d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
768a720e95
freebsd-dev/sys/netinet/tcp_syncache.c

2258 lines
66 KiB
C
Raw Normal View History

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 clause three from tcp_syncache.c license per permission of McAfee. Update copyright to McAfee from NETA.
2005-01-30 19:28:27 +00:00
* Copyright (c) 2001 McAfee, Inc.
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* Copyright (c) 2006,2013 Andre Oppermann, Internet Business Solutions AG
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
* All rights reserved.
*
* This software was developed for the FreeBSD Project by Jonathan Lemon
Remove clause three from tcp_syncache.c license per permission of McAfee. Update copyright to McAfee from NETA.
2005-01-30 19:28:27 +00:00
* and McAfee Research, the Security Research Division of McAfee, Inc. under
* DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS"), as part of the
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* DARPA CHATS research program. [2001 McAfee, Inc.]
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
*
* 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.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
*/
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"
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
#include "opt_inet6.h"
#include "opt_ipsec.h"
Implement a CPU-affine TCP and UDP connection lookup data structure, struct inpcbgroup. pcbgroups, or "connection groups", supplement the existing inpcbinfo connection hash table, which when pcbgroups are enabled, might now be thought of more usefully as a per-protocol 4-tuple reservation table. Connections are assigned to connection groups base on a hash of their 4-tuple; wildcard sockets require special handling, and are members of all connection groups. During a connection lookup, a per-connection group lock is employed rather than the global pcbinfo lock. By aligning connection groups with input path processing, connection groups take on an effective CPU affinity, especially when aligned with RSS work placement (see a forthcoming commit for details). This eliminates cache line migration associated with global, protocol-layer data structures in steady state TCP and UDP processing (with the exception of protocol-layer statistics; further commit to follow). Elements of this approach were inspired by Willman, Rixner, and Cox's 2006 USENIX paper, "An Evaluation of Network Stack Parallelization Strategies in Modern Operating Systems". However, there are also significant differences: we maintain the inpcb lock, rather than using the connection group lock for per-connection state. Likewise, the focus of this implementation is alignment with NIC packet distribution strategies such as RSS, rather than pure software strategies. Despite that focus, software distribution is supported through the parallel netisr implementation, and works well in configurations where the number of hardware threads is greater than the number of NIC input queues, such as in the RMI XLR threaded MIPS architecture. Another important difference is the continued maintenance of existing hash tables as "reservation tables" -- these are useful both to distinguish the resource allocation aspect of protocol name management and the more common-case lookup aspect. In configurations where connection tables are aligned with hardware hashes, it is desirable to use the traditional lookup tables for loopback or encapsulated traffic rather than take the expense of hardware hashes that are hard to implement efficiently in software (such as RSS Toeplitz). Connection group support is enabled by compiling "options PCBGROUP" into your kernel configuration; for the time being, this is an experimental feature, and hence is not enabled by default. Subject to the limited MFCability of change dependencies in inpcb, and its change to the inpcbinfo init function signature, this change in principle could be merged to FreeBSD 8.x. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-06-06 12:55:02 +00:00
#include "opt_pcbgroup.h"
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
#include <sys/param.h>
#include <sys/systm.h>
Use Jenkins hash for TCP syncache. o Unlike xor, in Jenkins hash every bit of input affects virtually every bit of output, thus salting the hash actually works. With xor salting only provides a false sense of security, since if hash(x) collides with hash(y), then of course, hash(x) ^ salt would also collide with hash(y) ^ salt. [1] o Jenkins provides much better distribution than xor, very close to ideal. TCP connection setup/teardown benchmark has shown a 10% increase with default hash size, and with bigger hashes that still provide possibility for collisions. With enormous hash size, when dataset is by an order of magnitude smaller than hash size, the benchmark has shown 4% decrease in performance decrease, which is expected and acceptable. Noticed by: Jeffrey Knockel <jeffk cs.unm.edu> [1] Benchmarks by: jch Reviewed by: jch, pkelsey, delphij Security: strengthens protection against hash collision DoS Sponsored by: Nginx, Inc.
2015-09-05 10:15:19 +00:00
#include <sys/hash.h>
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
#include <sys/refcount.h>
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
#include <sys/kernel.h>
#include <sys/sysctl.h>
Fix bugs in the TCP syncache timeout code. including: When system ticks are positive, for entries in the cache bucket, syncache_timer() ran on every tick (doing nothing useful) instead of the supposed 3, 6, 12, and 24 seconds later (when it's time to retransmit SYN,ACK). When ticks are negative, syncache_timer() was scheduled for the too far future (up to ~25 days on systems with HZ=1000), no SYN,ACK retransmits were attempted at all, and syncache entries added in that period that correspond to non-established connections stay there forever. Only HEAD and RELENG_7 are affected. Reviewed by: silby, kmacy (earlier version) Submitted by: Maxim Dounin, ru
2007-12-19 16:56:28 +00:00
#include <sys/limits.h>
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
#include <sys/lock.h>
#include <sys/mutex.h>
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
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/proc.h> /* for proc0 declaration */
#include <sys/random.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
#include <sys/syslog.h>
Cache the cred locally in _syncache_add() while holding the locks, so we can be sure that it's valid. In case we abort early free it again else put it into the syncache. We need the cred in the syncache to be able to restrict what will be exportet by the sysctl helper function syncache_pcblist() (to netstat) within jails. PR: kern/126493 Reviewed by: rwatson (earlier versions) MFC after: 3 days
2008-08-23 14:22:12 +00:00
#include <sys/ucred.h>
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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
#include <sys/md5.h>
#include <crypto/siphash/siphash.h>
Remove unessary includes and follow common ordering style.
2006-09-13 13:21:17 +00:00
#include <vm/uma.h>
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
#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>
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
#include <net/route.h>
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
#include <net/vnet.h>
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
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/in_var.h>
#include <netinet/in_pcb.h>
#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>
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
#ifdef INET6
#include <netinet/ip6.h>
#include <netinet/icmp6.h>
#include <netinet6/nd6.h>
#include <netinet6/ip6_var.h>
#include <netinet6/in6_pcb.h>
#endif
#include <netinet/tcp.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
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
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
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>
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
#ifdef INET6
#include <netinet6/tcp6_var.h>
#endif
- 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/toecore.h>
#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
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#include <netipsec/ipsec_support.h>
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
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
#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>
After some off-list discussion, revert a number of changes to the DPCPU_DEFINE and VNET_DEFINE macros, as these cause problems for various people working on the affected files. A better long-term solution is still being considered. This reversal may give some modules empty set_pcpu or set_vnet sections, but these are harmless. Changes reverted: ------------------------------------------------------------------------ r215318 | dim | 2010-11-14 21:40:55 +0100 (Sun, 14 Nov 2010) | 4 lines Instead of unconditionally emitting .globl's for the __start_set_xxx and __stop_set_xxx symbols, only emit them when the set_vnet or set_pcpu sections are actually defined. ------------------------------------------------------------------------ r215317 | dim | 2010-11-14 21:38:11 +0100 (Sun, 14 Nov 2010) | 3 lines Apply the STATIC_VNET_DEFINE and STATIC_DPCPU_DEFINE macros throughout the tree. ------------------------------------------------------------------------ r215316 | dim | 2010-11-14 21:23:02 +0100 (Sun, 14 Nov 2010) | 2 lines Add macros to define static instances of VNET_DEFINE and DPCPU_DEFINE.
2010-11-22 19:32:54 +00:00
static VNET_DEFINE(int, tcp_syncookies) = 1;
Remove unused VNET_SET() and related macros; only VNET_GET() is ever actually used. Rename VNET_GET() to VNET() to shorten variable references. Discussed with: bz, julian Reviewed by: bz Approved by: re (kensmith, kib)
2009-07-16 21:13:04 +00:00
#define V_tcp_syncookies VNET(tcp_syncookies)
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, syncookies, 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_syncookies), 0,
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
"Use TCP SYN cookies if the syncache overflows");
After some off-list discussion, revert a number of changes to the DPCPU_DEFINE and VNET_DEFINE macros, as these cause problems for various people working on the affected files. A better long-term solution is still being considered. This reversal may give some modules empty set_pcpu or set_vnet sections, but these are harmless. Changes reverted: ------------------------------------------------------------------------ r215318 | dim | 2010-11-14 21:40:55 +0100 (Sun, 14 Nov 2010) | 4 lines Instead of unconditionally emitting .globl's for the __start_set_xxx and __stop_set_xxx symbols, only emit them when the set_vnet or set_pcpu sections are actually defined. ------------------------------------------------------------------------ r215317 | dim | 2010-11-14 21:38:11 +0100 (Sun, 14 Nov 2010) | 3 lines Apply the STATIC_VNET_DEFINE and STATIC_DPCPU_DEFINE macros throughout the tree. ------------------------------------------------------------------------ r215316 | dim | 2010-11-14 21:23:02 +0100 (Sun, 14 Nov 2010) | 2 lines Add macros to define static instances of VNET_DEFINE and DPCPU_DEFINE.
2010-11-22 19:32:54 +00:00
static VNET_DEFINE(int, tcp_syncookiesonly) = 0;
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_syncookiesonly VNET(tcp_syncookiesonly)
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, syncookies_only, 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_syncookiesonly), 0,
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
"Use only TCP SYN cookies");
Add a knob to change default behavior of inheriting listen socket's tcp stack regardless of what the default stack for the system is set to. With current/default behavior, after changing the default tcp stack, the application needs to be restarted to pick up that change. Setting this new knob net.inet.tcp.functions_inherit_listen_socket_stack to '0' would change that behavior and make any new connection use the newly selected default tcp stack. Reviewed by: rrs MFC after: 2 weeks Sponsored by: Limelight Networks
2017-01-27 23:10:46 +00:00
static VNET_DEFINE(int, functions_inherit_listen_socket_stack) = 1;
#define V_functions_inherit_listen_socket_stack \
VNET(functions_inherit_listen_socket_stack)
SYSCTL_INT(_net_inet_tcp, OID_AUTO, functions_inherit_listen_socket_stack,
CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(functions_inherit_listen_socket_stack), 0,
"Inherit listen socket's stack");
- 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
#define ADDED_BY_TOE(sc) ((sc)->sc_tod != NULL)
incorporate feedback since initial commit - rename tcp_ofld.[ch] to tcp_offload.[ch] - document usage and locking conventions of the functions in the toe_usrreqs function vector - document tcpcb, inpcb, and socket fields used by toe - widen the listen interface into 2 functions - rename DISABLE_TCP_OFFLOAD to TCP_OFFLOAD_DISABLE - shrink conditional compilation to reduce the likelihood of bitrot - replace sc->sc_toepcb checks in tcp_syncache.c with TOEPCB_ISSET
2007-12-17 07:56:27 +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
static void syncache_drop(struct syncache *, struct syncache_head *);
static void syncache_free(struct syncache *);
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
static void syncache_insert(struct syncache *, struct syncache_head *);
tcp/syncache: Set flowid and hash type properly for SYN|ACK So the underlying drivers can use it to select the sending queue properly for SYN|ACK instead of rolling their own hash. Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6120
2016-04-29 07:23:08 +00:00
static int syncache_respond(struct syncache *, struct syncache_head *, int,
const struct mbuf *);
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
static struct socket *syncache_socket(struct syncache *, struct socket *,
Modify the arguments to syncache_socket() to include the mbuf (m) that results in the syncache entry being turned into a socket. While it's not used in the main tree, this is required in the MAC tree so that labels can be propagated from the mbuf to the socket. This is also useful if you're doing things like transparent IP connection hijacking and you want to use the syncache/cookie mechanism, but we won't go there. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-05-14 18:57:55 +00:00
struct mbuf *m);
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
static void syncache_timeout(struct syncache *sc, struct syncache_head *sch,
int docallout);
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
static void syncache_timer(void *);
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
static uint32_t syncookie_mac(struct in_conninfo *, tcp_seq, uint8_t,
uint8_t *, uintptr_t);
static tcp_seq syncookie_generate(struct syncache_head *, struct syncache *);
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
static struct syncache
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
*syncookie_lookup(struct in_conninfo *, struct syncache_head *,
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
struct syncache *, struct tcphdr *, struct tcpopt *,
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
struct socket *);
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
static void syncookie_reseed(void *);
#ifdef INVARIANTS
static int syncookie_cmp(struct in_conninfo *inc, struct syncache_head *sch,
struct syncache *sc, struct tcphdr *th, struct tcpopt *to,
struct socket *lso);
#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
/*
* Transmit the SYN,ACK fewer times than TCP_MAXRXTSHIFT specifies.
Fix bugs in the TCP syncache timeout code. including: When system ticks are positive, for entries in the cache bucket, syncache_timer() ran on every tick (doing nothing useful) instead of the supposed 3, 6, 12, and 24 seconds later (when it's time to retransmit SYN,ACK). When ticks are negative, syncache_timer() was scheduled for the too far future (up to ~25 days on systems with HZ=1000), no SYN,ACK retransmits were attempted at all, and syncache entries added in that period that correspond to non-established connections stay there forever. Only HEAD and RELENG_7 are affected. Reviewed by: silby, kmacy (earlier version) Submitted by: Maxim Dounin, ru
2007-12-19 16:56:28 +00:00
* 3 retransmits corresponds to a timeout of 3 * (1 + 2 + 4 + 8) == 45 seconds,
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
* the odds are that the user has given up attempting to connect by then.
*/
#define SYNCACHE_MAXREXMTS 3
/* Arbitrary values */
#define TCP_SYNCACHE_HASHSIZE 512
#define TCP_SYNCACHE_BUCKETLIMIT 30
After some off-list discussion, revert a number of changes to the DPCPU_DEFINE and VNET_DEFINE macros, as these cause problems for various people working on the affected files. A better long-term solution is still being considered. This reversal may give some modules empty set_pcpu or set_vnet sections, but these are harmless. Changes reverted: ------------------------------------------------------------------------ r215318 | dim | 2010-11-14 21:40:55 +0100 (Sun, 14 Nov 2010) | 4 lines Instead of unconditionally emitting .globl's for the __start_set_xxx and __stop_set_xxx symbols, only emit them when the set_vnet or set_pcpu sections are actually defined. ------------------------------------------------------------------------ r215317 | dim | 2010-11-14 21:38:11 +0100 (Sun, 14 Nov 2010) | 3 lines Apply the STATIC_VNET_DEFINE and STATIC_DPCPU_DEFINE macros throughout the tree. ------------------------------------------------------------------------ r215316 | dim | 2010-11-14 21:23:02 +0100 (Sun, 14 Nov 2010) | 2 lines Add macros to define static instances of VNET_DEFINE and DPCPU_DEFINE.
2010-11-22 19:32:54 +00:00
static VNET_DEFINE(struct tcp_syncache, tcp_syncache);
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_syncache VNET(tcp_syncache)
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, syncache, CTLFLAG_RW, 0,
"TCP SYN cache");
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 SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_UINT(_net_inet_tcp_syncache, OID_AUTO, bucketlimit, CTLFLAG_VNET | CTLFLAG_RDTUN,
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_syncache.bucket_limit), 0,
"Per-bucket hash limit 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
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_UINT(_net_inet_tcp_syncache, OID_AUTO, cachelimit, CTLFLAG_VNET | CTLFLAG_RDTUN,
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_syncache.cache_limit), 0,
"Overall entry limit 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
Utilize SYSCTL_UMA_CUR() to export usage of syncache and tcp reassembly zones. Sponsored by: Nginx, Inc.
2014-02-07 14:31:51 +00:00
SYSCTL_UMA_CUR(_net_inet_tcp_syncache, OID_AUTO, count, CTLFLAG_VNET,
&VNET_NAME(tcp_syncache.zone), "Current number of entries in 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
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_UINT(_net_inet_tcp_syncache, OID_AUTO, hashsize, CTLFLAG_VNET | CTLFLAG_RDTUN,
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_syncache.hashsize), 0,
"Size of TCP syncache hashtable");
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 SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_UINT(_net_inet_tcp_syncache, OID_AUTO, rexmtlimit, 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_syncache.rexmt_limit), 0,
"Limit on SYN/ACK retransmissions");
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
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_sc_rst_sock_fail) = 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_syncache, OID_AUTO, rst_on_sock_fail,
CTLFLAG_VNET | CTLFLAG_RW, &VNET_NAME(tcp_sc_rst_sock_fail), 0,
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
"Send reset on socket allocation failure");
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
static MALLOC_DEFINE(M_SYNCACHE, "syncache", "TCP syncache");
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
#define SCH_LOCK(sch) mtx_lock(&(sch)->sch_mtx)
#define SCH_UNLOCK(sch) mtx_unlock(&(sch)->sch_mtx)
#define SCH_LOCK_ASSERT(sch) mtx_assert(&(sch)->sch_mtx, MA_OWNED)
/*
* Requires the syncache entry to be already removed from the bucket list.
*/
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
static void
syncache_free(struct syncache *sc)
{
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
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 (sc->sc_ipopts)
(void) m_free(sc->sc_ipopts);
Cache the cred locally in _syncache_add() while holding the locks, so we can be sure that it's valid. In case we abort early free it again else put it into the syncache. We need the cred in the syncache to be able to restrict what will be exportet by the sysctl helper function syncache_pcblist() (to netstat) within jails. PR: kern/126493 Reviewed by: rwatson (earlier versions) MFC after: 3 days
2008-08-23 14:22:12 +00:00
if (sc->sc_cred)
crfree(sc->sc_cred);
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#ifdef MAC
Normalize TCP syncache-related MAC Framework entry points to match most other entry points in the form mac_<object>_method(). Discussed with: csjp Obtained from: TrustedBSD Project
2007-10-25 14:37:37 +00:00
mac_syncache_destroy(&sc->sc_label);
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#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
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
uma_zfree(V_tcp_syncache.zone, sc);
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
}
void
syncache_init(void)
{
int i;
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_syncache.hashsize = TCP_SYNCACHE_HASHSIZE;
V_tcp_syncache.bucket_limit = TCP_SYNCACHE_BUCKETLIMIT;
V_tcp_syncache.rexmt_limit = SYNCACHE_MAXREXMTS;
V_tcp_syncache.hash_secret = arc4random();
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
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
TUNABLE_INT_FETCH("net.inet.tcp.syncache.hashsize",
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_syncache.hashsize);
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
TUNABLE_INT_FETCH("net.inet.tcp.syncache.bucketlimit",
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_syncache.bucket_limit);
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
if (!powerof2(V_tcp_syncache.hashsize) ||
V_tcp_syncache.hashsize == 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
printf("WARNING: syncache hash size is not a power of 2.\n");
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_syncache.hashsize = TCP_SYNCACHE_HASHSIZE;
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
}
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_syncache.hashmask = V_tcp_syncache.hashsize - 1;
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/* Set limits. */
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_syncache.cache_limit =
V_tcp_syncache.hashsize * V_tcp_syncache.bucket_limit;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
TUNABLE_INT_FETCH("net.inet.tcp.syncache.cachelimit",
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_syncache.cache_limit);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +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
/* Allocate the hash table. */
Fix a number of style issues in the MALLOC / FREE commit. I've tried to be careful not to fix anything that was already broken; the NFSv4 code is particularly bad in this respect.
2008-10-23 20:26:15 +00:00
V_tcp_syncache.hashbase = malloc(V_tcp_syncache.hashsize *
sizeof(struct syncache_head), M_SYNCACHE, M_WAITOK | M_ZERO);
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
Permit buiding kernels with options VIMAGE, restricted to only a single active network stack instance. Turning on options VIMAGE at compile time yields the following changes relative to default kernel build: 1) V_ accessor macros for virtualized variables resolve to structure fields via base pointers, instead of being resolved as fields in global structs or plain global variables. As an example, V_ifnet becomes: options VIMAGE: ((struct vnet_net *) vnet_net)->_ifnet default build: vnet_net_0._ifnet options VIMAGE_GLOBALS: ifnet 2) INIT_VNET_* macros will declare and set up base pointers to be used by V_ accessor macros, instead of resolving to whitespace: INIT_VNET_NET(ifp->if_vnet); becomes struct vnet_net *vnet_net = (ifp->if_vnet)->mod_data[VNET_MOD_NET]; 3) Memory for vnet modules registered via vnet_mod_register() is now allocated at run time in sys/kern/kern_vimage.c, instead of per vnet module structs being declared as globals. If required, vnet modules can now request the framework to provide them with allocated bzeroed memory by filling in the vmi_size field in their vmi_modinfo structures. 4) structs socket, ifnet, inpcbinfo, tcpcb and syncache_head are extended to hold a pointer to the parent vnet. options VIMAGE builds will fill in those fields as required. 5) curvnet is introduced as a new global variable in options VIMAGE builds, always pointing to the default and only struct vnet. 6) struct sysctl_oid has been extended with additional two fields to store major and minor virtualization module identifiers, oid_v_subs and oid_v_mod. SYSCTL_V_* family of macros will fill in those fields accordingly, and store the offset in the appropriate vnet container struct in oid_arg1. In sysctl handlers dealing with virtualized sysctls, the SYSCTL_RESOLVE_V_ARG1() macro will compute the address of the target variable and make it available in arg1 variable for further processing. Unused fields in structs vnet_inet, vnet_inet6 and vnet_ipfw have been deleted. Reviewed by: bz, rwatson Approved by: julian (mentor)
2009-04-30 13:36:26 +00:00
#ifdef VIMAGE
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
V_tcp_syncache.vnet = curvnet;
Permit buiding kernels with options VIMAGE, restricted to only a single active network stack instance. Turning on options VIMAGE at compile time yields the following changes relative to default kernel build: 1) V_ accessor macros for virtualized variables resolve to structure fields via base pointers, instead of being resolved as fields in global structs or plain global variables. As an example, V_ifnet becomes: options VIMAGE: ((struct vnet_net *) vnet_net)->_ifnet default build: vnet_net_0._ifnet options VIMAGE_GLOBALS: ifnet 2) INIT_VNET_* macros will declare and set up base pointers to be used by V_ accessor macros, instead of resolving to whitespace: INIT_VNET_NET(ifp->if_vnet); becomes struct vnet_net *vnet_net = (ifp->if_vnet)->mod_data[VNET_MOD_NET]; 3) Memory for vnet modules registered via vnet_mod_register() is now allocated at run time in sys/kern/kern_vimage.c, instead of per vnet module structs being declared as globals. If required, vnet modules can now request the framework to provide them with allocated bzeroed memory by filling in the vmi_size field in their vmi_modinfo structures. 4) structs socket, ifnet, inpcbinfo, tcpcb and syncache_head are extended to hold a pointer to the parent vnet. options VIMAGE builds will fill in those fields as required. 5) curvnet is introduced as a new global variable in options VIMAGE builds, always pointing to the default and only struct vnet. 6) struct sysctl_oid has been extended with additional two fields to store major and minor virtualization module identifiers, oid_v_subs and oid_v_mod. SYSCTL_V_* family of macros will fill in those fields accordingly, and store the offset in the appropriate vnet container struct in oid_arg1. In sysctl handlers dealing with virtualized sysctls, the SYSCTL_RESOLVE_V_ARG1() macro will compute the address of the target variable and make it available in arg1 variable for further processing. Unused fields in structs vnet_inet, vnet_inet6 and vnet_ipfw have been deleted. Reviewed by: bz, rwatson Approved by: julian (mentor)
2009-04-30 13:36:26 +00:00
#endif
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
/* Initialize the hash buckets. */
for (i = 0; i < V_tcp_syncache.hashsize; i++) {
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
TAILQ_INIT(&V_tcp_syncache.hashbase[i].sch_bucket);
mtx_init(&V_tcp_syncache.hashbase[i].sch_mtx, "tcp_sc_head",
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
NULL, MTX_DEF);
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
callout_init_mtx(&V_tcp_syncache.hashbase[i].sch_timer,
&V_tcp_syncache.hashbase[i].sch_mtx, 0);
V_tcp_syncache.hashbase[i].sch_length = 0;
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
V_tcp_syncache.hashbase[i].sch_sc = &V_tcp_syncache;
Represent "a syncache overflow hasn't happend yet" by using -(SYNCOOKIE_LIFETIME + 1) instead of INT64_MIN, since it is good enough and works when time_t is int32 or int64. This fixes the issue reported by cy@ on i386. Reported by: cy MFC after: 1 week Sponsored by: Netflix, Inc.
2017-04-21 06:05:34 +00:00
V_tcp_syncache.hashbase[i].sch_last_overflow =
-(SYNCOOKIE_LIFETIME + 1);
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/* Create the syncache entry zone. */
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_syncache.zone = uma_zcreate("syncache", sizeof(struct syncache),
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0);
uma_zone_set_max() directly returns the rounded effective zone limit. Use the return value directly instead of doing a second uma_zone_set_max() step. MFC after: 1 week
2013-02-01 14:21:09 +00:00
V_tcp_syncache.cache_limit = uma_zone_set_max(V_tcp_syncache.zone,
V_tcp_syncache.cache_limit);
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
/* Start the SYN cookie reseeder callout. */
callout_init(&V_tcp_syncache.secret.reseed, 1);
arc4rand(V_tcp_syncache.secret.key[0], SYNCOOKIE_SECRET_SIZE, 0);
arc4rand(V_tcp_syncache.secret.key[1], SYNCOOKIE_SECRET_SIZE, 0);
callout_reset(&V_tcp_syncache.secret.reseed, SYNCOOKIE_LIFETIME * hz,
syncookie_reseed, &V_tcp_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
}
Introduce an infrastructure for dismantling vnet instances. Vnet modules and protocol domains may now register destructor functions to clean up and release per-module state. The destructor mechanisms can be triggered by invoking "vimage -d", or a future equivalent command which will be provided via the new jail framework. While this patch introduces numerous placeholder destructor functions, many of those are currently incomplete, thus leaking memory or (even worse) failing to stop all running timers. Many of such issues are already known and will be incrementaly fixed over the next weeks in smaller incremental commits. Apart from introducing new fields in structs ifnet, domain, protosw and vnet_net, which requires the kernel and modules to be rebuilt, this change should have no impact on nooptions VIMAGE builds, since vnet destructors can only be called in VIMAGE kernels. Moreover, destructor functions should be in general compiled in only in options VIMAGE builds, except for kernel modules which can be safely kldunloaded at run time. Bump __FreeBSD_version to 800097. Reviewed by: bz, julian Approved by: rwatson, kib (re), julian (mentor)
2009-06-08 17:15:40 +00:00
#ifdef VIMAGE
void
syncache_destroy(void)
{
Upon virtual network stack teardown properly release the TCP syncache resources. Sponsored by: ISPsystem Reviewed by: rwatson MFC After: 5 days
2010-02-20 21:45:04 +00:00
struct syncache_head *sch;
struct syncache *sc, *nsc;
int i;
Mfp: r296309 While there is no dependency interaction, stopping the timer before freeing the rest of the resources seems more natural and avoids it being scheduled an extra time when it is no longer needed. Reviewed by: gnn, emaste MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D5733
2016-04-09 10:51:07 +00:00
/*
* Stop the re-seed timer before freeing resources. No need to
* possibly schedule it another time.
*/
callout_drain(&V_tcp_syncache.secret.reseed);
Upon virtual network stack teardown properly release the TCP syncache resources. Sponsored by: ISPsystem Reviewed by: rwatson MFC After: 5 days
2010-02-20 21:45:04 +00:00
/* Cleanup hash buckets: stop timers, free entries, destroy locks. */
for (i = 0; i < V_tcp_syncache.hashsize; i++) {
sch = &V_tcp_syncache.hashbase[i];
callout_drain(&sch->sch_timer);
SCH_LOCK(sch);
TAILQ_FOREACH_SAFE(sc, &sch->sch_bucket, sc_hash, nsc)
syncache_drop(sc, sch);
SCH_UNLOCK(sch);
KASSERT(TAILQ_EMPTY(&sch->sch_bucket),
("%s: sch->sch_bucket not empty", __func__));
KASSERT(sch->sch_length == 0, ("%s: sch->sch_length %d not 0",
__func__, sch->sch_length));
mtx_destroy(&sch->sch_mtx);
}
Introduce an infrastructure for dismantling vnet instances. Vnet modules and protocol domains may now register destructor functions to clean up and release per-module state. The destructor mechanisms can be triggered by invoking "vimage -d", or a future equivalent command which will be provided via the new jail framework. While this patch introduces numerous placeholder destructor functions, many of those are currently incomplete, thus leaking memory or (even worse) failing to stop all running timers. Many of such issues are already known and will be incrementaly fixed over the next weeks in smaller incremental commits. Apart from introducing new fields in structs ifnet, domain, protosw and vnet_net, which requires the kernel and modules to be rebuilt, this change should have no impact on nooptions VIMAGE builds, since vnet destructors can only be called in VIMAGE kernels. Moreover, destructor functions should be in general compiled in only in options VIMAGE builds, except for kernel modules which can be safely kldunloaded at run time. Bump __FreeBSD_version to 800097. Reviewed by: bz, julian Approved by: rwatson, kib (re), julian (mentor)
2009-06-08 17:15:40 +00:00
Change the syncache count reporting the current number of entries from an unprotected u_int that reports garbage on SMP to a function based sysctl obtaining the current value from UMA. Also read back the actual cache_limit after page size rounding by UMA. PR: kern/165879 MFC after: 2 weeks
2012-10-28 18:07:34 +00:00
KASSERT(uma_zone_get_cur(V_tcp_syncache.zone) == 0,
("%s: cache_count not 0", __func__));
Introduce an infrastructure for dismantling vnet instances. Vnet modules and protocol domains may now register destructor functions to clean up and release per-module state. The destructor mechanisms can be triggered by invoking "vimage -d", or a future equivalent command which will be provided via the new jail framework. While this patch introduces numerous placeholder destructor functions, many of those are currently incomplete, thus leaking memory or (even worse) failing to stop all running timers. Many of such issues are already known and will be incrementaly fixed over the next weeks in smaller incremental commits. Apart from introducing new fields in structs ifnet, domain, protosw and vnet_net, which requires the kernel and modules to be rebuilt, this change should have no impact on nooptions VIMAGE builds, since vnet destructors can only be called in VIMAGE kernels. Moreover, destructor functions should be in general compiled in only in options VIMAGE builds, except for kernel modules which can be safely kldunloaded at run time. Bump __FreeBSD_version to 800097. Reviewed by: bz, julian Approved by: rwatson, kib (re), julian (mentor)
2009-06-08 17:15:40 +00:00
Upon virtual network stack teardown properly release the TCP syncache resources. Sponsored by: ISPsystem Reviewed by: rwatson MFC After: 5 days
2010-02-20 21:45:04 +00:00
/* Free the allocated global resources. */
Introduce an infrastructure for dismantling vnet instances. Vnet modules and protocol domains may now register destructor functions to clean up and release per-module state. The destructor mechanisms can be triggered by invoking "vimage -d", or a future equivalent command which will be provided via the new jail framework. While this patch introduces numerous placeholder destructor functions, many of those are currently incomplete, thus leaking memory or (even worse) failing to stop all running timers. Many of such issues are already known and will be incrementaly fixed over the next weeks in smaller incremental commits. Apart from introducing new fields in structs ifnet, domain, protosw and vnet_net, which requires the kernel and modules to be rebuilt, this change should have no impact on nooptions VIMAGE builds, since vnet destructors can only be called in VIMAGE kernels. Moreover, destructor functions should be in general compiled in only in options VIMAGE builds, except for kernel modules which can be safely kldunloaded at run time. Bump __FreeBSD_version to 800097. Reviewed by: bz, julian Approved by: rwatson, kib (re), julian (mentor)
2009-06-08 17:15:40 +00:00
uma_zdestroy(V_tcp_syncache.zone);
Upon virtual network stack teardown properly release the TCP syncache resources. Sponsored by: ISPsystem Reviewed by: rwatson MFC After: 5 days
2010-02-20 21:45:04 +00:00
free(V_tcp_syncache.hashbase, M_SYNCACHE);
Introduce an infrastructure for dismantling vnet instances. Vnet modules and protocol domains may now register destructor functions to clean up and release per-module state. The destructor mechanisms can be triggered by invoking "vimage -d", or a future equivalent command which will be provided via the new jail framework. While this patch introduces numerous placeholder destructor functions, many of those are currently incomplete, thus leaking memory or (even worse) failing to stop all running timers. Many of such issues are already known and will be incrementaly fixed over the next weeks in smaller incremental commits. Apart from introducing new fields in structs ifnet, domain, protosw and vnet_net, which requires the kernel and modules to be rebuilt, this change should have no impact on nooptions VIMAGE builds, since vnet destructors can only be called in VIMAGE kernels. Moreover, destructor functions should be in general compiled in only in options VIMAGE builds, except for kernel modules which can be safely kldunloaded at run time. Bump __FreeBSD_version to 800097. Reviewed by: bz, julian Approved by: rwatson, kib (re), julian (mentor)
2009-06-08 17:15:40 +00:00
}
#endif
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/*
* Inserts a syncache entry into the specified bucket row.
* Locks and unlocks the syncache_head autonomously.
*/
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
static void
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
syncache_insert(struct syncache *sc, struct syncache_head *sch)
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 syncache *sc2;
o add missing inpcb locking in tcp_respond o replace spl's with lock assertions Supported by: FreeBSD Foundation
2003-11-11 17:54:47 +00:00
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
SCH_LOCK(sch);
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
/*
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
* Make sure that we don't overflow the per-bucket limit.
* If the bucket is full, toss the oldest element.
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
*/
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 (sch->sch_length >= V_tcp_syncache.bucket_limit) {
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
KASSERT(!TAILQ_EMPTY(&sch->sch_bucket),
("sch->sch_length incorrect"));
sc2 = TAILQ_LAST(&sch->sch_bucket, sch_head);
Syncoockies can be used in combination with the syncache. If the cache overflows, syncookies are used. This patch restricts the usage of syncookies in this case: accept syncookies only if there was an overflow of the syncache recently. This mitigates a problem reported in PR217637, where is syncookie was accepted without any recent drops. Thanks to glebius@ for suggesting an improvement. PR: 217637 Reviewed by: gnn, glebius MFC after: 1 week Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D10272
2017-04-20 19:19:33 +00:00
sch->sch_last_overflow = time_uptime;
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
syncache_drop(sc2, sch);
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_sc_bucketoverflow);
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
}
/* Put it into the bucket. */
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
TAILQ_INSERT_HEAD(&sch->sch_bucket, sc, sc_hash);
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
sch->sch_length++;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +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 (ADDED_BY_TOE(sc)) {
struct toedev *tod = sc->sc_tod;
tod->tod_syncache_added(tod, sc->sc_todctx);
}
#endif
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/* Reinitialize the bucket row's timer. */
Fix bugs in the TCP syncache timeout code. including: When system ticks are positive, for entries in the cache bucket, syncache_timer() ran on every tick (doing nothing useful) instead of the supposed 3, 6, 12, and 24 seconds later (when it's time to retransmit SYN,ACK). When ticks are negative, syncache_timer() was scheduled for the too far future (up to ~25 days on systems with HZ=1000), no SYN,ACK retransmits were attempted at all, and syncache entries added in that period that correspond to non-established connections stay there forever. Only HEAD and RELENG_7 are affected. Reviewed by: silby, kmacy (earlier version) Submitted by: Maxim Dounin, ru
2007-12-19 16:56:28 +00:00
if (sch->sch_length == 1)
sch->sch_nextc = ticks + INT_MAX;
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
syncache_timeout(sc, sch, 1);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
SCH_UNLOCK(sch);
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
TCPSTATES_INC(TCPS_SYN_RECEIVED);
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_sc_added);
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/*
* Remove and free entry from syncache bucket row.
* Expects locked syncache head.
*/
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
static void
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
syncache_drop(struct syncache *sc, struct syncache_head *sch)
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
{
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
SCH_LOCK_ASSERT(sch);
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
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
TCPSTATES_DEC(TCPS_SYN_RECEIVED);
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
TAILQ_REMOVE(&sch->sch_bucket, sc, sc_hash);
sch->sch_length--;
- 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 (ADDED_BY_TOE(sc)) {
struct toedev *tod = sc->sc_tod;
tod->tod_syncache_removed(tod, sc->sc_todctx);
}
#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
syncache_free(sc);
}
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
/*
* Engage/reengage time on bucket row.
*/
static void
syncache_timeout(struct syncache *sc, struct syncache_head *sch, int docallout)
{
sc->sc_rxttime = ticks +
For retransmits of SYN|ACK from the syncache use the slightly more aggressive special tcp_syn_backoff[] retransmit schedule instead of the normal tcp_backoff[] schedule for established connections. MFC after: 2 weeks
2012-10-28 19:02:07 +00:00
TCPTV_RTOBASE * (tcp_syn_backoff[sc->sc_rxmits]);
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
sc->sc_rxmits++;
Fix bugs in the TCP syncache timeout code. including: When system ticks are positive, for entries in the cache bucket, syncache_timer() ran on every tick (doing nothing useful) instead of the supposed 3, 6, 12, and 24 seconds later (when it's time to retransmit SYN,ACK). When ticks are negative, syncache_timer() was scheduled for the too far future (up to ~25 days on systems with HZ=1000), no SYN,ACK retransmits were attempted at all, and syncache entries added in that period that correspond to non-established connections stay there forever. Only HEAD and RELENG_7 are affected. Reviewed by: silby, kmacy (earlier version) Submitted by: Maxim Dounin, ru
2007-12-19 16:56:28 +00:00
if (TSTMP_LT(sc->sc_rxttime, sch->sch_nextc)) {
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
sch->sch_nextc = sc->sc_rxttime;
Fix bugs in the TCP syncache timeout code. including: When system ticks are positive, for entries in the cache bucket, syncache_timer() ran on every tick (doing nothing useful) instead of the supposed 3, 6, 12, and 24 seconds later (when it's time to retransmit SYN,ACK). When ticks are negative, syncache_timer() was scheduled for the too far future (up to ~25 days on systems with HZ=1000), no SYN,ACK retransmits were attempted at all, and syncache entries added in that period that correspond to non-established connections stay there forever. Only HEAD and RELENG_7 are affected. Reviewed by: silby, kmacy (earlier version) Submitted by: Maxim Dounin, ru
2007-12-19 16:56:28 +00:00
if (docallout)
callout_reset(&sch->sch_timer, sch->sch_nextc - ticks,
syncache_timer, (void *)sch);
}
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +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
/*
* Walk the timer queues, looking for SYN,ACKs that need to be retransmitted.
* If we have retransmitted an entry the maximum number of times, expire it.
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
* One separate timer for each bucket row.
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
*/
static void
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
syncache_timer(void *xsch)
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
{
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
struct syncache_head *sch = (struct syncache_head *)xsch;
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 syncache *sc, *nsc;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
int tick = ticks;
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
char *s;
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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
CURVNET_SET(sch->sch_sc->vnet);
Merge more of currently non-functional (i.e. resolving to whitespace) macros from p4/vimage branch. Do a better job at enclosing all instantiations of globals scheduled for virtualization in #ifdef VIMAGE_GLOBALS blocks. De-virtualize and mark as const saorder_state_alive and saorder_state_any arrays from ipsec code, given that they are never updated at runtime, so virtualizing them would be pointless. 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
2008-11-26 22:32:07 +00:00
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/* NB: syncache_head has already been locked by the callout. */
SCH_LOCK_ASSERT(sch);
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
Fix bugs in the TCP syncache timeout code. including: When system ticks are positive, for entries in the cache bucket, syncache_timer() ran on every tick (doing nothing useful) instead of the supposed 3, 6, 12, and 24 seconds later (when it's time to retransmit SYN,ACK). When ticks are negative, syncache_timer() was scheduled for the too far future (up to ~25 days on systems with HZ=1000), no SYN,ACK retransmits were attempted at all, and syncache entries added in that period that correspond to non-established connections stay there forever. Only HEAD and RELENG_7 are affected. Reviewed by: silby, kmacy (earlier version) Submitted by: Maxim Dounin, ru
2007-12-19 16:56:28 +00:00
/*
* In the following cycle we may remove some entries and/or
* advance some timeouts, so re-initialize the bucket timer.
*/
sch->sch_nextc = tick + INT_MAX;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
TAILQ_FOREACH_SAFE(sc, &sch->sch_bucket, sc_hash, nsc) {
/*
* We do not check if the listen socket still exists
* and accept the case where the listen socket may be
* gone by the time we resend the SYN/ACK. We do
* not expect this to happens often. If it does,
* then the RST will be sent by the time the remote
* host does the SYN/ACK->ACK.
*/
Fix bugs in the TCP syncache timeout code. including: When system ticks are positive, for entries in the cache bucket, syncache_timer() ran on every tick (doing nothing useful) instead of the supposed 3, 6, 12, and 24 seconds later (when it's time to retransmit SYN,ACK). When ticks are negative, syncache_timer() was scheduled for the too far future (up to ~25 days on systems with HZ=1000), no SYN,ACK retransmits were attempted at all, and syncache entries added in that period that correspond to non-established connections stay there forever. Only HEAD and RELENG_7 are affected. Reviewed by: silby, kmacy (earlier version) Submitted by: Maxim Dounin, ru
2007-12-19 16:56:28 +00:00
if (TSTMP_GT(sc->sc_rxttime, tick)) {
if (TSTMP_LT(sc->sc_rxttime, sch->sch_nextc))
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sch->sch_nextc = sc->sc_rxttime;
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 (sc->sc_rxmits > V_tcp_syncache.rexmt_limit) {
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
if ((s = tcp_log_addrs(&sc->sc_inc, NULL, NULL, NULL))) {
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
log(LOG_DEBUG, "%s; %s: Retransmits exhausted, "
"giving up and removing syncache entry\n",
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
s, __func__);
free(s, M_TCPLOG);
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
syncache_drop(sc, sch);
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_sc_stale);
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
continue;
}
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
if ((s = tcp_log_addrs(&sc->sc_inc, NULL, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: Response timeout, "
"retransmitting (%u) SYN|ACK\n",
s, __func__, sc->sc_rxmits);
free(s, M_TCPLOG);
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
tcp/syncache: Set flowid and hash type properly for SYN|ACK So the underlying drivers can use it to select the sending queue properly for SYN|ACK instead of rolling their own hash. Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6120
2016-04-29 07:23:08 +00:00
syncache_respond(sc, sch, 1, 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_sc_retransmitted);
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
syncache_timeout(sc, sch, 0);
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
if (!TAILQ_EMPTY(&(sch)->sch_bucket))
callout_reset(&(sch)->sch_timer, (sch)->sch_nextc - tick,
syncache_timer, (void *)(sch));
Merge more of currently non-functional (i.e. resolving to whitespace) macros from p4/vimage branch. Do a better job at enclosing all instantiations of globals scheduled for virtualization in #ifdef VIMAGE_GLOBALS blocks. De-virtualize and mark as const saorder_state_alive and saorder_state_any arrays from ipsec code, given that they are never updated at runtime, so virtualizing them would be pointless. 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
2008-11-26 22:32:07 +00:00
CURVNET_RESTORE();
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
}
/*
* Find an entry in the syncache.
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
* Returns always with locked syncache_head plus a matching entry or 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
*/
syncache_lookup() is a file local function. Make it static and take it out of the public KPI; seems it was never used elsewhere. MFC after: 2 weeks
2014-05-24 15:03:36 +00:00
static struct syncache *
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
syncache_lookup(struct in_conninfo *inc, struct syncache_head **schp)
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 syncache *sc;
struct syncache_head *sch;
Use Jenkins hash for TCP syncache. o Unlike xor, in Jenkins hash every bit of input affects virtually every bit of output, thus salting the hash actually works. With xor salting only provides a false sense of security, since if hash(x) collides with hash(y), then of course, hash(x) ^ salt would also collide with hash(y) ^ salt. [1] o Jenkins provides much better distribution than xor, very close to ideal. TCP connection setup/teardown benchmark has shown a 10% increase with default hash size, and with bigger hashes that still provide possibility for collisions. With enormous hash size, when dataset is by an order of magnitude smaller than hash size, the benchmark has shown 4% decrease in performance decrease, which is expected and acceptable. Noticed by: Jeffrey Knockel <jeffk cs.unm.edu> [1] Benchmarks by: jch Reviewed by: jch, pkelsey, delphij Security: strengthens protection against hash collision DoS Sponsored by: Nginx, Inc.
2015-09-05 10:15:19 +00:00
uint32_t hash;
o add missing inpcb locking in tcp_respond o replace spl's with lock assertions Supported by: FreeBSD Foundation
2003-11-11 17:54:47 +00:00
Use Jenkins hash for TCP syncache. o Unlike xor, in Jenkins hash every bit of input affects virtually every bit of output, thus salting the hash actually works. With xor salting only provides a false sense of security, since if hash(x) collides with hash(y), then of course, hash(x) ^ salt would also collide with hash(y) ^ salt. [1] o Jenkins provides much better distribution than xor, very close to ideal. TCP connection setup/teardown benchmark has shown a 10% increase with default hash size, and with bigger hashes that still provide possibility for collisions. With enormous hash size, when dataset is by an order of magnitude smaller than hash size, the benchmark has shown 4% decrease in performance decrease, which is expected and acceptable. Noticed by: Jeffrey Knockel <jeffk cs.unm.edu> [1] Benchmarks by: jch Reviewed by: jch, pkelsey, delphij Security: strengthens protection against hash collision DoS Sponsored by: Nginx, Inc.
2015-09-05 10:15:19 +00:00
/*
* The hash is built on foreign port + local port + foreign address.
* We rely on the fact that struct in_conninfo starts with 16 bits
* of foreign port, then 16 bits of local port then followed by 128
* bits of foreign address. In case of IPv4 address, the first 3
* 32-bit words of the address always are zeroes.
*/
hash = jenkins_hash32((uint32_t *)&inc->inc_ie, 5,
V_tcp_syncache.hash_secret) & V_tcp_syncache.hashmask;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
Use Jenkins hash for TCP syncache. o Unlike xor, in Jenkins hash every bit of input affects virtually every bit of output, thus salting the hash actually works. With xor salting only provides a false sense of security, since if hash(x) collides with hash(y), then of course, hash(x) ^ salt would also collide with hash(y) ^ salt. [1] o Jenkins provides much better distribution than xor, very close to ideal. TCP connection setup/teardown benchmark has shown a 10% increase with default hash size, and with bigger hashes that still provide possibility for collisions. With enormous hash size, when dataset is by an order of magnitude smaller than hash size, the benchmark has shown 4% decrease in performance decrease, which is expected and acceptable. Noticed by: Jeffrey Knockel <jeffk cs.unm.edu> [1] Benchmarks by: jch Reviewed by: jch, pkelsey, delphij Security: strengthens protection against hash collision DoS Sponsored by: Nginx, Inc.
2015-09-05 10:15:19 +00:00
sch = &V_tcp_syncache.hashbase[hash];
*schp = sch;
SCH_LOCK(sch);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
Use Jenkins hash for TCP syncache. o Unlike xor, in Jenkins hash every bit of input affects virtually every bit of output, thus salting the hash actually works. With xor salting only provides a false sense of security, since if hash(x) collides with hash(y), then of course, hash(x) ^ salt would also collide with hash(y) ^ salt. [1] o Jenkins provides much better distribution than xor, very close to ideal. TCP connection setup/teardown benchmark has shown a 10% increase with default hash size, and with bigger hashes that still provide possibility for collisions. With enormous hash size, when dataset is by an order of magnitude smaller than hash size, the benchmark has shown 4% decrease in performance decrease, which is expected and acceptable. Noticed by: Jeffrey Knockel <jeffk cs.unm.edu> [1] Benchmarks by: jch Reviewed by: jch, pkelsey, delphij Security: strengthens protection against hash collision DoS Sponsored by: Nginx, Inc.
2015-09-05 10:15:19 +00:00
/* Circle through bucket row to find matching entry. */
TAILQ_FOREACH(sc, &sch->sch_bucket, sc_hash)
if (bcmp(&inc->inc_ie, &sc->sc_inc.inc_ie,
sizeof(struct in_endpoints)) == 0)
break;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
Use Jenkins hash for TCP syncache. o Unlike xor, in Jenkins hash every bit of input affects virtually every bit of output, thus salting the hash actually works. With xor salting only provides a false sense of security, since if hash(x) collides with hash(y), then of course, hash(x) ^ salt would also collide with hash(y) ^ salt. [1] o Jenkins provides much better distribution than xor, very close to ideal. TCP connection setup/teardown benchmark has shown a 10% increase with default hash size, and with bigger hashes that still provide possibility for collisions. With enormous hash size, when dataset is by an order of magnitude smaller than hash size, the benchmark has shown 4% decrease in performance decrease, which is expected and acceptable. Noticed by: Jeffrey Knockel <jeffk cs.unm.edu> [1] Benchmarks by: jch Reviewed by: jch, pkelsey, delphij Security: strengthens protection against hash collision DoS Sponsored by: Nginx, Inc.
2015-09-05 10:15:19 +00:00
return (sc); /* Always returns with locked sch. */
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
}
/*
* This function is called when we get a RST for a
* non-existent connection, so that we can see if the
* connection is in the syn cache. If it is, zap it.
*/
void
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
syncache_chkrst(struct in_conninfo *inc, struct tcphdr *th)
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 syncache *sc;
struct syncache_head *sch;
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
char *s = 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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sc = syncache_lookup(inc, &sch); /* returns locked sch */
SCH_LOCK_ASSERT(sch);
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
/*
* Any RST to our SYN|ACK must not carry ACK, SYN or FIN flags.
* See RFC 793 page 65, section SEGMENT ARRIVES.
*/
if (th->th_flags & (TH_ACK|TH_SYN|TH_FIN)) {
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Spurious RST with ACK, SYN or "
"FIN flag set, segment ignored\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_badrst);
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
goto done;
}
/*
* No corresponding connection was found in syncache.
* If syncookies are enabled and possibly exclusively
* used, or we are under memory pressure, a valid RST
* may not find a syncache entry. In that case we're
* done and no SYN|ACK retransmissions will happen.
Fix typos - remove duplicate "the". PR: bin/154928 Submitted by: Eitan Adler <lists at eitanadler.com> MFC after: 3 days
2011-02-21 09:01:34 +00:00
* Otherwise the RST was misdirected or spoofed.
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
*/
if (sc == NULL) {
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Spurious RST without matching "
"syncache entry (possibly syncookie only), "
"segment ignored\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_badrst);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
goto done;
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +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 the RST bit is set, check the sequence number to see
* if this is a valid reset segment.
* RFC 793 page 37:
* In all states except SYN-SENT, all reset (RST) segments
* are validated by checking their SEQ-fields. A reset is
* valid if its sequence number is in the window.
*
* The sequence number in the reset segment is normally an
* echo of our outgoing acknowlegement numbers, but some hosts
* send a reset with the sequence number at the rightmost edge
* of our receive window, and we have to handle this case.
*/
if (SEQ_GEQ(th->th_seq, sc->sc_irs) &&
SEQ_LEQ(th->th_seq, sc->sc_irs + sc->sc_wnd)) {
syncache_drop(sc, sch);
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
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Our SYN|ACK was rejected, "
"connection attempt aborted by remote endpoint\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_sc_reset);
Always bump tcpstat.tcps_badrst if we get a RST for a connection in the syncache that has an invalid SEQ instead of only doing it when we suceed in mallocing space for the log message. MFC after: 1 week Reviewed by: sam, bz
2008-05-08 22:21:09 +00:00
} else {
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: RST with invalid SEQ %u != "
"IRS %u (+WND %u), segment ignored\n",
s, __func__, th->th_seq, sc->sc_irs, sc->sc_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_badrst);
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
}
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
done:
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
if (s != NULL)
free(s, M_TCPLOG);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
SCH_UNLOCK(sch);
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
}
void
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
syncache_badack(struct in_conninfo *inc)
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 syncache *sc;
struct syncache_head *sch;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sc = syncache_lookup(inc, &sch); /* returns locked sch */
SCH_LOCK_ASSERT(sch);
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 (sc != NULL) {
syncache_drop(sc, sch);
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_sc_badack);
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
SCH_UNLOCK(sch);
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
}
void
Fix the ICMP6 handling for TCP. The ICMP6 packets might not be contained in a single mbuf. So don't assume this. Keep the IPv4 and IPv6 code in sync and make explicit that the syncache code only need the TCP sequence number, not the complete TCP header. MFC after: 3 days Sponsored by: Netflix, Inc.
2017-06-03 21:53:58 +00:00
syncache_unreach(struct in_conninfo *inc, tcp_seq th_seq)
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 syncache *sc;
struct syncache_head *sch;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sc = syncache_lookup(inc, &sch); /* returns locked sch */
SCH_LOCK_ASSERT(sch);
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 (sc == NULL)
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
goto 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
/* If the sequence number != sc_iss, then it's a bogus ICMP msg */
Fix the ICMP6 handling for TCP. The ICMP6 packets might not be contained in a single mbuf. So don't assume this. Keep the IPv4 and IPv6 code in sync and make explicit that the syncache code only need the TCP sequence number, not the complete TCP header. MFC after: 3 days Sponsored by: Netflix, Inc.
2017-06-03 21:53:58 +00:00
if (ntohl(th_seq) != sc->sc_iss)
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
goto 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
/*
* If we've rertransmitted 3 times and this is our second error,
* we remove the entry. Otherwise, we allow it to continue on.
* This prevents us from incorrectly nuking an entry during a
* spurious network outage.
*
* See tcp_notify().
*/
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
if ((sc->sc_flags & SCF_UNREACH) == 0 || sc->sc_rxmits < 3 + 1) {
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
sc->sc_flags |= SCF_UNREACH;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
goto 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
}
syncache_drop(sc, sch);
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_sc_unreach);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
done:
SCH_UNLOCK(sch);
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
}
/*
* Build a new TCP socket structure from a syncache entry.
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
*
* On success return the newly created socket with its underlying inp locked.
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
*/
static struct socket *
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
syncache_socket(struct syncache *sc, struct socket *lso, struct mbuf *m)
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
{
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
struct tcp_function_block *blk;
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 inpcb *inp = NULL;
struct socket *so;
struct tcpcb *tp;
Initializing the new error variable to zero in syncache_socket() is not necessary. Noticed by: bz
2010-08-15 13:07:08 +00:00
int error;
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
char *s;
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
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);
o add missing inpcb locking in tcp_respond o replace spl's with lock assertions Supported by: FreeBSD Foundation
2003-11-11 17:54:47 +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
/*
* Ok, create the full blown connection, and set things up
* as they would have been set up if we had created the
* connection when the SYN arrived. If we can't create
* the connection, abort it.
*/
Decrease lock contention within the TCP accept case by removing the INP_INFO lock from tcp_usr_accept. As the PR/patch states this was following the advice already in the code. See the PR below for a full disucssion of this change and its measured effects. PR: 183659 Submitted by: Julian Charbon Reviewed by: jhb
2014-01-28 20:28:32 +00:00
so = sonewconn(lso, 0);
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 (so == NULL) {
/*
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
* Drop the connection; we will either send a RST or
* have the peer retransmit its SYN again after its
* RTO and try again.
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
*/
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_listendrop);
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
if ((s = tcp_log_addrs(&sc->sc_inc, NULL, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: Socket create failed "
"due to limits or memory shortage\n",
s, __func__);
free(s, M_TCPLOG);
}
o add missing inpcb locking in tcp_respond o replace spl's with lock assertions Supported by: FreeBSD Foundation
2003-11-11 17:54:47 +00:00
goto abort2;
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
}
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
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
inp = sotoinpcb(so);
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
inp->inp_inc.inc_fibnum = so->so_fibnum;
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(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
/*
* Exclusive pcbinfo lock is not required in syncache socket case even
* if two inpcb locks can be acquired simultaneously:
* - the inpcb in LISTEN state,
* - the newly created inp.
*
* In this case, an inp cannot be at same time in LISTEN state and
* just created by an accept() call.
*/
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_HASH_WLOCK(&V_tcbinfo);
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/* Insert new socket into PCB hash list. */
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
inp->inp_inc.inc_flags = sc->sc_inc.inc_flags;
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
#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 (sc->sc_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
inp->in6p_laddr = sc->sc_inc.inc6_laddr;
} else {
inp->inp_vflag &= ~INP_IPV6;
inp->inp_vflag |= INP_IPV4;
#endif
inp->inp_laddr = sc->sc_inc.inc_laddr;
#ifdef INET6
}
#endif
Implement a CPU-affine TCP and UDP connection lookup data structure, struct inpcbgroup. pcbgroups, or "connection groups", supplement the existing inpcbinfo connection hash table, which when pcbgroups are enabled, might now be thought of more usefully as a per-protocol 4-tuple reservation table. Connections are assigned to connection groups base on a hash of their 4-tuple; wildcard sockets require special handling, and are members of all connection groups. During a connection lookup, a per-connection group lock is employed rather than the global pcbinfo lock. By aligning connection groups with input path processing, connection groups take on an effective CPU affinity, especially when aligned with RSS work placement (see a forthcoming commit for details). This eliminates cache line migration associated with global, protocol-layer data structures in steady state TCP and UDP processing (with the exception of protocol-layer statistics; further commit to follow). Elements of this approach were inspired by Willman, Rixner, and Cox's 2006 USENIX paper, "An Evaluation of Network Stack Parallelization Strategies in Modern Operating Systems". However, there are also significant differences: we maintain the inpcb lock, rather than using the connection group lock for per-connection state. Likewise, the focus of this implementation is alignment with NIC packet distribution strategies such as RSS, rather than pure software strategies. Despite that focus, software distribution is supported through the parallel netisr implementation, and works well in configurations where the number of hardware threads is greater than the number of NIC input queues, such as in the RMI XLR threaded MIPS architecture. Another important difference is the continued maintenance of existing hash tables as "reservation tables" -- these are useful both to distinguish the resource allocation aspect of protocol name management and the more common-case lookup aspect. In configurations where connection tables are aligned with hardware hashes, it is desirable to use the traditional lookup tables for loopback or encapsulated traffic rather than take the expense of hardware hashes that are hard to implement efficiently in software (such as RSS Toeplitz). Connection group support is enabled by compiling "options PCBGROUP" into your kernel configuration; for the time being, this is an experimental feature, and hence is not enabled by default. Subject to the limited MFCability of change dependencies in inpcb, and its change to the inpcbinfo init function signature, this change in principle could be merged to FreeBSD 8.x. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-06-06 12:55:02 +00:00
If the flowid is available for the mbuf that finalised the creation of a syncache connection, copy it into the inp_flowid field. Without this, an incoming TCP connection won't have an inp_flowid marked until some data comes in, and this means that things like the per-CPU TCP timer option will choose a different CPU for the timer work. (It also means that if one grabbed the flowid via an ioctl from userland, it won't be available until some data has been received.) Sponsored by: Netflix, Inc.
2014-01-18 23:48:20 +00:00
/*
* If there's an mbuf and it has a flowid, then let's initialise the
* inp with that particular flowid.
*/
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 (m != NULL && M_HASHTYPE_GET(m) != M_HASHTYPE_NONE) {
If the flowid is available for the mbuf that finalised the creation of a syncache connection, copy it into the inp_flowid field. Without this, an incoming TCP connection won't have an inp_flowid marked until some data comes in, and this means that things like the per-CPU TCP timer option will choose a different CPU for the timer work. (It also means that if one grabbed the flowid via an ioctl from userland, it won't be available until some data has been received.) Sponsored by: Netflix, Inc.
2014-01-18 23:48:20 +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);
If the flowid is available for the mbuf that finalised the creation of a syncache connection, copy it into the inp_flowid field. Without this, an incoming TCP connection won't have an inp_flowid marked until some data comes in, and this means that things like the per-CPU TCP timer option will choose a different CPU for the timer work. (It also means that if one grabbed the flowid via an ioctl from userland, it won't be available until some data has been received.) Sponsored by: Netflix, Inc.
2014-01-18 23:48:20 +00:00
}
Implement a CPU-affine TCP and UDP connection lookup data structure, struct inpcbgroup. pcbgroups, or "connection groups", supplement the existing inpcbinfo connection hash table, which when pcbgroups are enabled, might now be thought of more usefully as a per-protocol 4-tuple reservation table. Connections are assigned to connection groups base on a hash of their 4-tuple; wildcard sockets require special handling, and are members of all connection groups. During a connection lookup, a per-connection group lock is employed rather than the global pcbinfo lock. By aligning connection groups with input path processing, connection groups take on an effective CPU affinity, especially when aligned with RSS work placement (see a forthcoming commit for details). This eliminates cache line migration associated with global, protocol-layer data structures in steady state TCP and UDP processing (with the exception of protocol-layer statistics; further commit to follow). Elements of this approach were inspired by Willman, Rixner, and Cox's 2006 USENIX paper, "An Evaluation of Network Stack Parallelization Strategies in Modern Operating Systems". However, there are also significant differences: we maintain the inpcb lock, rather than using the connection group lock for per-connection state. Likewise, the focus of this implementation is alignment with NIC packet distribution strategies such as RSS, rather than pure software strategies. Despite that focus, software distribution is supported through the parallel netisr implementation, and works well in configurations where the number of hardware threads is greater than the number of NIC input queues, such as in the RMI XLR threaded MIPS architecture. Another important difference is the continued maintenance of existing hash tables as "reservation tables" -- these are useful both to distinguish the resource allocation aspect of protocol name management and the more common-case lookup aspect. In configurations where connection tables are aligned with hardware hashes, it is desirable to use the traditional lookup tables for loopback or encapsulated traffic rather than take the expense of hardware hashes that are hard to implement efficiently in software (such as RSS Toeplitz). Connection group support is enabled by compiling "options PCBGROUP" into your kernel configuration; for the time being, this is an experimental feature, and hence is not enabled by default. Subject to the limited MFCability of change dependencies in inpcb, and its change to the inpcbinfo init function signature, this change in principle could be merged to FreeBSD 8.x. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-06-06 12:55:02 +00:00
/*
* Install in the reservation hash table for now, but don't yet
* install a connection group since the full 4-tuple isn't yet
* configured.
*/
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
inp->inp_lport = sc->sc_inc.inc_lport;
Implement a CPU-affine TCP and UDP connection lookup data structure, struct inpcbgroup. pcbgroups, or "connection groups", supplement the existing inpcbinfo connection hash table, which when pcbgroups are enabled, might now be thought of more usefully as a per-protocol 4-tuple reservation table. Connections are assigned to connection groups base on a hash of their 4-tuple; wildcard sockets require special handling, and are members of all connection groups. During a connection lookup, a per-connection group lock is employed rather than the global pcbinfo lock. By aligning connection groups with input path processing, connection groups take on an effective CPU affinity, especially when aligned with RSS work placement (see a forthcoming commit for details). This eliminates cache line migration associated with global, protocol-layer data structures in steady state TCP and UDP processing (with the exception of protocol-layer statistics; further commit to follow). Elements of this approach were inspired by Willman, Rixner, and Cox's 2006 USENIX paper, "An Evaluation of Network Stack Parallelization Strategies in Modern Operating Systems". However, there are also significant differences: we maintain the inpcb lock, rather than using the connection group lock for per-connection state. Likewise, the focus of this implementation is alignment with NIC packet distribution strategies such as RSS, rather than pure software strategies. Despite that focus, software distribution is supported through the parallel netisr implementation, and works well in configurations where the number of hardware threads is greater than the number of NIC input queues, such as in the RMI XLR threaded MIPS architecture. Another important difference is the continued maintenance of existing hash tables as "reservation tables" -- these are useful both to distinguish the resource allocation aspect of protocol name management and the more common-case lookup aspect. In configurations where connection tables are aligned with hardware hashes, it is desirable to use the traditional lookup tables for loopback or encapsulated traffic rather than take the expense of hardware hashes that are hard to implement efficiently in software (such as RSS Toeplitz). Connection group support is enabled by compiling "options PCBGROUP" into your kernel configuration; for the time being, this is an experimental feature, and hence is not enabled by default. Subject to the limited MFCability of change dependencies in inpcb, and its change to the inpcbinfo init function signature, this change in principle could be merged to FreeBSD 8.x. Reviewed by: bz Sponsored by: Juniper Networks, Inc.
2011-06-06 12:55:02 +00:00
if ((error = in_pcbinshash_nopcbgroup(inp)) != 0) {
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
/*
* Undo the assignments above if we failed to
* put the PCB on the hash lists.
*/
#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 (sc->sc_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
inp->in6p_laddr = in6addr_any;
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
else
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
#endif
inp->inp_laddr.s_addr = INADDR_ANY;
inp->inp_lport = 0;
Add more logging points for failures in syncache_socket() to report when a new socket couldn't be created because one of in_pcbinshash(), in6_pcbconnect() or in_pcbconnect() failed. Logging is conditional on net.inet.tcp.log_debug being enabled. MFC after: 1 week
2010-08-15 09:30:13 +00:00
if ((s = tcp_log_addrs(&sc->sc_inc, NULL, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: in_pcbinshash failed "
"with error %i\n",
s, __func__, error);
free(s, M_TCPLOG);
}
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_HASH_WUNLOCK(&V_tcbinfo);
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
goto abort;
}
#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 (sc->sc_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
struct inpcb *oinp = sotoinpcb(lso);
struct in6_addr laddr6;
Remove an unused variable and put the sockaddr_in6 onto the stack instead of malloc'ing it. Reviewed by: sam (mentor) Obtained from: DragonFlyBSD rev 1.6 (hsu)
2004-01-22 23:10:11 +00:00
struct sockaddr_in6 sin6;
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
/*
* Inherit socket options from the listening socket.
* Note that in6p_inputopts are not (and should not be)
* copied, since it stores previously received options and is
* used to detect if each new option is different than the
* previous one and hence should be passed to a user.
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 copied in6p_inputopts, a user would not be able to
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
* receive options just after calling the accept system call.
*/
inp->inp_flags |= oinp->inp_flags & INP_CONTROLOPTS;
if (oinp->in6p_outputopts)
inp->in6p_outputopts =
ip6_copypktopts(oinp->in6p_outputopts, M_NOWAIT);
Remove an unused variable and put the sockaddr_in6 onto the stack instead of malloc'ing it. Reviewed by: sam (mentor) Obtained from: DragonFlyBSD rev 1.6 (hsu)
2004-01-22 23:10:11 +00:00
sin6.sin6_family = AF_INET6;
sin6.sin6_len = sizeof(sin6);
sin6.sin6_addr = sc->sc_inc.inc6_faddr;
sin6.sin6_port = sc->sc_inc.inc_fport;
sin6.sin6_flowinfo = sin6.sin6_scope_id = 0;
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
laddr6 = inp->in6p_laddr;
if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr))
inp->in6p_laddr = sc->sc_inc.inc6_laddr;
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
if ((error = in6_pcbconnect_mbuf(inp, (struct sockaddr *)&sin6,
thread0.td_ucred, m)) != 0) {
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
inp->in6p_laddr = laddr6;
Add more logging points for failures in syncache_socket() to report when a new socket couldn't be created because one of in_pcbinshash(), in6_pcbconnect() or in_pcbconnect() failed. Logging is conditional on net.inet.tcp.log_debug being enabled. MFC after: 1 week
2010-08-15 09:30:13 +00:00
if ((s = tcp_log_addrs(&sc->sc_inc, NULL, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: in6_pcbconnect failed "
"with error %i\n",
s, __func__, error);
free(s, M_TCPLOG);
}
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_HASH_WUNLOCK(&V_tcbinfo);
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
goto abort;
}
The tcp syncache code was leaving the IPv6 flowlabel uninitialised for the SYN|ACK packet and then letting in6_pcbconnect set the flowlabel later. Arange for the syncache/syncookie code to set and recall the flow label so that the flowlabel used for the SYN|ACK is consistent. This is done by using some of the cookie (when tcp cookies are enabeled) and by stashing the flowlabel in syncache. Tested and Discovered by: Orla McGann <orly@cnri.dit.ie> Approved by: ume, silby MFC after: 1 month
2004-07-17 19:44:13 +00:00
/* Override flowlabel from in6_pcbconnect. */
Another step assimilating IPv[46] PCB code - directly use the inpcb names rather than the following IPv6 compat macros: in6pcb,in6p_sp, in6p_ip6_nxt,in6p_flowinfo,in6p_vflag, in6p_flags,in6p_socket,in6p_lport,in6p_fport,in6p_ppcb and sotoin6pcb(). Apart from removing duplicate code in netipsec, this is a pure whitespace, not a functional change. Discussed with: rwatson Reviewed by: rwatson (version before review requested changes) MFC after: 4 weeks (set the timer and see then)
2008-12-15 21:50:54 +00:00
inp->inp_flow &= ~IPV6_FLOWLABEL_MASK;
inp->inp_flow |= sc->sc_flowlabel;
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 */
#if defined(INET) && defined(INET6)
else
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
#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 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 in_addr laddr;
Remove an unused variable and put the sockaddr_in6 onto the stack instead of malloc'ing it. Reviewed by: sam (mentor) Obtained from: DragonFlyBSD rev 1.6 (hsu)
2004-01-22 23:10:11 +00:00
struct sockaddr_in sin;
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
incorporate feedback since initial commit - rename tcp_ofld.[ch] to tcp_offload.[ch] - document usage and locking conventions of the functions in the toe_usrreqs function vector - document tcpcb, inpcb, and socket fields used by toe - widen the listen interface into 2 functions - rename DISABLE_TCP_OFFLOAD to TCP_OFFLOAD_DISABLE - shrink conditional compilation to reduce the likelihood of bitrot - replace sc->sc_toepcb checks in tcp_syncache.c with TOEPCB_ISSET
2007-12-17 07:56:27 +00:00
inp->inp_options = (m) ? ip_srcroute(m) : 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
if (inp->inp_options == NULL) {
inp->inp_options = sc->sc_ipopts;
sc->sc_ipopts = NULL;
}
Remove an unused variable and put the sockaddr_in6 onto the stack instead of malloc'ing it. Reviewed by: sam (mentor) Obtained from: DragonFlyBSD rev 1.6 (hsu)
2004-01-22 23:10:11 +00:00
sin.sin_family = AF_INET;
sin.sin_len = sizeof(sin);
sin.sin_addr = sc->sc_inc.inc_faddr;
sin.sin_port = sc->sc_inc.inc_fport;
bzero((caddr_t)sin.sin_zero, sizeof(sin.sin_zero));
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
laddr = inp->inp_laddr;
if (inp->inp_laddr.s_addr == INADDR_ANY)
inp->inp_laddr = sc->sc_inc.inc_laddr;
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
if ((error = in_pcbconnect_mbuf(inp, (struct sockaddr *)&sin,
thread0.td_ucred, m)) != 0) {
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
inp->inp_laddr = laddr;
Add more logging points for failures in syncache_socket() to report when a new socket couldn't be created because one of in_pcbinshash(), in6_pcbconnect() or in_pcbconnect() failed. Logging is conditional on net.inet.tcp.log_debug being enabled. MFC after: 1 week
2010-08-15 09:30:13 +00:00
if ((s = tcp_log_addrs(&sc->sc_inc, NULL, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: in_pcbconnect failed "
"with error %i\n",
s, __func__, error);
free(s, M_TCPLOG);
}
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_HASH_WUNLOCK(&V_tcbinfo);
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
goto abort;
}
}
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 */
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC) || defined(IPSEC_SUPPORT)
/* Copy old policy into new socket's. */
if (ipsec_copy_pcbpolicy(sotoinpcb(lso), inp) != 0)
printf("syncache_socket: could not copy policy\n");
#endif
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_HASH_WUNLOCK(&V_tcbinfo);
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 = intotcpcb(inp);
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);
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->iss = sc->sc_iss;
tp->irs = sc->sc_irs;
tcp_rcvseqinit(tp);
tcp_sendseqinit(tp);
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
blk = sototcpcb(lso)->t_fb;
Add a knob to change default behavior of inheriting listen socket's tcp stack regardless of what the default stack for the system is set to. With current/default behavior, after changing the default tcp stack, the application needs to be restarted to pick up that change. Setting this new knob net.inet.tcp.functions_inherit_listen_socket_stack to '0' would change that behavior and make any new connection use the newly selected default tcp stack. Reviewed by: rrs MFC after: 2 weeks Sponsored by: Limelight Networks
2017-01-27 23:10:46 +00:00
if (V_functions_inherit_listen_socket_stack && blk != tp->t_fb) {
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
/*
* Our parents t_fb was not the default,
* we need to release our ref on tp->t_fb and
* pickup one on the new entry.
*/
struct tcp_function_block *rblk;
rblk = find_and_ref_tcp_fb(blk);
KASSERT(rblk != NULL,
("cannot find blk %p out of syncache?", blk));
if (tp->t_fb->tfb_tcp_fb_fini)
Here we update the modular tcp to be able to switch to an alternate TCP stack in other then the closed state (pre-listen/connect). The idea is that *if* that is supported by the alternate stack, it is asked if its ok to switch. If it approves the "handoff" then we allow the switch to happen. Also the fini() function now gets a flag to tell if you are switching away *or* the tcb is destroyed. The init() call into the alternate stack is moved to the end so the tcb is more fully formed before the init transpires. Sponsored by: Netflix Inc. Differential Revision: D6790
2016-08-16 15:11:46 +00:00
(*tp->t_fb->tfb_tcp_fb_fini)(tp, 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
refcount_release(&tp->t_fb->tfb_refcnt);
tp->t_fb = rblk;
if (tp->t_fb->tfb_tcp_fb_init) {
(*tp->t_fb->tfb_tcp_fb_init)(tp);
}
}
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->snd_wl1 = sc->sc_irs;
Move last tcpcb initialization for the inbound connection case from tcp_input() to syncache_socket() where it belongs and the majority of it already happens. The "tp->snd_up = tp->snd_una" is removed as it is done with the tcp_sendseqinit() macro a few lines earlier.
2007-04-04 16:13:45 +00:00
tp->snd_max = tp->iss + 1;
tp->snd_nxt = tp->iss + 1;
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->rcv_up = sc->sc_irs + 1;
tp->rcv_wnd = sc->sc_wnd;
tp->rcv_adv += tp->rcv_wnd;
Move last tcpcb initialization for the inbound connection case from tcp_input() to syncache_socket() where it belongs and the majority of it already happens. The "tp->snd_up = tp->snd_una" is removed as it is done with the tcp_sendseqinit() macro a few lines earlier.
2007-04-04 16:13:45 +00:00
tp->last_ack_sent = tp->rcv_nxt;
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
When expanding a syncache entry into a socket, inherit the socket options from the current listen socket instead of the cached (and possibly stale) TCB pointer.
2002-02-20 16:47:11 +00:00
tp->t_flags = sototcpcb(lso)->t_flags & (TF_NOPUSH|TF_NODELAY);
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 (sc->sc_flags & SCF_NOOPT)
tp->t_flags |= TF_NOOPT;
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
else {
if (sc->sc_flags & SCF_WINSCALE) {
tp->t_flags |= TF_REQ_SCALE|TF_RCVD_SCALE;
tp->snd_scale = sc->sc_requested_s_scale;
tp->request_r_scale = sc->sc_requested_r_scale;
}
if (sc->sc_flags & SCF_TIMESTAMP) {
tp->t_flags |= TF_REQ_TSTMP|TF_RCVD_TSTMP;
tp->ts_recent = sc->sc_tsreflect;
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();
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
tp->ts_offset = sc->sc_tsoff;
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
}
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC_SUPPORT) || defined(TCP_SIGNATURE)
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
if (sc->sc_flags & SCF_SIGNATURE)
tp->t_flags |= TF_SIGNATURE;
Brucification. Submitted by: bde
2004-02-13 18:21:45 +00:00
#endif
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 (sc->sc_flags & SCF_SACK)
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
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +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
if (sc->sc_flags & SCF_ECN)
tp->t_flags |= TF_ECN_PERMIT;
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
/*
* Set up MSS and get cached values from tcp_hostcache.
* This might overwrite some of the defaults we just set.
*/
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
tcp_mss(tp, sc->sc_peer_mss);
/*
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 the SYN,ACK was retransmitted, indicate that CWND to be
* limited to one segment in cc_conn_init().
Fix a bug in syncache where the initial CWND for new incoming connections was limited to one segment under the faulty assumption of a retransmit. Due to this the opportunity to initialize the increased congestion window according to RFC3390 was missed. Support for RFC3465 introduced in r187289 uncovered the bug as the ACK to SYN/ACK no longer caused snd_cwnd increase by MSS (actually, this increase shouldn't happen as it's explicitly forbidden by RFC3390, but it's another issue). Snd_cwnd remains really small (1*MSS + 1) and this causes really bad interaction with delayed acks on other side. The variable name sc_rxmits is a bit misleading as it counts all transmits, not just retransmits. Submitted by: Maxim Dounin <mdounin-at-mdounin-dot-ru> MFC after: 10 days
2010-07-30 21:45:53 +00:00
* NB: sc_rxmits counts all SYN,ACK transmits, not just retransmits.
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
*/
Fix a bug in syncache where the initial CWND for new incoming connections was limited to one segment under the faulty assumption of a retransmit. Due to this the opportunity to initialize the increased congestion window according to RFC3390 was missed. Support for RFC3465 introduced in r187289 uncovered the bug as the ACK to SYN/ACK no longer caused snd_cwnd increase by MSS (actually, this increase shouldn't happen as it's explicitly forbidden by RFC3390, but it's another issue). Snd_cwnd remains really small (1*MSS + 1) and this causes really bad interaction with delayed acks on other side. The variable name sc_rxmits is a bit misleading as it counts all transmits, not just retransmits. Submitted by: Maxim Dounin <mdounin-at-mdounin-dot-ru> MFC after: 10 days
2010-07-30 21:45:53 +00:00
if (sc->sc_rxmits > 1)
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
tp->snd_cwnd = 1;
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
- 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
/*
* Allow a TOE driver to install its hooks. Note that we hold the
* pcbinfo lock too and that prevents tcp_usr_accept from accepting a
* new connection before the TOE driver has done its thing.
*/
if (ADDED_BY_TOE(sc)) {
struct toedev *tod = sc->sc_tod;
tod->tod_offload_socket(tod, sc->sc_todctx, so);
}
#endif
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
/*
* Copy and activate timers.
*/
tp->t_keepinit = sototcpcb(lso)->t_keepinit;
tp->t_keepidle = sototcpcb(lso)->t_keepidle;
tp->t_keepintvl = sototcpcb(lso)->t_keepintvl;
tp->t_keepcnt = sototcpcb(lso)->t_keepcnt;
tcp_timer_activate(tp, TT_KEEP, TP_KEEPINIT(tp));
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
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_accepts);
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
return (so);
abort:
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);
o add missing inpcb locking in tcp_respond o replace spl's with lock assertions Supported by: FreeBSD Foundation
2003-11-11 17:54:47 +00:00
abort2:
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 (so != NULL)
Change soabort() from returning int to returning void, since all consumers ignore the return value, soabort() is required to succeed, and protocols produce errors here to report multiple freeing of the pcb, which we hope to eliminate.
2006-03-16 07:03:14 +00:00
soabort(so);
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
return (NULL);
}
/*
* This function gets called when we receive an ACK for a
* socket in the LISTEN state. We look up the connection
* in the syncache, and if its there, we pull it out of
* the cache and turn it into a full-blown connection in
* the SYN-RECEIVED state.
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
*
* On syncache_socket() success the newly created socket
* has its underlying inp locked.
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
*/
int
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
syncache_expand(struct in_conninfo *inc, struct tcpopt *to, struct tcphdr *th,
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
struct socket **lsop, struct mbuf *m)
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 syncache *sc;
struct syncache_head *sch;
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
struct syncache scs;
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
char *s;
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/*
* Global TCP locks are held because we manipulate the PCB lists
* and create a new socket.
*/
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);
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
KASSERT((th->th_flags & (TH_RST|TH_ACK|TH_SYN)) == TH_ACK,
("%s: can handle only ACK", __func__));
o add missing inpcb locking in tcp_respond o replace spl's with lock assertions Supported by: FreeBSD Foundation
2003-11-11 17:54:47 +00:00
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sc = syncache_lookup(inc, &sch); /* returns locked sch */
SCH_LOCK_ASSERT(sch);
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
#ifdef INVARIANTS
/*
* Test code for syncookies comparing the syncache stored
* values with the reconstructed values from the cookie.
*/
if (sc != NULL)
syncookie_cmp(inc, sch, sc, th, to, *lsop);
#endif
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
if (sc == NULL) {
/*
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
* There is no syncache entry, so see if this ACK is
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
* a returning syncookie. To do this, first:
Syncoockies can be used in combination with the syncache. If the cache overflows, syncookies are used. This patch restricts the usage of syncookies in this case: accept syncookies only if there was an overflow of the syncache recently. This mitigates a problem reported in PR217637, where is syncookie was accepted without any recent drops. Thanks to glebius@ for suggesting an improvement. PR: 217637 Reviewed by: gnn, glebius MFC after: 1 week Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D10272
2017-04-20 19:19:33 +00:00
* A. Check if syncookies are used in case of syncache
* overflows
* B. See if this socket has had a syncache entry dropped in
* the recent past. We don't want to accept a bogus
* syncookie if we've never received a SYN or accept it
* twice.
* C. check that the syncookie is valid. If it is, then
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
* cobble up a fake syncache entry, and return.
*/
Merge more of currently non-functional (i.e. resolving to whitespace) macros from p4/vimage branch. Do a better job at enclosing all instantiations of globals scheduled for virtualization in #ifdef VIMAGE_GLOBALS blocks. De-virtualize and mark as const saorder_state_alive and saorder_state_any arrays from ipsec code, given that they are never updated at runtime, so virtualizing them would be pointless. 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
2008-11-26 22:32:07 +00:00
if (!V_tcp_syncookies) {
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
SCH_UNLOCK(sch);
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
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
log(LOG_DEBUG, "%s; %s: Spurious ACK, "
"segment rejected (syncookies disabled)\n",
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
s, __func__);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
goto failed;
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
}
Syncoockies can be used in combination with the syncache. If the cache overflows, syncookies are used. This patch restricts the usage of syncookies in this case: accept syncookies only if there was an overflow of the syncache recently. This mitigates a problem reported in PR217637, where is syncookie was accepted without any recent drops. Thanks to glebius@ for suggesting an improvement. PR: 217637 Reviewed by: gnn, glebius MFC after: 1 week Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D10272
2017-04-20 19:19:33 +00:00
if (!V_tcp_syncookiesonly &&
sch->sch_last_overflow < time_uptime - SYNCOOKIE_LIFETIME) {
SCH_UNLOCK(sch);
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Spurious ACK, "
"segment rejected (no syncache entry)\n",
s, __func__);
goto failed;
}
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
bzero(&scs, sizeof(scs));
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
sc = syncookie_lookup(inc, sch, &scs, th, to, *lsop);
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
SCH_UNLOCK(sch);
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
if (sc == NULL) {
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
log(LOG_DEBUG, "%s; %s: Segment failed "
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
"SYNCOOKIE authentication, segment rejected "
"(probably spoofed)\n", s, __func__);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
goto failed;
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
}
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC_SUPPORT) || defined(TCP_SIGNATURE)
/* If received ACK has MD5 signature, check it. */
if ((to->to_flags & TOF_SIGNATURE) != 0 &&
(!TCPMD5_ENABLED() ||
TCPMD5_INPUT(m, th, to->to_signature) != 0)) {
/* Drop the ACK. */
if ((s = tcp_log_addrs(inc, th, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: Segment rejected, "
"MD5 signature doesn't match.\n",
s, __func__);
free(s, M_TCPLOG);
}
TCPSTAT_INC(tcps_sig_err_sigopt);
return (-1); /* Do not send RST */
}
#endif /* TCP_SIGNATURE */
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
} else {
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC_SUPPORT) || defined(TCP_SIGNATURE)
/*
* If listening socket requested TCP digests, check that
* received ACK has signature and it is correct.
* If not, drop the ACK and leave sc entry in th cache,
* because SYN was received with correct signature.
*/
if (sc->sc_flags & SCF_SIGNATURE) {
if ((to->to_flags & TOF_SIGNATURE) == 0) {
/* No signature */
TCPSTAT_INC(tcps_sig_err_nosigopt);
SCH_UNLOCK(sch);
if ((s = tcp_log_addrs(inc, th, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: Segment "
"rejected, MD5 signature wasn't "
"provided.\n", s, __func__);
free(s, M_TCPLOG);
}
return (-1); /* Do not send RST */
}
if (!TCPMD5_ENABLED() ||
TCPMD5_INPUT(m, th, to->to_signature) != 0) {
/* Doesn't match or no SA */
SCH_UNLOCK(sch);
if ((s = tcp_log_addrs(inc, th, NULL, NULL))) {
log(LOG_DEBUG, "%s; %s: Segment "
"rejected, MD5 signature doesn't "
"match.\n", s, __func__);
free(s, M_TCPLOG);
}
return (-1); /* Do not send RST */
}
}
#endif /* TCP_SIGNATURE */
Augment struct tcpstat with tcps_states[], which is used for book-keeping the amount of TCP connections by state. Provides a cheap way to get connection count without traversing the whole pcb list. Sponsored by: Netflix
2016-01-27 00:45:46 +00:00
/*
* Pull out the entry to unlock the bucket row.
*
* NOTE: We must decrease TCPS_SYN_RECEIVED count here, not
* tcp_state_change(). The tcpcb is not existent at this
* moment. A new one will be allocated via syncache_socket->
* sonewconn->tcp_usr_attach in TCPS_CLOSED state, then
* syncache_socket() will change it to TCPS_SYN_RECEIVED.
*/
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
TCPSTATES_DEC(TCPS_SYN_RECEIVED);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
TAILQ_REMOVE(&sch->sch_bucket, sc, sc_hash);
sch->sch_length--;
- 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 (ADDED_BY_TOE(sc)) {
struct toedev *tod = sc->sc_tod;
tod->tod_syncache_removed(tod, sc->sc_todctx);
}
#endif
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
SCH_UNLOCK(sch);
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +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
/*
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
* Segment validation:
* ACK must match our initial sequence number + 1 (the SYN|ACK).
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
*/
- 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
if (th->th_ack != sc->sc_iss + 1) {
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
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
log(LOG_DEBUG, "%s; %s: ACK %u != ISS+1 %u, segment "
"rejected\n", s, __func__, th->th_ack, sc->sc_iss);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
goto failed;
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
}
Minor style tweaks.
2008-08-05 21:59:20 +00:00
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
/*
Minor style tweaks.
2008-08-05 21:59:20 +00:00
* The SEQ must fall in the window starting at the received
* initial receive sequence number + 1 (the SYN).
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +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
if (SEQ_LEQ(th->th_seq, sc->sc_irs) ||
SEQ_GT(th->th_seq, sc->sc_irs + sc->sc_wnd)) {
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
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
log(LOG_DEBUG, "%s; %s: SEQ %u != IRS+1 %u, segment "
Correctly print SEQ and IRS in the corresponding log message in syncache_expand().
2007-06-06 22:10:12 +00:00
"rejected\n", s, __func__, th->th_seq, sc->sc_irs);
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
goto failed;
}
Remove spurious timestamp check. RFC 1323 explicitly states that timestamps MAY be transmitted if negotiated.
2007-12-12 06:11:50 +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 timestamps were not negotiated during SYN/ACK they
* must not appear on any segment during this session.
*/
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
if (!(sc->sc_flags & SCF_TIMESTAMP) && (to->to_flags & TOF_TS)) {
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
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
log(LOG_DEBUG, "%s; %s: Timestamp not expected, "
"segment rejected\n", s, __func__);
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
goto failed;
}
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.
* XXXAO: This is only informal as there have been unverified
* reports of non-compliants stacks.
*/
if ((sc->sc_flags & SCF_TIMESTAMP) && !(to->to_flags & TOF_TS)) {
Free the non-fatal "timestamp missing" debug string manually as it is not covered by the catch-all free for the error cases. Found by: Coverity
2013-07-16 16:37:08 +00:00
if ((s = tcp_log_addrs(inc, th, NULL, NULL))) {
Extend debug logging of TCP timestamp related specification violations. Update related comments and style.
2013-07-10 12:06:01 +00:00
log(LOG_DEBUG, "%s; %s: Timestamp missing, "
"no action\n", s, __func__);
Free the non-fatal "timestamp missing" debug string manually as it is not covered by the catch-all free for the error cases. Found by: Coverity
2013-07-16 16:37:08 +00:00
free(s, M_TCPLOG);
s = NULL;
}
Extend debug logging of TCP timestamp related specification violations. Update related comments and style.
2013-07-10 12:06:01 +00:00
}
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
/*
For RTT calculations mid-session, we explicitly ignore ACKs with tsecr of 0 as many borken middle-boxes tend to do that. But during 3whs, in syncache_expand(), we don't do that which causes us to send a RST to such a client. Relax this constraint by only using tsecr to compare against timestamp that we sent when it is not 0. As a result, we'd now accept the final ACK of 3whs with tsecr of 0. Reviewed by: jtl, gnn Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D8552
2016-11-21 20:53:11 +00:00
* If timestamps were negotiated, the reflected timestamp
* must be equal to what we actually sent in the SYN|ACK
* except in the case of 0. Some boxes are known for sending
* broken timestamp replies during the 3whs (and potentially
* during the connection also).
*
* Accept the final ACK of 3whs with reflected timestamp of 0
* instead of sending a RST and deleting the syncache entry.
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
*/
For RTT calculations mid-session, we explicitly ignore ACKs with tsecr of 0 as many borken middle-boxes tend to do that. But during 3whs, in syncache_expand(), we don't do that which causes us to send a RST to such a client. Relax this constraint by only using tsecr to compare against timestamp that we sent when it is not 0. As a result, we'd now accept the final ACK of 3whs with tsecr of 0. Reviewed by: jtl, gnn Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D8552
2016-11-21 20:53:11 +00:00
if ((to->to_flags & TOF_TS) && to->to_tsecr &&
to->to_tsecr != sc->sc_ts) {
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
if ((s = tcp_log_addrs(inc, th, NULL, NULL)))
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
log(LOG_DEBUG, "%s; %s: TSECR %u != TS %u, "
"segment rejected\n",
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
s, __func__, to->to_tsecr, sc->sc_ts);
goto failed;
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
Simplifly syncache_expand() and clarify its semantics. Zero is returned when the ACK is invalid and doesn't belong to any registered connection, either in syncache or through SYN cookies. True but a NULL struct socket is returned when the 3WHS completed but the socket could not be created due to insufficient resources or limits reached. For both cases an RST is sent back in tcp_input(). A logic error leading to a panic is fixed where syncache_expand() would free the mbuf on socket allocation failure but tcp_input() later supplies it to tcp_dropwithreset() to issue a RST to the peer. Reported by: kris (the panic)
2007-04-20 13:51:34 +00:00
*lsop = syncache_socket(sc, *lsop, m);
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
Simplifly syncache_expand() and clarify its semantics. Zero is returned when the ACK is invalid and doesn't belong to any registered connection, either in syncache or through SYN cookies. True but a NULL struct socket is returned when the 3WHS completed but the socket could not be created due to insufficient resources or limits reached. For both cases an RST is sent back in tcp_input(). A logic error leading to a panic is fixed where syncache_expand() would free the mbuf on socket allocation failure but tcp_input() later supplies it to tcp_dropwithreset() to issue a RST to the peer. Reported by: kris (the panic)
2007-04-20 13:51:34 +00:00
if (*lsop == 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_sc_aborted);
Simplifly syncache_expand() and clarify its semantics. Zero is returned when the ACK is invalid and doesn't belong to any registered connection, either in syncache or through SYN cookies. True but a NULL struct socket is returned when the 3WHS completed but the socket could not be created due to insufficient resources or limits reached. For both cases an RST is sent back in tcp_input(). A logic error leading to a panic is fixed where syncache_expand() would free the mbuf on socket allocation failure but tcp_input() later supplies it to tcp_dropwithreset() to issue a RST to the peer. Reported by: kris (the panic)
2007-04-20 13:51:34 +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_sc_completed);
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
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
/* how do we find the inp for the new socket? */
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 (sc != &scs)
syncache_free(sc);
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
return (1);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
failed:
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 (sc != NULL && sc != &scs)
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
syncache_free(sc);
o Add syslog logging under LOG_DEBUG to various failures caused by bogus segments o Add more KASSERT()s o Update comments
2007-05-18 21:13:01 +00:00
if (s != NULL)
free(s, M_TCPLOG);
Simplifly syncache_expand() and clarify its semantics. Zero is returned when the ACK is invalid and doesn't belong to any registered connection, either in syncache or through SYN cookies. True but a NULL struct socket is returned when the 3WHS completed but the socket could not be created due to insufficient resources or limits reached. For both cases an RST is sent back in tcp_input(). A logic error leading to a panic is fixed where syncache_expand() would free the mbuf on socket allocation failure but tcp_input() later supplies it to tcp_dropwithreset() to issue a RST to the peer. Reported by: kris (the panic)
2007-04-20 13:51:34 +00:00
*lsop = NULL;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
return (0);
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
}
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
static void
syncache_tfo_expand(struct syncache *sc, struct socket **lsop, struct mbuf *m,
uint64_t response_cookie)
{
struct inpcb *inp;
struct tcpcb *tp;
unsigned int *pending_counter;
/*
* Global TCP locks are held because we manipulate the PCB lists
* and create a new socket.
*/
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
pending_counter = intotcpcb(sotoinpcb(*lsop))->t_tfo_pending;
*lsop = syncache_socket(sc, *lsop, m);
if (*lsop == NULL) {
TCPSTAT_INC(tcps_sc_aborted);
atomic_subtract_int(pending_counter, 1);
} else {
inp = sotoinpcb(*lsop);
tp = intotcpcb(inp);
tp->t_flags |= TF_FASTOPEN;
tp->t_tfo_cookie = response_cookie;
tp->snd_max = tp->iss;
tp->snd_nxt = tp->iss;
tp->t_tfo_pending = pending_counter;
TCPSTAT_INC(tcps_sc_completed);
}
}
#endif /* TCP_RFC7413 */
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
/*
* Given a LISTEN socket and an inbound SYN request, add
* this to the syn cache, and send back a segment:
* <SEQ=ISS><ACK=RCV_NXT><CTL=SYN,ACK>
* to the source.
*
* IMPORTANT NOTE: We do _NOT_ ACK data that might accompany the SYN.
* Doing so would require that we hold onto the data and deliver it
* to the application. However, if we are the target of a SYN-flood
* DoS attack, an attacker could send data which would eventually
* consume all available buffer space if it were ACKed. By not ACKing
* the data, we avoid this DoS scenario.
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
*
* The exception to the above is when a SYN with a valid TCP Fast Open (TFO)
Fix cases where the TFO pending counter would leak references, and eventually, memory. Also renamed some tfo labels and added/reworked comments for clarity. Based on an initial patch from jtl. PR: 213424 Reviewed by: jtl MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D8235
2016-10-15 01:41:28 +00:00
* cookie is processed and a new socket is created. In this case, any data
* accompanying the SYN will be queued to the socket by tcp_input() and will
* be ACKed either when the application sends response data or the delayed
* ACK timer expires, whichever comes first.
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
*/
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
- 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(struct in_conninfo *inc, struct tcpopt *to, struct tcphdr *th,
struct inpcb *inp, struct socket **lsop, struct mbuf *m, void *tod,
void *todctx)
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 tcpcb *tp;
struct socket *so;
struct syncache *sc = NULL;
struct syncache_head *sch;
struct mbuf *ipopts = NULL;
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
u_int ltflags;
Remove assigned only variable.
2016-12-21 22:47:10 +00:00
int win, ip_ttl, ip_tos;
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
char *s;
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 rv = 0;
Fix the !INET6 compile. Reported by: alc
2006-06-17 18:42:07 +00:00
#ifdef INET6
int autoflowlabel = 0;
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#endif
#ifdef MAC
struct label *maclabel;
Fix the !INET6 compile. Reported by: alc
2006-06-17 18:42:07 +00:00
#endif
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
struct syncache scs;
Cache the cred locally in _syncache_add() while holding the locks, so we can be sure that it's valid. In case we abort early free it again else put it into the syncache. We need the cred in the syncache to be able to restrict what will be exportet by the sysctl helper function syncache_pcblist() (to netstat) within jails. PR: kern/126493 Reviewed by: rwatson (earlier versions) MFC after: 3 days
2008-08-23 14:22:12 +00:00
struct ucred *cred;
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
uint64_t tfo_response_cookie;
Fix cases where the TFO pending counter would leak references, and eventually, memory. Also renamed some tfo labels and added/reworked comments for clarity. Based on an initial patch from jtl. PR: 213424 Reviewed by: jtl MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D8235
2016-10-15 01:41:28 +00:00
unsigned int *tfo_pending = 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
int tfo_cookie_valid = 0;
int tfo_response_cookie_valid = 0;
#endif
o add missing inpcb locking in tcp_respond o replace spl's with lock assertions Supported by: FreeBSD Foundation
2003-11-11 17:54:47 +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(inp); /* listen socket */
Fix a panic introduced in rev 1.126. Approved by: re (rwatson)
2007-07-28 20:13:40 +00:00
KASSERT((th->th_flags & (TH_RST|TH_ACK|TH_SYN)) == TH_SYN,
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
("%s: unexpected tcp flags", __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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/*
* Combine all so/tp operations very early to drop the INP lock as
* soon as possible.
*/
so = *lsop;
Listening sockets improvements. o Separate fields of struct socket that belong to listening from fields that belong to normal dataflow, and unionize them. This shrinks the structure a bit. - Take out selinfo's from the socket buffers into the socket. The first reason is to support braindamaged scenario when a socket is added to kevent(2) and then listen(2) is cast on it. The second reason is that there is future plan to make socket buffers pluggable, so that for a dataflow socket a socket buffer can be changed, and in this case we also want to keep same selinfos through the lifetime of a socket. - Remove struct struct so_accf. Since now listening stuff no longer affects struct socket size, just move its fields into listening part of the union. - Provide sol_upcall field and enforce that so_upcall_set() may be called only on a dataflow socket, which has buffers, and for listening sockets provide solisten_upcall_set(). o Remove ACCEPT_LOCK() global. - Add a mutex to socket, to be used instead of socket buffer lock to lock fields of struct socket that don't belong to a socket buffer. - Allow to acquire two socket locks, but the first one must belong to a listening socket. - Make soref()/sorele() to use atomic(9). This allows in some situations to do soref() without owning socket lock. There is place for improvement here, it is possible to make sorele() also to lock optionally. - Most protocols aren't touched by this change, except UNIX local sockets. See below for more information. o Reduce copy-and-paste in kernel modules that accept connections from listening sockets: provide function solisten_dequeue(), and use it in the following modules: ctl(4), iscsi(4), ng_btsocket(4), ng_ksocket(4), infiniband, rpc. o UNIX local sockets. - Removal of ACCEPT_LOCK() global uncovered several races in the UNIX local sockets. Most races exist around spawning a new socket, when we are connecting to a local listening socket. To cover them, we need to hold locks on both PCBs when spawning a third one. This means holding them across sonewconn(). This creates a LOR between pcb locks and unp_list_lock. - To fix the new LOR, abandon the global unp_list_lock in favor of global unp_link_lock. Indeed, separating these two locks didn't provide us any extra parralelism in the UNIX sockets. - Now call into uipc_attach() may happen with unp_link_lock hold if, we are accepting, or without unp_link_lock in case if we are just creating a socket. - Another problem in UNIX sockets is that uipc_close() basicly did nothing for a listening socket. The vnode remained opened for connections. This is fixed by removing vnode in uipc_close(). Maybe the right way would be to do it for all sockets (not only listening), simply move the vnode teardown from uipc_detach() to uipc_close()? Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D9770
2017-06-08 21:30:34 +00:00
KASSERT(SOLISTENING(so), ("%s: %p not listening", __func__, so));
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 = sototcpcb(so);
Cache the cred locally in _syncache_add() while holding the locks, so we can be sure that it's valid. In case we abort early free it again else put it into the syncache. We need the cred in the syncache to be able to restrict what will be exportet by the sysctl helper function syncache_pcblist() (to netstat) within jails. PR: kern/126493 Reviewed by: rwatson (earlier versions) MFC after: 3 days
2008-08-23 14:22:12 +00:00
cred = crhold(so->so_cred);
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +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) &&
Another step assimilating IPv[46] PCB code - directly use the inpcb names rather than the following IPv6 compat macros: in6pcb,in6p_sp, in6p_ip6_nxt,in6p_flowinfo,in6p_vflag, in6p_flags,in6p_socket,in6p_lport,in6p_fport,in6p_ppcb and sotoin6pcb(). Apart from removing duplicate code in netipsec, this is a pure whitespace, not a functional change. Discussed with: rwatson Reviewed by: rwatson (version before review requested changes) MFC after: 4 weeks (set the timer and see then)
2008-12-15 21:50:54 +00:00
(inp->inp_flags & IN6P_AUTOFLOWLABEL))
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
autoflowlabel = 1;
#endif
ip_ttl = inp->inp_ip_ttl;
ip_tos = inp->inp_ip_tos;
Listening sockets improvements. o Separate fields of struct socket that belong to listening from fields that belong to normal dataflow, and unionize them. This shrinks the structure a bit. - Take out selinfo's from the socket buffers into the socket. The first reason is to support braindamaged scenario when a socket is added to kevent(2) and then listen(2) is cast on it. The second reason is that there is future plan to make socket buffers pluggable, so that for a dataflow socket a socket buffer can be changed, and in this case we also want to keep same selinfos through the lifetime of a socket. - Remove struct struct so_accf. Since now listening stuff no longer affects struct socket size, just move its fields into listening part of the union. - Provide sol_upcall field and enforce that so_upcall_set() may be called only on a dataflow socket, which has buffers, and for listening sockets provide solisten_upcall_set(). o Remove ACCEPT_LOCK() global. - Add a mutex to socket, to be used instead of socket buffer lock to lock fields of struct socket that don't belong to a socket buffer. - Allow to acquire two socket locks, but the first one must belong to a listening socket. - Make soref()/sorele() to use atomic(9). This allows in some situations to do soref() without owning socket lock. There is place for improvement here, it is possible to make sorele() also to lock optionally. - Most protocols aren't touched by this change, except UNIX local sockets. See below for more information. o Reduce copy-and-paste in kernel modules that accept connections from listening sockets: provide function solisten_dequeue(), and use it in the following modules: ctl(4), iscsi(4), ng_btsocket(4), ng_ksocket(4), infiniband, rpc. o UNIX local sockets. - Removal of ACCEPT_LOCK() global uncovered several races in the UNIX local sockets. Most races exist around spawning a new socket, when we are connecting to a local listening socket. To cover them, we need to hold locks on both PCBs when spawning a third one. This means holding them across sonewconn(). This creates a LOR between pcb locks and unp_list_lock. - To fix the new LOR, abandon the global unp_list_lock in favor of global unp_link_lock. Indeed, separating these two locks didn't provide us any extra parralelism in the UNIX sockets. - Now call into uipc_attach() may happen with unp_link_lock hold if, we are accepting, or without unp_link_lock in case if we are just creating a socket. - Another problem in UNIX sockets is that uipc_close() basicly did nothing for a listening socket. The vnode remained opened for connections. This is fixed by removing vnode in uipc_close(). Maybe the right way would be to do it for all sockets (not only listening), simply move the vnode teardown from uipc_detach() to uipc_close()? Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D9770
2017-06-08 21:30:34 +00:00
win = so->sol_sbrcv_hiwat;
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
ltflags = (tp->t_flags & (TF_NOOPT | TF_SIGNATURE));
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
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
The TFO server-side code contains some changes that are not conditioned on the TCP_RFC7413 kernel option. This change removes those few instructions from the packet processing path. While not strictly necessary, for the sake of consistency, I applied the new IS_FASTOPEN macro to all places in the packet processing path that used the (t_flags & TF_FASTOPEN) check. Reviewed by: hiren Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D8219
2016-10-12 19:06:50 +00:00
if (V_tcp_fastopen_enabled && IS_FASTOPEN(tp->t_flags) &&
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_tfo_pending != NULL) && (to->to_flags & TOF_FASTOPEN)) {
/*
* Limit the number of pending TFO connections to
* approximately half of the queue limit. This prevents TFO
* SYN floods from starving the service by filling the
* listen queue with bogus TFO connections.
*/
if (atomic_fetchadd_int(tp->t_tfo_pending, 1) <=
Listening sockets improvements. o Separate fields of struct socket that belong to listening from fields that belong to normal dataflow, and unionize them. This shrinks the structure a bit. - Take out selinfo's from the socket buffers into the socket. The first reason is to support braindamaged scenario when a socket is added to kevent(2) and then listen(2) is cast on it. The second reason is that there is future plan to make socket buffers pluggable, so that for a dataflow socket a socket buffer can be changed, and in this case we also want to keep same selinfos through the lifetime of a socket. - Remove struct struct so_accf. Since now listening stuff no longer affects struct socket size, just move its fields into listening part of the union. - Provide sol_upcall field and enforce that so_upcall_set() may be called only on a dataflow socket, which has buffers, and for listening sockets provide solisten_upcall_set(). o Remove ACCEPT_LOCK() global. - Add a mutex to socket, to be used instead of socket buffer lock to lock fields of struct socket that don't belong to a socket buffer. - Allow to acquire two socket locks, but the first one must belong to a listening socket. - Make soref()/sorele() to use atomic(9). This allows in some situations to do soref() without owning socket lock. There is place for improvement here, it is possible to make sorele() also to lock optionally. - Most protocols aren't touched by this change, except UNIX local sockets. See below for more information. o Reduce copy-and-paste in kernel modules that accept connections from listening sockets: provide function solisten_dequeue(), and use it in the following modules: ctl(4), iscsi(4), ng_btsocket(4), ng_ksocket(4), infiniband, rpc. o UNIX local sockets. - Removal of ACCEPT_LOCK() global uncovered several races in the UNIX local sockets. Most races exist around spawning a new socket, when we are connecting to a local listening socket. To cover them, we need to hold locks on both PCBs when spawning a third one. This means holding them across sonewconn(). This creates a LOR between pcb locks and unp_list_lock. - To fix the new LOR, abandon the global unp_list_lock in favor of global unp_link_lock. Indeed, separating these two locks didn't provide us any extra parralelism in the UNIX sockets. - Now call into uipc_attach() may happen with unp_link_lock hold if, we are accepting, or without unp_link_lock in case if we are just creating a socket. - Another problem in UNIX sockets is that uipc_close() basicly did nothing for a listening socket. The vnode remained opened for connections. This is fixed by removing vnode in uipc_close(). Maybe the right way would be to do it for all sockets (not only listening), simply move the vnode teardown from uipc_detach() to uipc_close()? Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D9770
2017-06-08 21:30:34 +00:00
(so->sol_qlimit / 2)) {
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 result;
result = tcp_fastopen_check_cookie(inc,
to->to_tfo_cookie, to->to_tfo_len,
&tfo_response_cookie);
tfo_cookie_valid = (result > 0);
tfo_response_cookie_valid = (result >= 0);
Fix cases where the TFO pending counter would leak references, and eventually, memory. Also renamed some tfo labels and added/reworked comments for clarity. Based on an initial patch from jtl. PR: 213424 Reviewed by: jtl MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D8235
2016-10-15 01:41:28 +00:00
}
/*
* Remember the TFO pending counter as it will have to be
* decremented below if we don't make it to syncache_tfo_expand().
*/
tfo_pending = tp->t_tfo_pending;
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
Add an explicit comment why we NULLify the two variables. Reviewed by: rwatson MFC after: 3 days
2008-08-23 12:27:18 +00:00
/* By the time we drop the lock these should no longer be used. */
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
so = NULL;
tp = NULL;
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#ifdef MAC
Normalize TCP syncache-related MAC Framework entry points to match most other entry points in the form mac_<object>_method(). Discussed with: csjp Obtained from: TrustedBSD Project
2007-10-25 14:37:37 +00:00
if (mac_syncache_init(&maclabel) != 0) {
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);
o Plug memory leak in syncache_add() on MAC label allocation failure. o Simplify code flow with 'done' goto label. o Remove mbuf argument from syncache_respond(). It doesn't make use of it.
2007-04-20 13:30:08 +00:00
goto done;
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
} else
Normalize TCP syncache-related MAC Framework entry points to match most other entry points in the form mac_<object>_method(). Discussed with: csjp Obtained from: TrustedBSD Project
2007-10-25 14:37:37 +00:00
mac_syncache_create(maclabel, inp);
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#endif
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 (!tfo_cookie_valid)
#endif
INP_WUNLOCK(inp);
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +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
/*
* Remember the IP options, if any.
*/
#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))
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
#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
incorporate feedback since initial commit - rename tcp_ofld.[ch] to tcp_offload.[ch] - document usage and locking conventions of the functions in the toe_usrreqs function vector - document tcpcb, inpcb, and socket fields used by toe - widen the listen interface into 2 functions - rename DISABLE_TCP_OFFLOAD to TCP_OFFLOAD_DISABLE - shrink conditional compilation to reduce the likelihood of bitrot - replace sc->sc_toepcb checks in tcp_syncache.c with TOEPCB_ISSET
2007-12-17 07:56:27 +00:00
ipopts = (m) ? ip_srcroute(m) : 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
#else
ipopts = NULL;
#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
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC_SUPPORT) || defined(TCP_SIGNATURE)
/*
* If listening socket requested TCP digests, check that received
* SYN has signature and it is correct. If signature doesn't match
* or TCP_SIGNATURE support isn't enabled, drop the packet.
*/
if (ltflags & TF_SIGNATURE) {
if ((to->to_flags & TOF_SIGNATURE) == 0) {
TCPSTAT_INC(tcps_sig_err_nosigopt);
goto done;
}
if (!TCPMD5_ENABLED() ||
TCPMD5_INPUT(m, th, to->to_signature) != 0)
goto done;
}
#endif /* TCP_SIGNATURE */
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
/*
* See if we already have an entry for this connection.
* If we do, resend the SYN,ACK, and reset the retransmit timer.
*
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
* XXX: should the syncache be re-initialized with the contents
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
* of the new SYN here (which may have different options?)
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
*
* XXX: We do not check the sequence number to see if this is a
* real retransmit or a new connection attempt. The question is
* how to handle such a case; either ignore it as spoofed, or
* drop the current entry and create a new one?
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
*/
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sc = syncache_lookup(inc, &sch); /* returns locked entry */
SCH_LOCK_ASSERT(sch);
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 (sc != 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
#ifdef TCP_RFC7413
if (tfo_cookie_valid)
INP_WUNLOCK(inp);
#endif
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_sc_dupsyn);
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 (ipopts) {
/*
* If we were remembering a previous source route,
* forget it and use the new one we've been given.
*/
if (sc->sc_ipopts)
(void) m_free(sc->sc_ipopts);
sc->sc_ipopts = ipopts;
}
/*
* Update timestamp if present.
*/
Only update TCP timestamp on SYN duplication if it is present on current SYN in syncache_add(). Otherwise disable timestamps.
2007-04-20 13:36:48 +00:00
if ((sc->sc_flags & SCF_TIMESTAMP) && (to->to_flags & TOF_TS))
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
sc->sc_tsreflect = to->to_tsval;
Only update TCP timestamp on SYN duplication if it is present on current SYN in syncache_add(). Otherwise disable timestamps.
2007-04-20 13:36:48 +00:00
else
sc->sc_flags &= ~SCF_TIMESTAMP;
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#ifdef MAC
/*
* Since we have already unconditionally allocated label
* storage, free it up. The syncache entry will already
* have an initialized label we can use.
*/
Normalize TCP syncache-related MAC Framework entry points to match most other entry points in the form mac_<object>_method(). Discussed with: csjp Obtained from: TrustedBSD Project
2007-10-25 14:37:37 +00:00
mac_syncache_destroy(&maclabel);
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#endif
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
/* Retransmit SYN|ACK and reset retransmit count. */
if ((s = tcp_log_addrs(&sc->sc_inc, th, NULL, NULL))) {
Fix a typo in a log message: s/Reveived/Received/. Approved by: re (rwatson)
2007-07-29 20:13:22 +00:00
log(LOG_DEBUG, "%s; %s: Received duplicate SYN, "
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
"resetting timer and retransmitting SYN|ACK\n",
s, __func__);
free(s, M_TCPLOG);
}
tcp/syncache: Set flowid and hash type properly for SYN|ACK So the underlying drivers can use it to select the sending queue properly for SYN|ACK instead of rolling their own hash. Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6120
2016-04-29 07:23:08 +00:00
if (syncache_respond(sc, sch, 1, m) == 0) {
o Move setting/resetting logic of syncache timer from macro SYNCACHE_TIMEOUT to new function syncache_timeout(). o Fix inverted timeout callout engagement logic to actually enable the timer for the bucket row. Before SYN|ACK was not retransmitted. o Simplify SYN|ACK retransmit timeout backoff calculation. o Improve logging of retransmit and timeout events. o Reset timeout when duplicate SYN arrives. o Add comments. o Rearrange SYN cookie statistics counting. Bug found by: silby Submitted by: silby (different version) Approved by: re (rwatson)
2007-07-28 12:02:05 +00:00
sc->sc_rxmits = 0;
syncache_timeout(sc, sch, 1);
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_sndacks);
TCPSTAT_INC(tcps_sndtotal);
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
SCH_UNLOCK(sch);
goto 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
}
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 (tfo_cookie_valid) {
bzero(&scs, sizeof(scs));
sc = &scs;
goto skip_alloc;
}
#endif
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
sc = uma_zalloc(V_tcp_syncache.zone, M_NOWAIT | M_ZERO);
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 (sc == NULL) {
/*
* The zone allocator couldn't provide more entries.
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
* Treat this as if the cache was full; drop the oldest
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
* entry and insert the new one.
*/
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_sc_zonefail);
Syncoockies can be used in combination with the syncache. If the cache overflows, syncookies are used. This patch restricts the usage of syncookies in this case: accept syncookies only if there was an overflow of the syncache recently. This mitigates a problem reported in PR217637, where is syncookie was accepted without any recent drops. Thanks to glebius@ for suggesting an improvement. PR: 217637 Reviewed by: gnn, glebius MFC after: 1 week Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D10272
2017-04-20 19:19:33 +00:00
if ((sc = TAILQ_LAST(&sch->sch_bucket, sch_head)) != NULL) {
sch->sch_last_overflow = time_uptime;
When we run into the syncache entry limits syncache_add() tries to free the oldest entry in the current bucket row. The global entry limit may be smaller than the bucket rows and their limit combined however. Thus only try to free a syncache entry if we found one in this bucket row. Reported by: kris
2007-04-17 15:25:14 +00:00
syncache_drop(sc, sch);
Syncoockies can be used in combination with the syncache. If the cache overflows, syncookies are used. This patch restricts the usage of syncookies in this case: accept syncookies only if there was an overflow of the syncache recently. This mitigates a problem reported in PR217637, where is syncookie was accepted without any recent drops. Thanks to glebius@ for suggesting an improvement. PR: 217637 Reviewed by: gnn, glebius MFC after: 1 week Sponsored by: Netflix, Inc. Differential Revision: https://reviews.freebsd.org/D10272
2017-04-20 19:19:33 +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
sc = uma_zalloc(V_tcp_syncache.zone, M_NOWAIT | M_ZERO);
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 (sc == NULL) {
Merge more of currently non-functional (i.e. resolving to whitespace) macros from p4/vimage branch. Do a better job at enclosing all instantiations of globals scheduled for virtualization in #ifdef VIMAGE_GLOBALS blocks. De-virtualize and mark as const saorder_state_alive and saorder_state_any arrays from ipsec code, given that they are never updated at runtime, so virtualizing them would be pointless. 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
2008-11-26 22:32:07 +00:00
if (V_tcp_syncookies) {
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
bzero(&scs, sizeof(scs));
sc = &scs;
} else {
SCH_UNLOCK(sch);
if (ipopts)
(void) m_free(ipopts);
goto 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
}
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
}
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
skip_alloc:
if (!tfo_cookie_valid && tfo_response_cookie_valid)
sc->sc_tfo_cookie = &tfo_response_cookie;
#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
/*
* Fill in the syncache values.
*/
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#ifdef MAC
sc->sc_label = maclabel;
#endif
Cache the cred locally in _syncache_add() while holding the locks, so we can be sure that it's valid. In case we abort early free it again else put it into the syncache. We need the cred in the syncache to be able to restrict what will be exportet by the sysctl helper function syncache_pcblist() (to netstat) within jails. PR: kern/126493 Reviewed by: rwatson (earlier versions) MFC after: 3 days
2008-08-23 14:22:12 +00:00
sc->sc_cred = cred;
cred = 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
sc->sc_ipopts = ipopts;
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
bcopy(inc, &sc->sc_inc, sizeof(struct in_conninfo));
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
#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))
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
#endif
{
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sc->sc_ip_tos = ip_tos;
sc->sc_ip_ttl = ip_ttl;
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
}
- 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
sc->sc_tod = tod;
sc->sc_todctx = todctx;
Add interface for tcp offload to syncache: - make neccessary changes to release offload resources when a syncache entry is removed before connection establishment - disable checks for offloaded connection where insufficient information is available Reviewed by: silby
2007-12-12 20:35:59 +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
sc->sc_irs = th->th_seq;
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
sc->sc_iss = arc4random();
Fix a bug with syncookies; previously, the syncache's MSS size was not initialized until after a syncookie was generated. As a result, all connections resulting from a returned cookie would end up using a MSS of ~512 bytes. Now larger packets will be used where possible. MFC after: 5 days
2003-01-29 03:49:49 +00:00
sc->sc_flags = 0;
The tcp syncache code was leaving the IPv6 flowlabel uninitialised for the SYN|ACK packet and then letting in6_pcbconnect set the flowlabel later. Arange for the syncache/syncookie code to set and recall the flow label so that the flowlabel used for the SYN|ACK is consistent. This is done by using some of the cookie (when tcp cookies are enabeled) and by stashing the flowlabel in syncache. Tested and Discovered by: Orla McGann <orly@cnri.dit.ie> Approved by: ume, silby MFC after: 1 month
2004-07-17 19:44:13 +00:00
sc->sc_flowlabel = 0;
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/*
* Initial receive window: clip sbspace to [0 .. TCP_MAXWIN].
* win was derived from socket earlier in the function.
*/
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
win = imax(win, 0);
win = imin(win, TCP_MAXWIN);
sc->sc_wnd = win;
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_rfc1323) {
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 timestamp received in a SYN makes
* it ok to send timestamp requests and replies.
*/
if (to->to_flags & TOF_TS) {
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
sc->sc_tsreflect = 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
sc->sc_ts = 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
sc->sc_flags |= SCF_TIMESTAMP;
}
if (to->to_flags & TOF_SCALE) {
int wscale = 0;
Change the way the advertized TCP window scaling is computed. Instead of upper-bounding it to the size of the initial socket buffer lower-bound it to the smallest MSS we accept. Ideally we'd use the actual MSS information here but it is not available yet. For socket buffer auto sizing to be effective we need room to grow the receive window. The window scale shift is determined at connection setup and can't be changed afterwards. The previous, original, method effectively just did a power of two roundup of the socket buffer size at connection setup severely limiting the headroom for larger socket buffers. Tested by: many (as part of the socket buffer auto sizing patch) MFC after: 1 month
2007-02-01 17:39:18 +00:00
/*
Pick the smallest possible TCP window scaling factor that will still allow us to scale up to sb_max, aka kern.ipc.maxsockbuf. We do this because there are broken firewalls that will corrupt the window scale option, leading to the other endpoint believing that our advertised window is unscaled. At scale factors larger than 5 the unscaled window will drop below 1500 bytes, leading to serious problems when traversing these broken firewalls. With the default maxsockbuf of 256K, a scale factor of 3 will be chosen by this algorithm. Those who choose a larger maxsockbuf should watch out for the compatiblity problems mentioned above. Reviewed by: andre
2007-10-19 08:53:14 +00:00
* Pick the smallest possible scaling factor that
* will still allow us to scale up to sb_max, aka
* kern.ipc.maxsockbuf.
*
* We do this because there are broken firewalls that
* will corrupt the window scale option, leading to
* the other endpoint believing that our advertised
* window is unscaled. At scale factors larger than
* 5 the unscaled window will drop below 1500 bytes,
* leading to serious problems when traversing these
* broken firewalls.
*
* With the default maxsockbuf of 256K, a scale factor
* of 3 will be chosen by this algorithm. Those who
* choose a larger maxsockbuf should watch out
sys/net*: minor spelling fixes. No functional change.
2016-05-03 18:05:43 +00:00
* for the compatibility problems mentioned above.
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
*
* RFC1323: The Window field in a SYN (i.e., a <SYN>
* or <SYN,ACK>) segment itself is never scaled.
Change the way the advertized TCP window scaling is computed. Instead of upper-bounding it to the size of the initial socket buffer lower-bound it to the smallest MSS we accept. Ideally we'd use the actual MSS information here but it is not available yet. For socket buffer auto sizing to be effective we need room to grow the receive window. The window scale shift is determined at connection setup and can't be changed afterwards. The previous, original, method effectively just did a power of two roundup of the socket buffer size at connection setup severely limiting the headroom for larger socket buffers. Tested by: many (as part of the socket buffer auto sizing patch) MFC after: 1 month
2007-02-01 17:39:18 +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
while (wscale < TCP_MAX_WINSHIFT &&
Pick the smallest possible TCP window scaling factor that will still allow us to scale up to sb_max, aka kern.ipc.maxsockbuf. We do this because there are broken firewalls that will corrupt the window scale option, leading to the other endpoint believing that our advertised window is unscaled. At scale factors larger than 5 the unscaled window will drop below 1500 bytes, leading to serious problems when traversing these broken firewalls. With the default maxsockbuf of 256K, a scale factor of 3 will be chosen by this algorithm. Those who choose a larger maxsockbuf should watch out for the compatiblity problems mentioned above. Reviewed by: andre
2007-10-19 08:53:14 +00:00
(TCP_MAXWIN << wscale) < sb_max)
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
wscale++;
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
sc->sc_requested_r_scale = wscale;
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
sc->sc_requested_s_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
sc->sc_flags |= SCF_WINSCALE;
}
}
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC_SUPPORT) || defined(TCP_SIGNATURE)
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
/*
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
* If listening socket requested TCP digests, flag this in the
* syncache so that syncache_respond() will do the right thing
* with the SYN+ACK.
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
*/
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
if (ltflags & TF_SIGNATURE)
Do not ignore all other TCP options (eg. timestamp, window scaling) when responding to TCP SYN packets with TCP_MD5 enabled and set. PR: kern/82963 Submitted by: <demizu at dd.iij4u.or.jp> MFC after: 3 days
2005-09-14 15:06:22 +00:00
sc->sc_flags |= SCF_SIGNATURE;
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#endif /* TCP_SIGNATURE */
Fix SACK negotiation that was broken in rev 1.105. Before this fix, FreeBSD would negotiate SACK on outgoing connections, but would always fail to negotiate it on incoming connections. Discovered by: James Healy and Lawrence Stewart Submitted by: James Healy and Lawrence Stewart MFC after: 3 days
2007-12-04 07:11:13 +00:00
if (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
sc->sc_flags |= SCF_SACK;
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_MSS)
sc->sc_peer_mss = to->to_mss; /* peer mss may be zero */
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
if (ltflags & TF_NOOPT)
Do not access syncache entry before it was allocated for the TF_NOOPT case in syncache_add(). Found by: Coverity Prevent CID: 1473
2006-06-18 13:03:42 +00:00
sc->sc_flags |= SCF_NOOPT;
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 ((th->th_flags & (TH_ECE|TH_CWR)) && 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
sc->sc_flags |= SCF_ECN;
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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
if (V_tcp_syncookies)
sc->sc_iss = syncookie_generate(sch, sc);
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
#ifdef INET6
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
if (autoflowlabel) {
if (V_tcp_syncookies)
sc->sc_flowlabel = sc->sc_iss;
else
sc->sc_flowlabel = ip6_randomflowlabel();
sc->sc_flowlabel = htonl(sc->sc_flowlabel) & IPV6_FLOWLABEL_MASK;
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
}
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
#endif
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
SCH_UNLOCK(sch);
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 (tfo_cookie_valid) {
syncache_tfo_expand(sc, lsop, m, tfo_response_cookie);
Fix cases where the TFO pending counter would leak references, and eventually, memory. Also renamed some tfo labels and added/reworked comments for clarity. Based on an initial patch from jtl. PR: 213424 Reviewed by: jtl MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D8235
2016-10-15 01:41:28 +00:00
/* INP_WUNLOCK(inp) will be performed by the caller */
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
rv = 1;
Fix cases where the TFO pending counter would leak references, and eventually, memory. Also renamed some tfo labels and added/reworked comments for clarity. Based on an initial patch from jtl. PR: 213424 Reviewed by: jtl MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D8235
2016-10-15 01:41:28 +00:00
goto tfo_expanded;
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
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 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 a standard 3-way handshake.
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
*/
tcp/syncache: Set flowid and hash type properly for SYN|ACK So the underlying drivers can use it to select the sending queue properly for SYN|ACK instead of rolling their own hash. Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6120
2016-04-29 07:23:08 +00:00
if (syncache_respond(sc, sch, 0, m) == 0) {
Merge more of currently non-functional (i.e. resolving to whitespace) macros from p4/vimage branch. Do a better job at enclosing all instantiations of globals scheduled for virtualization in #ifdef VIMAGE_GLOBALS blocks. De-virtualize and mark as const saorder_state_alive and saorder_state_any arrays from ipsec code, given that they are never updated at runtime, so virtualizing them would be pointless. 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
2008-11-26 22:32:07 +00:00
if (V_tcp_syncookies && V_tcp_syncookiesonly && sc != &scs)
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
syncache_free(sc);
else if (sc != &scs)
syncache_insert(sc, sch); /* locks and unlocks sch */
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_sndacks);
TCPSTAT_INC(tcps_sndtotal);
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
} else {
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
if (sc != &scs)
syncache_free(sc);
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_sc_dropped);
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
}
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
done:
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 (m) {
*lsop = NULL;
m_freem(m);
}
#ifdef TCP_RFC7413
Fix cases where the TFO pending counter would leak references, and eventually, memory. Also renamed some tfo labels and added/reworked comments for clarity. Based on an initial patch from jtl. PR: 213424 Reviewed by: jtl MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D8235
2016-10-15 01:41:28 +00:00
/*
* If tfo_pending is not NULL here, then a TFO SYN that did not
* result in a new socket was processed and the associated pending
* counter has not yet been decremented. All such TFO processing paths
* transit this point.
*/
if (tfo_pending != NULL)
tcp_fastopen_decrement_counter(tfo_pending);
tfo_expanded:
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
Cache the cred locally in _syncache_add() while holding the locks, so we can be sure that it's valid. In case we abort early free it again else put it into the syncache. We need the cred in the syncache to be able to restrict what will be exportet by the sysctl helper function syncache_pcblist() (to netstat) within jails. PR: kern/126493 Reviewed by: rwatson (earlier versions) MFC after: 3 days
2008-08-23 14:22:12 +00:00
if (cred != NULL)
crfree(cred);
o Plug memory leak in syncache_add() on MAC label allocation failure. o Simplify code flow with 'done' goto label. o Remove mbuf argument from syncache_respond(). It doesn't make use of it.
2007-04-20 13:30:08 +00:00
#ifdef MAC
if (sc == &scs)
Normalize TCP syncache-related MAC Framework entry points to match most other entry points in the form mac_<object>_method(). Discussed with: csjp Obtained from: TrustedBSD Project
2007-10-25 14:37:37 +00:00
mac_syncache_destroy(&maclabel);
o Plug memory leak in syncache_add() on MAC label allocation failure. o Simplify code flow with 'done' goto label. o Remove mbuf argument from syncache_respond(). It doesn't make use of it.
2007-04-20 13:30:08 +00:00
#endif
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
return (rv);
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
}
tcp/syncache: Add comment for syncache_respond Suggested by: hiren, hps Reviewed by: sbruno Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6148
2016-05-10 04:59:04 +00:00
/*
* Send SYN|ACK to the peer. Either in response to the peer's SYN,
* i.e. m0 != NULL, or upon 3WHS ACK timeout, i.e. m0 == 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
static int
tcp/syncache: Set flowid and hash type properly for SYN|ACK So the underlying drivers can use it to select the sending queue properly for SYN|ACK instead of rolling their own hash. Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6120
2016-04-29 07:23:08 +00:00
syncache_respond(struct syncache *sc, struct syncache_head *sch, int locked,
const struct mbuf *m0)
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 ip *ip = NULL;
o Plug memory leak in syncache_add() on MAC label allocation failure. o Simplify code flow with 'done' goto label. o Remove mbuf argument from syncache_respond(). It doesn't make use of it.
2007-04-20 13:30:08 +00:00
struct mbuf *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
struct tcphdr *th = NULL;
int optlen, error = 0; /* Make compiler happy */
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
u_int16_t hlen, tlen, mssopt;
struct tcpopt to;
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
#ifdef INET6
struct ip6_hdr *ip6 = NULL;
#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
hlen =
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
#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
(sc->sc_inc.inc_flags & INC_ISIPV6) ? sizeof(struct ip6_hdr) :
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
#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
sizeof(struct ip);
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
tlen = hlen + sizeof(struct tcphdr);
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
ANSIfy and tidy up comments. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:49:11 +00:00
/* Determine MSS we advertize to other end of connection. */
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
mssopt = tcp_mssopt(&sc->sc_inc);
In syncache_respond() do not reply with a MSS that is larger than what the peer announced to us but make it at least tcp_minmss in size. Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-26 17:54:53 +00:00
if (sc->sc_peer_mss)
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
mssopt = max( min(sc->sc_peer_mss, mssopt), V_tcp_minmss);
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
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
/* XXX: Assume that the entire packet will fit in a header mbuf. */
o Remove unused and redundant TCP option definitions o Replace usage of MAX_TCPOPTLEN with the correctly constructed and derived MAX_TCPOPTLEN
2007-04-20 15:08:09 +00:00
KASSERT(max_linkhdr + tlen + TCP_MAXOLEN <= MHLEN,
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
("syncache: mbuf too small"));
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
/* Create the IP+TCP header from scratch. */
Mechanically substitute flags from historic mbuf allocator with malloc(9) flags within sys. Exceptions: - sys/contrib not touched - sys/mbuf.h edited manually
2012-12-05 08:04:20 +00:00
m = m_gethdr(M_NOWAIT, MT_DATA);
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 (m == NULL)
return (ENOBUFS);
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +00:00
#ifdef MAC
Normalize TCP syncache-related MAC Framework entry points to match most other entry points in the form mac_<object>_method(). Discussed with: csjp Obtained from: TrustedBSD Project
2007-10-25 14:37:37 +00:00
mac_syncache_create_mbuf(sc->sc_label, m);
Fix LOR between the syncache and inpcb locks when MAC is present in the kernel. This LOR snuck in with some of the recent syncache changes. To fix this, the inpcb handling was changed: - Hang a MAC label off the syncache object - When the syncache entry is initially created, we pickup the PCB lock is held because we extract information from it while initializing the syncache entry. While we do this, copy the MAC label associated with the PCB and use it for the syncache entry. - When the packet is transmitted, copy the label from the syncache entry to the mbuf so it can be processed by security policies which analyze mbuf labels. This change required that the MAC framework be extended to support the label copy operations from the PCB to the syncache entry, and then from the syncache entry to the mbuf. These functions really should be referencing the syncache structure instead of the label. However, due to some of the complexities associated with exposing this syncache structure we operate directly on it's label pointer. This should be OK since we aren't making any access control decisions within this code directly, we are merely allocating and copying label storage so we can properly initialize mbuf labels for any packets the syncache code might create. This also has a nice side effect of caching. Prior to this change, the PCB would be looked up/locked for each packet transmitted. Now the label is cached at the time the syncache entry is initialized. Submitted by: andre [1] Discussed with: rwatson [1] andre submitted the tcp_syncache.c changes
2006-12-13 06:00:57 +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
m->m_data += max_linkhdr;
m->m_len = tlen;
m->m_pkthdr.len = tlen;
m->m_pkthdr.rcvif = NULL;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +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
#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 (sc->sc_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
ip6 = mtod(m, struct ip6_hdr *);
ip6->ip6_vfc = IPV6_VERSION;
ip6->ip6_nxt = IPPROTO_TCP;
ip6->ip6_src = sc->sc_inc.inc6_laddr;
ip6->ip6_dst = sc->sc_inc.inc6_faddr;
ip6->ip6_plen = htons(tlen - hlen);
/* ip6_hlim is set after checksum */
The tcp syncache code was leaving the IPv6 flowlabel uninitialised for the SYN|ACK packet and then letting in6_pcbconnect set the flowlabel later. Arange for the syncache/syncookie code to set and recall the flow label so that the flowlabel used for the SYN|ACK is consistent. This is done by using some of the cookie (when tcp cookies are enabeled) and by stashing the flowlabel in syncache. Tested and Discovered by: Orla McGann <orly@cnri.dit.ie> Approved by: ume, silby MFC after: 1 month
2004-07-17 19:44:13 +00:00
ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK;
ip6->ip6_flow |= sc->sc_flowlabel;
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
th = (struct tcphdr *)(ip6 + 1);
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 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
#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(INET6) && defined(INET)
else
#endif
#ifdef INET
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
{
ip = mtod(m, struct ip *);
ip->ip_v = IPVERSION;
ip->ip_hl = sizeof(struct ip) >> 2;
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_len = htons(tlen);
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
ip->ip_id = 0;
ip->ip_off = 0;
ip->ip_sum = 0;
ip->ip_p = IPPROTO_TCP;
ip->ip_src = sc->sc_inc.inc_laddr;
ip->ip_dst = sc->sc_inc.inc_faddr;
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
ip->ip_ttl = sc->sc_ip_ttl;
ip->ip_tos = sc->sc_ip_tos;
Re-commit w/fix: Ensure that the syn cache's syn-ack packets contain the same ip_tos, ip_ttl, and DF bits as all other tcp packets. PR: 39141 MFC after: 2 weeks This time, make sure that ipv4 specific code (aka all of the above) is only run in the ipv4 case.
2002-06-14 03:08:05 +00:00
/*
Eliminate a goto. Fix some line breaks.
2002-12-20 11:24:02 +00:00
* See if we should do MTU discovery. Route lookups are
* expensive, so we will only unset the DF bit if:
Handle PMTU discovery in syn-ack packets slightly differently; rely on syncache flags instead of directly accessing the route entry. MFC after: 3 days
2002-08-05 22:34:15 +00:00
*
* 1) path_mtu_discovery is disabled
* 2) the SCF_UNREACH flag has been set
Re-commit w/fix: Ensure that the syn cache's syn-ack packets contain the same ip_tos, ip_ttl, and DF bits as all other tcp packets. PR: 39141 MFC after: 2 weeks This time, make sure that ipv4 specific code (aka all of the above) is only run in the ipv4 case.
2002-06-14 03:08:05 +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 (V_path_mtu_discovery && ((sc->sc_flags & SCF_UNREACH) == 0))
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_off |= htons(IP_DF);
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
th = (struct tcphdr *)(ip + 1);
}
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 */
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
th->th_sport = sc->sc_inc.inc_lport;
th->th_dport = sc->sc_inc.inc_fport;
th->th_seq = htonl(sc->sc_iss);
th->th_ack = htonl(sc->sc_irs + 1);
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
th->th_off = sizeof(struct tcphdr) >> 2;
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
th->th_x2 = 0;
th->th_flags = TH_SYN|TH_ACK;
th->th_win = htons(sc->sc_wnd);
th->th_urp = 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
if (sc->sc_flags & SCF_ECN) {
th->th_flags |= TH_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_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
}
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
/* Tack on the TCP options. */
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
if ((sc->sc_flags & SCF_NOOPT) == 0) {
to.to_flags = 0;
Eliminate a goto. Fix some line breaks.
2002-12-20 11:24:02 +00:00
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_mss = mssopt;
to.to_flags = TOF_MSS;
Eliminate a goto. Fix some line breaks.
2002-12-20 11:24:02 +00:00
if (sc->sc_flags & SCF_WINSCALE) {
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 = sc->sc_requested_r_scale;
to.to_flags |= TOF_SCALE;
Eliminate a goto. Fix some line breaks.
2002-12-20 11:24:02 +00:00
}
if (sc->sc_flags & SCF_TIMESTAMP) {
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
/* Virgin timestamp or TCP cookie enhanced one. */
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
to.to_tsval = sc->sc_ts;
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_tsecr = sc->sc_tsreflect;
to.to_flags |= TOF_TS;
Eliminate a goto. Fix some line breaks.
2002-12-20 11:24:02 +00:00
}
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
if (sc->sc_flags & SCF_SACK)
to.to_flags |= TOF_SACKPERM;
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC_SUPPORT) || defined(TCP_SIGNATURE)
if (sc->sc_flags & SCF_SIGNATURE)
to.to_flags |= TOF_SIGNATURE;
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
#endif
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 (sc->sc_tfo_cookie) {
to.to_flags |= TOF_FASTOPEN;
to.to_tfo_len = TCP_FASTOPEN_COOKIE_LEN;
to.to_tfo_cookie = sc->sc_tfo_cookie;
/* don't send cookie again when retransmitting response */
sc->sc_tfo_cookie = NULL;
}
#endif
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
optlen = tcp_addoptions(&to, (u_char *)(th + 1));
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
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
/* Adjust headers by option size. */
th->th_off = (sizeof(struct tcphdr) + optlen) >> 2;
m->m_len += optlen;
m->m_pkthdr.len += optlen;
#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 (sc->sc_inc.inc_flags & INC_ISIPV6)
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
ip6->ip6_plen = htons(ntohs(ip6->ip6_plen) + optlen);
Fix the most obvious of the bugs introduced by recent syncache changes - *ip is not initialized in the case of inet6 connection, but ip->ip_len is being changed anyway Now the question is, why does it think an ipv4 connection is an ipv6 connection? xemacs still doesn't work over X11 forwarding, but the kernel no longer panics.
2007-03-17 06:40:09 +00:00
else
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
#endif
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_len = htons(ntohs(ip->ip_len) + optlen);
Merge projects/ipsec into head/. Small summary ------------- o Almost all IPsec releated code was moved into sys/netipsec. o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel option IPSEC_SUPPORT added. It enables support for loading and unloading of ipsec.ko and tcpmd5.ko kernel modules. o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type support was removed. Added TCP/UDP checksum handling for inbound packets that were decapsulated by transport mode SAs. setkey(8) modified to show run-time NAT-T configuration of SA. o New network pseudo interface if_ipsec(4) added. For now it is build as part of ipsec.ko module (or with IPSEC kernel). It implements IPsec virtual tunnels to create route-based VPNs. o The network stack now invokes IPsec functions using special methods. The only one header file <netipsec/ipsec_support.h> should be included to declare all the needed things to work with IPsec. o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed. Now these protocols are handled directly via IPsec methods. o TCP_SIGNATURE support was reworked to be more close to RFC. o PF_KEY SADB was reworked: - now all security associations stored in the single SPI namespace, and all SAs MUST have unique SPI. - several hash tables added to speed up lookups in SADB. - SADB now uses rmlock to protect access, and concurrent threads can do SA lookups in the same time. - many PF_KEY message handlers were reworked to reflect changes in SADB. - SADB_UPDATE message was extended to support new PF_KEY headers: SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They can be used by IKE daemon to change SA addresses. o ipsecrequest and secpolicy structures were cardinally changed to avoid locking protection for ipsecrequest. Now we support only limited number (4) of bundled SAs, but they are supported for both INET and INET6. o INPCB security policy cache was introduced. Each PCB now caches used security policies to avoid SP lookup for each packet. o For inbound security policies added the mode, when the kernel does check for full history of applied IPsec transforms. o References counting rules for security policies and security associations were changed. The proper SA locking added into xform code. o xform code was also changed. Now it is possible to unregister xforms. tdb_xxx structures were changed and renamed to reflect changes in SADB/SPDB, and changed rules for locking and refcounting. Reviewed by: gnn, wblock Obtained from: Yandex LLC Relnotes: yes Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D9352
2017-02-06 08:49:57 +00:00
#if defined(IPSEC_SUPPORT) || defined(TCP_SIGNATURE)
if (sc->sc_flags & SCF_SIGNATURE) {
KASSERT(to.to_flags & TOF_SIGNATURE,
("tcp_addoptions() didn't set tcp_signature"));
/* NOTE: to.to_signature is inside of mbuf */
if (!TCPMD5_ENABLED() ||
TCPMD5_OUTPUT(m, th, to.to_signature) != 0) {
m_freem(m);
return (EACCES);
}
}
#endif
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
} else
optlen = 0;
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
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
M_SETFIB(m, sc->sc_inc.inc_fibnum);
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
m->m_pkthdr.csum_data = offsetof(struct tcphdr, th_sum);
tcp/syncache: Add comment for syncache_respond Suggested by: hiren, hps Reviewed by: sbruno Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6148
2016-05-10 04:59:04 +00:00
/*
* If we have peer's SYN and it has a flowid, then let's assign it to
* our SYN|ACK. ip6_output() and ip_output() will not assign flowid
* to SYN|ACK due to lack of inp here.
*/
tcp/syncache: Set flowid and hash type properly for SYN|ACK So the underlying drivers can use it to select the sending queue properly for SYN|ACK instead of rolling their own hash. Sponsored by: Microsoft OSTC Differential Revision: https://reviews.freebsd.org/D6120
2016-04-29 07:23:08 +00:00
if (m0 != NULL && M_HASHTYPE_GET(m0) != M_HASHTYPE_NONE) {
m->m_pkthdr.flowid = m0->m_pkthdr.flowid;
M_HASHTYPE_SET(m, M_HASHTYPE_GET(m0));
}
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
#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 (sc->sc_inc.inc_flags & INC_ISIPV6) {
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
m->m_pkthdr.csum_flags = CSUM_TCP_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
th->th_sum = in6_cksum_pseudo(ip6, tlen + optlen - hlen,
IPPROTO_TCP, 0);
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
ip6->ip6_hlim = in6_selecthlim(NULL, NULL);
Add TCP_OFFLOAD hook in syncache_respond for IPv6 too, just like the one that exists for IPv4. Reviewed by: bz@
2013-01-25 22:16:35 +00:00
#ifdef TCP_OFFLOAD
if (ADDED_BY_TOE(sc)) {
struct toedev *tod = sc->sc_tod;
error = tod->tod_syncache_respond(tod, sc->sc_todctx, m);
return (error);
}
#endif
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
error = ip6_output(m, NULL, NULL, 0, NULL, NULL, 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
}
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
#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(INET6) && defined(INET)
else
#endif
#ifdef INET
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
{
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
m->m_pkthdr.csum_flags = CSUM_TCP;
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
th->th_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr,
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
htons(tlen + optlen - hlen + IPPROTO_TCP));
- 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 (ADDED_BY_TOE(sc)) {
struct toedev *tod = sc->sc_tod;
error = tod->tod_syncache_respond(tod, sc->sc_todctx, m);
return (error);
}
#endif
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
error = ip_output(m, sc->sc_ipopts, NULL, 0, NULL, 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
}
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
return (error);
}
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
/*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* The purpose of syncookies is to handle spoofed SYN flooding DoS attacks
* that exceed the capacity of the syncache by avoiding the storage of any
* of the SYNs we receive. Syncookies defend against blind SYN flooding
* attacks where the attacker does not have access to our responses.
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* Syncookies encode and include all necessary information about the
* connection setup within the SYN|ACK that we send back. That way we
* can avoid keeping any local state until the ACK to our SYN|ACK returns
* (if ever). Normally the syncache and syncookies are running in parallel
* with the latter taking over when the former is exhausted. When matching
* syncache entry is found the syncookie is ignored.
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
*
sys/net*: minor spelling fixes. No functional change.
2016-05-03 18:05:43 +00:00
* The only reliable information persisting the 3WHS is our initial sequence
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* number ISS of 32 bits. Syncookies embed a cryptographically sufficient
* strong hash (MAC) value and a few bits of TCP SYN options in the ISS
* of our SYN|ACK. The MAC can be recomputed when the ACK to our SYN|ACK
* returns and signifies a legitimate connection if it matches the ACK.
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
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* The available space of 32 bits to store the hash and to encode the SYN
* option information is very tight and we should have at least 24 bits for
* the MAC to keep the number of guesses by blind spoofing reasonably high.
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
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* SYN option information we have to encode to fully restore a connection:
* MSS: is imporant to chose an optimal segment size to avoid IP level
* fragmentation along the path. The common MSS values can be encoded
* in a 3-bit table. Uncommon values are captured by the next lower value
* in the table leading to a slight increase in packetization overhead.
* WSCALE: is necessary to allow large windows to be used for high delay-
* bandwidth product links. Not scaling the window when it was initially
* negotiated is bad for performance as lack of scaling further decreases
* the apparent available send window. We only need to encode the WSCALE
* we received from the remote end. Our end can be recalculated at any
* time. The common WSCALE values can be encoded in a 3-bit table.
* Uncommon values are captured by the next lower value in the table
* making us under-estimate the available window size halving our
* theoretically possible maximum throughput for that connection.
* SACK: Greatly assists in packet loss recovery and requires 1 bit.
* TIMESTAMP and SIGNATURE is not encoded because they are permanent options
* that are included in all segments on a connection. We enable them when
* the ACK has them.
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
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* Security of syncookies and attack vectors:
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
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* The MAC is computed over (faddr||laddr||fport||lport||irs||flags||secmod)
* together with the gloabl secret to make it unique per connection attempt.
* Thus any change of any of those parameters results in a different MAC output
* in an unpredictable way unless a collision is encountered. 24 bits of the
* MAC are embedded into the ISS.
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
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* To prevent replay attacks two rotating global secrets are updated with a
* new random value every 15 seconds. The life-time of a syncookie is thus
* 15-30 seconds.
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
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* Vector 1: Attacking the secret. This requires finding a weakness in the
* MAC itself or the way it is used here. The attacker can do a chosen plain
* text attack by varying and testing the all parameters under his control.
* The strength depends on the size and randomness of the secret, and the
* cryptographic security of the MAC function. Due to the constant updating
* of the secret the attacker has at most 29.999 seconds to find the secret
* and launch spoofed connections. After that he has to start all over again.
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
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* Vector 2: Collision attack on the MAC of a single ACK. With a 24 bit MAC
* size an average of 4,823 attempts are required for a 50% chance of success
* to spoof a single syncookie (birthday collision paradox). However the
* attacker is blind and doesn't know if one of his attempts succeeded unless
* he has a side channel to interfere success from. A single connection setup
* success average of 90% requires 8,790 packets, 99.99% requires 17,578 packets.
* This many attempts are required for each one blind spoofed connection. For
* every additional spoofed connection he has to launch another N attempts.
* Thus for a sustained rate 100 spoofed connections per second approximately
* 1,800,000 packets per second would have to be sent.
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* NB: The MAC function should be fast so that it doesn't become a CPU
* exhaustion attack vector itself.
*
* References:
* RFC4987 TCP SYN Flooding Attacks and Common Mitigations
* SYN cookies were first proposed by cryptographer Dan J. Bernstein in 1996
* http://cr.yp.to/syncookies.html (overview)
* http://cr.yp.to/syncookies/archive (details)
*
*
* Schematic construction of a syncookie enabled Initial Sequence Number:
* 0 1 2 3
* 12345678901234567890123456789012
* |xxxxxxxxxxxxxxxxxxxxxxxxWWWMMMSP|
*
* x 24 MAC (truncated)
* W 3 Send Window Scale index
* M 3 MSS index
* S 1 SACK permitted
* P 1 Odd/even secret
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
*/
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
/*
* Distribution and probability of certain MSS values. Those in between are
* rounded down to the next lower one.
* [An Analysis of TCP Maximum Segment Sizes, S. Alcock and R. Nelson, 2011]
* .2% .3% 5% 7% 7% 20% 15% 45%
*/
static int tcp_sc_msstab[] = { 216, 536, 1200, 1360, 1400, 1440, 1452, 1460 };
/*
* Distribution and probability of certain WSCALE values. We have to map the
* (send) window scale (shift) option with a range of 0-14 from 4 bits into 3
* bits based on prevalence of certain values. Where we don't have an exact
* match for are rounded down to the next lower one letting us under-estimate
* the true available window. At the moment this would happen only for the
* very uncommon values 3, 5 and those above 8 (more than 16MB socket buffer
* and window size). The absence of the WSCALE option (no scaling in either
* direction) is encoded with index zero.
* [WSCALE values histograms, Allman, 2012]
* X 10 10 35 5 6 14 10% by host
* X 11 4 5 5 18 49 3% by connections
*/
static int tcp_sc_wstab[] = { 0, 0, 1, 2, 4, 6, 7, 8 };
/*
* Compute the MAC for the SYN cookie. SIPHASH-2-4 is chosen for its speed
* and good cryptographic properties.
*/
static uint32_t
syncookie_mac(struct in_conninfo *inc, tcp_seq irs, uint8_t flags,
uint8_t *secbits, uintptr_t secmod)
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
{
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
SIPHASH_CTX ctx;
uint32_t siphash[2];
SipHash24_Init(&ctx);
SipHash_SetKey(&ctx, secbits);
switch (inc->inc_flags & INC_ISIPV6) {
#ifdef INET
case 0:
SipHash_Update(&ctx, &inc->inc_faddr, sizeof(inc->inc_faddr));
SipHash_Update(&ctx, &inc->inc_laddr, sizeof(inc->inc_laddr));
break;
#endif
#ifdef INET6
case INC_ISIPV6:
SipHash_Update(&ctx, &inc->inc6_faddr, sizeof(inc->inc6_faddr));
SipHash_Update(&ctx, &inc->inc6_laddr, sizeof(inc->inc6_laddr));
break;
#endif
}
SipHash_Update(&ctx, &inc->inc_fport, sizeof(inc->inc_fport));
SipHash_Update(&ctx, &inc->inc_lport, sizeof(inc->inc_lport));
Make syncookie_mac() use 'tcp_seq irs' in computing hash. This fixes what seems like a simple oversight when the function was added in r253210. Reported by: Daniel Borkmann <dborkman@redhat.com> Florian Westphal <fw@strlen.de> Differential Revision: https://reviews.freebsd.org/D1628 Reviewed by: gnn MFC after: 1 month Sponsored by: Limelight Networks
2015-01-30 17:29:07 +00:00
SipHash_Update(&ctx, &irs, sizeof(irs));
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
SipHash_Update(&ctx, &flags, sizeof(flags));
SipHash_Update(&ctx, &secmod, sizeof(secmod));
SipHash_Final((u_int8_t *)&siphash, &ctx);
return (siphash[0] ^ siphash[1]);
}
static tcp_seq
syncookie_generate(struct syncache_head *sch, struct syncache *sc)
{
u_int i, mss, secbit, wscale;
uint32_t iss, hash;
uint8_t *secbits;
union syncookie cookie;
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
SCH_LOCK_ASSERT(sch);
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
cookie.cookie = 0;
/* Map our computed MSS into the 3-bit index. */
mss = min(tcp_mssopt(&sc->sc_inc), max(sc->sc_peer_mss, V_tcp_minmss));
Indentation issues. Contract some lines leftover from r298310. Mea culpa.
2016-04-20 16:19:44 +00:00
for (i = nitems(tcp_sc_msstab) - 1; tcp_sc_msstab[i] > mss && i > 0;
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
i--)
;
cookie.flags.mss_idx = i;
/*
* Map the send window scale into the 3-bit index but only if
* the wscale option was received.
*/
if (sc->sc_flags & SCF_WINSCALE) {
wscale = sc->sc_requested_s_scale;
kernel: use our nitems() macro when it is available through param.h. No functional change, only trivial cases are done in this sweep, Discussed in: freebsd-current
2016-04-19 23:48:27 +00:00
for (i = nitems(tcp_sc_wstab) - 1;
Indentation issues. Contract some lines leftover from r298310. Mea culpa.
2016-04-20 16:19:44 +00:00
tcp_sc_wstab[i] > wscale && i > 0;
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
i--)
;
cookie.flags.wscale_idx = i;
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
/* Can we do SACK? */
if (sc->sc_flags & SCF_SACK)
cookie.flags.sack_ok = 1;
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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
/* Which of the two secrets to use. */
secbit = sch->sch_sc->secret.oddeven & 0x1;
cookie.flags.odd_even = secbit;
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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
secbits = sch->sch_sc->secret.key[secbit];
hash = syncookie_mac(&sc->sc_inc, sc->sc_irs, cookie.cookie, secbits,
(uintptr_t)sch);
/*
* Put the flags into the hash and XOR them to get better ISS number
* variance. This doesn't enhance the cryptographic strength and is
* done to prevent the 8 cookie bits from showing up directly on the
* wire.
*/
iss = hash & ~0xff;
iss |= cookie.cookie ^ (hash >> 24);
/* Randomize the timestamp. */
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 (sc->sc_flags & SCF_TIMESTAMP) {
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
sc->sc_ts = arc4random();
sc->sc_tsoff = sc->sc_ts - tcp_ts_getticks();
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +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
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_sc_sendcookie);
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
return (iss);
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
}
static struct syncache *
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
syncookie_lookup(struct in_conninfo *inc, struct syncache_head *sch,
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
struct syncache *sc, struct tcphdr *th, struct tcpopt *to,
struct socket *lso)
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
{
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
uint32_t hash;
uint8_t *secbits;
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
tcp_seq ack, seq;
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
int wnd, wscale = 0;
union syncookie cookie;
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
SCH_LOCK_ASSERT(sch);
/*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* Pull information out of SYN-ACK/ACK and revert sequence number
* advances.
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
*/
ack = th->th_ack - 1;
seq = th->th_seq - 1;
/*
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
* Unpack the flags containing enough information to restore the
* connection.
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
*/
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
cookie.cookie = (ack & 0xff) ^ (ack >> 24);
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
/* Which of the two secrets to use. */
secbits = sch->sch_sc->secret.key[cookie.flags.odd_even];
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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
hash = syncookie_mac(inc, seq, cookie.cookie, secbits, (uintptr_t)sch);
/* The recomputed hash matches the ACK if this was a genuine cookie. */
if ((ack & ~0xff) != (hash & ~0xff))
return (NULL);
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
/* Fill in the syncache values. */
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
sc->sc_flags = 0;
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
bcopy(inc, &sc->sc_inc, sizeof(struct in_conninfo));
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
sc->sc_ipopts = NULL;
sc->sc_irs = seq;
sc->sc_iss = ack;
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
switch (inc->inc_flags & INC_ISIPV6) {
#ifdef INET
case 0:
sc->sc_ip_ttl = sotoinpcb(lso)->inp_ip_ttl;
sc->sc_ip_tos = sotoinpcb(lso)->inp_ip_tos;
break;
#endif
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
#ifdef INET6
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
case INC_ISIPV6:
if (sotoinpcb(lso)->inp_flags & IN6P_AUTOFLOWLABEL)
sc->sc_flowlabel = sc->sc_iss & IPV6_FLOWLABEL_MASK;
break;
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
#endif
}
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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
sc->sc_peer_mss = tcp_sc_msstab[cookie.flags.mss_idx];
/* We can simply recompute receive window scale we sent earlier. */
while (wscale < TCP_MAX_WINSHIFT && (TCP_MAXWIN << wscale) < sb_max)
wscale++;
/* Only use wscale if it was enabled in the orignal SYN. */
if (cookie.flags.wscale_idx > 0) {
sc->sc_requested_r_scale = wscale;
sc->sc_requested_s_scale = tcp_sc_wstab[cookie.flags.wscale_idx];
sc->sc_flags |= SCF_WINSCALE;
}
Listening sockets improvements. o Separate fields of struct socket that belong to listening from fields that belong to normal dataflow, and unionize them. This shrinks the structure a bit. - Take out selinfo's from the socket buffers into the socket. The first reason is to support braindamaged scenario when a socket is added to kevent(2) and then listen(2) is cast on it. The second reason is that there is future plan to make socket buffers pluggable, so that for a dataflow socket a socket buffer can be changed, and in this case we also want to keep same selinfos through the lifetime of a socket. - Remove struct struct so_accf. Since now listening stuff no longer affects struct socket size, just move its fields into listening part of the union. - Provide sol_upcall field and enforce that so_upcall_set() may be called only on a dataflow socket, which has buffers, and for listening sockets provide solisten_upcall_set(). o Remove ACCEPT_LOCK() global. - Add a mutex to socket, to be used instead of socket buffer lock to lock fields of struct socket that don't belong to a socket buffer. - Allow to acquire two socket locks, but the first one must belong to a listening socket. - Make soref()/sorele() to use atomic(9). This allows in some situations to do soref() without owning socket lock. There is place for improvement here, it is possible to make sorele() also to lock optionally. - Most protocols aren't touched by this change, except UNIX local sockets. See below for more information. o Reduce copy-and-paste in kernel modules that accept connections from listening sockets: provide function solisten_dequeue(), and use it in the following modules: ctl(4), iscsi(4), ng_btsocket(4), ng_ksocket(4), infiniband, rpc. o UNIX local sockets. - Removal of ACCEPT_LOCK() global uncovered several races in the UNIX local sockets. Most races exist around spawning a new socket, when we are connecting to a local listening socket. To cover them, we need to hold locks on both PCBs when spawning a third one. This means holding them across sonewconn(). This creates a LOR between pcb locks and unp_list_lock. - To fix the new LOR, abandon the global unp_list_lock in favor of global unp_link_lock. Indeed, separating these two locks didn't provide us any extra parralelism in the UNIX sockets. - Now call into uipc_attach() may happen with unp_link_lock hold if, we are accepting, or without unp_link_lock in case if we are just creating a socket. - Another problem in UNIX sockets is that uipc_close() basicly did nothing for a listening socket. The vnode remained opened for connections. This is fixed by removing vnode in uipc_close(). Maybe the right way would be to do it for all sockets (not only listening), simply move the vnode teardown from uipc_detach() to uipc_close()? Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D9770
2017-06-08 21:30:34 +00:00
wnd = lso->sol_sbrcv_hiwat;
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
wnd = imax(wnd, 0);
wnd = imin(wnd, TCP_MAXWIN);
sc->sc_wnd = wnd;
if (cookie.flags.sack_ok)
sc->sc_flags |= SCF_SACK;
if (to->to_flags & TOF_TS) {
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
sc->sc_flags |= SCF_TIMESTAMP;
sc->sc_tsreflect = to->to_tsval;
Be more restrictive with segment validity checks in syncache_expand() and log check failures to syslog at LOG_DEBUG level. Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
sc->sc_ts = 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
sc->sc_tsoff = to->to_tsecr - tcp_ts_getticks();
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +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
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
if (to->to_flags & TOF_SIGNATURE)
sc->sc_flags |= SCF_SIGNATURE;
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
Add locking to TCP syncache and drop the global tcpinfo lock as early as possible for the syncache_add() case. The syncache timer no longer aquires the tcpinfo lock and timeout/retransmit runs can happen in parallel with bucket granularity. On a P4 the additional locks cause a slight degression of 0.7% in tcp connections per second. When IP and TCP input are deserialized and can run in parallel this little overhead can be neglected. The syncookie handling still leaves room for improvement and its random salts may be moved to the syncache bucket head structures to remove the second lock operation currently required for it. However this would be a more involved change from the way syncookies work at the moment. Reviewed by: rwatson Tested by: rwatson, ps (earlier version) Sponsored by: TCP/IP Optimization Fundraise 2005
2006-06-17 17:32:38 +00:00
sc->sc_rxmits = 0;
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
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_sc_recvcookie);
Extend the SYN DoS defense by adding syncookies to the syncache. All TCP ISNs that are sent out are valid cookies, which allows entries in the syncache to be dropped and still have the ACK accepted later. As all entries pass through the syncache, there is no sudden switchover from cache -> cookies when the cache is full; instead, syncache entries simply have a reduced lifetime. More details may be found in the "Resisting DoS attacks with a SYN cache" paper in the Usenix BSDCon 2002 conference proceedings. Sponsored by: DARPA, NAI Labs
2001-12-19 06:12:14 +00:00
return (sc);
}
Export the contents of the syncache to netstat. Approved by: re (kensmith) MFC after: 2 weeks
2007-07-27 00:57:06 +00:00
Improve SYN cookies by encoding the MSS, WSCALE (window scaling) and SACK information into the ISN (initial sequence number) without the additional use of timestamp bits and switching to the very fast and cryptographically strong SipHash-2-4 MAC hash algorithm to protect the SYN cookie against forgeries. The purpose of SYN cookies is to encode all necessary session state in the 32 bits of our initial sequence number to avoid storing any information locally in memory. This is especially important when under heavy spoofed SYN attacks where we would either run out of memory or the syncache would fill with bogus connection attempts swamping out legitimate connections. The original SYN cookies method only stored an indexed MSS values in the cookie. This isn't sufficient anymore and breaks down in the presence of WSCALE information which is only exchanged during SYN and SYN-ACK. If we can't keep track of it then we may severely underestimate the available send or receive window. This is compounded with large windows whose size information on the TCP segment header is even lower numerically. A number of years back SYN cookies were extended to store the additional state in the TCP timestamp fields, if available on a connection. While timestamps are common among the BSD, Linux and other *nix systems Windows never enabled them by default and thus are not present for the vast majority of clients seen on the Internet. The common parameters used on TCP sessions have changed quite a bit since SYN cookies very invented some 17 years ago. Today we have a lot more bandwidth available making the use window scaling almost mandatory. Also SACK has become standard making recovering from packet loss much more efficient. This change moves all necessary information into the ISS removing the need for timestamps. Both the MSS (16 bits) and send WSCALE (4 bits) are stored in 3 bit indexed form together with a single bit for SACK. While this is significantly less than the original range, it is sufficient to encode all common values with minimal rounding. The MSS depends on the MTU of the path and with the dominance of ethernet the main value seen is around 1460 bytes. Encapsulations for DSL lines and some other overheads reduce it by a few more bytes for many connections seen. Rounding down to the next lower value in some cases isn't a problem as we send only slightly more packets for the same amount of data. The send WSCALE index is bit more tricky as rounding down under-estimates the available send space available towards the remote host, however a small number values dominate and are carefully selected again. The receive WSCALE isn't encoded at all but recalculated based on the local receive socket buffer size when a valid SYN cookie returns. A listen socket buffer size is unlikely to change while active. The index values for MSS and WSCALE are selected for minimal rounding errors based on large traffic surveys. These values have to be periodically validated against newer traffic surveys adjusting the arrays tcp_sc_msstab[] and tcp_sc_wstab[] if necessary. In addition the hash MAC to protect the SYN cookies is changed from MD5 to SipHash-2-4, a much faster and cryptographically secure algorithm. Reviewed by: dwmalone Tested by: Fabian Keil <fk@fabiankeil.de>
2013-07-11 15:29:25 +00:00
#ifdef INVARIANTS
static int
syncookie_cmp(struct in_conninfo *inc, struct syncache_head *sch,
struct syncache *sc, struct tcphdr *th, struct tcpopt *to,
struct socket *lso)
{
struct syncache scs, *scx;
char *s;
bzero(&scs, sizeof(scs));
scx = syncookie_lookup(inc, sch, &scs, th, to, lso);
if ((s = tcp_log_addrs(inc, th, NULL, NULL)) == NULL)
return (0);
if (scx != NULL) {
if (sc->sc_peer_mss != scx->sc_peer_mss)
log(LOG_DEBUG, "%s; %s: mss different %i vs %i\n",
s, __func__, sc->sc_peer_mss, scx->sc_peer_mss);
if (sc->sc_requested_r_scale != scx->sc_requested_r_scale)
log(LOG_DEBUG, "%s; %s: rwscale different %i vs %i\n",
s, __func__, sc->sc_requested_r_scale,
scx->sc_requested_r_scale);
if (sc->sc_requested_s_scale != scx->sc_requested_s_scale)
log(LOG_DEBUG, "%s; %s: swscale different %i vs %i\n",
s, __func__, sc->sc_requested_s_scale,
scx->sc_requested_s_scale);
if ((sc->sc_flags & SCF_SACK) != (scx->sc_flags & SCF_SACK))
log(LOG_DEBUG, "%s; %s: SACK different\n", s, __func__);
}
if (s != NULL)
free(s, M_TCPLOG);
return (0);
}
#endif /* INVARIANTS */
static void
syncookie_reseed(void *arg)
{
struct tcp_syncache *sc = arg;
uint8_t *secbits;
int secbit;
/*
* Reseeding the secret doesn't have to be protected by a lock.
* It only must be ensured that the new random values are visible
* to all CPUs in a SMP environment. The atomic with release
* semantics ensures that.
*/
secbit = (sc->secret.oddeven & 0x1) ? 0 : 1;
secbits = sc->secret.key[secbit];
arc4rand(secbits, SYNCOOKIE_SECRET_SIZE, 0);
atomic_add_rel_int(&sc->secret.oddeven, 1);
/* Reschedule ourself. */
callout_schedule(&sc->secret.reseed, SYNCOOKIE_LIFETIME * hz);
}
Export the contents of the syncache to netstat. Approved by: re (kensmith) MFC after: 2 weeks
2007-07-27 00:57:06 +00:00
/*
* Exports the syncache entries to userland so that netstat can display
* them alongside the other sockets. This function is intended to be
* called only from tcp_pcblist.
*
* Due to concurrency on an active system, the number of pcbs exported
* may have no relation to max_pcbs. max_pcbs merely indicates the
* amount of space the caller allocated for this function to use.
*/
int
syncache_pcblist(struct sysctl_req *req, int max_pcbs, int *pcbs_exported)
{
struct xtcpcb xt;
struct syncache *sc;
struct syncache_head *sch;
int count, error, i;
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
for (count = 0, error = 0, i = 0; i < V_tcp_syncache.hashsize; i++) {
sch = &V_tcp_syncache.hashbase[i];
Export the contents of the syncache to netstat. Approved by: re (kensmith) MFC after: 2 weeks
2007-07-27 00:57:06 +00:00
SCH_LOCK(sch);
TAILQ_FOREACH(sc, &sch->sch_bucket, sc_hash) {
if (count >= max_pcbs) {
SCH_UNLOCK(sch);
goto exit;
}
Cache the cred locally in _syncache_add() while holding the locks, so we can be sure that it's valid. In case we abort early free it again else put it into the syncache. We need the cred in the syncache to be able to restrict what will be exportet by the sysctl helper function syncache_pcblist() (to netstat) within jails. PR: kern/126493 Reviewed by: rwatson (earlier versions) MFC after: 3 days
2008-08-23 14:22:12 +00:00
if (cr_cansee(req->td->td_ucred, sc->sc_cred) != 0)
continue;
Export the contents of the syncache to netstat. Approved by: re (kensmith) MFC after: 2 weeks
2007-07-27 00:57:06 +00:00
bzero(&xt, sizeof(xt));
xt.xt_len = sizeof(xt);
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 (sc->sc_inc.inc_flags & INC_ISIPV6)
Export the contents of the syncache to netstat. Approved by: re (kensmith) MFC after: 2 weeks
2007-07-27 00:57:06 +00:00
xt.xt_inp.inp_vflag = INP_IPV6;
else
xt.xt_inp.inp_vflag = INP_IPV4;
Hide struct inpcb, struct tcpcb from the userland. This is a painful change, but it is needed. On the one hand, we avoid modifying them, and this slows down some ideas, on the other hand we still eventually modify them and tools like netstat(1) never work on next version of FreeBSD. We maintain a ton of spares in them, and we already got some ifdef hell at the end of tcpcb. Details: - Hide struct inpcb, struct tcpcb under _KERNEL || _WANT_FOO. - Make struct xinpcb, struct xtcpcb pure API structures, not including kernel structures inpcb and tcpcb inside. Export into these structures the fields from inpcb and tcpcb that are known to be used, and put there a ton of spare space. - Make kernel and userland utilities compilable after these changes. - Bump __FreeBSD_version. Reviewed by: rrs, gnn Differential Revision: D10018
2017-03-21 06:39:49 +00:00
bcopy(&sc->sc_inc, &xt.xt_inp.inp_inc,
sizeof (struct in_conninfo));
xt.t_state = TCPS_SYN_RECEIVED;
xt.xt_inp.xi_socket.xso_protocol = IPPROTO_TCP;
xt.xt_inp.xi_socket.xso_len = sizeof (struct xsocket);
xt.xt_inp.xi_socket.so_type = SOCK_STREAM;
xt.xt_inp.xi_socket.so_state = SS_ISCONNECTING;
Export the contents of the syncache to netstat. Approved by: re (kensmith) MFC after: 2 weeks
2007-07-27 00:57:06 +00:00
error = SYSCTL_OUT(req, &xt, sizeof xt);
if (error) {
SCH_UNLOCK(sch);
goto exit;
}
count++;
}
SCH_UNLOCK(sch);
}
exit:
*pcbs_exported = count;
return error;
}
Reference in New Issue Copy Permalink