linux/net/ipv4/inet_connection_sock.c

940 lines
26 KiB
C
Raw Normal View History

/*
* INET An implementation of the TCP/IP protocol suite for the LINUX
* operating system. INET is implemented using the BSD Socket
* interface as the means of communication with the user level.
*
* Support for INET connection oriented protocols.
*
* Authors: See the TCP sources
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or(at your option) any later version.
*/
#include <linux/module.h>
#include <linux/jhash.h>
#include <net/inet_connection_sock.h>
#include <net/inet_hashtables.h>
#include <net/inet_timewait_sock.h>
#include <net/ip.h>
#include <net/route.h>
#include <net/tcp_states.h>
#include <net/xfrm.h>
#ifdef INET_CSK_DEBUG
const char inet_csk_timer_bug_msg[] = "inet_csk BUG: unknown timer value\n";
EXPORT_SYMBOL(inet_csk_timer_bug_msg);
#endif
unsigned long *sysctl_local_reserved_ports;
EXPORT_SYMBOL(sysctl_local_reserved_ports);
void inet_get_local_port_range(struct net *net, int *low, int *high)
{
unsigned int seq;
do {
seq = read_seqbegin(&net->ipv4.sysctl_local_ports.lock);
*low = net->ipv4.sysctl_local_ports.range[0];
*high = net->ipv4.sysctl_local_ports.range[1];
} while (read_seqretry(&net->ipv4.sysctl_local_ports.lock, seq));
}
EXPORT_SYMBOL(inet_get_local_port_range);
int inet_csk_bind_conflict(const struct sock *sk,
const struct inet_bind_bucket *tb, bool relax)
{
struct sock *sk2;
int reuse = sk->sk_reuse;
int reuseport = sk->sk_reuseport;
kuid_t uid = sock_i_uid((struct sock *)sk);
/*
* Unlike other sk lookup places we do not check
* for sk_net here, since _all_ the socks listed
* in tb->owners list belong to the same net - the
* one this bucket belongs to.
*/
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
sk_for_each_bound(sk2, &tb->owners) {
if (sk != sk2 &&
!inet_v6_ipv6only(sk2) &&
(!sk->sk_bound_dev_if ||
!sk2->sk_bound_dev_if ||
sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) {
if ((!reuse || !sk2->sk_reuse ||
sk2->sk_state == TCP_LISTEN) &&
(!reuseport || !sk2->sk_reuseport ||
(sk2->sk_state != TCP_TIME_WAIT &&
!uid_eq(uid, sock_i_uid(sk2))))) {
if (!sk2->sk_rcv_saddr || !sk->sk_rcv_saddr ||
sk2->sk_rcv_saddr == sk->sk_rcv_saddr)
break;
}
if (!relax && reuse && sk2->sk_reuse &&
sk2->sk_state != TCP_LISTEN) {
if (!sk2->sk_rcv_saddr || !sk->sk_rcv_saddr ||
sk2->sk_rcv_saddr == sk->sk_rcv_saddr)
break;
}
}
}
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
return sk2 != NULL;
}
EXPORT_SYMBOL_GPL(inet_csk_bind_conflict);
/* Obtain a reference to a local port for the given sock,
* if snum is zero it means select any available local port.
*/
[SOCK] proto: Add hashinfo member to struct proto This way we can remove TCP and DCCP specific versions of sk->sk_prot->get_port: both v4 and v6 use inet_csk_get_port sk->sk_prot->hash: inet_hash is directly used, only v6 need a specific version to deal with mapped sockets sk->sk_prot->unhash: both v4 and v6 use inet_hash directly struct inet_connection_sock_af_ops also gets a new member, bind_conflict, so that inet_csk_get_port can find the per family routine. Now only the lookup routines receive as a parameter a struct inet_hashtable. With this we further reuse code, reducing the difference among INET transport protocols. Eventually work has to be done on UDP and SCTP to make them share this infrastructure and get as a bonus inet_diag interfaces so that iproute can be used with these protocols. net-2.6/net/ipv4/inet_hashtables.c: struct proto | +8 struct inet_connection_sock_af_ops | +8 2 structs changed __inet_hash_nolisten | +18 __inet_hash | -210 inet_put_port | +8 inet_bind_bucket_create | +1 __inet_hash_connect | -8 5 functions changed, 27 bytes added, 218 bytes removed, diff: -191 net-2.6/net/core/sock.c: proto_seq_show | +3 1 function changed, 3 bytes added, diff: +3 net-2.6/net/ipv4/inet_connection_sock.c: inet_csk_get_port | +15 1 function changed, 15 bytes added, diff: +15 net-2.6/net/ipv4/tcp.c: tcp_set_state | -7 1 function changed, 7 bytes removed, diff: -7 net-2.6/net/ipv4/tcp_ipv4.c: tcp_v4_get_port | -31 tcp_v4_hash | -48 tcp_v4_destroy_sock | -7 tcp_v4_syn_recv_sock | -2 tcp_unhash | -179 5 functions changed, 267 bytes removed, diff: -267 net-2.6/net/ipv6/inet6_hashtables.c: __inet6_hash | +8 1 function changed, 8 bytes added, diff: +8 net-2.6/net/ipv4/inet_hashtables.c: inet_unhash | +190 inet_hash | +242 2 functions changed, 432 bytes added, diff: +432 vmlinux: 16 functions changed, 485 bytes added, 492 bytes removed, diff: -7 /home/acme/git/net-2.6/net/ipv6/tcp_ipv6.c: tcp_v6_get_port | -31 tcp_v6_hash | -7 tcp_v6_syn_recv_sock | -9 3 functions changed, 47 bytes removed, diff: -47 /home/acme/git/net-2.6/net/dccp/proto.c: dccp_destroy_sock | -7 dccp_unhash | -179 dccp_hash | -49 dccp_set_state | -7 dccp_done | +1 5 functions changed, 1 bytes added, 242 bytes removed, diff: -241 /home/acme/git/net-2.6/net/dccp/ipv4.c: dccp_v4_get_port | -31 dccp_v4_request_recv_sock | -2 2 functions changed, 33 bytes removed, diff: -33 /home/acme/git/net-2.6/net/dccp/ipv6.c: dccp_v6_get_port | -31 dccp_v6_hash | -7 dccp_v6_request_recv_sock | +5 3 functions changed, 5 bytes added, 38 bytes removed, diff: -33 Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-02-03 12:06:04 +00:00
int inet_csk_get_port(struct sock *sk, unsigned short snum)
{
struct inet_hashinfo *hashinfo = sk->sk_prot->h.hashinfo;
struct inet_bind_hashbucket *head;
struct inet_bind_bucket *tb;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
int ret, attempts = 5;
struct net *net = sock_net(sk);
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
int smallest_size = -1, smallest_rover;
kuid_t uid = sock_i_uid(sk);
local_bh_disable();
if (!snum) {
int remaining, rover, low, high;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
again:
inet_get_local_port_range(net, &low, &high);
remaining = (high - low) + 1;
smallest_rover = rover = prandom_u32() % remaining + low;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
smallest_size = -1;
do {
if (inet_is_reserved_local_port(rover))
goto next_nolock;
head = &hashinfo->bhash[inet_bhashfn(net, rover,
hashinfo->bhash_size)];
spin_lock(&head->lock);
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
inet_bind_bucket_for_each(tb, &head->chain)
if (net_eq(ib_net(tb), net) && tb->port == rover) {
if (((tb->fastreuse > 0 &&
sk->sk_reuse &&
sk->sk_state != TCP_LISTEN) ||
(tb->fastreuseport > 0 &&
sk->sk_reuseport &&
uid_eq(tb->fastuid, uid))) &&
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
(tb->num_owners < smallest_size || smallest_size == -1)) {
smallest_size = tb->num_owners;
smallest_rover = rover;
if (atomic_read(&hashinfo->bsockets) > (high - low) + 1 &&
!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, false)) {
snum = smallest_rover;
goto tb_found;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
}
}
if (!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, false)) {
snum = rover;
goto tb_found;
}
goto next;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
}
break;
next:
spin_unlock(&head->lock);
next_nolock:
if (++rover > high)
rover = low;
} while (--remaining > 0);
/* Exhausted local port range during search? It is not
* possible for us to be holding one of the bind hash
* locks if this test triggers, because if 'remaining'
* drops to zero, we broke out of the do/while loop at
* the top level, not from the 'break;' statement.
*/
ret = 1;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
if (remaining <= 0) {
if (smallest_size != -1) {
snum = smallest_rover;
goto have_snum;
}
goto fail;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
}
/* OK, here is the one we will use. HEAD is
* non-NULL and we hold it's mutex.
*/
snum = rover;
} else {
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
have_snum:
head = &hashinfo->bhash[inet_bhashfn(net, snum,
hashinfo->bhash_size)];
spin_lock(&head->lock);
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
inet_bind_bucket_for_each(tb, &head->chain)
if (net_eq(ib_net(tb), net) && tb->port == snum)
goto tb_found;
}
tb = NULL;
goto tb_not_found;
tb_found:
if (!hlist_empty(&tb->owners)) {
if (sk->sk_reuse == SK_FORCE_REUSE)
goto success;
if (((tb->fastreuse > 0 &&
sk->sk_reuse && sk->sk_state != TCP_LISTEN) ||
(tb->fastreuseport > 0 &&
sk->sk_reuseport && uid_eq(tb->fastuid, uid))) &&
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
smallest_size == -1) {
goto success;
} else {
ret = 1;
if (inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, true)) {
if (((sk->sk_reuse && sk->sk_state != TCP_LISTEN) ||
(tb->fastreuseport > 0 &&
sk->sk_reuseport && uid_eq(tb->fastuid, uid))) &&
smallest_size != -1 && --attempts >= 0) {
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
spin_unlock(&head->lock);
goto again;
}
goto fail_unlock;
inet: Allowing more than 64k connections and heavily optimize bind(0) time. With simple extension to the binding mechanism, which allows to bind more than 64k sockets (or smaller amount, depending on sysctl parameters), we have to traverse the whole bind hash table to find out empty bucket. And while it is not a problem for example for 32k connections, bind() completion time grows exponentially (since after each successful binding we have to traverse one bucket more to find empty one) even if we start each time from random offset inside the hash table. So, when hash table is full, and we want to add another socket, we have to traverse the whole table no matter what, so effectivelly this will be the worst case performance and it will be constant. Attached picture shows bind() time depending on number of already bound sockets. Green area corresponds to the usual binding to zero port process, which turns on kernel port selection as described above. Red area is the bind process, when number of reuse-bound sockets is not limited by 64k (or sysctl parameters). The same exponential growth (hidden by the green area) before number of ports reaches sysctl limit. At this time bind hash table has exactly one reuse-enbaled socket in a bucket, but it is possible that they have different addresses. Actually kernel selects the first port to try randomly, so at the beginning bind will take roughly constant time, but with time number of port to check after random start will increase. And that will have exponential growth, but because of above random selection, not every next port selection will necessary take longer time than previous. So we have to consider the area below in the graph (if you could zoom it, you could find, that there are many different times placed there), so area can hide another. Blue area corresponds to the port selection optimization. This is rather simple design approach: hashtable now maintains (unprecise and racely updated) number of currently bound sockets, and when number of such sockets becomes greater than predefined value (I use maximum port range defined by sysctls), we stop traversing the whole bind hash table and just stop at first matching bucket after random start. Above limit roughly corresponds to the case, when bind hash table is full and we turned on mechanism of allowing to bind more reuse-enabled sockets, so it does not change behaviour of other sockets. Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net> Tested-by: Denys Fedoryschenko <denys@visp.net.lb> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 00:46:02 +00:00
}
}
}
tb_not_found:
ret = 1;
if (!tb && (tb = inet_bind_bucket_create(hashinfo->bind_bucket_cachep,
net, head, snum)) == NULL)
goto fail_unlock;
if (hlist_empty(&tb->owners)) {
if (sk->sk_reuse && sk->sk_state != TCP_LISTEN)
tb->fastreuse = 1;
else
tb->fastreuse = 0;
if (sk->sk_reuseport) {
tb->fastreuseport = 1;
tb->fastuid = uid;
} else
tb->fastreuseport = 0;
} else {
if (tb->fastreuse &&
(!sk->sk_reuse || sk->sk_state == TCP_LISTEN))
tb->fastreuse = 0;
if (tb->fastreuseport &&
(!sk->sk_reuseport || !uid_eq(tb->fastuid, uid)))
tb->fastreuseport = 0;
}
success:
if (!inet_csk(sk)->icsk_bind_hash)
inet_bind_hash(sk, tb, snum);
WARN_ON(inet_csk(sk)->icsk_bind_hash != tb);
ret = 0;
fail_unlock:
spin_unlock(&head->lock);
fail:
local_bh_enable();
return ret;
}
EXPORT_SYMBOL_GPL(inet_csk_get_port);
/*
* Wait for an incoming connection, avoid race conditions. This must be called
* with the socket locked.
*/
static int inet_csk_wait_for_connect(struct sock *sk, long timeo)
{
struct inet_connection_sock *icsk = inet_csk(sk);
DEFINE_WAIT(wait);
int err;
/*
* True wake-one mechanism for incoming connections: only
* one process gets woken up, not the 'whole herd'.
* Since we do not 'race & poll' for established sockets
* anymore, the common case will execute the loop only once.
*
* Subtle issue: "add_wait_queue_exclusive()" will be added
* after any current non-exclusive waiters, and we know that
* it will always _stay_ after any new non-exclusive waiters
* because all non-exclusive waiters are added at the
* beginning of the wait-queue. As such, it's ok to "drop"
* our exclusiveness temporarily when we get woken up without
* having to remove and re-insert us on the wait queue.
*/
for (;;) {
prepare_to_wait_exclusive(sk_sleep(sk), &wait,
TASK_INTERRUPTIBLE);
release_sock(sk);
if (reqsk_queue_empty(&icsk->icsk_accept_queue))
timeo = schedule_timeout(timeo);
lock_sock(sk);
err = 0;
if (!reqsk_queue_empty(&icsk->icsk_accept_queue))
break;
err = -EINVAL;
if (sk->sk_state != TCP_LISTEN)
break;
err = sock_intr_errno(timeo);
if (signal_pending(current))
break;
err = -EAGAIN;
if (!timeo)
break;
}
finish_wait(sk_sleep(sk), &wait);
return err;
}
/*
* This will accept the next outstanding connection.
*/
struct sock *inet_csk_accept(struct sock *sk, int flags, int *err)
{
struct inet_connection_sock *icsk = inet_csk(sk);
struct request_sock_queue *queue = &icsk->icsk_accept_queue;
struct sock *newsk;
struct request_sock *req;
int error;
lock_sock(sk);
/* We need to make sure that this socket is listening,
* and that it has something pending.
*/
error = -EINVAL;
if (sk->sk_state != TCP_LISTEN)
goto out_err;
/* Find already established connection */
if (reqsk_queue_empty(queue)) {
long timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
/* If this is a non blocking socket don't sleep */
error = -EAGAIN;
if (!timeo)
goto out_err;
error = inet_csk_wait_for_connect(sk, timeo);
if (error)
goto out_err;
}
req = reqsk_queue_remove(queue);
newsk = req->sk;
sk_acceptq_removed(sk);
tcp: fix TFO regression Fengguang Wu reported various panics and bisected to commit 8336886f786fdac (tcp: TCP Fast Open Server - support TFO listeners) Fix this by making sure socket is a TCP socket before accessing TFO data structures. [ 233.046014] kfree_debugcheck: out of range ptr ea6000000bb8h. [ 233.047399] ------------[ cut here ]------------ [ 233.048393] kernel BUG at /c/kernel-tests/src/stable/mm/slab.c:3074! [ 233.048393] invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC [ 233.048393] Modules linked in: [ 233.048393] CPU 0 [ 233.048393] Pid: 3929, comm: trinity-watchdo Not tainted 3.6.0-rc3+ #4192 Bochs Bochs [ 233.048393] RIP: 0010:[<ffffffff81169653>] [<ffffffff81169653>] kfree_debugcheck+0x27/0x2d [ 233.048393] RSP: 0018:ffff88000facbca8 EFLAGS: 00010092 [ 233.048393] RAX: 0000000000000031 RBX: 0000ea6000000bb8 RCX: 00000000a189a188 [ 233.048393] RDX: 000000000000a189 RSI: ffffffff8108ad32 RDI: ffffffff810d30f9 [ 233.048393] RBP: ffff88000facbcb8 R08: 0000000000000002 R09: ffffffff843846f0 [ 233.048393] R10: ffffffff810ae37c R11: 0000000000000908 R12: 0000000000000202 [ 233.048393] R13: ffffffff823dbd5a R14: ffff88000ec5bea8 R15: ffffffff8363c780 [ 233.048393] FS: 00007faa6899c700(0000) GS:ffff88001f200000(0000) knlGS:0000000000000000 [ 233.048393] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 233.048393] CR2: 00007faa6841019c CR3: 0000000012c82000 CR4: 00000000000006f0 [ 233.048393] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 233.048393] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 233.048393] Process trinity-watchdo (pid: 3929, threadinfo ffff88000faca000, task ffff88000faec600) [ 233.048393] Stack: [ 233.048393] 0000000000000000 0000ea6000000bb8 ffff88000facbce8 ffffffff8116ad81 [ 233.048393] ffff88000ff588a0 ffff88000ff58850 ffff88000ff588a0 0000000000000000 [ 233.048393] ffff88000facbd08 ffffffff823dbd5a ffffffff823dbcb0 ffff88000ff58850 [ 233.048393] Call Trace: [ 233.048393] [<ffffffff8116ad81>] kfree+0x5f/0xca [ 233.048393] [<ffffffff823dbd5a>] inet_sock_destruct+0xaa/0x13c [ 233.048393] [<ffffffff823dbcb0>] ? inet_sk_rebuild_header +0x319/0x319 [ 233.048393] [<ffffffff8231c307>] __sk_free+0x21/0x14b [ 233.048393] [<ffffffff8231c4bd>] sk_free+0x26/0x2a [ 233.048393] [<ffffffff825372db>] sctp_close+0x215/0x224 [ 233.048393] [<ffffffff810d6835>] ? lock_release+0x16f/0x1b9 [ 233.048393] [<ffffffff823daf12>] inet_release+0x7e/0x85 [ 233.048393] [<ffffffff82317d15>] sock_release+0x1f/0x77 [ 233.048393] [<ffffffff82317d94>] sock_close+0x27/0x2b [ 233.048393] [<ffffffff81173bbe>] __fput+0x101/0x20a [ 233.048393] [<ffffffff81173cd5>] ____fput+0xe/0x10 [ 233.048393] [<ffffffff810a3794>] task_work_run+0x5d/0x75 [ 233.048393] [<ffffffff8108da70>] do_exit+0x290/0x7f5 [ 233.048393] [<ffffffff82707415>] ? retint_swapgs+0x13/0x1b [ 233.048393] [<ffffffff8108e23f>] do_group_exit+0x7b/0xba [ 233.048393] [<ffffffff8108e295>] sys_exit_group+0x17/0x17 [ 233.048393] [<ffffffff8270de10>] tracesys+0xdd/0xe2 [ 233.048393] Code: 59 01 5d c3 55 48 89 e5 53 41 50 0f 1f 44 00 00 48 89 fb e8 d4 b0 f0 ff 84 c0 75 11 48 89 de 48 c7 c7 fc fa f7 82 e8 0d 0f 57 01 <0f> 0b 5f 5b 5d c3 55 48 89 e5 0f 1f 44 00 00 48 63 87 d8 00 00 [ 233.048393] RIP [<ffffffff81169653>] kfree_debugcheck+0x27/0x2d [ 233.048393] RSP <ffff88000facbca8> Reported-by: Fengguang Wu <wfg@linux.intel.com> Tested-by: Fengguang Wu <wfg@linux.intel.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: "H.K. Jerry Chu" <hkchu@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Acked-by: H.K. Jerry Chu <hkchu@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-09-06 08:07:13 +00:00
if (sk->sk_protocol == IPPROTO_TCP && queue->fastopenq != NULL) {
spin_lock_bh(&queue->fastopenq->lock);
if (tcp_rsk(req)->listener) {
/* We are still waiting for the final ACK from 3WHS
* so can't free req now. Instead, we set req->sk to
* NULL to signify that the child socket is taken
* so reqsk_fastopen_remove() will free the req
* when 3WHS finishes (or is aborted).
*/
req->sk = NULL;
req = NULL;
}
spin_unlock_bh(&queue->fastopenq->lock);
}
out:
release_sock(sk);
if (req)
__reqsk_free(req);
return newsk;
out_err:
newsk = NULL;
req = NULL;
*err = error;
goto out;
}
EXPORT_SYMBOL(inet_csk_accept);
/*
* Using different timers for retransmit, delayed acks and probes
* We may wish use just one timer maintaining a list of expire jiffies
* to optimize.
*/
void inet_csk_init_xmit_timers(struct sock *sk,
void (*retransmit_handler)(unsigned long),
void (*delack_handler)(unsigned long),
void (*keepalive_handler)(unsigned long))
{
struct inet_connection_sock *icsk = inet_csk(sk);
setup_timer(&icsk->icsk_retransmit_timer, retransmit_handler,
(unsigned long)sk);
setup_timer(&icsk->icsk_delack_timer, delack_handler,
(unsigned long)sk);
setup_timer(&sk->sk_timer, keepalive_handler, (unsigned long)sk);
icsk->icsk_pending = icsk->icsk_ack.pending = 0;
}
EXPORT_SYMBOL(inet_csk_init_xmit_timers);
void inet_csk_clear_xmit_timers(struct sock *sk)
{
struct inet_connection_sock *icsk = inet_csk(sk);
icsk->icsk_pending = icsk->icsk_ack.pending = icsk->icsk_ack.blocked = 0;
sk_stop_timer(sk, &icsk->icsk_retransmit_timer);
sk_stop_timer(sk, &icsk->icsk_delack_timer);
sk_stop_timer(sk, &sk->sk_timer);
}
EXPORT_SYMBOL(inet_csk_clear_xmit_timers);
void inet_csk_delete_keepalive_timer(struct sock *sk)
{
sk_stop_timer(sk, &sk->sk_timer);
}
EXPORT_SYMBOL(inet_csk_delete_keepalive_timer);
void inet_csk_reset_keepalive_timer(struct sock *sk, unsigned long len)
{
sk_reset_timer(sk, &sk->sk_timer, jiffies + len);
}
EXPORT_SYMBOL(inet_csk_reset_keepalive_timer);
struct dst_entry *inet_csk_route_req(struct sock *sk,
struct flowi4 *fl4,
const struct request_sock *req)
{
struct rtable *rt;
const struct inet_request_sock *ireq = inet_rsk(req);
struct ip_options_rcu *opt = inet_rsk(req)->opt;
struct net *net = sock_net(sk);
int flags = inet_sk_flowi_flags(sk);
flowi4_init_output(fl4, sk->sk_bound_dev_if, sk->sk_mark,
RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE,
tcp: do not create inetpeer on SYNACK message Another problem on SYNFLOOD/DDOS attack is the inetpeer cache getting larger and larger, using lots of memory and cpu time. tcp_v4_send_synack() ->inet_csk_route_req() ->ip_route_output_flow() ->rt_set_nexthop() ->rt_init_metrics() ->inet_getpeer( create = true) This is a side effect of commit a4daad6b09230 (net: Pre-COW metrics for TCP) added in 2.6.39 Possible solution : Instruct inet_csk_route_req() to remove FLOWI_FLAG_PRECOW_METRICS Before patch : # grep peer /proc/slabinfo inet_peer_cache 4175430 4175430 192 42 2 : tunables 0 0 0 : slabdata 99415 99415 0 Samples: 41K of event 'cycles', Event count (approx.): 30716565122 + 20,24% ksoftirqd/0 [kernel.kallsyms] [k] inet_getpeer + 8,19% ksoftirqd/0 [kernel.kallsyms] [k] peer_avl_rebalance.isra.1 + 4,81% ksoftirqd/0 [kernel.kallsyms] [k] sha_transform + 3,64% ksoftirqd/0 [kernel.kallsyms] [k] fib_table_lookup + 2,36% ksoftirqd/0 [ixgbe] [k] ixgbe_poll + 2,16% ksoftirqd/0 [kernel.kallsyms] [k] __ip_route_output_key + 2,11% ksoftirqd/0 [kernel.kallsyms] [k] kernel_map_pages + 2,11% ksoftirqd/0 [kernel.kallsyms] [k] ip_route_input_common + 2,01% ksoftirqd/0 [kernel.kallsyms] [k] __inet_lookup_established + 1,83% ksoftirqd/0 [kernel.kallsyms] [k] md5_transform + 1,75% ksoftirqd/0 [kernel.kallsyms] [k] check_leaf.isra.9 + 1,49% ksoftirqd/0 [kernel.kallsyms] [k] ipt_do_table + 1,46% ksoftirqd/0 [kernel.kallsyms] [k] hrtimer_interrupt + 1,45% ksoftirqd/0 [kernel.kallsyms] [k] kmem_cache_alloc + 1,29% ksoftirqd/0 [kernel.kallsyms] [k] inet_csk_search_req + 1,29% ksoftirqd/0 [kernel.kallsyms] [k] __netif_receive_skb + 1,16% ksoftirqd/0 [kernel.kallsyms] [k] copy_user_generic_string + 1,15% ksoftirqd/0 [kernel.kallsyms] [k] kmem_cache_free + 1,02% ksoftirqd/0 [kernel.kallsyms] [k] tcp_make_synack + 0,93% ksoftirqd/0 [kernel.kallsyms] [k] _raw_spin_lock_bh + 0,87% ksoftirqd/0 [kernel.kallsyms] [k] __call_rcu + 0,84% ksoftirqd/0 [kernel.kallsyms] [k] rt_garbage_collect + 0,84% ksoftirqd/0 [kernel.kallsyms] [k] fib_rules_lookup Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Hans Schillstrom <hans.schillstrom@ericsson.com> Cc: Jesper Dangaard Brouer <brouer@redhat.com> Cc: Neal Cardwell <ncardwell@google.com> Cc: Tom Herbert <therbert@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-05-31 21:00:26 +00:00
sk->sk_protocol,
flags,
(opt && opt->opt.srr) ? opt->opt.faddr : ireq->ir_rmt_addr,
ireq->ir_loc_addr, ireq->ir_rmt_port, inet_sk(sk)->inet_sport);
security_req_classify_flow(req, flowi4_to_flowi(fl4));
rt = ip_route_output_flow(net, fl4, sk);
if (IS_ERR(rt))
goto no_route;
if (opt && opt->opt.is_strictroute && rt->rt_uses_gateway)
goto route_err;
return &rt->dst;
route_err:
ip_rt_put(rt);
no_route:
IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES);
return NULL;
}
EXPORT_SYMBOL_GPL(inet_csk_route_req);
struct dst_entry *inet_csk_route_child_sock(struct sock *sk,
struct sock *newsk,
const struct request_sock *req)
{
const struct inet_request_sock *ireq = inet_rsk(req);
struct inet_sock *newinet = inet_sk(newsk);
struct ip_options_rcu *opt;
struct net *net = sock_net(sk);
struct flowi4 *fl4;
struct rtable *rt;
fl4 = &newinet->cork.fl.u.ip4;
rcu_read_lock();
opt = rcu_dereference(newinet->inet_opt);
flowi4_init_output(fl4, sk->sk_bound_dev_if, sk->sk_mark,
RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE,
sk->sk_protocol, inet_sk_flowi_flags(sk),
(opt && opt->opt.srr) ? opt->opt.faddr : ireq->ir_rmt_addr,
ireq->ir_loc_addr, ireq->ir_rmt_port, inet_sk(sk)->inet_sport);
security_req_classify_flow(req, flowi4_to_flowi(fl4));
rt = ip_route_output_flow(net, fl4, sk);
if (IS_ERR(rt))
goto no_route;
if (opt && opt->opt.is_strictroute && rt->rt_uses_gateway)
goto route_err;
rcu_read_unlock();
return &rt->dst;
route_err:
ip_rt_put(rt);
no_route:
rcu_read_unlock();
IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES);
return NULL;
}
EXPORT_SYMBOL_GPL(inet_csk_route_child_sock);
static inline u32 inet_synq_hash(const __be32 raddr, const __be16 rport,
const u32 rnd, const u32 synq_hsize)
{
return jhash_2words((__force u32)raddr, (__force u32)rport, rnd) & (synq_hsize - 1);
}
#if IS_ENABLED(CONFIG_IPV6)
#define AF_INET_FAMILY(fam) ((fam) == AF_INET)
#else
#define AF_INET_FAMILY(fam) 1
#endif
struct request_sock *inet_csk_search_req(const struct sock *sk,
struct request_sock ***prevp,
const __be16 rport, const __be32 raddr,
const __be32 laddr)
{
const struct inet_connection_sock *icsk = inet_csk(sk);
struct listen_sock *lopt = icsk->icsk_accept_queue.listen_opt;
struct request_sock *req, **prev;
for (prev = &lopt->syn_table[inet_synq_hash(raddr, rport, lopt->hash_rnd,
lopt->nr_table_entries)];
(req = *prev) != NULL;
prev = &req->dl_next) {
const struct inet_request_sock *ireq = inet_rsk(req);
if (ireq->ir_rmt_port == rport &&
ireq->ir_rmt_addr == raddr &&
ireq->ir_loc_addr == laddr &&
AF_INET_FAMILY(req->rsk_ops->family)) {
WARN_ON(req->sk);
*prevp = prev;
break;
}
}
return req;
}
EXPORT_SYMBOL_GPL(inet_csk_search_req);
void inet_csk_reqsk_queue_hash_add(struct sock *sk, struct request_sock *req,
unsigned long timeout)
{
struct inet_connection_sock *icsk = inet_csk(sk);
struct listen_sock *lopt = icsk->icsk_accept_queue.listen_opt;
const u32 h = inet_synq_hash(inet_rsk(req)->ir_rmt_addr,
inet_rsk(req)->ir_rmt_port,
lopt->hash_rnd, lopt->nr_table_entries);
reqsk_queue_hash_req(&icsk->icsk_accept_queue, h, req, timeout);
inet_csk_reqsk_queue_added(sk, timeout);
}
EXPORT_SYMBOL_GPL(inet_csk_reqsk_queue_hash_add);
/* Only thing we need from tcp.h */
extern int sysctl_tcp_synack_retries;
/* Decide when to expire the request and when to resend SYN-ACK */
static inline void syn_ack_recalc(struct request_sock *req, const int thresh,
const int max_retries,
const u8 rskq_defer_accept,
int *expire, int *resend)
{
if (!rskq_defer_accept) {
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
*expire = req->num_timeout >= thresh;
*resend = 1;
return;
}
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
*expire = req->num_timeout >= thresh &&
(!inet_rsk(req)->acked || req->num_timeout >= max_retries);
/*
* Do not resend while waiting for data after ACK,
* start to resend on end of deferring period to give
* last chance for data or ACK to create established socket.
*/
*resend = !inet_rsk(req)->acked ||
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
req->num_timeout >= rskq_defer_accept - 1;
}
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
int inet_rtx_syn_ack(struct sock *parent, struct request_sock *req)
{
int err = req->rsk_ops->rtx_syn_ack(parent, req);
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
if (!err)
req->num_retrans++;
return err;
}
EXPORT_SYMBOL(inet_rtx_syn_ack);
void inet_csk_reqsk_queue_prune(struct sock *parent,
const unsigned long interval,
const unsigned long timeout,
const unsigned long max_rto)
{
struct inet_connection_sock *icsk = inet_csk(parent);
struct request_sock_queue *queue = &icsk->icsk_accept_queue;
struct listen_sock *lopt = queue->listen_opt;
tcp: Revert 'process defer accept as established' changes. This reverts two changesets, ec3c0982a2dd1e671bad8e9d26c28dcba0039d87 ("[TCP]: TCP_DEFER_ACCEPT updates - process as established") and the follow-on bug fix 9ae27e0adbf471c7a6b80102e38e1d5a346b3b38 ("tcp: Fix slab corruption with ipv6 and tcp6fuzz"). This change causes several problems, first reported by Ingo Molnar as a distcc-over-loopback regression where connections were getting stuck. Ilpo Järvinen first spotted the locking problems. The new function added by this code, tcp_defer_accept_check(), only has the child socket locked, yet it is modifying state of the parent listening socket. Fixing that is non-trivial at best, because we can't simply just grab the parent listening socket lock at this point, because it would create an ABBA deadlock. The normal ordering is parent listening socket --> child socket, but this code path would require the reverse lock ordering. Next is a problem noticed by Vitaliy Gusev, he noted: ---------------------------------------- >--- a/net/ipv4/tcp_timer.c >+++ b/net/ipv4/tcp_timer.c >@@ -481,6 +481,11 @@ static void tcp_keepalive_timer (unsigned long data) > goto death; > } > >+ if (tp->defer_tcp_accept.request && sk->sk_state == TCP_ESTABLISHED) { >+ tcp_send_active_reset(sk, GFP_ATOMIC); >+ goto death; Here socket sk is not attached to listening socket's request queue. tcp_done() will not call inet_csk_destroy_sock() (and tcp_v4_destroy_sock() which should release this sk) as socket is not DEAD. Therefore socket sk will be lost for freeing. ---------------------------------------- Finally, Alexey Kuznetsov argues that there might not even be any real value or advantage to these new semantics even if we fix all of the bugs: ---------------------------------------- Hiding from accept() sockets with only out-of-order data only is the only thing which is impossible with old approach. Is this really so valuable? My opinion: no, this is nothing but a new loophole to consume memory without control. ---------------------------------------- So revert this thing for now. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-06-12 23:31:35 +00:00
int max_retries = icsk->icsk_syn_retries ? : sysctl_tcp_synack_retries;
int thresh = max_retries;
unsigned long now = jiffies;
struct request_sock **reqp, *req;
int i, budget;
if (lopt == NULL || lopt->qlen == 0)
return;
/* Normally all the openreqs are young and become mature
* (i.e. converted to established socket) for first timeout.
* If synack was not acknowledged for 1 second, it means
* one of the following things: synack was lost, ack was lost,
* rtt is high or nobody planned to ack (i.e. synflood).
* When server is a bit loaded, queue is populated with old
* open requests, reducing effective size of queue.
* When server is well loaded, queue size reduces to zero
* after several minutes of work. It is not synflood,
* it is normal operation. The solution is pruning
* too old entries overriding normal timeout, when
* situation becomes dangerous.
*
* Essentially, we reserve half of room for young
* embrions; and abort old ones without pity, if old
* ones are about to clog our table.
*/
if (lopt->qlen>>(lopt->max_qlen_log-1)) {
int young = (lopt->qlen_young<<1);
while (thresh > 2) {
if (lopt->qlen < young)
break;
thresh--;
young <<= 1;
}
}
tcp: Revert 'process defer accept as established' changes. This reverts two changesets, ec3c0982a2dd1e671bad8e9d26c28dcba0039d87 ("[TCP]: TCP_DEFER_ACCEPT updates - process as established") and the follow-on bug fix 9ae27e0adbf471c7a6b80102e38e1d5a346b3b38 ("tcp: Fix slab corruption with ipv6 and tcp6fuzz"). This change causes several problems, first reported by Ingo Molnar as a distcc-over-loopback regression where connections were getting stuck. Ilpo Järvinen first spotted the locking problems. The new function added by this code, tcp_defer_accept_check(), only has the child socket locked, yet it is modifying state of the parent listening socket. Fixing that is non-trivial at best, because we can't simply just grab the parent listening socket lock at this point, because it would create an ABBA deadlock. The normal ordering is parent listening socket --> child socket, but this code path would require the reverse lock ordering. Next is a problem noticed by Vitaliy Gusev, he noted: ---------------------------------------- >--- a/net/ipv4/tcp_timer.c >+++ b/net/ipv4/tcp_timer.c >@@ -481,6 +481,11 @@ static void tcp_keepalive_timer (unsigned long data) > goto death; > } > >+ if (tp->defer_tcp_accept.request && sk->sk_state == TCP_ESTABLISHED) { >+ tcp_send_active_reset(sk, GFP_ATOMIC); >+ goto death; Here socket sk is not attached to listening socket's request queue. tcp_done() will not call inet_csk_destroy_sock() (and tcp_v4_destroy_sock() which should release this sk) as socket is not DEAD. Therefore socket sk will be lost for freeing. ---------------------------------------- Finally, Alexey Kuznetsov argues that there might not even be any real value or advantage to these new semantics even if we fix all of the bugs: ---------------------------------------- Hiding from accept() sockets with only out-of-order data only is the only thing which is impossible with old approach. Is this really so valuable? My opinion: no, this is nothing but a new loophole to consume memory without control. ---------------------------------------- So revert this thing for now. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-06-12 23:31:35 +00:00
if (queue->rskq_defer_accept)
max_retries = queue->rskq_defer_accept;
budget = 2 * (lopt->nr_table_entries / (timeout / interval));
i = lopt->clock_hand;
do {
reqp=&lopt->syn_table[i];
while ((req = *reqp) != NULL) {
if (time_after_eq(now, req->expires)) {
int expire = 0, resend = 0;
syn_ack_recalc(req, thresh, max_retries,
queue->rskq_defer_accept,
&expire, &resend);
req->rsk_ops->syn_ack_timeout(parent, req);
if (!expire &&
(!resend ||
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
!inet_rtx_syn_ack(parent, req) ||
inet_rsk(req)->acked)) {
unsigned long timeo;
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
if (req->num_timeout++ == 0)
lopt->qlen_young--;
tcp: better retrans tracking for defer-accept For passive TCP connections using TCP_DEFER_ACCEPT facility, we incorrectly increment req->retrans each time timeout triggers while no SYNACK is sent. SYNACK are not sent for TCP_DEFER_ACCEPT that were established (for which we received the ACK from client). Only the last SYNACK is sent so that we can receive again an ACK from client, to move the req into accept queue. We plan to change this later to avoid the useless retransmit (and potential problem as this SYNACK could be lost) TCP_INFO later gives wrong information to user, claiming imaginary retransmits. Decouple req->retrans field into two independent fields : num_retrans : number of retransmit num_timeout : number of timeouts num_timeout is the counter that is incremented at each timeout, regardless of actual SYNACK being sent or not, and used to compute the exponential timeout. Introduce inet_rtx_syn_ack() helper to increment num_retrans only if ->rtx_syn_ack() succeeded. Use inet_rtx_syn_ack() from tcp_check_req() to increment num_retrans when we re-send a SYNACK in answer to a (retransmitted) SYN. Prior to this patch, we were not counting these retransmits. Change tcp_v[46]_rtx_synack() to increment TCP_MIB_RETRANSSEGS only if a synack packet was successfully queued. Reported-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Julian Anastasov <ja@ssi.bg> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Cc: Elliott Hughes <enh@google.com> Cc: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-27 23:16:46 +00:00
timeo = min(timeout << req->num_timeout,
max_rto);
req->expires = now + timeo;
reqp = &req->dl_next;
continue;
}
/* Drop this request */
inet_csk_reqsk_queue_unlink(parent, req, reqp);
reqsk_queue_removed(queue, req);
reqsk_free(req);
continue;
}
reqp = &req->dl_next;
}
i = (i + 1) & (lopt->nr_table_entries - 1);
} while (--budget > 0);
lopt->clock_hand = i;
if (lopt->qlen)
inet_csk_reset_keepalive_timer(parent, interval);
}
EXPORT_SYMBOL_GPL(inet_csk_reqsk_queue_prune);
/**
* inet_csk_clone_lock - clone an inet socket, and lock its clone
* @sk: the socket to clone
* @req: request_sock
* @priority: for allocation (%GFP_KERNEL, %GFP_ATOMIC, etc)
*
* Caller must unlock socket even in error path (bh_unlock_sock(newsk))
*/
struct sock *inet_csk_clone_lock(const struct sock *sk,
const struct request_sock *req,
const gfp_t priority)
{
struct sock *newsk = sk_clone_lock(sk, priority);
if (newsk != NULL) {
struct inet_connection_sock *newicsk = inet_csk(newsk);
newsk->sk_state = TCP_SYN_RECV;
newicsk->icsk_bind_hash = NULL;
inet_sk(newsk)->inet_dport = inet_rsk(req)->ir_rmt_port;
inet_sk(newsk)->inet_num = inet_rsk(req)->ir_num;
inet_sk(newsk)->inet_sport = htons(inet_rsk(req)->ir_num);
newsk->sk_write_space = sk_stream_write_space;
newicsk->icsk_retransmits = 0;
newicsk->icsk_backoff = 0;
newicsk->icsk_probes_out = 0;
/* Deinitialize accept_queue to trap illegal accesses. */
memset(&newicsk->icsk_accept_queue, 0, sizeof(newicsk->icsk_accept_queue));
security_inet_csk_clone(newsk, req);
}
return newsk;
}
EXPORT_SYMBOL_GPL(inet_csk_clone_lock);
/*
* At this point, there should be no process reference to this
* socket, and thus no user references at all. Therefore we
* can assume the socket waitqueue is inactive and nobody will
* try to jump onto it.
*/
void inet_csk_destroy_sock(struct sock *sk)
{
WARN_ON(sk->sk_state != TCP_CLOSE);
WARN_ON(!sock_flag(sk, SOCK_DEAD));
/* It cannot be in hash table! */
WARN_ON(!sk_unhashed(sk));
/* If it has not 0 inet_sk(sk)->inet_num, it must be bound */
WARN_ON(inet_sk(sk)->inet_num && !inet_csk(sk)->icsk_bind_hash);
sk->sk_prot->destroy(sk);
sk_stream_kill_queues(sk);
xfrm_sk_free_policy(sk);
sk_refcnt_debug_release(sk);
percpu_counter_dec(sk->sk_prot->orphan_count);
sock_put(sk);
}
EXPORT_SYMBOL(inet_csk_destroy_sock);
inet: Fix kmemleak in tcp_v4/6_syn_recv_sock and dccp_v4/6_request_recv_sock If in either of the above functions inet_csk_route_child_sock() or __inet_inherit_port() fails, the newsk will not be freed: unreferenced object 0xffff88022e8a92c0 (size 1592): comm "softirq", pid 0, jiffies 4294946244 (age 726.160s) hex dump (first 32 bytes): 0a 01 01 01 0a 01 01 02 00 00 00 00 a7 cc 16 00 ................ 02 00 03 01 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<ffffffff8153d190>] kmemleak_alloc+0x21/0x3e [<ffffffff810ab3e7>] kmem_cache_alloc+0xb5/0xc5 [<ffffffff8149b65b>] sk_prot_alloc.isra.53+0x2b/0xcd [<ffffffff8149b784>] sk_clone_lock+0x16/0x21e [<ffffffff814d711a>] inet_csk_clone_lock+0x10/0x7b [<ffffffff814ebbc3>] tcp_create_openreq_child+0x21/0x481 [<ffffffff814e8fa5>] tcp_v4_syn_recv_sock+0x3a/0x23b [<ffffffff814ec5ba>] tcp_check_req+0x29f/0x416 [<ffffffff814e8e10>] tcp_v4_do_rcv+0x161/0x2bc [<ffffffff814eb917>] tcp_v4_rcv+0x6c9/0x701 [<ffffffff814cea9f>] ip_local_deliver_finish+0x70/0xc4 [<ffffffff814cec20>] ip_local_deliver+0x4e/0x7f [<ffffffff814ce9f8>] ip_rcv_finish+0x1fc/0x233 [<ffffffff814cee68>] ip_rcv+0x217/0x267 [<ffffffff814a7bbe>] __netif_receive_skb+0x49e/0x553 [<ffffffff814a7cc3>] netif_receive_skb+0x50/0x82 This happens, because sk_clone_lock initializes sk_refcnt to 2, and thus a single sock_put() is not enough to free the memory. Additionally, things like xfrm, memcg, cookie_values,... may have been initialized. We have to free them properly. This is fixed by forcing a call to tcp_done(), ending up in inet_csk_destroy_sock, doing the final sock_put(). tcp_done() is necessary, because it ends up doing all the cleanup on xfrm, memcg, cookie_values, xfrm,... Before calling tcp_done, we have to set the socket to SOCK_DEAD, to force it entering inet_csk_destroy_sock. To avoid the warning in inet_csk_destroy_sock, inet_num has to be set to 0. As inet_csk_destroy_sock does a dec on orphan_count, we first have to increase it. Calling tcp_done() allows us to remove the calls to tcp_clear_xmit_timer() and tcp_cleanup_congestion_control(). A similar approach is taken for dccp by calling dccp_done(). This is in the kernel since 093d282321 (tproxy: fix hash locking issue when using port redirection in __inet_inherit_port()), thus since version >= 2.6.37. Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-14 04:07:58 +00:00
/* This function allows to force a closure of a socket after the call to
* tcp/dccp_create_openreq_child().
*/
void inet_csk_prepare_forced_close(struct sock *sk)
__releases(&sk->sk_lock.slock)
inet: Fix kmemleak in tcp_v4/6_syn_recv_sock and dccp_v4/6_request_recv_sock If in either of the above functions inet_csk_route_child_sock() or __inet_inherit_port() fails, the newsk will not be freed: unreferenced object 0xffff88022e8a92c0 (size 1592): comm "softirq", pid 0, jiffies 4294946244 (age 726.160s) hex dump (first 32 bytes): 0a 01 01 01 0a 01 01 02 00 00 00 00 a7 cc 16 00 ................ 02 00 03 01 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<ffffffff8153d190>] kmemleak_alloc+0x21/0x3e [<ffffffff810ab3e7>] kmem_cache_alloc+0xb5/0xc5 [<ffffffff8149b65b>] sk_prot_alloc.isra.53+0x2b/0xcd [<ffffffff8149b784>] sk_clone_lock+0x16/0x21e [<ffffffff814d711a>] inet_csk_clone_lock+0x10/0x7b [<ffffffff814ebbc3>] tcp_create_openreq_child+0x21/0x481 [<ffffffff814e8fa5>] tcp_v4_syn_recv_sock+0x3a/0x23b [<ffffffff814ec5ba>] tcp_check_req+0x29f/0x416 [<ffffffff814e8e10>] tcp_v4_do_rcv+0x161/0x2bc [<ffffffff814eb917>] tcp_v4_rcv+0x6c9/0x701 [<ffffffff814cea9f>] ip_local_deliver_finish+0x70/0xc4 [<ffffffff814cec20>] ip_local_deliver+0x4e/0x7f [<ffffffff814ce9f8>] ip_rcv_finish+0x1fc/0x233 [<ffffffff814cee68>] ip_rcv+0x217/0x267 [<ffffffff814a7bbe>] __netif_receive_skb+0x49e/0x553 [<ffffffff814a7cc3>] netif_receive_skb+0x50/0x82 This happens, because sk_clone_lock initializes sk_refcnt to 2, and thus a single sock_put() is not enough to free the memory. Additionally, things like xfrm, memcg, cookie_values,... may have been initialized. We have to free them properly. This is fixed by forcing a call to tcp_done(), ending up in inet_csk_destroy_sock, doing the final sock_put(). tcp_done() is necessary, because it ends up doing all the cleanup on xfrm, memcg, cookie_values, xfrm,... Before calling tcp_done, we have to set the socket to SOCK_DEAD, to force it entering inet_csk_destroy_sock. To avoid the warning in inet_csk_destroy_sock, inet_num has to be set to 0. As inet_csk_destroy_sock does a dec on orphan_count, we first have to increase it. Calling tcp_done() allows us to remove the calls to tcp_clear_xmit_timer() and tcp_cleanup_congestion_control(). A similar approach is taken for dccp by calling dccp_done(). This is in the kernel since 093d282321 (tproxy: fix hash locking issue when using port redirection in __inet_inherit_port()), thus since version >= 2.6.37. Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-14 04:07:58 +00:00
{
/* sk_clone_lock locked the socket and set refcnt to 2 */
bh_unlock_sock(sk);
sock_put(sk);
/* The below has to be done to allow calling inet_csk_destroy_sock */
sock_set_flag(sk, SOCK_DEAD);
percpu_counter_inc(sk->sk_prot->orphan_count);
inet_sk(sk)->inet_num = 0;
}
EXPORT_SYMBOL(inet_csk_prepare_forced_close);
int inet_csk_listen_start(struct sock *sk, const int nr_table_entries)
{
struct inet_sock *inet = inet_sk(sk);
struct inet_connection_sock *icsk = inet_csk(sk);
int rc = reqsk_queue_alloc(&icsk->icsk_accept_queue, nr_table_entries);
if (rc != 0)
return rc;
sk->sk_max_ack_backlog = 0;
sk->sk_ack_backlog = 0;
inet_csk_delack_init(sk);
/* There is race window here: we announce ourselves listening,
* but this transition is still not validated by get_port().
* It is OK, because this socket enters to hash table only
* after validation is complete.
*/
sk->sk_state = TCP_LISTEN;
if (!sk->sk_prot->get_port(sk, inet->inet_num)) {
inet->inet_sport = htons(inet->inet_num);
sk_dst_reset(sk);
sk->sk_prot->hash(sk);
return 0;
}
sk->sk_state = TCP_CLOSE;
__reqsk_queue_destroy(&icsk->icsk_accept_queue);
return -EADDRINUSE;
}
EXPORT_SYMBOL_GPL(inet_csk_listen_start);
/*
* This routine closes sockets which have been at least partially
* opened, but not yet accepted.
*/
void inet_csk_listen_stop(struct sock *sk)
{
struct inet_connection_sock *icsk = inet_csk(sk);
struct request_sock_queue *queue = &icsk->icsk_accept_queue;
struct request_sock *acc_req;
struct request_sock *req;
inet_csk_delete_keepalive_timer(sk);
/* make all the listen_opt local to us */
acc_req = reqsk_queue_yank_acceptq(queue);
/* Following specs, it would be better either to send FIN
* (and enter FIN-WAIT-1, it is normal close)
* or to send active reset (abort).
* Certainly, it is pretty dangerous while synflood, but it is
* bad justification for our negligence 8)
* To be honest, we are not able to make either
* of the variants now. --ANK
*/
reqsk_queue_destroy(queue);
while ((req = acc_req) != NULL) {
struct sock *child = req->sk;
acc_req = req->dl_next;
local_bh_disable();
bh_lock_sock(child);
WARN_ON(sock_owned_by_user(child));
sock_hold(child);
sk->sk_prot->disconnect(child, O_NONBLOCK);
sock_orphan(child);
percpu_counter_inc(sk->sk_prot->orphan_count);
tcp: fix TFO regression Fengguang Wu reported various panics and bisected to commit 8336886f786fdac (tcp: TCP Fast Open Server - support TFO listeners) Fix this by making sure socket is a TCP socket before accessing TFO data structures. [ 233.046014] kfree_debugcheck: out of range ptr ea6000000bb8h. [ 233.047399] ------------[ cut here ]------------ [ 233.048393] kernel BUG at /c/kernel-tests/src/stable/mm/slab.c:3074! [ 233.048393] invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC [ 233.048393] Modules linked in: [ 233.048393] CPU 0 [ 233.048393] Pid: 3929, comm: trinity-watchdo Not tainted 3.6.0-rc3+ #4192 Bochs Bochs [ 233.048393] RIP: 0010:[<ffffffff81169653>] [<ffffffff81169653>] kfree_debugcheck+0x27/0x2d [ 233.048393] RSP: 0018:ffff88000facbca8 EFLAGS: 00010092 [ 233.048393] RAX: 0000000000000031 RBX: 0000ea6000000bb8 RCX: 00000000a189a188 [ 233.048393] RDX: 000000000000a189 RSI: ffffffff8108ad32 RDI: ffffffff810d30f9 [ 233.048393] RBP: ffff88000facbcb8 R08: 0000000000000002 R09: ffffffff843846f0 [ 233.048393] R10: ffffffff810ae37c R11: 0000000000000908 R12: 0000000000000202 [ 233.048393] R13: ffffffff823dbd5a R14: ffff88000ec5bea8 R15: ffffffff8363c780 [ 233.048393] FS: 00007faa6899c700(0000) GS:ffff88001f200000(0000) knlGS:0000000000000000 [ 233.048393] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 233.048393] CR2: 00007faa6841019c CR3: 0000000012c82000 CR4: 00000000000006f0 [ 233.048393] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 233.048393] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 233.048393] Process trinity-watchdo (pid: 3929, threadinfo ffff88000faca000, task ffff88000faec600) [ 233.048393] Stack: [ 233.048393] 0000000000000000 0000ea6000000bb8 ffff88000facbce8 ffffffff8116ad81 [ 233.048393] ffff88000ff588a0 ffff88000ff58850 ffff88000ff588a0 0000000000000000 [ 233.048393] ffff88000facbd08 ffffffff823dbd5a ffffffff823dbcb0 ffff88000ff58850 [ 233.048393] Call Trace: [ 233.048393] [<ffffffff8116ad81>] kfree+0x5f/0xca [ 233.048393] [<ffffffff823dbd5a>] inet_sock_destruct+0xaa/0x13c [ 233.048393] [<ffffffff823dbcb0>] ? inet_sk_rebuild_header +0x319/0x319 [ 233.048393] [<ffffffff8231c307>] __sk_free+0x21/0x14b [ 233.048393] [<ffffffff8231c4bd>] sk_free+0x26/0x2a [ 233.048393] [<ffffffff825372db>] sctp_close+0x215/0x224 [ 233.048393] [<ffffffff810d6835>] ? lock_release+0x16f/0x1b9 [ 233.048393] [<ffffffff823daf12>] inet_release+0x7e/0x85 [ 233.048393] [<ffffffff82317d15>] sock_release+0x1f/0x77 [ 233.048393] [<ffffffff82317d94>] sock_close+0x27/0x2b [ 233.048393] [<ffffffff81173bbe>] __fput+0x101/0x20a [ 233.048393] [<ffffffff81173cd5>] ____fput+0xe/0x10 [ 233.048393] [<ffffffff810a3794>] task_work_run+0x5d/0x75 [ 233.048393] [<ffffffff8108da70>] do_exit+0x290/0x7f5 [ 233.048393] [<ffffffff82707415>] ? retint_swapgs+0x13/0x1b [ 233.048393] [<ffffffff8108e23f>] do_group_exit+0x7b/0xba [ 233.048393] [<ffffffff8108e295>] sys_exit_group+0x17/0x17 [ 233.048393] [<ffffffff8270de10>] tracesys+0xdd/0xe2 [ 233.048393] Code: 59 01 5d c3 55 48 89 e5 53 41 50 0f 1f 44 00 00 48 89 fb e8 d4 b0 f0 ff 84 c0 75 11 48 89 de 48 c7 c7 fc fa f7 82 e8 0d 0f 57 01 <0f> 0b 5f 5b 5d c3 55 48 89 e5 0f 1f 44 00 00 48 63 87 d8 00 00 [ 233.048393] RIP [<ffffffff81169653>] kfree_debugcheck+0x27/0x2d [ 233.048393] RSP <ffff88000facbca8> Reported-by: Fengguang Wu <wfg@linux.intel.com> Tested-by: Fengguang Wu <wfg@linux.intel.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: "H.K. Jerry Chu" <hkchu@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Acked-by: H.K. Jerry Chu <hkchu@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-09-06 08:07:13 +00:00
if (sk->sk_protocol == IPPROTO_TCP && tcp_rsk(req)->listener) {
BUG_ON(tcp_sk(child)->fastopen_rsk != req);
BUG_ON(sk != tcp_rsk(req)->listener);
/* Paranoid, to prevent race condition if
* an inbound pkt destined for child is
* blocked by sock lock in tcp_v4_rcv().
* Also to satisfy an assertion in
* tcp_v4_destroy_sock().
*/
tcp_sk(child)->fastopen_rsk = NULL;
sock_put(sk);
}
inet_csk_destroy_sock(child);
bh_unlock_sock(child);
local_bh_enable();
sock_put(child);
sk_acceptq_removed(sk);
__reqsk_free(req);
}
if (queue->fastopenq != NULL) {
/* Free all the reqs queued in rskq_rst_head. */
spin_lock_bh(&queue->fastopenq->lock);
acc_req = queue->fastopenq->rskq_rst_head;
queue->fastopenq->rskq_rst_head = NULL;
spin_unlock_bh(&queue->fastopenq->lock);
while ((req = acc_req) != NULL) {
acc_req = req->dl_next;
__reqsk_free(req);
}
}
WARN_ON(sk->sk_ack_backlog);
}
EXPORT_SYMBOL_GPL(inet_csk_listen_stop);
void inet_csk_addr2sockaddr(struct sock *sk, struct sockaddr *uaddr)
{
struct sockaddr_in *sin = (struct sockaddr_in *)uaddr;
const struct inet_sock *inet = inet_sk(sk);
sin->sin_family = AF_INET;
sin->sin_addr.s_addr = inet->inet_daddr;
sin->sin_port = inet->inet_dport;
}
EXPORT_SYMBOL_GPL(inet_csk_addr2sockaddr);
#ifdef CONFIG_COMPAT
int inet_csk_compat_getsockopt(struct sock *sk, int level, int optname,
char __user *optval, int __user *optlen)
{
const struct inet_connection_sock *icsk = inet_csk(sk);
if (icsk->icsk_af_ops->compat_getsockopt != NULL)
return icsk->icsk_af_ops->compat_getsockopt(sk, level, optname,
optval, optlen);
return icsk->icsk_af_ops->getsockopt(sk, level, optname,
optval, optlen);
}
EXPORT_SYMBOL_GPL(inet_csk_compat_getsockopt);
int inet_csk_compat_setsockopt(struct sock *sk, int level, int optname,
char __user *optval, unsigned int optlen)
{
const struct inet_connection_sock *icsk = inet_csk(sk);
if (icsk->icsk_af_ops->compat_setsockopt != NULL)
return icsk->icsk_af_ops->compat_setsockopt(sk, level, optname,
optval, optlen);
return icsk->icsk_af_ops->setsockopt(sk, level, optname,
optval, optlen);
}
EXPORT_SYMBOL_GPL(inet_csk_compat_setsockopt);
#endif
static struct dst_entry *inet_csk_rebuild_route(struct sock *sk, struct flowi *fl)
{
const struct inet_sock *inet = inet_sk(sk);
const struct ip_options_rcu *inet_opt;
__be32 daddr = inet->inet_daddr;
struct flowi4 *fl4;
struct rtable *rt;
rcu_read_lock();
inet_opt = rcu_dereference(inet->inet_opt);
if (inet_opt && inet_opt->opt.srr)
daddr = inet_opt->opt.faddr;
fl4 = &fl->u.ip4;
rt = ip_route_output_ports(sock_net(sk), fl4, sk, daddr,
inet->inet_saddr, inet->inet_dport,
inet->inet_sport, sk->sk_protocol,
RT_CONN_FLAGS(sk), sk->sk_bound_dev_if);
if (IS_ERR(rt))
rt = NULL;
if (rt)
sk_setup_caps(sk, &rt->dst);
rcu_read_unlock();
return &rt->dst;
}
struct dst_entry *inet_csk_update_pmtu(struct sock *sk, u32 mtu)
{
struct dst_entry *dst = __sk_dst_check(sk, 0);
struct inet_sock *inet = inet_sk(sk);
if (!dst) {
dst = inet_csk_rebuild_route(sk, &inet->cork.fl);
if (!dst)
goto out;
}
dst->ops->update_pmtu(dst, sk, NULL, mtu);
dst = __sk_dst_check(sk, 0);
if (!dst)
dst = inet_csk_rebuild_route(sk, &inet->cork.fl);
out:
return dst;
}
EXPORT_SYMBOL_GPL(inet_csk_update_pmtu);