: Netfilter update/fixes
Subsystems (4)
cover : Netfilter update/fixes
Hi Dave,
following are a number of fixes for netfilter, fixing RCU issues all over the place, including the invalid preemption assumptions in the NF_CT_STAT_INC macro. Additionally there are a few minor enhancements for things I noticed while going over the code: replacement of a few unnecessary spinlocks by mutexes, some cleanup for nf_log.c and some return value cleanup for unregistration functions.
With these patches I think there should be nothing left holding back the preemptable RCU merge from netfilter's perspective.
Please apply, thanks.
include/linux/netfilter.h | 4 include/linux/netfilter_ipv4/ip_conntrack.h | 6 include/net/netfilter/nf_conntrack.h | 6 include/net/netfilter/nf_conntrack_l3proto.h | 4 include/net/netfilter/nf_conntrack_l4proto.h | 2 net/bridge/netfilter/ebt_log.c | 2 net/bridge/netfilter/ebt_ulog.c | 2 net/ipv4/netfilter/Kconfig | 8 net/ipv4/netfilter/ip_conntrack_core.c | 53 +++--- net/ipv4/netfilter/ip_conntrack_standalone.c | 9 - net/ipv4/netfilter/ip_nat_core.c | 74 ++++---- net/ipv4/netfilter/ipt_LOG.c | 2 net/ipv4/netfilter/ipt_ULOG.c | 2 net/ipv4/netfilter/nf_conntrack_proto_icmp.c | 2 net/ipv4/netfilter/nf_nat_core.c | 75 ++++---- net/ipv6/netfilter/Kconfig | 3 net/ipv6/netfilter/ip6t_LOG.c | 2 net/ipv6/netfilter/ip6t_mh.c | 7 net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c | 4 net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c | 1 net/netfilter/Kconfig | 12 + net/netfilter/core.c | 45 +++-- net/netfilter/nf_conntrack_core.c | 72 +++++--- net/netfilter/nf_conntrack_expect.c | 2 net/netfilter/nf_conntrack_ftp.c | 14 - net/netfilter/nf_conntrack_h323_main.c | 20 +- net/netfilter/nf_conntrack_irc.c | 2 net/netfilter/nf_conntrack_l3proto_generic.c | 2 net/netfilter/nf_conntrack_netbios_ns.c | 2 net/netfilter/nf_conntrack_netlink.c | 106 ++++++------ net/netfilter/nf_conntrack_pptp.c | 2 net/netfilter/nf_conntrack_proto.c | 73 ++------ net/netfilter/nf_conntrack_proto_sctp.c | 72 ++++---- net/netfilter/nf_conntrack_proto_tcp.c | 210 ++++++++++++------------- net/netfilter/nf_conntrack_sip.c | 6 net/netfilter/nf_conntrack_standalone.c | 2 net/netfilter/nf_conntrack_tftp.c | 2 net/netfilter/nf_internals.h | 2 net/netfilter/nf_log.c | 69 ++++---- net/netfilter/nf_queue.c | 22 +- net/netfilter/nf_sockopt.c | 12 - net/netfilter/nfnetlink.c | 10 - net/netfilter/nfnetlink_log.c | 54 +++--- net/netfilter/nfnetlink_queue.c | 84 +++++----- net/netfilter/x_tables.c | 12 - net/netfilter/xt_CLASSIFY.c | 12 - net/netfilter/xt_MARK.c | 4 net/netfilter/xt_NFQUEUE.c | 4 net/netfilter/xt_NOTRACK.c | 4 net/netfilter/xt_SECMARK.c | 2 net/netfilter/xt_conntrack.c | 32 +-- net/netfilter/xt_dccp.c | 20 +- net/netfilter/xt_hashlimit.c | 8 net/netfilter/xt_helper.c | 18 +- net/netfilter/xt_length.c | 4 net/netfilter/xt_limit.c | 2 net/netfilter/xt_mac.c | 7 net/netfilter/xt_mark.c | 6 net/netfilter/xt_multiport.c | 2 net/netfilter/xt_physdev.c | 2 net/netfilter/xt_policy.c | 24 +- net/netfilter/xt_quota.c | 4 net/netfilter/xt_realm.c | 2 net/netfilter/xt_sctp.c | 30 +-- net/netfilter/xt_string.c | 8 net/netfilter/xt_tcpmss.c | 4 66 files changed, 709 insertions(+), 667 deletions(-)
Masahide NAKAMURA:
[NETFILTER]: ip6t_mh: drop piggyback payload packet on MH packets
Patrick McHardy:
[NETFILTER]: Properly use RCU in nf_ct_attach
[NETFILTER]: Remove unnecessary synchronize_net() in nf_register_hook [NETFILTER]: Switch nf_register_afinfo/nf_unregister_afinfo to mutex [NETFILTER]: Switch nf_register_hook/nf_unregister_hook to mutex [NETFILTER]: nf_log: use rcu_assign_pointer for RCU protected pointer [NETFILTER]: nf_log: make nf_log_unregister_pf return void
[NETFILTER]: nf_log: switch logger registration/unregistration to mutex [NETFILTER]: nf_log: minor cleanups
[NETFILTER]: ip_nat: properly use RCU API for ip_nat_protos array [NETFILTER]: nf_nat: properly use RCU API for nf_nat_protos array [NETFILTER]: ip_conntrack: properly use RCU API for ip_ct_protos array [NETFILTER]: nf_conntrack: properly use RCU API for nf_ct_protos/nf_ct_l3protos arrays [NETFILTER]: ip_conntrack: fix invalid conntrack statistics RCU assumption [NETFILTER]: nf_conntrack: fix invalid conntrack statistics RCU assumption [NETFILTER]: ip_conntrack: properly use RCU for ip_conntrack_destroyed callback [NETFILTER]: nf_conntrack: properly use RCU for nf_conntrack_destroyed callback [NETFILTER]: nf_conntrack: change nf_conntrack_l[34]proto_unregister to void [NETFILTER]: xt_mac/xt_CLASSIFY: use IPv6 hook names for IPv6 registration [NETFILTER]: Kconfig: improve dependency handling
[NETFILTER]: nf_conntrack_tcp: make sysctl variables static
YOSHIFUJI Hideaki:
[NETFILTER]: Fix whitespace errors
PATCH 1/22 : Properly use RCU in nf_ct_attach
[NETFILTER]: Properly use RCU in nf_ct_attach
Use rcu_assign_pointer/rcu_dereference for ip_ct_attach pointer instead of self-made RCU and use rcu_read_lock to make sure the conntrack module doesn't disappear below us while calling it, since this function can be called from outside the netfilter hooks.
---
commit aa70e344f883b6fb4283024afa9af13376162ed3
tree f353ef51191c23f210f508b4edf1a7008e164e53
parent 70cdc55ed5432048c25025283f064323e3ae94f4
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:35:23 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:35:23 +0100
net/ipv4/netfilter/ip_conntrack_core.c | 4 ++-- net/netfilter/core.c | 9 ++++++--- net/netfilter/nf_conntrack_core.c | 4 ++-- 3 files changed, 10 insertions(+), 7 deletions(-)
net/ipv4/netfilter/ip_conntrack_core.c +2 -2
@@ -1355,7 +1355,7 @@ static void free_conntrack_hash(struct lsupposed to kill the mall. */ void ip_conntrack_cleanup (void ) { - ip_ct_attach =NULL ; +rcu_assign_pointer (ip_ct_attach,NULL );/* This makes sure all current packets have passed through netfilter framework. Roll on, two-stage module
@@ -1516,7 +1516,7 @@ int __init ip_conntrack_init(void)write_unlock_bh (&ip_conntrack_lock);/* For use by ipt_REJECT */ - ip_ct_attach = ip_conntrack_attach; +rcu_assign_pointer (ip_ct_attach, ip_conntrack_attach);/* Set up fake conntrack: - to never be deleted, not in any hashes */
net/netfilter/core.c +6 -3
@@ -248,9 +248,12 @@ void nf_ct_attach(struct sk_buff *new, s {void (*attach)(struct sk_buff *,struct sk_buff *); -if (skb->nfct && (attach = ip_ct_attach) !=NULL ) { -mb ();/* Just to be sure: must be read before executing this */ -attach (new, skb); +if (skb->nfct) { +rcu_read_lock (); + attach =rcu_dereference (ip_ct_attach); +if (attach) +attach (new, skb); +rcu_read_unlock (); } }EXPORT_SYMBOL (nf_ct_attach);
net/netfilter/nf_conntrack_core.c +2 -2
@@ -1105,7 +1105,7 @@ void nf_conntrack_cleanup(void) {int i; - ip_ct_attach =NULL ; +rcu_assign_pointer (ip_ct_attach,NULL );/* This makes sure all current packets have passed through netfilter framework. Roll on, two-stage module
@@ -1273,7 +1273,7 @@ int __init nf_conntrack_init(void)write_unlock_bh (&nf_conntrack_lock);/* For use by REJECT target */ - ip_ct_attach = __nf_conntrack_attach; +rcu_assign_pointer (ip_ct_attach, __nf_conntrack_attach);/* Set up fake conntrack: - to never be deleted, not in any hashes */
PATCH 2/22 : Remove unnecessary synchronize_net() in nf_register_hook
In reply to Patrick McHardy[NETFILTER]: Remove unnecessary synchronize_net() in nf_register_hook
We're only adding to the list, no need to synchronize.
---
commit 4e6bba5661fe72e2c570701be4fcf4f2fabfdff2
tree 12e74ddf2a81a31aa4b5d59ba39c6fd0b14d7f5d
parent aa70e344f883b6fb4283024afa9af13376162ed3
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:35:30 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:35:30 +0100
net/netfilter/core.c | 2 -- 1 files changed, 0 insertions(+), 2 deletions(-)
net/netfilter/core.c +0 -2
@@ -69,8 +69,6 @@ int nf_register_hook(struct nf_hook_ops }list_add_rcu (®->list, i->prev);spin_unlock_bh (&nf_hook_lock); - -synchronize_net ();return 0 ; }EXPORT_SYMBOL (nf_register_hook);
PATCH 3/22 : Switch nf_register_afinfo/nf_unregister_afinfo to mutex
In reply to Patrick McHardy[NETFILTER]: Switch nf_register_afinfo/nf_unregister_afinfo to mutex
The spinlock is only used in process context (register/unregister), switch to a mutex.
---
commit ccead672bf5800d7418940aea9c0e6a231bc1f09
tree 9908ac59b324b764b4e7caaf17dc8472df8e52d9
parent 4e6bba5661fe72e2c570701be4fcf4f2fabfdff2
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:36:15 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:36:15 +0100
net/netfilter/core.c | 15 ++++++++++----- 1 files changed, 10 insertions(+), 5 deletions(-)
net/netfilter/core.c +10 -5
@@ -22,29 +22,34 @@ #include <linux/if.h>#include <linux/netdevice.h> #include <linux/inetdevice.h> #include <linux/proc_fs.h> +#include <linux/mutex.h> #include <net/sock.h> #include "nf_internals.h" -static DEFINE_SPINLOCK (afinfo_lock); +static DEFINE_MUTEX (afinfo_mutex);struct nf_afinfo *nf_afinfo[NPROTO ]__read_mostly ;EXPORT_SYMBOL (nf_afinfo);int nf_register_afinfo (struct nf_afinfo *afinfo) { -spin_lock (&afinfo_lock); +int err; + + err =mutex_lock_interruptible (&afinfo_mutex); +if (err <0 ) +return err;rcu_assign_pointer (nf_afinfo[afinfo->family], afinfo); -spin_unlock (&afinfo_lock); +mutex_unlock (&afinfo_mutex);return 0 ; }EXPORT_SYMBOL_GPL (nf_register_afinfo);void nf_unregister_afinfo (struct nf_afinfo *afinfo) { -spin_lock (&afinfo_lock); +mutex_lock (&afinfo_mutex);rcu_assign_pointer (nf_afinfo[afinfo->family],NULL ); -spin_unlock (&afinfo_lock); +mutex_unlock (&afinfo_mutex);synchronize_rcu (); }EXPORT_SYMBOL_GPL (nf_unregister_afinfo);
PATCH 4/22 : Switch nf_register_hook/nf_unregister_hook to mutex
In reply to Patrick McHardy[NETFILTER]: Switch nf_register_hook/nf_unregister_hook to mutex
The spinlock is only used in process context (register/unregister) since RCU is used for the nf_hook lists, switch to a mutex.
---
commit 4694a2599998f08af77aa1a1ee2c6d6bc2020913
tree b5d2752f7e5f66afd4b7d4bcd2402f1e1c9bfae9
parent ccead672bf5800d7418940aea9c0e6a231bc1f09
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:36:19 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:36:19 +0100
net/netfilter/core.c | 13 ++++++++----- 1 files changed, 8 insertions(+), 5 deletions(-)
net/netfilter/core.c +8 -5
@@ -61,28 +61,31 @@ EXPORT_SYMBOL_GPL(nf_unregister_afinfo);* packets come back: if the hook is gone, the packet is discarded. */ struct list_head nf_hooks[NPROTO ][NF_MAX_HOOKS ]__read_mostly ;EXPORT_SYMBOL (nf_hooks); -static DEFINE_SPINLOCK (nf_hook_lock); +static DEFINE_MUTEX (nf_hook_mutex);int nf_register_hook (struct nf_hook_ops *reg) {struct list_head *i; +int err; -spin_lock_bh (&nf_hook_lock); + err =mutex_lock_interruptible (&nf_hook_mutex); +if (err <0 ) +return err;list_for_each (i, &nf_hooks[reg->pf][reg->hooknum]) {if (reg->priority < ((struct nf_hook_ops *)i)->priority)break ; }list_add_rcu (®->list, i->prev); -spin_unlock_bh (&nf_hook_lock); +mutex_unlock (&nf_hook_mutex);return 0 ; }EXPORT_SYMBOL (nf_register_hook);void nf_unregister_hook (struct nf_hook_ops *reg) { -spin_lock_bh (&nf_hook_lock); +mutex_lock (&nf_hook_mutex);list_del_rcu (®->list); -spin_unlock_bh (&nf_hook_lock); +mutex_unlock (&nf_hook_mutex);synchronize_net (); }
PATCH 5/22 : nf_log: use rcu_assign_pointer for RCU protected pointer
In reply to Patrick McHardy[NETFILTER]: nf_log: use rcu_assign_pointer for RCU protected pointer
Also replace synchronize_net() calls by synchronize_rcu() since the RCU protected data is also used for sysfs.
---
commit a3d379f952d0861a1e0d876f69f301d5174f7990
tree 2f439287b3e9388ef7f186ea108868a05f119f4d
parent 4694a2599998f08af77aa1a1ee2c6d6bc2020913
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:36:34 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:36:34 +0100
net/netfilter/nf_log.c | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-)
net/netfilter/nf_log.c +4 -4
@@ -46,11 +46,11 @@ int nf_log_unregister_pf(int pf)return -EINVAL ;spin_lock (&nf_log_lock); - nf_logging[pf] =NULL ; +rcu_assign_pointer (nf_logging[pf],NULL );spin_unlock (&nf_log_lock);/* Give time to concurrent readers. */ -synchronize_net (); +synchronize_rcu ();return 0 ; }
@@ -63,11 +63,11 @@ void nf_log_unregister_logger(struct nf_spin_lock (&nf_log_lock);for (i =0 ; i <NPROTO ; i++) {if (nf_logging[i] == logger) - nf_logging[i] =NULL ; +rcu_assign_pointer (nf_logging[i],NULL ); }spin_unlock (&nf_log_lock); -synchronize_net (); +synchronize_rcu (); }EXPORT_SYMBOL (nf_log_unregister_logger);
PATCH 6/22 : nf_log: make nf_log_unregister_pf return void
In reply to Patrick McHardy[NETFILTER]: nf_log: make nf_log_unregister_pf return void
Since the only user of nf_log_unregister_pf (nfnetlink_log) doesn't check the return value, change it to void and bail out silently when a non-existant address family is supplied.
---
commit ca4c6357b4568c1b8644385828a542475a86c5fa
tree f5068e54682875208ff5d05cf7af434094daf545
parent a3d379f952d0861a1e0d876f69f301d5174f7990
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:37:02 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:37:02 +0100
include/linux/netfilter.h | 2 +- net/netfilter/nf_log.c | 7 ++----- 2 files changed, 3 insertions(+), 6 deletions(-)
include/linux/netfilter.h +1 -1
@@ -172,7 +172,7 @@ struct nf_logger {/* Function to register/unregister log function. */ int nf_log_register (int pf,struct nf_logger *logger); -int nf_log_unregister_pf (int pf); +void nf_log_unregister_pf (int pf);void nf_log_unregister_logger (struct nf_logger *logger);/* Calls the registered backend logging function */
net/netfilter/nf_log.c +2 -5
@@ -40,19 +40,16 @@ int nf_log_register(int pf, struct nf_lo }EXPORT_SYMBOL (nf_log_register); -int nf_log_unregister_pf (int pf) +void nf_log_unregister_pf (int pf) {if (pf >=NPROTO ) -return -EINVAL ; - +return ;spin_lock (&nf_log_lock);rcu_assign_pointer (nf_logging[pf],NULL );spin_unlock (&nf_log_lock);/* Give time to concurrent readers. */ synchronize_rcu (); - -return 0 ; }EXPORT_SYMBOL (nf_log_unregister_pf);
PATCH 7/22 : nf_log: switch logger registration/unregistration to mutex
In reply to Patrick McHardy[NETFILTER]: nf_log: switch logger registration/unregistration to mutex
The spinlock is only used in process context (register/unregister), switch to a mutex.
---
commit 99c46fbd857a8a2d01987650bd1492b5379ac250
tree c91f6d98b4d9035255bf89e8153745eec7cb9349
parent ca4c6357b4568c1b8644385828a542475a86c5fa
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:37:13 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:37:13 +0100
net/netfilter/nf_log.c | 26 +++++++++++++++----------- 1 files changed, 15 insertions(+), 11 deletions(-)
net/netfilter/nf_log.c +15 -11
@@ -15,27 +15,31 @@ #include "nf_internals.h"#define NF_LOG_PREFIXLEN 128static struct nf_logger *nf_logging[NPROTO ];/* = NULL */ -static DEFINE_SPINLOCK (nf_log_lock); +static DEFINE_MUTEX (nf_log_mutex);/* return EBUSY if somebody else is registered, EEXIST if the same logger * is registred, 0 on success. */ int nf_log_register (int pf,struct nf_logger *logger) { -int ret = -EBUSY ; +int ret;if (pf >=NPROTO )return -EINVAL ;/* Any setup of logging members must be done before * substituting pointer. */ -spin_lock (&nf_log_lock); -if (!nf_logging[pf]) { + ret =mutex_lock_interruptible (&nf_log_mutex); +if (ret <0 ) +return ret; + +if (!nf_logging[pf])rcu_assign_pointer (nf_logging[pf], logger); - ret =0 ; - }else if (nf_logging[pf] == logger) +else if (nf_logging[pf] == logger) ret = -EEXIST ; +else + ret = -EBUSY ; -spin_unlock (&nf_log_lock); +mutex_unlock (&nf_log_mutex);return ret; }EXPORT_SYMBOL (nf_log_register);
@@ -44,9 +48,9 @@ void nf_log_unregister_pf(int pf) {if (pf >=NPROTO )return ; -spin_lock (&nf_log_lock); +mutex_lock (&nf_log_mutex);rcu_assign_pointer (nf_logging[pf],NULL ); -spin_unlock (&nf_log_lock); +mutex_unlock (&nf_log_mutex);/* Give time to concurrent readers. */ synchronize_rcu ();
@@ -57,12 +61,12 @@ void nf_log_unregister_logger(struct nf_ {int i; -spin_lock (&nf_log_lock); +mutex_lock (&nf_log_mutex);for (i =0 ; i <NPROTO ; i++) {if (nf_logging[i] == logger)rcu_assign_pointer (nf_logging[i],NULL ); } -spin_unlock (&nf_log_lock); +mutex_unlock (&nf_log_mutex);synchronize_rcu (); }
PATCH 8/22 : nf_log: minor cleanups
In reply to Patrick McHardy[NETFILTER]: nf_log: minor cleanups
- rename nf_logging to nf_loggers since its an array of registered loggers
- rename nf_log_unregister_logger() to nf_log_unregister() to make it symetrical to nf_log_register() and convert all users
---
commit 5187c6343d6eac0b2fc995bfb607f817565ec258
tree c0290f3d2891089ca4f7d231cd21d3e081890ddc
parent 99c46fbd857a8a2d01987650bd1492b5379ac250
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:37:47 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:37:47 +0100
include/linux/netfilter.h | 2 +- net/bridge/netfilter/ebt_log.c | 2 +- net/bridge/netfilter/ebt_ulog.c | 2 +- net/ipv4/netfilter/ipt_LOG.c | 2 +- net/ipv4/netfilter/ipt_ULOG.c | 2 +- net/ipv6/netfilter/ip6t_LOG.c | 2 +- net/netfilter/nf_log.c | 22 +++++++++++----------- net/netfilter/nfnetlink_log.c | 2 +- 8 files changed, 18 insertions(+), 18 deletions(-)
include/linux/netfilter.h +1 -1
@@ -172,8 +172,8 @@ struct nf_logger {/* Function to register/unregister log function. */ int nf_log_register (int pf,struct nf_logger *logger); +void nf_log_unregister (struct nf_logger *logger);void nf_log_unregister_pf (int pf); -void nf_log_unregister_logger (struct nf_logger *logger);/* Calls the registered backend logging function */ void nf_log_packet(int pf,
net/bridge/netfilter/ebt_log.c +1 -1
@@ -208,7 +208,7 @@ static int __init ebt_log_init(void) static void __exit ebt_log_fini(void) { -nf_log_unregister_logger (&ebt_log_logger); +nf_log_unregister (&ebt_log_logger);ebt_unregister_watcher (&log); }
net/bridge/netfilter/ebt_ulog.c +1 -1
@@ -323,7 +323,7 @@ static void __exit ebt_ulog_fini(void)ebt_ulog_buff_t *ub;int i; -nf_log_unregister_logger (&ebt_ulog_logger); +nf_log_unregister (&ebt_ulog_logger);ebt_unregister_watcher (&ulog);for (i =0 ; i <EBT_ULOG_MAXNLGROUPS ; i++) { ub = &ulog_buffers[i];
net/ipv4/netfilter/ipt_LOG.c +1 -1
@@ -489,7 +489,7 @@ static int __init ipt_log_init(void) static void __exit ipt_log_fini(void) { -nf_log_unregister_logger (&ipt_log_logger); +nf_log_unregister (&ipt_log_logger);xt_unregister_target (&ipt_log_reg); }
net/ipv4/netfilter/ipt_ULOG.c +1 -1
@@ -419,7 +419,7 @@ static void __exit ipt_ulog_fini(void)DEBUGP ("ipt_ULOG: cleanup_module\n" );if (nflog) -nf_log_unregister_logger (&ipt_ulog_logger); +nf_log_unregister (&ipt_ulog_logger);xt_unregister_target (&ipt_ulog_reg);sock_release (nflognl->sk_socket);
net/ipv6/netfilter/ip6t_LOG.c +1 -1
@@ -501,7 +501,7 @@ static int __init ip6t_log_init(void) static void __exit ip6t_log_fini(void) { -nf_log_unregister_logger (&ip6t_logger); +nf_log_unregister (&ip6t_logger);xt_unregister_target (&ip6t_log_reg); }
net/netfilter/nf_log.c +11 -11
@@ -14,7 +14,7 @@ #include "nf_internals.h"#define NF_LOG_PREFIXLEN 128 -static struct nf_logger *nf_logging[NPROTO ];/* = NULL */ +static struct nf_logger *nf_loggers[NPROTO ];static DEFINE_MUTEX (nf_log_mutex);/* return EBUSY if somebody else is registered, EEXIST if the same logger
@@ -32,9 +32,9 @@ int nf_log_register(int pf, struct nf_loif (ret <0 )return ret; -if (!nf_logging[pf]) -rcu_assign_pointer (nf_logging[pf], logger); -else if (nf_logging[pf] == logger) +if (!nf_loggers[pf]) +rcu_assign_pointer (nf_loggers[pf], logger); +else if (nf_loggers[pf] == logger) ret = -EEXIST ;else ret = -EBUSY ;
@@ -49,7 +49,7 @@ void nf_log_unregister_pf(int pf)if (pf >=NPROTO )return ;mutex_lock (&nf_log_mutex); -rcu_assign_pointer (nf_logging[pf],NULL ); +rcu_assign_pointer (nf_loggers[pf],NULL );mutex_unlock (&nf_log_mutex);/* Give time to concurrent readers. */
@@ -57,20 +57,20 @@ void nf_log_unregister_pf(int pf) }EXPORT_SYMBOL (nf_log_unregister_pf); -void nf_log_unregister_logger (struct nf_logger *logger) +void nf_log_unregister (struct nf_logger *logger) {int i;mutex_lock (&nf_log_mutex);for (i =0 ; i <NPROTO ; i++) { -if (nf_logging[i] == logger) -rcu_assign_pointer (nf_logging[i],NULL ); +if (nf_loggers[i] == logger) +rcu_assign_pointer (nf_loggers[i],NULL ); }mutex_unlock (&nf_log_mutex);synchronize_rcu (); } -EXPORT_SYMBOL (nf_log_unregister_logger); +EXPORT_SYMBOL (nf_log_unregister); void nf_log_packet(int pf, unsigned int hooknum,
@@ -85,7 +85,7 @@ void nf_log_packet(int pf,struct nf_logger *logger;rcu_read_lock (); - logger =rcu_dereference (nf_logging[pf]); + logger =rcu_dereference (nf_loggers[pf]);if (logger) {va_start (args, fmt);vsnprintf (prefix,sizeof (prefix), fmt, args);
@@ -132,7 +132,7 @@ static int seq_show(struct seq_file *s,loff_t *pos = v;const struct nf_logger *logger; - logger =rcu_dereference (nf_logging[*pos]); + logger =rcu_dereference (nf_loggers[*pos]);if (!logger)return seq_printf (s,"%2lld NONE\n" , *pos);
net/netfilter/nfnetlink_log.c +1 -1
@@ -1077,7 +1077,7 @@ cleanup_netlink_notifier: static void __exit nfnetlink_log_fini(void) { -nf_log_unregister_logger (&nfulnl_logger); +nf_log_unregister (&nfulnl_logger);#ifdef CONFIG_PROC_FS remove_proc_entry ("nfnetlink_log" , proc_net_netfilter);#endif
PATCH 9/22 : ip_nat: properly use RCU API for ip_nat_protos array
In reply to Patrick McHardy[NETFILTER]: ip_nat: properly use RCU API for ip_nat_protos array
Replace preempt_{enable,disable} based RCU by proper use of the RCU API and add missing rcu_read_lock/rcu_read_unlock calls in paths used outside of packet processing context (nfnetlink_conntrack).
---
commit 6930ac084caa4aff8d702716b80a2b59212efb78
tree 527cf137f820f9aeecacc7fbfe6621bef49a333f
parent 5187c6343d6eac0b2fc995bfb607f817565ec258
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:43:27 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:43:27 +0100
net/ipv4/netfilter/ip_nat_core.c | 61 ++++++++++++++++++-------------------- 1 files changed, 29 insertions(+), 32 deletions(-)
net/ipv4/netfilter/ip_nat_core.c +29 -32
@@ -50,7 +50,7 @@ static struct ip_nat_protocol *ip_nat_prstatic inline struct ip_nat_protocol *__ip_nat_proto_find (u_int8_t protonum) { -return ip_nat_protos[protonum]; +return rcu_dereference (ip_nat_protos[protonum]); } struct ip_nat_protocol *
@@ -58,13 +58,11 @@ ip_nat_proto_find_get(u_int8_t protonum) {struct ip_nat_protocol *p; -/* we need to disable preemption to make sure 'p' doesn't get -* removed until we've grabbed the reference */ -preempt_disable (); +rcu_read_lock (); p =__ip_nat_proto_find (protonum);if (!try_module_get (p->me)) p = &ip_nat_unknown_protocol; -preempt_enable (); +rcu_read_unlock ();return p; }
@@ -120,8 +118,8 @@ static int in_range(const struct ip_conntrack_tuple *tuple, const struct ip_nat_range *range) { - struct ip_nat_protocol *proto = -__ip_nat_proto_find (tuple->dst.protonum); + struct ip_nat_protocol *proto; +int ret =0 ;/* If we are supposed to map IPs, then we must be in the range specified, otherwise let this drag us onto a new src IP. */
@@ -131,12 +129,15 @@ in_range(const struct ip_conntrack_tuplereturn 0 ; } +rcu_read_lock (); + proto =__ip_nat_proto_find (tuple->dst.protonum);if (!(range->flags &IP_NAT_RANGE_PROTO_SPECIFIED ) || proto->in_range (tuple,IP_NAT_MANIP_SRC , &range->min, &range->max)) -return 1 ; + ret =1 ; +rcu_read_unlock (); -return 0 ; +return ret; } static inline int
@@ -260,27 +261,25 @@ get_unique_tuple(struct ip_conntrack_tup/* 3) The per-protocol part of the manip is made to map into the range to make a unique tuple. */ - proto =ip_nat_proto_find_get (orig_tuple->dst.protonum); +rcu_read_lock (); + proto =__ip_nat_proto_find (orig_tuple->dst.protonum);/* Change protocol info to have some randomization */ if (range->flags &IP_NAT_RANGE_PROTO_RANDOM ) { proto->unique_tuple (tuple, range, maniptype, conntrack); -ip_nat_proto_put (proto); -return ; + goto out; }/* Only bother mapping if it's not already in range and unique */ if ((!(range->flags &IP_NAT_RANGE_PROTO_SPECIFIED ) || proto->in_range (tuple, maniptype, &range->min, &range->max)) - && !ip_nat_used_tuple (tuple, conntrack)) { -ip_nat_proto_put (proto); -return ; - } + && !ip_nat_used_tuple (tuple, conntrack)) + goto out;/* Last change: get protocol to try to obtain unique tuple. */ proto->unique_tuple (tuple, range, maniptype, conntrack); - -ip_nat_proto_put (proto); +out: +rcu_read_unlock (); }unsigned int
@@ -360,12 +359,11 @@ manip_pkt(u_int16_t proto, iph = (void *)(*pskb)->data + iphdroff;/* Manipulate protcol part. */ - p =ip_nat_proto_find_get (proto); -if (!p->manip_pkt (pskb, iphdroff, target, maniptype)) { -ip_nat_proto_put (p); + +/* rcu_read_lock()ed by nf_hook_slow */ + p =__ip_nat_proto_find (proto); +if (!p->manip_pkt (pskb, iphdroff, target, maniptype))return 0 ; - } -ip_nat_proto_put (p); iph = (void *)(*pskb)->data + iphdroff;
@@ -515,7 +513,7 @@ int ip_nat_protocol_register(struct ip_n ret = -EBUSY ; goto out; } - ip_nat_protos[proto->protonum] = proto; +rcu_assign_pointer (ip_nat_protos[proto->protonum], proto); out:write_unlock_bh (&ip_nat_lock);return ret;
@@ -526,11 +524,10 @@ EXPORT_SYMBOL(ip_nat_protocol_register);void ip_nat_protocol_unregister (struct ip_nat_protocol *proto) {write_lock_bh (&ip_nat_lock); - ip_nat_protos[proto->protonum] = &ip_nat_unknown_protocol; +rcu_assign_pointer (ip_nat_protos[proto->protonum], + &ip_nat_unknown_protocol);write_unlock_bh (&ip_nat_lock); - -/* Someone could be still looking at the proto in a bh. */ -synchronize_net (); +synchronize_rcu (); }EXPORT_SYMBOL (ip_nat_protocol_unregister);
@@ -594,10 +591,10 @@ static int __init ip_nat_init(void)/* Sew in builtin protocols. */ write_lock_bh (&ip_nat_lock);for (i =0 ; i <MAX_IP_NAT_PROTO ; i++) - ip_nat_protos[i] = &ip_nat_unknown_protocol; - ip_nat_protos[IPPROTO_TCP ] = &ip_nat_protocol_tcp; - ip_nat_protos[IPPROTO_UDP ] = &ip_nat_protocol_udp; - ip_nat_protos[IPPROTO_ICMP ] = &ip_nat_protocol_icmp; +rcu_assign_pointer (ip_nat_protos[i], &ip_nat_unknown_protocol); +rcu_assign_pointer (ip_nat_protos[IPPROTO_TCP ], &ip_nat_protocol_tcp); +rcu_assign_pointer (ip_nat_protos[IPPROTO_UDP ], &ip_nat_protocol_udp); +rcu_assign_pointer (ip_nat_protos[IPPROTO_ICMP ], &ip_nat_protocol_icmp);write_unlock_bh (&ip_nat_lock);for (i =0 ; i < ip_nat_htable_size; i++) {
PATCH 10/22 : nf_nat: properly use RCU API for nf_nat_protos array
In reply to Patrick McHardy[NETFILTER]: nf_nat: properly use RCU API for nf_nat_protos array
Replace preempt_{enable,disable} based RCU by proper use of the RCU API and add missing rcu_read_lock/rcu_read_unlock calls in paths used outside of packet processing context (nfnetlink_conntrack).
---
commit 8592a4e696c2d452053b005a7b8a84e29a7462af
tree c900a38cb8b11b3f4c78776f416cec8b7adbfd6c
parent 6930ac084caa4aff8d702716b80a2b59212efb78
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:44:58 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:44:58 +0100
net/ipv4/netfilter/nf_nat_core.c | 59 ++++++++++++++++++-------------------- 1 files changed, 28 insertions(+), 31 deletions(-)
net/ipv4/netfilter/nf_nat_core.c +28 -31
@@ -53,7 +53,7 @@ static struct nf_nat_protocol *nf_nat_prstatic inline struct nf_nat_protocol *__nf_nat_proto_find (u_int8_t protonum) { -return nf_nat_protos[protonum]; +return rcu_dereference (nf_nat_protos[protonum]); } struct nf_nat_protocol *
@@ -61,13 +61,11 @@ nf_nat_proto_find_get(u_int8_t protonum) {struct nf_nat_protocol *p; -/* we need to disable preemption to make sure 'p' doesn't get -* removed until we've grabbed the reference */ -preempt_disable (); +rcu_read_lock (); p =__nf_nat_proto_find (protonum);if (!try_module_get (p->me)) p = &nf_nat_unknown_protocol; -preempt_enable (); +rcu_read_unlock ();return p; }
@@ -126,8 +124,8 @@ in_range(const struct nf_conntrack_tuple const struct nf_nat_range *range) {struct nf_nat_protocol *proto; +int ret =0 ; - proto =__nf_nat_proto_find (tuple->dst.protonum);/* If we are supposed to map IPs, then we must be in the range specified, otherwise let this drag us onto a new src IP. */ if (range->flags &IP_NAT_RANGE_MAP_IPS ) {
@@ -136,12 +134,15 @@ in_range(const struct nf_conntrack_tuplereturn 0 ; } +rcu_read_lock (); + proto =__nf_nat_proto_find (tuple->dst.protonum);if (!(range->flags &IP_NAT_RANGE_PROTO_SPECIFIED ) || proto->in_range (tuple,IP_NAT_MANIP_SRC , &range->min, &range->max)) -return 1 ; + ret =1 ; +rcu_read_unlock (); -return 0 ; +return ret; } static inline int
@@ -268,27 +269,25 @@ get_unique_tuple(struct nf_conntrack_tup/* 3) The per-protocol part of the manip is made to map into the range to make a unique tuple. */ - proto =nf_nat_proto_find_get (orig_tuple->dst.protonum); +rcu_read_lock (); + proto =__nf_nat_proto_find (orig_tuple->dst.protonum);/* Change protocol info to have some randomization */ if (range->flags &IP_NAT_RANGE_PROTO_RANDOM ) { proto->unique_tuple (tuple, range, maniptype, ct); -nf_nat_proto_put (proto); -return ; + goto out; }/* Only bother mapping if it's not already in range and unique */ if ((!(range->flags &IP_NAT_RANGE_PROTO_SPECIFIED ) || proto->in_range (tuple, maniptype, &range->min, &range->max)) && - !nf_nat_used_tuple (tuple, ct)) { -nf_nat_proto_put (proto); -return ; - } + !nf_nat_used_tuple (tuple, ct)) + goto out;/* Last change: get protocol to try to obtain unique tuple. */ proto->unique_tuple (tuple, range, maniptype, ct); - -nf_nat_proto_put (proto); +out: +rcu_read_unlock (); }unsigned int
@@ -369,12 +368,11 @@ manip_pkt(u_int16_t proto, iph = (void *)(*pskb)->data + iphdroff;/* Manipulate protcol part. */ - p =nf_nat_proto_find_get (proto); -if (!p->manip_pkt (pskb, iphdroff, target, maniptype)) { -nf_nat_proto_put (p); + +/* rcu_read_lock()ed by nf_hook_slow */ + p =__nf_nat_proto_find (proto); +if (!p->manip_pkt (pskb, iphdroff, target, maniptype))return 0 ; - } -nf_nat_proto_put (p); iph = (void *)(*pskb)->data + iphdroff;
@@ -529,7 +527,7 @@ int nf_nat_protocol_register(struct nf_n ret = -EBUSY ; goto out; } - nf_nat_protos[proto->protonum] = proto; +rcu_assign_pointer (nf_nat_protos[proto->protonum], proto); out:write_unlock_bh (&nf_nat_lock);return ret;
@@ -540,11 +538,10 @@ EXPORT_SYMBOL(nf_nat_protocol_register);void nf_nat_protocol_unregister (struct nf_nat_protocol *proto) {write_lock_bh (&nf_nat_lock); - nf_nat_protos[proto->protonum] = &nf_nat_unknown_protocol; +rcu_assign_pointer (nf_nat_protos[proto->protonum], + &nf_nat_unknown_protocol);write_unlock_bh (&nf_nat_lock); - -/* Someone could be still looking at the proto in a bh. */ -synchronize_net (); +synchronize_rcu (); }EXPORT_SYMBOL (nf_nat_protocol_unregister);
@@ -608,10 +605,10 @@ static int __init nf_nat_init(void)/* Sew in builtin protocols. */ write_lock_bh (&nf_nat_lock);for (i =0 ; i <MAX_IP_NAT_PROTO ; i++) - nf_nat_protos[i] = &nf_nat_unknown_protocol; - nf_nat_protos[IPPROTO_TCP ] = &nf_nat_protocol_tcp; - nf_nat_protos[IPPROTO_UDP ] = &nf_nat_protocol_udp; - nf_nat_protos[IPPROTO_ICMP ] = &nf_nat_protocol_icmp; +rcu_assign_pointer (nf_nat_protos[i], &nf_nat_unknown_protocol); +rcu_assign_pointer (nf_nat_protos[IPPROTO_TCP ], &nf_nat_protocol_tcp); +rcu_assign_pointer (nf_nat_protos[IPPROTO_UDP ], &nf_nat_protocol_udp); +rcu_assign_pointer (nf_nat_protos[IPPROTO_ICMP ], &nf_nat_protocol_icmp);write_unlock_bh (&nf_nat_lock);for (i =0 ; i < nf_nat_htable_size; i++) {
PATCH 11/22 : ip_conntrack: properly use RCU API for ip_ct_protos array
In reply to Patrick McHardy[NETFILTER]: ip_conntrack: properly use RCU API for ip_ct_protos array
Replace preempt_{enable,disable} based RCU by proper use of the RCU API and add missing rcu_read_lock/rcu_read_unlock calls in all paths not obviously only used within packet process context (nfnetlink_conntrack).
---
commit f99ec74e0c460f0969f028c2853d2d31178ba3aa
tree 4a0534a350b1f4067cc2c4b1ef450dace12b9da7
parent 8592a4e696c2d452053b005a7b8a84e29a7462af
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:56:31 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 10:56:31 +0100
net/ipv4/netfilter/ip_conntrack_core.c | 26 ++++++++++++++++++-------- net/ipv4/netfilter/ip_conntrack_standalone.c | 9 ++++----- net/ipv4/netfilter/ip_nat_core.c | 6 ++++-- 3 files changed, 26 insertions(+), 15 deletions(-)
net/ipv4/netfilter/ip_conntrack_core.c +18 -8
@@ -318,9 +318,11 @@ destroy_conntrack(struct nf_conntrack *n/* To make sure we don't get any weird locking issues here: * destroy_conntrack() MUST NOT be called with a write lock * to ip_conntrack_lock!!! -HW */ +rcu_read_lock (); proto =__ip_conntrack_proto_find (ct->tuplehash[IP_CT_DIR_REPLY ].tuple.dst.protonum);if (proto && proto->destroy) proto->destroy (ct); +rcu_read_unlock ();if (ip_conntrack_destroyed)ip_conntrack_destroyed (ct);
@@ -595,13 +597,13 @@ ip_conntrack_proto_find_get(u_int8_t pro {struct ip_conntrack_protocol *p; -preempt_disable (); +rcu_read_lock (); p =__ip_conntrack_proto_find (protocol);if (p) {if (!try_module_get (p->me)) p = &ip_conntrack_generic_protocol; } -preempt_enable (); +rcu_read_unlock ();return p; }
@@ -831,6 +833,7 @@ #if 0 }#endif +/* rcu_read_lock()ed by nf_hook_slow */ proto =__ip_conntrack_proto_find ((*pskb)->nh.iph->protocol);/* It may be an special packet, error, unclean...
@@ -876,8 +879,15 @@ #endifint invert_tuplepr (struct ip_conntrack_tuple *inverse,const struct ip_conntrack_tuple *orig) { -return ip_ct_invert_tuple (inverse, orig, -__ip_conntrack_proto_find (orig->dst.protonum)); +struct ip_conntrack_protocol *proto; +int ret; + +rcu_read_lock (); + proto =__ip_conntrack_proto_find (orig->dst.protonum); + ret =ip_ct_invert_tuple (inverse, orig, proto); +rcu_read_unlock (); + +return ret; }/* Would two expected things clash? */
@@ -1508,11 +1518,11 @@ int __init ip_conntrack_init(void)/* Don't NEED lock here, but good form anyway. */ write_lock_bh (&ip_conntrack_lock);for (i =0 ; i <MAX_IP_CT_PROTO ; i++) - ip_ct_protos[i] = &ip_conntrack_generic_protocol; +rcu_assign_pointer (ip_ct_protos[i], &ip_conntrack_generic_protocol);/* Sew in builtin protocols. */ - ip_ct_protos[IPPROTO_TCP ] = &ip_conntrack_protocol_tcp; - ip_ct_protos[IPPROTO_UDP ] = &ip_conntrack_protocol_udp; - ip_ct_protos[IPPROTO_ICMP ] = &ip_conntrack_protocol_icmp; +rcu_assign_pointer (ip_ct_protos[IPPROTO_TCP ], &ip_conntrack_protocol_tcp); +rcu_assign_pointer (ip_ct_protos[IPPROTO_UDP ], &ip_conntrack_protocol_udp); +rcu_assign_pointer (ip_ct_protos[IPPROTO_ICMP ], &ip_conntrack_protocol_icmp);write_unlock_bh (&ip_conntrack_lock);/* For use by ipt_REJECT */
net/ipv4/netfilter/ip_conntrack_standalone.c +4 -5
@@ -796,7 +796,7 @@ int ip_conntrack_protocol_register(struc ret = -EBUSY ; goto out; } - ip_ct_protos[proto->proto] = proto; +rcu_assign_pointer (ip_ct_protos[proto->proto], proto); out:write_unlock_bh (&ip_conntrack_lock);return ret;
@@ -805,11 +805,10 @@ int ip_conntrack_protocol_register(strucvoid ip_conntrack_protocol_unregister (struct ip_conntrack_protocol *proto) {write_lock_bh (&ip_conntrack_lock); - ip_ct_protos[proto->proto] = &ip_conntrack_generic_protocol; +rcu_assign_pointer (ip_ct_protos[proto->proto], + &ip_conntrack_generic_protocol);write_unlock_bh (&ip_conntrack_lock); - -/* Somebody could be still looking at the proto in bh. */ -synchronize_net (); +synchronize_rcu ();/* Remove all contrack entries for this protocol */ ip_ct_iterate_cleanup (kill_proto, &proto->proto);
net/ipv4/netfilter/ip_nat_core.c +4 -2
@@ -420,6 +420,7 @@ int ip_nat_icmp_reply_translation(structstruct icmphdr icmp;struct iphdr ip; } *inside; +struct ip_conntrack_protocol *proto;struct ip_conntrack_tuple inner, target;int hdrlen = (*pskb)->nh.iph->ihl *4 ;enum ip_conntrack_dir dir =CTINFO2DIR (ctinfo);
@@ -455,10 +456,11 @@ int ip_nat_icmp_reply_translation(structDEBUGP ("icmp_reply_translation: translating error %p manp %u dir %s\n" , *pskb, manip, dir ==IP_CT_DIR_ORIGINAL ?"ORIG" :"REPLY" ); +/* rcu_read_lock()ed by nf_hook_slow */ + proto =__ip_conntrack_proto_find (inside->ip.protocol);if (!ip_ct_get_tuple (&inside->ip, *pskb, (*pskb)->nh.iph->ihl*4 +sizeof (struct icmphdr ) + inside->ip.ihl*4 , - &inner, -__ip_conntrack_proto_find (inside->ip.protocol))) + &inner, proto))return 0 ;/* Change inner back to look like incoming packet. We do the
PATCH 12/22 : nf_conntrack: properly use RCU API for nf_ct_protos/nf_ct_l3protos arrays
In reply to Patrick McHardy[NETFILTER]: nf_conntrack: properly use RCU API for nf_ct_protos/nf_ct_l3protos arrays
Replace preempt_{enable,disable} based RCU by proper use of the RCU API and add missing rcu_read_lock/rcu_read_unlock calls in all paths not obviously only used within packet process context (nfnetlink_conntrack).
---
commit 2f34e7966f36432f6915ab7eee85116439a3944e
tree 829a240f1c040ef8a712d7e159adb473934320f7
parent f99ec74e0c460f0969f028c2853d2d31178ba3aa
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:01:11 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:01:11 +0100
include/net/netfilter/nf_conntrack_l3proto.h | 2 + net/ipv4/netfilter/nf_conntrack_proto_icmp.c | 2 + net/ipv4/netfilter/nf_nat_core.c | 9 ++++--- net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c | 1 + net/netfilter/nf_conntrack_core.c | 27 +++++++++++++++----- net/netfilter/nf_conntrack_proto.c | 33 ++++++++++-------------- 6 files changed, 44 insertions(+), 30 deletions(-)
include/net/netfilter/nf_conntrack_l3proto.h +1 -1
@@ -106,7 +106,7 @@ __nf_ct_l3proto_find(u_int16_t l3proto) {if (unlikely (l3proto >=AF_MAX ))return &nf_conntrack_l3proto_generic; -return nf_ct_l3protos[l3proto]; +return rcu_dereference (nf_ct_l3protos[l3proto]); }#endif /*_NF_CONNTRACK_L3PROTO_H*/
net/ipv4/netfilter/nf_conntrack_proto_icmp.c +2 -0
@@ -170,7 +170,9 @@ icmp_error_message(struct sk_buff *skb,return -NF_ACCEPT ; } +/* rcu_read_lock()ed by nf_hook_slow */ innerproto =__nf_ct_l4proto_find (PF_INET , inside->ip.protocol); + dataoff = skb->nh.iph->ihl*4 +sizeof (inside->icmp);/* Are they talking about one of our connections? */ if (!nf_ct_get_tuple(skb, dataoff, dataoff + inside->ip.ihl*4, PF_INET,
net/ipv4/netfilter/nf_nat_core.c +5 -4
@@ -429,6 +429,7 @@ int nf_nat_icmp_reply_translation(structstruct icmphdr icmp;struct iphdr ip; } *inside; +struct nf_conntrack_l4proto *l4proto;struct nf_conntrack_tuple inner, target;int hdrlen = (*pskb)->nh.iph->ihl *4 ;enum ip_conntrack_dir dir =CTINFO2DIR (ctinfo);
@@ -464,16 +465,16 @@ int nf_nat_icmp_reply_translation(structDEBUGP ("icmp_reply_translation: translating error %p manp %u dir %s\n" , *pskb, manip, dir ==IP_CT_DIR_ORIGINAL ?"ORIG" :"REPLY" ); +/* rcu_read_lock()ed by nf_hook_slow */ + l4proto =__nf_ct_l4proto_find (PF_INET , inside->ip.protocol); +if (!nf_ct_get_tuple (*pskb, (*pskb)->nh.iph->ihl*4 +sizeof (struct icmphdr ), (*pskb)->nh.iph->ihl*4 +sizeof (struct icmphdr ) + inside->ip.ihl*4 , (u_int16_t )AF_INET , inside->ip.protocol, - &inner, - l3proto, -__nf_ct_l4proto_find ((u_int16_t )PF_INET , - inside->ip.protocol))) + &inner, l3proto, l4proto))return 0 ;/* Change inner back to look like incoming packet. We do the
net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c +1 -0
@@ -182,6 +182,7 @@ icmpv6_error_message(struct sk_buff *skbreturn -NF_ACCEPT ; } +/* rcu_read_lock()ed by nf_hook_slow */ inproto =__nf_ct_l4proto_find (PF_INET6 , inprotonum);/* Are they talking about one of our connections? */
net/netfilter/nf_conntrack_core.c +21 -6
@@ -332,13 +332,16 @@ destroy_conntrack(struct nf_conntrack *n/* To make sure we don't get any weird locking issues here: * destroy_conntrack() MUST NOT be called with a write lock * to nf_conntrack_lock!!! -HW */ +rcu_read_lock (); l3proto =__nf_ct_l3proto_find (ct->tuplehash[IP_CT_DIR_REPLY ].tuple.src.l3num);if (l3proto && l3proto->destroy) l3proto->destroy (ct); - l4proto =__nf_ct_l4proto_find (ct->tuplehash[IP_CT_DIR_REPLY ].tuple.src.l3num, ct->tuplehash[IP_CT_DIR_REPLY ].tuple.dst.protonum); + l4proto =__nf_ct_l4proto_find (ct->tuplehash[IP_CT_DIR_REPLY ].tuple.src.l3num, + ct->tuplehash[IP_CT_DIR_REPLY ].tuple.dst.protonum);if (l4proto && l4proto->destroy) l4proto->destroy (ct); +rcu_read_unlock ();if (nf_conntrack_destroyed)nf_conntrack_destroyed (ct);
@@ -647,9 +650,14 @@ struct nf_conn *nf_conntrack_alloc(const const struct nf_conntrack_tuple *repl) {struct nf_conntrack_l3proto *l3proto; +struct nf_conn *ct; +rcu_read_lock (); l3proto =__nf_ct_l3proto_find (orig->src.l3num); -return __nf_conntrack_alloc (orig, repl, l3proto,0 ); + ct =__nf_conntrack_alloc (orig, repl, l3proto,0 ); +rcu_read_unlock (); + +return ct; }EXPORT_SYMBOL_GPL (nf_conntrack_alloc);
@@ -817,7 +825,9 @@ nf_conntrack_in(int pf, unsigned int hooreturn NF_ACCEPT ; } +/* rcu_read_lock()ed by nf_hook_slow */ l3proto =__nf_ct_l3proto_find ((u_int16_t )pf); +if ((ret = l3proto->prepare (pskb, hooknum, &dataoff, &protonum)) <=0 ) {DEBUGP ("not prepared to track yet or error occured\n" );return -ret;
@@ -872,10 +882,15 @@ EXPORT_SYMBOL_GPL(nf_conntrack_in);int nf_ct_invert_tuplepr (struct nf_conntrack_tuple *inverse,const struct nf_conntrack_tuple *orig) { -return nf_ct_invert_tuple (inverse, orig, -__nf_ct_l3proto_find (orig->src.l3num), -__nf_ct_l4proto_find (orig->src.l3num, - orig->dst.protonum)); +int ret; + +rcu_read_lock (); + ret =nf_ct_invert_tuple (inverse, orig, +__nf_ct_l3proto_find (orig->src.l3num), +__nf_ct_l4proto_find (orig->src.l3num, + orig->dst.protonum)); +rcu_read_unlock (); +return ret; }EXPORT_SYMBOL_GPL (nf_ct_invert_tuplepr);
net/netfilter/nf_conntrack_proto.c +14 -19
@@ -66,7 +66,7 @@ __nf_ct_l4proto_find(u_int16_t l3proto,if (unlikely (l3proto >=AF_MAX || nf_ct_protos[l3proto] ==NULL ))return &nf_conntrack_l4proto_generic; -return nf_ct_protos[l3proto][l4proto]; +return rcu_dereference (nf_ct_protos[l3proto][l4proto]); }EXPORT_SYMBOL_GPL (__nf_ct_l4proto_find);
@@ -77,11 +77,11 @@ nf_ct_l4proto_find_get(u_int16_t l3proto {struct nf_conntrack_l4proto *p; -preempt_disable (); +rcu_read_lock (); p =__nf_ct_l4proto_find (l3proto, l4proto);if (!try_module_get (p->me)) p = &nf_conntrack_l4proto_generic; -preempt_enable (); +rcu_read_unlock ();return p; }
@@ -98,11 +98,11 @@ nf_ct_l3proto_find_get(u_int16_t l3proto {struct nf_conntrack_l3proto *p; -preempt_disable (); +rcu_read_lock (); p =__nf_ct_l3proto_find (l3proto);if (!try_module_get (p->me)) p = &nf_conntrack_l3proto_generic; -preempt_enable (); +rcu_read_unlock ();return p; }
@@ -137,10 +137,8 @@ void nf_ct_l3proto_module_put(unsigned s {struct nf_conntrack_l3proto *p; -preempt_disable (); +/* rcu_read_lock not necessary since the caller holds a reference */ p =__nf_ct_l3proto_find (l3proto); -preempt_enable (); -module_put (p->me); }EXPORT_SYMBOL_GPL (nf_ct_l3proto_module_put);
@@ -202,7 +200,7 @@ int nf_conntrack_l3proto_register(struct ret = -EBUSY ; goto out_unlock; } - nf_ct_l3protos[proto->l3proto] = proto; +rcu_assign_pointer (nf_ct_l3protos[proto->l3proto], proto);write_unlock_bh (&nf_conntrack_lock); ret =nf_ct_l3proto_register_sysctl (proto);
@@ -233,14 +231,13 @@ int nf_conntrack_l3proto_unregister(stru goto out; } - nf_ct_l3protos[proto->l3proto] = &nf_conntrack_l3proto_generic; +rcu_assign_pointer (nf_ct_l3protos[proto->l3proto], + &nf_conntrack_l3proto_generic);write_unlock_bh (&nf_conntrack_lock); +synchronize_rcu ();nf_ct_l3proto_unregister_sysctl (proto); -/* Somebody could be still looking at the proto in bh. */ -synchronize_net (); -/* Remove all contrack entries for this protocol */ nf_ct_iterate_cleanup (kill_l3proto, proto);
@@ -356,7 +353,7 @@ retry: goto retry; } - nf_ct_protos[l4proto->l3proto][l4proto->l4proto] = l4proto; +rcu_assign_pointer (nf_ct_protos[l4proto->l3proto][l4proto->l4proto], l4proto);write_unlock_bh (&nf_conntrack_lock); ret =nf_ct_l4proto_register_sysctl (l4proto);
@@ -392,15 +389,13 @@ int nf_conntrack_l4proto_unregister(stru ret = -EBUSY ; goto out; } - nf_ct_protos[l4proto->l3proto][l4proto->l4proto] - = &nf_conntrack_l4proto_generic; +rcu_assign_pointer (nf_ct_protos[l4proto->l3proto][l4proto->l4proto], + &nf_conntrack_l4proto_generic);write_unlock_bh (&nf_conntrack_lock); +synchronize_rcu ();nf_ct_l4proto_unregister_sysctl (l4proto); -/* Somebody could be still looking at the proto in bh. */ -synchronize_net (); -/* Remove all contrack entries for this protocol */ nf_ct_iterate_cleanup (kill_l4proto, l4proto);
PATCH 13/22 : ip_conntrack: fix invalid conntrack statistics RCU assumption
In reply to Patrick McHardy[NETFILTER]: ip_conntrack: fix invalid conntrack statistics RCU assumption
CONNTRACK_STAT_INC assumes rcu_read_lock in nf_hook_slow disables preemption as well, making it legal to use __get_cpu_var without disabling preemption manually. The assumption is not correct anymore with preemptable RCU, additionally we need to protect against softirqs when not holding ip_conntrack_lock.
Add CONNTRACK_STAT_INC_ATOMIC macro, which disables local softirqs, and use where necessary.
---
commit ff0856c0c2e24e278b0611df2d66f66d5c0d1fd6
tree d5899481f22f436368fa121ac617a08fc53941f4
parent 2f34e7966f36432f6915ab7eee85116439a3944e
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:03:24 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:03:24 +0100
include/linux/netfilter_ipv4/ip_conntrack.h | 6 ++++++ net/ipv4/netfilter/ip_conntrack_core.c | 14 +++++++------- 2 files changed, 13 insertions(+), 7 deletions(-)
include/linux/netfilter_ipv4/ip_conntrack.h +6 -0
@@ -301,6 +301,12 @@ extern unsigned int ip_conntrack_htable_extern int ip_conntrack_checksum;#define CONNTRACK_STAT_INC (count) (__get_cpu_var(ip_conntrack_stat).count++) +#define CONNTRACK_STAT_INC_ATOMIC (count) \ +do { \ + local_bh_disable(); \ + __get_cpu_var(ip_conntrack_stat).count++; \ + local_bh_enable(); \ +} while (0)#ifdef CONFIG_IP_NF_CONNTRACK_EVENTS #include <linux/notifier.h>
net/ipv4/netfilter/ip_conntrack_core.c +7 -7
@@ -538,7 +538,7 @@ static int early_drop(struct list_head *if (del_timer (&ct->timeout)) {death_by_timeout ((unsigned long )ct); dropped =1 ; -CONNTRACK_STAT_INC (early_drop); +CONNTRACK_STAT_INC_ATOMIC (early_drop); }ip_conntrack_put (ct);return dropped;
@@ -805,7 +805,7 @@ unsigned int ip_conntrack_in(unsigned in/* Previously seen (loopback or untracked)? Ignore. */ if ((*pskb)->nfct) { -CONNTRACK_STAT_INC (ignore); +CONNTRACK_STAT_INC_ATOMIC (ignore);return NF_ACCEPT ; }
@@ -841,20 +841,20 @@ #endif* core what to do with the packet. */ if (proto->error !=NULL && (ret = proto->error (*pskb, &ctinfo, hooknum)) <=0 ) { -CONNTRACK_STAT_INC (error); -CONNTRACK_STAT_INC (invalid); +CONNTRACK_STAT_INC_ATOMIC (error); +CONNTRACK_STAT_INC_ATOMIC (invalid);return -ret; }if (!(ct =resolve_normal_ct (*pskb, proto,&set_reply,hooknum,&ctinfo))) {/* Not valid part of a connection */ -CONNTRACK_STAT_INC (invalid); +CONNTRACK_STAT_INC_ATOMIC (invalid);return NF_ACCEPT ; }if (IS_ERR (ct)) {/* Too stressed to deal. */ -CONNTRACK_STAT_INC (drop); +CONNTRACK_STAT_INC_ATOMIC (drop);return NF_DROP ; }
@@ -866,7 +866,7 @@ #endif* the netfilter core what to do*/ nf_conntrack_put ((*pskb)->nfct); (*pskb)->nfct =NULL ; -CONNTRACK_STAT_INC (invalid); +CONNTRACK_STAT_INC_ATOMIC (invalid);return -ret; }
PATCH 14/22 : nf_conntrack: fix invalid conntrack statistics RCU assumption
In reply to Patrick McHardy[NETFILTER]: nf_conntrack: fix invalid conntrack statistics RCU assumption
NF_CT_STAT_INC assumes rcu_read_lock in nf_hook_slow disables preemption as well, making it legal to use __get_cpu_var without disabling preemption manually. The assumption is not correct anymore with preemptable RCU, additionally we need to protect against softirqs when not holding nf_conntrack_lock.
Add NF_CT_STAT_INC_ATOMIC macro, which disables local softirqs, and use where necessary.
---
commit bab7afccc8d458cc09b3a5bb4224358cbd4f188a
tree 537777d83e0b45a4fa17aa148113c9b275823806
parent ff0856c0c2e24e278b0611df2d66f66d5c0d1fd6
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:04:50 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:04:50 +0100
include/net/netfilter/nf_conntrack.h | 6 ++++++ net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c | 4 ++-- net/netfilter/nf_conntrack_core.c | 14 +++++++------- 3 files changed, 15 insertions(+), 9 deletions(-)
include/net/netfilter/nf_conntrack.h +6 -0
@@ -257,6 +257,12 @@ extern int nf_conntrack_max; DECLARE_PER_CPU(struct ip_conntrack_stat, nf_conntrack_stat);#define NF_CT_STAT_INC (count) (__get_cpu_var(nf_conntrack_stat).count++) +#define NF_CT_STAT_INC_ATOMIC (count) \ +do { \ + local_bh_disable(); \ + __get_cpu_var(nf_conntrack_stat).count++; \ + local_bh_enable(); \ +} while (0)/* no helper, no nat */ #define NF_CT_F_BASIC 0
net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c +2 -2
@@ -154,8 +154,8 @@ ipv6_prepare(struct sk_buff **pskb, unsi*/ if ((protoff <0 ) || (protoff > (*pskb)->len)) {DEBUGP ("ip6_conntrack_core: can't find proto in pkt\n" ); -NF_CT_STAT_INC (error); -NF_CT_STAT_INC (invalid); +NF_CT_STAT_INC_ATOMIC (error); +NF_CT_STAT_INC_ATOMIC (invalid);return -NF_ACCEPT ; }
net/netfilter/nf_conntrack_core.c +7 -7
@@ -563,7 +563,7 @@ static int early_drop(struct list_head *if (del_timer (&ct->timeout)) {death_by_timeout ((unsigned long )ct); dropped =1 ; -NF_CT_STAT_INC (early_drop); +NF_CT_STAT_INC_ATOMIC (early_drop); }nf_ct_put (ct);return dropped;
@@ -821,7 +821,7 @@ nf_conntrack_in(int pf, unsigned int hoo/* Previously seen (loopback or untracked)? Ignore. */ if ((*pskb)->nfct) { -NF_CT_STAT_INC (ignore); +NF_CT_STAT_INC_ATOMIC (ignore);return NF_ACCEPT ; }
@@ -840,8 +840,8 @@ nf_conntrack_in(int pf, unsigned int hoo* core what to do with the packet. */ if (l4proto->error !=NULL && (ret = l4proto->error (*pskb, dataoff, &ctinfo, pf, hooknum)) <=0 ) { -NF_CT_STAT_INC (error); -NF_CT_STAT_INC (invalid); +NF_CT_STAT_INC_ATOMIC (error); +NF_CT_STAT_INC_ATOMIC (invalid);return -ret; }
@@ -849,13 +849,13 @@ nf_conntrack_in(int pf, unsigned int hoo &set_reply, &ctinfo);if (!ct) {/* Not valid part of a connection */ -NF_CT_STAT_INC (invalid); +NF_CT_STAT_INC_ATOMIC (invalid);return NF_ACCEPT ; }if (IS_ERR (ct)) {/* Too stressed to deal. */ -NF_CT_STAT_INC (drop); +NF_CT_STAT_INC_ATOMIC (drop);return NF_DROP ; }
@@ -868,7 +868,7 @@ nf_conntrack_in(int pf, unsigned int hooDEBUGP ("nf_conntrack_in: Can't track with proto module\n" );nf_conntrack_put ((*pskb)->nfct); (*pskb)->nfct =NULL ; -NF_CT_STAT_INC (invalid); +NF_CT_STAT_INC_ATOMIC (invalid);return -ret; }
PATCH 15/22 : ip_conntrack: properly use RCU for ip_conntrack_destroyed callback
In reply to Patrick McHardy[NETFILTER]: ip_conntrack: properly use RCU for ip_conntrack_destroyed callback
---
commit 26462f9a42bcf2230a68d84dec3a048bc5a9a1af
tree 89f615281c88a2f541583337f1fb700a7a3c1874
parent bab7afccc8d458cc09b3a5bb4224358cbd4f188a
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:06:55 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:06:55 +0100
net/ipv4/netfilter/ip_conntrack_core.c | 9 ++++++--- net/ipv4/netfilter/ip_nat_core.c | 7 ++++--- 2 files changed, 10 insertions(+), 6 deletions(-)
net/ipv4/netfilter/ip_conntrack_core.c +6 -3
@@ -303,6 +303,7 @@ destroy_conntrack(struct nf_conntrack *nstruct ip_conntrack *ct = (struct ip_conntrack *)nfct;struct ip_conntrack_protocol *proto;struct ip_conntrack_helper *helper; +typeof (ip_conntrack_destroyed) destroyed;DEBUGP ("destroy_conntrack(%p)\n" , ct);IP_NF_ASSERT (atomic_read (&nfct->use) ==0 );
@@ -322,10 +323,12 @@ destroy_conntrack(struct nf_conntrack *n proto =__ip_conntrack_proto_find (ct->tuplehash[IP_CT_DIR_REPLY ].tuple.dst.protonum);if (proto && proto->destroy) proto->destroy (ct); -rcu_read_unlock (); -if (ip_conntrack_destroyed) -ip_conntrack_destroyed (ct); + destroyed =rcu_dereference (ip_conntrack_destroyed); +if (destroyed) +destroyed (ct); + +rcu_read_unlock ();write_lock_bh (&ip_conntrack_lock);/* Expectations will have been removed in clean_from_lists,
net/ipv4/netfilter/ip_nat_core.c +4 -3
@@ -604,8 +604,8 @@ static int __init ip_nat_init(void) }/* FIXME: Man, this is a hack. <SIGH> */ -IP_NF_ASSERT (ip_conntrack_destroyed ==NULL ); - ip_conntrack_destroyed = &ip_nat_cleanup_conntrack; +IP_NF_ASSERT (rcu_dereference (ip_conntrack_destroyed) ==NULL ); +rcu_assign_pointer (ip_conntrack_destroyed, ip_nat_cleanup_conntrack);/* Initialize fake conntrack so that NAT will skip it */ ip_conntrack_untracked.status |=IPS_NAT_DONE_MASK ;
@@ -623,7 +623,8 @@ static int clean_nat(struct ip_conntrack static void __exit ip_nat_cleanup(void) {ip_ct_iterate_cleanup (&clean_nat,NULL ); - ip_conntrack_destroyed =NULL ; +rcu_assign_pointer (ip_conntrack_destroyed,NULL ); +synchronize_rcu ();vfree (bysource); }
PATCH 16/22 : nf_conntrack: properly use RCU for nf_conntrack_destroyed callback
In reply to Patrick McHardy[NETFILTER]: nf_conntrack: properly use RCU for nf_conntrack_destroyed callback
---
commit 2428570e4fc9d689faeb4a96ba7b34352f0a6e2c
tree 7224fd0620ed1a49075a55da0c3b08cd77334313
parent 26462f9a42bcf2230a68d84dec3a048bc5a9a1af
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:07:18 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:07:18 +0100
net/ipv4/netfilter/nf_nat_core.c | 7 ++++--- net/netfilter/nf_conntrack_core.c | 9 ++++++--- 2 files changed, 10 insertions(+), 6 deletions(-)
net/ipv4/netfilter/nf_nat_core.c +4 -3
@@ -617,8 +617,8 @@ static int __init nf_nat_init(void) }/* FIXME: Man, this is a hack. <SIGH> */ -NF_CT_ASSERT (nf_conntrack_destroyed ==NULL ); - nf_conntrack_destroyed = &nf_nat_cleanup_conntrack; +NF_CT_ASSERT (rcu_dereference (nf_conntrack_destroyed) ==NULL ); +rcu_assign_pointer (nf_conntrack_destroyed, nf_nat_cleanup_conntrack);/* Initialize fake conntrack so that NAT will skip it */ nf_conntrack_untracked.status |=IPS_NAT_DONE_MASK ;
@@ -642,7 +642,8 @@ static int clean_nat(struct nf_conn *i, static void __exit nf_nat_cleanup(void) {nf_ct_iterate_cleanup (&clean_nat,NULL ); - nf_conntrack_destroyed =NULL ; +rcu_assign_pointer (nf_conntrack_destroyed,NULL ); +synchronize_rcu ();vfree (bysource);nf_ct_l3proto_put (l3proto); }
net/netfilter/nf_conntrack_core.c +6 -3
@@ -318,6 +318,7 @@ destroy_conntrack(struct nf_conntrack *nstruct nf_conn_help *help =nfct_help (ct);struct nf_conntrack_l3proto *l3proto;struct nf_conntrack_l4proto *l4proto; +typeof (nf_conntrack_destroyed) destroyed;DEBUGP ("destroy_conntrack(%p)\n" , ct);NF_CT_ASSERT (atomic_read (&nfct->use) ==0 );
@@ -341,10 +342,12 @@ destroy_conntrack(struct nf_conntrack *n ct->tuplehash[IP_CT_DIR_REPLY].tuple.dst.protonum);if (l4proto && l4proto->destroy) l4proto->destroy (ct); -rcu_read_unlock (); -if (nf_conntrack_destroyed) -nf_conntrack_destroyed (ct); + destroyed =rcu_dereference (nf_conntrack_destroyed); +if (destroyed) +destroyed (ct); + +rcu_read_unlock ();write_lock_bh (&nf_conntrack_lock);/* Expectations will have been removed in clean_from_lists,
PATCH 17/22 : nf_conntrack: change nf_conntrack_l[34]proto_unregister to void
In reply to Patrick McHardy[NETFILTER]: nf_conntrack: change nf_conntrack_l[34]proto_unregister to void
No caller checks the return value, and since its usually called within the module unload path there's nothing a module could do about errors anyway, so BUG on invalid conditions and return void.
---
commit 4b4109fff045b66b7bc0a696fc020f0cc22bad0e
tree 88928fbaa762aca7dfce1d0b4138fc9372487e4e
parent 2428570e4fc9d689faeb4a96ba7b34352f0a6e2c
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:07:30 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:07:30 +0100
include/net/netfilter/nf_conntrack_l3proto.h | 2 + include/net/netfilter/nf_conntrack_l4proto.h | 2 + net/netfilter/nf_conntrack_proto.c | 40 +++++--------------------- 3 files changed, 9 insertions(+), 35 deletions(-)
include/net/netfilter/nf_conntrack_l3proto.h +1 -1
@@ -89,7 +89,7 @@ extern struct nf_conntrack_l3proto *nf_c/* Protocol registration. */ extern int nf_conntrack_l3proto_register (struct nf_conntrack_l3proto *proto); -extern int nf_conntrack_l3proto_unregister (struct nf_conntrack_l3proto *proto); +extern void nf_conntrack_l3proto_unregister (struct nf_conntrack_l3proto *proto);extern struct nf_conntrack_l3proto *nf_ct_l3proto_find_get (u_int16_t l3proto);
include/net/netfilter/nf_conntrack_l4proto.h +1 -1
@@ -109,7 +109,7 @@ extern void nf_ct_l4proto_put(struct nf_/* Protocol registration. */ extern int nf_conntrack_l4proto_register (struct nf_conntrack_l4proto *proto); -extern int nf_conntrack_l4proto_unregister (struct nf_conntrack_l4proto *proto); +extern void nf_conntrack_l4proto_unregister (struct nf_conntrack_l4proto *proto);/* Generic netlink helpers */ extern int nf_ct_port_tuple_to_nfattr(struct sk_buff *skb,
net/netfilter/nf_conntrack_proto.c +7 -33
@@ -215,22 +215,12 @@ out: }EXPORT_SYMBOL_GPL (nf_conntrack_l3proto_register); -int nf_conntrack_l3proto_unregister (struct nf_conntrack_l3proto *proto) +void nf_conntrack_l3proto_unregister (struct nf_conntrack_l3proto *proto) { -int ret =0 ; - -if (proto->l3proto >=AF_MAX ) { - ret = -EBUSY ; - goto out; - } +BUG_ON (proto->l3proto >=AF_MAX );write_lock_bh (&nf_conntrack_lock); -if (nf_ct_l3protos[proto->l3proto] != proto) { -write_unlock_bh (&nf_conntrack_lock); - ret = -EBUSY ; - goto out; - } - +BUG_ON (nf_ct_l3protos[proto->l3proto] != proto);rcu_assign_pointer (nf_ct_l3protos[proto->l3proto], &nf_conntrack_l3proto_generic);write_unlock_bh (&nf_conntrack_lock);
@@ -240,9 +230,6 @@ int nf_conntrack_l3proto_unregister(stru/* Remove all contrack entries for this protocol */ nf_ct_iterate_cleanup (kill_l3proto, proto); - -out: -return ret; }EXPORT_SYMBOL_GPL (nf_conntrack_l3proto_unregister);
@@ -368,27 +355,17 @@ out: }EXPORT_SYMBOL_GPL (nf_conntrack_l4proto_register); -int nf_conntrack_l4proto_unregister (struct nf_conntrack_l4proto *l4proto) +void nf_conntrack_l4proto_unregister (struct nf_conntrack_l4proto *l4proto) { -int ret =0 ; - -if (l4proto->l3proto >=PF_MAX ) { - ret = -EBUSY ; - goto out; - } +BUG_ON (l4proto->l3proto >=PF_MAX );if (l4proto == &nf_conntrack_l4proto_generic) {nf_ct_l4proto_unregister_sysctl (l4proto); - goto out; +return ; }write_lock_bh (&nf_conntrack_lock); -if (nf_ct_protos[l4proto->l3proto][l4proto->l4proto] - != l4proto) { -write_unlock_bh (&nf_conntrack_lock); - ret = -EBUSY ; - goto out; - } +BUG_ON (nf_ct_protos[l4proto->l3proto][l4proto->l4proto] != l4proto);rcu_assign_pointer (nf_ct_protos[l4proto->l3proto][l4proto->l4proto], &nf_conntrack_l4proto_generic);write_unlock_bh (&nf_conntrack_lock);
@@ -398,8 +375,5 @@ int nf_conntrack_l4proto_unregister(stru/* Remove all contrack entries for this protocol */ nf_ct_iterate_cleanup (kill_l4proto, l4proto); - -out: -return ret; }EXPORT_SYMBOL_GPL (nf_conntrack_l4proto_unregister);
PATCH 18/22 : xt_mac/xt_CLASSIFY: use IPv6 hook names for IPv6 registration
In reply to Patrick McHardy[NETFILTER]: xt_mac/xt_CLASSIFY: use IPv6 hook names for IPv6 registration
Use NF_IP6_ instead of NF_IP_. The values are identical, this is merely cleanup.
---
commit 0a03715d7befc80612212ee80723dcce8b75f3c1
tree 1ddf22e671682d94ad017f0a9f7f4a74566c60bd
parent 4b4109fff045b66b7bc0a696fc020f0cc22bad0e
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:07:51 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:07:51 +0100
net/netfilter/xt_CLASSIFY.c | 8 +++++--- net/netfilter/xt_mac.c | 7 ++++--- 2 files changed, 9 insertions(+), 6 deletions(-)
net/netfilter/xt_CLASSIFY.c +5 -3
@@ -15,6 +15,8 @@ #include <linux/skbuff.h>#include <linux/ip.h> #include <net/checksum.h> +#include <linux/netfilter_ipv4.h> +#include <linux/netfilter_ipv6.h> #include <linux/netfilter/x_tables.h> #include <linux/netfilter/xt_CLASSIFY.h>
@@ -55,9 +57,9 @@ static struct xt_target xt_classify_targ .target = target, .targetsize = sizeof(struct xt_classify_target_info), .table = "mangle", - .hooks = (1 <<NF_IP_LOCAL_OUT ) | - (1 <<NF_IP_FORWARD ) | - (1 << NF_IP_POST_ROUTING), + .hooks = (1 <<NF_IP6_LOCAL_OUT ) | + (1 <<NF_IP6_FORWARD ) | + (1 << NF_IP6_POST_ROUTING), .me = THIS_MODULE, }, };
net/netfilter/xt_mac.c +4 -3
@@ -14,6 +14,7 @@ #include <linux/if_ether.h>#include <linux/etherdevice.h> #include <linux/netfilter_ipv4.h> +#include <linux/netfilter_ipv6.h> #include <linux/netfilter/xt_mac.h> #include <linux/netfilter/x_tables.h>
@@ -59,9 +60,9 @@ static struct xt_match xt_mac_match[] = .family = AF_INET6, .match = match, .matchsize = sizeof(struct xt_mac_info), - .hooks = (1 <<NF_IP_PRE_ROUTING ) | - (1 <<NF_IP_LOCAL_IN ) | - (1 << NF_IP_FORWARD), + .hooks = (1 <<NF_IP6_PRE_ROUTING ) | + (1 <<NF_IP6_LOCAL_IN ) | + (1 << NF_IP6_FORWARD), .me = THIS_MODULE, }, };
PATCH 19/22 : Kconfig: improve dependency handling
In reply to Patrick McHardy[NETFILTER]: Kconfig: improve dependency handling
Instead of depending on internally needed options and letting users figure out what is needed, select them when needed:
- IP_NF_IPTABLES, IP_NF_ARPTABLES and IP6_NF_IPTABLES select NETFILTER_XTABLES
- NETFILTER_XT_TARGET_CONNMARK, NETFILTER_XT_MATCH_CONNMARK and IP_NF_TARGET_CLUSTERIP select NF_CONNTRACK_MARK
- NETFILTER_XT_MATCH_CONNBYTES selects NF_CT_ACCT
---
commit a3987b41ed8c0275393942316c502e4cffbc0126
tree e5f23f16e4bdd9b454356a75269f8a3fb161082a
parent 0a03715d7befc80612212ee80723dcce8b75f3c1
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:08:00 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:08:00 +0100
net/ipv4/netfilter/Kconfig | 8 +++++--- net/ipv6/netfilter/Kconfig | 3 ++- net/netfilter/Kconfig | 12 +++++++++--- 3 files changed, 16 insertions(+), 7 deletions(-)
net/ipv4/netfilter/Kconfig +5 -3
@@ -226,7 +226,7 @@ config IP_NF_QUEUE config IP_NF_IPTABLES tristate "IP tables support (required for filtering/masq/NAT)" - depends on NETFILTER_XTABLES + select NETFILTER_XTABLES help iptables is a general, extensible packet identification framework. The packet filtering and full NAT (masquerading, port forwarding,
@@ -606,7 +606,9 @@ config IP_NF_TARGET_TTL config IP_NF_TARGET_CLUSTERIP tristate "CLUSTERIP target support (EXPERIMENTAL)" depends on IP_NF_MANGLE && EXPERIMENTAL - depends on (IP_NF_CONNTRACK && IP_NF_CONNTRACK_MARK) || (NF_CONNTRACK_MARK && NF_CONNTRACK_IPV4) + depends on IP_NF_CONNTRACK || NF_CONNTRACK_IPV4 + select IP_NF_CONNTRACK_MARK if IP_NF_CONNTRACK + select NF_CONNTRACK_MARK if NF_CONNTRACK_IPV4 help The CLUSTERIP target allows you to build load-balancing clusters of network servers without having a dedicated load-balancing
@@ -629,7 +631,7 @@ config IP_NF_RAW # ARP tables config IP_NF_ARPTABLES tristate "ARP tables support" - depends on NETFILTER_XTABLES + select NETFILTER_XTABLES help arptables is a general, extensible packet identification framework. The ARP packet filtering and mangling (manipulation)subsystems
net/ipv6/netfilter/Kconfig +2 -1
@@ -42,7 +42,8 @@ config IP6_NF_QUEUE config IP6_NF_IPTABLES tristate "IP6 tables support (required for filtering)" - depends on INET && IPV6 && EXPERIMENTAL && NETFILTER_XTABLES + depends on INET && IPV6 && EXPERIMENTAL + select NETFILTER_XTABLES help ip6tables is a general, extensible packet identification framework. Currently only the packet filtering and packet mangling subsystem
net/netfilter/Kconfig +9 -3
@@ -302,7 +302,9 @@ config NETFILTER_XT_TARGET_CONNMARK tristate '"CONNMARK" target support' depends on NETFILTER_XTABLES depends on IP_NF_MANGLE || IP6_NF_MANGLE - depends on (IP_NF_CONNTRACK && IP_NF_CONNTRACK_MARK) || (NF_CONNTRACK_MARK && NF_CONNTRACK) + depends on IP_NF_CONNTRACK || NF_CONNTRACK + select IP_NF_CONNTRACK_MARK if IP_NF_CONNTRACK + select NF_CONNTRACK_MARK if NF_CONNTRACK help This option adds a `CONNMARK' target, which allows one to manipulate the connection mark value. Similar to the MARK target, but
@@ -434,7 +436,9 @@ config NETFILTER_XT_MATCH_COMMENT config NETFILTER_XT_MATCH_CONNBYTES tristate '"connbytes" per-connection counter match support' depends on NETFILTER_XTABLES - depends on (IP_NF_CONNTRACK && IP_NF_CT_ACCT) || (NF_CT_ACCT && NF_CONNTRACK) + depends on IP_NF_CONNTRACK || NF_CONNTRACK + select IP_NF_CT_ACCT if IP_NF_CONNTRACK + select NF_CT_ACCT if NF_CONNTRACK help This option adds a `connbytes' match, which allows you to match the number of bytes and/or packets for each direction within a connection.
@@ -445,7 +449,9 @@ config NETFILTER_XT_MATCH_CONNBYTES config NETFILTER_XT_MATCH_CONNMARK tristate '"connmark" connection mark match support' depends on NETFILTER_XTABLES - depends on (IP_NF_CONNTRACK && IP_NF_CONNTRACK_MARK) || (NF_CONNTRACK_MARK && NF_CONNTRACK) + depends on IP_NF_CONNTRACK || NF_CONNTRACK + select IP_NF_CONNTRACK_MARK if IP_NF_CONNTRACK + select NF_CONNTRACK_MARK if NF_CONNTRACK help This option adds a `connmark' match, which allows you to match the connection mark value previously set for the session by `CONNMARK'.
PATCH 20/22 : Fix whitespace errors
In reply to Patrick McHardy[NETFILTER]: Fix whitespace errors
---
commit 353c941b0735b883510bfd7bb98fcd4fc8a0ee08
tree 14c9e7a24feed39641b2f97b6e2b247bd229062c
parent a3987b41ed8c0275393942316c502e4cffbc0126
author YOSHIFUJI Hideaki <[email protected]> Mon, 12 Feb 2007 11:08:07 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:08:07 +0100
net/netfilter/core.c | 6 - net/netfilter/nf_conntrack_core.c | 18 +- net/netfilter/nf_conntrack_expect.c | 2 net/netfilter/nf_conntrack_ftp.c | 14 +- net/netfilter/nf_conntrack_h323_main.c | 20 +-- net/netfilter/nf_conntrack_irc.c | 2 net/netfilter/nf_conntrack_l3proto_generic.c | 2 net/netfilter/nf_conntrack_netbios_ns.c | 2 net/netfilter/nf_conntrack_netlink.c | 106 +++++++------- net/netfilter/nf_conntrack_pptp.c | 2 net/netfilter/nf_conntrack_proto_sctp.c | 72 +++++---- net/netfilter/nf_conntrack_proto_tcp.c | 204 +++++++++++++------------- net/netfilter/nf_conntrack_sip.c | 6 - net/netfilter/nf_conntrack_standalone.c | 2 net/netfilter/nf_conntrack_tftp.c | 2 net/netfilter/nf_internals.h | 2 net/netfilter/nf_log.c | 6 - net/netfilter/nf_queue.c | 22 +-- net/netfilter/nf_sockopt.c | 12 +- net/netfilter/nfnetlink.c | 10 + net/netfilter/nfnetlink_log.c | 52 +++---- net/netfilter/nfnetlink_queue.c | 84 +++++------ net/netfilter/x_tables.c | 12 +- net/netfilter/xt_CLASSIFY.c | 4 - net/netfilter/xt_MARK.c | 4 - net/netfilter/xt_NFQUEUE.c | 4 - net/netfilter/xt_NOTRACK.c | 4 - net/netfilter/xt_SECMARK.c | 2 net/netfilter/xt_conntrack.c | 32 ++-- net/netfilter/xt_dccp.c | 20 +-- net/netfilter/xt_hashlimit.c | 8 + net/netfilter/xt_helper.c | 18 +- net/netfilter/xt_length.c | 4 - net/netfilter/xt_limit.c | 2 net/netfilter/xt_mark.c | 6 - net/netfilter/xt_multiport.c | 2 net/netfilter/xt_physdev.c | 2 net/netfilter/xt_policy.c | 24 ++- net/netfilter/xt_quota.c | 4 - net/netfilter/xt_realm.c | 2 net/netfilter/xt_sctp.c | 30 ++-- net/netfilter/xt_string.c | 8 + net/netfilter/xt_tcpmss.c | 4 - 43 files changed, 422 insertions(+), 422 deletions(-)
net/netfilter/core.c +3 -3
@@ -1,4 +1,4 @@ -/* netfilter.c: look after the filters for various protocols. +/* netfilter.c: look after the filters for various protocols. * Heavily influenced by the old firewall.c by David Bonn and Alan Cox. * * Thanks to Rob `CmdrTaco' Malda for not influencing this code in any
@@ -141,14 +141,14 @@ unsigned int nf_iterate(struct list_headcontinue ;/* Optimization: we don't need to hold module -reference here, since function can't sleep. --RR */ +reference here, since function can't sleep. --RR */ verdict = elem->hook (hook, skb, indev, outdev, okfn);if (verdict !=NF_ACCEPT ) {#ifdef CONFIG_NETFILTER_DEBUG if (unlikely ((verdict &NF_VERDICT_MASK ) >NF_MAX_VERDICT )) {NFDEBUG ("Evil return from %p(%u).\n" , - elem->hook, hook); + elem->hook, hook);continue ; }#endif
net/netfilter/nf_conntrack_core.c +9 -9
@@ -424,7 +424,7 @@ EXPORT_SYMBOL_GPL(nf_conntrack_find_get)static void __nf_conntrack_hash_insert (struct nf_conn *ct,unsigned int hash, -unsigned int repl_hash) +unsigned int repl_hash) { ct->id = ++nf_conntrack_next_id; list_add(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list,
@@ -1066,7 +1066,7 @@ get_next_corpse(int (*iter)(struct nf_coif (iter (ct, data)) goto found; } - } + }list_for_each_entry (h, &unconfirmed, list) { ct =nf_ct_tuplehash_to_ctrack (h);if (iter (ct, data))
@@ -1107,7 +1107,7 @@ static void free_conntrack_hash(struct lif (vmalloced)vfree (hash);else -free_pages ((unsigned long )hash, +free_pages ((unsigned long )hash,get_order (sizeof (struct list_head ) * size)); }
@@ -1168,18 +1168,18 @@ static struct list_head *alloc_hashtablestruct list_head *hash;unsigned int i; - *vmalloced =0 ; - hash = (void *)__get_free_pages (GFP_KERNEL , + *vmalloced =0 ; + hash = (void *)__get_free_pages (GFP_KERNEL ,get_order (sizeof (struct list_head ) * size)); -if (!hash) { +if (!hash) { *vmalloced =1 ;printk (KERN_WARNING "nf_conntrack: falling back to vmalloc.\n" ); hash =vmalloc (sizeof (struct list_head ) * size); }if (hash) -for (i =0 ; i < size; i++) +for (i =0 ; i < size; i++)INIT_LIST_HEAD (&hash[i]);return hash;
@@ -1286,9 +1286,9 @@ int __init nf_conntrack_init(void)/* Don't NEED lock here, but good form anyway. */ write_lock_bh (&nf_conntrack_lock); -for (i =0 ; i <AF_MAX ; i++) +for (i =0 ; i <AF_MAX ; i++) nf_ct_l3protos[i] = &nf_conntrack_l3proto_generic; -write_unlock_bh (&nf_conntrack_lock); +write_unlock_bh (&nf_conntrack_lock);/* For use by REJECT target */ rcu_assign_pointer (ip_ct_attach, __nf_conntrack_attach);
net/netfilter/nf_conntrack_expect.c +1 -1
@@ -130,7 +130,7 @@ void nf_ct_remove_expectations(struct nfif (i->master == ct &&del_timer (&i->timeout)) {nf_ct_unlink_expect (i);nf_conntrack_expect_put (i); - } + } } }EXPORT_SYMBOL_GPL (nf_ct_remove_expectations);
net/netfilter/nf_conntrack_ftp.c +7 -7
@@ -126,7 +126,7 @@ get_ipv6_addr(const char *src, size_t dl }static int try_number (const char *data,size_t dlen,u_int32_t array[], -int array_size,char sep,char term) +int array_size,char sep,char term) {u_int32_t i, len;
@@ -413,8 +413,8 @@ static int help(struct sk_buff **pskb, goto out_update_nl; } -/* Initialize IP/IPv6 addr to expected address (it's not mentioned -in EPSV responses) */ +/* Initialize IP/IPv6 addr to expected address (it's not mentioned +in EPSV responses) */ cmd.l3num = ct->tuplehash[dir].tuple.src.l3num;memcpy (cmd.u3.all, &ct->tuplehash[dir].tuple.src.u3.all,sizeof (cmd.u3.all));
@@ -466,11 +466,11 @@ static int help(struct sk_buff **pskb, memcmp(&cmd.u3.all, &ct->tuplehash[dir].tuple.src.u3.all, sizeof(cmd.u3.all))) {/* Enrico Scholz's passive FTP to partially RNAT'd ftp -server: it really wants us to connect to a -different IP address. Simply don't record it for -NAT. */ +server: it really wants us to connect to a +different IP address. Simply don't record it for +NAT. */ if (cmd.l3num ==PF_INET ) { -DEBUGP ("conntrack_ftp: NOT RECORDING: " NIPQUAD_FMT " != " NIPQUAD_FMT "\n" , +DEBUGP ("conntrack_ftp: NOT RECORDING: " NIPQUAD_FMT " != " NIPQUAD_FMT "\n" ,NIPQUAD (cmd.u3.ip),NIPQUAD (ct->tuplehash[dir].tuple.src.u3.ip)); }else {
net/netfilter/nf_conntrack_h323_main.c +10 -10
@@ -49,7 +49,7 @@ MODULE_PARM_DESC(gkrouted_only, "only acstatic int callforward_filter__read_mostly =1 ;module_param (callforward_filter, bool,0600 );MODULE_PARM_DESC (callforward_filter,"only create call forwarding expectations " -"if both endpoints are on different sides " +"if both endpoints are on different sides " "(determined by routing information)" );/* Hooks for NAT */
@@ -300,7 +300,7 @@ static int expect_rtp_rtcp(struct sk_buf IPPROTO_UDP, NULL, &rtcp_port);if (memcmp (&ct->tuplehash[dir].tuple.src.u3, - &ct->tuplehash[!dir].tuple.dst.u3, + &ct->tuplehash[!dir].tuple.dst.u3,sizeof (ct->tuplehash[dir].tuple.src.u3)) && (nat_rtp_rtcp =rcu_dereference (nat_rtp_rtcp_hook)) && ct->status &IPS_NAT_MASK ) {
@@ -743,7 +743,7 @@ #if defined(CONFIG_IPV6) || defined(CONF rt2 = (struct rt6_info *)ip6_route_output (NULL , &fl2);if (rt2) {if (!memcmp (&rt1->rt6i_gateway, &rt2->rt6i_gateway, -sizeof (rt1->rt6i_gateway)) && +sizeof (rt1->rt6i_gateway)) && rt1->u.dst.dev == rt2->u.dst.dev) ret =1 ;dst_release (&rt2->u.dst);
@@ -780,7 +780,7 @@ static int expect_callforwarding(struct* we don't need to track the second call */ if (callforward_filter &&callforward_do_filter (&addr, &ct->tuplehash[!dir].tuple.src.u3, - ct->tuplehash[!dir].tuple.src.l3num)) { + ct->tuplehash[!dir].tuple.src.l3num)) {DEBUGP ("nf_ct_q931: Call Forwarding not tracked\n" );return 0 ; }
@@ -840,7 +840,7 @@ static int process_setup(struct sk_buffif ((setup->options & eSetup_UUIE_destCallSignalAddress) && (set_h225_addr) && ct->status &&IPS_NAT_MASK &&get_h225_addr (ct, *data, &setup->destCallSignalAddress, - &addr, &port) && + &addr, &port) &&memcmp (&addr, &ct->tuplehash[!dir].tuple.src.u3,sizeof (addr))) { DEBUGP("nf_ct_q931: set destCallSignalAddress " NIP6_FMT ":%hu->" NIP6_FMT ":%hu\n",
@@ -858,7 +858,7 @@ static int process_setup(struct sk_buffif ((setup->options & eSetup_UUIE_sourceCallSignalAddress) && (set_h225_addr) && ct->status &IPS_NAT_MASK &&get_h225_addr (ct, *data, &setup->sourceCallSignalAddress, - &addr, &port) && + &addr, &port) &&memcmp (&addr, &ct->tuplehash[!dir].tuple.dst.u3,sizeof (addr))) { DEBUGP("nf_ct_q931: set sourceCallSignalAddress " NIP6_FMT ":%hu->" NIP6_FMT ":%hu\n",
@@ -1282,7 +1282,7 @@ static int expect_q931(struct sk_buff **for (i =0 ; i < count; i++) {if (get_h225_addr (ct, *data, &taddr[i], &addr, &port) &&memcmp (&addr, &ct->tuplehash[dir].tuple.src.u3, -sizeof (addr)) ==0 && port !=0 ) +sizeof (addr)) ==0 && port !=0 )break ; }
@@ -1294,7 +1294,7 @@ static int expect_q931(struct sk_buff **return -1 ;nf_conntrack_expect_init (exp, ct->tuplehash[!dir].tuple.src.l3num, gkrouted_only ?/* only accept calls from GK? */ - &ct->tuplehash[!dir].tuple.src.u3 : + &ct->tuplehash[!dir].tuple.src.u3 :NULL , &ct->tuplehash[!dir].tuple.dst.u3,IPPROTO_TCP ,NULL , &port);
@@ -1513,7 +1513,7 @@ static int process_arq(struct sk_buff ** set_h225_addr =rcu_dereference (set_h225_addr_hook);if ((arq->options & eAdmissionRequest_destCallSignalAddress) &&get_h225_addr (ct, *data, &arq->destCallSignalAddress, - &addr, &port) && + &addr, &port) && !memcmp (&addr, &ct->tuplehash[dir].tuple.src.u3,sizeof (addr)) && port == info->sig_port[dir] && set_h225_addr && ct->status &IPS_NAT_MASK ) {
@@ -1526,7 +1526,7 @@ static int process_arq(struct sk_buff **if ((arq->options & eAdmissionRequest_srcCallSignalAddress) &&get_h225_addr (ct, *data, &arq->srcCallSignalAddress, - &addr, &port) && + &addr, &port) && !memcmp (&addr, &ct->tuplehash[dir].tuple.src.u3,sizeof (addr)) && set_h225_addr && ct->status &IPS_NAT_MASK ) {/* Calling ARQ */
net/netfilter/nf_conntrack_irc.c +1 -1
@@ -57,7 +57,7 @@ #define MINMATCHLEN 5#if 0 #define DEBUGP(format, args...) printk(KERN_DEBUG "%s:%s:" format, \ - __FILE__, __FUNCTION__ , ## args) + __FILE__, __FUNCTION__ , ## args)#else #define DEBUGP(format, args...)#endif
net/netfilter/nf_conntrack_l3proto_generic.c +1 -1
@@ -77,7 +77,7 @@ generic_prepare(struct sk_buff **pskb, ustatic u_int32_t generic_get_features (const struct nf_conntrack_tuple *tuple) - + {return NF_CT_F_BASIC ; }
net/netfilter/nf_conntrack_netbios_ns.c +1 -1
@@ -43,7 +43,7 @@ module_param(timeout, uint, 0400);MODULE_PARM_DESC (timeout,"timeout for master connection/replies in seconds" );static int help (struct sk_buff **pskb,unsigned int protoff, -struct nf_conn *ct,enum ip_conntrack_info ctinfo) +struct nf_conn *ct,enum ip_conntrack_info ctinfo) {struct nf_conntrack_expect *exp;struct iphdr *iph = (*pskb)->nh.iph;
net/netfilter/nf_conntrack_netlink.c +53 -53
385 lines not on this page · raw message
net/netfilter/nf_conntrack_pptp.c +1 -1
@@ -520,7 +520,7 @@ conntrack_pptp_help(struct sk_buff **psk tcph =skb_header_pointer (*pskb, nexthdr_off,sizeof (_tcph), &_tcph);BUG_ON (!tcph); nexthdr_off += tcph->doff *4 ; - datalen = tcplen - tcph->doff *4 ; + datalen = tcplen - tcph->doff *4 ; pptph =skb_header_pointer (*pskb, nexthdr_off,sizeof (_pptph), &_pptph);if (!pptph) {
net/netfilter/nf_conntrack_proto_sctp.c +36 -36
192 lines not on this page · raw message
net/netfilter/nf_conntrack_proto_tcp.c +102 -102
547 lines not on this page · raw message
net/netfilter/nf_conntrack_sip.c +3 -3
@@ -341,7 +341,7 @@ int ct_sip_get_info(struct nf_conn *ct,continue ; } aux =ct_sip_search (hnfo->ln_str, dptr, hnfo->ln_strlen, -ct_sip_lnlen (dptr, limit), +ct_sip_lnlen (dptr, limit), hnfo->case_sensitive);if (!aux) { DEBUGP("'%s' not found in '%s'.\n", hnfo->ln_str,
@@ -451,12 +451,12 @@ static int sip_help(struct sk_buff **psk/* We'll drop only if there are parse problems. */ if (!parse_addr (ct, dptr + matchoff,NULL , &addr, - dptr + datalen)) { + dptr + datalen)) { ret =NF_DROP ; goto out; }if (ct_sip_get_info (ct, dptr, datalen, &matchoff, &matchlen, -POS_MEDIA ) >0 ) { +POS_MEDIA ) >0 ) { port =simple_strtoul (dptr + matchoff,NULL ,10 );if (port <1024 ) {
net/netfilter/nf_conntrack_standalone.c +1 -1
@@ -472,7 +472,7 @@ #endif /* CNFIG_PROC_FS */ static void __exit nf_conntrack_standalone_fini(void) {#ifdef CONFIG_SYSCTL -unregister_sysctl_table (nf_ct_sysctl_header); +unregister_sysctl_table (nf_ct_sysctl_header);#endif #ifdef CONFIG_PROC_FS remove_proc_entry ("nf_conntrack" , proc_net_stat);
net/netfilter/nf_conntrack_tftp.c +1 -1
@@ -31,7 +31,7 @@ MODULE_PARM_DESC(ports, "Port numbers of#if 0 #define DEBUGP(format, args...) printk("%s:%s:" format, \ - __FILE__, __FUNCTION__ , ## args) + __FILE__, __FUNCTION__ , ## args)#else #define DEBUGP(format, args...)#endif
net/netfilter/nf_internals.h +1 -1
@@ -24,7 +24,7 @@ extern unsigned int nf_iterate(struct li/* nf_queue.c */ extern int nf_queue(struct sk_buff *skb, - struct list_head *elem, + struct list_head *elem, int pf, unsigned int hook, struct net_device *indev, struct net_device *outdev,
net/netfilter/nf_log.c +3 -3
@@ -41,7 +41,7 @@ int nf_log_register(int pf, struct nf_lomutex_unlock (&nf_log_mutex);return ret; -} +}EXPORT_SYMBOL (nf_log_register);void nf_log_unregister_pf (int pf)
@@ -83,7 +83,7 @@ void nf_log_packet(int pf,va_list args;char prefix[NF_LOG_PREFIXLEN ];struct nf_logger *logger; - +rcu_read_lock (); logger =rcu_dereference (nf_loggers[pf]);if (logger) {
@@ -136,7 +136,7 @@ static int seq_show(struct seq_file *s,if (!logger)return seq_printf (s,"%2lld NONE\n" , *pos); - +return seq_printf (s,"%2lld %s\n" , *pos, logger->name); }
net/netfilter/nf_queue.c +11 -11
@@ -10,7 +10,7 @@ #include <net/protocol.h>#include "nf_internals.h" -/* +/* * A queue handler may be registered for each protocol. Each is protected by * long term mutex. The handler must provide an an outfn() to accept packets * for queueing and must reinject all packets it receives, no matter what.
@@ -22,7 +22,7 @@ static DEFINE_RWLOCK(queue_handler_lock)/* return EBUSY when somebody else is registered, return EEXIST if the * same handler is registered, return 0 in case of success. */ int nf_register_queue_handler (int pf,struct nf_queue_handler *qh) -{ +{int ret;if (pf >=NPROTO )
@@ -52,7 +52,7 @@ int nf_unregister_queue_handler(int pf)write_lock_bh (&queue_handler_lock); queue_handler[pf] =NULL ;write_unlock_bh (&queue_handler_lock); - +return 0 ; }EXPORT_SYMBOL (nf_unregister_queue_handler);
@@ -70,8 +70,8 @@ void nf_unregister_queue_handlers(struct }EXPORT_SYMBOL_GPL (nf_unregister_queue_handlers); -/* -* Any packet that leaves via this function must come back +/* +* Any packet that leaves via this function must come back * through nf_reinject(). */ static int __nf_queue(struct sk_buff *skb,
@@ -115,7 +115,7 @@ #endifreturn 1 ; } - *info = (struct nf_info ) { + *info = (struct nf_info ) { (struct nf_hook_ops *)elem, pf, hook, indev, outdev, okfn };/* If it's going away, ignore hook. */
@@ -226,10 +226,10 @@ #endifmodule_put (info->elem->owner);list_for_each_rcu (i, &nf_hooks[info->pf][info->hook]) { -if (i == elem) -break ; - } - +if (i == elem) +break ; + } +if (i == &nf_hooks[info->pf][info->hook]) {/* The module which sent it to userspace is gone. */ NFDEBUG("%s: module disappeared, dropping packet.\n",
@@ -252,7 +252,7 @@ #endifif (verdict ==NF_ACCEPT ) { next_hook: verdict =nf_iterate (&nf_hooks[info->pf][info->hook], - &skb, info->hook, + &skb, info->hook, info->indev, info->outdev, &elem, info->okfn,INT_MIN ); }
net/netfilter/nf_sockopt.c +6 -6
@@ -32,13 +32,13 @@ int nf_register_sockopt(struct nf_sockoplist_for_each (i, &nf_sockopts) {struct nf_sockopt_ops *ops = (struct nf_sockopt_ops *)i;if (ops->pf == reg->pf - && (overlap (ops->set_optmin, ops->set_optmax, + && (overlap (ops->set_optmin, ops->set_optmax, reg->set_optmin, reg->set_optmax) - ||overlap (ops->get_optmin, ops->get_optmax, + ||overlap (ops->get_optmin, ops->get_optmax, reg->get_optmin, reg->get_optmax))) {NFDEBUG ("nf_sock overlap: %u-%u/%u-%u v %u-%u/%u-%u\n" , - ops->set_optmin, ops->set_optmax, - ops->get_optmin, ops->get_optmax, + ops->set_optmin, ops->set_optmax, + ops->get_optmin, ops->get_optmax, reg->set_optmin, reg->set_optmax, reg->get_optmin, reg->get_optmax); ret = -EBUSY ;
@@ -73,7 +73,7 @@ void nf_unregister_sockopt(struct nf_socEXPORT_SYMBOL (nf_unregister_sockopt);/* Call get/setsockopt() */ -static int nf_sockopt (struct sock *sk,int pf,int val, +static int nf_sockopt (struct sock *sk,int pf,int val,char __user *opt,int *len,int get) {struct list_head *i;
@@ -107,7 +107,7 @@ static int nf_sockopt(struct sock *sk, i }mutex_unlock (&nf_sockopt_mutex);return -ENOPROTOOPT ; - + out:mutex_lock (&nf_sockopt_mutex); ops->use--;
net/netfilter/nfnetlink.c +5 -5
@@ -105,7 +105,7 @@ static inline struct nfnl_callback * nfnetlink_find_client(u_int16_t type, struct nfnetlink_subsystem *ss) { u_int8_t cb_id = NFNL_MSG_TYPE(type); - +if (cb_id >= ss->cb_count) {DEBUGP ("msgtype %u >= %u, returning\n" , type, ss->cb_count);return NULL ;
@@ -187,7 +187,7 @@ nfnetlink_check_attributes(struct nfnetl/* implicit: if nlmsg_len == min_len, we return 0, and an empty * (zeroed) cda[] array. The message is valid, but empty. */ -return 0 ; +return 0 ; }int nfnetlink_has_listeners (unsigned int group)
@@ -268,12 +268,12 @@ #endif } { -u_int16_t attr_count = +u_int16_t attr_count = ss->cb[NFNL_MSG_TYPE (nlh->nlmsg_type)].attr_count;struct nfattr *cda[attr_count];memset (cda,0 ,sizeof (struct nfattr *) * attr_count); - + err =nfnetlink_check_attributes (ss, nlh, cda);if (err <0 ) goto err_inval;
@@ -357,7 +357,7 @@ static int __init nfnetlink_init(void)printk ("Netfilter messages via NETLINK v%s.\n" , nfversion); nfnl =netlink_kernel_create (NETLINK_NETFILTER ,NFNLGRP_MAX , - nfnetlink_rcv,THIS_MODULE ); + nfnetlink_rcv,THIS_MODULE );if (!nfnl) {printk (KERN_ERR "cannot initialize nfnetlink!\n" );return -1 ;
net/netfilter/nfnetlink_log.c +26 -26
174 lines not on this page · raw message
net/netfilter/nfnetlink_queue.c +42 -42
266 lines not on this page · raw message
net/netfilter/x_tables.c +6 -6
@@ -305,7 +305,7 @@ int xt_find_revision(int af, const charEXPORT_SYMBOL_GPL (xt_find_revision);int xt_check_match (const struct xt_match *match,unsigned short family, -unsigned int size,const char *table,unsigned int hook_mask, +unsigned int size,const char *table,unsigned int hook_mask,unsigned short proto,int inv_proto) {if (XT_ALIGN (match->matchsize) != size) {
@@ -377,7 +377,7 @@ int xt_compat_match_to_user(struct xt_enif (copy_to_user (cm, m,sizeof (*cm)) ||put_user (msize, &cm->u.user.match_size)) -return -EFAULT ; +return -EFAULT ;if (match->compat_to_user) {if (match->compat_to_user ((void __user *)cm->data, m->data))
@@ -432,7 +432,7 @@ int xt_compat_target_offset(struct xt_taEXPORT_SYMBOL_GPL (xt_compat_target_offset);void xt_compat_target_from_user (struct xt_entry_target *t,void **dstptr, -int *size) +int *size) {struct xt_target *target = t->u.kernel.target;struct compat_xt_entry_target *ct = (struct compat_xt_entry_target *)t;
@@ -467,7 +467,7 @@ int xt_compat_target_to_user(struct xt_eif (copy_to_user (ct, t,sizeof (*ct)) ||put_user (tsize, &ct->u.user.target_size)) -return -EFAULT ; +return -EFAULT ;if (target->compat_to_user) {if (target->compat_to_user ((void __user *)ct->data, t->data))
@@ -710,7 +710,7 @@ static void *xt_tgt_seq_start(struct seqif (mutex_lock_interruptible (&xt[af].mutex) !=0 )return NULL ; - +return xt_get_idx (list, seq, *pos); }
@@ -723,7 +723,7 @@ static void *xt_tgt_seq_next(struct seq_if (af >=NPROTO )return NULL ; - + list =type2list (af, type);if (!list)return NULL ;
net/netfilter/xt_CLASSIFY.c +2 -2
@@ -48,7 +48,7 @@ static struct xt_target xt_classify_targ .table = "mangle", .hooks = (1 << NF_IP_LOCAL_OUT) | (1 << NF_IP_FORWARD) | - (1 << NF_IP_POST_ROUTING), + (1 << NF_IP_POST_ROUTING), .me = THIS_MODULE, }, {
@@ -59,7 +59,7 @@ static struct xt_target xt_classify_targ .table = "mangle", .hooks = (1 <<NF_IP6_LOCAL_OUT ) | (1 <<NF_IP6_FORWARD ) | - (1 << NF_IP6_POST_ROUTING), + (1 << NF_IP6_POST_ROUTING), .me = THIS_MODULE, }, };
net/netfilter/xt_MARK.c +2 -2
@@ -50,11 +50,11 @@ target_v1(struct sk_buff **pskb,case XT_MARK_SET : mark = markinfo->mark;break ; - +case XT_MARK_AND : mark = (*pskb)->mark & markinfo->mark;break ; - +case XT_MARK_OR : mark = (*pskb)->mark | markinfo->mark;break ;
net/netfilter/xt_NFQUEUE.c +2 -2
@@ -3,9 +3,9 @@* (C) 2005 by Harald Welte <[email protected]> * * This program is free software; you can redistribute it and/or modify -* it under the terms of the GNU General Public License version 2 as +* it under the terms of the GNU General Public License version 2 as * published by the Free Software Foundation. -* +* */ #include <linux/module.h>
net/netfilter/xt_NOTRACK.c +2 -2
@@ -22,8 +22,8 @@ target(struct sk_buff **pskb,if ((*pskb)->nfct !=NULL )return XT_CONTINUE ; -/* Attach fake conntrack entry. -If there is a real ct entry correspondig to this packet, +/* Attach fake conntrack entry. +If there is a real ct entry correspondig to this packet, it'll hang aroun till timing out. We don't deal with it for performance reasons. JK */ nf_ct_untrack (*pskb);
net/netfilter/xt_SECMARK.c +1 -1
@@ -55,7 +55,7 @@ static int checkentry_selinux(struct xt_ {int err;struct xt_secmark_target_selinux_info *sel = &info->u.sel; - + sel->selctx[SECMARK_SELCTX_MAX -1 ] ='\0' ; err =selinux_string_to_sid (sel->selctx, &sel->selsid);
net/netfilter/xt_conntrack.c +16 -16
@@ -51,10 +51,10 @@ #define FWINV(bool, invflg) ((bool) ^ !!if (ct == &ip_conntrack_untracked) statebit =XT_CONNTRACK_STATE_UNTRACKED ;else if (ct) - statebit =XT_CONNTRACK_STATE_BIT (ctinfo); -else - statebit =XT_CONNTRACK_STATE_INVALID ; - + statebit =XT_CONNTRACK_STATE_BIT (ctinfo); +else + statebit =XT_CONNTRACK_STATE_INVALID ; +if (sinfo->flags &XT_CONNTRACK_STATE ) {if (ct) {if (test_bit (IPS_SRC_NAT_BIT , &ct->status))
@@ -77,7 +77,7 @@ #define FWINV(bool, invflg) ((bool) ^ !!FWINV (ct->tuplehash[IP_CT_DIR_ORIGINAL ].tuple.dst.protonum != sinfo->tuple[IP_CT_DIR_ORIGINAL ].dst.protonum, XT_CONNTRACK_PROTO)) - return 0; + return 0;if (sinfo->flags &XT_CONNTRACK_ORIGSRC && FWINV((ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.ip &
@@ -147,10 +147,10 @@ #define FWINV(bool,invflg) ((bool) ^ !!(if (ct == &nf_conntrack_untracked) statebit =XT_CONNTRACK_STATE_UNTRACKED ;else if (ct) - statebit =XT_CONNTRACK_STATE_BIT (ctinfo); -else - statebit =XT_CONNTRACK_STATE_INVALID ; - + statebit =XT_CONNTRACK_STATE_BIT (ctinfo); +else + statebit =XT_CONNTRACK_STATE_INVALID ; +if (sinfo->flags &XT_CONNTRACK_STATE ) {if (ct) {if (test_bit (IPS_SRC_NAT_BIT , &ct->status))
@@ -171,41 +171,41 @@ #define FWINV(bool,invflg) ((bool) ^ !!(if (sinfo->flags &XT_CONNTRACK_PROTO &&FWINV (ct->tuplehash[IP_CT_DIR_ORIGINAL ].tuple.dst.protonum != - sinfo->tuple[IP_CT_DIR_ORIGINAL ].dst.protonum, + sinfo->tuple[IP_CT_DIR_ORIGINAL ].dst.protonum,XT_CONNTRACK_PROTO )) -return 0 ; +return 0 ;if (sinfo->flags &XT_CONNTRACK_ORIGSRC &&FWINV ((ct->tuplehash[IP_CT_DIR_ORIGINAL ].tuple.src.u3.ip & - sinfo->sipmsk[IP_CT_DIR_ORIGINAL ].s_addr) != + sinfo->sipmsk[IP_CT_DIR_ORIGINAL ].s_addr) != sinfo->tuple[IP_CT_DIR_ORIGINAL ].src.ip,XT_CONNTRACK_ORIGSRC ))return 0 ;if (sinfo->flags &XT_CONNTRACK_ORIGDST &&FWINV ((ct->tuplehash[IP_CT_DIR_ORIGINAL ].tuple.dst.u3.ip & - sinfo->dipmsk[IP_CT_DIR_ORIGINAL ].s_addr) != + sinfo->dipmsk[IP_CT_DIR_ORIGINAL ].s_addr) != sinfo->tuple[IP_CT_DIR_ORIGINAL ].dst.ip,XT_CONNTRACK_ORIGDST ))return 0 ;if (sinfo->flags &XT_CONNTRACK_REPLSRC &&FWINV ((ct->tuplehash[IP_CT_DIR_REPLY ].tuple.src.u3.ip & - sinfo->sipmsk[IP_CT_DIR_REPLY ].s_addr) != + sinfo->sipmsk[IP_CT_DIR_REPLY ].s_addr) != sinfo->tuple[IP_CT_DIR_REPLY ].src.ip,XT_CONNTRACK_REPLSRC ))return 0 ;if (sinfo->flags &XT_CONNTRACK_REPLDST &&FWINV ((ct->tuplehash[IP_CT_DIR_REPLY ].tuple.dst.u3.ip & - sinfo->dipmsk[IP_CT_DIR_REPLY ].s_addr) != + sinfo->dipmsk[IP_CT_DIR_REPLY ].s_addr) != sinfo->tuple[IP_CT_DIR_REPLY ].dst.ip,XT_CONNTRACK_REPLDST ))return 0 ;if (sinfo->flags &XT_CONNTRACK_STATUS &&FWINV ((ct->status & sinfo->statusmask) ==0 , -XT_CONNTRACK_STATUS )) +XT_CONNTRACK_STATUS ))return 0 ;if (sinfo->flags &XT_CONNTRACK_EXPIRES ) {
net/netfilter/xt_dccp.c +10 -10
@@ -26,7 +26,7 @@ MODULE_DESCRIPTION("Match for DCCP protoMODULE_ALIAS ("ipt_dccp" );#define DCCHECK (cond, option, flag, invflag) (!((flag) & (option)) \ - || (!!((invflag) & (option)) ^ (cond))) + || (!!((invflag) & (option)) ^ (cond)))static unsigned char *dccp_optbuf;static DEFINE_SPINLOCK (dccp_buflock);
@@ -67,9 +67,9 @@ dccp_find_option(u_int8_t option,return 1 ; } -if (op[i] <2 ) +if (op[i] <2 ) i++; -else +else i += op[i+1 ]?:1 ; }
@@ -106,18 +106,18 @@ match(const struct sk_buff *skb,if (offset)return 0 ; - + dh =skb_header_pointer (skb, protoff,sizeof (_dh), &_dh);if (dh ==NULL ) { *hotdrop =1 ;return 0 ; - } + } -return DCCHECK (((ntohs (dh->dccph_sport) >= info->spts[0 ]) - && (ntohs (dh->dccph_sport) <= info->spts[1 ])), -XT_DCCP_SRC_PORTS , info->flags, info->invflags) - &&DCCHECK (((ntohs (dh->dccph_dport) >= info->dpts[0 ]) - && (ntohs (dh->dccph_dport) <= info->dpts[1 ])), +return DCCHECK (((ntohs (dh->dccph_sport) >= info->spts[0 ]) + && (ntohs (dh->dccph_sport) <= info->spts[1 ])), +XT_DCCP_SRC_PORTS , info->flags, info->invflags) + &&DCCHECK (((ntohs (dh->dccph_dport) >= info->dpts[0 ]) + && (ntohs (dh->dccph_dport) <= info->dpts[1 ])),XT_DCCP_DEST_PORTS , info->flags, info->invflags) &&DCCHECK (match_types (dh, info->typemask),XT_DCCP_TYPE , info->flags, info->invflags)
net/netfilter/xt_hashlimit.c +4 -4
@@ -208,7 +208,7 @@ static int htable_create(struct xt_hashlspin_lock_init (&hinfo->lock); hinfo->pde =create_proc_entry (minfo->name,0 , family ==AF_INET ? hashlimit_procdir4 : - hashlimit_procdir6); + hashlimit_procdir6);if (!hinfo->pde) {vfree (hinfo);return -1 ;
@@ -240,7 +240,7 @@ static int select_gc(struct xt_hashlimit }static void htable_selective_cleanup (struct xt_hashlimit_htable *ht, -int (*select)(struct xt_hashlimit_htable *ht, +int (*select)(struct xt_hashlimit_htable *ht,struct dsthash_ent *he)) {unsigned int i;
@@ -279,7 +279,7 @@ static void htable_destroy(struct xt_has/* remove proc entry */ remove_proc_entry (hinfo->pde->name, hinfo->family ==AF_INET ? hashlimit_procdir4 : - hashlimit_procdir6); + hashlimit_procdir6);htable_selective_cleanup (hinfo, select_all);vfree (hinfo); }
@@ -483,7 +483,7 @@ hashlimit_match(const struct sk_buff *skreturn 1 ; } -spin_unlock_bh (&hinfo->lock); +spin_unlock_bh (&hinfo->lock);/* default case: we're overlimit, thus don't match */ return 0 ;
net/netfilter/xt_helper.c +9 -9
@@ -53,7 +53,7 @@ match(const struct sk_buff *skb,struct ip_conntrack *ct;enum ip_conntrack_info ctinfo;int ret = info->invert; - + ct =ip_conntrack_get ((struct sk_buff *)skb, &ctinfo);if (!ct) {DEBUGP ("xt_helper: Eek! invalid conntrack?\n" );
@@ -67,19 +67,19 @@ match(const struct sk_buff *skb,read_lock_bh (&ip_conntrack_lock);if (!ct->master->helper) { -DEBUGP ("xt_helper: master ct %p has no helper\n" , +DEBUGP ("xt_helper: master ct %p has no helper\n" , exp->expectant); goto out_unlock; } -DEBUGP ("master's name = %s , info->name = %s\n" , +DEBUGP ("master's name = %s , info->name = %s\n" , ct->master->helper->name, info->name);if (info->name[0 ] =='\0' ) ret ^=1 ;else - ret ^= !strncmp (ct->master->helper->name, info->name, -strlen (ct->master->helper->name)); + ret ^= !strncmp (ct->master->helper->name, info->name, +strlen (ct->master->helper->name)); out_unlock:read_unlock_bh (&ip_conntrack_lock);return ret;
@@ -102,7 +102,7 @@ match(const struct sk_buff *skb,struct nf_conn_help *master_help;enum ip_conntrack_info ctinfo;int ret = info->invert; - + ct =nf_ct_get ((struct sk_buff *)skb, &ctinfo);if (!ct) {DEBUGP ("xt_helper: Eek! invalid conntrack?\n" );
@@ -117,19 +117,19 @@ match(const struct sk_buff *skb,read_lock_bh (&nf_conntrack_lock); master_help =nfct_help (ct->master);if (!master_help || !master_help->helper) { -DEBUGP ("xt_helper: master ct %p has no helper\n" , +DEBUGP ("xt_helper: master ct %p has no helper\n" , exp->expectant); goto out_unlock; } -DEBUGP ("master's name = %s , info->name = %s\n" , +DEBUGP ("master's name = %s , info->name = %s\n" , ct->master->helper->name, info->name);if (info->name[0 ] =='\0' ) ret ^=1 ;else ret ^= !strncmp (master_help->helper->name, info->name, -strlen (master_help->helper->name)); +strlen (master_help->helper->name)); out_unlock:read_unlock_bh (&nf_conntrack_lock);return ret;
net/netfilter/xt_length.c +2 -2
@@ -32,7 +32,7 @@ match(const struct sk_buff *skb, {const struct xt_length_info *info = matchinfo;u_int16_t pktlen =ntohs (skb->nh.iph->tot_len); - +return (pktlen >= info->min && pktlen <= info->max) ^ info->invert; }
@@ -48,7 +48,7 @@ match6(const struct sk_buff *skb, {const struct xt_length_info *info = matchinfo;u_int16_t pktlen =ntohs (skb->nh.ipv6h->payload_len) +sizeof (struct ipv6hdr ); - +return (pktlen >= info->min && pktlen <= info->max) ^ info->invert; }
net/netfilter/xt_limit.c +1 -1
@@ -89,7 +89,7 @@ ipt_limit_match(const struct sk_buff *skreturn 1 ; } -spin_unlock_bh (&limit_lock); +spin_unlock_bh (&limit_lock);return 0 ; }
net/netfilter/xt_mark.c +3 -3
@@ -36,10 +36,10 @@ match(const struct sk_buff *skb,static int checkentry (const char *tablename, -const void *entry, +const void *entry,const struct xt_match *match, -void *matchinfo, -unsigned int hook_mask) +void *matchinfo, +unsigned int hook_mask) {const struct xt_mark_info *minfo = matchinfo;
net/netfilter/xt_multiport.c +1 -1
@@ -91,7 +91,7 @@ ports_match_v1(const struct xt_multiport } } -return minfo->invert; +return minfo->invert; } static int
net/netfilter/xt_physdev.c +1 -1
@@ -117,7 +117,7 @@ checkentry(const char *tablename, (!(info->bitmask & XT_PHYSDEV_OP_BRIDGED) || info->invert & XT_PHYSDEV_OP_BRIDGED) && hook_mask & ((1 << NF_IP_LOCAL_OUT) | (1 << NF_IP_FORWARD) | - (1 << NF_IP_POST_ROUTING))) { + (1 << NF_IP_POST_ROUTING))) {printk (KERN_WARNING "physdev match: using --physdev-out in the " "OUTPUT, FORWARD and POSTROUTING chains for non-bridged " "traffic is not supported anymore.\n" );
net/netfilter/xt_policy.c +12 -12
@@ -109,13 +109,13 @@ match_policy_out(const struct sk_buff *s }static int match (const struct sk_buff *skb, -const struct net_device *in, -const struct net_device *out, -const struct xt_match *match, -const void *matchinfo, -int offset, -unsigned int protoff, -int *hotdrop) +const struct net_device *in, +const struct net_device *out, +const struct xt_match *match, +const void *matchinfo, +int offset, +unsigned int protoff, +int *hotdrop) {const struct xt_policy_info *info = matchinfo;int ret;
@@ -134,27 +134,27 @@ static int match(const struct sk_buff *s }static int checkentry (const char *tablename,const void *ip_void, -const struct xt_match *match, -void *matchinfo,unsigned int hook_mask) +const struct xt_match *match, +void *matchinfo,unsigned int hook_mask) {struct xt_policy_info *info = matchinfo;if (!(info->flags & (XT_POLICY_MATCH_IN |XT_POLICY_MATCH_OUT ))) {printk (KERN_ERR "xt_policy: neither incoming nor " -"outgoing policy selected\n" ); +"outgoing policy selected\n" );return 0 ; }/* hook values are equal for IPv4 and IPv6 */ if (hook_mask & (1 <<NF_IP_PRE_ROUTING |1 <<NF_IP_LOCAL_IN ) && info->flags &XT_POLICY_MATCH_OUT ) {printk (KERN_ERR "xt_policy: output policy not valid in " -"PRE_ROUTING and INPUT\n" ); +"PRE_ROUTING and INPUT\n" );return 0 ; }if (hook_mask & (1 <<NF_IP_POST_ROUTING |1 <<NF_IP_LOCAL_OUT ) && info->flags &XT_POLICY_MATCH_IN ) {printk (KERN_ERR "xt_policy: input policy not valid in " -"POST_ROUTING and OUTPUT\n" ); +"POST_ROUTING and OUTPUT\n" );return 0 ; }if (info->len >XT_POLICY_MAX_ELEM ) {
net/netfilter/xt_quota.c +2 -2
@@ -30,8 +30,8 @@ match(const struct sk_buff *skb, q->quota -= skb->len; ret ^=1 ; }else { -/* we do not allow even small packets from now on */ - q->quota =0 ; +/* we do not allow even small packets from now on */ + q->quota =0 ; }spin_unlock_bh ("a_lock);
net/netfilter/xt_realm.c +1 -1
@@ -35,7 +35,7 @@ match(const struct sk_buff *skb, {const struct xt_realm_info *info = matchinfo;struct dst_entry *dst = skb->dst; - +return (info->id == (dst->tclassid & info->mask)) ^ info->invert; }
net/netfilter/xt_sctp.c +15 -15
@@ -66,9 +66,9 @@ #endifduprintf ("Dropping invalid SCTP packet.\n" ); *hotdrop =1 ;return 0 ; - } + } -duprintf ("Chunk num: %d\toffset: %d\ttype: %d\tlength: %d\tflags: %x\n" , +duprintf ("Chunk num: %d\toffset: %d\ttype: %d\tlength: %d\tflags: %x\n" , ++i, offset, sch->type,htons (sch->length), sch->flags); offset += (ntohs (sch->length) +3 ) & ~3 ;
@@ -78,21 +78,21 @@ #endifif (SCTP_CHUNKMAP_IS_SET (chunkmap, sch->type)) {switch (chunk_match_type) {case SCTP_CHUNK_MATCH_ANY : -if (match_flags (flag_info, flag_count, +if (match_flags (flag_info, flag_count, sch->type, sch->flags)) {return 1 ; }break ;case SCTP_CHUNK_MATCH_ALL : -if (match_flags (flag_info, flag_count, +if (match_flags (flag_info, flag_count, sch->type, sch->flags)) {SCTP_CHUNKMAP_CLEAR (chunkmapcopy, sch->type); }break ;case SCTP_CHUNK_MATCH_ONLY : -if (!match_flags (flag_info, flag_count, +if (!match_flags (flag_info, flag_count, sch->type, sch->flags)) {return 0 ; }
@@ -136,24 +136,24 @@ match(const struct sk_buff *skb,duprintf ("Dropping non-first fragment.. FIXME\n" );return 0 ; } - + sh =skb_header_pointer (skb, protoff,sizeof (_sh), &_sh);if (sh ==NULL ) {duprintf ("Dropping evil TCP offset=0 tinygram.\n" ); *hotdrop =1 ;return 0 ; - } + }duprintf ("spt: %d\tdpt: %d\n" ,ntohs (sh->source),ntohs (sh->dest)); -return SCCHECK (((ntohs (sh->source) >= info->spts[0 ]) - && (ntohs (sh->source) <= info->spts[1 ])), -XT_SCTP_SRC_PORTS , info->flags, info->invflags) - &&SCCHECK (((ntohs (sh->dest) >= info->dpts[0 ]) - && (ntohs (sh->dest) <= info->dpts[1 ])), +return SCCHECK (((ntohs (sh->source) >= info->spts[0 ]) + && (ntohs (sh->source) <= info->spts[1 ])), +XT_SCTP_SRC_PORTS , info->flags, info->invflags) + &&SCCHECK (((ntohs (sh->dest) >= info->dpts[0 ]) + && (ntohs (sh->dest) <= info->dpts[1 ])),XT_SCTP_DEST_PORTS , info->flags, info->invflags) &&SCCHECK (match_packet (skb, protoff +sizeof (sctp_sctphdr_t), info->chunkmap, info->chunk_match_type, - info->flag_info, info->flag_count, + info->flag_info, info->flag_count, hotdrop),XT_SCTP_CHUNK_TYPES , info->flags, info->invflags); }
@@ -170,9 +170,9 @@ checkentry(const char *tablename,return !(info->flags & ~XT_SCTP_VALID_FLAGS ) && !(info->invflags & ~XT_SCTP_VALID_FLAGS ) && !(info->invflags & ~info->flags) - && ((!(info->flags &XT_SCTP_CHUNK_TYPES )) || + && ((!(info->flags &XT_SCTP_CHUNK_TYPES )) || (info->chunk_match_type & - (SCTP_CHUNK_MATCH_ALL + (SCTP_CHUNK_MATCH_ALL |SCTP_CHUNK_MATCH_ANY |SCTP_CHUNK_MATCH_ONLY ))); }
net/netfilter/xt_string.c +4 -4
@@ -1,5 +1,5 @@/* String matching match for iptables -* +* * (C) 2005 Pablo Neira Ayuso <[email protected]> * * This program is free software; you can redistribute it and/or modify
@@ -35,8 +35,8 @@ static int match(const struct sk_buff *smemset (&state,0 ,sizeof (struct ts_state )); -return (skb_find_text ((struct sk_buff *)skb, conf->from_offset, - conf->to_offset, conf->config, &state) +return (skb_find_text ((struct sk_buff *)skb, conf->from_offset, + conf->to_offset, conf->config, &state) !=UINT_MAX ) ^ conf->invert; }
@@ -55,7 +55,7 @@ static int checkentry(const char *tablenif (conf->from_offset > conf->to_offset)return 0 ;if (conf->algo[XT_STRING_MAX_ALGO_NAME_SIZE -1 ] !='\0' ) -return 0 ; +return 0 ;if (conf->patlen >XT_STRING_MAX_PATTERN_SIZE )return 0 ; ts_conf = textsearch_prepare(conf->algo, conf->pattern, conf->patlen,
net/netfilter/xt_tcpmss.c +2 -2
@@ -64,9 +64,9 @@ match(const struct sk_buff *skb,u_int16_t mssval; mssval = (op[i+2 ] <<8 ) | op[i+3 ]; - +return (mssval >= info->mss_min && - mssval <= info->mss_max) ^ info->invert; + mssval <= info->mss_max) ^ info->invert; }if (op[i] <2 ) i++;
PATCH 21/22 : ip6t_mh: drop piggyback payload packet on MH packets
In reply to Patrick McHardy[NETFILTER]: ip6t_mh: drop piggyback payload packet on MH packets
Regarding RFC3775, MH payload proto field should be IPPROTO_NONE. Otherwise it must be discarded (and the receiver should send ICMP error).
We assume filter should drop such piggyback everytime to disallow slipping through firewall rules, even the final receiver will discard it.
---
commit a7b4ff1031d3e62fad2b7da2068bb5deb2041325
tree ba6d4b4517999ebf0876a5618640276a6a251503
parent 353c941b0735b883510bfd7bb98fcd4fc8a0ee08
author Masahide NAKAMURA <[email protected]> Mon, 12 Feb 2007 11:08:11 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:08:11 +0100
net/ipv6/netfilter/ip6t_mh.c | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-)
net/ipv6/netfilter/ip6t_mh.c +7 -0
@@ -66,6 +66,13 @@ match(const struct sk_buff *skb,return 0 ; } +if (mh->ip6mh_proto !=IPPROTO_NONE ) { +duprintf ("Dropping invalid MH Payload Proto: %u\n" , + mh->ip6mh_proto); + *hotdrop =1 ; +return 0 ; + } +return type_match (mhinfo->types[0 ], mhinfo->types[1 ], mh->ip6mh_type, !!(mhinfo->invflags &IP6T_MH_INV_TYPE )); }
PATCH 22/22 : nf_conntrack_tcp: make sysctl variables static
In reply to Patrick McHardy[NETFILTER]: nf_conntrack_tcp: make sysctl variables static
sysctls are registered by the protocol module itself since 2.6.19, no need to have them visible to others.
---
commit ebdb23c15670f1d49d7466856f4db8fd00f6c3a9
tree 005b35c07855e3723b9d94ea77f203fb4e829924
parent a7b4ff1031d3e62fad2b7da2068bb5deb2041325
author Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:08:17 +0100 committer Patrick McHardy <[email protected]> Mon, 12 Feb 2007 11:08:17 +0100
net/netfilter/nf_conntrack_proto_tcp.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-)
net/netfilter/nf_conntrack_proto_tcp.c +3 -3
@@ -58,16 +58,16 @@ static DEFINE_RWLOCK(tcp_lock);/* "Be conservative in what you do, be liberal in what you accept from others." If it's non-zero, we mark only out of window RST segments as INVALID. */ -int nf_ct_tcp_be_liberal__read_mostly =0 ; +static int nf_ct_tcp_be_liberal__read_mostly =0 ;/* If it is set to zero, we disable picking up already established connections. */ -int nf_ct_tcp_loose__read_mostly =1 ; +static int nf_ct_tcp_loose__read_mostly =1 ;/* Max number of the retransmitted packets without receiving an (acceptable) ACK from the destination. If this number is reached, a shorter timer will be started. */ -int nf_ct_tcp_max_retrans__read_mostly =3 ; +static int nf_ct_tcp_max_retrans__read_mostly =3 ;/* FIXME: Examine ipfilter's timeouts and conntrack transitions more closely. They're more complex. --RR */
David Miller
maintainerFrom: Patrick McHardy <[email protected]>
Date: Mon, 12 Feb 2007 11:36:21 +0100 (MET)
Quoted 8 lines
Hi Dave,
following are a number of fixes for netfilter, fixing RCU issues all over the place, including the invalid preemption assumptions in the
NF_CT_STAT_INCmacro. Additionally there are a few minor enhancements for things I noticed while going over the code: replacement of a few unnecessary spinlocks by mutexes, some cleanup fornf_log.cand some return value cleanup for unregistration functions.
Quoted 4 lines
With these patches I think there should be nothing left holding back the preemptable RCU merge from netfilter's perspective.
Please apply, thanks.
Thanks a lot for doing this work Patrick.
All applied, thanks again.