| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
Fixes: 0a8635183edd0 ("xtables-compat: ignore '+' interface name")
Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=1702
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
| |
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").
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
| |
Dump extension data if it differs.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
| |
When other extensions may have "garbage" appended to their data which
should not be considered for match comparison, among match is the
opposite in that it extends its data beyond the value in 'size' field.
Add special casing to cover for this, avoiding false-positive rule
comparison.
Fixes: 26753888720d8 ("nft: bridge: Rudimental among extension support")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
| |
Extract all code dealing with parsing from struct nftnl_rule into struct
iptables_command_state from nft-shared.c into a separate source file.
Basically this is nft_rule_to_iptables_command_state() and the functions
it calls, plus family-independent parsers called from family-specific
callbacks.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
Code does not refer to struct xt_comment_info anymore.
Fixes: 3bb497c61d743 ("xtables: Fix for deleting rules with comment")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use new 'meta broute set 1' to emulate -t broute. If '-t broute' is given,
automatically translate -j DROP to 'meta broute set 1 accept' internally.
Reverse translation zaps the broute and pretends verdict was DROP.
Note that BROUTING is internally handled via PREROUTING, i.e. 'redirect'
and 'nat' targets are not available, they will need to be emulated via
nft expressions.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
| |
Perform the nft_family_ops::parse_match call from inside
nft_create_match(). It frees callers from having to access the match
itself.
Then return a pointer to match data instead of the match itself.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
By dropping the per-family 'cs->matches' selection (which is the default
anyway), code becomes identical to the function's body.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Some matches may turn into multiple nft statements (naturally or via
translation). Such statements must parse into a single extension again
in order to rebuild the rule as it was.
Introduce nft_find_match_in_cs() to iterate through the lists and drop
tcp/udp port match caching in struct nft_xt_ctx which is not needed
anymore.
Note: Match reuse is not enabled unconditionally for all matches,
because iptables supports having multiple instances of the same
extension.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
Finish parsing the rule, thereby printing all potential problems and
abort the program.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
| |
Catch more unexpected conditions.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
| |
These were previously ignored.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
| |
It is just a hack to translate '! -i +' into a never matching nft rule,
but recognize it anyway for completeness' sake and to make xlate replay
test pass.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's actually nonsense since it will never match, but iptables accepts
it and the resulting nftables rule must behave identically. Reuse the
solution implemented into xtables-translate (by commit e179e87a1179e)
and turn the above match into 'iifname INVAL/D'.
The commit this fixes merely ignored the fact that "any interface" match
might be inverted.
Fixes: 0a8635183edd0 ("xtables-compat: ignore '+' interface name")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
| |
Since nft adds the interface name as fixed-size string of 16 bytes,
filling a mask based on the length value will not match the mask nft
set.
Fixes: 652b98e793711 ("xtables-compat: fix wildcard detection")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
| |
This is nft_clear_iptables_command_state() but in a location reachable
by legacy iptables, too.
Changes callers in non-family-specific code to use clear_cs callback
instead of directly calling it - ebtables still has a custom variant.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
| |
We always return 0, even if we printed some error message half-way.
Increment an error counter whenever an error message was printed so that
the chain-loop can exit with an error if this counter is nonzero.
Another effect is that iptables-save will no longer print the COMMIT line anmore.
Reported-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
Replace nftnl_expr_get_data() alias by real function call.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
The same algorithm was present four times, outsource it. Also use
max()/min() macros for a more readable boundary notation.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The function nft_meta_set_to_target() would always bail since nothing
sets 'sreg->meta_sreg.set' to true. This is obvious, as the immediate
expression "filling" the source register does not indicate its purpose.
The whole source register purpose storing in meta_sreg seems to be
pointless, so drop it altogether.
Fixes: f315af1cf8871 ("nft: track each register individually")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixed commit broke among match in two ways:
1) The two lookup sizes are 12 and 6, not 12 and 4 - among supports
either ether+IP or ether only, not IP only.
2) Adding two to sreg_count to get the second register is too simple: It
works only for four byte regs, not the 16 byte ones. The first
register is always a 16 byte one, though.
Fixing (1) is trivial, fix (2) by introduction of nft_get_next_reg()
doing the right thing. For consistency, use it for among match creation,
too.
Fixes: f315af1cf8871 ("nft: track each register individually")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
This is an inner function to get_cmp_data() returning the op value as-is
for caller examination.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of assuming only one register is used, track all 16 regs
individually.
This avoids need for the 'PREV_PAYLOAD' hack and also avoids the need to
clear out old flags:
When we see that register 'x' will be written to, that register state is
reset automatically.
Existing dissector decodes
ip saddr 1.2.3.4 meta l4proto tcp
... as
-s 6.0.0.0 -p tcp
iptables-nft -s 1.2.3.4 -p tcp is decoded correctly because the expressions
are ordered like:
meta l4proto tcp ip saddr 1.2.3.4
|
... and 'meta l4proto' did clear the PAYLOAD flag.
The simpler fix is:
ctx->flags &= ~NFT_XT_CTX_PAYLOAD;
in nft_parse_cmp(), but that breaks dissection of '1-42', because
the second compare ('cmp lte 42') will not find the
payload expression anymore.
Link: https://lore.kernel.org/netfilter-devel/20220922143544.GA22541@breakpoint.cc/T/#t
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
xlate raw "nft ... ttl eq 1" and so on to the ttl/hl matches.
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
Makes iptables-nft-save dump 'nft meta pkttype' rules.
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
| |
In some code-paths, 'reg' pointer remaining uninitialized is used later
so at least minimal error checking is necessary. Given that a call to
nftnl_alloc_expr() with sane argument should never fail, complain and
exit if it happens.
Fixes: 7e38890c6b4fb ("nft: prepare for dynamic register allocation")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
| |
Store the register that has been allocated and pass it on to the next
expression. NFT_REG_1 is still used.
No functional changes are expected.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
| |
Pass struct nft_handle to helper functions in preparation for the
dynamic register allocation.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
| |
Use meta mark + bitwise + cmp instead of nft_compat mark match.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
| |
In preparation for native mark match support.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
| |
Update the destination register, otherwise nft_parse_cmp() gives up on
interpreting the cmp expression when bitwise sreg != dreg.
Fixes: 2c4a34c30cb4 ("iptables-compat: fix address prefix")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
| |
There are no family-specific versions of struct iptables_command_state
anymore, so no need to hide it behind void pointer. Pass the type as-is
and save a few casts.
While at it, drop unused callbacks parse_bitwise and parse_cmp.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Parsing of rules which jump to a chain pointlessly causes a call to
xtables_find_target() despite the code already knowing the outcome.
Avoid the significant delay for rulesets with many chain jumps by
performing the (standard) target lookup only for accept/drop/return
verdicts.
From a biased test-case on my VM:
| # iptables-nft-save | grep -c -- '-j'
| 133943
| # time ./old/iptables-nft-save >/dev/null
| real 0m45.566s
| user 0m1.308s
| sys 0m8.430s
| # time ./new/iptables-nft-save >/dev/null
| real 0m3.547s
| user 0m0.762s
| sys 0m2.476s
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
| |
Implementations of parse_immediate callback are mostly trivial, the only
relevant part is access to family-specific parts of struct
iptables_command_state when setting goto flag for iptables and
ip6tables. Refactor them into simple set_goto_flag callbacks.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
| |
Detect payload load of th->flags and convert it to xt tcp match
structure.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
| |
same as previous patch, but for udp.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
| |
adds support for
nft ... tcp dport != min-max
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This extends iptables-nft dissector to decode native tcp
port matching. nft ruleset:
table ip filter {
chain INPUT {
type filter hook input priority filter; policy accept;
tcp sport 12345
tcp sport 12345 tcp dport 6789
tcp sport < 1024
tcp dport >= 1024
}
}
$ iptables-nft-save
-A INPUT -p tcp -m tcp --sport 12345
-A INPUT -p tcp -m tcp --sport 12345 --dport 6789
-A INPUT -p tcp -m tcp --sport 0:1023
-A INPUT -p tcp -m tcp --dport 1024:65535
This would allow to extend iptables-nft to prefer
native payload expressions for --sport,dport in the future.
Also, parse_cmp must not clear the "payload" flag, this is because
cmp-based range expressions will contain following sequence:
payload => reg1
cmp reg1 > minv
cmp reg1 < maxv
... so second cmp would work.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When parsing the rule, use a struct with a layout compatible to that of
struct xt_nflog_info, but with a buffer large enough to contain the
whole 128-character nft prefix.
We always send the nflog-group to the kernel since, for nft, log and
nflog targets are handled by the same kernel module, and are
distinguished by whether they define an nflog-group. Therefore, we must
send the group even if it is zero, or the kernel will configure the
target as a log, not an nflog.
Changes to nft_is_expr_compatible were made since only targets which
have an `nflog-group` are compatible. Since nflog targets are
distinguished by having an nflog-group, we ignore targets without one.
We also set the copy-len flag if the snap-len is set since without this,
iptables will mistake `nflog-size` for `nflog-range`.
Signed-off-by: Kyle Bowman <kbowman@cloudflare.com>
Signed-off-by: Alex Forster <aforster@cloudflare.com>
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
| |
NFTNL_EXPR_META_DREG equals NFTNL_EXPR_PAYLOAD_BASE, so we set
dreg to the payload base instead.
It "works" because the simple nft rules currently generated via
ipables-nft have base == register-number but this is a
coincidence.
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
| |
The last users vanished back in 2013. There is identical code in
save_rule_details(), but with only a single user there's not much point
in keeping the function.
Fixes: cdc78b1d6bd7b ("nft: convert rule into a command state structure")
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Legacy iptables fetches the relevant data via libiptc before calling the
shared routine which merely prints data as requested.
Drop the 'basechain' parameter, instead make sure a policy name is
passed only with base chains. Since the function is not shared with
ebtables (which uses a very rudimental header instead), this is safe.
In order to support legacy iptables' checking of iptc_get_references()
return code (printing an error message instead of the reference count),
make refs parameter signed and print the error message if it's negative.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
| |
Have to pass pointer to counters directly since different fields are
being used for some reason.
Since proto_to_name() is not used outside of xshared.c anymore, make it
static.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
| |
The function combines printing of input and output interfaces and
protocol parameter, all being IP family independent. Extend the function
to print fragment option ('-f'), too if requested. While being at it,
drop unused iptables_command_state parameter and reorder the remaining
ones a bit.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
| |
Merge the three identical copies into one and name it 'save_iface' (as
the printed syntax is for "save"-format). Leave arptables alone for now,
its rather complicated whitespace printing doesn't allow for use of the
shared function. Also keep ebtables' custom implementation, it is used
for the --logical-in/--logical-out long-options, too. Apart from that,
ebtables-nft does not use a mask, at all.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
This aligns whitespace printing with legacy iptables' print_rule4() in
order to prepare for further code-sharing.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
Of course there is no such thing as *_tables_names for ebtables, so no
legacy tables checking for ebtables-nft.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
| |
Share the common proto name lookup code. While being at it, make proto
number variable 16bit, values may exceed 256.
This aligns iptables-nft '-p' argument printing with legacy iptables. In
practice, this should make a difference only in corner cases.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|