Phil Sutter [Fri, 17 Nov 2023 16:27:56 +0000 (17:27 +0100)]
nft: Leave interface masks alone when parsing from kernel
The mask is entirely unused by nft-variants in general and legacy ones
when printing. It is relevant only when inserting a legacy rule into
kernel as it doesn't detect the '+'-suffix.
Phil Sutter [Sat, 18 Nov 2023 03:28:25 +0000 (04:28 +0100)]
xshared: Do not populate interface masks per default
These are needed by legacy variants only, so introduce a simplified
xtables_parse_interface() replacement which does not deal with them and
a small function which sets the mask based on given interface name for
use by legacy tools.
Replace the awkward inverse_for_options array with basically a few
switch() statements clearly identifying the relation between option and
inverse values and relieve callers from having to find the option flag
bit's position.
Phil Sutter [Tue, 14 Nov 2023 17:51:59 +0000 (18:51 +0100)]
xshared: Introduce xt_cmd_parse_ops::option_name
The old opt2char() function was flawed: Since not every field in
optflags contains a printable character, typical use of its return value
in print statements could lead to garbage on screen.
Replace this by a mechanism to retrieve an option's long name which
supports family-specific overrides. and get rid of optflags field
altogether and define NUMBER_OF_OPT similar to NUMBER_OF_CMD.
Phil Sutter [Fri, 3 Sep 2021 17:23:43 +0000 (19:23 +0200)]
extensions: MARK: arptables: Use guided option parser
It expects mark values in hex which is possible by setting the base
field.
The only adjustment needed to use the revision 2 parser is to fill the
mask for --set-mark: With XTTYPE_MARKMASK32, an omitted mask sets all
mask bits, XTTYPE_UINT32 leaves it uninitialized, though.
Phil Sutter [Thu, 2 Sep 2021 14:53:26 +0000 (16:53 +0200)]
extensions: libarpt_mangle: Use guided option parser
Sadly not the best conversion, struct arpt_mangle is not ideal for use
as storage backend: With MAC addresses, xtopt_parse_ethermac() refuses
to write into *_devaddr fields as they are larger than expected. With
XTTYPE_HOSTMASK OTOH, XTOPT_PUT is not supported in the first place.
As a side-effect, network names (from /etc/networks) are no longer
accepted. But earlier migrations to guided option parser had this
side-effect as well, so probably not a frequently used feature.
Phil Sutter [Wed, 22 Nov 2023 19:41:34 +0000 (20:41 +0100)]
libxtables: Introduce xtables_strtoul_base()
Semantically identical to xtables_strtoul() but accepts the base as
parameter so callers may force it irrespective of number prefix. The old
xtables_strtoul() becomes a shallow wrapper.
Phil Sutter [Thu, 2 Sep 2021 14:51:15 +0000 (16:51 +0200)]
libxtables: Fix guided option parser for use with arptables
With an unexpected value in afinfo->family, guided option parser was
rather useless when called from arptables extensions. Introduce
afinfo_family() wrapper to sanitize at least NFPROTO_ARP value.
Phil Sutter [Fri, 3 Sep 2021 21:43:41 +0000 (23:43 +0200)]
libxtables: Combine the two extension option mergers
For extending the command parser's struct option array, there is
xtables_merge_options() and xtables_options_xfrm(). Since their bodies
were almost identical, make the latter a wrapper of the former by
transforming the passed struct xt_option_entry array into a temporary
struct option one before handing over.
Phil Sutter [Tue, 21 Nov 2023 22:14:47 +0000 (23:14 +0100)]
ebtables: Implement --change-counters command
Treat it like --replace against the same rule with changed counters.
The operation is obviously not atomic, so rule counters may change in
kernel while the rule is fetched, modified and replaced.
Phil Sutter [Wed, 22 Nov 2023 02:58:18 +0000 (03:58 +0100)]
xshared: do_parse: Ignore '-j CONTINUE'
While iptables does not support his NOP, arptables man page claims it
does (although legacy arptables rejects it) and ebtables prefers to
print it instead of omitting the '-j' option.
Accept and ignore the target when parsing to at least fix for
arptables-nft and prepare for ebtables-nft using do_parse() as well.
Phil Sutter [Sun, 19 Nov 2023 10:23:09 +0000 (11:23 +0100)]
tests: xlate: Print failing command line
If the command segfaults, 'error' variable is empty and the resulting
error message is even misleading as the called program may not have been
iptables-translate.
Phil Sutter [Thu, 16 Nov 2023 16:03:10 +0000 (17:03 +0100)]
xshared: Drop pointless CMD_REPLACE check
All current users set default source and destination addresses in their
post_parse callbacks, so legacy variants are safe and nft variants don't
have this restriction anyway.
Phil Sutter [Sun, 19 Nov 2023 12:18:26 +0000 (13:18 +0100)]
xshared: struct xt_cmd_parse::xlate is unused
Drop the boolean, it was meant to disable some existence checks in
do_parse() prior to the caching rework. Now that do_parse() runs before
any caching is done, the checks in question don't exist anymore so drop
this relict.
Fixes: a7f1e208cdf9c ("nft: split parsing from netlink commands") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Wed, 15 Nov 2023 16:41:50 +0000 (17:41 +0100)]
nft-bridge: nft_bridge_add() uses wrong flags
When checking whether -s or -d was given, invflags were used by
accident. This change has no functional effect since the values remain
the same, but this way it's clear where the previously assigned flags
are used.
Jan Engelhardt [Mon, 13 Nov 2023 10:17:35 +0000 (11:17 +0100)]
man: more backslash-encoding of characters
"-" is the dash, "\-" is minus as we know, but groff lists some more
characters: "^" is "modifier circumflex" and "~" is "modifier tilde",
which, too, need to be escaped for our use.
Phil Sutter [Wed, 8 Nov 2023 03:08:44 +0000 (04:08 +0100)]
arptables: Fix --proto-type mask formatting
Arptables accepts numeric --proto-type values and masks in any numeral
system identified by (absence of) prefix. Yet it prints the mask value
in hex without '0x'-prefix, breaking save and restore the same way
numeric --h-type output did.
In theory, this could be fixed either by adding the missing prefix or
printing the mask in decimal (like most other builtin matches do), but
since the value is printed in hex with prefix already, align mask output
with that.
Also a day 1 bug and consistent with legacy, so no Fixes: tag here as
well.
Signed-off-by: Phil Sutter <phil@nwl.cc> Acked-by: Florian Westphal <fw@strlen.de>
Phil Sutter [Wed, 8 Nov 2023 02:58:42 +0000 (03:58 +0100)]
arptables: Fix formatting of numeric --h-type output
Arptables expects numeric arguments to --h-type option in hexadecimal
form, even if no '0x'-prefix is present. In contrast, it prints such
values in decimal. This is not just inconsistent, but makes it
impossible to save and later restore a ruleset without fixing up the
values in between.
Assuming that the parser side can't be changed for compatibility
reasons, fix the output side instead.
This is a day 1 bug and present in legacy arptables as well, so treat
this as a "feature" of arptables-nft and omit a Fixes: tag.
Signed-off-by: Phil Sutter <phil@nwl.cc> Acked-by: Florian Westphal <fw@strlen.de>
Add test cases for libarpt_mangle and extend the generic
tests to cover basic arptables matches.
Note that there are several historic artefacts that could be revised.
For example, arptables-legacy and arptables-nft both ignore "-p"
instead of returning an error about an unsupported option.
The ptype could be hard-wired to 0x800 and set unconditionally.
OTOH, this should always match for ethernet arp packets anyway.
Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Tue, 7 Nov 2023 18:12:14 +0000 (19:12 +0100)]
ebtables: Fix corner-case noflush restore bug
Report came from firwalld, but this is actually rather hard to trigger.
Since a regular chain line prevents it, typical dump/restore use-cases
are unaffected.
Fixes: 73611d5582e72 ("ebtables-nft: add broute table emulation") Cc: Eric Garver <eric@garver.life> Signed-off-by: Phil Sutter <phil@nwl.cc>
ARPT_ and IPT_INV flags are not interchangeable, e.g.:
define IPT_INV_SRCDEVADDR 0x0080
define ARPT_INV_SRCDEVADDR 0x0010
as these flags can be tested by libarp_foo.so such checks can yield
incorrect results.
Because arptables-nft uses existing code, e.g. xt_mark, it makes
sense to unify this completely by converting the last users of
ARPT_INV_ constants.
Note that arptables-legacy does not do run-time module loading via
dlopen(). Functionaliy implemented by "extensions" in the
arptables-legacy git tree are built-in, so this doesn't break
arptables-legacy binaries.
Fixes: 44457c080590 ("xtables-arp: Don't use ARPT_INV_*") Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Phil Sutter <phil@nwl.cc>
Jan Engelhardt [Tue, 24 Oct 2023 13:04:03 +0000 (15:04 +0200)]
man: encode hyphens the way groff/man requires it
Edit a few spots where indeed a hyphens (U+2010) rather than U+002D is desired.
("set-name" is not something you input, it is a placeholder in the context of
documentation. "out-of-flow" is part of the regular flowed text, so should not
use anything but hyphens.)
Signed-off-by: Jan Engelhardt <jengelh@inai.de> Signed-off-by: Phil Sutter <phil@nwl.cc>
Jan Engelhardt [Tue, 24 Oct 2023 12:58:06 +0000 (14:58 +0200)]
man: encode minushyphen the way groff/man requires it
Sparked by a recent LWN article[1], sweeps over the iptables manpages
for incorrectly encoded dashes was made by Phil Sutter and myself.
An ASCII minushyphen in the source manpage translates to a hyphen in
output, so one has to use the sequence "\-" to get a minushyphen in
the output, as groff_char(7) explains.
[1] https://lwn.net/Articles/947941/ (paywalled until about 2023-11-06)
Signed-off-by: Jan Engelhardt <jengelh@inai.de> Signed-off-by: Phil Sutter <phil@nwl.cc>
Since kernel commit c4eee56e14fe ("net: skb_find_text: Ignore patterns
extending past 'to'"), pattern scanning no longer happens past --to
offset even if skb_seq_read() returned a larger block. Point this out in
the description and also drop the '-1' offset which is not true as
kernel's selftest in tools/testing/selftests/netfilter/xt_string.sh
shows.
Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=1707 Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Thu, 12 Oct 2023 15:54:53 +0000 (17:54 +0200)]
extensions: string: Clarify description of --to
String match indeed returns a match as long as the given pattern starts
in the range of --from and --to, update the text accordingly.
Also add a note regarding fragment boundaries.
Phil Sutter [Thu, 12 Oct 2023 15:27:42 +0000 (17:27 +0200)]
libiptc: Fix for another segfault due to chain index NULL pointer
Chain rename code missed to adjust the num_chains value which is used to
calculate the number of chain index buckets to allocate during an index
rebuild. So with the right number of chains present, the last chain in a
middle bucket being renamed (and ending up in another bucket) triggers
an index rebuild based on false data. The resulting NULL pointer index
bucket then causes a segfault upon reinsertion.
Phil Sutter [Wed, 6 Sep 2023 17:02:52 +0000 (19:02 +0200)]
include: linux: Update kernel.h
Its contents were moved into const.h and sysinfo.h, apply these changes
to the cached copies. Fixes for the following warning when compiling
xtables-monitor.c with new kernel headers in /usr/include:
| In file included from ../include/linux/netfilter/x_tables.h:3,
| from ../include/xtables.h:19,
| from xtables-monitor.c:36:
| ../include/linux/kernel.h:7: warning: "__ALIGN_KERNEL" redefined
| 7 | #define __ALIGN_KERNEL(x, a) __ALIGN_KERNEL_MASK(x, (typeof(x))(a) - 1)
| |
| In file included from /usr/include/linux/netlink.h:5,
| from /home/n0-1/git/libmnl/install/include/libmnl/libmnl.h:9,
| from xtables-monitor.c:30:
| /usr/include/linux/const.h:31: note: this is the location of the previous definition
| 31 | #define __ALIGN_KERNEL(x, a) __ALIGN_KERNEL_MASK(x, (__typeof__(x))(a) - 1)
| |
Phil Sutter [Wed, 6 Sep 2023 14:32:47 +0000 (16:32 +0200)]
nft: Fix for useless meta expressions in rule
A relict of legacy iptables' mandatory matching on interfaces and IP
addresses is support for the '-i +' notation, basically a "match any
input interface". Trying to make things better than its predecessor,
iptables-nft boldly optimizes that nop away - not entirely though, the
meta expression loading the interface name was left in place. While not
a problem (apart from pointless overhead) in current HEAD, v1.8.7 would
trip over this as a following cmp expression (for another match) was
incorrectly linked to that stale meta expression, loading strange values
into the respective interface name field.
While being at it, merge and generalize the functions into a common one
for use with ebtables' NFT_META_BRI_(I|O)IFNAME matches, too.
Phil Sutter [Fri, 1 Sep 2023 11:16:56 +0000 (13:16 +0200)]
tests: shell: Fix for ineffective 0007-mid-restore-flush_0
The test did not catch non-zero exit status of the spawned coprocess. To
make it happen, Drop the line killing it (it will exit anyway) and pass
its PID to 'wait'.
While being at it, put the sleep into the correct spot (otherwise the
check for chain 'foo' existence fails as it runs too early) and make
said chain existence check effective.
Fixes: 4e3c11a6f5a94 ("nft: Fix for ruleset flush while restoring") Signed-off-by: Phil Sutter <phil@nwl.cc>
Quentin Armitage [Sat, 23 Nov 2013 08:41:58 +0000 (08:41 +0000)]
extensions: Fix checking of conntrack --ctproto 0
There are three issues in the code:
1) the check (sinfo->invflags & XT_INV_PROTO) is using the wrong mask
2) in conntrack_mt_parse it is testing (info->invert_flags &
XT_INV_PROTO) before the invert bit has been set.
3) the sense of the error message is the wrong way round
1) To get the error, ! -ctstatus XXX has to be specified, since
XT_INV_PROTO == XT_CONNTRACK_STATUS e.g.
| iptables -I CHAIN -m conntrack ! --ctstatus ASSURED --ctproto 0 ...
3) Unlike --proto 0 (where 0 means all protocols), in the conntrack
match --ctproto 0 appears to mean protocol 0, which can never be.
Therefore --ctproto 0 could never match and ! --ctproto 0 will always
match. Both of these should be rejected, since the user clearly
cannot be intending what was specified.
The attached patch resolves the issue, and also produces an error
message if --ctproto 0 is specified (as well as ! --ctproto 0 ), since
--ctproto 0 will never match, and ! --ctproto 0 will always match.
[Phil: - Added Fixes: tag - it's a day 1 bug
- Copied patch description from Bugzilla
- Reorganized changes to reduce diff
- Added test cases]
Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=874 Fixes: 5054e85be3068 ("general conntrack match module userspace support files") Signed-off-by: Quentin Armitage <quentin@armitage.org.uk> Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Tue, 15 Aug 2023 11:47:28 +0000 (13:47 +0200)]
Revert --compat option related commits
This reverts the following commits:
b14c971db6db0 ("tests: Test compat mode") 11c464ed015b5 ("Add --compat option to *tables-nft and *-nft-restore commands") ca709b5784c98 ("nft: Introduce and use bool nft_handle::compat") 402b9b3c07c81 ("nft: Pass nft_handle to add_{target,action}()")
This implementation of a compatibility mode implements rules using
xtables extensions if possible and thus relies upon existence of those
in kernel space. Assuming no viable replacement for the internal
mechanics of this mode will be found in foreseeable future, it will
effectively block attempts at deprecating and removing of these xtables
extensions in favor of nftables expressions and thus hinder upstream's
future plans for iptables.
Phil Sutter [Fri, 5 May 2023 18:04:41 +0000 (20:04 +0200)]
Add --compat option to *tables-nft and *-nft-restore commands
The flag sets nft_handle::compat boolean, indicating a compatible rule
implementation is wanted. Users expecting their created rules to be
fetched from kernel by an older version of *tables-nft may use this to
avoid potential compatibility issues.
Changes since v1:
- Expect short option '-C' in {ip,ip6,eb}tables-nft-restore command line
parser
- Support -C/--compat in arptables-nft-restore, too
- Update man pages with the new flag
Phil Sutter [Fri, 5 May 2023 15:39:08 +0000 (17:39 +0200)]
nft: Introduce and use bool nft_handle::compat
If set, create rules using compat expressions where possible and disable
the bitwise expression avoidance introduced in 323259001d617 ("nft:
Optimize class-based IP prefix matches").
The change can't be right: A simple rule append call will reset all
built-in chains' counters. The old code works fine even given the
mentioned "empty restore" use-case, at least if counters don't change on
the fly in-kernel.
Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=912 Fixes: 7c4d668c9c2ee ("libiptc: fix wrong maptype of base chain counters on restore") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Thu, 10 Aug 2023 09:30:59 +0000 (11:30 +0200)]
nft: Create builtin chains with counters enabled
The kernel enables policy counters for nftables chains only if
NFTA_CHAIN_COUNTERS attribute is present. For this to be generated, one
has to set NFTNL_CHAIN_PACKETS and NFTNL_CHAIN_BYTES attributes in the
allocated nftnl_chain object.
The above happened for base chains only with iptables-nft-restore if
called with --counters flag. Since this is very unintuitive to users,
fix the situation by adding counters to base chains in any case.
Fixes: 384958620abab ("use nf_tables and nf_tables compatibility interface") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Tue, 31 Jan 2023 18:55:57 +0000 (19:55 +0100)]
nft-ruleparse: Introduce nft_create_target()
Like nft_create_match(), this is a small wrapper around the typical
target extension lookup and (standard) init code.
To use it from nft_parse_target() and nft_parse_log(), introduce an
inner variant which accepts the target payload size as parameter.
The call to rule_parse_ops::target callback was problematic with
standard target, because the callbacks initialized
iptables_command_state::jumpto with the target name, "standard" in that
case. Perform its tasks in nft_create_target(), keep it only for bridge
family's special handling of watcher "targets".
Jan Palus [Mon, 28 Dec 2020 09:59:42 +0000 (10:59 +0100)]
nft: move processing logic out of asserts
[Phil: Introduce assert_nft_restart() to keep things clean, also add
fallback returns to nft_action() and nft_prepare(), sanitizing
things at least a bit.]
Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=1487 Signed-off-by: Jan Palus <atler@pld-linux.org> Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Tue, 1 Aug 2023 23:55:08 +0000 (01:55 +0200)]
man: iptables-save.8: Clarify 'available tables'
This appears to be confusing. Since a missing table is also not flushed
("restored") when feeding the dump into iptables-restore, such a restore
call may be considered incomplete.
Phil Sutter [Tue, 1 Aug 2023 22:28:03 +0000 (00:28 +0200)]
man: iptables.8: Clarify --goto description
Text speaks about behaviour of RETURN target when used in chains
redirected to using --goto instead of --jump, not the difference between
--jump option and "return".
Fixes: 17fc163babc34 ("add 'goto' support (Henrik Nordstrom <hno@marasystems.com>)") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Tue, 1 Aug 2023 22:05:45 +0000 (00:05 +0200)]
man: iptables.8: Trivial spelling fixes
- Missing "and" as well as full stop
- Missing comma in enumeration
- Duplicate "previous"
- Confusions are avoided rather than simplified
- Missing space after comma
Phil Sutter [Tue, 1 Aug 2023 19:24:15 +0000 (21:24 +0200)]
man: iptables.8: Extend exit code description
Codes 3 and 4 were missing.
Reported-by: Steven Barre <steven.barre@dxcas.com> Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=1353 Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Tue, 1 Aug 2023 21:28:20 +0000 (23:28 +0200)]
extensions: libipt_icmp: Fix confusion between 255/255 and any
Per definition, ICMP type "any" is type 255 and the full range of codes
(0-255). Save callback though ignored the actual code values, printing
"any" for every type 255 match. This at least confuses users as they
can't find their rule added as '--icmp-type 255/255' anymore.
It is not entirely clear what the fixed commit was trying to establish,
but the save output is certainly not correct (especially since print
callback gets things right).
Phil Sutter [Tue, 1 Aug 2023 14:56:42 +0000 (16:56 +0200)]
iptables-apply: Eliminate shellcheck warnings
Actual warnings were only about use of '-a' in bracket expressions
(replace by '&&' pipeline) and the immediate evaluation of the variable
in trap command.
The remaining changes silence info-level messages: missing quoting
around variables, pointless '$' in arithmetic expressions, backticks
instead of $(...), missing '-r' parameter when calling read and an
awkward negated '-z' check.
Phil Sutter [Thu, 13 Jul 2023 16:32:02 +0000 (18:32 +0200)]
iptables-restore: Drop dead code
Handle initialization is guarded by 'in_table' boolean, so there can't
be a handle already (because the branch which unsets 'in_table' also
frees the handle).
Phil Sutter [Fri, 28 Jul 2023 11:58:46 +0000 (13:58 +0200)]
tests: shell: Fix and extend chain rename test
The old version exited unintentionally before testing ip6tables. Replace
it by a more complete variant testing for all tools, creating and
renaming of,chains with various illegal names instead of just renaming
to a clashing name.
Fixes: ed9cfe1b48526 ("tests: add initial save/restore test cases") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Fri, 28 Jul 2023 11:50:11 +0000 (13:50 +0200)]
ebtables: Improve invalid chain name detection
Fix several issues:
- Most importantly, --new-chain command accepted any name. Introduce
ebt_assert_valid_chain_name() for use with both --new-chain and
--rename-chain.
- Restrict maximum name length to what legacy ebtables allows - this is
a bit more than iptables-nft, subject to be unified.
- Like iptables, legacy ebtables rejects names prefixed by '-' or '!'.
- Use xs_has_arg() for consistency, keep the check for extra args for
now.
Fixes: da871de2a6efb ("nft: bootstrap ebtables-compat") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Fri, 21 Jul 2023 18:14:09 +0000 (20:14 +0200)]
*tables: Reject invalid chain names when renaming
While given chain name was sanity checked with --new-chain command,
--rename-chain command allowed to choose an invalid name. Keep things
consistent by adding the missing check.
Fixes: e6869a8f59d77 ("reorganized tree after kernel merge") Signed-off-by: Phil Sutter <phil@nwl.cc>
Phil Sutter [Fri, 21 Jul 2023 17:40:30 +0000 (19:40 +0200)]
*tables-restore: Enforce correct counters syntax if present
If '--counters' option was not given, restore parsers would ignore
anything following the policy word. Make them more strict, rejecting
anything in that spot which does not look like counter values even if
not restoring counters.
projects / thirdparty / iptables.git / log
