tcp.c

GNU Hurd 源代码

/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Implementation of the Transmission Control Protocol(TCP).
 *
 * Version:	$Id: tcp.c,v 1.140.2.5 1999/09/23 19:21:16 davem Exp $
 *
 * Authors:	Ross Biro, <[e-mail]>
 *		Fred N. van Kempen, <[e-mail]>
 *		Mark Evans, <[e-mail]>
 *		Corey Minyard <wf-rch![e-mail]>
 *		Florian La Roche, <[e-mail]>
 *		Charles Hedrick, <[e-mail]>
 *		Linus Torvalds, <[e-mail]>
 *		Alan Cox, <[e-mail]>
 *		Matthew Dillon, <[e-mail]>
 *		Arnt Gulbrandsen, <[e-mail]>
 *		Jorge Cwik, <[e-mail]>
 *
 * Fixes:
 *		Alan Cox	:	Numerous verify_area() calls
 *		Alan Cox	:	Set the ACK bit on a reset
 *		Alan Cox	:	Stopped it crashing if it closed while
 *					sk->inuse=1 and was trying to connect
 *					(tcp_err()).
 *		Alan Cox	:	All icmp error handling was broken
 *					pointers passed where wrong and the
 *					socket was looked up backwards. Nobody
 *					tested any icmp error code obviously.
 *		Alan Cox	:	tcp_err() now handled properly. It
 *					wakes people on errors. poll
 *					behaves and the icmp error race
 *					has gone by moving it into sock.c
 *		Alan Cox	:	tcp_send_reset() fixed to work for
 *					everything not just packets for
 *					unknown sockets.
 *		Alan Cox	:	tcp option processing.
 *		Alan Cox	:	Reset tweaked (still not 100%) [Had
 *					syn rule wrong]
 *		Herp Rosmanith  :	More reset fixes
 *		Alan Cox	:	No longer acks invalid rst frames.
 *					Acking any kind of RST is right out.
 *		Alan Cox	:	Sets an ignore me flag on an rst
 *					receive otherwise odd bits of prattle
 *					escape still
 *		Alan Cox	:	Fixed another acking RST frame bug.
 *					Should stop LAN workplace lockups.
 *		Alan Cox	: 	Some tidyups using the new skb list
 *					facilities
 *		Alan Cox	:	sk->keepopen now seems to work
 *		Alan Cox	:	Pulls options out correctly on accepts
 *		Alan Cox	:	Fixed assorted sk->rqueue->next errors
 *		Alan Cox	:	PSH doesn't end a TCP read. Switched a
 *					bit to skb ops.
 *		Alan Cox	:	Tidied tcp_data to avoid a potential
 *					nasty.
 *		Alan Cox	:	Added some better commenting, as the
 *					tcp is hard to follow
 *		Alan Cox	:	Removed incorrect check for 20 * psh
 *	Michael O'Reilly	:	ack < copied bug fix.
 *	Johannes Stille		:	Misc tcp fixes (not all in yet).
 *		Alan Cox	:	FIN with no memory -> CRASH
 *		Alan Cox	:	Added socket option proto entries.
 *					Also added awareness of them to accept.
 *		Alan Cox	:	Added TCP options (SOL_TCP)
 *		Alan Cox	:	Switched wakeup calls to callbacks,
 *					so the kernel can layer network
 *					sockets.
 *		Alan Cox	:	Use ip_tos/ip_ttl settings.
 *		Alan Cox	:	Handle FIN (more) properly (we hope).
 *		Alan Cox	:	RST frames sent on unsynchronised
 *					state ack error.
 *		Alan Cox	:	Put in missing check for SYN bit.
 *		Alan Cox	:	Added tcp_select_window() aka NET2E
 *					window non shrink trick.
 *		Alan Cox	:	Added a couple of small NET2E timer
 *					fixes
 *		Charles Hedrick :	TCP fixes
 *		Toomas Tamm	:	TCP window fixes
 *		Alan Cox	:	Small URG fix to rlogin ^C ack fight
 *		Charles Hedrick	:	Rewrote most of it to actually work
 *		Linus		:	Rewrote tcp_read() and URG handling
 *					completely
 *		Gerhard Koerting:	Fixed some missing timer handling
 *		Matthew Dillon  :	Reworked TCP machine states as per RFC
 *		Gerhard Koerting:	PC/TCP workarounds
 *		Adam Caldwell	:	Assorted timer/timing errors
 *		Matthew Dillon	:	Fixed another RST bug
 *		Alan Cox	:	Move to kernel side addressing changes.
 *		Alan Cox	:	Beginning work on TCP fastpathing
 *					(not yet usable)
 *		Arnt Gulbrandsen:	Turbocharged tcp_check() routine.
 *		Alan Cox	:	TCP fast path debugging
 *		Alan Cox	:	Window clamping
 *		Michael Riepe	:	Bug in tcp_check()
 *		Matt Dillon	:	More TCP improvements and RST bug fixes
 *		Matt Dillon	:	Yet more small nasties remove from the
 *					TCP code (Be very nice to this man if
 *					tcp finally works 100%) 8)
 *		Alan Cox	:	BSD accept semantics.
 *		Alan Cox	:	Reset on closedown bug.
 *	Peter De Schrijver	:	ENOTCONN check missing in tcp_sendto().
 *		Michael Pall	:	Handle poll() after URG properly in
 *					all cases.
 *		Michael Pall	:	Undo the last fix in tcp_read_urg()
 *					(multi URG PUSH broke rlogin).
 *		Michael Pall	:	Fix the multi URG PUSH problem in
 *					tcp_readable(), poll() after URG
 *					works now.
 *		Michael Pall	:	recv(...,MSG_OOB) never blocks in the
 *					BSD api.
 *		Alan Cox	:	Changed the semantics of sk->socket to
 *					fix a race and a signal problem with
 *					accept() and async I/O.
 *		Alan Cox	:	Relaxed the rules on tcp_sendto().
 *		Yury Shevchuk	:	Really fixed accept() blocking problem.
 *		Craig I. Hagan  :	Allow for BSD compatible TIME_WAIT for
 *					clients/servers which listen in on
 *					fixed ports.
 *		Alan Cox	:	Cleaned the above up and shrank it to
 *					a sensible code size.
 *		Alan Cox	:	Self connect lockup fix.
 *		Alan Cox	:	No connect to multicast.
 *		Ross Biro	:	Close unaccepted children on master
 *					socket close.
 *		Alan Cox	:	Reset tracing code.
 *		Alan Cox	:	Spurious resets on shutdown.
 *		Alan Cox	:	Giant 15 minute/60 second timer error
 *		Alan Cox	:	Small whoops in polling before an
 *					accept.
 *		Alan Cox	:	Kept the state trace facility since
 *					it's handy for debugging.
 *		Alan Cox	:	More reset handler fixes.
 *		Alan Cox	:	Started rewriting the code based on
 *					the RFC's for other useful protocol
 *					references see: Comer, KA9Q NOS, and
 *					for a reference on the difference
 *					between specifications and how BSD
 *					works see the 4.4lite source.
 *		A.N.Kuznetsov	:	Don't time wait on completion of tidy
 *					close.
 *		Linus Torvalds	:	Fin/Shutdown & copied_seq changes.
 *		Linus Torvalds	:	Fixed BSD port reuse to work first syn
 *		Alan Cox	:	Reimplemented timers as per the RFC
 *					and using multiple timers for sanity.
 *		Alan Cox	:	Small bug fixes, and a lot of new
 *					comments.
 *		Alan Cox	:	Fixed dual reader crash by locking
 *					the buffers (much like datagram.c)
 *		Alan Cox	:	Fixed stuck sockets in probe. A probe
 *					now gets fed up of retrying without
 *					(even a no space) answer.
 *		Alan Cox	:	Extracted closing code better
 *		Alan Cox	:	Fixed the closing state machine to
 *					resemble the RFC.
 *		Alan Cox	:	More 'per spec' fixes.
 *		Jorge Cwik	:	Even faster checksumming.
 *		Alan Cox	:	tcp_data() doesn't ack illegal PSH
 *					only frames. At least one pc tcp stack
 *					generates them.
 *		Alan Cox	:	Cache last socket.
 *		Alan Cox	:	Per route irtt.
 *		Matt Day	:	poll()->select() match BSD precisely on error
 *		Alan Cox	:	New buffers
 *		Marc Tamsky	:	Various sk->prot->retransmits and
 *					sk->retransmits misupdating fixed.
 *					Fixed tcp_write_timeout: stuck close,
 *					and TCP syn retries gets used now.
 *		Mark Yarvis	:	In tcp_read_wakeup(), don't send an
 *					ack if state is TCP_CLOSED.
 *		Alan Cox	:	Look up device on a retransmit - routes may
 *					change. Doesn't yet cope with MSS shrink right
 *					but its a start!
 *		Marc Tamsky	:	Closing in closing fixes.
 *		Mike Shaver	:	RFC1122 verifications.
 *		Alan Cox	:	rcv_saddr errors.
 *		Alan Cox	:	Block double connect().
 *		Alan Cox	:	Small hooks for enSKIP.
 *		Alexey Kuznetsov:	Path MTU discovery.
 *		Alan Cox	:	Support soft errors.
 *		Alan Cox	:	Fix MTU discovery pathological case
 *					when the remote claims no mtu!
 *		Marc Tamsky	:	TCP_CLOSE fix.
 *		Colin (G3TNE)	:	Send a reset on syn ack replies in
 *					window but wrong (fixes NT lpd problems)
 *		Pedro Roque	:	Better TCP window handling, delayed ack.
 *		Joerg Reuter	:	No modification of locked buffers in
 *					tcp_do_retransmit()
 *		Eric Schenk	:	Changed receiver side silly window
 *					avoidance algorithm to BSD style
 *					algorithm. This doubles throughput
 *					against machines running Solaris,
 *					and seems to result in general
 *					improvement.
 *	Stefan Magdalinski	:	adjusted tcp_readable() to fix FIONREAD
 *	Willy Konynenberg	:	Transparent proxying support.
 *	Mike McLagan		:	Routing by source
 *		Keith Owens	:	Do proper merging with partial SKB's in
 *					tcp_do_sendmsg to avoid burstiness.
 *		Eric Schenk	:	Fix fast close down bug with
 *					shutdown() followed by close().
 *		Andi Kleen :	Make poll agree with SIGIO
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or(at your option) any later version.
 *
 * Description of States:
 *
 *	TCP_SYN_SENT		sent a connection request, waiting for ack
 *
 *	TCP_SYN_RECV		received a connection request, sent ack,
 *				waiting for final ack in three-way handshake.
 *
 *	TCP_ESTABLISHED		connection established
 *
 *	TCP_FIN_WAIT1		our side has shutdown, waiting to complete
 *				transmission of remaining buffered data
 *
 *	TCP_FIN_WAIT2		all buffered data sent, waiting for remote
 *				to shutdown
 *
 *	TCP_CLOSING		both sides have shutdown but we still have
 *				data we have to finish sending
 *
 *	TCP_TIME_WAIT		timeout to catch resent junk before entering
 *				closed, can only be entered from FIN_WAIT2
 *				or CLOSING.  Required because the other end
 *				may not have gotten our last ACK causing it
 *				to retransmit the data packet (which we ignore)
 *
 *	TCP_CLOSE_WAIT		remote side has shutdown and is waiting for
 *				us to finish writing our data and to shutdown
 *				(we have to close() to move on to LAST_ACK)
 *
 *	TCP_LAST_ACK		out side has shutdown after remote has
 *				shutdown.  There may still be data in our
 *				buffer that we have to finish sending
 *
 *	TCP_CLOSE		socket is finished
 */

/*
 * RFC1122 status:
 * NOTE: I'm not going to be doing comments in the code for this one except
 * for violations and the like.  tcp.c is just too big... If I say something
 * "does?" or "doesn't?", it means I'm not sure, and will have to hash it out
 * with Alan. -- MS 950903
 * [Note: Most of the TCP code has been rewriten/redesigned since this
 *  RFC1122 check. It is probably not correct anymore. It should be redone
 *  before 2.2. -AK]
 *
 * Use of PSH (4.2.2.2)
 *   MAY aggregate data sent without the PSH flag. (does)
 *   MAY queue data received without the PSH flag. (does)
 *   SHOULD collapse successive PSH flags when it packetizes data. (doesn't)
 *   MAY implement PSH on send calls. (doesn't, thus:)
 *     MUST NOT buffer data indefinitely (doesn't [1 second])
 *     MUST set PSH on last segment (does)
 *   MAY pass received PSH to application layer (doesn't)
 *   SHOULD send maximum-sized segment whenever possible. (almost always does)
 *
 * Window Size (4.2.2.3, 4.2.2.16)
 *   MUST treat window size as an unsigned number (does)
 *   SHOULD treat window size as a 32-bit number (does not)
 *   MUST NOT shrink window once it is offered (does not normally)
 *
 * Urgent Pointer (4.2.2.4)
 * **MUST point urgent pointer to last byte of urgent data (not right
 *     after). (doesn't, to be like BSD. That's configurable, but defaults
 *	to off)
 *   MUST inform application layer asynchronously of incoming urgent
 *     data. (does)
 *   MUST provide application with means of determining the amount of
 *     urgent data pending. (does)
 * **MUST support urgent data sequence of arbitrary length. (doesn't, but
 *   it's sort of tricky to fix, as urg_ptr is a 16-bit quantity)
 *	[Follows BSD 1 byte of urgent data]
 *
 * TCP Options (4.2.2.5)
 *   MUST be able to receive TCP options in any segment. (does)
 *   MUST ignore unsupported options (does)
 *
 * Maximum Segment Size Option (4.2.2.6)
 *   MUST implement both sending and receiving MSS. (does, but currently
 *	only uses the smaller of both of them)
 *   SHOULD send an MSS with every SYN where receive MSS != 536 (MAY send
 *     it always). (does, even when MSS == 536, which is legal)
 *   MUST assume MSS == 536 if no MSS received at connection setup (does)
 *   MUST calculate "effective send MSS" correctly:
 *     min(physical_MTU, remote_MSS+20) - sizeof(tcphdr) - sizeof(ipopts)
 *     (does - but allows operator override)
 *
 * TCP Checksum (4.2.2.7)
 *   MUST generate and check TCP checksum. (does)
 *
 * Initial Sequence Number Selection (4.2.2.8)
 *   MUST use the RFC 793 clock selection mechanism.  (doesn't, but it's
 *     OK: RFC 793 specifies a 250KHz clock, while we use 1MHz, which is
 *     necessary for 10Mbps networks - and harder than BSD to spoof!
 *     With syncookies we don't)
 *
 * Simultaneous Open Attempts (4.2.2.10)
 *   MUST support simultaneous open attempts (does)
 *
 * Recovery from Old Duplicate SYN (4.2.2.11)
 *   MUST keep track of active vs. passive open (does)
 *
 * RST segment (4.2.2.12)
 *   SHOULD allow an RST segment to contain data (does, but doesn't do
 *     anything with it, which is standard)
 *
 * Closing a Connection (4.2.2.13)
 *   MUST inform application of whether connection was closed by RST or
 *     normal close. (does)
 *   MAY allow "half-duplex" close (treat connection as closed for the
 *     local app, even before handshake is done). (does)
 *   MUST linger in TIME_WAIT for 2 * MSL (does)
 *
 * Retransmission Timeout (4.2.2.15)
 *   MUST implement Jacobson's slow start and congestion avoidance
 *     stuff. (does)
 *
 * Probing Zero Windows (4.2.2.17)
 *   MUST support probing of zero windows. (does)
 *   MAY keep offered window closed indefinitely. (does)
 *   MUST allow remote window to stay closed indefinitely. (does)
 *
 * Passive Open Calls (4.2.2.18)
 *   MUST NOT let new passive open affect other connections. (doesn't)
 *   MUST support passive opens (LISTENs) concurrently. (does)
 *
 * Time to Live (4.2.2.19)
 *   MUST make TCP TTL configurable. (does - IP_TTL option)
 *
 * Event Processing (4.2.2.20)
 *   SHOULD queue out-of-order segments. (does)
 *   MUST aggregate ACK segments whenever possible. (does but badly)
 *
 * Retransmission Timeout Calculation (4.2.3.1)
 *   MUST implement Karn's algorithm and Jacobson's algorithm for RTO
 *     calculation. (does, or at least explains them in the comments 8*b)
 *  SHOULD initialize RTO to 0 and RTT to 3. (does)
 *
 * When to Send an ACK Segment (4.2.3.2)
 *   SHOULD implement delayed ACK. (does)
 *   MUST keep ACK delay < 0.5 sec. (does)
 *
 * When to Send a Window Update (4.2.3.3)
 *   MUST implement receiver-side SWS. (does)
 *
 * When to Send Data (4.2.3.4)
 *   MUST implement sender-side SWS. (does)
 *   SHOULD implement Nagle algorithm. (does)
 *
 * TCP Connection Failures (4.2.3.5)
 *  MUST handle excessive retransmissions "properly" (see the RFC). (does)
 *   SHOULD inform application layer of soft errors. (does)
 *
 * TCP Keep-Alives (4.2.3.6)
 *   MAY provide keep-alives. (does)
 *   MUST make keep-alives configurable on a per-connection basis. (does)
 *   MUST default to no keep-alives. (does)
 *   MUST make keep-alive interval configurable. (does)
 *   MUST make default keep-alive interval > 2 hours. (does)
 *   MUST NOT interpret failure to ACK keep-alive packet as dead
 *     connection. (doesn't)
 *   SHOULD send keep-alive with no data. (does)
 *
 * TCP Multihoming (4.2.3.7)
 *   MUST get source address from IP layer before sending first
 *     SYN. (does)
 *   MUST use same local address for all segments of a connection. (does)
 *
 * IP Options (4.2.3.8)
 *   MUST ignore unsupported IP options. (does)
 *   MAY support Time Stamp and Record Route. (does)
 *   MUST allow application to specify a source route. (does)
 *   MUST allow received Source Route option to set route for all future
 *     segments on this connection. (does not (security issues))
 *
 * ICMP messages (4.2.3.9)
 *   MUST act on ICMP errors. (does)
 *   MUST slow transmission upon receipt of a Source Quench. (doesn't anymore
 *   because that is deprecated now by the IETF, can be turned on)
 *   MUST NOT abort connection upon receipt of soft Destination
 *     Unreachables (0, 1, 5), Time Exceededs and Parameter
 *     Problems. (doesn't)
 *   SHOULD report soft Destination Unreachables etc. to the
 *     application. (does, except during SYN_RECV and may drop messages
 *     in some rare cases before accept() - ICMP is unreliable)
 *   SHOULD abort connection upon receipt of hard Destination Unreachable
 *     messages (2, 3, 4). (does, but see above)
 *
 * Remote Address Validation (4.2.3.10)
 *   MUST reject as an error OPEN for invalid remote IP address. (does)
 *   MUST ignore SYN with invalid source address. (does)
 *   MUST silently discard incoming SYN for broadcast/multicast
 *     address. (does)
 *
 * Asynchronous Reports (4.2.4.1)
 * MUST provide mechanism for reporting soft errors to application
 *     layer. (does)
 *
 * Type of Service (4.2.4.2)
 *   MUST allow application layer to set Type of Service. (does IP_TOS)
 *
 * (Whew. -- MS 950903)
 * (Updated by AK, but not complete yet.)
 **/

#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/poll.h>
#include <linux/init.h>

#include <net/icmp.h>
#include <net/tcp.h>

#include <asm/uaccess.h>

int sysctl_tcp_fin_timeout = TCP_FIN_TIMEOUT;

struct tcp_mib	tcp_statistics;

kmem_cache_t *tcp_openreq_cachep;
kmem_cache_t *tcp_bucket_cachep;
kmem_cache_t *tcp_timewait_cachep;

/*
 *	Find someone to 'accept'. Must be called with
 *	the socket locked or with interrupts disabled
 */

static struct open_request *tcp_find_established(struct tcp_opt *tp,
						 struct open_request **prevp)
{
	struct open_request *req = tp->syn_wait_queue;
	struct open_request *prev = (struct open_request *)&tp->syn_wait_queue;
	while(req) {
		if (req->sk &&
		    ((1 << req->sk->state) &
		     ~(TCPF_SYN_SENT|TCPF_SYN_RECV)))
			break;
		prev = req;
		req = req->dl_next;
	}
	*prevp = prev;
	return req;
}

/*
 *	Walk down the receive queue counting readable data.
 *
 *	Must be called with the socket lock held.
 */

static int tcp_readable(struct sock *sk)
{
	unsigned long counted;
	unsigned long amount;
	struct sk_buff *skb;
	int sum;

	SOCK_DEBUG(sk, "tcp_readable: %p - ",sk);

	skb = skb_peek(&sk->receive_queue);
	if (skb == NULL) {
		SOCK_DEBUG(sk, "empty\n");