| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
All ebtables extensions are loaded upon program start as due to the lack
of '-m' parameters, loading on demand is not possible. Introduce
nft_fini_eb() to counteract nft_init_eb() and free dynamic memory in
matches and targets from there.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Iterate over command list and collect chains to cache. Insert them into
a sorted list to pass to __nft_build_cache().
If a command is interested in all chains (e.g., --list), cmd->chain
remains unset. To record this case reliably, use a boolean
('all_chains'). Otherwise, it is hard to distinguish between first call
to nft_cache_level_set() and previous command with NULL cmd->chain
value.
When caching only specific chains, manually add builtin ones for the
given table as well - otherwise nft_xt_builtin_init() will act as if
they don't exist and possibly override non-default chain policies.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Restore per-table operation of cache routines as initially implemented
in commit e2883c5531e6e ("nft-cache: Support partial cache per table").
As before, this doesn't limit fetching of tables (their number is
supposed to be low) but instead limits fetching of sets, chains and
rules to the specified table.
For this to behave correctly when restoring without flushing over
multiple tables, cache must be freed fully after each commit - otherwise
the previous table's cache level is reused for the current one. The
exception being fake cache, used for flushing restore: NFT_CL_FAKE is
set just once at program startup, so it must stay set otherwise
consecutive tables cause pointless cache fetching.
The sole use-case requiring a multi-table cache, iptables-save, is
indicated by req->table being NULL. Therefore, req->table assignment is
a bit sloppy: All calls to nft_cache_level_set() are assumed to set the
same table value, collision detection exists merely to catch programming
mistakes.
Make nft_fini() call nft_release_cache() instead of flush_chain_cache(),
the former does a full cache deinit including cache_req contents.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
| |
This embedded struct collects cache requirement info gathered from parsed
nft_cmds and is interpreted by __nft_build_cache().
While being at it, remove unused parameters passed to the latter
function, nft_handle pointer is sufficient.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
| |
Update among support to work again with the new parser and cache logic.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch uses the new list of commands to calculate the cache
requirements, the rationale after this updates is the following:
#1 Parsing, that builds the list of commands and it also calculates
cache level requirements.
#2 Cache building.
#3 Translate commands to jobs
#4 Translate jobs to netlink
This patch removes the pre-parsing code in xtables-restore.c to
calculate the cache.
After this patch, cache is calculated only once, there is no need
to cancel and refetch for an in-transit transaction.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch updates the parser to generate a list of command objects.
This list of commands is then transformed to a list of netlink jobs.
This new command object stores the rule using the nftnl representation
via nft_rule_new().
To reduce the number of updates in this patch, the nft_*_rule_find()
functions have been updated to restore the native representation to
skip the update of the rule comparison code.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
| |
At least since flushing xtables-restore doesn't fetch chains from kernel
anymore, problems with pending policy rule delete jobs can't happen
anymore.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
| |
If transaction needed a refresh in nft_action(), restore with flush
would fetch a full cache instead of merely refreshing table list
contained in "fake" cache.
To fix this, nft_rebuild_cache() must distinguish between fake cache and
full rule cache. Therefore introduce NFT_CL_FAKE to be distinguished
from NFT_CL_RULES.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move common code into nft_init(), such as:
* initial zeroing nft_handle fields
* family ops lookup and assignment to 'ops' field
* setting of 'family' field
This requires minor adjustments in xtables_restore_main() so extra field
initialization doesn't happen before nft_init() call.
As a side-effect, this fixes segfaulting xtables-monitor binary when
printing rules for trace event as in that code-path 'ops' field wasn't
initialized.
Signed-off-by: Phil Sutter <phil@nwl.cc>
|
|
|
|
|
|
|
|
|
|
| |
In order to support anonymous sets, introduce an intermediate cache
level between NFT_CL_CHAINS and NFT_CL_RULES. Actually chains are not
needed to fetch sets, but given that sets are only needed for rules, put
it late to not slow down fetching chains.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
| |
This is the actual callback used to parse nftables rules. Pass
nft_handle to it so it can access the cache (and possible sets therein).
Having to pass nft_handle to nft_rule_print_save() allows to simplify it
a bit since no family ops lookup has to be done anymore.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
| |
In order for add_match() to create anonymous sets when converting
xtables matches it needs access to nft handle. So pass it along from
callers of family ops' add callback.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When operating on a single chain only, compatibility checking causes
unwanted overhead by checking all chains of the current table. Avoid
this by accepting the current chain name as parameter and pass it along
to nft_chain_list_get().
While being at it, introduce nft_assert_table_compatible() which
calls xtables_error() in case compatibility check fails. If a chain name
was given, include that in error message.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Replace the simple have_cache boolean by a cache level indicator
defining how complete the cache is. Since have_cache indicated full
cache (including rules), make code depending on it check for cache level
NFT_CL_RULES.
Core cache fetching routine __nft_build_cache() accepts a new level via
parameter and raises cache completeness to that level.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
| |
The amount of code dealing with caching only is considerable and hence
deserves an own source file.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Unless --noflush was given, xtables-restore merely needs the list of
tables to decide whether to delete it or not. Introduce nft_fake_cache()
function which populates table list, initializes chain lists (so
nft_chain_list_get() returns an empty list instead of NULL) and sets
'have_cache' to turn any later calls to nft_build_cache() into nops.
If --noflush was given, call nft_build_cache() just once instead of for
each table line in input.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
| |
No need to check family value from nft_commit() if we can have a
dedicated callback for bridge family.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
| |
As decided upon at NFWS2019, drop support for configurable nftables base
chains to use with iptables-nft.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
| |
Support passing arbitrary data (via void pointer) to the callback.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
| |
Store them next to the mnl_socket pointer. While being at it, add a
comment to mnl_set_rcvbuffer() explaining why the buffer size is
changed.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Phil Sutter says:
"The problem is that data in h->obj_list potentially sits in cache, too.
At least rules have to be there so insert with index works correctly. If
the cache is flushed before regenerating the batch, use-after-free
occurs which crashes the program."
This patch keeps around the original cache until we have refreshed the
batch.
Fixes: 862818ac3a0de ("xtables: add and use nft_build_cache")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
| |
nft_table_find() uses the table list cache to look up for existing
tables.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
| |
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
| |
Add new structure that encloses the cache and update the code to use it.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We currently race when several xtables-nft-restore processes attempt to
handle rules in parallel. For instance, when no rules are present at
all, then
iptables-nft-restore < X & iptables-nft-restore < X
... can cause rules to be restored twice.
Reason is that both processes might detect 'no rules exist', so
neither issues a flush operation.
We can't unconditionally issue the flush, because it would
cause kernel to fail with -ENOENT unless the to-be-flushed table
exists.
This change passes the generation id that was used to build
the transaction to the kernel.
In case another process changed *any* rule somewhere, the transaction
will now fail with -ERESTART.
We then flush the cache, re-fetch the ruleset and refresh
our transaction.
For example, in the above 'parallel restore' case, the iptables-restore
instance that lost the race would detect that the table has been created
already, and would add the needed flush.
In a similar vein, in case --noflush is used, we will add the flush
op for user-defined chains that were created in the mean-time.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Will be used with the "generation id" infrastructure.
When we're told that the commit failed because someone else made
changes, we can use this to re-initialize the cache and then
revalidate the transaction list (e.g. to detect that we now have
to flush the user-defined chain 'foo' that we wanted to create, but
was added just now by someone else).
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The idea here is to move the 'flush' decision into the core, rather than
have the decision in the frontend.
This will be required later when "generation id" is passed to kernel.
In this case, we might have to add the flush when re-trying the
transaction.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Legacy ebtables supports policies for user-defined chains - and what's
worse, they default to ACCEPT unlike anywhere else. So lack of support
for this braindead feature in ebtables-nft is actually a change of
behaviour which very likely affects all ebtables users out there.
The solution implemented here uses an implicit (and transparent) last
rule in all user-defined ebtables-nft chains with policy other than
RETURN. This rule is identified by an nft comment
"XTABLES_EB_INTERNAL_POLICY_RULE" (since commit ccf154d7420c0 ("xtables:
Don't use native nftables comments") nft comments are not used
otherwise).
To minimize interference with existing code, this policy rule is removed
from chains during cache population and the policy is saved in
NFTNL_CHAIN_POLICY attribute. When committing changes to the kernel,
nft_commit() traverses through the list of chains and (re-)creates
policy rules if required.
In ebtables-nft-restore, table flushes are problematic. To avoid weird
kernel error responses, introduce a custom 'table_flush' callback which
removes any pending policy rule add/remove jobs prior to creating the
NFT_COMPAT_TABLE_FLUSH one.
I've hidden all this mess behind checks for h->family, so hopefully
impact on {ip,ip6,arp}tables-nft should be negligible.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
iptables-restore allows to insert rules at a certain position which is
problematic for iptables-nft to realize since rule position is not
determined by number but handle of previous or following rule and in
case the rules surrounding the new one are new as well, they don't have
a handle to refer to yet.
Fix this by making use of NFTNL_RULE_POSITION_ID attribute: When
inserting before a rule which does not have a handle, refer to it using
its NFTNL_RULE_ID value. If the latter doesn't exist either, assign a
new one to it.
The last used rule ID value is tracked in a new field of struct
nft_handle which is incremented before each use.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
| |
When replacing a rule, the replacement was simply appended to the
chain's rule list. Instead, insert it where the rule it replaces was.
This also fixes for zero counters command to remove the old rule from
cache.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
| |
Replace the function by nftnl_chain_list_lookup_byname() as provided by
libnftnl.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use recently introduced support for rules inside chains in libnftnl to
introduce a rule cache per chain instead of a global one.
A tricky bit is to decide if cache should be updated or not. Previously,
the global rule cache was populated just once and then reused unless
being flushed completely (via call to flush_rule_cache() with
NULL-pointer table argument). Resemble this behaviour by introducing a
boolean indicating cache status and fetch rules for all chains when
updating the chain cache in nft_chain_list_get().
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The problem with converting libxt_comment into nftables comment is that
rules change when parsing from kernel due to comment match being moved
to the end of the match list. And since match ordering matters, the rule
may not be found anymore when checking or deleting. Apart from that,
iptables-nft didn't support multiple comments per rule anymore. This is
a compatibility issue without technical reason.
Leave conversion from nftables comment to libxt_comment in place so we
don't break running systems during an update.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
| |
These definitions should be const, propagate this to all existing users.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
| |
Move this to the structure that stores, stateful information. Introduce
nft_table_initialized() and use it.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
| |
Place this back into the structure that stores the state information.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
| |
Use enum nft_table_type to set the new type field in the structure that
define tables.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Being able to omit the previously obligatory table name check when
iterating over the chain cache might help restore performance with large
rulesets in xtables-save and -restore.
There is one subtle quirk in the code: flush_chain_cache() did free the
global chain cache if not called with a table name but didn't if a table
name was given even if it emptied the chain cache. In other places,
chain_cache being non-NULL prevented a cache update from happening, so
this patch establishes the same behaviour (for each individual chain
cache) since otherwise unexpected cache updates lead to weird problems.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
| |
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
| |
So we can remove nft_chain_dump() and replace nftnl_chain_get_list().
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
|
|
|
|
|
|
|
|
| |
Make sure return codes match legacy ones at least for a few selected
commands typically used to check ruleset state.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
| |
This adds C implementations for arptables-save and -restore in compat
layer based on the two perl scripts in legacy arptables repository.
To share common code, introduce nft_init_arp() analogous to
nft_init_eb() introduced earlier.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
| |
This drops the dedicated input parser (which was broken in many ways
anyway) and replaces it by the common one now that all required knobs
are in place.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Legacy ebtables-save does not use a policy string of '-' to denote
user-defined chains but instead lists them with a policy of ACCEPT.
In order to use ebtables_restore_parse() for ebtables-save
implementation, make use of builtin table definitions to decide whether
a given chain is a builtin one or not.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Legacy ip{,6}tables prints feedback for various commands if in verbose
mode, make sure nft variants do the same.
There is one difference, namely when checking a rule (-C command):
Legacy ip{,6}tables print the rule in any case, nft variants don't in
case the rule wasn't found. Changing this though would require to
populate the nftnl_rule object just for printing, which is probably not
feasible.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
| |
Since nft_init_eb() is shared among standalone ebtables and
ebtables-restore, allow for callers to pass the program name.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The code for ebtables-restore was derived from legacy code,
ebtables-save is actually a new implementation using the existing
infrastructure and trying to adhere to legacy perl script output
formatting as much as possible.
This introduces a new format flag (FMT_EBT_SAVE) to allow
nft_bridge_save_rule() to distinguish between ruleset listing (i.e.,
ebtables -L) and saving via ebtables-save - the two differ in how
counters are being formatted. Odd, but that's how it is.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
| |
This wraps nft_init(), adding required things needed for ebtables.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|
|
|
|
|
|
|
|
| |
Preparing ebtables-save implementation, allow for callers to pass format
bits to nft_rule_save() instead of just the 'counters' boolean.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Florian Westphal <fw@strlen.de>
|