summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* Don't overwrite errno with return value of setsockopt (which is -1 on error).Patrick McHardy2006-04-221-6/+2
| | | | Fixes "Unknown error 4294967295" message (bugzilla #460).
* Revert incorrect fix for "Unknown error 4294967295" problemPatrick McHardyHarald Welte2006-04-221-2/+0
|
* When entering an invalid command (such as iptables -A INPUT -j MARK --set-markHarald Welte2006-04-211-0/+2
| | | | 1), the error message "Unknown error 4294967295" is displayed; (Closes: #460)
* In ip[6]tables.c, NUMBER_OF_OPT was increased to 12 for the OPT_COUNTERSPatrick McHardyHarald Welte2006-04-222-30/+32
| | | | | | option. However, the new array element is not initialized in either commands_v_options[NUMBER_OF_CMD][NUMBER_OF_OPT] or inverse_for_options[NUMBER_OF_OPT]. (Closes: #462)
* cmdflags is used in cmd2char() to return the option for a command. It uses theHarald Welte2006-04-212-4/+2
| | | | | | bit position of the command mask as an index in the array. There's no entry for CMD_CHECK (0x0800U), so lookups for CMD_RENAME_CHAIN (0x1000U) index outside the array. (Closes: #463)
* [IPTABLES,IP6TABLES]: check invalid esp spi rangeYasuyuki KOZAKAI2006-04-152-0/+6
|
* [IP6TABLES] kill manual comparing protocol name with "ipv6-icmp".Yasuyuki KOZAKAI2006-04-151-3/+1
|
* fix loading shared library of ICMPv6 match.Yasuyuki KOZAKAI2006-04-153-1/+1
| | | | | | | | | | | | | The current ip6tables tries to load libip6t_icmp6.so when user types 'ip6tables -p icmpv6 ...' or 'ip6tables ... -m icmpv6' ...', and it fails. This patch renames libip6t_icmpv6.c to libip6t_icmp6.c so that ip6tables can load it. Now kernel module and user library has same name 'icmp6'. It can reduce confusion about name mismatch. That's why I renamed it instead of reverting change in find_match() which brought this bug. This patch keeps compatibiity and we can use '-p icmpv6', '-p ipv6-icmpv6', '-m icmpv6', '-m ipv6-icmpv6', and '-m icmp6', as ever.
* [IPTABLES,IP6TABLES]: fix the path to detect esp/connbytes support in kernelHarald Welte2006-04-122-2/+2
| | | | The recent kernels don't have ipt_connbytes.c and ip6t_esp.c.
* Correct iptables-save output of osf module (Daniel De Graaf)Daniel De Graaf2006-03-311-0/+8
|
* don't allow to specify protocol of IPv6 extension header (Yasuyuki Kozakai)Yasuyuki KOZAKAI2006-03-291-0/+16
| | | | | | | Sometimes I hear that people do 'ip6tables -p ah ...' which never matches any packet. IPv6 extension headers except of ESP are skipped and invalid as argument of '-p'. Then I propose that ip6tables exits with error in such case.
* Multiple matches of the same type can be specified on the commandline.Joszef Kadlecsik2006-03-034-32/+84
| | | | | | | | | | If two or more matches of the same type are detected then the options are assumed to be grouped in order to tell which option belongs to which match: ... -m foo ... <options0> ... -m foo ... <options1> ... Otherwise the commandline parsing is unmodified.
* Make '-p all' a special case that is handled before calling getprotoent() ↵Harald Welte2006-02-112-2/+14
| | | | (Closes: #446)
* fix double-free if a single match is used multiple times within a signle ruleHarald Welte2006-02-112-2/+6
| | | | | | (Closes: #440). However, while this fixes the double-free, it still doesn't make iptables support two of the same matches within one rule. Apparently the last matchinfo is copied into all the previous matchinfo instances.
* don't install libiptc.aHarald Welte2006-02-091-1/+2
|
* fix segfault or loading of invalid counters in ip[6]tables-restore (Olaf ↵Harald Welte2006-02-092-2/+8
| | | | Rempel) (Closes: #437)
* make policy match compile independant of kernel headersv1.3.5Harald Welte2006-02-013-2/+6
|
* Some !%$!*##$@ has modified the kernel include/linux/netfilter_ipv4/ipt_sctp.hHarald Welte2006-02-011-0/+13
| | | | file in a way that breaks userspace :(
* fix ipt_conntrack compilation against very early (2.4.0) kernel releasesHarald Welte2006-02-011-1/+1
|
* remove other bits of old ip pool code, people should use ipset ↵Harald Welte2006-02-015-323/+2
| | | | (ipset.netfilter.org) these days
* remove ippoolHarald Welte2006-02-013-679/+0
|
* Prepare policy match for x_tables unification by making sure bothPatrick McHardy2006-01-317-16/+126
| | | | ipt_policy and ip6t_policy use the same data structure.
* fix 'save' (Michael Rash)Michael Rash2006-01-301-2/+2
|
* major manpage update (Yasuyuki Kozakai)Yasuyuki KOZAKAI2006-01-3027-88/+149
|
* Add 'copy+paste' support for 'state' and 'connmark' match, as well asHarald Welte2006-01-264-1/+535
| | | | 'CONNMARK' target for ip6tables / nf_conntrack_l3proto_ipv6. This is a temporary solution for the iptables-1.3.x branch, since the 1.4.x branch will have proper support.
* add note about deprecated stateHarald Welte2006-01-261-0/+2
|
* fix spelling 'adress' -> 'address' (Closes: #431) (MJ Anthony)Harald Welte2006-01-222-2/+2
|
* Fix "empty policy element" complaining in non-strict mode.Noticed by Tom Eastep2006-01-222-2/+4
| | | | Noticed by Tom Eastep <teastep@shorewall.net>.
* Clarify --tunnel-src/--tunnel-dst optionsPatrick McHardy2006-01-122-6/+10
|
* Move empty policy element check to also catch last elementPatrick McHardy2006-01-122-10/+12
|
* Don't allow using --next option without specifying a policy elementPatrick McHardy2006-01-122-4/+14
|
* Fix invalid assignment of tunnel-src to dest address (Patrick McHardy)Patrick McHardy2006-01-091-2/+2
|
* Add documentation for string match (Pablo Neira)Pablo Neira2006-01-031-0/+15
|
* Fix probing for supported revisions (Jones Desougi <jones@ingate.com>)Jones Desougi2005-12-224-8/+10
| | | | Bugzilla #413
* fix iptables-save of 'goto' target (Closes: #410)Harald Welte2005-12-051-2/+2
|
* Add note that TCPMSS is only valid in the mangle table (not true today, but ↵Patrick McHardy2005-12-051-1/+4
| | | | maybe someday)
* fix compilation of iptables on [old] systems that don't have IPT_F_GOTOHarald Welte2005-11-242-0/+6
|
* note that we can only delete chains that are emptyHarald Welte2005-11-221-3/+4
|
* tcp-rst is the alias, not tcp-reset (Torsten Hilbrich)Harald Welte2005-11-221-1/+1
|
* Add policy match extensions from patch-o-maticPatrick McHardy2005-11-196-0/+998
|
* Fix some gcc-4 warningsPatrick McHardy2005-11-184-7/+7
|
* Don't eat numeric arguments for other extensionsPatrick McHardy2005-11-181-4/+12
|
* The conntrack match does not print any info for --ctproto, thusPhil Oester2005-11-171-0/+7
| | | | | breaking iptables-restore of any rules using this option. Below patch adds output and closes bug #398. (Phil Oester)
* only set revisions on real targets, not on jumps. (Pablo Neira)Pablo Neira2005-11-171-1/+3
|
* - Fix memory leak in TC_COMMIT() (Markus Sundberg)Harald Welte2005-11-121-23/+25
| | | | | - Cleanup error path of TC_COMMIT() - Correctly propagate errors of setsockopt to calling function
* add 'goto' support (Henrik Nordstrom <hno@marasystems.com>)Henrik Nordstrom2005-11-053-3/+33
|
* fix connmark, it's now only 32bits (Deti Fliegl <deti@fliegl.de)v1.3.4Deti Fliegl2005-11-032-53/+10
| | | | | | We'ver screwed this up with the 2.6.14 release. It refuses any mask that extends 32bits. We should have fixed this by adding a new target/match revision, but now it's too late anyway :(
* about to release 1.3.4Harald Welte2005-10-311-2/+2
|
* The conntrack match extension doesn't handle address inversion correctly. ↵Tom Eastep2005-09-191-2/+2
| | | | (Tom Eastep)
* Kernels higher than 2.6.10 don't support multiple --to arguments inPhil Oester2005-09-194-0/+41
| | | | | | | | | | | | | | | | | | | | | DNAT and SNAT targets. At present, the error is somewhat vague: # iptables -t nat -A foo -j SNAT --to 1.2.3.4 --to 2.3.4.5 iptables: Invalid argument But if we want current iptables to work with kernels <= 2.6.10, we cannot simply disallow this in all cases. So the below patch adds kernel version checking to iptables, and utilizes it in [DS]NAT. Now, users will see a more informative error: # iptables -t nat -A foo -j SNAT --to 1.2.3.4 --to 2.3.4.5 iptables v1.3.3: Multiple --to-source not supported This generic infrastructure (shamelessly lifted from procps btw) may come in handy in the future for other changes. This fixes bugzilla #367. (Phil Oester)