Vitezslav Samel discovered that since 2.6.30.4+ active FTP can not work
over NAT. The "cause" of the problem was a fix of unacknowledged data
detection with NAT (commit a3a9f79e36).
However, actually, that fix uncovered a long standing bug in TCP conntrack:
when NAT was enabled, we simply updated the max of the right edge of
the segments we have seen (td_end), by the offset NAT produced with
changing IP/port in the data. However, we did not update the other parameter
(td_maxend) which is affected by the NAT offset. Thus that could drift
away from the correct value and thus resulted breaking active FTP.
The patch below fixes the issue by *not* updating the conntrack parameters
from NAT, but instead taking into account the NAT offsets in conntrack in a
consistent way. (Updating from NAT would be more harder and expensive because
it'd need to re-calculate parameters we already calculated in conntrack.)
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
When NAT helpers change the TCP packet size, the highest seen sequence
number needs to be corrected. This is currently only done upwards, when
the packet size is reduced the sequence number is unchanged. This causes
TCP conntrack to falsely detect unacknowledged data and decrease the
timeout.
Fix by updating the highest seen sequence number in both directions after
packet mangling.
Tested-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Introduce per-conntrack locks and use them instead of the global protocol
locks to avoid contention. Especially tcp_lock shows up very high in
profiles on larger machines.
This will also allow to simplify the upcoming reliable event delivery patches.
Signed-off-by: Patrick McHardy <kaber@trash.net>
This patch simplifies the conntrack event caching system by removing
several events:
* IPCT_[*]_VOLATILE, IPCT_HELPINFO and IPCT_NATINFO has been deleted
since the have no clients.
* IPCT_COUNTER_FILLING which is a leftover of the 32-bits counter
days.
* IPCT_REFRESH which is not of any use since we always include the
timeout in the messages.
After this patch, the existing events are:
* IPCT_NEW, IPCT_RELATED and IPCT_DESTROY, that are used to identify
addition and deletion of entries.
* IPCT_STATUS, that notes that the status bits have changes,
eg. IPS_SEEN_REPLY and IPS_ASSURED.
* IPCT_PROTOINFO, that reports that internal protocol information has
changed, eg. the TCP, DCCP and SCTP protocol state.
* IPCT_HELPER, that a helper has been assigned or unassigned to this
entry.
* IPCT_MARK and IPCT_SECMARK, that reports that the mark has changed, this
covers the case when a mark is set to zero.
* IPCT_NATSEQADJ, to report that there's updates in the NAT sequence
adjustment.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The patch below adds supporting TCP simultaneous open to conntrack. The
unused LISTEN state is replaced by a new state (SYN_SENT2) denoting the
second SYN sent from the reply direction in the new case. The state table
is updated and the function tcp_in_window is modified to handle
simultaneous open.
The functionality can fairly easily be tested by socat. A sample tcpdump
recording
23:21:34.244733 IP (tos 0x0, ttl 64, id 49224, offset 0, flags [DF], proto TCP (6), length 60) 192.168.0.254.2020 > 192.168.0.1.2020: S, cksum 0xe75f (correct), 3383710133:3383710133(0) win 5840 <mss 1460,sackOK,timestamp 173445629 0,nop,wscale 7>
23:21:34.244783 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto TCP (6), length 40) 192.168.0.1.2020 > 192.168.0.254.2020: R, cksum 0x0253 (correct), 0:0(0) ack 3383710134 win 0
23:21:36.038680 IP (tos 0x0, ttl 64, id 28092, offset 0, flags [DF], proto TCP (6), length 60) 192.168.0.1.2020 > 192.168.0.254.2020: S, cksum 0x704b (correct), 2634546729:2634546729(0) win 5840 <mss 1460,sackOK,timestamp 824213 0,nop,wscale 1>
23:21:36.038777 IP (tos 0x0, ttl 64, id 49225, offset 0, flags [DF], proto TCP (6), length 60) 192.168.0.254.2020 > 192.168.0.1.2020: S, cksum 0xb179 (correct), 3383710133:3383710133(0) ack 2634546730 win 5840 <mss 1460,sackOK,timestamp 173447423 824213,nop,wscale 7>
23:21:36.038847 IP (tos 0x0, ttl 64, id 28093, offset 0, flags [DF], proto TCP (6), length 52) 192.168.0.1.2020 > 192.168.0.254.2020: ., cksum 0xebad (correct), ack 3383710134 win 2920 <nop,nop,timestamp 824213 173447423>
and the corresponding netlink events:
[NEW] tcp 6 120 SYN_SENT src=192.168.0.254 dst=192.168.0.1 sport=2020 dport=2020 [UNREPLIED] src=192.168.0.1 dst=192.168.0.254 sport=2020 dport=2020
[UPDATE] tcp 6 120 LISTEN src=192.168.0.254 dst=192.168.0.1 sport=2020 dport=2020 src=192.168.0.1 dst=192.168.0.254 sport=2020 dport=2020
[UPDATE] tcp 6 60 SYN_RECV src=192.168.0.254 dst=192.168.0.1 sport=2020 dport=2020 src=192.168.0.1 dst=192.168.0.254 sport=2020 dport=2020
[UPDATE] tcp 6 432000 ESTABLISHED src=192.168.0.254 dst=192.168.0.1 sport=2020 dport=2020 src=192.168.0.1 dst=192.168.0.254 sport=2020 dport=2020 [ASSURED]
The RST packet was dropped in the raw table, thus it did not reach
conntrack. nfnetlink_conntrack is unpatched so it shows the new SYN_SENT2
state as the old unused LISTEN.
With TCP simultaneous open support we satisfy REQ-2 in RFC 5382 ;-) .
Additional minor correction in this patch is that in order to catch
uninitialized reply directions, "td_maxwin == 0" is used instead of
"td_end == 0" because the former can't be true except in uninitialized
state while td_end may accidentally be equal to zero in the mid of a
connection.
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Robert L Mathews discovered that some clients send evil TCP RST segments,
which are accepted by netfilter conntrack but discarded by the
destination. Thus the conntrack entry is destroyed but the destination
retransmits data until timeout.
The same technique, i.e. sending properly crafted RST segments, can easily
be used to bypass connlimit/connbytes based restrictions (the sample
script written by Robert can be found in the netfilter mailing list
archives).
The patch below adds a new flag and new field to struct ip_ct_tcp_state so
that checking RST segments can be made more strict and thus TCP conntrack
can catch the invalid ones: the RST segment is accepted only if its
sequence number higher than or equal to the highest ack we seen from the
other direction. (The last_ack field cannot be reused because it is used
to catch resent packets.)
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
This patch fixes an unaligned memory access in tcp_sack while reading
sequence numbers from TCP selective acknowledgement options. Prior to
applying this patch, upstream linux-2.6.27.20 was occasionally
generating messages like this on my sparc64 system:
[54678.532071] Kernel unaligned access at TPC[6b17d4] tcp_packet+0xcd4/0xd00
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Patrick McHardy <kaber@trash.net>
We currently use the negative value in the conntrack code to encode
the packet verdict in the error. As NF_DROP is equal to 0, inverting
NF_DROP makes no sense and, as a result, no packets are ever dropped.
Signed-off-by: Christoph Paasch <christoph.paasch@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
I want to compile out proc_* and sysctl_* handlers totally and
stub them to NULL depending on config options, however usage of &
will prevent this, since taking adress of NULL pointer will break
compilation.
So, drop & in front of every ->proc_handler and every ->strategy
handler, it was never needed in fact.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is cleaner, we already know conntrack to which event is relevant.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Again, it's deducible from skb, but we're going to use it for
nf_conntrack_checksum and statistics, so just pass it from upper layer.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
and (try to) consistently use u_int8_t for the L3 family.
Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>
In order to time out dead connections quicker, keep track of outstanding data
and cap the timeout.
Suggested by Herbert Xu.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a conntrack entry is destroyed in process context and destruction
is interrupted by packet processing and the packet is an attempt to
reopen a closed connection, TCP conntrack tries to kill the old entry
itself and returns NF_REPEAT to pass the packet through the hook
again. This may lead to an endless loop: TCP conntrack repeatedly
finds the old entry, but can not kill it itself since destruction
is already in progress, but destruction in process context can not
complete since TCP conntrack is keeping the CPU busy.
Drop the packet in TCP conntrack if we can't kill the connection
ourselves to avoid this.
Reported by: hemao77@gmail.com [ Kernel bugzilla #11058 ]
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Lost connections was reported by Thomas Bätzler (running 2.6.25 kernel) on
the netfilter mailing list (see the thread "Weird nat/conntrack Problem
with PASV FTP upload"). He provided tcpdump recordings which helped to
find a long lingering bug in conntrack.
In TCP connection tracking, checking the lower bound of valid ACK could
lead to mark valid packets as INVALID because:
- We have got a "higher or equal" inequality, but the test checked
the "higher" condition only; fixed.
- If the packet contains a SACK option, it could occur that the ACK
value was before the left edge of our (S)ACK "window": if a previous
packet from the other party intersected the right edge of the window
of the receiver, we could move forward the window parameters beyond
accepting a valid ack. Therefore in this patch we check the rightmost
SACK edge instead of the ACK value in the lower bound of valid (S)ACK
test.
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently the last packet of a connection isn't accounted when its causing
abnormal termination.
Introduces nf_ct_kill_acct() which increments the accounting counters on
conntrack kill. The new function was necessary, because there are calls
to nf_ct_kill() which don't need accounting:
nf_conntrack_proto_tcp.c line ~847:
Kills ct and returns NF_REPEAT. We don't want to count twice.
nf_conntrack_proto_tcp.c line ~880:
Kills ct and returns NF_DROP. I think we don't want to count dropped
packets.
nf_conntrack_netlink.c line ~824:
As far as I can see ctnetlink_del_conntrack() is used to destroy a
conntrack on behalf of the user. There is an sk_buff, but I don't think
this is an actual packet. Incrementing counters here is therefore not
desired.
Signed-off-by: Fabian Hugelshofer <hugelshofer2006@gmx.ch>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Encapsulate the common
if (del_timer(&ct->timeout))
ct->timeout.function((unsigned long)ct)
sequence in a new function.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Directly call IPv4 and IPv6 variants where the address family is
easily known.
Signed-off-by: Jan Engelhardt <jengelh@computergmbh.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Invalid states can cause out-of-bound memory accesses of the state table.
Also don't insist on having a new state contained in the netlink message.
Signed-off-by: Patrick McHardy <kaber@trash.net>
TCP connection tracking in netfilter did not handle TCP reopening
properly: active close was taken into account for one side only and
not for any side, which is fixed now. The patch includes more comments
to explain the logic how the different cases are handled.
The bug was discovered by Jeff Chua.
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Constify a few data tables use const qualifiers on variables where
possible in the nf_*_proto_tcp sources.
Signed-off-by: Jan Engelhardt <jengelh@computergmbh.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rename all "conntrack" variables to "ct" for more consistency and
avoiding some overly long lines.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Teach sparse about locking here, and fix signed/unsigned warnings.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of keeping pointers to the timeout values in a table, simply
put the timeout values in the table directly.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The TCP and SCTP conntrack state transition tables only holds
small numbers, but gcc uses 4 byte per entry for the enum. Switching
to an u8 reduces the size from 480 to 120 bytes for TCP and from
576 to 144 bytes for SCTP.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The IPv4 and IPv6 hook values are identical, yet some code tries to figure
out the "correct" value by looking at the address family. Introduce NF_INET_*
values for both IPv4 and IPv6. The old values are kept in a #ifndef __KERNEL__
section for userspace compatibility.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6: (51 commits)
[IPV6]: Fix again the fl6_sock_lookup() fixed locking
[NETFILTER]: nf_conntrack_tcp: fix connection reopening fix
[IPV6]: Fix race in ipv6_flowlabel_opt() when inserting two labels
[IPV6]: Lost locking in fl6_sock_lookup
[IPV6]: Lost locking when inserting a flowlabel in ipv6_fl_list
[NETFILTER]: xt_sctp: fix mistake to pass a pointer where array is required
[NET]: Fix OOPS due to missing check in dev_parse_header().
[TCP]: Remove lost_retrans zero seqno special cases
[NET]: fix carrier-on bug?
[NET]: Fix uninitialised variable in ip_frag_reasm()
[IPSEC]: Rename mode to outer_mode and add inner_mode
[IPSEC]: Disallow combinations of RO and AH/ESP/IPCOMP
[IPSEC]: Use the top IPv4 route's peer instead of the bottom
[IPSEC]: Store afinfo pointer in xfrm_mode
[IPSEC]: Add missing BEET checks
[IPSEC]: Move type and mode map into xfrm_state.c
[IPSEC]: Fix length check in xfrm_parse_spi
[IPSEC]: Move ip_summed zapping out of xfrm6_rcv_spi
[IPSEC]: Get nexthdr from caller in xfrm6_rcv_spi
[IPSEC]: Move tunnel parsing for IPv4 out of xfrm4_input
...
No one has bothered to set strategy routine for the the netfilter sysctls that
return jiffies to be sysctl_jiffies.
So it appears the sys_sysctl path is unused and untested, so this patch
removes the binary sysctl numbers.
Which fixes the netfilter oops in 2.6.23-rc2-mm2 for me.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Patrick McHardy <kaber@trash.net>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If one side aborts an established connection, the entry still lingers
for 10s in conntrack for the late packets. Allow to open up the
connection again for the party which sent the RST packet.
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Tested-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
With your description I could reproduce the bug and actually you were
completely right: the code above is incorrect. Somehow I was able to
misread RFC1122 and mixed the roles :-(:
When a connection is >>closed actively<<, it MUST linger in
TIME-WAIT state for a time 2xMSL (Maximum Segment Lifetime).
However, it MAY >>accept<< a new SYN from the remote TCP to
reopen the connection directly from TIME-WAIT state, if it:
[...]
The fix is as follows: if the receiver initiated an active close, then the
sender may reopen the connection - otherwise try to figure out if we hold
a dead connection.
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Tested-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is no struct nfattr anymore, rename functions to 'nlattr'.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Get rid of the duplicated rtnetlink macros and use the generic netlink
attribute functions. The old duplicated stuff is moved to a new header
file that exists just for userspace.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch cleans up duplicate includes in
net/netfilter/
Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>