2006-11-28 03:10:57 +08:00
|
|
|
/*
|
|
|
|
* UDPLITE An implementation of the UDP-Lite protocol (RFC 3828).
|
|
|
|
*
|
|
|
|
* Authors: Gerrit Renker <gerrit@erg.abdn.ac.uk>
|
|
|
|
*
|
|
|
|
* Changes:
|
|
|
|
* Fixes:
|
|
|
|
* 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.
|
|
|
|
*/
|
2011-07-15 23:47:34 +08:00
|
|
|
#include <linux/export.h>
|
2006-11-28 03:10:57 +08:00
|
|
|
#include "udp_impl.h"
|
|
|
|
|
2009-10-07 08:37:59 +08:00
|
|
|
struct udp_table udplite_table __read_mostly;
|
2008-10-29 16:41:45 +08:00
|
|
|
EXPORT_SYMBOL(udplite_table);
|
2006-11-28 03:10:57 +08:00
|
|
|
|
2006-12-01 09:22:29 +08:00
|
|
|
static int udplite_rcv(struct sk_buff *skb)
|
2006-11-28 03:10:57 +08:00
|
|
|
{
|
2008-10-29 16:41:45 +08:00
|
|
|
return __udp4_lib_rcv(skb, &udplite_table, IPPROTO_UDPLITE);
|
2006-11-28 03:10:57 +08:00
|
|
|
}
|
|
|
|
|
2006-12-01 09:22:29 +08:00
|
|
|
static void udplite_err(struct sk_buff *skb, u32 info)
|
2006-11-28 03:10:57 +08:00
|
|
|
{
|
2008-10-29 16:41:45 +08:00
|
|
|
__udp4_lib_err(skb, info, &udplite_table);
|
2006-11-28 03:10:57 +08:00
|
|
|
}
|
|
|
|
|
2009-09-14 20:21:47 +08:00
|
|
|
static const struct net_protocol udplite_protocol = {
|
2006-11-28 03:10:57 +08:00
|
|
|
.handler = udplite_rcv,
|
|
|
|
.err_handler = udplite_err,
|
|
|
|
.no_policy = 1,
|
2008-03-25 06:34:06 +08:00
|
|
|
.netns_ok = 1,
|
2006-11-28 03:10:57 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
struct proto udplite_prot = {
|
|
|
|
.name = "UDP-Lite",
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.close = udp_lib_close,
|
|
|
|
.connect = ip4_datagram_connect,
|
|
|
|
.disconnect = udp_disconnect,
|
|
|
|
.ioctl = udp_ioctl,
|
|
|
|
.init = udplite_sk_init,
|
|
|
|
.destroy = udp_destroy_sock,
|
|
|
|
.setsockopt = udp_setsockopt,
|
|
|
|
.getsockopt = udp_getsockopt,
|
|
|
|
.sendmsg = udp_sendmsg,
|
|
|
|
.recvmsg = udp_recvmsg,
|
|
|
|
.sendpage = udp_sendpage,
|
|
|
|
.backlog_rcv = udp_queue_rcv_skb,
|
|
|
|
.hash = udp_lib_hash,
|
|
|
|
.unhash = udp_lib_unhash,
|
2008-03-23 07:51:21 +08:00
|
|
|
.get_port = udp_v4_get_port,
|
2006-11-28 03:10:57 +08:00
|
|
|
.obj_size = sizeof(struct udp_sock),
|
udp: RCU handling for Unicast packets.
Goals are :
1) Optimizing handling of incoming Unicast UDP frames, so that no memory
writes should happen in the fast path.
Note: Multicasts and broadcasts still will need to take a lock,
because doing a full lockless lookup in this case is difficult.
2) No expensive operations in the socket bind/unhash phases :
- No expensive synchronize_rcu() calls.
- No added rcu_head in socket structure, increasing memory needs,
but more important, forcing us to use call_rcu() calls,
that have the bad property of making sockets structure cold.
(rcu grace period between socket freeing and its potential reuse
make this socket being cold in CPU cache).
David did a previous patch using call_rcu() and noticed a 20%
impact on TCP connection rates.
Quoting Cristopher Lameter :
"Right. That results in cacheline cooldown. You'd want to recycle
the object as they are cache hot on a per cpu basis. That is screwed
up by the delayed regular rcu processing. We have seen multiple
regressions due to cacheline cooldown.
The only choice in cacheline hot sensitive areas is to deal with the
complexity that comes with SLAB_DESTROY_BY_RCU or give up on RCU."
- Because udp sockets are allocated from dedicated kmem_cache,
use of SLAB_DESTROY_BY_RCU can help here.
Theory of operation :
---------------------
As the lookup is lockfree (using rcu_read_lock()/rcu_read_unlock()),
special attention must be taken by readers and writers.
Use of SLAB_DESTROY_BY_RCU is tricky too, because a socket can be freed,
reused, inserted in a different chain or in worst case in the same chain
while readers could do lookups in the same time.
In order to avoid loops, a reader must check each socket found in a chain
really belongs to the chain the reader was traversing. If it finds a
mismatch, lookup must start again at the begining. This *restart* loop
is the reason we had to use rdlock for the multicast case, because
we dont want to send same message several times to the same socket.
We use RCU only for fast path.
Thus, /proc/net/udp still takes spinlocks.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-29 17:11:14 +08:00
|
|
|
.slab_flags = SLAB_DESTROY_BY_RCU,
|
2008-10-29 16:41:45 +08:00
|
|
|
.h.udp_table = &udplite_table,
|
2006-11-28 03:10:57 +08:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
.compat_setsockopt = compat_udp_setsockopt,
|
|
|
|
.compat_getsockopt = compat_udp_getsockopt,
|
|
|
|
#endif
|
2010-12-17 06:26:56 +08:00
|
|
|
.clear_sk = sk_prot_clear_portaddr_nulls,
|
2006-11-28 03:10:57 +08:00
|
|
|
};
|
2010-07-10 05:22:10 +08:00
|
|
|
EXPORT_SYMBOL(udplite_prot);
|
2006-11-28 03:10:57 +08:00
|
|
|
|
|
|
|
static struct inet_protosw udplite4_protosw = {
|
|
|
|
.type = SOCK_DGRAM,
|
|
|
|
.protocol = IPPROTO_UDPLITE,
|
|
|
|
.prot = &udplite_prot,
|
|
|
|
.ops = &inet_dgram_ops,
|
|
|
|
.no_check = 0, /* must checksum (RFC 3828) */
|
|
|
|
.flags = INET_PROTOSW_PERMANENT,
|
|
|
|
};
|
|
|
|
|
|
|
|
#ifdef CONFIG_PROC_FS
|
2011-10-30 14:46:30 +08:00
|
|
|
|
|
|
|
static const struct file_operations udplite_afinfo_seq_fops = {
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.open = udp_seq_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.release = seq_release_net
|
|
|
|
};
|
|
|
|
|
2006-11-28 03:10:57 +08:00
|
|
|
static struct udp_seq_afinfo udplite4_seq_afinfo = {
|
|
|
|
.name = "udplite",
|
|
|
|
.family = AF_INET,
|
2008-10-29 16:41:45 +08:00
|
|
|
.udp_table = &udplite_table,
|
2011-10-30 14:46:30 +08:00
|
|
|
.seq_fops = &udplite_afinfo_seq_fops,
|
2008-03-29 09:24:26 +08:00
|
|
|
.seq_ops = {
|
|
|
|
.show = udp4_seq_show,
|
|
|
|
},
|
2006-11-28 03:10:57 +08:00
|
|
|
};
|
2008-03-25 05:56:34 +08:00
|
|
|
|
2010-01-17 11:35:32 +08:00
|
|
|
static int __net_init udplite4_proc_init_net(struct net *net)
|
2008-03-25 05:56:57 +08:00
|
|
|
{
|
|
|
|
return udp_proc_register(net, &udplite4_seq_afinfo);
|
|
|
|
}
|
|
|
|
|
2010-01-17 11:35:32 +08:00
|
|
|
static void __net_exit udplite4_proc_exit_net(struct net *net)
|
2008-03-25 05:56:57 +08:00
|
|
|
{
|
|
|
|
udp_proc_unregister(net, &udplite4_seq_afinfo);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct pernet_operations udplite4_net_ops = {
|
|
|
|
.init = udplite4_proc_init_net,
|
|
|
|
.exit = udplite4_proc_exit_net,
|
|
|
|
};
|
|
|
|
|
2008-03-25 05:56:34 +08:00
|
|
|
static __init int udplite4_proc_init(void)
|
|
|
|
{
|
2008-03-25 05:56:57 +08:00
|
|
|
return register_pernet_subsys(&udplite4_net_ops);
|
2008-03-25 05:56:34 +08:00
|
|
|
}
|
|
|
|
#else
|
|
|
|
static inline int udplite4_proc_init(void)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
2006-11-28 03:10:57 +08:00
|
|
|
#endif
|
|
|
|
|
|
|
|
void __init udplite4_register(void)
|
|
|
|
{
|
2009-10-07 08:37:59 +08:00
|
|
|
udp_table_init(&udplite_table, "UDP-Lite");
|
2006-11-28 03:10:57 +08:00
|
|
|
if (proto_register(&udplite_prot, 1))
|
|
|
|
goto out_register_err;
|
|
|
|
|
|
|
|
if (inet_add_protocol(&udplite_protocol, IPPROTO_UDPLITE) < 0)
|
|
|
|
goto out_unregister_proto;
|
|
|
|
|
|
|
|
inet_register_protosw(&udplite4_protosw);
|
|
|
|
|
2008-03-25 05:56:34 +08:00
|
|
|
if (udplite4_proc_init())
|
2012-03-12 02:36:11 +08:00
|
|
|
pr_err("%s: Cannot register /proc!\n", __func__);
|
2006-11-28 03:10:57 +08:00
|
|
|
return;
|
|
|
|
|
|
|
|
out_unregister_proto:
|
|
|
|
proto_unregister(&udplite_prot);
|
|
|
|
out_register_err:
|
2012-03-12 02:36:11 +08:00
|
|
|
pr_crit("%s: Cannot add UDP-Lite protocol\n", __func__);
|
2006-11-28 03:10:57 +08:00
|
|
|
}
|