Skip to content

L
linux
Commit 0f7ff927 authored by Arnaldo Carvalho de Melo's avatar Arnaldo Carvalho de Melo Committed by David S. Miller
Browse files
Options

[INET]: Just rename the TCP hashtable functions/structs to inet_ 

This is to break down the complexity of the series of patches,
making it very clear that this one just does:

1. renames tcp_ prefixed hashtable functions and data structures that
   were already mostly generic to inet_ to share it with DCCP and
   other INET transport protocols.

2. Removes not used functions (__tb_head & tb_head)

3. Removes some leftover prototypes in the headers (tcp_bucket_unlock &
   tcp_v4_build_header)

Next changesets will move tcp_sk(sk)->bind_hash to inet_sock so that we can
make functions such as tcp_inherit_port, __tcp_inherit_port, tcp_v4_get_port,
__tcp_put_port,  generic and get others like tcp_destroy_sock closer to generic
(tcp_orphan_count will go to sk->sk_prot to allow this).

Eventually most of these functions will be used passing the transport protocol
inet_hashinfo structure.
Signed-off-by: default avatarArnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent 304a1618
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 139 additions and 139 deletions
include/linux/tcp.h
View file @ 0f7ff927
......@@ -258,7 +258,7 @@ struct tcp_sock {
__u32 snd_sml; /* Last byte of the most recently transmitted small packet */
__u32 rcv_tstamp; /* timestamp of last received ACK (for keepalives) */
__u32 lsndtime; /* timestamp of last sent data packet (for restart window) */
struct tcp_bind_bucket *bind_hash;
struct inet_bind_bucket *bind_hash;
/* Delayed ACK control data */
struct {
__u8 pending; /* ACK is pending */
......
include/net/tcp.h
View file @ 0f7ff927
......@@ -44,13 +44,13 @@
* New scheme, half the table is for TIME_WAIT, the other half is
* for the rest. I'll experiment with dynamic table growth later.
*/
struct tcp_ehash_bucket {
struct inet_ehash_bucket {
rwlock_t lock;
struct hlist_head chain;
} __attribute__((__aligned__(8)));
/* This is for listening sockets, thus all sockets which possess wildcards. */
#define TCP_LHTABLE_SIZE 32 /* Yes, really, this is all you need. */
#define INET_LHTABLE_SIZE 32 /* Yes, really, this is all you need. */
/* There are a few simple rules, which allow for local port reuse by
* an application. In essence:
......@@ -83,31 +83,22 @@ struct tcp_ehash_bucket {
* users logged onto your box, isn't it nice to know that new data
* ports are created in O(1) time? I thought so. ;-) -DaveM
*/
struct tcp_bind_bucket {
struct inet_bind_bucket {
unsigned short port;
signed short fastreuse;
struct hlist_node node;
struct hlist_head owners;
};
#define tb_for_each(tb, node, head) hlist_for_each_entry(tb, node, head, node)
#define inet_bind_bucket_for_each(tb, node, head) \
hlist_for_each_entry(tb, node, head, node)
struct tcp_bind_hashbucket {
struct inet_bind_hashbucket {
spinlock_t lock;
struct hlist_head chain;
};
static inline struct tcp_bind_bucket *__tb_head(struct tcp_bind_hashbucket *head)
{
return hlist_entry(head->chain.first, struct tcp_bind_bucket, node);
}
static inline struct tcp_bind_bucket *tb_head(struct tcp_bind_hashbucket *head)
{
return hlist_empty(&head->chain) ? NULL : __tb_head(head);
}
extern struct tcp_hashinfo {
struct inet_hashinfo {
/* This is for sockets with full identity only. Sockets here will
* always be without wildcards and will have the following invariant:
*
......@@ -116,21 +107,21 @@ extern struct tcp_hashinfo {
* First half of the table is for sockets not in TIME_WAIT, second half
* is for TIME_WAIT sockets only.
*/
struct tcp_ehash_bucket *__tcp_ehash;
struct inet_ehash_bucket *ehash;
/* Ok, let's try this, I give up, we do need a local binding
* TCP hash as well as the others for fast bind/connect.
*/
struct tcp_bind_hashbucket *__tcp_bhash;
struct inet_bind_hashbucket *bhash;
int __tcp_bhash_size;
int __tcp_ehash_size;
int bhash_size;
int ehash_size;
/* All sockets in TCP_LISTEN state will be in here. This is the only
* table where wildcard'd TCP sockets can exist. Hash function here
* is just local port number.
*/
struct hlist_head __tcp_listening_hash[TCP_LHTABLE_SIZE];
struct hlist_head listening_hash[INET_LHTABLE_SIZE];
/* All the above members are written once at bootup and
* never written again _or_ are predominantly read-access.
......@@ -138,36 +129,39 @@ extern struct tcp_hashinfo {
* Now align to a new cache line as all the following members
* are often dirty.
*/
rwlock_t __tcp_lhash_lock ____cacheline_aligned;
atomic_t __tcp_lhash_users;
wait_queue_head_t __tcp_lhash_wait;
spinlock_t __tcp_portalloc_lock;
} tcp_hashinfo;
#define tcp_ehash (tcp_hashinfo.__tcp_ehash)
#define tcp_bhash (tcp_hashinfo.__tcp_bhash)
#define tcp_ehash_size (tcp_hashinfo.__tcp_ehash_size)
#define tcp_bhash_size (tcp_hashinfo.__tcp_bhash_size)
#define tcp_listening_hash (tcp_hashinfo.__tcp_listening_hash)
#define tcp_lhash_lock (tcp_hashinfo.__tcp_lhash_lock)
#define tcp_lhash_users (tcp_hashinfo.__tcp_lhash_users)
#define tcp_lhash_wait (tcp_hashinfo.__tcp_lhash_wait)
#define tcp_portalloc_lock (tcp_hashinfo.__tcp_portalloc_lock)
rwlock_t lhash_lock ____cacheline_aligned;
atomic_t lhash_users;
wait_queue_head_t lhash_wait;
spinlock_t portalloc_lock;
};
extern struct inet_hashinfo tcp_hashinfo;
#define tcp_ehash (tcp_hashinfo.ehash)
#define tcp_bhash (tcp_hashinfo.bhash)
#define tcp_ehash_size (tcp_hashinfo.ehash_size)
#define tcp_bhash_size (tcp_hashinfo.bhash_size)
#define tcp_listening_hash (tcp_hashinfo.listening_hash)
#define tcp_lhash_lock (tcp_hashinfo.lhash_lock)
#define tcp_lhash_users (tcp_hashinfo.lhash_users)
#define tcp_lhash_wait (tcp_hashinfo.lhash_wait)
#define tcp_portalloc_lock (tcp_hashinfo.portalloc_lock)
extern kmem_cache_t *tcp_bucket_cachep;
extern struct tcp_bind_bucket *tcp_bucket_create(struct tcp_bind_hashbucket *head,
unsigned short snum);
extern void tcp_bucket_destroy(struct tcp_bind_bucket *tb);
extern void tcp_bucket_unlock(struct sock *sk);
extern struct inet_bind_bucket *
inet_bind_bucket_create(kmem_cache_t *cachep,
struct inet_bind_hashbucket *head,
const unsigned short snum);
extern void inet_bind_bucket_destroy(kmem_cache_t *cachep,
struct inet_bind_bucket *tb);
extern int tcp_port_rover;
/* These are AF independent. */
static __inline__ int tcp_bhashfn(__u16 lport)
static inline int inet_bhashfn(const __u16 lport, const int bhash_size)
{
return (lport & (tcp_bhash_size - 1));
return lport & (bhash_size - 1);
}
extern void tcp_bind_hash(struct sock *sk, struct tcp_bind_bucket *tb,
extern void tcp_bind_hash(struct sock *sk, struct inet_bind_bucket *tb,
unsigned short snum);
#if (BITS_PER_LONG == 64)
......@@ -212,7 +206,7 @@ struct tcp_tw_bucket {
__u32 tw_ts_recent;
long tw_ts_recent_stamp;
unsigned long tw_ttd;
struct tcp_bind_bucket *tw_tb;
struct inet_bind_bucket *tw_tb;
struct hlist_node tw_death_node;
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
struct in6_addr tw_v6_daddr;
......@@ -366,14 +360,14 @@ extern void tcp_tw_deschedule(struct tcp_tw_bucket *tw);
(!((__sk)->sk_bound_dev_if) || ((__sk)->sk_bound_dev_if == (__dif))))
/* These can have wildcards, don't try too hard. */
static __inline__ int tcp_lhashfn(unsigned short num)
static inline int inet_lhashfn(const unsigned short num)
{
return num & (TCP_LHTABLE_SIZE - 1);
return num & (INET_LHTABLE_SIZE - 1);
}
static __inline__ int tcp_sk_listen_hashfn(struct sock *sk)
static inline int inet_sk_listen_hashfn(const struct sock *sk)
{
return tcp_lhashfn(inet_sk(sk)->num);
return inet_lhashfn(inet_sk(sk)->num);
}
#define MAX_TCP_HEADER (128 + MAX_HEADER)
......@@ -799,9 +793,6 @@ extern void tcp_parse_options(struct sk_buff *skb,
* TCP v4 functions exported for the inet6 API
*/
extern int tcp_v4_build_header(struct sock *sk,
struct sk_buff *skb);
extern void tcp_v4_send_check(struct sock *sk,
struct tcphdr *th, int len,
struct sk_buff *skb);
......
net/ipv4/tcp.c
View file @ 0f7ff927
......@@ -272,6 +272,9 @@ int sysctl_tcp_fin_timeout = TCP_FIN_TIMEOUT;
DEFINE_SNMP_STAT(struct tcp_mib, tcp_statistics);
kmem_cache_t *tcp_bucket_cachep;
EXPORT_SYMBOL_GPL(tcp_bucket_cachep);
kmem_cache_t *tcp_timewait_cachep;
atomic_t tcp_orphan_count = ATOMIC_INIT(0);
......@@ -2259,7 +2262,7 @@ void __init tcp_init(void)
sizeof(skb->cb));
tcp_bucket_cachep = kmem_cache_create("tcp_bind_bucket",
sizeof(struct tcp_bind_bucket),
sizeof(struct inet_bind_bucket),
0, SLAB_HWCACHE_ALIGN,
NULL, NULL);
if (!tcp_bucket_cachep)
......@@ -2277,9 +2280,9 @@ void __init tcp_init(void)
*
* The methodology is similar to that of the buffer cache.
*/
tcp_ehash = (struct tcp_ehash_bucket *)
tcp_ehash =
alloc_large_system_hash("TCP established",
sizeof(struct tcp_ehash_bucket),
sizeof(struct inet_ehash_bucket),
thash_entries,
(num_physpages >= 128 * 1024) ?
(25 - PAGE_SHIFT) :
......@@ -2294,9 +2297,9 @@ void __init tcp_init(void)
INIT_HLIST_HEAD(&tcp_ehash[i].chain);
}
tcp_bhash = (struct tcp_bind_hashbucket *)
tcp_bhash =
alloc_large_system_hash("TCP bind",
sizeof(struct tcp_bind_hashbucket),
sizeof(struct inet_bind_hashbucket),
tcp_ehash_size,
(num_physpages >= 128 * 1024) ?
(25 - PAGE_SHIFT) :
......@@ -2315,7 +2318,7 @@ void __init tcp_init(void)
* on available memory.
*/
for (order = 0; ((1 << order) << PAGE_SHIFT) <
(tcp_bhash_size * sizeof(struct tcp_bind_hashbucket));
(tcp_bhash_size * sizeof(struct inet_bind_hashbucket));
order++)
;
if (order >= 4) {
......
net/ipv4/tcp_diag.c
View file @ 0f7ff927
......@@ -590,7 +590,7 @@ static int tcpdiag_dump(struct sk_buff *skb, struct netlink_callback *cb)
if (!(r->tcpdiag_states&(TCPF_LISTEN|TCPF_SYN_RECV)))
goto skip_listen_ht;
tcp_listen_lock();
for (i = s_i; i < TCP_LHTABLE_SIZE; i++) {
for (i = s_i; i < INET_LHTABLE_SIZE; i++) {
struct sock *sk;
struct hlist_node *node;
......@@ -646,7 +646,7 @@ static int tcpdiag_dump(struct sk_buff *skb, struct netlink_callback *cb)
return skb->len;
for (i = s_i; i < tcp_ehash_size; i++) {
struct tcp_ehash_bucket *head = &tcp_ehash[i];
struct inet_ehash_bucket *head = &tcp_ehash[i];
struct sock *sk;
struct hlist_node *node;
......
net/ipv4/tcp_ipv4.c
View file @ 0f7ff927
This diff is collapsed.
net/ipv4/tcp_minisocks.c
View file @ 0f7ff927
......@@ -60,9 +60,9 @@ int tcp_tw_count;
/* Must be called with locally disabled BHs. */
static void tcp_timewait_kill(struct tcp_tw_bucket *tw)
{
struct tcp_ehash_bucket *ehead;
struct tcp_bind_hashbucket *bhead;
struct tcp_bind_bucket *tb;
struct inet_ehash_bucket *ehead;
struct inet_bind_hashbucket *bhead;
struct inet_bind_bucket *tb;
/* Unlink from established hashes. */
ehead = &tcp_ehash[tw->tw_hashent];
......@@ -76,12 +76,12 @@ static void tcp_timewait_kill(struct tcp_tw_bucket *tw)
write_unlock(&ehead->lock);
/* Disassociate with bind bucket. */
bhead = &tcp_bhash[tcp_bhashfn(tw->tw_num)];
bhead = &tcp_bhash[inet_bhashfn(tw->tw_num, tcp_bhash_size)];
spin_lock(&bhead->lock);
tb = tw->tw_tb;
__hlist_del(&tw->tw_bind_node);
tw->tw_tb = NULL;
tcp_bucket_destroy(tb);
inet_bind_bucket_destroy(tcp_bucket_cachep, tb);
spin_unlock(&bhead->lock);
#ifdef SOCK_REFCNT_DEBUG
......@@ -296,14 +296,14 @@ tcp_timewait_state_process(struct tcp_tw_bucket *tw, struct sk_buff *skb,
*/
static void __tcp_tw_hashdance(struct sock *sk, struct tcp_tw_bucket *tw)
{
struct tcp_ehash_bucket *ehead = &tcp_ehash[sk->sk_hashent];
struct tcp_bind_hashbucket *bhead;
struct inet_ehash_bucket *ehead = &tcp_ehash[sk->sk_hashent];
struct inet_bind_hashbucket *bhead;
/* Step 1: Put TW into bind hash. Original socket stays there too.
Note, that any socket with inet_sk(sk)->num != 0 MUST be bound in
binding cache, even if it is closed.
*/
bhead = &tcp_bhash[tcp_bhashfn(inet_sk(sk)->num)];
bhead = &tcp_bhash[inet_bhashfn(inet_sk(sk)->num, tcp_bhash_size)];
spin_lock(&bhead->lock);
tw->tw_tb = tcp_sk(sk)->bind_hash;
BUG_TRAP(tcp_sk(sk)->bind_hash);
......
net/ipv6/tcp_ipv6.c
View file @ 0f7ff927
......@@ -98,11 +98,11 @@ static __inline__ int tcp_v6_sk_hashfn(struct sock *sk)
return tcp_v6_hashfn(laddr, lport, faddr, fport);
}
static inline int tcp_v6_bind_conflict(struct sock *sk,
struct tcp_bind_bucket *tb)
static inline int tcp_v6_bind_conflict(const struct sock *sk,
const struct inet_bind_bucket *tb)
{
struct sock *sk2;
struct hlist_node *node;
const struct sock *sk2;
const struct hlist_node *node;
/* We must walk the whole port owner list in this case. -DaveM */
sk_for_each_bound(sk2, node, &tb->owners) {
......@@ -126,8 +126,8 @@ static inline int tcp_v6_bind_conflict(struct sock *sk,
*/
static int tcp_v6_get_port(struct sock *sk, unsigned short snum)
{
struct tcp_bind_hashbucket *head;
struct tcp_bind_bucket *tb;
struct inet_bind_hashbucket *head;
struct inet_bind_bucket *tb;
struct hlist_node *node;
int ret;
......@@ -146,9 +146,9 @@ static int tcp_v6_get_port(struct sock *sk, unsigned short snum)
do { rover++;
if (rover > high)
rover = low;
head = &tcp_bhash[tcp_bhashfn(rover)];
head = &tcp_bhash[inet_bhashfn(rover, tcp_bhash_size)];
spin_lock(&head->lock);
tb_for_each(tb, node, &head->chain)
inet_bind_bucket_for_each(tb, node, &head->chain)
if (tb->port == rover)
goto next;
break;
......@@ -171,9 +171,9 @@ static int tcp_v6_get_port(struct sock *sk, unsigned short snum)
/* OK, here is the one we will use. */
snum = rover;
} else {
head = &tcp_bhash[tcp_bhashfn(snum)];
head = &tcp_bhash[inet_bhashfn(snum, tcp_bhash_size)];
spin_lock(&head->lock);
tb_for_each(tb, node, &head->chain)
inet_bind_bucket_for_each(tb, node, &head->chain)
if (tb->port == snum)
goto tb_found;
}
......@@ -192,7 +192,7 @@ static int tcp_v6_get_port(struct sock *sk, unsigned short snum)
}
tb_not_found:
ret = 1;
if (!tb && (tb = tcp_bucket_create(head, snum)) == NULL)
if (!tb && (tb = inet_bind_bucket_create(tcp_bucket_cachep, head, snum)) == NULL)
goto fail_unlock;
if (hlist_empty(&tb->owners)) {
if (sk->sk_reuse && sk->sk_state != TCP_LISTEN)
......@@ -224,7 +224,7 @@ static __inline__ void __tcp_v6_hash(struct sock *sk)
BUG_TRAP(sk_unhashed(sk));
if (sk->sk_state == TCP_LISTEN) {
list = &tcp_listening_hash[tcp_sk_listen_hashfn(sk)];
list = &tcp_listening_hash[inet_sk_listen_hashfn(sk)];
lock = &tcp_lhash_lock;
tcp_listen_wlock();
} else {
......@@ -264,7 +264,7 @@ static struct sock *tcp_v6_lookup_listener(struct in6_addr *daddr, unsigned shor
hiscore=0;
read_lock(&tcp_lhash_lock);
sk_for_each(sk, node, &tcp_listening_hash[tcp_lhashfn(hnum)]) {
sk_for_each(sk, node, &tcp_listening_hash[inet_lhashfn(hnum)]) {
if (inet_sk(sk)->num == hnum && sk->sk_family == PF_INET6) {
struct ipv6_pinfo *np = inet6_sk(sk);
......@@ -305,7 +305,7 @@ static inline struct sock *__tcp_v6_lookup_established(struct in6_addr *saddr, u
struct in6_addr *daddr, u16 hnum,
int dif)
{
struct tcp_ehash_bucket *head;
struct inet_ehash_bucket *head;
struct sock *sk;
struct hlist_node *node;
__u32 ports = TCP_COMBINED_PORTS(sport, hnum);
......@@ -461,7 +461,7 @@ static int __tcp_v6_check_established(struct sock *sk, __u16 lport,
int dif = sk->sk_bound_dev_if;
u32 ports = TCP_COMBINED_PORTS(inet->dport, lport);
int hash = tcp_v6_hashfn(daddr, inet->num, saddr, inet->dport);
struct tcp_ehash_bucket *head = &tcp_ehash[hash];
struct inet_ehash_bucket *head = &tcp_ehash[hash];
struct sock *sk2;
struct hlist_node *node;
struct tcp_tw_bucket *tw;
......@@ -540,8 +540,8 @@ static inline u32 tcpv6_port_offset(const struct sock *sk)
static int tcp_v6_hash_connect(struct sock *sk)
{
unsigned short snum = inet_sk(sk)->num;
struct tcp_bind_hashbucket *head;
struct tcp_bind_bucket *tb;
struct inet_bind_hashbucket *head;
struct inet_bind_bucket *tb;
int ret;
if (!snum) {
......@@ -558,14 +558,14 @@ static int tcp_v6_hash_connect(struct sock *sk)
local_bh_disable();
for (i = 1; i <= range; i++) {
port = low + (i + offset) % range;
head = &tcp_bhash[tcp_bhashfn(port)];
head = &tcp_bhash[inet_bhashfn(port, tcp_bhash_size)];
spin_lock(&head->lock);
/* Does not bother with rcv_saddr checks,
* because the established check is already
* unique enough.
*/
tb_for_each(tb, node, &head->chain) {
inet_bind_bucket_for_each(tb, node, &head->chain) {
if (tb->port == port) {
BUG_TRAP(!hlist_empty(&tb->owners));
if (tb->fastreuse >= 0)
......@@ -578,7 +578,7 @@ static int tcp_v6_hash_connect(struct sock *sk)
}
}
tb = tcp_bucket_create(head, port);
tb = inet_bind_bucket_create(tcp_bucket_cachep, head, port);
if (!tb) {
spin_unlock(&head->lock);
break;
......@@ -613,7 +613,7 @@ static int tcp_v6_hash_connect(struct sock *sk)
goto out;
}
head = &tcp_bhash[tcp_bhashfn(snum)];
head = &tcp_bhash[inet_bhashfn(snum, tcp_bhash_size)];
tb = tcp_sk(sk)->bind_hash;
spin_lock_bh(&head->lock);
......
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
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7