Commit d3d40f23 authored by Mark Tomlinson's avatar Mark Tomlinson Committed by Pablo Neira Ayuso

Revert "netfilter: x_tables: Switch synchronization to RCU"

This reverts commit cc00bcaa.

This (and the preceding) patch basically re-implemented the RCU
mechanisms of patch 78454473. That patch was replaced because of the
performance problems that it created when replacing tables. Now, we have
the same issue: the call to synchronize_rcu() makes replacing tables
slower by as much as an order of magnitude.

Prior to using RCU a script calling "iptables" approx. 200 times was
taking 1.16s. With RCU this increased to 11.59s.

Revert these patches and fix the issue in a different way.
Signed-off-by: default avatarMark Tomlinson <mark.tomlinson@alliedtelesis.co.nz>
Signed-off-by: default avatarPablo Neira Ayuso <pablo@netfilter.org>
parent abe7034b
...@@ -227,7 +227,7 @@ struct xt_table { ...@@ -227,7 +227,7 @@ struct xt_table {
unsigned int valid_hooks; unsigned int valid_hooks;
/* Man behind the curtain... */ /* Man behind the curtain... */
struct xt_table_info __rcu *private; struct xt_table_info *private;
/* Set this to THIS_MODULE if you are a module, otherwise NULL */ /* Set this to THIS_MODULE if you are a module, otherwise NULL */
struct module *me; struct module *me;
...@@ -448,9 +448,6 @@ xt_get_per_cpu_counter(struct xt_counters *cnt, unsigned int cpu) ...@@ -448,9 +448,6 @@ xt_get_per_cpu_counter(struct xt_counters *cnt, unsigned int cpu)
struct nf_hook_ops *xt_hook_ops_alloc(const struct xt_table *, nf_hookfn *); struct nf_hook_ops *xt_hook_ops_alloc(const struct xt_table *, nf_hookfn *);
struct xt_table_info
*xt_table_get_private_protected(const struct xt_table *table);
#ifdef CONFIG_COMPAT #ifdef CONFIG_COMPAT
#include <net/compat.h> #include <net/compat.h>
......
...@@ -203,7 +203,7 @@ unsigned int arpt_do_table(struct sk_buff *skb, ...@@ -203,7 +203,7 @@ unsigned int arpt_do_table(struct sk_buff *skb,
local_bh_disable(); local_bh_disable();
addend = xt_write_recseq_begin(); addend = xt_write_recseq_begin();
private = rcu_access_pointer(table->private); private = READ_ONCE(table->private); /* Address dependency. */
cpu = smp_processor_id(); cpu = smp_processor_id();
table_base = private->entries; table_base = private->entries;
jumpstack = (struct arpt_entry **)private->jumpstack[cpu]; jumpstack = (struct arpt_entry **)private->jumpstack[cpu];
...@@ -649,7 +649,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table) ...@@ -649,7 +649,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table)
{ {
unsigned int countersize; unsigned int countersize;
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
/* We need atomic snapshot of counters: rest doesn't change /* We need atomic snapshot of counters: rest doesn't change
* (other than comefrom, which userspace doesn't care * (other than comefrom, which userspace doesn't care
...@@ -673,7 +673,7 @@ static int copy_entries_to_user(unsigned int total_size, ...@@ -673,7 +673,7 @@ static int copy_entries_to_user(unsigned int total_size,
unsigned int off, num; unsigned int off, num;
const struct arpt_entry *e; const struct arpt_entry *e;
struct xt_counters *counters; struct xt_counters *counters;
struct xt_table_info *private = xt_table_get_private_protected(table); struct xt_table_info *private = table->private;
int ret = 0; int ret = 0;
void *loc_cpu_entry; void *loc_cpu_entry;
...@@ -807,7 +807,7 @@ static int get_info(struct net *net, void __user *user, const int *len) ...@@ -807,7 +807,7 @@ static int get_info(struct net *net, void __user *user, const int *len)
t = xt_request_find_table_lock(net, NFPROTO_ARP, name); t = xt_request_find_table_lock(net, NFPROTO_ARP, name);
if (!IS_ERR(t)) { if (!IS_ERR(t)) {
struct arpt_getinfo info; struct arpt_getinfo info;
const struct xt_table_info *private = xt_table_get_private_protected(t); const struct xt_table_info *private = t->private;
#ifdef CONFIG_COMPAT #ifdef CONFIG_COMPAT
struct xt_table_info tmp; struct xt_table_info tmp;
...@@ -860,7 +860,7 @@ static int get_entries(struct net *net, struct arpt_get_entries __user *uptr, ...@@ -860,7 +860,7 @@ static int get_entries(struct net *net, struct arpt_get_entries __user *uptr,
t = xt_find_table_lock(net, NFPROTO_ARP, get.name); t = xt_find_table_lock(net, NFPROTO_ARP, get.name);
if (!IS_ERR(t)) { if (!IS_ERR(t)) {
const struct xt_table_info *private = xt_table_get_private_protected(t); const struct xt_table_info *private = t->private;
if (get.size == private->size) if (get.size == private->size)
ret = copy_entries_to_user(private->size, ret = copy_entries_to_user(private->size,
...@@ -1017,7 +1017,7 @@ static int do_add_counters(struct net *net, sockptr_t arg, unsigned int len) ...@@ -1017,7 +1017,7 @@ static int do_add_counters(struct net *net, sockptr_t arg, unsigned int len)
} }
local_bh_disable(); local_bh_disable();
private = xt_table_get_private_protected(t); private = t->private;
if (private->number != tmp.num_counters) { if (private->number != tmp.num_counters) {
ret = -EINVAL; ret = -EINVAL;
goto unlock_up_free; goto unlock_up_free;
...@@ -1330,7 +1330,7 @@ static int compat_copy_entries_to_user(unsigned int total_size, ...@@ -1330,7 +1330,7 @@ static int compat_copy_entries_to_user(unsigned int total_size,
void __user *userptr) void __user *userptr)
{ {
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
void __user *pos; void __user *pos;
unsigned int size; unsigned int size;
int ret = 0; int ret = 0;
......
...@@ -258,7 +258,7 @@ ipt_do_table(struct sk_buff *skb, ...@@ -258,7 +258,7 @@ ipt_do_table(struct sk_buff *skb,
WARN_ON(!(table->valid_hooks & (1 << hook))); WARN_ON(!(table->valid_hooks & (1 << hook)));
local_bh_disable(); local_bh_disable();
addend = xt_write_recseq_begin(); addend = xt_write_recseq_begin();
private = rcu_access_pointer(table->private); private = READ_ONCE(table->private); /* Address dependency. */
cpu = smp_processor_id(); cpu = smp_processor_id();
table_base = private->entries; table_base = private->entries;
jumpstack = (struct ipt_entry **)private->jumpstack[cpu]; jumpstack = (struct ipt_entry **)private->jumpstack[cpu];
...@@ -791,7 +791,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table) ...@@ -791,7 +791,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table)
{ {
unsigned int countersize; unsigned int countersize;
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
/* We need atomic snapshot of counters: rest doesn't change /* We need atomic snapshot of counters: rest doesn't change
(other than comefrom, which userspace doesn't care (other than comefrom, which userspace doesn't care
...@@ -815,7 +815,7 @@ copy_entries_to_user(unsigned int total_size, ...@@ -815,7 +815,7 @@ copy_entries_to_user(unsigned int total_size,
unsigned int off, num; unsigned int off, num;
const struct ipt_entry *e; const struct ipt_entry *e;
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
int ret = 0; int ret = 0;
const void *loc_cpu_entry; const void *loc_cpu_entry;
...@@ -964,7 +964,7 @@ static int get_info(struct net *net, void __user *user, const int *len) ...@@ -964,7 +964,7 @@ static int get_info(struct net *net, void __user *user, const int *len)
t = xt_request_find_table_lock(net, AF_INET, name); t = xt_request_find_table_lock(net, AF_INET, name);
if (!IS_ERR(t)) { if (!IS_ERR(t)) {
struct ipt_getinfo info; struct ipt_getinfo info;
const struct xt_table_info *private = xt_table_get_private_protected(t); const struct xt_table_info *private = t->private;
#ifdef CONFIG_COMPAT #ifdef CONFIG_COMPAT
struct xt_table_info tmp; struct xt_table_info tmp;
...@@ -1018,7 +1018,7 @@ get_entries(struct net *net, struct ipt_get_entries __user *uptr, ...@@ -1018,7 +1018,7 @@ get_entries(struct net *net, struct ipt_get_entries __user *uptr,
t = xt_find_table_lock(net, AF_INET, get.name); t = xt_find_table_lock(net, AF_INET, get.name);
if (!IS_ERR(t)) { if (!IS_ERR(t)) {
const struct xt_table_info *private = xt_table_get_private_protected(t); const struct xt_table_info *private = t->private;
if (get.size == private->size) if (get.size == private->size)
ret = copy_entries_to_user(private->size, ret = copy_entries_to_user(private->size,
t, uptr->entrytable); t, uptr->entrytable);
...@@ -1173,7 +1173,7 @@ do_add_counters(struct net *net, sockptr_t arg, unsigned int len) ...@@ -1173,7 +1173,7 @@ do_add_counters(struct net *net, sockptr_t arg, unsigned int len)
} }
local_bh_disable(); local_bh_disable();
private = xt_table_get_private_protected(t); private = t->private;
if (private->number != tmp.num_counters) { if (private->number != tmp.num_counters) {
ret = -EINVAL; ret = -EINVAL;
goto unlock_up_free; goto unlock_up_free;
...@@ -1543,7 +1543,7 @@ compat_copy_entries_to_user(unsigned int total_size, struct xt_table *table, ...@@ -1543,7 +1543,7 @@ compat_copy_entries_to_user(unsigned int total_size, struct xt_table *table,
void __user *userptr) void __user *userptr)
{ {
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
void __user *pos; void __user *pos;
unsigned int size; unsigned int size;
int ret = 0; int ret = 0;
......
...@@ -280,7 +280,7 @@ ip6t_do_table(struct sk_buff *skb, ...@@ -280,7 +280,7 @@ ip6t_do_table(struct sk_buff *skb,
local_bh_disable(); local_bh_disable();
addend = xt_write_recseq_begin(); addend = xt_write_recseq_begin();
private = rcu_access_pointer(table->private); private = READ_ONCE(table->private); /* Address dependency. */
cpu = smp_processor_id(); cpu = smp_processor_id();
table_base = private->entries; table_base = private->entries;
jumpstack = (struct ip6t_entry **)private->jumpstack[cpu]; jumpstack = (struct ip6t_entry **)private->jumpstack[cpu];
...@@ -807,7 +807,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table) ...@@ -807,7 +807,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table)
{ {
unsigned int countersize; unsigned int countersize;
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
/* We need atomic snapshot of counters: rest doesn't change /* We need atomic snapshot of counters: rest doesn't change
(other than comefrom, which userspace doesn't care (other than comefrom, which userspace doesn't care
...@@ -831,7 +831,7 @@ copy_entries_to_user(unsigned int total_size, ...@@ -831,7 +831,7 @@ copy_entries_to_user(unsigned int total_size,
unsigned int off, num; unsigned int off, num;
const struct ip6t_entry *e; const struct ip6t_entry *e;
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
int ret = 0; int ret = 0;
const void *loc_cpu_entry; const void *loc_cpu_entry;
...@@ -980,7 +980,7 @@ static int get_info(struct net *net, void __user *user, const int *len) ...@@ -980,7 +980,7 @@ static int get_info(struct net *net, void __user *user, const int *len)
t = xt_request_find_table_lock(net, AF_INET6, name); t = xt_request_find_table_lock(net, AF_INET6, name);
if (!IS_ERR(t)) { if (!IS_ERR(t)) {
struct ip6t_getinfo info; struct ip6t_getinfo info;
const struct xt_table_info *private = xt_table_get_private_protected(t); const struct xt_table_info *private = t->private;
#ifdef CONFIG_COMPAT #ifdef CONFIG_COMPAT
struct xt_table_info tmp; struct xt_table_info tmp;
...@@ -1035,7 +1035,7 @@ get_entries(struct net *net, struct ip6t_get_entries __user *uptr, ...@@ -1035,7 +1035,7 @@ get_entries(struct net *net, struct ip6t_get_entries __user *uptr,
t = xt_find_table_lock(net, AF_INET6, get.name); t = xt_find_table_lock(net, AF_INET6, get.name);
if (!IS_ERR(t)) { if (!IS_ERR(t)) {
struct xt_table_info *private = xt_table_get_private_protected(t); struct xt_table_info *private = t->private;
if (get.size == private->size) if (get.size == private->size)
ret = copy_entries_to_user(private->size, ret = copy_entries_to_user(private->size,
t, uptr->entrytable); t, uptr->entrytable);
...@@ -1189,7 +1189,7 @@ do_add_counters(struct net *net, sockptr_t arg, unsigned int len) ...@@ -1189,7 +1189,7 @@ do_add_counters(struct net *net, sockptr_t arg, unsigned int len)
} }
local_bh_disable(); local_bh_disable();
private = xt_table_get_private_protected(t); private = t->private;
if (private->number != tmp.num_counters) { if (private->number != tmp.num_counters) {
ret = -EINVAL; ret = -EINVAL;
goto unlock_up_free; goto unlock_up_free;
...@@ -1552,7 +1552,7 @@ compat_copy_entries_to_user(unsigned int total_size, struct xt_table *table, ...@@ -1552,7 +1552,7 @@ compat_copy_entries_to_user(unsigned int total_size, struct xt_table *table,
void __user *userptr) void __user *userptr)
{ {
struct xt_counters *counters; struct xt_counters *counters;
const struct xt_table_info *private = xt_table_get_private_protected(table); const struct xt_table_info *private = table->private;
void __user *pos; void __user *pos;
unsigned int size; unsigned int size;
int ret = 0; int ret = 0;
......
...@@ -1351,14 +1351,6 @@ struct xt_counters *xt_counters_alloc(unsigned int counters) ...@@ -1351,14 +1351,6 @@ struct xt_counters *xt_counters_alloc(unsigned int counters)
} }
EXPORT_SYMBOL(xt_counters_alloc); EXPORT_SYMBOL(xt_counters_alloc);
struct xt_table_info
*xt_table_get_private_protected(const struct xt_table *table)
{
return rcu_dereference_protected(table->private,
mutex_is_locked(&xt[table->af].mutex));
}
EXPORT_SYMBOL(xt_table_get_private_protected);
struct xt_table_info * struct xt_table_info *
xt_replace_table(struct xt_table *table, xt_replace_table(struct xt_table *table,
unsigned int num_counters, unsigned int num_counters,
...@@ -1366,6 +1358,7 @@ xt_replace_table(struct xt_table *table, ...@@ -1366,6 +1358,7 @@ xt_replace_table(struct xt_table *table,
int *error) int *error)
{ {
struct xt_table_info *private; struct xt_table_info *private;
unsigned int cpu;
int ret; int ret;
ret = xt_jumpstack_alloc(newinfo); ret = xt_jumpstack_alloc(newinfo);
...@@ -1375,20 +1368,47 @@ xt_replace_table(struct xt_table *table, ...@@ -1375,20 +1368,47 @@ xt_replace_table(struct xt_table *table,
} }
/* Do the substitution. */ /* Do the substitution. */
private = xt_table_get_private_protected(table); local_bh_disable();
private = table->private;
/* Check inside lock: is the old number correct? */ /* Check inside lock: is the old number correct? */
if (num_counters != private->number) { if (num_counters != private->number) {
pr_debug("num_counters != table->private->number (%u/%u)\n", pr_debug("num_counters != table->private->number (%u/%u)\n",
num_counters, private->number); num_counters, private->number);
local_bh_enable();
*error = -EAGAIN; *error = -EAGAIN;
return NULL; return NULL;
} }
newinfo->initial_entries = private->initial_entries; newinfo->initial_entries = private->initial_entries;
/*
* Ensure contents of newinfo are visible before assigning to
* private.
*/
smp_wmb();
table->private = newinfo;
/* make sure all cpus see new ->private value */
smp_wmb();
/*
* Even though table entries have now been swapped, other CPU's
* may still be using the old entries...
*/
local_bh_enable();
/* ... so wait for even xt_recseq on all cpus */
for_each_possible_cpu(cpu) {
seqcount_t *s = &per_cpu(xt_recseq, cpu);
u32 seq = raw_read_seqcount(s);
rcu_assign_pointer(table->private, newinfo); if (seq & 1) {
synchronize_rcu(); do {
cond_resched();
cpu_relax();
} while (seq == raw_read_seqcount(s));
}
}
audit_log_nfcfg(table->name, table->af, private->number, audit_log_nfcfg(table->name, table->af, private->number,
!private->number ? AUDIT_XT_OP_REGISTER : !private->number ? AUDIT_XT_OP_REGISTER :
...@@ -1424,12 +1444,12 @@ struct xt_table *xt_register_table(struct net *net, ...@@ -1424,12 +1444,12 @@ struct xt_table *xt_register_table(struct net *net,
} }
/* Simplifies replace_table code. */ /* Simplifies replace_table code. */
rcu_assign_pointer(table->private, bootstrap); table->private = bootstrap;
if (!xt_replace_table(table, 0, newinfo, &ret)) if (!xt_replace_table(table, 0, newinfo, &ret))
goto unlock; goto unlock;
private = xt_table_get_private_protected(table); private = table->private;
pr_debug("table->private->number = %u\n", private->number); pr_debug("table->private->number = %u\n", private->number);
/* save number of initial entries */ /* save number of initial entries */
...@@ -1452,8 +1472,7 @@ void *xt_unregister_table(struct xt_table *table) ...@@ -1452,8 +1472,7 @@ void *xt_unregister_table(struct xt_table *table)
struct xt_table_info *private; struct xt_table_info *private;
mutex_lock(&xt[table->af].mutex); mutex_lock(&xt[table->af].mutex);
private = xt_table_get_private_protected(table); private = table->private;
RCU_INIT_POINTER(table->private, NULL);
list_del(&table->list); list_del(&table->list);
mutex_unlock(&xt[table->af].mutex); mutex_unlock(&xt[table->af].mutex);
audit_log_nfcfg(table->name, table->af, private->number, audit_log_nfcfg(table->name, table->af, private->number,
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment