freebsd-skq/sys/netinet/ipfw/ip_fw_private.h

302 lines
9.3 KiB
C
Raw Normal View History

/*-
* Copyright (c) 2002-2009 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.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* 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.
*
* $FreeBSD$
*/
#ifndef _IPFW2_PRIVATE_H
#define _IPFW2_PRIVATE_H
/*
* Internal constants and data structures used by ipfw components
* and not meant to be exported outside the kernel.
*/
#ifdef _KERNEL
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
/*
* For platforms that do not have SYSCTL support, we wrap the
* SYSCTL_* into a function (one per file) to collect the values
* into an array at module initialization. The wrapping macros,
* SYSBEGIN() and SYSEND, are empty in the default case.
*/
#ifndef SYSBEGIN
#define SYSBEGIN(x)
#endif
#ifndef SYSEND
#define SYSEND
#endif
/* Return values from ipfw_chk() */
enum {
IP_FW_PASS = 0,
IP_FW_DENY,
IP_FW_DIVERT,
IP_FW_TEE,
IP_FW_DUMMYNET,
IP_FW_NETGRAPH,
IP_FW_NGTEE,
IP_FW_NAT,
IP_FW_REASS,
};
/*
* Structure for collecting parameters to dummynet for ip6_output forwarding
*/
struct _ip6dn_args {
struct ip6_pktopts *opt_or;
struct route_in6 ro_or;
int flags_or;
struct ip6_moptions *im6o_or;
struct ifnet *origifp_or;
struct ifnet *ifp_or;
struct sockaddr_in6 dst_or;
u_long mtu_or;
struct route_in6 ro_pmtu_or;
};
/*
* Arguments for calling ipfw_chk() and dummynet_io(). We put them
* all into a structure because this way it is easier and more
* efficient to pass variables around and extend the interface.
*/
struct ip_fw_args {
struct mbuf *m; /* the mbuf chain */
struct ifnet *oif; /* output interface */
struct sockaddr_in *next_hop; /* forward address */
/*
* On return, it points to the matching rule.
* On entry, rule.slot > 0 means the info is valid and
* contains the the starting rule for an ipfw search.
* If chain_id == chain->id && slot >0 then jump to that slot.
* Otherwise, we locate the first rule >= rulenum:rule_id
merge code from ipfw3-head to reduce contention on the ipfw lock and remove all O(N) sequences from kernel critical sections in ipfw. In detail: 1. introduce a IPFW_UH_LOCK to arbitrate requests from the upper half of the kernel. Some things, such as 'ipfw show', can be done holding this lock in read mode, whereas insert and delete require IPFW_UH_WLOCK. 2. introduce a mapping structure to keep rules together. This replaces the 'next' chain currently used in ipfw rules. At the moment the map is a simple array (sorted by rule number and then rule_id), so we can find a rule quickly instead of having to scan the list. This reduces many expensive lookups from O(N) to O(log N). 3. when an expensive operation (such as insert or delete) is done by userland, we grab IPFW_UH_WLOCK, create a new copy of the map without blocking the bottom half of the kernel, then acquire IPFW_WLOCK and quickly update pointers to the map and related info. After dropping IPFW_LOCK we can then continue the cleanup protected by IPFW_UH_LOCK. So userland still costs O(N) but the kernel side is only blocked for O(1). 4. do not pass pointers to rules through dummynet, netgraph, divert etc, but rather pass a <slot, chain_id, rulenum, rule_id> tuple. We validate the slot index (in the array of #2) with chain_id, and if successful do a O(1) dereference; otherwise, we can find the rule in O(log N) through <rulenum, rule_id> All the above does not change the userland/kernel ABI, though there are some disgusting casts between pointers and uint32_t Operation costs now are as follows: Function Old Now Planned ------------------------------------------------------------------- + skipto X, non cached O(N) O(log N) + skipto X, cached O(1) O(1) XXX dynamic rule lookup O(1) O(log N) O(1) + skipto tablearg O(N) O(1) + reinject, non cached O(N) O(log N) + reinject, cached O(1) O(1) + kernel blocked during setsockopt() O(N) O(1) ------------------------------------------------------------------- The only (very small) regression is on dynamic rule lookup and this will be fixed in a day or two, without changing the userland/kernel ABI Supported by: Valeria Paoli MFC after: 1 month
2009-12-22 19:01:47 +00:00
*/
struct ipfw_rule_ref rule; /* match/restart info */
struct ether_header *eh; /* for bridged packets */
struct ipfw_flow_id f_id; /* grabbed from IP header */
//uint32_t cookie; /* a cookie depending on rule action */
struct inpcb *inp;
struct _ip6dn_args dummypar; /* dummynet->ip6_output */
struct sockaddr_in hopstore; /* store here if cannot use a pointer */
};
MALLOC_DECLARE(M_IPFW);
/*
* Hooks sometime need to know the direction of the packet
* (divert, dummynet, netgraph, ...)
* We use a generic definition here, with bit0-1 indicating the
* direction, bit 2 indicating layer2 or 3, bit 3-4 indicating the
* specific protocol
* indicating the protocol (if necessary)
*/
enum {
DIR_MASK = 0x3,
DIR_OUT = 0,
DIR_IN = 1,
DIR_FWD = 2,
DIR_DROP = 3,
PROTO_LAYER2 = 0x4, /* set for layer 2 */
/* PROTO_DEFAULT = 0, */
PROTO_IPV4 = 0x08,
PROTO_IPV6 = 0x10,
PROTO_IFB = 0x0c, /* layer2 + ifbridge */
/* PROTO_OLDBDG = 0x14, unused, old bridge */
};
/* wrapper for freeing a packet, in case we need to do more work */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#ifndef FREE_PKT
#if defined(__linux__) || defined(_WIN32)
#define FREE_PKT(m) netisr_dispatch(-1, m)
#else
#define FREE_PKT(m) m_freem(m)
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#endif
#endif /* !FREE_PKT */
/*
* Function definitions.
*/
/* attach (arg = 1) or detach (arg = 0) hooks */
int ipfw_attach_hooks(int);
#ifdef NOTYET
void ipfw_nat_destroy(void);
#endif
/* In ip_fw_log.c */
struct ip;
void ipfw_log_bpf(int);
void ipfw_log(struct ip_fw *f, u_int hlen, struct ip_fw_args *args,
struct mbuf *m, struct ifnet *oif, u_short offset, uint32_t tablearg,
struct ip *ip);
VNET_DECLARE(u_int64_t, norule_counter);
#define V_norule_counter VNET(norule_counter)
VNET_DECLARE(int, verbose_limit);
#define V_verbose_limit VNET(verbose_limit)
/* In ip_fw_dynamic.c */
enum { /* result for matching dynamic rules */
MATCH_REVERSE = 0,
MATCH_FORWARD,
MATCH_NONE,
MATCH_UNKNOWN,
};
/*
* The lock for dynamic rules is only used once outside the file,
* and only to release the result of lookup_dyn_rule().
* Eventually we may implement it with a callback on the function.
*/
void ipfw_dyn_unlock(void);
struct tcphdr;
struct mbuf *ipfw_send_pkt(struct mbuf *, struct ipfw_flow_id *,
u_int32_t, u_int32_t, int);
int ipfw_install_state(struct ip_fw *rule, ipfw_insn_limit *cmd,
struct ip_fw_args *args, uint32_t tablearg);
ipfw_dyn_rule *ipfw_lookup_dyn_rule(struct ipfw_flow_id *pkt,
int *match_direction, struct tcphdr *tcp);
void ipfw_remove_dyn_children(struct ip_fw *rule);
void ipfw_get_dynamic(char **bp, const char *ep);
void ipfw_dyn_attach(void); /* uma_zcreate .... */
void ipfw_dyn_detach(void); /* uma_zdestroy ... */
void ipfw_dyn_init(void); /* per-vnet initialization */
void ipfw_dyn_uninit(int); /* per-vnet deinitialization */
int ipfw_dyn_len(void);
/* common variables */
VNET_DECLARE(int, fw_one_pass);
#define V_fw_one_pass VNET(fw_one_pass)
VNET_DECLARE(int, fw_verbose);
#define V_fw_verbose VNET(fw_verbose)
VNET_DECLARE(struct ip_fw_chain, layer3_chain);
#define V_layer3_chain VNET(layer3_chain)
VNET_DECLARE(u_int32_t, set_disable);
#define V_set_disable VNET(set_disable)
VNET_DECLARE(int, autoinc_step);
#define V_autoinc_step VNET(autoinc_step)
struct ip_fw_chain {
struct ip_fw *rules; /* list of rules */
struct ip_fw *reap; /* list of rules to reap */
struct ip_fw *default_rule;
int n_rules; /* number of static rules */
int static_len; /* total len of static rules */
struct ip_fw **map; /* array of rule ptrs to ease lookup */
LIST_HEAD(nat_list, cfg_nat) nat; /* list of nat entries */
struct radix_node_head *tables[IPFW_TABLES_MAX];
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#if defined( __linux__ ) || defined( _WIN32 )
spinlock_t rwmtx;
spinlock_t uh_lock;
#else
struct rwlock rwmtx;
merge code from ipfw3-head to reduce contention on the ipfw lock and remove all O(N) sequences from kernel critical sections in ipfw. In detail: 1. introduce a IPFW_UH_LOCK to arbitrate requests from the upper half of the kernel. Some things, such as 'ipfw show', can be done holding this lock in read mode, whereas insert and delete require IPFW_UH_WLOCK. 2. introduce a mapping structure to keep rules together. This replaces the 'next' chain currently used in ipfw rules. At the moment the map is a simple array (sorted by rule number and then rule_id), so we can find a rule quickly instead of having to scan the list. This reduces many expensive lookups from O(N) to O(log N). 3. when an expensive operation (such as insert or delete) is done by userland, we grab IPFW_UH_WLOCK, create a new copy of the map without blocking the bottom half of the kernel, then acquire IPFW_WLOCK and quickly update pointers to the map and related info. After dropping IPFW_LOCK we can then continue the cleanup protected by IPFW_UH_LOCK. So userland still costs O(N) but the kernel side is only blocked for O(1). 4. do not pass pointers to rules through dummynet, netgraph, divert etc, but rather pass a <slot, chain_id, rulenum, rule_id> tuple. We validate the slot index (in the array of #2) with chain_id, and if successful do a O(1) dereference; otherwise, we can find the rule in O(log N) through <rulenum, rule_id> All the above does not change the userland/kernel ABI, though there are some disgusting casts between pointers and uint32_t Operation costs now are as follows: Function Old Now Planned ------------------------------------------------------------------- + skipto X, non cached O(N) O(log N) + skipto X, cached O(1) O(1) XXX dynamic rule lookup O(1) O(log N) O(1) + skipto tablearg O(N) O(1) + reinject, non cached O(N) O(log N) + reinject, cached O(1) O(1) + kernel blocked during setsockopt() O(N) O(1) ------------------------------------------------------------------- The only (very small) regression is on dynamic rule lookup and this will be fixed in a day or two, without changing the userland/kernel ABI Supported by: Valeria Paoli MFC after: 1 month
2009-12-22 19:01:47 +00:00
struct rwlock uh_lock; /* lock for upper half */
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
#endif
uint32_t id; /* ruleset id */
};
struct sockopt; /* used by tcp_var.h */
/*
* The lock is heavily used by ip_fw2.c (the main file) and ip_fw_nat.c
* so the variable and the macros must be here.
*/
merge code from ipfw3-head to reduce contention on the ipfw lock and remove all O(N) sequences from kernel critical sections in ipfw. In detail: 1. introduce a IPFW_UH_LOCK to arbitrate requests from the upper half of the kernel. Some things, such as 'ipfw show', can be done holding this lock in read mode, whereas insert and delete require IPFW_UH_WLOCK. 2. introduce a mapping structure to keep rules together. This replaces the 'next' chain currently used in ipfw rules. At the moment the map is a simple array (sorted by rule number and then rule_id), so we can find a rule quickly instead of having to scan the list. This reduces many expensive lookups from O(N) to O(log N). 3. when an expensive operation (such as insert or delete) is done by userland, we grab IPFW_UH_WLOCK, create a new copy of the map without blocking the bottom half of the kernel, then acquire IPFW_WLOCK and quickly update pointers to the map and related info. After dropping IPFW_LOCK we can then continue the cleanup protected by IPFW_UH_LOCK. So userland still costs O(N) but the kernel side is only blocked for O(1). 4. do not pass pointers to rules through dummynet, netgraph, divert etc, but rather pass a <slot, chain_id, rulenum, rule_id> tuple. We validate the slot index (in the array of #2) with chain_id, and if successful do a O(1) dereference; otherwise, we can find the rule in O(log N) through <rulenum, rule_id> All the above does not change the userland/kernel ABI, though there are some disgusting casts between pointers and uint32_t Operation costs now are as follows: Function Old Now Planned ------------------------------------------------------------------- + skipto X, non cached O(N) O(log N) + skipto X, cached O(1) O(1) XXX dynamic rule lookup O(1) O(log N) O(1) + skipto tablearg O(N) O(1) + reinject, non cached O(N) O(log N) + reinject, cached O(1) O(1) + kernel blocked during setsockopt() O(N) O(1) ------------------------------------------------------------------- The only (very small) regression is on dynamic rule lookup and this will be fixed in a day or two, without changing the userland/kernel ABI Supported by: Valeria Paoli MFC after: 1 month
2009-12-22 19:01:47 +00:00
#define IPFW_LOCK_INIT(_chain) do { \
rw_init(&(_chain)->rwmtx, "IPFW static rules"); \
rw_init(&(_chain)->uh_lock, "IPFW UH lock"); \
} while (0)
#define IPFW_LOCK_DESTROY(_chain) do { \
rw_destroy(&(_chain)->rwmtx); \
rw_destroy(&(_chain)->uh_lock); \
} while (0)
#define IPFW_WLOCK_ASSERT(_chain) rw_assert(&(_chain)->rwmtx, RA_WLOCKED)
#define IPFW_RLOCK(p) rw_rlock(&(p)->rwmtx)
#define IPFW_RUNLOCK(p) rw_runlock(&(p)->rwmtx)
#define IPFW_WLOCK(p) rw_wlock(&(p)->rwmtx)
#define IPFW_WUNLOCK(p) rw_wunlock(&(p)->rwmtx)
merge code from ipfw3-head to reduce contention on the ipfw lock and remove all O(N) sequences from kernel critical sections in ipfw. In detail: 1. introduce a IPFW_UH_LOCK to arbitrate requests from the upper half of the kernel. Some things, such as 'ipfw show', can be done holding this lock in read mode, whereas insert and delete require IPFW_UH_WLOCK. 2. introduce a mapping structure to keep rules together. This replaces the 'next' chain currently used in ipfw rules. At the moment the map is a simple array (sorted by rule number and then rule_id), so we can find a rule quickly instead of having to scan the list. This reduces many expensive lookups from O(N) to O(log N). 3. when an expensive operation (such as insert or delete) is done by userland, we grab IPFW_UH_WLOCK, create a new copy of the map without blocking the bottom half of the kernel, then acquire IPFW_WLOCK and quickly update pointers to the map and related info. After dropping IPFW_LOCK we can then continue the cleanup protected by IPFW_UH_LOCK. So userland still costs O(N) but the kernel side is only blocked for O(1). 4. do not pass pointers to rules through dummynet, netgraph, divert etc, but rather pass a <slot, chain_id, rulenum, rule_id> tuple. We validate the slot index (in the array of #2) with chain_id, and if successful do a O(1) dereference; otherwise, we can find the rule in O(log N) through <rulenum, rule_id> All the above does not change the userland/kernel ABI, though there are some disgusting casts between pointers and uint32_t Operation costs now are as follows: Function Old Now Planned ------------------------------------------------------------------- + skipto X, non cached O(N) O(log N) + skipto X, cached O(1) O(1) XXX dynamic rule lookup O(1) O(log N) O(1) + skipto tablearg O(N) O(1) + reinject, non cached O(N) O(log N) + reinject, cached O(1) O(1) + kernel blocked during setsockopt() O(N) O(1) ------------------------------------------------------------------- The only (very small) regression is on dynamic rule lookup and this will be fixed in a day or two, without changing the userland/kernel ABI Supported by: Valeria Paoli MFC after: 1 month
2009-12-22 19:01:47 +00:00
#define IPFW_UH_RLOCK(p) rw_rlock(&(p)->uh_lock)
#define IPFW_UH_RUNLOCK(p) rw_runlock(&(p)->uh_lock)
#define IPFW_UH_WLOCK(p) rw_wlock(&(p)->uh_lock)
#define IPFW_UH_WUNLOCK(p) rw_wunlock(&(p)->uh_lock)
/* In ip_fw_sockopt.c */
merge code from ipfw3-head to reduce contention on the ipfw lock and remove all O(N) sequences from kernel critical sections in ipfw. In detail: 1. introduce a IPFW_UH_LOCK to arbitrate requests from the upper half of the kernel. Some things, such as 'ipfw show', can be done holding this lock in read mode, whereas insert and delete require IPFW_UH_WLOCK. 2. introduce a mapping structure to keep rules together. This replaces the 'next' chain currently used in ipfw rules. At the moment the map is a simple array (sorted by rule number and then rule_id), so we can find a rule quickly instead of having to scan the list. This reduces many expensive lookups from O(N) to O(log N). 3. when an expensive operation (such as insert or delete) is done by userland, we grab IPFW_UH_WLOCK, create a new copy of the map without blocking the bottom half of the kernel, then acquire IPFW_WLOCK and quickly update pointers to the map and related info. After dropping IPFW_LOCK we can then continue the cleanup protected by IPFW_UH_LOCK. So userland still costs O(N) but the kernel side is only blocked for O(1). 4. do not pass pointers to rules through dummynet, netgraph, divert etc, but rather pass a <slot, chain_id, rulenum, rule_id> tuple. We validate the slot index (in the array of #2) with chain_id, and if successful do a O(1) dereference; otherwise, we can find the rule in O(log N) through <rulenum, rule_id> All the above does not change the userland/kernel ABI, though there are some disgusting casts between pointers and uint32_t Operation costs now are as follows: Function Old Now Planned ------------------------------------------------------------------- + skipto X, non cached O(N) O(log N) + skipto X, cached O(1) O(1) XXX dynamic rule lookup O(1) O(log N) O(1) + skipto tablearg O(N) O(1) + reinject, non cached O(N) O(log N) + reinject, cached O(1) O(1) + kernel blocked during setsockopt() O(N) O(1) ------------------------------------------------------------------- The only (very small) regression is on dynamic rule lookup and this will be fixed in a day or two, without changing the userland/kernel ABI Supported by: Valeria Paoli MFC after: 1 month
2009-12-22 19:01:47 +00:00
int ipfw_find_rule(struct ip_fw_chain *chain, uint32_t key, uint32_t id);
int ipfw_add_rule(struct ip_fw_chain *chain, struct ip_fw *input_rule);
int ipfw_ctl(struct sockopt *sopt);
int ipfw_chk(struct ip_fw_args *args);
void ipfw_reap_rules(struct ip_fw *head);
Bring in the most recent version of ipfw and dummynet, developed and tested over the past two months in the ipfw3-head branch. This also happens to be the same code available in the Linux and Windows ports of ipfw and dummynet. The major enhancement is a completely restructured version of dummynet, with support for different packet scheduling algorithms (loadable at runtime), faster queue/pipe lookup, and a much cleaner internal architecture and kernel/userland ABI which simplifies future extensions. In addition to the existing schedulers (FIFO and WF2Q+), we include a Deficit Round Robin (DRR or RR for brevity) scheduler, and a new, very fast version of WF2Q+ called QFQ. Some test code is also present (in sys/netinet/ipfw/test) that lets you build and test schedulers in userland. Also, we have added a compatibility layer that understands requests from the RELENG_7 and RELENG_8 versions of the /sbin/ipfw binaries, and replies correctly (at least, it does its best; sometimes you just cannot tell who sent the request and how to answer). The compatibility layer should make it possible to MFC this code in a relatively short time. Some minor glitches (e.g. handling of ipfw set enable/disable, and a workaround for a bug in RELENG_7's /sbin/ipfw) will be fixed with separate commits. CREDITS: This work has been partly supported by the ONELAB2 project, and mostly developed by Riccardo Panicucci and myself. The code for the qfq scheduler is mostly from Fabio Checconi, and Marta Carbone and Francesco Magno have helped with testing, debugging and some bug fixes.
2010-03-02 17:40:48 +00:00
/* In ip_fw_pfil */
int ipfw_check_hook(void *arg, struct mbuf **m0, struct ifnet *ifp, int dir,
struct inpcb *inp);
/* In ip_fw_table.c */
struct radix_node;
int ipfw_lookup_table(struct ip_fw_chain *ch, uint16_t tbl, in_addr_t addr,
uint32_t *val);
int ipfw_init_tables(struct ip_fw_chain *ch);
void ipfw_destroy_tables(struct ip_fw_chain *ch);
int ipfw_flush_table(struct ip_fw_chain *ch, uint16_t tbl);
int ipfw_add_table_entry(struct ip_fw_chain *ch, uint16_t tbl, in_addr_t addr,
uint8_t mlen, uint32_t value);
int ipfw_dump_table_entry(struct radix_node *rn, void *arg);
int ipfw_del_table_entry(struct ip_fw_chain *ch, uint16_t tbl, in_addr_t addr,
uint8_t mlen);
int ipfw_count_table(struct ip_fw_chain *ch, uint32_t tbl, uint32_t *cnt);
int ipfw_dump_table(struct ip_fw_chain *ch, ipfw_table *tbl);
/* In ip_fw_nat.c -- XXX to be moved to ip_var.h */
extern struct cfg_nat *(*lookup_nat_ptr)(struct nat_list *, int);
typedef int ipfw_nat_t(struct ip_fw_args *, struct cfg_nat *, struct mbuf *);
typedef int ipfw_nat_cfg_t(struct sockopt *);
extern ipfw_nat_t *ipfw_nat_ptr;
#define IPFW_NAT_LOADED (ipfw_nat_ptr != NULL)
extern ipfw_nat_cfg_t *ipfw_nat_cfg_ptr;
extern ipfw_nat_cfg_t *ipfw_nat_del_ptr;
extern ipfw_nat_cfg_t *ipfw_nat_get_cfg_ptr;
extern ipfw_nat_cfg_t *ipfw_nat_get_log_ptr;
#endif /* _KERNEL */
#endif /* _IPFW2_PRIVATE_H */