netlink: dynset: set compound expr dtype based on set key definition
"nft add rule ... add @t { ip saddr . 22 ..." will be listed as
'ip saddr . 0x16 [ invalid type]".
This is a display bug, the compound expression created during netlink
deserialization lacks correct datatypes for the value expression.
Avoid this by setting the individual expressions' datatype.
The set key has the needed information, so walk over the types and set
them in the dynset statment.
Also add a test case.
Reported-by: Paulo Ricardo Bruck <paulobruck1@gmail.com> Signed-off-by: Florian Westphal <fw@strlen.de>
payload: don't adjust offsets of autogenerated dependency expressions
Pablo says:
user reports that this is broken:
nft --debug=netlink add rule bridge filter forward vlan id 100 vlan id set 200
[..]
[ payload load 2b @ link header + 14 => reg 1 ]
[..]
[ payload load 2b @ link header + 28 => reg 1 ]
[ bitwise reg 1 = ( reg 1 & 0x000000f0 ) ^ 0x0000c800 ]
[ payload write reg 1 => 2b @ link header + 14 csum_type 0 csum_off 0 csum_flags 0x0 ]
offset says 28, it is assuming q-in-q, in this case it is mangling the
existing header.
The problem here is that 'vlan id set 200' needs a read-modify-write
cycle because 'vlan id set' has to preserve bits located in the same byte area
as the vlan id.
The first 'payload load' at offset 14 is generated via 'vlan id 100',
this part is ok.
The second 'payload load' at offset 28 is the bogus one.
Its added as a dependency, but then adjusted because nft evaluation
considers this identical to 'vlan id 1 vlan id '2, where nft assumes
q-in-q.
To fix this, skip offset adjustments for raw expressions and mark the
dependency-generated payload instruction as such.
This is fine because raw payload operations assume that user specifies
base/offset/length manually.
Also add a test case for this.
Reported-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Florian Westphal <fw@strlen.de>
Since 309785674b25 ("datatype: time_print() ignores -T"), time_type
honors -T option. Given tests/py run in numeric format, this patch
fixes a warning since the ct expiration is now expressed in seconds.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Partially revert 913979f882d1 ("src: add expression handler hashtable")
which is causing a crash with two instances of the nftables handler.
$ sudo python
[sudo] password for echerkashin:
Python 3.9.7 (default, Sep 3 2021, 06:18:44)
[GCC 11.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from nftables import Nftables
>>> n1=Nftables()
>>> n2=Nftables()
>>> <Ctrl-D>
double free or corruption (top)
Aborted
src: Check range bounds before converting to prefix
The lower bound must be the first value of the prefix to be coverted.
For example, range "10.0.0.15-10.0.0.240" can not be converted to
"10.0.0.15/24". Validate it by checking if the lower bound value has
enough trailing zeros.
Signed-off-by: Xiao Liang <shaw.leon@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
# nft list ruleset
table ip x {
set y {
type ipv4_addr
flags timeout
elements = { 1.1.1.1 timeout 5m expires 1m49s40ms }
}
}
# sudo nft -T list ruleset
table ip x {
set y {
type ipv4_addr
flags timeout
elements = { 1.1.1.1 timeout 300s expires 108s }
}
}
Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=1561 Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netlink_delinearize: incorrect meta protocol dependency kill again
This patch adds __meta_dependency_may_kill() to consolidate inspection
of the meta protocol, nfproto and ether type expression to validate
dependency removal on listings.
Phil reports that 567ea4774e13 includes an update on the ip and ip6
families that is not described in the patch, moreover, it flips the
default verdict from true to false.
Fixes: 567ea4774e13 ("netlink_delinearize: incorrect meta protocol dependency kill") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Phil Sutter [Wed, 1 Sep 2021 14:41:44 +0000 (16:41 +0200)]
parser_json: Fix error reporting for invalid syntax
Errors emitted by the JSON parser caused BUG() in erec_print() due to
input descriptor values being bogus.
Due to lack of 'include' support, JSON parser uses a single input
descriptor only and it lived inside the json_ctx object on stack of
nft_parse_json_*() functions.
By the time errors are printed though, that scope is not valid anymore.
Move the static input descriptor object to avoid this.
Fixes: 586ad210368b7 ("libnftables: Implement JSON parser") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Wed, 11 Aug 2021 16:14:06 +0000 (18:14 +0200)]
tests: json_echo: Print errors to stderr
Apart from the obvious, this fixes exit_dump() which tried to dump the
wrong variable ('out' instead of 'obj') and missed that json.dumps()
doesn't print but just returns a string. Make it call exit_err() to
share some code, which changes the prefix from 'FAIL' to 'Error' as a
side-effect.
While being at it, fix for a syntax warning with newer Python in
unrelated code.
Fixes: bb32d8db9a125 ("JSON: Add support for echo option") Signed-off-by: Phil Sutter <phil@nwl.cc>
Print queue statement using the 'queue ... to' syntax to consolidate the
syntax around Florian's proposal introduced in 6cf0f2c17bfb ("src:
queue: allow use of arbitrary queue expressions").
Retain backward compatibility, 'queue num' syntax is still allowed.
Update and add new tests.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Fri, 20 Aug 2021 09:52:35 +0000 (11:52 +0200)]
parser: permit symbolic define for 'queue num' again
WHen I simplified the parser to restrict 'queue num' I forgot that
instead of range and immediate value its also allowed to pass in
a variable expression, e.g.
Update this command to display the hook datapath for a packet depending
on its family.
This patch also includes:
- Group of existing hooks based on the hook location.
- Order hooks by priority, from INT_MIN to INT_MAX.
- Do not add sign to priority zero.
- Refresh include/linux/netfilter/nfnetlink_hook.h cache copy.
- Use NFNLA_CHAIN_* attributes to print the chain family, table and name.
If NFNLA_CHAIN_* attributes are not available, display the hookfn name.
- Update syntax: remove optional hook parameter, promote the 'device'
argument.
The following example shows the hook datapath for IPv4 packets coming in
from netdevice 'eth0':
# nft list hooks ip device eth0
family ip {
hook ingress {
+0000000010 chain netdev x y [nf_tables]
+0000000300 chain inet m w [nf_tables]
}
hook input {
-0000000100 chain ip a b [nf_tables]
+0000000300 chain inet m z [nf_tables]
}
hook forward {
-0000000225 selinux_ipv4_forward 0000000000 chain ip a c [nf_tables]
}
hook output {
-0000000225 selinux_ipv4_output
}
hook postrouting {
+0000000225 selinux_ipv4_postroute
}
}
Note that the listing above includes the existing netdev and inet
hooks/chains which *might* interfer in the travel of an incoming IPv4
packet. This allows users to debug the pipeline, basically, to
understand in what order the hooks/chains are evaluated for the IPv4
packets.
If the netdevice is not specified, then the ingress hooks are not
shown.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The negation was introduced to provide a simple shortcut. Extend e6c32b2fa0b8 ("src: add negation match on singleton bitmask value") to
disallow negation with binary operations too.
evaluate: error reporting for missing statements in set/map declaration
Assuming this map:
map y {
type ipv4_addr : verdict
}
This patch slightly improves error reporting to refer to the missing
'counter' statement in the map declaration.
# nft 'add element x y { 1.2.3.4 counter packets 1 bytes 1 : accept, * counter : drop }'
Error: missing statement in map declaration
add element x y { 1.2.3.4 counter packets 10 bytes 640 : accept, * counter : drop }
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The kernel already assumes that that ICMP type to reject a packet is
destination-unreachable, hence the user specifies the *ICMP code*.
Simplify the syntax to:
... reject with icmp port-unreachable
this removes the 'type' keyword before the ICMP code to reject the
packet with.
IIRC, the original intention is to leave room for future extensions that
allow to specify both the ICMP type and the ICMP code, this is however
not possible with the current inconsistent syntax.
Phil Sutter [Mon, 26 Jul 2021 13:27:32 +0000 (15:27 +0200)]
tests: shell: Fix bogus testsuite failure with 100Hz
On kernels with CONFIG_HZ=100, clock granularity does not allow tracking
timeouts in single digit ms range. Change sets/0031set_timeout_size_0 to
not expose this detail.
Signed-off-by: Phil Sutter <phil@nwl.cc> Acked-by: Florian Westphal <fw@strlen.de>
parser_bison: missing initialization of ct timeout policy list
rule.c:1715:3: runtime error: member access within null pointer of type 'struct timeout_state'
AddressSanitizer:DEADLYSIGNAL
=================================================================
==29500==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x7f5bfd43c2a4 bp 0x7ffcb82f13b0 sp 0x7ffcb82f1360 T0)
==29500==The signal is caused by a READ memory access.
==29500==Hint: address points to the zero page.
#0 0x7f5bfd43c2a3 in obj_free /home/test/nftables/src/rule.c:1715
#1 0x7f5bfd43875d in cmd_free /home/test/nftables/src/rule.c:1447
#2 0x7f5bfd58e6f2 in nft_run_cmd_from_filename /home/test/nftables/src/libnftables.c:628
#3 0x5645c48762b1 in main /home/test/nftables/src/main.c:512
#4 0x7f5bfc0eb09a in __libc_start_main ../csu/libc-start.c:308
#5 0x5645c4873459 in _start (/home/test/nftables/src/.libs/nft+0x9459)
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /home/test/nftables/src/rule.c:1715 in obj_free
==29500==ABORTING
Fixes: 7a0e26723496 ("rule: memleak of list of timeout policies") Signed-off-by: Pablo Neira Ayuso <test@netfilter.org>
IP address (min), source port (min), IP address (max), source port (max)
to be obtained from the map. This representation simplifies the
delinearize path, since the datatype is specified as:
ipv4_addr . inet_service.
A few more notes on this update:
- alloc_nftnl_setelem() needs a variant netlink_gen_data() to deal with
the representation of the range on the rhs of the mapping. In contrast
to interval concatenation in the key side, where the range is expressed
as two netlink attributes, the data side of the set element mapping
stores the interval concatenation in a contiguos memory area, see
__netlink_gen_concat_expand() for reference.
- add range_expr_postprocess() to postprocess the data mapping range.
If either one single IP address or port is used, then the minimum and
maximum value in the range is the same value, e.g. to avoid listing
80-80, this round simplify the range. This also invokes the range
to prefix conversion routine.
- add concat_elem_expr() helper function to consolidate code to build
the concatenation expression on the rhs element data side.
This patch also adds tests/py and tests/shell.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
src: infer NAT mapping with concatenation from set
If the map is anonymous, infer it from the set elements. Otherwise, the
set definition already have an explicit concatenation definition in the
data side of the mapping.
This update simplifies the NAT mapping syntax with concatenations, e.g.
snat ip to ip saddr map { 10.141.11.4 : 192.168.2.3 . 80 }
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Wed, 30 Jun 2021 15:45:18 +0000 (17:45 +0200)]
netlink_delinarize: don't check for set element if set is not populated
0065_icmp_postprocessing: line 13: Segmentation fault $NFT insert rule ip x foo index 1 accept
Since no listing is done, cache isn't populated and 'nft insert' will trip over
set->init == NULL during postprocessing of the existing 'icmp id 42' expression.
Fixes: 9a5574e2d4e9 ("netlink_delinearize: add missing icmp id/sequence support") Reported-by: Eric Garver <eric@garver.life> Reported-by: Phil Sutter <phil@nwl.cc> Signed-off-by: Florian Westphal <fw@strlen.de>
If the command refers to an inexisting table, then use the table location.
ruleset.nft:3:12-12: Error: No such file or directory; did you mean table ‘filter’ in family ip?
add rule x x ip saddr @x log prefix "Anti SSH-Bruteforce: " drop
^
before this patch location is not correct:
ruleset.nft:3:12-12: Error: No such file or directory; did you mean table ‘filter’ in family ip?
add rule x x ip saddr @x log prefix "Anti SSH-Bruteforce: " drop
^
Fixes: 0276c2fee939 ("cmd: check for table mismatch first in error reporting") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Direct leak of 13 byte(s) in 1 object(s) allocated from:
#0 0x7fb49ad79810 in strdup (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x3a810)
#1 0x7fb496b8f63a in xstrdup /home/pablo/nftables/src/utils.c:85
#2 0x7fb496c9a79d in nft_lex /home/pablo/nftables/src/scanner.l:740
[...]
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Release dynamically allocated string by lex from the YYERROR path, e.g.
# cat test.nft
table x {
map test {
type ipv4_addr . foo . inet_service : ipv4_addr . inet_service
}
}
# nft -f test.nft
test.nft:3:20-22: Error: unknown datatype foo
type ipv4_addr . foo . inet_service : ipv4_addr . inet_service
^^^
test.nft:6-9: Error: set definition does not specify key
map test {
^^^^
==29692==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 5 byte(s) in 1 object(s) allocated from:
#0 0x7f6c869e8810 in strdup (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x3a810)
#1 0x7f6c8637f63a in xstrdup /home/test/nftables/src/utils.c:85
#2 0x7f6c8648a4d3 in nft_lex /home/test/nftables/src/scanner.l:740
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
libnftables: fix memleak when first message in batch is used to report error
The err->seqnum == batch_seqnum case results in a memleak of mnl_err
objects under some scenarios such as nf_tables kernel support is not
available or user runs the nft executable as non-root.
Fixes: f930cc500318 ("nftables: fix supression of "permission denied" errors") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
evaluate: fix maps with key and data concatenations
expr_evaluate_concat() is overloaded, it deals with two cases:
#1 set key and data definitions, this case uses the special
dynamically created concatenation datatype which is taken
from the context.
#2 set elements, this case iterates over the set key and data
expressions that are components of the concatenation tuple,
to fetch the corresponding datatype.
Add a new function to deal with case #1 specifically.
This patch is implicitly fixing up map that include arbitrary
concatenations. This is failing with a spurious error report such as:
# cat bug.nft
table x {
map test {
type ipv4_addr . inet_proto . inet_service : ipv4_addr . inet_service
}
}
# nft -f bug.nft
bug.nft:3:48-71: Error: datatype mismatch, expected concatenation of (IPv4 address, Internet protocol, internet network service), expression has type concatenation of (IPv4 address, internet network service)
type ipv4_addr . inet_proto . inet_service : ipv4_addr . inet_service
^^^^^^^^^^^^^^^^^^^^^^^^
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Tue, 15 Jun 2021 12:57:08 +0000 (14:57 +0200)]
src: queue: allow use of arbitrary queue expressions
back in 2016 Liping Zhang added support to kernel and libnftnl to
specify a source register containing the queue number to use.
This was never added to nft itself, so allow this.
On linearization side, check if attached expression is a range.
If its not, allocate a new register and set NFTNL_EXPR_QUEUE_SREG_QNUM
attribute after generating the lowlevel expressions for the kernel.
On delinarization we need to check for presence of
NFTNL_EXPR_QUEUE_SREG_QNUM and decode the expression(s) when present.
Also need to do postprocessing for STMT_QUEUE so that the protocol
context is set correctly, without this only raw payload expressions
will be shown (@nh,32,...) instead of 'ip ...'.
Florian Westphal [Tue, 15 Jun 2021 22:43:46 +0000 (00:43 +0200)]
parser: restrict queue num expressiveness
Else we run into trouble once we allow
queue num symhash mod 4 and 1
and so on. Example problem:
queue num jhash ip saddr mod 4 and 1 bypass
This will fail to parse because the scanner is in the wrong state
(ip, not queue), so 'bypass' is parsed as a string.
Currently, while nft will eat the above just fine (minus 'bypass'),
nft rejects this from the evaluation phase with
Error: queue number is not constant
So seems we are lucky and can restrict the supported expressions
to integer and range.
Furthermore, the line looks wrong because this statement:
queue num jhash ip saddr mod 4 and 1 bypass
doesn't specifiy a number, "queue num 4" does, or "queue num 1-2" do.
For arbitrary expr support it seems sensible to enforce stricter
ordering to avoid any problems with the flags, for example:
queue bypass,futurekeyword to jhash ip saddr mod 42
Release list of ct timeout policy when object is freed.
Direct leak of 160 byte(s) in 2 object(s) allocated from:
#0 0x7fc0273ad330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
#1 0x7fc0231377c4 in xmalloc /home/.../devel/nftables/src/utils.c:36
#2 0x7fc023137983 in xzalloc /home/.../devel/nftables/src/utils.c:75
#3 0x7fc0231f64d6 in nft_parse /home/.../devel/nftables/src/parser_bison.y:4448
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netlink_delinearize: memleak when listing ct event rule
listing a ruleset containing:
ct event set new,related,destroy,label
results in memleak:
Direct leak of 3672 byte(s) in 27 object(s) allocated from:
#0 0x7fa5465c0330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
#1 0x7fa54233772c in xmalloc /home/.../devel/nftables/src/utils.c:36
#2 0x7fa5423378eb in xzalloc /home/.../devel/nftables/src/utils.c:75
#3 0x7fa5422488c6 in expr_alloc /home/.../devel/nftables/src/expression.c:45
#4 0x7fa54224fb91 in binop_expr_alloc /home/.../devel/nftables/src/expression.c:698
#5 0x7fa54224ddf8 in bitmask_expr_to_binops /home/.../devel/nftables/src/expression.c:512
#6 0x7fa5423102ca in expr_postprocess /home/.../devel/nftables/src/netlink_delinearize.c:2448
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netlink_delinearize: memleak in string netlink postprocessing
Listing a matching wilcard string results in a memleak: ifname "dummy*"
Direct leak of 136 byte(s) in 1 object(s) allocated from:
#0 0x7f27ba52e330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
#1 0x7f27b9e1d434 in xmalloc /home/.../devel/nftables/src/utils.c:36
#2 0x7f27b9e1d5f3 in xzalloc /home/.../devel/nftables/src/utils.c:75
#3 0x7f27b9d2e8c6 in expr_alloc /home/.../devel/nftables/src/expression.c:45
#4 0x7f27b9d326e9 in constant_expr_alloc /home/.../devel/nftables/src/expression.c:419
#5 0x7f27b9db9318 in netlink_alloc_value /home/.../devel/nftables/src/netlink.c:390
#6 0x7f27b9de0433 in netlink_parse_cmp /home/.../devel/nftables/src/netlink_delinearize.c:321
#7 0x7f27b9deb025 in netlink_parse_expr /home/.../devel/nftables/src/netlink_delinearize.c:1764
#8 0x7f27b9deb0de in netlink_parse_rule_expr /home/.../devel/nftables/src/netlink_delinearize.c:1776
#9 0x7f27b860af7b in nftnl_expr_foreach /home/.../devel/libnftnl/src/rule.c:690
Direct leak of 8 byte(s) in 1 object(s) allocated from:
#0 0x7f27ba52e330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
#1 0x7f27b9e1d434 in xmalloc /home/.../devel/nftables/src/utils.c:36
#2 0x7f27b96975c5 in __gmpz_init2 (/usr/lib/x86_64-linux-gnu/libgmp.so.10+0x1c5c5)
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
evaluate: memleak in binary operation transfer to RHS
Remove useless reference count grabbing on constant expression that
results in a memleak.
Direct leak of 136 byte(s) in 1 object(s) allocated from:
#0 0x7f4cd54af330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
#1 0x7f4cd4d9e489 in xmalloc /home/.../devel/nftables/src/utils.c:36
#2 0x7f4cd4d9e648 in xzalloc /home/.../devel/nftables/src/utils.c:75
#3 0x7f4cd4caf8c6 in expr_alloc /home/.../devel/nftables/src/expression.c:45
#4 0x7f4cd4cb36e9 in constant_expr_alloc /home/.../devel/nftables/src/expression.c:419
#5 0x7f4cd4ca714c in integer_type_parse /home/.../devel/nftables/src/datatype.c:397
#6 0x7f4cd4ca4bee in symbolic_constant_parse /home/.../devel/nftables/src/datatype.c:165
#7 0x7f4cd4ca4572 in symbol_parse /home/.../devel/nftables/src/datatype.c:135
#8 0x7f4cd4cc333f in expr_evaluate_symbol /home/.../devel/nftables/src/evaluate.c:251
[...]
Indirect leak of 8 byte(s) in 1 object(s) allocated from:
#0 0x7f4cd54af330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
#1 0x7f4cd4d9e489 in xmalloc /home/.../devel/nftables/src/utils.c:36
#2 0x7f4cd46185c5 in __gmpz_init2 (/usr/lib/x86_64-linux-gnu/libgmp.so.10+0x1c5c5)
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Set element keys are of EXPR_SET_ELEM expression type, however, mappings
use the EXPR_MAPPING expression to wrap the EXPR_SET_ELEM key
(mapping->left) and the corresponding data (mapping->right).
This patch adds a wrapper function to fetch the EXPR_SET_ELEM expression
from the key in case of mappings and use it.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Tue, 15 Jun 2021 16:01:51 +0000 (18:01 +0200)]
tests: add a icmp-reply only and icmpv6 id test cases
Check that nft doesn't remove the dependency in these cases:
icmp type echo-reply icmp id 1
("icmp id" matches both echo request and reply).
Add icmpv6 test cases. These fail without the previous patches:
add rule ip6 test-ip6 input icmpv6 id 1:
'icmpv6 id 1' mismatches
'icmpv6 type { echo-request, echo-reply} icmpv6 parameter-problem 65536/16'
add rule ip6 test-ip6 input icmpv6 type echo-reply icmpv6 id 65534':
'icmpv6 type echo-reply icmpv6 id 65534' mismatches
'icmpv6 type echo-reply @th,32,16 65534'
tests: shell: cover split chain reference across tables
Add a test to cover table T1 containing the definition of chain C1, and
table T1' (actually the same definition as T1) that contains a (jump)
reference to chain C1.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
projects / thirdparty / nftables.git / log
