linux/net/ipv4/inetpeer.c
Eric Dumazet 2c1409a0a2 inetpeer: Optimize inet_getid()
While investigating for network latencies, I found inet_getid() was a
contention point for some workloads, as inet_peer_idlock is shared
by all inet_getid() users regardless of peers.

One way to fix this is to make ip_id_count an atomic_t instead
of __u16, and use atomic_add_return().

In order to keep sizeof(struct inet_peer) = 64 on 64bit arches
tcp_ts_stamp is also converted to __u32 instead of "unsigned long".

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-11-13 20:46:58 -08:00

461 lines
14 KiB
C

/*
* INETPEER - A storage for permanent information about peers
*
* This source is covered by the GNU GPL, the same as all kernel sources.
*
* Authors: Andrey V. Savochkin <saw@msu.ru>
*/
#include <linux/module.h>
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/spinlock.h>
#include <linux/random.h>
#include <linux/timer.h>
#include <linux/time.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/net.h>
#include <net/ip.h>
#include <net/inetpeer.h>
/*
* Theory of operations.
* We keep one entry for each peer IP address. The nodes contains long-living
* information about the peer which doesn't depend on routes.
* At this moment this information consists only of ID field for the next
* outgoing IP packet. This field is incremented with each packet as encoded
* in inet_getid() function (include/net/inetpeer.h).
* At the moment of writing this notes identifier of IP packets is generated
* to be unpredictable using this code only for packets subjected
* (actually or potentially) to defragmentation. I.e. DF packets less than
* PMTU in size uses a constant ID and do not use this code (see
* ip_select_ident() in include/net/ip.h).
*
* Route cache entries hold references to our nodes.
* New cache entries get references via lookup by destination IP address in
* the avl tree. The reference is grabbed only when it's needed i.e. only
* when we try to output IP packet which needs an unpredictable ID (see
* __ip_select_ident() in net/ipv4/route.c).
* Nodes are removed only when reference counter goes to 0.
* When it's happened the node may be removed when a sufficient amount of
* time has been passed since its last use. The less-recently-used entry can
* also be removed if the pool is overloaded i.e. if the total amount of
* entries is greater-or-equal than the threshold.
*
* Node pool is organised as an AVL tree.
* Such an implementation has been chosen not just for fun. It's a way to
* prevent easy and efficient DoS attacks by creating hash collisions. A huge
* amount of long living nodes in a single hash slot would significantly delay
* lookups performed with disabled BHs.
*
* Serialisation issues.
* 1. Nodes may appear in the tree only with the pool write lock held.
* 2. Nodes may disappear from the tree only with the pool write lock held
* AND reference count being 0.
* 3. Nodes appears and disappears from unused node list only under
* "inet_peer_unused_lock".
* 4. Global variable peer_total is modified under the pool lock.
* 5. struct inet_peer fields modification:
* avl_left, avl_right, avl_parent, avl_height: pool lock
* unused: unused node list lock
* refcnt: atomically against modifications on other CPU;
* usually under some other lock to prevent node disappearing
* dtime: unused node list lock
* v4daddr: unchangeable
* ip_id_count: idlock
*/
static struct kmem_cache *peer_cachep __read_mostly;
#define node_height(x) x->avl_height
static struct inet_peer peer_fake_node = {
.avl_left = &peer_fake_node,
.avl_right = &peer_fake_node,
.avl_height = 0
};
#define peer_avl_empty (&peer_fake_node)
static struct inet_peer *peer_root = peer_avl_empty;
static DEFINE_RWLOCK(peer_pool_lock);
#define PEER_MAXDEPTH 40 /* sufficient for about 2^27 nodes */
static int peer_total;
/* Exported for sysctl_net_ipv4. */
int inet_peer_threshold __read_mostly = 65536 + 128; /* start to throw entries more
* aggressively at this stage */
int inet_peer_minttl __read_mostly = 120 * HZ; /* TTL under high load: 120 sec */
int inet_peer_maxttl __read_mostly = 10 * 60 * HZ; /* usual time to live: 10 min */
int inet_peer_gc_mintime __read_mostly = 10 * HZ;
int inet_peer_gc_maxtime __read_mostly = 120 * HZ;
static LIST_HEAD(unused_peers);
static DEFINE_SPINLOCK(inet_peer_unused_lock);
static void peer_check_expire(unsigned long dummy);
static DEFINE_TIMER(peer_periodic_timer, peer_check_expire, 0, 0);
/* Called from ip_output.c:ip_init */
void __init inet_initpeers(void)
{
struct sysinfo si;
/* Use the straight interface to information about memory. */
si_meminfo(&si);
/* The values below were suggested by Alexey Kuznetsov
* <kuznet@ms2.inr.ac.ru>. I don't have any opinion about the values
* myself. --SAW
*/
if (si.totalram <= (32768*1024)/PAGE_SIZE)
inet_peer_threshold >>= 1; /* max pool size about 1MB on IA32 */
if (si.totalram <= (16384*1024)/PAGE_SIZE)
inet_peer_threshold >>= 1; /* about 512KB */
if (si.totalram <= (8192*1024)/PAGE_SIZE)
inet_peer_threshold >>= 2; /* about 128KB */
peer_cachep = kmem_cache_create("inet_peer_cache",
sizeof(struct inet_peer),
0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
NULL);
/* All the timers, started at system startup tend
to synchronize. Perturb it a bit.
*/
peer_periodic_timer.expires = jiffies
+ net_random() % inet_peer_gc_maxtime
+ inet_peer_gc_maxtime;
add_timer(&peer_periodic_timer);
}
/* Called with or without local BH being disabled. */
static void unlink_from_unused(struct inet_peer *p)
{
spin_lock_bh(&inet_peer_unused_lock);
list_del_init(&p->unused);
spin_unlock_bh(&inet_peer_unused_lock);
}
/*
* Called with local BH disabled and the pool lock held.
* _stack is known to be NULL or not at compile time,
* so compiler will optimize the if (_stack) tests.
*/
#define lookup(_daddr, _stack) \
({ \
struct inet_peer *u, **v; \
if (_stack != NULL) { \
stackptr = _stack; \
*stackptr++ = &peer_root; \
} \
for (u = peer_root; u != peer_avl_empty; ) { \
if (_daddr == u->v4daddr) \
break; \
if ((__force __u32)_daddr < (__force __u32)u->v4daddr) \
v = &u->avl_left; \
else \
v = &u->avl_right; \
if (_stack != NULL) \
*stackptr++ = v; \
u = *v; \
} \
u; \
})
/* Called with local BH disabled and the pool write lock held. */
#define lookup_rightempty(start) \
({ \
struct inet_peer *u, **v; \
*stackptr++ = &start->avl_left; \
v = &start->avl_left; \
for (u = *v; u->avl_right != peer_avl_empty; ) { \
v = &u->avl_right; \
*stackptr++ = v; \
u = *v; \
} \
u; \
})
/* Called with local BH disabled and the pool write lock held.
* Variable names are the proof of operation correctness.
* Look into mm/map_avl.c for more detail description of the ideas. */
static void peer_avl_rebalance(struct inet_peer **stack[],
struct inet_peer ***stackend)
{
struct inet_peer **nodep, *node, *l, *r;
int lh, rh;
while (stackend > stack) {
nodep = *--stackend;
node = *nodep;
l = node->avl_left;
r = node->avl_right;
lh = node_height(l);
rh = node_height(r);
if (lh > rh + 1) { /* l: RH+2 */
struct inet_peer *ll, *lr, *lrl, *lrr;
int lrh;
ll = l->avl_left;
lr = l->avl_right;
lrh = node_height(lr);
if (lrh <= node_height(ll)) { /* ll: RH+1 */
node->avl_left = lr; /* lr: RH or RH+1 */
node->avl_right = r; /* r: RH */
node->avl_height = lrh + 1; /* RH+1 or RH+2 */
l->avl_left = ll; /* ll: RH+1 */
l->avl_right = node; /* node: RH+1 or RH+2 */
l->avl_height = node->avl_height + 1;
*nodep = l;
} else { /* ll: RH, lr: RH+1 */
lrl = lr->avl_left; /* lrl: RH or RH-1 */
lrr = lr->avl_right; /* lrr: RH or RH-1 */
node->avl_left = lrr; /* lrr: RH or RH-1 */
node->avl_right = r; /* r: RH */
node->avl_height = rh + 1; /* node: RH+1 */
l->avl_left = ll; /* ll: RH */
l->avl_right = lrl; /* lrl: RH or RH-1 */
l->avl_height = rh + 1; /* l: RH+1 */
lr->avl_left = l; /* l: RH+1 */
lr->avl_right = node; /* node: RH+1 */
lr->avl_height = rh + 2;
*nodep = lr;
}
} else if (rh > lh + 1) { /* r: LH+2 */
struct inet_peer *rr, *rl, *rlr, *rll;
int rlh;
rr = r->avl_right;
rl = r->avl_left;
rlh = node_height(rl);
if (rlh <= node_height(rr)) { /* rr: LH+1 */
node->avl_right = rl; /* rl: LH or LH+1 */
node->avl_left = l; /* l: LH */
node->avl_height = rlh + 1; /* LH+1 or LH+2 */
r->avl_right = rr; /* rr: LH+1 */
r->avl_left = node; /* node: LH+1 or LH+2 */
r->avl_height = node->avl_height + 1;
*nodep = r;
} else { /* rr: RH, rl: RH+1 */
rlr = rl->avl_right; /* rlr: LH or LH-1 */
rll = rl->avl_left; /* rll: LH or LH-1 */
node->avl_right = rll; /* rll: LH or LH-1 */
node->avl_left = l; /* l: LH */
node->avl_height = lh + 1; /* node: LH+1 */
r->avl_right = rr; /* rr: LH */
r->avl_left = rlr; /* rlr: LH or LH-1 */
r->avl_height = lh + 1; /* r: LH+1 */
rl->avl_right = r; /* r: LH+1 */
rl->avl_left = node; /* node: LH+1 */
rl->avl_height = lh + 2;
*nodep = rl;
}
} else {
node->avl_height = (lh > rh ? lh : rh) + 1;
}
}
}
/* Called with local BH disabled and the pool write lock held. */
#define link_to_pool(n) \
do { \
n->avl_height = 1; \
n->avl_left = peer_avl_empty; \
n->avl_right = peer_avl_empty; \
**--stackptr = n; \
peer_avl_rebalance(stack, stackptr); \
} while(0)
/* May be called with local BH enabled. */
static void unlink_from_pool(struct inet_peer *p)
{
int do_free;
do_free = 0;
write_lock_bh(&peer_pool_lock);
/* Check the reference counter. It was artificially incremented by 1
* in cleanup() function to prevent sudden disappearing. If the
* reference count is still 1 then the node is referenced only as `p'
* here and from the pool. So under the exclusive pool lock it's safe
* to remove the node and free it later. */
if (atomic_read(&p->refcnt) == 1) {
struct inet_peer **stack[PEER_MAXDEPTH];
struct inet_peer ***stackptr, ***delp;
if (lookup(p->v4daddr, stack) != p)
BUG();
delp = stackptr - 1; /* *delp[0] == p */
if (p->avl_left == peer_avl_empty) {
*delp[0] = p->avl_right;
--stackptr;
} else {
/* look for a node to insert instead of p */
struct inet_peer *t;
t = lookup_rightempty(p);
BUG_ON(*stackptr[-1] != t);
**--stackptr = t->avl_left;
/* t is removed, t->v4daddr > x->v4daddr for any
* x in p->avl_left subtree.
* Put t in the old place of p. */
*delp[0] = t;
t->avl_left = p->avl_left;
t->avl_right = p->avl_right;
t->avl_height = p->avl_height;
BUG_ON(delp[1] != &p->avl_left);
delp[1] = &t->avl_left; /* was &p->avl_left */
}
peer_avl_rebalance(stack, stackptr);
peer_total--;
do_free = 1;
}
write_unlock_bh(&peer_pool_lock);
if (do_free)
kmem_cache_free(peer_cachep, p);
else
/* The node is used again. Decrease the reference counter
* back. The loop "cleanup -> unlink_from_unused
* -> unlink_from_pool -> putpeer -> link_to_unused
* -> cleanup (for the same node)"
* doesn't really exist because the entry will have a
* recent deletion time and will not be cleaned again soon. */
inet_putpeer(p);
}
/* May be called with local BH enabled. */
static int cleanup_once(unsigned long ttl)
{
struct inet_peer *p = NULL;
/* Remove the first entry from the list of unused nodes. */
spin_lock_bh(&inet_peer_unused_lock);
if (!list_empty(&unused_peers)) {
__u32 delta;
p = list_first_entry(&unused_peers, struct inet_peer, unused);
delta = (__u32)jiffies - p->dtime;
if (delta < ttl) {
/* Do not prune fresh entries. */
spin_unlock_bh(&inet_peer_unused_lock);
return -1;
}
list_del_init(&p->unused);
/* Grab an extra reference to prevent node disappearing
* before unlink_from_pool() call. */
atomic_inc(&p->refcnt);
}
spin_unlock_bh(&inet_peer_unused_lock);
if (p == NULL)
/* It means that the total number of USED entries has
* grown over inet_peer_threshold. It shouldn't really
* happen because of entry limits in route cache. */
return -1;
unlink_from_pool(p);
return 0;
}
/* Called with or without local BH being disabled. */
struct inet_peer *inet_getpeer(__be32 daddr, int create)
{
struct inet_peer *p, *n;
struct inet_peer **stack[PEER_MAXDEPTH], ***stackptr;
/* Look up for the address quickly. */
read_lock_bh(&peer_pool_lock);
p = lookup(daddr, NULL);
if (p != peer_avl_empty)
atomic_inc(&p->refcnt);
read_unlock_bh(&peer_pool_lock);
if (p != peer_avl_empty) {
/* The existing node has been found. */
/* Remove the entry from unused list if it was there. */
unlink_from_unused(p);
return p;
}
if (!create)
return NULL;
/* Allocate the space outside the locked region. */
n = kmem_cache_alloc(peer_cachep, GFP_ATOMIC);
if (n == NULL)
return NULL;
n->v4daddr = daddr;
atomic_set(&n->refcnt, 1);
atomic_set(&n->rid, 0);
atomic_set(&n->ip_id_count, secure_ip_id(daddr));
n->tcp_ts_stamp = 0;
write_lock_bh(&peer_pool_lock);
/* Check if an entry has suddenly appeared. */
p = lookup(daddr, stack);
if (p != peer_avl_empty)
goto out_free;
/* Link the node. */
link_to_pool(n);
INIT_LIST_HEAD(&n->unused);
peer_total++;
write_unlock_bh(&peer_pool_lock);
if (peer_total >= inet_peer_threshold)
/* Remove one less-recently-used entry. */
cleanup_once(0);
return n;
out_free:
/* The appropriate node is already in the pool. */
atomic_inc(&p->refcnt);
write_unlock_bh(&peer_pool_lock);
/* Remove the entry from unused list if it was there. */
unlink_from_unused(p);
/* Free preallocated the preallocated node. */
kmem_cache_free(peer_cachep, n);
return p;
}
/* Called with local BH disabled. */
static void peer_check_expire(unsigned long dummy)
{
unsigned long now = jiffies;
int ttl;
if (peer_total >= inet_peer_threshold)
ttl = inet_peer_minttl;
else
ttl = inet_peer_maxttl
- (inet_peer_maxttl - inet_peer_minttl) / HZ *
peer_total / inet_peer_threshold * HZ;
while (!cleanup_once(ttl)) {
if (jiffies != now)
break;
}
/* Trigger the timer after inet_peer_gc_mintime .. inet_peer_gc_maxtime
* interval depending on the total number of entries (more entries,
* less interval). */
if (peer_total >= inet_peer_threshold)
peer_periodic_timer.expires = jiffies + inet_peer_gc_mintime;
else
peer_periodic_timer.expires = jiffies
+ inet_peer_gc_maxtime
- (inet_peer_gc_maxtime - inet_peer_gc_mintime) / HZ *
peer_total / inet_peer_threshold * HZ;
add_timer(&peer_periodic_timer);
}
void inet_putpeer(struct inet_peer *p)
{
spin_lock_bh(&inet_peer_unused_lock);
if (atomic_dec_and_test(&p->refcnt)) {
list_add_tail(&p->unused, &unused_peers);
p->dtime = (__u32)jiffies;
}
spin_unlock_bh(&inet_peer_unused_lock);
}