mirror/freebsd
1
0
Fork 0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-22 11:17:19 +00:00
Code Issues Releases Activity
facc67676f
freebsd/sys/netinet/ip_fw2.c

2765 lines
70 KiB
C
Raw Normal View History

The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* Copyright (c) 2002 Luigi Rizzo, Universita` di Pisa
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
* 2. Redistributions in binary form must reproduce the above copyright
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
*
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
#define DEB(x)
#define DDB(x) x
/*
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
* Implement IP packet firewall (new version)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
#if !defined(KLD_MODULE)
#include "opt_ipfw.h"
#include "opt_ipdn.h"
#include "opt_ipdivert.h"
#include "opt_inet.h"
#ifndef INET
#error IPFIREWALL requires INET.
#endif /* INET */
#endif
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
#define IPFW2 1
#if IPFW2
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/kernel.h>
#include <sys/proc.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/sysctl.h>
#include <sys/syslog.h>
#include <sys/ucred.h>
#include <net/if.h>
#include <net/route.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/in_var.h>
#include <netinet/in_pcb.h>
#include <netinet/ip.h>
#include <netinet/ip_var.h>
#include <netinet/ip_icmp.h>
#include <netinet/ip_fw.h>
#include <netinet/ip_dummynet.h>
#include <netinet/tcp.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#include <netinet/tcpip.h>
#include <netinet/udp.h>
#include <netinet/udp_var.h>
#include <netinet/if_ether.h> /* XXX for ETHERTYPE_IP */
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
#include <machine/in_cksum.h> /* XXX for in_cksum */
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
/*
* XXX This one should go in sys/mbuf.h. It is used to avoid that
* a firewall-generated packet loops forever through the firewall.
*/
#ifndef M_SKIP_FIREWALL
#define M_SKIP_FIREWALL 0x4000
#endif
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
/*
* set_disable contains one bit per set value (0..31).
* If the bit is set, all rules with the corresponding set
* are disabled. Set 31 is reserved for the default rule
* and CANNOT be disabled.
*/
static u_int32_t set_disable;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
static int fw_verbose;
static int verbose_limit;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
static struct callout_handle ipfw_timeout_h;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#define IPFW_DEFAULT_RULE 65535
/*
* list of rules for layer 3
*/
static struct ip_fw *layer3_chain;
MALLOC_DEFINE(M_IPFW, "IpFw/IpAcct", "IpFw/IpAcct chain's");
static int fw_debug = 1;
static int autoinc_step = 100; /* bounded to 1..1000 in add_rule() */
#ifdef SYSCTL_NODE
SYSCTL_NODE(_net_inet_ip, OID_AUTO, fw, CTLFLAG_RW, 0, "Firewall");
Lock the sysctl(8) knobs that turn ip{,6}fw(8) firewalling and firewall logging on and off when at elevated securelevel(8). It would be nice to be able to only lock these at securelevel >= 3, like rules are, but there is no such functionality at present. I don't see reason to be adding features to securelevel(8) with MAC being merged into 5.0. PR: kern/39396 Reviewed by: luigi MFC after: 1 week
2002-08-25 03:50:29 +00:00
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, enable,
CTLFLAG_RW | CTLFLAG_SECURE,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
&fw_enable, 0, "Enable ipfw");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, autoinc_step, CTLFLAG_RW,
&autoinc_step, 0, "Rule number autincrement step");
Lock the sysctl(8) knobs that turn ip{,6}fw(8) firewalling and firewall logging on and off when at elevated securelevel(8). It would be nice to be able to only lock these at securelevel >= 3, like rules are, but there is no such functionality at present. I don't see reason to be adding features to securelevel(8) with MAC being merged into 5.0. PR: kern/39396 Reviewed by: luigi MFC after: 1 week
2002-08-25 03:50:29 +00:00
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, one_pass,
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
CTLFLAG_RW | CTLFLAG_SECURE,
&fw_one_pass, 0,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
"Only do a single pass through ipfw when using dummynet(4)");
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, debug, CTLFLAG_RW,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
&fw_debug, 0, "Enable printing of debug ip_fw statements");
Lock the sysctl(8) knobs that turn ip{,6}fw(8) firewalling and firewall logging on and off when at elevated securelevel(8). It would be nice to be able to only lock these at securelevel >= 3, like rules are, but there is no such functionality at present. I don't see reason to be adding features to securelevel(8) with MAC being merged into 5.0. PR: kern/39396 Reviewed by: luigi MFC after: 1 week
2002-08-25 03:50:29 +00:00
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, verbose,
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
CTLFLAG_RW | CTLFLAG_SECURE,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
&fw_verbose, 0, "Log matches to ipfw rules");
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, verbose_limit, CTLFLAG_RW,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
&verbose_limit, 0, "Set upper limit of matches of ipfw rules logged");
/*
* Description of dynamic rules.
*
* Dynamic rules are stored in lists accessed through a hash table
* (ipfw_dyn_v) whose size is curr_dyn_buckets. This value can
* be modified through the sysctl variable dyn_buckets which is
* updated when the table becomes empty.
*
* XXX currently there is only one list, ipfw_dyn.
*
* When a packet is received, its address fields are first masked
* with the mask defined for the rule, then hashed, then matched
* against the entries in the corresponding list.
* Dynamic rules can be used for different purposes:
* + stateful rules;
* + enforcing limits on the number of sessions;
* + in-kernel NAT (not implemented yet)
*
* The lifetime of dynamic rules is regulated by dyn_*_lifetime,
* measured in seconds and depending on the flags.
*
* The total number of dynamic rules is stored in dyn_count.
* The max number of dynamic rules is dyn_max. When we reach
* the maximum number of rules we do not create anymore. This is
* done to avoid consuming too much memory, but also too much
* time when searching on each packet (ideally, we should try instead
* to put a limit on the length of the list on each bucket...).
*
* Each dynamic rule holds a pointer to the parent ipfw rule so
* we know what action to perform. Dynamic rules are removed when
* the parent rule is deleted. XXX we should make them survive.
*
* There are some limitations with dynamic rules -- we do not
* obey the 'randomized match', and we do not do multiple
* passes through the firewall. XXX check the latter!!!
*/
static ipfw_dyn_rule **ipfw_dyn_v = NULL;
static u_int32_t dyn_buckets = 256; /* must be power of 2 */
static u_int32_t curr_dyn_buckets = 256; /* must be power of 2 */
/*
* Timeouts for various events in handing dynamic rules.
*/
static u_int32_t dyn_ack_lifetime = 300;
static u_int32_t dyn_syn_lifetime = 20;
static u_int32_t dyn_fin_lifetime = 1;
static u_int32_t dyn_rst_lifetime = 1;
static u_int32_t dyn_udp_lifetime = 10;
static u_int32_t dyn_short_lifetime = 5;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
/*
* Keepalives are sent if dyn_keepalive is set. They are sent every
* dyn_keepalive_period seconds, in the last dyn_keepalive_interval
* seconds of lifetime of a rule.
* dyn_rst_lifetime and dyn_fin_lifetime should be strictly lower
* than dyn_keepalive_period.
*/
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
static u_int32_t dyn_keepalive_interval = 20;
static u_int32_t dyn_keepalive_period = 5;
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
static u_int32_t dyn_keepalive = 1; /* do send keepalives */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
static u_int32_t static_count; /* # of static rules */
static u_int32_t static_len; /* size in bytes of static rules */
static u_int32_t dyn_count; /* # of dynamic rules */
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
static u_int32_t dyn_max = 4096; /* max # of dynamic rules */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_buckets, CTLFLAG_RW,
&dyn_buckets, 0, "Number of dyn. buckets");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, curr_dyn_buckets, CTLFLAG_RD,
&curr_dyn_buckets, 0, "Current Number of dyn. buckets");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_count, CTLFLAG_RD,
&dyn_count, 0, "Number of dyn. rules");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_max, CTLFLAG_RW,
&dyn_max, 0, "Max number of dyn. rules");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, static_count, CTLFLAG_RD,
&static_count, 0, "Number of static rules");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_ack_lifetime, CTLFLAG_RW,
&dyn_ack_lifetime, 0, "Lifetime of dyn. rules for acks");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_syn_lifetime, CTLFLAG_RW,
&dyn_syn_lifetime, 0, "Lifetime of dyn. rules for syn");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_fin_lifetime, CTLFLAG_RW,
&dyn_fin_lifetime, 0, "Lifetime of dyn. rules for fin");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_rst_lifetime, CTLFLAG_RW,
&dyn_rst_lifetime, 0, "Lifetime of dyn. rules for rst");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_udp_lifetime, CTLFLAG_RW,
&dyn_udp_lifetime, 0, "Lifetime of dyn. rules for UDP");
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_short_lifetime, CTLFLAG_RW,
&dyn_short_lifetime, 0, "Lifetime of dyn. rules for other situations");
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
SYSCTL_INT(_net_inet_ip_fw, OID_AUTO, dyn_keepalive, CTLFLAG_RW,
&dyn_keepalive, 0, "Enable keepalives for dyn. rules");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#endif /* SYSCTL_NODE */
static ip_fw_chk_t ipfw_chk;
ip_dn_ruledel_t *ip_dn_ruledel_ptr = NULL; /* hook into dummynet */
/*
* This macro maps an ip pointer into a layer3 header pointer of type T
*/
#define L3HDR(T, ip) ((T *)((u_int32_t *)(ip) + (ip)->ip_hl))
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
static __inline int
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
icmptype_match(struct ip *ip, ipfw_insn_u32 *cmd)
{
int type = L3HDR(struct icmp,ip)->icmp_type;
return (type <= ICMP_MAXTYPE && (cmd->d[0] & (1<<type)) );
}
#define TT ( (1 << ICMP_ECHO) | (1 << ICMP_ROUTERSOLICIT) | \
(1 << ICMP_TSTAMP) | (1 << ICMP_IREQ) | (1 << ICMP_MASKREQ) )
static int
is_icmp_query(struct ip *ip)
{
int type = L3HDR(struct icmp, ip)->icmp_type;
return (type <= ICMP_MAXTYPE && (TT & (1<<type)) );
}
#undef TT
/*
* The following checks use two arrays of 8 or 16 bits to store the
* bits that we want set or clear, respectively. They are in the
* low and high half of cmd->arg1 or cmd->d[0].
*
* We scan options and store the bits we find set. We succeed if
*
* (want_set & ~bits) == 0 && (want_clear & ~bits) == want_clear
*
* The code is sometimes optimized not to store additional variables.
*/
static int
flags_match(ipfw_insn *cmd, u_int8_t bits)
{
u_char want_clear;
bits = ~bits;
if ( ((cmd->arg1 & 0xff) & bits) != 0)
return 0; /* some bits we want set were clear */
want_clear = (cmd->arg1 >> 8) & 0xff;
if ( (want_clear & bits) != want_clear)
return 0; /* some bits we want clear were set */
return 1;
}
static int
ipopts_match(struct ip *ip, ipfw_insn *cmd)
{
int optlen, bits = 0;
u_char *cp = (u_char *)(ip + 1);
int x = (ip->ip_hl << 2) - sizeof (struct ip);
for (; x > 0; x -= optlen, cp += optlen) {
int opt = cp[IPOPT_OPTVAL];
if (opt == IPOPT_EOL)
break;
if (opt == IPOPT_NOP)
optlen = 1;
else {
optlen = cp[IPOPT_OLEN];
if (optlen <= 0 || optlen > x)
return 0; /* invalid or truncated */
}
switch (opt) {
default:
break;
case IPOPT_LSRR:
bits |= IP_FW_IPOPT_LSRR;
break;
case IPOPT_SSRR:
bits |= IP_FW_IPOPT_SSRR;
break;
case IPOPT_RR:
bits |= IP_FW_IPOPT_RR;
break;
case IPOPT_TS:
bits |= IP_FW_IPOPT_TS;
break;
}
}
return (flags_match(cmd, bits));
}
static int
tcpopts_match(struct ip *ip, ipfw_insn *cmd)
{
int optlen, bits = 0;
struct tcphdr *tcp = L3HDR(struct tcphdr,ip);
u_char *cp = (u_char *)(tcp + 1);
int x = (tcp->th_off << 2) - sizeof(struct tcphdr);
for (; x > 0; x -= optlen, cp += optlen) {
int opt = cp[0];
if (opt == TCPOPT_EOL)
break;
if (opt == TCPOPT_NOP)
optlen = 1;
else {
optlen = cp[1];
if (optlen <= 0)
break;
}
switch (opt) {
default:
break;
case TCPOPT_MAXSEG:
bits |= IP_FW_TCPOPT_MSS;
break;
case TCPOPT_WINDOW:
bits |= IP_FW_TCPOPT_WINDOW;
break;
case TCPOPT_SACK_PERMITTED:
case TCPOPT_SACK:
bits |= IP_FW_TCPOPT_SACK;
break;
case TCPOPT_TIMESTAMP:
bits |= IP_FW_TCPOPT_TS;
break;
case TCPOPT_CC:
case TCPOPT_CCNEW:
case TCPOPT_CCECHO:
bits |= IP_FW_TCPOPT_CC;
break;
}
}
return (flags_match(cmd, bits));
}
static int
iface_match(struct ifnet *ifp, ipfw_insn_if *cmd)
{
if (ifp == NULL) /* no iface with this packet, match fails */
return 0;
/* Check by name or by IP address */
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
if (cmd->name[0] != '\0') { /* match by name */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/* Check unit number (-1 is wildcard) */
if (cmd->p.unit != -1 && cmd->p.unit != ifp->if_unit)
return(0);
/* Check name */
if (!strncmp(ifp->if_name, cmd->name, IFNAMSIZ))
return(1);
} else {
struct ifaddr *ia;
TAILQ_FOREACH(ia, &ifp->if_addrhead, ifa_link) {
if (ia->ifa_addr == NULL)
continue;
if (ia->ifa_addr->sa_family != AF_INET)
continue;
if (cmd->p.ip.s_addr == ((struct sockaddr_in *)
(ia->ifa_addr))->sin_addr.s_addr)
return(1); /* match */
}
}
return(0); /* no match, fail ... */
}
static u_int64_t norule_counter; /* counter for ipfw_log(NULL...) */
#define SNPARGS(buf, len) buf + len, sizeof(buf) > len ? sizeof(buf) - len : 0
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
#define SNP(buf) buf, sizeof(buf)
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* We enter here when we have a rule with O_LOG.
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
* XXX this function alone takes about 2Kbytes of code!
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
static void
ipfw_log(struct ip_fw *f, u_int hlen, struct ether_header *eh,
struct mbuf *m, struct ifnet *oif)
{
char *action;
int limit_reached = 0;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
char action2[40], proto[48], fragment[28];
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
fragment[0] = '\0';
proto[0] = '\0';
if (f == NULL) { /* bogus pkt */
if (verbose_limit != 0 && norule_counter >= verbose_limit)
return;
norule_counter++;
if (norule_counter == verbose_limit)
limit_reached = verbose_limit;
action = "Refuse";
} else { /* O_LOG is the first action, find the real one */
ipfw_insn *cmd = ACTION_PTR(f);
ipfw_insn_log *l = (ipfw_insn_log *)cmd;
if (l->max_log != 0 && l->log_left == 0)
return;
l->log_left--;
if (l->log_left == 0)
limit_reached = l->max_log;
cmd += F_LEN(cmd); /* point to first action */
if (cmd->opcode == O_PROB)
cmd += F_LEN(cmd);
action = action2;
switch (cmd->opcode) {
case O_DENY:
action = "Deny";
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_REJECT:
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
if (cmd->arg1==ICMP_REJECT_RST)
action = "Reset";
else if (cmd->arg1==ICMP_UNREACH_HOST)
action = "Reject";
else
snprintf(SNPARGS(action2, 0), "Unreach %d",
cmd->arg1);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_ACCEPT:
action = "Accept";
break;
case O_COUNT:
action = "Count";
break;
case O_DIVERT:
snprintf(SNPARGS(action2, 0), "Divert %d",
cmd->arg1);
break;
case O_TEE:
snprintf(SNPARGS(action2, 0), "Tee %d",
cmd->arg1);
break;
case O_SKIPTO:
snprintf(SNPARGS(action2, 0), "SkipTo %d",
cmd->arg1);
break;
case O_PIPE:
snprintf(SNPARGS(action2, 0), "Pipe %d",
cmd->arg1);
break;
case O_QUEUE:
snprintf(SNPARGS(action2, 0), "Queue %d",
cmd->arg1);
break;
case O_FORWARD_IP: {
ipfw_insn_sa *sa = (ipfw_insn_sa *)cmd;
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
int len;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
len = snprintf(SNPARGS(action2, 0), "Forward to %s",
inet_ntoa(sa->sa.sin_addr));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (sa->sa.sin_port)
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
snprintf(SNPARGS(action2, len), ":%d",
o Fix byte order logging issue: sa.sin_port is already in host byte order. PR: kern/45964 Submitted by: Sascha Blank <sblank@tiscali.de> Reviewed by: luigi MFC after: 1 week
2002-12-15 09:44:02 +00:00
sa->sa.sin_port);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
break;
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
default:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
action = "UNKNOWN";
break;
}
}
if (hlen == 0) { /* non-ip */
snprintf(SNPARGS(proto, 0), "MAC");
} else {
struct ip *ip = mtod(m, struct ip *);
/* these three are all aliases to the same thing */
struct icmp *const icmp = L3HDR(struct icmp, ip);
struct tcphdr *const tcp = (struct tcphdr *)icmp;
struct udphdr *const udp = (struct udphdr *)icmp;
int ip_off, offset, ip_len;
int len;
if (eh != NULL) { /* layer 2 packets are as on the wire */
ip_off = ntohs(ip->ip_off);
ip_len = ntohs(ip->ip_len);
} else {
ip_off = ip->ip_off;
ip_len = ip->ip_len;
}
offset = ip_off & IP_OFFMASK;
switch (ip->ip_p) {
case IPPROTO_TCP:
len = snprintf(SNPARGS(proto, 0), "TCP %s",
inet_ntoa(ip->ip_src));
if (offset == 0)
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
snprintf(SNPARGS(proto, len), ":%d %s:%d",
ntohs(tcp->th_sport),
inet_ntoa(ip->ip_dst),
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ntohs(tcp->th_dport));
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
else
snprintf(SNPARGS(proto, len), " %s",
inet_ntoa(ip->ip_dst));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case IPPROTO_UDP:
len = snprintf(SNPARGS(proto, 0), "UDP %s",
inet_ntoa(ip->ip_src));
if (offset == 0)
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
snprintf(SNPARGS(proto, len), ":%d %s:%d",
ntohs(udp->uh_sport),
inet_ntoa(ip->ip_dst),
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ntohs(udp->uh_dport));
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
else
snprintf(SNPARGS(proto, len), " %s",
inet_ntoa(ip->ip_dst));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
break;
case IPPROTO_ICMP:
if (offset == 0)
len = snprintf(SNPARGS(proto, 0),
"ICMP:%u.%u ",
icmp->icmp_type, icmp->icmp_code);
else
len = snprintf(SNPARGS(proto, 0), "ICMP ");
len += snprintf(SNPARGS(proto, len), "%s",
inet_ntoa(ip->ip_src));
snprintf(SNPARGS(proto, len), " %s",
inet_ntoa(ip->ip_dst));
break;
default:
len = snprintf(SNPARGS(proto, 0), "P:%d %s", ip->ip_p,
inet_ntoa(ip->ip_src));
snprintf(SNPARGS(proto, len), " %s",
inet_ntoa(ip->ip_dst));
break;
}
if (ip_off & (IP_MF | IP_OFFMASK))
snprintf(SNPARGS(fragment, 0), " (frag %d:%d@%d%s)",
ntohs(ip->ip_id), ip_len - (ip->ip_hl << 2),
offset << 3,
(ip_off & IP_MF) ? "+" : "");
}
if (oif || m->m_pkthdr.rcvif)
log(LOG_SECURITY | LOG_INFO,
"ipfw: %d %s %s %s via %s%d%s\n",
f ? f->rulenum : -1,
action, proto, oif ? "out" : "in",
oif ? oif->if_name : m->m_pkthdr.rcvif->if_name,
oif ? oif->if_unit : m->m_pkthdr.rcvif->if_unit,
fragment);
else
log(LOG_SECURITY | LOG_INFO,
"ipfw: %d %s %s [no if info]%s\n",
f ? f->rulenum : -1,
action, proto, fragment);
if (limit_reached)
log(LOG_SECURITY | LOG_NOTICE,
"ipfw: limit %d reached on entry %d\n",
limit_reached, f ? f->rulenum : -1);
}
/*
* IMPORTANT: the hash function for dynamic rules must be commutative
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
* in source and destination (ip,port), because rules are bidirectional
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* and we want to find both in the same bucket.
*/
static __inline int
hash_packet(struct ipfw_flow_id *id)
{
u_int32_t i;
i = (id->dst_ip) ^ (id->src_ip) ^ (id->dst_port) ^ (id->src_port);
i &= (curr_dyn_buckets - 1);
return i;
}
/**
* unlink a dynamic rule from a chain. prev is a pointer to
* the previous one, q is a pointer to the rule to delete,
* head is a pointer to the head of the queue.
* Modifies q and potentially also head.
*/
#define UNLINK_DYN_RULE(prev, head, q) { \
ipfw_dyn_rule *old_q = q; \
\
/* remove a refcount to the parent */ \
if (q->dyn_type == O_LIMIT) \
q->parent->count--; \
DEB(printf("-- unlink entry 0x%08x %d -> 0x%08x %d, %d left\n", \
(q->id.src_ip), (q->id.src_port), \
(q->id.dst_ip), (q->id.dst_port), dyn_count-1 ); ) \
if (prev != NULL) \
prev->next = q = q->next; \
else \
head = q = q->next; \
dyn_count--; \
free(old_q, M_IPFW); }
#define TIME_LEQ(a,b) ((int)((a)-(b)) <= 0)
/**
* Remove dynamic rules pointing to "rule", or all of them if rule == NULL.
*
* If keep_me == NULL, rules are deleted even if not expired,
* otherwise only expired rules are removed.
*
* The value of the second parameter is also used to point to identify
* a rule we absolutely do not want to remove (e.g. because we are
* holding a reference to it -- this is the case with O_LIMIT_PARENT
* rules). The pointer is only used for comparison, so any non-null
* value will do.
*/
static void
remove_dyn_rule(struct ip_fw *rule, ipfw_dyn_rule *keep_me)
{
static u_int32_t last_remove = 0;
#define FORCE (keep_me == NULL)
ipfw_dyn_rule *prev, *q;
int i, pass = 0, max_pass = 0;
if (ipfw_dyn_v == NULL || dyn_count == 0)
return;
/* do not expire more than once per second, it is useless */
if (!FORCE && last_remove == time_second)
return;
last_remove = time_second;
/*
* because O_LIMIT refer to parent rules, during the first pass only
* remove child and mark any pending LIMIT_PARENT, and remove
* them in a second pass.
*/
next_pass:
for (i = 0 ; i < curr_dyn_buckets ; i++) {
for (prev=NULL, q = ipfw_dyn_v[i] ; q ; ) {
/*
* Logic can become complex here, so we split tests.
*/
if (q == keep_me)
goto next;
if (rule != NULL && rule != q->rule)
goto next; /* not the one we are looking for */
if (q->dyn_type == O_LIMIT_PARENT) {
/*
* handle parent in the second pass,
* record we need one.
*/
max_pass = 1;
if (pass == 0)
goto next;
if (FORCE && q->count != 0 ) {
/* XXX should not happen! */
printf( "OUCH! cannot remove rule,"
" count %d\n", q->count);
}
} else {
if (!FORCE &&
!TIME_LEQ( q->expire, time_second ))
goto next;
}
UNLINK_DYN_RULE(prev, ipfw_dyn_v[i], q);
continue;
next:
prev=q;
q=q->next;
}
}
if (pass++ < max_pass)
goto next_pass;
}
/**
* lookup a dynamic rule.
*/
static ipfw_dyn_rule *
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
lookup_dyn_rule(struct ipfw_flow_id *pkt, int *match_direction,
struct tcphdr *tcp)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
/*
* stateful ipfw extensions.
* Lookup into dynamic session queue
*/
#define MATCH_REVERSE 0
#define MATCH_FORWARD 1
#define MATCH_NONE 2
#define MATCH_UNKNOWN 3
int i, dir = MATCH_NONE;
ipfw_dyn_rule *prev, *q=NULL;
if (ipfw_dyn_v == NULL)
goto done; /* not found */
i = hash_packet( pkt );
for (prev=NULL, q = ipfw_dyn_v[i] ; q != NULL ; ) {
if (q->dyn_type == O_LIMIT_PARENT)
goto next;
if (TIME_LEQ( q->expire, time_second)) { /* expire entry */
UNLINK_DYN_RULE(prev, ipfw_dyn_v[i], q);
continue;
}
if ( pkt->proto == q->id.proto) {
if (pkt->src_ip == q->id.src_ip &&
pkt->dst_ip == q->id.dst_ip &&
pkt->src_port == q->id.src_port &&
pkt->dst_port == q->id.dst_port ) {
dir = MATCH_FORWARD;
break;
}
if (pkt->src_ip == q->id.dst_ip &&
pkt->dst_ip == q->id.src_ip &&
pkt->src_port == q->id.dst_port &&
pkt->dst_port == q->id.src_port ) {
dir = MATCH_REVERSE;
break;
}
}
next:
prev = q;
q = q->next;
}
if (q == NULL)
goto done; /* q = NULL, not found */
if ( prev != NULL) { /* found and not in front */
prev->next = q->next;
q->next = ipfw_dyn_v[i];
ipfw_dyn_v[i] = q;
}
if (pkt->proto == IPPROTO_TCP) { /* update state according to flags */
u_char flags = pkt->flags & (TH_FIN|TH_SYN|TH_RST);
#define BOTH_SYN (TH_SYN | (TH_SYN << 8))
#define BOTH_FIN (TH_FIN | (TH_FIN << 8))
q->state |= (dir == MATCH_FORWARD ) ? flags : (flags << 8);
switch (q->state) {
case TH_SYN: /* opening */
q->expire = time_second + dyn_syn_lifetime;
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case BOTH_SYN: /* move to established */
case BOTH_SYN | TH_FIN : /* one side tries to close */
case BOTH_SYN | (TH_FIN << 8) :
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
if (tcp) {
#define _SEQ_GE(a,b) ((int)(a) - (int)(b) >= 0)
u_int32_t ack = ntohl(tcp->th_ack);
if (dir == MATCH_FORWARD) {
if (q->ack_fwd == 0 || _SEQ_GE(ack, q->ack_fwd))
q->ack_fwd = ack;
else { /* ignore out-of-sequence */
break;
}
} else {
if (q->ack_rev == 0 || _SEQ_GE(ack, q->ack_rev))
q->ack_rev = ack;
else { /* ignore out-of-sequence */
break;
}
}
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
q->expire = time_second + dyn_ack_lifetime;
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case BOTH_SYN | BOTH_FIN: /* both sides closed */
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (dyn_fin_lifetime >= dyn_keepalive_period)
dyn_fin_lifetime = dyn_keepalive_period - 1;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
q->expire = time_second + dyn_fin_lifetime;
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default:
#if 0
/*
* reset or some invalid combination, but can also
* occur if we use keep-state the wrong way.
*/
if ( (q->state & ((TH_RST << 8)|TH_RST)) == 0)
printf("invalid state: 0x%x\n", q->state);
#endif
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (dyn_rst_lifetime >= dyn_keepalive_period)
dyn_rst_lifetime = dyn_keepalive_period - 1;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
q->expire = time_second + dyn_rst_lifetime;
break;
}
} else if (pkt->proto == IPPROTO_UDP) {
q->expire = time_second + dyn_udp_lifetime;
} else {
/* other protocols */
q->expire = time_second + dyn_short_lifetime;
}
done:
if (match_direction)
*match_direction = dir;
return q;
}
static void
realloc_dynamic_table(void)
{
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
/*
* Try reallocation, make sure we have a power of 2 and do
* not allow more than 64k entries. In case of overflow,
* default to 1024.
*/
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (dyn_buckets > 65536)
dyn_buckets = 1024;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if ((dyn_buckets & (dyn_buckets-1)) != 0) { /* not a power of 2 */
dyn_buckets = curr_dyn_buckets; /* reset */
return;
}
curr_dyn_buckets = dyn_buckets;
if (ipfw_dyn_v != NULL)
free(ipfw_dyn_v, M_IPFW);
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
for (;;) {
ipfw_dyn_v = malloc(curr_dyn_buckets * sizeof(ipfw_dyn_rule *),
Several malloc() calls were passing the M_DONTWAIT flag which is an mbuf allocation flag. Use the correct M_NOWAIT malloc() flag. Fortunately, both were defined to 1, so this commit is a no-op.
2002-10-19 11:31:50 +00:00
M_IPFW, M_NOWAIT | M_ZERO);
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (ipfw_dyn_v != NULL || curr_dyn_buckets <= 2)
break;
curr_dyn_buckets /= 2;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/**
* Install state of type 'type' for a dynamic session.
* The hash table contains two type of rules:
* - regular rules (O_KEEP_STATE)
* - rules for sessions with limited number of sess per user
* (O_LIMIT). When they are created, the parent is
* increased by 1, and decreased on delete. In this case,
* the third parameter is the parent rule and not the chain.
* - "parent" rules for the above (O_LIMIT_PARENT).
*/
static ipfw_dyn_rule *
add_dyn_rule(struct ipfw_flow_id *id, u_int8_t dyn_type, struct ip_fw *rule)
{
ipfw_dyn_rule *r;
int i;
if (ipfw_dyn_v == NULL ||
(dyn_count == 0 && dyn_buckets != curr_dyn_buckets)) {
realloc_dynamic_table();
if (ipfw_dyn_v == NULL)
return NULL; /* failed ! */
}
i = hash_packet(id);
Several malloc() calls were passing the M_DONTWAIT flag which is an mbuf allocation flag. Use the correct M_NOWAIT malloc() flag. Fortunately, both were defined to 1, so this commit is a no-op.
2002-10-19 11:31:50 +00:00
r = malloc(sizeof *r, M_IPFW, M_NOWAIT | M_ZERO);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (r == NULL) {
printf ("sorry cannot allocate state\n");
return NULL;
}
/* increase refcount on parent, and set pointer */
if (dyn_type == O_LIMIT) {
ipfw_dyn_rule *parent = (ipfw_dyn_rule *)rule;
if ( parent->dyn_type != O_LIMIT_PARENT)
panic("invalid parent");
parent->count++;
r->parent = parent;
rule = parent->rule;
}
r->id = *id;
r->expire = time_second + dyn_syn_lifetime;
r->rule = rule;
r->dyn_type = dyn_type;
r->pcnt = r->bcnt = 0;
r->count = 0;
r->bucket = i;
r->next = ipfw_dyn_v[i];
ipfw_dyn_v[i] = r;
dyn_count++;
DEB(printf("-- add dyn entry ty %d 0x%08x %d -> 0x%08x %d, total %d\n",
dyn_type,
(r->id.src_ip), (r->id.src_port),
(r->id.dst_ip), (r->id.dst_port),
dyn_count ); )
return r;
}
/**
* lookup dynamic parent rule using pkt and rule as search keys.
* If the lookup fails, then install one.
*/
static ipfw_dyn_rule *
lookup_dyn_parent(struct ipfw_flow_id *pkt, struct ip_fw *rule)
{
ipfw_dyn_rule *q;
int i;
if (ipfw_dyn_v) {
i = hash_packet( pkt );
for (q = ipfw_dyn_v[i] ; q != NULL ; q=q->next)
if (q->dyn_type == O_LIMIT_PARENT &&
rule== q->rule &&
pkt->proto == q->id.proto &&
pkt->src_ip == q->id.src_ip &&
pkt->dst_ip == q->id.dst_ip &&
pkt->src_port == q->id.src_port &&
pkt->dst_port == q->id.dst_port) {
q->expire = time_second + dyn_short_lifetime;
DEB(printf("lookup_dyn_parent found 0x%p\n",q);)
return q;
}
}
return add_dyn_rule(pkt, O_LIMIT_PARENT, rule);
}
/**
* Install dynamic state for rule type cmd->o.opcode
*
* Returns 1 (failure) if state is not installed because of errors or because
* session limitations are enforced.
*/
static int
install_state(struct ip_fw *rule, ipfw_insn_limit *cmd,
struct ip_fw_args *args)
{
static int last_log;
ipfw_dyn_rule *q;
DEB(printf("-- install state type %d 0x%08x %u -> 0x%08x %u\n",
cmd->o.opcode,
(args->f_id.src_ip), (args->f_id.src_port),
(args->f_id.dst_ip), (args->f_id.dst_port) );)
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
q = lookup_dyn_rule(&args->f_id, NULL, NULL);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (q != NULL) { /* should never occur */
if (last_log != time_second) {
last_log = time_second;
printf(" install_state: entry already present, done\n");
}
return 0;
}
if (dyn_count >= dyn_max)
/*
* Run out of slots, try to remove any expired rule.
*/
remove_dyn_rule(NULL, (ipfw_dyn_rule *)1);
if (dyn_count >= dyn_max) {
if (last_log != time_second) {
last_log = time_second;
printf("install_state: Too many dynamic rules\n");
}
return 1; /* cannot install, notify caller */
}
switch (cmd->o.opcode) {
case O_KEEP_STATE: /* bidir rule */
add_dyn_rule(&args->f_id, O_KEEP_STATE, rule);
break;
case O_LIMIT: /* limit number of sessions */
{
u_int16_t limit_mask = cmd->limit_mask;
struct ipfw_flow_id id;
ipfw_dyn_rule *parent;
DEB(printf("installing dyn-limit rule %d\n", cmd->conn_limit);)
id.dst_ip = id.src_ip = 0;
id.dst_port = id.src_port = 0;
id.proto = args->f_id.proto;
if (limit_mask & DYN_SRC_ADDR)
id.src_ip = args->f_id.src_ip;
if (limit_mask & DYN_DST_ADDR)
id.dst_ip = args->f_id.dst_ip;
if (limit_mask & DYN_SRC_PORT)
id.src_port = args->f_id.src_port;
if (limit_mask & DYN_DST_PORT)
id.dst_port = args->f_id.dst_port;
parent = lookup_dyn_parent(&id, rule);
if (parent == NULL) {
printf("add parent failed\n");
return 1;
}
if (parent->count >= cmd->conn_limit) {
/*
* See if we can remove some expired rule.
*/
remove_dyn_rule(rule, parent);
if (parent->count >= cmd->conn_limit) {
if (fw_verbose && last_log != time_second) {
last_log = time_second;
Lower a priority of "session drop" messages. Requested by: Eugene Grosbein <eugen@kuzbass.ru> MFC after: 3 days
2002-10-29 08:53:14 +00:00
log(LOG_SECURITY | LOG_DEBUG,
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
"drop session, too many entries\n");
}
return 1;
}
}
add_dyn_rule(&args->f_id, O_LIMIT, (struct ip_fw *)parent);
}
break;
default:
printf("unknown dynamic rule type %u\n", cmd->o.opcode);
return 1;
}
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
lookup_dyn_rule(&args->f_id, NULL, NULL); /* XXX just set lifetime */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return 0;
}
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
/*
* Transmit a TCP packet, containing either a RST or a keepalive.
* When flags & TH_RST, we are sending a RST packet, because of a
* "reset" action matched the packet.
* Otherwise we are sending a keepalive, and flags & TH_
*/
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
static void
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
send_pkt(struct ipfw_flow_id *id, u_int32_t seq, u_int32_t ack, int flags)
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
{
struct mbuf *m;
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
struct ip *ip;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
struct tcphdr *tcp;
struct route sro; /* fake route */
MGETHDR(m, M_DONTWAIT, MT_HEADER);
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
if (m == 0)
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
return;
m->m_pkthdr.rcvif = (struct ifnet *)0;
m->m_pkthdr.len = m->m_len = sizeof(struct ip) + sizeof(struct tcphdr);
m->m_data += max_linkhdr;
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
ip = mtod(m, struct ip *);
bzero(ip, m->m_len);
tcp = (struct tcphdr *)(ip + 1); /* no IP options */
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
ip->ip_p = IPPROTO_TCP;
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
tcp->th_off = 5;
/*
* Assume we are sending a RST (or a keepalive in the reverse
* direction), swap src and destination addresses and ports.
*/
ip->ip_src.s_addr = htonl(id->dst_ip);
ip->ip_dst.s_addr = htonl(id->src_ip);
tcp->th_sport = htons(id->dst_port);
tcp->th_dport = htons(id->src_port);
if (flags & TH_RST) { /* we are sending a RST */
if (flags & TH_ACK) {
tcp->th_seq = htonl(ack);
tcp->th_ack = htonl(0);
tcp->th_flags = TH_RST;
} else {
if (flags & TH_SYN)
seq++;
tcp->th_seq = htonl(0);
tcp->th_ack = htonl(seq);
tcp->th_flags = TH_RST | TH_ACK;
}
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
} else {
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
/*
* We are sending a keepalive. flags & TH_SYN determines
* the direction, forward if set, reverse if clear.
* NOTE: seq and ack are always assumed to be correct
* as set by the caller. This may be confusing...
*/
if (flags & TH_SYN) {
/*
* we have to rewrite the correct addresses!
*/
ip->ip_dst.s_addr = htonl(id->dst_ip);
ip->ip_src.s_addr = htonl(id->src_ip);
tcp->th_dport = htons(id->dst_port);
tcp->th_sport = htons(id->src_port);
}
tcp->th_seq = htonl(seq);
tcp->th_ack = htonl(ack);
tcp->th_flags = TH_ACK;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
}
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
/*
* set ip_len to the payload size so we can compute
* the tcp checksum on the pseudoheader
* XXX check this, could save a couple of words ?
*/
ip->ip_len = htons(sizeof(struct tcphdr));
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
tcp->th_sum = in_cksum(m, m->m_pkthdr.len);
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
/*
* now fill fields left out earlier
*/
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
ip->ip_ttl = ip_defttl;
ip->ip_len = m->m_pkthdr.len;
bzero (&sro, sizeof (sro));
ip_rtaddr(ip->ip_dst, &sro);
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
m->m_flags |= M_SKIP_FIREWALL;
Replace aux mbufs with packet tags: o instead of a list of mbufs use a list of m_tag structures a la openbsd o for netgraph et. al. extend the stock openbsd m_tag to include a 32-bit ABI/module number cookie o for openbsd compatibility define a well-known cookie MTAG_ABI_COMPAT and use this in defining openbsd-compatible m_tag_find and m_tag_get routines o rewrite KAME use of aux mbufs in terms of packet tags o eliminate the most heavily used aux mbufs by adding an additional struct inpcb parameter to ip_output and ip6_output to allow the IPsec code to locate the security policy to apply to outbound packets o bump __FreeBSD_version so code can be conditionalized o fixup ipfilter's call to ip_output based on __FreeBSD_version Reviewed by: julian, luigi (silent), -arch, -net, darren Approved by: julian, silence from everyone else Obtained from: openbsd (mostly) MFC after: 1 month
2002-10-16 01:54:46 +00:00
ip_output(m, NULL, &sro, 0, NULL, NULL);
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
if (sro.ro_rt)
RTFREE(sro.ro_rt);
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* sends a reject message, consuming the mbuf passed as an argument.
*/
static void
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
send_reject(struct ip_fw_args *args, int code, int offset, int ip_len)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
remove spurious printf
2002-08-13 19:13:23 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (code != ICMP_REJECT_RST) /* Send an ICMP unreach */
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
icmp_error(args->m, ICMP_UNREACH, code, 0L, 0);
else if (offset == 0 && args->f_id.proto == IPPROTO_TCP) {
struct tcphdr *const tcp =
L3HDR(struct tcphdr, mtod(args->m, struct ip *));
if ( (tcp->th_flags & TH_RST) == 0)
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
send_pkt(&(args->f_id), ntohl(tcp->th_seq),
ntohl(tcp->th_ack),
tcp->th_flags | TH_RST);
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
m_freem(args->m);
} else
m_freem(args->m);
args->m = NULL;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
/**
*
* Given an ip_fw *, lookup_next_rule will return a pointer
* to the next rule, which can be either the jump
* target (for skipto instructions) or the next one in the list (in
* all other cases including a missing jump target).
* The result is also written in the "next_rule" field of the rule.
* Backward jumps are not allowed, so start looking from the next
* rule...
*
* This never returns NULL -- in case we do not have an exact match,
* the next rule is returned. When the ruleset is changed,
* pointers are flushed so we are always correct.
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
*/
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
static struct ip_fw *
lookup_next_rule(struct ip_fw *me)
{
struct ip_fw *rule = NULL;
ipfw_insn *cmd;
/* look for action, in case it is a skipto */
cmd = ACTION_PTR(me);
if ( cmd->opcode == O_SKIPTO )
for (rule = me->next; rule ; rule = rule->next)
if (rule->rulenum >= cmd->arg1)
break;
if (rule == NULL) /* failure or not a skipto */
rule = me->next;
me->next_rule = rule;
return rule;
}
/*
* The main check routine for the firewall.
*
* All arguments are in args so we can modify them and return them
* back to the caller.
*
* Parameters:
*
* args->m (in/out) The packet; we set to NULL when/if we nuke it.
* Starts with the IP header.
* args->eh (in) Mac header if present, or NULL for layer3 packet.
* args->oif Outgoing interface, or NULL if packet is incoming.
* The incoming interface is in the mbuf. (in)
* args->divert_rule (in/out)
* Skip up to the first rule past this rule number;
* upon return, non-zero port number for divert or tee.
*
* args->rule Pointer to the last matching rule (in/out)
* args->next_hop Socket we are forwarding to (out).
* args->f_id Addresses grabbed from the packet (out)
*
* Return value:
*
* IP_FW_PORT_DENY_FLAG the packet must be dropped.
* 0 The packet is to be accepted and routed normally OR
* the packet was denied/rejected and has been dropped;
* in the latter case, *m is equal to NULL upon return.
* port Divert the packet to port, with these caveats:
*
* - If IP_FW_PORT_TEE_FLAG is set, tee the packet instead
* of diverting it (ie, 'ipfw tee').
*
* - If IP_FW_PORT_DYNT_FLAG is set, interpret the lower
* 16 bits as a dummynet pipe number instead of diverting
*/
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
static int
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ipfw_chk(struct ip_fw_args *args)
{
/*
* Local variables hold state during the processing of a packet.
*
* IMPORTANT NOTE: to speed up the processing of rules, there
* are some assumption on the values of the variables, which
* are documented here. Should you change them, please check
* the implementation of the various instructions to make sure
* that they still work.
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
*
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* args->eh The MAC header. It is non-null for a layer2
* packet, it is NULL for a layer-3 packet.
*
* m | args->m Pointer to the mbuf, as received from the caller.
* It may change if ipfw_chk() does an m_pullup, or if it
* consumes the packet because it calls send_reject().
* XXX This has to change, so that ipfw_chk() never modifies
* or consumes the buffer.
* ip is simply an alias of the value of m, and it is kept
* in sync with it (the packet is supposed to start with
* the ip header).
*/
struct mbuf *m = args->m;
struct ip *ip = mtod(m, struct ip *);
/*
* oif | args->oif If NULL, ipfw_chk has been called on the
* inbound path (ether_input, bdg_forward, ip_input).
* If non-NULL, ipfw_chk has been called on the outbound path
* (ether_output, ip_output).
*/
struct ifnet *oif = args->oif;
struct ip_fw *f = NULL; /* matching rule */
int retval = 0;
/*
* hlen The length of the IPv4 header.
* hlen >0 means we have an IPv4 packet.
*/
u_int hlen = 0; /* hlen >0 means we have an IP pkt */
/*
* offset The offset of a fragment. offset != 0 means that
* we have a fragment at this offset of an IPv4 packet.
* offset == 0 means that (if this is an IPv4 packet)
* this is the first or only fragment.
*/
u_short offset = 0;
/*
* Local copies of addresses. They are only valid if we have
* an IP packet.
*
* proto The protocol. Set to 0 for non-ip packets,
* or to the protocol read from the packet otherwise.
* proto != 0 means that we have an IPv4 packet.
*
* src_port, dst_port port numbers, in HOST format. Only
* valid for TCP and UDP packets.
*
* src_ip, dst_ip ip addresses, in NETWORK format.
* Only valid for IPv4 packets.
*/
u_int8_t proto;
u_int16_t src_port = 0, dst_port = 0; /* NOTE: host format */
struct in_addr src_ip, dst_ip; /* NOTE: network format */
u_int16_t ip_len=0;
int dyn_dir = MATCH_UNKNOWN;
ipfw_dyn_rule *q = NULL;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (m->m_flags & M_SKIP_FIREWALL)
return 0; /* accept */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* dyn_dir = MATCH_UNKNOWN when rules unchecked,
* MATCH_NONE when checked and not matched (q = NULL),
* MATCH_FORWARD or MATCH_REVERSE otherwise (q != NULL)
*/
if (args->eh == NULL || /* layer 3 packet */
( m->m_pkthdr.len >= sizeof(struct ip) &&
ntohs(args->eh->ether_type) == ETHERTYPE_IP))
hlen = ip->ip_hl << 2;
/*
* Collect parameters into local variables for faster matching.
*/
if (hlen == 0) { /* do not grab addresses for non-ip pkts */
proto = args->f_id.proto = 0; /* mark f_id invalid */
goto after_ip_checks;
}
proto = args->f_id.proto = ip->ip_p;
src_ip = ip->ip_src;
dst_ip = ip->ip_dst;
if (args->eh != NULL) { /* layer 2 packets are as on the wire */
offset = ntohs(ip->ip_off) & IP_OFFMASK;
ip_len = ntohs(ip->ip_len);
} else {
offset = ip->ip_off & IP_OFFMASK;
ip_len = ip->ip_len;
}
#define PULLUP_TO(len) \
do { \
if ((m)->m_len < (len)) { \
args->m = m = m_pullup(m, (len)); \
if (m == 0) \
goto pullup_failed; \
ip = mtod(m, struct ip *); \
} \
} while (0)
if (offset == 0) {
switch (proto) {
case IPPROTO_TCP:
{
struct tcphdr *tcp;
PULLUP_TO(hlen + sizeof(struct tcphdr));
tcp = L3HDR(struct tcphdr, ip);
dst_port = tcp->th_dport;
src_port = tcp->th_sport;
args->f_id.flags = tcp->th_flags;
}
break;
case IPPROTO_UDP:
{
struct udphdr *udp;
PULLUP_TO(hlen + sizeof(struct udphdr));
udp = L3HDR(struct udphdr, ip);
dst_port = udp->uh_dport;
src_port = udp->uh_sport;
}
break;
case IPPROTO_ICMP:
PULLUP_TO(hlen + 4); /* type, code and checksum. */
args->f_id.flags = L3HDR(struct icmp, ip)->icmp_type;
break;
default:
break;
}
#undef PULLUP_TO
}
args->f_id.src_ip = ntohl(src_ip.s_addr);
args->f_id.dst_ip = ntohl(dst_ip.s_addr);
args->f_id.src_port = src_port = ntohs(src_port);
args->f_id.dst_port = dst_port = ntohs(dst_port);
after_ip_checks:
if (args->rule) {
/*
* Packet has already been tagged. Look for the next rule
* to restart processing.
*
* If fw_one_pass != 0 then just accept it.
* XXX should not happen here, but optimized out in
* the caller.
*/
if (fw_one_pass)
return 0;
f = args->rule->next_rule;
if (f == NULL)
f = lookup_next_rule(args->rule);
} else {
/*
* Find the starting rule. It can be either the first
* one, or the one after divert_rule if asked so.
*/
int skipto = args->divert_rule;
f = layer3_chain;
if (args->eh == NULL && skipto != 0) {
if (skipto >= IPFW_DEFAULT_RULE)
return(IP_FW_PORT_DENY_FLAG); /* invalid */
while (f && f->rulenum <= skipto)
f = f->next;
if (f == NULL) /* drop packet */
return(IP_FW_PORT_DENY_FLAG);
}
}
args->divert_rule = 0; /* reset to avoid confusion later */
/*
* Now scan the rules, and parse microinstructions for each rule.
*/
for (; f; f = f->next) {
int l, cmdlen;
ipfw_insn *cmd;
int skip_or; /* skip rest of OR block */
again:
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
if (set_disable & (1 << f->set) )
continue;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
skip_or = 0;
for (l = f->cmd_len, cmd = f->cmd ; l > 0 ;
l -= cmdlen, cmd += cmdlen) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
int match;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* check_body is a jump target used when we find a
* CHECK_STATE, and need to jump to the body of
* the target rule.
*/
check_body:
cmdlen = F_LEN(cmd);
/*
* An OR block (insn_1 || .. || insn_n) has the
* F_OR bit set in all but the last instruction.
* The first match will set "skip_or", and cause
* the following instructions to be skipped until
* past the one with the F_OR bit clear.
*/
if (skip_or) { /* skip this instruction */
if ((cmd->len & F_OR) == 0)
skip_or = 0; /* next one is good */
continue;
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = 0; /* set to 1 if we succeed */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
switch (cmd->opcode) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
/*
* The first set of opcodes compares the packet's
* fields with some pattern, setting 'match' if a
* match is found. At the end of the loop there is
* logic to deal with F_NOT and F_OR flags associated
* with the opcode.
*/
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_NOP:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = 1;
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_FORWARD_MAC:
printf("ipfw: opcode %d unimplemented\n",
cmd->opcode);
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_GID:
case O_UID:
/*
* We only check offset == 0 && proto != 0,
* as this ensures that we have an IPv4
* packet with the ports info.
*/
if (offset!=0)
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
struct inpcbinfo *pi;
int wildcard;
struct inpcb *pcb;
if (proto == IPPROTO_TCP) {
wildcard = 0;
pi = &tcbinfo;
} else if (proto == IPPROTO_UDP) {
wildcard = 1;
pi = &udbinfo;
} else
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
pcb = (oif) ?
in_pcblookup_hash(pi,
dst_ip, htons(dst_port),
src_ip, htons(src_port),
wildcard, oif) :
in_pcblookup_hash(pi,
src_ip, htons(src_port),
dst_ip, htons(dst_port),
wildcard, NULL);
if (pcb == NULL || pcb->inp_socket == NULL)
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
#if __FreeBSD_version < 500034
#define socheckuid(a,b) ((a)->so_cred->cr_uid == (b))
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#endif
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if (cmd->opcode == O_UID) {
match =
socheckuid(pcb->inp_socket,
(uid_t)((ipfw_insn_u32 *)cmd)->d[0]);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
} else {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = groupmember(
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
(uid_t)((ipfw_insn_u32 *)cmd)->d[0],
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
pcb->inp_socket->so_cred);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_RECV:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = iface_match(m->m_pkthdr.rcvif,
(ipfw_insn_if *)cmd);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_XMIT:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = iface_match(oif, (ipfw_insn_if *)cmd);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_VIA:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = iface_match(oif ? oif :
m->m_pkthdr.rcvif, (ipfw_insn_if *)cmd);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_MACADDR2:
if (args->eh != NULL) { /* have MAC header */
u_int32_t *want = (u_int32_t *)
((ipfw_insn_mac *)cmd)->addr;
u_int32_t *mask = (u_int32_t *)
((ipfw_insn_mac *)cmd)->mask;
u_int32_t *hdr = (u_int32_t *)args->eh;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match =
( want[0] == (hdr[0] & mask[0]) &&
want[1] == (hdr[1] & mask[1]) &&
want[2] == (hdr[2] & mask[2]) );
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_MAC_TYPE:
if (args->eh != NULL) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
u_int16_t t =
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ntohs(args->eh->ether_type);
u_int16_t *p =
((ipfw_insn_u16 *)cmd)->ports;
int i;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
for (i = cmdlen - 1; !match && i>0;
i--, p += 2)
match = (t>=p[0] && t<=p[1]);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_FRAG:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 && offset != 0);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IN: /* "out" is "not in" */
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (oif == NULL);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_LAYER2:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (args->eh != NULL);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_PROTO:
/*
* We do not allow an arg of 0 so the
* check of "proto" only suffices.
*/
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (proto == cmd->arg1);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_SRC:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 &&
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
((ipfw_insn_ip *)cmd)->addr.s_addr ==
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
src_ip.s_addr);
break;
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_SRC_MASK:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 &&
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
((ipfw_insn_ip *)cmd)->addr.s_addr ==
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
(src_ip.s_addr &
((ipfw_insn_ip *)cmd)->mask.s_addr));
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_SRC_ME:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if (hlen > 0) {
struct ifnet *tif;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
INADDR_TO_IFP(src_ip, tif);
match = (tif != NULL);
}
break;
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_DST_SET:
case O_IP_SRC_SET:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if (hlen > 0) {
u_int32_t *d = (u_int32_t *)(cmd+1);
u_int32_t addr =
cmd->opcode == O_IP_DST_SET ?
src and dst address were erroneously swapped in SRC_SET and DST_SET commands. Use the correct one. Also affects ipfw2 in -stable.
2002-10-24 18:01:53 +00:00
args->f_id.dst_ip :
args->f_id.src_ip;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if (addr < d[0])
break;
addr -= d[0]; /* subtract base */
match = (addr < cmd->arg1) &&
( d[ 1 + (addr>>5)] &
(1<<(addr & 0x1f)) );
}
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_DST:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 &&
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
((ipfw_insn_ip *)cmd)->addr.s_addr ==
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
dst_ip.s_addr);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_DST_MASK:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0) &&
(((ipfw_insn_ip *)cmd)->addr.s_addr ==
(dst_ip.s_addr &
((ipfw_insn_ip *)cmd)->mask.s_addr));
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_DST_ME:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if (hlen > 0) {
struct ifnet *tif;
INADDR_TO_IFP(dst_ip, tif);
match = (tif != NULL);
}
break;
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IP_SRCPORT:
case O_IP_DSTPORT:
/*
* offset == 0 && proto != 0 is enough
* to guarantee that we have an IPv4
* packet with port info.
*/
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if ((proto==IPPROTO_UDP || proto==IPPROTO_TCP)
&& offset == 0) {
u_int16_t x =
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
(cmd->opcode == O_IP_SRCPORT) ?
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
src_port : dst_port ;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
u_int16_t *p =
((ipfw_insn_u16 *)cmd)->ports;
int i;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
for (i = cmdlen - 1; !match && i>0;
i--, p += 2)
match = (x>=p[0] && x<=p[1]);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_ICMPTYPE:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (offset == 0 && proto==IPPROTO_ICMP &&
icmptype_match(ip, (ipfw_insn_u32 *)cmd) );
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPOPT:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 && ipopts_match(ip, cmd) );
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPVER:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 && cmd->arg1 == ip->ip_v);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPTTL:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 && cmd->arg1 == ip->ip_ttl);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPID:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 &&
cmd->arg1 == ntohs(ip->ip_id));
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPLEN:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 && cmd->arg1 == ip_len);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_IPPRECEDENCE:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 &&
(cmd->arg1 == (ip->ip_tos & 0xe0)) );
break;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPTOS:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (hlen > 0 &&
flags_match(cmd, ip->ip_tos));
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_TCPFLAGS:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (proto == IPPROTO_TCP && offset == 0 &&
flags_match(cmd,
L3HDR(struct tcphdr,ip)->th_flags));
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_TCPOPTS:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (proto == IPPROTO_TCP && offset == 0 &&
tcpopts_match(ip, cmd));
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_TCPSEQ:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (proto == IPPROTO_TCP && offset == 0 &&
((ipfw_insn_u32 *)cmd)->d[0] ==
L3HDR(struct tcphdr,ip)->th_seq);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_TCPACK:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (proto == IPPROTO_TCP && offset == 0 &&
((ipfw_insn_u32 *)cmd)->d[0] ==
L3HDR(struct tcphdr,ip)->th_ack);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_TCPWIN:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (proto == IPPROTO_TCP && offset == 0 &&
cmd->arg1 ==
L3HDR(struct tcphdr,ip)->th_win);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_ESTAB:
/* reject packets which have SYN only */
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
/* XXX should i also check for TH_ACK ? */
match = (proto == IPPROTO_TCP && offset == 0 &&
(L3HDR(struct tcphdr,ip)->th_flags &
(TH_RST | TH_ACK | TH_SYN)) != TH_SYN);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_LOG:
Only log things net.inet.ip.fw.verbose is set
2002-07-24 02:41:19 +00:00
if (fw_verbose)
ipfw_log(f, hlen, args->eh, m, oif);
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = 1;
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_PROB:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = (random()<((ipfw_insn_u32 *)cmd)->d[0]);
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
/*
* The second set of opcodes represents 'actions',
* i.e. the terminal part of a rule once the packet
* matches all previous patterns.
* Typically there is only one action for each rule,
* and the opcode is stored at the end of the rule
* (but there are exceptions -- see below).
*
* In general, here we set retval and terminate the
* outer loop (would be a 'break 3' in some language,
* but we need to do a 'goto done').
*
* Exceptions:
* O_COUNT and O_SKIPTO actions:
* instead of terminating, we jump to the next rule
* ('goto next_rule', equivalent to a 'break 2'),
* or to the SKIPTO target ('goto again' after
* having set f, cmd and l), respectively.
*
* O_LIMIT and O_KEEP_STATE: these opcodes are
* not real 'actions', and are stored right
* before the 'action' part of the rule.
* These opcodes try to install an entry in the
* state tables; if successful, we continue with
* the next opcode (match=1; break;), otherwise
* the packet * must be dropped
* ('goto done' after setting retval);
*
* O_PROBE_STATE and O_CHECK_STATE: these opcodes
* cause a lookup of the state table, and a jump
* to the 'action' part of the parent rule
* ('goto check_body') if an entry is found, or
* (CHECK_STATE only) a jump to the next rule if
* the entry is not found ('goto next_rule').
* The result of the lookup is cached to make
* further instances of these opcodes are
* effectively NOPs.
*/
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_LIMIT:
case O_KEEP_STATE:
if (install_state(f,
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
(ipfw_insn_limit *)cmd, args)) {
retval = IP_FW_PORT_DENY_FLAG;
goto done; /* error/limit violation */
}
match = 1;
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_PROBE_STATE:
case O_CHECK_STATE:
/*
* dynamic rules are checked at the first
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
* keep-state or check-state occurrence,
* with the result being stored in dyn_dir.
* The compiler introduces a PROBE_STATE
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* instruction for us when we have a
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
* KEEP_STATE (because PROBE_STATE needs
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* to be run first).
*/
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if (dyn_dir == MATCH_UNKNOWN &&
(q = lookup_dyn_rule(&args->f_id,
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
&dyn_dir, proto == IPPROTO_TCP ?
L3HDR(struct tcphdr, ip) : NULL))
!= NULL) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
/*
* Found dynamic entry, update stats
* and jump to the 'action' part of
* the parent rule.
*/
q->pcnt++;
q->bcnt += ip_len;
f = q->rule;
cmd = ACTION_PTR(f);
l = f->cmd_len - f->act_ofs;
goto check_body;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
/*
* Dynamic entry not found. If CHECK_STATE,
* skip to next rule, if PROBE_STATE just
* ignore and continue with next opcode.
*/
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (cmd->opcode == O_CHECK_STATE)
goto next_rule;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
match = 1;
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_ACCEPT:
retval = 0; /* accept */
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
goto done;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_PIPE:
case O_QUEUE:
args->rule = f; /* report matching rule */
retval = cmd->arg1 | IP_FW_PORT_DYNT_FLAG;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
goto done;
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_DIVERT:
case O_TEE:
if (args->eh) /* not on layer 2 */
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
args->divert_rule = f->rulenum;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
retval = (cmd->opcode == O_DIVERT) ?
cmd->arg1 :
cmd->arg1 | IP_FW_PORT_TEE_FLAG;
goto done;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_COUNT:
case O_SKIPTO:
f->pcnt++; /* update stats */
f->bcnt += ip_len;
f->timestamp = time_second;
if (cmd->opcode == O_COUNT)
goto next_rule;
/* handle skipto */
if (f->next_rule == NULL)
lookup_next_rule(f);
f = f->next_rule;
goto again;
case O_REJECT:
/*
* Drop the packet and send a reject notice
* if the packet is not ICMP (or is an ICMP
* query), and it is not multicast/broadcast.
*/
if (hlen > 0 &&
(proto != IPPROTO_ICMP ||
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
is_icmp_query(ip)) &&
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
!(m->m_flags & (M_BCAST|M_MCAST)) &&
!IN_MULTICAST(dst_ip.s_addr)) {
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
send_reject(args, cmd->arg1,
offset,ip_len);
m = args->m;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
/* FALLTHROUGH */
case O_DENY:
retval = IP_FW_PORT_DENY_FLAG;
goto done;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_FORWARD_IP:
if (args->eh) /* not valid on layer2 pkts */
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
break;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (!q || dyn_dir == MATCH_FORWARD)
args->next_hop =
&((ipfw_insn_sa *)cmd)->sa;
retval = 0;
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
goto done;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default:
panic("-- unknown opcode %d\n", cmd->opcode);
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
} /* end of switch() on opcodes */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
if (cmd->len & F_NOT)
match = !match;
if (match) {
if (cmd->len & F_OR)
skip_or = 1;
} else {
if (!(cmd->len & F_OR)) /* not an OR block, */
break; /* try next rule */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
} /* end of inner for, scan opcodes */
Fix warning. Reviewed by: luigi
2002-06-28 08:36:26 +00:00
next_rule:; /* try next rule */
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
} /* end of outer for, scan rules */
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
printf("+++ ipfw: ouch!, skip past end of rules, denying packet\n");
return(IP_FW_PORT_DENY_FLAG);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
done:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/* Update statistics */
f->pcnt++;
f->bcnt += ip_len;
f->timestamp = time_second;
return retval;
pullup_failed:
if (fw_verbose)
printf("pullup failed\n");
return(IP_FW_PORT_DENY_FLAG);
}
/*
* When a rule is added/deleted, clear the next_rule pointers in all rules.
* These will be reconstructed on the fly as packets are matched.
* Must be called at splimp().
*/
static void
flush_rule_ptrs(void)
{
struct ip_fw *rule;
for (rule = layer3_chain; rule; rule = rule->next)
rule->next_rule = NULL;
}
/*
* When pipes/queues are deleted, clear the "pipe_ptr" pointer to a given
* pipe/queue, or to all of them (match == NULL).
* Must be called at splimp().
*/
void
flush_pipe_ptrs(struct dn_flow_set *match)
{
struct ip_fw *rule;
for (rule = layer3_chain; rule; rule = rule->next) {
ipfw_insn_pipe *cmd = (ipfw_insn_pipe *)ACTION_PTR(rule);
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (cmd->o.opcode != O_PIPE && cmd->o.opcode != O_QUEUE)
continue;
if (match == NULL || cmd->pipe_ptr == match)
cmd->pipe_ptr = NULL;
}
}
/*
* Add a new rule to the list. Copy the rule into a malloc'ed area, then
* possibly create a rule number and add the rule to the list.
* Update the rule_number in the input struct so the caller knows it as well.
*/
static int
add_rule(struct ip_fw **head, struct ip_fw *input_rule)
{
struct ip_fw *rule, *f, *prev;
int s;
int l = RULESIZE(input_rule);
if (*head == NULL && input_rule->rulenum != IPFW_DEFAULT_RULE)
return (EINVAL);
Several malloc() calls were passing the M_DONTWAIT flag which is an mbuf allocation flag. Use the correct M_NOWAIT malloc() flag. Fortunately, both were defined to 1, so this commit is a no-op.
2002-10-19 11:31:50 +00:00
rule = malloc(l, M_IPFW, M_NOWAIT | M_ZERO);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (rule == NULL)
return (ENOSPC);
bcopy(input_rule, rule, l);
rule->next = NULL;
rule->next_rule = NULL;
rule->pcnt = 0;
rule->bcnt = 0;
rule->timestamp = 0;
s = splimp();
if (*head == NULL) { /* default rule */
*head = rule;
goto done;
}
/*
* If rulenum is 0, find highest numbered rule before the
* default rule, and add autoinc_step
*/
if (autoinc_step < 1)
autoinc_step = 1;
else if (autoinc_step > 1000)
autoinc_step = 1000;
if (rule->rulenum == 0) {
/*
* locate the highest numbered rule before default
*/
for (f = *head; f; f = f->next) {
if (f->rulenum == IPFW_DEFAULT_RULE)
break;
rule->rulenum = f->rulenum;
}
if (rule->rulenum < IPFW_DEFAULT_RULE - autoinc_step)
rule->rulenum += autoinc_step;
input_rule->rulenum = rule->rulenum;
}
/*
* Now insert the new rule in the right place in the sorted list.
*/
for (prev = NULL, f = *head; f; prev = f, f = f->next) {
if (f->rulenum > rule->rulenum) { /* found the location */
if (prev) {
rule->next = f;
prev->next = rule;
} else { /* head insert */
rule->next = *head;
*head = rule;
}
break;
}
}
flush_rule_ptrs();
done:
static_count++;
static_len += l;
splx(s);
DEB(printf("++ installed rule %d, static count now %d\n",
rule->rulenum, static_count);)
return (0);
}
/**
* Free storage associated with a static rule (including derived
* dynamic rules).
* The caller is in charge of clearing rule pointers to avoid
* dangling pointers.
* @return a pointer to the next entry.
* Arguments are not checked, so they better be correct.
* Must be called at splimp().
*/
static struct ip_fw *
delete_rule(struct ip_fw **head, struct ip_fw *prev, struct ip_fw *rule)
{
struct ip_fw *n;
int l = RULESIZE(rule);
n = rule->next;
remove_dyn_rule(rule, NULL /* force removal */);
if (prev == NULL)
*head = n;
else
prev->next = n;
static_count--;
static_len -= l;
if (DUMMYNET_LOADED)
ip_dn_ruledel_ptr(rule);
free(rule, M_IPFW);
return n;
}
/*
* Deletes all rules from a chain (including the default rule
* if the second argument is set).
* Must be called at splimp().
*/
static void
free_chain(struct ip_fw **chain, int kill_default)
{
struct ip_fw *rule;
flush_rule_ptrs(); /* more efficient to do outside the loop */
while ( (rule = *chain) != NULL &&
(kill_default || rule->rulenum != IPFW_DEFAULT_RULE) )
delete_rule(chain, NULL, rule);
}
/**
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
* Remove all rules with given number, and also do set manipulation.
*
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
* The argument is an u_int32_t. The low 16 bit are the rule or set number,
* the next 8 bits are the new set, the top 8 bits are the command:
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
*
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
* 0 delete rules with given number
* 1 delete rules with given set number
* 2 move rules with given number to new set
* 3 move rules with given set number to new set
* 4 swap sets with given numbers
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
*/
static int
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
del_entry(struct ip_fw **chain, u_int32_t arg)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
{
struct ip_fw *prev, *rule;
int s;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
u_int16_t rulenum;
u_int8_t cmd, new_set;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
rulenum = arg & 0xffff;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
cmd = (arg >> 24) & 0xff;
new_set = (arg >> 16) & 0xff;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (cmd > 4)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return EINVAL;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (new_set > 30)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return EINVAL;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (cmd == 0 || cmd == 2) {
if (rulenum == IPFW_DEFAULT_RULE)
return EINVAL;
} else {
if (rulenum > 30)
return EINVAL;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
}
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
switch (cmd) {
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
case 0: /* delete rules with given number */
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
/*
* locate first rule to delete
*/
for (prev = NULL, rule = *chain;
rule && rule->rulenum < rulenum;
prev = rule, rule = rule->next)
;
if (rule->rulenum != rulenum)
return EINVAL;
s = splimp(); /* no access to rules while removing */
/*
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
* flush pointers outside the loop, then delete all matching
* rules. prev remains the same throughout the cycle.
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
*/
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
flush_rule_ptrs();
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
while (rule && rule->rulenum == rulenum)
rule = delete_rule(chain, prev, rule);
splx(s);
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
case 1: /* delete all rules with given set number */
s = splimp();
flush_rule_ptrs();
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
for (prev = NULL, rule = *chain; rule ; )
if (rule->set == rulenum)
rule = delete_rule(chain, prev, rule);
else {
prev = rule;
rule = rule->next;
}
splx(s);
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
case 2: /* move rules with given number to new set */
s = splimp();
for (rule = *chain; rule ; rule = rule->next)
if (rule->rulenum == rulenum)
rule->set = new_set;
splx(s);
break;
case 3: /* move rules with given set number to new set */
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
s = splimp();
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
for (rule = *chain; rule ; rule = rule->next)
if (rule->set == rulenum)
rule->set = new_set;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
splx(s);
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
case 4: /* swap two sets */
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
s = splimp();
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
for (rule = *chain; rule ; rule = rule->next)
if (rule->set == rulenum)
rule->set = new_set;
else if (rule->set == new_set)
rule->set = rulenum;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
splx(s);
break;
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return 0;
}
/*
* Clear counters for a specific rule.
*/
static void
clear_counters(struct ip_fw *rule, int log_only)
{
ipfw_insn_log *l = (ipfw_insn_log *)ACTION_PTR(rule);
if (log_only == 0) {
rule->bcnt = rule->pcnt = 0;
rule->timestamp = 0;
}
if (l->o.opcode == O_LOG)
l->log_left = l->max_log;
}
/**
* Reset some or all counters on firewall rules.
* @arg frwl is null to clear all entries, or contains a specific
* rule number.
* @arg log_only is 1 if we only want to reset logs, zero otherwise.
*/
static int
zero_entry(int rulenum, int log_only)
{
struct ip_fw *rule;
int s;
char *msg;
if (rulenum == 0) {
s = splimp();
norule_counter = 0;
for (rule = layer3_chain; rule; rule = rule->next)
clear_counters(rule, log_only);
splx(s);
msg = log_only ? "ipfw: All logging counts reset.\n" :
"ipfw: Accounting cleared.\n";
} else {
int cleared = 0;
/*
* We can have multiple rules with the same number, so we
* need to clear them all.
*/
for (rule = layer3_chain; rule; rule = rule->next)
if (rule->rulenum == rulenum) {
s = splimp();
while (rule && rule->rulenum == rulenum) {
clear_counters(rule, log_only);
rule = rule->next;
}
splx(s);
cleared = 1;
break;
}
if (!cleared) /* we did not find any matching rules */
return (EINVAL);
msg = log_only ? "ipfw: Entry %d logging count reset.\n" :
"ipfw: Entry %d cleared.\n";
}
if (fw_verbose)
log(LOG_SECURITY | LOG_NOTICE, msg, rulenum);
return (0);
}
/*
* Check validity of the structure before insert.
* Fortunately rules are simple, so this mostly need to check rule sizes.
*/
static int
check_ipfw_struct(struct ip_fw *rule, int size)
{
int l, cmdlen = 0;
int have_action=0;
ipfw_insn *cmd;
if (size < sizeof(*rule)) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: rule too short\n");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return (EINVAL);
}
/* first, check for valid size */
l = RULESIZE(rule);
if (l != size) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: size mismatch (have %d want %d)\n", size, l);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return (EINVAL);
}
/*
* Now go for the individual checks. Very simple ones, basically only
* instruction sizes.
*/
for (l = rule->cmd_len, cmd = rule->cmd ;
l > 0 ; l -= cmdlen, cmd += cmdlen) {
cmdlen = F_LEN(cmd);
if (cmdlen > l) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: opcode %d size truncated\n",
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->opcode);
return EINVAL;
}
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
DEB(printf("ipfw: opcode %d\n", cmd->opcode);)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
switch (cmd->opcode) {
case O_NOP:
case O_PROBE_STATE:
case O_KEEP_STATE:
case O_PROTO:
case O_IP_SRC_ME:
case O_IP_DST_ME:
case O_LAYER2:
case O_IN:
case O_FRAG:
case O_IPOPT:
case O_IPLEN:
case O_IPID:
case O_IPTOS:
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_IPPRECEDENCE:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_IPTTL:
case O_IPVER:
case O_TCPWIN:
case O_TCPFLAGS:
case O_TCPOPTS:
case O_ESTAB:
if (cmdlen != F_INSN_SIZE(ipfw_insn))
goto bad_size;
break;
case O_UID:
case O_GID:
case O_IP_SRC:
case O_IP_DST:
case O_TCPSEQ:
case O_TCPACK:
case O_PROB:
case O_ICMPTYPE:
if (cmdlen != F_INSN_SIZE(ipfw_insn_u32))
goto bad_size;
break;
case O_LIMIT:
if (cmdlen != F_INSN_SIZE(ipfw_insn_limit))
goto bad_size;
break;
case O_LOG:
if (cmdlen != F_INSN_SIZE(ipfw_insn_log))
goto bad_size;
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
((ipfw_insn_log *)cmd)->log_left =
((ipfw_insn_log *)cmd)->max_log;
break;
case O_IP_SRC_MASK:
case O_IP_DST_MASK:
if (cmdlen != F_INSN_SIZE(ipfw_insn_ip))
goto bad_size;
if (((ipfw_insn_ip *)cmd)->mask.s_addr == 0) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: opcode %d, useless rule\n",
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->opcode);
return EINVAL;
}
break;
case O_IP_SRC_SET:
case O_IP_DST_SET:
if (cmd->arg1 == 0 || cmd->arg1 > 256) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: invalid set size %d\n",
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->arg1);
return EINVAL;
}
if (cmdlen != F_INSN_SIZE(ipfw_insn_u32) +
(cmd->arg1+31)/32 )
goto bad_size;
break;
case O_MACADDR2:
if (cmdlen != F_INSN_SIZE(ipfw_insn_mac))
goto bad_size;
break;
case O_MAC_TYPE:
case O_IP_SRCPORT:
Raise limit for port lists to 30 entries/ranges. Remove a duplicate "logging" message, and identify the firewall as ipfw2 in the boot message.
2002-08-19 04:45:01 +00:00
case O_IP_DSTPORT: /* XXX artificial limit, 30 port pairs */
if (cmdlen < 2 || cmdlen > 31)
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
goto bad_size;
break;
case O_RECV:
case O_XMIT:
case O_VIA:
if (cmdlen != F_INSN_SIZE(ipfw_insn_if))
goto bad_size;
break;
case O_PIPE:
case O_QUEUE:
if (cmdlen != F_INSN_SIZE(ipfw_insn_pipe))
goto bad_size;
goto check_action;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_FORWARD_IP:
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (cmdlen != F_INSN_SIZE(ipfw_insn_sa))
goto bad_size;
goto check_action;
Implement the last 2-3 missing instructions for ipfw, now it should support all the instructions of the old ipfw. Fix some bugs in the user interface, /sbin/ipfw. Please check this code against your rulesets, so i can fix the remaining bugs (if any, i think they will be mostly in /sbin/ipfw). Once we have done a bit of testing, this code is ready to be MFC'ed, together with a bunch of other changes (glue to ipfw, and also the removal of some global variables) which have been in -current for a couple of weeks now. MFC after: 7 days
2002-07-05 22:43:06 +00:00
case O_FORWARD_MAC: /* XXX not implemented yet */
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
case O_CHECK_STATE:
case O_COUNT:
case O_ACCEPT:
case O_DENY:
case O_REJECT:
case O_SKIPTO:
case O_DIVERT:
case O_TEE:
if (cmdlen != F_INSN_SIZE(ipfw_insn))
goto bad_size;
check_action:
if (have_action) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: opcode %d, multiple actions"
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
" not allowed\n",
cmd->opcode);
return EINVAL;
}
have_action = 1;
if (l != cmdlen) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: opcode %d, action must be"
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
" last opcode\n",
cmd->opcode);
return EINVAL;
}
break;
default:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: opcode %d, unknown opcode\n",
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->opcode);
return EINVAL;
}
}
if (have_action == 0) {
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: missing action\n");
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
return EINVAL;
}
return 0;
bad_size:
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
printf("ipfw: opcode %d size %d wrong\n",
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
cmd->opcode, cmdlen);
return EINVAL;
}
/**
* {set|get}sockopt parser.
*/
static int
ipfw_ctl(struct sockopt *sopt)
{
int error, s, rulenum;
size_t size;
struct ip_fw *bp , *buf, *rule;
static u_int32_t rule_buf[255]; /* we copy the data here */
/*
* Disallow modifications in really-really secure mode, but still allow
* the logging counters to be reset.
*/
if (sopt->sopt_name == IP_FW_ADD ||
(sopt->sopt_dir == SOPT_SET && sopt->sopt_name != IP_FW_RESETLOG)) {
#if __FreeBSD_version >= 500034
error = securelevel_ge(sopt->sopt_td->td_ucred, 3);
if (error)
return (error);
#else /* FreeBSD 4.x */
if (securelevel >= 3)
return (EPERM);
#endif
}
error = 0;
switch (sopt->sopt_name) {
case IP_FW_GET:
/*
* pass up a copy of the current rules. Static rules
* come first (the last of which has number IPFW_DEFAULT_RULE),
* followed by a possibly empty list of dynamic rule.
* The last dynamic rule has NULL in the "next" field.
*/
s = splimp();
size = static_len; /* size of static rules */
if (ipfw_dyn_v) /* add size of dyn.rules */
size += (dyn_count * sizeof(ipfw_dyn_rule));
/*
* XXX todo: if the user passes a short length just to know
* how much room is needed, do not bother filling up the
* buffer, just jump to the sooptcopyout.
*/
buf = malloc(size, M_TEMP, M_WAITOK);
if (buf == 0) {
splx(s);
error = ENOBUFS;
break;
}
bp = buf;
for (rule = layer3_chain; rule ; rule = rule->next) {
int i = RULESIZE(rule);
bcopy(rule, bp, i);
Fix ipfw2 panics on 64-bit platforms. Quoting luigi: In order to make the userland code fully 64-bit clean it may be necessary to commit other changes that may or may not cause a minor change in the ABI. Reviewed by: luigi
2002-10-24 18:04:44 +00:00
((struct ip_fw *)bp)->set_disable = set_disable;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
bp = (struct ip_fw *)((char *)bp + i);
}
if (ipfw_dyn_v) {
int i;
ipfw_dyn_rule *p, *dst, *last = NULL;
dst = (ipfw_dyn_rule *)bp;
for (i = 0 ; i < curr_dyn_buckets ; i++ )
for ( p = ipfw_dyn_v[i] ; p != NULL ;
p = p->next, dst++ ) {
bcopy(p, dst, sizeof *p);
Fix ipfw2 panics on 64-bit platforms. Quoting luigi: In order to make the userland code fully 64-bit clean it may be necessary to commit other changes that may or may not cause a minor change in the ABI. Reviewed by: luigi
2002-10-24 18:04:44 +00:00
dst->rulenum = p->rule->rulenum;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
/*
* store a non-null value in "next".
* The userland code will interpret a
* NULL here as a marker
* for the last dynamic rule.
*/
dst->next = dst ;
last = dst ;
dst->expire =
TIME_LEQ(dst->expire, time_second) ?
0 : dst->expire - time_second ;
}
if (last != NULL) /* mark last dynamic rule */
last->next = NULL;
}
splx(s);
error = sooptcopyout(sopt, buf, size);
free(buf, M_TEMP);
break;
case IP_FW_FLUSH:
/*
* Normally we cannot release the lock on each iteration.
* We could do it here only because we start from the head all
* the times so there is no risk of missing some entries.
* On the other hand, the risk is that we end up with
* a very inconsistent ruleset, so better keep the lock
* around the whole cycle.
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
*
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
* XXX this code can be improved by resetting the head of
* the list to point to the default rule, and then freeing
* the old list without the need for a lock.
*/
s = splimp();
free_chain(&layer3_chain, 0 /* keep default rule */);
splx(s);
break;
case IP_FW_ADD:
rule = (struct ip_fw *)rule_buf; /* XXX do a malloc */
error = sooptcopyin(sopt, rule, sizeof(rule_buf),
sizeof(struct ip_fw) );
size = sopt->sopt_valsize;
if (error || (error = check_ipfw_struct(rule, size)))
break;
error = add_rule(&layer3_chain, rule);
size = RULESIZE(rule);
if (!error && sopt->sopt_dir == SOPT_GET)
error = sooptcopyout(sopt, rule, size);
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
case IP_FW_DEL:
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
/*
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
* IP_FW_DEL is used for deleting single rules or sets,
* and (ab)used to atomically manipulate sets. Argument size
* is used to distinguish between the two:
* sizeof(u_int32_t)
* delete single rule or set of rules,
* or reassign rules (or sets) to a different set.
* 2*sizeof(u_int32_t)
* atomic disable/enable sets.
* first u_int32_t contains sets to be disabled,
* second u_int32_t contains sets to be enabled.
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
*/
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
error = sooptcopyin(sopt, rule_buf,
2*sizeof(u_int32_t), sizeof(u_int32_t));
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
if (error)
break;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
size = sopt->sopt_valsize;
if (size == sizeof(u_int32_t)) /* delete or reassign */
error = del_entry(&layer3_chain, rule_buf[0]);
else if (size == 2*sizeof(u_int32_t)) /* set enable/disable */
set_disable =
(set_disable | rule_buf[0]) & ~rule_buf[1] &
~(1<<31); /* set 31 always enabled */
else
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
error = EINVAL;
break;
case IP_FW_ZERO:
case IP_FW_RESETLOG: /* argument is an int, the rule number */
rulenum=0;
if (sopt->sopt_val != 0) {
error = sooptcopyin(sopt, &rulenum,
sizeof(int), sizeof(int));
if (error)
break;
}
error = zero_entry(rulenum, sopt->sopt_name == IP_FW_RESETLOG);
break;
default:
printf("ipfw_ctl invalid option %d\n", sopt->sopt_name);
error = EINVAL;
}
return (error);
}
/**
* dummynet needs a reference to the default rule, because rules can be
* deleted while packets hold a reference to them. When this happens,
* dummynet changes the reference to the default rule (it could well be a
* NULL pointer, but this way we do not need to check for the special
* case, plus here he have info on the default behaviour).
*/
struct ip_fw *ip_fw_default_rule;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
/*
* This procedure is only used to handle keepalives. It is invoked
* every dyn_keepalive_period
*/
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
static void
ipfw_tick(void * __unused unused)
{
int i;
int s;
ipfw_dyn_rule *q;
if (dyn_keepalive == 0 || ipfw_dyn_v == NULL || dyn_count == 0)
goto done;
s = splimp();
for (i = 0 ; i < curr_dyn_buckets ; i++) {
for (q = ipfw_dyn_v[i] ; q ; q = q->next ) {
if (q->dyn_type == O_LIMIT_PARENT)
continue;
if (q->id.proto != IPPROTO_TCP)
continue;
if ( (q->state & BOTH_SYN) != BOTH_SYN)
continue;
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
if (TIME_LEQ( time_second+dyn_keepalive_interval,
q->expire))
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
continue; /* too early */
if (TIME_LEQ(q->expire, time_second))
continue; /* too late, rule expired */
send_pkt(&(q->id), q->ack_rev - 1, q->ack_fwd, TH_SYN);
send_pkt(&(q->id), q->ack_fwd - 1, q->ack_rev, 0);
}
}
splx(s);
done:
sys/netinet/ip_fw2.c: Implement the M_SKIP_FIREWALL bit in m_flags to avoid loops for firewall-generated packets (the constant has to go in sys/mbuf.h). Better comments on keepalive generation, and enforce dyn_rst_lifetime and dyn_fin_lifetime to be less than dyn_keepalive_period. Enforce limits (up to 64k) on the number of dynamic buckets, and retry allocation with smaller sizes. Raise default number of dynamic rules to 4096. Improved handling of set of rules -- now you can atomically enable/disable multiple sets, move rules from one set to another, and swap sets. sbin/ipfw/ipfw2.c: userland support for "noerror" pipe attribute. userland support for sets of rules. minor improvements on rule parsing and printing. sbin/ipfw/ipfw.8: more documentation on ipfw2 extensions, differences from ipfw1 (so we can use the same manpage for both), stateful rules, and some additional examples. Feedback and more examples needed here.
2002-08-16 10:31:47 +00:00
ipfw_timeout_h = timeout(ipfw_tick, NULL, dyn_keepalive_period*hz);
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
}
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
static void
ipfw_init(void)
{
struct ip_fw default_rule;
ip_fw_chk_ptr = ipfw_chk;
ip_fw_ctl_ptr = ipfw_ctl;
layer3_chain = NULL;
bzero(&default_rule, sizeof default_rule);
default_rule.act_ofs = 0;
default_rule.rulenum = IPFW_DEFAULT_RULE;
default_rule.cmd_len = 1;
One bugfix and one new feature. The bugfix (ipfw2.c) makes the handling of port numbers with a dash in the name, e.g. ftp-data, consistent with old ipfw: use \\ before the - to consider it as part of the name and not a range separator. The new feature (all this description will go in the manpage): each rule now belongs to one of 32 different sets, which can be optionally specified in the following form: ipfw add 100 set 23 allow ip from any to any If "set N" is not specified, the rule belongs to set 0. Individual sets can be disabled, enabled, and deleted with the commands: ipfw disable set N ipfw enable set N ipfw delete set N Enabling/disabling of a set is atomic. Rules belonging to a disabled set are skipped during packet matching, and they are not listed unless you use the '-S' flag in the show/list commands. Note that dynamic rules, once created, are always active until they expire or their parent rule is deleted. Set 31 is reserved for the default rule and cannot be disabled. All sets are enabled by default. The enable/disable status of the sets can be shown with the command ipfw show sets Hopefully, this feature will make life easier to those who want to have atomic ruleset addition/deletion/tests. Examples: To add a set of rules atomically: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 To delete a set of rules atomically ipfw disable set 18 ipfw delete set 18 ipfw enable set 18 To test a ruleset and disable it and regain control if something goes wrong: ipfw disable set 18 ipfw add ... set 18 ... # repeat as needed ipfw enable set 18 ; echo "done "; sleep 30 && ipfw disable set 18 here if everything goes well, you press control-C before the "sleep" terminates, and your ruleset will be left active. Otherwise, e.g. if you cannot access your box, the ruleset will be disabled after the sleep terminates. I think there is only one more thing that one might want, namely a command to assign all rules in set X to set Y, so one can test a ruleset using the above mechanisms, and once it is considered acceptable, make it part of an existing ruleset.
2002-08-10 04:37:32 +00:00
default_rule.set = 31;
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
default_rule.cmd[0].len = 1;
default_rule.cmd[0].opcode =
#ifdef IPFIREWALL_DEFAULT_TO_ACCEPT
1 ? O_ACCEPT :
#endif
O_DENY;
add_rule(&layer3_chain, &default_rule);
ip_fw_default_rule = layer3_chain;
Raise limit for port lists to 30 entries/ranges. Remove a duplicate "logging" message, and identify the firewall as ipfw2 in the boot message.
2002-08-19 04:45:01 +00:00
printf("ipfw2 initialized, divert %s, "
"rule-based forwarding enabled, default to %s, logging ",
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
#ifdef IPDIVERT
"enabled",
#else
"disabled",
#endif
default_rule.cmd[0].opcode == O_ACCEPT ? "accept" : "deny");
#ifdef IPFIREWALL_VERBOSE
fw_verbose = 1;
#endif
#ifdef IPFIREWALL_VERBOSE_LIMIT
verbose_limit = IPFIREWALL_VERBOSE_LIMIT;
#endif
if (fw_verbose == 0)
printf("disabled\n");
else if (verbose_limit == 0)
printf("unlimited\n");
else
printf("limited to %d packets/entry by default\n",
verbose_limit);
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
bzero(&ipfw_timeout_h, sizeof(struct callout_handle));
ipfw_timeout_h = timeout(ipfw_tick, NULL, hz);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
}
static int
ipfw_modevent(module_t mod, int type, void *unused)
{
int s;
int err = 0;
Kill EOL spaces. Approved by: luigi MFC after: 1 week
2002-10-23 10:07:55 +00:00
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
switch (type) {
case MOD_LOAD:
s = splimp();
if (IPFW_LOADED) {
splx(s);
printf("IP firewall already loaded\n");
err = EEXIST;
} else {
ipfw_init();
splx(s);
}
break;
case MOD_UNLOAD:
#if !defined(KLD_MODULE)
printf("ipfw statically compiled, cannot unload\n");
err = EBUSY;
#else
s = splimp();
Implement keepalives for dynamic rules, so they will not expire just because you leave your session idle. Also, put in a fix for 64-bit architectures (to be revised). In detail: ip_fw.h * Reorder fields in struct ip_fw to avoid alignment problems on 64-bit machines. This only masks the problem, I am still not sure whether I am doing something wrong in the code or there is a problem elsewhere (e.g. different aligmnent of structures between userland and kernel because of pragmas etc.) * added fields in dyn_rule to store ack numbers, so we can generate keepalives when the dynamic rule is about to expire ip_fw2.c * use a local function, send_pkt(), to generate TCP RST for Reset rules; * save about 250 bytes by cleaning up the various snprintf() in ipfw_log() ... * ... and use twice as many bytes to implement keepalives (this seems to be working, but i have not tested it extensively). Keepalives are generated once every 5 seconds for the last 20 seconds of the lifetime of a dynamic rule for an established TCP flow. The packets are sent to both sides, so if at least one of the endpoints is responding, the timeout is refreshed and the rule will not expire. You can disable this feature with sysctl net.inet.ip.fw.dyn_keepalive=0 (the default is 1, to have them enabled). MFC after: 1 day (just kidding... I will supply an updated version of ipfw2 for RELENG_4 tomorrow).
2002-07-14 23:47:18 +00:00
untimeout(ipfw_tick, NULL, ipfw_timeout_h);
The new ipfw code. This code makes use of variable-size kernel representation of rules (exactly the same concept of BPF instructions, as used in the BSDI's firewall), which makes firewall operation a lot faster, and the code more readable and easier to extend and debug. The interface with the rest of the system is unchanged, as witnessed by this commit. The only extra kernel files that I am touching are if_fw.h and ip_dummynet.c, which is quite tied to ipfw. In userland I only had to touch those programs which manipulate the internal representation of firewall rules). The code is almost entirely new (and I believe I have written the vast majority of those sections which were taken from the former ip_fw.c), so rather than modifying the old ip_fw.c I decided to create a new file, sys/netinet/ip_fw2.c . Same for the user interface, which is in sbin/ipfw/ipfw2.c (it still compiles to /sbin/ipfw). The old files are still there, and will be removed in due time. I have not renamed the header file because it would have required touching a one-line change to a number of kernel files. In terms of user interface, the new "ipfw" is supposed to accepts the old syntax for ipfw rules (and produce the same output with "ipfw show". Only a couple of the old options (out of some 30 of them) has not been implemented, but they will be soon. On the other hand, the new code has some very powerful extensions. First, you can put "or" connectives between match fields (and soon also between options), and write things like ipfw add allow ip from { 1.2.3.4/27 or 5.6.7.8/30 } 10-23,25,1024-3000 to any This should make rulesets slightly more compact (and lines longer!), by condensing 2 or more of the old rules into single ones. Also, as an example of how easy the rules can be extended, I have implemented an 'address set' match pattern, where you can specify an IP address in a format like this: 10.20.30.0/26{18,44,33,22,9} which will match the set of hosts listed in braces belonging to the subnet 10.20.30.0/26 . The match is done using a bitmap, so it is essentially a constant time operation requiring a handful of CPU instructions (and a very small amount of memmory -- for a full /24 subnet, the instruction only consumes 40 bytes). Again, in this commit I have focused on functionality and tried to minimize changes to the other parts of the system. Some performance improvement can be achieved with minor changes to the interface of ip_fw_chk_t. This will be done later when this code is settled. The code is meant to compile unmodified on RELENG_4 (once the PACKET_TAG_* changes have been merged), for this reason you will see #ifdef __FreeBSD_version in a couple of places. This should minimize errors when (hopefully soon) it will be time to do the MFC.
2002-06-27 23:02:18 +00:00
ip_fw_chk_ptr = NULL;
ip_fw_ctl_ptr = NULL;
free_chain(&layer3_chain, 1 /* kill default rule */);
splx(s);
printf("IP firewall unloaded\n");
#endif
break;
default:
break;
}
return err;
}
static moduledata_t ipfwmod = {
"ipfw",
ipfw_modevent,
0
};
DECLARE_MODULE(ipfw, ipfwmod, SI_SUB_PSEUDO, SI_ORDER_ANY);
MODULE_VERSION(ipfw, 1);
No functional changes, but: Following Darren's suggestion, make Dijkstra happy and rewrite the ipfw_chk() main loop removing a lot of goto's and using instead a variable to store match status. Add a lot of comments to explain what instructions are supposed to do and how -- this should ease auditing of the code and make people more confident with it. In terms of code size: the entire file takes about 12700 bytes of text, about 3K of which are for the main function, ipfw_chk(), and 2K (ouch!) for ipfw_log().
2002-07-08 22:46:01 +00:00
#endif /* IPFW2 */
Reference in New Issue Copy Permalink