It doesn't log a packet if sysctl_log_invalid isn't equal to protonum
OR sysctl_log_invalid isn't equal to IPPROTO_RAW. This sentence is
always true. I believe we need to replace OR to AND.
Cc: Florian Westphal <fw@strlen.de>
Fixes: c4f3db1595 ("netfilter: conntrack: add and use nf_l4proto_log_invalid")
Signed-off-by: Andrei Vagin <avagin@gmail.com>
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
setting net.netfilter.nf_conntrack_timestamp=1 breaks xmit with fq
scheduler. skb->tstamp might be "refreshed" using ktime_get_real(),
but fq expects CLOCK_MONOTONIC.
This patch removes all places in netfilter that check/set skb->tstamp:
1. To fix the bogus "start" time seen with conntrack timestamping for
outgoing packets, never use skb->tstamp and always use current time.
2. In nfqueue and nflog, only use skb->tstamp for incoming packets,
as determined by current hook (prerouting, input, forward).
3. xt_time has to use system clock as well rather than skb->tstamp.
We could still use skb->tstamp for prerouting/input/foward, but
I see no advantage to make this conditional.
Fixes: fb420d5d91 ("tcp/fq: move back to CLOCK_MONOTONIC")
Cc: Eric Dumazet <edumazet@google.com>
Reported-by: Michal Soltys <soltys@ziu.info>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Sven Auhagen reported that a 2nd ping request will fail if 'fully-random'
mode is used.
Reason is that if no proto information is given, min/max are both 0,
so we set the icmp id to 0 instead of chosing a random value between
0 and 65535.
Update test case as well to catch this, without fix this yields:
[..]
ERROR: cannot ping ns1 from ns2 with ip masquerade fully-random (attempt 2)
ERROR: cannot ping ns1 from ns2 with ipv6 masquerade fully-random (attempt 2)
... becaus 2nd ping clashes with existing 'id 0' icmp conntrack and gets
dropped.
Fixes: 203f2e7820 ("netfilter: nat: remove l4proto->unique_tuple")
Reported-by: Sven Auhagen <sven.auhagen@voleatech.de>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
I believe that "hook->num" can be up to UINT_MAX. Shifting more than
31 bits would is undefined in C but in practice it would lead to shift
wrapping. That would lead to an array overflow in nf_tables_addchain():
ops->hook = hook.type->hooks[ops->hooknum];
Fixes: fe19c04ca1 ("netfilter: nf_tables: remove nhooks field from struct nft_af_info")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
else, we leak the addresses to userspace via ctnetlink events
and dumps.
Compute an ID on demand based on the immutable parts of nf_conn struct.
Another advantage compared to using an address is that there is no
immediate re-use of the same ID in case the conntrack entry is freed and
reallocated again immediately.
Fixes: 3583240249 ("[NETFILTER]: nf_conntrack_expect: kill unique ID")
Fixes: 7f85f91472 ("[NETFILTER]: nf_conntrack: kill unique ID")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
We can receive ICMP errors from client or from
tunneling real server. While the former can be
scheduled to real server, the latter should
not be scheduled, they are decapsulated only when
existing connection is found.
Fixes: 6044eeffaf ("ipvs: attempt to schedule icmp packets")
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Luca Moro says:
------
The issue lies in the filtering of ICMP and ICMPv6 errors that include an
inner IP datagram.
For these packets, icmp_error_message() extract the ICMP error and inner
layer to search of a known state.
If a state is found the packet is tagged as related (IP_CT_RELATED).
The problem is that there is no correlation check between the inner and
outer layer of the packet.
So one can encapsulate an error with an inner layer matching a known state,
while its outer layer is directed to a filtered host.
In this case the whole packet will be tagged as related.
This has various implications from a rule bypass (if a rule to related
trafic is allow), to a known state oracle.
Unfortunately, we could not find a real statement in a RFC on how this case
should be filtered.
The closest we found is RFC5927 (Section 4.3) but it is not very clear.
A possible fix would be to check that the inner IP source is the same than
the outer destination.
We believed this kind of attack was not documented yet, so we started to
write a blog post about it.
You can find it attached to this mail (sorry for the extract quality).
It contains more technical details, PoC and discussion about the identified
behavior.
We discovered later that
https://www.gont.com.ar/papers/filtering-of-icmp-error-messages.pdf
described a similar attack concept in 2004 but without the stateful
filtering in mind.
-----
This implements above suggested fix:
In icmp(v6) error handler, take outer destination address, then pass
that into the common function that does the "related" association.
After obtaining the nf_conn of the matching inner-headers connection,
check that the destination address of the opposite direction tuple
is the same as the outer address and only set RELATED if thats the case.
Reported-by: Luca Moro <luca.moro@synacktiv.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Replace NF_HOOK() based invocation of the netfilter hooks with a private
copy of nf_hook_slow().
This copy has one difference: it can return the rx handler value expected
by the stack, i.e. RX_HANDLER_CONSUMED or RX_HANDLER_PASS.
This is needed by the next patch to invoke the ebtables
"broute" table via the standard netfilter hooks rather than the custom
"br_should_route_hook" indirection that is used now.
When the skb is to be "brouted", we must return RX_HANDLER_PASS from the
bridge rx input handler, but there is no way to indicate this via
NF_HOOK(), unless perhaps by some hack such as exposing bridge_cb in the
netfilter core or a percpu flag.
text data bss dec filename
3369 56 0 3425 net/bridge/br_input.o.before
3458 40 0 3498 net/bridge/br_input.o.after
This allows removal of the "br_should_route_hook" in the next patch.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: David S. Miller <davem@davemloft.net>
Acked-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Only reason for having two different register functions was because of
ipt_MASQUERADE and ip6t_MASQUERADE being two different modules.
Previous patch merged those into xt_MASQUERADE, so we can merge this too.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
No need to have separate modules for this.
before:
text data bss dec filename
2038 1168 0 3206 net/ipv4/netfilter/ipt_MASQUERADE.ko
1526 1024 0 2550 net/ipv6/netfilter/ip6t_MASQUERADE.ko
after:
text data bss dec filename
2521 1296 0 3817 net/netfilter/xt_MASQUERADE.ko
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Both are now implemented by nf_nat_masquerade.c, so no need to keep
different headers.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Add version option support to the nftables "osf" expression.
Signed-off-by: Fernando Fernandez Mancera <ffmancera@riseup.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
allows to redirect both ipv4 and ipv6 with a single rule in an
inet nat table.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This allows use of a single masquerade rule in nat inet family
to handle both ipv4 and ipv6.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
NF_NAT_NEEDED is true whenever nat support for either ipv4 or ipv6 is
enabled. Now that the af-specific nat configuration switches have been
removed, IS_ENABLED(CONFIG_NF_NAT) has the same effect.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
very little code, so it really doesn't make sense to have extra
modules or even a kconfig knob for this.
Merge them and make functionality available unconditionally.
The merge makes inet family route support trivial, so add it
as well here.
Before:
text data bss dec hex filename
835 832 0 1667 683 nft_chain_route_ipv4.ko
870 832 0 1702 6a6 nft_chain_route_ipv6.ko
111568 2556 529 114653 1bfdd nf_tables.ko
After:
text data bss dec hex filename
113133 2556 529 116218 1c5fa nf_tables.ko
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
We need minimal support from the nat core for this, as we do not
want to register additional base hooks.
When an inet hook is registered, interally register ipv4 and ipv6
hooks for them and unregister those when inet hooks are removed.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
ipip packets are blocked in some public cloud environments, this patch
allows gue encapsulation with the tunneling method, which would make
tunneling working in those environments.
Signed-off-by: Jacky Hu <hengqing.hu@gmail.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Fix sparse warning:
net/netfilter/nft_redir.c:85:5:
warning: symbol 'nft_redir_dump' was not declared. Should it be static?
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Function nf_tables_set_desc_parse parameter ctx is not being used
so remove it as it is redundant.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
there is a similar helper in net/netfilter/nf_tables_api.c,
this maybe become a common request someday, so move it to
time.c
Signed-off-by: Zhang Yu <zhangyu31@baidu.com>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
Acked-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This patch changes rhashtables to use a bit_spin_lock on BIT(1) of the
bucket pointer to lock the hash chain for that bucket.
The benefits of a bit spin_lock are:
- no need to allocate a separate array of locks.
- no need to have a configuration option to guide the
choice of the size of this array
- locking cost is often a single test-and-set in a cache line
that will have to be loaded anyway. When inserting at, or removing
from, the head of the chain, the unlock is free - writing the new
address in the bucket head implicitly clears the lock bit.
For __rhashtable_insert_fast() we ensure this always happens
when adding a new key.
- even when lockings costs 2 updates (lock and unlock), they are
in a cacheline that needs to be read anyway.
The cost of using a bit spin_lock is a little bit of code complexity,
which I think is quite manageable.
Bit spin_locks are sometimes inappropriate because they are not fair -
if multiple CPUs repeatedly contend of the same lock, one CPU can
easily be starved. This is not a credible situation with rhashtable.
Multiple CPUs may want to repeatedly add or remove objects, but they
will typically do so at different buckets, so they will attempt to
acquire different locks.
As we have more bit-locks than we previously had spinlocks (by at
least a factor of two) we can expect slightly less contention to
go with the slightly better cache behavior and reduced memory
consumption.
To enhance type checking, a new struct is introduced to represent the
pointer plus lock-bit
that is stored in the bucket-table. This is "struct rhash_lock_head"
and is empty. A pointer to this needs to be cast to either an
unsigned lock, or a "struct rhash_head *" to be useful.
Variables of this type are most often called "bkt".
Previously "pprev" would sometimes point to a bucket, and sometimes a
->next pointer in an rhash_head. As these are now different types,
pprev is NULL when it would have pointed to the bucket. In that case,
'blk' is used, together with correct locking protocol.
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch exports nf_ct_set_timeout() and nf_ct_destroy_timeout().
The two functions are derived from xt_ct_destroy_timeout() and
xt_ct_set_timeout() in xt_CT.c, and moved to nf_conntrack_timeout.c
without any functional change.
It would be useful for other users (i.e. OVS) that utilizes the
finer-grain conntrack timeout feature.
CC: Pablo Neira Ayuso <pablo@netfilter.org>
CC: Pravin Shelar <pshelar@ovn.org>
Signed-off-by: Yi-Hung Wei <yihung.wei@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since maxattr is common, the policy can't really differ sanely,
so make it common as well.
The only user that did in fact manage to make a non-common policy
is taskstats, which has to be really careful about it (since it's
still using a common maxattr!). This is no longer supported, but
we can fake it using pre_doit.
This reduces the size of e.g. nl80211.o (which has lots of commands):
text data bss dec hex filename
398745 14323 2240 415308 6564c net/wireless/nl80211.o (before)
397913 14331 2240 414484 65314 net/wireless/nl80211.o (after)
--------------------------------
-832 +8 0 -824
Which is obviously just 8 bytes for each command, and an added 8
bytes for the new policy pointer. I'm not sure why the ops list is
counted as .text though.
Most of the code transformations were done using the following spatch:
@ops@
identifier OPS;
expression POLICY;
@@
struct genl_ops OPS[] = {
...,
{
- .policy = POLICY,
},
...
};
@@
identifier ops.OPS;
expression ops.POLICY;
identifier fam;
expression M;
@@
struct genl_family fam = {
.ops = OPS,
.maxattr = M,
+ .policy = POLICY,
...
};
This also gets rid of devlink_nl_cmd_region_read_dumpit() accessing
the cb->data as ops, which we want to change in a later genl patch.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
->release_ops() callback releases resources and this is used in error path.
If nf_tables_newrule() fails after ->select_ops(), it should release
resources. but it can not call ->destroy() because that should be called
after ->init().
At this point, ->release_ops() should be used for releasing resources.
Test commands:
modprobe -rv xt_tcpudp
iptables-nft -I INPUT -m tcp <-- error command
lsmod
Result:
Module Size Used by
xt_tcpudp 20480 2 <-- it should be 0
Fixes: b8e2040063 ("netfilter: nft_compat: use .release_ops and remove list of extension")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
AF_INET4 does not exist.
Fixes: c78efc99c7 ("netfilter: nf_tables: nat: merge nft_redir protocol specific modules)"
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Proper use counter updates when activating and deactivating the object,
otherwise, this hits bogus EBUSY error.
Fixes: cd5125d8f5 ("netfilter: nf_tables: split set destruction in deactivate and destroy phase")
Reported-by: Laura Garcia <nevola@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
With NETFILTER_XT_TARGET_TEE=y and IP6_NF_IPTABLES=m, we get a link
error when referencing the NF_DUP_IPV6 module:
net/netfilter/xt_TEE.o: In function `tee_tg6':
xt_TEE.c:(.text+0x14): undefined reference to `nf_dup_ipv6'
The problem here is the 'select NF_DUP_IPV6 if IP6_NF_IPTABLES'
that forces NF_DUP_IPV6 to be =m as well rather than setting it
to =y as was intended here. Adding a soft dependency on
IP6_NF_IPTABLES avoids that broken configuration.
Fixes: 5d400a4933 ("netfilter: Kconfig: Change select IPv6 dependencies")
Cc: Máté Eckl <ecklm94@gmail.com>
Cc: Taehee Yoo <ap420073@gmail.com>
Link: https://patchwork.ozlabs.org/patch/999498/
Link: https://lore.kernel.org/patchwork/patch/960062/
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Otherwise, we hit bogus ENOENT when removing elements.
Fixes: e701001e7c ("netfilter: nft_rbtree: allow adjacent intervals with dynamic updates")
Reported-by: Václav Zindulka <vaclav.zindulka@tlapnet.cz>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
When running 'nft flush ruleset' while no rules exist, we will increment
the generation counter and announce a new genid to userspace, yet
nothing had changed in the first place.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Set deletion after flush coming in the same batch results in EBUSY. Add
set use counter to track the number of references to this set from
rules. We cannot rely on the list of bindings for this since such list
is still populated from the preparation phase.
Reported-by: Václav Zindulka <vaclav.zindulka@tlapnet.cz>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The abort path can cause a double-free of an anonymous set.
Added-and-to-be-aborted rule looks like this:
udp dport { 137, 138 } drop
The to-be-aborted transaction list looks like this:
newset
newsetelem
newsetelem
rule
This gets walked in reverse order, so first pass disables the rule, the
set elements, then the set.
After synchronize_rcu(), we then destroy those in same order: rule, set
element, set element, newset.
Problem is that the anonymous set has already been bound to the rule, so
the rule (lookup expression destructor) already frees the set, when then
cause use-after-free when trying to delete the elements from this set,
then try to free the set again when handling the newset expression.
Rule releases the bound set in first place from the abort path, this
causes the use-after-free on set element removal when undoing the new
element transactions. To handle this, skip new element transaction if
set is bound from the abort path.
This is still causes the use-after-free on set element removal. To
handle this, remove transaction from the list when the set is already
bound.
Joint work with Florian Westphal.
Fixes: f6ac858589 ("netfilter: nf_tables: unbind set in rule from commit path")
Bugzilla: https://bugzilla.netfilter.org/show_bug.cgi?id=1325
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Otherwise, we get notifier list corruption.
This is the most simple fix: remove the device notifier call chain
from the ipv6 masquerade register function and handle it only
in the ipv4 version.
The better fix is merge
nf_nat_masquerade_ipv4/6_(un)register_notifier
into a single
nf_nat_masquerade_(un)register_notifiers
but to do this its needed to first merge the two masquerade modules
into a single xt_MASQUERADE.
Furthermore, we need to use different refcounts for ipv4/ipv6
until we can merge MASQUERADE.
Fixes: d1aca8ab31 ("netfilter: nat: merge ipv4 and ipv6 masquerade functionality")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Merge the ipv4 and ipv6 nat chain type. This is the last
missing piece which allows to provide inet family support
for nat in a follow patch.
The kconfig knobs for ipv4/ipv6 nat chain are removed, the
nat chain type will be built unconditionally if NFT_NAT
expression is enabled.
Before:
text data bss dec hex filename
1576 896 0 2472 9a8 nft_chain_nat_ipv4.ko
1697 896 0 2593 a21 nft_chain_nat_ipv6.ko
After:
text data bss dec hex filename
1832 896 0 2728 aa8 nft_chain_nat.ko
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The family specific masq modules are way too small to warrant
an extra module, just place all of them in nft_masq.
before:
text data bss dec hex filename
1001 832 0 1833 729 nft_masq.ko
766 896 0 1662 67e nft_masq_ipv4.ko
764 896 0 1660 67c nft_masq_ipv6.ko
after:
2010 960 0 2970 b9a nft_masq.ko
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
before:
text data bss dec hex filename
990 832 0 1822 71e nft_redir.ko
697 896 0 1593 639 nft_redir_ipv4.ko
713 896 0 1609 649 nft_redir_ipv6.ko
after:
text data bss dec hex filename
1910 960 0 2870 b36 nft_redir.ko
size is reduced, all helpers from nft_redir.ko can be made static.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Use struct device_attribute instead of struct idletimer_tg_attr, and
the correct callback function type to avoid indirect call mismatches
with Control Flow Integrity checking.
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
CONNTRACK_LOCKS is divisor when computer array index, if it is power of
2, compiler will optimize modulo operation as bitwise AND, or else
modulo will lower performance.
Suggested-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Check the result of dereferencing base_chain->stats, instead of result
of this_cpu_ptr with NULL.
base_chain->stats maybe be changed to NULL when a chain is updated and a
new NULL counter can be attached.
And we do not need to check returning of this_cpu_ptr since
base_chain->stats is from percpu allocator if it is non-NULL,
this_cpu_ptr returns a valid value.
And fix two sparse error by replacing rcu_access_pointer and
rcu_dereference with READ_ONCE under rcu_read_lock.
Thanks for Eric's help to finish this patch.
Fixes: 009240940e ("netfilter: nf_tables: don't assume chain stats are set when jumplabel is set")
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Zhang Yu <zhangyu31@baidu.com>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
sctp_csum_check() is called by sctp_s/dnat_handler() where it calls
skb_make_writable() to ensure sctphdr to be linearized.
So there's no need to get sctphdr by calling skb_header_pointer()
in sctp_csum_check().
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The proto in struct xt_match and struct xt_target is u16, when
calling xt_check_target/match, their proto argument is u8,
and will cause truncation, it is harmless to ip packet, since
ip proto is u8
if a etable's match/target has proto that is u16, will cause
the check failure.
and convert be16 to short in bridge/netfilter/ebtables.c
Signed-off-by: Zhang Yu <zhangyu31@baidu.com>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The metadata_dst does not initialize the dst_cache field, this causes
problems to ip_md_tunnel_xmit() since it cannot use this cache, hence,
Triggering a route lookup for every packet.
Signed-off-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
TCP resets cause instant transition from established to closed state
provided the reset is in-window. Endpoints that implement RFC 5961
require resets to match the next expected sequence number.
RST segments that are in-window (but that do not match RCV.NXT) are
ignored, and a "challenge ACK" is sent back.
Main problem for conntrack is that its a middlebox, i.e. whereas an end
host might have ACK'd SEQ (and would thus accept an RST with this
sequence number), conntrack might not have seen this ACK (yet).
Therefore we can't simply flag RSTs with non-exact match as invalid.
This updates RST processing as follows:
1. If the connection is in a state other than ESTABLISHED, nothing is
changed, RST is subject to normal in-window check.
2. If the RSTs sequence number either matches exactly RCV.NXT,
connection state moves to CLOSE.
3. The same applies if the RST sequence number aligns with a previous
packet in the same direction.
In all other cases, the connection remains in ESTABLISHED state.
If the normal-in-window check passes, the timeout will be lowered
to that of CLOSE.
If the peer sends a challenge ack, connection timeout will be reset.
If the challenge ACK triggers another RST (RST was valid after all),
this 2nd RST will match expected sequence and conntrack state changes to
CLOSE.
If no challenge ACK is received, the connection will time out after
CLOSE seconds (10 seconds by default), just like without this patch.
Packetdrill test case:
0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
0.000 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
0.000 bind(3, ..., ...) = 0
0.000 listen(3, 1) = 0
0.100 < S 0:0(0) win 32792 <mss 1460,sackOK,nop,nop,nop,wscale 7>
0.100 > S. 0:0(0) ack 1 win 64240 <mss 1460,nop,nop,sackOK,nop,wscale 7>
0.200 < . 1:1(0) ack 1 win 257
0.200 accept(3, ..., ...) = 4
// Receive a segment.
0.210 < P. 1:1001(1000) ack 1 win 46
0.210 > . 1:1(0) ack 1001
// Application writes 1000 bytes.
0.250 write(4, ..., 1000) = 1000
0.250 > P. 1:1001(1000) ack 1001
// First reset, old sequence. Conntrack (correctly) considers this
// invalid due to failed window validation (regardless of this patch).
0.260 < R 2:2(0) ack 1001 win 260
// 2nd reset, but too far ahead sequence. Same: correctly handled
// as invalid.
0.270 < R 99990001:99990001(0) ack 1001 win 260
// in-window, but not exact sequence.
// Current Linux kernels might reply with a challenge ack, and do not
// remove connection.
// Without this patch, conntrack state moves to CLOSE.
// With patch, timeout is lowered like CLOSE, but connection stays
// in ESTABLISHED state.
0.280 < R 1010:1010(0) ack 1001 win 260
// Expect challenge ACK
0.281 > . 1001:1001(0) ack 1001 win 501
// With or without this patch, RST will cause connection
// to move to CLOSE (sequence number matches)
// 0.282 < R 1001:1001(0) ack 1001 win 260
// ACK
0.300 < . 1001:1001(0) ack 1001 win 257
// more data could be exchanged here, connection
// is still established
// Client closes the connection.
0.610 < F. 1001:1001(0) ack 1001 win 260
0.650 > . 1001:1001(0) ack 1002
// Close the connection without reading outstanding data
0.700 close(4) = 0
// so one more reset. Will be deemed acceptable with patch as well:
// connection is already closing.
0.701 > R. 1001:1001(0) ack 1002 win 501
// End packetdrill test case.
With patch, this generates following conntrack events:
[NEW] 120 SYN_SENT src=10.0.2.1 dst=10.0.0.1 sport=5437 dport=80 [UNREPLIED]
[UPDATE] 60 SYN_RECV src=10.0.2.1 dst=10.0.0.1 sport=5437 dport=80
[UPDATE] 432000 ESTABLISHED src=10.0.2.1 dst=10.0.0.1 sport=5437 dport=80 [ASSURED]
[UPDATE] 120 FIN_WAIT src=10.0.2.1 dst=10.0.0.1 sport=5437 dport=80 [ASSURED]
[UPDATE] 60 CLOSE_WAIT src=10.0.2.1 dst=10.0.0.1 sport=5437 dport=80 [ASSURED]
[UPDATE] 10 CLOSE src=10.0.2.1 dst=10.0.0.1 sport=5437 dport=80 [ASSURED]
Without patch, first RST moves connection to close, whereas socket state
does not change until FIN is received.
[NEW] 120 SYN_SENT src=10.0.2.1 dst=10.0.0.1 sport=5141 dport=80 [UNREPLIED]
[UPDATE] 60 SYN_RECV src=10.0.2.1 dst=10.0.0.1 sport=5141 dport=80
[UPDATE] 432000 ESTABLISHED src=10.0.2.1 dst=10.0.0.1 sport=5141 dport=80 [ASSURED]
[UPDATE] 10 CLOSE src=10.0.2.1 dst=10.0.0.1 sport=5141 dport=80 [ASSURED]
Cc: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
David S. Miller