summaryrefslogtreecommitdiff
path: root/net/ipv4/route.c
diff options
context:
space:
mode:
authorEric Dumazet <edumazet@google.com>2014-06-02 05:26:03 -0700
committerDavid S. Miller <davem@davemloft.net>2014-06-02 11:00:41 -0700
commit73f156a6e8c1074ac6327e0abd1169e95eb66463 (patch)
tree2c8b222f21784e738c397ba95dee70a8f256ea64 /net/ipv4/route.c
parente067ee336a9d3f038ffa9699c59f2abec3376bf7 (diff)
inetpeer: get rid of ip_id_count
Ideally, we would need to generate IP ID using a per destination IP generator. linux kernels used inet_peer cache for this purpose, but this had a huge cost on servers disabling MTU discovery. 1) each inet_peer struct consumes 192 bytes 2) inetpeer cache uses a binary tree of inet_peer structs, with a nominal size of ~66000 elements under load. 3) lookups in this tree are hitting a lot of cache lines, as tree depth is about 20. 4) If server deals with many tcp flows, we have a high probability of not finding the inet_peer, allocating a fresh one, inserting it in the tree with same initial ip_id_count, (cf secure_ip_id()) 5) We garbage collect inet_peer aggressively. IP ID generation do not have to be 'perfect' Goal is trying to avoid duplicates in a short period of time, so that reassembly units have a chance to complete reassembly of fragments belonging to one message before receiving other fragments with a recycled ID. We simply use an array of generators, and a Jenkin hash using the dst IP as a key. ipv6_select_ident() is put back into net/ipv6/ip6_output.c where it belongs (it is only used from this file) secure_ip_id() and secure_ipv6_id() no longer are needed. Rename ip_select_ident_more() to ip_select_ident_segs() to avoid unnecessary decrement/increment of the number of segments. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv4/route.c')
-rw-r--r--net/ipv4/route.c45
1 files changed, 16 insertions, 29 deletions
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 4154eb76b0ad..082239ffe34a 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -89,6 +89,7 @@
#include <linux/rcupdate.h>
#include <linux/times.h>
#include <linux/slab.h>
+#include <linux/jhash.h>
#include <net/dst.h>
#include <net/net_namespace.h>
#include <net/protocol.h>
@@ -456,39 +457,19 @@ static struct neighbour *ipv4_neigh_lookup(const struct dst_entry *dst,
return neigh_create(&arp_tbl, pkey, dev);
}
-/*
- * Peer allocation may fail only in serious out-of-memory conditions. However
- * we still can generate some output.
- * Random ID selection looks a bit dangerous because we have no chances to
- * select ID being unique in a reasonable period of time.
- * But broken packet identifier may be better than no packet at all.
- */
-static void ip_select_fb_ident(struct iphdr *iph)
-{
- static DEFINE_SPINLOCK(ip_fb_id_lock);
- static u32 ip_fallback_id;
- u32 salt;
+atomic_t *ip_idents __read_mostly;
+EXPORT_SYMBOL(ip_idents);
- spin_lock_bh(&ip_fb_id_lock);
- salt = secure_ip_id((__force __be32)ip_fallback_id ^ iph->daddr);
- iph->id = htons(salt & 0xFFFF);
- ip_fallback_id = salt;
- spin_unlock_bh(&ip_fb_id_lock);
-}
-
-void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more)
+void __ip_select_ident(struct iphdr *iph, int segs)
{
- struct net *net = dev_net(dst->dev);
- struct inet_peer *peer;
+ static u32 ip_idents_hashrnd __read_mostly;
+ u32 hash, id;
- peer = inet_getpeer_v4(net->ipv4.peers, iph->daddr, 1);
- if (peer) {
- iph->id = htons(inet_getid(peer, more));
- inet_putpeer(peer);
- return;
- }
+ net_get_random_once(&ip_idents_hashrnd, sizeof(ip_idents_hashrnd));
- ip_select_fb_ident(iph);
+ hash = jhash_1word((__force u32)iph->daddr, ip_idents_hashrnd);
+ id = ip_idents_reserve(hash, segs);
+ iph->id = htons(id);
}
EXPORT_SYMBOL(__ip_select_ident);
@@ -2711,6 +2692,12 @@ int __init ip_rt_init(void)
{
int rc = 0;
+ ip_idents = kmalloc(IP_IDENTS_SZ * sizeof(*ip_idents), GFP_KERNEL);
+ if (!ip_idents)
+ panic("IP: failed to allocate ip_idents\n");
+
+ prandom_bytes(ip_idents, IP_IDENTS_SZ * sizeof(*ip_idents));
+
#ifdef CONFIG_IP_ROUTE_CLASSID
ip_rt_acct = __alloc_percpu(256 * sizeof(struct ip_rt_acct), __alignof__(struct ip_rt_acct));
if (!ip_rt_acct)