freebsd-skq/sys/netinet/tcp_timewait.c

741 lines
18 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
1994-05-24 10:09:53 +00:00
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)tcp_subr.c 8.2 (Berkeley) 5/24/95
1994-05-24 10:09:53 +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"
#include "opt_inet6.h"
1997-09-16 18:36:06 +00:00
#include "opt_tcpdebug.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/callout.h>
#include <sys/kernel.h>
#include <sys/sysctl.h>
1994-05-24 10:09:53 +00:00
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/priv.h>
#include <sys/proc.h>
1994-05-24 10:09:53 +00:00
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/protosw.h>
#include <sys/random.h>
#include <vm/uma.h>
1994-05-24 10:09:53 +00:00
#include <net/route.h>
#include <net/if.h>
#include <net/if_var.h>
#include <net/vnet.h>
1994-05-24 10:09:53 +00:00
#include <netinet/in.h>
#include <netinet/in_pcb.h>
1994-05-24 10:09:53 +00:00
#include <netinet/in_systm.h>
#include <netinet/in_var.h>
1994-05-24 10:09:53 +00:00
#include <netinet/ip.h>
#include <netinet/ip_icmp.h>
#include <netinet/ip_var.h>
#ifdef INET6
#include <netinet/ip6.h>
#include <netinet6/in6_pcb.h>
#include <netinet6/ip6_var.h>
#include <netinet6/scope6_var.h>
#include <netinet6/nd6.h>
#endif
1994-05-24 10:09:53 +00:00
#include <netinet/tcp.h>
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#ifdef INET6
#include <netinet6/tcp6_var.h>
#endif
1994-05-24 10:09:53 +00:00
#include <netinet/tcpip.h>
#ifdef TCPDEBUG
#include <netinet/tcp_debug.h>
#endif
#ifdef INET6
#include <netinet6/ip6protosw.h>
#endif
#include <machine/in_cksum.h>
#include <security/mac/mac_framework.h>
static VNET_DEFINE(uma_zone_t, tcptw_zone);
#define V_tcptw_zone VNET(tcptw_zone)
static int maxtcptw;
/*
* The timed wait queue contains references to each of the TCP sessions
* currently in the TIME_WAIT state. The queue pointers, including the
* queue pointers in each tcptw structure, are protected using the global
* timewait lock, which must be held over queue iteration and modification.
*
* Rules on tcptw usage:
* - a inpcb is always freed _after_ its tcptw
* - a tcptw relies on its inpcb reference counting for memory stability
* - a tcptw is dereferenceable only while its inpcb is locked
*/
static VNET_DEFINE(TAILQ_HEAD(, tcptw), twq_2msl);
#define V_twq_2msl VNET(twq_2msl)
/* Global timewait lock */
static VNET_DEFINE(struct rwlock, tw_lock);
#define V_tw_lock VNET(tw_lock)
#define TW_LOCK_INIT(tw, d) rw_init_flags(&(tw), (d), 0)
#define TW_LOCK_DESTROY(tw) rw_destroy(&(tw))
#define TW_RLOCK(tw) rw_rlock(&(tw))
#define TW_WLOCK(tw) rw_wlock(&(tw))
#define TW_RUNLOCK(tw) rw_runlock(&(tw))
#define TW_WUNLOCK(tw) rw_wunlock(&(tw))
#define TW_LOCK_ASSERT(tw) rw_assert(&(tw), RA_LOCKED)
#define TW_RLOCK_ASSERT(tw) rw_assert(&(tw), RA_RLOCKED)
#define TW_WLOCK_ASSERT(tw) rw_assert(&(tw), RA_WLOCKED)
#define TW_UNLOCK_ASSERT(tw) rw_assert(&(tw), RA_UNLOCKED)
static void tcp_tw_2msl_reset(struct tcptw *, int);
static void tcp_tw_2msl_stop(struct tcptw *, int);
static int tcp_twrespond(struct tcptw *, int);
static int
tcptw_auto_size(void)
We currently does not react to ICMP administratively prohibited messages send by routers when they deny our traffic, this causes a timeout when trying to connect to TCP ports/services on a remote host, which is blocked by routers or firewalls. rfc1122 (Requirements for Internet Hosts) section 3.2.2.1 actually requi re that we treat such a message for a TCP session, that we treat it like if we had recieved a RST. quote begin. A Destination Unreachable message that is received MUST be reported to the transport layer. The transport layer SHOULD use the information appropriately; for example, see Sections 4.1.3.3, 4.2.3.9, and 4.2.4 below. A transport protocol that has its own mechanism for notifying the sender that a port is unreachable (e.g., TCP, which sends RST segments) MUST nevertheless accept an ICMP Port Unreachable for the same purpose. quote end. I've written a small extension that implement this, it also create a sysctl "net.inet.tcp.icmp_admin_prohib_like_rst" to control if this new behaviour is activated. When it's activated (set to 1) we'll treat a ICMP administratively prohibited message (icmp type 3 code 9, 10 and 13) for a TCP sessions, as if we recived a TCP RST, but only if the TCP session is in SYN_SENT state. The reason for only reacting when in SYN_SENT state, is that this will solve the problem, and at the same time minimize the risk of this being abused. I suggest that we enable this new behaviour by default, but it would be a change of current behaviour, so if people prefer to leave it disabled by default, at least for now, this would be ok for me, the attached diff actually have the sysctl set to 0 by default. PR: 23086 Submitted by: Jesper Skriver <jesper@skriver.dk>
2000-12-16 19:42:06 +00:00
{
int halfrange;
/*
* Max out at half the ephemeral port range so that TIME_WAIT
* sockets don't tie up too many ephemeral ports.
*/
if (V_ipport_lastauto > V_ipport_firstauto)
halfrange = (V_ipport_lastauto - V_ipport_firstauto) / 2;
else
halfrange = (V_ipport_firstauto - V_ipport_lastauto) / 2;
/* Protect against goofy port ranges smaller than 32. */
return (imin(imax(halfrange, 32), maxsockets / 5));
}
static int
sysctl_maxtcptw(SYSCTL_HANDLER_ARGS)
{
int error, new;
if (maxtcptw == 0)
new = tcptw_auto_size();
else
new = maxtcptw;
error = sysctl_handle_int(oidp, &new, 0, req);
if (error == 0 && req->newptr)
if (new >= 32) {
maxtcptw = new;
uma_zone_set_max(V_tcptw_zone, maxtcptw);
}
return (error);
}
SYSCTL_PROC(_net_inet_tcp, OID_AUTO, maxtcptw, CTLTYPE_INT|CTLFLAG_RW,
&maxtcptw, 0, sysctl_maxtcptw, "IU",
"Maximum number of compressed TCP TIME_WAIT entries");
VNET_DEFINE(int, nolocaltimewait) = 0;
#define V_nolocaltimewait VNET(nolocaltimewait)
SYSCTL_INT(_net_inet_tcp, OID_AUTO, nolocaltimewait, 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(nolocaltimewait), 0,
"Do not create compressed TCP TIME_WAIT entries for local connections");
void
tcp_tw_zone_change(void)
{
if (maxtcptw == 0)
uma_zone_set_max(V_tcptw_zone, tcptw_auto_size());
}
void
tcp_tw_init(void)
{
V_tcptw_zone = uma_zcreate("tcptw", sizeof(struct tcptw),
NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0);
TUNABLE_INT_FETCH("net.inet.tcp.maxtcptw", &maxtcptw);
if (maxtcptw == 0)
uma_zone_set_max(V_tcptw_zone, tcptw_auto_size());
else
uma_zone_set_max(V_tcptw_zone, maxtcptw);
TAILQ_INIT(&V_twq_2msl);
TW_LOCK_INIT(V_tw_lock, "tcptw");
}
#ifdef VIMAGE
void
tcp_tw_destroy(void)
{
struct tcptw *tw;
INP_INFO_RLOCK(&V_tcbinfo);
while ((tw = TAILQ_FIRST(&V_twq_2msl)) != NULL)
tcp_twclose(tw, 0);
INP_INFO_RUNLOCK(&V_tcbinfo);
TW_LOCK_DESTROY(V_tw_lock);
uma_zdestroy(V_tcptw_zone);
}
#endif
/*
* Move a TCP connection into TIME_WAIT state.
* tcbinfo is locked.
* inp is locked, and is unlocked before returning.
*/
void
tcp_twstart(struct tcpcb *tp)
{
struct tcptw *tw;
struct inpcb *inp = tp->t_inpcb;
int acknow;
struct socket *so;
#ifdef INET6
int isipv6 = inp->inp_inc.inc_flags & INC_ISIPV6;
#endif
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
INP_WLOCK_ASSERT(inp);
/* A dropped inp should never transition to TIME_WAIT state. */
KASSERT((inp->inp_flags & INP_DROPPED) == 0, ("tcp_twstart: "
"(inp->inp_flags & INP_DROPPED) != 0"));
if (V_nolocaltimewait) {
int error = 0;
#ifdef INET6
if (isipv6)
error = in6_localaddr(&inp->in6p_faddr);
#endif
#if defined(INET6) && defined(INET)
else
#endif
#ifdef INET
error = in_localip(inp->inp_faddr);
#endif
if (error) {
tp = tcp_close(tp);
if (tp != NULL)
INP_WUNLOCK(inp);
return;
}
}
/*
* For use only by DTrace. We do not reference the state
* after this point so modifying it in place is not a problem.
*/
tcp_state_change(tp, TCPS_TIME_WAIT);
tw = uma_zalloc(V_tcptw_zone, M_NOWAIT);
if (tw == NULL) {
/*
* Reached limit on total number of TIMEWAIT connections
* allowed. Remove a connection from TIMEWAIT queue in LRU
* fashion to make room for this connection.
*
* XXX: Check if it possible to always have enough room
* in advance based on guarantees provided by uma_zalloc().
*/
tw = tcp_tw_2msl_scan(1);
if (tw == NULL) {
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
tp = tcp_close(tp);
if (tp != NULL)
INP_WUNLOCK(inp);
return;
}
}
/*
* The tcptw will hold a reference on its inpcb until tcp_twclose
* is called
*/
tw->tw_inpcb = inp;
in_pcbref(inp); /* Reference from tw */
/*
* Recover last window size sent.
*/
if (SEQ_GT(tp->rcv_adv, tp->rcv_nxt))
tw->last_win = (tp->rcv_adv - tp->rcv_nxt) >> tp->rcv_scale;
else
tw->last_win = 0;
/*
* Set t_recent if timestamps are used on the connection.
*/
if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) ==
(TF_REQ_TSTMP|TF_RCVD_TSTMP)) {
tw->t_recent = tp->ts_recent;
tw->ts_offset = tp->ts_offset;
} else {
tw->t_recent = 0;
tw->ts_offset = 0;
}
tw->snd_nxt = tp->snd_nxt;
tw->rcv_nxt = tp->rcv_nxt;
tw->iss = tp->iss;
tw->irs = tp->irs;
tw->t_starttime = tp->t_starttime;
tw->tw_time = 0;
/* XXX
* If this code will
* be used for fin-wait-2 state also, then we may need
* a ts_recent from the last segment.
*/
acknow = tp->t_flags & TF_ACKNOW;
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
/*
* First, discard tcpcb state, which includes stopping its timers and
* freeing it. tcp_discardcb() used to also release the inpcb, but
* that work is now done in the caller.
*
* Note: soisdisconnected() call used to be made in tcp_discardcb(),
* and might not be needed here any longer.
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
*/
tcp_discardcb(tp);
so = inp->inp_socket;
soisdisconnected(so);
tw->tw_cred = crhold(so->so_cred);
SOCK_LOCK(so);
tw->tw_so_options = so->so_options;
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
SOCK_UNLOCK(so);
if (acknow)
tcp_twrespond(tw, TH_ACK);
inp->inp_ppcb = tw;
inp->inp_flags |= INP_TIMEWAIT;
tcp_tw_2msl_reset(tw, 0);
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
/*
* If the inpcb owns the sole reference to the socket, then we can
* detach and free the socket as it is not needed in time wait.
*/
if (inp->inp_flags & INP_SOCKREF) {
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
KASSERT(so->so_state & SS_PROTOREF,
("tcp_twstart: !SS_PROTOREF"));
inp->inp_flags &= ~INP_SOCKREF;
INP_WUNLOCK(inp);
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
ACCEPT_LOCK();
SOCK_LOCK(so);
so->so_state &= ~SS_PROTOREF;
sofree(so);
} else
INP_WUNLOCK(inp);
}
/*
* Returns 1 if the TIME_WAIT state was killed and we should start over,
* looking for a pcb in the listen state. Returns 0 otherwise.
*/
int
tcp_twcheck(struct inpcb *inp, struct tcpopt *to __unused, struct tcphdr *th,
struct mbuf *m, int tlen)
{
struct tcptw *tw;
int thflags;
tcp_seq seq;
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
INP_WLOCK_ASSERT(inp);
/*
* XXXRW: Time wait state for inpcb has been recycled, but inpcb is
* still present. This is undesirable, but temporarily necessary
* until we work out how to handle inpcb's who's timewait state has
* been removed.
*/
tw = intotw(inp);
if (tw == NULL)
goto drop;
thflags = th->th_flags;
/*
* NOTE: for FIN_WAIT_2 (to be added later),
* must validate sequence number before accepting RST
*/
/*
* If the segment contains RST:
* Drop the segment - see Stevens, vol. 2, p. 964 and
* RFC 1337.
*/
if (thflags & TH_RST)
goto drop;
#if 0
/* PAWS not needed at the moment */
/*
* RFC 1323 PAWS: If we have a timestamp reply on this segment
* and it's less than ts_recent, drop it.
*/
if ((to.to_flags & TOF_TS) != 0 && tp->ts_recent &&
TSTMP_LT(to.to_tsval, tp->ts_recent)) {
if ((thflags & TH_ACK) == 0)
goto drop;
goto ack;
}
/*
* ts_recent is never updated because we never accept new segments.
*/
#endif
/*
* If a new connection request is received
* while in TIME_WAIT, drop the old connection
* and start over if the sequence numbers
* are above the previous ones.
*/
if ((thflags & TH_SYN) && SEQ_GT(th->th_seq, tw->rcv_nxt)) {
tcp_twclose(tw, 0);
return (1);
}
/*
* Drop the segment if it does not contain an ACK.
*/
if ((thflags & TH_ACK) == 0)
goto drop;
/*
* Reset the 2MSL timer if this is a duplicate FIN.
*/
if (thflags & TH_FIN) {
seq = th->th_seq + tlen + (thflags & TH_SYN ? 1 : 0);
if (seq + 1 == tw->rcv_nxt)
tcp_tw_2msl_reset(tw, 1);
}
/*
* Acknowledge the segment if it has data or is not a duplicate ACK.
*/
if (thflags != TH_ACK || tlen != 0 ||
th->th_seq != tw->rcv_nxt || th->th_ack != tw->snd_nxt)
tcp_twrespond(tw, TH_ACK);
drop:
INP_WUNLOCK(inp);
m_freem(m);
return (0);
}
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
void
tcp_twclose(struct tcptw *tw, int reuse)
{
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
struct socket *so;
struct inpcb *inp;
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
/*
* At this point, we are in one of two situations:
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
*
* (1) We have no socket, just an inpcb<->twtcp pair. We can free
* all state.
*
* (2) We have a socket -- if we own a reference, release it and
* notify the socket layer.
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
*/
inp = tw->tw_inpcb;
KASSERT((inp->inp_flags & INP_TIMEWAIT), ("tcp_twclose: !timewait"));
KASSERT(intotw(inp) == tw, ("tcp_twclose: inp_ppcb != tw"));
INP_INFO_RLOCK_ASSERT(&V_tcbinfo); /* in_pcbfree() */
INP_WLOCK_ASSERT(inp);
tcp_tw_2msl_stop(tw, reuse);
inp->inp_ppcb = NULL;
in_pcbdrop(inp);
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
so = inp->inp_socket;
if (so != NULL) {
/*
* If there's a socket, handle two cases: first, we own a
* strong reference, which we will now release, or we don't
* in which case another reference exists (XXXRW: think
* about this more), and we don't need to take action.
*/
if (inp->inp_flags & INP_SOCKREF) {
inp->inp_flags &= ~INP_SOCKREF;
INP_WUNLOCK(inp);
ACCEPT_LOCK();
SOCK_LOCK(so);
KASSERT(so->so_state & SS_PROTOREF,
("tcp_twclose: INP_SOCKREF && !SS_PROTOREF"));
so->so_state &= ~SS_PROTOREF;
sofree(so);
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
} else {
/*
* If we don't own the only reference, the socket and
* inpcb need to be left around to be handled by
* tcp_usr_detach() later.
*/
INP_WUNLOCK(inp);
Update TCP for infrastructural changes to the socket/pcb refcount model, pru_abort(), pru_detach(), and in_pcbdetach(): - Universally support and enforce the invariant that so_pcb is never NULL, converting dozens of unnecessary NULL checks into assertions, and eliminating dozens of unnecessary error handling cases in protocol code. - In some cases, eliminate unnecessary pcbinfo locking, as it is no longer required to ensure so_pcb != NULL. For example, the receive code no longer requires the pcbinfo lock, and the send code only requires it if building a new connection on an otherwise unconnected socket triggered via sendto() with an address. This should significnatly reduce tcbinfo lock contention in the receive and send cases. - In order to support the invariant that so_pcb != NULL, it is now necessary for the TCP code to not discard the tcpcb any time a connection is dropped, but instead leave the tcpcb until the socket is shutdown. This case is handled by setting INP_DROPPED, to substitute for using a NULL so_pcb to indicate that the connection has been dropped. This requires the inpcb lock, but not the pcbinfo lock. - Unlike all other protocols in the tree, TCP may need to retain access to the socket after the file descriptor has been closed. Set SS_PROTOREF in tcp_detach() in order to prevent the socket from being freed, and add a flag, INP_SOCKREF, so that the TCP code knows whether or not it needs to free the socket when the connection finally does close. The typical case where this occurs is if close() is called on a TCP socket before all sent data in the send socket buffer has been transmitted or acknowledged. If INP_SOCKREF is found when the connection is dropped, we release the inpcb, tcpcb, and socket instead of flagging INP_DROPPED. - Abort and detach protocol switch methods no longer return failures, nor attempt to free sockets, as the socket layer does this. - Annotate the existence of a long-standing race in the TCP timer code, in which timers are stopped but not drained when the socket is freed, as waiting for drain may lead to deadlocks, or have to occur in a context where waiting is not permitted. This race has been handled by testing to see if the tcpcb pointer in the inpcb is NULL (and vice versa), which is not normally permitted, but may be true of a inpcb and tcpcb have been freed. Add a counter to test how often this race has actually occurred, and a large comment for each instance where we compare potentially freed memory with NULL. This will have to be fixed in the near future, but requires is to further address how to handle the timer shutdown shutdown issue. - Several TCP calls no longer potentially free the passed inpcb/tcpcb, so no longer need to return a pointer to indicate whether the argument passed in is still valid. - Un-macroize debugging and locking setup for various protocol switch methods for TCP, as it lead to more obscurity, and as locking becomes more customized to the methods, offers less benefit. - Assert copyright on tcp_usrreq.c due to significant modifications that have been made as part of this work. These changes significantly modify the memory management and connection logic of our TCP implementation, and are (as such) High Risk Changes, and likely to contain serious bugs. Please report problems to the current@ mailing list ASAP, ideally with simple test cases, and optionally, packet traces. MFC after: 3 months
2006-04-01 16:36:36 +00:00
}
} else {
/*
* The socket has been already cleaned-up for us, only free the
* inpcb.
*/
in_pcbfree(inp);
}
TCPSTAT_INC(tcps_closed);
}
static int
tcp_twrespond(struct tcptw *tw, int flags)
{
struct inpcb *inp = tw->tw_inpcb;
#if defined(INET6) || defined(INET)
struct tcphdr *th = NULL;
#endif
struct mbuf *m;
#ifdef INET
struct ip *ip = NULL;
#endif
u_int hdrlen, optlen;
int error = 0; /* Keep compiler happy */
struct tcpopt to;
#ifdef INET6
struct ip6_hdr *ip6 = NULL;
int isipv6 = inp->inp_inc.inc_flags & INC_ISIPV6;
#endif
hdrlen = 0; /* Keep compiler happy */
INP_WLOCK_ASSERT(inp);
m = m_gethdr(M_NOWAIT, MT_DATA);
if (m == NULL)
return (ENOBUFS);
m->m_data += max_linkhdr;
#ifdef MAC
mac_inpcb_create_mbuf(inp, m);
#endif
#ifdef INET6
if (isipv6) {
hdrlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr);
ip6 = mtod(m, struct ip6_hdr *);
th = (struct tcphdr *)(ip6 + 1);
tcpip_fillheaders(inp, ip6, th);
}
#endif
#if defined(INET6) && defined(INET)
else
#endif
#ifdef INET
{
hdrlen = sizeof(struct tcpiphdr);
ip = mtod(m, struct ip *);
th = (struct tcphdr *)(ip + 1);
tcpip_fillheaders(inp, ip, th);
}
#endif
to.to_flags = 0;
/*
* Send a timestamp and echo-reply if both our side and our peer
* have sent timestamps in our SYN's and this is not a RST.
*/
if (tw->t_recent && flags == TH_ACK) {
to.to_flags |= TOF_TS;
to.to_tsval = tcp_ts_getticks() + tw->ts_offset;
to.to_tsecr = tw->t_recent;
}
optlen = tcp_addoptions(&to, (u_char *)(th + 1));
m->m_len = hdrlen + optlen;
m->m_pkthdr.len = m->m_len;
KASSERT(max_linkhdr + m->m_len <= MHLEN, ("tcptw: mbuf too small"));
th->th_seq = htonl(tw->snd_nxt);
th->th_ack = htonl(tw->rcv_nxt);
th->th_off = (sizeof(struct tcphdr) + optlen) >> 2;
th->th_flags = flags;
th->th_win = htons(tw->last_win);
m->m_pkthdr.csum_data = offsetof(struct tcphdr, th_sum);
#ifdef INET6
if (isipv6) {
m->m_pkthdr.csum_flags = CSUM_TCP_IPV6;
th->th_sum = in6_cksum_pseudo(ip6,
sizeof(struct tcphdr) + optlen, IPPROTO_TCP, 0);
ip6->ip6_hlim = in6_selecthlim(inp, NULL);
error = ip6_output(m, inp->in6p_outputopts, NULL,
(tw->tw_so_options & SO_DONTROUTE), NULL, NULL, inp);
}
#endif
#if defined(INET6) && defined(INET)
else
#endif
#ifdef INET
{
m->m_pkthdr.csum_flags = CSUM_TCP;
th->th_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr,
htons(sizeof(struct tcphdr) + optlen + IPPROTO_TCP));
ip->ip_len = htons(m->m_pkthdr.len);
if (V_path_mtu_discovery)
ip->ip_off |= htons(IP_DF);
error = ip_output(m, inp->inp_options, NULL,
((tw->tw_so_options & SO_DONTROUTE) ? IP_ROUTETOIF : 0),
NULL, inp);
}
#endif
if (flags & TH_ACK)
TCPSTAT_INC(tcps_sndacks);
else
TCPSTAT_INC(tcps_sndctrl);
TCPSTAT_INC(tcps_sndtotal);
return (error);
}
static void
tcp_tw_2msl_reset(struct tcptw *tw, int rearm)
{
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
INP_WLOCK_ASSERT(tw->tw_inpcb);
TW_WLOCK(V_tw_lock);
if (rearm)
TAILQ_REMOVE(&V_twq_2msl, tw, tw_2msl);
tw->tw_time = ticks + 2 * tcp_msl;
TAILQ_INSERT_TAIL(&V_twq_2msl, tw, tw_2msl);
TW_WUNLOCK(V_tw_lock);
}
static void
tcp_tw_2msl_stop(struct tcptw *tw, int reuse)
{
struct ucred *cred;
struct inpcb *inp;
int released;
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
TW_WLOCK(V_tw_lock);
inp = tw->tw_inpcb;
tw->tw_inpcb = NULL;
TAILQ_REMOVE(&V_twq_2msl, tw, tw_2msl);
cred = tw->tw_cred;
tw->tw_cred = NULL;
TW_WUNLOCK(V_tw_lock);
if (cred != NULL)
crfree(cred);
released = in_pcbrele_wlocked(inp);
KASSERT(!released, ("%s: inp should not be released here", __func__));
if (!reuse)
uma_zfree(V_tcptw_zone, tw);
TCPSTATES_DEC(TCPS_TIME_WAIT);
}
struct tcptw *
tcp_tw_2msl_scan(int reuse)
{
struct tcptw *tw;
struct inpcb *inp;
#ifdef INVARIANTS
if (reuse) {
/*
* Exclusive pcbinfo lock is not required in reuse case even if
* two inpcb locks can be acquired simultaneously:
* - the inpcb transitioning to TIME_WAIT state in
* tcp_tw_start(),
* - the inpcb closed by tcp_twclose().
*
* It is because only inpcbs in FIN_WAIT2 or CLOSING states can
* transition in TIME_WAIT state. Then a pcbcb cannot be in
* TIME_WAIT list and transitioning to TIME_WAIT state at same
* time.
*/
INP_INFO_RLOCK_ASSERT(&V_tcbinfo);
}
#endif
for (;;) {
TW_RLOCK(V_tw_lock);
tw = TAILQ_FIRST(&V_twq_2msl);
if (tw == NULL || (!reuse && (tw->tw_time - ticks) > 0)) {
TW_RUNLOCK(V_tw_lock);
break;
}
KASSERT(tw->tw_inpcb != NULL, ("%s: tw->tw_inpcb == NULL",
__func__));
inp = tw->tw_inpcb;
in_pcbref(inp);
TW_RUNLOCK(V_tw_lock);
if (INP_INFO_TRY_RLOCK(&V_tcbinfo)) {
INP_WLOCK(inp);
tw = intotw(inp);
if (in_pcbrele_wlocked(inp)) {
KASSERT(tw == NULL, ("%s: held last inp "
"reference but tw not NULL", __func__));
INP_INFO_RUNLOCK(&V_tcbinfo);
continue;
}
if (tw == NULL) {
/* tcp_twclose() has already been called */
INP_WUNLOCK(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
continue;
}
tcp_twclose(tw, reuse);
INP_INFO_RUNLOCK(&V_tcbinfo);
if (reuse)
return tw;
} else {
/* INP_INFO lock is busy, continue later. */
INP_WLOCK(inp);
if (!in_pcbrele_wlocked(inp))
INP_WUNLOCK(inp);
break;
}
}
return NULL;
}