possible which in turn may cause slower recovery from losses. As an example, one of the current calculations looks something like this: snd_nxt - snd_fack + sackhint.sack_bytes_rexmit New proposal from RFC 6675 is: snd_max - snd_una - sackhint.sacked_bytes + sackhint.sack_bytes_rexmit which takes sacked bytes into account which is a new addition to the sackhint struct. Only thing we are missing from RFC 6675 is isLost() i.e. segment being considered lost and thus adjusting pipe based on that which makes this calculation a bit on conservative side. The approach is very simple. We already process each ack with sack info in tcp_sack_doack() and extract sack blocks/holes out of it. We'd now also track this new variable sacked_bytes which keeps track of total sacked bytes reported. One downside to this approach is that we may get incorrect count of sacked_bytes if the other end decides to drop sack info in the ack because of memory pressure or some other reasons. But in this (not very likely) case also the pipe calculation would be conservative which is okay as opposed to being aggressive in sending packets into the network. Next step is to use this more accurate pipe estimation to drive congestion window adjustments. In collaboration with: rrs Reviewed by: jason_eggnet dot com, rrs MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D3971
2015-10-28 22:57:51 +00:00
&VNET_NAME(tcp_do_rfc6675_pipe), 0,
"Use calculated pipe/in-flight bytes per RFC 6675");
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_do_rfc3042) = 1;
#define V_tcp_do_rfc3042 VNET(tcp_do_rfc3042)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, rfc3042, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_do_rfc3042), 0,
"Enable RFC 3042 (Limited Transmit)");
Implement the Limited Transmit algorithm (RFC 3042).
2003-03-12 20:27:28 +00:00
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_do_rfc3390) = 1;
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, rfc3390, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_do_rfc3390), 0,
Add support for RFC 3390, which allows for a variable-sized initial congestion window.
2003-03-13 01:43:45 +00:00
"Enable RFC 3390 (Increasing TCP's Initial Congestion Window)");
Add sysctl tunable net.inet.tcp.initcwnd_segments to specify initial congestion window in number of segments on fly. It is set to 10 segments by default. Remove net.inet.tcp.experimental.initcwnd10 which is now redundant. Also remove the parent node net.inet.tcp.experimental as it's not needed anymore and also because it was not well thought out. Differential Revision: https://reviews.freebsd.org/D3858 In collaboration with: lstewart Reviewed by: gnn (prev version), rwatson, allanjude, wblock (man page) MFC after: 2 weeks Relnotes: yes Sponsored by: Limelight Networks
2015-10-27 09:43:05 +00:00
VNET_DEFINE(int, tcp_initcwnd_segments) = 10;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, initcwnd_segments,
CTLFLAG_VNET | CTLFLAG_RW, &VNET_NAME(tcp_initcwnd_segments), 0,
"Slow-start flight size (initial congestion window) in number of segments");
Increase the initial CWND to 10 segments as defined in IETF TCPM draft-ietf-tcpm-initcwnd-05. It explains why the increased initial window improves the overall performance of many web services without risking congestion collapse. As long as it remains a draft it is placed under a sysctl marking it as experimental: net.inet.tcp.experimental.initcwnd10 = 1 When it becomes an official RFC soon the sysctl will be changed to the RFC number and moved to net.inet.tcp. This implementation differs from the RFC draft in that it is a bit more conservative in the case of packet loss on SYN or SYN|ACK because we haven't reduced the default RTO to 1 second yet. Also the restart window isn't yet increased as allowed. Both will be adjusted with upcoming changes. Is is enabled by default. In Linux it is enabled since kernel 3.0. MFC after: 2 weeks
2012-10-28 19:47:46 +00:00
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_do_rfc3465) = 1;
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, rfc3465, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_do_rfc3465), 0,
Add TCP Appropriate Byte Counting (RFC 3465) support to kernel. The new behaviour is on by default, and can be disabled by setting the net.inet.tcp.rfc3465 sysctl to 0 to obtain previous behaviour. The patch changes struct tcpcb in sys/netinet/tcp_var.h which breaks the ABI. Bump __FreeBSD_version to 800061 accordingly. User space tools that rely on the size of struct tcpcb (e.g. sockstat) need to be recompiled. Reviewed by: rpaulo, gnn Approved by: gnn, kmacy (mentors) Sponsored by: FreeBSD Foundation
2009-01-15 06:44:22 +00:00
"Enable RFC 3465 (Appropriate Byte Counting)");
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_abc_l_var) = 2;
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, abc_l_var, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_abc_l_var), 2,
Add TCP Appropriate Byte Counting (RFC 3465) support to kernel. The new behaviour is on by default, and can be disabled by setting the net.inet.tcp.rfc3465 sysctl to 0 to obtain previous behaviour. The patch changes struct tcpcb in sys/netinet/tcp_var.h which breaks the ABI. Bump __FreeBSD_version to 800061 accordingly. User space tools that rely on the size of struct tcpcb (e.g. sockstat) need to be recompiled. Reviewed by: rpaulo, gnn Approved by: gnn, kmacy (mentors) Sponsored by: FreeBSD Foundation
2009-01-15 06:44:22 +00:00
"Cap the max cwnd increment during slow-start to this number of segments");
Mark all SYSCTL_NODEs static that have no corresponding SYSCTL_DECLs. The SYSCTL_NODE macro defines a list that stores all child-elements of that node. If there's no SYSCTL_DECL macro anywhere else, there's no reason why it shouldn't be static.
2011-11-07 15:43:11 +00:00
static SYSCTL_NODE(_net_inet_tcp, OID_AUTO, ecn, CTLFLAG_RW, 0, "TCP ECN");
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
Change net.inet.tcp.ecn.enable sysctl mib from a binary off/on control to a three way setting. 0 - Totally disable ECN. (no change) 1 - Enable ECN if incoming connections request it. Outgoing connections will request ECN. (no change from present != 0 setting) 2 - Enable ECN if incoming connections request it. Outgoing conections will not request ECN. Change the default value of net.inet.tcp.ecn.enable from 0 to 2. Linux version 2.4.20 and newer, Solaris, and Mac OS X 10.5 and newer have similar capabilities. The actual values above match Linux, and the default matches the current Linux default. Reviewed by: eadler MFC after: 1 month MFH: yes Sponsored by: https://reviews.freebsd.org/D6386
2016-05-19 22:20:35 +00:00
VNET_DEFINE(int, tcp_do_ecn) = 2;
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp_ecn, OID_AUTO, enable, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_do_ecn), 0,
"TCP ECN support");
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_ecn_maxretries) = 1;
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp_ecn, OID_AUTO, maxretries, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_ecn_maxretries), 0,
"Max retries before giving up on ECN");
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
VNET_DEFINE(int, tcp_insecure_syn) = 0;
#define V_tcp_insecure_syn VNET(tcp_insecure_syn)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, insecure_syn, CTLFLAG_VNET | CTLFLAG_RW,
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
&VNET_NAME(tcp_insecure_syn), 0,
"Follow RFC793 instead of RFC5961 criteria for accepting SYN packets");
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_insecure_rst) = 0;
#define V_tcp_insecure_rst VNET(tcp_insecure_rst)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, insecure_rst, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_insecure_rst), 0,
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
"Follow RFC793 instead of RFC5961 criteria for accepting RST packets");
Add a sysctl (net.inet.tcp.insecure_rst) which allows one to specify that the RFC 793 specification for accepting RST packets should be following. When followed, this makes one vulnerable to the attacks described in "slipping in the window", but it may be necessary in some odd circumstances.
2005-01-03 07:08:37 +00:00
Add syntactic sugar missed in r226437 and then not added either when moving things around in r226448 but desperately needed to always make things compile successfully. MFC after: 1 week
2011-10-17 00:05:31 +00:00
VNET_DEFINE(int, tcp_recvspace) = 1024*64;
Move the tcp_sendspace and tcp_recvspace sysctl's from the middle of tcp_usrreq.c to the top of tcp_output.c and tcp_input.c respectively next to the socket buffer autosizing controls. MFC after: 1 week
2011-10-16 20:18:39 +00:00
#define V_tcp_recvspace VNET(tcp_recvspace)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, TCPCTL_RECVSPACE, recvspace, CTLFLAG_VNET | CTLFLAG_RW,
Move the tcp_sendspace and tcp_recvspace sysctl's from the middle of tcp_usrreq.c to the top of tcp_output.c and tcp_input.c respectively next to the socket buffer autosizing controls. MFC after: 1 week
2011-10-16 20:18:39 +00:00
&VNET_NAME(tcp_recvspace), 0, "Initial receive socket buffer size");
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_do_autorcvbuf) = 1;
#define V_tcp_do_autorcvbuf VNET(tcp_do_autorcvbuf)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, recvbuf_auto, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_do_autorcvbuf), 0,
Step 1.5 of importing the network stack virtualization infrastructure from the vimage project, as per plan established at devsummit 08/08: http://wiki.freebsd.org/Image/Notes200808DevSummit Introduce INIT_VNET_*() initializer macros, VNET_FOREACH() iterator macros, and CURVNET_SET() context setting macros, all currently resolving to NOPs. Prepare for virtualization of selected SYSCTL objects by introducing a family of SYSCTL_V_*() macros, currently resolving to their global counterparts, i.e. SYSCTL_V_INT() == SYSCTL_INT(). Move selected #defines from sys/sys/vimage.h to newly introduced header files specific to virtualized subsystems (sys/net/vnet.h, sys/netinet/vinet.h etc.). All the changes are verified to have zero functional impact at this point in time by doing MD5 comparision between pre- and post-change object files(*). (*) netipsec/keysock.c did not validate depending on compile time options. Implemented by: julian, bz, brooks, zec Reviewed by: julian, bz, brooks, kris, rwatson, ... Approved by: julian (mentor) Obtained from: //depot/projects/vimage-commit2/... X-MFC after: never Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-10-02 15:37:58 +00:00
"Enable automatic receive buffer sizing");
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(int, tcp_autorcvbuf_inc) = 16*1024;
#define V_tcp_autorcvbuf_inc VNET(tcp_autorcvbuf_inc)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, recvbuf_inc, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_autorcvbuf_inc), 0,
Match up SYSCTL declaration style.
2007-03-19 19:00:51 +00:00
"Incrementor step size of automatic receive buffer");
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
Increase the defaults for the maximum socket buffer limit, and the maximum TCP send and receive buffer limits from 256kB to 2MB. For sb_max_adj we need to add the cast as already used in the sysctl handler to not overflow the type doing the maths. Note that this is just the defaults. They will allow more memory to be consumed per socket/connection if needed but not change the default "idle" memory consumption. All values are still tunable by sysctls. Suggested by: gnn Discussed on: arch (Mar and Aug 2011) MFC after: 3 weeks Approved by: re (kib)
2011-08-25 09:20:13 +00:00
VNET_DEFINE(int, tcp_autorcvbuf_max) = 2*1024*1024;
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
#define V_tcp_autorcvbuf_max VNET(tcp_autorcvbuf_max)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_inet_tcp, OID_AUTO, recvbuf_max, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(tcp_autorcvbuf_max), 0,
Step 1.5 of importing the network stack virtualization infrastructure from the vimage project, as per plan established at devsummit 08/08: http://wiki.freebsd.org/Image/Notes200808DevSummit Introduce INIT_VNET_*() initializer macros, VNET_FOREACH() iterator macros, and CURVNET_SET() context setting macros, all currently resolving to NOPs. Prepare for virtualization of selected SYSCTL objects by introducing a family of SYSCTL_V_*() macros, currently resolving to their global counterparts, i.e. SYSCTL_V_INT() == SYSCTL_INT(). Move selected #defines from sys/sys/vimage.h to newly introduced header files specific to virtualized subsystems (sys/net/vnet.h, sys/netinet/vinet.h etc.). All the changes are verified to have zero functional impact at this point in time by doing MD5 comparision between pre- and post-change object files(*). (*) netipsec/keysock.c did not validate depending on compile time options. Implemented by: julian, bz, brooks, zec Reviewed by: julian, bz, brooks, kris, rwatson, ... Approved by: julian (mentor) Obtained from: //depot/projects/vimage-commit2/... X-MFC after: never Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-10-02 15:37:58 +00:00
"Max size of automatic receive buffer");
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
VNET_DEFINE(struct inpcbhead, tcb);
Change the initialization methodology for global variables scheduled for virtualization. Instead of initializing the affected global variables at instatiation, assign initial values to them in initializer functions. As a rule, initialization at instatiation for such variables should never be introduced again from now on. Furthermore, enclose all instantiations of such global variables in #ifdef VIMAGE_GLOBALS blocks. Essentialy, this change should have zero functional impact. In the next phase of merging network stack virtualization infrastructure from p4/vimage branch, the new initialization methology will allow us to switch between using global variables and their counterparts residing in virtualization containers with minimum code churn, and in the long run allow us to intialize multiple instances of such container structures. Discussed at: devsummit Strassburg Reviewed by: bz, julian Approved by: julian (mentor) Obtained from: //depot/projects/vimage-commit2/... X-MFC after: never Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-11-19 09:39:34 +00:00
#define tcb6 tcb /* for KAME src sync over BSD*'s */
MFP4: @176978-176982, 176984, 176990-176994, 177441 "Whitspace" churn after the VIMAGE/VNET whirls. Remove the need for some "init" functions within the network stack, like pim6_init(), icmp_init() or significantly shorten others like ip6_init() and nd6_init(), using static initialization again where possible and formerly missed. Move (most) variables back to the place they used to be before the container structs and VIMAGE_GLOABLS (before r185088) and try to reduce the diff to stable/7 and earlier as good as possible, to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9. This also removes some header file pollution for putatively static global variables. Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are no longer needed. Reviewed by: jhb Discussed with: rwatson Sponsored by: The FreeBSD Foundation Sponsored by: CK Software GmbH MFC after: 6 days
2010-04-29 11:52:42 +00:00
VNET_DEFINE(struct inpcbinfo, tcbinfo);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Merge from projects/counters: TCP/IP stats. Convert 'struct ipstat' and 'struct tcpstat' to counter(9). This speeds up IP forwarding at extreme packet rates, and makes accounting more precise. Sponsored by: Nginx, Inc.
2013-04-08 19:57:21 +00:00
/*
Redo r294869. The array of counters for TCP states doesn't belong to struct tcpstat, because the structure can be zeroed out by netstat(1) -z, and of course running connection counts shouldn't be touched. Place running connection counts into separate array, and provide separate read-only sysctl oid for it.
2016-03-15 00:15:10 +00:00
* TCP statistics are stored in an array of counter(9)s, which size matches
* size of struct tcpstat. TCP running connection count is a regular array.
Merge from projects/counters: TCP/IP stats. Convert 'struct ipstat' and 'struct tcpstat' to counter(9). This speeds up IP forwarding at extreme packet rates, and makes accounting more precise. Sponsored by: Nginx, Inc.
2013-04-08 19:57:21 +00:00
*/
Use new macros to implement ipstat and tcpstat using PCPU counters. Change interface of kread_counters() similar ot kread() in the netstat(1).
2013-07-09 09:43:03 +00:00
VNET_PCPUSTAT_DEFINE(struct tcpstat, tcpstat);
SYSCTL_VNET_PCPUSTAT(_net_inet_tcp, TCPCTL_STATS, stats, struct tcpstat,
tcpstat, "TCP statistics (struct tcpstat, netinet/tcp_var.h)");
Redo r294869. The array of counters for TCP states doesn't belong to struct tcpstat, because the structure can be zeroed out by netstat(1) -z, and of course running connection counts shouldn't be touched. Place running connection counts into separate array, and provide separate read-only sysctl oid for it.
2016-03-15 00:15:10 +00:00
VNET_DEFINE(counter_u64_t, tcps_states[TCP_NSTATES]);
SYSCTL_COUNTER_U64_ARRAY(_net_inet_tcp, TCPCTL_STATES, states, CTLFLAG_RD |
Fix kernel build after adding new sysctl asserts in r296933.
2016-03-16 10:42:24 +00:00
CTLFLAG_VNET, &VNET_NAME(tcps_states)[0], TCP_NSTATES,
Redo r294869. The array of counters for TCP states doesn't belong to struct tcpstat, because the structure can be zeroed out by netstat(1) -z, and of course running connection counts shouldn't be touched. Place running connection counts into separate array, and provide separate read-only sysctl oid for it.
2016-03-15 00:15:10 +00:00
"TCP connection counts by TCP state");
static void
tcp_vnet_init(const void *unused)
{
Tiny refactor of r294869/r296881: use defines to mask the VNET() macro. Suggested by: bz
2016-05-17 23:14:17 +00:00
COUNTER_ARRAY_ALLOC(V_tcps_states, TCP_NSTATES, M_WAITOK);
Redo r294869. The array of counters for TCP states doesn't belong to struct tcpstat, because the structure can be zeroed out by netstat(1) -z, and of course running connection counts shouldn't be touched. Place running connection counts into separate array, and provide separate read-only sysctl oid for it.
2016-03-15 00:15:10 +00:00
VNET_PCPUSTAT_ALLOC(tcpstat, M_WAITOK);
}
VNET_SYSINIT(tcp_vnet_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
tcp_vnet_init, NULL);
Merge from projects/counters: TCP/IP stats. Convert 'struct ipstat' and 'struct tcpstat' to counter(9). This speeds up IP forwarding at extreme packet rates, and makes accounting more precise. Sponsored by: Nginx, Inc.
2013-04-08 19:57:21 +00:00
#ifdef VIMAGE
Redo r294869. The array of counters for TCP states doesn't belong to struct tcpstat, because the structure can be zeroed out by netstat(1) -z, and of course running connection counts shouldn't be touched. Place running connection counts into separate array, and provide separate read-only sysctl oid for it.
2016-03-15 00:15:10 +00:00
static void
tcp_vnet_uninit(const void *unused)
{
Tiny refactor of r294869/r296881: use defines to mask the VNET() macro. Suggested by: bz
2016-05-17 23:14:17 +00:00
COUNTER_ARRAY_FREE(V_tcps_states, TCP_NSTATES);
Redo r294869. The array of counters for TCP states doesn't belong to struct tcpstat, because the structure can be zeroed out by netstat(1) -z, and of course running connection counts shouldn't be touched. Place running connection counts into separate array, and provide separate read-only sysctl oid for it.
2016-03-15 00:15:10 +00:00
VNET_PCPUSTAT_FREE(tcpstat);
}
VNET_SYSUNINIT(tcp_vnet_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
tcp_vnet_uninit, NULL);
Merge from projects/counters: TCP/IP stats. Convert 'struct ipstat' and 'struct tcpstat' to counter(9). This speeds up IP forwarding at extreme packet rates, and makes accounting more precise. Sponsored by: Nginx, Inc.
2013-04-08 19:57:21 +00:00
#endif /* VIMAGE */
Redo r294869. The array of counters for TCP states doesn't belong to struct tcpstat, because the structure can be zeroed out by netstat(1) -z, and of course running connection counts shouldn't be touched. Place running connection counts into separate array, and provide separate read-only sysctl oid for it.
2016-03-15 00:15:10 +00:00
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
/*
* Kernel module interface for updating tcpstat. The argument is an index
Merge from projects/counters: TCP/IP stats. Convert 'struct ipstat' and 'struct tcpstat' to counter(9). This speeds up IP forwarding at extreme packet rates, and makes accounting more precise. Sponsored by: Nginx, Inc.
2013-04-08 19:57:21 +00:00
* into tcpstat treated as an array.
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
*/
void
kmod_tcpstat_inc(int statnum)
{
Use new macros to implement ipstat and tcpstat using PCPU counters. Change interface of kread_counters() similar ot kread() in the netstat(1).
2013-07-09 09:43:03 +00:00
counter_u64_add(VNET(tcpstat)[statnum], 1);
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
}
- Add some helper hook points to the TCP stack. The hooks allow Khelp modules to access inbound/outbound events and associated data for established TCP connections. The hooks only run if at least one hook function is registered for the hook point, ensuring the impact on the stack is effectively nil when no TCP Khelp modules are loaded. struct tcp_hhook_data is passed as contextual data to any registered Khelp module hook functions. - Add an OSD (Object Specific Data) pointer to struct tcpcb to allow Khelp modules to associate per-connection data with the TCP control block. - Bump __FreeBSD_version and add a note to UPDATING regarding to ABI changes introduced by this commit and r216753. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: FreeBSD Foundation Reviewed by: bz, others along the way MFC after: 3 months
2010-12-28 12:13:30 +00:00
/*
* Wrapper for the TCP established input helper hook.
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
- Add some helper hook points to the TCP stack. The hooks allow Khelp modules to access inbound/outbound events and associated data for established TCP connections. The hooks only run if at least one hook function is registered for the hook point, ensuring the impact on the stack is effectively nil when no TCP Khelp modules are loaded. struct tcp_hhook_data is passed as contextual data to any registered Khelp module hook functions. - Add an OSD (Object Specific Data) pointer to struct tcpcb to allow Khelp modules to associate per-connection data with the TCP control block. - Bump __FreeBSD_version and add a note to UPDATING regarding to ABI changes introduced by this commit and r216753. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: FreeBSD Foundation Reviewed by: bz, others along the way MFC after: 3 months
2010-12-28 12:13:30 +00:00
hhook_run_tcp_est_in(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to)
{
struct tcp_hhook_data hhook_data;
if (V_tcp_hhh[HHOOK_TCP_EST_IN]->hhh_nhooks > 0) {
hhook_data.tp = tp;
hhook_data.th = th;
hhook_data.to = to;
hhook_run_hooks(V_tcp_hhh[HHOOK_TCP_EST_IN], &hhook_data,
tp->osd);
}
}
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
/*
* CC wrapper hook functions
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
cc_ack_received(struct tcpcb *tp, struct tcphdr *th, uint16_t nsegs,
uint16_t type)
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
{
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
tp->ccv->nsegs = nsegs;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
tp->ccv->bytes_this_ack = BYTES_THIS_ACK(tp, th);
Simplify and fix a bug in cc_ack_received()'s "are we congestion window limited" logic (refer to [1] for associated discussion). snd_cwnd and snd_wnd are unsigned long and on 64 bit hosts, min() will truncate them to 32 bits and could therefore potentially corrupt the result (although under normal operation, neither variable should legitmately exceed 32 bits). [1] http://lists.freebsd.org/pipermail/freebsd-net/2013-January/034297.html Submitted by: jhb MFC after: 1 week
2013-01-22 09:44:21 +00:00
if (tp->snd_cwnd <= tp->snd_wnd)
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
tp->ccv->flags |= CCF_CWND_LIMITED;
else
tp->ccv->flags &= ~CCF_CWND_LIMITED;
if (type == CC_ACK) {
if (tp->snd_cwnd > tp->snd_ssthresh) {
tp->t_bytes_acked += min(tp->ccv->bytes_this_ack,
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
nsegs * V_tcp_abc_l_var * tcp_maxseg(tp));
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
if (tp->t_bytes_acked >= tp->snd_cwnd) {
tp->t_bytes_acked -= tp->snd_cwnd;
tp->ccv->flags |= CCF_ABC_SENTAWND;
}
} else {
tp->ccv->flags &= ~CCF_ABC_SENTAWND;
tp->t_bytes_acked = 0;
}
}
if (CC_ALGO(tp)->ack_received != NULL) {
/* XXXLAS: Find a way to live without this */
tp->ccv->curack = th->th_ack;
CC_ALGO(tp)->ack_received(tp->ccv, type);
}
}
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
cc_conn_init(struct tcpcb *tp)
{
struct hc_metrics_lite metrics;
struct inpcb *inp = tp->t_inpcb;
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
u_int maxseg;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
int rtt;
INP_WLOCK_ASSERT(tp->t_inpcb);
tcp_hc_get(&inp->inp_inc, &metrics);
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
maxseg = tcp_maxseg(tp);
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
if (tp->t_srtt == 0 && (rtt = metrics.rmx_rtt)) {
tp->t_srtt = rtt;
tp->t_rttbest = tp->t_srtt + TCP_RTT_SCALE;
TCPSTAT_INC(tcps_usedrtt);
if (metrics.rmx_rttvar) {
tp->t_rttvar = metrics.rmx_rttvar;
TCPSTAT_INC(tcps_usedrttvar);
} else {
/* default variation is +- 1 rtt */
tp->t_rttvar =
tp->t_srtt * TCP_RTTVAR_SCALE / TCP_RTT_SCALE;
}
TCPT_RANGESET(tp->t_rxtcur,
((tp->t_srtt >> 2) + tp->t_rttvar) >> 1,
tp->t_rttmin, TCPTV_REXMTMAX);
}
if (metrics.rmx_ssthresh) {
/*
* There's some sort of gateway or interface
* buffer limit on the path. Use this to set
sys/net*: minor spelling fixes. No functional change.
2016-05-03 18:05:43 +00:00
* the slow start threshold, but set the
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
* threshold to no less than 2*mss.
*/
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_ssthresh = max(2 * maxseg, metrics.rmx_ssthresh);
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
TCPSTAT_INC(tcps_usedssthresh);
}
/*
Remove the ss_fltsz and ss_fltsz_local sysctl's which have long been superseded by the RFC3390 initial CWND sizing. Also remove the remnants of TCP_METRICS_CWND which used the TCP hostcache to set the initial CWND in a non-RFC compliant way. MFC after: 1 week
2011-10-16 20:06:44 +00:00
* Set the initial slow-start flight size.
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
*
When SYN or SYN/ACK had to be retransmitted RFC5681 requires us to reduce the initial CWND to one segment. This reduction got lost some time ago due to a change in initialization ordering. Additionally in tcp_timer_rexmt() avoid entering fast recovery when we're still in TCPS_SYN_SENT state. MFC after: 2 weeks
2012-10-28 17:25:08 +00:00
* RFC5681 Section 3.1 specifies the default conservative values.
* RFC3390 specifies slightly more aggressive values.
Draft-ietf-tcpm-initcwnd-05 became RFC6928. MFC after: 1 week
2013-12-26 04:24:08 +00:00
* RFC6928 increases it to ten segments.
Add sysctl tunable net.inet.tcp.initcwnd_segments to specify initial congestion window in number of segments on fly. It is set to 10 segments by default. Remove net.inet.tcp.experimental.initcwnd10 which is now redundant. Also remove the parent node net.inet.tcp.experimental as it's not needed anymore and also because it was not well thought out. Differential Revision: https://reviews.freebsd.org/D3858 In collaboration with: lstewart Reviewed by: gnn (prev version), rwatson, allanjude, wblock (man page) MFC after: 2 weeks Relnotes: yes Sponsored by: Limelight Networks
2015-10-27 09:43:05 +00:00
* Support for user specified value for initial flight size.
When SYN or SYN/ACK had to be retransmitted RFC5681 requires us to reduce the initial CWND to one segment. This reduction got lost some time ago due to a change in initialization ordering. Additionally in tcp_timer_rexmt() avoid entering fast recovery when we're still in TCPS_SYN_SENT state. MFC after: 2 weeks
2012-10-28 17:25:08 +00:00
*
* If a SYN or SYN/ACK was lost and retransmitted, we have to
* reduce the initial CWND to one segment as congestion is likely
* requiring us to be cautious.
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
*/
When SYN or SYN/ACK had to be retransmitted RFC5681 requires us to reduce the initial CWND to one segment. This reduction got lost some time ago due to a change in initialization ordering. Additionally in tcp_timer_rexmt() avoid entering fast recovery when we're still in TCPS_SYN_SENT state. MFC after: 2 weeks
2012-10-28 17:25:08 +00:00
if (tp->snd_cwnd == 1)
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd = maxseg; /* SYN(-ACK) lost */
Add sysctl tunable net.inet.tcp.initcwnd_segments to specify initial congestion window in number of segments on fly. It is set to 10 segments by default. Remove net.inet.tcp.experimental.initcwnd10 which is now redundant. Also remove the parent node net.inet.tcp.experimental as it's not needed anymore and also because it was not well thought out. Differential Revision: https://reviews.freebsd.org/D3858 In collaboration with: lstewart Reviewed by: gnn (prev version), rwatson, allanjude, wblock (man page) MFC after: 2 weeks Relnotes: yes Sponsored by: Limelight Networks
2015-10-27 09:43:05 +00:00
else if (V_tcp_initcwnd_segments)
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd = min(V_tcp_initcwnd_segments * maxseg,
max(2 * maxseg, V_tcp_initcwnd_segments * 1460));
When SYN or SYN/ACK had to be retransmitted RFC5681 requires us to reduce the initial CWND to one segment. This reduction got lost some time ago due to a change in initialization ordering. Additionally in tcp_timer_rexmt() avoid entering fast recovery when we're still in TCPS_SYN_SENT state. MFC after: 2 weeks
2012-10-28 17:25:08 +00:00
else if (V_tcp_do_rfc3390)
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd = min(4 * maxseg, max(2 * maxseg, 4380));
Adjust the initial default CWND upon connection establishment to the new and increased values specified by RFC5681 Section 3.1. The even larger initial CWND per RFC3390, if enabled, is not affected. MFC after: 2 weeks
2012-10-28 17:16:09 +00:00
else {
/* Per RFC5681 Section 3.1 */
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
if (maxseg > 2190)
tp->snd_cwnd = 2 * maxseg;
else if (maxseg > 1095)
tp->snd_cwnd = 3 * maxseg;
Adjust the initial default CWND upon connection establishment to the new and increased values specified by RFC5681 Section 3.1. The even larger initial CWND per RFC3390, if enabled, is not affected. MFC after: 2 weeks
2012-10-28 17:16:09 +00:00
else
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd = 4 * maxseg;
Adjust the initial default CWND upon connection establishment to the new and increased values specified by RFC5681 Section 3.1. The even larger initial CWND per RFC3390, if enabled, is not affected. MFC after: 2 weeks
2012-10-28 17:16:09 +00:00
}
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
if (CC_ALGO(tp)->conn_init != NULL)
CC_ALGO(tp)->conn_init(tp->ccv);
}
void inline
cc_cong_signal(struct tcpcb *tp, struct tcphdr *th, uint32_t type)
{
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
u_int maxseg;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
switch(type) {
case CC_NDUPACK:
if (!IN_FASTRECOVERY(tp->t_flags)) {
tp->snd_recover = tp->snd_max;
if (tp->t_flags & TF_ECN_PERMIT)
tp->t_flags |= TF_ECN_SND_CWR;
}
break;
case CC_ECN:
if (!IN_CONGRECOVERY(tp->t_flags)) {
TCPSTAT_INC(tcps_ecn_rcwnd);
tp->snd_recover = tp->snd_max;
if (tp->t_flags & TF_ECN_PERMIT)
tp->t_flags |= TF_ECN_SND_CWR;
}
break;
case CC_RTO:
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
maxseg = tcp_maxseg(tp);
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
tp->t_dupacks = 0;
tp->t_bytes_acked = 0;
EXIT_RECOVERY(tp->t_flags);
Set ssthresh appropriately on RTO. This change was accidentally not ported from the pre modular CC stack. Sponsored by: FreeBSD Foundation Submitted by: David Hayes <dahayes at swin edu au> MFC after: 9 weeks X-MFC with: r215166
2010-12-02 01:01:37 +00:00
tp->snd_ssthresh = max(2, min(tp->snd_wnd, tp->snd_cwnd) / 2 /
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
maxseg) * maxseg;
tp->snd_cwnd = maxseg;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
break;
case CC_RTO_ERR:
TCPSTAT_INC(tcps_sndrexmitbad);
/* RTO was unnecessary, so reset everything. */
tp->snd_cwnd = tp->snd_cwnd_prev;
tp->snd_ssthresh = tp->snd_ssthresh_prev;
tp->snd_recover = tp->snd_recover_prev;
if (tp->t_flags & TF_WASFRECOVERY)
ENTER_FASTRECOVERY(tp->t_flags);
if (tp->t_flags & TF_WASCRECOVERY)
ENTER_CONGRECOVERY(tp->t_flags);
tp->snd_nxt = tp->snd_max;
TCP reuses t_rxtshift to determine the backoff timer used for both the persist state and the retransmit timer. However, the code that implements "bad retransmit recovery" only checks t_rxtshift to see if an ACK has been received in during the first retransmit timeout window. As a result, if ticks has wrapped over to a negative value and a socket is in the persist state, it can incorrectly treat an ACK from the remote peer as a "bad retransmit recovery" and restore saved values such as snd_ssthresh and snd_cwnd. However, if the socket has never had a retransmit timeout, then these saved values will be zero, so snd_ssthresh and snd_cwnd will be set to 0. If the socket is in fast recovery (this can be caused by excessive duplicate ACKs such as those fixed by 220794), then each ACK that arrives triggers either NewReno or SACK partial ACK handling which clamps snd_cwnd to be no larger than snd_ssthresh. In effect, the socket's send window is permamently stuck at 0 even though the remote peer is advertising a much larger window and pending data is only sent via TCP window probes (so one byte every few seconds). Fix this by adding a new TCP pcb flag (TF_PREVVALID) that indicates that the various snd_*_prev fields in the pcb are valid and only perform "bad retransmit recovery" if this flag is set in the pcb. The flag is set on the first retransmit timeout that occurs and is cleared on subsequent retransmit timeouts or when entering the persist state. Reviewed by: bz MFC after: 2 weeks
2011-04-29 15:40:12 +00:00
tp->t_flags &= ~TF_PREVVALID;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
tp->t_badrxtwin = 0;
break;
}
if (CC_ALGO(tp)->cong_signal != NULL) {
if (th != NULL)
tp->ccv->curack = th->th_ack;
CC_ALGO(tp)->cong_signal(tp->ccv, type);
}
}
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void inline
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
cc_post_recovery(struct tcpcb *tp, struct tcphdr *th)
{
INP_WLOCK_ASSERT(tp->t_inpcb);
/* XXXLAS: KASSERT that we're in recovery? */
if (CC_ALGO(tp)->post_recovery != NULL) {
tp->ccv->curack = th->th_ack;
CC_ALGO(tp)->post_recovery(tp->ccv);
}
/* XXXLAS: EXIT_RECOVERY ? */
tp->t_bytes_acked = 0;
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
}
New style sysctl & staticize alot of stuff.
1995-11-14 20:34:56 +00:00
Add the possibility to verify MD5 hash of incoming TCP packets. As long as this is a costy function, even when compiled in (along with the option TCP_SIGNATURE), it can be disabled via the net.inet.tcp.signature_verify_input sysctl. Sponsored by: Sandvine Incorporated Reviewed by: emaste, bz MFC after: 2 weeks
2011-04-25 17:13:40 +00:00
#ifdef TCP_SIGNATURE
static inline int
tcp_signature_verify_input(struct mbuf *m, int off0, int tlen, int optlen,
struct tcpopt *to, struct tcphdr *th, u_int tcpbflag)
{
int ret;
tcp_fields_to_net(th);
ret = tcp_signature_verify(m, off0, tlen, optlen, to, th, tcpbflag);
tcp_fields_to_host(th);
return (ret);
}
#endif
Do not delay a new ack if there already is a delayed ack pending on the connection, but send it immediately. Prior to this change, it was possible to delay a delayed-ack for multiple times, resulting in degraded TCP behavior in certain corner cases.
2001-02-25 15:17:24 +00:00
/*
Fix a bug with transmitter restart after receiving a 0 window. The receiver was not sending an immediate ack with delayed acks turned on when the input buffer is drained, preventing the transmitter from restarting immediately. Propogate the TCP_NODELAY option to accept()ed sockets. (Helps tbench and is a good idea anyway). Some cleanup. Identify additonal issues in comments. MFC after: 1 day
2001-12-02 08:49:29 +00:00
* Indicate whether this ack should be delayed. We can delay the ack if
Improve readability of comments for DELAY_ACK() macro.
2014-04-03 01:46:03 +00:00
* following conditions are met:
* - There is no delayed ack timer in progress.
* - Our last ack wasn't a 0-sized window. We never want to delay
* the ack that opens up a 0-sized window.
* - LRO wasn't used for this segment. We make sure by checking that the
* segment size is not larger than the MSS.
Do not delay a new ack if there already is a delayed ack pending on the connection, but send it immediately. Prior to this change, it was possible to delay a delayed-ack for multiple times, resulting in degraded TCP behavior in certain corner cases.
2001-02-25 15:17:24 +00:00
*/
The TCP delayed ACK logic isn't aware of LRO passing up large aggregated segments thinking it received only one segment. This causes it to enable the delay the ACK for 100ms to wait for another segment which may never come because all the data was received already. Doing delayed ACK for LRO segments is bogus for two reasons: a) it pushes us further away from acking every other packet; b) it introduces additional delay in responding to the sender. The latter is especially bad because it is in the nature of LRO to aggregated all segments of a burst with no more coming until an ACK is sent back. Change the delayed ACK logic to detect LRO segments by being larger than the MSS for this connection and issuing an immediate ACK for them to keep the ACK clock ticking without interruption. Reported by: julian, cperciva Tested by: cperciva Reviewed by: lstewart MFC after: 3 days
2013-10-22 18:24:34 +00:00
#define DELAY_ACK(tp, tlen) \
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
((!tcp_timer_active(tp, TT_DELACK) && \
Check to see if the TF_DELACK flag is set before returning from tcp_input(). This unbreaks delack handling, while still preserving correct T/TCP behavior Tested by: maxim Sponsored by: DARPA, NAI Labs
2003-02-22 21:54:57 +00:00
(tp->t_flags & TF_RXWIN0SENT) == 0) && \
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
(tlen <= tp->t_maxseg) && \
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
(V_tcp_delack_enabled || (tp->t_flags & TF_NEEDSYN)))
Do not delay a new ack if there already is a delayed ack pending on the connection, but send it immediately. Prior to this change, it was possible to delay a delayed-ack for multiple times, resulting in degraded TCP behavior in certain corner cases.
2001-02-25 15:17:24 +00:00
DCTCP (Data Center TCP) implementation. DCTCP congestion control algorithm aims to maximise throughput and minimise latency in data center networks by utilising the proportion of Explicit Congestion Notification (ECN) marked packets received from capable hardware as a congestion signal. Highlights: Implemented as a mod_cc(4) module. ECN (Explicit congestion notification) processing is done differently from RFC3168. Takes one-sided DCTCP into consideration where only one of the sides is using DCTCP and other is using standard ECN. IETF draft: http://tools.ietf.org/html/draft-bensley-tcpm-dctcp-00 Thesis report by Midori Kato: https://eggert.org/students/kato-thesis.pdf Submitted by: Midori Kato <katoon@sfc.wide.ad.jp> and Lars Eggert <lars@netapp.com> with help and modifications from hiren Differential Revision: https://reviews.freebsd.org/D604 Reviewed by: gnn
2015-01-12 08:33:04 +00:00
static void inline
cc_ecnpkt_handler(struct tcpcb *tp, struct tcphdr *th, uint8_t iptos)
{
INP_WLOCK_ASSERT(tp->t_inpcb);
if (CC_ALGO(tp)->ecnpkt_handler != NULL) {
switch (iptos & IPTOS_ECN_MASK) {
case IPTOS_ECN_CE:
tp->ccv->flags |= CCF_IPHDR_CE;
break;
case IPTOS_ECN_ECT0:
tp->ccv->flags &= ~CCF_IPHDR_CE;
break;
case IPTOS_ECN_ECT1:
tp->ccv->flags &= ~CCF_IPHDR_CE;
break;
}
if (th->th_flags & TH_CWR)
tp->ccv->flags |= CCF_TCPHDR_CWR;
else
tp->ccv->flags &= ~CCF_TCPHDR_CWR;
if (tp->t_flags & TF_DELACK)
tp->ccv->flags |= CCF_DELACK;
else
tp->ccv->flags &= ~CCF_DELACK;
CC_ALGO(tp)->ecnpkt_handler(tp->ccv);
if (tp->ccv->flags & CCF_ACKNOW)
tcp_timer_activate(tp, TT_DELACK, tcp_delacktime);
}
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* TCP input handling is split into multiple parts:
* tcp6_input is a thin wrapper around tcp_input for the extended
* ip6_protox[] call format in ip6_input
* tcp_input handles primary segment validation, inpcb lookup and
* SYN processing on listen sockets
* tcp_do_segment processes the ACK and text of the segment for
* establishing, established and closing connections
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#ifdef INET6
int
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
tcp6_input(struct mbuf **mp, int *offp, int proto)
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
{
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
struct mbuf *m = *mp;
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
struct in6_ifaddr *ia6;
Remove ip6_getdstifaddr() and all functions to work with auxiliary data. It isn't safe to keep unreferenced ifaddrs. Use in6ifa_ifwithaddr() to determine ifaddr corresponding to destination address. Since currently we keep addresses with embedded scope zone, in6ifa_ifwithaddr is called with zero zoneid and marked with XXX. Also remove route and lle lookups from ip6_input. Use in6ifa_ifwithaddr() instead. Sponsored by: Yandex LLC
2014-11-08 19:38:34 +00:00
struct ip6_hdr *ip6;
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
IP6_EXTHDR_CHECK(m, *offp, sizeof(struct tcphdr), IPPROTO_DONE);
/*
* draft-itojun-ipv6-tcp-to-anycast
* better place to put this in?
*/
Remove ip6_getdstifaddr() and all functions to work with auxiliary data. It isn't safe to keep unreferenced ifaddrs. Use in6ifa_ifwithaddr() to determine ifaddr corresponding to destination address. Since currently we keep addresses with embedded scope zone, in6ifa_ifwithaddr is called with zero zoneid and marked with XXX. Also remove route and lle lookups from ip6_input. Use in6ifa_ifwithaddr() instead. Sponsored by: Yandex LLC
2014-11-08 19:38:34 +00:00
ip6 = mtod(m, struct ip6_hdr *);
ia6 = in6ifa_ifwithaddr(&ip6->ip6_dst, 0 /* XXX */);
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
if (ia6 && (ia6->ia6_flags & IN6_IFF_ANYCAST)) {
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
struct ip6_hdr *ip6;
Modify most routines returning 'struct ifaddr *' to return references rather than pointers, requiring callers to properly dispose of those references. The following routines now return references: ifaddr_byindex ifa_ifwithaddr ifa_ifwithbroadaddr ifa_ifwithdstaddr ifa_ifwithnet ifaof_ifpforaddr ifa_ifwithroute ifa_ifwithroute_fib rt_getifa rt_getifa_fib IFP_TO_IA ip_rtaddr in6_ifawithifp in6ifa_ifpforlinklocal in6ifa_ifpwithaddr in6_ifadd carp_iamatch6 ip6_getdstifaddr Remove unused macro which didn't have required referencing: IFP_TO_IA6 This closes many small races in which changes to interface or address lists while an ifaddr was in use could lead to use of freed memory (etc). In a few cases, add missing if_addr_list locking required to safely acquire references. Because of a lack of deep copying support, we accept a race in which an in6_ifaddr pointed to by mbuf tags and extracted with ip6_getdstifaddr() doesn't hold a reference while in transmit. Once we have mbuf tag deep copy support, this can be fixed. Reviewed by: bz Obtained from: Apple, Inc. (portions) MFC after: 6 weeks (portions)
2009-06-23 20:19:09 +00:00
ifa_free(&ia6->ia_ifa);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
ip6 = mtod(m, struct ip6_hdr *);
icmp6_error(m, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR,
(caddr_t)&ip6->ip6_dst - (caddr_t)ip6);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
}
Plug more refcount leaks and possible NULL deref for interface address list. Submitted by: scottl@ MFC after: 3 days
2012-06-04 18:43:51 +00:00
if (ia6)
ifa_free(&ia6->ia_ifa);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (tcp_input(mp, offp, proto));
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif /* INET6 */
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
int
tcp_input(struct mbuf **mp, int *offp, int proto)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
struct mbuf *m = *mp;
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
struct tcphdr *th = NULL;
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
struct ip *ip = NULL;
struct inpcb *inp = NULL;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
struct tcpcb *tp = NULL;
struct socket *so = NULL;
Finish 4.4-Lite-2 merge: randomize TCP initial sequence numbers to make ISS-guessing spoofing attacks harder.
1995-10-03 16:54:17 +00:00
u_char *optp = NULL;
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
int off0;
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int optlen = 0;
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#ifdef INET
int len;
#endif
int tlen = 0, off;
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
int drop_hdrlen;
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
int thflags;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
int rstreason = 0; /* For badport_bandlim accounting purposes */
Add the possibility to verify MD5 hash of incoming TCP packets. As long as this is a costy function, even when compiled in (along with the option TCP_SIGNATURE), it can be disabled via the net.inet.tcp.signature_verify_input sysctl. Sponsored by: Sandvine Incorporated Reviewed by: emaste, bz MFC after: 2 weeks
2011-04-25 17:13:40 +00:00
#ifdef TCP_SIGNATURE
uint8_t sig_checked = 0;
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
uint8_t iptos = 0;
Remove the IPFIREWALL_FORWARD kernel option and make possible to turn on the related functionality in the runtime via the sysctl variable net.pfil.forward. It is turned off by default. Sponsored by: Yandex LLC Discussed with: net@ MFC after: 2 weeks
2012-10-25 09:39:14 +00:00
struct m_tag *fwd_tag = NULL;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
#ifdef INET6
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
struct ip6_hdr *ip6 = NULL;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
int isipv6;
#else
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
const void *ip6 = NULL;
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif /* INET6 */
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
struct tcpopt to; /* options in this segment */
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
char *s = NULL; /* address and port logging */
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
int ti_locked;
Made TCPDEBUG truely optional. Based on changes I made in FreeBSD 1.1.5. Fixed somebody's idea of a joke - about the first half of the lines in in_proto.c were spaced over by one space.
1994-09-15 10:36:56 +00:00
#ifdef TCPDEBUG
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
/*
* The size of tcp_saveipgen must be the size of the max ip header,
* now IPv6.
*/
o Use a define for a buffer size. Prodded by: db o Add missed vars for TCPDEBUG in tcp_do_segment(). Prodded by: tinderbox
2007-03-24 22:15:02 +00:00
u_char tcp_saveipgen[IP6_HDR_LEN];
Move two global variables to automatic variables within the only function where they are used (they are used with TCPDEBUG only).
2002-06-23 21:22:56 +00:00
struct tcphdr tcp_savetcp;
Made TCPDEBUG truely optional. Based on changes I made in FreeBSD 1.1.5. Fixed somebody's idea of a joke - about the first half of the lines in in_proto.c were spaced over by one space.
1994-09-15 10:36:56 +00:00
short ostate = 0;
#endif
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#ifdef INET6
isipv6 = (mtod(m, struct ip *)->ip_v == 6) ? 1 : 0;
#endif
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
off0 = *offp;
m = *mp;
*mp = NULL;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
to.to_flags = 0;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvtotal);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
fix -O0 compilation without INET6. Pointed out by: ru
2004-03-01 19:10:31 +00:00
#ifdef INET6
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
if (isipv6) {
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
/* IP6_EXTHDR_CHECK() is already done at tcp6_input(). */
MFp4 bz_ipv6_fast: Add code to handle pre-checked TCP checksums as indicated by mbuf flags to save the entire computation for validation if not needed. In the IPv6 TCP output path only compute the pseudo-header checksum, set the checksum offset in the mbuf field along the appropriate flag as done in IPv4. In tcp_respond() just initialize the IPv6 payload length to 0 as ip6_output() will properly set it. Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems Reviewed by: gnn (as part of the whole) MFC After: 3 days
2012-05-25 02:23:26 +00:00
if (m->m_len < (sizeof(*ip6) + sizeof(*th))) {
m = m_pullup(m, sizeof(*ip6) + sizeof(*th));
if (m == NULL) {
TCPSTAT_INC(tcps_rcvshort);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
MFp4 bz_ipv6_fast: Add code to handle pre-checked TCP checksums as indicated by mbuf flags to save the entire computation for validation if not needed. In the IPv6 TCP output path only compute the pseudo-header checksum, set the checksum offset in the mbuf field along the appropriate flag as done in IPv4. In tcp_respond() just initialize the IPv6 payload length to 0 as ip6_output() will properly set it. Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems Reviewed by: gnn (as part of the whole) MFC After: 3 days
2012-05-25 02:23:26 +00:00
}
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
ip6 = mtod(m, struct ip6_hdr *);
MFp4 bz_ipv6_fast: Add code to handle pre-checked TCP checksums as indicated by mbuf flags to save the entire computation for validation if not needed. In the IPv6 TCP output path only compute the pseudo-header checksum, set the checksum offset in the mbuf field along the appropriate flag as done in IPv4. In tcp_respond() just initialize the IPv6 payload length to 0 as ip6_output() will properly set it. Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems Reviewed by: gnn (as part of the whole) MFC After: 3 days
2012-05-25 02:23:26 +00:00
th = (struct tcphdr *)((caddr_t)ip6 + off0);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tlen = sizeof(*ip6) + ntohs(ip6->ip6_plen) - off0;
It turns out that too many drivers are not only parsing the L2/3/4 headers for TSO but also for generic checksum offloading. Ideally we would only have one common function shared amongst all drivers, and perhaps when updating them for IPv6 we should introduce that. Eventually we should provide the meta information along with mbufs to avoid (re-)parsing entirely. To not break IPv6 (checksums and offload) and to be able to MFC the changes without risking to hurt 3rd party drivers, duplicate the v4 framework, as other OSes have done as well. Introduce interface capability flags for TX/RX checksum offload with IPv6, to allow independent toggling (where possible). Add CSUM_*_IPV6 flags for UDP/TCP over IPv6, and reserve further for SCTP, and IPv6 fragmentation. Define CSUM_DELAY_DATA_IPV6 as we do for legacy IP and add an alias for CSUM_DATA_VALID_IPV6. This pretty much brings IPv6 handling in line with IPv4. TSO is still handled in a different way and not via if_hwassist. Update ifconfig to allow (un)setting of the new capability flags. Update loopback to announce the new capabilities and if_hwassist flags. Individual driver updates will have to follow, as will SCTP. Reported by: gallatin, dim, .. Reviewed by: gallatin (glanced at?) MFC after: 3 days X-MFC with: r235961,235959,235958
2012-05-28 09:30:13 +00:00
if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID_IPV6) {
MFp4 bz_ipv6_fast: Add code to handle pre-checked TCP checksums as indicated by mbuf flags to save the entire computation for validation if not needed. In the IPv6 TCP output path only compute the pseudo-header checksum, set the checksum offset in the mbuf field along the appropriate flag as done in IPv4. In tcp_respond() just initialize the IPv6 payload length to 0 as ip6_output() will properly set it. Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems Reviewed by: gnn (as part of the whole) MFC After: 3 days
2012-05-25 02:23:26 +00:00
if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
th->th_sum = m->m_pkthdr.csum_data;
else
th->th_sum = in6_cksum_pseudo(ip6, tlen,
IPPROTO_TCP, m->m_pkthdr.csum_data);
th->th_sum ^= 0xffff;
} else
th->th_sum = in6_cksum(m, IPPROTO_TCP, off0, tlen);
if (th->th_sum) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvbadsum);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
goto drop;
}
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
/*
* Be proactive about unspecified IPv6 address in source.
* As we use all-zero to indicate unbounded/unconnected pcb,
* unspecified IPv6 address can be used to confuse us.
*
* Note that packets with unspecified IPv6 destination is
* already dropped in ip6_input.
*/
if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) {
/* XXX stat */
goto drop;
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
}
fix -O0 compilation without INET6. Pointed out by: ru
2004-03-01 19:10:31 +00:00
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#if defined(INET) && defined(INET6)
else
#endif
#ifdef INET
{
Add support for offloading IP/TCP/UDP checksums to NIC hardware which supports them.
2000-03-27 19:14:27 +00:00
/*
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
* Get IP and TCP header together in first mbuf.
* Note: IP leaves IP header in first mbuf.
Add support for offloading IP/TCP/UDP checksums to NIC hardware which supports them.
2000-03-27 19:14:27 +00:00
*/
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
if (off0 > sizeof (struct ip)) {
In ip_stripoptions(): - Remove unused argument and incorrect comment. - Fixup ip_len after stripping.
2012-10-12 09:24:24 +00:00
ip_stripoptions(m);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
off0 = sizeof(struct ip);
}
if (m->m_len < sizeof (struct tcpiphdr)) {
Tidy up some code to conform better to surroundings and style(9), 0 = NULL and space/tab.
2007-03-23 19:11:22 +00:00
if ((m = m_pullup(m, sizeof (struct tcpiphdr)))
== NULL) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvshort);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
}
}
ip = mtod(m, struct ip *);
th = (struct tcphdr *)((caddr_t)ip + off0);
Do not reduce ip_len by size of IP header in the ip_input() before passing a packet to protocol input routines. For several protocols this mean that now protocol needs to do subtraction itself, and for another half this means that we do not need to add header length back to the packet. Make ip_stripoptions() to adjust ip_len, since now we enter this function with a packet header whose ip_len does represent length of entire packet, not payload only.
2012-10-23 08:33:13 +00:00
tlen = ntohs(ip->ip_len) - off0;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) {
if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
th->th_sum = m->m_pkthdr.csum_data;
else
th->th_sum = in_pseudo(ip->ip_src.s_addr,
Switch the entire IPv4 stack to keep the IP packet header in network byte order. Any host byte order processing is done in local variables and host byte order values are never[1] written to a packet. After this change a packet processed by the stack isn't modified at all[2] except for TTL. After this change a network stack hacker doesn't need to scratch his head trying to figure out what is the byte order at the given place in the stack. [1] One exception still remains. The raw sockets convert host byte order before pass a packet to an application. Probably this would remain for ages for compatibility. [2] The ip_input() still subtructs header len from ip->ip_len, but this is planned to be fixed soon. Reviewed by: luigi, Maxim Dounin <mdounin mdounin.ru> Tested by: ray, Olivier Cochard-Labbe <olivier cochard.me>
2012-10-22 21:09:03 +00:00
ip->ip_dst.s_addr,
htonl(m->m_pkthdr.csum_data + tlen +
IPPROTO_TCP));
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
th->th_sum ^= 0xffff;
} else {
Switch the entire IPv4 stack to keep the IP packet header in network byte order. Any host byte order processing is done in local variables and host byte order values are never[1] written to a packet. After this change a packet processed by the stack isn't modified at all[2] except for TTL. After this change a network stack hacker doesn't need to scratch his head trying to figure out what is the byte order at the given place in the stack. [1] One exception still remains. The raw sockets convert host byte order before pass a packet to an application. Probably this would remain for ages for compatibility. [2] The ip_input() still subtructs header len from ip->ip_len, but this is planned to be fixed soon. Reviewed by: luigi, Maxim Dounin <mdounin mdounin.ru> Tested by: ray, Olivier Cochard-Labbe <olivier cochard.me>
2012-10-22 21:09:03 +00:00
struct ipovly *ipov = (struct ipovly *)ip;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
/*
* Checksum extended TCP header and data.
*/
Switch the entire IPv4 stack to keep the IP packet header in network byte order. Any host byte order processing is done in local variables and host byte order values are never[1] written to a packet. After this change a packet processed by the stack isn't modified at all[2] except for TTL. After this change a network stack hacker doesn't need to scratch his head trying to figure out what is the byte order at the given place in the stack. [1] One exception still remains. The raw sockets convert host byte order before pass a packet to an application. Probably this would remain for ages for compatibility. [2] The ip_input() still subtructs header len from ip->ip_len, but this is planned to be fixed soon. Reviewed by: luigi, Maxim Dounin <mdounin mdounin.ru> Tested by: ray, Olivier Cochard-Labbe <olivier cochard.me>
2012-10-22 21:09:03 +00:00
len = off0 + tlen;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
bzero(ipov->ih_x1, sizeof(ipov->ih_x1));
Switch the entire IPv4 stack to keep the IP packet header in network byte order. Any host byte order processing is done in local variables and host byte order values are never[1] written to a packet. After this change a packet processed by the stack isn't modified at all[2] except for TTL. After this change a network stack hacker doesn't need to scratch his head trying to figure out what is the byte order at the given place in the stack. [1] One exception still remains. The raw sockets convert host byte order before pass a packet to an application. Probably this would remain for ages for compatibility. [2] The ip_input() still subtructs header len from ip->ip_len, but this is planned to be fixed soon. Reviewed by: luigi, Maxim Dounin <mdounin mdounin.ru> Tested by: ray, Olivier Cochard-Labbe <olivier cochard.me>
2012-10-22 21:09:03 +00:00
ipov->ih_len = htons(tlen);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
th->th_sum = in_cksum(m, len);
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
/* Reset length for SDT probes. */
ip->ip_len = htons(tlen + off0);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
}
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
if (th->th_sum) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvbadsum);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
goto drop;
}
/* Re-initialization for later version check */
ip->ip_v = IPVERSION;
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif /* INET */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
#ifdef INET6
if (isipv6)
iptos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
#if defined(INET) && defined(INET6)
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
else
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#ifdef INET
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
iptos = ip->ip_tos;
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Check that TCP offset makes sense,
* pull out TCP options and adjust length. XXX
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
off = th->th_off << 2;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (off < sizeof (struct tcphdr) || off > tlen) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvbadoff);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto drop;
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tlen -= off; /* tlen is used instead of ti->ti_len */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (off > sizeof (struct tcphdr)) {
fix -O0 compilation without INET6. Pointed out by: ru
2004-03-01 19:10:31 +00:00
#ifdef INET6
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
if (isipv6) {
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
IP6_EXTHDR_CHECK(m, off0, off, IPPROTO_DONE);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
ip6 = mtod(m, struct ip6_hdr *);
th = (struct tcphdr *)((caddr_t)ip6 + off0);
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
}
fix -O0 compilation without INET6. Pointed out by: ru
2004-03-01 19:10:31 +00:00
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#if defined(INET) && defined(INET6)
else
#endif
#ifdef INET
{
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
if (m->m_len < sizeof(struct ip) + off) {
if ((m = m_pullup(m, sizeof (struct ip) + off))
Tidy up some code to conform better to surroundings and style(9), 0 = NULL and space/tab.
2007-03-23 19:11:22 +00:00
== NULL) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvshort);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
}
ip = mtod(m, struct ip *);
th = (struct tcphdr *)((caddr_t)ip + off0);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
optlen = off - sizeof (struct tcphdr);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
optp = (u_char *)(th + 1);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
thflags = th->th_flags;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Convert TCP protocol specific fields to host format.
*/
Add the possibility to verify MD5 hash of incoming TCP packets. As long as this is a costy function, even when compiled in (along with the option TCP_SIGNATURE), it can be disabled via the net.inet.tcp.signature_verify_input sysctl. Sponsored by: Sandvine Incorporated Reviewed by: emaste, bz MFC after: 2 weeks
2011-04-25 17:13:40 +00:00
tcp_fields_to_host(th);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Fix a bug in tcp_input reported by Rick Jones <raj@hpisrdq.cup.hp.com>. If a goto findpcb occurred during the processing of a segment, the TCP and IP headers were dropped twice from the mbuf which resulted in data acked by TCP but not delivered to the user. Reviewed by: davidg
1995-04-05 10:32:14 +00:00
/*
Update and clarify comments in first section of tcp_input().
2007-03-21 18:52:58 +00:00
* Delay dropping TCP, IP headers, IPv6 ext headers, and TCP options.
Fix a bug in tcp_input reported by Rick Jones <raj@hpisrdq.cup.hp.com>. If a goto findpcb occurred during the processing of a segment, the TCP and IP headers were dropped twice from the mbuf which resulted in data acked by TCP but not delivered to the user. Reviewed by: davidg
1995-04-05 10:32:14 +00:00
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
drop_hdrlen = off0 + off;
Fix a bug in tcp_input reported by Rick Jones <raj@hpisrdq.cup.hp.com>. If a goto findpcb occurred during the processing of a segment, the TCP and IP headers were dropped twice from the mbuf which resulted in data acked by TCP but not delivered to the user. Reviewed by: davidg
1995-04-05 10:32:14 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
* Locate pcb for segment; if we're likely to add or remove a
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
* connection then first acquire pcbinfo lock. There are three cases
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
* where we might discover later we need a write lock despite the
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
* flags: ACKs moving a connection out of the syncache, ACKs for a
* connection in TIMEWAIT and SYNs not targeting a listening socket.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
if ((thflags & (TH_FIN | TH_RST)) != 0) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK(&V_tcbinfo);
ti_locked = TI_RLOCKED;
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
} else
ti_locked = TI_UNLOCKED;
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
/*
* Grab info from PACKET_TAG_IPFORWARD tag prepended to the chain.
*/
Fix !INET6 build after r244365.
2012-12-18 08:14:16 +00:00
if (
#ifdef INET6
(isipv6 && (m->m_flags & M_IP6_NEXTHOP))
#ifdef INET
|| (!isipv6 && (m->m_flags & M_IP_NEXTHOP))
#endif
#endif
#if defined(INET) && !defined(INET6)
(m->m_flags & M_IP_NEXTHOP)
#endif
)
Remove the IPFIREWALL_FORWARD kernel option and make possible to turn on the related functionality in the runtime via the sysctl variable net.pfil.forward. It is turned off by default. Sponsored by: Yandex LLC Discussed with: net@ MFC after: 2 weeks
2012-10-25 09:39:14 +00:00
fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL);
Convert ipfw to use PFIL_HOOKS. This is change is transparent to userland and preserves the ipfw ABI. The ipfw core packet inspection and filtering functions have not been changed, only how ipfw is invoked is different. However there are many changes how ipfw is and its add-on's are handled: In general ipfw is now called through the PFIL_HOOKS and most associated magic, that was in ip_input() or ip_output() previously, is now done in ipfw_check_[in|out]() in the ipfw PFIL handler. IPDIVERT is entirely handled within the ipfw PFIL handlers. A packet to be diverted is checked if it is fragmented, if yes, ip_reass() gets in for reassembly. If not, or all fragments arrived and the packet is complete, divert_packet is called directly. For 'tee' no reassembly attempt is made and a copy of the packet is sent to the divert socket unmodified. The original packet continues its way through ip_input/output(). ipfw 'forward' is done via m_tag's. The ipfw PFIL handlers tag the packet with the new destination sockaddr_in. A check if the new destination is a local IP address is made and the m_flags are set appropriately. ip_input() and ip_output() have some more work to do here. For ip_input() the m_flags are checked and a packet for us is directly sent to the 'ours' section for further processing. Destination changes on the input path are only tagged and the 'srcrt' flag to ip_forward() is set to disable destination checks and ICMP replies at this stage. The tag is going to be handled on output. ip_output() again checks for m_flags and the 'ours' tag. If found, the packet will be dropped back to the IP netisr where it is going to be picked up by ip_input() again and the directly sent to the 'ours' section. When only the destination changes, the route's 'dst' is overwritten with the new destination from the forward m_tag. Then it jumps back at the route lookup again and skips the firewall check because it has been marked with M_SKIP_FIREWALL. ipfw 'forward' has to be compiled into the kernel with 'option IPFIREWALL_FORWARD' to enable it. DUMMYNET is entirely handled within the ipfw PFIL handlers. A packet for a dummynet pipe or queue is directly sent to dummynet_io(). Dummynet will then inject it back into ip_input/ip_output() after it has served its time. Dummynet packets are tagged and will continue from the next rule when they hit the ipfw PFIL handlers again after re-injection. BRIDGING and IPFW_ETHER are not changed yet and use ipfw_chk() directly as they did before. Later this will be changed to dedicated ETHER PFIL_HOOKS. More detailed changes to the code: conf/files Add netinet/ip_fw_pfil.c. conf/options Add IPFIREWALL_FORWARD option. modules/ipfw/Makefile Add ip_fw_pfil.c. net/bridge.c Disable PFIL_HOOKS if ipfw for bridging is active. Bridging ipfw is still directly invoked to handle layer2 headers and packets would get a double ipfw when run through PFIL_HOOKS as well. netinet/ip_divert.c Removed divert_clone() function. It is no longer used. netinet/ip_dummynet.[ch] Neither the route 'ro' nor the destination 'dst' need to be stored while in dummynet transit. Structure members and associated macros are removed. netinet/ip_fastfwd.c Removed all direct ipfw handling code and replace it with the new 'ipfw forward' handling code. netinet/ip_fw.h Removed 'ro' and 'dst' from struct ip_fw_args. netinet/ip_fw2.c (Re)moved some global variables and the module handling. netinet/ip_fw_pfil.c New file containing the ipfw PFIL handlers and module initialization. netinet/ip_input.c Removed all direct ipfw handling code and replace it with the new 'ipfw forward' handling code. ip_forward() does not longer require the 'next_hop' struct sockaddr_in argument. Disable early checks if 'srcrt' is set. netinet/ip_output.c Removed all direct ipfw handling code and replace it with the new 'ipfw forward' handling code. netinet/ip_var.h Add ip_reass() as general function. (Used from ipfw PFIL handlers for IPDIVERT.) netinet/raw_ip.c Directly check if ipfw and dummynet control pointers are active. netinet/tcp_input.c Rework the 'ipfw forward' to local code to work with the new way of forward tags. netinet/tcp_sack.c Remove include 'opt_ipfw.h' which is not needed here. sys/mbuf.h Remove m_claim_next() macro which was exclusively for ipfw 'forward' and is no longer needed. Approved by: re (scottl)
2004-08-17 22:05:54 +00:00
Keep fwd_tag around for subsequent pcb lookups For TIMEWAIT handling tcp_input may have to jump back for an additional pass through pcblookup. Prior to this change the fwd_tag had been discarded after the first lookup, so a new connection attempt delivered locally via 'ipfw fwd' would fail to find a match. As of r248886 the tag will be detached and freed when passed to the socket buffer.
2013-03-29 20:51:44 +00:00
findpcb:
#ifdef INVARIANTS
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED) {
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Keep fwd_tag around for subsequent pcb lookups For TIMEWAIT handling tcp_input may have to jump back for an additional pass through pcblookup. Prior to this change the fwd_tag had been discarded after the first lookup, so a new connection attempt delivered locally via 'ipfw fwd' would fail to find a match. As of r248886 the tag will be detached and freed when passed to the socket buffer.
2013-03-29 20:51:44 +00:00
} else {
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
}
#endif
Add support for IPv6 to ipfw fwd: Distinguish IPv4 and IPv6 addresses and optional port numbers in user space to set the option for the correct protocol family. Add support in the kernel for carrying the new IPv6 destination address and port. Add support to TCP and UDP for IPv6 and fix UDP IPv4 to not change the address in the IP header. Add support for IPv6 forwarding to a non-local destination. Add a regession test uitilizing VIMAGE to check all 20 possible combinations I could think of. Obtained from: David Dolson at Sandvine Incorporated (original version for ipfw fwd IPv6 support) Sponsored by: Sandvine Incorporated PR: bin/117214 MFC after: 4 weeks Approved by: re (kib)
2011-08-20 17:05:11 +00:00
#ifdef INET6
if (isipv6 && fwd_tag != NULL) {
struct sockaddr_in6 *next_hop6;
next_hop6 = (struct sockaddr_in6 *)(fwd_tag + 1);
/*
* Transparently forwarded. Pretend to be the destination.
* Already got one like this?
*/
inp = in6_pcblookup_mbuf(&V_tcbinfo,
&ip6->ip6_src, th->th_sport, &ip6->ip6_dst, th->th_dport,
INPLOOKUP_WLOCKPCB, m->m_pkthdr.rcvif, m);
if (!inp) {
/*
* It's new. Try to find the ambushing socket.
* Because we've rewritten the destination address,
* any hardware-generated hash is ignored.
*/
inp = in6_pcblookup(&V_tcbinfo, &ip6->ip6_src,
th->th_sport, &next_hop6->sin6_addr,
next_hop6->sin6_port ? ntohs(next_hop6->sin6_port) :
th->th_dport, INPLOOKUP_WILDCARD |
INPLOOKUP_WLOCKPCB, m->m_pkthdr.rcvif);
}
Remove the IPFIREWALL_FORWARD kernel option and make possible to turn on the related functionality in the runtime via the sysctl variable net.pfil.forward. It is turned off by default. Sponsored by: Yandex LLC Discussed with: net@ MFC after: 2 weeks
2012-10-25 09:39:14 +00:00
} else if (isipv6) {
Add support for IPv6 to ipfw fwd: Distinguish IPv4 and IPv6 addresses and optional port numbers in user space to set the option for the correct protocol family. Add support in the kernel for carrying the new IPv6 destination address and port. Add support to TCP and UDP for IPv6 and fix UDP IPv4 to not change the address in the IP header. Add support for IPv6 forwarding to a non-local destination. Add a regession test uitilizing VIMAGE to check all 20 possible combinations I could think of. Obtained from: David Dolson at Sandvine Incorporated (original version for ipfw fwd IPv6 support) Sponsored by: Sandvine Incorporated PR: bin/117214 MFC after: 4 weeks Approved by: re (kib)
2011-08-20 17:05:11 +00:00
inp = in6_pcblookup_mbuf(&V_tcbinfo, &ip6->ip6_src,
th->th_sport, &ip6->ip6_dst, th->th_dport,
INPLOOKUP_WILDCARD | INPLOOKUP_WLOCKPCB,
m->m_pkthdr.rcvif, m);
}
#endif /* INET6 */
#if defined(INET6) && defined(INET)
else
#endif
#ifdef INET
if (fwd_tag != NULL) {
Convert ipfw to use PFIL_HOOKS. This is change is transparent to userland and preserves the ipfw ABI. The ipfw core packet inspection and filtering functions have not been changed, only how ipfw is invoked is different. However there are many changes how ipfw is and its add-on's are handled: In general ipfw is now called through the PFIL_HOOKS and most associated magic, that was in ip_input() or ip_output() previously, is now done in ipfw_check_[in|out]() in the ipfw PFIL handler. IPDIVERT is entirely handled within the ipfw PFIL handlers. A packet to be diverted is checked if it is fragmented, if yes, ip_reass() gets in for reassembly. If not, or all fragments arrived and the packet is complete, divert_packet is called directly. For 'tee' no reassembly attempt is made and a copy of the packet is sent to the divert socket unmodified. The original packet continues its way through ip_input/output(). ipfw 'forward' is done via m_tag's. The ipfw PFIL handlers tag the packet with the new destination sockaddr_in. A check if the new destination is a local IP address is made and the m_flags are set appropriately. ip_input() and ip_output() have some more work to do here. For ip_input() the m_flags are checked and a packet for us is directly sent to the 'ours' section for further processing. Destination changes on the input path are only tagged and the 'srcrt' flag to ip_forward() is set to disable destination checks and ICMP replies at this stage. The tag is going to be handled on output. ip_output() again checks for m_flags and the 'ours' tag. If found, the packet will be dropped back to the IP netisr where it is going to be picked up by ip_input() again and the directly sent to the 'ours' section. When only the destination changes, the route's 'dst' is overwritten with the new destination from the forward m_tag. Then it jumps back at the route lookup again and skips the firewall check because it has been marked with M_SKIP_FIREWALL. ipfw 'forward' has to be compiled into the kernel with 'option IPFIREWALL_FORWARD' to enable it. DUMMYNET is entirely handled within the ipfw PFIL handlers. A packet for a dummynet pipe or queue is directly sent to dummynet_io(). Dummynet will then inject it back into ip_input/ip_output() after it has served its time. Dummynet packets are tagged and will continue from the next rule when they hit the ipfw PFIL handlers again after re-injection. BRIDGING and IPFW_ETHER are not changed yet and use ipfw_chk() directly as they did before. Later this will be changed to dedicated ETHER PFIL_HOOKS. More detailed changes to the code: conf/files Add netinet/ip_fw_pfil.c. conf/options Add IPFIREWALL_FORWARD option. modules/ipfw/Makefile Add ip_fw_pfil.c. net/bridge.c Disable PFIL_HOOKS if ipfw for bridging is active. Bridging ipfw is still directly invoked to handle layer2 headers and packets would get a double ipfw when run through PFIL_HOOKS as well. netinet/ip_divert.c Removed divert_clone() function. It is no longer used. netinet/ip_dummynet.[ch] Neither the route 'ro' nor the destination 'dst' need to be stored while in dummynet transit. Structure members and associated macros are removed. netinet/ip_fastfwd.c Removed all direct ipfw handling code and replace it with the new 'ipfw forward' handling code. netinet/ip_fw.h Removed 'ro' and 'dst' from struct ip_fw_args. netinet/ip_fw2.c (Re)moved some global variables and the module handling. netinet/ip_fw_pfil.c New file containing the ipfw PFIL handlers and module initialization. netinet/ip_input.c Removed all direct ipfw handling code and replace it with the new 'ipfw forward' handling code. ip_forward() does not longer require the 'next_hop' struct sockaddr_in argument. Disable early checks if 'srcrt' is set. netinet/ip_output.c Removed all direct ipfw handling code and replace it with the new 'ipfw forward' handling code. netinet/ip_var.h Add ip_reass() as general function. (Used from ipfw PFIL handlers for IPDIVERT.) netinet/raw_ip.c Directly check if ipfw and dummynet control pointers are active. netinet/tcp_input.c Rework the 'ipfw forward' to local code to work with the new way of forward tags. netinet/tcp_sack.c Remove include 'opt_ipfw.h' which is not needed here. sys/mbuf.h Remove m_claim_next() macro which was exclusively for ipfw 'forward' and is no longer needed. Approved by: re (scottl)
2004-08-17 22:05:54 +00:00
struct sockaddr_in *next_hop;
next_hop = (struct sockaddr_in *)(fwd_tag+1);
Support for IPFW based transparent forwarding. Any packet that can be matched by a ipfw rule can be redirected transparently to another port or machine. Redirection to another port mostly makes sense with tcp, where a session can be set up between a proxy and an unsuspecting client. Redirection to another machine requires that the other machine also be expecting to receive the forwarded packets, as their headers will not have been modified. /sbin/ipfw must be recompiled!!! Reviewed by: Peter Wemm <peter@freebsd.org> Submitted by: Chrisy Luke <chrisy@flix.net>
1998-07-06 03:20:19 +00:00
/*
Remove (almost all) global variables that were used to hold packet forwarding state ("annotations") during ip processing. The code is considerably cleaner now. The variables removed by this change are: ip_divert_cookie used by divert sockets ip_fw_fwd_addr used for transparent ip redirection last_pkt used by dynamic pipes in dummynet Removal of the first two has been done by carrying the annotations into volatile structs prepended to the mbuf chains, and adding appropriate code to add/remove annotations in the routines which make use of them, i.e. ip_input(), ip_output(), tcp_input(), bdg_forward(), ether_demux(), ether_output_frame(), div_output(). On passing, remove a bug in divert handling of fragmented packet. Now it is the fragment at offset 0 which sets the divert status of the whole packet, whereas formerly it was the last incoming fragment to decide. Removal of last_pkt required a change in the interface of ip_fw_chk() and dummynet_io(). On passing, use the same mechanism for dummynet annotations and for divert/forward annotations. option IPFIREWALL_FORWARD is effectively useless, the code to implement it is very small and is now in by default to avoid the obfuscation of conditionally compiled code. NOTES: * there is at least one global variable left, sro_fwd, in ip_output(). I am not sure if/how this can be removed. * I have deliberately avoided gratuitous style changes in this commit to avoid cluttering the diffs. Minor stule cleanup will likely be necessary * this commit only focused on the IP layer. I am sure there is a number of global variables used in the TCP and maybe UDP stack. * despite the number of files touched, there are absolutely no API's or data structures changed by this commit (except the interfaces of ip_fw_chk() and dummynet_io(), which are internal anyways), so an MFC is quite safe and unintrusive (and desirable, given the improved readability of the code). MFC after: 10 days
2002-06-22 11:51:02 +00:00
* Transparently forwarded. Pretend to be the destination.
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
* already got one like this?
Support for IPFW based transparent forwarding. Any packet that can be matched by a ipfw rule can be redirected transparently to another port or machine. Redirection to another port mostly makes sense with tcp, where a session can be set up between a proxy and an unsuspecting client. Redirection to another machine requires that the other machine also be expecting to receive the forwarded packets, as their headers will not have been modified. /sbin/ipfw must be recompiled!!! Reviewed by: Peter Wemm <peter@freebsd.org> Submitted by: Chrisy Luke <chrisy@flix.net>
1998-07-06 03:20:19 +00:00
*/
Add _mbuf() variants of various inpcb-related interfaces, including lookup, hash install, etc. For now, these are arguments are unused, but as we add RSS support, we will want to use hashes extracted from mbufs, rather than manually calculated hashes of header fields, due to the expensive of the software version of Toeplitz (and similar hashes). Add notes that it would be nice to be able to pass mbufs into lookup routines in pf(4), optimising firewall lookup in the same way, but the code structure there doesn't facilitate that currently. (In principle there is no reason this couldn't be MFCed -- the change extends rather than modifies the KBI. However, it won't be useful without other previous possibly less MFCable changes.) Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-06-04 16:33:06 +00:00
inp = in_pcblookup_mbuf(&V_tcbinfo, ip->ip_src, th->th_sport,
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
ip->ip_dst, th->th_dport, INPLOOKUP_WLOCKPCB,
Add _mbuf() variants of various inpcb-related interfaces, including lookup, hash install, etc. For now, these are arguments are unused, but as we add RSS support, we will want to use hashes extracted from mbufs, rather than manually calculated hashes of header fields, due to the expensive of the software version of Toeplitz (and similar hashes). Add notes that it would be nice to be able to pass mbufs into lookup routines in pf(4), optimising firewall lookup in the same way, but the code structure there doesn't facilitate that currently. (In principle there is no reason this couldn't be MFCed -- the change extends rather than modifies the KBI. However, it won't be useful without other previous possibly less MFCable changes.) Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-06-04 16:33:06 +00:00
m->m_pkthdr.rcvif, m);
Support for IPFW based transparent forwarding. Any packet that can be matched by a ipfw rule can be redirected transparently to another port or machine. Redirection to another port mostly makes sense with tcp, where a session can be set up between a proxy and an unsuspecting client. Redirection to another machine requires that the other machine also be expecting to receive the forwarded packets, as their headers will not have been modified. /sbin/ipfw must be recompiled!!! Reviewed by: Peter Wemm <peter@freebsd.org> Submitted by: Chrisy Luke <chrisy@flix.net>
1998-07-06 03:20:19 +00:00
if (!inp) {
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
/*
* It's new. Try to find the ambushing socket.
Add _mbuf() variants of various inpcb-related interfaces, including lookup, hash install, etc. For now, these are arguments are unused, but as we add RSS support, we will want to use hashes extracted from mbufs, rather than manually calculated hashes of header fields, due to the expensive of the software version of Toeplitz (and similar hashes). Add notes that it would be nice to be able to pass mbufs into lookup routines in pf(4), optimising firewall lookup in the same way, but the code structure there doesn't facilitate that currently. (In principle there is no reason this couldn't be MFCed -- the change extends rather than modifies the KBI. However, it won't be useful without other previous possibly less MFCable changes.) Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-06-04 16:33:06 +00:00
* Because we've rewritten the destination address,
* any hardware-generated hash is ignored.
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
*/
inp = in_pcblookup(&V_tcbinfo, ip->ip_src,
th->th_sport, next_hop->sin_addr,
next_hop->sin_port ? ntohs(next_hop->sin_port) :
th->th_dport, INPLOOKUP_WILDCARD |
INPLOOKUP_WLOCKPCB, m->m_pkthdr.rcvif);
Support for IPFW based transparent forwarding. Any packet that can be matched by a ipfw rule can be redirected transparently to another port or machine. Redirection to another port mostly makes sense with tcp, where a session can be set up between a proxy and an unsuspecting client. Redirection to another machine requires that the other machine also be expecting to receive the forwarded packets, as their headers will not have been modified. /sbin/ipfw must be recompiled!!! Reviewed by: Peter Wemm <peter@freebsd.org> Submitted by: Chrisy Luke <chrisy@flix.net>
1998-07-06 03:20:19 +00:00
}
Tidy up IPFIREWALL_FORWARD sections and comments.
2007-03-21 18:56:03 +00:00
} else
Add support for IPv6 to ipfw fwd: Distinguish IPv4 and IPv6 addresses and optional port numbers in user space to set the option for the correct protocol family. Add support in the kernel for carrying the new IPv6 destination address and port. Add support to TCP and UDP for IPv6 and fix UDP IPv4 to not change the address in the IP header. Add support for IPv6 forwarding to a non-local destination. Add a regession test uitilizing VIMAGE to check all 20 possible combinations I could think of. Obtained from: David Dolson at Sandvine Incorporated (original version for ipfw fwd IPv6 support) Sponsored by: Sandvine Incorporated PR: bin/117214 MFC after: 4 weeks Approved by: re (kib)
2011-08-20 17:05:11 +00:00
inp = in_pcblookup_mbuf(&V_tcbinfo, ip->ip_src,
th->th_sport, ip->ip_dst, th->th_dport,
INPLOOKUP_WILDCARD | INPLOOKUP_WLOCKPCB,
m->m_pkthdr.rcvif, m);
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif /* INET */
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Tidy up tcp_log_in_vain and blackhole.
2007-03-21 18:36:49 +00:00
* If the INPCB does not exist then all data in the incoming
* segment is discarded and an appropriate RST is sent back.
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
* XXX MRT Send RST using which routing table?
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Log TCP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.tcp.log_in_vain: 1 Log UDP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.udp.log_in_vain: 1 Suggested by: Warren Toomey <wkt@cs.adfa.oz.au>
1996-04-04 10:46:44 +00:00
if (inp == NULL) {
Tidy up tcp_log_in_vain and blackhole.
2007-03-21 18:36:49 +00:00
/*
* Log communication attempts to ports that are not
* in use.
*/
if ((tcp_log_in_vain == 1 && (thflags & TH_SYN)) ||
tcp_log_in_vain == 2) {
Untangle the net.inet.tcp.log_in_vain and net.inet.tcp.log_debug sysctl's and remove any side effects. Both sysctl's share the same backend infrastructure and due to the way it was implemented enabling net.inet.tcp.log_in_vain would also cause log_debug output to be generated. This was surprising and eventually annoying to the user. The log output backend is kept the same but a little shim is inserted to properly separate log_in_vain and log_debug and to remove any side effects. PR: kern/137317 MFC after: 1 week
2010-08-18 17:39:47 +00:00
if ((s = tcp_log_vain(NULL, th, (void *)ip, ip6)))
Add tcp_log_addrs() function to generate and standardized TCP log line for use thoughout the tcp subsystem. It is IPv4 and IPv6 aware creates a line in the following format: "TCP: [1.2.3.4]:50332 to [1.2.3.4]:80 tcpflags <RST>" A "\n" is not included at the end. The caller is supposed to add further information after the standard tcp log header. The function returns a NUL terminated string which the caller has to free(s, M_TCPLOG) after use. All memory allocation is done with M_NOWAIT and the return value may be NULL in memory shortage situations. Either struct in_conninfo || (struct tcphdr && (struct ip || struct ip6_hdr) have to be supplied. Due to ip[6].h header inclusion limitations and ordering issues the struct ip and struct ip6_hdr parameters have to be casted and passed as void * pointers. tcp_log_addrs(struct in_conninfo *inc, struct tcphdr *th, void *ip4hdr, void *ip6hdr) Usage example: struct ip *ip; char *tcplog; if (tcplog = tcp_log_addrs(NULL, th, (void *)ip, NULL)) { log(LOG_DEBUG, "%s; %s: Connection attempt to closed port\n", tcplog, __func__); free(s, M_TCPLOG); }
2007-05-18 19:58:37 +00:00
log(LOG_INFO, "%s; %s: Connection attempt "
"to closed port\n", s, __func__);
Fix breakage if blackhole=1 and tiflags & TH_SYN, plus style(9) fixes Submitted by: Jonathon Lemon
1999-08-19 05:22:12 +00:00
}
Tidy up tcp_log_in_vain and blackhole.
2007-03-21 18:36:49 +00:00
/*
* When blackholing do not respond with a RST but
* completely ignore the segment and drop it.
*/
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if ((V_blackhole == 1 && (thflags & TH_SYN)) ||
V_blackhole == 2)
When blackholing do a 'dropunlock' in the new world order to prevent the INP_INFO_LOCK from leaking. Reported by: ache Found by: rwatson
2007-03-28 12:58:13 +00:00
goto dropunlock;
Tidy up tcp_log_in_vain and blackhole.
2007-03-21 18:36:49 +00:00
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
rstreason = BANDLIM_RST_CLOSEDPORT;
goto dropwithreset;
Log TCP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.tcp.log_in_vain: 1 Log UDP syn packets for ports we don't listen on. Controlled by: sysctl net.inet.udp.log_in_vain: 1 Suggested by: Warren Toomey <wkt@cs.adfa.oz.au>
1996-04-04 10:46:44 +00:00
}
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
INP_WLOCK_ASSERT(inp);
Start process of removing the use of the deprecated "M_FLOWID" flag from the FreeBSD network code. The flag is still kept around in the "sys/mbuf.h" header file, but does no longer have any users. Instead the "m_pkthdr.rsstype" field in the mbuf structure is now used to decide the meaning of the "m_pkthdr.flowid" field. To modify the "m_pkthdr.rsstype" field please use the existing "M_HASHTYPE_XXX" macros as defined in the "sys/mbuf.h" header file. This patch introduces new behaviour in the transmit direction. Previously network drivers checked if "M_FLOWID" was set in "m_flags" before using the "m_pkthdr.flowid" field. This check has now now been replaced by checking if "M_HASHTYPE_GET(m)" is different from "M_HASHTYPE_NONE". In the future more hashtypes will be added, for example hashtypes for hardware dedicated flows. "M_HASHTYPE_OPAQUE" indicates that the "m_pkthdr.flowid" value is valid and has no particular type. This change removes the need for an "if" statement in TCP transmit code checking for the presence of a valid flowid value. The "if" statement mentioned above is now a direct variable assignment which is then later checked by the respective network drivers like before. Additional notes: - The SCTP code changes will be committed as a separate patch. - Removal of the "M_FLOWID" flag will also be done separately. - The FreeBSD version has been bumped. MFC after: 1 month Sponsored by: Mellanox Technologies
2014-12-01 11:45:24 +00:00
if ((inp->inp_flowtype == M_HASHTYPE_NONE) &&
(M_HASHTYPE_GET(m) != M_HASHTYPE_NONE) &&
((inp->inp_socket == NULL) ||
(inp->inp_socket->so_options & SO_ACCEPTCONN) == 0)) {
Import "flowid" support for serializing flows across transmit queues Reviewed by: rwatson and jeli
2009-04-10 06:16:14 +00:00
inp->inp_flowid = m->m_pkthdr.flowid;
Ensure that the flowid hashtype is assigned to the inp if the flowid is also assigned.
2014-05-18 22:34:06 +00:00
inp->inp_flowtype = M_HASHTYPE_GET(m);
Import "flowid" support for serializing flows across transmit queues Reviewed by: rwatson and jeli
2009-04-10 06:16:14 +00:00
}
Make sure that either inp is NULL or we have obtained a lock on it before jumping to dropunlock to avoid a panic. While here move the calls to ipsec4_in_reject() and ipsec6_in_reject() so they are after we obtain the lock on inp. Original patch to avoid panic: pjd Review of locking adjustments: gnn, sam Approved by: re (rwatson)
2007-09-10 14:49:32 +00:00
#ifdef IPSEC
#ifdef INET6
if (isipv6 && ipsec6_in_reject(m, inp)) {
goto dropunlock;
} else
#endif /* INET6 */
if (ipsec4_in_reject(m, inp) != 0) {
goto dropunlock;
}
#endif /* IPSEC */
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
/*
* Check the minimum TTL for socket.
*/
Check the correct TTL in both the IPv6 and IPv4 cases. Submitted by: glebius Reviewed by: gnn, bz Found with: Coverity Prevent(tm)
2006-01-14 16:39:31 +00:00
if (inp->inp_ip_minttl != 0) {
#ifdef INET6
When checking the inp_ip_minttl restriction for IPv6 packets, don't check the IPv4 header. CID: 1017920 Differential Revision: https://reviews.freebsd.org/D4727 Reviewed by: bz MFC after: 2 weeks Sponsored by: Juniper Networks
2015-12-29 19:20:39 +00:00
if (isipv6) {
if (inp->inp_ip_minttl > ip6->ip6_hlim)
goto dropunlock;
} else
Check the correct TTL in both the IPv6 and IPv4 cases. Submitted by: glebius Reviewed by: gnn, bz Found with: Coverity Prevent(tm)
2006-01-14 16:39:31 +00:00
#endif
if (inp->inp_ip_minttl > ip->ip_ttl)
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
goto dropunlock;
Check the correct TTL in both the IPv6 and IPv4 cases. Submitted by: glebius Reviewed by: gnn, bz Found with: Coverity Prevent(tm)
2006-01-14 16:39:31 +00:00
}
Add socketoption IP_MINTTL. May be used to set the minimum acceptable TTL a packet must have when received on a socket. All packets with a lower TTL are silently dropped. Works on already connected/connecting and listening sockets for RAW/UDP/TCP. This option is only really useful when set to 255 preventing packets from outside the directly connected networks reaching local listeners on sockets. Allows userland implementation of 'The Generalized TTL Security Mechanism (GTSM)' according to RFC3682. Examples of such use include the Cisco IOS BGP implementation command "neighbor ttl-security". MFC after: 2 weeks Sponsored by: TCP/IP Optimization Fundraise 2005
2005-08-22 16:13:08 +00:00
Update and clarify comments in first section of tcp_input().
2007-03-21 18:52:58 +00:00
/*
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
* A previous connection in TIMEWAIT state is supposed to catch stray
* or duplicate segments arriving late. If this segment was a
Update some stale comments regarding tcbinfo locking in the TCP input path: read locks on tcbinfo are no longer used, so won't happen. No functional change. MFC after: 3 days
2012-07-22 17:31:36 +00:00
* legitimate new connection attempt, the old INPCB gets removed and
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
* we can try again to find a listening socket.
*
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
* At this point, due to earlier optimism, we may hold only an inpcb
* lock, and not the inpcbinfo write lock. If so, we need to try to
* acquire it, or if that fails, acquire a reference on the inpcb,
* drop all locks, acquire a global write lock, and then re-acquire
* the inpcb lock. We may at that point discover that another thread
* has tried to free the inpcb, in which case we need to loop back
* and try to find a new inpcb to deliver to.
*
* XXXRW: It may be time to rethink timewait locking.
Update and clarify comments in first section of tcp_input().
2007-03-21 18:52:58 +00:00
*/
In tcp_input(), we acquire a global write lock at first only if a segment is likely to trigger a TCP state change (i.e., FIN/RST/SYN). If we later have to upgrade the lock, we acquire an inpcb reference and drop both global/inpcb locks before reacquiring in-order. In that gap, the connection may transition into TIMEWAIT, so we need to loop back and reevaluate the inpcb after relocking. MFC after: 3 days Reported by: Kamigishi Rei <spambox at haruhiism.net> Reviewed by: bz
2009-10-05 22:24:13 +00:00
relocked:
Correct a number of evolved problems with inp_vflag and inp_flags: certain flags that should have been in inp_flags ended up in inp_vflag, meaning that they were inconsistently locked, and in one case, interpreted. Move the following flags from inp_vflag to gaps in the inp_flags space (and clean up the inp_flags constants to make gaps more obvious to future takers): INP_TIMEWAIT INP_SOCKREF INP_ONESBCAST INP_DROPPED Some aspects of this change have no effect on kernel ABI at all, as these are UDP/TCP/IP-internal uses; however, netstat and sockstat detect INP_TIMEWAIT when listing TCP sockets, so any MFC will need to take this into account. MFC after: 1 week (or after dependencies are MFC'd) Reviewed by: bz
2009-03-15 09:58:31 +00:00
if (inp->inp_flags & INP_TIMEWAIT) {
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
if (ti_locked == TI_UNLOCKED) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (INP_INFO_TRY_RLOCK(&V_tcbinfo) == 0) {
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
in_pcbref(inp);
INP_WUNLOCK(inp);
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK(&V_tcbinfo);
ti_locked = TI_RLOCKED;
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
INP_WLOCK(inp);
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
if (in_pcbrele_wlocked(inp)) {
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
inp = NULL;
goto findpcb;
}
Remove tcp_input lock statistics; these are intended for debugging only and are not intended to ship in 8.0 as they dirty additional cache lines in a performance-critical per-packet path. MFC after: 3 days
2009-10-06 20:35:41 +00:00
} else
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
ti_locked = TI_RLOCKED;
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
}
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
Add a TCP TIMEWAIT state which uses less space than a fullblown TCP control block. Allow the socket and tcpcb structures to be freed earlier than inpcb. Update code to understand an inp w/o a socket. Reviewed by: hsu, silby, jayanth Sponsored by: DARPA, NAI Labs
2003-02-19 22:32:43 +00:00
if (thflags & TH_SYN)
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
tcp_dooptions(&to, optp, optlen, TO_SYN);
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
/*
* NB: tcp_twcheck unlocks the INP and frees the mbuf.
*/
Move TIME_WAIT related functions and timer handling from files other than repo copied tcp_subr.c into tcp_timewait.c#1.284: tcp_input.c#1.350 tcp_timewait() -> tcp_twcheck() tcp_timer.c#1.92 tcp_timer_2msl_reset() -> tcp_tw_2msl_reset() tcp_timer.c#1.92 tcp_timer_2msl_stop() -> tcp_tw_2msl_stop() tcp_timer.c#1.92 tcp_timer_2msl_tw() -> tcp_tw_2msl_scan() This is a mechanical move with appropriate renames and making them static if used only locally. The tcp_tw_2msl_scan() cleanup function is still run from the tcp_slowtimo() in tcp_timer.c.
2007-05-16 17:14:25 +00:00
if (tcp_twcheck(inp, &to, th, m, tlen))
Add a TCP TIMEWAIT state which uses less space than a fullblown TCP control block. Allow the socket and tcpcb structures to be freed earlier than inpcb. Update code to understand an inp w/o a socket. Reviewed by: hsu, silby, jayanth Sponsored by: DARPA, NAI Labs
2003-02-19 22:32:43 +00:00
goto findpcb;
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RUNLOCK(&V_tcbinfo);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
Add a TCP TIMEWAIT state which uses less space than a fullblown TCP control block. Allow the socket and tcpcb structures to be freed earlier than inpcb. Update code to understand an inp w/o a socket. Reviewed by: hsu, silby, jayanth Sponsored by: DARPA, NAI Labs
2003-02-19 22:32:43 +00:00
}
Update and clarify comments in first section of tcp_input().
2007-03-21 18:52:58 +00:00
/*
* The TCPCB may no longer exist if the connection is winding
* down or it is in the CLOSED state. Either way we drop the
* segment and send an appropriate response.
*/
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp = intotcpcb(inp);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
if (tp == NULL || tp->t_state == TCPS_CLOSED) {
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
rstreason = BANDLIM_RST_CLOSEDPORT;
goto dropwithreset;
Change the following: 1. ICMP ECHO and TSTAMP replies are now rate limited. 2. RSTs generated due to packets sent to open and unopen ports are now limited by seperate counters. 3. Each rate limiting queue now has its own description, as follows: Limiting icmp unreach response from 439 to 200 packets per second Limiting closed port RST response from 283 to 200 packets per second Limiting open port RST response from 18724 to 200 packets per second Limiting icmp ping response from 211 to 200 packets per second Limiting icmp tstamp response from 394 to 200 packets per second Submitted by: Mike Silbersack <silby@silby.com>
2000-12-15 21:45:49 +00:00
}
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
- Updated TOE support in the kernel. - Stateful TCP offload drivers for Terminator 3 and 4 (T3 and T4) ASICs. These are available as t3_tom and t4_tom modules that augment cxgb(4) and cxgbe(4) respectively. The cxgb/cxgbe drivers continue to work as usual with or without these extra features. - iWARP driver for Terminator 3 ASIC (kernel verbs). T4 iWARP in the works and will follow soon. Build-tested with make universe. 30s overview ============ What interfaces support TCP offload? Look for TOE4 and/or TOE6 in the capabilities of an interface: # ifconfig -m | grep TOE Enable/disable TCP offload on an interface (just like any other ifnet capability): # ifconfig cxgbe0 toe # ifconfig cxgbe0 -toe Which connections are offloaded? Look for toe4 and/or toe6 in the output of netstat and sockstat: # netstat -np tcp | grep toe # sockstat -46c | grep toe Reviewed by: bz, gnn Sponsored by: Chelsio communications. MFC after: ~3 months (after 9.1, and after ensuring MFC is feasible)
2012-06-19 07:34:13 +00:00
#ifdef TCP_OFFLOAD
if (tp->t_flags & TF_TOE) {
tcp_offload_input(tp, m);
m = NULL; /* consumed by the TOE driver */
goto dropunlock;
}
#endif
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
/*
* We've identified a valid inpcb, but it could be that we need an
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
* inpcbinfo write lock but don't hold it. In this case, attempt to
* acquire using the same strategy as the TIMEWAIT case above. If we
* relock, we have to jump back to 'relocked' as the connection might
* now be in TIMEWAIT.
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
*/
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
#ifdef INVARIANTS
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
if ((thflags & (TH_FIN | TH_RST)) != 0)
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
#endif
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
if (!((tp->t_state == TCPS_ESTABLISHED && (thflags & TH_SYN) == 0) ||
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
(tp->t_state == TCPS_LISTEN && (thflags & TH_SYN) &&
!(tp->t_flags & TF_FASTOPEN)))) {
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
if (ti_locked == TI_UNLOCKED) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (INP_INFO_TRY_RLOCK(&V_tcbinfo) == 0) {
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
in_pcbref(inp);
INP_WUNLOCK(inp);
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK(&V_tcbinfo);
ti_locked = TI_RLOCKED;
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
INP_WLOCK(inp);
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
if (in_pcbrele_wlocked(inp)) {
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
inp = NULL;
goto findpcb;
}
In tcp_input(), we acquire a global write lock at first only if a segment is likely to trigger a TCP state change (i.e., FIN/RST/SYN). If we later have to upgrade the lock, we acquire an inpcb reference and drop both global/inpcb locks before reacquiring in-order. In that gap, the connection may transition into TIMEWAIT, so we need to loop back and reevaluate the inpcb after relocking. MFC after: 3 days Reported by: Kamigishi Rei <spambox at haruhiism.net> Reviewed by: bz
2009-10-05 22:24:13 +00:00
goto relocked;
Remove tcp_input lock statistics; these are intended for debugging only and are not intended to ship in 8.0 as they dirty additional cache lines in a performance-critical per-packet path. MFC after: 3 days
2009-10-06 20:35:41 +00:00
} else
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
ti_locked = TI_RLOCKED;
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
}
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
}
Introduce support for Mandatory Access Control and extensible kernel access control. Instrument the TCP socket code for packet generation and delivery: label outgoing mbufs with the label of the socket, and check socket and mbuf labels before permitting delivery to a socket. Assign labels to newly accepted connections when the syncache/cookie code has done its business. Also set peer labels as convenient. Currently, MAC policies cannot influence the PCB matching algorithm, so cannot implement polyinstantiation. Note that there is at least one case where a PCB is not available due to the TCP packet not being associated with any socket, so we don't label in that case, but need to handle it in a special manner. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-31 19:06:49 +00:00
#ifdef MAC
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(inp);
Merge first in a series of TrustedBSD MAC Framework KPI changes from Mac OS X Leopard--rationalize naming for entry points to the following general forms: mac_<object>_<method/action> mac_<object>_check_<method/action> The previous naming scheme was inconsistent and mostly reversed from the new scheme. Also, make object types more consistent and remove spaces from object types that contain multiple parts ("posix_sem" -> "posixsem") to make mechanical parsing easier. Introduce a new "netinet" object type for certain IPv4/IPv6-related methods. Also simplify, slightly, some entry point names. All MAC policy modules will need to be recompiled, and modules not updates as part of this commit will need to be modified to conform to the new KPI. Sponsored by: SPARTA (original patches against Mac OS X) Obtained from: TrustedBSD Project, Apple Computer
2007-10-24 19:04:04 +00:00
if (mac_inpcb_check_deliver(inp, m))
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
goto dropunlock;
Introduce support for Mandatory Access Control and extensible kernel access control. Instrument the TCP socket code for packet generation and delivery: label outgoing mbufs with the label of the socket, and check socket and mbuf labels before permitting delivery to a socket. Assign labels to newly accepted connections when the syncache/cookie code has done its business. Also set peer labels as convenient. Currently, MAC policies cannot influence the PCB matching algorithm, so cannot implement polyinstantiation. Note that there is at least one case where a PCB is not available due to the TCP packet not being associated with any socket, so we don't label in that case, but need to handle it in a special manner. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-31 19:06:49 +00:00
#endif
Introduce a MAC label reference in 'struct inpcb', which caches the MAC label referenced from 'struct socket' in the IPv4 and IPv6-based protocols. This permits MAC labels to be checked during network delivery operations without dereferencing inp->inp_socket to get to so->so_label, which will eventually avoid our having to grab the socket lock during delivery at the network layer. This change introduces 'struct inpcb' as a labeled object to the MAC Framework, along with the normal circus of entry points: initialization, creation from socket, destruction, as well as a delivery access control check. For most policies, the inpcb label will simply be a cache of the socket label, so a new protocol switch method is introduced, pr_sosetlabel() to notify protocols that the socket layer label has been updated so that the cache can be updated while holding appropriate locks. Most protocols implement this using pru_sosetlabel_null(), but IPv4/IPv6 protocols using inpcbs use the the worker function in_pcbsosetlabel(), which calls into the MAC Framework to perform a cache update. Biba, LOMAC, and MLS implement these entry points, as do the stub policy, and test policy. Reviewed by: sam, bms Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-11-18 00:39:07 +00:00
so = inp->inp_socket;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
KASSERT(so != NULL, ("%s: so == NULL", __func__));
Made TCPDEBUG truely optional. Based on changes I made in FreeBSD 1.1.5. Fixed somebody's idea of a joke - about the first half of the lines in in_proto.c were spaced over by one space.
1994-09-15 10:36:56 +00:00
#ifdef TCPDEBUG
Unravel a nested conditional. Remove an unneeded local variable.
2002-12-20 11:16:52 +00:00
if (so->so_options & SO_DEBUG) {
ostate = tp->t_state;
o Unbreak "options TCPDEBUG" && "nooptions INET6" kernel build. PR: kern/112517 Submitted by: vd
2007-05-09 06:09:40 +00:00
#ifdef INET6
Fix compilation in case of defined(INET) && defined(IPFIREWALL_FORWARD) but no INET6. Reported by: avg Tested by: avg MFC after: 4 weeks X-MFC with: r225044 Approved by: re (kib)
2011-08-20 18:45:38 +00:00
if (isipv6) {
Unravel a nested conditional. Remove an unneeded local variable.
2002-12-20 11:16:52 +00:00
bcopy((char *)ip6, (char *)tcp_saveipgen, sizeof(*ip6));
o Fix style(9) bugs introduced in the last commit. Pointed out by: bde
2007-05-09 11:39:46 +00:00
} else
Fix compilation in case of defined(INET) && defined(IPFIREWALL_FORWARD) but no INET6. Reported by: avg Tested by: avg MFC after: 4 weeks X-MFC with: r225044 Approved by: re (kib)
2011-08-20 18:45:38 +00:00
#endif
Unravel a nested conditional. Remove an unneeded local variable.
2002-12-20 11:16:52 +00:00
bcopy((char *)ip, (char *)tcp_saveipgen, sizeof(*ip));
tcp_savetcp = *th;
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif /* TCPDEBUG */
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
/*
* When the socket is accepting connections (the INPCB is in LISTEN
* state) we look into the SYN cache if this is a new connection
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
* attempt or the completion of a previous one.
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
*/
Unravel a nested conditional. Remove an unneeded local variable.
2002-12-20 11:16:52 +00:00
if (so->so_options & SO_ACCEPTCONN) {
struct in_conninfo inc;
Add more KASSERT's.
2007-04-20 15:21:29 +00:00
KASSERT(tp->t_state == TCPS_LISTEN, ("%s: so accepting but "
"tp not listening", __func__));
Some cleanups and janitorial work to tcp_syncache: o don't assign remote/local host/port information manually between provided struct in_conninfo and struct syncache, bcopy() it instead o rename sc_tsrecent to sc_tsreflect in struct syncache to better capture the purpose of this field o rename sc_request_r_scale to sc_requested_r_scale for ditto reasons o fix IPSEC error case printf's to report correct function name o in syncache_socket() only transpose enhanced tcp options parameters to struct tcpcb when the inpcb doesn't has TF_NOOPT set o in syncache_respond() reorder stack variables o in syncache_respond() remove bogus KASSERT() No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 16:14:19 +00:00
bzero(&inc, sizeof(inc));
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
#ifdef INET6
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
if (isipv6) {
Use inc_flags instead of the inc_isipv6 alias which so far had been the only flag with random usage patterns. Switch inc_flags to be used as a real bit field by using INC_ISIPV6 with bitops to check for the 'isipv6' condition. While here fix a place or two where in case of v4 inc_flags were not properly initialized before.[1] Found by: rwatson during review [1] Discussed with: rwatson Reviewed by: rwatson MFC after: 4 weeks
2008-12-17 12:52:34 +00:00
inc.inc_flags |= INC_ISIPV6;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
inc.inc6_faddr = ip6->ip6_src;
inc.inc6_laddr = ip6->ip6_dst;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
} else
#endif
{
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
inc.inc_faddr = ip->ip_src;
inc.inc_laddr = ip->ip_dst;
}
inc.inc_fport = th->th_sport;
inc.inc_lport = th->th_dport;
Somewhere along the line accept sockets stopped honoring the FIB selected for them. Fix this. Reviewed by: ambrisko Approved by: re (kib) MFC after: 3 days
2009-07-28 19:43:27 +00:00
inc.inc_fibnum = so->so_fibnum;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
Tidy up some code to conform better to surroundings and style(9), 0 = NULL and space/tab.
2007-03-23 19:11:22 +00:00
/*
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* Check for an existing connection attempt in syncache if
* the flag is only ACK. A successful lookup creates a new
* socket appended to the listen queue in SYN_RECEIVED state.
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
*/
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
if ((thflags & (TH_RST|TH_ACK|TH_SYN)) == TH_ACK) {
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
/*
* Parse the TCP options here because
* syncookies need access to the reflected
* timestamp.
*/
tcp_dooptions(&to, optp, optlen, 0);
/*
* NB: syncache_expand() doesn't unlock
* inp and tcpinfo locks.
*/
if (!syncache_expand(&inc, &to, th, &so, m)) {
Rewrite of TCP syncookies to remove locking requirements and to enhance functionality: - Remove a rwlock aquisition/release per generated syncookie. Locking is now integrated with the bucket row locking of syncache itself and syncookies no longer add any additional lock overhead. - Syncookie secrets are different for and stored per syncache buck row. Secrets expire after 16 seconds and are reseeded on-demand. - The computational overhead for syncookie generation and verification is one MD5 hash computation as before. - Syncache can be turned off and run with syncookies only by setting the sysctl net.inet.tcp.syncookies_only=1. This implementation extends the orginal idea and first implementation of FreeBSD by using not only the initial sequence number field to store information but also the timestamp field if present. This way we can keep track of the entire state we need to know to recreate the session in its original form. Almost all TCP speakers implement RFC1323 timestamps these days. For those that do not we still have to live with the known shortcomings of the ISN only SYN cookies. The use of the timestamp field causes the timestamps to be randomized if syncookies are enabled. The idea of SYN cookies is to encode and include all necessary information about the connection setup state within the SYN-ACK we send back and thus to get along without keeping any local state until the ACK to the SYN-ACK arrives (if ever). Everything we need to know should be available from the information we encoded in the SYN-ACK. A detailed description of the inner working of the syncookies mechanism is included in the comments in tcp_syncache.c. Reviewed by: silby (slightly earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-09-13 13:08:27 +00:00
/*
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
* No syncache entry or ACK was not
* for our SYN/ACK. Send a RST.
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* NB: syncache did its own logging
* of the failure cause.
Rewrite of TCP syncookies to remove locking requirements and to enhance functionality: - Remove a rwlock aquisition/release per generated syncookie. Locking is now integrated with the bucket row locking of syncache itself and syncookies no longer add any additional lock overhead. - Syncookie secrets are different for and stored per syncache buck row. Secrets expire after 16 seconds and are reseeded on-demand. - The computational overhead for syncookie generation and verification is one MD5 hash computation as before. - Syncache can be turned off and run with syncookies only by setting the sysctl net.inet.tcp.syncookies_only=1. This implementation extends the orginal idea and first implementation of FreeBSD by using not only the initial sequence number field to store information but also the timestamp field if present. This way we can keep track of the entire state we need to know to recreate the session in its original form. Almost all TCP speakers implement RFC1323 timestamps these days. For those that do not we still have to live with the known shortcomings of the ISN only SYN cookies. The use of the timestamp field causes the timestamps to be randomized if syncookies are enabled. The idea of SYN cookies is to encode and include all necessary information about the connection setup state within the SYN-ACK we send back and thus to get along without keeping any local state until the ACK to the SYN-ACK arrives (if ever). Everything we need to know should be available from the information we encoded in the SYN-ACK. A detailed description of the inner working of the syncookies mechanism is included in the comments in tcp_syncache.c. Reviewed by: silby (slightly earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-09-13 13:08:27 +00:00
*/
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
rstreason = BANDLIM_RST_OPENPORT;
goto dropwithreset;
}
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
#ifdef TCP_RFC7413
new_tfo_socket:
#endif
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
if (so == NULL) {
o Fix INP lock leak in the minttl case o Remove indirection in the decision of unlocking inp o Further annotation of locking in tcp_input()
2007-04-23 19:41:47 +00:00
/*
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
* We completed the 3-way handshake
* but could not allocate a socket
* either due to memory shortage,
* listen queue length limits or
* global socket limits. Send RST
* or wait and have the remote end
* retransmit the ACK for another
* try.
o Fix INP lock leak in the minttl case o Remove indirection in the decision of unlocking inp o Further annotation of locking in tcp_input()
2007-04-23 19:41:47 +00:00
*/
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"Socket allocation failed due to "
"limits or memory shortage, %s\n",
A bunch of formatting fixes brough to light by, or created by the Vimage commit a few days ago.
2008-08-20 01:05:56 +00:00
s, __func__,
V_tcp_sc_rst_sock_fail ?
"sending RST" : "try again");
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if (V_tcp_sc_rst_sock_fail) {
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
rstreason = BANDLIM_UNLIMITED;
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
goto dropwithreset;
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
} else
goto dropunlock;
}
/*
* Socket is created in state SYN_RECEIVED.
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* Unlock the listen socket, lock the newly
* created socket and update the tp variable.
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
*/
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WUNLOCK(inp); /* listen socket */
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
inp = sotoinpcb(so);
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
/*
* New connection inpcb is already locked by
* syncache_expand().
*/
INP_WLOCK_ASSERT(inp);
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
tp = intotcpcb(inp);
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
KASSERT(tp->t_state == TCPS_SYN_RECEIVED,
("%s: ", __func__));
Add the possibility to verify MD5 hash of incoming TCP packets. As long as this is a costy function, even when compiled in (along with the option TCP_SIGNATURE), it can be disabled via the net.inet.tcp.signature_verify_input sysctl. Sponsored by: Sandvine Incorporated Reviewed by: emaste, bz MFC after: 2 weeks
2011-04-25 17:13:40 +00:00
#ifdef TCP_SIGNATURE
if (sig_checked == 0) {
tcp_dooptions(&to, optp, optlen,
(thflags & TH_SYN) ? TO_SYN : 0);
if (!tcp_signature_verify_input(m, off0, tlen,
optlen, &to, th, tp->t_flags)) {
/*
* In SYN_SENT state if it receives an
* RST, it is allowed for further
* processing.
*/
if ((thflags & TH_RST) == 0 ||
(tp->t_state == TCPS_SYN_SENT) == 0)
goto dropunlock;
}
sig_checked = 1;
}
#endif
Fix indentation of the syncache_expand() section in tcp_input().
2007-05-28 11:35:40 +00:00
/*
* Process the segment and the data it
* contains. tcp_do_segment() consumes
* the mbuf chain and unlocks the inpcb.
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
tp->t_fb->tfb_tcp_do_segment(m, th, so, tp, drop_hdrlen, tlen,
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
iptos, ti_locked);
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
}
/*
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* Segment flag validation for new connection attempts:
*
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
* Our (SYN|ACK) response was rejected.
* Check with syncache and remove entry to prevent
* retransmits.
o Move all detailed checks for RST in LISTEN state from tcp_input() to syncache_rst(). o Fix tests for flag combinations of RST and SYN, ACK, FIN. Before a RST for a connection in syncache did not properly free the entry. o Add more detailed logging. Approved by: re (rwatson)
2007-07-28 11:51:44 +00:00
*
* NB: syncache_chkrst does its own logging of failure
* causes.
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
*/
if (thflags & TH_RST) {
o Move all detailed checks for RST in LISTEN state from tcp_input() to syncache_rst(). o Fix tests for flag combinations of RST and SYN, ACK, FIN. Before a RST for a connection in syncache did not properly free the entry. o Add more detailed logging. Approved by: re (rwatson)
2007-07-28 11:51:44 +00:00
syncache_chkrst(&inc, th);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
goto dropunlock;
}
/*
* We can't do anything without SYN.
*/
if ((thflags & TH_SYN) == 0) {
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
"SYN is missing, segment ignored\n",
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
s, __func__);
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_badsyn);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
goto dropunlock;
}
/*
* (SYN|ACK) is bogus on a listen socket.
*/
if (thflags & TH_ACK) {
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"SYN|ACK invalid, segment rejected\n",
s, __func__);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
syncache_badack(&inc); /* XXX: Not needed! */
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_badsyn);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
rstreason = BANDLIM_RST_OPENPORT;
goto dropwithreset;
}
/*
* If the drop_synfin option is enabled, drop all
* segments with both the SYN and FIN bits set.
* This prevents e.g. nmap from identifying the
* TCP/IP stack.
* XXX: Poor reasoning. nmap has other methods
* and is constantly refining its stack detection
* strategies.
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* XXX: This is a violation of the TCP specification
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
* and was used by RFC1644.
*/
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if ((thflags & TH_FIN) && V_drop_synfin) {
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
"SYN|FIN segment ignored (based on "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"sysctl setting)\n", s, __func__);
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_badsyn);
Trim extra spaces before tabs.
2011-01-07 21:40:34 +00:00
goto dropunlock;
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
}
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
/*
* Segment's flags are (SYN) or (SYN|FIN).
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
*
* TH_PUSH, TH_URG, TH_ECE, TH_CWR are ignored
* as they do not affect the state of the TCP FSM.
* The data pointed to by TH_URG and th_urp is ignored.
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
*/
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
KASSERT((thflags & (TH_RST|TH_ACK)) == 0,
("%s: Listen socket: TH_RST or TH_ACK set", __func__));
KASSERT(thflags & (TH_SYN),
("%s: Listen socket: TH_SYN not set", __func__));
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
#ifdef INET6
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
/*
* If deprecated address is forbidden,
* we do not accept SYN to deprecated interface
* address to prevent any new inbound connection from
* getting established.
* When we do not accept SYN, we send a TCP RST,
* with deprecated source address (instead of dropping
* it). We compromise it as it is much better for peer
* to send a RST, and RST will be the final packet
* for the exchange.
*
* If we do not forbid deprecated addresses, we accept
* the SYN packet. RFC2462 does not suggest dropping
* SYN in this case.
* If we decipher RFC2462 5.5.4, it says like this:
* 1. use of deprecated addr with existing
* communication is okay - "SHOULD continue to be
* used"
* 2. use of it with new communication:
* (2a) "SHOULD NOT be used if alternate address
* with sufficient scope is available"
* (2b) nothing mentioned otherwise.
* Here we fall into (2b) case as we have no choice in
* our source address selection - we must obey the peer.
*
* The wording in RFC2462 is confusing, and there are
* multiple description text for deprecated address
* handling - worse, they are not exactly the same.
* I believe 5.5.4 is the best one, so we follow 5.5.4.
*/
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if (isipv6 && !V_ip6_use_deprecated) {
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
struct in6_ifaddr *ia6;
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
Remove ip6_getdstifaddr() and all functions to work with auxiliary data. It isn't safe to keep unreferenced ifaddrs. Use in6ifa_ifwithaddr() to determine ifaddr corresponding to destination address. Since currently we keep addresses with embedded scope zone, in6ifa_ifwithaddr is called with zero zoneid and marked with XXX. Also remove route and lle lookups from ip6_input. Use in6ifa_ifwithaddr() instead. Sponsored by: Yandex LLC
2014-11-08 19:38:34 +00:00
ia6 = in6ifa_ifwithaddr(&ip6->ip6_dst, 0 /* XXX */);
Modify most routines returning 'struct ifaddr *' to return references rather than pointers, requiring callers to properly dispose of those references. The following routines now return references: ifaddr_byindex ifa_ifwithaddr ifa_ifwithbroadaddr ifa_ifwithdstaddr ifa_ifwithnet ifaof_ifpforaddr ifa_ifwithroute ifa_ifwithroute_fib rt_getifa rt_getifa_fib IFP_TO_IA ip_rtaddr in6_ifawithifp in6ifa_ifpforlinklocal in6ifa_ifpwithaddr in6_ifadd carp_iamatch6 ip6_getdstifaddr Remove unused macro which didn't have required referencing: IFP_TO_IA6 This closes many small races in which changes to interface or address lists while an ifaddr was in use could lead to use of freed memory (etc). In a few cases, add missing if_addr_list locking required to safely acquire references. Because of a lack of deep copying support, we accept a race in which an in6_ifaddr pointed to by mbuf tags and extracted with ip6_getdstifaddr() doesn't hold a reference while in transmit. Once we have mbuf tag deep copy support, this can be fixed. Reviewed by: bz Obtained from: Apple, Inc. (portions) MFC after: 6 weeks (portions)
2009-06-23 20:19:09 +00:00
if (ia6 != NULL &&
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
(ia6->ia6_flags & IN6_IFF_DEPRECATED)) {
Modify most routines returning 'struct ifaddr *' to return references rather than pointers, requiring callers to properly dispose of those references. The following routines now return references: ifaddr_byindex ifa_ifwithaddr ifa_ifwithbroadaddr ifa_ifwithdstaddr ifa_ifwithnet ifaof_ifpforaddr ifa_ifwithroute ifa_ifwithroute_fib rt_getifa rt_getifa_fib IFP_TO_IA ip_rtaddr in6_ifawithifp in6ifa_ifpforlinklocal in6ifa_ifpwithaddr in6_ifadd carp_iamatch6 ip6_getdstifaddr Remove unused macro which didn't have required referencing: IFP_TO_IA6 This closes many small races in which changes to interface or address lists while an ifaddr was in use could lead to use of freed memory (etc). In a few cases, add missing if_addr_list locking required to safely acquire references. Because of a lack of deep copying support, we accept a race in which an in6_ifaddr pointed to by mbuf tags and extracted with ip6_getdstifaddr() doesn't hold a reference while in transmit. Once we have mbuf tag deep copy support, this can be fixed. Reviewed by: bz Obtained from: Apple, Inc. (portions) MFC after: 6 weeks (portions)
2009-06-23 20:19:09 +00:00
ifa_free(&ia6->ia_ifa);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"Connection attempt to deprecated "
"IPv6 address rejected\n",
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
s, __func__);
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
rstreason = BANDLIM_RST_OPENPORT;
goto dropwithreset;
Add a counter for the number of times the listen queue was overflowed to the tcpstat structure. (netstat -s) Reviewed by: wollman Obtained from: Steves, TCP/IP Ill. vol.3, page 189
1996-02-26 21:47:13 +00:00
}
Plug more refcount leaks and possible NULL deref for interface address list. Submitted by: scottl@ MFC after: 3 days
2012-06-04 18:43:51 +00:00
if (ia6)
ifa_free(&ia6->ia_ifa);
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif /* INET6 */
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
/*
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
* Basic sanity checks on incoming SYN requests:
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* Don't respond if the destination is a link layer
* broadcast according to RFC1122 4.2.3.10, p. 104.
* If it is from this socket it must be forged.
* Don't respond if the source or destination is a
* global or subnet broad- or multicast address.
* Note that it is quite possible to receive unicast
* link-layer packets with a broadcast IP address. Use
* in_broadcast() to find them.
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
*/
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
if (m->m_flags & (M_BCAST|M_MCAST)) {
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
"Connection attempt from broad- or multicast "
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
"link layer address ignored\n", s, __func__);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
goto dropunlock;
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
}
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
#ifdef INET6
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
if (isipv6) {
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
if (th->th_dport == th->th_sport &&
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &ip6->ip6_src)) {
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"Connection attempt to/from self "
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
"ignored\n", s, __func__);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
goto dropunlock;
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
}
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src)) {
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"Connection attempt from/to multicast "
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
"address ignored\n", s, __func__);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
goto dropunlock;
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
}
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#if defined(INET) && defined(INET6)
else
#endif
#ifdef INET
{
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
if (th->th_dport == th->th_sport &&
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
ip->ip_dst.s_addr == ip->ip_src.s_addr) {
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"Connection attempt from/to self "
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
"ignored\n", s, __func__);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
goto dropunlock;
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
}
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif)) {
if ((s = tcp_log_addrs(&inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Listen socket: "
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
"Connection attempt from/to broad- "
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
"or multicast address ignored\n",
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
s, __func__);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
goto dropunlock;
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
}
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
/*
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
* SYN appears to be valid. Create compressed TCP state
* for syncache.
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
*/
A number of patches in the last years have created new return paths in tcp_input that leave the function before hitting the tcp_trace function call for the TCPDEBUG option. This has made TCPDEBUG mostly useless (and tools like ports/benchmarks/dbs not working). Add tcp_trace calls to the return paths that could be identified in this maze. This is a NOP unless you compile with TCPDEBUG.
2003-08-13 08:46:54 +00:00
#ifdef TCPDEBUG
Remove bogus check for accept queue length and associated failure handling from the incoming SYN handling section of tcp_input(). Enforcement of the accept queue limits is done by sonewconn() after the 3WHS is completed. It is not necessary to have an earlier check before a connection request enters the SYN cache awaiting the full handshake. It rather limits the effectiveness of the syncache by preventing legit and illegit connections from entering it and having them shaken out before we hit the real limit which may have vanished by then. Change return value of syncache_add() to void. No status communication is required.
2007-04-20 14:34:54 +00:00
if (so->so_options & SO_DEBUG)
tcp_trace(TA_INPUT, ostate, tp,
(void *)tcp_saveipgen, &tcp_savetcp, 0);
A number of patches in the last years have created new return paths in tcp_input that leave the function before hitting the tcp_trace function call for the TCPDEBUG option. This has made TCPDEBUG mostly useless (and tools like ports/benchmarks/dbs not working). Add tcp_trace calls to the return paths that could be identified in this maze. This is a NOP unless you compile with TCPDEBUG.
2003-08-13 08:46:54 +00:00
#endif
dd DTrace probe points, translators and a corresponding script to provide the TCPDEBUG functionality with pure DTrace. Reviewed by: rwatson MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: D3530
2015-09-13 15:50:55 +00:00
TCP_PROBE3(debug__input, tp, th, mtod(m, const char *));
Remove bogus check for accept queue length and associated failure handling from the incoming SYN handling section of tcp_input(). Enforcement of the accept queue limits is done by sonewconn() after the 3WHS is completed. It is not necessary to have an earlier check before a connection request enters the SYN cache awaiting the full handshake. It rather limits the effectiveness of the syncache by preventing legit and illegit connections from entering it and having them shaken out before we hit the real limit which may have vanished by then. Change return value of syncache_add() to void. No status communication is required.
2007-04-20 14:34:54 +00:00
tcp_dooptions(&to, optp, optlen, TO_SYN);
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
#ifdef TCP_RFC7413
if (syncache_add(&inc, &to, th, inp, &so, m, NULL, NULL))
goto new_tfo_socket;
#else
- Updated TOE support in the kernel. - Stateful TCP offload drivers for Terminator 3 and 4 (T3 and T4) ASICs. These are available as t3_tom and t4_tom modules that augment cxgb(4) and cxgbe(4) respectively. The cxgb/cxgbe drivers continue to work as usual with or without these extra features. - iWARP driver for Terminator 3 ASIC (kernel verbs). T4 iWARP in the works and will follow soon. Build-tested with make universe. 30s overview ============ What interfaces support TCP offload? Look for TOE4 and/or TOE6 in the capabilities of an interface: # ifconfig -m | grep TOE Enable/disable TCP offload on an interface (just like any other ifnet capability): # ifconfig cxgbe0 toe # ifconfig cxgbe0 -toe Which connections are offloaded? Look for toe4 and/or toe6 in the output of netstat and sockstat: # netstat -np tcp | grep toe # sockstat -46c | grep toe Reviewed by: bz, gnn Sponsored by: Chelsio communications. MFC after: ~3 months (after 9.1, and after ensuring MFC is feasible)
2012-06-19 07:34:13 +00:00
syncache_add(&inc, &to, th, inp, &so, m, NULL, NULL);
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
#endif
Remove bogus check for accept queue length and associated failure handling from the incoming SYN handling section of tcp_input(). Enforcement of the accept queue limits is done by sonewconn() after the 3WHS is completed. It is not necessary to have an earlier check before a connection request enters the SYN cache awaiting the full handshake. It rather limits the effectiveness of the syncache by preventing legit and illegit connections from entering it and having them shaken out before we hit the real limit which may have vanished by then. Change return value of syncache_add() to void. No status communication is required.
2007-04-20 14:34:54 +00:00
/*
* Entry added to syncache and mbuf consumed.
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
* Only the listen socket is unlocked by syncache_add().
Remove bogus check for accept queue length and associated failure handling from the incoming SYN handling section of tcp_input(). Enforcement of the accept queue limits is done by sonewconn() after the 3WHS is completed. It is not necessary to have an earlier check before a connection request enters the SYN cache awaiting the full handshake. It rather limits the effectiveness of the syncache by preventing legit and illegit connections from entering it and having them shaken out before we hit the real limit which may have vanished by then. Change return value of syncache_add() to void. No status communication is required.
2007-04-20 14:34:54 +00:00
*/
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED) {
INP_INFO_RUNLOCK(&V_tcbinfo);
In tcp_input(), don't acquire the pcbinfo global write lock for SYN packets targeting a listening socket. Permit to reduce TCP input processing starvation in context of high SYN load (e.g. short-lived TCP connections or SYN flood). Submitted by: Julien Charbon <jcharbon@verisign.com> Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
2014-09-04 19:09:08 +00:00
ti_locked = TI_UNLOCKED;
}
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
Fix a race condition on tcp listen socket teardown with pending connections in the accept queue and contiguous new incoming SYNs. Compared to the original submitters patch I've moved the test next to the SYN handling to have it together in a logical unit and reworded the comment explaining the issue. Submitted by: Matt Miller <matt@matthewjmiller.net> Submitted by: Juan Mojica <jmojica@gmail.com> Reviewed by: Matt Miller (changes) Tested by: pho MFC after: 1 week
2013-04-09 20:52:26 +00:00
} else if (tp->t_state == TCPS_LISTEN) {
/*
* When a listen socket is torn down the SO_ACCEPTCONN
* flag is removed first while connections are drained
* from the accept queue in a unlock/lock cycle of the
* ACCEPT_LOCK, opening a race condition allowing a SYN
* attempt go through unhandled.
*/
goto dropunlock;
Back out my lats commit of locking down a socket, it conflicts with hsu's work. Requested by: hsu
2002-05-31 11:52:35 +00:00
}
Tidy up the ACCEPTCONN section of tcp_input(), ajust comments and remove old dead T/TCP code.
2007-03-21 18:49:43 +00:00
Add the possibility to verify MD5 hash of incoming TCP packets. As long as this is a costy function, even when compiled in (along with the option TCP_SIGNATURE), it can be disabled via the net.inet.tcp.signature_verify_input sysctl. Sponsored by: Sandvine Incorporated Reviewed by: emaste, bz MFC after: 2 weeks
2011-04-25 17:13:40 +00:00
#ifdef TCP_SIGNATURE
if (sig_checked == 0) {
tcp_dooptions(&to, optp, optlen,
(thflags & TH_SYN) ? TO_SYN : 0);
if (!tcp_signature_verify_input(m, off0, tlen, optlen, &to,
th, tp->t_flags)) {
/*
* In SYN_SENT state if it receives an RST, it is
* allowed for further processing.
*/
if ((thflags & TH_RST) == 0 ||
(tp->t_state == TCPS_SYN_SENT) == 0)
goto dropunlock;
}
sig_checked = 1;
}
#endif
Convert over the TCP probes to use mtod() rather than directly dereferencing m->m_data. Sponsored by: Netflix, Inc.
2013-11-25 22:55:06 +00:00
TCP_PROBE5(receive, NULL, tp, mtod(m, const char *), tp, th);
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
/*
Make log messages more verbose and simpler to understand for non-experts. Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
* Segment belongs to a connection in SYN_SENT, ESTABLISHED or later
* state. tcp_do_segment() always consumes the mbuf chain, unlocks
* the inpcb, and unlocks pcbinfo.
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
tp->t_fb->tfb_tcp_do_segment(m, th, so, tp, drop_hdrlen, tlen, iptos, ti_locked);
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
dropwithreset:
Convert over the TCP probes to use mtod() rather than directly dereferencing m->m_data. Sponsored by: Netflix, Inc.
2013-11-25 22:55:06 +00:00
TCP_PROBE5(receive, NULL, tp, mtod(m, const char *), tp, th);
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED) {
INP_INFO_RUNLOCK(&V_tcbinfo);
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
ti_locked = TI_UNLOCKED;
}
#ifdef INVARIANTS
else {
KASSERT(ti_locked == TI_UNLOCKED, ("%s: dropwithreset "
"ti_locked: %d", __func__, ti_locked));
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
}
#endif
As a follow-on to r183323, correct another case where ip_output() was called without an inpcb pointer despite holding the tcbinfo global lock, which lead to a deadlock or panic when ipfw tried to further acquire it recursively. Reported by: Stefan Ehmann <shoesoft at gmx dot net> MFC after: 3 days
2008-09-25 17:26:54 +00:00
if (inp != NULL) {
tcp_dropwithreset(m, th, tp, tlen, rstreason);
INP_WUNLOCK(inp);
In both dropwithreset paths in tcp_input.c, drop the tcbinfo lock sooner to decomplicate locking and eliminate the need for a rather chatty comment about why we have to handle the global lock in a special way for the benefit of ipfw and pf cred rules. MFC after: 3 days
2008-10-26 22:03:52 +00:00
} else
As a follow-on to r183323, correct another case where ip_output() was called without an inpcb pointer despite holding the tcbinfo global lock, which lead to a deadlock or panic when ipfw tried to further acquire it recursively. Reported by: Stefan Ehmann <shoesoft at gmx dot net> MFC after: 3 days
2008-09-25 17:26:54 +00:00
tcp_dropwithreset(m, th, NULL, tlen, rstreason);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
m = NULL; /* mbuf chain got consumed. */
As a follow-on to r183323, correct another case where ip_output() was called without an inpcb pointer despite holding the tcbinfo global lock, which lead to a deadlock or panic when ipfw tried to further acquire it recursively. Reported by: Stefan Ehmann <shoesoft at gmx dot net> MFC after: 3 days
2008-09-25 17:26:54 +00:00
goto drop;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
dropunlock:
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
if (m != NULL)
Convert over the TCP probes to use mtod() rather than directly dereferencing m->m_data. Sponsored by: Netflix, Inc.
2013-11-25 22:55:06 +00:00
TCP_PROBE5(receive, NULL, tp, mtod(m, const char *), tp, th);
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED) {
INP_INFO_RUNLOCK(&V_tcbinfo);
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
ti_locked = TI_UNLOCKED;
}
#ifdef INVARIANTS
else {
KASSERT(ti_locked == TI_UNLOCKED, ("%s: dropunlock "
"ti_locked: %d", __func__, ti_locked));
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
}
#endif
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
o Fix INP lock leak in the minttl case o Remove indirection in the decision of unlocking inp o Further annotation of locking in tcp_input()
2007-04-23 19:41:47 +00:00
if (inp != NULL)
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WUNLOCK(inp);
As a follow-on to r183323, correct another case where ip_output() was called without an inpcb pointer despite holding the tcbinfo global lock, which lead to a deadlock or panic when ipfw tried to further acquire it recursively. Reported by: Stefan Ehmann <shoesoft at gmx dot net> MFC after: 3 days
2008-09-25 17:26:54 +00:00
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
drop:
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
Refactor and rewrite in parts the SYN handling code on listen sockets in tcp_input(): o tighten the checks on allowed TCP flags to be RFC793 and tcp-secure conform o log check failures to syslog at LOG_DEBUG level o rearrange the code flow to be easier to follow o add KASSERTs to validate assumptions of the code flow Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable that controls the behavior on socket creation failure for a otherwise successful 3-way handshake. The socket creation can fail due to global memory shortage, listen queue limits and file descriptor limits. The sysctl allows to chose between two options to deal with this. One is to send a reset to the other endpoint to notify it about the failure (default). The other one is to ignore and treat the failure as a transient error and have the other endpoint retransmit for another try. Reviewed by: rwatson (in general)
2007-05-28 11:03:53 +00:00
if (s != NULL)
free(s, M_TCPLOG);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
if (m != NULL)
m_freem(m);
Merge 'struct ip6protosw' and 'struct protosw' into one. Now we have only one protocol switch structure that is shared between ipv4 and ipv6. Phabric: D476 Reviewed by: jhb
2014-08-08 01:57:15 +00:00
return (IPPROTO_DONE);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
}
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
tcp_do_segment(struct mbuf *m, struct tcphdr *th, struct socket *so,
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
struct tcpcb *tp, int drop_hdrlen, int tlen, uint8_t iptos,
int ti_locked)
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
{
One of the ways to detect loss is to count duplicate acks coming back from the other end till it reaches predetermined threshold which is 3 for us right now. Once that happens, we trigger fast-retransmit to do loss recovery. Main problem with the current implementation is that we don't honor SACK information well to detect whether an incoming ack is a dupack or not. RFC6675 has latest recommendations for that. According to it, dupack is a segment that arrives carrying a SACK block that identifies previously unknown information between snd_una and snd_max even if it carries new data, changes the advertised window, or moves the cumulative acknowledgment point. With the prevalence of Selective ACK (SACK) these days, improper handling can lead to delayed loss recovery. With the fix, new behavior looks like following: 0) th_ack < snd_una --> ignore Old acks are ignored. 1) th_ack == snd_una, !sack_changed --> ignore Acks with SACK enabled but without any new SACK info in them are ignored. 2) th_ack == snd_una, window == old_window --> increment Increment on a good dupack. 3) th_ack == snd_una, window != old_window, sack_changed --> increment When SACK enabled, it's okay to have advertized window changed if the ack has new SACK info. 4) th_ack > snd_una --> reset to 0 Reset to 0 when left edge moves. 5) th_ack > snd_una, sack_changed --> increment Increment if left edge moves but there is new SACK info. Here, sack_changed is the indicator that incoming ack has previously unknown SACK info in it. Note: This fix is not fully compliant to RFC6675. That may require a few changes to current implementation in order to keep per-sackhole dupack counter and change to the way we mark/handle sack holes. PR: 203663 Reviewed by: jtl MFC after: 3 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D4225
2015-12-08 21:21:48 +00:00
int thflags, acked, ourfinisacked, needoutput = 0, sack_changed;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
int rstreason, todrop, win;
u_long tiwin;
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
uint16_t nsegs;
Extend debug logging of TCP timestamp related specification violations. Update related comments and style.
2013-07-10 12:06:01 +00:00
char *s;
struct in_conninfo *inc;
When processing ACK in tcp_do_segment, use sbcut_locked() instead of sbdrop_locked() to cut acked mbufs from the socket buffer. Free this chain a batch manner after the socket buffer lock is dropped. This measurably reduces contention on socket buffer. Sponsored by: Netflix Sponsored by: Nginx, Inc. Approved by: re (marius)
2013-10-09 12:00:38 +00:00
struct mbuf *mfree;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
struct tcpopt to;
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
int tfo_syn;
o Use a define for a buffer size. Prodded by: db o Add missed vars for TCPDEBUG in tcp_do_segment(). Prodded by: tinderbox
2007-03-24 22:15:02 +00:00
#ifdef TCPDEBUG
/*
* The size of tcp_saveipgen must be the size of the max ip header,
* now IPv6.
*/
u_char tcp_saveipgen[IP6_HDR_LEN];
struct tcphdr tcp_savetcp;
short ostate = 0;
#endif
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
thflags = th->th_flags;
Extend debug logging of TCP timestamp related specification violations. Update related comments and style.
2013-07-10 12:06:01 +00:00
inc = &tp->t_inpcb->inp_inc;
Reset the last_sack_ack SACK hint for TCP input processing to ensure that the hint is 0 when no SACK data is received to update the hint with. This was accidentally omitted from r216753. Sponsored by: FreeBSD Foundation MFC after: 10 weeks X-MFC with: 216753
2011-01-10 06:12:01 +00:00
tp->sackhint.last_sack_ack = 0;
One of the ways to detect loss is to count duplicate acks coming back from the other end till it reaches predetermined threshold which is 3 for us right now. Once that happens, we trigger fast-retransmit to do loss recovery. Main problem with the current implementation is that we don't honor SACK information well to detect whether an incoming ack is a dupack or not. RFC6675 has latest recommendations for that. According to it, dupack is a segment that arrives carrying a SACK block that identifies previously unknown information between snd_una and snd_max even if it carries new data, changes the advertised window, or moves the cumulative acknowledgment point. With the prevalence of Selective ACK (SACK) these days, improper handling can lead to delayed loss recovery. With the fix, new behavior looks like following: 0) th_ack < snd_una --> ignore Old acks are ignored. 1) th_ack == snd_una, !sack_changed --> ignore Acks with SACK enabled but without any new SACK info in them are ignored. 2) th_ack == snd_una, window == old_window --> increment Increment on a good dupack. 3) th_ack == snd_una, window != old_window, sack_changed --> increment When SACK enabled, it's okay to have advertized window changed if the ack has new SACK info. 4) th_ack > snd_una --> reset to 0 Reset to 0 when left edge moves. 5) th_ack > snd_una, sack_changed --> increment Increment if left edge moves but there is new SACK info. Here, sack_changed is the indicator that incoming ack has previously unknown SACK info in it. Note: This fix is not fully compliant to RFC6675. That may require a few changes to current implementation in order to keep per-sackhole dupack counter and change to the way we mark/handle sack holes. PR: 203663 Reviewed by: jtl MFC after: 3 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D4225
2015-12-08 21:21:48 +00:00
sack_changed = 0;
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
nsegs = max(1, m->m_pkthdr.lro_nsegs);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
/*
* If this is either a state-changing packet or current state isn't
* established, we require a write lock on tcbinfo. Otherwise, we
Update some stale comments regarding tcbinfo locking in the TCP input path: read locks on tcbinfo are no longer used, so won't happen. No functional change. MFC after: 3 days
2012-07-22 17:31:36 +00:00
* allow the tcbinfo to be in either alocked or unlocked, as the
* caller may have unnecessarily acquired a write lock due to a race.
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
*/
if ((thflags & (TH_SYN | TH_FIN | TH_RST)) != 0 ||
tp->t_state != TCPS_ESTABLISHED) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
KASSERT(ti_locked == TI_RLOCKED, ("%s ti_locked %d for "
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
"SYN/FIN/RST/!EST", __func__, ti_locked));
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
} else {
#ifdef INVARIANTS
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED)
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
else {
KASSERT(ti_locked == TI_UNLOCKED, ("%s: EST "
"ti_locked: %d", __func__, ti_locked));
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
}
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
#endif
}
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Remove the defunct remains of the TCPS_TIME_WAIT cases from tcp_do_segment and change it to a void function. We use a compressed structure for TCPS_TIME_WAIT to save memory. Any late late segments arriving for such a connection is handled directly in the TW code.
2007-05-06 15:16:05 +00:00
KASSERT(tp->t_state > TCPS_LISTEN, ("%s: TCPS_LISTEN",
__func__));
KASSERT(tp->t_state != TCPS_TIME_WAIT, ("%s: TCPS_TIME_WAIT",
__func__));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
There are times when it would be really nice to have a record of the last few packets and/or state transitions from each TCP socket. That would help with narrowing down certain problems we see in the field that are hard to reproduce without understanding the history of how we got into a certain state. This change provides just that. It saves copies of the last N packets in a list in the tcpcb. When the tcpcb is destroyed, the list is freed. I thought this was likely to be more performance-friendly than saving copies of the tcpcb. Plus, with the packets, you should be able to reverse-engineer what happened to the tcpcb. To enable the feature, you will need to compile a kernel with the TCPPCAP option. Even then, the feature defaults to being deactivated. You can activate it by setting a positive value for the number of captured packets. You can do that on either a global basis or on a per-socket basis (via a setsockopt call). There is no way to get the packets out of the kernel other than using kmem or getting a coredump. I thought that would help some of the legal/privacy concerns regarding such a feature. However, it should be possible to add a future effort to export them in PCAP format. I tested this at low scale, and found that there were no mbuf leaks and the peak mbuf usage appeared to be unchanged with and without the feature. The main performance concern I can envision is the number of mbufs that would be used on systems with a large number of sockets. If you save five packets per direction per socket and have 3,000 sockets, that will consume at least 30,000 mbufs just to keep these packets. I tried to reduce the concerns associated with this by limiting the number of clusters (not mbufs) that could be used for this feature. Again, in my testing, that appears to work correctly. Differential Revision: D3100 Submitted by: Jonathan Looney <jlooney at juniper dot net> Reviewed by: gnn, hiren
2015-10-14 00:35:37 +00:00
#ifdef TCPPCAP
/* Save segment, if requested. */
tcp_pcap_add(th, m, &(tp->t_inpkts));
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Segment received on connection.
* Reset idle time and keep-alive timer.
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
* XXX: This should be done after segment
* validation to ignore broken/spoofed segs.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
tp->t_rcvtime = ticks;
Receipt of two SYN's are sufficient to set the t_timer[TCPT_KEEP] to "keepidle". this should not occur unless the connection has been established via the 3-way handshake which requires an ACK Submitted by: jmb Obtained from: problem discussed in Stevens vol. 3
1996-09-13 18:47:03 +00:00
if (TCPS_HAVEESTABLISHED(tp->t_state))
Add new socket options: TCP_KEEPINIT, TCP_KEEPIDLE, TCP_KEEPINTVL and TCP_KEEPCNT, that allow to control initial timeout, idle time, idle re-send interval and idle send count on a per-socket basis. Reviewed by: andre, bz, lstewart
2012-02-05 16:53:02 +00:00
tcp_timer_activate(tp, TT_KEEP, TP_KEEPIDLE(tp));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Rework TCP window scaling (RFC1323) to properly scale the send window right from the beginning and partly clean up the differences in handling between SYN_SENT and SYN_RCVD (syncache). Further changes to this code to come. This is a first incremental step to a general overhaul and streamlining of the TCP code. PR: kern/15095 PR: kern/92690 (partly) Reviewed by: qingli (and tested with ANVL) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-28 23:05:59 +00:00
/*
Correct the wording as we are increasing the window size. Reviewed by: jhb Sponsored by: Limelight Networks
2015-05-19 19:17:20 +00:00
* Scale up the window into a 32-bit value.
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
* For the SYN_SENT state the scale is zero.
Rework TCP window scaling (RFC1323) to properly scale the send window right from the beginning and partly clean up the differences in handling between SYN_SENT and SYN_RCVD (syncache). Further changes to this code to come. This is a first incremental step to a general overhaul and streamlining of the TCP code. PR: kern/15095 PR: kern/92690 (partly) Reviewed by: qingli (and tested with ANVL) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-28 23:05:59 +00:00
*/
tiwin = th->th_win << tp->snd_scale;
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
/*
* TCP ECN processing.
*/
if (tp->t_flags & TF_ECN_PERMIT) {
Honor the CE bit even when the CWR bit is set. PR: 145600 Submitted by: Richard Scheffenegger <rs at netapp.com> MFC after: 1 week
2010-04-10 12:47:06 +00:00
if (thflags & TH_CWR)
tp->t_flags &= ~TF_ECN_SND_ECE;
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
switch (iptos & IPTOS_ECN_MASK) {
case IPTOS_ECN_CE:
tp->t_flags |= TF_ECN_SND_ECE;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_ecn_ce);
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
break;
case IPTOS_ECN_ECT0:
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_ecn_ect0);
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
break;
case IPTOS_ECN_ECT1:
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_ecn_ect1);
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
break;
}
DCTCP (Data Center TCP) implementation. DCTCP congestion control algorithm aims to maximise throughput and minimise latency in data center networks by utilising the proportion of Explicit Congestion Notification (ECN) marked packets received from capable hardware as a congestion signal. Highlights: Implemented as a mod_cc(4) module. ECN (Explicit congestion notification) processing is done differently from RFC3168. Takes one-sided DCTCP into consideration where only one of the sides is using DCTCP and other is using standard ECN. IETF draft: http://tools.ietf.org/html/draft-bensley-tcpm-dctcp-00 Thesis report by Midori Kato: https://eggert.org/students/kato-thesis.pdf Submitted by: Midori Kato <katoon@sfc.wide.ad.jp> and Lars Eggert <lars@netapp.com> with help and modifications from hiren Differential Revision: https://reviews.freebsd.org/D604 Reviewed by: gnn
2015-01-12 08:33:04 +00:00
/* Process a packet differently from RFC3168. */
cc_ecnpkt_handler(tp, th, iptos);
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
/* Congestion experienced. */
if (thflags & TH_ECE) {
cc_cong_signal(tp, th, CC_ECN);
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
}
}
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
/*
* Parse options on any incoming segment.
*/
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
tcp_dooptions(&to, (u_char *)(th + 1),
(th->th_off << 2) - sizeof(struct tcphdr),
(thflags & TH_SYN) ? TO_SYN : 0);
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
/*
* If echoed timestamp is later than the current time,
Rewrite of TCP syncookies to remove locking requirements and to enhance functionality: - Remove a rwlock aquisition/release per generated syncookie. Locking is now integrated with the bucket row locking of syncache itself and syncookies no longer add any additional lock overhead. - Syncookie secrets are different for and stored per syncache buck row. Secrets expire after 16 seconds and are reseeded on-demand. - The computational overhead for syncookie generation and verification is one MD5 hash computation as before. - Syncache can be turned off and run with syncookies only by setting the sysctl net.inet.tcp.syncookies_only=1. This implementation extends the orginal idea and first implementation of FreeBSD by using not only the initial sequence number field to store information but also the timestamp field if present. This way we can keep track of the entire state we need to know to recreate the session in its original form. Almost all TCP speakers implement RFC1323 timestamps these days. For those that do not we still have to live with the known shortcomings of the ISN only SYN cookies. The use of the timestamp field causes the timestamps to be randomized if syncookies are enabled. The idea of SYN cookies is to encode and include all necessary information about the connection setup state within the SYN-ACK we send back and thus to get along without keeping any local state until the ACK to the SYN-ACK arrives (if ever). Everything we need to know should be available from the information we encoded in the SYN-ACK. A detailed description of the inner working of the syncookies mechanism is included in the comments in tcp_syncache.c. Reviewed by: silby (slightly earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-09-13 13:08:27 +00:00
* fall back to non RFC1323 RTT calculation. Normalize
* timestamp if syncookies were used when this connection
* was established.
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
*/
Rewrite of TCP syncookies to remove locking requirements and to enhance functionality: - Remove a rwlock aquisition/release per generated syncookie. Locking is now integrated with the bucket row locking of syncache itself and syncookies no longer add any additional lock overhead. - Syncookie secrets are different for and stored per syncache buck row. Secrets expire after 16 seconds and are reseeded on-demand. - The computational overhead for syncookie generation and verification is one MD5 hash computation as before. - Syncache can be turned off and run with syncookies only by setting the sysctl net.inet.tcp.syncookies_only=1. This implementation extends the orginal idea and first implementation of FreeBSD by using not only the initial sequence number field to store information but also the timestamp field if present. This way we can keep track of the entire state we need to know to recreate the session in its original form. Almost all TCP speakers implement RFC1323 timestamps these days. For those that do not we still have to live with the known shortcomings of the ISN only SYN cookies. The use of the timestamp field causes the timestamps to be randomized if syncookies are enabled. The idea of SYN cookies is to encode and include all necessary information about the connection setup state within the SYN-ACK we send back and thus to get along without keeping any local state until the ACK to the SYN-ACK arrives (if ever). Everything we need to know should be available from the information we encoded in the SYN-ACK. A detailed description of the inner working of the syncookies mechanism is included in the comments in tcp_syncache.c. Reviewed by: silby (slightly earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-09-13 13:08:27 +00:00
if ((to.to_flags & TOF_TS) && (to.to_tsecr != 0)) {
fix calculating to_tsecr... This prevents the rtt calculations from going all wonky...
2006-09-26 01:21:46 +00:00
to.to_tsecr -= tp->ts_offset;
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
if (TSTMP_GT(to.to_tsecr, tcp_ts_getticks()))
Rewrite of TCP syncookies to remove locking requirements and to enhance functionality: - Remove a rwlock aquisition/release per generated syncookie. Locking is now integrated with the bucket row locking of syncache itself and syncookies no longer add any additional lock overhead. - Syncookie secrets are different for and stored per syncache buck row. Secrets expire after 16 seconds and are reseeded on-demand. - The computational overhead for syncookie generation and verification is one MD5 hash computation as before. - Syncache can be turned off and run with syncookies only by setting the sysctl net.inet.tcp.syncookies_only=1. This implementation extends the orginal idea and first implementation of FreeBSD by using not only the initial sequence number field to store information but also the timestamp field if present. This way we can keep track of the entire state we need to know to recreate the session in its original form. Almost all TCP speakers implement RFC1323 timestamps these days. For those that do not we still have to live with the known shortcomings of the ISN only SYN cookies. The use of the timestamp field causes the timestamps to be randomized if syncookies are enabled. The idea of SYN cookies is to encode and include all necessary information about the connection setup state within the SYN-ACK we send back and thus to get along without keeping any local state until the ACK to the SYN-ACK arrives (if ever). Everything we need to know should be available from the information we encoded in the SYN-ACK. A detailed description of the inner working of the syncookies mechanism is included in the comments in tcp_syncache.c. Reviewed by: silby (slightly earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-09-13 13:08:27 +00:00
to.to_tsecr = 0;
}
Extend debug logging of TCP timestamp related specification violations. Update related comments and style.
2013-07-10 12:06:01 +00:00
/*
* If timestamps were negotiated during SYN/ACK they should
* appear on every segment during this session and vice versa.
*/
if ((tp->t_flags & TF_RCVD_TSTMP) && !(to.to_flags & TOF_TS)) {
if ((s = tcp_log_addrs(inc, th, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: Timestamp missing, "
"no action\n", s, __func__);
free(s, M_TCPLOG);
}
}
if (!(tp->t_flags & TF_RCVD_TSTMP) && (to.to_flags & TOF_TS)) {
if ((s = tcp_log_addrs(inc, th, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: Timestamp not expected, "
"no action\n", s, __func__);
free(s, M_TCPLOG);
}
}
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
* Process options only when we get SYN/ACK back. The SYN case
* for incoming connections is handled in tcp_syncache.
Remove some bogosity from the SYN_SENT case in tcp_do_segment and simplify handling of the send/receive window scaling. No change in effective behavour. RFC1323 requires the window field in a SYN (i.e., a <SYN> or <SYN,ACK>) segment itself never be scaled. Noticed by: yar
2007-06-09 21:09:49 +00:00
* According to RFC1323 the window field in a SYN (i.e., a <SYN>
* or <SYN,ACK>) segment itself is never scaled.
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
* XXX this is traditional behavior, may need to be cleaned up.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Fix ipfw packet matching errors with address tables. The ipfw tables lookup code caches the result of the last query. The kernel may process multiple packets concurrently, performing several concurrent table lookups. Due to an insufficient locking, a cached result can become corrupted that could cause some addresses to be incorrectly matched against a lookup table. Submitted by: ru Reviewed by: csjp, mlaier Security: CAN-2005-2019 Security: FreeBSD-SA-05:13.ipfw Correct bzip2 permission race condition vulnerability. Obtained from: Steve Grubb via RedHat Security: CAN-2005-0953 Security: FreeBSD-SA-05:14.bzip2 Approved by: obrien Correct TCP connection stall denial of service vulnerability. A TCP packets with the SYN flag set is accepted for established connections, allowing an attacker to overwrite certain TCP options. Submitted by: Noritoshi Demizu Reviewed by: andre, Mohan Srinivasan Security: CAN-2005-2068 Security: FreeBSD-SA-05:15.tcp Approved by: re (security blanket), cperciva
2005-06-29 21:36:49 +00:00
if (tp->t_state == TCPS_SYN_SENT && (thflags & TH_SYN)) {
Rework TCP window scaling (RFC1323) to properly scale the send window right from the beginning and partly clean up the differences in handling between SYN_SENT and SYN_RCVD (syncache). Further changes to this code to come. This is a first incremental step to a general overhaul and streamlining of the TCP code. PR: kern/15095 PR: kern/92690 (partly) Reviewed by: qingli (and tested with ANVL) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-28 23:05:59 +00:00
if ((to.to_flags & TOF_SCALE) &&
(tp->t_flags & TF_REQ_SCALE)) {
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
tp->t_flags |= TF_RCVD_SCALE;
Consolidate insertion of TCP options into a segment from within tcp_output() and syncache_respond() into its own generic function tcp_addoptions(). tcp_addoptions() is alignment agnostic and does optimal packing in all cases. In struct tcpopt rename to_requested_s_scale to just to_wscale. Add a comment with quote from RFC1323: "The Window field in a SYN (i.e., a <SYN> or <SYN,ACK>) segment itself is never scaled." Reviewed by: silby, mohans, julian Sponsored by: TCP/IP Optimization Fundraise 2005
2007-03-15 15:59:28 +00:00
tp->snd_scale = to.to_wscale;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
}
Remove some bogosity from the SYN_SENT case in tcp_do_segment and simplify handling of the send/receive window scaling. No change in effective behavour. RFC1323 requires the window field in a SYN (i.e., a <SYN> or <SYN,ACK>) segment itself never be scaled. Noticed by: yar
2007-06-09 21:09:49 +00:00
/*
* Initial send window. It will be updated with
* the next incoming segment to the scaled value.
*/
tp->snd_wnd = th->th_win;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
if (to.to_flags & TOF_TS) {
tp->t_flags |= TF_RCVD_TSTMP;
tp->ts_recent = to.to_tsval;
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
tp->ts_recent_age = tcp_ts_getticks();
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
}
if (to.to_flags & TOF_MSS)
tcp_mss(tp, to.to_mss);
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if ((tp->t_flags & TF_SACK_PERMIT) &&
(to.to_flags & TOF_SACKPERM) == 0)
tp->t_flags &= ~TF_SACK_PERMIT;
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
}
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
/*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* Header prediction: check for the two common cases
* of a uni-directional data xfer. If the packet has
* no control flags, is in-sequence, the window didn't
* change and we're not retransmitting, it's a
* candidate. If the length is zero and the ack moved
* forward, we're the sender side of the xfer. Just
* free the data acked & wake any higher level process
* that was blocked waiting for space. If the length
* is non-zero and the ack didn't move, we're the
* receiver side. If we're getting packets in-order
* (the reassembly queue is empty), add the data to
* the socket buffer and note that we need a delayed ack.
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
* Make sure that the hidden state-flags are also off.
Reorder the TCP header prediction test to check for the most volatile values first to spend less time on a fallback to normal processing.
2007-05-06 15:23:51 +00:00
* Since we check for TCPS_ESTABLISHED first, it can only
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
* be TH_NEEDSYN.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
if (tp->t_state == TCPS_ESTABLISHED &&
Reorder the TCP header prediction test to check for the most volatile values first to spend less time on a fallback to normal processing.
2007-05-06 15:23:51 +00:00
th->th_seq == tp->rcv_nxt &&
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
(thflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK &&
Reorder the TCP header prediction test to check for the most volatile values first to spend less time on a fallback to normal processing.
2007-05-06 15:23:51 +00:00
tp->snd_nxt == tp->snd_max &&
tiwin && tiwin == tp->snd_wnd &&
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
((tp->t_flags & (TF_NEEDSYN|TF_NEEDFIN)) == 0) &&
Revert r265338, r271089 and r271123 as those changes do not handle non-inline urgent data and introduce an mbuf exhaustion attack vector similar to FreeBSD-SA-15:15.tcp, but not requiring VNETs. Address the issue described in FreeBSD-SA-15:15.tcp. Reviewed by: glebius Approved by: so Approved by: jmallett (mentor) Security: FreeBSD-SA-15:15.tcp Sponsored by: Norse Corp, Inc.
2015-07-29 17:59:13 +00:00
LIST_EMPTY(&tp->t_segq) &&
((to.to_flags & TOF_TS) == 0 ||
Reorder the TCP header prediction test to check for the most volatile values first to spend less time on a fallback to normal processing.
2007-05-06 15:23:51 +00:00
TSTMP_GEQ(to.to_tsval, tp->ts_recent)) ) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
/*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* If last ACK falls within this segment's sequence numbers,
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
* record the timestamp.
* NOTE that the test is modified according to the latest
* proposal of the tcplw@cray.com list (Braden 1993/04/26).
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
if ((to.to_flags & TOF_TS) != 0 &&
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
SEQ_LEQ(th->th_seq, tp->last_ack_sent)) {
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
tp->ts_recent_age = tcp_ts_getticks();
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tp->ts_recent = to.to_tsval;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (tlen == 0) {
if (SEQ_GT(th->th_ack, tp->snd_una) &&
SEQ_LEQ(th->th_ack, tp->snd_max) &&
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
!IN_RECOVERY(tp->t_flags) &&
(to.to_flags & TOF_SACK) == 0 &&
TAILQ_EMPTY(&tp->snd_holes)) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
* This is a pure ack for outstanding data.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED)
INP_INFO_RUNLOCK(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
ti_locked = TI_UNLOCKED;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_predack);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
/*
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
* "bad retransmit" recovery.
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
*/
if (tp->t_rxtshift == 1 &&
TCP reuses t_rxtshift to determine the backoff timer used for both the persist state and the retransmit timer. However, the code that implements "bad retransmit recovery" only checks t_rxtshift to see if an ACK has been received in during the first retransmit timeout window. As a result, if ticks has wrapped over to a negative value and a socket is in the persist state, it can incorrectly treat an ACK from the remote peer as a "bad retransmit recovery" and restore saved values such as snd_ssthresh and snd_cwnd. However, if the socket has never had a retransmit timeout, then these saved values will be zero, so snd_ssthresh and snd_cwnd will be set to 0. If the socket is in fast recovery (this can be caused by excessive duplicate ACKs such as those fixed by 220794), then each ACK that arrives triggers either NewReno or SACK partial ACK handling which clamps snd_cwnd to be no larger than snd_ssthresh. In effect, the socket's send window is permamently stuck at 0 even though the remote peer is advertising a much larger window and pending data is only sent via TCP window probes (so one byte every few seconds). Fix this by adding a new TCP pcb flag (TF_PREVVALID) that indicates that the various snd_*_prev fields in the pcb are valid and only perform "bad retransmit recovery" if this flag is set in the pcb. The flag is set on the first retransmit timeout that occurs and is cleared on subsequent retransmit timeouts or when entering the persist state. Reviewed by: bz MFC after: 2 weeks
2011-04-29 15:40:12 +00:00
tp->t_flags & TF_PREVVALID &&
Fix edge cases with ticks wrapping from INT_MAX to INT_MIN in the handling of the per-tcpcb t_badtrxtwin. Submitted by: bde
2009-06-16 19:00:12 +00:00
(int)(ticks - tp->t_badrxtwin) < 0) {
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
cc_cong_signal(tp, th, CC_RTO_ERR);
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
}
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
/*
* Recalculate the transmit timer / rtt.
*
* Some boxes send broken timestamp replies
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
* during the SYN+ACK phase, ignore
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
* timestamps of 0 or we could calculate a
* huge RTT and blow up the retransmit timer.
*/
if ((to.to_flags & TOF_TS) != 0 &&
to.to_tsecr) {
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
u_int t;
t = tcp_ts_getticks() - to.to_tsecr;
if (!tp->t_rttlow || tp->t_rttlow > t)
tp->t_rttlow = t;
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tcp_xmit_timer(tp,
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
TCP_TS_TO_TICKS(t) + 1);
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
} else if (tp->t_rtttime &&
Tidy up some code to conform better to surroundings and style(9), 0 = NULL and space/tab.
2007-03-23 19:11:22 +00:00
SEQ_GT(th->th_ack, tp->t_rtseq)) {
Have TCP Inflight disable itself if the RTT is below a certain threshold. Inflight doesn't make sense on a LAN as it has trouble figuring out the maximal bandwidth because of the coarse tick granularity. The sysctl net.inet.tcp.inflight.rttthresh specifies the threshold in milliseconds below which inflight will disengage. It defaults to 10ms. Tested by: Joao Barros <joao.barros-at-gmail.com>, Rich Murphey <rich-at-whiteoaklabs.com> Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-16 19:38:07 +00:00
if (!tp->t_rttlow ||
tp->t_rttlow > ticks - tp->t_rtttime)
tp->t_rttlow = ticks - tp->t_rtttime;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
tcp_xmit_timer(tp,
ticks - tp->t_rtttime);
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
}
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
acked = BYTES_THIS_ACK(tp, th);
- Add some helper hook points to the TCP stack. The hooks allow Khelp modules to access inbound/outbound events and associated data for established TCP connections. The hooks only run if at least one hook function is registered for the hook point, ensuring the impact on the stack is effectively nil when no TCP Khelp modules are loaded. struct tcp_hhook_data is passed as contextual data to any registered Khelp module hook functions. - Add an OSD (Object Specific Data) pointer to struct tcpcb to allow Khelp modules to associate per-connection data with the TCP control block. - Bump __FreeBSD_version and add a note to UPDATING regarding to ABI changes introduced by this commit and r216753. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: FreeBSD Foundation Reviewed by: bz, others along the way MFC after: 3 months
2010-12-28 12:13:30 +00:00
/* Run HHOOK_TCP_ESTABLISHED_IN helper hooks. */
hhook_run_tcp_est_in(tp, th, &to);
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
TCPSTAT_ADD(tcps_rcvackpack, nsegs);
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_ADD(tcps_rcvackbyte, acked);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
sbdrop(&so->so_snd, acked);
Unify the "send high" and "recover" variables as specified in the lastest rev of the spec. Use an explicit flag for Fast Recovery. [1] Fix bug with exiting Fast Recovery on a retransmit timeout diagnosed by Lu Guohan. [2] Reviewed by: Thomas Henderson <thomas.r.henderson@boeing.com> Reported and tested by: Lu Guohan <lguohan00@mails.tsinghua.edu.cn> [2] Approved by: Thomas Henderson <thomas.r.henderson@boeing.com>, Sally Floyd <floyd@acm.org> [1]
2003-07-15 21:49:53 +00:00
if (SEQ_GT(tp->snd_una, tp->snd_recover) &&
SEQ_LEQ(th->th_ack, tp->snd_recover))
tp->snd_recover = th->th_ack - 1;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
/*
* Let the congestion control algorithm update
* congestion control related information. This
* typically means increasing the congestion
* window.
*/
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
cc_ack_received(tp, th, nsegs, CC_ACK);
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
Unify the "send high" and "recover" variables as specified in the lastest rev of the spec. Use an explicit flag for Fast Recovery. [1] Fix bug with exiting Fast Recovery on a retransmit timeout diagnosed by Lu Guohan. [2] Reviewed by: Thomas Henderson <thomas.r.henderson@boeing.com> Reported and tested by: Lu Guohan <lguohan00@mails.tsinghua.edu.cn> [2] Approved by: Thomas Henderson <thomas.r.henderson@boeing.com>, Sally Floyd <floyd@acm.org> [1]
2003-07-15 21:49:53 +00:00
tp->snd_una = th->th_ack;
- Consistently update snd_wl1, snd_wl2, and rcv_up in the header prediction code. Previously, 2GB worth of header predicted data could leave these variables too far out of sequence which would cause problems after receiving a packet that did not match the header prediction. Submitted by: Bill Baumann <bbaumann@isilon.com> Sponsored by: Isilon Systems, Inc. Reviewed by: hsu, pete@isilon.com, neal@isilon.com, aaronp@isilon.com
2002-10-31 23:24:13 +00:00
/*
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
* Pull snd_wl2 up to prevent seq wrap relative
- Consistently update snd_wl1, snd_wl2, and rcv_up in the header prediction code. Previously, 2GB worth of header predicted data could leave these variables too far out of sequence which would cause problems after receiving a packet that did not match the header prediction. Submitted by: Bill Baumann <bbaumann@isilon.com> Sponsored by: Isilon Systems, Inc. Reviewed by: hsu, pete@isilon.com, neal@isilon.com, aaronp@isilon.com
2002-10-31 23:24:13 +00:00
* to th_ack.
*/
Fix NewReno. Reviewed by: Tom Henderson <thomas.r.henderson@boeing.com>
2003-01-13 11:01:20 +00:00
tp->snd_wl2 = th->th_ack;
Reset dupack count in header prediction. Follow-on to rev 1.39. Reviewed by: jayanth, Thomas R Henderson <thomas.r.henderson@boeing.com>, silby, dillon
2002-08-15 17:13:18 +00:00
tp->t_dupacks = 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
m_freem(m);
/*
* If all outstanding data are acked, stop
* retransmit timer, otherwise restart timer
* using current (possibly backed-off) value.
* If process is waiting for space,
* wakeup/selwakeup/signal. If data
* are ready to send, let tcp_output
* decide between more output or persist.
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
*/
A number of patches in the last years have created new return paths in tcp_input that leave the function before hitting the tcp_trace function call for the TCPDEBUG option. This has made TCPDEBUG mostly useless (and tools like ports/benchmarks/dbs not working). Add tcp_trace calls to the return paths that could be identified in this maze. This is a NOP unless you compile with TCPDEBUG.
2003-08-13 08:46:54 +00:00
#ifdef TCPDEBUG
if (so->so_options & SO_DEBUG)
tcp_trace(TA_INPUT, ostate, tp,
(void *)tcp_saveipgen,
&tcp_savetcp, 0);
#endif
dd DTrace probe points, translators and a corresponding script to provide the TCPDEBUG functionality with pure DTrace. Reviewed by: rwatson MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: D3530
2015-09-13 15:50:55 +00:00
TCP_PROBE3(debug__input, tp, th,
mtod(m, const char *));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (tp->snd_una == tp->snd_max)
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
tcp_timer_activate(tp, TT_REXMT, 0);
else if (!tcp_timer_active(tp, TT_PERSIST))
tcp_timer_activate(tp, TT_REXMT,
tp->t_rxtcur);
Let the sowwakeup macro decide when to call sowakeup rather than have tcp "know" about it. A pending upcall would be missed, eg: used by NFS. Obtained from: NetBSD
1998-05-31 18:42:49 +00:00
sowwakeup(so);
In preparation of merging projects/sendfile, transform bare access to sb_cc member of struct sockbuf to a couple of inline functions: sbavail() and sbused() Right now they are equal, but once notion of "not ready socket buffer data", will be checked in, they are going to be different. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-12 09:57:15 +00:00
if (sbavail(&so->so_snd))
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Check to see if the TF_DELACK flag is set before returning from tcp_input(). This unbreaks delack handling, while still preserving correct T/TCP behavior Tested by: maxim Sponsored by: DARPA, NAI Labs
2003-02-22 21:54:57 +00:00
goto check_delack;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
} else if (th->th_ack == tp->snd_una &&
tlen <= sbspace(&so->so_rcv)) {
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
int newsize = 0; /* automatic sockbuf scaling */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
* This is a pure, in-sequence data packet with
* nothing on the reassembly queue and we have enough
* buffer space to take it.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED)
INP_INFO_RUNLOCK(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
ti_locked = TI_UNLOCKED;
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
/* Clean receiver SACK report if present */
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if ((tp->t_flags & TF_SACK_PERMIT) && tp->rcv_numsacks)
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
tcp_clean_sackreport(tp);
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_preddat);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tp->rcv_nxt += tlen;
- Consistently update snd_wl1, snd_wl2, and rcv_up in the header prediction code. Previously, 2GB worth of header predicted data could leave these variables too far out of sequence which would cause problems after receiving a packet that did not match the header prediction. Submitted by: Bill Baumann <bbaumann@isilon.com> Sponsored by: Isilon Systems, Inc. Reviewed by: hsu, pete@isilon.com, neal@isilon.com, aaronp@isilon.com
2002-10-31 23:24:13 +00:00
/*
* Pull snd_wl1 up to prevent seq wrap relative to
* th_seq.
*/
Back out r242262. The simplified window change/update logic wasn't complete and ready for production use. PR: kern/173309
2012-11-05 09:13:06 +00:00
tp->snd_wl1 = th->th_seq;
- Consistently update snd_wl1, snd_wl2, and rcv_up in the header prediction code. Previously, 2GB worth of header predicted data could leave these variables too far out of sequence which would cause problems after receiving a packet that did not match the header prediction. Submitted by: Bill Baumann <bbaumann@isilon.com> Sponsored by: Isilon Systems, Inc. Reviewed by: hsu, pete@isilon.com, neal@isilon.com, aaronp@isilon.com
2002-10-31 23:24:13 +00:00
/*
* Pull rcv_up up to prevent seq wrap relative to
* rcv_nxt.
*/
tp->rcv_up = tp->rcv_nxt;
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
TCPSTAT_ADD(tcps_rcvpack, nsegs);
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_ADD(tcps_rcvbyte, tlen);
A number of patches in the last years have created new return paths in tcp_input that leave the function before hitting the tcp_trace function call for the TCPDEBUG option. This has made TCPDEBUG mostly useless (and tools like ports/benchmarks/dbs not working). Add tcp_trace calls to the return paths that could be identified in this maze. This is a NOP unless you compile with TCPDEBUG.
2003-08-13 08:46:54 +00:00
#ifdef TCPDEBUG
if (so->so_options & SO_DEBUG)
tcp_trace(TA_INPUT, ostate, tp,
(void *)tcp_saveipgen, &tcp_savetcp, 0);
#endif
dd DTrace probe points, translators and a corresponding script to provide the TCPDEBUG functionality with pure DTrace. Reviewed by: rwatson MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: D3530
2015-09-13 15:50:55 +00:00
TCP_PROBE3(debug__input, tp, th, mtod(m, const char *));
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
/*
* Automatic sizing of receive socket buffer. Often the send
* buffer size is not optimally adjusted to the actual network
* conditions at hand (delay bandwidth product). Setting the
* buffer size too small limits throughput on links with high
* bandwidth and high delay (eg. trans-continental/oceanic links).
*
* On the receive side the socket buffer memory is only rarely
* used to any significant extent. This allows us to be much
* more aggressive in scaling the receive socket buffer. For
* the case that the buffer space is actually used to a large
* extent and we run out of kernel memory we can simply drop
* the new segments; TCP on the sender will just retransmit it
* later. Setting the buffer size too big may only consume too
* much kernel memory if the application doesn't read() from
* the socket or packet loss or reordering makes use of the
* reassembly queue.
*
* The criteria to step up the receive buffer one notch are:
Improve comments by listing a criteria for automatic increment of receive socket buffer. Reviewed by: jmg
2014-08-09 21:01:24 +00:00
* 1. Application has not set receive buffer size with
* SO_RCVBUF. Setting SO_RCVBUF clears SB_AUTOSIZE.
* 2. the number of bytes received during the time it takes
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
* one timestamp to be reflected back to us (the RTT);
Improve comments by listing a criteria for automatic increment of receive socket buffer. Reviewed by: jmg
2014-08-09 21:01:24 +00:00
* 3. received bytes per RTT is within seven eighth of the
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
* current socket buffer size;
Improve comments by listing a criteria for automatic increment of receive socket buffer. Reviewed by: jmg
2014-08-09 21:01:24 +00:00
* 4. receive buffer size has not hit maximal automatic size;
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
*
* This algorithm does one step per RTT at most and only if
* we receive a bulk stream w/o packet losses or reorderings.
* Shrinking the buffer during idle times is not necessary as
* it doesn't consume any memory when idle.
*
* TODO: Only step up if the application is actually serving
* the buffer to better manage the socket buffer resources.
*/
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if (V_tcp_do_autorcvbuf &&
Check TCP timestamp option flag so that the automatic receive buffer scaling code does not use an uninitialized timestamp echo reply value from the stack when timestamps are not enabled. Differential Revision: https://reviews.freebsd.org/D3060 Reviewed by: hiren Approved by: jmallett (mentor) MFC after: 3 days Sponsored by: Norse Corp, Inc.
2015-07-17 17:36:33 +00:00
(to.to_flags & TOF_TS) &&
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
to.to_tsecr &&
(so->so_rcv.sb_flags & SB_AUTOSIZE)) {
Use timestamp modulo comparison macro for automatic receive buffer scaling to correctly handle wrapping of ticks value. MFC after: 1 week
2010-08-27 12:34:53 +00:00
if (TSTMP_GT(to.to_tsecr, tp->rfbuf_ts) &&
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
to.to_tsecr - tp->rfbuf_ts < hz) {
if (tp->rfbuf_cnt >
(so->so_rcv.sb_hiwat / 8 * 7) &&
so->so_rcv.sb_hiwat <
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
V_tcp_autorcvbuf_max) {
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
newsize =
min(so->so_rcv.sb_hiwat +
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
V_tcp_autorcvbuf_inc,
V_tcp_autorcvbuf_max);
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
}
/* Start over with next RTT. */
tp->rfbuf_ts = 0;
tp->rfbuf_cnt = 0;
} else
tp->rfbuf_cnt += tlen; /* add up */
}
/* Add data to socket buffer. */
Reduce the number of unnecessary unlock-relocks on socket buffer mutexes associated with performing a wakeup on the socket buffer: - When performing an sbappend*() followed by a so[rw]wakeup(), explicitly acquire the socket buffer lock and use the _locked() variants of both calls. Note that the _locked() sowakeup() versions unlock the mutex on return. This is done in uipc_send(), divert_packet(), mroute socket_send(), raw_append(), tcp_reass(), tcp_input(), and udp_append(). - When the socket buffer lock is dropped before a sowakeup(), remove the explicit unlock and use the _locked() sowakeup() variant. This is done in soisdisconnecting(), soisdisconnected() when setting the can't send/ receive flags and dropping data, and in uipc_rcvd() which adjusting back-pressure on the sockets. For UNIX domain sockets running mpsafe with a contention-intensive SMP mysql benchmark, this results in a 1.6% query rate improvement due to reduce mutex costs.
2004-06-26 19:10:39 +00:00
SOCKBUF_LOCK(&so->so_rcv);
The socket field so_state is used to hold a variety of socket related flags relating to several aspects of socket functionality. This change breaks out several bits relating to send and receive operation into a new per-socket buffer field, sb_state, in order to facilitate locking. This is required because, in order to provide more granular locking of sockets, different state fields have different locking properties. The following fields are moved to sb_state: SS_CANTRCVMORE (so_state) SS_CANTSENDMORE (so_state) SS_RCVATMARK (so_state) Rename respectively to: SBS_CANTRCVMORE (so_rcv.sb_state) SBS_CANTSENDMORE (so_snd.sb_state) SBS_RCVATMARK (so_rcv.sb_state) This facilitates locking by isolating fields to be located with other identically locked fields, and permits greater granularity in socket locking by avoiding storing fields with different locking semantics in the same short (avoiding locking conflicts). In the future, we may wish to coallesce sb_state and sb_flags; for the time being I leave them separate and there is no additional memory overhead due to the packing/alignment of shorts in the socket buffer structure.
2004-06-14 18:16:22 +00:00
if (so->so_rcv.sb_state & SBS_CANTRCVMORE) {
Fix issue where shutdown(socket, SHUT_RD) was effectively ignored for TCP sockets. NetBSD PR: 18185 Submitted by: Sean Boudreau <seanb@qnx.com> MFC after: 3 days
2002-09-22 02:54:07 +00:00
m_freem(m);
} else {
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
/*
* Set new socket buffer size.
* Give up when limit is reached.
*/
if (newsize)
if (!sbreserve_locked(&so->so_rcv,
Don't pass curthread to sbreserve_locked() in tcp_do_segment(), as the netisr or ithread's socket buffer size limit is not the right limit to use. Instead, pass NULL as the other two calls to sbreserve_locked() in the TCP input path (tcp_mss()) do. In practice, this is a no-op, as ithreads and the netisr run without a process limit on socket buffer use, and a NULL thread pointer leads to not using the process's limit, if any. However, if tcp_input() is called in other contexts that do have limits, this may prevent the incorrect limit from being used. MFC after: 3 days
2008-10-07 09:41:07 +00:00
newsize, so, NULL))
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
so->so_rcv.sb_flags &= ~SB_AUTOSIZE;
Fix issue where shutdown(socket, SHUT_RD) was effectively ignored for TCP sockets. NetBSD PR: 18185 Submitted by: Sean Boudreau <seanb@qnx.com> MFC after: 3 days
2002-09-22 02:54:07 +00:00
m_adj(m, drop_hdrlen); /* delayed header drop */
Merge from projects/sendfile: extend protocols API to support sending not ready data: o Add new flag to pru_send() flags - PRUS_NOTREADY. o Add new protocol method pru_ready(). Sponsored by: Nginx, Inc. Sponsored by: Netflix
2014-11-30 13:24:21 +00:00
sbappendstream_locked(&so->so_rcv, m, 0);
Fix issue where shutdown(socket, SHUT_RD) was effectively ignored for TCP sockets. NetBSD PR: 18185 Submitted by: Sean Boudreau <seanb@qnx.com> MFC after: 3 days
2002-09-22 02:54:07 +00:00
}
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
/* NB: sorwakeup_locked() does an implicit unlock. */
Reduce the number of unnecessary unlock-relocks on socket buffer mutexes associated with performing a wakeup on the socket buffer: - When performing an sbappend*() followed by a so[rw]wakeup(), explicitly acquire the socket buffer lock and use the _locked() variants of both calls. Note that the _locked() sowakeup() versions unlock the mutex on return. This is done in uipc_send(), divert_packet(), mroute socket_send(), raw_append(), tcp_reass(), tcp_input(), and udp_append(). - When the socket buffer lock is dropped before a sowakeup(), remove the explicit unlock and use the _locked() sowakeup() variant. This is done in soisdisconnecting(), soisdisconnected() when setting the can't send/ receive flags and dropping data, and in uipc_rcvd() which adjusting back-pressure on the sockets. For UNIX domain sockets running mpsafe with a contention-intensive SMP mysql benchmark, this results in a 1.6% query rate improvement due to reduce mutex costs.
2004-06-26 19:10:39 +00:00
sorwakeup_locked(so);
The TCP delayed ACK logic isn't aware of LRO passing up large aggregated segments thinking it received only one segment. This causes it to enable the delay the ACK for 100ms to wait for another segment which may never come because all the data was received already. Doing delayed ACK for LRO segments is bogus for two reasons: a) it pushes us further away from acking every other packet; b) it introduces additional delay in responding to the sender. The latter is especially bad because it is in the nature of LRO to aggregated all segments of a burst with no more coming until an ACK is sent back. Change the delayed ACK logic to detect LRO segments by being larger than the MSS for this connection and issuing an immediate ACK for them to keep the ACK clock ticking without interruption. Reported by: julian, cperciva Tested by: cperciva Reviewed by: lstewart MFC after: 3 days
2013-10-22 18:24:34 +00:00
if (DELAY_ACK(tp, tlen)) {
Clean up delayed acks and T/TCP interactions: - delay acks for T/TCP regardless of delack setting - fix bug where a single pass through tcp_input might not delay acks - use callout_active() instead of callout_pending() Sponsored by: DARPA, NAI Labs
2003-02-19 21:18:23 +00:00
tp->t_flags |= TF_DELACK;
Changes to support the addition of a new sysctl variable: net.inet.tcp.delack_enabled Which defaults to 1 and can be set to 0 to disable TCP delayed-ack processing (i.e. all acks are immediate).
1998-02-26 05:25:39 +00:00
} else {
Re-apply my "breakage" to the Nagel congestion avoidence. This version differs slightly in the logic from the previous version; packets are now acked immediately if the sender set PUSH.
1995-03-27 07:12:24 +00:00
tp->t_flags |= TF_ACKNOW;
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
tp->t_fb->tfb_tcp_output(tp);
Re-apply my "breakage" to the Nagel congestion avoidence. This version differs slightly in the logic from the previous version; packets are now acked immediately if the sender set PUSH.
1995-03-27 07:12:24 +00:00
}
Check to see if the TF_DELACK flag is set before returning from tcp_input(). This unbreaks delack handling, while still preserving correct T/TCP behavior Tested by: maxim Sponsored by: DARPA, NAI Labs
2003-02-22 21:54:57 +00:00
goto check_delack;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
/*
* Calculate amount of space in receive window,
* and then do TCP input processing.
* Receive window is amount of space in rcv queue,
* but not less than advertised window.
*/
win = sbspace(&so->so_rcv);
if (win < 0)
win = 0;
Fix a bug (apparently very old) that can cause a TCP connection to be dropped when it has an unusual traffic pattern. For full details as well as a test case that demonstrates the failure, see the referenced PR. Under certain circumstances involving the persist state, it is possible for the receive side's tp->rcv_nxt to advance beyond its tp->rcv_adv. This causes (tp->rcv_adv - tp->rcv_nxt) to become negative. However, in the code affected by this fix, that difference was interpreted as an unsigned number by max(). Since it was negative, it was taken as a huge unsigned number. The effect was to cause the receiver to believe that its receive window had negative size, thereby rejecting all received segments including ACKs. As the test case shows, this led to fruitless retransmissions and eventually to a dropped connection. Even connections using the loopback interface could be dropped. The fix substitutes the signed imax() for the unsigned max() function. PR: closes kern/3998 Reviewed by: davidg, fenner, wollman
1997-07-01 05:42:16 +00:00
tp->rcv_wnd = imax(win, (int)(tp->rcv_adv - tp->rcv_nxt));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Auto sizing TCP socket buffers. Normally the socket buffers are static (either derived from global defaults or set with setsockopt) and do not adapt to real network conditions. Two things happen: a) your socket buffers are too small and you can't reach the full potential of the network between both hosts; b) your socket buffers are too big and you waste a lot of kernel memory for data just sitting around. With automatic TCP send and receive socket buffers we can start with a small buffer and quickly grow it in parallel with the TCP congestion window to match real network conditions. FreeBSD has a default 32K send socket buffer. This supports a maximal transfer rate of only slightly more than 2Mbit/s on a 100ms RTT trans-continental link. Or at 200ms just above 1Mbit/s. With TCP send buffer auto scaling and the default values below it supports 20Mbit/s at 100ms and 10Mbit/s at 200ms. That's an improvement of factor 10, or 1000%. For the receive side it looks slightly better with a default of 64K buffer size. New sysctls are: net.inet.tcp.sendbuf_auto=1 (enabled) net.inet.tcp.sendbuf_inc=8192 (8K, step size) net.inet.tcp.sendbuf_max=262144 (256K, growth limit) net.inet.tcp.recvbuf_auto=1 (enabled) net.inet.tcp.recvbuf_inc=16384 (16K, step size) net.inet.tcp.recvbuf_max=262144 (256K, growth limit) Tested by: many (on HEAD and RELENG_6) Approved by: re MFC after: 1 month
2007-02-01 18:32:13 +00:00
/* Reset receive buffer auto scaling when not in bulk receive mode. */
tp->rfbuf_ts = 0;
tp->rfbuf_cnt = 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
switch (tp->t_state) {
A more complete fix for the "land" attack, removing the "quick fix" from rev 1.66. This fix contains both belt and suspenders. Belt: ignore packets where src == dst and srcport == dstport in TCPS_LISTEN. These packets can only legitimately occur when connecting a socket to itself, which doesn't go through TCPS_LISTEN (it goes CLOSED->SYN_SENT->SYN_RCVD-> ESTABLISHED). This prevents the "standard" "land" attack, although doesn't prevent the multi-homed variation. Suspenders: send a RST in response to a SYN/ACK in SYN_RECEIVED state. The only packets we should get in SYN_RECEIVED are 1. A retransmitted SYN, or 2. An ack of our SYN/ACK. The "land" attack depends on us accepting our own SYN/ACK as an ACK; in SYN_RECEIVED state; this should prevent all "land" attacks. We also move up the sequence number check for the ACK in SYN_RECEIVED. This neither helps nor hurts with respect to the "land" attack, but puts more of the validation checking in one spot. PR: kern/5103
1998-01-21 02:05:59 +00:00
/*
* If the state is SYN_RECEIVED:
* if seg contains an ACK, but not for our SYN/ACK, send a RST.
*/
case TCPS_SYN_RECEIVED:
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if ((thflags & TH_ACK) &&
(SEQ_LEQ(th->th_ack, tp->snd_una) ||
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
SEQ_GT(th->th_ack, tp->snd_max))) {
rstreason = BANDLIM_RST_OPENPORT;
goto dropwithreset;
}
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
#ifdef TCP_RFC7413
if (tp->t_flags & TF_FASTOPEN) {
/*
* When a TFO connection is in SYN_RECEIVED, the
* only valid packets are the initial SYN, a
* retransmit/copy of the initial SYN (possibly with
* a subset of the original data), a valid ACK, a
* FIN, or a RST.
*/
if ((thflags & (TH_SYN|TH_ACK)) == (TH_SYN|TH_ACK)) {
rstreason = BANDLIM_RST_OPENPORT;
goto dropwithreset;
} else if (thflags & TH_SYN) {
/* non-initial SYN is ignored */
if ((tcp_timer_active(tp, TT_DELACK) ||
tcp_timer_active(tp, TT_REXMT)))
goto drop;
} else if (!(thflags & (TH_ACK|TH_FIN|TH_RST))) {
goto drop;
}
}
#endif
A more complete fix for the "land" attack, removing the "quick fix" from rev 1.66. This fix contains both belt and suspenders. Belt: ignore packets where src == dst and srcport == dstport in TCPS_LISTEN. These packets can only legitimately occur when connecting a socket to itself, which doesn't go through TCPS_LISTEN (it goes CLOSED->SYN_SENT->SYN_RCVD-> ESTABLISHED). This prevents the "standard" "land" attack, although doesn't prevent the multi-homed variation. Suspenders: send a RST in response to a SYN/ACK in SYN_RECEIVED state. The only packets we should get in SYN_RECEIVED are 1. A retransmitted SYN, or 2. An ack of our SYN/ACK. The "land" attack depends on us accepting our own SYN/ACK as an ACK; in SYN_RECEIVED state; this should prevent all "land" attacks. We also move up the sequence number check for the ACK in SYN_RECEIVED. This neither helps nor hurts with respect to the "land" attack, but puts more of the validation checking in one spot. PR: kern/5103
1998-01-21 02:05:59 +00:00
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* If the state is SYN_SENT:
* if seg contains an ACK, but not for our SYN, drop the input.
* if seg contains a RST, then drop the connection.
* if seg does not contain SYN, then drop it.
* Otherwise this is an acceptable SYN segment
* initialize tp->rcv_nxt and tp->irs
* if seg contains ack then advance tp->snd_una
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
* if seg contains an ECE and ECN support is enabled, the stream
* is ECN capable.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* if SYN has been acked change to ESTABLISHED else SYN_RCVD state
* arrange for segment to be acked (eventually)
* continue processing rest of data/controls, beginning with URG
*/
case TCPS_SYN_SENT:
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if ((thflags & TH_ACK) &&
(SEQ_LEQ(th->th_ack, tp->iss) ||
SEQ_GT(th->th_ack, tp->snd_max))) {
Remove RFC1644 T/TCP support from the TCP side of the network stack. A complete rationale and discussion is given in this message and the resulting discussion: http://docs.freebsd.org/cgi/mid.cgi?4177C8AD.6060706 Note that this commit removes only the functional part of T/TCP from the tcp_* related functions in the kernel. Other features introduced with RFC1644 are left intact (socket layer changes, sendmsg(2) on connection oriented protocols) and are meant to be reused by a simpler and less intrusive reimplemention of the previous T/TCP functionality. Discussed on: -arch
2004-11-02 22:22:22 +00:00
rstreason = BANDLIM_UNLIMITED;
goto dropwithreset;
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
}
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
if ((thflags & (TH_ACK|TH_RST)) == (TH_ACK|TH_RST)) {
dtrace sdt: remove the ugly sname parameter of SDT_PROBE_DEFINE In its stead use the Solaris / illumos approach of emulating '-' (dash) in probe names with '__' (two consecutive underscores). Reviewed by: markj MFC after: 3 weeks
2013-11-26 08:46:27 +00:00
TCP_PROBE5(connect__refused, NULL, tp,
Convert over the TCP probes to use mtod() rather than directly dereferencing m->m_data. Sponsored by: Netflix, Inc.
2013-11-25 22:55:06 +00:00
mtod(m, const char *), tp, th);
o Remove redundant tcp reassembly check in header prediction code o Rearrange code to make intent in TCPS_SYN_SENT case more clear o Assorted style cleanup o Comment clarification for tcp_dropwithreset()
2007-05-06 15:41:06 +00:00
tp = tcp_drop(tp, ECONNREFUSED);
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
}
o Remove redundant tcp reassembly check in header prediction code o Rearrange code to make intent in TCPS_SYN_SENT case more clear o Assorted style cleanup o Comment clarification for tcp_dropwithreset()
2007-05-06 15:41:06 +00:00
if (thflags & TH_RST)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto drop;
o Remove redundant tcp reassembly check in header prediction code o Rearrange code to make intent in TCPS_SYN_SENT case more clear o Assorted style cleanup o Comment clarification for tcp_dropwithreset()
2007-05-06 15:41:06 +00:00
if (!(thflags & TH_SYN))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto drop;
Rework TCP window scaling (RFC1323) to properly scale the send window right from the beginning and partly clean up the differences in handling between SYN_SENT and SYN_RCVD (syncache). Further changes to this code to come. This is a first incremental step to a general overhaul and streamlining of the TCP code. PR: kern/15095 PR: kern/92690 (partly) Reviewed by: qingli (and tested with ANVL) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-28 23:05:59 +00:00
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tp->irs = th->th_seq;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tcp_rcvseqinit(tp);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (thflags & TH_ACK) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_connects);
Cosmetic changes to processing of segments in the SYN_SENT state: - remove a redundant condition; - complete all validity checks on segment before calling soisconnected(so). Reviewed by: Richard Stevens, davidg, wollman
1995-11-03 22:31:54 +00:00
soisconnected(so);
Introduce support for Mandatory Access Control and extensible kernel access control. Instrument the TCP socket code for packet generation and delivery: label outgoing mbufs with the label of the socket, and check socket and mbuf labels before permitting delivery to a socket. Assign labels to newly accepted connections when the syncache/cookie code has done its business. Also set peer labels as convenient. Currently, MAC policies cannot influence the PCB matching algorithm, so cannot implement polyinstantiation. Note that there is at least one case where a PCB is not available due to the TCP packet not being associated with any socket, so we don't label in that case, but need to handle it in a special manner. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-31 19:06:49 +00:00
#ifdef MAC
Merge first in a series of TrustedBSD MAC Framework KPI changes from Mac OS X Leopard--rationalize naming for entry points to the following general forms: mac_<object>_<method/action> mac_<object>_check_<method/action> The previous naming scheme was inconsistent and mostly reversed from the new scheme. Also, make object types more consistent and remove spaces from object types that contain multiple parts ("posix_sem" -> "posixsem") to make mechanical parsing easier. Introduce a new "netinet" object type for certain IPv4/IPv6-related methods. Also simplify, slightly, some entry point names. All MAC policy modules will need to be recompiled, and modules not updates as part of this commit will need to be modified to conform to the new KPI. Sponsored by: SPARTA (original patches against Mac OS X) Obtained from: TrustedBSD Project, Apple Computer
2007-10-24 19:04:04 +00:00
mac_socketpeer_set_from_mbuf(m, so);
Introduce support for Mandatory Access Control and extensible kernel access control. Instrument the TCP socket code for packet generation and delivery: label outgoing mbufs with the label of the socket, and check socket and mbuf labels before permitting delivery to a socket. Assign labels to newly accepted connections when the syncache/cookie code has done its business. Also set peer labels as convenient. Currently, MAC policies cannot influence the PCB matching algorithm, so cannot implement polyinstantiation. Note that there is at least one case where a PCB is not available due to the TCP packet not being associated with any socket, so we don't label in that case, but need to handle it in a special manner. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-31 19:06:49 +00:00
#endif
Cosmetic changes to processing of segments in the SYN_SENT state: - remove a redundant condition; - complete all validity checks on segment before calling soisconnected(so). Reviewed by: Richard Stevens, davidg, wollman
1995-11-03 22:31:54 +00:00
/* Do window scaling on this connection? */
if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) ==
(TF_RCVD_SCALE|TF_REQ_SCALE)) {
tp->rcv_scale = tp->request_r_scale;
}
Clamp the initial advertised receive window when responding to a SYN/ACK to the maximum allowed window. Growing the window too large would cause an underflow in the calculations in tcp_output() to decide if a window update should be sent which would prevent the persist timer from being started if data was pending and the other end of the connection advertised an initial window size of 0. PR: kern/154006 Submitted by: Stefan `Sec` Zehl sec 42 org Reviewed by: bz MFC after: 1 week
2011-03-30 12:35:39 +00:00
tp->rcv_adv += imin(tp->rcv_wnd,
TCP_MAXWIN << tp->rcv_scale);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tp->snd_una++; /* SYN is acked */
/*
* If there's data, delay ACK; if there's also a FIN
* ACKNOW will be turned on later.
*/
The TCP delayed ACK logic isn't aware of LRO passing up large aggregated segments thinking it received only one segment. This causes it to enable the delay the ACK for 100ms to wait for another segment which may never come because all the data was received already. Doing delayed ACK for LRO segments is bogus for two reasons: a) it pushes us further away from acking every other packet; b) it introduces additional delay in responding to the sender. The latter is especially bad because it is in the nature of LRO to aggregated all segments of a burst with no more coming until an ACK is sent back. Change the delayed ACK logic to detect LRO segments by being larger than the MSS for this connection and issuing an immediate ACK for them to keep the ACK clock ticking without interruption. Reported by: julian, cperciva Tested by: cperciva Reviewed by: lstewart MFC after: 3 days
2013-10-22 18:24:34 +00:00
if (DELAY_ACK(tp, tlen) && tlen != 0)
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
tcp_timer_activate(tp, TT_DELACK,
tcp_delacktime);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
else
tp->t_flags |= TF_ACKNOW;
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if ((thflags & TH_ECE) && V_tcp_do_ecn) {
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
tp->t_flags |= TF_ECN_PERMIT;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_ecn_shs);
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
* Received <SYN,ACK> in SYN_SENT[*] state.
* Transitions:
* SYN_SENT --> ESTABLISHED
* SYN_SENT* --> FIN_WAIT_1
*/
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
tp->t_starttime = ticks;
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
if (tp->t_flags & TF_NEEDFIN) {
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_FIN_WAIT_1);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tp->t_flags &= ~TF_NEEDFIN;
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
thflags &= ~TH_SYN;
Receipt of two SYN's are sufficient to set the t_timer[TCPT_KEEP] to "keepidle". this should not occur unless the connection has been established via the 3-way handshake which requires an ACK Submitted by: jmb Obtained from: problem discussed in Stevens vol. 3
1996-09-13 18:47:03 +00:00
} else {
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_ESTABLISHED);
dtrace sdt: remove the ugly sname parameter of SDT_PROBE_DEFINE In its stead use the Solaris / illumos approach of emulating '-' (dash) in probe names with '__' (two consecutive underscores). Reviewed by: markj MFC after: 3 weeks
2013-11-26 08:46:27 +00:00
TCP_PROBE5(connect__established, NULL, tp,
Convert over the TCP probes to use mtod() rather than directly dereferencing m->m_data. Sponsored by: Netflix, Inc.
2013-11-25 22:55:06 +00:00
mtod(m, const char *), tp, th);
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
cc_conn_init(tp);
Add new socket options: TCP_KEEPINIT, TCP_KEEPIDLE, TCP_KEEPINTVL and TCP_KEEPCNT, that allow to control initial timeout, idle time, idle re-send interval and idle send count on a per-socket basis. Reviewed by: andre, bz, lstewart
2012-02-05 16:53:02 +00:00
tcp_timer_activate(tp, TT_KEEP,
TP_KEEPIDLE(tp));
Receipt of two SYN's are sufficient to set the t_timer[TCPT_KEEP] to "keepidle". this should not occur unless the connection has been established via the 3-way handshake which requires an ACK Submitted by: jmb Obtained from: problem discussed in Stevens vol. 3
1996-09-13 18:47:03 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
} else {
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
/*
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
* Received initial SYN in SYN-SENT[*] state =>
Correct the comments as support for RFC 1644 has been removed for a long time.
2014-03-25 21:57:50 +00:00
* simultaneous open.
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
* If it succeeds, connection is * half-synchronized.
* Otherwise, do 3-way handshake:
* SYN-SENT -> SYN-RECEIVED
* SYN-SENT* -> SYN-RECEIVED*
*/
This patch fixes the problem where the current TCP code can not handle simultaneous open. Both the bug and the patch were verified using the ANVL test suite. PR: kern/74935 Submitted by: qingli (before I became committer) Reviewed by: andre MFC after: 5 days
2006-02-23 21:14:34 +00:00
tp->t_flags |= (TF_ACKNOW | TF_NEEDSYN);
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
tcp_timer_activate(tp, TT_REXMT, 0);
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_SYN_RECEIVED);
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
KASSERT(ti_locked == TI_RLOCKED, ("%s: trimthenstep6: "
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
"ti_locked %d", __func__, ti_locked));
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
After each label in tcp_input(), assert the inpcbinfo and inpcb lock state that we expect.
2004-07-12 19:28:07 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
* Advance th->th_seq to correspond to first data byte.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* If data, trim to stay within window,
* dropping FIN if necessary.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
th->th_seq++;
if (tlen > tp->rcv_wnd) {
todrop = tlen - tp->rcv_wnd;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
m_adj(m, -todrop);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tlen = tp->rcv_wnd;
thflags &= ~TH_FIN;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvpackafterwin);
TCPSTAT_ADD(tcps_rcvbyteafterwin, todrop);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tp->snd_wl1 = th->th_seq - 1;
tp->rcv_up = th->th_seq;
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
* Client side of transaction: already sent SYN and data.
* If the remote host used T/TCP to validate the SYN,
* our data will be ACK'd; if so, enter normal data segment
* processing in the middle of step 5, ack processing.
* Otherwise, goto step 6.
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
*/
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
if (thflags & TH_ACK)
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
goto process_ACK;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto step6;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
* If the state is LAST_ACK or CLOSING or TIME_WAIT:
Remove RFC1644 T/TCP support from the TCP side of the network stack. A complete rationale and discussion is given in this message and the resulting discussion: http://docs.freebsd.org/cgi/mid.cgi?4177C8AD.6060706 Note that this commit removes only the functional part of T/TCP from the tcp_* related functions in the kernel. Other features introduced with RFC1644 are left intact (socket layer changes, sendmsg(2) on connection oriented protocols) and are meant to be reused by a simpler and less intrusive reimplemention of the previous T/TCP functionality. Discussed on: -arch
2004-11-02 22:22:22 +00:00
* do normal processing.
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
*
Remove RFC1644 T/TCP support from the TCP side of the network stack. A complete rationale and discussion is given in this message and the resulting discussion: http://docs.freebsd.org/cgi/mid.cgi?4177C8AD.6060706 Note that this commit removes only the functional part of T/TCP from the tcp_* related functions in the kernel. Other features introduced with RFC1644 are left intact (socket layer changes, sendmsg(2) on connection oriented protocols) and are meant to be reused by a simpler and less intrusive reimplemention of the previous T/TCP functionality. Discussed on: -arch
2004-11-02 22:22:22 +00:00
* NB: Leftover from RFC1644 T/TCP. Cases to be reused later.
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
*/
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
case TCPS_LAST_ACK:
case TCPS_CLOSING:
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
break; /* continue normal processing */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
/*
* States other than LISTEN or SYN_SENT.
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
* First check the RST flag and sequence number since reset segments
* are exempt from the timestamp and connection count tests. This
* fixes a bug introduced by the Stevens, vol. 2, p. 960 bugfix
* below which allowed reset segments in half the sequence space
* to fall though and be processed (which gives forged reset
* segments with a random sequence number a 50 percent chance of
* killing a connection).
* Then check timestamp, if present.
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
* Then check the connection count, if present.
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
* Then check that at least some bytes of segment are within
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* receive window. If segment begins before rcv_nxt,
* drop leading data (and SYN); if nothing left, just ack.
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (thflags & TH_RST) {
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
/*
* RFC5961 Section 3.2
*
* - RST drops connection only if SEG.SEQ == RCV.NXT.
* - If RST is in window, we send challenge ACK.
*
* Note: to take into account delayed ACKs, we should
* test against last_ack_sent instead of rcv_nxt.
* Note 2: we handle special case of closed window, not
* covered by the RFC.
*/
if ((SEQ_GEQ(th->th_seq, tp->last_ack_sent) &&
SEQ_LT(th->th_seq, tp->last_ack_sent + tp->rcv_wnd)) ||
(tp->rcv_wnd == 0 && tp->last_ack_sent == th->th_seq)) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
KASSERT(ti_locked == TI_RLOCKED,
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
("%s: TH_RST ti_locked %d, th %p tp %p",
__func__, ti_locked, th, tp));
KASSERT(tp->t_state != TCPS_SYN_SENT,
("%s: TH_RST for TCPS_SYN_SENT th %p tp %p",
__func__, th, tp));
if (V_tcp_insecure_rst ||
tp->last_ack_sent == th->th_seq) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_drops);
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
/* Drop the connection. */
switch (tp->t_state) {
case TCPS_SYN_RECEIVED:
so->so_error = ECONNREFUSED;
goto close;
case TCPS_ESTABLISHED:
case TCPS_FIN_WAIT_1:
case TCPS_FIN_WAIT_2:
case TCPS_CLOSE_WAIT:
so->so_error = ECONNRESET;
close:
tcp_state_change(tp, TCPS_CLOSED);
/* FALLTHROUGH */
default:
tp = tcp_close(tp);
}
} else {
TCPSTAT_INC(tcps_badrst);
/* Send challenge ACK. */
tcp_respond(tp, mtod(m, void *), th, m,
tp->rcv_nxt, tp->snd_nxt, TH_ACK);
tp->last_ack_sent = tp->rcv_nxt;
m = NULL;
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
}
}
goto drop;
}
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
/*
* RFC5961 Section 4.2
* Send challenge ACK for any SYN in synchronized state.
*/
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
if ((thflags & TH_SYN) && tp->t_state != TCPS_SYN_SENT &&
tp->t_state != TCPS_SYN_RECEIVED) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
KASSERT(ti_locked == TI_RLOCKED,
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
("tcp_do_segment: TH_SYN ti_locked %d", ti_locked));
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
FreeBSD-SA-14:19.tcp raised attention to the state of our stack towards blind SYN/RST spoofed attack. Originally our stack used in-window checks for incoming SYN/RST as proposed by RFC793. Later, circa 2003 the RST attack was mitigated using the technique described in P. Watson "Slipping in the window" paper [1]. After that, the checks were only relaxed for the sake of compatibility with some buggy TCP stacks. First, r192912 introduced the vulnerability, just fixed by aforementioned SA. Second, r167310 had slightly relaxed the default RST checks, instead of utilizing net.inet.tcp.insecure_rst sysctl. In 2010 a new technique for mitigation of these attacks was proposed in RFC5961 [2]. The idea is to send a "challenge ACK" packet to the peer, to verify that packet arrived isn't spoofed. If peer receives challenge ACK it should regenerate its RST or SYN with correct sequence number. This should not only protect against attacks, but also improve communication with broken stacks, so authors of reverted r167310 and r192912 won't be disappointed. [1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf [2] http://www.rfc-editor.org/rfc/rfc5961.txt Changes made: o Revert r167310. o Implement "challenge ACK" protection as specificed in RFC5961 against RST attack. On by default. - Carefully preserve r138098, which handles empty window edge case, not described by the RFC. - Update net.inet.tcp.insecure_rst description. o Implement "challenge ACK" protection as specificed in RFC5961 against SYN attack. On by default. - Provide net.inet.tcp.insecure_syn sysctl, to turn off RFC5961 protection. The changes were tested at Netflix. The tested box didn't show any anomalies compared to control box, except slightly increased number of TCP connection in LAST_ACK state. Reviewed by: rrs Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-09-16 11:07:25 +00:00
TCPSTAT_INC(tcps_badsyn);
if (V_tcp_insecure_syn &&
SEQ_GEQ(th->th_seq, tp->last_ack_sent) &&
SEQ_LT(th->th_seq, tp->last_ack_sent + tp->rcv_wnd)) {
tp = tcp_drop(tp, ECONNRESET);
rstreason = BANDLIM_UNLIMITED;
} else {
/* Send challenge ACK. */
tcp_respond(tp, mtod(m, void *), th, m, tp->rcv_nxt,
tp->snd_nxt, TH_ACK);
tp->last_ack_sent = tp->rcv_nxt;
m = NULL;
}
goto drop;
}
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
/*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* RFC 1323 PAWS: If we have a timestamp reply on this segment
* and it's less than ts_recent, drop it.
*/
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
if ((to.to_flags & TOF_TS) != 0 && tp->ts_recent &&
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
TSTMP_LT(to.to_tsval, tp->ts_recent)) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/* Check to see if ts_recent is over 24 days old. */
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
if (tcp_ts_getticks() - tp->ts_recent_age > TCP_PAWS_IDLE) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Invalidate ts_recent. If this segment updates
* ts_recent, the age will be reset later and ts_recent
* will get a valid value. If it does not, setting
* ts_recent to zero will at least satisfy the
* requirement that zero be placed in the timestamp
* echo reply when ts_recent isn't valid. The
* age isn't reset until we get a valid ts_recent
* because we don't want out-of-order segments to be
* dropped when ts_recent is old.
*/
tp->ts_recent = 0;
} else {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvduppack);
TCPSTAT_ADD(tcps_rcvdupbyte, tlen);
TCPSTAT_INC(tcps_pawsdrop);
Remove the PAWS ack-on-ack debugging printf(). Note that the original RFC 1323 (PAWS) says in 4.2.1 that the out of order / reverse-time-indexed packet should be acknowledged as specified in RFC-793 page 69 then dropped. The original PAWS code in FreeBSD (1994) simply acknowledged the segment unconditionally, which is incorrect, and was fixed in 1.183 (2002). At the moment we do not do checks for SYN or FIN in addition to (tlen != 0), which may or may not be correct, but the worst that ought to happen should be a retry by the sender.
2002-12-30 19:31:04 +00:00
if (tlen)
Fix syntax in last commit.
2002-12-17 00:24:48 +00:00
goto dropafterack;
Remove the PAWS ack-on-ack debugging printf(). Note that the original RFC 1323 (PAWS) says in 4.2.1 that the out of order / reverse-time-indexed packet should be acknowledged as specified in RFC-793 page 69 then dropped. The original PAWS code in FreeBSD (1994) simply acknowledged the segment unconditionally, which is incorrect, and was fixed in 1.183 (2002). At the moment we do not do checks for SYN or FIN in addition to (tlen != 0), which may or may not be correct, but the worst that ought to happen should be a retry by the sender.
2002-12-30 19:31:04 +00:00
goto drop;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
/*
* In the SYN-RECEIVED state, validate that the packet belongs to
* this connection before trimming the data to fit the receive
* window. Check the sequence number versus IRS since we know
* the sequence numbers haven't wrapped. This is a partial fix
* for the "LAND" DoS attack.
*/
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
if (tp->t_state == TCPS_SYN_RECEIVED && SEQ_LT(th->th_seq, tp->irs)) {
rstreason = BANDLIM_RST_OPENPORT;
goto dropwithreset;
}
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
todrop = tp->rcv_nxt - th->th_seq;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (todrop > 0) {
Fix Denial of Service in TCP packet processing. Submitted by: glebius Security: FreeBSD-SA-14:19.tcp
2014-09-16 09:48:24 +00:00
if (thflags & TH_SYN) {
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
thflags &= ~TH_SYN;
th->th_seq++;
if (th->th_urp > 1)
th->th_urp--;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
else
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
thflags &= ~TH_URG;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
todrop--;
}
Avoid deadlock situation described by Stevens using his suggested replacement code. Obtained from: Stevens, vol. 2, pp. 959-960
1995-02-16 01:39:19 +00:00
/*
* Following if statement from Stevens, vol. 2, p. 960.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (todrop > tlen
|| (todrop == tlen && (thflags & TH_FIN) == 0)) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Avoid deadlock situation described by Stevens using his suggested replacement code. Obtained from: Stevens, vol. 2, pp. 959-960
1995-02-16 01:39:19 +00:00
* Any valid FIN must be to the left of the window.
* At this point the FIN must be a duplicate or out
* of sequence; drop it.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
thflags &= ~TH_FIN;
Avoid deadlock situation described by Stevens using his suggested replacement code. Obtained from: Stevens, vol. 2, pp. 959-960
1995-02-16 01:39:19 +00:00
/*
* Send an ACK to resynchronize and drop any data.
* But keep on processing for RST or ACK.
*/
tp->t_flags |= TF_ACKNOW;
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
todrop = tlen;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvduppack);
TCPSTAT_ADD(tcps_rcvdupbyte, todrop);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
} else {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvpartduppack);
TCPSTAT_ADD(tcps_rcvpartdupbyte, todrop);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
drop_hdrlen += todrop; /* drop from the top afterwards */
th->th_seq += todrop;
tlen -= todrop;
if (th->th_urp > todrop)
th->th_urp -= todrop;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
else {
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
thflags &= ~TH_URG;
th->th_urp = 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
/*
* If new data are received on a connection after the
* user processes are gone, then RST the other end.
*/
if ((so->so_state & SS_NOFDREF) &&
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tp->t_state > TCPS_CLOSE_WAIT && tlen) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
KASSERT(ti_locked == TI_RLOCKED, ("%s: SS_NOFDEREF && "
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
"CLOSE_WAIT && tlen ti_locked %d", __func__, ti_locked));
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
Extend debug logging of TCP timestamp related specification violations. Update related comments and style.
2013-07-10 12:06:01 +00:00
if ((s = tcp_log_addrs(inc, th, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: %s: Received %d bytes of data "
"after socket was closed, "
"sending RST and removing tcpcb\n",
Improve the debugging message: TCP: [X.X.X.X]:X to [X.X.X.X]:X tcpflags 0x18<PUSH,ACK>; tcp_do_segment: FIN_WAIT_2: Received data after socket was closed, sending RST and removing tcpcb So that it also includes how many bytes of data were received. It now looks like this: TCP: [X.X.X.X]:X to [X.X.X.X]:X tcpflags 0x18<PUSH,ACK>; tcp_do_segment: FIN_WAIT_2: Received X bytes of data after socket was closed, sending RST and removing tcpcb Approved by: re (gnn)
2007-10-07 00:07:27 +00:00
s, __func__, tcpstates[tp->t_state], tlen);
Provide a sysctl to toggle reporting of TCP debug logging: sys.net.inet.tcp.log_debug = 1 It defaults to enabled for the moment and is to be turned off for the next release like other diagnostics from development branches. It is important to note that sysctl sys.net.inet.tcp.log_in_vain uses the same logging function as log_debug. Enabling of the former also causes the latter to engage, but not vice versa. Use consistent terminology in tcp log messages: "ignored" means a segment contains invalid flags/information and is dropped without changing state or issuing a reply. "rejected" means a segments contains invalid flags/information but is causing a reply (usually RST) and may cause a state change. Approved by: re (rwatson)
2007-07-28 12:20:39 +00:00
free(s, M_TCPLOG);
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp = tcp_close(tp);
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvafterclose);
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
rstreason = BANDLIM_UNLIMITED;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto dropwithreset;
}
/*
* If segment ends after window, drop trailing data
* (and PUSH and FIN); if nothing left, just ACK.
*/
o Remove redundant tcp reassembly check in header prediction code o Rearrange code to make intent in TCPS_SYN_SENT case more clear o Assorted style cleanup o Comment clarification for tcp_dropwithreset()
2007-05-06 15:41:06 +00:00
todrop = (th->th_seq + tlen) - (tp->rcv_nxt + tp->rcv_wnd);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (todrop > 0) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvpackafterwin);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (todrop >= tlen) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_ADD(tcps_rcvbyteafterwin, tlen);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* If window is closed can only take segments at
* window edge, and have to drop data and PUSH from
* incoming segments. Continue processing, but
* remember to ack. Otherwise, drop segment
* and ack.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (tp->rcv_wnd == 0 && th->th_seq == tp->rcv_nxt) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->t_flags |= TF_ACKNOW;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvwinprobe);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
} else
goto dropafterack;
} else
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_ADD(tcps_rcvbyteafterwin, todrop);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
m_adj(m, -todrop);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tlen -= todrop;
thflags &= ~(TH_PUSH|TH_FIN);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
/*
* If last ACK falls within this segment's sequence numbers,
* record its timestamp.
- Tighten up the Timestamp checks to prevent a spoofed segment from setting ts_recent to an arbitrary value, stopping further communication between the two hosts. - If the Echoed Timestamp is greater than the current time, fall back to the non RFC 1323 RTT calculation. Submitted by: Raja Mukerji (raja at moselle dot com) Reviewed by: Noritoshi Demizu, Mohan Srinivasan
2005-04-10 05:24:59 +00:00
* NOTE:
* 1) That the test incorporates suggestions from the latest
* proposal of the tcplw@cray.com list (Braden 1993/04/26).
* 2) That updating only on newer timestamps interferes with
* our earlier PAWS tests, so this check should be solely
* predicated on the sequence space of this segment.
* 3) That we modify the segment boundary check to be
* Last.ACK.Sent <= SEG.SEQ + SEG.Len
* instead of RFC1323's
* Last.ACK.Sent < SEG.SEQ + SEG.Len,
* This modified check allows us to overcome RFC1323's
* limitations as described in Stevens TCP/IP Illustrated
* Vol. 2 p.869. In such cases, we can still calculate the
* RTT correctly when RCV.NXT == Last.ACK.Sent.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
if ((to.to_flags & TOF_TS) != 0 &&
- Tighten up the Timestamp checks to prevent a spoofed segment from setting ts_recent to an arbitrary value, stopping further communication between the two hosts. - If the Echoed Timestamp is greater than the current time, fall back to the non RFC 1323 RTT calculation. Submitted by: Raja Mukerji (raja at moselle dot com) Reviewed by: Noritoshi Demizu, Mohan Srinivasan
2005-04-10 05:24:59 +00:00
SEQ_LEQ(th->th_seq, tp->last_ack_sent) &&
SEQ_LEQ(tp->last_ack_sent, th->th_seq + tlen +
((thflags & (TH_SYN|TH_FIN)) != 0))) {
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
tp->ts_recent_age = tcp_ts_getticks();
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tp->ts_recent = to.to_tsval;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
* If the ACK bit is off: if in SYN-RECEIVED state or SENDSYN
* flag is on (half-synchronized state), then queue data for
* later processing; else drop segment and return.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if ((thflags & TH_ACK) == 0) {
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
if (tp->t_state == TCPS_SYN_RECEIVED ||
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
(tp->t_flags & TF_NEEDSYN)) {
#ifdef TCP_RFC7413
if (tp->t_state == TCPS_SYN_RECEIVED &&
tp->t_flags & TF_FASTOPEN) {
tp->snd_wnd = tiwin;
cc_conn_init(tp);
}
#endif
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
goto step6;
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
} else if (tp->t_flags & TF_ACKNOW)
When we receive an out-of-window SYN for an "ESTABLISHED" connection, ACK the SYN as required by RFC793, rather than ignoring it. NetBSD have had a similar change since 1999. PR: 93236 Submitted by: Grant Edwards <grante@visi.com> MFC after: 1 month
2006-06-19 12:33:52 +00:00
goto dropafterack;
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
else
goto drop;
}
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Ack processing.
*/
switch (tp->t_state) {
/*
A more complete fix for the "land" attack, removing the "quick fix" from rev 1.66. This fix contains both belt and suspenders. Belt: ignore packets where src == dst and srcport == dstport in TCPS_LISTEN. These packets can only legitimately occur when connecting a socket to itself, which doesn't go through TCPS_LISTEN (it goes CLOSED->SYN_SENT->SYN_RCVD-> ESTABLISHED). This prevents the "standard" "land" attack, although doesn't prevent the multi-homed variation. Suspenders: send a RST in response to a SYN/ACK in SYN_RECEIVED state. The only packets we should get in SYN_RECEIVED are 1. A retransmitted SYN, or 2. An ack of our SYN/ACK. The "land" attack depends on us accepting our own SYN/ACK as an ACK; in SYN_RECEIVED state; this should prevent all "land" attacks. We also move up the sequence number check for the ACK in SYN_RECEIVED. This neither helps nor hurts with respect to the "land" attack, but puts more of the validation checking in one spot. PR: kern/5103
1998-01-21 02:05:59 +00:00
* In SYN_RECEIVED state, the ack ACKs our SYN, so enter
* ESTABLISHED state and continue processing.
* The ACK was checked above.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
case TCPS_SYN_RECEIVED:
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_connects);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
soisconnected(so);
/* Do window scaling? */
if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) ==
(TF_RCVD_SCALE|TF_REQ_SCALE)) {
tp->rcv_scale = tp->request_r_scale;
Rework TCP window scaling (RFC1323) to properly scale the send window right from the beginning and partly clean up the differences in handling between SYN_SENT and SYN_RCVD (syncache). Further changes to this code to come. This is a first incremental step to a general overhaul and streamlining of the TCP code. PR: kern/15095 PR: kern/92690 (partly) Reviewed by: qingli (and tested with ANVL) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-28 23:05:59 +00:00
tp->snd_wnd = tiwin;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
* Make transitions:
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
* SYN-RECEIVED -> ESTABLISHED
* SYN-RECEIVED* -> FIN-WAIT-1
*/
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
tp->t_starttime = ticks;
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
if (tp->t_flags & TF_NEEDFIN) {
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_FIN_WAIT_1);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tp->t_flags &= ~TF_NEEDFIN;
Receipt of two SYN's are sufficient to set the t_timer[TCPT_KEEP] to "keepidle". this should not occur unless the connection has been established via the 3-way handshake which requires an ACK Submitted by: jmb Obtained from: problem discussed in Stevens vol. 3
1996-09-13 18:47:03 +00:00
} else {
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_ESTABLISHED);
dtrace sdt: remove the ugly sname parameter of SDT_PROBE_DEFINE In its stead use the Solaris / illumos approach of emulating '-' (dash) in probe names with '__' (two consecutive underscores). Reviewed by: markj MFC after: 3 weeks
2013-11-26 08:46:27 +00:00
TCP_PROBE5(accept__established, NULL, tp,
Convert over the TCP probes to use mtod() rather than directly dereferencing m->m_data. Sponsored by: Netflix, Inc.
2013-11-25 22:55:06 +00:00
mtod(m, const char *), tp, th);
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
#ifdef TCP_RFC7413
if (tp->t_tfo_pending) {
tcp_fastopen_decrement_counter(tp->t_tfo_pending);
tp->t_tfo_pending = NULL;
/*
* Account for the ACK of our SYN prior to
* regular ACK processing below.
*/
tp->snd_una++;
}
/*
* TFO connections call cc_conn_init() during SYN
* processing. Calling it again here for such
* connections is not harmless as it would undo the
* snd_cwnd reduction that occurs when a TFO SYN|ACK
* is retransmitted.
*/
if (!(tp->t_flags & TF_FASTOPEN))
#endif
cc_conn_init(tp);
Add new socket options: TCP_KEEPINIT, TCP_KEEPIDLE, TCP_KEEPINTVL and TCP_KEEPCNT, that allow to control initial timeout, idle time, idle re-send interval and idle send count on a per-socket basis. Reviewed by: andre, bz, lstewart
2012-02-05 16:53:02 +00:00
tcp_timer_activate(tp, TT_KEEP, TP_KEEPIDLE(tp));
Receipt of two SYN's are sufficient to set the t_timer[TCPT_KEEP] to "keepidle". this should not occur unless the connection has been established via the 3-way handshake which requires an ACK Submitted by: jmb Obtained from: problem discussed in Stevens vol. 3
1996-09-13 18:47:03 +00:00
}
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
/*
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
* If segment contains data or ACK, will call tcp_reass()
* later; if not, do so now to pass queued data to user.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (tlen == 0 && (thflags & TH_FIN) == 0)
(void) tcp_reass(tp, (struct tcphdr *)0, 0,
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
(struct mbuf *)0);
Back out r242262. The simplified window change/update logic wasn't complete and ready for production use. PR: kern/173309
2012-11-05 09:13:06 +00:00
tp->snd_wl1 = th->th_seq - 1;
Replace various spelling with FALLTHROUGH which is lint()able
2002-08-25 13:23:09 +00:00
/* FALLTHROUGH */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* In ESTABLISHED state: drop duplicate ACKs; ACK out of range
* ACKs. If the ack is in the range
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
* tp->snd_una < th->th_ack <= tp->snd_max
* then advance tp->snd_una to th->th_ack and drop
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* data from the retransmission queue. If this ACK reflects
* more up to date window information we update our window information.
*/
case TCPS_ESTABLISHED:
case TCPS_FIN_WAIT_1:
case TCPS_FIN_WAIT_2:
case TCPS_CLOSE_WAIT:
case TCPS_CLOSING:
case TCPS_LAST_ACK:
- Postpone SACK option processing until after PAWS checks. SACK option processing is now done in the ACK processing case. - Merge tcp_sack_option() and tcp_del_sackholes() into a new function called tcp_sack_doack(). - Test (SEG.ACK < SND.MAX) before processing the ACK. Submitted by: Noritoshi Demizu Reveiewed by: Mohan Srinivasan, Raja Mukerji Approved by: re
2005-06-27 22:27:42 +00:00
if (SEQ_GT(th->th_ack, tp->snd_max)) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvacktoomuch);
- Postpone SACK option processing until after PAWS checks. SACK option processing is now done in the ACK processing case. - Merge tcp_sack_option() and tcp_del_sackholes() into a new function called tcp_sack_doack(). - Test (SEG.ACK < SND.MAX) before processing the ACK. Submitted by: Noritoshi Demizu Reveiewed by: Mohan Srinivasan, Raja Mukerji Approved by: re
2005-06-27 22:27:42 +00:00
goto dropafterack;
}
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if ((tp->t_flags & TF_SACK_PERMIT) &&
Bring SACK option handling in tcp_dooptions() in line with all other options and ajust users accordingly.
2007-03-23 18:33:21 +00:00
((to.to_flags & TOF_SACK) ||
!TAILQ_EMPTY(&tp->snd_holes)))
One of the ways to detect loss is to count duplicate acks coming back from the other end till it reaches predetermined threshold which is 3 for us right now. Once that happens, we trigger fast-retransmit to do loss recovery. Main problem with the current implementation is that we don't honor SACK information well to detect whether an incoming ack is a dupack or not. RFC6675 has latest recommendations for that. According to it, dupack is a segment that arrives carrying a SACK block that identifies previously unknown information between snd_una and snd_max even if it carries new data, changes the advertised window, or moves the cumulative acknowledgment point. With the prevalence of Selective ACK (SACK) these days, improper handling can lead to delayed loss recovery. With the fix, new behavior looks like following: 0) th_ack < snd_una --> ignore Old acks are ignored. 1) th_ack == snd_una, !sack_changed --> ignore Acks with SACK enabled but without any new SACK info in them are ignored. 2) th_ack == snd_una, window == old_window --> increment Increment on a good dupack. 3) th_ack == snd_una, window != old_window, sack_changed --> increment When SACK enabled, it's okay to have advertized window changed if the ack has new SACK info. 4) th_ack > snd_una --> reset to 0 Reset to 0 when left edge moves. 5) th_ack > snd_una, sack_changed --> increment Increment if left edge moves but there is new SACK info. Here, sack_changed is the indicator that incoming ack has previously unknown SACK info in it. Note: This fix is not fully compliant to RFC6675. That may require a few changes to current implementation in order to keep per-sackhole dupack counter and change to the way we mark/handle sack holes. PR: 203663 Reviewed by: jtl MFC after: 3 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D4225
2015-12-08 21:21:48 +00:00
sack_changed = tcp_sack_doack(tp, &to, th->th_ack);
Calculate the correct amount of bytes that are in-flight for a connection as suggested by RFC 6675. Currently differnt places in the stack tries to guess this in suboptimal ways. The main problem is that current calculations don't take sacked bytes into account. Sacked bytes are the bytes receiver acked via SACK option. This is suboptimal because it assumes that network has more outstanding (unacked) bytes than the actual value and thus sends less data by setting congestion window lower than what's possible which in turn may cause slower recovery from losses. As an example, one of the current calculations looks something like this: snd_nxt - snd_fack + sackhint.sack_bytes_rexmit New proposal from RFC 6675 is: snd_max - snd_una - sackhint.sacked_bytes + sackhint.sack_bytes_rexmit which takes sacked bytes into account which is a new addition to the sackhint struct. Only thing we are missing from RFC 6675 is isLost() i.e. segment being considered lost and thus adjusting pipe based on that which makes this calculation a bit on conservative side. The approach is very simple. We already process each ack with sack info in tcp_sack_doack() and extract sack blocks/holes out of it. We'd now also track this new variable sacked_bytes which keeps track of total sacked bytes reported. One downside to this approach is that we may get incorrect count of sacked_bytes if the other end decides to drop sack info in the ack because of memory pressure or some other reasons. But in this (not very likely) case also the pipe calculation would be conservative which is okay as opposed to being aggressive in sending packets into the network. Next step is to use this more accurate pipe estimation to drive congestion window adjustments. In collaboration with: rrs Reviewed by: jason_eggnet dot com, rrs MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D3971
2015-10-28 22:57:51 +00:00
else
/*
* Reset the value so that previous (valid) value
* from the last ack with SACK doesn't get used.
*/
tp->sackhint.sacked_bytes = 0;
- Add some helper hook points to the TCP stack. The hooks allow Khelp modules to access inbound/outbound events and associated data for established TCP connections. The hooks only run if at least one hook function is registered for the hook point, ensuring the impact on the stack is effectively nil when no TCP Khelp modules are loaded. struct tcp_hhook_data is passed as contextual data to any registered Khelp module hook functions. - Add an OSD (Object Specific Data) pointer to struct tcpcb to allow Khelp modules to associate per-connection data with the TCP control block. - Bump __FreeBSD_version and add a note to UPDATING regarding to ABI changes introduced by this commit and r216753. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: FreeBSD Foundation Reviewed by: bz, others along the way MFC after: 3 months
2010-12-28 12:13:30 +00:00
/* Run HHOOK_TCP_ESTABLISHED_IN helper hooks. */
hhook_run_tcp_est_in(tp, th, &to);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (SEQ_LEQ(th->th_ack, tp->snd_una)) {
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
u_int maxseg;
maxseg = tcp_maxseg(tp);
One of the ways to detect loss is to count duplicate acks coming back from the other end till it reaches predetermined threshold which is 3 for us right now. Once that happens, we trigger fast-retransmit to do loss recovery. Main problem with the current implementation is that we don't honor SACK information well to detect whether an incoming ack is a dupack or not. RFC6675 has latest recommendations for that. According to it, dupack is a segment that arrives carrying a SACK block that identifies previously unknown information between snd_una and snd_max even if it carries new data, changes the advertised window, or moves the cumulative acknowledgment point. With the prevalence of Selective ACK (SACK) these days, improper handling can lead to delayed loss recovery. With the fix, new behavior looks like following: 0) th_ack < snd_una --> ignore Old acks are ignored. 1) th_ack == snd_una, !sack_changed --> ignore Acks with SACK enabled but without any new SACK info in them are ignored. 2) th_ack == snd_una, window == old_window --> increment Increment on a good dupack. 3) th_ack == snd_una, window != old_window, sack_changed --> increment When SACK enabled, it's okay to have advertized window changed if the ack has new SACK info. 4) th_ack > snd_una --> reset to 0 Reset to 0 when left edge moves. 5) th_ack > snd_una, sack_changed --> increment Increment if left edge moves but there is new SACK info. Here, sack_changed is the indicator that incoming ack has previously unknown SACK info in it. Note: This fix is not fully compliant to RFC6675. That may require a few changes to current implementation in order to keep per-sackhole dupack counter and change to the way we mark/handle sack holes. PR: 203663 Reviewed by: jtl MFC after: 3 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D4225
2015-12-08 21:21:48 +00:00
if (tlen == 0 &&
(tiwin == tp->snd_wnd ||
(tp->t_flags & TF_SACK_PERMIT))) {
Adjust r239672 from rrs and r258821 from eadler. By definition, the very first FIN is not a duplicate. Process it normally and don't feed it to congestion control as though it were a dupe. Don't prevent CC from seeing later dupe acks while in a half close state.
2014-01-28 21:13:15 +00:00
/*
* If this is the first time we've seen a
* FIN from the remote, this is not a
* duplicate and it needs to be processed
* normally. This happens during a
* simultaneous close.
*/
if ((thflags & TH_FIN) &&
(TCPS_HAVERCVDFIN(tp->t_state) == 0)) {
tp->t_dupacks = 0;
break;
}
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvdupack);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* If we have outstanding data (other than
* a window probe), this is a completely
* duplicate ack (ie, window info didn't
In a situation where: - The remote host sends a FIN - in an ACK for a sequence number for which an ACK has already been received - There is still unacked data on route to the remote host - The packet does not contain a window update The packet may be dropped without processing the FIN flag. PR: kern/99188 Submitted by: Staffan Ulfberg <staffan@ulfberg.se> Discussed with: andre MFC after: never
2013-12-02 03:11:25 +00:00
* change and FIN isn't set),
* the ack is the biggest we've
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* seen and we've seen exactly our rexmt
sys/net*: minor spelling fixes. No functional change.
2016-05-03 18:05:43 +00:00
* threshold of them, assume a packet
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* has been dropped and retransmit it.
* Kludge snd_nxt & the congestion
* window so we send only this one
* packet.
*
* We know we're losing at the current
* window size so do congestion avoidance
* (set ssthresh to half the current window
* and pull our congestion window back to
* the new ssthresh).
*
* Dup acks mean that packets have left the
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
* network (they're now cached at the receiver)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* so bump cwnd by the amount in the receiver
* to keep a constant cwnd packets in the
* network.
MFp4 (//depot/projects/tcpecn/): TCP ECN support. Merge of my GSoC 2006 work for NetBSD. TCP ECN is defined in RFC 3168. Partly reviewed by: dwmalone, silby Obtained from: NetBSD
2008-07-31 15:10:09 +00:00
*
* When using TCP ECN, notify the peer that
* we reduced the cwnd.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
One of the ways to detect loss is to count duplicate acks coming back from the other end till it reaches predetermined threshold which is 3 for us right now. Once that happens, we trigger fast-retransmit to do loss recovery. Main problem with the current implementation is that we don't honor SACK information well to detect whether an incoming ack is a dupack or not. RFC6675 has latest recommendations for that. According to it, dupack is a segment that arrives carrying a SACK block that identifies previously unknown information between snd_una and snd_max even if it carries new data, changes the advertised window, or moves the cumulative acknowledgment point. With the prevalence of Selective ACK (SACK) these days, improper handling can lead to delayed loss recovery. With the fix, new behavior looks like following: 0) th_ack < snd_una --> ignore Old acks are ignored. 1) th_ack == snd_una, !sack_changed --> ignore Acks with SACK enabled but without any new SACK info in them are ignored. 2) th_ack == snd_una, window == old_window --> increment Increment on a good dupack. 3) th_ack == snd_una, window != old_window, sack_changed --> increment When SACK enabled, it's okay to have advertized window changed if the ack has new SACK info. 4) th_ack > snd_una --> reset to 0 Reset to 0 when left edge moves. 5) th_ack > snd_una, sack_changed --> increment Increment if left edge moves but there is new SACK info. Here, sack_changed is the indicator that incoming ack has previously unknown SACK info in it. Note: This fix is not fully compliant to RFC6675. That may require a few changes to current implementation in order to keep per-sackhole dupack counter and change to the way we mark/handle sack holes. PR: 203663 Reviewed by: jtl MFC after: 3 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D4225
2015-12-08 21:21:48 +00:00
/*
* Following 2 kinds of acks should not affect
* dupack counting:
* 1) Old acks
* 2) Acks with SACK but without any new SACK
* information in them. These could result from
* any anomaly in the network like a switch
* duplicating packets or a possible DoS attack.
*/
if (th->th_ack != tp->snd_una ||
((tp->t_flags & TF_SACK_PERMIT) &&
!sack_changed))
break;
else if (!tcp_timer_active(tp, TT_REXMT))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->t_dupacks = 0;
Fix NewReno. Reviewed by: Tom Henderson <thomas.r.henderson@boeing.com>
2003-01-13 11:01:20 +00:00
else if (++tp->t_dupacks > tcprexmtthresh ||
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
IN_FASTRECOVERY(tp->t_flags)) {
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
cc_ack_received(tp, th, nsegs,
CC_DUPACK);
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if ((tp->t_flags & TF_SACK_PERMIT) &&
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
IN_FASTRECOVERY(tp->t_flags)) {
This is conform with the terminology in M.Mathis and J.Mahdavi, "Forward Acknowledgement: Refining TCP Congestion Control" SIGCOMM'96, August 1996. Submitted by: Noritoshi Demizu, Raja Mukerji
2005-05-25 17:55:27 +00:00
int awnd;
Fix for a TCP SACK bug where more than (win/2) bytes could have been in flight in SACK recovery. Found by: Noritoshi Demizu Submitted by: Mohan Srinivasan <mohans at yahoo-inc dot com> Noritoshi Demizu <demizu at dd dot ij4u dot or dot jp> Raja Mukerji <raja at moselle dot com>
2005-04-14 20:09:52 +00:00
/*
* Compute the amount of data in flight first.
Reverting r284710. Today I learned: iff == if and only if. Suggested by: many
2015-06-22 22:16:06 +00:00
* We can inject new data into the pipe iff
Trim extra spaces before tabs.
2011-01-07 21:40:34 +00:00
* we have less than 1/2 the original window's
Fix for a TCP SACK bug where more than (win/2) bytes could have been in flight in SACK recovery. Found by: Noritoshi Demizu Submitted by: Mohan Srinivasan <mohans at yahoo-inc dot com> Noritoshi Demizu <demizu at dd dot ij4u dot or dot jp> Raja Mukerji <raja at moselle dot com>
2005-04-14 20:09:52 +00:00
* worth of data in flight.
*/
Calculate the correct amount of bytes that are in-flight for a connection as suggested by RFC 6675. Currently differnt places in the stack tries to guess this in suboptimal ways. The main problem is that current calculations don't take sacked bytes into account. Sacked bytes are the bytes receiver acked via SACK option. This is suboptimal because it assumes that network has more outstanding (unacked) bytes than the actual value and thus sends less data by setting congestion window lower than what's possible which in turn may cause slower recovery from losses. As an example, one of the current calculations looks something like this: snd_nxt - snd_fack + sackhint.sack_bytes_rexmit New proposal from RFC 6675 is: snd_max - snd_una - sackhint.sacked_bytes + sackhint.sack_bytes_rexmit which takes sacked bytes into account which is a new addition to the sackhint struct. Only thing we are missing from RFC 6675 is isLost() i.e. segment being considered lost and thus adjusting pipe based on that which makes this calculation a bit on conservative side. The approach is very simple. We already process each ack with sack info in tcp_sack_doack() and extract sack blocks/holes out of it. We'd now also track this new variable sacked_bytes which keeps track of total sacked bytes reported. One downside to this approach is that we may get incorrect count of sacked_bytes if the other end decides to drop sack info in the ack because of memory pressure or some other reasons. But in this (not very likely) case also the pipe calculation would be conservative which is okay as opposed to being aggressive in sending packets into the network. Next step is to use this more accurate pipe estimation to drive congestion window adjustments. In collaboration with: rrs Reviewed by: jason_eggnet dot com, rrs MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D3971
2015-10-28 22:57:51 +00:00
if (V_tcp_do_rfc6675_pipe)
awnd = tcp_compute_pipe(tp);
else
awnd = (tp->snd_nxt - tp->snd_fack) +
tp->sackhint.sack_bytes_rexmit;
This is conform with the terminology in M.Mathis and J.Mahdavi, "Forward Acknowledgement: Refining TCP Congestion Control" SIGCOMM'96, August 1996. Submitted by: Noritoshi Demizu, Raja Mukerji
2005-05-25 17:55:27 +00:00
if (awnd < tp->snd_ssthresh) {
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd += maxseg;
Fix for a TCP SACK bug where more than (win/2) bytes could have been in flight in SACK recovery. Found by: Noritoshi Demizu Submitted by: Mohan Srinivasan <mohans at yahoo-inc dot com> Noritoshi Demizu <demizu at dd dot ij4u dot or dot jp> Raja Mukerji <raja at moselle dot com>
2005-04-14 20:09:52 +00:00
if (tp->snd_cwnd > tp->snd_ssthresh)
tp->snd_cwnd = tp->snd_ssthresh;
}
} else
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd += maxseg;
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Fix NewReno. Reviewed by: Tom Henderson <thomas.r.henderson@boeing.com>
2003-01-13 11:01:20 +00:00
goto drop;
} else if (tp->t_dupacks == tcprexmtthresh) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tcp_seq onxt = tp->snd_nxt;
On receiving 3 duplicate acknowledgements, SACK recovery was not being entered correctly. Fix this problem by separating out the SACK and the newreno cases. Also, check if we are in FASTRECOVERY for the sack case and if so, turn off dupacks. Fix an issue where the congestion window was not being incremented by ssthresh. Thanks to Mohan Srinivasan for finding this problem.
2004-07-01 23:34:06 +00:00
/*
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
* If we're doing sack, check to
* see if we're already in sack
On receiving 3 duplicate acknowledgements, SACK recovery was not being entered correctly. Fix this problem by separating out the SACK and the newreno cases. Also, check if we are in FASTRECOVERY for the sack case and if so, turn off dupacks. Fix an issue where the congestion window was not being incremented by ssthresh. Thanks to Mohan Srinivasan for finding this problem.
2004-07-01 23:34:06 +00:00
* recovery. If we're not doing sack,
* check to see if we're in newreno
* recovery.
*/
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if (tp->t_flags & TF_SACK_PERMIT) {
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
if (IN_FASTRECOVERY(tp->t_flags)) {
On receiving 3 duplicate acknowledgements, SACK recovery was not being entered correctly. Fix this problem by separating out the SACK and the newreno cases. Also, check if we are in FASTRECOVERY for the sack case and if so, turn off dupacks. Fix an issue where the congestion window was not being incremented by ssthresh. Thanks to Mohan Srinivasan for finding this problem.
2004-07-01 23:34:06 +00:00
tp->t_dupacks = 0;
break;
}
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
} else {
On receiving 3 duplicate acknowledgements, SACK recovery was not being entered correctly. Fix this problem by separating out the SACK and the newreno cases. Also, check if we are in FASTRECOVERY for the sack case and if so, turn off dupacks. Fix an issue where the congestion window was not being incremented by ssthresh. Thanks to Mohan Srinivasan for finding this problem.
2004-07-01 23:34:06 +00:00
if (SEQ_LEQ(th->th_ack,
tp->snd_recover)) {
tp->t_dupacks = 0;
break;
}
Implement TCP NewReno, as documented in RFC 2582. This allows better recovery for multiple packet losses in a single window. The algorithm can be toggled via the sysctl net.inet.tcp.newreno, which defaults to "on". Submitted by: Jayanth Vijayaraghavan <jayanth@yahoo-inc.com>
2000-05-06 03:31:09 +00:00
}
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
/* Congestion signal before ack. */
cc_cong_signal(tp, th, CC_NDUPACK);
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
cc_ack_received(tp, th, nsegs,
CC_DUPACK);
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
tcp_timer_activate(tp, TT_REXMT, 0);
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
tp->t_rtttime = 0;
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if (tp->t_flags & TF_SACK_PERMIT) {
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(
tcps_sack_recovery_episode);
- Estimate the amount of data in flight in sack recovery and use it to control the packets injected while in sack recovery (for both retransmissions and new data). - Cleanups to the sack codepaths in tcp_output.c and tcp_sack.c. - Add a new sysctl (net.inet.tcp.sack.initburst) that controls the number of sack retransmissions done upon initiation of sack recovery. Submitted by: Mohan Srinivasan <mohans@yahoo-inc.com>
2004-10-05 18:36:24 +00:00
tp->sack_newdata = tp->snd_nxt;
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd = maxseg;
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
goto drop;
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tp->snd_nxt = th->th_ack;
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd = maxseg;
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Observe conservation of packets when entering Fast Recovery while doing Limited Transmit. Only artificially inflate the congestion window by 1 segment instead of the usual 3 to take into account the 2 already sent by Limited Transmit. Approved in principle by: Mark Allman <mallman@grc.nasa.gov>, Hari Balakrishnan <hari@nms.lcs.mit.edu>, Sally Floyd <floyd@icir.org>
2003-04-01 21:16:46 +00:00
KASSERT(tp->snd_limited <= 2,
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
("%s: tp->snd_limited too big",
__func__));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->snd_cwnd = tp->snd_ssthresh +
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
maxseg *
Observe conservation of packets when entering Fast Recovery while doing Limited Transmit. Only artificially inflate the congestion window by 1 segment instead of the usual 3 to take into account the 2 already sent by Limited Transmit. Approved in principle by: Mark Allman <mallman@grc.nasa.gov>, Hari Balakrishnan <hari@nms.lcs.mit.edu>, Sally Floyd <floyd@icir.org>
2003-04-01 21:16:46 +00:00
(tp->t_dupacks - tp->snd_limited);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
goto drop;
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
} else if (V_tcp_do_rfc3042) {
Add a comment specifying how we implement rfc3042. Differential Revision: D3746 MFC after: 1 week Sponsored by: Limelight Networks
2015-10-06 07:46:19 +00:00
/*
* Process first and second duplicate
* ACKs. Each indicates a segment
* leaving the network, creating room
* for more. Make sure we can send a
* packet on reception of each duplicate
* ACK by increasing snd_cwnd by one
* segment. Restore the original
* snd_cwnd after packet transmission.
*/
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
cc_ack_received(tp, th, nsegs,
CC_DUPACK);
Implement the Limited Transmit algorithm (RFC 3042).
2003-03-12 20:27:28 +00:00
u_long oldcwnd = tp->snd_cwnd;
Observe conservation of packets when entering Fast Recovery while doing Limited Transmit. Only artificially inflate the congestion window by 1 segment instead of the usual 3 to take into account the 2 already sent by Limited Transmit. Approved in principle by: Mark Allman <mallman@grc.nasa.gov>, Hari Balakrishnan <hari@nms.lcs.mit.edu>, Sally Floyd <floyd@icir.org>
2003-04-01 21:16:46 +00:00
tcp_seq oldsndmax = tp->snd_max;
u_int sent;
When doing RFC3042 limited transmit on the first on second duplicate ACK make sure we actually have new data to send. This prevents us from sending unneccessary pure ACKs. Reported by: Matt Miller <matt@matthewjmiller.net> Tested by: Matt Miller <matt@matthewjmiller.net> MFC after: 2 weeks
2013-04-23 14:06:32 +00:00
int avail;
Relax a KASSERT condition to allow for a valid corner case where the FIN on the last segment consumes an extra sequence number. Spurious panic reported by Mike Silbersack <silby@silby.com>.
2004-02-25 08:53:17 +00:00
Implement the Limited Transmit algorithm (RFC 3042).
2003-03-12 20:27:28 +00:00
KASSERT(tp->t_dupacks == 1 ||
tp->t_dupacks == 2,
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
("%s: dupacks not 1 or 2",
__func__));
Merge from DragonFlyBSD rev 1.10: date: 2003/09/02 10:04:47; author: hsu; state: Exp; lines: +5 -6 Account for when Limited Transmit is not congestion window limited. Obtained from: DragonFlyBSD
2004-01-20 21:40:25 +00:00
if (tp->t_dupacks == 1)
Observe conservation of packets when entering Fast Recovery while doing Limited Transmit. Only artificially inflate the congestion window by 1 segment instead of the usual 3 to take into account the 2 already sent by Limited Transmit. Approved in principle by: Mark Allman <mallman@grc.nasa.gov>, Hari Balakrishnan <hari@nms.lcs.mit.edu>, Sally Floyd <floyd@icir.org>
2003-04-01 21:16:46 +00:00
tp->snd_limited = 0;
Merge from DragonFlyBSD rev 1.10: date: 2003/09/02 10:04:47; author: hsu; state: Exp; lines: +5 -6 Account for when Limited Transmit is not congestion window limited. Obtained from: DragonFlyBSD
2004-01-20 21:40:25 +00:00
tp->snd_cwnd =
(tp->snd_nxt - tp->snd_una) +
(tp->t_dupacks - tp->snd_limited) *
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
maxseg;
When doing RFC3042 limited transmit on the first on second duplicate ACK make sure we actually have new data to send. This prevents us from sending unneccessary pure ACKs. Reported by: Matt Miller <matt@matthewjmiller.net> Tested by: Matt Miller <matt@matthewjmiller.net> MFC after: 2 weeks
2013-04-23 14:06:32 +00:00
/*
* Only call tcp_output when there
* is new data available to be sent.
* Otherwise we would send pure ACKs.
*/
SOCKBUF_LOCK(&so->so_snd);
In preparation of merging projects/sendfile, transform bare access to sb_cc member of struct sockbuf to a couple of inline functions: sbavail() and sbused() Right now they are equal, but once notion of "not ready socket buffer data", will be checked in, they are going to be different. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-12 09:57:15 +00:00
avail = sbavail(&so->so_snd) -
When doing RFC3042 limited transmit on the first on second duplicate ACK make sure we actually have new data to send. This prevents us from sending unneccessary pure ACKs. Reported by: Matt Miller <matt@matthewjmiller.net> Tested by: Matt Miller <matt@matthewjmiller.net> MFC after: 2 weeks
2013-04-23 14:06:32 +00:00
(tp->snd_nxt - tp->snd_una);
SOCKBUF_UNLOCK(&so->so_snd);
if (avail > 0)
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Observe conservation of packets when entering Fast Recovery while doing Limited Transmit. Only artificially inflate the congestion window by 1 segment instead of the usual 3 to take into account the 2 already sent by Limited Transmit. Approved in principle by: Mark Allman <mallman@grc.nasa.gov>, Hari Balakrishnan <hari@nms.lcs.mit.edu>, Sally Floyd <floyd@icir.org>
2003-04-01 21:16:46 +00:00
sent = tp->snd_max - oldsndmax;
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
if (sent > maxseg) {
Relax a KASSERT condition to allow for a valid corner case where the FIN on the last segment consumes an extra sequence number. Spurious panic reported by Mike Silbersack <silby@silby.com>.
2004-02-25 08:53:17 +00:00
KASSERT((tp->t_dupacks == 2 &&
tp->snd_limited == 0) ||
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
(sent == maxseg + 1 &&
Relax a KASSERT condition to allow for a valid corner case where the FIN on the last segment consumes an extra sequence number. Spurious panic reported by Mike Silbersack <silby@silby.com>.
2004-02-25 08:53:17 +00:00
tp->t_flags & TF_SENTFIN),
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
("%s: sent too much",
__func__));
Observe conservation of packets when entering Fast Recovery while doing Limited Transmit. Only artificially inflate the congestion window by 1 segment instead of the usual 3 to take into account the 2 already sent by Limited Transmit. Approved in principle by: Mark Allman <mallman@grc.nasa.gov>, Hari Balakrishnan <hari@nms.lcs.mit.edu>, Sally Floyd <floyd@icir.org>
2003-04-01 21:16:46 +00:00
tp->snd_limited = 2;
} else if (sent > 0)
++tp->snd_limited;
Implement the Limited Transmit algorithm (RFC 3042).
2003-03-12 20:27:28 +00:00
tp->snd_cwnd = oldcwnd;
goto drop;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
One of the ways to detect loss is to count duplicate acks coming back from the other end till it reaches predetermined threshold which is 3 for us right now. Once that happens, we trigger fast-retransmit to do loss recovery. Main problem with the current implementation is that we don't honor SACK information well to detect whether an incoming ack is a dupack or not. RFC6675 has latest recommendations for that. According to it, dupack is a segment that arrives carrying a SACK block that identifies previously unknown information between snd_una and snd_max even if it carries new data, changes the advertised window, or moves the cumulative acknowledgment point. With the prevalence of Selective ACK (SACK) these days, improper handling can lead to delayed loss recovery. With the fix, new behavior looks like following: 0) th_ack < snd_una --> ignore Old acks are ignored. 1) th_ack == snd_una, !sack_changed --> ignore Acks with SACK enabled but without any new SACK info in them are ignored. 2) th_ack == snd_una, window == old_window --> increment Increment on a good dupack. 3) th_ack == snd_una, window != old_window, sack_changed --> increment When SACK enabled, it's okay to have advertized window changed if the ack has new SACK info. 4) th_ack > snd_una --> reset to 0 Reset to 0 when left edge moves. 5) th_ack > snd_una, sack_changed --> increment Increment if left edge moves but there is new SACK info. Here, sack_changed is the indicator that incoming ack has previously unknown SACK info in it. Note: This fix is not fully compliant to RFC6675. That may require a few changes to current implementation in order to keep per-sackhole dupack counter and change to the way we mark/handle sack holes. PR: 203663 Reviewed by: jtl MFC after: 3 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D4225
2015-12-08 21:21:48 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
One of the ways to detect loss is to count duplicate acks coming back from the other end till it reaches predetermined threshold which is 3 for us right now. Once that happens, we trigger fast-retransmit to do loss recovery. Main problem with the current implementation is that we don't honor SACK information well to detect whether an incoming ack is a dupack or not. RFC6675 has latest recommendations for that. According to it, dupack is a segment that arrives carrying a SACK block that identifies previously unknown information between snd_una and snd_max even if it carries new data, changes the advertised window, or moves the cumulative acknowledgment point. With the prevalence of Selective ACK (SACK) these days, improper handling can lead to delayed loss recovery. With the fix, new behavior looks like following: 0) th_ack < snd_una --> ignore Old acks are ignored. 1) th_ack == snd_una, !sack_changed --> ignore Acks with SACK enabled but without any new SACK info in them are ignored. 2) th_ack == snd_una, window == old_window --> increment Increment on a good dupack. 3) th_ack == snd_una, window != old_window, sack_changed --> increment When SACK enabled, it's okay to have advertized window changed if the ack has new SACK info. 4) th_ack > snd_una --> reset to 0 Reset to 0 when left edge moves. 5) th_ack > snd_una, sack_changed --> increment Increment if left edge moves but there is new SACK info. Here, sack_changed is the indicator that incoming ack has previously unknown SACK info in it. Note: This fix is not fully compliant to RFC6675. That may require a few changes to current implementation in order to keep per-sackhole dupack counter and change to the way we mark/handle sack holes. PR: 203663 Reviewed by: jtl MFC after: 3 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D4225
2015-12-08 21:21:48 +00:00
} else {
/*
* This ack is advancing the left edge, reset the
* counter.
*/
tp->t_dupacks = 0;
/*
* If this ack also has new SACK info, increment the
* counter as per rfc6675.
*/
if ((tp->t_flags & TF_SACK_PERMIT) && sack_changed)
tp->t_dupacks++;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Fix NewReno. Reviewed by: Tom Henderson <thomas.r.henderson@boeing.com>
2003-01-13 11:01:20 +00:00
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
KASSERT(SEQ_GT(th->th_ack, tp->snd_una),
("%s: th_ack <= snd_una", __func__));
Fix NewReno. Reviewed by: Tom Henderson <thomas.r.henderson@boeing.com>
2003-01-13 11:01:20 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* If the congestion window was inflated to account
* for the other side's cached packets, retract it.
*/
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
if (IN_FASTRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (tp->t_flags & TF_SACK_PERMIT)
tcp_sack_partialack(tp, th);
else
tcp_newreno_partial_ack(tp, th);
} else
cc_post_recovery(tp, th);
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
* If we reach this point, ACK is not a duplicate,
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
* i.e., it ACKs something we sent.
*/
if (tp->t_flags & TF_NEEDSYN) {
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
/*
Fix a bug related to the interworking of T/TCP and window scaling: when a connection enters the ESTBLS state using T/TCP, then window scaling wasn't properly handled. The fix is twofold. 1) When the 3WHS completes, make sure that we update our window scaling state variables. 2) When setting the `virtual advertized window', then make sure that we do not try to offer a window that is larger than the maximum window without scaling (TCP_MAXWIN). Reviewed by: davidg Reported by: Jerry Chen <chen@Ipsilon.COM>
1996-01-31 08:22:24 +00:00
* T/TCP: Connection was half-synchronized, and our
* SYN has been ACK'd (so connection is now fully
* synchronized). Go to non-starred state,
* increment snd_una for ACK of SYN, and check if
* we can do window scaling.
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
*/
tp->t_flags &= ~TF_NEEDSYN;
tp->snd_una++;
Fix a bug related to the interworking of T/TCP and window scaling: when a connection enters the ESTBLS state using T/TCP, then window scaling wasn't properly handled. The fix is twofold. 1) When the 3WHS completes, make sure that we update our window scaling state variables. 2) When setting the `virtual advertized window', then make sure that we do not try to offer a window that is larger than the maximum window without scaling (TCP_MAXWIN). Reviewed by: davidg Reported by: Jerry Chen <chen@Ipsilon.COM>
1996-01-31 08:22:24 +00:00
/* Do window scaling? */
if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) ==
(TF_RCVD_SCALE|TF_REQ_SCALE)) {
tp->rcv_scale = tp->request_r_scale;
Rework TCP window scaling (RFC1323) to properly scale the send window right from the beginning and partly clean up the differences in handling between SYN_SENT and SYN_RCVD (syncache). Further changes to this code to come. This is a first incremental step to a general overhaul and streamlining of the TCP code. PR: kern/15095 PR: kern/92690 (partly) Reviewed by: qingli (and tested with ANVL) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-28 23:05:59 +00:00
/* Send window already scaled. */
Fix a bug related to the interworking of T/TCP and window scaling: when a connection enters the ESTBLS state using T/TCP, then window scaling wasn't properly handled. The fix is twofold. 1) When the 3WHS completes, make sure that we update our window scaling state variables. 2) When setting the `virtual advertized window', then make sure that we do not try to offer a window that is larger than the maximum window without scaling (TCP_MAXWIN). Reviewed by: davidg Reported by: Jerry Chen <chen@Ipsilon.COM>
1996-01-31 08:22:24 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
}
process_ACK:
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
After each label in tcp_input(), assert the inpcbinfo and inpcb lock state that we expect.
2004-07-12 19:28:07 +00:00
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
acked = BYTES_THIS_ACK(tp, th);
Prevent underflows in tp->snd_wnd if the remote side ACKs more than tp->snd_wnd. This can happen, for example, when the remote side responds to a window probe by ACKing the one byte it contains. Differential Revision: https://reviews.freebsd.org/D5625 Reviewed by: hiren Obtained from: Juniper Networks (earlier version) MFC after: 2 weeks Sponsored by: Juniper Networks
2016-04-21 15:06:53 +00:00
KASSERT(acked >= 0, ("%s: acked unexepectedly negative "
"(tp->snd_una=%u, th->th_ack=%u, tp=%p, m=%p)", __func__,
tp->snd_una, th->th_ack, tp, m));
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
TCPSTAT_ADD(tcps_rcvackpack, nsegs);
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_ADD(tcps_rcvackbyte, acked);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
/*
* If we just performed our first retransmit, and the ACK
* arrives within our recovery window, then it was a mistake
* to do the retransmit in the first place. Recover our
* original cwnd and ssthresh, and proceed to transmit where
* we left off.
*/
TCP reuses t_rxtshift to determine the backoff timer used for both the persist state and the retransmit timer. However, the code that implements "bad retransmit recovery" only checks t_rxtshift to see if an ACK has been received in during the first retransmit timeout window. As a result, if ticks has wrapped over to a negative value and a socket is in the persist state, it can incorrectly treat an ACK from the remote peer as a "bad retransmit recovery" and restore saved values such as snd_ssthresh and snd_cwnd. However, if the socket has never had a retransmit timeout, then these saved values will be zero, so snd_ssthresh and snd_cwnd will be set to 0. If the socket is in fast recovery (this can be caused by excessive duplicate ACKs such as those fixed by 220794), then each ACK that arrives triggers either NewReno or SACK partial ACK handling which clamps snd_cwnd to be no larger than snd_ssthresh. In effect, the socket's send window is permamently stuck at 0 even though the remote peer is advertising a much larger window and pending data is only sent via TCP window probes (so one byte every few seconds). Fix this by adding a new TCP pcb flag (TF_PREVVALID) that indicates that the various snd_*_prev fields in the pcb are valid and only perform "bad retransmit recovery" if this flag is set in the pcb. The flag is set on the first retransmit timeout that occurs and is cleared on subsequent retransmit timeouts or when entering the persist state. Reviewed by: bz MFC after: 2 weeks
2011-04-29 15:40:12 +00:00
if (tp->t_rxtshift == 1 && tp->t_flags & TF_PREVVALID &&
(int)(ticks - tp->t_badrxtwin) < 0)
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
cc_cong_signal(tp, th, CC_RTO_ERR);
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* If we have a timestamp reply, update smoothed
* round trip time. If no timestamp is present but
* transmit timer is running and timed sequence
* number was acked, update smoothed round trip time.
* Since we now have an rtt measurement, cancel the
* timer backoff (cf., Phil Karn's retransmit alg.).
* Recompute the initial retransmit timer.
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
*
* Some boxes send broken timestamp replies
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
* during the SYN+ACK phase, ignore
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
* timestamps of 0 or we could calculate a
* huge RTT and blow up the retransmit timer.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Fix PAWS (Protect Against Wrapped Sequence numbers) in cases when hz >> 1000 and thus getting outside the timestamp clock frequenceny of 1ms < x < 1s per tick as mandated by RFC1323, leading to connection resets on idle connections. Always use a granularity of 1ms using getmicrouptime() making all but relevant callouts independent of hz. Use getmicrouptime(), not getmicrotime() as the latter may make a jump possibly breaking TCP nfsroot mounts having our timestamps move forward for more than 24.8 days in a second without having been idle for that long. PR: kern/61404 Reviewed by: jhb, mav, rrs Discussed with: silby, lstewart Sponsored by: Sandvine Incorporated (originally in 2011) MFC after: 6 weeks
2012-02-15 16:09:56 +00:00
if ((to.to_flags & TOF_TS) != 0 && to.to_tsecr) {
u_int t;
t = tcp_ts_getticks() - to.to_tsecr;
if (!tp->t_rttlow || tp->t_rttlow > t)
tp->t_rttlow = t;
tcp_xmit_timer(tp, TCP_TS_TO_TICKS(t) + 1);
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
} else if (tp->t_rtttime && SEQ_GT(th->th_ack, tp->t_rtseq)) {
Have TCP Inflight disable itself if the RTT is below a certain threshold. Inflight doesn't make sense on a LAN as it has trouble figuring out the maximal bandwidth because of the coarse tick granularity. The sysctl net.inet.tcp.inflight.rttthresh specifies the threshold in milliseconds below which inflight will disengage. It defaults to 10ms. Tested by: Joao Barros <joao.barros-at-gmail.com>, Rich Murphey <rich-at-whiteoaklabs.com> Sponsored by: TCP/IP Optimization Fundraise 2005
2006-02-16 19:38:07 +00:00
if (!tp->t_rttlow || tp->t_rttlow > ticks - tp->t_rtttime)
tp->t_rttlow = ticks - tp->t_rtttime;
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
tcp_xmit_timer(tp, ticks - tp->t_rtttime);
Guido reported an interesting bug where an FTP connection between a Windows 2000 box and a FreeBSD box could stall. The problem turned out to be a timestamp reply bug in the W2K TCP stack. FreeBSD sends a timestamp with the SYN, W2K returns a timestamp of 0 in the SYN+ACK causing FreeBSD to calculate an insane SRTT and RTT, resulting in a maximal retransmit timeout (60 seconds). If there is any packet loss on the connection for the first six or so packets the retransmit case may be hit (the window will still be too small for fast-retransmit), causing a 60+ second pause. The W2K box gives up and closes the connection. This commit works around the W2K bug. 15:04:59.374588 FREEBSD.20 > W2K.1036: S 1420807004:1420807004(0) win 65535 <mss 1460,nop,wscale 2,nop,nop,timestamp 188297344 0> (DF) [tos 0x8] 15:04:59.377558 W2K.1036 > FREEBSD.20: S 4134611565:4134611565(0) ack 1420807005 win 17520 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) Bug reported by: Guido van Rooij <guido@gvr.org>
2002-09-17 22:21:37 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* If all outstanding data is acked, stop retransmit
* timer and remember to restart (more output or persist).
* If there is more data to be acked, restart retransmit
* timer, using current (possibly backed-off) value.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (th->th_ack == tp->snd_max) {
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
tcp_timer_activate(tp, TT_REXMT, 0);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
needoutput = 1;
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
} else if (!tcp_timer_active(tp, TT_PERSIST))
tcp_timer_activate(tp, TT_REXMT, tp->t_rxtcur);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
* If no data (only SYN) was ACK'd,
* skip rest of ACK processing.
*/
if (acked == 0)
goto step6;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
* Let the congestion control algorithm update congestion
* control related information. This typically means increasing
* the congestion window.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Pass the number of segments coalesced by LRO up the stack by repurposing the tso_segsz pkthdr field during RX processing, and use the information in TCP for more correct accounting and as a congestion control input. This is only a start, and an audit of other uses for the data is left as future work. Reviewed by: gallatin, rrs Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D7564
2016-08-25 13:33:32 +00:00
cc_ack_received(tp, th, nsegs, CC_ACK);
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
Broaden scope of the socket buffer lock when processing an ACK so that the read and write of sb_cc are atomic. Call sbdrop_locked() instead of sbdrop() since we already hold the socket buffer lock.
2004-06-24 03:07:27 +00:00
SOCKBUF_LOCK(&so->so_snd);
In preparation of merging projects/sendfile, transform bare access to sb_cc member of struct sockbuf to a couple of inline functions: sbavail() and sbused() Right now they are equal, but once notion of "not ready socket buffer data", will be checked in, they are going to be different. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-12 09:57:15 +00:00
if (acked > sbavail(&so->so_snd)) {
Prevent underflows in tp->snd_wnd if the remote side ACKs more than tp->snd_wnd. This can happen, for example, when the remote side responds to a window probe by ACKing the one byte it contains. Differential Revision: https://reviews.freebsd.org/D5625 Reviewed by: hiren Obtained from: Juniper Networks (earlier version) MFC after: 2 weeks Sponsored by: Juniper Networks
2016-04-21 15:06:53 +00:00
if (tp->snd_wnd >= sbavail(&so->so_snd))
tp->snd_wnd -= sbavail(&so->so_snd);
else
tp->snd_wnd = 0;
When processing ACK in tcp_do_segment, use sbcut_locked() instead of sbdrop_locked() to cut acked mbufs from the socket buffer. Free this chain a batch manner after the socket buffer lock is dropped. This measurably reduces contention on socket buffer. Sponsored by: Netflix Sponsored by: Nginx, Inc. Approved by: re (marius)
2013-10-09 12:00:38 +00:00
mfree = sbcut_locked(&so->so_snd,
In preparation of merging projects/sendfile, transform bare access to sb_cc member of struct sockbuf to a couple of inline functions: sbavail() and sbused() Right now they are equal, but once notion of "not ready socket buffer data", will be checked in, they are going to be different. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-12 09:57:15 +00:00
(int)sbavail(&so->so_snd));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
ourfinisacked = 1;
} else {
When processing ACK in tcp_do_segment, use sbcut_locked() instead of sbdrop_locked() to cut acked mbufs from the socket buffer. Free this chain a batch manner after the socket buffer lock is dropped. This measurably reduces contention on socket buffer. Sponsored by: Netflix Sponsored by: Nginx, Inc. Approved by: re (marius)
2013-10-09 12:00:38 +00:00
mfree = sbcut_locked(&so->so_snd, acked);
Prevent underflows in tp->snd_wnd if the remote side ACKs more than tp->snd_wnd. This can happen, for example, when the remote side responds to a window probe by ACKing the one byte it contains. Differential Revision: https://reviews.freebsd.org/D5625 Reviewed by: hiren Obtained from: Juniper Networks (earlier version) MFC after: 2 weeks Sponsored by: Juniper Networks
2016-04-21 15:06:53 +00:00
if (tp->snd_wnd >= (u_long) acked)
tp->snd_wnd -= acked;
else
tp->snd_wnd = 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
ourfinisacked = 0;
}
Fix comments in tcp_do_segment(). Approved by: re (kensmith)
2007-07-25 18:48:24 +00:00
/* NB: sowwakeup_locked() does an implicit unlock. */
Reduce the number of unnecessary unlock-relocks on socket buffer mutexes associated with performing a wakeup on the socket buffer: - When performing an sbappend*() followed by a so[rw]wakeup(), explicitly acquire the socket buffer lock and use the _locked() variants of both calls. Note that the _locked() sowakeup() versions unlock the mutex on return. This is done in uipc_send(), divert_packet(), mroute socket_send(), raw_append(), tcp_reass(), tcp_input(), and udp_append(). - When the socket buffer lock is dropped before a sowakeup(), remove the explicit unlock and use the _locked() sowakeup() variant. This is done in soisdisconnecting(), soisdisconnected() when setting the can't send/ receive flags and dropping data, and in uipc_rcvd() which adjusting back-pressure on the sockets. For UNIX domain sockets running mpsafe with a contention-intensive SMP mysql benchmark, this results in a 1.6% query rate improvement due to reduce mutex costs.
2004-06-26 19:10:39 +00:00
sowwakeup_locked(so);
When processing ACK in tcp_do_segment, use sbcut_locked() instead of sbdrop_locked() to cut acked mbufs from the socket buffer. Free this chain a batch manner after the socket buffer lock is dropped. This measurably reduces contention on socket buffer. Sponsored by: Netflix Sponsored by: Nginx, Inc. Approved by: re (marius)
2013-10-09 12:00:38 +00:00
m_freem(mfree);
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
/* Detect una wraparound. */
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
if (!IN_RECOVERY(tp->t_flags) &&
Unify the "send high" and "recover" variables as specified in the lastest rev of the spec. Use an explicit flag for Fast Recovery. [1] Fix bug with exiting Fast Recovery on a retransmit timeout diagnosed by Lu Guohan. [2] Reviewed by: Thomas Henderson <thomas.r.henderson@boeing.com> Reported and tested by: Lu Guohan <lguohan00@mails.tsinghua.edu.cn> [2] Approved by: Thomas Henderson <thomas.r.henderson@boeing.com>, Sally Floyd <floyd@acm.org> [1]
2003-07-15 21:49:53 +00:00
SEQ_GT(tp->snd_una, tp->snd_recover) &&
SEQ_LEQ(th->th_ack, tp->snd_recover))
tp->snd_recover = th->th_ack - 1;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
/* XXXLAS: Can this be moved up into cc_post_recovery? */
if (IN_RECOVERY(tp->t_flags) &&
Add TCP Appropriate Byte Counting (RFC 3465) support to kernel. The new behaviour is on by default, and can be disabled by setting the net.inet.tcp.rfc3465 sysctl to 0 to obtain previous behaviour. The patch changes struct tcpcb in sys/netinet/tcp_var.h which breaks the ABI. Bump __FreeBSD_version to 800061 accordingly. User space tools that rely on the size of struct tcpcb (e.g. sockstat) need to be recompiled. Reviewed by: rpaulo, gnn Approved by: gnn, kmacy (mentors) Sponsored by: FreeBSD Foundation
2009-01-15 06:44:22 +00:00
SEQ_GEQ(th->th_ack, tp->snd_recover)) {
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
EXIT_RECOVERY(tp->t_flags);
Add TCP Appropriate Byte Counting (RFC 3465) support to kernel. The new behaviour is on by default, and can be disabled by setting the net.inet.tcp.rfc3465 sysctl to 0 to obtain previous behaviour. The patch changes struct tcpcb in sys/netinet/tcp_var.h which breaks the ABI. Bump __FreeBSD_version to 800061 accordingly. User space tools that rely on the size of struct tcpcb (e.g. sockstat) need to be recompiled. Reviewed by: rpaulo, gnn Approved by: gnn, kmacy (mentors) Sponsored by: FreeBSD Foundation
2009-01-15 06:44:22 +00:00
}
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tp->snd_una = th->th_ack;
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if (tp->t_flags & TF_SACK_PERMIT) {
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
if (SEQ_GT(tp->snd_una, tp->snd_recover))
tp->snd_recover = tp->snd_una;
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (SEQ_LT(tp->snd_nxt, tp->snd_una))
tp->snd_nxt = tp->snd_una;
switch (tp->t_state) {
/*
* In FIN_WAIT_1 STATE in addition to the processing
* for the ESTABLISHED state if our FIN is now acknowledged
* then enter FIN_WAIT_2.
*/
case TCPS_FIN_WAIT_1:
if (ourfinisacked) {
/*
* If we can't receive any more
* data, then closing user can proceed.
* Starting the timer is contrary to the
* specification, but if we don't get a FIN
* we'll hang forever.
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
*
* XXXjl:
* we should release the tp also, and use a
* compressed state.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
The socket field so_state is used to hold a variety of socket related flags relating to several aspects of socket functionality. This change breaks out several bits relating to send and receive operation into a new per-socket buffer field, sb_state, in order to facilitate locking. This is required because, in order to provide more granular locking of sockets, different state fields have different locking properties. The following fields are moved to sb_state: SS_CANTRCVMORE (so_state) SS_CANTSENDMORE (so_state) SS_RCVATMARK (so_state) Rename respectively to: SBS_CANTRCVMORE (so_rcv.sb_state) SBS_CANTSENDMORE (so_snd.sb_state) SBS_RCVATMARK (so_rcv.sb_state) This facilitates locking by isolating fields to be located with other identically locked fields, and permits greater granularity in socket locking by avoiding storing fields with different locking semantics in the same short (avoiding locking conflicts). In the future, we may wish to coallesce sb_state and sb_flags; for the time being I leave them separate and there is no additional memory overhead due to the packing/alignment of shorts in the socket buffer structure.
2004-06-14 18:16:22 +00:00
if (so->so_rcv.sb_state & SBS_CANTRCVMORE) {
Remove so*_locked(), which were backed out by mistake.
2002-06-18 07:42:02 +00:00
soisdisconnected(so);
Add new socket options: TCP_KEEPINIT, TCP_KEEPIDLE, TCP_KEEPINTVL and TCP_KEEPCNT, that allow to control initial timeout, idle time, idle re-send interval and idle send count on a per-socket basis. Reviewed by: andre, bz, lstewart
2012-02-05 16:53:02 +00:00
tcp_timer_activate(tp, TT_2MSL,
(tcp_fast_finwait2_recycle ?
tcp_finwait2_timeout :
TP_MAXIDLE(tp)));
Back out my lats commit of locking down a socket, it conflicts with hsu's work. Requested by: hsu
2002-05-31 11:52:35 +00:00
}
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_FIN_WAIT_2);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
break;
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
/*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* In CLOSING STATE in addition to the processing for
* the ESTABLISHED state if the ACK acknowledges our FIN
* then enter the TIME-WAIT state, otherwise ignore
* the segment.
*/
case TCPS_CLOSING:
if (ourfinisacked) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
tcp_twstart() need to be called with the TCP protocol lock held to avoid a race condition with the TCP timer routines.
2003-02-24 00:52:03 +00:00
tcp_twstart(tp);
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RUNLOCK(&V_tcbinfo);
Add a TCP TIMEWAIT state which uses less space than a fullblown TCP control block. Allow the socket and tcpcb structures to be freed earlier than inpcb. Update code to understand an inp w/o a socket. Reviewed by: hsu, silby, jayanth Sponsored by: DARPA, NAI Labs
2003-02-19 22:32:43 +00:00
m_freem(m);
Remove the defunct remains of the TCPS_TIME_WAIT cases from tcp_do_segment and change it to a void function. We use a compressed structure for TCPS_TIME_WAIT to save memory. Any late late segments arriving for such a connection is handled directly in the TW code.
2007-05-06 15:16:05 +00:00
return;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
break;
/*
* In LAST_ACK, we may still be waiting for data to drain
* and/or to be acked, as well as for the ack of our FIN.
* If our FIN is now acknowledged, delete the TCB,
* enter the closed state and return.
*/
case TCPS_LAST_ACK:
if (ourfinisacked) {
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp = tcp_close(tp);
goto drop;
}
break;
}
}
step6:
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
After each label in tcp_input(), assert the inpcbinfo and inpcb lock state that we expect.
2004-07-12 19:28:07 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Back out r242262. The simplified window change/update logic wasn't complete and ready for production use. PR: kern/173309
2012-11-05 09:13:06 +00:00
* Update window information.
* Don't look at window if no ACK: TAC's send garbage on first SYN.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Back out r242262. The simplified window change/update logic wasn't complete and ready for production use. PR: kern/173309
2012-11-05 09:13:06 +00:00
if ((thflags & TH_ACK) &&
(SEQ_LT(tp->snd_wl1, th->th_seq) ||
(tp->snd_wl1 == th->th_seq && (SEQ_LT(tp->snd_wl2, th->th_ack) ||
(tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd))))) {
/* keep track of pure window updates */
if (tlen == 0 &&
tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd)
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvwinupd);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->snd_wnd = tiwin;
Back out r242262. The simplified window change/update logic wasn't complete and ready for production use. PR: kern/173309
2012-11-05 09:13:06 +00:00
tp->snd_wl1 = th->th_seq;
tp->snd_wl2 = th->th_ack;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (tp->snd_wnd > tp->max_sndwnd)
tp->max_sndwnd = tp->snd_wnd;
Back out r242262. The simplified window change/update logic wasn't complete and ready for production use. PR: kern/173309
2012-11-05 09:13:06 +00:00
needoutput = 1;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
/*
* Process segments with URG.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if ((thflags & TH_URG) && th->th_urp &&
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
/*
* This is a kludge, but if we receive and accept
* random urgent pointers, we'll crash in
* soreceive. It's hard to imagine someone
* actually wanting to send this much urgent data.
*/
Expand coverage of the receive socket buffer lock when handling urgent pointer updates: test available space while holding the socket buffer mutex, and continue to hold until until the pointer update has been performed. MFC after: 2 weeks
2004-11-28 11:01:31 +00:00
SOCKBUF_LOCK(&so->so_rcv);
In preparation of merging projects/sendfile, transform bare access to sb_cc member of struct sockbuf to a couple of inline functions: sbavail() and sbused() Right now they are equal, but once notion of "not ready socket buffer data", will be checked in, they are going to be different. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-12 09:57:15 +00:00
if (th->th_urp + sbavail(&so->so_rcv) > sb_max) {
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
th->th_urp = 0; /* XXX */
thflags &= ~TH_URG; /* XXX */
Expand coverage of the receive socket buffer lock when handling urgent pointer updates: test available space while holding the socket buffer mutex, and continue to hold until until the pointer update has been performed. MFC after: 2 weeks
2004-11-28 11:01:31 +00:00
SOCKBUF_UNLOCK(&so->so_rcv); /* XXX */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto dodata; /* XXX */
}
/*
* If this segment advances the known urgent pointer,
* then mark the data stream. This should not happen
* in CLOSE_WAIT, CLOSING, LAST_ACK or TIME_WAIT STATES since
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
* a FIN has been received from the remote side.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* In these states we ignore the URG.
*
* According to RFC961 (Assigned Protocols),
* the urgent pointer points to the last octet
* of urgent data. We continue, however,
* to consider it to indicate the first octet
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
* of data past the urgent section as the original
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* spec states (in one of two places).
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (SEQ_GT(th->th_seq+th->th_urp, tp->rcv_up)) {
tp->rcv_up = th->th_seq + th->th_urp;
In preparation of merging projects/sendfile, transform bare access to sb_cc member of struct sockbuf to a couple of inline functions: sbavail() and sbused() Right now they are equal, but once notion of "not ready socket buffer data", will be checked in, they are going to be different. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-12 09:57:15 +00:00
so->so_oobmark = sbavail(&so->so_rcv) +
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
(tp->rcv_up - tp->rcv_nxt) - 1;
Protect so_oobmark with with SOCKBUF_LOCK(&so->so_rcv), and broaden locking in tcp_input() for TCP packets with urgent data pointers to hold the socket buffer lock across testing and updating oobmark from just protecting sb_state. Update socket locking annotations
2004-06-24 02:57:12 +00:00
if (so->so_oobmark == 0)
The socket field so_state is used to hold a variety of socket related flags relating to several aspects of socket functionality. This change breaks out several bits relating to send and receive operation into a new per-socket buffer field, sb_state, in order to facilitate locking. This is required because, in order to provide more granular locking of sockets, different state fields have different locking properties. The following fields are moved to sb_state: SS_CANTRCVMORE (so_state) SS_CANTSENDMORE (so_state) SS_RCVATMARK (so_state) Rename respectively to: SBS_CANTRCVMORE (so_rcv.sb_state) SBS_CANTSENDMORE (so_snd.sb_state) SBS_RCVATMARK (so_rcv.sb_state) This facilitates locking by isolating fields to be located with other identically locked fields, and permits greater granularity in socket locking by avoiding storing fields with different locking semantics in the same short (avoiding locking conflicts). In the future, we may wish to coallesce sb_state and sb_flags; for the time being I leave them separate and there is no additional memory overhead due to the packing/alignment of shorts in the socket buffer structure.
2004-06-14 18:16:22 +00:00
so->so_rcv.sb_state |= SBS_RCVATMARK;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
sohasoutofband(so);
tp->t_oobflags &= ~(TCPOOB_HAVEDATA | TCPOOB_HADDATA);
}
Expand coverage of the receive socket buffer lock when handling urgent pointer updates: test available space while holding the socket buffer mutex, and continue to hold until until the pointer update has been performed. MFC after: 2 weeks
2004-11-28 11:01:31 +00:00
SOCKBUF_UNLOCK(&so->so_rcv);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Remove out of band data so doesn't get presented to user.
* This can happen independent of advancing the URG pointer,
* but if two URG's are pending at once, some out-of-band
* data may creep in... ick.
*/
Don't need to check if SO_OOBINLINE is defined. Don't need to protect isipv6 conditional with INET6. Fix leading indentation in 2 lines.
2002-10-30 08:32:19 +00:00
if (th->th_urp <= (u_long)tlen &&
!(so->so_options & SO_OOBINLINE)) {
/* hdr drop is delayed */
tcp_pulloutofband(so, th, m, drop_hdrlen);
}
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
} else {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* If no out of band data is expected,
* pull receive urgent pointer along
* with the receive window.
*/
if (SEQ_GT(tp->rcv_nxt, tp->rcv_up))
tp->rcv_up = tp->rcv_nxt;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
dodata: /* XXX */
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
After each label in tcp_input(), assert the inpcbinfo and inpcb lock state that we expect.
2004-07-12 19:28:07 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Process the segment text, merging it into the TCP sequencing queue,
* and arranging for acknowledgment of receipt if necessary.
* This process logically involves adjusting tp->rcv_wnd as data
* is presented to the user (this happens in tcp_usrreq.c,
* case PRU_RCVD). If a FIN has already been received on this
* connection then we just ignore the text.
*/
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
tfo_syn = ((tp->t_state == TCPS_SYN_RECEIVED) &&
(tp->t_flags & TF_FASTOPEN));
if ((tlen || (thflags & TH_FIN) || tfo_syn) &&
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
Fix for a SACK crash caused by a bug in tcp_reass(). tcp_reass() does not clear tlen and frees the mbuf (leaving th pointing at freed memory), if the data segment is a complete duplicate. This change works around that bug. A fix for the tcp_reass() bug will appear later (that bug is benign for now, as neither th nor tlen is referenced in tcp_input() after the call to tcp_reass()). Found by: Pawel Jakub Dawidek. Submitted by: Raja Mukerji, Noritoshi Demizu. Approved by: re
2005-07-01 22:52:46 +00:00
tcp_seq save_start = th->th_seq;
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
m_adj(m, drop_hdrlen); /* delayed header drop */
Inline TCP_REASS() in the single location where it's used, just as OpenBSD and NetBSD has done. No functional difference. MFC after: 2 weeks
2001-05-29 19:54:45 +00:00
/*
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
* Insert segment which includes th into TCP reassembly queue
* with control block tp. Set thflags to whether reassembly now
* includes a segment with FIN. This handles the common case
* inline (segment is the next to be received on an established
* connection, and the queue is empty), avoiding linkage into
* and removal from the queue and repetition of various
* conversions.
* Set DELACK for segments received in order, but ack
* immediately when segments are out of order (so
* fast retransmit can work).
Inline TCP_REASS() in the single location where it's used, just as OpenBSD and NetBSD has done. No functional difference. MFC after: 2 weeks
2001-05-29 19:54:45 +00:00
*/
Revert r265338, r271089 and r271123 as those changes do not handle non-inline urgent data and introduce an mbuf exhaustion attack vector similar to FreeBSD-SA-15:15.tcp, but not requiring VNETs. Address the issue described in FreeBSD-SA-15:15.tcp. Reviewed by: glebius Approved by: so Approved by: jmallett (mentor) Security: FreeBSD-SA-15:15.tcp Sponsored by: Norse Corp, Inc.
2015-07-29 17:59:13 +00:00
if (th->th_seq == tp->rcv_nxt &&
LIST_EMPTY(&tp->t_segq) &&
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
(TCPS_HAVEESTABLISHED(tp->t_state) ||
tfo_syn)) {
if (DELAY_ACK(tp, tlen) || tfo_syn)
Clean up delayed acks and T/TCP interactions: - delay acks for T/TCP regardless of delack setting - fix bug where a single pass through tcp_input might not delay acks - use callout_active() instead of callout_pending() Sponsored by: DARPA, NAI Labs
2003-02-19 21:18:23 +00:00
tp->t_flags |= TF_DELACK;
Inline TCP_REASS() in the single location where it's used, just as OpenBSD and NetBSD has done. No functional difference. MFC after: 2 weeks
2001-05-29 19:54:45 +00:00
else
tp->t_flags |= TF_ACKNOW;
tp->rcv_nxt += tlen;
thflags = th->th_flags & TH_FIN;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rcvpack);
TCPSTAT_ADD(tcps_rcvbyte, tlen);
Reduce the number of unnecessary unlock-relocks on socket buffer mutexes associated with performing a wakeup on the socket buffer: - When performing an sbappend*() followed by a so[rw]wakeup(), explicitly acquire the socket buffer lock and use the _locked() variants of both calls. Note that the _locked() sowakeup() versions unlock the mutex on return. This is done in uipc_send(), divert_packet(), mroute socket_send(), raw_append(), tcp_reass(), tcp_input(), and udp_append(). - When the socket buffer lock is dropped before a sowakeup(), remove the explicit unlock and use the _locked() sowakeup() variant. This is done in soisdisconnecting(), soisdisconnected() when setting the can't send/ receive flags and dropping data, and in uipc_rcvd() which adjusting back-pressure on the sockets. For UNIX domain sockets running mpsafe with a contention-intensive SMP mysql benchmark, this results in a 1.6% query rate improvement due to reduce mutex costs.
2004-06-26 19:10:39 +00:00
SOCKBUF_LOCK(&so->so_rcv);
The socket field so_state is used to hold a variety of socket related flags relating to several aspects of socket functionality. This change breaks out several bits relating to send and receive operation into a new per-socket buffer field, sb_state, in order to facilitate locking. This is required because, in order to provide more granular locking of sockets, different state fields have different locking properties. The following fields are moved to sb_state: SS_CANTRCVMORE (so_state) SS_CANTSENDMORE (so_state) SS_RCVATMARK (so_state) Rename respectively to: SBS_CANTRCVMORE (so_rcv.sb_state) SBS_CANTSENDMORE (so_snd.sb_state) SBS_RCVATMARK (so_rcv.sb_state) This facilitates locking by isolating fields to be located with other identically locked fields, and permits greater granularity in socket locking by avoiding storing fields with different locking semantics in the same short (avoiding locking conflicts). In the future, we may wish to coallesce sb_state and sb_flags; for the time being I leave them separate and there is no additional memory overhead due to the packing/alignment of shorts in the socket buffer structure.
2004-06-14 18:16:22 +00:00
if (so->so_rcv.sb_state & SBS_CANTRCVMORE)
Fix issue where shutdown(socket, SHUT_RD) was effectively ignored for TCP sockets. NetBSD PR: 18185 Submitted by: Sean Boudreau <seanb@qnx.com> MFC after: 3 days
2002-09-22 02:54:07 +00:00
m_freem(m);
else
Merge from projects/sendfile: extend protocols API to support sending not ready data: o Add new flag to pru_send() flags - PRUS_NOTREADY. o Add new protocol method pru_ready(). Sponsored by: Nginx, Inc. Sponsored by: Netflix
2014-11-30 13:24:21 +00:00
sbappendstream_locked(&so->so_rcv, m, 0);
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
/* NB: sorwakeup_locked() does an implicit unlock. */
Reduce the number of unnecessary unlock-relocks on socket buffer mutexes associated with performing a wakeup on the socket buffer: - When performing an sbappend*() followed by a so[rw]wakeup(), explicitly acquire the socket buffer lock and use the _locked() variants of both calls. Note that the _locked() sowakeup() versions unlock the mutex on return. This is done in uipc_send(), divert_packet(), mroute socket_send(), raw_append(), tcp_reass(), tcp_input(), and udp_append(). - When the socket buffer lock is dropped before a sowakeup(), remove the explicit unlock and use the _locked() sowakeup() variant. This is done in soisdisconnecting(), soisdisconnected() when setting the can't send/ receive flags and dropping data, and in uipc_rcvd() which adjusting back-pressure on the sockets. For UNIX domain sockets running mpsafe with a contention-intensive SMP mysql benchmark, this results in a 1.6% query rate improvement due to reduce mutex costs.
2004-06-26 19:10:39 +00:00
sorwakeup_locked(so);
Inline TCP_REASS() in the single location where it's used, just as OpenBSD and NetBSD has done. No functional difference. MFC after: 2 weeks
2001-05-29 19:54:45 +00:00
} else {
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
/*
* XXX: Due to the header drop above "th" is
* theoretically invalid by now. Fortunately
* m_adj() doesn't actually frees any mbufs
* when trimming from the head.
*/
Inline TCP_REASS() in the single location where it's used, just as OpenBSD and NetBSD has done. No functional difference. MFC after: 2 weeks
2001-05-29 19:54:45 +00:00
thflags = tcp_reass(tp, th, &tlen, m);
tp->t_flags |= TF_ACKNOW;
}
Use existing TF_SACK_PERMIT flag in struct tcpcb t_flags field instead of a decdicated sack_enable int for this bool. Change all users accordingly.
2007-05-06 15:56:31 +00:00
if (tlen > 0 && (tp->t_flags & TF_SACK_PERMIT))
Fix a case in tcp_do_segment() where tcp_update_sack_list() would be called with an incorrect segment end value. tcp_reass() may trim segments when they overlap with already existing ones in the reassembly queue. Instead of saving the segment end value before the call to tcp_reass() compute it on the fly based on the effective segment length afterwards. This bug was not really problematic as no information got lost and the eventual SACK information computation was correct nontheless. MFC after: 1 week
2007-06-10 21:07:21 +00:00
tcp_update_sack_list(tp, save_start, save_start + tlen);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
#if 0
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Note the amount of data that peer has sent into
* our window, in order to estimate the sender's
* buffer size.
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
* XXX: Unused.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Handle a rare edge case with nearly full TCP receive buffers. If a TCP buffer fills up causing the remote sender to enter into persist mode, but there is still room available in the receive buffer when a window probe arrives (either due to window scaling, or due to the local application very slowing draining data from the receive buffer), then the single byte of data in the window probe is accepted. However, this can cause rcv_nxt to be greater than rcv_adv. This condition will only last until the next ACK packet is pushed out via tcp_output(), and since the previous ACK advertised a zero window, the ACK should be pushed out while the TCP pcb is write-locked. During the window while rcv_nxt is greather than rcv_adv, a few places would compute the remaining receive window via rcv_adv - rcv_nxt. However, this value was then (uint32_t)-1. On a 64 bit machine this could expand to a positive 2^32 - 1 when cast to a long. In particular, when calculating the receive window in tcp_output(), the result would be that the receive window was computed as 2^32 - 1 resulting in advertising a far larger window to the remote peer than actually existed. Fix various places that compute the remaining receive window to either assert that it is not negative (i.e. rcv_nxt <= rcv_adv), or treat the window as full if rcv_nxt is greather than rcv_adv. Reviewed by: bz MFC after: 1 month
2011-05-02 21:05:52 +00:00
if (SEQ_GT(tp->rcv_adv, tp->rcv_nxt))
len = so->so_rcv.sb_hiwat - (tp->rcv_adv - tp->rcv_nxt);
else
len = so->so_rcv.sb_hiwat;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
} else {
m_freem(m);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
thflags &= ~TH_FIN;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
/*
* If FIN is received ACK the FIN and let the user know
* that the connection is closing.
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (thflags & TH_FIN) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (TCPS_HAVERCVDFIN(tp->t_state) == 0) {
socantrcvmore(so);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
* If connection is half-synchronized
* (ie NEEDSYN flag on) then delay ACK,
* so it may be piggybacked when SYN is sent.
* Otherwise, since we received a FIN then no
* more input can be expected, send ACK now.
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
*/
Clean up delayed acks and T/TCP interactions: - delay acks for T/TCP regardless of delack setting - fix bug where a single pass through tcp_input might not delay acks - use callout_active() instead of callout_pending() Sponsored by: DARPA, NAI Labs
2003-02-19 21:18:23 +00:00
if (tp->t_flags & TF_NEEDSYN)
tp->t_flags |= TF_DELACK;
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
else
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tp->t_flags |= TF_ACKNOW;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->rcv_nxt++;
}
switch (tp->t_state) {
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
/*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* In SYN_RECEIVED and ESTABLISHED STATES
* enter the CLOSE_WAIT state.
*/
case TCPS_SYN_RECEIVED:
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
tp->t_starttime = ticks;
Fix comments in tcp_do_segment(). Approved by: re (kensmith)
2007-07-25 18:48:24 +00:00
/* FALLTHROUGH */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case TCPS_ESTABLISHED:
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_CLOSE_WAIT);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
/*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* If still in FIN_WAIT_1 STATE FIN has not been acked so
* enter the CLOSING state.
*/
case TCPS_FIN_WAIT_1:
Implement the ip, tcp, and udp DTrace providers. The probe definitions use dynamic translation so that their arguments match the definitions for these providers in Solaris and illumos. Thus, existing scripts for these providers should work unmodified on FreeBSD. Tested by: gnn, hiren MFC after: 1 month
2013-08-25 21:54:41 +00:00
tcp_state_change(tp, TCPS_CLOSING);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
/*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* In FIN_WAIT_2 state enter the TIME_WAIT state,
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
* starting the time-wait timer, turning off the other
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* standard timers.
*/
case TCPS_FIN_WAIT_2:
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
KASSERT(ti_locked == TI_RLOCKED, ("%s: dodata "
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
"TCP_FIN_WAIT_2 ti_locked: %d", __func__,
ti_locked));
Add a TCP TIMEWAIT state which uses less space than a fullblown TCP control block. Allow the socket and tcpcb structures to be freed earlier than inpcb. Update code to understand an inp w/o a socket. Reviewed by: hsu, silby, jayanth Sponsored by: DARPA, NAI Labs
2003-02-19 22:32:43 +00:00
tcp_twstart(tp);
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
INP_INFO_RUNLOCK(&V_tcbinfo);
Remove the defunct remains of the TCPS_TIME_WAIT cases from tcp_do_segment and change it to a void function. We use a compressed structure for TCPS_TIME_WAIT to save memory. Any late late segments arriving for such a connection is handled directly in the TW code.
2007-05-06 15:16:05 +00:00
return;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED)
INP_INFO_RUNLOCK(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
ti_locked = TI_UNLOCKED;
Made TCPDEBUG truely optional. Based on changes I made in FreeBSD 1.1.5. Fixed somebody's idea of a joke - about the first half of the lines in in_proto.c were spaced over by one space.
1994-09-15 10:36:56 +00:00
#ifdef TCPDEBUG
Back out my lats commit of locking down a socket, it conflicts with hsu's work. Requested by: hsu
2002-05-31 11:52:35 +00:00
if (so->so_options & SO_DEBUG)
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tcp_trace(TA_INPUT, ostate, tp, (void *)tcp_saveipgen,
&tcp_savetcp, 0);
Made TCPDEBUG truely optional. Based on changes I made in FreeBSD 1.1.5. Fixed somebody's idea of a joke - about the first half of the lines in in_proto.c were spaced over by one space.
1994-09-15 10:36:56 +00:00
#endif
dd DTrace probe points, translators and a corresponding script to provide the TCPDEBUG functionality with pure DTrace. Reviewed by: rwatson MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: D3530
2015-09-13 15:50:55 +00:00
TCP_PROBE3(debug__input, tp, th, mtod(m, const char *));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Return any desired output.
*/
if (needoutput || (tp->t_flags & TF_ACKNOW))
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Greatly simplify the unlocking logic by holding the TCP protocol lock until after FIN_WAIT_2 processing. Helped with debugging: Doug Barton
2003-03-13 11:46:57 +00:00
Check to see if the TF_DELACK flag is set before returning from tcp_input(). This unbreaks delack handling, while still preserving correct T/TCP behavior Tested by: maxim Sponsored by: DARPA, NAI Labs
2003-02-22 21:54:57 +00:00
check_delack:
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
KASSERT(ti_locked == TI_UNLOCKED, ("%s: check_delack ti_locked %d",
__func__, ti_locked));
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
Check to see if the TF_DELACK flag is set before returning from tcp_input(). This unbreaks delack handling, while still preserving correct T/TCP behavior Tested by: maxim Sponsored by: DARPA, NAI Labs
2003-02-22 21:54:57 +00:00
if (tp->t_flags & TF_DELACK) {
Clean up delayed acks and T/TCP interactions: - delay acks for T/TCP regardless of delack setting - fix bug where a single pass through tcp_input might not delay acks - use callout_active() instead of callout_pending() Sponsored by: DARPA, NAI Labs
2003-02-19 21:18:23 +00:00
tp->t_flags &= ~TF_DELACK;
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
tcp_timer_activate(tp, TT_DELACK, tcp_delacktime);
Clean up delayed acks and T/TCP interactions: - delay acks for T/TCP regardless of delack setting - fix bug where a single pass through tcp_input might not delay acks - use callout_active() instead of callout_pending() Sponsored by: DARPA, NAI Labs
2003-02-19 21:18:23 +00:00
}
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WUNLOCK(tp->t_inpcb);
Remove the defunct remains of the TCPS_TIME_WAIT cases from tcp_do_segment and change it to a void function. We use a compressed structure for TCPS_TIME_WAIT to save memory. Any late late segments arriving for such a connection is handled directly in the TW code.
2007-05-06 15:16:05 +00:00
return;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
dropafterack:
/*
* Generate an ACK dropping incoming segment if it occupies
* sequence space, where the ACK reflects our state.
Fix RST validation. PR: 7892 Submitted by: Don.Lewis@tsc.tdk.com
1998-09-11 16:04:03 +00:00
*
* We can now skip the test for the RST flag since all
* paths to this code happen after packets containing
* RST have been dropped.
*
* In the SYN-RECEIVED state, don't send an ACK unless the
* segment we received passes the SYN-RECEIVED ACK test.
* If it fails send a RST. This breaks the loop in the
* "LAND" DoS attack, and also prevents an ACK storm
* between two listening ports that have been sent forged
* SYN segments, each with the source address of the other.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (tp->t_state == TCPS_SYN_RECEIVED && (thflags & TH_ACK) &&
(SEQ_GT(tp->snd_una, th->th_ack) ||
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
SEQ_GT(th->th_ack, tp->snd_max)) ) {
rstreason = BANDLIM_RST_OPENPORT;
goto dropwithreset;
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
#ifdef TCPDEBUG
Back out my lats commit of locking down a socket, it conflicts with hsu's work. Requested by: hsu
2002-05-31 11:52:35 +00:00
if (so->so_options & SO_DEBUG)
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tcp_trace(TA_DROP, ostate, tp, (void *)tcp_saveipgen,
&tcp_savetcp, 0);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
#endif
dd DTrace probe points, translators and a corresponding script to provide the TCPDEBUG functionality with pure DTrace. Reviewed by: rwatson MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: D3530
2015-09-13 15:50:55 +00:00
TCP_PROBE3(debug__input, tp, th, mtod(m, const char *));
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED)
INP_INFO_RUNLOCK(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
ti_locked = TI_UNLOCKED;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->t_flags |= TF_ACKNOW;
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WUNLOCK(tp->t_inpcb);
Do some re-sorting of TCP pcbinfo locking and assertions: make sure to retain the pcbinfo lock until we're done using a pcb in the in-bound path, as the pcbinfo lock acts as a pseuo-reference to prevent the pcb from potentially being recycled. Clean up assertions and make sure to assert that the pcbinfo is locked at the head of code subsections where it is needed. Free the mbuf at the end of tcp_input after releasing any held locks to reduce the time the locks are held. MFC after: 3 weeks
2004-11-07 19:19:35 +00:00
m_freem(m);
Remove the defunct remains of the TCPS_TIME_WAIT cases from tcp_do_segment and change it to a void function. We use a compressed structure for TCPS_TIME_WAIT to save memory. Any late late segments arriving for such a connection is handled directly in the TW code.
2007-05-06 15:16:05 +00:00
return;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
dropwithreset:
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED)
INP_INFO_RUNLOCK(&V_tcbinfo);
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
ti_locked = TI_UNLOCKED;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
When dropping a packet and issuing a reset during TCP segment handling, unconditionally drop the tcbinfo lock (after all, we assert it lines before), but call tcp_dropwithreset() under both inpcb and inpcbinfo locks only if we pass in an tcpcb. Otherwise, if the pointer is NULL, firewall code may later recurse the global tcbinfo lock trying to look up an inpcb. This is an instance where a layering violation leads not only potentially to code reentrace and recursion, but also to lock recursion, and was revealed by the conversion to rwlocks because acquiring a read lock on an rwlock already held with a write lock is forbidden. When these locks were mutexes, they simply recursed. Reported by: Stefan Ehmann <shoesoft at gmx dot net> MFC after: 3 days
2008-09-24 11:07:03 +00:00
if (tp != NULL) {
tcp_dropwithreset(m, th, tp, tlen, rstreason);
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WUNLOCK(tp->t_inpcb);
In both dropwithreset paths in tcp_input.c, drop the tcbinfo lock sooner to decomplicate locking and eliminate the need for a rather chatty comment about why we have to handle the global lock in a special way for the benefit of ipfw and pf cred rules. MFC after: 3 days
2008-10-26 22:03:52 +00:00
} else
When dropping a packet and issuing a reset during TCP segment handling, unconditionally drop the tcbinfo lock (after all, we assert it lines before), but call tcp_dropwithreset() under both inpcb and inpcbinfo locks only if we pass in an tcpcb. Otherwise, if the pointer is NULL, firewall code may later recurse the global tcbinfo lock trying to look up an inpcb. This is an instance where a layering violation leads not only potentially to code reentrace and recursion, but also to lock recursion, and was revealed by the conversion to rwlocks because acquiring a read lock on an rwlock already held with a write lock is forbidden. When these locks were mutexes, they simply recursed. Reported by: Stefan Ehmann <shoesoft at gmx dot net> MFC after: 3 days
2008-09-24 11:07:03 +00:00
tcp_dropwithreset(m, th, NULL, tlen, rstreason);
Remove the defunct remains of the TCPS_TIME_WAIT cases from tcp_do_segment and change it to a void function. We use a compressed structure for TCPS_TIME_WAIT to save memory. Any late late segments arriving for such a connection is handled directly in the TW code.
2007-05-06 15:16:05 +00:00
return;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
drop:
Decompose TCP INP_INFO lock to increase short-lived TCP connections scalability: - The existing TCP INP_INFO lock continues to protect the global inpcb list stability during full list traversal (e.g. tcp_pcblist()). - A new INP_LIST lock protects inpcb list actual modifications (inp allocation and free) and inpcb global counters. It allows to use TCP INP_INFO_RLOCK lock in critical paths (e.g. tcp_input()) and INP_INFO_WLOCK only in occasional operations that walk all connections. PR: 183659 Differential Revision: https://reviews.freebsd.org/D2599 Reviewed by: jhb, adrian Tested by: adrian, nitroboost-gmail.com Sponsored by: Verisign, Inc.
2015-08-03 12:13:54 +00:00
if (ti_locked == TI_RLOCKED) {
INP_INFO_RUNLOCK(&V_tcbinfo);
Decompose the current single inpcbinfo lock into two locks: - The existing ipi_lock continues to protect the global inpcb list and inpcb counter. This lock is now relegated to a small number of allocation and free operations, and occasional operations that walk all connections (including, awkwardly, certain UDP multicast receive operations -- something to revisit). - A new ipi_hash_lock protects the two inpcbinfo hash tables for looking up connections and bound sockets, manipulated using new INP_HASH_*() macros. This lock, combined with inpcb locks, protects the 4-tuple address space. Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb connection locks, so may be acquired while manipulating a connection on which a lock is already held, avoiding the need to acquire the inpcbinfo lock preemptively when a binding change might later be required. As a result, however, lookup operations necessarily go through a reference acquire while holding the lookup lock, later acquiring an inpcb lock -- if required. A new function in_pcblookup() looks up connections, and accepts flags indicating how to return the inpcb. Due to lock order changes, callers no longer need acquire locks before performing a lookup: the lookup routine will acquire the ipi_hash_lock as needed. In the future, it will also be able to use alternative lookup and locking strategies transparently to callers, such as pcbgroup lookup. New lookup flags are, supplementing the existing INPLOOKUP_WILDCARD flag: INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb Callers must pass exactly one of these flags (for the time being). Some notes: - All protocols are updated to work within the new regime; especially, TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely eliminated, and global hash lock hold times are dramatically reduced compared to previous locking. - The TCP syncache still relies on the pcbinfo lock, something that we may want to revisit. - Support for reverting to the FreeBSD 7.x locking strategy in TCP input is no longer available -- hash lookup locks are now held only very briefly during inpcb lookup, rather than for potentially extended periods. However, the pcbinfo ipi_lock will still be acquired if a connection state might change such that a connection is added or removed. - Raw IP sockets continue to use the pcbinfo ipi_lock for protection, due to maintaining their own hash tables. - The interface in6_pcblookup_hash_locked() is maintained, which allows callers to acquire hash locks and perform one or more lookups atomically with 4-tuple allocation: this is required only for TCPv6, as there is no in6_pcbconnect_setup(), which there should be. - UDPv6 locking remains significantly more conservative than UDPv4 locking, which relates to source address selection. This needs attention, as it likely significantly reduces parallelism in this code for multithreaded socket use (such as in BIND). - In the UDPv4 and UDPv6 multicast cases, we need to revisit locking somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which is no longer sufficient. A second check once the inpcb lock is held should do the trick, keeping the general case from requiring the inpcb lock for every inpcb visited. - This work reminds us that we need to revisit locking of the v4/v6 flags, which may be accessed lock-free both before and after this change. - Right now, a single lock name is used for the pcbhash lock -- this is undesirable, and probably another argument is required to take care of this (or a char array name field in the pcbinfo?). This is not an MFC candidate for 8.x due to its impact on lookup and locking semantics. It's possible some of these issues could be worked around with compatibility wrappers, if necessary. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-05-30 09:43:55 +00:00
ti_locked = TI_UNLOCKED;
}
Move from solely write-locking the global tcbinfo in tcp_input() to read-locking in the TCP input path, allowing greater TCP input parallelism where multiple ithreads or ithread and netisr are able to run in parallel. Previously, most TCP input paths held a write lock on the global tcbinfo lock, effectively serializing TCP input. Before looking up the connection, acquire a write lock if a potentially state-changing flag is set on the TCP segment header (FIN, RST, SYN), and otherwise a read lock. We may later have to upgrade to a write lock in certain cases (ACKs received by the syncache or during TIMEWAIT) in order to support global state transitions, but this is never required for steady-state packets. Upgrading from a write lock to a read lock must be done as a trylock operation to avoid deadlocks, and actually violates the lock order as the tcbinfo lock preceeds the inpcb lock held at the time of upgrade. If the trylock fails, we bump the refcount on the inpcb, drop both locks, and re-acquire in-order. If another thread has freed the connection while the locks are dropped, we free the inpcb and repeat the lookup (this should hardly ever or never happen in practice). For now, maintain a number of new counters measuring how many times various cases execute, and in particular whether various optimistic assumptions about when read locks can be used, whether upgrades are done using the fast path, and whether connections close in practice in the above-described race, actually occur. MFC after: 6 weeks Discussed with: kmacy Reviewed by: bz, gnn, kmacy Tested by: kmacy
2008-12-08 20:27:00 +00:00
#ifdef INVARIANTS
else
INP_INFO_UNLOCK_ASSERT(&V_tcbinfo);
#endif
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
/*
* Drop space held by incoming segment and return.
*/
#ifdef TCPDEBUG
if (tp == NULL || (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
tcp_trace(TA_DROP, ostate, tp, (void *)tcp_saveipgen,
&tcp_savetcp, 0);
#endif
dd DTrace probe points, translators and a corresponding script to provide the TCPDEBUG functionality with pure DTrace. Reviewed by: rwatson MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: D3530
2015-09-13 15:50:55 +00:00
TCP_PROBE3(debug__input, tp, th, mtod(m, const char *));
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
if (tp != NULL)
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WUNLOCK(tp->t_inpcb);
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
m_freem(m);
}
/*
o Remove redundant tcp reassembly check in header prediction code o Rearrange code to make intent in TCPS_SYN_SENT case more clear o Assorted style cleanup o Comment clarification for tcp_dropwithreset()
2007-05-06 15:41:06 +00:00
* Issue RST and make ACK acceptable to originator of segment.
* The mbuf must still include the original packet header.
* tp may be NULL.
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
tcp_dropwithreset(struct mbuf *m, struct tcphdr *th, struct tcpcb *tp,
int tlen, int rstreason)
{
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#ifdef INET
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
struct ip *ip;
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
#ifdef INET6
struct ip6_hdr *ip6;
#endif
Add further TCP inpcb locking assertions to some TCP input code paths. MFC after: 1 month
2008-04-07 12:41:45 +00:00
if (tp != NULL) {
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Add further TCP inpcb locking assertions to some TCP input code paths. MFC after: 1 month
2008-04-07 12:41:45 +00:00
}
o Remove redundant tcp reassembly check in header prediction code o Rearrange code to make intent in TCPS_SYN_SENT case more clear o Assorted style cleanup o Comment clarification for tcp_dropwithreset()
2007-05-06 15:41:06 +00:00
/* Don't bother if destination was broadcast/multicast. */
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
if ((th->th_flags & TH_RST) || m->m_flags & (M_BCAST|M_MCAST))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto drop;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
#ifdef INET6
if (mtod(m, struct ip *)->ip_v == 6) {
ip6 = mtod(m, struct ip6_hdr *);
Mitigate the stream.c attacks o Drop all broadcast and multicast source addresses in tcp_input. o Enable ICMP_BANDLIM in GENERIC. o Change default to 200/s from 100/s. This will still stop the attack, but is conservative enough to do this close to code freeze. This is not the optimal patch for the problem, but is likely the least intrusive patch that can be made for this. Obtained from: Don Lewis and Matt Dillon. Reviewed by: freebsd-security
2000-01-28 06:13:09 +00:00
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src))
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
goto drop;
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
/* IPv6 anycast check is done at tcp6_input() */
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
}
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#if defined(INET) && defined(INET6)
else
#endif
#ifdef INET
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
{
ip = mtod(m, struct ip *);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
ip->ip_src.s_addr == htonl(INADDR_BROADCAST) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif))
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
goto drop;
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
/* Perform bandwidth limiting. */
Clean up RST ratelimiting. Previously, ratelimiting occured before tests were performed to determine if the received packet should be reset. This created erroneous ratelimiting and false alarms in some cases. The code has now been reorganized so that the checks for validity come before the call to badport_bandlim. Additionally, a few changes in the symbolic names of the bandlim types have been made, as well as a clarification of exactly which type each RST case falls under. Submitted by: Mike Silbersack <silby@silby.com>
2001-02-11 07:39:51 +00:00
if (badport_bandlim(rstreason) < 0)
goto drop;
White space cleanup for netinet before branch: - Trailing tab/space cleanup - Remove spurious spaces between or before tabs This change avoids touching files that Andre likely has in his working set for PFIL hooks changes for IPFW/DUMMYNET. Approved by: re (scottl) Submitted by: Xin LI <delphij@frontfree.net>
2004-08-16 18:32:07 +00:00
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
/* tcp_respond consumes the mbuf chain. */
if (th->th_flags & TH_ACK) {
tcp_respond(tp, mtod(m, void *), th, m, (tcp_seq)0,
th->th_ack, TH_RST);
} else {
if (th->th_flags & TH_SYN)
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
tlen++;
tcp_respond(tp, mtod(m, void *), th, m, th->th_seq+tlen,
Split tcp_input() into its two functional parts: o tcp_input() now handles TCP segment sanity checks and preparations including the INPCB lookup and syncache. o tcp_do_segment() handles all data and ACK processing and is IPv4/v6 agnostic. Change all KASSERT() messages to ("%s: ", __func__). The changes in this commit are primarily of mechanical nature and no functional changes besides the function split are made. Discussed with: rwatson
2007-03-23 20:16:50 +00:00
(tcp_seq)0, TH_RST|TH_ACK);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
return;
drop:
Do some re-sorting of TCP pcbinfo locking and assertions: make sure to retain the pcbinfo lock until we're done using a pcb in the in-bound path, as the pcbinfo lock acts as a pseuo-reference to prevent the pcb from potentially being recycled. Clean up assertions and make sure to assert that the pcbinfo is locked at the head of code subsections where it is needed. Free the mbuf at the end of tcp_input after releasing any held locks to reduce the time the locks are held. MFC after: 3 weeks
2004-11-07 19:19:35 +00:00
m_freem(m);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
/*
* Parse TCP options and place in tcpopt.
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
tcp_dooptions(struct tcpopt *to, u_char *cp, int cnt, int flags)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
int opt, optlen;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
to->to_flags = 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
for (; cnt > 0; cnt -= optlen, cp += optlen) {
opt = cp[0];
if (opt == TCPOPT_EOL)
break;
if (opt == TCPOPT_NOP)
optlen = 1;
else {
be more cautious about tcp option length field. drop bogus ones earlier. not sure if there is a real threat or not, but it seems that there's possibility for overrun/underrun (like non-NOP option with optlen > cnt).
2000-07-09 13:01:59 +00:00
if (cnt < 2)
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
optlen = cp[1];
be more cautious about tcp option length field. drop bogus ones earlier. not sure if there is a real threat or not, but it seems that there's possibility for overrun/underrun (like non-NOP option with optlen > cnt).
2000-07-09 13:01:59 +00:00
if (optlen < 2 || optlen > cnt)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
}
switch (opt) {
case TCPOPT_MAXSEG:
if (optlen != TCPOLEN_MAXSEG)
continue;
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
if (!(flags & TO_SYN))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
continue;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
to->to_flags |= TOF_MSS;
bcopy((char *)cp + 2,
(char *)&to->to_mss, sizeof(to->to_mss));
o Move NTOHL() and associated macros into <sys/param.h>. These are deprecated in favor of the POSIX-defined lowercase variants. o Change all occurrences of NTOHL() and associated marcros in the source tree to use the lowercase function variants. o Add missing license bits to sparc64's <machine/endian.h>. Approved by: jake o Clean up <machine/endian.h> files. o Remove unused __uint16_swap_uint32() from i386's <machine/endian.h>. o Remove prototypes for non-existent bswapXX() functions. o Include <machine/endian.h> in <arpa/inet.h> to define the POSIX-required ntohl() family of functions. o Do similar things to expose the ntohl() family in libstand, <netinet/in.h>, and <sys/param.h>. o Prepend underscores to the ntohl() family to help deal with complexities associated with having MD (asm and inline) versions, and having to prevent exposure of these functions in other headers that happen to make use of endian-specific defines. o Create weak aliases to the canonical function name to help deal with third-party software forgetting to include an appropriate header. o Remove some now unneeded pollution from <sys/types.h>. o Add missing <arpa/inet.h> includes in userland. Tested on: alpha, i386 Reviewed by: bde, jake, tmm
2002-02-18 20:35:27 +00:00
to->to_mss = ntohs(to->to_mss);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case TCPOPT_WINDOW:
if (optlen != TCPOLEN_WINDOW)
continue;
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
if (!(flags & TO_SYN))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
continue;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
to->to_flags |= TOF_SCALE;
Consolidate insertion of TCP options into a segment from within tcp_output() and syncache_respond() into its own generic function tcp_addoptions(). tcp_addoptions() is alignment agnostic and does optimal packing in all cases. In struct tcpopt rename to_requested_s_scale to just to_wscale. Add a comment with quote from RFC1323: "The Window field in a SYN (i.e., a <SYN> or <SYN,ACK>) segment itself is never scaled." Reviewed by: silby, mohans, julian Sponsored by: TCP/IP Optimization Fundraise 2005
2007-03-15 15:59:28 +00:00
to->to_wscale = min(cp[2], TCP_MAX_WINSHIFT);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case TCPOPT_TIMESTAMP:
if (optlen != TCPOLEN_TIMESTAMP)
continue;
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
to->to_flags |= TOF_TS;
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
bcopy((char *)cp + 2,
(char *)&to->to_tsval, sizeof(to->to_tsval));
o Move NTOHL() and associated macros into <sys/param.h>. These are deprecated in favor of the POSIX-defined lowercase variants. o Change all occurrences of NTOHL() and associated marcros in the source tree to use the lowercase function variants. o Add missing license bits to sparc64's <machine/endian.h>. Approved by: jake o Clean up <machine/endian.h> files. o Remove unused __uint16_swap_uint32() from i386's <machine/endian.h>. o Remove prototypes for non-existent bswapXX() functions. o Include <machine/endian.h> in <arpa/inet.h> to define the POSIX-required ntohl() family of functions. o Do similar things to expose the ntohl() family in libstand, <netinet/in.h>, and <sys/param.h>. o Prepend underscores to the ntohl() family to help deal with complexities associated with having MD (asm and inline) versions, and having to prevent exposure of these functions in other headers that happen to make use of endian-specific defines. o Create weak aliases to the canonical function name to help deal with third-party software forgetting to include an appropriate header. o Remove some now unneeded pollution from <sys/types.h>. o Add missing <arpa/inet.h> includes in userland. Tested on: alpha, i386 Reviewed by: bde, jake, tmm
2002-02-18 20:35:27 +00:00
to->to_tsval = ntohl(to->to_tsval);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
bcopy((char *)cp + 6,
(char *)&to->to_tsecr, sizeof(to->to_tsecr));
o Move NTOHL() and associated macros into <sys/param.h>. These are deprecated in favor of the POSIX-defined lowercase variants. o Change all occurrences of NTOHL() and associated marcros in the source tree to use the lowercase function variants. o Add missing license bits to sparc64's <machine/endian.h>. Approved by: jake o Clean up <machine/endian.h> files. o Remove unused __uint16_swap_uint32() from i386's <machine/endian.h>. o Remove prototypes for non-existent bswapXX() functions. o Include <machine/endian.h> in <arpa/inet.h> to define the POSIX-required ntohl() family of functions. o Do similar things to expose the ntohl() family in libstand, <netinet/in.h>, and <sys/param.h>. o Prepend underscores to the ntohl() family to help deal with complexities associated with having MD (asm and inline) versions, and having to prevent exposure of these functions in other headers that happen to make use of endian-specific defines. o Create weak aliases to the canonical function name to help deal with third-party software forgetting to include an appropriate header. o Remove some now unneeded pollution from <sys/types.h>. o Add missing <arpa/inet.h> includes in userland. Tested on: alpha, i386 Reviewed by: bde, jake, tmm
2002-02-18 20:35:27 +00:00
to->to_tsecr = ntohl(to->to_tsecr);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
Initial import of RFC 2385 (TCP-MD5) digest support. This is the first of two commits; bringing in the kernel support first. This can be enabled by compiling a kernel with options TCP_SIGNATURE and FAST_IPSEC. For the uninitiated, this is a TCP option which provides for a means of authenticating TCP sessions which came into being before IPSEC. It is still relevant today, however, as it is used by many commercial router vendors, particularly with BGP, and as such has become a requirement for interconnect at many major Internet points of presence. Several parts of the TCP and IP headers, including the segment payload, are digested with MD5, including a shared secret. The PF_KEY interface is used to manage the secrets using security associations in the SADB. There is a limitation here in that as there is no way to map a TCP flow per-port back to an SPI without polluting tcpcb or using the SPD; the code to do the latter is unstable at this time. Therefore this code only supports per-host keying granularity. Whilst FAST_IPSEC is mutually exclusive with KAME IPSEC (and thus IPv6), TCP_SIGNATURE applies only to IPv4. For the vast majority of prospective users of this feature, this will not pose any problem. This implementation is output-only; that is, the option is honoured when responding to a host initiating a TCP session, but no effort is made [yet] to authenticate inbound traffic. This is, however, sufficient to interwork with Cisco equipment. Tested with a Cisco 2501 running IOS 12.0(27), and Quagga 0.96.4 with local patches. Patches for tcpdump to validate TCP-MD5 sessions are also available from me upon request. Sponsored by: sentex.net
2004-02-11 04:26:04 +00:00
#ifdef TCP_SIGNATURE
/*
* XXX In order to reply to a host which has set the
* TCP_SIGNATURE option in its initial SYN, we have to
* record the fact that the option was observed here
* for the syncache code to perform the correct response.
*/
case TCPOPT_SIGNATURE:
if (optlen != TCPOLEN_SIGNATURE)
continue;
o Remove unncessary TOF_SIGLEN flag from struct tcpopt o Correctly set to->to_signature in tcp_dooptions() o Update comments
2007-04-20 15:28:01 +00:00
to->to_flags |= TOF_SIGNATURE;
to->to_signature = cp + 2;
Initial import of RFC 2385 (TCP-MD5) digest support. This is the first of two commits; bringing in the kernel support first. This can be enabled by compiling a kernel with options TCP_SIGNATURE and FAST_IPSEC. For the uninitiated, this is a TCP option which provides for a means of authenticating TCP sessions which came into being before IPSEC. It is still relevant today, however, as it is used by many commercial router vendors, particularly with BGP, and as such has become a requirement for interconnect at many major Internet points of presence. Several parts of the TCP and IP headers, including the segment payload, are digested with MD5, including a shared secret. The PF_KEY interface is used to manage the secrets using security associations in the SADB. There is a limitation here in that as there is no way to map a TCP flow per-port back to an SPI without polluting tcpcb or using the SPD; the code to do the latter is unstable at this time. Therefore this code only supports per-host keying granularity. Whilst FAST_IPSEC is mutually exclusive with KAME IPSEC (and thus IPv6), TCP_SIGNATURE applies only to IPv4. For the vast majority of prospective users of this feature, this will not pose any problem. This implementation is output-only; that is, the option is honoured when responding to a host initiating a TCP session, but no effort is made [yet] to authenticate inbound traffic. This is, however, sufficient to interwork with Cisco equipment. Tested with a Cisco 2501 running IOS 12.0(27), and Quagga 0.96.4 with local patches. Patches for tcpdump to validate TCP-MD5 sessions are also available from me upon request. Sponsored by: sentex.net
2004-02-11 04:26:04 +00:00
break;
Brucification. Submitted by: bde
2004-02-13 18:21:45 +00:00
#endif
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
case TCPOPT_SACK_PERMITTED:
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
if (optlen != TCPOLEN_SACK_PERMITTED)
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
continue;
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
if (!(flags & TO_SYN))
continue;
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if (!V_tcp_do_sack)
Some cleanups and janitorial work to tcp_dooptions(): o redefine the parameter 'is_syn' to 'flags', add TO_SYN flag and adjust its usage accordingly o update the comments to the tcp_dooptions() invocation in tcp_input():after_listen to reflect reality o move the logic checking the echoed timestamp out of tcp_dooptions() to the only place that uses it next to the invocation described in the previous item o adjust parsing of TCPOPT_SACK_PERMITTED to use the same style as the others o add comments in to struct tcpopt.to_flags #defines No functional changes. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 15:35:25 +00:00
continue;
Bring SACK option handling in tcp_dooptions() in line with all other options and ajust users accordingly.
2007-03-23 18:33:21 +00:00
to->to_flags |= TOF_SACKPERM;
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
break;
case TCPOPT_SACK:
- Postpone SACK option processing until after PAWS checks. SACK option processing is now done in the ACK processing case. - Merge tcp_sack_option() and tcp_del_sackholes() into a new function called tcp_sack_doack(). - Test (SEG.ACK < SND.MAX) before processing the ACK. Submitted by: Noritoshi Demizu Reveiewed by: Mohan Srinivasan, Raja Mukerji Approved by: re
2005-06-27 22:27:42 +00:00
if (optlen <= 2 || (optlen - 2) % TCPOLEN_SACK != 0)
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
continue;
In tcp_dooptions() skip over SACK options if it is a SYN segment.
2007-04-04 14:39:49 +00:00
if (flags & TO_SYN)
continue;
Bring SACK option handling in tcp_dooptions() in line with all other options and ajust users accordingly.
2007-03-23 18:33:21 +00:00
to->to_flags |= TOF_SACK;
- Postpone SACK option processing until after PAWS checks. SACK option processing is now done in the ACK processing case. - Merge tcp_sack_option() and tcp_del_sackholes() into a new function called tcp_sack_doack(). - Test (SEG.ACK < SND.MAX) before processing the ACK. Submitted by: Noritoshi Demizu Reveiewed by: Mohan Srinivasan, Raja Mukerji Approved by: re
2005-06-27 22:27:42 +00:00
to->to_nsacks = (optlen - 2) / TCPOLEN_SACK;
to->to_sacks = cp + 2;
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_sack_rcv_blocks);
Add support for TCP Selective Acknowledgements. The work for this originated on RELENG_4 and was ported to -CURRENT. The scoreboarding code was obtained from OpenBSD, and many of the remaining changes were inspired by OpenBSD, but not taken directly from there. You can enable/disable sack using net.inet.tcp.do_sack. You can also limit the number of sack holes that all senders can have in the scoreboard with net.inet.tcp.sackhole_limit. Reviewed by: gnn Obtained from: Yahoo! (Mohan Srinivasan, Jayanth Vijayaraghavan)
2004-06-23 21:04:37 +00:00
break;
Implementation of server-side TCP Fast Open (TFO) [RFC7413]. TFO is disabled by default in the kernel build. See the top comment in sys/netinet/tcp_fastopen.c for implementation particulars. Reviewed by: gnn, jch, stas MFC after: 3 days Sponsored by: Verisign, Inc. Differential Revision: https://reviews.freebsd.org/D4350
2015-12-24 19:09:48 +00:00
#ifdef TCP_RFC7413
case TCPOPT_FAST_OPEN:
if ((optlen != TCPOLEN_FAST_OPEN_EMPTY) &&
(optlen < TCPOLEN_FAST_OPEN_MIN) &&
(optlen > TCPOLEN_FAST_OPEN_MAX))
continue;
if (!(flags & TO_SYN))
continue;
if (!V_tcp_fastopen_enabled)
continue;
to->to_flags |= TOF_FASTOPEN;
to->to_tfo_len = optlen - 2;
to->to_tfo_cookie = to->to_tfo_len ? cp + 2 : NULL;
break;
#endif
Introduce a syncache, which enables FreeBSD to withstand a SYN flood DoS in an improved fashion over the existing code. Reviewed by: silby (in a previous iteration) Sponsored by: DARPA, NAI Labs
2001-11-22 04:50:44 +00:00
default:
continue;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
}
/*
* Pull out of band byte out of a segment so
* it doesn't appear in the user's data queue.
* It is still reflected in the segment length for
* sequencing purposes.
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
tcp_pulloutofband(struct socket *so, struct tcphdr *th, struct mbuf *m,
int off)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
int cnt = off + th->th_urp - 1;
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
while (cnt >= 0) {
if (m->m_len > cnt) {
char *cp = mtod(m, caddr_t) + cnt;
struct tcpcb *tp = sototcpcb(so);
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Add further TCP inpcb locking assertions to some TCP input code paths. MFC after: 1 month
2008-04-07 12:41:45 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->t_iobc = *cp;
tp->t_oobflags |= TCPOOB_HAVEDATA;
bcopy(cp+1, cp, (unsigned)(m->m_len - cnt - 1));
m->m_len--;
Avoid m_len and m_pkthdr.len inconsistency when changing m_len for an mbuf whose M_PKTHDR is set. PR: related to kern/15175 Reviewed by: archie
2000-01-25 01:26:47 +00:00
if (m->m_flags & M_PKTHDR)
m->m_pkthdr.len--;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return;
}
cnt -= m->m_len;
m = m->m_next;
Tidy up some code to conform better to surroundings and style(9), 0 = NULL and space/tab.
2007-03-23 19:11:22 +00:00
if (m == NULL)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
}
panic("tcp_pulloutofband");
}
/*
* Collect new round-trip time estimate
* and update averages and current timeout.
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
tcp_xmit_timer(struct tcpcb *tp, int rtt)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
int delta;
A number of performance-reducing flaws fixed based on comments from Larry Peterson &co. at Arizona: - Header prediction for ACKs did not exclude Fast Retransmit/Recovery. - srtt calculation tended to get ``stuck'' and could never decrease when below 8. It still can't, but the scaling factors are adjusted so that this artifact does not cause as bad an effect on the RTO value as it used to. The paper also points out the incr/8 error that has been long since fixed, and the problems with ACKing frequency resulting from the use of options which I suspect to be fixed already as well (as part of the T/TCP work). Obtained from: Brakmo & Peterson, ``Performance Problems in BSD4.4 TCP''
1996-03-22 18:09:21 +00:00
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Assert the inpcb lock in tcp_xmit_timer() as it performs read-modify- write of various time/rtt-related fields in the tcpcb.
2004-11-28 11:06:22 +00:00
Update stats in struct tcpstat using two new macros, TCPSTAT_ADD() and TCPSTAT_INC(), rather than directly manipulating the fields across the kernel. This will make it easier to change the implementation of these statistics, such as using per-CPU versions of the data structures. MFC after: 3 days
2009-04-11 22:07:19 +00:00
TCPSTAT_INC(tcps_rttupdated);
A number of performance-reducing flaws fixed based on comments from Larry Peterson &co. at Arizona: - Header prediction for ACKs did not exclude Fast Retransmit/Recovery. - srtt calculation tended to get ``stuck'' and could never decrease when below 8. It still can't, but the scaling factors are adjusted so that this artifact does not cause as bad an effect on the RTO value as it used to. The paper also points out the incr/8 error that has been long since fixed, and the problems with ACKing frequency resulting from the use of options which I suspect to be fixed already as well (as part of the T/TCP work). Obtained from: Brakmo & Peterson, ``Performance Problems in BSD4.4 TCP''
1996-03-22 18:09:21 +00:00
tp->t_rttupdated++;
if (tp->t_srtt != 0) {
/*
* srtt is stored as fixed point with 5 bits after the
* binary point (i.e., scaled by 8). The following magic
* is equivalent to the smoothing algorithm in rfc793 with
* an alpha of .875 (srtt = rtt/8 + srtt*7/8 in fixed
* point). Adjust rtt to origin 0.
*/
delta = ((rtt - 1) << TCP_DELTA_SHIFT)
- (tp->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT));
if ((tp->t_srtt += delta) <= 0)
tp->t_srtt = 1;
/*
* We accumulate a smoothed rtt variance (actually, a
* smoothed mean difference), then set the retransmit
* timer to smoothed rtt + 4 times the smoothed variance.
* rttvar is stored as fixed point with 4 bits after the
* binary point (scaled by 16). The following is
* equivalent to rfc793 smoothing with an alpha of .75
* (rttvar = rttvar*3/4 + |delta| / 4). This replaces
* rfc793's wired-in beta.
*/
if (delta < 0)
delta = -delta;
delta -= tp->t_rttvar >> (TCP_RTTVAR_SHIFT - TCP_DELTA_SHIFT);
if ((tp->t_rttvar += delta) <= 0)
tp->t_rttvar = 1;
Implement TCP bandwidth delay product window limiting, similar to (but not meant to duplicate) TCP/Vegas. Add four sysctls and default the implementation to 'off'. net.inet.tcp.inflight_enable enable algorithm (defaults to 0=off) net.inet.tcp.inflight_debug debugging (defaults to 1=on) net.inet.tcp.inflight_min minimum window limit net.inet.tcp.inflight_max maximum window limit MFC after: 1 week
2002-08-17 18:26:02 +00:00
if (tp->t_rttbest > tp->t_srtt + tp->t_rttvar)
tp->t_rttbest = tp->t_srtt + tp->t_rttvar;
A number of performance-reducing flaws fixed based on comments from Larry Peterson &co. at Arizona: - Header prediction for ACKs did not exclude Fast Retransmit/Recovery. - srtt calculation tended to get ``stuck'' and could never decrease when below 8. It still can't, but the scaling factors are adjusted so that this artifact does not cause as bad an effect on the RTO value as it used to. The paper also points out the incr/8 error that has been long since fixed, and the problems with ACKing frequency resulting from the use of options which I suspect to be fixed already as well (as part of the T/TCP work). Obtained from: Brakmo & Peterson, ``Performance Problems in BSD4.4 TCP''
1996-03-22 18:09:21 +00:00
} else {
/*
* No rtt measurement yet - use the unsmoothed rtt.
* Set the variance to half the rtt (so our first
* retransmit happens at 3*rtt).
*/
tp->t_srtt = rtt << TCP_RTT_SHIFT;
tp->t_rttvar = rtt << (TCP_RTTVAR_SHIFT - 1);
Implement TCP bandwidth delay product window limiting, similar to (but not meant to duplicate) TCP/Vegas. Add four sysctls and default the implementation to 'off'. net.inet.tcp.inflight_enable enable algorithm (defaults to 0=off) net.inet.tcp.inflight_debug debugging (defaults to 1=on) net.inet.tcp.inflight_min minimum window limit net.inet.tcp.inflight_max maximum window limit MFC after: 1 week
2002-08-17 18:26:02 +00:00
tp->t_rttbest = tp->t_srtt + tp->t_rttvar;
A number of performance-reducing flaws fixed based on comments from Larry Peterson &co. at Arizona: - Header prediction for ACKs did not exclude Fast Retransmit/Recovery. - srtt calculation tended to get ``stuck'' and could never decrease when below 8. It still can't, but the scaling factors are adjusted so that this artifact does not cause as bad an effect on the RTO value as it used to. The paper also points out the incr/8 error that has been long since fixed, and the problems with ACKing frequency resulting from the use of options which I suspect to be fixed already as well (as part of the T/TCP work). Obtained from: Brakmo & Peterson, ``Performance Problems in BSD4.4 TCP''
1996-03-22 18:09:21 +00:00
}
Restructure TCP timeout handling: - eliminate the fast/slow timeout lists for TCP and instead use a callout entry for each timer. - increase the TCP timer granularity to HZ - implement "bad retransmit" recovery, as presented in "On Estimating End-to-End Network Path Properties", by Allman and Paxson. Submitted by: jlemon, wollmann
1999-08-30 21:17:07 +00:00
tp->t_rtttime = 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
tp->t_rxtshift = 0;
/*
* the retransmit should happen at rtt + 4 * rttvar.
* Because of the way we do the smoothing, srtt and rttvar
* will each average +1/2 tick of bias. When we compute
* the retransmit timer, we want 1/2 tick of rounding and
* 1 extra tick because of +-1/2 tick uncertainty in the
* firing of the timer. The bias will give us exactly the
* 1.5 tick we need. But, because the bias is
* statistical, we have to test that we don't drop below
* the minimum feasible timer (which is 2 ticks).
*/
A number of performance-reducing flaws fixed based on comments from Larry Peterson &co. at Arizona: - Header prediction for ACKs did not exclude Fast Retransmit/Recovery. - srtt calculation tended to get ``stuck'' and could never decrease when below 8. It still can't, but the scaling factors are adjusted so that this artifact does not cause as bad an effect on the RTO value as it used to. The paper also points out the incr/8 error that has been long since fixed, and the problems with ACKing frequency resulting from the use of options which I suspect to be fixed already as well (as part of the T/TCP work). Obtained from: Brakmo & Peterson, ``Performance Problems in BSD4.4 TCP''
1996-03-22 18:09:21 +00:00
TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp),
Slight modification of RTO floor calculation.
1996-03-25 20:13:21 +00:00
max(tp->t_rttmin, rtt + 2), TCPTV_REXMTMAX);
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* We received an ack for a packet that wasn't retransmitted;
* it is probably safe to discard any error indications we've
* received recently. This isn't quite right, but close enough
* for now (a route might have failed after we sent a segment,
* and the return path might not be symmetrical).
*/
tp->t_softerror = 0;
}
/*
* Determine a reasonable value for maxseg size.
* If the route is known, check route for mtu.
Allow arbitrary MSS sizes and don't mind about the cluster size anymore. We've got more cluster sizes for quite some time now and the orginally imposed limits and the previously codified thoughts on efficiency gains are no longer true. MFC after: 2 weeks
2012-10-28 18:33:52 +00:00
* If none, use an mss that can be handled on the outgoing interface
* without forcing IP to fragment. If no route is found, route has no mtu,
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* or the destination isn't local, use a default, hopefully conservative
* size (usually 512 or the default IP max size, but no more than the mtu
* of the interface), as we can't discover anything about intervening
* gateways or networks. We also initialize the congestion/slow start
* window to be a single segment if the destination isn't local.
* While looking at the routing entry, we also initialize other path-dependent
* parameters from pre-set or cached values in the routing entry.
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
*
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
* NOTE that resulting t_maxseg doesn't include space for TCP options or
* IP options, e.g. IPSEC data, since length of this data may vary, and
* thus it is calculated for every segment separately in tcp_output().
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
*
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
* NOTE that this routine is only called when we process an incoming
When we receive an ICMP unreach need fragmentation datagram, we take proposed MTU value from it and update the TCP host cache. Then tcp_mss_update() is called on the corresponding tcpcb. It finds the just allocated entry in the TCP host cache and updates MSS on the tcpcb. And then we do a fast retransmit of what we have in the tcp send buffer. This sequence gets broken if the TCP host cache is exausted. In this case allocation fails, and later called tcp_mss_update() finds nothing in cache. The fast retransmit is done with not reduced MSS and is immidiately replied by remote host with new ICMP datagrams and the cycle repeats. This ping-pong can go up to wirespeed. To fix this: - tcp_mss_update() gets new parameter - mtuoffer, that is like offer, but needs to have min_protoh subtracted. - tcp_mtudisc() as notification method renamed to tcp_mtudisc_notify(). - tcp_mtudisc() now accepts not a useless error argument, but proposed MTU value, that is passed to tcp_mss_update() as mtuoffer. Reported by: az Reported by: Andrey Zonov <andrey zonov.org> Reviewed by: andre (previous version of patch)
2012-04-16 13:49:03 +00:00
* segment, or an ICMP need fragmentation datagram. Outgoing SYN/ACK MSS
* settings are handled in tcp_mssopt().
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
void
When we receive an ICMP unreach need fragmentation datagram, we take proposed MTU value from it and update the TCP host cache. Then tcp_mss_update() is called on the corresponding tcpcb. It finds the just allocated entry in the TCP host cache and updates MSS on the tcpcb. And then we do a fast retransmit of what we have in the tcp send buffer. This sequence gets broken if the TCP host cache is exausted. In this case allocation fails, and later called tcp_mss_update() finds nothing in cache. The fast retransmit is done with not reduced MSS and is immidiately replied by remote host with new ICMP datagrams and the cycle repeats. This ping-pong can go up to wirespeed. To fix this: - tcp_mss_update() gets new parameter - mtuoffer, that is like offer, but needs to have min_protoh subtracted. - tcp_mtudisc() as notification method renamed to tcp_mtudisc_notify(). - tcp_mtudisc() now accepts not a useless error argument, but proposed MTU value, that is passed to tcp_mss_update() as mtuoffer. Reported by: az Reported by: Andrey Zonov <andrey zonov.org> Reviewed by: andre (previous version of patch)
2012-04-16 13:49:03 +00:00
tcp_mss_update(struct tcpcb *tp, int offer, int mtuoffer,
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
struct hc_metrics_lite *metricptr, struct tcp_ifcap *cap)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
int mss = 0;
u_long maxmtu = 0;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
struct inpcb *inp = tp->t_inpcb;
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
struct hc_metrics_lite metrics;
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#ifdef INET6
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
int isipv6 = ((inp->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
size_t min_protoh = isipv6 ?
sizeof (struct ip6_hdr) + sizeof (struct tcphdr) :
sizeof (struct tcpiphdr);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#else
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
const size_t min_protoh = sizeof(struct tcpiphdr);
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
#endif
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Add further TCP inpcb locking assertions to some TCP input code paths. MFC after: 1 month
2008-04-07 12:41:45 +00:00
When we receive an ICMP unreach need fragmentation datagram, we take proposed MTU value from it and update the TCP host cache. Then tcp_mss_update() is called on the corresponding tcpcb. It finds the just allocated entry in the TCP host cache and updates MSS on the tcpcb. And then we do a fast retransmit of what we have in the tcp send buffer. This sequence gets broken if the TCP host cache is exausted. In this case allocation fails, and later called tcp_mss_update() finds nothing in cache. The fast retransmit is done with not reduced MSS and is immidiately replied by remote host with new ICMP datagrams and the cycle repeats. This ping-pong can go up to wirespeed. To fix this: - tcp_mss_update() gets new parameter - mtuoffer, that is like offer, but needs to have min_protoh subtracted. - tcp_mtudisc() as notification method renamed to tcp_mtudisc_notify(). - tcp_mtudisc() now accepts not a useless error argument, but proposed MTU value, that is passed to tcp_mss_update() as mtuoffer. Reported by: az Reported by: Andrey Zonov <andrey zonov.org> Reviewed by: andre (previous version of patch)
2012-04-16 13:49:03 +00:00
if (mtuoffer != -1) {
KASSERT(offer == -1, ("%s: conflict", __func__));
offer = mtuoffer - min_protoh;
}
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
/* Initialize. */
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
#ifdef INET6
if (isipv6) {
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
maxmtu = tcp_maxmtu6(&inp->inp_inc, cap);
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->t_maxseg = V_tcp_v6mssdflt;
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
}
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#if defined(INET) && defined(INET6)
else
#endif
#ifdef INET
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
{
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
maxmtu = tcp_maxmtu(&inp->inp_inc, cap);
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->t_maxseg = V_tcp_mssdflt;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
/*
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
* No route to sender, stay with default mss and return.
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
*/
Fix a bug introduced with r182851 splitting tcp_mss() into tcp_mss() and tcp_mss_update() so that tcp_mtudisc() could re-use the same code. In case we return early and got a metricptr to pass the hostcache info back to the caller we need to initialize the data to a defined state (zero it) as tcp_hc_get() would do if there was no hit. Without that the caller would check on random stack garbage which could lead to undefined results. This only affected tcp_mss() if there was no routing entry for the peer, tcp_mtudisc() was not affected. MFC after: 2 months (along with r182851)
2008-11-06 12:33:33 +00:00
if (maxmtu == 0) {
/*
Fix typo and while here another one. Reviewed by: keramida Reported by: keramida MFC after: 2 months (with r184720)
2008-11-06 16:30:20 +00:00
* In case we return early we need to initialize metrics
Fix a bug introduced with r182851 splitting tcp_mss() into tcp_mss() and tcp_mss_update() so that tcp_mtudisc() could re-use the same code. In case we return early and got a metricptr to pass the hostcache info back to the caller we need to initialize the data to a defined state (zero it) as tcp_hc_get() would do if there was no hit. Without that the caller would check on random stack garbage which could lead to undefined results. This only affected tcp_mss() if there was no routing entry for the peer, tcp_mtudisc() was not affected. MFC after: 2 months (along with r182851)
2008-11-06 12:33:33 +00:00
* to a defined state as tcp_hc_get() would do for us
* if there was no cache hit.
*/
if (metricptr != NULL)
bzero(metricptr, sizeof(struct hc_metrics_lite));
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
return;
Fix a bug introduced with r182851 splitting tcp_mss() into tcp_mss() and tcp_mss_update() so that tcp_mtudisc() could re-use the same code. In case we return early and got a metricptr to pass the hostcache info back to the caller we need to initialize the data to a defined state (zero it) as tcp_hc_get() would do if there was no hit. Without that the caller would check on random stack garbage which could lead to undefined results. This only affected tcp_mss() if there was no routing entry for the peer, tcp_mtudisc() was not affected. MFC after: 2 months (along with r182851)
2008-11-06 12:33:33 +00:00
}
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
/* What have we got? */
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
switch (offer) {
case 0:
/*
* Offer == 0 means that there was no MSS on the SYN
Some "cleanup" of tcp_mss(): - Move the assigment of the socket down before we first need it. No need to do it at the beginning and then drop out the function by one of the returns before using it 100 lines further down. - Use t_maxopd which was assigned the "tcp_mssdflt" for the corrrect AF already instead of another #ifdef ? : #endif block doing the same. - Remove an unneeded (duplicate) assignment of mss to t_maxseg just before we possibly change mss and re-do the assignment without using t_maxseg in between. Reviewed by: silby No objections: net@ (silence) MFC after: 5 days
2008-03-02 08:40:47 +00:00
* segment, in this case we use tcp_mssdflt as
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
* already assigned to t_maxseg above.
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
*/
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
offer = tp->t_maxseg;
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
break;
case -1:
/*
Remove RFC1644 T/TCP support from the TCP side of the network stack. A complete rationale and discussion is given in this message and the resulting discussion: http://docs.freebsd.org/cgi/mid.cgi?4177C8AD.6060706 Note that this commit removes only the functional part of T/TCP from the tcp_* related functions in the kernel. Other features introduced with RFC1644 are left intact (socket layer changes, sendmsg(2) on connection oriented protocols) and are meant to be reused by a simpler and less intrusive reimplemention of the previous T/TCP functionality. Discussed on: -arch
2004-11-02 22:22:22 +00:00
* Offer == -1 means that we didn't receive SYN yet.
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
*/
/* FALLTHROUGH */
default:
Limiters and sanity checks for TCP MSS (maximum segement size) resource exhaustion attacks. For network link optimization TCP can adjust its MSS and thus packet size according to the observed path MTU. This is done dynamically based on feedback from the remote host and network components along the packet path. This information can be abused to pretend an extremely low path MTU. The resource exhaustion works in two ways: o during tcp connection setup the advertized local MSS is exchanged between the endpoints. The remote endpoint can set this arbitrarily low (except for a minimum MTU of 64 octets enforced in the BSD code). When the local host is sending data it is forced to send many small IP packets instead of a large one. For example instead of the normal TCP payload size of 1448 it forces TCP payload size of 12 (MTU 64) and thus we have a 120 times increase in workload and packets. On fast links this quickly saturates the local CPU and may also hit pps processing limites of network components along the path. This type of attack is particularly effective for servers where the attacker can download large files (WWW and FTP). We mitigate it by enforcing a minimum MTU settable by sysctl net.inet.tcp.minmss defaulting to 256 octets. o the local host is reveiving data on a TCP connection from the remote host. The local host has no control over the packet size the remote host is sending. The remote host may chose to do what is described in the first attack and send the data in packets with an TCP payload of at least one byte. For each packet the tcp_input() function will be entered, the packet is processed and a sowakeup() is signalled to the connected process. For example an attack with 2 Mbit/s gives 4716 packets per second and the same amount of sowakeup()s to the process (and context switches). This type of attack is particularly effective for servers where the attacker can upload large amounts of data. Normally this is the case with WWW server where large POSTs can be made. We mitigate this by calculating the average MSS payload per second. If it goes below 'net.inet.tcp.minmss' and the pps rate is above 'net.inet.tcp.minmssoverload' defaulting to 1000 this particular TCP connection is resetted and dropped. MITRE CVE: CAN-2004-0002 Reviewed by: sam (mentor) MFC after: 1 day
2004-01-08 17:40:07 +00:00
/*
* Prevent DoS attack with too small MSS. Round up
* to at least minmss.
*/
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
offer = max(offer, V_tcp_minmss);
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Fix style for comments, be more verbose and add some more.
2007-06-10 20:59:22 +00:00
* rmx information is now retrieved from tcp_hostcache.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
tcp_hc_get(&inp->inp_inc, &metrics);
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
if (metricptr != NULL)
bcopy(&metrics, metricptr, sizeof(struct hc_metrics_lite));
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
2014-07-02 22:04:14 +00:00
* If there's a discovered mtu in tcp hostcache, use it.
* Else, use the link mtu.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
if (metrics.rmx_mtu)
Correct an edge case in tcp_mss() where the cached path MTU from tcp_hostcache would have overridden a (now) lower MTU of an interface or route that changed since first PMTU discovery. The bug would have caused TCP to redo the PMTU discovery when not strictly necessary. Make a comment about already pre-initialized default values more clear. Reviewed by: sam
2004-04-23 22:44:59 +00:00
mss = min(metrics.rmx_mtu, maxmtu) - min_protoh;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
else {
enclose IPv6 part with ifdef INET6. Obtained from: KAME
2003-10-20 16:19:01 +00:00
#ifdef INET6
tcp updates to support IPv6. also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change. Reviewed by: freebsd-arch, cvs-committers Obtained from: KAME project
2000-01-09 19:17:30 +00:00
if (isipv6) {
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
mss = maxmtu - min_protoh;
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if (!V_path_mtu_discovery &&
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
!in6_localaddr(&inp->in6p_faddr))
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
mss = min(mss, V_tcp_v6mssdflt);
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
}
enclose IPv6 part with ifdef INET6. Obtained from: KAME
2003-10-20 16:19:01 +00:00
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#if defined(INET) && defined(INET6)
else
#endif
#ifdef INET
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
{
mss = maxmtu - min_protoh;
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if (!V_path_mtu_discovery &&
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
!in_localaddr(inp->inp_faddr))
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
mss = min(mss, V_tcp_mssdflt);
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
/*
* XXX - The above conditional (mss = maxmtu - min_protoh)
* probably violates the TCP spec.
* The problem is that, since we don't know the
* other end's MSS, we are supposed to use a conservative
* default. But, if we do that, then MTU discovery will
* never actually take place, because the conservative
* default is much less than the MTUs typically seen
* on the Internet today. For the moment, we'll sweep
* this under the carpet.
*
* The conservative default might not actually be a problem
* if the only case this occurs is when sending an initial
* SYN with options and data to a host we've never talked
* to before. Then, they will reply with an MSS value which
* will get recorded and the new parameters should get
* recomputed. For Further Study.
*/
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
}
mss = min(mss, offer);
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
/*
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
* Sanity check: make sure that maxseg will be large
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
* enough to allow some data on segments even if the
* all the option space is used (40bytes). Otherwise
* funny things may happen in tcp_output.
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
*
* XXXGL: shouldn't we reserve space for IP/IPv6 options?
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
*/
mss = max(mss, 64);
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
tp->t_maxseg = mss;
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
}
void
tcp_mss(struct tcpcb *tp, int offer)
{
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
int mss;
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
u_long bufsize;
struct inpcb *inp;
struct socket *so;
struct hc_metrics_lite metrics;
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
struct tcp_ifcap cap;
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
KASSERT(tp != NULL, ("%s: tp == NULL", __func__));
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
bzero(&cap, sizeof(cap));
tcp_mss_update(tp, offer, -1, &metrics, &cap);
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former calls the latter. Merge tcp_mss_update() with code from tcp_mtudisc() basically doing the same thing. This gives us one central place where we calcuate and check mss values to update t_maxopd (maximum mss + options length) instead of two slightly different but almost equal implementations to maintain. PR: kern/118455 Reviewed by: silby (back in March) MFC after: 2 months
2008-09-07 18:50:25 +00:00
mss = tp->t_maxseg;
inp = tp->t_inpcb;
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
* If there's a pipesize, change the socket buffer to that size,
* don't change if sb_hiwat is different than default (then it
* has been changed on purpose with setsockopt).
* Make the socket buffers an integral number of mss units;
* if the mss is larger than the socket buffer, decrease the mss.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Some "cleanup" of tcp_mss(): - Move the assigment of the socket down before we first need it. No need to do it at the beginning and then drop out the function by one of the returns before using it 100 lines further down. - Use t_maxopd which was assigned the "tcp_mssdflt" for the corrrect AF already instead of another #ifdef ? : #endif block doing the same. - Remove an unneeded (duplicate) assignment of mss to t_maxseg just before we possibly change mss and re-do the assignment without using t_maxseg in between. Reviewed by: silby No objections: net@ (silence) MFC after: 5 days
2008-03-02 08:40:47 +00:00
so = inp->inp_socket;
Introduce sbreserve_locked(), which asserts the socket buffer lock on the socket buffer having its limits adjusted. sbreserve() now acquires the lock before calling sbreserve_locked(). In soreserve(), acquire socket buffer locks across read-modify-writes of socket buffer fields, and calls into sbreserve/sbrelease; make sure to acquire in keeping with the socket buffer lock order. In tcp_mss(), acquire the socket buffer lock in the calling context so that we have atomic read-modify -write on buffer sizes.
2004-06-24 01:37:04 +00:00
SOCKBUF_LOCK(&so->so_snd);
VNET virtualize tcp_sendspace/tcp_recvspace and change the type to INT. A long is not necessary as the TCP window is limited to 2**30. A larger initial window isn't useful. MFC after: 1 week
2011-10-16 15:08:43 +00:00
if ((so->so_snd.sb_hiwat == V_tcp_sendspace) && metrics.rmx_sendpipe)
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
bufsize = metrics.rmx_sendpipe;
else
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
bufsize = so->so_snd.sb_hiwat;
if (bufsize < mss)
mss = bufsize;
else {
bufsize = roundup(bufsize, mss);
if (bufsize > sb_max)
bufsize = sb_max;
Don't shrink socket buffers in tcp_mss(), application might have already configured them with setsockopt(SO_*BUF), for RFC1323's scaled windows. PR: kern/11966 MFC after: 1 week
2002-07-22 22:31:09 +00:00
if (bufsize > so->so_snd.sb_hiwat)
Introduce sbreserve_locked(), which asserts the socket buffer lock on the socket buffer having its limits adjusted. sbreserve() now acquires the lock before calling sbreserve_locked(). In soreserve(), acquire socket buffer locks across read-modify-writes of socket buffer fields, and calls into sbreserve/sbrelease; make sure to acquire in keeping with the socket buffer lock order. In tcp_mss(), acquire the socket buffer lock in the calling context so that we have atomic read-modify -write on buffer sizes.
2004-06-24 01:37:04 +00:00
(void)sbreserve_locked(&so->so_snd, bufsize, so, NULL);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
}
Introduce sbreserve_locked(), which asserts the socket buffer lock on the socket buffer having its limits adjusted. sbreserve() now acquires the lock before calling sbreserve_locked(). In soreserve(), acquire socket buffer locks across read-modify-writes of socket buffer fields, and calls into sbreserve/sbrelease; make sure to acquire in keeping with the socket buffer lock order. In tcp_mss(), acquire the socket buffer lock in the calling context so that we have atomic read-modify -write on buffer sizes.
2004-06-24 01:37:04 +00:00
SOCKBUF_UNLOCK(&so->so_snd);
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
tp->t_maxseg = mss;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Introduce sbreserve_locked(), which asserts the socket buffer lock on the socket buffer having its limits adjusted. sbreserve() now acquires the lock before calling sbreserve_locked(). In soreserve(), acquire socket buffer locks across read-modify-writes of socket buffer fields, and calls into sbreserve/sbrelease; make sure to acquire in keeping with the socket buffer lock order. In tcp_mss(), acquire the socket buffer lock in the calling context so that we have atomic read-modify -write on buffer sizes.
2004-06-24 01:37:04 +00:00
SOCKBUF_LOCK(&so->so_rcv);
VNET virtualize tcp_sendspace/tcp_recvspace and change the type to INT. A long is not necessary as the TCP window is limited to 2**30. A larger initial window isn't useful. MFC after: 1 week
2011-10-16 15:08:43 +00:00
if ((so->so_rcv.sb_hiwat == V_tcp_recvspace) && metrics.rmx_recvpipe)
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
bufsize = metrics.rmx_recvpipe;
else
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
bufsize = so->so_rcv.sb_hiwat;
if (bufsize > mss) {
bufsize = roundup(bufsize, mss);
if (bufsize > sb_max)
bufsize = sb_max;
Don't shrink socket buffers in tcp_mss(), application might have already configured them with setsockopt(SO_*BUF), for RFC1323's scaled windows. PR: kern/11966 MFC after: 1 week
2002-07-22 22:31:09 +00:00
if (bufsize > so->so_rcv.sb_hiwat)
Introduce sbreserve_locked(), which asserts the socket buffer lock on the socket buffer having its limits adjusted. sbreserve() now acquires the lock before calling sbreserve_locked(). In soreserve(), acquire socket buffer locks across read-modify-writes of socket buffer fields, and calls into sbreserve/sbrelease; make sure to acquire in keeping with the socket buffer lock order. In tcp_mss(), acquire the socket buffer lock in the calling context so that we have atomic read-modify -write on buffer sizes.
2004-06-24 01:37:04 +00:00
(void)sbreserve_locked(&so->so_rcv, bufsize, so, NULL);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Introduce sbreserve_locked(), which asserts the socket buffer lock on the socket buffer having its limits adjusted. sbreserve() now acquires the lock before calling sbreserve_locked(). In soreserve(), acquire socket buffer locks across read-modify-writes of socket buffer fields, and calls into sbreserve/sbrelease; make sure to acquire in keeping with the socket buffer lock order. In tcp_mss(), acquire the socket buffer lock in the calling context so that we have atomic read-modify -write on buffer sizes.
2004-06-24 01:37:04 +00:00
SOCKBUF_UNLOCK(&so->so_rcv);
Fix a bug introduced with r182851 splitting tcp_mss() into tcp_mss() and tcp_mss_update() so that tcp_mtudisc() could re-use the same code. Move the TSO logic back to tcp_mss() and out of tcp_mss_update(). We tried to avoid that initially but if were are called from tcp_output() with EMSGSIZE, we cleared the TSO flag on the tcpcb there, called into tcp_mtudisc() and tcp_mss_update() which then would reenable TSO on the tcpcb based on TSO capabilities of the interface as learnt in tcp_maxmtu/6(). So if TSO was enabled on the (possibly new) outgoing interface it was turned back on, which lead to an endless loop between tcp_output() and tcp_mtudisc() until we overflew the stack. Reported by: kmacy MFC after: 2 months (along with r182851)
2008-11-06 13:25:59 +00:00
/* Check the interface for TSO capabilities. */
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
if (cap.ifcap & CSUM_TSO) {
Fix a bug introduced with r182851 splitting tcp_mss() into tcp_mss() and tcp_mss_update() so that tcp_mtudisc() could re-use the same code. Move the TSO logic back to tcp_mss() and out of tcp_mss_update(). We tried to avoid that initially but if were are called from tcp_output() with EMSGSIZE, we cleared the TSO flag on the tcpcb there, called into tcp_mtudisc() and tcp_mss_update() which then would reenable TSO on the tcpcb based on TSO capabilities of the interface as learnt in tcp_maxmtu/6(). So if TSO was enabled on the (possibly new) outgoing interface it was turned back on, which lead to an endless loop between tcp_output() and tcp_mtudisc() until we overflew the stack. Reported by: kmacy MFC after: 2 months (along with r182851)
2008-11-06 13:25:59 +00:00
tp->t_flags |= TF_TSO;
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
tp->t_tsomax = cap.tsomax;
Improve transmit sending offload, TSO, algorithm in general. The current TSO limitation feature only takes the total number of bytes in an mbuf chain into account and does not limit by the number of mbufs in a chain. Some kinds of hardware is limited by two factors. One is the fragment length and the second is the fragment count. Both of these limits need to be taken into account when doing TSO. Else some kinds of hardware might have to drop completely valid mbuf chains because they cannot loaded into the given hardware's DMA engine. The new way of doing TSO limitation has been made backwards compatible as input from other FreeBSD developers and will use defaults for values not set. Reviewed by: adrian, rmacklem Sponsored by: Mellanox Technologies MFC after: 1 week
2014-09-22 08:27:27 +00:00
tp->t_tsomaxsegcount = cap.tsomaxsegcount;
tp->t_tsomaxsegsize = cap.tsomaxsegsize;
Allow drivers to specify a maximum TSO length in bytes if they are limited in the amount of data they can handle at once. Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to change the limit. The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything less wouldn't be very useful anymore. The upper limit is still at IP_MAXPACKET (65536 bytes). Raising it requires further auditing of the IPv4/v6 code path's as the length field in the IP header would overflow leading to confusion in firewalls and others packet handler on the real size of the packet. The placement into "struct ifnet" is a bit hackish but the best place that was found. When the stack/driver boundary is updated it should be handled in a better way. Submitted by: cperciva (earlier version) Reviewed by: cperciva Tested by: cperciva MFC after: 1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
}
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
}
/*
* Determine the MSS option to send on an outgoing SYN.
*/
int
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
tcp_mssopt(struct in_conninfo *inc)
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
{
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
int mss = 0;
u_long maxmtu = 0;
u_long thcmtu = 0;
size_t min_protoh;
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
KASSERT(inc != NULL, ("tcp_mssopt with NULL in_conninfo pointer"));
Merge Transaction TCP, courtesy of Andras Olah <olah@cs.utwente.nl> and Bob Braden <braden@isi.edu>. NB: This has not had David's TCP ACK hack re-integrated. It is not clear what the correct solution to this problem is, if any. If a better solution doesn't pop up in response to this message, I'll put David's code back in (or he's welcome to do so himself).
1995-02-09 23:13:27 +00:00
correct linkmtu handling. Obtained from: KAME
2003-10-20 15:27:48 +00:00
#ifdef INET6
Use inc_flags instead of the inc_isipv6 alias which so far had been the only flag with random usage patterns. Switch inc_flags to be used as a real bit field by using INC_ISIPV6 with bitops to check for the 'isipv6' condition. While here fix a place or two where in case of v4 inc_flags were not properly initialized before.[1] Found by: rwatson during review [1] Discussed with: rwatson Reviewed by: rwatson MFC after: 4 weeks
2008-12-17 12:52:34 +00:00
if (inc->inc_flags & INC_ISIPV6) {
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
mss = V_tcp_v6mssdflt;
First step of TSO (TCP segmentation offload) support in our network stack. o add IFCAP_TSO[46] for drivers to announce this capability for IPv4 and IPv6 o add CSUM_TSO flag to mbuf pkthdr csum_flags field o add tso_segsz field to mbuf pkthdr o enhance ip_output() packet length check to allow for large TSO packets o extend tcp_maxmtu[46]() with a flag pointer to pass interface capabilities o adjust all callers of tcp_maxmtu[46]() accordingly Discussed on: -current, -net Sponsored by: TCP/IP Optimization Fundraise 2005
2006-09-06 21:51:59 +00:00
maxmtu = tcp_maxmtu6(inc, NULL);
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
min_protoh = sizeof(struct ip6_hdr) + sizeof(struct tcphdr);
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
}
#endif
#if defined(INET) && defined(INET6)
else
correct linkmtu handling. Obtained from: KAME
2003-10-20 15:27:48 +00:00
#endif
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#ifdef INET
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
{
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
mss = V_tcp_mssdflt;
First step of TSO (TCP segmentation offload) support in our network stack. o add IFCAP_TSO[46] for drivers to announce this capability for IPv4 and IPv6 o add CSUM_TSO flag to mbuf pkthdr csum_flags field o add tso_segsz field to mbuf pkthdr o enhance ip_output() packet length check to allow for large TSO packets o extend tcp_maxmtu[46]() with a flag pointer to pass interface capabilities o adjust all callers of tcp_maxmtu[46]() accordingly Discussed on: -current, -net Sponsored by: TCP/IP Optimization Fundraise 2005
2006-09-06 21:51:59 +00:00
maxmtu = tcp_maxmtu(inc, NULL);
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
min_protoh = sizeof(struct tcpiphdr);
}
Make the TCP code compile without INET. Sort #includes and add #ifdef INETs. Add some comments at #endifs given more nestedness. To make the compiler happy, some default initializations were added in accordance with the style on the files. Reviewed by: gnn Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems MFC after: 4 days
2011-04-30 11:21:29 +00:00
#endif
MFp4 bz_ipv6_fast: Factor out the tcp_hc_getmtu() call. As the comments say it applies to both v4 and v6, so only write it once making it easier to read the protocol family specifc code. Sponsored by: The FreeBSD Foundation Sponsored by: iXsystems Reviewed by: gnn (as part of the whole) MFC After: 3 days
2012-05-25 01:13:39 +00:00
#if defined(INET6) || defined(INET)
thcmtu = tcp_hc_getmtu(inc); /* IPv4 and IPv6 */
#endif
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
if (maxmtu && thcmtu)
mss = min(maxmtu, thcmtu) - min_protoh;
else if (maxmtu || thcmtu)
mss = max(maxmtu, thcmtu) - min_protoh;
return (mss);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Implement TCP NewReno, as documented in RFC 2582. This allows better recovery for multiple packet losses in a single window. The algorithm can be toggled via the sysctl net.inet.tcp.newreno, which defaults to "on". Submitted by: Jayanth Vijayaraghavan <jayanth@yahoo-inc.com>
2000-05-06 03:31:09 +00:00
/*
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
* On a partial ack arrives, force the retransmission of the
* next unacknowledged segment. Do not clear tp->t_dupacks.
* By setting snd_nxt to ti_ack, this forces retransmission timer to
* be started again.
Implement TCP NewReno, as documented in RFC 2582. This allows better recovery for multiple packet losses in a single window. The algorithm can be toggled via the sysctl net.inet.tcp.newreno, which defaults to "on". Submitted by: Jayanth Vijayaraghavan <jayanth@yahoo-inc.com>
2000-05-06 03:31:09 +00:00
*/
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
void
ANSIfy function declarations and remove register keywords for variables. Consistently apply style to all function declarations.
2007-03-21 19:37:55 +00:00
tcp_newreno_partial_ack(struct tcpcb *tp, struct tcphdr *th)
Implement TCP NewReno, as documented in RFC 2582. This allows better recovery for multiple packet losses in a single window. The algorithm can be toggled via the sysctl net.inet.tcp.newreno, which defaults to "on". Submitted by: Jayanth Vijayaraghavan <jayanth@yahoo-inc.com>
2000-05-06 03:31:09 +00:00
{
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
tcp_seq onxt = tp->snd_nxt;
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
u_long ocwnd = tp->snd_cwnd;
u_int maxseg = tcp_maxseg(tp);
Implement TCP NewReno, as documented in RFC 2582. This allows better recovery for multiple packet losses in a single window. The algorithm can be toggled via the sysctl net.inet.tcp.newreno, which defaults to "on". Submitted by: Jayanth Vijayaraghavan <jayanth@yahoo-inc.com>
2000-05-06 03:31:09 +00:00
Convert pcbinfo and inpcb mutexes to rwlocks, and modify macros to explicitly select write locking for all use of the inpcb mutex. Update some pcbinfo lock assertions to assert locked rather than write-locked, although in practice almost all uses of the pcbinfo rwlock main exclusive, and all instances of inpcb lock acquisition are exclusive. This change should introduce (ideally) little functional change. However, it lays the groundwork for significantly increased parallelism in the TCP/IP code. MFC after: 3 months Tested by: kris (superset of committered patch)
2008-04-17 21:38:18 +00:00
INP_WLOCK_ASSERT(tp->t_inpcb);
Add further TCP inpcb locking assertions to some TCP input code paths. MFC after: 1 month
2008-04-07 12:41:45 +00:00
Change the TCP timer system from using the callout system five times directly to a merged model where only one callout, the next to fire, is registered. Instead of callout_reset(9) and callout_stop(9) the new function tcp_timer_activate() is used which then internally manages the callout. The single new callout is a mutex callout on inpcb simplifying the locking a bit. tcp_timer() is the called function which handles all race conditions in one place and then dispatches the individual timer functions. Reviewed by: rwatson (earlier version)
2007-04-11 09:45:16 +00:00
tcp_timer_activate(tp, TT_REXMT, 0);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
tp->t_rtttime = 0;
tp->snd_nxt = th->th_ack;
/*
* Set snd_cwnd to one segment beyond acknowledged offset.
* (tp->snd_una has not yet been updated when this function is called.)
*/
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd = maxseg + BYTES_THIS_ACK(tp, th);
Guido found another bug. There is a situation with timestamped TCP packets where FreeBSD will send DATA+FIN and A W2K box will ack just the DATA portion. If this occurs after FreeBSD has done a (NewReno) fast-retransmit and is recovering it (dupacks > threshold) it triggers a case in tcp_newreno_partial_ack() (tcp_newreno() in stable) where tcp_output() is called with the expectation that the retransmit timer will be reloaded. But tcp_output() falls through and returns without doing anything, causing the persist timer to be loaded instead. This causes the connection to hang until W2K gives up. This occurs because in the case where only the FIN must be acked, the 'len' calculation in tcp_output() will be 0, a lot of checks will be skipped, and the FIN check will also be skipped because it is designed to handle FIN retransmits, not forced transmits from tcp_newreno(). The solution is to simply set TF_ACKNOW before calling tcp_output() to absolute guarentee that it will run the send code and reset the retransmit timer. TF_ACKNOW is already used for this purpose in other cases. For some unknown reason this patch also seems to greatly reduce the number of duplicate acks received when Guido runs his tests over a lossy network. It is quite possible that there are other tcp_newreno{_partial_ack()} cases which were not generating the expected output which this patch also fixes. X-MFC after: Will be MFC'd after the freeze is over
2002-09-30 18:55:45 +00:00
tp->t_flags |= TF_ACKNOW;
First cut of the modularization of our TCP stack. Still to do is to clean up the timer handling using the async-drain. Other optimizations may be coming to go with this. Whats here will allow differnet tcp implementations (one included). Reviewed by: jtl, hiren, transports Sponsored by: Netflix Inc. Differential Revision: D4055
2015-12-16 00:56:45 +00:00
(void) tp->t_fb->tfb_tcp_output(tp);
Cosmetic-only changes for readability. Reviewed by: (early form passed by) bde Approved by: itojun (from core@kame.net)
2002-08-17 02:05:25 +00:00
tp->snd_cwnd = ocwnd;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
/*
* Partial window deflation. Relies on fact that tp->snd_una
* not updated yet.
*/
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
if (tp->snd_cwnd > BYTES_THIS_ACK(tp, th))
tp->snd_cwnd -= BYTES_THIS_ACK(tp, th);
Fix for a bug in newreno partial ack handling where if a large amount of data is partial acked, snd_cwnd underflows, causing a burst. Found, Submitted by: Noritoshi Demizu Approved by: re
2005-07-05 19:23:02 +00:00
else
tp->snd_cwnd = 0;
Historically we have two fields in tcpcb to describe sender MSS: t_maxopd, and t_maxseg. This dualism emerged with T/TCP, but was not properly cleaned up after T/TCP removal. After all permutations over the years the result is that t_maxopd stores a minimum of peer offered MSS and MTU reduced by minimum protocol header. And t_maxseg stores (t_maxopd - TCPOLEN_TSTAMP_APPA) if timestamps are in action, or is equal to t_maxopd otherwise. That's a very rough estimate of MSS reduced by options length. Throughout the code it was used in places, where preciseness was not important, like cwnd or ssthresh calculations. With this change: - t_maxopd goes away. - t_maxseg now stores MSS not adjusted by options. - new function tcp_maxseg() is provided, that calculates MSS reduced by options length. The functions gives a better estimate, since it takes into account SACK state as well. Reviewed by: jtl Differential Revision: https://reviews.freebsd.org/D3593
2016-01-07 00:14:42 +00:00
tp->snd_cwnd += maxseg;
Implement TCP NewReno, as documented in RFC 2582. This allows better recovery for multiple packet losses in a single window. The algorithm can be toggled via the sysctl net.inet.tcp.newreno, which defaults to "on". Submitted by: Jayanth Vijayaraghavan <jayanth@yahoo-inc.com>
2000-05-06 03:31:09 +00:00
}
Calculate the correct amount of bytes that are in-flight for a connection as suggested by RFC 6675. Currently differnt places in the stack tries to guess this in suboptimal ways. The main problem is that current calculations don't take sacked bytes into account. Sacked bytes are the bytes receiver acked via SACK option. This is suboptimal because it assumes that network has more outstanding (unacked) bytes than the actual value and thus sends less data by setting congestion window lower than what's possible which in turn may cause slower recovery from losses. As an example, one of the current calculations looks something like this: snd_nxt - snd_fack + sackhint.sack_bytes_rexmit New proposal from RFC 6675 is: snd_max - snd_una - sackhint.sacked_bytes + sackhint.sack_bytes_rexmit which takes sacked bytes into account which is a new addition to the sackhint struct. Only thing we are missing from RFC 6675 is isLost() i.e. segment being considered lost and thus adjusting pipe based on that which makes this calculation a bit on conservative side. The approach is very simple. We already process each ack with sack info in tcp_sack_doack() and extract sack blocks/holes out of it. We'd now also track this new variable sacked_bytes which keeps track of total sacked bytes reported. One downside to this approach is that we may get incorrect count of sacked_bytes if the other end decides to drop sack info in the ack because of memory pressure or some other reasons. But in this (not very likely) case also the pipe calculation would be conservative which is okay as opposed to being aggressive in sending packets into the network. Next step is to use this more accurate pipe estimation to drive congestion window adjustments. In collaboration with: rrs Reviewed by: jason_eggnet dot com, rrs MFC after: 2 weeks Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D3971
2015-10-28 22:57:51 +00:00
int
tcp_compute_pipe(struct tcpcb *tp)
{
return (tp->snd_max - tp->snd_una +
tp->sackhint.sack_bytes_rexmit -
tp->sackhint.sacked_bytes);
}
Reference in New Issue Copy Permalink