Commit Graph

262 Commits

Author SHA1 Message Date
Maxim Konovalov
acc03ac6bb o Convert w/spaces to tabs in the previous commit. 2006-09-29 06:46:31 +00:00
Mike Silbersack
d4bdcb16cc Rather than autoscaling the number of TIME_WAIT sockets to maxsockets / 5,
scale it to min(ephemeral port range / 2, maxsockets / 5) so that people
with large gobs of memory and/or large maxsockets settings will not
exhaust their entire ephemeral port range with sockets in the TIME_WAIT
state during periods of heavy load.

Those who wish to tweak the size of the TIME_WAIT zone can still do so with
net.inet.tcp.maxtcptw.

Reviewed by: glebius, ru
2006-09-29 06:24:26 +00:00
Gleb Smirnoff
3e630ef9a9 Add a sysctl net.inet.tcp.nolocaltimewait that allows to suppress
creating a compress TIME WAIT states, if both connection endpoints
are local. Default is off.
2006-09-08 13:09:15 +00:00
Ruslan Ermilov
751dea2935 Back when we had T/TCP support, we used to apply different
timeouts for TCP and T/TCP connections in the TIME_WAIT
state, and we had two separate timed wait queues for them.
Now that is has gone, the timeout is always 2*MSL again,
and there is no reason to keep two queues (the first was
unused anyway!).

Also, reimplement the remaining queue using a TAILQ (it
was technically impossible before, with two queues).
2006-09-07 13:06:00 +00:00
Andre Oppermann
233dcce118 First step of TSO (TCP segmentation offload) support in our network stack.
o add IFCAP_TSO[46] for drivers to announce this capability for IPv4 and IPv6
 o add CSUM_TSO flag to mbuf pkthdr csum_flags field
 o add tso_segsz field to mbuf pkthdr
 o enhance ip_output() packet length check to allow for large TSO packets
 o extend tcp_maxmtu[46]() with a flag pointer to pass interface capabilities
 o adjust all callers of tcp_maxmtu[46]() accordingly

Discussed on:	-current, -net
Sponsored by:	TCP/IP Optimization Fundraise 2005
2006-09-06 21:51:59 +00:00
Gleb Smirnoff
2c857a9be9 o Backout rev. 1.125 of in_pcb.c. It appeared to behave extremely
bad under high load. For example with 40k sockets and 25k tcptw
  entries, connect() syscall can run for seconds. Debugging showed
  that it iterates the cycle millions times and purges thousands of
  tcptw entries at a time.
  Besides practical unusability this change is architecturally
  wrong. First, in_pcblookup_local() is used in connect() and bind()
  syscalls. No stale entries purging shouldn't be done here. Second,
  it is a layering violation.
o Return back the tcptw purging cycle to tcp_timer_2msl_tw(),
  that was removed in rev. 1.78 by rwatson. The commit log of this
  revision tells nothing about the reason cycle was removed. Now
  we need this cycle, since major cleaner of stale tcptw structures
  is removed.
o Disable probably necessary, but now unused
  tcp_twrecycleable() function.

Reviewed by:	ru
2006-09-06 13:56:35 +00:00
Gleb Smirnoff
c3e07bf82a Finally fix rev. 1.256
Pointy hat to:	glebius
2006-09-05 14:00:59 +00:00
Gleb Smirnoff
23ebab416c Remove extra parenthesis in last commit.
Nitpicked by:	ru
2006-09-05 12:22:54 +00:00
Gleb Smirnoff
1f1f90c3a7 - Make net.inet.tcp.maxtcptw modifiable at run time.
- If net.inet.tcp.maxtcptw was ever set explicitly, do
  not change it if kern.ipc.maxsockets is changed.
2006-09-05 12:08:47 +00:00
Mohan Srinivasan
2374501ca4 Fix for a bug that causes the computation of "len" in tcp_output() to
get messed up, resulting in an inconsistency between the TCP state
and so_snd.
2006-08-26 17:53:19 +00:00
Mohan Srinivasan
464469c713 Fixes an edge case bug in timewait handling where ticks rolling over causing
the timewait expiry to be exactly 0 corrupts the timewait queues (and that entry).
Reviewed by:	silby
2006-08-11 21:15:23 +00:00
Robert Watson
e850475248 Move soisdisconnected() in tcp_discardcb() to one of its calling contexts,
tcp_twstart(), but not to the other, tcp_detach(), as the socket is
already being torn down and therefore there are no listeners.  This avoids
a panic if kqueue state is registered on the socket at close(), and
eliminates to XXX comments.  There is one case remaining in which
tcp_discardcb() reaches up to the socket layer as part of the TCP host
cache, which would be good to avoid.

Reported by:	Goran Gajic <ggajic at afrodita dot rcub dot bg dot ac dot yu>
2006-08-02 16:18:05 +00:00
Robert Watson
a152f8a361 Change semantics of socket close and detach. Add a new protocol switch
function, pru_close, to notify protocols that the file descriptor or
other consumer of a socket is closing the socket.  pru_abort is now a
notification of close also, and no longer detaches.  pru_detach is no
longer used to notify of close, and will be called during socket
tear-down by sofree() when all references to a socket evaporate after
an earlier call to abort or close the socket.  This means detach is now
an unconditional teardown of a socket, whereas previously sockets could
persist after detach of the protocol retained a reference.

This faciliates sharing mutexes between layers of the network stack as
the mutex is required during the checking and removal of references at
the head of sofree().  With this change, pru_detach can now assume that
the mutex will no longer be required by the socket layer after
completion, whereas before this was not necessarily true.

Reviewed by:	gnn
2006-07-21 17:11:15 +00:00
Stephan Uphoff
d915b28015 Fix race conditions on enumerating pcb lists by moving the initialization
( and where appropriate the destruction) of the pcb mutex to the init/finit
functions of the pcb zones.
This allows locking of the pcb entries and race condition free comparison
of the generation count.
Rearrange locking a bit to avoid extra locking operation to update the generation
count in in_pcballoc(). (in_pcballoc now returns the pcb locked)

I am planning to convert pcb list handling from a type safe to a reference count
model soon. ( As this allows really freeing the PCBs)

Reviewed by:	rwatson@, mohans@
MFC after:	1 week
2006-07-18 22:34:27 +00:00
Robert Watson
10702a2840 Abstract inpcb drop logic, previously just setting of INP_DROPPED in TCP,
into in_pcbdrop().  Expand logic to detach the inpcb from its bound
address/port so that dropping a TCP connection releases the inpcb resource
reservation, which since the introduction of socket/pcb reference count
updates, has been persisting until the socket closed rather than being
released implicitly due to prior freeing of the inpcb on TCP drop.

MFC after:	3 months
2006-04-25 11:17:35 +00:00
Robert Watson
9106a6d6b0 Replace isn_mtx direct use with ISN_*() lock macros so that locking
details/strategy can be changed without touching every use.

MFC after:	3 months
2006-04-23 12:27:42 +00:00
Robert Watson
4c0e8f41f6 Introduce a new TCP mutex, isn_mtx, which protects the initial sequence
number state, rather than re-using pcbinfo.  This introduces some
additional mutex operations during isn query, but avoids hitting the TCP
pcbinfo lock out of yet another frequently firing TCP timer.

MFC after:	3 months
2006-04-22 19:23:24 +00:00
Paul Saab
4f590175b7 Allow for nmbclusters and maxsockets to be increased via sysctl.
An eventhandler is used to update all the various zones that depend
on these values.
2006-04-21 09:25:40 +00:00
Gleb Smirnoff
a73b656763 Add a tunable net.inet.tcp.maxtcptw, that allows to set a limit
on tcptw zone independently from setting a limit on socket zone.
2006-04-04 14:31:37 +00:00
Robert Watson
ae0e714308 Before dereferencing intotw() when INP_TIMEWAIT, check for inp_ppcb being
NULL.  We currently do allow this to happen, but may want to remove that
possibility in the future.  This case can occur when a socket is left
open after TCP wraps up, and the timewait state is recycled.  This will
be cleaned up in the future.

Found by:	Kazuaki Oda <kaakun at highway dot ne dot jp>
MFC after:	3 months
2006-04-04 12:26:07 +00:00
Robert Watson
cb895fb9b0 In TCP notify routines, check inpcb for INP_TIMEWAIT and INP_DROPPED.
The INP_DROPPED check replaces the current NULL checks; the INP_TIMEWAIT
checks appear to have always been required, but not been there, which
is/was a bug.  This avoids unconditionally casting of in_ppcb to a tcpcb,
when it may be a twtcb, which may have resulted in obscure ICMP-related
panics in earlier releases.

MFC after:	3 months
2006-04-03 14:07:50 +00:00
Robert Watson
afa39e25c4 Change inp_ppcb from caddr_t to void *, fix/remove associated related
casts.

Consistently use intotw() to cast inp_ppcb pointers to struct tcptw *
pointers.

Consistently use intotcpcb() to cast inp_ppcb pointers to struct tcpcb *
pointers.

Don't assign tp to the results to intotcpcb() during variable declation
at the top of functions, as that is before the asserts relating to
locking have been performed.  Do this later in the function after
appropriate assertions have run to allow that operation to be conisdered
safe.

MFC after:	3 months
2006-04-03 13:33:55 +00:00
Robert Watson
43f56a32a0 Style tweaks: convert to ANSI from K&R function prototypes.
MFC after:	3 months
2006-04-03 12:59:27 +00:00
Robert Watson
2fc5ae87d0 Update comment on tcp_close() for new world order.
MFC after:	3 months
2006-04-03 12:52:13 +00:00
Robert Watson
fa38deac65 Fix up locking surrounding tcp_drop sysctl: in the new world order, we
don't free inpcbs until after the socket is closed, so we always need
to unlock an inpcb after calling tcp_drop() on it.

MFC after:	3 months
2006-04-03 11:57:12 +00:00
Robert Watson
34af7bae80 Properly handle an edge case previously not handled correctly: a
socket can have a tcp connection that has entered time wait
attached to it, in the event that shutdown() is called on the
socket and the FINs properly exchange before close().  In this
case we don't detach or free the inpcb, just leave the tcptw
detached and freed, but we must release the inpcb lock (which we
didn't previously).

MFC after:	3 months
2006-04-01 23:53:25 +00:00
Robert Watson
623dce13c6 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
Andre Oppermann
eaf80179e2 Have TCP Inflight disable itself if the RTT is below a certain
threshold.  Inflight doesn't make sense on a LAN as it has
trouble figuring out the maximal bandwidth because of the coarse
tick granularity.

The sysctl net.inet.tcp.inflight.rttthresh specifies the threshold
in milliseconds below which inflight will disengage.  It defaults
to 10ms.

Tested by:	Joao Barros <joao.barros-at-gmail.com>,
		Rich Murphey <rich-at-whiteoaklabs.com>
Sponsored by:	TCP/IP Optimization Fundraise 2005
2006-02-16 19:38:07 +00:00
Andre Oppermann
34333b16cd Retire MT_HEADER mbuf type and change its users to use MT_DATA.
Having an additional MT_HEADER mbuf type is superfluous and redundant
as nothing depends on it.  It only adds a layer of confusion.  The
distinction between header mbuf's and data mbuf's is solely done
through the m->m_flags M_PKTHDR flag.

Non-native code is not changed in this commit.  For compatibility
MT_HEADER is mapped to MT_DATA.

Sponsored by:	TCP/IP Optimization Fundraise 2005
2005-11-02 13:46:32 +00:00
Philip Paeps
7691747aac Unbreak the net.inet6.tcp6.getcred sysctl.
This makes inetd/auth work again in IPv6 setups.

Pointy hat to:	ume/KAME
2005-10-12 09:24:18 +00:00
Maxim Konovalov
ac827533df o Teach sysctl_drop() how to deal with the sockets in TIME_WAIT state.
This is a special case because tcp_twstart() destroys a tcp control
block via tcp_discardcb() so we cannot call tcp_drop(struct *tcpcb) on
such connections.  Use tcp_twclose() instead.

MFC after:	5 days
2005-10-02 08:43:57 +00:00
Andre Oppermann
ffabe3dce8 In tcp_ctlinput() do not swap ip->ip_len a second time. It
has been done in icmp_input() already.

This fixes the ICMP_UNREACH_NEEDFRAG case where no MTU was
proposed in the ICMP reply.

PR:		kern/81813
Submitted by:	Vitezslav Novy <vita at fio.cz>
MFC after:	3 days
2005-09-10 07:43:29 +00:00
Andre Oppermann
e0aec68255 Use the correct mbuf type for MGET(). 2005-08-30 16:35:27 +00:00
Hajimu UMEMOTO
4dad226e45 recover the line which was wrongly disappeared during scope cleanup.
tcpdrop(8) should work for IPv6, again.
2005-08-01 12:08:49 +00:00
Hajimu UMEMOTO
a1f7e5f8ee scope cleanup. with this change
- most of the kernel code will not care about the actual encoding of
  scope zone IDs and won't touch "s6_addr16[1]" directly.
- similarly, most of the kernel code will not care about link-local
  scoped addresses as a special case.
- scope boundary check will be stricter.  For example, the current
  *BSD code allows a packet with src=::1 and dst=(some global IPv6
  address) to be sent outside of the node, if the application do:
    s = socket(AF_INET6);
    bind(s, "::1");
    sendto(s, some_global_IPv6_addr);
  This is clearly wrong, since ::1 is only meaningful within a single
  node, but the current implementation of the *BSD kernel cannot
  reject this attempt.

Submitted by:	JINMEI Tatuya <jinmei__at__isl.rdc.toshiba.co.jp>
Obtained from:	KAME
2005-07-25 12:31:43 +00:00
Robert Watson
f59a9ebf10 Remove no-op spl's and most comment references to spls, as TCP locking
is believed to be basically done (modulo any remaining bugs).

MFC after:	3 days
2005-07-19 12:21:26 +00:00
Paul Saab
482ac96888 Fix for a bug in the change that defers sack option processing until
after PAWS checks. The symptom of this is an inconsistency in the cached
sack state, caused by the fact that the sack scoreboard was not being
updated for an ACK handled in the header prediction path.

Found by:	Andrey Chernov.
Submitted by:	Noritoshi Demizu, Raja Mukerji.
Approved by:	re
2005-07-01 22:54:18 +00:00
Robert Watson
e3d5315d01 Assert tcbinfo lock in tcp_drop() due to its call of tcp_close()
Assert tcbinfo lock in tcp_close() due to its call to in{,6}_detach()
Assert tcbinfo lock in tcp_drop_syn_sent() due to its call to tcp_drop()

MFC after:	7 days
2005-06-01 12:06:07 +00:00
Colin Percival
fe2eee8231 Fix two issues which were missed in FreeBSD-SA-05:08.kmem.
Reported by:	Uwe Doering
2005-05-07 00:41:36 +00:00
Andre Oppermann
9e4ca6315d If we don't get a suggested MTU during path MTU discovery
look up the packet size of the packet that generated the
response, step down the MTU by one step through ip_next_mtu()
and try again.

Suggested by:	dwmalone
2005-05-04 13:48:44 +00:00
Paul Saab
a6235da61e - Make the sack scoreboard logic use the TAILQ macros. This improves
code readability and facilitates some anticipated optimizations in
  tcp_sack_option().
- Remove tcp_print_holes() and TCP_SACK_DEBUG.

Submitted by:	Raja Mukerji.
Reviewed by:	Mohan Srinivasan, Noritoshi Demizu.
2005-04-21 20:11:01 +00:00
Andre Oppermann
1aedbd9c80 Move Path MTU discovery ICMP processing from icmp_input() to
tcp_ctlinput() and subject it to active tcpcb and sequence
number checking.  Previously any ICMP unreachable/needfrag
message would cause an update to the TCP hostcache.  Now only
ICMP PMTU messages belonging to an active TCP session with
the correct src/dst/port and sequence number will update the
hostcache and complete the path MTU discovery process.

Note that we don't entirely implement the recommended counter
measures of Section 7.2 of the paper.  However we close down
the possible degradation vector from trivially easy to really
complex and resource intensive.  In addition we have limited
the smallest acceptable MTU with net.inet.tcp.minmss sysctl
for some time already, further reducing the effect of any
degradation due to an attack.

Security:	draft-gont-tcpm-icmp-attacks-03.txt Section 7.2
MFC after:	3 days
2005-04-21 14:29:34 +00:00
Andre Oppermann
1600372b6b Ignore ICMP Source Quench messages for TCP sessions. Source Quench is
ineffective, depreciated and can be abused to degrade the performance
of active TCP sessions if spoofed.

Replace a bogus call to tcp_quench() in tcp_output() with the direct
equivalent tcpcb variable assignment.

Security:	draft-gont-tcpm-icmp-attacks-03.txt Section 7.1
MFC after:	3 days
2005-04-21 12:37:12 +00:00
Paul Saab
e346eeff65 - If the reassembly queue limit was reached or if we couldn't allocate
a reassembly queue state structure, don't update (receiver) sack
  report.
- Similarly, if tcp_drain() is called, freeing up all items on the
  reassembly queue, clean the sack report.

Found, Submitted by:	Noritoshi Demizu <demizu at dd dot iij4u dot or dot jp>
Reviewed by:	Mohan Srinivasan (mohans at yahoo-inc dot com),
		Raja Mukerji (raja at moselle dot com).
2005-04-10 05:21:29 +00:00
Gleb Smirnoff
31199c8463 Use NET_CALLOUT_MPSAFE macro. 2005-03-01 12:01:17 +00:00
Maxim Konovalov
9945c0e21f o Add handling of an IPv4-mapped IPv6 address.
o Use SYSCTL_IN() macro instead of direct call of copyin(9).

Submitted by:	ume

o Move sysctl_drop() implementation to sys/netinet/tcp_subr.c where
most of tcp sysctls live.
o There are net.inet[6].tcp[6].getcred sysctls already, no needs in
a separate struct tcp_ident_mapping.

Suggested by:	ume
2005-02-14 07:37:51 +00:00
Hajimu UMEMOTO
6d0a982bdf teach scope of IPv6 address to net.inet6.tcp6.getcred.
MFC after:	1 week
2005-02-04 14:43:05 +00:00
Robert Watson
06456da2c6 Update an additional reference to the rate of ISN tick callouts that was
missed in tcp_subr.c:1.216: projected_offset must also reflect how often
the tcp_isn_tick() callout will fire.

MFC after:	2 weeks
Submitted by:	silby
2005-01-31 01:35:01 +00:00
Robert Watson
54082796aa Have tcp_isn_tick() fire 100 times a second, rather than HZ times a
second; since the default hz has changed to 1000 times a second,
this resulted in unecessary work being performed.

MFC after:		2 weeks
Discussed with:		phk, cperciva
General head nod:	silby
2005-01-30 23:30:28 +00:00
Warner Losh
c398230b64 /* -> /*- for license, minor formatting changes 2005-01-07 01:45:51 +00:00