| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Derek Dai reports:
"If there are a lot of command in JSON node, seqnum_to_json() will slow
down application (eg: firewalld) dramatically since it iterate whole
command list every time."
He sent a patch implementing a lookup table, but we can do better: Speed
this up by introducing a hash table to store the struct json_cmd_assoc
objects in, taking their netlink sequence number as key.
Quickly tested restoring a ruleset containing about 19k rules:
| # time ./before/nft -jeaf large_ruleset.json >/dev/null
| 4.85user 0.47system 0:05.48elapsed 97%CPU (0avgtext+0avgdata 69732maxresident)k
| 0inputs+0outputs (15major+16937minor)pagefaults 0swaps
| # time ./after/nft -jeaf large_ruleset.json >/dev/null
| 0.18user 0.44system 0:00.70elapsed 89%CPU (0avgtext+0avgdata 68484maxresident)k
| 0inputs+0outputs (15major+16645minor)pagefaults 0swaps
Bugzilla: https://bugzilla.netfilter.org/show_bug.cgi?id=1479
Reported-by: Derek Dai <daiderek@gmail.com>
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
| |
|
|
|
|
|
|
|
| |
libnftnl has been changed to bring the format of registers in bitwise
dumps in line with those in other types of expression. Update the
expected output of Python test-cases.
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In ARP header, destination ether address sits between source IP and
destination IP addresses. Enum arp_hdr_fields had this wrong, which
in turn caused wrong ordering of entries in proto_arp->templates. When
expanding a combined payload expression, code assumes that template
entries are ordered by header offset, therefore the destination ether
address match was printed as raw if an earlier field was matched as
well:
| arp saddr ip 192.168.1.1 arp daddr ether 3e:d1:3f:d6:12:0b
was printed as:
| arp saddr ip 192.168.1.1 @nh,144,48 69068440080907
Note: Although strictly not necessary, reorder fields in
proto_arp->templates as well to match their actual ordering, just to
avoid confusion.
Fixes: 4b0f2a712b579 ("src: support for arp sender and target ethernet and IPv4 addresses")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
| |
|
|
|
|
|
|
|
| |
nft-test.py only needs one payload per rule, but a number of rules have
duplicates, typically one per address family, so just keep the last
payload for rules listed more than once.
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
| |
To similar change as in previous one, this time for the
jason (de)serialization.
Re-uses the raw payload match syntax, i.e. base,offset,length.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
| |
tcp option @42,16,4 (@kind,offset,length).
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
nft currently doesn't allow to check for presence of arbitrary tcp options.
Only known options where nft provides a template can be tested for.
This allows to test for presence of raw protocol values as well.
Example:
tcp option 42 exists
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently we're limited to ten template fields in exthdr_desc struct.
Using a single enum for all tpc option fields thus won't work
indefinitely (TCPOPTHDR_FIELD_TSECR is 9) when new option templates get
added.
Fortunately we can just use one enum per tcp option to avoid this.
As a side effect this also allows to simplify the sack offset
calculations. Rather than computing that on-the-fly, just add extra
fields to the SACK template.
expr->exthdr.offset now holds the 'raw' value, filled in from the option
template. This would ease implementation of 'raw option matching'
using offset and length to load from the option.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
| |
'nop' is the tcp padding "option". "noop" is retained for compatibility
on parser side.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
tcpopt template mapping is asymmetric:
one mapping is to match dumped netlink exthdr expression to the original
tcp option template.
This struct is indexed by the raw, on-write kind/type number.
The other mapping maps parsed options to the tcp option template.
Remove the latter. The parser is changed to translate the textual
option name, e.g. "maxseg" to the on-wire number.
This avoids the second mapping, it will also allow to more easily
support raw option matching in a followup patch.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
One was added by the tcp option parsing ocde, the other by synproxy.
So we have:
synproxy ... sack-perm
synproxy ... mss
and
tcp option maxseg
tcp option sack-permitted
This kills the extra tokens on the scanner/parser side,
so sack-perm and sack-permitted can both be used.
Likewise, 'synproxy maxseg' and 'tcp option mss size 42' will work too.
On the output side, the shorter form is now preferred, i.e. sack-perm
and mss.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
| |
Tests currently fail with
ip6/dnat.t: WARNING: line 8: ... because test still expects a range expression.
Fixes: ee4391d0ac1e7 ("nat: transform range to prefix expression when possible")
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
JSON in/output doesn't know about nat_type and thus cannot save/restore
nat mappings involving prefixes or concatenations because the snat
statement lacks the prefix/concat/interval type flags.
Furthermore, bison parser was extended to support netmap.
This is done with an internal 'netmap' flag that is passed to the
kernel. We need to dump/restore that as well.
Also make sure ip/snat.t passes in json mode.
Fixes: 35a6b10c1bc4 ("src: add netmap support")
Fixes: 9599d9d25a6b ("src: NAT support for intervals in maps")
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
| |
make dnat.t pass in json mode.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Fix warnings and errors when running nf-test.py -j due to missing json test case updates.
This also makes bridge/reject.t pass in json mode.
No code changes.
Fixes: 8615ed93f6e4c4 ("evaluate: enable reject with 802.1q")
Fixes: fae0a0972d7a71 ("tests: py: Enable anonymous set rule with concatenated ranges in inet/sets.t")
Fixes: 2a20b5bdbde8a1 ("datatype: add frag-needed (ipv4) to reject options")
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a prefix expression's length is on a byte-boundary, it is sufficient
to just reduce the length passed to "cmp" expression. No need for
explicit bitwise modification of data on LHS. The relevant code is
already there, used for string prefix matches. There is one exception
though, namely zero-length prefixes: Kernel doesn't accept zero-length
"cmp" expressions, so keep them in the old code-path for now.
This patch depends upon the previous one to correctly parse odd-sized
payload matches but has to extend support for non-payload LHS as well.
In practice, this is needed for "ct" expressions as they allow matching
against IP address prefixes, too.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
When expanding a payload match, don't disregard oversized templates at
the right offset. A more flexible user may extract less bytes from the
packet if only parts of a field are interesting, e.g. only the prefix of
source/destination address. Support that by using the template, but fix
the length. Later when creating a relational expression for it, detect
the unusually small payload expression length and turn the RHS value
into a prefix expression.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
| |
|
|
|
|
|
|
|
|
| |
Add unit tests for the use of reject with icmp inside netdev family.
reject.t from inet family couldn't be reused because it was using
meta nfproto which is not supported inside netdev.
Signed-off-by: Jose M. Guisado Gomez <guigom@riseup.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
| |
Enables not specifying any icmp type and code when using reject inside
netdev.
This patch completely enables using reject for the netdev family.
Signed-off-by: Jose M. Guisado Gomez <guigom@riseup.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
| |
The example adds a rule to the `output` chain, not the `input` chain.
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
| |
Explicitly eliminate the newgen message from output instead of just the
last line to make sure no other output is dropped by accident. This also
allows the test to pass in unpatched kernels which do not emit the
newgen message despite NLM_F_ECHO if no netlink listeners are present.
Fixes: 46b54fdcf266d ("Revert "monitor: do not print generation ID with --echo"")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
| |
|
|
|
|
|
| |
Using oif in fib from prerouting is not support, make sure -c reports an
error.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
| |
Update release name based on the Fearless Fosdick series: Anyface.
Bump dependencies on libnftnl.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Revert 0e258556f7f3 ("monitor: do not print generation ID with --echo").
There is actually a kernel bug which is preventing from displaying
this generation ID message.
Update the tests/shell to remove the last line of the --echo output
which displays the generation ID once the "netfilter: nftables: fix netlink
report logic in flowtable and genid" kernel fix is applied.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
| |
This fixes testcases/sets/0036add_set_element_expiration_0
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
| |
Turns out json_string() already dups the input, so the temporary dup
passed to it is lost.
Fixes: e70354f53e9f6 ("libnftables: Implement JSON output support")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
| |
reported by tests/monitor# bash run-tests.sh
...
SUMMARY: AddressSanitizer: heap-use-after-free /home/pablo/devel/scm/git-netfilter/nftables/src/expression.c:1385 in expr_ops
Due to incorrect structure layout when calling interval_expr_copy().
Fixes: c1f0476fd590 ("segtree: copy expr data to closing element")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Kernel provides information regarding expression since
83d9dcba06c5 ("netfilter: nf_tables: extended netlink error reporting for
expressions").
A common mistake is to refer a chain which does not exist, e.g.
# nft add rule x y jump test
Error: Could not process rule: No such file or directory
add rule x y jump test
^^^^
Use the existing netlink extended error reporting infrastructure to
provide better error reporting as in the example above.
Requires Linux kernel patch 83d9dcba06c5 ("netfilter: nf_tables:
extended netlink error reporting for expressions").
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
| |
Constify pointer to location object to compile check for unintentional
updates.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
| |
Statically store up to 32 locations per command, if the number of
locations is larger than 32, then skip rather than hit assertion.
Revisit this later to dynamically store location per command using a
hashtable.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
| |
Available since Linux kernel >= 5.10.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When last expr has no closing element we did not propagate
expr properties such as comment or expire date to the newly
allocated set elem.
Before:
nft create table t
nft 'add set t s { type ipv4_addr; flags interval; timeout 60s; }'
nft add element t s { 224.0.0.0/3 }
nft list set t s | grep -o 'elements.*'
elements = { 224.0.0.0-255.255.255.255 }
nft flush set t s
nft add element t s { 224.0.0.0/4, 240.0.0.0/4 }
nft list set t s | grep -o 'elements.*'
elements = { 224.0.0.0/4 expires 55s152ms, 240.0.0.0-255.255.255.255 }
nft delete set t s
nft 'add set t s { type ipv4_addr; flags interval; auto-merge; timeout 60s; }'
nft add element t s { 224.0.0.0/4, 240.0.0.0/4 }
nft list set t s | grep -o 'elements.*'
elements = { 224.0.0.0-255.255.255.255 }
After:
elements = { 224.0.0.0-255.255.255.255 expires 58s515ms }
elements = { 224.0.0.0/4 expires 54s622ms, 240.0.0.0-255.255.255.255 expires 54s622ms }
elements = { 224.0.0.0-255.255.255.255 expires 57s92ms }
Bugzilla: https://bugzilla.netfilter.org/show_bug.cgi?id=1454
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
| |
Stateless SCTP header mangling doesn't work reliably.
This tells the kernel to update the checksum field using
the sctp crc32 algorithm.
Note that this needs additional kernel support to work.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add support for inet ingress chains.
table inet filter {
chain ingress {
type filter hook ingress device "veth0" priority filter; policy accept;
}
chain input {
type filter hook input priority filter; policy accept;
}
chain forward {
type filter hook forward priority filter; policy accept;
}
}
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
| |
Fixes: 3926a3369bb5 ("mergesort: unbreak listing with binops")
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
| |
Fixes: 741a06ac15d2 ("mergesort: find base value expression type via recursion")
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
| |
Fixes: dcec7d57559a ("ct: Add support for the 'id' key")
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
| |
Element counters reside in 'stmt' field as counter statement. Append
them to 'elem' object as additional 'counter' property, generated by
counter_stmt_json().
Signed-off-by: Gopal Yadav <gopunop@gmail.com>
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix for an assertion fail when trying to match against an all-wildcard
interface name:
| % nft add rule t c iifname '"*"'
| nft: expression.c:402: constant_expr_alloc: Assertion `(((len) + (8) - 1) / (8)) > 0' failed.
| zsh: abort nft add rule t c iifname '"*"'
Fix this by detecting the string in expr_evaluate_string() and returning
an error message:
| % nft add rule t c iifname '"*"'
| Error: All-wildcard strings are not supported
| add rule t c iifname "*"
| ^^^
While being at it, drop the 'datalen >= 1' clause from the following
conditional as together with the added check for 'datalen == 0', all
possible other values have been caught already.
|
| |
|
|
|
|
|
| |
this updates the two examples in the man page that use the obsolete `meter` to
use sets. I also fixed a bit of formatting for the conntrack expressions.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch enables the user to specify a comment when adding a chain.
Relies on kernel space supporting userdata for chains.
> nft add table ip filter
> nft add chain ip filter input { comment "test"\; type filter hook input priority 0\; policy accept\; }
> list ruleset
table ip filter {
chain input {
comment "test"
type filter hook input priority filter; policy accept;
}
}
Signed-off-by: Jose M. Guisado Gomez <guigom@riseup.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
| |
--terse with --json is ignored, fix this. This patch also includes a test.
Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=1388
Signed-off-by: Gopal Yadav <gopunop@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch extends the protocol context infrastructure to track multiple
transport protocols when they are specified from sets.
This removes errors like:
"transport protocol mapping is only valid after transport protocol match"
when invoking:
# nft add rule x z meta l4proto { tcp, udp } dnat to 1.1.1.1:80
This patch also catches conflicts like:
# nft add rule x z ip protocol { tcp, udp } tcp dport 20 dnat to 1.1.1.1:80
Error: conflicting protocols specified: udp vs. tcp
add rule x z ip protocol { tcp, udp } tcp dport 20 dnat to 1.1.1.1:80
^^^^^^^^^
and:
# nft add rule x z meta l4proto { tcp, udp } tcp dport 20 dnat to 1.1.1.1:80
Error: conflicting protocols specified: udp vs. tcp
add rule x z meta l4proto { tcp, udp } tcp dport 20 dnat to 1.1.1.1:80
^^^^^^^^^
Note that:
- the singleton protocol context tracker is left in place until the
existing users are updated to use this new multiprotocol tracker.
Moving forward, it would be good to consolidate things around this new
multiprotocol context tracker infrastructure.
- link and network layers are not updated to use this infrastructure
yet. The code that deals with vlan conflicts relies on forcing
protocol context updates to the singleton protocol base.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
| |
If there is protocol context for this base, just return from function
to remove one level of indentation. This patch is cleanup.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
| |
If nft crashes or hits an assertion, the last command run shows in the
/tmp/nftables-test.log file.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
| |
Because the command seqnums are monotonic, repeated traversals
of the cmds list from the beginning are not necessary as long as
the error seqnums are also monotonic.
Signed-off-by: Jindrich Makovicka <makovick@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
| |
Assume each error in the batch will result in a 1k notification for the
non-echo flag set on case as described in 860671662d3f ("mnl: fix --echo
buffer size again").
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before this patch grammar supported specifying multiple comments, and
only the last value would be assigned.
This patch adds a function to test if an attribute is already assigned
and, if so, calls erec_queue with this attribute location.
Use this function in order to check for duplication (or more) of comments
for actions that support it.
> nft add table inet filter { flags "dormant"\; comment "test"\; comment "another"\;}
Error: You can only specify this once. This statement is duplicated.
add table inet filter { flags dormant; comment test; comment another;}
^^^^^^^^^^^^^^^^
Signed-off-by: Jose M. Guisado Gomez <guigom@riseup.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Enables specifying an optional comment when declaring named objects. The
comment is to be specified inside the object's block ({} block)
Relies on libnftnl exporting nftnl_obj_get_data and kernel space support
to store the comments.
For consistency, this patch makes the comment be printed first when
listing objects.
Adds a testcase importing all commented named objects except for secmark,
although it's supported.
Example: Adding a quota with a comment
> add table inet filter
> nft add quota inet filter q { over 1200 bytes \; comment "test_comment"\; }
> list ruleset
table inet filter {
quota q {
comment "test_comment"
over 1200 bytes
}
}
Signed-off-by: Jose M. Guisado Gomez <guigom@riseup.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Sets that store flags might contain a mixture of values and binary
operations. Find the base value type via recursion to compare the
expressions.
Make sure concatenations are listed in a deterministic way via
concat_expr_msort_value() which builds a mpz value with the tuple.
Adjust a few tests after this update since listing differs after this
update.
Fixes: 14ee0a979b62 ("src: sort set elements in netlink_get_setelems()")
Fixes: 3926a3369bb5 ("mergesort: unbreak listing with binops")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
