From 0eae750e6019a93643063924209c1daf9cb9b4a7 Mon Sep 17 00:00:00 2001 From: John Dykstra Date: Mon, 19 Oct 2009 21:53:53 -0700 Subject: IP: Cleanups Use symbols instead of magic constants while checking PMTU discovery setsockopt. Remove redundant test in ip_rt_frag_needed() (done by caller). Signed-off-by: John Dykstra Signed-off-by: David S. Miller --- net/ipv4/route.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'net/ipv4/route.c') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index bb4199252026..68fb22702051 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -1628,9 +1628,6 @@ unsigned short ip_rt_frag_needed(struct net *net, struct iphdr *iph, __be32 daddr = iph->daddr; unsigned short est_mtu = 0; - if (ipv4_config.no_pmtu_disc) - return 0; - for (k = 0; k < 2; k++) { for (i = 0; i < 2; i++) { unsigned hash = rt_hash(daddr, skeys[i], ikeys[k], -- cgit v1.2.1 From 2c1409a0a2b88585ec0c03f1de0aafa178c56313 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 12 Nov 2009 09:33:09 +0000 Subject: inetpeer: Optimize inet_getid() While investigating for network latencies, I found inet_getid() was a contention point for some workloads, as inet_peer_idlock is shared by all inet_getid() users regardless of peers. One way to fix this is to make ip_id_count an atomic_t instead of __u16, and use atomic_add_return(). In order to keep sizeof(struct inet_peer) = 64 on 64bit arches tcp_ts_stamp is also converted to __u32 instead of "unsigned long". Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/route.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv4/route.c') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index ff258b57680b..4284ceef7945 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -2852,7 +2852,7 @@ static int rt_fill_info(struct net *net, error = rt->u.dst.error; expires = rt->u.dst.expires ? rt->u.dst.expires - jiffies : 0; if (rt->peer) { - id = rt->peer->ip_id_count; + id = atomic_read(&rt->peer->ip_id_count) & 0xffff; if (rt->peer->tcp_ts_stamp) { ts = rt->peer->tcp_ts; tsage = get_seconds() - rt->peer->tcp_ts_stamp; -- cgit v1.2.1 From 9d4fb27db90043cd2640e4bc778f9c755d3c17c1 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Mon, 23 Nov 2009 10:41:23 -0800 Subject: net/ipv4: Move && and || to end of previous line On Sun, 2009-11-22 at 16:31 -0800, David Miller wrote: > It should be of the form: > if (x && > y) > > or: > if (x && y) > > Fix patches, rather than complaints, for existing cases where things > do not follow this pattern are certainly welcome. Also collapsed some multiple tabs to single space. Signed-off-by: Joe Perches Signed-off-by: David S. Miller --- net/ipv4/route.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) (limited to 'net/ipv4/route.c') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 4284ceef7945..7547944ea9bf 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -1346,9 +1346,9 @@ void ip_rt_redirect(__be32 old_gw, __be32 daddr, __be32 new_gw, return; net = dev_net(dev); - if (new_gw == old_gw || !IN_DEV_RX_REDIRECTS(in_dev) - || ipv4_is_multicast(new_gw) || ipv4_is_lbcast(new_gw) - || ipv4_is_zeronet(new_gw)) + if (new_gw == old_gw || !IN_DEV_RX_REDIRECTS(in_dev) || + ipv4_is_multicast(new_gw) || ipv4_is_lbcast(new_gw) || + ipv4_is_zeronet(new_gw)) goto reject_redirect; if (!rt_caching(net)) @@ -2311,10 +2311,11 @@ skip_cache: ip_hdr(skb)->protocol); if (our #ifdef CONFIG_IP_MROUTE - || (!ipv4_is_local_multicast(daddr) && - IN_DEV_MFORWARD(in_dev)) + || + (!ipv4_is_local_multicast(daddr) && + IN_DEV_MFORWARD(in_dev)) #endif - ) { + ) { rcu_read_unlock(); return ip_route_input_mc(skb, daddr, saddr, tos, dev, our); @@ -2511,9 +2512,9 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp, of another iface. --ANK */ - if (oldflp->oif == 0 - && (ipv4_is_multicast(oldflp->fl4_dst) || - oldflp->fl4_dst == htonl(0xFFFFFFFF))) { + if (oldflp->oif == 0 && + (ipv4_is_multicast(oldflp->fl4_dst) || + oldflp->fl4_dst == htonl(0xFFFFFFFF))) { /* It is equivalent to inet_addr_type(saddr) == RTN_LOCAL */ dev_out = ip_dev_find(net, oldflp->fl4_src); if (dev_out == NULL) -- cgit v1.2.1 From 09ad9bc752519cc167d0a573e1acf69b5c707c67 Mon Sep 17 00:00:00 2001 From: Octavian Purdila Date: Wed, 25 Nov 2009 15:14:13 -0800 Subject: net: use net_eq to compare nets Generated with the following semantic patch @@ struct net *n1; struct net *n2; @@ - n1 == n2 + net_eq(n1, n2) @@ struct net *n1; struct net *n2; @@ - n1 != n2 + !net_eq(n1, n2) applied over {include,net,drivers/net}. Signed-off-by: Octavian Purdila Signed-off-by: David S. Miller --- net/ipv4/route.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/ipv4/route.c') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 7547944ea9bf..aea7bb369cfa 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -703,7 +703,7 @@ static inline int compare_keys(struct flowi *fl1, struct flowi *fl2) static inline int compare_netns(struct rtable *rt1, struct rtable *rt2) { - return dev_net(rt1->u.dst.dev) == dev_net(rt2->u.dst.dev); + return net_eq(dev_net(rt1->u.dst.dev), dev_net(rt2->u.dst.dev)); } static inline int rt_is_expired(struct rtable *rth) @@ -3310,7 +3310,7 @@ static __net_init int sysctl_route_net_init(struct net *net) struct ctl_table *tbl; tbl = ipv4_route_flush_table; - if (net != &init_net) { + if (!net_eq(net, &init_net)) { tbl = kmemdup(tbl, sizeof(ipv4_route_flush_table), GFP_KERNEL); if (tbl == NULL) goto err_dup; -- cgit v1.2.1 From a661c4199b300503dae694efa21f1354f763b97c Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Wed, 25 Nov 2009 15:40:35 -0800 Subject: net: convert /proc/net/rt_acct to seq_file Rewrite statistics accumulation to be in terms of structure fields, not raw u32 additions. Keep them in same order, though. This is the last user of create_proc_read_entry() in net/, please NAK all new ones as well as all new ->write_proc, ->read_proc and create_proc_entry() users. Cc me if there are problems. :-) Signed-off-by: Alexey Dobriyan Signed-off-by: David S. Miller --- net/ipv4/route.c | 68 +++++++++++++++++++++++++++----------------------------- 1 file changed, 33 insertions(+), 35 deletions(-) (limited to 'net/ipv4/route.c') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index aea7bb369cfa..9889fbd96487 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -513,43 +513,42 @@ static const struct file_operations rt_cpu_seq_fops = { }; #ifdef CONFIG_NET_CLS_ROUTE -static int ip_rt_acct_read(char *buffer, char **start, off_t offset, - int length, int *eof, void *data) -{ - unsigned int i; - - if ((offset & 3) || (length & 3)) - return -EIO; - - if (offset >= sizeof(struct ip_rt_acct) * 256) { - *eof = 1; - return 0; - } - - if (offset + length >= sizeof(struct ip_rt_acct) * 256) { - length = sizeof(struct ip_rt_acct) * 256 - offset; - *eof = 1; +static int rt_acct_proc_show(struct seq_file *m, void *v) +{ + struct ip_rt_acct *dst, *src; + unsigned int i, j; + + dst = kcalloc(256, sizeof(struct ip_rt_acct), GFP_KERNEL); + if (!dst) + return -ENOMEM; + + for_each_possible_cpu(i) { + src = (struct ip_rt_acct *)per_cpu_ptr(ip_rt_acct, i); + for (j = 0; j < 256; j++) { + dst[j].o_bytes += src[j].o_bytes; + dst[j].o_packets += src[j].o_packets; + dst[j].i_bytes += src[j].i_bytes; + dst[j].i_packets += src[j].i_packets; + } } - offset /= sizeof(u32); - - if (length > 0) { - u32 *dst = (u32 *) buffer; - - *start = buffer; - memset(dst, 0, length); - - for_each_possible_cpu(i) { - unsigned int j; - u32 *src; + seq_write(m, dst, 256 * sizeof(struct ip_rt_acct)); + kfree(dst); + return 0; +} - src = ((u32 *) per_cpu_ptr(ip_rt_acct, i)) + offset; - for (j = 0; j < length/4; j++) - dst[j] += src[j]; - } - } - return length; +static int rt_acct_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, rt_acct_proc_show, NULL); } + +static const struct file_operations rt_acct_proc_fops = { + .owner = THIS_MODULE, + .open = rt_acct_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; #endif static int __net_init ip_rt_do_proc_init(struct net *net) @@ -567,8 +566,7 @@ static int __net_init ip_rt_do_proc_init(struct net *net) goto err2; #ifdef CONFIG_NET_CLS_ROUTE - pde = create_proc_read_entry("rt_acct", 0, net->proc_net, - ip_rt_acct_read, NULL); + pde = proc_create("rt_acct", 0, net->proc_net, &rt_acct_proc_fops); if (!pde) goto err3; #endif -- cgit v1.2.1 From a5ee155136b4a8f4ab0e4c9c064b661da475e298 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Sun, 29 Nov 2009 15:45:58 +0000 Subject: net: NETDEV_UNREGISTER_PERNET -> NETDEV_UNREGISTER_BATCH The motivation for an additional notifier in batched netdevice notification (rt_do_flush) only needs to be called once per batch not once per namespace. For further batching improvements I need a guarantee that the netdevices are unregistered in order allowing me to unregister an all of the network devices in a network namespace at the same time with the guarantee that the loopback device is really and truly unregistered last. Additionally it appears that we moved the route cache flush after the final synchronize_net, which seems wrong and there was no explanation. So I have restored the original location of the final synchronize_net. Cc: Octavian Purdila Signed-off-by: Eric W. Biederman Signed-off-by: David S. Miller --- net/ipv4/route.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'net/ipv4/route.c') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 9889fbd96487..90cdcfc32937 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -900,6 +900,12 @@ void rt_cache_flush(struct net *net, int delay) rt_do_flush(!in_softirq()); } +/* Flush previous cache invalidated entries from the cache */ +void rt_cache_flush_batch(void) +{ + rt_do_flush(!in_softirq()); +} + /* * We change rt_genid and let gc do the cleanup */ -- cgit v1.2.1