From d5a4502e9efa534212484fd691339f6469cf95ff Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 19 May 2008 13:49:52 -0700 Subject: netns: Register net/core/ sysctls at read-only root. Most of the net/core/xxx sysctls are read-only now, but this goal is achieved with excessive memory consumption in each namespace - the whole table is cloned and most of the entries in it are ~= 0222. Split it into two parts and register (the largest) one at the read-only root. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/core/sysctl_net_core.c | 39 +++++++++++++++++++-------------------- 1 file changed, 19 insertions(+), 20 deletions(-) (limited to 'net/core') diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c index 5fc801057244..a570e2af22cb 100644 --- a/net/core/sysctl_net_core.c +++ b/net/core/sysctl_net_core.c @@ -124,14 +124,6 @@ static struct ctl_table net_core_table[] = { }, #endif /* CONFIG_XFRM */ #endif /* CONFIG_NET */ - { - .ctl_name = NET_CORE_SOMAXCONN, - .procname = "somaxconn", - .data = &init_net.core.sysctl_somaxconn, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = &proc_dointvec - }, { .ctl_name = NET_CORE_BUDGET, .procname = "netdev_budget", @@ -151,6 +143,18 @@ static struct ctl_table net_core_table[] = { { .ctl_name = 0 } }; +static struct ctl_table netns_core_table[] = { + { + .ctl_name = NET_CORE_SOMAXCONN, + .procname = "somaxconn", + .data = &init_net.core.sysctl_somaxconn, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec + }, + { .ctl_name = 0 } +}; + static __net_initdata struct ctl_path net_core_path[] = { { .procname = "net", .ctl_name = CTL_NET, }, { .procname = "core", .ctl_name = NET_CORE, }, @@ -159,23 +163,17 @@ static __net_initdata struct ctl_path net_core_path[] = { static __net_init int sysctl_core_net_init(struct net *net) { - struct ctl_table *tbl, *tmp; + struct ctl_table *tbl; net->core.sysctl_somaxconn = SOMAXCONN; - tbl = net_core_table; + tbl = netns_core_table; if (net != &init_net) { - tbl = kmemdup(tbl, sizeof(net_core_table), GFP_KERNEL); + tbl = kmemdup(tbl, sizeof(netns_core_table), GFP_KERNEL); if (tbl == NULL) goto err_dup; - for (tmp = tbl; tmp->procname; tmp++) { - if (tmp->data >= (void *)&init_net && - tmp->data < (void *)(&init_net + 1)) - tmp->data += (char *)net - (char *)&init_net; - else - tmp->mode &= ~0222; - } + tbl[0].data = &net->core.sysctl_somaxconn; } net->core.sysctl_hdr = register_net_sysctl_table(net, @@ -186,7 +184,7 @@ static __net_init int sysctl_core_net_init(struct net *net) return 0; err_reg: - if (tbl != net_core_table) + if (tbl != netns_core_table) kfree(tbl); err_dup: return -ENOMEM; @@ -198,7 +196,7 @@ static __net_exit void sysctl_core_net_exit(struct net *net) tbl = net->core.sysctl_hdr->ctl_table_arg; unregister_net_sysctl_table(net->core.sysctl_hdr); - BUG_ON(tbl == net_core_table); + BUG_ON(tbl == netns_core_table); kfree(tbl); } @@ -209,6 +207,7 @@ static __net_initdata struct pernet_operations sysctl_core_ops = { static __init int sysctl_core_init(void) { + register_net_sysctl_rotable(net_core_path, net_core_table); return register_pernet_subsys(&sysctl_core_ops); } -- cgit v1.3-8-gc7d7 From 96e74088f1da4d9a53735a4a57a4f984f86b75c6 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:12:46 -0700 Subject: net: The dev->get_stats pointer is not NULL nowadays. And so does the pointer is returns, but sysfs and netlinks still check for both cases. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/core/net-sysfs.c | 9 ++++----- net/core/rtnetlink.c | 20 ++++++++------------ 2 files changed, 12 insertions(+), 17 deletions(-) (limited to 'net/core') diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 90e2177af081..dccd737ea2e3 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -242,11 +242,11 @@ static ssize_t netstat_show(const struct device *d, offset % sizeof(unsigned long) != 0); read_lock(&dev_base_lock); - if (dev_isalive(dev) && dev->get_stats && - (stats = (*dev->get_stats)(dev))) + if (dev_isalive(dev)) { + stats = dev->get_stats(dev); ret = sprintf(buf, fmt_ulong, *(unsigned long *)(((u8 *) stats) + offset)); - + } read_unlock(&dev_base_lock); return ret; } @@ -457,8 +457,7 @@ int netdev_register_kobject(struct net_device *net) strlcpy(dev->bus_id, net->name, BUS_ID_SIZE); #ifdef CONFIG_SYSFS - if (net->get_stats) - *groups++ = &netstat_group; + *groups++ = &netstat_group; #ifdef CONFIG_WIRELESS_EXT if (net->wireless_handlers && net->wireless_handlers->get_wireless_stats) diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index cf857c4dc7b1..ca32ddb8ad17 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -606,6 +606,8 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, { struct ifinfomsg *ifm; struct nlmsghdr *nlh; + struct net_device_stats *stats; + struct nlattr *attr; nlh = nlmsg_put(skb, pid, seq, type, sizeof(*ifm), flags); if (nlh == NULL) @@ -652,19 +654,13 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, NLA_PUT(skb, IFLA_BROADCAST, dev->addr_len, dev->broadcast); } - if (dev->get_stats) { - struct net_device_stats *stats = dev->get_stats(dev); - if (stats) { - struct nlattr *attr; + attr = nla_reserve(skb, IFLA_STATS, + sizeof(struct rtnl_link_stats)); + if (attr == NULL) + goto nla_put_failure; - attr = nla_reserve(skb, IFLA_STATS, - sizeof(struct rtnl_link_stats)); - if (attr == NULL) - goto nla_put_failure; - - copy_rtnl_link_stats(nla_data(attr), stats); - } - } + stats = dev->get_stats(dev); + copy_rtnl_link_stats(nla_data(attr), stats); if (dev->rtnl_link_ops) { if (rtnl_link_fill(skb, dev) < 0) -- cgit v1.3-8-gc7d7 From 0e12f848b337fc034ceb3c0d03d75f8de1b8cc96 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Mon, 12 May 2008 21:21:13 +0200 Subject: net: use performance variant for_each_cpu_mask_nr Change references from for_each_cpu_mask to for_each_cpu_mask_nr where appropriate Reviewed-by: Paul Jackson Reviewed-by: Christoph Lameter Signed-off-by: Mike Travis Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- net/core/dev.c | 4 ++-- net/iucv/iucv.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'net/core') diff --git a/net/core/dev.c b/net/core/dev.c index 582963077877..ee61b987f4d9 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2239,7 +2239,7 @@ out: */ if (!cpus_empty(net_dma.channel_mask)) { int chan_idx; - for_each_cpu_mask(chan_idx, net_dma.channel_mask) { + for_each_cpu_mask_nr(chan_idx, net_dma.channel_mask) { struct dma_chan *chan = net_dma.channels[chan_idx]; if (chan) dma_async_memcpy_issue_pending(chan); @@ -4300,7 +4300,7 @@ static void net_dma_rebalance(struct net_dma *net_dma) i = 0; cpu = first_cpu(cpu_online_map); - for_each_cpu_mask(chan_idx, net_dma->channel_mask) { + for_each_cpu_mask_nr(chan_idx, net_dma->channel_mask) { chan = net_dma->channels[chan_idx]; n = ((num_online_cpus() / cpus_weight(net_dma->channel_mask)) diff --git a/net/iucv/iucv.c b/net/iucv/iucv.c index 918970762131..8de511070593 100644 --- a/net/iucv/iucv.c +++ b/net/iucv/iucv.c @@ -497,7 +497,7 @@ static void iucv_setmask_up(void) /* Disable all cpu but the first in cpu_irq_cpumask. */ cpumask = iucv_irq_cpumask; cpu_clear(first_cpu(iucv_irq_cpumask), cpumask); - for_each_cpu_mask(cpu, cpumask) + for_each_cpu_mask_nr(cpu, cpumask) smp_call_function_single(cpu, iucv_block_cpu, NULL, 0, 1); } -- cgit v1.3-8-gc7d7 From 962cf36c5bf6d2840b8d66ee9a606fae2f540bbd Mon Sep 17 00:00:00 2001 From: "Carlos R. Mafra" Date: Thu, 15 May 2008 11:15:37 -0300 Subject: Remove argument from open_softirq which is always NULL As git-grep shows, open_softirq() is always called with the last argument being NULL block/blk-core.c: open_softirq(BLOCK_SOFTIRQ, blk_done_softirq, NULL); kernel/hrtimer.c: open_softirq(HRTIMER_SOFTIRQ, run_hrtimer_softirq, NULL); kernel/rcuclassic.c: open_softirq(RCU_SOFTIRQ, rcu_process_callbacks, NULL); kernel/rcupreempt.c: open_softirq(RCU_SOFTIRQ, rcu_process_callbacks, NULL); kernel/sched.c: open_softirq(SCHED_SOFTIRQ, run_rebalance_domains, NULL); kernel/softirq.c: open_softirq(TASKLET_SOFTIRQ, tasklet_action, NULL); kernel/softirq.c: open_softirq(HI_SOFTIRQ, tasklet_hi_action, NULL); kernel/timer.c: open_softirq(TIMER_SOFTIRQ, run_timer_softirq, NULL); net/core/dev.c: open_softirq(NET_TX_SOFTIRQ, net_tx_action, NULL); net/core/dev.c: open_softirq(NET_RX_SOFTIRQ, net_rx_action, NULL); This observation has already been made by Matthew Wilcox in June 2002 (http://www.cs.helsinki.fi/linux/linux-kernel/2002-25/0687.html) "I notice that none of the current softirq routines use the data element passed to them." and the situation hasn't changed since them. So it appears we can safely remove that extra argument to save 128 (54) bytes of kernel data (text). Signed-off-by: Carlos R. Mafra Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- block/blk-core.c | 2 +- include/linux/interrupt.h | 3 +-- kernel/hrtimer.c | 2 +- kernel/rcuclassic.c | 2 +- kernel/rcupreempt.c | 2 +- kernel/sched.c | 2 +- kernel/softirq.c | 7 +++---- kernel/timer.c | 2 +- net/core/dev.c | 4 ++-- 9 files changed, 12 insertions(+), 14 deletions(-) (limited to 'net/core') diff --git a/block/blk-core.c b/block/blk-core.c index 6a9cc0d22a61..75fdc65136e8 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -2048,7 +2048,7 @@ int __init blk_dev_init(void) for_each_possible_cpu(i) INIT_LIST_HEAD(&per_cpu(blk_cpu_done, i)); - open_softirq(BLOCK_SOFTIRQ, blk_done_softirq, NULL); + open_softirq(BLOCK_SOFTIRQ, blk_done_softirq); register_hotcpu_notifier(&blk_cpu_notifier); return 0; diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index f1fc7470d26c..a86186dd0474 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -285,12 +285,11 @@ enum struct softirq_action { void (*action)(struct softirq_action *); - void *data; }; asmlinkage void do_softirq(void); asmlinkage void __do_softirq(void); -extern void open_softirq(int nr, void (*action)(struct softirq_action*), void *data); +extern void open_softirq(int nr, void (*action)(struct softirq_action *)); extern void softirq_init(void); #define __raise_softirq_irqoff(nr) do { or_softirq_pending(1UL << (nr)); } while (0) extern void raise_softirq_irqoff(unsigned int nr); diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index 421be5fe5cc7..861b4088092a 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -1669,7 +1669,7 @@ void __init hrtimers_init(void) (void *)(long)smp_processor_id()); register_cpu_notifier(&hrtimers_nb); #ifdef CONFIG_HIGH_RES_TIMERS - open_softirq(HRTIMER_SOFTIRQ, run_hrtimer_softirq, NULL); + open_softirq(HRTIMER_SOFTIRQ, run_hrtimer_softirq); #endif } diff --git a/kernel/rcuclassic.c b/kernel/rcuclassic.c index f4ffbd0f306f..f6e01f3ae9c6 100644 --- a/kernel/rcuclassic.c +++ b/kernel/rcuclassic.c @@ -529,7 +529,7 @@ static void __cpuinit rcu_online_cpu(int cpu) rcu_init_percpu_data(cpu, &rcu_ctrlblk, rdp); rcu_init_percpu_data(cpu, &rcu_bh_ctrlblk, bh_rdp); - open_softirq(RCU_SOFTIRQ, rcu_process_callbacks, NULL); + open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); } static int __cpuinit rcu_cpu_notify(struct notifier_block *self, diff --git a/kernel/rcupreempt.c b/kernel/rcupreempt.c index e1cdf196a515..9dd827db359f 100644 --- a/kernel/rcupreempt.c +++ b/kernel/rcupreempt.c @@ -1125,7 +1125,7 @@ void __init __rcu_init(void) for_each_online_cpu(cpu) rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE, (void *)(long) cpu); - open_softirq(RCU_SOFTIRQ, rcu_process_callbacks, NULL); + open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); } /* diff --git a/kernel/sched.c b/kernel/sched.c index cfa222a91539..56ea3a203a5a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -8154,7 +8154,7 @@ void __init sched_init(void) #endif #ifdef CONFIG_SMP - open_softirq(SCHED_SOFTIRQ, run_rebalance_domains, NULL); + open_softirq(SCHED_SOFTIRQ, run_rebalance_domains); #endif #ifdef CONFIG_RT_MUTEXES diff --git a/kernel/softirq.c b/kernel/softirq.c index 36e061740047..059256874e9b 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -347,9 +347,8 @@ void raise_softirq(unsigned int nr) local_irq_restore(flags); } -void open_softirq(int nr, void (*action)(struct softirq_action*), void *data) +void open_softirq(int nr, void (*action)(struct softirq_action *)) { - softirq_vec[nr].data = data; softirq_vec[nr].action = action; } @@ -503,8 +502,8 @@ void __init softirq_init(void) &per_cpu(tasklet_hi_vec, cpu).head; } - open_softirq(TASKLET_SOFTIRQ, tasklet_action, NULL); - open_softirq(HI_SOFTIRQ, tasklet_hi_action, NULL); + open_softirq(TASKLET_SOFTIRQ, tasklet_action); + open_softirq(HI_SOFTIRQ, tasklet_hi_action); } static int ksoftirqd(void * __bind_cpu) diff --git a/kernel/timer.c b/kernel/timer.c index ceacc6626572..b4da888497fa 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -1502,7 +1502,7 @@ void __init init_timers(void) BUG_ON(err == NOTIFY_BAD); register_cpu_notifier(&timers_nb); - open_softirq(TIMER_SOFTIRQ, run_timer_softirq, NULL); + open_softirq(TIMER_SOFTIRQ, run_timer_softirq); } /** diff --git a/net/core/dev.c b/net/core/dev.c index 582963077877..cf0e16731dc7 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4563,8 +4563,8 @@ static int __init net_dev_init(void) dev_boot_phase = 0; - open_softirq(NET_TX_SOFTIRQ, net_tx_action, NULL); - open_softirq(NET_RX_SOFTIRQ, net_rx_action, NULL); + open_softirq(NET_TX_SOFTIRQ, net_tx_action); + open_softirq(NET_RX_SOFTIRQ, net_rx_action); hotcpu_notifier(dev_cpu_callback, 0); dst_init(); -- cgit v1.3-8-gc7d7 From 0b040829952d84bf2a62526f0e24b624e0699447 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Tue, 10 Jun 2008 22:46:50 -0700 Subject: net: remove CVS keywords This patch removes CVS keywords that weren't updated for a long time from comments. Signed-off-by: Adrian Bunk Signed-off-by: David S. Miller --- include/linux/if_bridge.h | 2 -- include/linux/if_ppp.h | 2 -- include/linux/if_tun.h | 2 -- include/linux/ip6_tunnel.h | 4 ---- include/linux/ppp-comp.h | 2 -- include/linux/ppp_defs.h | 2 -- include/linux/sunrpc/auth_gss.h | 2 -- include/linux/sunrpc/gss_api.h | 2 -- include/linux/sunrpc/svcauth_gss.h | 3 --- include/net/inetpeer.h | 2 -- include/net/ip6_tunnel.h | 4 ---- include/net/ipconfig.h | 2 -- include/net/ipv6.h | 2 -- include/net/snmp.h | 2 -- net/bluetooth/bnep/bnep.h | 4 ---- net/bluetooth/bnep/core.c | 4 ---- net/bluetooth/bnep/netdev.c | 4 ---- net/bluetooth/bnep/sock.c | 4 ---- net/bluetooth/rfcomm/core.c | 2 -- net/bluetooth/rfcomm/sock.c | 2 -- net/bluetooth/rfcomm/tty.c | 2 -- net/bridge/br.c | 2 -- net/bridge/br_device.c | 2 -- net/bridge/br_fdb.c | 2 -- net/bridge/br_forward.c | 2 -- net/bridge/br_if.c | 2 -- net/bridge/br_input.c | 2 -- net/bridge/br_ioctl.c | 2 -- net/bridge/br_notify.c | 2 -- net/bridge/br_private.h | 2 -- net/bridge/br_private_stp.h | 2 -- net/bridge/br_stp.c | 2 -- net/bridge/br_stp_bpdu.c | 2 -- net/bridge/br_stp_if.c | 2 -- net/bridge/br_stp_timer.c | 2 -- net/core/skbuff.c | 2 -- net/core/sock.c | 2 -- net/ipv4/af_inet.c | 2 -- net/ipv4/arp.c | 2 -- net/ipv4/devinet.c | 2 -- net/ipv4/fib_frontend.c | 2 -- net/ipv4/fib_hash.c | 2 -- net/ipv4/fib_semantics.c | 2 -- net/ipv4/fib_trie.c | 2 -- net/ipv4/icmp.c | 2 -- net/ipv4/igmp.c | 2 -- net/ipv4/inet_diag.c | 2 -- net/ipv4/inetpeer.c | 2 -- net/ipv4/ip_forward.c | 2 -- net/ipv4/ip_fragment.c | 2 -- net/ipv4/ip_input.c | 2 -- net/ipv4/ip_options.c | 2 -- net/ipv4/ip_output.c | 2 -- net/ipv4/ip_sockglue.c | 2 -- net/ipv4/ipconfig.c | 2 -- net/ipv4/ipip.c | 2 -- net/ipv4/ipmr.c | 2 -- net/ipv4/ipvs/ip_vs_app.c | 2 -- net/ipv4/ipvs/ip_vs_conn.c | 2 -- net/ipv4/ipvs/ip_vs_core.c | 2 -- net/ipv4/ipvs/ip_vs_ctl.c | 2 -- net/ipv4/ipvs/ip_vs_dh.c | 2 -- net/ipv4/ipvs/ip_vs_est.c | 2 -- net/ipv4/ipvs/ip_vs_ftp.c | 2 -- net/ipv4/ipvs/ip_vs_lblc.c | 2 -- net/ipv4/ipvs/ip_vs_lblcr.c | 2 -- net/ipv4/ipvs/ip_vs_lc.c | 2 -- net/ipv4/ipvs/ip_vs_nq.c | 2 -- net/ipv4/ipvs/ip_vs_proto.c | 2 -- net/ipv4/ipvs/ip_vs_proto_ah.c | 2 -- net/ipv4/ipvs/ip_vs_proto_esp.c | 2 -- net/ipv4/ipvs/ip_vs_proto_tcp.c | 2 -- net/ipv4/ipvs/ip_vs_proto_udp.c | 2 -- net/ipv4/ipvs/ip_vs_rr.c | 2 -- net/ipv4/ipvs/ip_vs_sched.c | 2 -- net/ipv4/ipvs/ip_vs_sed.c | 2 -- net/ipv4/ipvs/ip_vs_sh.c | 2 -- net/ipv4/ipvs/ip_vs_sync.c | 2 -- net/ipv4/ipvs/ip_vs_wlc.c | 2 -- net/ipv4/ipvs/ip_vs_wrr.c | 2 -- net/ipv4/ipvs/ip_vs_xmit.c | 2 -- net/ipv4/proc.c | 2 -- net/ipv4/protocol.c | 2 -- net/ipv4/raw.c | 2 -- net/ipv4/route.c | 2 -- net/ipv4/syncookies.c | 2 -- net/ipv4/sysctl_net_ipv4.c | 2 -- net/ipv4/tcp.c | 2 -- net/ipv4/tcp_diag.c | 2 -- net/ipv4/tcp_input.c | 2 -- net/ipv4/tcp_ipv4.c | 2 -- net/ipv4/tcp_minisocks.c | 2 -- net/ipv4/tcp_output.c | 2 -- net/ipv4/tcp_timer.c | 2 -- net/ipv4/udp.c | 2 -- net/ipv4/udplite.c | 2 -- net/ipv6/addrconf.c | 2 -- net/ipv6/af_inet6.c | 2 -- net/ipv6/datagram.c | 2 -- net/ipv6/exthdrs.c | 2 -- net/ipv6/icmp.c | 2 -- net/ipv6/ip6_fib.c | 2 -- net/ipv6/ip6_input.c | 2 -- net/ipv6/ip6_output.c | 2 -- net/ipv6/ip6_tunnel.c | 2 -- net/ipv6/ipv6_sockglue.c | 2 -- net/ipv6/mcast.c | 2 -- net/ipv6/proc.c | 2 -- net/ipv6/protocol.c | 2 -- net/ipv6/raw.c | 2 -- net/ipv6/reassembly.c | 2 -- net/ipv6/route.c | 2 -- net/ipv6/sit.c | 2 -- net/ipv6/tcp_ipv6.c | 2 -- net/ipv6/udp.c | 2 -- net/ipv6/udplite.c | 2 -- net/packet/af_packet.c | 2 -- net/sched/sch_htb.c | 2 -- net/sunrpc/auth_gss/auth_gss.c | 2 -- net/sysctl_net.c | 1 - net/unix/af_unix.c | 2 -- 121 files changed, 254 deletions(-) (limited to 'net/core') diff --git a/include/linux/if_bridge.h b/include/linux/if_bridge.h index 950e13d09e06..6badb3e2c4e4 100644 --- a/include/linux/if_bridge.h +++ b/include/linux/if_bridge.h @@ -4,8 +4,6 @@ * Authors: * Lennert Buytenhek * - * $Id: if_bridge.h,v 1.1 2000/02/18 16:47:01 davem Exp $ - * * 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 diff --git a/include/linux/if_ppp.h b/include/linux/if_ppp.h index 0f2f70d4e48c..c3b1f8562709 100644 --- a/include/linux/if_ppp.h +++ b/include/linux/if_ppp.h @@ -1,5 +1,3 @@ -/* $Id: if_ppp.h,v 1.21 2000/03/27 06:03:36 paulus Exp $ */ - /* * if_ppp.h - Point-to-Point Protocol definitions. * diff --git a/include/linux/if_tun.h b/include/linux/if_tun.h index 8c71fe2fb1f5..18f31b6187a3 100644 --- a/include/linux/if_tun.h +++ b/include/linux/if_tun.h @@ -11,8 +11,6 @@ * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. - * - * $Id: if_tun.h,v 1.2 2001/06/01 18:39:47 davem Exp $ */ #ifndef __IF_TUN_H diff --git a/include/linux/ip6_tunnel.h b/include/linux/ip6_tunnel.h index af3f4a70f3df..1e7cc4af40de 100644 --- a/include/linux/ip6_tunnel.h +++ b/include/linux/ip6_tunnel.h @@ -1,7 +1,3 @@ -/* - * $Id$ - */ - #ifndef _IP6_TUNNEL_H #define _IP6_TUNNEL_H diff --git a/include/linux/ppp-comp.h b/include/linux/ppp-comp.h index e86a7a5cf355..b8d4ddd22736 100644 --- a/include/linux/ppp-comp.h +++ b/include/linux/ppp-comp.h @@ -23,8 +23,6 @@ * ON AN "AS IS" BASIS, AND THE AUSTRALIAN NATIONAL UNIVERSITY HAS NO * OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, * OR MODIFICATIONS. - * - * $Id: ppp-comp.h,v 1.6 1997/11/27 06:04:44 paulus Exp $ */ /* diff --git a/include/linux/ppp_defs.h b/include/linux/ppp_defs.h index c6b13ff85028..6e8adc77522c 100644 --- a/include/linux/ppp_defs.h +++ b/include/linux/ppp_defs.h @@ -1,5 +1,3 @@ -/* $Id: ppp_defs.h,v 1.2 1994/09/21 01:31:06 paulus Exp $ */ - /* * ppp_defs.h - PPP definitions. * diff --git a/include/linux/sunrpc/auth_gss.h b/include/linux/sunrpc/auth_gss.h index fec6899bf355..d48d4e605f74 100644 --- a/include/linux/sunrpc/auth_gss.h +++ b/include/linux/sunrpc/auth_gss.h @@ -7,8 +7,6 @@ * Andy Adamson * Bruce Fields * Copyright (c) 2000 The Regents of the University of Michigan - * - * $Id$ */ #ifndef _LINUX_SUNRPC_AUTH_GSS_H diff --git a/include/linux/sunrpc/gss_api.h b/include/linux/sunrpc/gss_api.h index 459c5fc11d51..03f33330ece2 100644 --- a/include/linux/sunrpc/gss_api.h +++ b/include/linux/sunrpc/gss_api.h @@ -7,8 +7,6 @@ * Andy Adamson * Bruce Fields * Copyright (c) 2000 The Regents of the University of Michigan - * - * $Id$ */ #ifndef _LINUX_SUNRPC_GSS_API_H diff --git a/include/linux/sunrpc/svcauth_gss.h b/include/linux/sunrpc/svcauth_gss.h index 417a1def56db..c9165d9771a8 100644 --- a/include/linux/sunrpc/svcauth_gss.h +++ b/include/linux/sunrpc/svcauth_gss.h @@ -3,9 +3,6 @@ * * Bruce Fields * Copyright (c) 2002 The Regents of the Unviersity of Michigan - * - * $Id$ - * */ #ifndef _LINUX_SUNRPC_SVCAUTH_GSS_H diff --git a/include/net/inetpeer.h b/include/net/inetpeer.h index ad8404b56113..15e1f8fe4c1f 100644 --- a/include/net/inetpeer.h +++ b/include/net/inetpeer.h @@ -1,8 +1,6 @@ /* * INETPEER - A storage for permanent information about peers * - * Version: $Id: inetpeer.h,v 1.2 2002/01/12 07:54:56 davem Exp $ - * * Authors: Andrey V. Savochkin */ diff --git a/include/net/ip6_tunnel.h b/include/net/ip6_tunnel.h index 3780592ebe8e..83b4e008b16d 100644 --- a/include/net/ip6_tunnel.h +++ b/include/net/ip6_tunnel.h @@ -1,7 +1,3 @@ -/* - * $Id$ - */ - #ifndef _NET_IP6_TUNNEL_H #define _NET_IP6_TUNNEL_H diff --git a/include/net/ipconfig.h b/include/net/ipconfig.h index 3924d7d2cb11..c74cc1bd5a02 100644 --- a/include/net/ipconfig.h +++ b/include/net/ipconfig.h @@ -1,6 +1,4 @@ /* - * $Id: ipconfig.h,v 1.4 2001/04/30 04:51:46 davem Exp $ - * * Copyright (C) 1997 Martin Mares * * Automatic IP Layer Configuration diff --git a/include/net/ipv6.h b/include/net/ipv6.h index e0a612bc9c4e..7f7db8d57934 100644 --- a/include/net/ipv6.h +++ b/include/net/ipv6.h @@ -4,8 +4,6 @@ * Authors: * Pedro Roque * - * $Id: ipv6.h,v 1.1 2002/05/20 15:13:07 jgrimm Exp $ - * * 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 diff --git a/include/net/snmp.h b/include/net/snmp.h index ce2f48507510..57c93628695f 100644 --- a/include/net/snmp.h +++ b/include/net/snmp.h @@ -14,8 +14,6 @@ * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. * - * $Id: snmp.h,v 1.19 2001/06/14 13:40:46 davem Exp $ - * */ #ifndef _SNMP_H diff --git a/net/bluetooth/bnep/bnep.h b/net/bluetooth/bnep/bnep.h index e69244dd8de8..b69bf4e7c48b 100644 --- a/net/bluetooth/bnep/bnep.h +++ b/net/bluetooth/bnep/bnep.h @@ -16,10 +16,6 @@ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ -/* - * $Id: bnep.h,v 1.5 2002/08/04 21:23:58 maxk Exp $ - */ - #ifndef _BNEP_H #define _BNEP_H diff --git a/net/bluetooth/bnep/core.c b/net/bluetooth/bnep/core.c index f85d94643aaf..1d98a1b80da7 100644 --- a/net/bluetooth/bnep/core.c +++ b/net/bluetooth/bnep/core.c @@ -25,10 +25,6 @@ SOFTWARE IS DISCLAIMED. */ -/* - * $Id: core.c,v 1.20 2002/08/04 21:23:58 maxk Exp $ - */ - #include #include diff --git a/net/bluetooth/bnep/netdev.c b/net/bluetooth/bnep/netdev.c index 95e3837e4312..d9fa0ab2c87f 100644 --- a/net/bluetooth/bnep/netdev.c +++ b/net/bluetooth/bnep/netdev.c @@ -25,10 +25,6 @@ SOFTWARE IS DISCLAIMED. */ -/* - * $Id: netdev.c,v 1.8 2002/08/04 21:23:58 maxk Exp $ - */ - #include #include diff --git a/net/bluetooth/bnep/sock.c b/net/bluetooth/bnep/sock.c index 201e5b1ce473..8ffb57f2303a 100644 --- a/net/bluetooth/bnep/sock.c +++ b/net/bluetooth/bnep/sock.c @@ -24,10 +24,6 @@ SOFTWARE IS DISCLAIMED. */ -/* - * $Id: sock.c,v 1.4 2002/08/04 21:23:58 maxk Exp $ - */ - #include #include diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index 0c2c93735e93..b4fb84e398e5 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -23,8 +23,6 @@ /* * Bluetooth RFCOMM core. - * - * $Id: core.c,v 1.42 2002/10/01 23:26:25 maxk Exp $ */ #include diff --git a/net/bluetooth/rfcomm/sock.c b/net/bluetooth/rfcomm/sock.c index 5083adcbfae5..c9054487670a 100644 --- a/net/bluetooth/rfcomm/sock.c +++ b/net/bluetooth/rfcomm/sock.c @@ -23,8 +23,6 @@ /* * RFCOMM sockets. - * - * $Id: sock.c,v 1.24 2002/10/03 01:00:34 maxk Exp $ */ #include diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c index c9191871c1e0..be84f4fc1477 100644 --- a/net/bluetooth/rfcomm/tty.c +++ b/net/bluetooth/rfcomm/tty.c @@ -23,8 +23,6 @@ /* * RFCOMM TTY. - * - * $Id: tty.c,v 1.24 2002/10/03 01:54:38 holtmann Exp $ */ #include diff --git a/net/bridge/br.c b/net/bridge/br.c index 8f3c58e5f7a5..cede010f4ddd 100644 --- a/net/bridge/br.c +++ b/net/bridge/br.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br.c,v 1.47 2001/12/24 00:56:41 davem Exp $ - * * 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 diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c index 626c7795ae30..a6ffc6c2a69f 100644 --- a/net/bridge/br_device.c +++ b/net/bridge/br_device.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_device.c,v 1.6 2001/12/24 00:59:55 davem Exp $ - * * 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 diff --git a/net/bridge/br_fdb.c b/net/bridge/br_fdb.c index 72c5976a5ce3..4de74cdd091d 100644 --- a/net/bridge/br_fdb.c +++ b/net/bridge/br_fdb.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_fdb.c,v 1.6 2002/01/17 00:57:07 davem Exp $ - * * 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 diff --git a/net/bridge/br_forward.c b/net/bridge/br_forward.c index a4711674b3df..512645727f51 100644 --- a/net/bridge/br_forward.c +++ b/net/bridge/br_forward.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_forward.c,v 1.4 2001/08/14 22:05:57 davem Exp $ - * * 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 diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index c2397f503b0f..143c954681b8 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_if.c,v 1.7 2001/12/24 00:59:55 davem Exp $ - * * 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 diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c index fa0f5711a996..0145e9416714 100644 --- a/net/bridge/br_input.c +++ b/net/bridge/br_input.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_input.c,v 1.10 2001/12/24 04:50:20 davem Exp $ - * * 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 diff --git a/net/bridge/br_ioctl.c b/net/bridge/br_ioctl.c index 0655a5f07f58..eeee218eed80 100644 --- a/net/bridge/br_ioctl.c +++ b/net/bridge/br_ioctl.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_ioctl.c,v 1.4 2000/11/08 05:16:40 davem Exp $ - * * 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 diff --git a/net/bridge/br_notify.c b/net/bridge/br_notify.c index 00644a544e3c..88d8ec7b3142 100644 --- a/net/bridge/br_notify.c +++ b/net/bridge/br_notify.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_notify.c,v 1.2 2000/02/21 15:51:34 davem Exp $ - * * 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 diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 0243cb489edd..83ff5861c2d2 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -4,8 +4,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_private.h,v 1.7 2001/12/24 00:59:55 davem Exp $ - * * 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 diff --git a/net/bridge/br_private_stp.h b/net/bridge/br_private_stp.h index e29f01ac1adf..8b650f7fbfa0 100644 --- a/net/bridge/br_private_stp.h +++ b/net/bridge/br_private_stp.h @@ -4,8 +4,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_private_stp.h,v 1.3 2001/02/05 06:03:47 davem Exp $ - * * 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 diff --git a/net/bridge/br_stp.c b/net/bridge/br_stp.c index e38034aa56f5..284d1b2fa1ff 100644 --- a/net/bridge/br_stp.c +++ b/net/bridge/br_stp.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp.c,v 1.4 2000/06/19 10:13:35 davem Exp $ - * * 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 diff --git a/net/bridge/br_stp_bpdu.c b/net/bridge/br_stp_bpdu.c index ddeb6e5d45d6..9dc2de656965 100644 --- a/net/bridge/br_stp_bpdu.c +++ b/net/bridge/br_stp_bpdu.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp_bpdu.c,v 1.3 2001/11/10 02:35:25 davem Exp $ - * * 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 diff --git a/net/bridge/br_stp_if.c b/net/bridge/br_stp_if.c index 1a430eccec9b..1a4e5c37a0cf 100644 --- a/net/bridge/br_stp_if.c +++ b/net/bridge/br_stp_if.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp_if.c,v 1.4 2001/04/14 21:14:39 davem Exp $ - * * 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 diff --git a/net/bridge/br_stp_timer.c b/net/bridge/br_stp_timer.c index 77f5255e6915..772a140bfdf0 100644 --- a/net/bridge/br_stp_timer.c +++ b/net/bridge/br_stp_timer.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp_timer.c,v 1.3 2000/05/05 02:17:17 davem Exp $ - * * 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 diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 1e556d312117..3e18f8525e82 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -4,8 +4,6 @@ * Authors: Alan Cox * Florian La Roche * - * Version: $Id: skbuff.c,v 1.90 2001/11/07 05:56:19 davem Exp $ - * * Fixes: * Alan Cox : Fixed the worst of the load * balancer bugs. diff --git a/net/core/sock.c b/net/core/sock.c index 88094cb09c06..3879bf65897e 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -7,8 +7,6 @@ * handler for protocols to use and generic option handler. * * - * Version: $Id: sock.c,v 1.117 2002/02/01 22:01:03 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Florian La Roche, diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 24eca23c2db3..42bd24b64b57 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -5,8 +5,6 @@ * * PF_INET protocol family socket handler. * - * Version: $Id: af_inet.c,v 1.137 2002/02/01 22:01:03 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Florian La Roche, diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c index 9b539fa9fe18..20c515a1be28 100644 --- a/net/ipv4/arp.c +++ b/net/ipv4/arp.c @@ -1,6 +1,4 @@ /* linux/net/ipv4/arp.c - * - * Version: $Id: arp.c,v 1.99 2001/08/30 22:55:42 davem Exp $ * * Copyright (C) 1994 by Florian La Roche * diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index 61011e1d580e..f8c0b0aea93a 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -1,8 +1,6 @@ /* * NET3 IP device support routines. * - * Version: $Id: devinet.c,v 1.44 2001/10/31 21:55:54 davem Exp $ - * * 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 diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c index 0b2ac6a3d903..5ad01d63f83b 100644 --- a/net/ipv4/fib_frontend.c +++ b/net/ipv4/fib_frontend.c @@ -5,8 +5,6 @@ * * IPv4 Forwarding Information Base: FIB frontend. * - * Version: $Id: fib_frontend.c,v 1.26 2001/10/31 21:55:54 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/fib_hash.c b/net/ipv4/fib_hash.c index 2e2fc3376ac9..eeec4bf982b8 100644 --- a/net/ipv4/fib_hash.c +++ b/net/ipv4/fib_hash.c @@ -5,8 +5,6 @@ * * IPv4 FIB: lookup engine and maintenance routines. * - * Version: $Id: fib_hash.c,v 1.13 2001/10/31 21:55:54 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c index 3b83c34019fc..9335eba683c3 100644 --- a/net/ipv4/fib_semantics.c +++ b/net/ipv4/fib_semantics.c @@ -5,8 +5,6 @@ * * IPv4 Forwarding Information Base: semantics. * - * Version: $Id: fib_semantics.c,v 1.19 2002/01/12 07:54:56 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index 4b02d14e7ab9..394db9c941a1 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -22,8 +22,6 @@ * IP-address lookup using LC-tries. Stefan Nilsson and Gunnar Karlsson * IEEE Journal on Selected Areas in Communications, 17(6):1083-1092, June 1999 * - * Version: $Id: fib_trie.c,v 1.3 2005/06/08 14:20:01 robert Exp $ - * * * Code from fib_hash has been reused which includes the following header: * diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index 87397351ddac..aa7cf46853b7 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -3,8 +3,6 @@ * * Alan Cox, * - * Version: $Id: icmp.c,v 1.85 2002/02/01 22:01:03 davem Exp $ - * * 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 diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index 2769dc4a4c84..68e84a933e90 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -8,8 +8,6 @@ * the older version didn't come out right using gcc 2.5.8, the newer one * seems to fall out with gcc 2.6.2. * - * Version: $Id: igmp.c,v 1.47 2002/02/01 22:01:03 davem Exp $ - * * Authors: * Alan Cox * diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c index da97695e7096..c10036e7a463 100644 --- a/net/ipv4/inet_diag.c +++ b/net/ipv4/inet_diag.c @@ -1,8 +1,6 @@ /* * inet_diag.c Module for monitoring INET transport protocols sockets. * - * Version: $Id: inet_diag.c,v 1.3 2002/02/01 22:01:04 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/inetpeer.c b/net/ipv4/inetpeer.c index af995198f643..a456ceeac3f2 100644 --- a/net/ipv4/inetpeer.c +++ b/net/ipv4/inetpeer.c @@ -3,8 +3,6 @@ * * This source is covered by the GNU GPL, the same as all kernel sources. * - * Version: $Id: inetpeer.c,v 1.7 2001/09/20 21:22:50 davem Exp $ - * * Authors: Andrey V. Savochkin */ diff --git a/net/ipv4/ip_forward.c b/net/ipv4/ip_forward.c index 4813c39b438b..37d36a3f33cd 100644 --- a/net/ipv4/ip_forward.c +++ b/net/ipv4/ip_forward.c @@ -5,8 +5,6 @@ * * The IP forwarding functionality. * - * Version: $Id: ip_forward.c,v 1.48 2000/12/13 18:31:48 davem Exp $ - * * Authors: see ip.c * * Fixes: diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index be1cb89a8d5a..91e321407313 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -5,8 +5,6 @@ * * The IP fragmentation functionality. * - * Version: $Id: ip_fragment.c,v 1.59 2002/01/12 07:54:56 davem Exp $ - * * Authors: Fred N. van Kempen * Alan Cox * diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c index ff77a4a7f9ec..7c26428ea67b 100644 --- a/net/ipv4/ip_input.c +++ b/net/ipv4/ip_input.c @@ -5,8 +5,6 @@ * * The Internet Protocol (IP) module. * - * Version: $Id: ip_input.c,v 1.55 2002/01/12 07:39:45 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Donald Becker, diff --git a/net/ipv4/ip_options.c b/net/ipv4/ip_options.c index 33126ad2cfdc..be3f18a7a40e 100644 --- a/net/ipv4/ip_options.c +++ b/net/ipv4/ip_options.c @@ -5,8 +5,6 @@ * * The options processing module for ip.c * - * Version: $Id: ip_options.c,v 1.21 2001/09/01 00:31:50 davem Exp $ - * * Authors: A.N.Kuznetsov * */ diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index e527628f56cf..f1278eecf56d 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -5,8 +5,6 @@ * * The Internet Protocol (IP) output module. * - * Version: $Id: ip_output.c,v 1.100 2002/02/01 22:01:03 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Donald Becker, diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c index e0514e82308e..105d92a039b9 100644 --- a/net/ipv4/ip_sockglue.c +++ b/net/ipv4/ip_sockglue.c @@ -5,8 +5,6 @@ * * The IP to API glue. * - * Version: $Id: ip_sockglue.c,v 1.62 2002/02/01 22:01:04 davem Exp $ - * * Authors: see ip.c * * Fixes: diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c index ed45037ce9be..b88aa9afa42e 100644 --- a/net/ipv4/ipconfig.c +++ b/net/ipv4/ipconfig.c @@ -1,6 +1,4 @@ /* - * $Id: ipconfig.c,v 1.46 2002/02/01 22:01:04 davem Exp $ - * * Automatic Configuration of IP -- use DHCP, BOOTP, RARP, or * user-supplied information to configure own IP address and routes. * diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c index 86d8836551b9..4c6d2caf9203 100644 --- a/net/ipv4/ipip.c +++ b/net/ipv4/ipip.c @@ -1,8 +1,6 @@ /* * Linux NET3: IP/IP protocol decoder. * - * Version: $Id: ipip.c,v 1.50 2001/10/02 02:22:36 davem Exp $ - * * Authors: * Sam Lantinga (slouken@cs.ucdavis.edu) 02/01/95 * diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index a34da4977c73..300ab0c2919e 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -9,8 +9,6 @@ * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. * - * Version: $Id: ipmr.c,v 1.65 2001/10/31 21:55:54 davem Exp $ - * * Fixes: * Michael Chastain : Incorrect size of copying. * Alan Cox : Added the cache manager code diff --git a/net/ipv4/ipvs/ip_vs_app.c b/net/ipv4/ipvs/ip_vs_app.c index 535abe0c45e7..1f1897a1a702 100644 --- a/net/ipv4/ipvs/ip_vs_app.c +++ b/net/ipv4/ipvs/ip_vs_app.c @@ -1,8 +1,6 @@ /* * ip_vs_app.c: Application module support for IPVS * - * Version: $Id: ip_vs_app.c,v 1.17 2003/03/22 06:31:21 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_conn.c b/net/ipv4/ipvs/ip_vs_conn.c index 65f1ba112752..f8bdae47a77f 100644 --- a/net/ipv4/ipvs/ip_vs_conn.c +++ b/net/ipv4/ipvs/ip_vs_conn.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_conn.c,v 1.31 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * Julian Anastasov diff --git a/net/ipv4/ipvs/ip_vs_core.c b/net/ipv4/ipvs/ip_vs_core.c index 963981a9d501..bcf6276ba4b2 100644 --- a/net/ipv4/ipvs/ip_vs_core.c +++ b/net/ipv4/ipvs/ip_vs_core.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_core.c,v 1.34 2003/05/10 03:05:23 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * Julian Anastasov diff --git a/net/ipv4/ipvs/ip_vs_ctl.c b/net/ipv4/ipvs/ip_vs_ctl.c index 94c5767c8e01..9a5ace0b4dd6 100644 --- a/net/ipv4/ipvs/ip_vs_ctl.c +++ b/net/ipv4/ipvs/ip_vs_ctl.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_ctl.c,v 1.36 2003/06/08 09:31:19 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * Julian Anastasov diff --git a/net/ipv4/ipvs/ip_vs_dh.c b/net/ipv4/ipvs/ip_vs_dh.c index dcf5d46aaa5e..8afc1503ed20 100644 --- a/net/ipv4/ipvs/ip_vs_dh.c +++ b/net/ipv4/ipvs/ip_vs_dh.c @@ -1,8 +1,6 @@ /* * IPVS: Destination Hashing scheduling module * - * Version: $Id: ip_vs_dh.c,v 1.5 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * Inspired by the consistent hashing scheduler patch from diff --git a/net/ipv4/ipvs/ip_vs_est.c b/net/ipv4/ipvs/ip_vs_est.c index dfa0d713c801..bc04eedd6dbb 100644 --- a/net/ipv4/ipvs/ip_vs_est.c +++ b/net/ipv4/ipvs/ip_vs_est.c @@ -1,8 +1,6 @@ /* * ip_vs_est.c: simple rate estimator for IPVS * - * Version: $Id: ip_vs_est.c,v 1.4 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_ftp.c b/net/ipv4/ipvs/ip_vs_ftp.c index 59aa166b7678..c1c758e4f733 100644 --- a/net/ipv4/ipvs/ip_vs_ftp.c +++ b/net/ipv4/ipvs/ip_vs_ftp.c @@ -1,8 +1,6 @@ /* * ip_vs_ftp.c: IPVS ftp application module * - * Version: $Id: ip_vs_ftp.c,v 1.13 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * Changes: diff --git a/net/ipv4/ipvs/ip_vs_lblc.c b/net/ipv4/ipvs/ip_vs_lblc.c index 3888642706ad..0efa3db4b180 100644 --- a/net/ipv4/ipvs/ip_vs_lblc.c +++ b/net/ipv4/ipvs/ip_vs_lblc.c @@ -1,8 +1,6 @@ /* * IPVS: Locality-Based Least-Connection scheduling module * - * Version: $Id: ip_vs_lblc.c,v 1.10 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_lblcr.c b/net/ipv4/ipvs/ip_vs_lblcr.c index daa260eb21cf..8e3bbeb45138 100644 --- a/net/ipv4/ipvs/ip_vs_lblcr.c +++ b/net/ipv4/ipvs/ip_vs_lblcr.c @@ -1,8 +1,6 @@ /* * IPVS: Locality-Based Least-Connection with Replication scheduler * - * Version: $Id: ip_vs_lblcr.c,v 1.11 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_lc.c b/net/ipv4/ipvs/ip_vs_lc.c index d88fef90a641..ac9f08e065d5 100644 --- a/net/ipv4/ipvs/ip_vs_lc.c +++ b/net/ipv4/ipvs/ip_vs_lc.c @@ -1,8 +1,6 @@ /* * IPVS: Least-Connection Scheduling module * - * Version: $Id: ip_vs_lc.c,v 1.10 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_nq.c b/net/ipv4/ipvs/ip_vs_nq.c index bc2a9e5f2a7b..a46bf258d420 100644 --- a/net/ipv4/ipvs/ip_vs_nq.c +++ b/net/ipv4/ipvs/ip_vs_nq.c @@ -1,8 +1,6 @@ /* * IPVS: Never Queue scheduling module * - * Version: $Id: ip_vs_nq.c,v 1.2 2003/06/08 09:31:19 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_proto.c b/net/ipv4/ipvs/ip_vs_proto.c index 4b1c16cbb16b..876714f23d65 100644 --- a/net/ipv4/ipvs/ip_vs_proto.c +++ b/net/ipv4/ipvs/ip_vs_proto.c @@ -1,8 +1,6 @@ /* * ip_vs_proto.c: transport protocol load balancing support for IPVS * - * Version: $Id: ip_vs_proto.c,v 1.2 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/ipvs/ip_vs_proto_ah.c b/net/ipv4/ipvs/ip_vs_proto_ah.c index 4bf835e1d86d..73e0ea87c1f5 100644 --- a/net/ipv4/ipvs/ip_vs_proto_ah.c +++ b/net/ipv4/ipvs/ip_vs_proto_ah.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_ah.c: AH IPSec load balancing support for IPVS * - * Version: $Id: ip_vs_proto_ah.c,v 1.1 2003/07/04 15:04:37 wensong Exp $ - * * Authors: Julian Anastasov , February 2002 * Wensong Zhang * diff --git a/net/ipv4/ipvs/ip_vs_proto_esp.c b/net/ipv4/ipvs/ip_vs_proto_esp.c index db6a6b7b1a0b..21d70c8ffa54 100644 --- a/net/ipv4/ipvs/ip_vs_proto_esp.c +++ b/net/ipv4/ipvs/ip_vs_proto_esp.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_esp.c: ESP IPSec load balancing support for IPVS * - * Version: $Id: ip_vs_proto_esp.c,v 1.1 2003/07/04 15:04:37 wensong Exp $ - * * Authors: Julian Anastasov , February 2002 * Wensong Zhang * diff --git a/net/ipv4/ipvs/ip_vs_proto_tcp.c b/net/ipv4/ipvs/ip_vs_proto_tcp.c index b83dc14b0a4d..d0ea467986a0 100644 --- a/net/ipv4/ipvs/ip_vs_proto_tcp.c +++ b/net/ipv4/ipvs/ip_vs_proto_tcp.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_tcp.c: TCP load balancing support for IPVS * - * Version: $Id: ip_vs_proto_tcp.c,v 1.3 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/ipvs/ip_vs_proto_udp.c b/net/ipv4/ipvs/ip_vs_proto_udp.c index 75771cb3cd6f..c6be5d56823f 100644 --- a/net/ipv4/ipvs/ip_vs_proto_udp.c +++ b/net/ipv4/ipvs/ip_vs_proto_udp.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_udp.c: UDP load balancing support for IPVS * - * Version: $Id: ip_vs_proto_udp.c,v 1.3 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/ipvs/ip_vs_rr.c b/net/ipv4/ipvs/ip_vs_rr.c index 433f8a947924..c8db12d39e61 100644 --- a/net/ipv4/ipvs/ip_vs_rr.c +++ b/net/ipv4/ipvs/ip_vs_rr.c @@ -1,8 +1,6 @@ /* * IPVS: Round-Robin Scheduling module * - * Version: $Id: ip_vs_rr.c,v 1.9 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * diff --git a/net/ipv4/ipvs/ip_vs_sched.c b/net/ipv4/ipvs/ip_vs_sched.c index 121a32b1b756..b64767309855 100644 --- a/net/ipv4/ipvs/ip_vs_sched.c +++ b/net/ipv4/ipvs/ip_vs_sched.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_sched.c,v 1.13 2003/05/10 03:05:23 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * diff --git a/net/ipv4/ipvs/ip_vs_sed.c b/net/ipv4/ipvs/ip_vs_sed.c index dd7c128f9db3..2a7d31358181 100644 --- a/net/ipv4/ipvs/ip_vs_sed.c +++ b/net/ipv4/ipvs/ip_vs_sed.c @@ -1,8 +1,6 @@ /* * IPVS: Shortest Expected Delay scheduling module * - * Version: $Id: ip_vs_sed.c,v 1.1 2003/05/10 03:06:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_sh.c b/net/ipv4/ipvs/ip_vs_sh.c index 1b25b00ef1e1..b8fdfac65001 100644 --- a/net/ipv4/ipvs/ip_vs_sh.c +++ b/net/ipv4/ipvs/ip_vs_sh.c @@ -1,8 +1,6 @@ /* * IPVS: Source Hashing scheduling module * - * Version: $Id: ip_vs_sh.c,v 1.5 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index eff54efe0351..2d4a86f73325 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_sync.c,v 1.13 2003/06/08 09:31:19 wensong Exp $ - * * Authors: Wensong Zhang * * ip_vs_sync: sync connection info from master load balancer to backups diff --git a/net/ipv4/ipvs/ip_vs_wlc.c b/net/ipv4/ipvs/ip_vs_wlc.c index 8a9d913261d8..772c3cb4eca1 100644 --- a/net/ipv4/ipvs/ip_vs_wlc.c +++ b/net/ipv4/ipvs/ip_vs_wlc.c @@ -1,8 +1,6 @@ /* * IPVS: Weighted Least-Connection Scheduling module * - * Version: $Id: ip_vs_wlc.c,v 1.13 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * diff --git a/net/ipv4/ipvs/ip_vs_wrr.c b/net/ipv4/ipvs/ip_vs_wrr.c index 85c680add6df..1d6932d7dc97 100644 --- a/net/ipv4/ipvs/ip_vs_wrr.c +++ b/net/ipv4/ipvs/ip_vs_wrr.c @@ -1,8 +1,6 @@ /* * IPVS: Weighted Round-Robin Scheduling module * - * Version: $Id: ip_vs_wrr.c,v 1.12 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_xmit.c b/net/ipv4/ipvs/ip_vs_xmit.c index f63006caea03..9892d4aca42e 100644 --- a/net/ipv4/ipvs/ip_vs_xmit.c +++ b/net/ipv4/ipvs/ip_vs_xmit.c @@ -1,8 +1,6 @@ /* * ip_vs_xmit.c: various packet transmitters for IPVS * - * Version: $Id: ip_vs_xmit.c,v 1.2 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 552169b41b16..eb5cee279c5f 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -7,8 +7,6 @@ * PROC file system. It is mainly used for debugging and * statistics. * - * Version: $Id: proc.c,v 1.45 2001/05/16 16:45:35 davem Exp $ - * * Authors: Fred N. van Kempen, * Gerald J. Heim, * Fred Baumgarten, diff --git a/net/ipv4/protocol.c b/net/ipv4/protocol.c index 971ab9356e51..ea50da0649fd 100644 --- a/net/ipv4/protocol.c +++ b/net/ipv4/protocol.c @@ -5,8 +5,6 @@ * * INET protocol dispatch tables. * - * Version: $Id: protocol.c,v 1.14 2001/05/18 02:25:49 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index e7e091d365ff..1d0c97c8712d 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -5,8 +5,6 @@ * * RAW - implementation of IP "raw" sockets. * - * Version: $Id: raw.c,v 1.64 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 96be336064fb..fe3a02237286 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -5,8 +5,6 @@ * * ROUTE - implementation of the IP router. * - * Version: $Id: route.c,v 1.103 2002/01/12 07:44:09 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Alan Cox, diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c index 73ba98921d64..6317d3c8dc0d 100644 --- a/net/ipv4/syncookies.c +++ b/net/ipv4/syncookies.c @@ -8,8 +8,6 @@ * 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. - * - * $Id: syncookies.c,v 1.18 2002/02/01 22:01:04 davem Exp $ */ #include diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index c437f804ee38..901607003205 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -1,8 +1,6 @@ /* * sysctl_net_ipv4.c: sysctl interface to net IPV4 subsystem. * - * $Id: sysctl_net_ipv4.c,v 1.50 2001/10/20 00:00:11 davem Exp $ - * * Begun April 1, 1996, Mike Shaver. * Added /proc/sys/net/ipv4 directory entry (empty =) ). [MS] */ diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index ab66683b8043..ad66b09e0bcd 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp.c,v 1.216 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_diag.c b/net/ipv4/tcp_diag.c index 2fbcc7d1b1a0..838d491dfda7 100644 --- a/net/ipv4/tcp_diag.c +++ b/net/ipv4/tcp_diag.c @@ -1,8 +1,6 @@ /* * tcp_diag.c Module for monitoring TCP transport protocols sockets. * - * Version: $Id: tcp_diag.c,v 1.3 2002/02/01 22:01:04 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index eba873e9b560..b68c3c7d906b 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_input.c,v 1.243 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index cd601a866c2f..f2926ae1de57 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_ipv4.c,v 1.240 2002/02/01 22:01:04 davem Exp $ - * * IPv4 specific functions * * diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 019c8c16e5cc..1276cab85e3e 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_minisocks.c,v 1.15 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index ad993ecb4810..b171ac65ccab 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_output.c,v 1.146 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 4de68cf5f2aa..e77e7ae0bf2c 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_timer.c,v 1.88 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 56fcda3694ba..355e6d62d483 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -5,8 +5,6 @@ * * The User Datagram Protocol (UDP). * - * Version: $Id: udp.c,v 1.102 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Arnt Gulbrandsen, diff --git a/net/ipv4/udplite.c b/net/ipv4/udplite.c index 72ce26b6c4d3..4ad16b6d5138 100644 --- a/net/ipv4/udplite.c +++ b/net/ipv4/udplite.c @@ -1,8 +1,6 @@ /* * UDPLITE An implementation of the UDP-Lite protocol (RFC 3828). * - * Version: $Id: udplite.c,v 1.25 2006/10/19 07:22:36 gerrit Exp $ - * * Authors: Gerrit Renker * * Changes: diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 147588f4c7c0..deb38bf03376 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -6,8 +6,6 @@ * Pedro Roque * Alexey Kuznetsov * - * $Id: addrconf.c,v 1.69 2001/10/31 21:55:54 davem Exp $ - * * 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 diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index e84b3fd17fb4..350457c761e6 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -7,8 +7,6 @@ * * Adapted from linux/net/ipv4/af_inet.c * - * $Id: af_inet6.c,v 1.66 2002/02/01 22:01:04 davem Exp $ - * * Fixes: * piggy, Karl Knutson : Socket protocol table * Hideaki YOSHIFUJI : sin6_scope_id support diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c index b9c2de84a8a2..8cdb6b65ee96 100644 --- a/net/ipv6/datagram.c +++ b/net/ipv6/datagram.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: datagram.c,v 1.24 2002/02/01 22:01:04 davem Exp $ - * * 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 diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c index 3cd1c993d52b..602ea826f0a5 100644 --- a/net/ipv6/exthdrs.c +++ b/net/ipv6/exthdrs.c @@ -7,8 +7,6 @@ * Andi Kleen * Alexey Kuznetsov * - * $Id: exthdrs.c,v 1.13 2001/06/19 15:58:56 davem Exp $ - * * 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 diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c index d42dd16d3487..399d41f65437 100644 --- a/net/ipv6/icmp.c +++ b/net/ipv6/icmp.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: icmp.c,v 1.38 2002/02/08 03:57:19 davem Exp $ - * * Based on net/ipv4/icmp.c * * RFC 1885 diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c index 1ee4fa17c129..4de2b9efcacb 100644 --- a/net/ipv6/ip6_fib.c +++ b/net/ipv6/ip6_fib.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: ip6_fib.c,v 1.25 2001/10/31 21:55:55 davem Exp $ - * * 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 diff --git a/net/ipv6/ip6_input.c b/net/ipv6/ip6_input.c index 4e5c8615832c..f77a6011c302 100644 --- a/net/ipv6/ip6_input.c +++ b/net/ipv6/ip6_input.c @@ -6,8 +6,6 @@ * Pedro Roque * Ian P. Morris * - * $Id: ip6_input.c,v 1.19 2000/12/13 18:31:50 davem Exp $ - * * Based in linux/net/ipv4/ip_input.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 48cdce9c696c..40a2813a63d1 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: ip6_output.c,v 1.34 2002/02/01 22:01:04 davem Exp $ - * * Based on linux/net/ipv4/ip_output.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index 37814810ac49..17c7b098cdb0 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -6,8 +6,6 @@ * Ville Nuorvala * Yasuyuki Kozakai * - * $Id$ - * * Based on: * linux/net/ipv6/sit.c and linux/net/ipv4/ipip.c * diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c index 26b83e512a09..237ebbb9383b 100644 --- a/net/ipv6/ipv6_sockglue.c +++ b/net/ipv6/ipv6_sockglue.c @@ -7,8 +7,6 @@ * * Based on linux/net/ipv4/ip_sockglue.c * - * $Id: ipv6_sockglue.c,v 1.41 2002/02/01 22:01:04 davem Exp $ - * * 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 diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c index fd632dd7f98d..fbb2d12c41bc 100644 --- a/net/ipv6/mcast.c +++ b/net/ipv6/mcast.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: mcast.c,v 1.40 2002/02/08 03:57:19 davem Exp $ - * * Based on linux/ipv4/igmp.c and linux/ipv4/ip_sockglue.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c index df0736a4cafa..cbc7e514d3ec 100644 --- a/net/ipv6/proc.c +++ b/net/ipv6/proc.c @@ -7,8 +7,6 @@ * PROC file system. This is very similar to the IPv4 version, * except it reports the sockets in the INET6 address family. * - * Version: $Id: proc.c,v 1.17 2002/02/01 22:01:04 davem Exp $ - * * Authors: David S. Miller (davem@caip.rutgers.edu) * YOSHIFUJI Hideaki * diff --git a/net/ipv6/protocol.c b/net/ipv6/protocol.c index f929f47b925e..9ab789159913 100644 --- a/net/ipv6/protocol.c +++ b/net/ipv6/protocol.c @@ -5,8 +5,6 @@ * * PF_INET6 protocol dispatch tables. * - * Version: $Id: protocol.c,v 1.10 2001/05/18 02:25:49 davem Exp $ - * * Authors: Pedro Roque * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 8fee9a15b2d3..e03c1898ab2e 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -7,8 +7,6 @@ * * Adapted from linux/net/ipv4/raw.c * - * $Id: raw.c,v 1.51 2002/02/01 22:01:04 davem Exp $ - * * Fixes: * Hideaki YOSHIFUJI : sin6_scope_id support * YOSHIFUJI,H.@USAGI : raw checksum (RFC2292(bis) compliance) diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c index 9391a6949b96..13509f906d89 100644 --- a/net/ipv6/reassembly.c +++ b/net/ipv6/reassembly.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: reassembly.c,v 1.26 2001/03/07 22:00:57 davem Exp $ - * * Based on: net/ipv4/ip_fragment.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 220cffe9e63b..edae81319b51 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: route.c,v 1.56 2001/10/31 21:55:55 davem Exp $ - * * 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 diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c index 6b8f0583b637..b0c5080420a8 100644 --- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -6,8 +6,6 @@ * Pedro Roque * Alexey Kuznetsov * - * $Id: sit.c,v 1.53 2001/09/25 05:09:53 davem Exp $ - * * 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 diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 715965f0fac0..155499197fc5 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: tcp_ipv6.c,v 1.144 2002/02/01 22:01:04 davem Exp $ - * * Based on: * linux/net/ipv4/tcp.c * linux/net/ipv4/tcp_input.c diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index dd309626ae9a..e0693fffc9bd 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -7,8 +7,6 @@ * * Based on linux/ipv4/udp.c * - * $Id: udp.c,v 1.65 2002/02/01 22:01:04 davem Exp $ - * * Fixes: * Hideaki YOSHIFUJI : sin6_scope_id support * YOSHIFUJI Hideaki @USAGI and: Support IPV6_V6ONLY socket option, which diff --git a/net/ipv6/udplite.c b/net/ipv6/udplite.c index 491efd00a866..f6cdcb348e05 100644 --- a/net/ipv6/udplite.c +++ b/net/ipv6/udplite.c @@ -2,8 +2,6 @@ * UDPLITEv6 An implementation of the UDP-Lite protocol over IPv6. * See also net/ipv4/udplite.c * - * Version: $Id: udplite.c,v 1.9 2006/10/19 08:28:10 gerrit Exp $ - * * Authors: Gerrit Renker * * Changes: diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 2cee87da4441..beca6402f1cf 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -5,8 +5,6 @@ * * PACKET - implements raw packet sockets. * - * Version: $Id: af_packet.c,v 1.61 2002/02/08 03:57:19 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Alan Cox, diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 5bc1ed490180..213071859030 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -24,8 +24,6 @@ * Jiri Fojtasek * fixed requeue routine * and many others. thanks. - * - * $Id: sch_htb.c,v 1.25 2003/12/07 11:08:25 devik Exp devik $ */ #include #include diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c index cc12d5f5d5da..019d4b4478c9 100644 --- a/net/sunrpc/auth_gss/auth_gss.c +++ b/net/sunrpc/auth_gss/auth_gss.c @@ -33,8 +33,6 @@ * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. - * - * $Id$ */ diff --git a/net/sysctl_net.c b/net/sysctl_net.c index d8e79162724c..007c1a6708ee 100644 --- a/net/sysctl_net.c +++ b/net/sysctl_net.c @@ -4,7 +4,6 @@ * Begun April 1, 1996, Mike Shaver. * Added /proc/sys/net directories for each protocol family. [MS] * - * $Log: sysctl_net.c,v $ * Revision 1.2 1996/05/08 20:24:40 shaver * Added bits for NET_BRIDGE and the NET_IPV4_ARP stuff and * NET_IPV4_IP_FORWARD. diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index e18cd3628db4..392e80e3268d 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -8,8 +8,6 @@ * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. * - * Version: $Id: af_unix.c,v 1.133 2002/02/08 03:57:19 davem Exp $ - * * Fixes: * Linus Torvalds : Assorted bug cures. * Niibe Yutaka : async I/O support. -- cgit v1.3-8-gc7d7 From c1da4ac752b8b0411791d26c678fcf23d2eed242 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Fri, 13 Jun 2008 18:12:00 -0700 Subject: net/core: add NETDEV_BONDING_FAILOVER event Add NETDEV_BONDING_FAILOVER event to be used in a successive patch by bonding to announce fail-over for the active-backup mode through the netdev events notifier chain mechanism. Such an event can be of use for the RDMA CM (communication manager) to let native RDMA ULPs (eg NFS-RDMA, iSER) always be aligned with the IP stack, in the sense that they use the same ports/links as the stack does. More usages can be done to allow monitoring tools based on netlink events being aware to bonding fail-over. Signed-off-by: Or Gerlitz Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- include/linux/netdevice.h | 1 + include/linux/notifier.h | 1 + net/core/dev.c | 6 ++++++ 3 files changed, 8 insertions(+) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index f27fd2009334..e92fc839ab1d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1479,6 +1479,7 @@ extern void __dev_addr_unsync(struct dev_addr_list **to, int *to_count, struct extern void dev_set_promiscuity(struct net_device *dev, int inc); extern void dev_set_allmulti(struct net_device *dev, int inc); extern void netdev_state_change(struct net_device *dev); +extern void netdev_bonding_change(struct net_device *dev); extern void netdev_features_change(struct net_device *dev); /* Load a device via the kmod */ extern void dev_load(struct net *net, const char *name); diff --git a/include/linux/notifier.h b/include/linux/notifier.h index 0ff6224d172a..bd3d72ddf333 100644 --- a/include/linux/notifier.h +++ b/include/linux/notifier.h @@ -197,6 +197,7 @@ static inline int notifier_to_errno(int ret) #define NETDEV_GOING_DOWN 0x0009 #define NETDEV_CHANGENAME 0x000A #define NETDEV_FEAT_CHANGE 0x000B +#define NETDEV_BONDING_FAILOVER 0x000C #define SYS_DOWN 0x0001 /* Notify of system down */ #define SYS_RESTART SYS_DOWN diff --git a/net/core/dev.c b/net/core/dev.c index 68d8df0992ab..0e45742e7158 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -961,6 +961,12 @@ void netdev_state_change(struct net_device *dev) } } +void netdev_bonding_change(struct net_device *dev) +{ + call_netdevice_notifiers(NETDEV_BONDING_FAILOVER, dev); +} +EXPORT_SYMBOL(netdev_bonding_change); + /** * dev_load - load a network module * @net: the applicable net namespace -- cgit v1.3-8-gc7d7 From b8a9787eddb0e4665f31dd1d64584732b2b5d051 Mon Sep 17 00:00:00 2001 From: Jay Vosburgh Date: Fri, 13 Jun 2008 18:12:04 -0700 Subject: bonding: Allow setting max_bonds to zero Permit bonding to function rationally if max_bonds is set to zero. This will load the module, but create no master devices (which can be created via sysfs). Requires some change to bond_create_sysfs; currently, the netdev sysfs directory is determined from the first bonding device created, but this is no longer possible. Instead, an interface from net/core is created to create and destroy files in net_class. Based on a patch submitted by Phil Oester . Modified by Jay Vosburgh to fix the sysfs issue mentioned above and to update the documentation. Signed-off-by: Phil Oester Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- Documentation/networking/bonding.txt | 3 ++- drivers/net/bonding/bond_main.c | 6 +++--- drivers/net/bonding/bond_sysfs.c | 22 +++------------------- include/linux/netdevice.h | 3 +++ net/core/net-sysfs.c | 13 +++++++++++++ 5 files changed, 24 insertions(+), 23 deletions(-) (limited to 'net/core') diff --git a/Documentation/networking/bonding.txt b/Documentation/networking/bonding.txt index 370b7da73ab4..7fa7fe71d7a8 100644 --- a/Documentation/networking/bonding.txt +++ b/Documentation/networking/bonding.txt @@ -376,7 +376,8 @@ max_bonds Specifies the number of bonding devices to create for this instance of the bonding driver. E.g., if max_bonds is 3, and the bonding driver is not already loaded, then bond0, bond1 - and bond2 will be created. The default value is 1. + and bond2 will be created. The default value is 1. Specifying + a value of 0 will load bonding, but will not create any devices. miimon diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 3b6d66a8ab98..d57b65dc2c72 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -4750,11 +4750,11 @@ static int bond_check_params(struct bond_params *params) } } - if (max_bonds < 1 || max_bonds > INT_MAX) { + if (max_bonds < 0 || max_bonds > INT_MAX) { printk(KERN_WARNING DRV_NAME ": Warning: max_bonds (%d) not in range %d-%d, so it " "was reset to BOND_DEFAULT_MAX_BONDS (%d)\n", - max_bonds, 1, INT_MAX, BOND_DEFAULT_MAX_BONDS); + max_bonds, 0, INT_MAX, BOND_DEFAULT_MAX_BONDS); max_bonds = BOND_DEFAULT_MAX_BONDS; } @@ -4953,7 +4953,7 @@ static int bond_check_params(struct bond_params *params) printk("\n"); - } else { + } else if (max_bonds) { /* miimon and arp_interval not set, we need one so things * work as expected, see bonding.txt for details */ diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c index dd265c69b0df..6caac0ffb2f2 100644 --- a/drivers/net/bonding/bond_sysfs.c +++ b/drivers/net/bonding/bond_sysfs.c @@ -53,7 +53,6 @@ extern struct bond_parm_tbl arp_validate_tbl[]; extern struct bond_parm_tbl fail_over_mac_tbl[]; static int expected_refcount = -1; -static struct class *netdev_class; /*--------------------------- Data Structures -----------------------------*/ /* Bonding sysfs lock. Why can't we just use the subsystem lock? @@ -1447,19 +1446,9 @@ static struct attribute_group bonding_group = { */ int bond_create_sysfs(void) { - int ret = 0; - struct bonding *firstbond; - - /* get the netdev class pointer */ - firstbond = container_of(bond_dev_list.next, struct bonding, bond_list); - if (!firstbond) - return -ENODEV; - - netdev_class = firstbond->dev->dev.class; - if (!netdev_class) - return -ENODEV; + int ret; - ret = class_create_file(netdev_class, &class_attr_bonding_masters); + ret = netdev_class_create_file(&class_attr_bonding_masters); /* * Permit multiple loads of the module by ignoring failures to * create the bonding_masters sysfs file. Bonding devices @@ -1478,10 +1467,6 @@ int bond_create_sysfs(void) printk(KERN_ERR "network device named %s already exists in sysfs", class_attr_bonding_masters.attr.name); - else { - netdev_class = NULL; - return 0; - } } return ret; @@ -1493,8 +1478,7 @@ int bond_create_sysfs(void) */ void bond_destroy_sysfs(void) { - if (netdev_class) - class_remove_file(netdev_class, &class_attr_bonding_masters); + netdev_class_remove_file(&class_attr_bonding_masters); } /* diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index e92fc839ab1d..9ccbfac3fd95 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1506,6 +1506,9 @@ extern void *dev_seq_next(struct seq_file *seq, void *v, loff_t *pos); extern void dev_seq_stop(struct seq_file *seq, void *v); #endif +extern int netdev_class_create_file(struct class_attribute *class_attr); +extern void netdev_class_remove_file(struct class_attribute *class_attr); + extern void linkwatch_run_queue(void); extern int netdev_compute_features(unsigned long all, unsigned long one); diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index dccd737ea2e3..3f7941319217 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -468,6 +468,19 @@ int netdev_register_kobject(struct net_device *net) return device_add(dev); } +int netdev_class_create_file(struct class_attribute *class_attr) +{ + return class_create_file(&net_class, class_attr); +} + +void netdev_class_remove_file(struct class_attribute *class_attr) +{ + class_remove_file(&net_class, class_attr); +} + +EXPORT_SYMBOL(netdev_class_create_file); +EXPORT_SYMBOL(netdev_class_remove_file); + void netdev_initialize_kobject(struct net_device *net) { struct device *device = &(net->dev); -- cgit v1.3-8-gc7d7 From 972692e0db9b0a62329ca394062b58917ddbd03c Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 22:41:38 -0700 Subject: net: Add sk_set_socket() helper. In order to more easily grep for all things that set sk->sk_socket, add sk_set_socket() helper inline function. Suggested (although only half-seriously) by Evgeniy Polyakov. Signed-off-by: David S. Miller --- include/net/sock.h | 9 +++++++-- net/core/sock.c | 4 ++-- 2 files changed, 9 insertions(+), 4 deletions(-) (limited to 'net/core') diff --git a/include/net/sock.h b/include/net/sock.h index a7c30412de66..3f4897ab432e 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -990,6 +990,11 @@ static inline void sock_put(struct sock *sk) extern int sk_receive_skb(struct sock *sk, struct sk_buff *skb, const int nested); +static inline void sk_set_socket(struct sock *sk, struct socket *sock) +{ + sk->sk_socket = sock; +} + /* Detach socket from process context. * Announce socket dead, detach it from wait queue and inode. * Note that parent inode held reference count on this struct sock, @@ -1001,7 +1006,7 @@ static inline void sock_orphan(struct sock *sk) { write_lock_bh(&sk->sk_callback_lock); sock_set_flag(sk, SOCK_DEAD); - sk->sk_socket = NULL; + sk_set_socket(sk, NULL); sk->sk_sleep = NULL; write_unlock_bh(&sk->sk_callback_lock); } @@ -1011,7 +1016,7 @@ static inline void sock_graft(struct sock *sk, struct socket *parent) write_lock_bh(&sk->sk_callback_lock); sk->sk_sleep = &parent->wait; parent->sk = sk; - sk->sk_socket = parent; + sk_set_socket(sk, parent); security_sock_graft(sk, parent); write_unlock_bh(&sk->sk_callback_lock); } diff --git a/net/core/sock.c b/net/core/sock.c index 3879bf65897e..2c0ba52e5303 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -1066,7 +1066,7 @@ struct sock *sk_clone(const struct sock *sk, const gfp_t priority) * to be taken into account in all callers. -acme */ sk_refcnt_debug_inc(newsk); - newsk->sk_socket = NULL; + sk_set_socket(newsk, NULL); newsk->sk_sleep = NULL; if (newsk->sk_prot->sockets_allocated) @@ -1702,7 +1702,7 @@ void sock_init_data(struct socket *sock, struct sock *sk) sk->sk_rcvbuf = sysctl_rmem_default; sk->sk_sndbuf = sysctl_wmem_default; sk->sk_state = TCP_CLOSE; - sk->sk_socket = sock; + sk_set_socket(sk, sock); sock_set_flag(sk, SOCK_ZAPPED); -- cgit v1.3-8-gc7d7 From dad9b335c6940de2746a9788eb456d09cf102f81 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Wed, 18 Jun 2008 01:48:28 -0700 Subject: netdevice: Fix promiscuity and allmulti overflow Max of promiscuity and allmulti plus positive @inc can cause overflow. Fox example: when allmulti=0xFFFFFFFF, any caller give dev_set_allmulti() a positive @inc will cause allmulti be off. This is not what we want, though it's rare case. The fix is that only negative @inc will cause allmulti or promiscuity be off and when any caller makes the counters touch the roof, we return error. Change of v2: Change void function dev_set_promiscuity/allmulti to return int. So callers can get the overflow error. Caller's fix will be done later. Change of v3: 1. Since we return error to caller, we don't need to print KERN_ERROR, KERN_WARNING is enough. 2. In dev_set_promiscuity(), if __dev_set_promiscuity() failed, we return at once. Signed-off-by: Wang Chen Acked-by: Patrick McHardy Signed-off-by: David S. Miller --- include/linux/netdevice.h | 4 ++-- net/core/dev.c | 55 ++++++++++++++++++++++++++++++++++++++--------- 2 files changed, 47 insertions(+), 12 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 4bf613cd9e2d..45dce2b58d4c 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1476,8 +1476,8 @@ extern int __dev_addr_delete(struct dev_addr_list **list, int *count, void *ad extern int __dev_addr_add(struct dev_addr_list **list, int *count, void *addr, int alen, int newonly); extern int __dev_addr_sync(struct dev_addr_list **to, int *to_count, struct dev_addr_list **from, int *from_count); extern void __dev_addr_unsync(struct dev_addr_list **to, int *to_count, struct dev_addr_list **from, int *from_count); -extern void dev_set_promiscuity(struct net_device *dev, int inc); -extern void dev_set_allmulti(struct net_device *dev, int inc); +extern int dev_set_promiscuity(struct net_device *dev, int inc); +extern int dev_set_allmulti(struct net_device *dev, int inc); extern void netdev_state_change(struct net_device *dev); extern void netdev_bonding_change(struct net_device *dev); extern void netdev_features_change(struct net_device *dev); diff --git a/net/core/dev.c b/net/core/dev.c index 0e45742e7158..a495f712d38c 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2771,16 +2771,29 @@ int netdev_set_master(struct net_device *slave, struct net_device *master) return 0; } -static void __dev_set_promiscuity(struct net_device *dev, int inc) +static int __dev_set_promiscuity(struct net_device *dev, int inc) { unsigned short old_flags = dev->flags; ASSERT_RTNL(); - if ((dev->promiscuity += inc) == 0) - dev->flags &= ~IFF_PROMISC; - else - dev->flags |= IFF_PROMISC; + dev->flags |= IFF_PROMISC; + dev->promiscuity += inc; + if (dev->promiscuity == 0) { + /* + * Avoid overflow. + * If inc causes overflow, untouch promisc and return error. + */ + if (inc < 0) + dev->flags &= ~IFF_PROMISC; + else { + dev->promiscuity -= inc; + printk(KERN_WARNING "%s: promiscuity touches roof, " + "set promiscuity failed, promiscuity feature " + "of device might be broken.\n", dev->name); + return -EOVERFLOW; + } + } if (dev->flags != old_flags) { printk(KERN_INFO "device %s %s promiscuous mode\n", dev->name, (dev->flags & IFF_PROMISC) ? "entered" : @@ -2798,6 +2811,7 @@ static void __dev_set_promiscuity(struct net_device *dev, int inc) if (dev->change_rx_flags) dev->change_rx_flags(dev, IFF_PROMISC); } + return 0; } /** @@ -2809,14 +2823,19 @@ static void __dev_set_promiscuity(struct net_device *dev, int inc) * remains above zero the interface remains promiscuous. Once it hits zero * the device reverts back to normal filtering operation. A negative inc * value is used to drop promiscuity on the device. + * Return 0 if successful or a negative errno code on error. */ -void dev_set_promiscuity(struct net_device *dev, int inc) +int dev_set_promiscuity(struct net_device *dev, int inc) { unsigned short old_flags = dev->flags; + int err; - __dev_set_promiscuity(dev, inc); + err = __dev_set_promiscuity(dev, inc); + if (!err) + return err; if (dev->flags != old_flags) dev_set_rx_mode(dev); + return err; } /** @@ -2829,22 +2848,38 @@ void dev_set_promiscuity(struct net_device *dev, int inc) * to all interfaces. Once it hits zero the device reverts back to normal * filtering operation. A negative @inc value is used to drop the counter * when releasing a resource needing all multicasts. + * Return 0 if successful or a negative errno code on error. */ -void dev_set_allmulti(struct net_device *dev, int inc) +int dev_set_allmulti(struct net_device *dev, int inc) { unsigned short old_flags = dev->flags; ASSERT_RTNL(); dev->flags |= IFF_ALLMULTI; - if ((dev->allmulti += inc) == 0) - dev->flags &= ~IFF_ALLMULTI; + dev->allmulti += inc; + if (dev->allmulti == 0) { + /* + * Avoid overflow. + * If inc causes overflow, untouch allmulti and return error. + */ + if (inc < 0) + dev->flags &= ~IFF_ALLMULTI; + else { + dev->allmulti -= inc; + printk(KERN_WARNING "%s: allmulti touches roof, " + "set allmulti failed, allmulti feature of " + "device might be broken.\n", dev->name); + return -EOVERFLOW; + } + } if (dev->flags ^ old_flags) { if (dev->change_rx_flags) dev->change_rx_flags(dev, IFF_ALLMULTI); dev_set_rx_mode(dev); } + return 0; } /* -- cgit v1.3-8-gc7d7 From 0187bdfb05674147774ca79a79942537f3ad54bd Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Thu, 19 Jun 2008 16:15:47 -0700 Subject: net: Disable LRO on devices that are forwarding Large Receive Offload (LRO) is only appropriate for packets that are destined for the host, and should be disabled if received packets may be forwarded. It can also confuse the GSO on output. Add dev_disable_lro() function which uses the appropriate ethtool ops to disable LRO if enabled. Add calls to dev_disable_lro() in br_add_if() and functions that enable IPv4 and IPv6 forwarding. Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- include/linux/netdevice.h | 1 + net/bridge/br_if.c | 1 + net/core/dev.c | 24 ++++++++++++++++++++++++ net/ipv4/devinet.c | 21 ++++++++++++++++----- net/ipv6/addrconf.c | 6 ++++++ 5 files changed, 48 insertions(+), 5 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 45dce2b58d4c..1304ad2d7105 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -886,6 +886,7 @@ extern struct net_device *__dev_get_by_name(struct net *net, const char *name); extern int dev_alloc_name(struct net_device *dev, const char *name); extern int dev_open(struct net_device *dev); extern int dev_close(struct net_device *dev); +extern void dev_disable_lro(struct net_device *dev); extern int dev_queue_xmit(struct sk_buff *skb); extern int register_netdevice(struct net_device *dev); extern void unregister_netdevice(struct net_device *dev); diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index 143c954681b8..832a561500d9 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -387,6 +387,7 @@ int br_add_if(struct net_bridge *br, struct net_device *dev) goto err2; rcu_assign_pointer(dev->br_port, p); + dev_disable_lro(dev); dev_set_promiscuity(dev, 1); list_add_rcu(&p->list, &br->port_list); diff --git a/net/core/dev.c b/net/core/dev.c index a495f712d38c..f6944ecd5b2e 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -90,6 +90,7 @@ #include #include #include +#include #include #include #include @@ -1123,6 +1124,29 @@ int dev_close(struct net_device *dev) } +/** + * dev_disable_lro - disable Large Receive Offload on a device + * @dev: device + * + * Disable Large Receive Offload (LRO) on a net device. Must be + * called under RTNL. This is needed if received packets may be + * forwarded to another interface. + */ +void dev_disable_lro(struct net_device *dev) +{ + if (dev->ethtool_ops && dev->ethtool_ops->get_flags && + dev->ethtool_ops->set_flags) { + u32 flags = dev->ethtool_ops->get_flags(dev); + if (flags & ETH_FLAG_LRO) { + flags &= ~ETH_FLAG_LRO; + dev->ethtool_ops->set_flags(dev, flags); + } + } + WARN_ON(dev->features & NETIF_F_LRO); +} +EXPORT_SYMBOL(dev_disable_lro); + + static int dev_boot_phase = 1; /* diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index f8c0b0aea93a..9de2514946ca 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -168,6 +168,8 @@ static struct in_device *inetdev_init(struct net_device *dev) in_dev->dev = dev; if ((in_dev->arp_parms = neigh_parms_alloc(dev, &arp_tbl)) == NULL) goto out_kfree; + if (IPV4_DEVCONF(in_dev->cnf, FORWARDING)) + dev_disable_lro(dev); /* Reference in_dev->dev */ dev_hold(dev); /* Account for reference dev->ip_ptr (below) */ @@ -1241,6 +1243,8 @@ static void inet_forward_change(struct net *net) read_lock(&dev_base_lock); for_each_netdev(net, dev) { struct in_device *in_dev; + if (on) + dev_disable_lro(dev); rcu_read_lock(); in_dev = __in_dev_get_rcu(dev); if (in_dev) @@ -1248,8 +1252,6 @@ static void inet_forward_change(struct net *net) rcu_read_unlock(); } read_unlock(&dev_base_lock); - - rt_cache_flush(0); } static int devinet_conf_proc(ctl_table *ctl, int write, @@ -1335,10 +1337,19 @@ static int devinet_sysctl_forward(ctl_table *ctl, int write, if (write && *valp != val) { struct net *net = ctl->extra2; - if (valp == &IPV4_DEVCONF_ALL(net, FORWARDING)) - inet_forward_change(net); - else if (valp != &IPV4_DEVCONF_DFLT(net, FORWARDING)) + if (valp != &IPV4_DEVCONF_DFLT(net, FORWARDING)) { + rtnl_lock(); + if (valp == &IPV4_DEVCONF_ALL(net, FORWARDING)) { + inet_forward_change(net); + } else if (*valp) { + struct ipv4_devconf *cnf = ctl->extra1; + struct in_device *idev = + container_of(cnf, struct in_device, cnf); + dev_disable_lro(idev->dev); + } + rtnl_unlock(); rt_cache_flush(0); + } } return ret; diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 9be6be3a7ff3..84127d854cfc 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -348,6 +348,8 @@ static struct inet6_dev * ipv6_add_dev(struct net_device *dev) kfree(ndev); return NULL; } + if (ndev->cnf.forwarding) + dev_disable_lro(dev); /* We refer to the device */ dev_hold(dev); @@ -442,6 +444,8 @@ static void dev_forward_change(struct inet6_dev *idev) if (!idev) return; dev = idev->dev; + if (idev->cnf.forwarding) + dev_disable_lro(dev); if (dev && (dev->flags & IFF_MULTICAST)) { if (idev->cnf.forwarding) ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters); @@ -487,12 +491,14 @@ static void addrconf_fixup_forwarding(struct ctl_table *table, int *p, int old) if (p == &net->ipv6.devconf_dflt->forwarding) return; + rtnl_lock(); if (p == &net->ipv6.devconf_all->forwarding) { __s32 newf = net->ipv6.devconf_all->forwarding; net->ipv6.devconf_dflt->forwarding = newf; addrconf_forward_change(net, newf); } else if ((!*p) ^ (!old)) dev_forward_change((struct inet6_dev *)table->extra1); + rtnl_unlock(); if (*p) rt6_purge_dflt_routers(net); -- cgit v1.3-8-gc7d7 From 4497b0763cb1afae463f5e144c28b5d806e28b60 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Thu, 19 Jun 2008 16:22:28 -0700 Subject: net: Discard and warn about LRO'd skbs received for forwarding Add skb_warn_if_lro() to test whether an skb was received with LRO and warn if so. Change br_forward(), ip_forward() and ip6_forward() to call it) and discard the skb if it returns true. Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- include/linux/skbuff.h | 14 ++++++++++++++ net/bridge/br_forward.c | 2 +- net/core/skbuff.c | 8 ++++++++ net/ipv4/ip_forward.c | 3 +++ net/ipv6/ip6_output.c | 3 +++ 5 files changed, 29 insertions(+), 1 deletion(-) (limited to 'net/core') diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 299ec4b31412..2220b9e2dab0 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -1702,6 +1702,20 @@ static inline int skb_is_gso_v6(const struct sk_buff *skb) return skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6; } +extern void __skb_warn_lro_forwarding(const struct sk_buff *skb); + +static inline bool skb_warn_if_lro(const struct sk_buff *skb) +{ + /* LRO sets gso_size but not gso_type, whereas if GSO is really + * wanted then gso_type will be set. */ + struct skb_shared_info *shinfo = skb_shinfo(skb); + if (shinfo->gso_size != 0 && unlikely(shinfo->gso_type == 0)) { + __skb_warn_lro_forwarding(skb); + return true; + } + return false; +} + static inline void skb_forward_csum(struct sk_buff *skb) { /* Unfortunately we don't support this one. Any brave souls? */ diff --git a/net/bridge/br_forward.c b/net/bridge/br_forward.c index 512645727f51..bdd9ccea17ce 100644 --- a/net/bridge/br_forward.c +++ b/net/bridge/br_forward.c @@ -89,7 +89,7 @@ void br_deliver(const struct net_bridge_port *to, struct sk_buff *skb) /* called with rcu_read_lock */ void br_forward(const struct net_bridge_port *to, struct sk_buff *skb) { - if (should_deliver(to, skb)) { + if (!skb_warn_if_lro(skb) && should_deliver(to, skb)) { __br_forward(to, skb); return; } diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 3e18f8525e82..2df012be973d 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2583,6 +2583,13 @@ bool skb_partial_csum_set(struct sk_buff *skb, u16 start, u16 off) return true; } +void __skb_warn_lro_forwarding(const struct sk_buff *skb) +{ + if (net_ratelimit()) + pr_warning("%s: received packets cannot be forwarded" + " while LRO is enabled\n", skb->dev->name); +} + EXPORT_SYMBOL(___pskb_trim); EXPORT_SYMBOL(__kfree_skb); EXPORT_SYMBOL(kfree_skb); @@ -2616,6 +2623,7 @@ EXPORT_SYMBOL(skb_seq_read); EXPORT_SYMBOL(skb_abort_seq_read); EXPORT_SYMBOL(skb_find_text); EXPORT_SYMBOL(skb_append_datato_frags); +EXPORT_SYMBOL(__skb_warn_lro_forwarding); EXPORT_SYMBOL_GPL(skb_to_sgvec); EXPORT_SYMBOL_GPL(skb_cow_data); diff --git a/net/ipv4/ip_forward.c b/net/ipv4/ip_forward.c index 37d36a3f33cd..da14725916d3 100644 --- a/net/ipv4/ip_forward.c +++ b/net/ipv4/ip_forward.c @@ -56,6 +56,9 @@ int ip_forward(struct sk_buff *skb) struct rtable *rt; /* Route we use */ struct ip_options * opt = &(IPCB(skb)->opt); + if (skb_warn_if_lro(skb)) + goto drop; + if (!xfrm4_policy_check(NULL, XFRM_POLICY_FWD, skb)) goto drop; diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 40a2813a63d1..fd7cd1bfe151 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -407,6 +407,9 @@ int ip6_forward(struct sk_buff *skb) if (ipv6_devconf.forwarding == 0) goto error; + if (skb_warn_if_lro(skb)) + goto drop; + if (!xfrm6_policy_check(NULL, XFRM_POLICY_FWD, skb)) { IP6_INC_STATS(ip6_dst_idev(dst), IPSTATS_MIB_INDISCARDS); goto drop; -- cgit v1.3-8-gc7d7 From 8691e5a8f691cc2a4fda0651e8d307aaba0e7d68 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Fri, 6 Jun 2008 11:18:06 +0200 Subject: smp_call_function: get rid of the unused nonatomic/retry argument It's never used and the comments refer to nonatomic and retry interchangably. So get rid of it. Acked-by: Jeremy Fitzhardinge Signed-off-by: Jens Axboe --- arch/alpha/kernel/core_marvel.c | 2 +- arch/alpha/kernel/smp.c | 6 +++--- arch/alpha/oprofile/common.c | 6 +++--- arch/arm/oprofile/op_model_mpcore.c | 2 +- arch/arm/vfp/vfpmodule.c | 2 +- arch/cris/arch-v32/kernel/smp.c | 5 ++--- arch/ia64/kernel/mca.c | 2 +- arch/ia64/kernel/palinfo.c | 2 +- arch/ia64/kernel/perfmon.c | 2 +- arch/ia64/kernel/process.c | 2 +- arch/ia64/kernel/smpboot.c | 2 +- arch/ia64/kernel/uncached.c | 5 ++--- arch/ia64/sn/kernel/sn2/sn_hwperf.c | 2 +- arch/m32r/kernel/smp.c | 4 ++-- arch/mips/kernel/smp.c | 4 ++-- arch/mips/mm/c-r4k.c | 18 +++++++++--------- arch/mips/pmc-sierra/yosemite/prom.c | 2 +- arch/mips/sibyte/cfe/setup.c | 2 +- arch/mips/sibyte/sb1250/prom.c | 2 +- arch/powerpc/kernel/smp.c | 2 +- arch/s390/appldata/appldata_base.c | 4 ++-- arch/s390/kernel/smp.c | 16 ++++++---------- arch/s390/kernel/time.c | 4 ++-- arch/sh/kernel/smp.c | 10 +++++----- arch/sparc64/kernel/smp.c | 12 ++++-------- arch/um/kernel/smp.c | 3 +-- arch/x86/kernel/cpu/mtrr/main.c | 4 ++-- arch/x86/kernel/cpuid.c | 2 +- arch/x86/kernel/ldt.c | 2 +- arch/x86/kernel/nmi_32.c | 2 +- arch/x86/kernel/nmi_64.c | 2 +- arch/x86/kernel/smp.c | 2 +- arch/x86/kernel/vsyscall_64.c | 2 +- arch/x86/kvm/vmx.c | 2 +- arch/x86/kvm/x86.c | 2 +- arch/x86/lib/msr-on-cpu.c | 8 ++++---- arch/x86/mach-voyager/voyager_smp.c | 2 +- arch/x86/xen/smp.c | 2 +- drivers/acpi/processor_idle.c | 2 +- drivers/cpuidle/cpuidle.c | 2 +- include/asm-alpha/smp.h | 2 +- include/asm-sparc/smp.h | 2 +- include/linux/smp.h | 8 ++++---- kernel/smp.c | 6 ++---- kernel/softirq.c | 2 +- kernel/time/tick-broadcast.c | 2 +- net/core/flow.c | 2 +- net/iucv/iucv.c | 14 +++++++------- virt/kvm/kvm_main.c | 6 +++--- 49 files changed, 95 insertions(+), 108 deletions(-) (limited to 'net/core') diff --git a/arch/alpha/kernel/core_marvel.c b/arch/alpha/kernel/core_marvel.c index ced4aae8b804..04dcc5e5d4c1 100644 --- a/arch/alpha/kernel/core_marvel.c +++ b/arch/alpha/kernel/core_marvel.c @@ -662,7 +662,7 @@ __marvel_rtc_io(u8 b, unsigned long addr, int write) if (smp_processor_id() != boot_cpuid) smp_call_function_single(boot_cpuid, __marvel_access_rtc, - &rtc_access, 1, 1); + &rtc_access, 1); else __marvel_access_rtc(&rtc_access); #else diff --git a/arch/alpha/kernel/smp.c b/arch/alpha/kernel/smp.c index 95c905be9154..44114c8dbb2a 100644 --- a/arch/alpha/kernel/smp.c +++ b/arch/alpha/kernel/smp.c @@ -710,7 +710,7 @@ flush_tlb_mm(struct mm_struct *mm) } } - if (smp_call_function(ipi_flush_tlb_mm, mm, 1, 1)) { + if (smp_call_function(ipi_flush_tlb_mm, mm, 1)) { printk(KERN_CRIT "flush_tlb_mm: timed out\n"); } @@ -763,7 +763,7 @@ flush_tlb_page(struct vm_area_struct *vma, unsigned long addr) data.mm = mm; data.addr = addr; - if (smp_call_function(ipi_flush_tlb_page, &data, 1, 1)) { + if (smp_call_function(ipi_flush_tlb_page, &data, 1)) { printk(KERN_CRIT "flush_tlb_page: timed out\n"); } @@ -815,7 +815,7 @@ flush_icache_user_range(struct vm_area_struct *vma, struct page *page, } } - if (smp_call_function(ipi_flush_icache_page, mm, 1, 1)) { + if (smp_call_function(ipi_flush_icache_page, mm, 1)) { printk(KERN_CRIT "flush_icache_page: timed out\n"); } diff --git a/arch/alpha/oprofile/common.c b/arch/alpha/oprofile/common.c index 9fc0eeb4f0ab..7c3d5ec6ec67 100644 --- a/arch/alpha/oprofile/common.c +++ b/arch/alpha/oprofile/common.c @@ -65,7 +65,7 @@ op_axp_setup(void) model->reg_setup(®, ctr, &sys); /* Configure the registers on all cpus. */ - (void)smp_call_function(model->cpu_setup, ®, 0, 1); + (void)smp_call_function(model->cpu_setup, ®, 1); model->cpu_setup(®); return 0; } @@ -86,7 +86,7 @@ op_axp_cpu_start(void *dummy) static int op_axp_start(void) { - (void)smp_call_function(op_axp_cpu_start, NULL, 0, 1); + (void)smp_call_function(op_axp_cpu_start, NULL, 1); op_axp_cpu_start(NULL); return 0; } @@ -101,7 +101,7 @@ op_axp_cpu_stop(void *dummy) static void op_axp_stop(void) { - (void)smp_call_function(op_axp_cpu_stop, NULL, 0, 1); + (void)smp_call_function(op_axp_cpu_stop, NULL, 1); op_axp_cpu_stop(NULL); } diff --git a/arch/arm/oprofile/op_model_mpcore.c b/arch/arm/oprofile/op_model_mpcore.c index 74fae6045650..4458705021e0 100644 --- a/arch/arm/oprofile/op_model_mpcore.c +++ b/arch/arm/oprofile/op_model_mpcore.c @@ -201,7 +201,7 @@ static int em_call_function(int (*fn)(void)) data.ret = 0; preempt_disable(); - smp_call_function(em_func, &data, 1, 1); + smp_call_function(em_func, &data, 1); em_func(&data); preempt_enable(); diff --git a/arch/arm/vfp/vfpmodule.c b/arch/arm/vfp/vfpmodule.c index 32455c633f1c..c0d2c9bb952b 100644 --- a/arch/arm/vfp/vfpmodule.c +++ b/arch/arm/vfp/vfpmodule.c @@ -352,7 +352,7 @@ static int __init vfp_init(void) else if (vfpsid & FPSID_NODOUBLE) { printk("no double precision support\n"); } else { - smp_call_function(vfp_enable, NULL, 1, 1); + smp_call_function(vfp_enable, NULL, 1); VFP_arch = (vfpsid & FPSID_ARCH_MASK) >> FPSID_ARCH_BIT; /* Extract the architecture version */ printk("implementor %02x architecture %d part %02x variant %x rev %x\n", diff --git a/arch/cris/arch-v32/kernel/smp.c b/arch/cris/arch-v32/kernel/smp.c index a9c3334e46c9..952a24b2f5a9 100644 --- a/arch/cris/arch-v32/kernel/smp.c +++ b/arch/cris/arch-v32/kernel/smp.c @@ -194,7 +194,7 @@ void stop_this_cpu(void* dummy) /* Other calls */ void smp_send_stop(void) { - smp_call_function(stop_this_cpu, NULL, 1, 0); + smp_call_function(stop_this_cpu, NULL, 0); } int setup_profiling_timer(unsigned int multiplier) @@ -316,8 +316,7 @@ int send_ipi(int vector, int wait, cpumask_t cpu_mask) * You must not call this function with disabled interrupts or from a * hardware interrupt handler or from a bottom half handler. */ -int smp_call_function(void (*func)(void *info), void *info, - int nonatomic, int wait) +int smp_call_function(void (*func)(void *info), void *info, int wait) { cpumask_t cpu_mask = CPU_MASK_ALL; struct call_data_struct data; diff --git a/arch/ia64/kernel/mca.c b/arch/ia64/kernel/mca.c index 705176b434b3..9cd818cc7008 100644 --- a/arch/ia64/kernel/mca.c +++ b/arch/ia64/kernel/mca.c @@ -1881,7 +1881,7 @@ static int __cpuinit mca_cpu_callback(struct notifier_block *nfb, case CPU_ONLINE: case CPU_ONLINE_FROZEN: smp_call_function_single(hotcpu, ia64_mca_cmc_vector_adjust, - NULL, 1, 0); + NULL, 0); break; } return NOTIFY_OK; diff --git a/arch/ia64/kernel/palinfo.c b/arch/ia64/kernel/palinfo.c index 9dc00f7fe10e..e5c57f413ca2 100644 --- a/arch/ia64/kernel/palinfo.c +++ b/arch/ia64/kernel/palinfo.c @@ -921,7 +921,7 @@ int palinfo_handle_smp(pal_func_cpu_u_t *f, char *page) /* will send IPI to other CPU and wait for completion of remote call */ - if ((ret=smp_call_function_single(f->req_cpu, palinfo_smp_call, &ptr, 0, 1))) { + if ((ret=smp_call_function_single(f->req_cpu, palinfo_smp_call, &ptr, 1))) { printk(KERN_ERR "palinfo: remote CPU call from %d to %d on function %d: " "error %d\n", smp_processor_id(), f->req_cpu, f->func_id, ret); return 0; diff --git a/arch/ia64/kernel/perfmon.c b/arch/ia64/kernel/perfmon.c index 7714a97b0104..9baa48255c12 100644 --- a/arch/ia64/kernel/perfmon.c +++ b/arch/ia64/kernel/perfmon.c @@ -1820,7 +1820,7 @@ pfm_syswide_cleanup_other_cpu(pfm_context_t *ctx) int ret; DPRINT(("calling CPU%d for cleanup\n", ctx->ctx_cpu)); - ret = smp_call_function_single(ctx->ctx_cpu, pfm_syswide_force_stop, ctx, 0, 1); + ret = smp_call_function_single(ctx->ctx_cpu, pfm_syswide_force_stop, ctx, 1); DPRINT(("called CPU%d for cleanup ret=%d\n", ctx->ctx_cpu, ret)); } #endif /* CONFIG_SMP */ diff --git a/arch/ia64/kernel/process.c b/arch/ia64/kernel/process.c index a3a34b4eb038..fabaf08d9a69 100644 --- a/arch/ia64/kernel/process.c +++ b/arch/ia64/kernel/process.c @@ -286,7 +286,7 @@ void cpu_idle_wait(void) { smp_mb(); /* kick all the CPUs so that they exit out of pm_idle */ - smp_call_function(do_nothing, NULL, 0, 1); + smp_call_function(do_nothing, NULL, 1); } EXPORT_SYMBOL_GPL(cpu_idle_wait); diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c index eaa1b6795a13..9d1d429c6c59 100644 --- a/arch/ia64/kernel/smpboot.c +++ b/arch/ia64/kernel/smpboot.c @@ -317,7 +317,7 @@ ia64_sync_itc (unsigned int master) go[MASTER] = 1; - if (smp_call_function_single(master, sync_master, NULL, 1, 0) < 0) { + if (smp_call_function_single(master, sync_master, NULL, 0) < 0) { printk(KERN_ERR "sync_itc: failed to get attention of CPU %u!\n", master); return; } diff --git a/arch/ia64/kernel/uncached.c b/arch/ia64/kernel/uncached.c index e77995a6e3ed..8eff8c1d40a6 100644 --- a/arch/ia64/kernel/uncached.c +++ b/arch/ia64/kernel/uncached.c @@ -123,8 +123,7 @@ static int uncached_add_chunk(struct uncached_pool *uc_pool, int nid) status = ia64_pal_prefetch_visibility(PAL_VISIBILITY_PHYSICAL); if (status == PAL_VISIBILITY_OK_REMOTE_NEEDED) { atomic_set(&uc_pool->status, 0); - status = smp_call_function(uncached_ipi_visibility, uc_pool, - 0, 1); + status = smp_call_function(uncached_ipi_visibility, uc_pool, 1); if (status || atomic_read(&uc_pool->status)) goto failed; } else if (status != PAL_VISIBILITY_OK) @@ -146,7 +145,7 @@ static int uncached_add_chunk(struct uncached_pool *uc_pool, int nid) if (status != PAL_STATUS_SUCCESS) goto failed; atomic_set(&uc_pool->status, 0); - status = smp_call_function(uncached_ipi_mc_drain, uc_pool, 0, 1); + status = smp_call_function(uncached_ipi_mc_drain, uc_pool, 1); if (status || atomic_read(&uc_pool->status)) goto failed; diff --git a/arch/ia64/sn/kernel/sn2/sn_hwperf.c b/arch/ia64/sn/kernel/sn2/sn_hwperf.c index 8cc0c4753d89..636588e7e068 100644 --- a/arch/ia64/sn/kernel/sn2/sn_hwperf.c +++ b/arch/ia64/sn/kernel/sn2/sn_hwperf.c @@ -629,7 +629,7 @@ static int sn_hwperf_op_cpu(struct sn_hwperf_op_info *op_info) if (use_ipi) { /* use an interprocessor interrupt to call SAL */ smp_call_function_single(cpu, sn_hwperf_call_sal, - op_info, 1, 1); + op_info, 1); } else { /* migrate the task before calling SAL */ diff --git a/arch/m32r/kernel/smp.c b/arch/m32r/kernel/smp.c index 74eb7bcd5a40..7577f971ea4e 100644 --- a/arch/m32r/kernel/smp.c +++ b/arch/m32r/kernel/smp.c @@ -212,7 +212,7 @@ void smp_flush_tlb_all(void) local_irq_save(flags); __flush_tlb_all(); local_irq_restore(flags); - smp_call_function(flush_tlb_all_ipi, NULL, 1, 1); + smp_call_function(flush_tlb_all_ipi, NULL, 1); preempt_enable(); } @@ -505,7 +505,7 @@ void smp_invalidate_interrupt(void) *==========================================================================*/ void smp_send_stop(void) { - smp_call_function(stop_this_cpu, NULL, 1, 0); + smp_call_function(stop_this_cpu, NULL, 0); } /*==========================================================================* diff --git a/arch/mips/kernel/smp.c b/arch/mips/kernel/smp.c index c75b26cb61df..7a9ae830be86 100644 --- a/arch/mips/kernel/smp.c +++ b/arch/mips/kernel/smp.c @@ -167,7 +167,7 @@ static void stop_this_cpu(void *dummy) void smp_send_stop(void) { - smp_call_function(stop_this_cpu, NULL, 1, 0); + smp_call_function(stop_this_cpu, NULL, 0); } void __init smp_cpus_done(unsigned int max_cpus) @@ -266,7 +266,7 @@ static void flush_tlb_mm_ipi(void *mm) static inline void smp_on_other_tlbs(void (*func) (void *info), void *info) { #ifndef CONFIG_MIPS_MT_SMTC - smp_call_function(func, info, 1, 1); + smp_call_function(func, info, 1); #endif } diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c index 27096751ddce..71df3390c07b 100644 --- a/arch/mips/mm/c-r4k.c +++ b/arch/mips/mm/c-r4k.c @@ -43,12 +43,12 @@ * primary cache. */ static inline void r4k_on_each_cpu(void (*func) (void *info), void *info, - int retry, int wait) + int wait) { preempt_disable(); #if !defined(CONFIG_MIPS_MT_SMP) && !defined(CONFIG_MIPS_MT_SMTC) - smp_call_function(func, info, retry, wait); + smp_call_function(func, info, wait); #endif func(info); preempt_enable(); @@ -350,7 +350,7 @@ static inline void local_r4k___flush_cache_all(void * args) static void r4k___flush_cache_all(void) { - r4k_on_each_cpu(local_r4k___flush_cache_all, NULL, 1, 1); + r4k_on_each_cpu(local_r4k___flush_cache_all, NULL, 1); } static inline int has_valid_asid(const struct mm_struct *mm) @@ -397,7 +397,7 @@ static void r4k_flush_cache_range(struct vm_area_struct *vma, int exec = vma->vm_flags & VM_EXEC; if (cpu_has_dc_aliases || (exec && !cpu_has_ic_fills_f_dc)) - r4k_on_each_cpu(local_r4k_flush_cache_range, vma, 1, 1); + r4k_on_each_cpu(local_r4k_flush_cache_range, vma, 1); } static inline void local_r4k_flush_cache_mm(void * args) @@ -429,7 +429,7 @@ static void r4k_flush_cache_mm(struct mm_struct *mm) if (!cpu_has_dc_aliases) return; - r4k_on_each_cpu(local_r4k_flush_cache_mm, mm, 1, 1); + r4k_on_each_cpu(local_r4k_flush_cache_mm, mm, 1); } struct flush_cache_page_args { @@ -521,7 +521,7 @@ static void r4k_flush_cache_page(struct vm_area_struct *vma, args.addr = addr; args.pfn = pfn; - r4k_on_each_cpu(local_r4k_flush_cache_page, &args, 1, 1); + r4k_on_each_cpu(local_r4k_flush_cache_page, &args, 1); } static inline void local_r4k_flush_data_cache_page(void * addr) @@ -535,7 +535,7 @@ static void r4k_flush_data_cache_page(unsigned long addr) local_r4k_flush_data_cache_page((void *)addr); else r4k_on_each_cpu(local_r4k_flush_data_cache_page, (void *) addr, - 1, 1); + 1); } struct flush_icache_range_args { @@ -571,7 +571,7 @@ static void r4k_flush_icache_range(unsigned long start, unsigned long end) args.start = start; args.end = end; - r4k_on_each_cpu(local_r4k_flush_icache_range, &args, 1, 1); + r4k_on_each_cpu(local_r4k_flush_icache_range, &args, 1); instruction_hazard(); } @@ -672,7 +672,7 @@ static void local_r4k_flush_cache_sigtramp(void * arg) static void r4k_flush_cache_sigtramp(unsigned long addr) { - r4k_on_each_cpu(local_r4k_flush_cache_sigtramp, (void *) addr, 1, 1); + r4k_on_each_cpu(local_r4k_flush_cache_sigtramp, (void *) addr, 1); } static void r4k_flush_icache_all(void) diff --git a/arch/mips/pmc-sierra/yosemite/prom.c b/arch/mips/pmc-sierra/yosemite/prom.c index 35dc435846a6..cf4c868715ac 100644 --- a/arch/mips/pmc-sierra/yosemite/prom.c +++ b/arch/mips/pmc-sierra/yosemite/prom.c @@ -64,7 +64,7 @@ static void prom_exit(void) #ifdef CONFIG_SMP if (smp_processor_id()) /* CPU 1 */ - smp_call_function(prom_cpu0_exit, NULL, 1, 1); + smp_call_function(prom_cpu0_exit, NULL, 1); #endif prom_cpu0_exit(NULL); } diff --git a/arch/mips/sibyte/cfe/setup.c b/arch/mips/sibyte/cfe/setup.c index 33fce826f8bf..fd9604d5555a 100644 --- a/arch/mips/sibyte/cfe/setup.c +++ b/arch/mips/sibyte/cfe/setup.c @@ -74,7 +74,7 @@ static void __noreturn cfe_linux_exit(void *arg) if (!reboot_smp) { /* Get CPU 0 to do the cfe_exit */ reboot_smp = 1; - smp_call_function(cfe_linux_exit, arg, 1, 0); + smp_call_function(cfe_linux_exit, arg, 0); } } else { printk("Passing control back to CFE...\n"); diff --git a/arch/mips/sibyte/sb1250/prom.c b/arch/mips/sibyte/sb1250/prom.c index cf8f6b3de86c..65b1af66b674 100644 --- a/arch/mips/sibyte/sb1250/prom.c +++ b/arch/mips/sibyte/sb1250/prom.c @@ -66,7 +66,7 @@ static void prom_linux_exit(void) { #ifdef CONFIG_SMP if (smp_processor_id()) { - smp_call_function(prom_cpu0_exit, NULL, 1, 1); + smp_call_function(prom_cpu0_exit, NULL, 1); } #endif while(1); diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c index 37a5ab410dcc..5191b46a611e 100644 --- a/arch/powerpc/kernel/smp.c +++ b/arch/powerpc/kernel/smp.c @@ -168,7 +168,7 @@ static void stop_this_cpu(void *dummy) void smp_send_stop(void) { - smp_call_function(stop_this_cpu, NULL, 0, 0); + smp_call_function(stop_this_cpu, NULL, 0); } extern struct gettimeofday_struct do_gtod; diff --git a/arch/s390/appldata/appldata_base.c b/arch/s390/appldata/appldata_base.c index ad40729bec3d..837a3b3e7759 100644 --- a/arch/s390/appldata/appldata_base.c +++ b/arch/s390/appldata/appldata_base.c @@ -209,7 +209,7 @@ __appldata_vtimer_setup(int cmd) per_cpu(appldata_timer, i).expires = per_cpu_interval; smp_call_function_single(i, add_virt_timer_periodic, &per_cpu(appldata_timer, i), - 0, 1); + 1); } appldata_timer_active = 1; P_INFO("Monitoring timer started.\n"); @@ -236,7 +236,7 @@ __appldata_vtimer_setup(int cmd) args.timer = &per_cpu(appldata_timer, i); args.expires = per_cpu_interval; smp_call_function_single(i, __appldata_mod_vtimer_wrap, - &args, 0, 1); + &args, 1); } } } diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c index 5d4fa4b1c74c..276b105fb2a4 100644 --- a/arch/s390/kernel/smp.c +++ b/arch/s390/kernel/smp.c @@ -109,7 +109,7 @@ static void do_call_function(void) } static void __smp_call_function_map(void (*func) (void *info), void *info, - int nonatomic, int wait, cpumask_t map) + int wait, cpumask_t map) { struct call_data_struct data; int cpu, local = 0; @@ -162,7 +162,6 @@ out: * smp_call_function: * @func: the function to run; this must be fast and non-blocking * @info: an arbitrary pointer to pass to the function - * @nonatomic: unused * @wait: if true, wait (atomically) until function has completed on other CPUs * * Run a function on all other CPUs. @@ -170,15 +169,14 @@ out: * You must not call this function with disabled interrupts, from a * hardware interrupt handler or from a bottom half. */ -int smp_call_function(void (*func) (void *info), void *info, int nonatomic, - int wait) +int smp_call_function(void (*func) (void *info), void *info, int wait) { cpumask_t map; spin_lock(&call_lock); map = cpu_online_map; cpu_clear(smp_processor_id(), map); - __smp_call_function_map(func, info, nonatomic, wait, map); + __smp_call_function_map(func, info, wait, map); spin_unlock(&call_lock); return 0; } @@ -189,7 +187,6 @@ EXPORT_SYMBOL(smp_call_function); * @cpu: the CPU where func should run * @func: the function to run; this must be fast and non-blocking * @info: an arbitrary pointer to pass to the function - * @nonatomic: unused * @wait: if true, wait (atomically) until function has completed on other CPUs * * Run a function on one processor. @@ -198,11 +195,10 @@ EXPORT_SYMBOL(smp_call_function); * hardware interrupt handler or from a bottom half. */ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, - int nonatomic, int wait) + int wait) { spin_lock(&call_lock); - __smp_call_function_map(func, info, nonatomic, wait, - cpumask_of_cpu(cpu)); + __smp_call_function_map(func, info, wait, cpumask_of_cpu(cpu)); spin_unlock(&call_lock); return 0; } @@ -228,7 +224,7 @@ int smp_call_function_mask(cpumask_t mask, void (*func)(void *), void *info, { spin_lock(&call_lock); cpu_clear(smp_processor_id(), mask); - __smp_call_function_map(func, info, 0, wait, mask); + __smp_call_function_map(func, info, wait, mask); spin_unlock(&call_lock); return 0; } diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c index 7aec676fefd5..bf7bf2c2236a 100644 --- a/arch/s390/kernel/time.c +++ b/arch/s390/kernel/time.c @@ -690,7 +690,7 @@ static int etr_sync_clock(struct etr_aib *aib, int port) */ memset(&etr_sync, 0, sizeof(etr_sync)); preempt_disable(); - smp_call_function(etr_sync_cpu_start, NULL, 0, 0); + smp_call_function(etr_sync_cpu_start, NULL, 0); local_irq_disable(); etr_enable_sync_clock(); @@ -729,7 +729,7 @@ static int etr_sync_clock(struct etr_aib *aib, int port) rc = -EAGAIN; } local_irq_enable(); - smp_call_function(etr_sync_cpu_end,NULL,0,0); + smp_call_function(etr_sync_cpu_end,NULL,0); preempt_enable(); return rc; } diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c index 2ed8dceb297b..71781ba2675b 100644 --- a/arch/sh/kernel/smp.c +++ b/arch/sh/kernel/smp.c @@ -168,7 +168,7 @@ static void stop_this_cpu(void *unused) void smp_send_stop(void) { - smp_call_function(stop_this_cpu, 0, 1, 0); + smp_call_function(stop_this_cpu, 0, 0); } void arch_send_call_function_ipi(cpumask_t mask) @@ -223,7 +223,7 @@ void flush_tlb_mm(struct mm_struct *mm) preempt_disable(); if ((atomic_read(&mm->mm_users) != 1) || (current->mm != mm)) { - smp_call_function(flush_tlb_mm_ipi, (void *)mm, 1, 1); + smp_call_function(flush_tlb_mm_ipi, (void *)mm, 1); } else { int i; for (i = 0; i < num_online_cpus(); i++) @@ -260,7 +260,7 @@ void flush_tlb_range(struct vm_area_struct *vma, fd.vma = vma; fd.addr1 = start; fd.addr2 = end; - smp_call_function(flush_tlb_range_ipi, (void *)&fd, 1, 1); + smp_call_function(flush_tlb_range_ipi, (void *)&fd, 1); } else { int i; for (i = 0; i < num_online_cpus(); i++) @@ -303,7 +303,7 @@ void flush_tlb_page(struct vm_area_struct *vma, unsigned long page) fd.vma = vma; fd.addr1 = page; - smp_call_function(flush_tlb_page_ipi, (void *)&fd, 1, 1); + smp_call_function(flush_tlb_page_ipi, (void *)&fd, 1); } else { int i; for (i = 0; i < num_online_cpus(); i++) @@ -327,6 +327,6 @@ void flush_tlb_one(unsigned long asid, unsigned long vaddr) fd.addr1 = asid; fd.addr2 = vaddr; - smp_call_function(flush_tlb_one_ipi, (void *)&fd, 1, 1); + smp_call_function(flush_tlb_one_ipi, (void *)&fd, 1); local_flush_tlb_one(asid, vaddr); } diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c index b82d017a1744..c099d96f1239 100644 --- a/arch/sparc64/kernel/smp.c +++ b/arch/sparc64/kernel/smp.c @@ -807,7 +807,6 @@ extern unsigned long xcall_call_function; * smp_call_function(): Run a function on all other CPUs. * @func: The function to run. This must be fast and non-blocking. * @info: An arbitrary pointer to pass to the function. - * @nonatomic: currently unused. * @wait: If true, wait (atomically) until function has completed on other CPUs. * * Returns 0 on success, else a negative status code. Does not return until @@ -817,8 +816,7 @@ extern unsigned long xcall_call_function; * hardware interrupt handler or from a bottom half handler. */ static int sparc64_smp_call_function_mask(void (*func)(void *info), void *info, - int nonatomic, int wait, - cpumask_t mask) + int wait, cpumask_t mask) { struct call_data_struct data; int cpus; @@ -853,11 +851,9 @@ out_unlock: return 0; } -int smp_call_function(void (*func)(void *info), void *info, - int nonatomic, int wait) +int smp_call_function(void (*func)(void *info), void *info, int wait) { - return sparc64_smp_call_function_mask(func, info, nonatomic, wait, - cpu_online_map); + return sparc64_smp_call_function_mask(func, info, wait, cpu_online_map); } void smp_call_function_client(int irq, struct pt_regs *regs) @@ -894,7 +890,7 @@ static void tsb_sync(void *info) void smp_tsb_sync(struct mm_struct *mm) { - sparc64_smp_call_function_mask(tsb_sync, mm, 0, 1, mm->cpu_vm_mask); + sparc64_smp_call_function_mask(tsb_sync, mm, 1, mm->cpu_vm_mask); } extern unsigned long xcall_flush_tlb_mm; diff --git a/arch/um/kernel/smp.c b/arch/um/kernel/smp.c index e1062ec36d40..be2d50c3aa95 100644 --- a/arch/um/kernel/smp.c +++ b/arch/um/kernel/smp.c @@ -214,8 +214,7 @@ void smp_call_function_slave(int cpu) atomic_inc(&scf_finished); } -int smp_call_function(void (*_func)(void *info), void *_info, int nonatomic, - int wait) +int smp_call_function(void (*_func)(void *info), void *_info, int wait) { int cpus = num_online_cpus() - 1; int i; diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index 6a1e278d9323..290652cefddb 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -222,7 +222,7 @@ static void set_mtrr(unsigned int reg, unsigned long base, atomic_set(&data.gate,0); /* Start the ball rolling on other CPUs */ - if (smp_call_function(ipi_handler, &data, 1, 0) != 0) + if (smp_call_function(ipi_handler, &data, 0) != 0) panic("mtrr: timed out waiting for other CPUs\n"); local_irq_save(flags); @@ -822,7 +822,7 @@ void mtrr_ap_init(void) */ void mtrr_save_state(void) { - smp_call_function_single(0, mtrr_save_fixed_ranges, NULL, 1, 1); + smp_call_function_single(0, mtrr_save_fixed_ranges, NULL, 1); } static int __init mtrr_init_finialize(void) diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c index daff52a62248..336dd43c9158 100644 --- a/arch/x86/kernel/cpuid.c +++ b/arch/x86/kernel/cpuid.c @@ -95,7 +95,7 @@ static ssize_t cpuid_read(struct file *file, char __user *buf, for (; count; count -= 16) { cmd.eax = pos; cmd.ecx = pos >> 32; - smp_call_function_single(cpu, cpuid_smp_cpuid, &cmd, 1, 1); + smp_call_function_single(cpu, cpuid_smp_cpuid, &cmd, 1); if (copy_to_user(tmp, &cmd, 16)) return -EFAULT; tmp += 16; diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c index 0224c3637c73..cb0a6398c64b 100644 --- a/arch/x86/kernel/ldt.c +++ b/arch/x86/kernel/ldt.c @@ -68,7 +68,7 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload) load_LDT(pc); mask = cpumask_of_cpu(smp_processor_id()); if (!cpus_equal(current->mm->cpu_vm_mask, mask)) - smp_call_function(flush_ldt, NULL, 1, 1); + smp_call_function(flush_ldt, NULL, 1); preempt_enable(); #else load_LDT(pc); diff --git a/arch/x86/kernel/nmi_32.c b/arch/x86/kernel/nmi_32.c index 84160f74eeb0..5562dab0bd20 100644 --- a/arch/x86/kernel/nmi_32.c +++ b/arch/x86/kernel/nmi_32.c @@ -87,7 +87,7 @@ int __init check_nmi_watchdog(void) #ifdef CONFIG_SMP if (nmi_watchdog == NMI_LOCAL_APIC) - smp_call_function(nmi_cpu_busy, (void *)&endflag, 0, 0); + smp_call_function(nmi_cpu_busy, (void *)&endflag, 0); #endif for_each_possible_cpu(cpu) diff --git a/arch/x86/kernel/nmi_64.c b/arch/x86/kernel/nmi_64.c index 5a29ded994fa..2f1e4f503c9e 100644 --- a/arch/x86/kernel/nmi_64.c +++ b/arch/x86/kernel/nmi_64.c @@ -96,7 +96,7 @@ int __init check_nmi_watchdog(void) #ifdef CONFIG_SMP if (nmi_watchdog == NMI_LOCAL_APIC) - smp_call_function(nmi_cpu_busy, (void *)&endflag, 0, 0); + smp_call_function(nmi_cpu_busy, (void *)&endflag, 0); #endif for (cpu = 0; cpu < NR_CPUS; cpu++) diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 575aa3d7248a..56546e8a13ac 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -164,7 +164,7 @@ static void native_smp_send_stop(void) if (reboot_force) return; - smp_call_function(stop_this_cpu, NULL, 0, 0); + smp_call_function(stop_this_cpu, NULL, 0); local_irq_save(flags); disable_local_APIC(); local_irq_restore(flags); diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c index 61efa2f7d564..0a03d57f9b3b 100644 --- a/arch/x86/kernel/vsyscall_64.c +++ b/arch/x86/kernel/vsyscall_64.c @@ -278,7 +278,7 @@ cpu_vsyscall_notifier(struct notifier_block *n, unsigned long action, void *arg) { long cpu = (long)arg; if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) - smp_call_function_single(cpu, cpu_vsyscall_init, NULL, 0, 1); + smp_call_function_single(cpu, cpu_vsyscall_init, NULL, 1); return NOTIFY_DONE; } diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 540e95179074..5534fe59b5fc 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -335,7 +335,7 @@ static void vcpu_clear(struct vcpu_vmx *vmx) { if (vmx->vcpu.cpu == -1) return; - smp_call_function_single(vmx->vcpu.cpu, __vcpu_clear, vmx, 0, 1); + smp_call_function_single(vmx->vcpu.cpu, __vcpu_clear, vmx, 1); vmx->launched = 0; } diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 63a77caa59f1..0faa2546b1cd 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -4044,6 +4044,6 @@ void kvm_vcpu_kick(struct kvm_vcpu *vcpu) * So need not to call smp_call_function_single() in that case. */ if (vcpu->guest_mode && vcpu->cpu != cpu) - smp_call_function_single(ipi_pcpu, vcpu_kick_intr, vcpu, 0, 0); + smp_call_function_single(ipi_pcpu, vcpu_kick_intr, vcpu, 0); put_cpu(); } diff --git a/arch/x86/lib/msr-on-cpu.c b/arch/x86/lib/msr-on-cpu.c index 57d043fa893e..d5a2b39f882b 100644 --- a/arch/x86/lib/msr-on-cpu.c +++ b/arch/x86/lib/msr-on-cpu.c @@ -30,10 +30,10 @@ static int _rdmsr_on_cpu(unsigned int cpu, u32 msr_no, u32 *l, u32 *h, int safe) rv.msr_no = msr_no; if (safe) { - smp_call_function_single(cpu, __rdmsr_safe_on_cpu, &rv, 0, 1); + smp_call_function_single(cpu, __rdmsr_safe_on_cpu, &rv, 1); err = rv.err; } else { - smp_call_function_single(cpu, __rdmsr_on_cpu, &rv, 0, 1); + smp_call_function_single(cpu, __rdmsr_on_cpu, &rv, 1); } *l = rv.l; *h = rv.h; @@ -64,10 +64,10 @@ static int _wrmsr_on_cpu(unsigned int cpu, u32 msr_no, u32 l, u32 h, int safe) rv.l = l; rv.h = h; if (safe) { - smp_call_function_single(cpu, __wrmsr_safe_on_cpu, &rv, 0, 1); + smp_call_function_single(cpu, __wrmsr_safe_on_cpu, &rv, 1); err = rv.err; } else { - smp_call_function_single(cpu, __wrmsr_on_cpu, &rv, 0, 1); + smp_call_function_single(cpu, __wrmsr_on_cpu, &rv, 1); } return err; diff --git a/arch/x86/mach-voyager/voyager_smp.c b/arch/x86/mach-voyager/voyager_smp.c index cb34407a9930..04f596eab749 100644 --- a/arch/x86/mach-voyager/voyager_smp.c +++ b/arch/x86/mach-voyager/voyager_smp.c @@ -1113,7 +1113,7 @@ int safe_smp_processor_id(void) /* broadcast a halt to all other CPUs */ static void voyager_smp_send_stop(void) { - smp_call_function(smp_stop_cpu_function, NULL, 1, 1); + smp_call_function(smp_stop_cpu_function, NULL, 1); } /* this function is triggered in time.c when a clock tick fires diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index b3786e749b8e..a1651d029ea8 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -331,7 +331,7 @@ static void stop_self(void *v) void xen_smp_send_stop(void) { - smp_call_function(stop_self, NULL, 0, 0); + smp_call_function(stop_self, NULL, 0); } void xen_smp_send_reschedule(int cpu) diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index 556ee1585192..4976e5db2b3f 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -1339,7 +1339,7 @@ static void smp_callback(void *v) static int acpi_processor_latency_notify(struct notifier_block *b, unsigned long l, void *v) { - smp_call_function(smp_callback, NULL, 0, 1); + smp_call_function(smp_callback, NULL, 1); return NOTIFY_OK; } diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c index 23554b676d6e..5405769020a1 100644 --- a/drivers/cpuidle/cpuidle.c +++ b/drivers/cpuidle/cpuidle.c @@ -340,7 +340,7 @@ static void smp_callback(void *v) static int cpuidle_latency_notify(struct notifier_block *b, unsigned long l, void *v) { - smp_call_function(smp_callback, NULL, 0, 1); + smp_call_function(smp_callback, NULL, 1); return NOTIFY_OK; } diff --git a/include/asm-alpha/smp.h b/include/asm-alpha/smp.h index 2f60a362d75e..544c69af8168 100644 --- a/include/asm-alpha/smp.h +++ b/include/asm-alpha/smp.h @@ -53,7 +53,7 @@ extern void arch_send_call_function_ipi(cpumask_t mask); #else /* CONFIG_SMP */ #define hard_smp_processor_id() 0 -#define smp_call_function_on_cpu(func,info,retry,wait,cpu) ({ 0; }) +#define smp_call_function_on_cpu(func,info,wait,cpu) ({ 0; }) #endif /* CONFIG_SMP */ diff --git a/include/asm-sparc/smp.h b/include/asm-sparc/smp.h index e6d561599726..b61e74bea06a 100644 --- a/include/asm-sparc/smp.h +++ b/include/asm-sparc/smp.h @@ -72,7 +72,7 @@ static inline void xc5(smpfunc_t func, unsigned long arg1, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5) { smp_cross_call(func, arg1, arg2, arg3, arg4, arg5); } -static inline int smp_call_function(void (*func)(void *info), void *info, int nonatomic, int wait) +static inline int smp_call_function(void (*func)(void *info), void *info, int wait) { xc1((smpfunc_t)func, (unsigned long)info); return 0; diff --git a/include/linux/smp.h b/include/linux/smp.h index eac3e062250f..338cad1b9548 100644 --- a/include/linux/smp.h +++ b/include/linux/smp.h @@ -62,11 +62,11 @@ extern void smp_cpus_done(unsigned int max_cpus); /* * Call a function on all other processors */ -int smp_call_function(void(*func)(void *info), void *info, int retry, int wait); +int smp_call_function(void(*func)(void *info), void *info, int wait); int smp_call_function_mask(cpumask_t mask, void(*func)(void *info), void *info, int wait); int smp_call_function_single(int cpuid, void (*func) (void *info), void *info, - int retry, int wait); + int wait); void __smp_call_function_single(int cpuid, struct call_single_data *data); /* @@ -119,7 +119,7 @@ static inline int up_smp_call_function(void (*func)(void *), void *info) { return 0; } -#define smp_call_function(func, info, retry, wait) \ +#define smp_call_function(func, info, wait) \ (up_smp_call_function(func, info)) #define on_each_cpu(func,info,retry,wait) \ ({ \ @@ -131,7 +131,7 @@ static inline int up_smp_call_function(void (*func)(void *), void *info) static inline void smp_send_reschedule(int cpu) { } #define num_booting_cpus() 1 #define smp_prepare_boot_cpu() do {} while (0) -#define smp_call_function_single(cpuid, func, info, retry, wait) \ +#define smp_call_function_single(cpuid, func, info, wait) \ ({ \ WARN_ON(cpuid != 0); \ local_irq_disable(); \ diff --git a/kernel/smp.c b/kernel/smp.c index f77b75c027ad..7e0432a4a0e2 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -195,7 +195,6 @@ void generic_smp_call_function_single_interrupt(void) * smp_call_function_single - Run a function on a specific CPU * @func: The function to run. This must be fast and non-blocking. * @info: An arbitrary pointer to pass to the function. - * @retry: Unused * @wait: If true, wait until function has completed on other CPUs. * * Returns 0 on success, else a negative status code. Note that @wait @@ -203,7 +202,7 @@ void generic_smp_call_function_single_interrupt(void) * we fall back to on-stack allocation. */ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, - int retry, int wait) + int wait) { struct call_single_data d; unsigned long flags; @@ -339,7 +338,6 @@ EXPORT_SYMBOL(smp_call_function_mask); * smp_call_function(): Run a function on all other CPUs. * @func: The function to run. This must be fast and non-blocking. * @info: An arbitrary pointer to pass to the function. - * @natomic: Unused * @wait: If true, wait (atomically) until function has completed on other CPUs. * * Returns 0 on success, else a negative status code. @@ -351,7 +349,7 @@ EXPORT_SYMBOL(smp_call_function_mask); * You must not call this function with disabled interrupts or from a * hardware interrupt handler or from a bottom half handler. */ -int smp_call_function(void (*func)(void *), void *info, int natomic, int wait) +int smp_call_function(void (*func)(void *), void *info, int wait) { int ret; diff --git a/kernel/softirq.c b/kernel/softirq.c index 36e061740047..d73afb4764ef 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -679,7 +679,7 @@ int on_each_cpu(void (*func) (void *info), void *info, int retry, int wait) int ret = 0; preempt_disable(); - ret = smp_call_function(func, info, retry, wait); + ret = smp_call_function(func, info, wait); local_irq_disable(); func(info); local_irq_enable(); diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 57a1f02e5ec0..75e718539dcb 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -266,7 +266,7 @@ void tick_broadcast_on_off(unsigned long reason, int *oncpu) "offline CPU #%d\n", *oncpu); else smp_call_function_single(*oncpu, tick_do_broadcast_on_off, - &reason, 1, 1); + &reason, 1); } /* diff --git a/net/core/flow.c b/net/core/flow.c index 19991175fdeb..5cf81052d044 100644 --- a/net/core/flow.c +++ b/net/core/flow.c @@ -298,7 +298,7 @@ void flow_cache_flush(void) init_completion(&info.completion); local_bh_disable(); - smp_call_function(flow_cache_flush_per_cpu, &info, 1, 0); + smp_call_function(flow_cache_flush_per_cpu, &info, 0); flow_cache_flush_tasklet((unsigned long)&info); local_bh_enable(); diff --git a/net/iucv/iucv.c b/net/iucv/iucv.c index 918970762131..94d5a45c3a57 100644 --- a/net/iucv/iucv.c +++ b/net/iucv/iucv.c @@ -480,7 +480,7 @@ static void iucv_setmask_mp(void) if (cpu_isset(cpu, iucv_buffer_cpumask) && !cpu_isset(cpu, iucv_irq_cpumask)) smp_call_function_single(cpu, iucv_allow_cpu, - NULL, 0, 1); + NULL, 1); preempt_enable(); } @@ -498,7 +498,7 @@ static void iucv_setmask_up(void) cpumask = iucv_irq_cpumask; cpu_clear(first_cpu(iucv_irq_cpumask), cpumask); for_each_cpu_mask(cpu, cpumask) - smp_call_function_single(cpu, iucv_block_cpu, NULL, 0, 1); + smp_call_function_single(cpu, iucv_block_cpu, NULL, 1); } /** @@ -523,7 +523,7 @@ static int iucv_enable(void) rc = -EIO; preempt_disable(); for_each_online_cpu(cpu) - smp_call_function_single(cpu, iucv_declare_cpu, NULL, 0, 1); + smp_call_function_single(cpu, iucv_declare_cpu, NULL, 1); preempt_enable(); if (cpus_empty(iucv_buffer_cpumask)) /* No cpu could declare an iucv buffer. */ @@ -580,7 +580,7 @@ static int __cpuinit iucv_cpu_notify(struct notifier_block *self, case CPU_ONLINE_FROZEN: case CPU_DOWN_FAILED: case CPU_DOWN_FAILED_FROZEN: - smp_call_function_single(cpu, iucv_declare_cpu, NULL, 0, 1); + smp_call_function_single(cpu, iucv_declare_cpu, NULL, 1); break; case CPU_DOWN_PREPARE: case CPU_DOWN_PREPARE_FROZEN: @@ -589,10 +589,10 @@ static int __cpuinit iucv_cpu_notify(struct notifier_block *self, if (cpus_empty(cpumask)) /* Can't offline last IUCV enabled cpu. */ return NOTIFY_BAD; - smp_call_function_single(cpu, iucv_retrieve_cpu, NULL, 0, 1); + smp_call_function_single(cpu, iucv_retrieve_cpu, NULL, 1); if (cpus_empty(iucv_irq_cpumask)) smp_call_function_single(first_cpu(iucv_buffer_cpumask), - iucv_allow_cpu, NULL, 0, 1); + iucv_allow_cpu, NULL, 1); break; } return NOTIFY_OK; @@ -652,7 +652,7 @@ static void iucv_cleanup_queue(void) * pending interrupts force them to the work queue by calling * an empty function on all cpus. */ - smp_call_function(__iucv_cleanup_queue, NULL, 0, 1); + smp_call_function(__iucv_cleanup_queue, NULL, 1); spin_lock_irq(&iucv_queue_lock); list_for_each_entry_safe(p, n, &iucv_task_queue, list) { /* Remove stale work items from the task queue. */ diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 2d29e260da3d..ea1f595f8a87 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -1266,12 +1266,12 @@ static int kvm_cpu_hotplug(struct notifier_block *notifier, unsigned long val, case CPU_UP_CANCELED: printk(KERN_INFO "kvm: disabling virtualization on CPU%d\n", cpu); - smp_call_function_single(cpu, hardware_disable, NULL, 0, 1); + smp_call_function_single(cpu, hardware_disable, NULL, 1); break; case CPU_ONLINE: printk(KERN_INFO "kvm: enabling virtualization on CPU%d\n", cpu); - smp_call_function_single(cpu, hardware_enable, NULL, 0, 1); + smp_call_function_single(cpu, hardware_enable, NULL, 1); break; } return NOTIFY_OK; @@ -1474,7 +1474,7 @@ int kvm_init(void *opaque, unsigned int vcpu_size, for_each_online_cpu(cpu) { smp_call_function_single(cpu, kvm_arch_check_processor_compat, - &r, 0, 1); + &r, 1); if (r < 0) goto out_free_1; } -- cgit v1.3-8-gc7d7 From 0853ad66b14feb12acde7ac13b7c3b75770a0adc Mon Sep 17 00:00:00 2001 From: Santwona Behera Date: Wed, 2 Jul 2008 03:47:41 -0700 Subject: netdev: Add support for rx flow hash configuration, using ethtool. Added new interfaces to ethtool to configure receive network flow distribution across multiple rx rings using hashing. Signed-off-by: Santwona Behera Signed-off-by: David S. Miller --- include/linux/ethtool.h | 33 +++++++++++++++++++++++++++++++++ net/core/ethtool.c | 37 +++++++++++++++++++++++++++++++++++++ 2 files changed, 70 insertions(+) (limited to 'net/core') diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h index c8d216357865..8bb5e87df365 100644 --- a/include/linux/ethtool.h +++ b/include/linux/ethtool.h @@ -272,6 +272,12 @@ enum ethtool_flags { ETH_FLAG_LRO = (1 << 15), /* LRO is enabled */ }; +struct ethtool_rxnfc { + __u32 cmd; + __u32 flow_type; + __u64 data; +}; + #ifdef __KERNEL__ struct net_device; @@ -396,6 +402,8 @@ struct ethtool_ops { /* the following hooks are obsolete */ int (*self_test_count)(struct net_device *);/* use get_sset_count */ int (*get_stats_count)(struct net_device *);/* use get_sset_count */ + int (*get_rxhash)(struct net_device *, struct ethtool_rxnfc *); + int (*set_rxhash)(struct net_device *, struct ethtool_rxnfc *); }; #endif /* __KERNEL__ */ @@ -442,6 +450,9 @@ struct ethtool_ops { #define ETHTOOL_GPFLAGS 0x00000027 /* Get driver-private flags bitmap */ #define ETHTOOL_SPFLAGS 0x00000028 /* Set driver-private flags bitmap */ +#define ETHTOOL_GRXFH 0x00000029 /* Get RX flow hash configuration */ +#define ETHTOOL_SRXFH 0x0000002a /* Set RX flow hash configuration */ + /* compatibility with older code */ #define SPARC_ETH_GSET ETHTOOL_GSET #define SPARC_ETH_SSET ETHTOOL_SSET @@ -528,4 +539,26 @@ struct ethtool_ops { #define WAKE_MAGIC (1 << 5) #define WAKE_MAGICSECURE (1 << 6) /* only meaningful if WAKE_MAGIC */ +/* L3-L4 network traffic flow types */ +#define TCP_V4_FLOW 0x01 +#define UDP_V4_FLOW 0x02 +#define SCTP_V4_FLOW 0x03 +#define AH_ESP_V4_FLOW 0x04 +#define TCP_V6_FLOW 0x05 +#define UDP_V6_FLOW 0x06 +#define SCTP_V6_FLOW 0x07 +#define AH_ESP_V6_FLOW 0x08 + +/* L3-L4 network traffic flow hash options */ +#define RXH_DEV_PORT (1 << 0) +#define RXH_L2DA (1 << 1) +#define RXH_VLAN (1 << 2) +#define RXH_L3_PROTO (1 << 3) +#define RXH_IP_SRC (1 << 4) +#define RXH_IP_DST (1 << 5) +#define RXH_L4_B_0_1 (1 << 6) /* src port in case of TCP/UDP/SCTP */ +#define RXH_L4_B_2_3 (1 << 7) /* dst port in case of TCP/UDP/SCTP */ +#define RXH_DISCARD (1 << 31) + + #endif /* _LINUX_ETHTOOL_H */ diff --git a/net/core/ethtool.c b/net/core/ethtool.c index 0133b5ebd545..14ada537f895 100644 --- a/net/core/ethtool.c +++ b/net/core/ethtool.c @@ -209,6 +209,36 @@ static int ethtool_get_drvinfo(struct net_device *dev, void __user *useraddr) return 0; } +static int ethtool_set_rxhash(struct net_device *dev, void __user *useraddr) +{ + struct ethtool_rxnfc cmd; + + if (!dev->ethtool_ops->set_rxhash) + return -EOPNOTSUPP; + + if (copy_from_user(&cmd, useraddr, sizeof(cmd))) + return -EFAULT; + + return dev->ethtool_ops->set_rxhash(dev, &cmd); +} + +static int ethtool_get_rxhash(struct net_device *dev, void __user *useraddr) +{ + struct ethtool_rxnfc info; + + if (!dev->ethtool_ops->get_rxhash) + return -EOPNOTSUPP; + + if (copy_from_user(&info, useraddr, sizeof(info))) + return -EFAULT; + + dev->ethtool_ops->get_rxhash(dev, &info); + + if (copy_to_user(useraddr, &info, sizeof(info))) + return -EFAULT; + return 0; +} + static int ethtool_get_regs(struct net_device *dev, char __user *useraddr) { struct ethtool_regs regs; @@ -826,6 +856,7 @@ int dev_ethtool(struct net *net, struct ifreq *ifr) case ETHTOOL_GGSO: case ETHTOOL_GFLAGS: case ETHTOOL_GPFLAGS: + case ETHTOOL_GRXFH: break; default: if (!capable(CAP_NET_ADMIN)) @@ -977,6 +1008,12 @@ int dev_ethtool(struct net *net, struct ifreq *ifr) rc = ethtool_set_value(dev, useraddr, dev->ethtool_ops->set_priv_flags); break; + case ETHTOOL_GRXFH: + rc = ethtool_get_rxhash(dev, useraddr); + break; + case ETHTOOL_SRXFH: + rc = ethtool_set_rxhash(dev, useraddr); + break; default: rc = -EOPNOTSUPP; } -- cgit v1.3-8-gc7d7 From ae299fc051aa68ca6ef1807c37bb92d9b6ff817c Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:01:28 -0700 Subject: net: add fib_rules_ops to flush_cache method This is required to pass namespace context into rt_cache_flush called from ->flush_cache. Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- include/net/fib_rules.h | 2 +- net/core/fib_rules.c | 2 +- net/decnet/dn_rules.c | 2 +- net/ipv4/fib_rules.c | 4 ++-- 4 files changed, 5 insertions(+), 5 deletions(-) (limited to 'net/core') diff --git a/include/net/fib_rules.h b/include/net/fib_rules.h index a5c6ccc5bb19..c2bb5cae6515 100644 --- a/include/net/fib_rules.h +++ b/include/net/fib_rules.h @@ -62,7 +62,7 @@ struct fib_rules_ops /* Called after modifications to the rules set, must flush * the route cache if one exists. */ - void (*flush_cache)(void); + void (*flush_cache)(struct fib_rules_ops *ops); int nlgroup; const struct nla_policy *policy; diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c index e3e9ab0f74e3..1c2943a119f3 100644 --- a/net/core/fib_rules.c +++ b/net/core/fib_rules.c @@ -69,7 +69,7 @@ static void rules_ops_put(struct fib_rules_ops *ops) static void flush_route_cache(struct fib_rules_ops *ops) { if (ops->flush_cache) - ops->flush_cache(); + ops->flush_cache(ops); } int fib_rules_register(struct fib_rules_ops *ops) diff --git a/net/decnet/dn_rules.c b/net/decnet/dn_rules.c index 5b7539b7fe0c..14fbca55e908 100644 --- a/net/decnet/dn_rules.c +++ b/net/decnet/dn_rules.c @@ -229,7 +229,7 @@ static u32 dn_fib_rule_default_pref(struct fib_rules_ops *ops) return 0; } -static void dn_fib_rule_flush_cache(void) +static void dn_fib_rule_flush_cache(struct fib_rules_ops *ops) { dn_rt_cache_flush(-1); } diff --git a/net/ipv4/fib_rules.c b/net/ipv4/fib_rules.c index bc05de413087..6080d7120821 100644 --- a/net/ipv4/fib_rules.c +++ b/net/ipv4/fib_rules.c @@ -258,9 +258,9 @@ static size_t fib4_rule_nlmsg_payload(struct fib_rule *rule) + nla_total_size(4); /* flow */ } -static void fib4_rule_flush_cache(void) +static void fib4_rule_flush_cache(struct fib_rules_ops *ops) { - rt_cache_flush(&init_net, -1); + rt_cache_flush(ops->fro_net, -1); } static struct fib_rules_ops fib4_rules_ops_template = { -- cgit v1.3-8-gc7d7 From 4b5a698ef423eebc37cfacc6d3376d6dffd5bf83 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sun, 6 Jul 2008 15:49:08 -0700 Subject: net: fix dev_set_promiscuity() breakage Commit dad9b335 (netdevice: Fix promiscuity and allmulti overflow) broke dev_set_promiscuity() by returning on success without reprogramming the device. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/core/dev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core') diff --git a/net/core/dev.c b/net/core/dev.c index bfa9a6a951dd..75933932463d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2859,7 +2859,7 @@ int dev_set_promiscuity(struct net_device *dev, int inc) int err; err = __dev_set_promiscuity(dev, inc); - if (!err) + if (err < 0) return err; if (dev->flags != old_flags) dev_set_rx_mode(dev); -- cgit v1.3-8-gc7d7 From bb949fbd1878973c3539d9aecff52f284482a937 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 16:55:56 -0700 Subject: netdev: Create netdev_queue abstraction. A netdev_queue is an entity managed by a qdisc. Currently there is one RX and one TX queue, and a netdev_queue merely contains a backpointer to the net_device. The Qdisc struct is augmented with a netdev_queue pointer as well. Eventually the 'dev' Qdisc member will go away and we will have the resulting hierarchy: net_device --> netdev_queue --> Qdisc Also, qdisc_alloc() and qdisc_create_dflt() now take a netdev_queue pointer argument. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 7 +++++++ include/net/sch_generic.h | 6 +++++- net/core/dev.c | 8 ++++++++ net/mac80211/wme.c | 6 ++++-- net/sched/sch_api.c | 12 +++++++----- net/sched/sch_atm.c | 6 ++++-- net/sched/sch_cbq.c | 9 ++++++--- net/sched/sch_dsmark.c | 6 ++++-- net/sched/sch_fifo.c | 3 ++- net/sched/sch_generic.c | 14 ++++++++++---- net/sched/sch_hfsc.c | 9 ++++++--- net/sched/sch_htb.c | 11 +++++++---- net/sched/sch_netem.c | 3 ++- net/sched/sch_prio.c | 3 ++- 14 files changed, 74 insertions(+), 29 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index e009c6fbf5cd..515fd25bf0fc 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -448,6 +448,10 @@ static inline void napi_synchronize(const struct napi_struct *n) # define napi_synchronize(n) barrier() #endif +struct netdev_queue { + struct net_device *dev; +}; + /* * The DEVICE structure. * Actually, this whole structure is a big mistake. It mixes I/O @@ -624,6 +628,9 @@ struct net_device unsigned char broadcast[MAX_ADDR_LEN]; /* hw bcast add */ + struct netdev_queue rx_queue; + struct netdev_queue tx_queue; + /* ingress path synchronizer */ spinlock_t ingress_lock; struct Qdisc *qdisc_ingress; diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index 073f2580b83b..0ab53c575f87 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -37,6 +37,7 @@ struct Qdisc u32 parent; atomic_t refcnt; struct sk_buff_head q; + struct netdev_queue *dev_queue; struct net_device *dev; struct list_head list; @@ -216,8 +217,11 @@ extern void dev_deactivate(struct net_device *dev); extern void qdisc_reset(struct Qdisc *qdisc); extern void qdisc_destroy(struct Qdisc *qdisc); extern void qdisc_tree_decrease_qlen(struct Qdisc *qdisc, unsigned int n); -extern struct Qdisc *qdisc_alloc(struct net_device *dev, struct Qdisc_ops *ops); +extern struct Qdisc *qdisc_alloc(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc_ops *ops); extern struct Qdisc *qdisc_create_dflt(struct net_device *dev, + struct netdev_queue *dev_queue, struct Qdisc_ops *ops, u32 parentid); extern void tcf_destroy(struct tcf_proto *tp); extern void tcf_destroy_chain(struct tcf_proto **fl); diff --git a/net/core/dev.c b/net/core/dev.c index 75933932463d..9b281c906eb0 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4072,6 +4072,12 @@ static struct net_device_stats *internal_stats(struct net_device *dev) return &dev->stats; } +static void netdev_init_queues(struct net_device *dev) +{ + dev->rx_queue.dev = dev; + dev->tx_queue.dev = dev; +} + /** * alloc_netdev_mq - allocate network device * @sizeof_priv: size of private data to allocate space for @@ -4124,6 +4130,8 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, dev->egress_subqueue_count = queue_count; dev->gso_max_size = GSO_MAX_SIZE; + netdev_init_queues(dev); + dev->get_stats = internal_stats; netpoll_netdev_init(dev); setup(dev); diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 5c666f7eda8f..770f1c09b793 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -359,7 +359,8 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) /* create child queues */ for (i = 0; i < QD_NUM(hw); i++) { skb_queue_head_init(&q->requeued[i]); - q->queues[i] = qdisc_create_dflt(qd->dev, &pfifo_qdisc_ops, + q->queues[i] = qdisc_create_dflt(qd->dev, qd->dev_queue, + &pfifo_qdisc_ops, qd->handle); if (!q->queues[i]) { q->queues[i] = &noop_qdisc; @@ -575,7 +576,8 @@ void ieee80211_install_qdisc(struct net_device *dev) { struct Qdisc *qdisc; - qdisc = qdisc_create_dflt(dev, &wme_qdisc_ops, TC_H_ROOT); + qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + &wme_qdisc_ops, TC_H_ROOT); if (!qdisc) { printk(KERN_ERR "%s: qdisc installation failed\n", dev->name); return; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 69e918bb4278..b86c98bd06a3 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -552,8 +552,8 @@ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, */ static struct Qdisc * -qdisc_create(struct net_device *dev, u32 parent, u32 handle, - struct nlattr **tca, int *errp) +qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, + u32 parent, u32 handle, struct nlattr **tca, int *errp) { int err; struct nlattr *kind = tca[TCA_KIND]; @@ -593,7 +593,7 @@ qdisc_create(struct net_device *dev, u32 parent, u32 handle, if (ops == NULL) goto err_out; - sch = qdisc_alloc(dev, ops); + sch = qdisc_alloc(dev, dev_queue, ops); if (IS_ERR(sch)) { err = PTR_ERR(sch); goto err_out2; @@ -892,10 +892,12 @@ create_n_graft: if (!(n->nlmsg_flags&NLM_F_CREATE)) return -ENOENT; if (clid == TC_H_INGRESS) - q = qdisc_create(dev, tcm->tcm_parent, tcm->tcm_parent, + q = qdisc_create(dev, &dev->rx_queue, + tcm->tcm_parent, tcm->tcm_parent, tca, &err); else - q = qdisc_create(dev, tcm->tcm_parent, tcm->tcm_handle, + q = qdisc_create(dev, &dev->tx_queue, + tcm->tcm_parent, tcm->tcm_handle, tca, &err); if (q == NULL) { if (err == -EAGAIN) diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index db0e23ae85f8..3dddab531d5a 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -296,7 +296,8 @@ static int atm_tc_change(struct Qdisc *sch, u32 classid, u32 parent, goto err_out; } flow->filter_list = NULL; - flow->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid); + flow->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid); if (!flow->q) flow->q = &noop_qdisc; pr_debug("atm_tc_change: qdisc %p\n", flow->q); @@ -555,7 +556,8 @@ static int atm_tc_init(struct Qdisc *sch, struct nlattr *opt) pr_debug("atm_tc_init(sch %p,[qdisc %p],opt %p)\n", sch, p, opt); p->flows = &p->link; - p->link.q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, sch->handle); + p->link.q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (!p->link.q) p->link.q = &noop_qdisc; pr_debug("atm_tc_init: link (%p) qdisc %p\n", &p->link, p->link.q); diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 968b4c73c9c1..d360dcd0818b 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -1401,7 +1401,8 @@ static int cbq_init(struct Qdisc *sch, struct nlattr *opt) q->link.sibling = &q->link; q->link.common.classid = sch->handle; q->link.qdisc = sch; - if (!(q->link.q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + if (!(q->link.q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle))) q->link.q = &noop_qdisc; @@ -1645,7 +1646,8 @@ static int cbq_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl) { if (new == NULL) { - new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, cl->common.classid); if (new == NULL) return -ENOBUFS; @@ -1877,7 +1879,8 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t cl->R_tab = rtab; rtab = NULL; cl->refcnt = 1; - if (!(cl->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid))) + if (!(cl->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid))) cl->q = &noop_qdisc; cl->common.classid = classid; cl->tparent = parent; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index c4c1317cd47d..c955ba24e5cf 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -60,7 +60,8 @@ static int dsmark_graft(struct Qdisc *sch, unsigned long arg, sch, p, new, old); if (new == NULL) { - new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (new == NULL) new = &noop_qdisc; @@ -390,7 +391,8 @@ static int dsmark_init(struct Qdisc *sch, struct nlattr *opt) p->default_index = default_index; p->set_tc_index = nla_get_flag(tb[TCA_DSMARK_SET_TC_INDEX]); - p->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, sch->handle); + p->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (p->q == NULL) p->q = &noop_qdisc; diff --git a/net/sched/sch_fifo.c b/net/sched/sch_fifo.c index 82d7d7bbbb16..779eae85faf0 100644 --- a/net/sched/sch_fifo.c +++ b/net/sched/sch_fifo.c @@ -137,7 +137,8 @@ struct Qdisc *fifo_create_dflt(struct Qdisc *sch, struct Qdisc_ops *ops, struct Qdisc *q; int err = -ENOMEM; - q = qdisc_create_dflt(sch->dev, ops, TC_H_MAKE(sch->handle, 1)); + q = qdisc_create_dflt(sch->dev, sch->dev_queue, + ops, TC_H_MAKE(sch->handle, 1)); if (q) { err = fifo_set_limit(q, limit); if (err < 0) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 13afa7214392..d97086480893 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -440,7 +440,9 @@ static struct Qdisc_ops pfifo_fast_ops __read_mostly = { .owner = THIS_MODULE, }; -struct Qdisc *qdisc_alloc(struct net_device *dev, struct Qdisc_ops *ops) +struct Qdisc *qdisc_alloc(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc_ops *ops) { void *p; struct Qdisc *sch; @@ -462,6 +464,7 @@ struct Qdisc *qdisc_alloc(struct net_device *dev, struct Qdisc_ops *ops) sch->ops = ops; sch->enqueue = ops->enqueue; sch->dequeue = ops->dequeue; + sch->dev_queue = dev_queue; sch->dev = dev; dev_hold(dev); atomic_set(&sch->refcnt, 1); @@ -471,12 +474,14 @@ errout: return ERR_PTR(err); } -struct Qdisc * qdisc_create_dflt(struct net_device *dev, struct Qdisc_ops *ops, +struct Qdisc * qdisc_create_dflt(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc_ops *ops, unsigned int parentid) { struct Qdisc *sch; - sch = qdisc_alloc(dev, ops); + sch = qdisc_alloc(dev, dev_queue, ops); if (IS_ERR(sch)) goto errout; sch->stats_lock = &dev->queue_lock; @@ -545,7 +550,8 @@ void dev_activate(struct net_device *dev) if (dev->qdisc_sleeping == &noop_qdisc) { struct Qdisc *qdisc; if (dev->tx_queue_len) { - qdisc = qdisc_create_dflt(dev, &pfifo_fast_ops, + qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + &pfifo_fast_ops, TC_H_ROOT); if (qdisc == NULL) { printk(KERN_INFO "%s: activation failed\n", dev->name); diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 3a8267246a4f..5a22fec4eadd 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1083,7 +1083,8 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, cl->refcnt = 1; cl->sched = q; cl->cl_parent = parent; - cl->qdisc = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid); + cl->qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid); if (cl->qdisc == NULL) cl->qdisc = &noop_qdisc; INIT_LIST_HEAD(&cl->children); @@ -1201,7 +1202,8 @@ hfsc_graft_class(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl->level > 0) return -EINVAL; if (new == NULL) { - new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, cl->cl_common.classid); if (new == NULL) new = &noop_qdisc; @@ -1443,7 +1445,8 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt) q->root.cl_common.classid = sch->handle; q->root.refcnt = 1; q->root.sched = q; - q->root.qdisc = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + q->root.qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (q->root.qdisc == NULL) q->root.qdisc = &noop_qdisc; diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index ee8b4ffe110c..956a67f66b9c 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1129,7 +1129,8 @@ static int htb_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl && !cl->level) { if (new == NULL && - (new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + (new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, cl->common.classid)) == NULL) return -ENOBUFS; @@ -1256,8 +1257,9 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) return -EBUSY; if (!cl->level && htb_parent_last_child(cl)) { - new_q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, - cl->parent->common.classid); + new_q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, + cl->parent->common.classid); last_child = 1; } @@ -1376,7 +1378,8 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, /* create leaf qdisc early because it uses kmalloc(GFP_KERNEL) so that can't be used inside of sch_tree_lock -- thanks to Karlis Peisenieks */ - new_q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid); + new_q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid); sch_tree_lock(sch); if (parent && !parent->level) { unsigned int qlen = parent->un.leaf.q->q.qlen; diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 24697667247c..aa7a04e32ae9 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -536,7 +536,8 @@ static int netem_init(struct Qdisc *sch, struct nlattr *opt) qdisc_watchdog_init(&q->watchdog, sch); - q->qdisc = qdisc_create_dflt(sch->dev, &tfifo_qdisc_ops, + q->qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + &tfifo_qdisc_ops, TC_H_MAKE(sch->handle, 1)); if (!q->qdisc) { pr_debug("netem: qdisc create failed\n"); diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index 5532f1031ab5..ca58a039208e 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -281,7 +281,8 @@ static int prio_tune(struct Qdisc *sch, struct nlattr *opt) for (i=0; ibands; i++) { if (q->queues[i] == &noop_qdisc) { struct Qdisc *child; - child = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + child = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, TC_H_MAKE(sch->handle, i + 1)); if (child) { sch_tree_lock(sch); -- cgit v1.3-8-gc7d7 From dc2b48475a0a36f8b3bbb2da60d3a006dc5c2c84 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 17:18:23 -0700 Subject: netdev: Move queue_lock into struct netdev_queue. The lock is now an attribute of the device queue. One thing to notice is that "suspicious" places emerge which will need specific training about multiple queue handling. They are so marked with explicit "netdev->rx_queue" and "netdev->tx_queue" references. Signed-off-by: David S. Miller --- drivers/net/ifb.c | 8 ++++---- include/linux/netdevice.h | 4 ++-- net/core/dev.c | 33 ++++++++++++++++++++++----------- net/mac80211/main.c | 10 +++++----- net/mac80211/wme.c | 2 +- net/sched/sch_api.c | 2 +- net/sched/sch_cbq.c | 8 ++++---- net/sched/sch_generic.c | 40 ++++++++++++++++++++-------------------- net/sched/sch_hfsc.c | 4 ++-- net/sched/sch_htb.c | 16 ++++++++-------- net/sched/sch_netem.c | 4 ++-- net/sched/sch_teql.c | 4 ++-- 12 files changed, 73 insertions(+), 62 deletions(-) (limited to 'net/core') diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c index af233b591534..bc3de272a829 100644 --- a/drivers/net/ifb.c +++ b/drivers/net/ifb.c @@ -229,12 +229,12 @@ module_param(numifbs, int, 0); MODULE_PARM_DESC(numifbs, "Number of ifb devices"); /* - * dev_ifb->queue_lock is usually taken after dev->ingress_lock, + * dev_ifb->tx_queue.lock is usually taken after dev->ingress_lock, * reversely to e.g. qdisc_lock_tree(). It should be safe until - * ifb doesn't take dev->queue_lock with dev_ifb->ingress_lock. + * ifb doesn't take dev->tx_queue.lock with dev_ifb->ingress_lock. * But lockdep should know that ifb has different locks from dev. */ -static struct lock_class_key ifb_queue_lock_key; +static struct lock_class_key ifb_tx_queue_lock_key; static struct lock_class_key ifb_ingress_lock_key; @@ -258,7 +258,7 @@ static int __init ifb_init_one(int index) if (err < 0) goto err; - lockdep_set_class(&dev_ifb->queue_lock, &ifb_queue_lock_key); + lockdep_set_class(&dev_ifb->tx_queue.lock, &ifb_tx_queue_lock_key); lockdep_set_class(&dev_ifb->ingress_lock, &ifb_ingress_lock_key); return 0; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 515fd25bf0fc..e835acacb479 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -449,6 +449,7 @@ static inline void napi_synchronize(const struct napi_struct *n) #endif struct netdev_queue { + spinlock_t lock; struct net_device *dev; }; @@ -629,7 +630,7 @@ struct net_device unsigned char broadcast[MAX_ADDR_LEN]; /* hw bcast add */ struct netdev_queue rx_queue; - struct netdev_queue tx_queue; + struct netdev_queue tx_queue ____cacheline_aligned_in_smp; /* ingress path synchronizer */ spinlock_t ingress_lock; @@ -639,7 +640,6 @@ struct net_device * Cache line mostly used on queue transmit path (qdisc) */ /* device queue lock */ - spinlock_t queue_lock ____cacheline_aligned_in_smp; struct Qdisc *qdisc; struct Qdisc *qdisc_sleeping; struct list_head qdisc_list; diff --git a/net/core/dev.c b/net/core/dev.c index 9b281c906eb0..05011048b86c 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1667,6 +1667,7 @@ out_kfree_skb: int dev_queue_xmit(struct sk_buff *skb) { struct net_device *dev = skb->dev; + struct netdev_queue *txq; struct Qdisc *q; int rc = -ENOMEM; @@ -1699,14 +1700,15 @@ int dev_queue_xmit(struct sk_buff *skb) } gso: - spin_lock_prefetch(&dev->queue_lock); + txq = &dev->tx_queue; + spin_lock_prefetch(&txq->lock); /* Disable soft irqs for various locks below. Also * stops preemption for RCU. */ rcu_read_lock_bh(); - /* Updates of qdisc are serialized by queue_lock. + /* Updates of qdisc are serialized by queue->lock. * The struct Qdisc which is pointed to by qdisc is now a * rcu structure - it may be accessed without acquiring * a lock (but the structure may be stale.) The freeing of the @@ -1714,7 +1716,7 @@ gso: * more references to it. * * If the qdisc has an enqueue function, we still need to - * hold the queue_lock before calling it, since queue_lock + * hold the queue->lock before calling it, since queue->lock * also serializes access to the device queue. */ @@ -1724,19 +1726,19 @@ gso: #endif if (q->enqueue) { /* Grab device queue */ - spin_lock(&dev->queue_lock); + spin_lock(&txq->lock); q = dev->qdisc; if (q->enqueue) { /* reset queue_mapping to zero */ skb_set_queue_mapping(skb, 0); rc = q->enqueue(skb, q); qdisc_run(dev); - spin_unlock(&dev->queue_lock); + spin_unlock(&txq->lock); rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; goto out; } - spin_unlock(&dev->queue_lock); + spin_unlock(&txq->lock); } /* The device has no queue. Common case for software devices: @@ -1919,14 +1921,17 @@ static void net_tx_action(struct softirq_action *h) while (head) { struct net_device *dev = head; + struct netdev_queue *txq; head = head->next_sched; + txq = &dev->tx_queue; + smp_mb__before_clear_bit(); clear_bit(__LINK_STATE_SCHED, &dev->state); - if (spin_trylock(&dev->queue_lock)) { + if (spin_trylock(&txq->lock)) { qdisc_run(dev); - spin_unlock(&dev->queue_lock); + spin_unlock(&txq->lock); } else { netif_schedule(dev); } @@ -3787,7 +3792,6 @@ int register_netdevice(struct net_device *dev) BUG_ON(!dev_net(dev)); net = dev_net(dev); - spin_lock_init(&dev->queue_lock); spin_lock_init(&dev->_xmit_lock); netdev_set_lockdep_class(&dev->_xmit_lock, dev->type); dev->xmit_lock_owner = -1; @@ -4072,10 +4076,17 @@ static struct net_device_stats *internal_stats(struct net_device *dev) return &dev->stats; } +static void netdev_init_one_queue(struct net_device *dev, + struct netdev_queue *queue) +{ + spin_lock_init(&queue->lock); + queue->dev = dev; +} + static void netdev_init_queues(struct net_device *dev) { - dev->rx_queue.dev = dev; - dev->tx_queue.dev = dev; + netdev_init_one_queue(dev, &dev->rx_queue); + netdev_init_one_queue(dev, &dev->tx_queue); } /** diff --git a/net/mac80211/main.c b/net/mac80211/main.c index cf477ad39dac..12aeaf78ae75 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -636,7 +636,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* ensure that TX flow won't interrupt us * until the end of the call to requeue function */ - spin_lock_bh(&local->mdev->queue_lock); + spin_lock_bh(&local->mdev->tx_queue.lock); /* create a new queue for this aggregation */ ret = ieee80211_ht_agg_queue_add(local, sta, tid); @@ -675,7 +675,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* Will put all the packets in the new SW queue */ ieee80211_requeue(local, ieee802_1d_to_ac[tid]); - spin_unlock_bh(&local->mdev->queue_lock); + spin_unlock_bh(&local->mdev->tx_queue.lock); spin_unlock_bh(&sta->lock); /* send an addBA request */ @@ -701,7 +701,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) err_unlock_queue: kfree(sta->ampdu_mlme.tid_tx[tid]); sta->ampdu_mlme.tid_tx[tid] = NULL; - spin_unlock_bh(&local->mdev->queue_lock); + spin_unlock_bh(&local->mdev->tx_queue.lock); ret = -EBUSY; err_unlock_sta: spin_unlock_bh(&sta->lock); @@ -875,10 +875,10 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) /* avoid ordering issues: we are the only one that can modify * the content of the qdiscs */ - spin_lock_bh(&local->mdev->queue_lock); + spin_lock_bh(&local->mdev->tx_queue.lock); /* remove the queue for this aggregation */ ieee80211_ht_agg_queue_remove(local, sta, tid, 1); - spin_unlock_bh(&local->mdev->queue_lock); + spin_unlock_bh(&local->mdev->tx_queue.lock); /* we just requeued the all the frames that were in the removed * queue, and since we might miss a softirq we do netif_schedule. diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 2fbc171130bf..59ed9cae66b9 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -648,7 +648,7 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, } /** - * the caller needs to hold local->mdev->queue_lock + * the caller needs to hold local->mdev->tx_queue.lock */ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, struct sta_info *sta, u16 tid, diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 1f893082a4f6..2a1834f8c7d8 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -606,7 +606,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, sch->stats_lock = &dev->ingress_lock; handle = TC_H_MAKE(TC_H_INGRESS, 0); } else { - sch->stats_lock = &dev->queue_lock; + sch->stats_lock = &dev_queue->lock; if (handle == 0) { handle = qdisc_alloc_handle(dev); err = -ENOMEM; diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 9f2ace585fd6..99ce3da2b0a4 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -1746,10 +1746,10 @@ static void cbq_put(struct Qdisc *sch, unsigned long arg) #ifdef CONFIG_NET_CLS_ACT struct cbq_sched_data *q = qdisc_priv(sch); - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); if (q->rx_class == cl) q->rx_class = NULL; - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); #endif cbq_destroy_class(sch, cl); @@ -1828,7 +1828,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE]); return 0; } @@ -1919,7 +1919,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); + &sch->dev_queue->lock, tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index b626a4f32b6b..ee8f9f78a095 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -29,31 +29,31 @@ /* Main transmission queue. */ /* Modifications to data participating in scheduling must be protected with - * dev->queue_lock spinlock. + * queue->lock spinlock. * * The idea is the following: * - enqueue, dequeue are serialized via top level device - * spinlock dev->queue_lock. + * spinlock queue->lock. * - ingress filtering is serialized via top level device * spinlock dev->ingress_lock. * - updates to tree and tree walking are only done under the rtnl mutex. */ void qdisc_lock_tree(struct net_device *dev) - __acquires(dev->queue_lock) + __acquires(dev->tx_queue.lock) __acquires(dev->ingress_lock) { - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); spin_lock(&dev->ingress_lock); } EXPORT_SYMBOL(qdisc_lock_tree); void qdisc_unlock_tree(struct net_device *dev) __releases(dev->ingress_lock) - __releases(dev->queue_lock) + __releases(dev->tx_queue.lock) { spin_unlock(&dev->ingress_lock); - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); } EXPORT_SYMBOL(qdisc_unlock_tree); @@ -118,15 +118,15 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, } /* - * NOTE: Called under dev->queue_lock with locally disabled BH. + * NOTE: Called under queue->lock with locally disabled BH. * * __LINK_STATE_QDISC_RUNNING guarantees only one CPU can process this - * device at a time. dev->queue_lock serializes queue accesses for + * device at a time. queue->lock serializes queue accesses for * this device AND dev->qdisc pointer itself. * * netif_tx_lock serializes accesses to device driver. * - * dev->queue_lock and netif_tx_lock are mutually exclusive, + * queue->lock and netif_tx_lock are mutually exclusive, * if one is grabbed, another must be free. * * Note, that this procedure can be called by a watchdog timer @@ -148,14 +148,14 @@ static inline int qdisc_restart(struct net_device *dev) /* And release queue */ - spin_unlock(&dev->queue_lock); + spin_unlock(&q->dev_queue->lock); HARD_TX_LOCK(dev, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); HARD_TX_UNLOCK(dev); - spin_lock(&dev->queue_lock); + spin_lock(&q->dev_queue->lock); q = dev->qdisc; switch (ret) { @@ -482,7 +482,7 @@ struct Qdisc * qdisc_create_dflt(struct net_device *dev, sch = qdisc_alloc(dev_queue, ops); if (IS_ERR(sch)) goto errout; - sch->stats_lock = &dev->queue_lock; + sch->stats_lock = &dev_queue->lock; sch->parent = parentid; if (!ops->init || ops->init(sch, NULL) == 0) @@ -494,7 +494,7 @@ errout: } EXPORT_SYMBOL(qdisc_create_dflt); -/* Under dev->queue_lock and BH! */ +/* Under queue->lock and BH! */ void qdisc_reset(struct Qdisc *qdisc) { @@ -514,7 +514,7 @@ static void __qdisc_destroy(struct rcu_head *head) kfree((char *) qdisc - qdisc->padded); } -/* Under dev->queue_lock and BH! */ +/* Under queue->lock and BH! */ void qdisc_destroy(struct Qdisc *qdisc) { @@ -566,13 +566,13 @@ void dev_activate(struct net_device *dev) /* Delay activation until next carrier-on event */ return; - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); rcu_assign_pointer(dev->qdisc, dev->qdisc_sleeping); if (dev->qdisc != &noqueue_qdisc) { dev->trans_start = jiffies; dev_watchdog_up(dev); } - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); } void dev_deactivate(struct net_device *dev) @@ -581,7 +581,7 @@ void dev_deactivate(struct net_device *dev) struct sk_buff *skb; int running; - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); qdisc = dev->qdisc; dev->qdisc = &noop_qdisc; @@ -589,7 +589,7 @@ void dev_deactivate(struct net_device *dev) skb = dev->gso_skb; dev->gso_skb = NULL; - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); kfree_skb(skb); @@ -607,9 +607,9 @@ void dev_deactivate(struct net_device *dev) * Double-check inside queue lock to ensure that all effects * of the queue run are visible when we return. */ - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); running = test_bit(__LINK_STATE_QDISC_RUNNING, &dev->state); - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); /* * The running flag should never be set at this point because diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 333525422f45..997d520ca580 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1045,7 +1045,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE]); return 0; } @@ -1104,7 +1104,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); + &sch->dev_queue->lock, tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; } diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 31f7d1536e6d..c8ca54cc26b0 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1043,7 +1043,7 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) struct nlattr *nest; struct tc_htb_glob gopt; - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); gopt.direct_pkts = q->direct_pkts; gopt.version = HTB_VER; @@ -1057,11 +1057,11 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) NLA_PUT(skb, TCA_HTB_INIT, sizeof(gopt), &gopt); nla_nest_end(skb, nest); - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); return skb->len; nla_put_failure: - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); nla_nest_cancel(skb, nest); return -1; } @@ -1073,7 +1073,7 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, struct nlattr *nest; struct tc_htb_opt opt; - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); tcm->tcm_parent = cl->parent ? cl->parent->common.classid : TC_H_ROOT; tcm->tcm_handle = cl->common.classid; if (!cl->level && cl->un.leaf.q) @@ -1095,11 +1095,11 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, NLA_PUT(skb, TCA_HTB_PARMS, sizeof(opt), &opt); nla_nest_end(skb, nest); - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); return skb->len; nla_put_failure: - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); nla_nest_cancel(skb, nest); return -1; } @@ -1365,7 +1365,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, goto failure; gen_new_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE] ? : &est.nla); cl->refcnt = 1; cl->children = 0; @@ -1420,7 +1420,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, } else { if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE]); sch_tree_lock(sch); } diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 790582960444..71b73c528f9b 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -333,9 +333,9 @@ static int get_dist_table(struct Qdisc *sch, const struct nlattr *attr) for (i = 0; i < n; i++) d->table[i] = data[i]; - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); d = xchg(&q->delay_dist, d); - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); kfree(d); return 0; diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index b3fc82623fc6..4f3054e8e1ab 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -154,9 +154,9 @@ teql_destroy(struct Qdisc* sch) master->slaves = NEXT_SLAVE(q); if (q == master->slaves) { master->slaves = NULL; - spin_lock_bh(&master->dev->queue_lock); + spin_lock_bh(&master->dev->tx_queue.lock); qdisc_reset(master->dev->qdisc); - spin_unlock_bh(&master->dev->queue_lock); + spin_unlock_bh(&master->dev->tx_queue.lock); } } skb_queue_purge(&dat->q); -- cgit v1.3-8-gc7d7 From 555353cfa1aee293de445bfa6de43276138ddd82 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 17:33:13 -0700 Subject: netdev: The ingress_lock member is no longer needed. Every qdisc is assosciated with a queue, and in the case of ingress qdiscs that will now be netdev->rx_queue so using that queue's lock is the thing to do. Signed-off-by: David S. Miller --- drivers/net/ifb.c | 8 ++++---- include/linux/netdevice.h | 2 -- net/core/dev.c | 12 +++++++----- net/sched/sch_api.c | 3 +-- net/sched/sch_generic.c | 10 +++++----- 5 files changed, 17 insertions(+), 18 deletions(-) (limited to 'net/core') diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c index bc3de272a829..ccbd6554f6eb 100644 --- a/drivers/net/ifb.c +++ b/drivers/net/ifb.c @@ -229,13 +229,13 @@ module_param(numifbs, int, 0); MODULE_PARM_DESC(numifbs, "Number of ifb devices"); /* - * dev_ifb->tx_queue.lock is usually taken after dev->ingress_lock, + * dev_ifb->tx_queue.lock is usually taken after dev->rx_queue.lock, * reversely to e.g. qdisc_lock_tree(). It should be safe until - * ifb doesn't take dev->tx_queue.lock with dev_ifb->ingress_lock. + * ifb doesn't take dev->tx_queue.lock with dev_ifb->rx_queue.lock. * But lockdep should know that ifb has different locks from dev. */ static struct lock_class_key ifb_tx_queue_lock_key; -static struct lock_class_key ifb_ingress_lock_key; +static struct lock_class_key ifb_rx_queue_lock_key; static int __init ifb_init_one(int index) @@ -259,7 +259,7 @@ static int __init ifb_init_one(int index) goto err; lockdep_set_class(&dev_ifb->tx_queue.lock, &ifb_tx_queue_lock_key); - lockdep_set_class(&dev_ifb->ingress_lock, &ifb_ingress_lock_key); + lockdep_set_class(&dev_ifb->rx_queue.lock, &ifb_rx_queue_lock_key); return 0; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index e835acacb479..633a44c6fa5e 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -632,8 +632,6 @@ struct net_device struct netdev_queue rx_queue; struct netdev_queue tx_queue ____cacheline_aligned_in_smp; - /* ingress path synchronizer */ - spinlock_t ingress_lock; struct Qdisc *qdisc_ingress; /* diff --git a/net/core/dev.c b/net/core/dev.c index 05011048b86c..2322fb69fd53 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2014,10 +2014,11 @@ static inline struct sk_buff *handle_macvlan(struct sk_buff *skb, */ static int ing_filter(struct sk_buff *skb) { - struct Qdisc *q; struct net_device *dev = skb->dev; - int result = TC_ACT_OK; u32 ttl = G_TC_RTTL(skb->tc_verd); + struct netdev_queue *rxq; + int result = TC_ACT_OK; + struct Qdisc *q; if (MAX_RED_LOOP < ttl++) { printk(KERN_WARNING @@ -2029,10 +2030,12 @@ static int ing_filter(struct sk_buff *skb) skb->tc_verd = SET_TC_RTTL(skb->tc_verd, ttl); skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_INGRESS); - spin_lock(&dev->ingress_lock); + rxq = &dev->rx_queue; + + spin_lock(&rxq->lock); if ((q = dev->qdisc_ingress) != NULL) result = q->enqueue(skb, q); - spin_unlock(&dev->ingress_lock); + spin_unlock(&rxq->lock); return result; } @@ -3795,7 +3798,6 @@ int register_netdevice(struct net_device *dev) spin_lock_init(&dev->_xmit_lock); netdev_set_lockdep_class(&dev->_xmit_lock, dev->type); dev->xmit_lock_owner = -1; - spin_lock_init(&dev->ingress_lock); dev->iflink = -1; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 2a1834f8c7d8..570cef2a9c5f 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -601,12 +601,11 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, sch->parent = parent; + sch->stats_lock = &dev_queue->lock; if (handle == TC_H_INGRESS) { sch->flags |= TCQ_F_INGRESS; - sch->stats_lock = &dev->ingress_lock; handle = TC_H_MAKE(TC_H_INGRESS, 0); } else { - sch->stats_lock = &dev_queue->lock; if (handle == 0) { handle = qdisc_alloc_handle(dev); err = -ENOMEM; diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index ee8f9f78a095..804d44b00348 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -35,24 +35,24 @@ * - enqueue, dequeue are serialized via top level device * spinlock queue->lock. * - ingress filtering is serialized via top level device - * spinlock dev->ingress_lock. + * spinlock dev->rx_queue.lock. * - updates to tree and tree walking are only done under the rtnl mutex. */ void qdisc_lock_tree(struct net_device *dev) __acquires(dev->tx_queue.lock) - __acquires(dev->ingress_lock) + __acquires(dev->rx_queue.lock) { spin_lock_bh(&dev->tx_queue.lock); - spin_lock(&dev->ingress_lock); + spin_lock(&dev->rx_queue.lock); } EXPORT_SYMBOL(qdisc_lock_tree); void qdisc_unlock_tree(struct net_device *dev) - __releases(dev->ingress_lock) + __releases(dev->rx_queue.lock) __releases(dev->tx_queue.lock) { - spin_unlock(&dev->ingress_lock); + spin_unlock(&dev->rx_queue.lock); spin_unlock_bh(&dev->tx_queue.lock); } EXPORT_SYMBOL(qdisc_unlock_tree); -- cgit v1.3-8-gc7d7 From b0e1e6462df3c5944010b3328a546d8fe5d932cd Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 17:42:10 -0700 Subject: netdev: Move rest of qdisc state into struct netdev_queue Now qdisc, qdisc_sleeping, and qdisc_list also live there. Signed-off-by: David S. Miller --- drivers/isdn/i4l/isdn_net.c | 2 +- include/linux/netdevice.h | 10 ++--- include/net/irda/irda_device.h | 2 +- net/core/dev.c | 4 +- net/core/link_watch.c | 8 +++- net/core/rtnetlink.c | 6 ++- net/ipv6/addrconf.c | 3 +- net/mac80211/wme.c | 20 ++++++---- net/sched/cls_api.c | 7 +++- net/sched/sch_api.c | 34 ++++++++++------ net/sched/sch_generic.c | 90 ++++++++++++++++++++++++++---------------- net/sched/sch_netem.c | 2 +- net/sched/sch_teql.c | 14 ++++--- 13 files changed, 125 insertions(+), 77 deletions(-) (limited to 'net/core') diff --git a/drivers/isdn/i4l/isdn_net.c b/drivers/isdn/i4l/isdn_net.c index ef1a300068dc..457bbd119f9b 100644 --- a/drivers/isdn/i4l/isdn_net.c +++ b/drivers/isdn/i4l/isdn_net.c @@ -287,7 +287,7 @@ isdn_net_unbind_channel(isdn_net_local * lp) BEWARE! This chunk of code cannot be called from hardware interrupt handler. I hope it is true. --ANK */ - qdisc_reset(lp->netdev->dev->qdisc); + qdisc_reset(lp->netdev->dev->tx_queue.qdisc); } lp->dialstate = 0; dev->rx_netdev[isdn_dc2minor(lp->isdn_device, lp->isdn_channel)] = NULL; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 633a44c6fa5e..df702a7b3db5 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -451,6 +451,9 @@ static inline void napi_synchronize(const struct napi_struct *n) struct netdev_queue { spinlock_t lock; struct net_device *dev; + struct Qdisc *qdisc; + struct Qdisc *qdisc_sleeping; + struct list_head qdisc_list; }; /* @@ -634,13 +637,6 @@ struct net_device struct Qdisc *qdisc_ingress; -/* - * Cache line mostly used on queue transmit path (qdisc) - */ - /* device queue lock */ - struct Qdisc *qdisc; - struct Qdisc *qdisc_sleeping; - struct list_head qdisc_list; unsigned long tx_queue_len; /* Max frames per queue allowed */ /* Partially transmitted GSO packet. */ diff --git a/include/net/irda/irda_device.h b/include/net/irda/irda_device.h index f70e9b39ebaf..16fbf672e0b2 100644 --- a/include/net/irda/irda_device.h +++ b/include/net/irda/irda_device.h @@ -223,7 +223,7 @@ int irda_device_is_receiving(struct net_device *dev); /* Interface for internal use */ static inline int irda_device_txqueue_empty(const struct net_device *dev) { - return skb_queue_empty(&dev->qdisc->q); + return skb_queue_empty(&dev->tx_queue.qdisc->q); } int irda_device_set_raw_mode(struct net_device* self, int status); struct net_device *alloc_irdadev(int sizeof_priv); diff --git a/net/core/dev.c b/net/core/dev.c index 2322fb69fd53..ce79c28d739d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1720,14 +1720,14 @@ gso: * also serializes access to the device queue. */ - q = rcu_dereference(dev->qdisc); + q = rcu_dereference(txq->qdisc); #ifdef CONFIG_NET_CLS_ACT skb->tc_verd = SET_TC_AT(skb->tc_verd,AT_EGRESS); #endif if (q->enqueue) { /* Grab device queue */ spin_lock(&txq->lock); - q = dev->qdisc; + q = txq->qdisc; if (q->enqueue) { /* reset queue_mapping to zero */ skb_set_queue_mapping(skb, 0); diff --git a/net/core/link_watch.c b/net/core/link_watch.c index a5e372b9ec4d..50218218445b 100644 --- a/net/core/link_watch.c +++ b/net/core/link_watch.c @@ -79,8 +79,10 @@ static void rfc2863_policy(struct net_device *dev) static int linkwatch_urgent_event(struct net_device *dev) { + struct netdev_queue *txq = &dev->tx_queue; + return netif_running(dev) && netif_carrier_ok(dev) && - dev->qdisc != dev->qdisc_sleeping; + txq->qdisc != txq->qdisc_sleeping; } @@ -181,7 +183,9 @@ static void __linkwatch_run_queue(int urgent_only) rfc2863_policy(dev); if (dev->flags & IFF_UP) { if (netif_carrier_ok(dev)) { - WARN_ON(dev->qdisc_sleeping == &noop_qdisc); + struct netdev_queue *txq = &dev->tx_queue; + + WARN_ON(txq->qdisc_sleeping == &noop_qdisc); dev_activate(dev); } else dev_deactivate(dev); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 6c8d7f0ea01a..8ef9f1db610e 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -605,6 +605,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, int type, u32 pid, u32 seq, u32 change, unsigned int flags) { + struct netdev_queue *txq; struct ifinfomsg *ifm; struct nlmsghdr *nlh; struct net_device_stats *stats; @@ -635,8 +636,9 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, if (dev->master) NLA_PUT_U32(skb, IFLA_MASTER, dev->master->ifindex); - if (dev->qdisc_sleeping) - NLA_PUT_STRING(skb, IFLA_QDISC, dev->qdisc_sleeping->ops->id); + txq = &dev->tx_queue; + if (txq->qdisc_sleeping) + NLA_PUT_STRING(skb, IFLA_QDISC, txq->qdisc_sleeping->ops->id); if (1) { struct rtnl_link_ifmap map = { diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 8572cb05fc21..5c84c798331d 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -231,7 +231,8 @@ const struct in6_addr in6addr_linklocal_allrouters = IN6ADDR_LINKLOCAL_ALLROUTER /* Check if a valid qdisc is available */ static inline int addrconf_qdisc_ok(struct net_device *dev) { - return (dev->qdisc != &noop_qdisc); + struct netdev_queue *txq = &dev->tx_queue; + return (txq->qdisc != &noop_qdisc); } /* Check if a route is valid prefix route */ diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 59ed9cae66b9..6ae43a3c7726 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -574,9 +574,10 @@ static struct Qdisc_ops wme_qdisc_ops __read_mostly = void ieee80211_install_qdisc(struct net_device *dev) { + struct netdev_queue *txq = &dev->tx_queue; struct Qdisc *qdisc; - qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + qdisc = qdisc_create_dflt(dev, txq, &wme_qdisc_ops, TC_H_ROOT); if (!qdisc) { printk(KERN_ERR "%s: qdisc installation failed\n", dev->name); @@ -587,15 +588,17 @@ void ieee80211_install_qdisc(struct net_device *dev) qdisc->handle = 0x80010000; qdisc_lock_tree(dev); - list_add_tail(&qdisc->list, &dev->qdisc_list); - dev->qdisc_sleeping = qdisc; + list_add_tail(&qdisc->list, &txq->qdisc_list); + txq->qdisc_sleeping = qdisc; qdisc_unlock_tree(dev); } int ieee80211_qdisc_installed(struct net_device *dev) { - return dev->qdisc_sleeping->ops == &wme_qdisc_ops; + struct netdev_queue *txq = &dev->tx_queue; + + return txq->qdisc_sleeping->ops == &wme_qdisc_ops; } @@ -614,8 +617,9 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, struct sta_info *sta, u16 tid) { int i; + struct netdev_queue *txq = &local->mdev->tx_queue; struct ieee80211_sched_data *q = - qdisc_priv(local->mdev->qdisc_sleeping); + qdisc_priv(txq->qdisc_sleeping); DECLARE_MAC_BUF(mac); /* prepare the filter and save it for the SW queue @@ -655,8 +659,9 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, u8 requeue) { struct ieee80211_hw *hw = &local->hw; + struct netdev_queue *txq = &local->mdev->tx_queue; struct ieee80211_sched_data *q = - qdisc_priv(local->mdev->qdisc_sleeping); + qdisc_priv(txq->qdisc_sleeping); int agg_queue = sta->tid_to_tx_q[tid]; /* return the qdisc to the pool */ @@ -671,7 +676,8 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, void ieee80211_requeue(struct ieee80211_local *local, int queue) { - struct Qdisc *root_qd = local->mdev->qdisc_sleeping; + struct netdev_queue *txq = &local->mdev->tx_queue; + struct Qdisc *root_qd = txq->qdisc_sleeping; struct ieee80211_sched_data *q = qdisc_priv(root_qd); struct Qdisc *qdisc = q->queues[queue]; struct sk_buff *skb = NULL; diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index e2389f161e46..b483bbea6118 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -166,7 +166,8 @@ replay: /* Find qdisc */ if (!parent) { - q = dev->qdisc_sleeping; + struct netdev_queue *dev_queue = &dev->tx_queue; + q = dev_queue->qdisc_sleeping; parent = q->handle; } else { q = qdisc_lookup(dev, TC_H_MAJ(t->tcm_parent)); @@ -390,6 +391,7 @@ static int tcf_node_dump(struct tcf_proto *tp, unsigned long n, static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); + struct netdev_queue *dev_queue; int t; int s_t; struct net_device *dev; @@ -408,8 +410,9 @@ static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb) if ((dev = dev_get_by_index(&init_net, tcm->tcm_ifindex)) == NULL) return skb->len; + dev_queue = &dev->tx_queue; if (!tcm->tcm_parent) - q = dev->qdisc_sleeping; + q = dev_queue->qdisc_sleeping; else q = qdisc_lookup(dev, TC_H_MAJ(tcm->tcm_parent)); if (!q) diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 570cef2a9c5f..2313fa7c97be 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -185,9 +185,10 @@ EXPORT_SYMBOL(unregister_qdisc); struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) { + struct netdev_queue *dev_queue = &dev->tx_queue; struct Qdisc *q; - list_for_each_entry(q, &dev->qdisc_list, list) { + list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (q->handle == handle) return q; } @@ -441,6 +442,7 @@ static u32 qdisc_alloc_handle(struct net_device *dev) static struct Qdisc * dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) { + struct netdev_queue *dev_queue; struct Qdisc *oqdisc; if (dev->flags & IFF_UP) @@ -459,8 +461,8 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) } } else { - - oqdisc = dev->qdisc_sleeping; + dev_queue = &dev->tx_queue; + oqdisc = dev_queue->qdisc_sleeping; /* Prune old scheduler */ if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) @@ -469,8 +471,8 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) /* ... and graft new one */ if (qdisc == NULL) qdisc = &noop_qdisc; - dev->qdisc_sleeping = qdisc; - dev->qdisc = &noop_qdisc; + dev_queue->qdisc_sleeping = qdisc; + dev_queue->qdisc = &noop_qdisc; } qdisc_unlock_tree(dev); @@ -633,7 +635,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, } } qdisc_lock_tree(dev); - list_add_tail(&sch->list, &dev->qdisc_list); + list_add_tail(&sch->list, &dev_queue->qdisc_list); qdisc_unlock_tree(dev); return sch; @@ -740,7 +742,8 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) q = dev->qdisc_ingress; } } else { - q = dev->qdisc_sleeping; + struct netdev_queue *dev_queue = &dev->tx_queue; + q = dev_queue->qdisc_sleeping; } if (!q) return -ENOENT; @@ -814,7 +817,8 @@ replay: q = dev->qdisc_ingress; } } else { - q = dev->qdisc_sleeping; + struct netdev_queue *dev_queue = &dev->tx_queue; + q = dev_queue->qdisc_sleeping; } /* It may be default qdisc, ignore it */ @@ -1015,12 +1019,14 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) read_lock(&dev_base_lock); idx = 0; for_each_netdev(&init_net, dev) { + struct netdev_queue *dev_queue; if (idx < s_idx) goto cont; if (idx > s_idx) s_q_idx = 0; q_idx = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { + dev_queue = &dev->tx_queue; + list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (q_idx < s_q_idx) { q_idx++; continue; @@ -1054,6 +1060,7 @@ done: static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) { struct net *net = sock_net(skb->sk); + struct netdev_queue *dev_queue; struct tcmsg *tcm = NLMSG_DATA(n); struct nlattr *tca[TCA_MAX + 1]; struct net_device *dev; @@ -1091,6 +1098,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) /* Step 1. Determine qdisc handle X:0 */ + dev_queue = &dev->tx_queue; if (pid != TC_H_ROOT) { u32 qid1 = TC_H_MAJ(pid); @@ -1101,7 +1109,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) } else if (qid1) { qid = qid1; } else if (qid == 0) - qid = dev->qdisc_sleeping->handle; + qid = dev_queue->qdisc_sleeping->handle; /* Now qid is genuine qdisc handle consistent both with parent and child. @@ -1112,7 +1120,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) pid = TC_H_MAKE(qid, pid); } else { if (qid == 0) - qid = dev->qdisc_sleeping->handle; + qid = dev_queue->qdisc_sleeping->handle; } /* OK. Locate qdisc */ @@ -1248,6 +1256,7 @@ static int qdisc_class_dump(struct Qdisc *q, unsigned long cl, struct qdisc_walk static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); + struct netdev_queue *dev_queue; int t; int s_t; struct net_device *dev; @@ -1266,7 +1275,8 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { + dev_queue = &dev->tx_queue; + list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (t < s_t || !q->ops->cl_ops || (tcm->tcm_parent && TC_H_MAJ(tcm->tcm_parent) != q->handle)) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 804d44b00348..3223e5ba76aa 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -122,7 +122,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * * __LINK_STATE_QDISC_RUNNING guarantees only one CPU can process this * device at a time. queue->lock serializes queue accesses for - * this device AND dev->qdisc pointer itself. + * this device AND txq->qdisc pointer itself. * * netif_tx_lock serializes accesses to device driver. * @@ -138,7 +138,8 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, */ static inline int qdisc_restart(struct net_device *dev) { - struct Qdisc *q = dev->qdisc; + struct netdev_queue *txq = &dev->tx_queue; + struct Qdisc *q = txq->qdisc; struct sk_buff *skb; int ret = NETDEV_TX_BUSY; @@ -148,15 +149,15 @@ static inline int qdisc_restart(struct net_device *dev) /* And release queue */ - spin_unlock(&q->dev_queue->lock); + spin_unlock(&txq->lock); HARD_TX_LOCK(dev, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); HARD_TX_UNLOCK(dev); - spin_lock(&q->dev_queue->lock); - q = dev->qdisc; + spin_lock(&txq->lock); + q = txq->qdisc; switch (ret) { case NETDEV_TX_OK: @@ -207,9 +208,10 @@ void __qdisc_run(struct net_device *dev) static void dev_watchdog(unsigned long arg) { struct net_device *dev = (struct net_device *)arg; + struct netdev_queue *txq = &dev->tx_queue; netif_tx_lock(dev); - if (dev->qdisc != &noop_qdisc) { + if (txq->qdisc != &noop_qdisc) { if (netif_device_present(dev) && netif_running(dev) && netif_carrier_ok(dev)) { @@ -539,53 +541,63 @@ EXPORT_SYMBOL(qdisc_destroy); void dev_activate(struct net_device *dev) { + struct netdev_queue *txq = &dev->tx_queue; + /* No queueing discipline is attached to device; create default one i.e. pfifo_fast for devices, which need queueing and noqueue_qdisc for virtual interfaces */ - if (dev->qdisc_sleeping == &noop_qdisc) { + if (txq->qdisc_sleeping == &noop_qdisc) { struct Qdisc *qdisc; if (dev->tx_queue_len) { - qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + qdisc = qdisc_create_dflt(dev, txq, &pfifo_fast_ops, TC_H_ROOT); if (qdisc == NULL) { printk(KERN_INFO "%s: activation failed\n", dev->name); return; } - list_add_tail(&qdisc->list, &dev->qdisc_list); + list_add_tail(&qdisc->list, &txq->qdisc_list); } else { qdisc = &noqueue_qdisc; } - dev->qdisc_sleeping = qdisc; + txq->qdisc_sleeping = qdisc; } if (!netif_carrier_ok(dev)) /* Delay activation until next carrier-on event */ return; - spin_lock_bh(&dev->tx_queue.lock); - rcu_assign_pointer(dev->qdisc, dev->qdisc_sleeping); - if (dev->qdisc != &noqueue_qdisc) { + spin_lock_bh(&txq->lock); + rcu_assign_pointer(txq->qdisc, txq->qdisc_sleeping); + if (txq->qdisc != &noqueue_qdisc) { dev->trans_start = jiffies; dev_watchdog_up(dev); } - spin_unlock_bh(&dev->tx_queue.lock); + spin_unlock_bh(&txq->lock); +} + +static void dev_deactivate_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc *qdisc_default) +{ + struct Qdisc *qdisc = dev_queue->qdisc; + + if (qdisc) { + dev_queue->qdisc = qdisc_default; + qdisc_reset(qdisc); + } } void dev_deactivate(struct net_device *dev) { - struct Qdisc *qdisc; struct sk_buff *skb; int running; spin_lock_bh(&dev->tx_queue.lock); - qdisc = dev->qdisc; - dev->qdisc = &noop_qdisc; - - qdisc_reset(qdisc); + dev_deactivate_queue(dev, &dev->tx_queue, &noop_qdisc); skb = dev->gso_skb; dev->gso_skb = NULL; @@ -622,32 +634,44 @@ void dev_deactivate(struct net_device *dev) } while (WARN_ON_ONCE(running)); } +static void dev_init_scheduler_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc *qdisc) +{ + dev_queue->qdisc = qdisc; + dev_queue->qdisc_sleeping = qdisc; + INIT_LIST_HEAD(&dev_queue->qdisc_list); +} + void dev_init_scheduler(struct net_device *dev) { qdisc_lock_tree(dev); - dev->qdisc = &noop_qdisc; - dev->qdisc_sleeping = &noop_qdisc; - INIT_LIST_HEAD(&dev->qdisc_list); + dev_init_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); + dev_init_scheduler_queue(dev, &dev->rx_queue, NULL); qdisc_unlock_tree(dev); setup_timer(&dev->watchdog_timer, dev_watchdog, (unsigned long)dev); } -void dev_shutdown(struct net_device *dev) +static void dev_shutdown_scheduler_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc *qdisc_default) { - struct Qdisc *qdisc; + struct Qdisc *qdisc = dev_queue->qdisc_sleeping; + + if (qdisc) { + dev_queue->qdisc = qdisc_default; + dev_queue->qdisc_sleeping = qdisc_default; - qdisc_lock_tree(dev); - qdisc = dev->qdisc_sleeping; - dev->qdisc = &noop_qdisc; - dev->qdisc_sleeping = &noop_qdisc; - qdisc_destroy(qdisc); -#if defined(CONFIG_NET_SCH_INGRESS) || defined(CONFIG_NET_SCH_INGRESS_MODULE) - if ((qdisc = dev->qdisc_ingress) != NULL) { - dev->qdisc_ingress = NULL; qdisc_destroy(qdisc); } -#endif +} + +void dev_shutdown(struct net_device *dev) +{ + qdisc_lock_tree(dev); + dev_shutdown_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); + dev_shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); BUG_TRAP(!timer_pending(&dev->watchdog_timer)); qdisc_unlock_tree(dev); } diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 71b73c528f9b..4093f1eaaf60 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -180,7 +180,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) * skb will be queued. */ if (count > 1 && (skb2 = skb_clone(skb, GFP_ATOMIC)) != NULL) { - struct Qdisc *rootq = qdisc_dev(sch)->qdisc; + struct Qdisc *rootq = qdisc_dev(sch)->tx_queue.qdisc; u32 dupsave = q->duplicate; /* prevent duplicating a dup... */ q->duplicate = 0; diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 4f3054e8e1ab..8ac05981be20 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -107,17 +107,19 @@ static struct sk_buff * teql_dequeue(struct Qdisc* sch) { struct teql_sched_data *dat = qdisc_priv(sch); + struct netdev_queue *dat_queue; struct sk_buff *skb; skb = __skb_dequeue(&dat->q); + dat_queue = &dat->m->dev->tx_queue; if (skb == NULL) { - struct net_device *m = qdisc_dev(dat->m->dev->qdisc); + struct net_device *m = qdisc_dev(dat_queue->qdisc); if (m) { dat->m->slaves = sch; netif_wake_queue(m); } } - sch->q.qlen = dat->q.qlen + dat->m->dev->qdisc->q.qlen; + sch->q.qlen = dat->q.qlen + dat_queue->qdisc->q.qlen; return skb; } @@ -155,7 +157,7 @@ teql_destroy(struct Qdisc* sch) if (q == master->slaves) { master->slaves = NULL; spin_lock_bh(&master->dev->tx_queue.lock); - qdisc_reset(master->dev->qdisc); + qdisc_reset(master->dev->tx_queue.qdisc); spin_unlock_bh(&master->dev->tx_queue.lock); } } @@ -216,7 +218,7 @@ static int teql_qdisc_init(struct Qdisc *sch, struct nlattr *opt) static int __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - struct teql_sched_data *q = qdisc_priv(dev->qdisc); + struct teql_sched_data *q = qdisc_priv(dev->tx_queue.qdisc); struct neighbour *mn = skb->dst->neighbour; struct neighbour *n = q->ncache; @@ -252,7 +254,7 @@ __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device * static inline int teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - if (dev->qdisc == &noop_qdisc) + if (dev->tx_queue.qdisc == &noop_qdisc) return -ENODEV; if (dev->header_ops == NULL || @@ -284,7 +286,7 @@ restart: do { struct net_device *slave = qdisc_dev(q); - if (slave->qdisc_sleeping != q) + if (slave->tx_queue.qdisc_sleeping != q) continue; if (netif_queue_stopped(slave) || __netif_subqueue_stopped(slave, subq) || -- cgit v1.3-8-gc7d7 From 816f3258e70db38d6d92c8d871377179fd69160f Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 22:49:00 -0700 Subject: netdev: Kill qdisc_ingress, use netdev->rx_queue.qdisc instead. Now that our qdisc management is bi-directional, per-queue, and fully orthogonal, there is no reason to have a special ingress qdisc pointer in struct net_device. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 3 --- net/core/dev.c | 4 ++-- net/sched/sch_api.c | 11 ++++++----- 3 files changed, 8 insertions(+), 10 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index df702a7b3db5..e7c49246fd88 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -634,9 +634,6 @@ struct net_device struct netdev_queue rx_queue; struct netdev_queue tx_queue ____cacheline_aligned_in_smp; - - struct Qdisc *qdisc_ingress; - unsigned long tx_queue_len; /* Max frames per queue allowed */ /* Partially transmitted GSO packet. */ diff --git a/net/core/dev.c b/net/core/dev.c index ce79c28d739d..ab760a954d99 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2033,7 +2033,7 @@ static int ing_filter(struct sk_buff *skb) rxq = &dev->rx_queue; spin_lock(&rxq->lock); - if ((q = dev->qdisc_ingress) != NULL) + if ((q = rxq->qdisc) != NULL) result = q->enqueue(skb, q); spin_unlock(&rxq->lock); @@ -2044,7 +2044,7 @@ static inline struct sk_buff *handle_ing(struct sk_buff *skb, struct packet_type **pt_prev, int *ret, struct net_device *orig_dev) { - if (!skb->dev->qdisc_ingress) + if (!skb->dev->rx_queue.qdisc) goto out; if (*pt_prev) { diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 2313fa7c97be..4003c280b69f 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -450,14 +450,15 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) qdisc_lock_tree(dev); if (qdisc && qdisc->flags&TCQ_F_INGRESS) { - oqdisc = dev->qdisc_ingress; + dev_queue = &dev->rx_queue; + oqdisc = dev_queue->qdisc; /* Prune old scheduler */ if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) { /* delete */ qdisc_reset(oqdisc); - dev->qdisc_ingress = NULL; + dev_queue->qdisc = NULL; } else { /* new */ - dev->qdisc_ingress = qdisc; + dev_queue->qdisc = qdisc; } } else { @@ -739,7 +740,7 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) return -ENOENT; q = qdisc_leaf(p, clid); } else { /* ingress */ - q = dev->qdisc_ingress; + q = dev->rx_queue.qdisc; } } else { struct netdev_queue *dev_queue = &dev->tx_queue; @@ -814,7 +815,7 @@ replay: return -ENOENT; q = qdisc_leaf(p, clid); } else { /*ingress */ - q = dev->qdisc_ingress; + q = dev->rx_queue.qdisc; } } else { struct netdev_queue *dev_queue = &dev->tx_queue; -- cgit v1.3-8-gc7d7 From ee609cb36220d18c0cf476b066a5ab7e6f6d3a69 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 22:58:37 -0700 Subject: netdev: Move next_sched into struct netdev_queue. We schedule queues, not the device, for output queue processing in BH. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 5 ++--- net/core/dev.c | 15 +++++++-------- 2 files changed, 9 insertions(+), 11 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index e7c49246fd88..1379c822e51d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -454,6 +454,7 @@ struct netdev_queue { struct Qdisc *qdisc; struct Qdisc *qdisc_sleeping; struct list_head qdisc_list; + struct netdev_queue *next_sched; }; /* @@ -545,8 +546,6 @@ struct net_device #define NETIF_F_V6_CSUM (NETIF_F_GEN_CSUM | NETIF_F_IPV6_CSUM) #define NETIF_F_ALL_CSUM (NETIF_F_V4_CSUM | NETIF_F_V6_CSUM) - struct net_device *next_sched; - /* Interface index. Unique device identifier */ int ifindex; int iflink; @@ -940,7 +939,7 @@ static inline int unregister_gifconf(unsigned int family) */ struct softnet_data { - struct net_device *output_queue; + struct netdev_queue *output_queue; struct sk_buff_head input_pkt_queue; struct list_head poll_list; struct sk_buff *completion_queue; diff --git a/net/core/dev.c b/net/core/dev.c index ab760a954d99..d6b8d3c3e6ec 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1323,13 +1323,14 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) void __netif_schedule(struct net_device *dev) { if (!test_and_set_bit(__LINK_STATE_SCHED, &dev->state)) { + struct netdev_queue *txq = &dev->tx_queue; unsigned long flags; struct softnet_data *sd; local_irq_save(flags); sd = &__get_cpu_var(softnet_data); - dev->next_sched = sd->output_queue; - sd->output_queue = dev; + txq->next_sched = sd->output_queue; + sd->output_queue = txq; raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_restore(flags); } @@ -1912,7 +1913,7 @@ static void net_tx_action(struct softirq_action *h) } if (sd->output_queue) { - struct net_device *head; + struct netdev_queue *head; local_irq_disable(); head = sd->output_queue; @@ -1920,12 +1921,10 @@ static void net_tx_action(struct softirq_action *h) local_irq_enable(); while (head) { - struct net_device *dev = head; - struct netdev_queue *txq; + struct netdev_queue *txq = head; + struct net_device *dev = txq->dev; head = head->next_sched; - txq = &dev->tx_queue; - smp_mb__before_clear_bit(); clear_bit(__LINK_STATE_SCHED, &dev->state); @@ -4346,7 +4345,7 @@ static int dev_cpu_callback(struct notifier_block *nfb, void *ocpu) { struct sk_buff **list_skb; - struct net_device **list_net; + struct netdev_queue **list_net; struct sk_buff *skb; unsigned int cpu, oldcpu = (unsigned long)ocpu; struct softnet_data *sd, *oldsd; -- cgit v1.3-8-gc7d7 From 6fa9864b53f0680e432a2c431c2cf2055daa3a88 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:01:06 -0700 Subject: net: Clean up explicit ->tx_queue references in link watch. First, we add a qdisc_tx_changing() helper which returns true if the qdisc attachment is in transition. Second, we remove an assertion warning which is of limited value and is hard to express precisely in a multiqueue environment. Signed-off-by: David S. Miller --- include/net/sch_generic.h | 8 ++++++++ net/core/link_watch.c | 15 +++++---------- 2 files changed, 13 insertions(+), 10 deletions(-) (limited to 'net/core') diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index 8cfdaebbbab3..bf8f7264a778 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -242,6 +242,14 @@ static inline bool qdisc_all_tx_empty(const struct net_device *dev) return (q->q.qlen == 0); } +/* Are any of the TX qdiscs changing? */ +static inline bool qdisc_tx_changing(struct net_device *dev) +{ + struct netdev_queue *txq = &dev->tx_queue; + + return (txq->qdisc != txq->qdisc_sleeping); +} + static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch, struct sk_buff_head *list) { diff --git a/net/core/link_watch.c b/net/core/link_watch.c index 50218218445b..bf8f7af699d7 100644 --- a/net/core/link_watch.c +++ b/net/core/link_watch.c @@ -77,12 +77,10 @@ static void rfc2863_policy(struct net_device *dev) } -static int linkwatch_urgent_event(struct net_device *dev) +static bool linkwatch_urgent_event(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; - return netif_running(dev) && netif_carrier_ok(dev) && - txq->qdisc != txq->qdisc_sleeping; + qdisc_tx_changing(dev); } @@ -182,12 +180,9 @@ static void __linkwatch_run_queue(int urgent_only) rfc2863_policy(dev); if (dev->flags & IFF_UP) { - if (netif_carrier_ok(dev)) { - struct netdev_queue *txq = &dev->tx_queue; - - WARN_ON(txq->qdisc_sleeping == &noop_qdisc); + if (netif_carrier_ok(dev)) dev_activate(dev); - } else + else dev_deactivate(dev); netdev_state_change(dev); @@ -218,7 +213,7 @@ static void linkwatch_event(struct work_struct *dummy) void linkwatch_fire_event(struct net_device *dev) { - int urgent = linkwatch_urgent_event(dev); + bool urgent = linkwatch_urgent_event(dev); if (!test_and_set_bit(__LINK_STATE_LINKWATCH_PENDING, &dev->state)) { dev_hold(dev); -- cgit v1.3-8-gc7d7 From 86d804e10a37cd86f16bf72386c37e843a98a74b Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:11:25 -0700 Subject: netdev: Make netif_schedule() routines work with netdev_queue objects. Only plain netif_schedule() remains taking a net_device, mostly as a compatability item while we transition the rest of these interfaces. Everything else calls netif_schedule_queue() or __netif_schedule(), both of which take a netdev_queue pointer. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 17 ++++++++++++----- net/core/dev.c | 9 +++++---- net/mac80211/main.c | 4 ++-- net/sched/sch_api.c | 4 ++-- net/sched/sch_cbq.c | 2 +- net/sched/sch_generic.c | 10 +++++----- 6 files changed, 27 insertions(+), 19 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index aae6c6d153f2..28aa8e77cee9 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -952,12 +952,19 @@ DECLARE_PER_CPU(struct softnet_data,softnet_data); #define HAVE_NETIF_QUEUE -extern void __netif_schedule(struct net_device *dev); +extern void __netif_schedule(struct netdev_queue *txq); -static inline void netif_schedule(struct net_device *dev) +static inline void netif_schedule_queue(struct netdev_queue *txq) { + struct net_device *dev = txq->dev; + if (!test_bit(__LINK_STATE_XOFF, &dev->state)) - __netif_schedule(dev); + __netif_schedule(txq); +} + +static inline void netif_schedule(struct net_device *dev) +{ + netif_schedule_queue(&dev->tx_queue); } /** @@ -987,7 +994,7 @@ static inline void netif_wake_queue(struct net_device *dev) } #endif if (test_and_clear_bit(__LINK_STATE_XOFF, &dev->state)) - __netif_schedule(dev); + __netif_schedule(&dev->tx_queue); } /** @@ -1103,7 +1110,7 @@ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index) #endif if (test_and_clear_bit(__LINK_STATE_XOFF, &dev->egress_subqueue[queue_index].state)) - __netif_schedule(dev); + __netif_schedule(&dev->tx_queue); #endif } diff --git a/net/core/dev.c b/net/core/dev.c index d6b8d3c3e6ec..0dc888ad4217 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1320,12 +1320,13 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) } -void __netif_schedule(struct net_device *dev) +void __netif_schedule(struct netdev_queue *txq) { + struct net_device *dev = txq->dev; + if (!test_and_set_bit(__LINK_STATE_SCHED, &dev->state)) { - struct netdev_queue *txq = &dev->tx_queue; - unsigned long flags; struct softnet_data *sd; + unsigned long flags; local_irq_save(flags); sd = &__get_cpu_var(softnet_data); @@ -1932,7 +1933,7 @@ static void net_tx_action(struct softirq_action *h) qdisc_run(dev); spin_unlock(&txq->lock); } else { - netif_schedule(dev); + netif_schedule_queue(txq); } } } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 2968baa66b91..1c4d3ba6b878 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -885,10 +885,10 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) spin_unlock_bh(&txq->lock); /* we just requeued the all the frames that were in the removed - * queue, and since we might miss a softirq we do netif_schedule. + * queue, and since we might miss a softirq we do netif_schedule_queue. * ieee80211_wake_queue is not used here as this queue is not * necessarily stopped */ - netif_schedule(local->mdev); + netif_schedule_queue(txq); spin_lock_bh(&sta->lock); *state = HT_AGG_STATE_IDLE; sta->ampdu_mlme.addba_req_num[tid] = 0; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index e73bd68aa7ae..95873f8dd37c 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -282,11 +282,11 @@ static enum hrtimer_restart qdisc_watchdog(struct hrtimer *timer) { struct qdisc_watchdog *wd = container_of(timer, struct qdisc_watchdog, timer); - struct net_device *dev = qdisc_dev(wd->qdisc); + struct netdev_queue *txq = wd->qdisc->dev_queue; wd->qdisc->flags &= ~TCQ_F_THROTTLED; smp_wmb(); - netif_schedule(dev); + netif_schedule_queue(txq); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 99ce3da2b0a4..4efc836cbf38 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -650,7 +650,7 @@ static enum hrtimer_restart cbq_undelay(struct hrtimer *timer) } sch->flags &= ~TCQ_F_THROTTLED; - netif_schedule(qdisc_dev(sch)); + netif_schedule_queue(sch->dev_queue); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 8247a406a401..407dfdb142a4 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -62,7 +62,7 @@ static inline int qdisc_qlen(struct Qdisc *q) return q->q.qlen; } -static inline int dev_requeue_skb(struct sk_buff *skb, struct net_device *dev, +static inline int dev_requeue_skb(struct sk_buff *skb, struct netdev_queue *dev_queue, struct Qdisc *q) { @@ -71,7 +71,7 @@ static inline int dev_requeue_skb(struct sk_buff *skb, struct net_device *dev, else q->ops->requeue(skb, q); - netif_schedule(dev); + netif_schedule_queue(dev_queue); return 0; } @@ -114,7 +114,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * some time. */ __get_cpu_var(netdev_rx_stat).cpu_collision++; - ret = dev_requeue_skb(skb, dev, dev_queue, q); + ret = dev_requeue_skb(skb, dev_queue, q); } return ret; @@ -179,7 +179,7 @@ static inline int qdisc_restart(struct net_device *dev) printk(KERN_WARNING "BUG %s code %d qlen %d\n", dev->name, ret, q->q.qlen); - ret = dev_requeue_skb(skb, dev, txq, q); + ret = dev_requeue_skb(skb, txq, q); break; } @@ -200,7 +200,7 @@ void __qdisc_run(struct net_device *dev) * 2. we've been doing it for too long. */ if (need_resched() || jiffies != start_time) { - netif_schedule(dev); + netif_schedule_queue(&dev->tx_queue); break; } } -- cgit v1.3-8-gc7d7 From eb6aafe3f843cb0e939546c03540a3b4911b6964 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:12:38 -0700 Subject: pkt_sched: Make qdisc_run take a netdev_queue. This allows us to use this calling convention all the way down into qdisc_restart(). Signed-off-by: David S. Miller --- include/net/pkt_sched.h | 8 +++++--- net/core/dev.c | 4 ++-- net/sched/sch_generic.c | 26 ++++++++++++++------------ 3 files changed, 21 insertions(+), 17 deletions(-) (limited to 'net/core') diff --git a/include/net/pkt_sched.h b/include/net/pkt_sched.h index 8e3a0c4e9d97..2311d242bb35 100644 --- a/include/net/pkt_sched.h +++ b/include/net/pkt_sched.h @@ -84,13 +84,15 @@ extern struct qdisc_rate_table *qdisc_get_rtab(struct tc_ratespec *r, struct nlattr *tab); extern void qdisc_put_rtab(struct qdisc_rate_table *tab); -extern void __qdisc_run(struct net_device *dev); +extern void __qdisc_run(struct netdev_queue *txq); -static inline void qdisc_run(struct net_device *dev) +static inline void qdisc_run(struct netdev_queue *txq) { + struct net_device *dev = txq->dev; + if (!netif_queue_stopped(dev) && !test_and_set_bit(__LINK_STATE_QDISC_RUNNING, &dev->state)) - __qdisc_run(dev); + __qdisc_run(txq); } extern int tc_classify_compat(struct sk_buff *skb, struct tcf_proto *tp, diff --git a/net/core/dev.c b/net/core/dev.c index 0dc888ad4217..0218b0b9be80 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1734,7 +1734,7 @@ gso: /* reset queue_mapping to zero */ skb_set_queue_mapping(skb, 0); rc = q->enqueue(skb, q); - qdisc_run(dev); + qdisc_run(txq); spin_unlock(&txq->lock); rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; @@ -1930,7 +1930,7 @@ static void net_tx_action(struct softirq_action *h) clear_bit(__LINK_STATE_SCHED, &dev->state); if (spin_trylock(&txq->lock)) { - qdisc_run(dev); + qdisc_run(txq); spin_unlock(&txq->lock); } else { netif_schedule_queue(txq); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 407dfdb142a4..fcc7533f0bcc 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -75,9 +75,8 @@ static inline int dev_requeue_skb(struct sk_buff *skb, return 0; } -static inline struct sk_buff *dev_dequeue_skb(struct net_device *dev, - struct netdev_queue *dev_queue, - struct Qdisc *q) +static inline struct sk_buff *dequeue_skb(struct netdev_queue *dev_queue, + struct Qdisc *q) { struct sk_buff *skb; @@ -90,10 +89,10 @@ static inline struct sk_buff *dev_dequeue_skb(struct net_device *dev, } static inline int handle_dev_cpu_collision(struct sk_buff *skb, - struct net_device *dev, struct netdev_queue *dev_queue, struct Qdisc *q) { + struct net_device *dev = dev_queue->dev; int ret; if (unlikely(dev->xmit_lock_owner == smp_processor_id())) { @@ -139,21 +138,23 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * >0 - queue is not empty. * */ -static inline int qdisc_restart(struct net_device *dev) +static inline int qdisc_restart(struct netdev_queue *txq) { - struct netdev_queue *txq = &dev->tx_queue; struct Qdisc *q = txq->qdisc; - struct sk_buff *skb; int ret = NETDEV_TX_BUSY; + struct net_device *dev; + struct sk_buff *skb; /* Dequeue packet */ - if (unlikely((skb = dev_dequeue_skb(dev, txq, q)) == NULL)) + if (unlikely((skb = dequeue_skb(txq, q)) == NULL)) return 0; /* And release queue */ spin_unlock(&txq->lock); + dev = txq->dev; + HARD_TX_LOCK(dev, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); @@ -170,7 +171,7 @@ static inline int qdisc_restart(struct net_device *dev) case NETDEV_TX_LOCKED: /* Driver try lock failed */ - ret = handle_dev_cpu_collision(skb, dev, txq, q); + ret = handle_dev_cpu_collision(skb, txq, q); break; default: @@ -186,11 +187,12 @@ static inline int qdisc_restart(struct net_device *dev) return ret; } -void __qdisc_run(struct net_device *dev) +void __qdisc_run(struct netdev_queue *txq) { + struct net_device *dev = txq->dev; unsigned long start_time = jiffies; - while (qdisc_restart(dev)) { + while (qdisc_restart(txq)) { if (netif_queue_stopped(dev)) break; @@ -200,7 +202,7 @@ void __qdisc_run(struct net_device *dev) * 2. we've been doing it for too long. */ if (need_resched() || jiffies != start_time) { - netif_schedule_queue(&dev->tx_queue); + netif_schedule_queue(txq); break; } } -- cgit v1.3-8-gc7d7 From c773e847ea8f6812804e40f52399c6921a00eab1 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:13:53 -0700 Subject: netdev: Move _xmit_lock and xmit_lock_owner into netdev_queue. Accesses are mostly structured such that when there are multiple TX queues the code transformations will be a little bit simpler. Signed-off-by: David S. Miller --- drivers/net/bonding/bond_main.c | 13 ++++++- drivers/net/hamradio/bpqether.c | 12 ++++++- drivers/net/macvlan.c | 14 +++++++- drivers/net/wireless/hostap/hostap_hw.c | 12 ++++++- include/linux/netdevice.h | 62 +++++++++++++++++++++------------ net/8021q/vlan_dev.c | 15 ++++++-- net/core/dev.c | 28 ++++++++++----- net/netrom/af_netrom.c | 12 ++++++- net/rose/af_rose.c | 12 ++++++- net/sched/sch_generic.c | 9 +++-- 10 files changed, 145 insertions(+), 44 deletions(-) (limited to 'net/core') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index d57b65dc2c72..dc733d75a5e9 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -5019,6 +5019,17 @@ static int bond_check_params(struct bond_params *params) static struct lock_class_key bonding_netdev_xmit_lock_key; +static void bond_set_lockdep_class_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, + &bonding_netdev_xmit_lock_key); +} + +static void bond_set_lockdep_class(struct net_device *dev) +{ + bond_set_lockdep_class_one(&dev->tx_queue); +} + /* Create a new bond based on the specified name and bonding parameters. * If name is NULL, obtain a suitable "bond%d" name for us. * Caller must NOT hold rtnl_lock; we need to release it here before we @@ -5076,7 +5087,7 @@ int bond_create(char *name, struct bond_params *params) goto out_bond; } - lockdep_set_class(&bond_dev->_xmit_lock, &bonding_netdev_xmit_lock_key); + bond_set_lockdep_class(bond_dev); netif_carrier_off(bond_dev); diff --git a/drivers/net/hamradio/bpqether.c b/drivers/net/hamradio/bpqether.c index 5f4b4c6c9f76..fb186b8c3d4d 100644 --- a/drivers/net/hamradio/bpqether.c +++ b/drivers/net/hamradio/bpqether.c @@ -124,6 +124,16 @@ static LIST_HEAD(bpq_devices); */ static struct lock_class_key bpq_netdev_xmit_lock_key; +static void bpq_set_lockdep_class_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, &bpq_netdev_xmit_lock_key); +} + +static void bpq_set_lockdep_class(struct net_device *dev) +{ + bpq_set_lockdep_class_one(&dev->tx_queue); +} + /* ------------------------------------------------------------------------ */ @@ -523,7 +533,7 @@ static int bpq_new_device(struct net_device *edev) err = register_netdevice(ndev); if (err) goto error; - lockdep_set_class(&ndev->_xmit_lock, &bpq_netdev_xmit_lock_key); + bpq_set_lockdep_class(ndev); /* List protected by RTNL */ list_add_rcu(&bpq->bpq_list, &bpq_devices); diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c index c36a03ae9bfb..c02ceaa4a216 100644 --- a/drivers/net/macvlan.c +++ b/drivers/net/macvlan.c @@ -277,6 +277,17 @@ static struct lock_class_key macvlan_netdev_xmit_lock_key; #define MACVLAN_STATE_MASK \ ((1<<__LINK_STATE_NOCARRIER) | (1<<__LINK_STATE_DORMANT)) +static void macvlan_set_lockdep_class_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, + &macvlan_netdev_xmit_lock_key); +} + +static void macvlan_set_lockdep_class(struct net_device *dev) +{ + macvlan_set_lockdep_class_one(&dev->tx_queue); +} + static int macvlan_init(struct net_device *dev) { struct macvlan_dev *vlan = netdev_priv(dev); @@ -287,7 +298,8 @@ static int macvlan_init(struct net_device *dev) dev->features = lowerdev->features & MACVLAN_FEATURES; dev->iflink = lowerdev->ifindex; - lockdep_set_class(&dev->_xmit_lock, &macvlan_netdev_xmit_lock_key); + macvlan_set_lockdep_class(dev); + return 0; } diff --git a/drivers/net/wireless/hostap/hostap_hw.c b/drivers/net/wireless/hostap/hostap_hw.c index 09004a632ae7..c1f4bb005d92 100644 --- a/drivers/net/wireless/hostap/hostap_hw.c +++ b/drivers/net/wireless/hostap/hostap_hw.c @@ -3102,6 +3102,16 @@ static void prism2_clear_set_tim_queue(local_info_t *local) */ static struct lock_class_key hostap_netdev_xmit_lock_key; +static void prism2_set_lockdep_class_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, + &hostap_netdev_xmit_lock_key); +} + +static void prism2_set_lockdep_class(struct net_device *dev) +{ + prism2_set_lockdep_class_one(&dev->tx_queue); +} static struct net_device * prism2_init_local_data(struct prism2_helper_functions *funcs, int card_idx, @@ -3268,7 +3278,7 @@ while (0) if (ret >= 0) ret = register_netdevice(dev); - lockdep_set_class(&dev->_xmit_lock, &hostap_netdev_xmit_lock_key); + prism2_set_lockdep_class(dev); rtnl_unlock(); if (ret < 0) { printk(KERN_WARNING "%s: register netdevice failed!\n", diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 28aa8e77cee9..c8d5f128858d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -453,6 +453,8 @@ struct netdev_queue { struct net_device *dev; struct Qdisc *qdisc; struct sk_buff *gso_skb; + spinlock_t _xmit_lock; + int xmit_lock_owner; struct Qdisc *qdisc_sleeping; struct list_head qdisc_list; struct netdev_queue *next_sched; @@ -639,12 +641,6 @@ struct net_device /* * One part is mostly used on xmit path (device) */ - /* hard_start_xmit synchronizer */ - spinlock_t _xmit_lock ____cacheline_aligned_in_smp; - /* cpu id of processor entered to hard_start_xmit or -1, - if nobody entered there. - */ - int xmit_lock_owner; void *priv; /* pointer to private data */ int (*hard_start_xmit) (struct sk_buff *skb, struct net_device *dev); @@ -1402,52 +1398,72 @@ static inline void netif_rx_complete(struct net_device *dev, * * Get network device transmit lock */ -static inline void __netif_tx_lock(struct net_device *dev, int cpu) +static inline void __netif_tx_lock(struct netdev_queue *txq, int cpu) { - spin_lock(&dev->_xmit_lock); - dev->xmit_lock_owner = cpu; + spin_lock(&txq->_xmit_lock); + txq->xmit_lock_owner = cpu; } static inline void netif_tx_lock(struct net_device *dev) { - __netif_tx_lock(dev, smp_processor_id()); + __netif_tx_lock(&dev->tx_queue, smp_processor_id()); +} + +static inline void __netif_tx_lock_bh(struct netdev_queue *txq) +{ + spin_lock_bh(&txq->_xmit_lock); + txq->xmit_lock_owner = smp_processor_id(); } static inline void netif_tx_lock_bh(struct net_device *dev) { - spin_lock_bh(&dev->_xmit_lock); - dev->xmit_lock_owner = smp_processor_id(); + __netif_tx_lock_bh(&dev->tx_queue); } -static inline int netif_tx_trylock(struct net_device *dev) +static inline int __netif_tx_trylock(struct netdev_queue *txq) { - int ok = spin_trylock(&dev->_xmit_lock); + int ok = spin_trylock(&txq->_xmit_lock); if (likely(ok)) - dev->xmit_lock_owner = smp_processor_id(); + txq->xmit_lock_owner = smp_processor_id(); return ok; } +static inline int netif_tx_trylock(struct net_device *dev) +{ + return __netif_tx_trylock(&dev->tx_queue); +} + +static inline void __netif_tx_unlock(struct netdev_queue *txq) +{ + txq->xmit_lock_owner = -1; + spin_unlock(&txq->_xmit_lock); +} + static inline void netif_tx_unlock(struct net_device *dev) { - dev->xmit_lock_owner = -1; - spin_unlock(&dev->_xmit_lock); + __netif_tx_unlock(&dev->tx_queue); +} + +static inline void __netif_tx_unlock_bh(struct netdev_queue *txq) +{ + txq->xmit_lock_owner = -1; + spin_unlock_bh(&txq->_xmit_lock); } static inline void netif_tx_unlock_bh(struct net_device *dev) { - dev->xmit_lock_owner = -1; - spin_unlock_bh(&dev->_xmit_lock); + __netif_tx_unlock_bh(&dev->tx_queue); } -#define HARD_TX_LOCK(dev, cpu) { \ +#define HARD_TX_LOCK(dev, txq, cpu) { \ if ((dev->features & NETIF_F_LLTX) == 0) { \ - __netif_tx_lock(dev, cpu); \ + __netif_tx_lock(txq, cpu); \ } \ } -#define HARD_TX_UNLOCK(dev) { \ +#define HARD_TX_UNLOCK(dev, txq) { \ if ((dev->features & NETIF_F_LLTX) == 0) { \ - netif_tx_unlock(dev); \ + __netif_tx_unlock(txq); \ } \ } diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index b6e52c025fd8..8efa399823e3 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -627,6 +627,18 @@ static void vlan_dev_set_rx_mode(struct net_device *vlan_dev) */ static struct lock_class_key vlan_netdev_xmit_lock_key; +static void vlan_dev_set_lockdep_one(struct netdev_queue *txq, + int subclass) +{ + lockdep_set_class_and_subclass(&txq->_xmit_lock, + &vlan_netdev_xmit_lock_key, subclass); +} + +static void vlan_dev_set_lockdep_class(struct net_device *dev, int subclass) +{ + vlan_dev_set_lockdep_one(&dev->tx_queue, subclass); +} + static const struct header_ops vlan_header_ops = { .create = vlan_dev_hard_header, .rebuild = vlan_dev_rebuild_header, @@ -668,8 +680,7 @@ static int vlan_dev_init(struct net_device *dev) if (is_vlan_dev(real_dev)) subclass = 1; - lockdep_set_class_and_subclass(&dev->_xmit_lock, - &vlan_netdev_xmit_lock_key, subclass); + vlan_dev_set_lockdep_class(dev, subclass); return 0; } diff --git a/net/core/dev.c b/net/core/dev.c index 0218b0b9be80..a29a359b15d1 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -258,7 +258,7 @@ DEFINE_PER_CPU(struct softnet_data, softnet_data); #ifdef CONFIG_DEBUG_LOCK_ALLOC /* - * register_netdevice() inits dev->_xmit_lock and sets lockdep class + * register_netdevice() inits txq->_xmit_lock and sets lockdep class * according to dev->type */ static const unsigned short netdev_lock_type[] = @@ -1758,19 +1758,19 @@ gso: if (dev->flags & IFF_UP) { int cpu = smp_processor_id(); /* ok because BHs are off */ - if (dev->xmit_lock_owner != cpu) { + if (txq->xmit_lock_owner != cpu) { - HARD_TX_LOCK(dev, cpu); + HARD_TX_LOCK(dev, txq, cpu); if (!netif_queue_stopped(dev) && !netif_subqueue_stopped(dev, skb)) { rc = 0; if (!dev_hard_start_xmit(skb, dev)) { - HARD_TX_UNLOCK(dev); + HARD_TX_UNLOCK(dev, txq); goto out; } } - HARD_TX_UNLOCK(dev); + HARD_TX_UNLOCK(dev, txq); if (net_ratelimit()) printk(KERN_CRIT "Virtual device %s asks to " "queue packet!\n", dev->name); @@ -3761,6 +3761,20 @@ static void rollback_registered(struct net_device *dev) dev_put(dev); } +static void __netdev_init_queue_locks_one(struct netdev_queue *dev_queue, + struct net_device *dev) +{ + spin_lock_init(&dev_queue->_xmit_lock); + netdev_set_lockdep_class(&dev_queue->_xmit_lock, dev->type); + dev_queue->xmit_lock_owner = -1; +} + +static void netdev_init_queue_locks(struct net_device *dev) +{ + __netdev_init_queue_locks_one(&dev->tx_queue, dev); + __netdev_init_queue_locks_one(&dev->rx_queue, dev); +} + /** * register_netdevice - register a network device * @dev: device to register @@ -3795,9 +3809,7 @@ int register_netdevice(struct net_device *dev) BUG_ON(!dev_net(dev)); net = dev_net(dev); - spin_lock_init(&dev->_xmit_lock); - netdev_set_lockdep_class(&dev->_xmit_lock, dev->type); - dev->xmit_lock_owner = -1; + netdev_init_queue_locks(dev); dev->iflink = -1; diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index 74884f4a6255..819afc449e1e 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -74,6 +74,16 @@ static const struct proto_ops nr_proto_ops; */ static struct lock_class_key nr_netdev_xmit_lock_key; +static void nr_set_lockdep_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, &nr_netdev_xmit_lock_key); +} + +static void nr_set_lockdep_key(struct net_device *dev) +{ + nr_set_lockdep_one(&dev->tx_queue); +} + /* * Socket removal during an interrupt is now safe. */ @@ -1430,7 +1440,7 @@ static int __init nr_proto_init(void) free_netdev(dev); goto fail; } - lockdep_set_class(&dev->_xmit_lock, &nr_netdev_xmit_lock_key); + nr_set_lockdep_key(dev); dev_nr[i] = dev; } diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index 46461a69cd0f..7dbbc0891623 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -75,6 +75,16 @@ ax25_address rose_callsign; */ static struct lock_class_key rose_netdev_xmit_lock_key; +static void rose_set_lockdep_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, &rose_netdev_xmit_lock_key); +} + +static void rose_set_lockdep_key(struct net_device *dev) +{ + rose_set_lockdep_one(&dev->tx_queue); +} + /* * Convert a ROSE address into text. */ @@ -1576,7 +1586,7 @@ static int __init rose_proto_init(void) free_netdev(dev); goto fail; } - lockdep_set_class(&dev->_xmit_lock, &rose_netdev_xmit_lock_key); + rose_set_lockdep_key(dev); dev_rose[i] = dev; } diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index fcc7533f0bcc..b6a36d394663 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -92,10 +92,9 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, struct netdev_queue *dev_queue, struct Qdisc *q) { - struct net_device *dev = dev_queue->dev; int ret; - if (unlikely(dev->xmit_lock_owner == smp_processor_id())) { + if (unlikely(dev_queue->xmit_lock_owner == smp_processor_id())) { /* * Same CPU holding the lock. It may be a transient * configuration error, when hard_start_xmit() recurses. We @@ -105,7 +104,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, kfree_skb(skb); if (net_ratelimit()) printk(KERN_WARNING "Dead loop on netdevice %s, " - "fix it urgently!\n", dev->name); + "fix it urgently!\n", dev_queue->dev->name); ret = qdisc_qlen(q); } else { /* @@ -155,10 +154,10 @@ static inline int qdisc_restart(struct netdev_queue *txq) dev = txq->dev; - HARD_TX_LOCK(dev, smp_processor_id()); + HARD_TX_LOCK(dev, txq, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); - HARD_TX_UNLOCK(dev); + HARD_TX_UNLOCK(dev, txq); spin_lock(&txq->lock); q = txq->qdisc; -- cgit v1.3-8-gc7d7 From 22bb1be4d271961846cd0889b0f8d671db773080 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 10 Jul 2008 11:16:47 +0200 Subject: wext: make sysfs bits optional and deprecate them The /sys/class/net/*/wireless/ direcory is, as far as I know, not used by anyone. Additionally, the same data is available via wext ioctls. Hence the sysfs files are pretty much useless. This patch makes them optional and schedules them for removal. Signed-off-by: Johannes Berg Cc: Jean Tourrilhes Signed-off-by: John W. Linville --- Documentation/feature-removal-schedule.txt | 10 ++++++++++ net/core/net-sysfs.c | 4 ++-- net/wireless/Kconfig | 11 +++++++++++ 3 files changed, 23 insertions(+), 2 deletions(-) (limited to 'net/core') diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt index 8319c462c9f0..db300e09c9ac 100644 --- a/Documentation/feature-removal-schedule.txt +++ b/Documentation/feature-removal-schedule.txt @@ -333,3 +333,13 @@ Why: This option was introduced just to allow older lm-sensors userspace to keep working over the upgrade to 2.6.26. At the scheduled time of removal fixed lm-sensors (2.x or 3.x) should be readily available. Who: Rene Herman + +--------------------------- + +What: Code that is now under CONFIG_WIRELESS_EXT_SYSFS + (in net/core/net-sysfs.c) +When: After the only user (hal) has seen a release with the patches + for enough time, probably some time in 2010. +Why: Over 1K .text/.data size reduction, data is available in other + ways (ioctls) +Who: Johannes Berg diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 3f7941319217..c1f4e0d428c0 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -318,7 +318,7 @@ static struct attribute_group netstat_group = { .attrs = netstat_attrs, }; -#ifdef CONFIG_WIRELESS_EXT +#ifdef CONFIG_WIRELESS_EXT_SYSFS /* helper function that does all the locking etc for wireless stats */ static ssize_t wireless_show(struct device *d, char *buf, ssize_t (*format)(const struct iw_statistics *, @@ -459,7 +459,7 @@ int netdev_register_kobject(struct net_device *net) #ifdef CONFIG_SYSFS *groups++ = &netstat_group; -#ifdef CONFIG_WIRELESS_EXT +#ifdef CONFIG_WIRELESS_EXT_SYSFS if (net->wireless_handlers && net->wireless_handlers->get_wireless_stats) *groups++ = &wireless_group; #endif diff --git a/net/wireless/Kconfig b/net/wireless/Kconfig index 79270903bda6..ab015c62d561 100644 --- a/net/wireless/Kconfig +++ b/net/wireless/Kconfig @@ -29,3 +29,14 @@ config WIRELESS_EXT Say N (if you can) unless you know you need wireless extensions for external modules. + +config WIRELESS_EXT_SYSFS + bool "Wireless extensions sysfs files" + default y + depends on WIRELESS_EXT && SYSFS + help + This option enables the deprecated wireless statistics + files in /sys/class/net/*/wireless/. The same information + is available via the ioctls as well. + + Say Y if you have programs using it (we don't know of any). -- cgit v1.3-8-gc7d7 From 6aa895b047720f71ec4eb11452f7c3ce8426941f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:49:06 -0700 Subject: vlan: Don't store VLAN tag in cb Use a real skb member to store the skb to avoid clashes with qdiscs, which are allowed to use the cb area themselves. As currently only real devices that consume the skb set the NETIF_F_HW_VLAN_TX flag, no explicit invalidation is neccessary. The new member fills a hole on 64 bit, the skb layout changes from: __u32 mark; /* 172 4 */ sk_buff_data_t transport_header; /* 176 4 */ sk_buff_data_t network_header; /* 180 4 */ sk_buff_data_t mac_header; /* 184 4 */ sk_buff_data_t tail; /* 188 4 */ /* --- cacheline 3 boundary (192 bytes) --- */ sk_buff_data_t end; /* 192 4 */ /* XXX 4 bytes hole, try to pack */ to __u32 mark; /* 172 4 */ __u16 vlan_tci; /* 176 2 */ /* XXX 2 bytes hole, try to pack */ sk_buff_data_t transport_header; /* 180 4 */ sk_buff_data_t network_header; /* 184 4 */ Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- include/linux/if_vlan.h | 31 +++++++------------------------ include/linux/skbuff.h | 3 +++ net/core/skbuff.c | 3 +++ 3 files changed, 13 insertions(+), 24 deletions(-) (limited to 'net/core') diff --git a/include/linux/if_vlan.h b/include/linux/if_vlan.h index 93f5d9b0e9f9..9e7b49b8062d 100644 --- a/include/linux/if_vlan.h +++ b/include/linux/if_vlan.h @@ -105,17 +105,8 @@ static inline void vlan_group_set_device(struct vlan_group *vg, array[vlan_id % VLAN_GROUP_ARRAY_PART_LEN] = dev; } -/* VLAN tx hw acceleration helpers. */ -struct vlan_skb_tx_cookie { - u32 magic; - u32 vlan_tag; -}; - -#define VLAN_TX_COOKIE_MAGIC 0x564c414e /* "VLAN" in ascii. */ -#define VLAN_TX_SKB_CB(__skb) ((struct vlan_skb_tx_cookie *)&((__skb)->cb[0])) -#define vlan_tx_tag_present(__skb) \ - (VLAN_TX_SKB_CB(__skb)->magic == VLAN_TX_COOKIE_MAGIC) -#define vlan_tx_tag_get(__skb) (VLAN_TX_SKB_CB(__skb)->vlan_tag) +#define vlan_tx_tag_present(__skb) ((__skb)->vlan_tci) +#define vlan_tx_tag_get(__skb) ((__skb)->vlan_tci) #if defined(CONFIG_VLAN_8021Q) || defined(CONFIG_VLAN_8021Q_MODULE) extern struct net_device *vlan_dev_real_dev(const struct net_device *dev); @@ -210,17 +201,12 @@ static inline struct sk_buff *__vlan_put_tag(struct sk_buff *skb, u16 vlan_tci) * @skb: skbuff to tag * @vlan_tci: VLAN TCI to insert * - * Puts the VLAN TCI in @skb->cb[] and lets the device do the rest + * Puts the VLAN TCI in @skb->vlan_tci and lets the device do the rest */ static inline struct sk_buff *__vlan_hwaccel_put_tag(struct sk_buff *skb, u16 vlan_tci) { - struct vlan_skb_tx_cookie *cookie; - - cookie = VLAN_TX_SKB_CB(skb); - cookie->magic = VLAN_TX_COOKIE_MAGIC; - cookie->vlan_tag = vlan_tci; - + skb->vlan_tci = vlan_tci; return skb; } @@ -267,16 +253,13 @@ static inline int __vlan_get_tag(const struct sk_buff *skb, u16 *vlan_tci) * @skb: skbuff to query * @vlan_tci: buffer to store vlaue * - * Returns error if @skb->cb[] is not set correctly + * Returns error if @skb->vlan_tci is not set correctly */ static inline int __vlan_hwaccel_get_tag(const struct sk_buff *skb, u16 *vlan_tci) { - struct vlan_skb_tx_cookie *cookie; - - cookie = VLAN_TX_SKB_CB(skb); - if (cookie->magic == VLAN_TX_COOKIE_MAGIC) { - *vlan_tci = cookie->vlan_tag; + if (vlan_tx_tag_present(skb)) { + *vlan_tci = skb->vlan_tci; return 0; } else { *vlan_tci = 0; diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 8f10e3d08fd9..7ea44f6621f2 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -246,6 +246,7 @@ typedef unsigned char *sk_buff_data_t; * @dma_cookie: a cookie to one of several possible DMA operations * done by skb DMA functions * @secmark: security marking + * @vlan_tci: vlan tag control information */ struct sk_buff { @@ -326,6 +327,8 @@ struct sk_buff { __u32 mark; + __u16 vlan_tci; + sk_buff_data_t transport_header; sk_buff_data_t network_header; sk_buff_data_t mac_header; diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 7c571560e9d2..50a853f7cd8e 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -459,6 +459,8 @@ static void __copy_skb_header(struct sk_buff *new, const struct sk_buff *old) new->tc_verd = old->tc_verd; #endif #endif + new->vlan_tci = old->vlan_tci; + skb_copy_secmark(new, old); } @@ -2286,6 +2288,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, int features) skb_copy_queue_mapping(nskb, skb); nskb->priority = skb->priority; nskb->protocol = skb->protocol; + nskb->vlan_tci = skb->vlan_tci; nskb->dst = dst_clone(skb->dst); memcpy(nskb->cb, skb->cb, sizeof(skb->cb)); nskb->pkt_type = skb->pkt_type; -- cgit v1.3-8-gc7d7 From bc1d0411b804ad190cdadabac48a10067f17b9e6 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:49:30 -0700 Subject: vlan: deliver packets received with VLAN acceleration to network taps When VLAN header stripping is used, packets currently bypass packet sockets (and other network taps) completely. For locally existing VLANs, they appear directly on the VLAN device, for unknown VLANs they are silently dropped. Add a new function netif_nit_deliver() to deliver incoming packets to all network interface taps and use it in __vlan_hwaccel_rx() to make VLAN packets visible on the underlying device. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- include/linux/netdevice.h | 1 + net/8021q/vlan_core.c | 4 ++++ net/core/dev.c | 27 +++++++++++++++++++++++++++ 3 files changed, 32 insertions(+) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index b54ec16dfbda..ba5c4639ea91 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1165,6 +1165,7 @@ extern int netif_rx(struct sk_buff *skb); extern int netif_rx_ni(struct sk_buff *skb); #define HAVE_NETIF_RECEIVE_SKB 1 extern int netif_receive_skb(struct sk_buff *skb); +extern void netif_nit_deliver(struct sk_buff *skb); extern int dev_valid_name(const char *name); extern int dev_ioctl(struct net *net, unsigned int cmd, void __user *); extern int dev_ethtool(struct net *net, struct ifreq *); diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c index 68df12d3664b..916061f681b6 100644 --- a/net/8021q/vlan_core.c +++ b/net/8021q/vlan_core.c @@ -14,6 +14,9 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, return NET_RX_DROP; } + skb->vlan_tci = vlan_tci; + netif_nit_deliver(skb); + skb->dev = vlan_group_get_device(grp, vlan_tci & VLAN_VID_MASK); if (skb->dev == NULL) { dev_kfree_skb_any(skb); @@ -22,6 +25,7 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, return NET_RX_SUCCESS; } skb->dev->last_rx = jiffies; + skb->vlan_tci = 0; stats = &skb->dev->stats; stats->rx_packets++; diff --git a/net/core/dev.c b/net/core/dev.c index a29a359b15d1..feaab4898a5b 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2068,6 +2068,33 @@ out: } #endif +/* + * netif_nit_deliver - deliver received packets to network taps + * @skb: buffer + * + * This function is used to deliver incoming packets to network + * taps. It should be used when the normal netif_receive_skb path + * is bypassed, for example because of VLAN acceleration. + */ +void netif_nit_deliver(struct sk_buff *skb) +{ + struct packet_type *ptype; + + if (list_empty(&ptype_all)) + return; + + skb_reset_network_header(skb); + skb_reset_transport_header(skb); + skb->mac_len = skb->network_header - skb->mac_header; + + rcu_read_lock(); + list_for_each_entry_rcu(ptype, &ptype_all, list) { + if (!ptype->dev || ptype->dev == skb->dev) + deliver_skb(skb, ptype, skb->dev); + } + rcu_read_unlock(); +} + /** * netif_receive_skb - process receive buffer from network * @skb: buffer to process -- cgit v1.3-8-gc7d7 From f1f28aa3510ddb84c966bac65611bb866c77a092 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 00:08:33 -0700 Subject: netdev: Add addr_list_lock to struct net_device. This will be used to protect the per-device unicast and multicast address lists, as well as the callbacks into the drivers which configure such state such as ->set_rx_mode() and ->set_multicast_list(). Signed-off-by: David S. Miller --- include/linux/netdevice.h | 1 + net/core/dev.c | 1 + 2 files changed, 2 insertions(+) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index ba5c4639ea91..fd0365219181 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -609,6 +609,7 @@ struct net_device unsigned char addr_len; /* hardware address length */ unsigned short dev_id; /* for shared network cards */ + spinlock_t addr_list_lock; struct dev_addr_list *uc_list; /* Secondary unicast mac addresses */ int uc_count; /* Number of installed ucasts */ int uc_promisc; diff --git a/net/core/dev.c b/net/core/dev.c index feaab4898a5b..d933d1bfa6fa 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3836,6 +3836,7 @@ int register_netdevice(struct net_device *dev) BUG_ON(!dev_net(dev)); net = dev_net(dev); + spin_lock_init(&dev->addr_list_lock); netdev_init_queue_locks(dev); dev->iflink = -1; -- cgit v1.3-8-gc7d7 From e308a5d806c852f56590ffdd3834d0df0cbed8d7 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 00:13:44 -0700 Subject: netdev: Add netdev->addr_list_lock protection. Add netif_addr_{lock,unlock}{,_bh}() helpers. Use them to protect operations that operate on or read the network device unicast and multicast address lists. Also use them in cases where the code simply wants to block calls into the driver's ->set_rx_mode() and ->set_multicast_list() methods. Signed-off-by: David S. Miller --- drivers/infiniband/ulp/ipoib/ipoib_multicast.c | 2 ++ drivers/media/dvb/dvb-core/dvb_net.c | 2 ++ drivers/net/bonding/bond_main.c | 8 ++++++++ drivers/net/forcedeth.c | 16 ++++++++++++++++ drivers/net/hamradio/6pack.c | 2 ++ drivers/net/hamradio/mkiss.c | 2 ++ drivers/net/ibm_newemac/core.c | 4 ++++ drivers/net/sfc/efx.c | 2 ++ drivers/net/wireless/libertas/main.c | 2 ++ include/linux/netdevice.h | 20 ++++++++++++++++++++ net/core/dev.c | 14 ++++++++++++++ net/core/dev_mcast.c | 12 ++++++++++++ net/mac80211/main.c | 4 ++++ net/mac80211/mlme.c | 4 ++++ 14 files changed, 94 insertions(+) (limited to 'net/core') diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c index 3f663fb852c1..261ab7150431 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c @@ -775,6 +775,7 @@ void ipoib_mcast_restart_task(struct work_struct *work) local_irq_save(flags); netif_tx_lock(dev); + netif_addr_lock(dev); spin_lock(&priv->lock); /* @@ -851,6 +852,7 @@ void ipoib_mcast_restart_task(struct work_struct *work) } spin_unlock(&priv->lock); + netif_addr_unlock(dev); netif_tx_unlock(dev); local_irq_restore(flags); diff --git a/drivers/media/dvb/dvb-core/dvb_net.c b/drivers/media/dvb/dvb-core/dvb_net.c index c2334aef4143..809d18c663bc 100644 --- a/drivers/media/dvb/dvb-core/dvb_net.c +++ b/drivers/media/dvb/dvb-core/dvb_net.c @@ -1134,6 +1134,7 @@ static void wq_set_multicast_list (struct work_struct *work) dvb_net_feed_stop(dev); priv->rx_mode = RX_MODE_UNI; netif_tx_lock_bh(dev); + netif_addr_lock(dev); if (dev->flags & IFF_PROMISC) { dprintk("%s: promiscuous mode\n", dev->name); @@ -1158,6 +1159,7 @@ static void wq_set_multicast_list (struct work_struct *work) } } + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); dvb_net_feed_start(dev); } diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 8ae7ff313218..ea71abd6f728 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1568,10 +1568,12 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) } netif_tx_lock_bh(bond_dev); + netif_addr_lock(bond_dev); /* upload master's mc_list to new slave */ for (dmi = bond_dev->mc_list; dmi; dmi = dmi->next) { dev_mc_add (slave_dev, dmi->dmi_addr, dmi->dmi_addrlen, 0); } + netif_addr_unlock(bond_dev); netif_tx_unlock_bh(bond_dev); } @@ -1937,7 +1939,9 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev) /* flush master's mc_list from slave */ netif_tx_lock_bh(bond_dev); + netif_addr_lock(bond_dev); bond_mc_list_flush(bond_dev, slave_dev); + netif_addr_unlock(bond_dev); netif_tx_unlock_bh(bond_dev); } @@ -2060,7 +2064,9 @@ static int bond_release_all(struct net_device *bond_dev) /* flush master's mc_list from slave */ netif_tx_lock_bh(bond_dev); + netif_addr_lock(bond_dev); bond_mc_list_flush(bond_dev, slave_dev); + netif_addr_unlock(bond_dev); netif_tx_unlock_bh(bond_dev); } @@ -4674,7 +4680,9 @@ static void bond_free_all(void) bond_work_cancel_all(bond); netif_tx_lock_bh(bond_dev); + netif_addr_lock(bond_dev); bond_mc_list_destroy(bond); + netif_addr_unlock(bond_dev); netif_tx_unlock_bh(bond_dev); /* Release the bonded slaves */ bond_release_all(bond_dev); diff --git a/drivers/net/forcedeth.c b/drivers/net/forcedeth.c index 786d668c612e..4ed89fa9ae46 100644 --- a/drivers/net/forcedeth.c +++ b/drivers/net/forcedeth.c @@ -2831,6 +2831,7 @@ static int nv_change_mtu(struct net_device *dev, int new_mtu) */ nv_disable_irq(dev); netif_tx_lock_bh(dev); + netif_addr_lock(dev); spin_lock(&np->lock); /* stop engines */ nv_stop_rxtx(dev); @@ -2855,6 +2856,7 @@ static int nv_change_mtu(struct net_device *dev, int new_mtu) /* restart rx engine */ nv_start_rxtx(dev); spin_unlock(&np->lock); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); nv_enable_irq(dev); } @@ -2891,6 +2893,7 @@ static int nv_set_mac_address(struct net_device *dev, void *addr) if (netif_running(dev)) { netif_tx_lock_bh(dev); + netif_addr_lock(dev); spin_lock_irq(&np->lock); /* stop rx engine */ @@ -2902,6 +2905,7 @@ static int nv_set_mac_address(struct net_device *dev, void *addr) /* restart rx engine */ nv_start_rx(dev); spin_unlock_irq(&np->lock); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } else { nv_copy_mac_to_hw(dev); @@ -3971,6 +3975,7 @@ static void nv_do_nic_poll(unsigned long data) printk(KERN_INFO "forcedeth: MAC in recoverable error state\n"); if (netif_running(dev)) { netif_tx_lock_bh(dev); + netif_addr_lock(dev); spin_lock(&np->lock); /* stop engines */ nv_stop_rxtx(dev); @@ -3995,6 +4000,7 @@ static void nv_do_nic_poll(unsigned long data) /* restart rx engine */ nv_start_rxtx(dev); spin_unlock(&np->lock); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } } @@ -4202,6 +4208,7 @@ static int nv_set_settings(struct net_device *dev, struct ethtool_cmd *ecmd) nv_disable_irq(dev); netif_tx_lock_bh(dev); + netif_addr_lock(dev); /* with plain spinlock lockdep complains */ spin_lock_irqsave(&np->lock, flags); /* stop engines */ @@ -4215,6 +4222,7 @@ static int nv_set_settings(struct net_device *dev, struct ethtool_cmd *ecmd) */ nv_stop_rxtx(dev); spin_unlock_irqrestore(&np->lock, flags); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } @@ -4360,10 +4368,12 @@ static int nv_nway_reset(struct net_device *dev) if (netif_running(dev)) { nv_disable_irq(dev); netif_tx_lock_bh(dev); + netif_addr_lock(dev); spin_lock(&np->lock); /* stop engines */ nv_stop_rxtx(dev); spin_unlock(&np->lock); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); printk(KERN_INFO "%s: link down.\n", dev->name); } @@ -4471,6 +4481,7 @@ static int nv_set_ringparam(struct net_device *dev, struct ethtool_ringparam* ri if (netif_running(dev)) { nv_disable_irq(dev); netif_tx_lock_bh(dev); + netif_addr_lock(dev); spin_lock(&np->lock); /* stop engines */ nv_stop_rxtx(dev); @@ -4519,6 +4530,7 @@ static int nv_set_ringparam(struct net_device *dev, struct ethtool_ringparam* ri /* restart engines */ nv_start_rxtx(dev); spin_unlock(&np->lock); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); nv_enable_irq(dev); } @@ -4556,10 +4568,12 @@ static int nv_set_pauseparam(struct net_device *dev, struct ethtool_pauseparam* if (netif_running(dev)) { nv_disable_irq(dev); netif_tx_lock_bh(dev); + netif_addr_lock(dev); spin_lock(&np->lock); /* stop engines */ nv_stop_rxtx(dev); spin_unlock(&np->lock); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } @@ -4946,6 +4960,7 @@ static void nv_self_test(struct net_device *dev, struct ethtool_test *test, u64 napi_disable(&np->napi); #endif netif_tx_lock_bh(dev); + netif_addr_lock(dev); spin_lock_irq(&np->lock); nv_disable_hw_interrupts(dev, np->irqmask); if (!(np->msi_flags & NV_MSI_X_ENABLED)) { @@ -4959,6 +4974,7 @@ static void nv_self_test(struct net_device *dev, struct ethtool_test *test, u64 /* drain rx queue */ nv_drain_rxtx(dev); spin_unlock_irq(&np->lock); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } diff --git a/drivers/net/hamradio/6pack.c b/drivers/net/hamradio/6pack.c index 06ad9f302b5a..ffc937f5d15d 100644 --- a/drivers/net/hamradio/6pack.c +++ b/drivers/net/hamradio/6pack.c @@ -300,7 +300,9 @@ static int sp_set_mac_address(struct net_device *dev, void *addr) struct sockaddr_ax25 *sa = addr; netif_tx_lock_bh(dev); + netif_addr_lock(dev); memcpy(dev->dev_addr, &sa->sax25_call, AX25_ADDR_LEN); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return 0; diff --git a/drivers/net/hamradio/mkiss.c b/drivers/net/hamradio/mkiss.c index 65166035aca0..b8740e6a5cec 100644 --- a/drivers/net/hamradio/mkiss.c +++ b/drivers/net/hamradio/mkiss.c @@ -356,7 +356,9 @@ static int ax_set_mac_address(struct net_device *dev, void *addr) struct sockaddr_ax25 *sa = addr; netif_tx_lock_bh(dev); + netif_addr_lock(dev); memcpy(dev->dev_addr, &sa->sax25_call, AX25_ADDR_LEN); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return 0; diff --git a/drivers/net/ibm_newemac/core.c b/drivers/net/ibm_newemac/core.c index babc79ad490b..9ca57d365599 100644 --- a/drivers/net/ibm_newemac/core.c +++ b/drivers/net/ibm_newemac/core.c @@ -295,7 +295,9 @@ static void emac_rx_disable(struct emac_instance *dev) static inline void emac_netif_stop(struct emac_instance *dev) { netif_tx_lock_bh(dev->ndev); + netif_addr_lock(dev->ndev); dev->no_mcast = 1; + netif_addr_unlock(dev->ndev); netif_tx_unlock_bh(dev->ndev); dev->ndev->trans_start = jiffies; /* prevent tx timeout */ mal_poll_disable(dev->mal, &dev->commac); @@ -305,9 +307,11 @@ static inline void emac_netif_stop(struct emac_instance *dev) static inline void emac_netif_start(struct emac_instance *dev) { netif_tx_lock_bh(dev->ndev); + netif_addr_lock(dev->ndev); dev->no_mcast = 0; if (dev->mcast_pending && netif_running(dev->ndev)) __emac_set_multicast_list(dev); + netif_addr_unlock(dev->ndev); netif_tx_unlock_bh(dev->ndev); netif_wake_queue(dev->ndev); diff --git a/drivers/net/sfc/efx.c b/drivers/net/sfc/efx.c index 74265d8553b8..e1257e556e48 100644 --- a/drivers/net/sfc/efx.c +++ b/drivers/net/sfc/efx.c @@ -697,6 +697,8 @@ static void efx_stop_port(struct efx_nic *efx) /* Serialise against efx_set_multicast_list() */ if (efx_dev_registered(efx)) { netif_tx_lock_bh(efx->net_dev); + netif_addr_lock(efx->net_dev); + netif_addr_unlock(efx->net_dev); netif_tx_unlock_bh(efx->net_dev); } } diff --git a/drivers/net/wireless/libertas/main.c b/drivers/net/wireless/libertas/main.c index abd6d9ed8f4b..42e9b2771eab 100644 --- a/drivers/net/wireless/libertas/main.c +++ b/drivers/net/wireless/libertas/main.c @@ -594,6 +594,7 @@ static int lbs_add_mcast_addrs(struct cmd_ds_mac_multicast_adr *cmd, return nr_addrs; netif_tx_lock_bh(dev); + netif_addr_lock(dev); for (mc_list = dev->mc_list; mc_list; mc_list = mc_list->next) { if (mac_in_list(cmd->maclist, nr_addrs, mc_list->dmi_addr)) { lbs_deb_net("mcast address %s:%s skipped\n", dev->name, @@ -608,6 +609,7 @@ static int lbs_add_mcast_addrs(struct cmd_ds_mac_multicast_adr *cmd, print_mac(mac, mc_list->dmi_addr)); i++; } + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); if (mc_list) return -EOVERFLOW; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index fd0365219181..570cf7affa72 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1498,6 +1498,26 @@ static inline void netif_tx_disable(struct net_device *dev) netif_tx_unlock_bh(dev); } +static inline void netif_addr_lock(struct net_device *dev) +{ + spin_lock(&dev->addr_list_lock); +} + +static inline void netif_addr_lock_bh(struct net_device *dev) +{ + spin_lock_bh(&dev->addr_list_lock); +} + +static inline void netif_addr_unlock(struct net_device *dev) +{ + spin_unlock(&dev->addr_list_lock); +} + +static inline void netif_addr_unlock_bh(struct net_device *dev) +{ + spin_unlock_bh(&dev->addr_list_lock); +} + /* These functions live elsewhere (drivers/net/net_init.c, but related) */ extern void ether_setup(struct net_device *dev); diff --git a/net/core/dev.c b/net/core/dev.c index d933d1bfa6fa..ef1502d71f25 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2982,7 +2982,9 @@ void __dev_set_rx_mode(struct net_device *dev) void dev_set_rx_mode(struct net_device *dev) { netif_tx_lock_bh(dev); + netif_addr_lock(dev); __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } @@ -3062,9 +3064,11 @@ int dev_unicast_delete(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_delete(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -3088,9 +3092,11 @@ int dev_unicast_add(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_add(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -3159,10 +3165,12 @@ int dev_unicast_sync(struct net_device *to, struct net_device *from) int err = 0; netif_tx_lock_bh(to); + netif_addr_lock(to); err = __dev_addr_sync(&to->uc_list, &to->uc_count, &from->uc_list, &from->uc_count); if (!err) __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); return err; } @@ -3180,13 +3188,17 @@ EXPORT_SYMBOL(dev_unicast_sync); void dev_unicast_unsync(struct net_device *to, struct net_device *from) { netif_tx_lock_bh(from); + netif_addr_lock(from); netif_tx_lock_bh(to); + netif_addr_lock(to); __dev_addr_unsync(&to->uc_list, &to->uc_count, &from->uc_list, &from->uc_count); __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); + netif_addr_unlock(from); netif_tx_unlock_bh(from); } EXPORT_SYMBOL(dev_unicast_unsync); @@ -3208,6 +3220,7 @@ static void __dev_addr_discard(struct dev_addr_list **list) static void dev_addr_discard(struct net_device *dev) { netif_tx_lock_bh(dev); + netif_addr_lock(dev); __dev_addr_discard(&dev->uc_list); dev->uc_count = 0; @@ -3215,6 +3228,7 @@ static void dev_addr_discard(struct net_device *dev) __dev_addr_discard(&dev->mc_list); dev->mc_count = 0; + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } diff --git a/net/core/dev_mcast.c b/net/core/dev_mcast.c index f8a3455f4493..b6b2a129971a 100644 --- a/net/core/dev_mcast.c +++ b/net/core/dev_mcast.c @@ -73,6 +73,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) int err; netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_delete(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) { @@ -83,6 +84,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) __dev_set_rx_mode(dev); } + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -96,9 +98,11 @@ int dev_mc_add(struct net_device *dev, void *addr, int alen, int glbl) int err; netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_add(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -120,10 +124,12 @@ int dev_mc_sync(struct net_device *to, struct net_device *from) int err = 0; netif_tx_lock_bh(to); + netif_addr_lock(to); err = __dev_addr_sync(&to->mc_list, &to->mc_count, &from->mc_list, &from->mc_count); if (!err) __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); return err; @@ -144,13 +150,17 @@ EXPORT_SYMBOL(dev_mc_sync); void dev_mc_unsync(struct net_device *to, struct net_device *from) { netif_tx_lock_bh(from); + netif_addr_lock(from); netif_tx_lock_bh(to); + netif_addr_lock(to); __dev_addr_unsync(&to->mc_list, &to->mc_count, &from->mc_list, &from->mc_count); __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); + netif_addr_unlock(from); netif_tx_unlock_bh(from); } EXPORT_SYMBOL(dev_mc_unsync); @@ -165,6 +175,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) return 0; netif_tx_lock_bh(dev); + netif_addr_lock(dev); for (m = dev->mc_list; m; m = m->next) { int i; @@ -176,6 +187,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) seq_putc(seq, '\n'); } + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return 0; } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 36859e794928..095b7d928d64 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -292,7 +292,9 @@ static int ieee80211_open(struct net_device *dev) local->fif_other_bss++; netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); ieee80211_configure_filter(local); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_STA: @@ -491,7 +493,9 @@ static int ieee80211_stop(struct net_device *dev) local->fif_other_bss--; netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); ieee80211_configure_filter(local); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_MESH_POINT: diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 8f51375317dd..1232ba25e1e9 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -3869,6 +3869,7 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw) netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); local->filter_flags &= ~FIF_BCN_PRBRESP_PROMISC; local->ops->configure_filter(local_to_hw(local), FIF_BCN_PRBRESP_PROMISC, @@ -3876,6 +3877,7 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw) local->mdev->mc_count, local->mdev->mc_list); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); rcu_read_lock(); @@ -4063,12 +4065,14 @@ static int ieee80211_sta_start_scan(struct net_device *dev, local->scan_dev = dev; netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); local->filter_flags |= FIF_BCN_PRBRESP_PROMISC; local->ops->configure_filter(local_to_hw(local), FIF_BCN_PRBRESP_PROMISC, &local->filter_flags, local->mdev->mc_count, local->mdev->mc_list); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); /* TODO: start scan as soon as all nullfunc frames are ACKed */ -- cgit v1.3-8-gc7d7 From b9e40857682ecfc5bcd0356a23ff409883ffb982 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 00:15:08 -0700 Subject: netdev: Do not use TX lock to protect address lists. Now that we have a specific lock to protect the network device unicast and multicast lists, remove extraneous grabs of the TX lock in cases where the code only needs address list protection. Signed-off-by: David S. Miller --- drivers/infiniband/ulp/ipoib/ipoib_multicast.c | 2 -- drivers/media/dvb/dvb-core/dvb_net.c | 6 ++-- drivers/net/bonding/bond_main.c | 24 ++++++---------- drivers/net/sfc/efx.c | 6 ++-- drivers/net/wireless/libertas/main.c | 6 ++-- net/core/dev.c | 38 ++++++++------------------ net/core/dev_mcast.c | 32 +++++++--------------- net/mac80211/main.c | 12 +++----- net/mac80211/mlme.c | 6 ++-- 9 files changed, 42 insertions(+), 90 deletions(-) (limited to 'net/core') diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c index 261ab7150431..cd2fb955040f 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c @@ -774,7 +774,6 @@ void ipoib_mcast_restart_task(struct work_struct *work) ipoib_mcast_stop_thread(dev, 0); local_irq_save(flags); - netif_tx_lock(dev); netif_addr_lock(dev); spin_lock(&priv->lock); @@ -853,7 +852,6 @@ void ipoib_mcast_restart_task(struct work_struct *work) spin_unlock(&priv->lock); netif_addr_unlock(dev); - netif_tx_unlock(dev); local_irq_restore(flags); /* We have to cancel outside of the spinlock */ diff --git a/drivers/media/dvb/dvb-core/dvb_net.c b/drivers/media/dvb/dvb-core/dvb_net.c index 809d18c663bc..c2c033722a93 100644 --- a/drivers/media/dvb/dvb-core/dvb_net.c +++ b/drivers/media/dvb/dvb-core/dvb_net.c @@ -1133,8 +1133,7 @@ static void wq_set_multicast_list (struct work_struct *work) dvb_net_feed_stop(dev); priv->rx_mode = RX_MODE_UNI; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); if (dev->flags & IFF_PROMISC) { dprintk("%s: promiscuous mode\n", dev->name); @@ -1159,8 +1158,7 @@ static void wq_set_multicast_list (struct work_struct *work) } } - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); dvb_net_feed_start(dev); } diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index ea71abd6f728..fd87dbe7999a 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1567,14 +1567,12 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) goto err_close; } - netif_tx_lock_bh(bond_dev); - netif_addr_lock(bond_dev); + netif_addr_lock_bh(bond_dev); /* upload master's mc_list to new slave */ for (dmi = bond_dev->mc_list; dmi; dmi = dmi->next) { dev_mc_add (slave_dev, dmi->dmi_addr, dmi->dmi_addrlen, 0); } - netif_addr_unlock(bond_dev); - netif_tx_unlock_bh(bond_dev); + netif_addr_unlock_bh(bond_dev); } if (bond->params.mode == BOND_MODE_8023AD) { @@ -1938,11 +1936,9 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev) } /* flush master's mc_list from slave */ - netif_tx_lock_bh(bond_dev); - netif_addr_lock(bond_dev); + netif_addr_lock_bh(bond_dev); bond_mc_list_flush(bond_dev, slave_dev); - netif_addr_unlock(bond_dev); - netif_tx_unlock_bh(bond_dev); + netif_addr_unlock_bh(bond_dev); } netdev_set_master(slave_dev, NULL); @@ -2063,11 +2059,9 @@ static int bond_release_all(struct net_device *bond_dev) } /* flush master's mc_list from slave */ - netif_tx_lock_bh(bond_dev); - netif_addr_lock(bond_dev); + netif_addr_lock_bh(bond_dev); bond_mc_list_flush(bond_dev, slave_dev); - netif_addr_unlock(bond_dev); - netif_tx_unlock_bh(bond_dev); + netif_addr_unlock_bh(bond_dev); } netdev_set_master(slave_dev, NULL); @@ -4679,11 +4673,9 @@ static void bond_free_all(void) struct net_device *bond_dev = bond->dev; bond_work_cancel_all(bond); - netif_tx_lock_bh(bond_dev); - netif_addr_lock(bond_dev); + netif_addr_lock_bh(bond_dev); bond_mc_list_destroy(bond); - netif_addr_unlock(bond_dev); - netif_tx_unlock_bh(bond_dev); + netif_addr_unlock_bh(bond_dev); /* Release the bonded slaves */ bond_release_all(bond_dev); bond_destroy(bond); diff --git a/drivers/net/sfc/efx.c b/drivers/net/sfc/efx.c index e1257e556e48..7b2015f9e469 100644 --- a/drivers/net/sfc/efx.c +++ b/drivers/net/sfc/efx.c @@ -696,10 +696,8 @@ static void efx_stop_port(struct efx_nic *efx) /* Serialise against efx_set_multicast_list() */ if (efx_dev_registered(efx)) { - netif_tx_lock_bh(efx->net_dev); - netif_addr_lock(efx->net_dev); - netif_addr_unlock(efx->net_dev); - netif_tx_unlock_bh(efx->net_dev); + netif_addr_lock_bh(efx->net_dev); + netif_addr_unlock_bh(efx->net_dev); } } diff --git a/drivers/net/wireless/libertas/main.c b/drivers/net/wireless/libertas/main.c index 42e9b2771eab..14d5d61cec4c 100644 --- a/drivers/net/wireless/libertas/main.c +++ b/drivers/net/wireless/libertas/main.c @@ -593,8 +593,7 @@ static int lbs_add_mcast_addrs(struct cmd_ds_mac_multicast_adr *cmd, if ((dev->flags & (IFF_UP|IFF_MULTICAST)) != (IFF_UP|IFF_MULTICAST)) return nr_addrs; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); for (mc_list = dev->mc_list; mc_list; mc_list = mc_list->next) { if (mac_in_list(cmd->maclist, nr_addrs, mc_list->dmi_addr)) { lbs_deb_net("mcast address %s:%s skipped\n", dev->name, @@ -609,8 +608,7 @@ static int lbs_add_mcast_addrs(struct cmd_ds_mac_multicast_adr *cmd, print_mac(mac, mc_list->dmi_addr)); i++; } - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); if (mc_list) return -EOVERFLOW; diff --git a/net/core/dev.c b/net/core/dev.c index ef1502d71f25..9b49f74a9820 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2981,11 +2981,9 @@ void __dev_set_rx_mode(struct net_device *dev) void dev_set_rx_mode(struct net_device *dev) { - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); } int __dev_addr_delete(struct dev_addr_list **list, int *count, @@ -3063,13 +3061,11 @@ int dev_unicast_delete(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_delete(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } EXPORT_SYMBOL(dev_unicast_delete); @@ -3091,13 +3087,11 @@ int dev_unicast_add(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_add(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } EXPORT_SYMBOL(dev_unicast_add); @@ -3164,14 +3158,12 @@ int dev_unicast_sync(struct net_device *to, struct net_device *from) { int err = 0; - netif_tx_lock_bh(to); - netif_addr_lock(to); + netif_addr_lock_bh(to); err = __dev_addr_sync(&to->uc_list, &to->uc_count, &from->uc_list, &from->uc_count); if (!err) __dev_set_rx_mode(to); - netif_addr_unlock(to); - netif_tx_unlock_bh(to); + netif_addr_unlock_bh(to); return err; } EXPORT_SYMBOL(dev_unicast_sync); @@ -3187,9 +3179,7 @@ EXPORT_SYMBOL(dev_unicast_sync); */ void dev_unicast_unsync(struct net_device *to, struct net_device *from) { - netif_tx_lock_bh(from); - netif_addr_lock(from); - netif_tx_lock_bh(to); + netif_addr_lock_bh(from); netif_addr_lock(to); __dev_addr_unsync(&to->uc_list, &to->uc_count, @@ -3197,9 +3187,7 @@ void dev_unicast_unsync(struct net_device *to, struct net_device *from) __dev_set_rx_mode(to); netif_addr_unlock(to); - netif_tx_unlock_bh(to); - netif_addr_unlock(from); - netif_tx_unlock_bh(from); + netif_addr_unlock_bh(from); } EXPORT_SYMBOL(dev_unicast_unsync); @@ -3219,8 +3207,7 @@ static void __dev_addr_discard(struct dev_addr_list **list) static void dev_addr_discard(struct net_device *dev) { - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); __dev_addr_discard(&dev->uc_list); dev->uc_count = 0; @@ -3228,8 +3215,7 @@ static void dev_addr_discard(struct net_device *dev) __dev_addr_discard(&dev->mc_list); dev->mc_count = 0; - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); } unsigned dev_get_flags(const struct net_device *dev) diff --git a/net/core/dev_mcast.c b/net/core/dev_mcast.c index b6b2a129971a..5402b3b38e0d 100644 --- a/net/core/dev_mcast.c +++ b/net/core/dev_mcast.c @@ -72,8 +72,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) { int err; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_delete(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) { @@ -84,8 +83,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) __dev_set_rx_mode(dev); } - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } @@ -97,13 +95,11 @@ int dev_mc_add(struct net_device *dev, void *addr, int alen, int glbl) { int err; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_add(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } @@ -123,14 +119,12 @@ int dev_mc_sync(struct net_device *to, struct net_device *from) { int err = 0; - netif_tx_lock_bh(to); - netif_addr_lock(to); + netif_addr_lock_bh(to); err = __dev_addr_sync(&to->mc_list, &to->mc_count, &from->mc_list, &from->mc_count); if (!err) __dev_set_rx_mode(to); - netif_addr_unlock(to); - netif_tx_unlock_bh(to); + netif_addr_unlock_bh(to); return err; } @@ -149,9 +143,7 @@ EXPORT_SYMBOL(dev_mc_sync); */ void dev_mc_unsync(struct net_device *to, struct net_device *from) { - netif_tx_lock_bh(from); - netif_addr_lock(from); - netif_tx_lock_bh(to); + netif_addr_lock_bh(from); netif_addr_lock(to); __dev_addr_unsync(&to->mc_list, &to->mc_count, @@ -159,9 +151,7 @@ void dev_mc_unsync(struct net_device *to, struct net_device *from) __dev_set_rx_mode(to); netif_addr_unlock(to); - netif_tx_unlock_bh(to); - netif_addr_unlock(from); - netif_tx_unlock_bh(from); + netif_addr_unlock_bh(from); } EXPORT_SYMBOL(dev_mc_unsync); @@ -174,8 +164,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) if (v == SEQ_START_TOKEN) return 0; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); for (m = dev->mc_list; m; m = m->next) { int i; @@ -187,8 +176,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) seq_putc(seq, '\n'); } - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return 0; } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 095b7d928d64..af0056e7e5b3 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -291,11 +291,9 @@ static int ieee80211_open(struct net_device *dev) if (sdata->u.mntr_flags & MONITOR_FLAG_OTHER_BSS) local->fif_other_bss++; - netif_tx_lock_bh(local->mdev); - netif_addr_lock(local->mdev); + netif_addr_lock_bh(local->mdev); ieee80211_configure_filter(local); - netif_addr_unlock(local->mdev); - netif_tx_unlock_bh(local->mdev); + netif_addr_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_STA: case IEEE80211_IF_TYPE_IBSS: @@ -492,11 +490,9 @@ static int ieee80211_stop(struct net_device *dev) if (sdata->u.mntr_flags & MONITOR_FLAG_OTHER_BSS) local->fif_other_bss--; - netif_tx_lock_bh(local->mdev); - netif_addr_lock(local->mdev); + netif_addr_lock_bh(local->mdev); ieee80211_configure_filter(local); - netif_addr_unlock(local->mdev); - netif_tx_unlock_bh(local->mdev); + netif_addr_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_MESH_POINT: case IEEE80211_IF_TYPE_STA: diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 1232ba25e1e9..d7c371e36bf0 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -4064,16 +4064,14 @@ static int ieee80211_sta_start_scan(struct net_device *dev, local->scan_band = IEEE80211_BAND_2GHZ; local->scan_dev = dev; - netif_tx_lock_bh(local->mdev); - netif_addr_lock(local->mdev); + netif_addr_lock_bh(local->mdev); local->filter_flags |= FIF_BCN_PRBRESP_PROMISC; local->ops->configure_filter(local_to_hw(local), FIF_BCN_PRBRESP_PROMISC, &local->filter_flags, local->mdev->mc_count, local->mdev->mc_list); - netif_addr_unlock(local->mdev); - netif_tx_unlock_bh(local->mdev); + netif_addr_unlock_bh(local->mdev); /* TODO: start scan as soon as all nullfunc frames are ACKed */ queue_delayed_work(local->hw.workqueue, &local->scan_work, -- cgit v1.3-8-gc7d7 From 2870c43d1795bcb40b12bad6456f07ad8e64b3de Mon Sep 17 00:00:00 2001 From: Octavian Purdila Date: Tue, 15 Jul 2008 00:49:11 -0700 Subject: net: refactor tcp splice receive path to improve readability - move all of the details on offsets, lengths and buffers into a single function instead of doing these operation from multiple places - use a bottom up approach: try to avoid details in the high level functions, introduce them gradually as we go deeper in the function call stack With helpful feedback from Jarek Poplawski. Signed-off-by: Octavian Purdila Acked-by: Jarek Poplawski Signed-off-by: David S. Miller --- net/core/skbuff.c | 153 ++++++++++++++++++++++-------------------------------- 1 file changed, 61 insertions(+), 92 deletions(-) (limited to 'net/core') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 50a853f7cd8e..48a43727d12f 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -1282,114 +1282,83 @@ static inline int spd_fill_page(struct splice_pipe_desc *spd, struct page *page, return 0; } -/* - * Map linear and fragment data from the skb to spd. Returns number of - * pages mapped. - */ -static int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, - unsigned int *total_len, - struct splice_pipe_desc *spd) -{ - unsigned int nr_pages = spd->nr_pages; - unsigned int poff, plen, len, toff, tlen; - int headlen, seg, error = 0; - - toff = *offset; - tlen = *total_len; - if (!tlen) { - error = 1; - goto err; +static inline void __segment_seek(struct page **page, unsigned int *poff, + unsigned int *plen, unsigned int off) +{ + *poff += off; + *page += *poff / PAGE_SIZE; + *poff = *poff % PAGE_SIZE; + *plen -= off; +} + +static inline int __splice_segment(struct page *page, unsigned int poff, + unsigned int plen, unsigned int *off, + unsigned int *len, struct sk_buff *skb, + struct splice_pipe_desc *spd) +{ + if (!*len) + return 1; + + /* skip this segment if already processed */ + if (*off >= plen) { + *off -= plen; + return 0; } - /* - * if the offset is greater than the linear part, go directly to - * the fragments. - */ - headlen = skb_headlen(skb); - if (toff >= headlen) { - toff -= headlen; - goto map_frag; + /* ignore any bits we already processed */ + if (*off) { + __segment_seek(&page, &poff, &plen, *off); + *off = 0; } - /* - * first map the linear region into the pages/partial map, skipping - * any potential initial offset. - */ - len = 0; - while (len < headlen) { - void *p = skb->data + len; - - poff = (unsigned long) p & (PAGE_SIZE - 1); - plen = min_t(unsigned int, headlen - len, PAGE_SIZE - poff); - len += plen; - - if (toff) { - if (plen <= toff) { - toff -= plen; - continue; - } - plen -= toff; - poff += toff; - toff = 0; - } + do { + unsigned int flen = min(*len, plen); - plen = min(plen, tlen); - if (!plen) - break; + /* the linear region may spread across several pages */ + flen = min_t(unsigned int, flen, PAGE_SIZE - poff); - /* - * just jump directly to update and return, no point - * in going over fragments when the output is full. - */ - error = spd_fill_page(spd, virt_to_page(p), plen, poff, skb); - if (error) - goto done; + if (spd_fill_page(spd, page, flen, poff, skb)) + return 1; - tlen -= plen; - } + __segment_seek(&page, &poff, &plen, flen); + *len -= flen; + + } while (*len && plen); + + return 0; +} + +/* + * Map linear and fragment data from the skb to spd. It reports failure if the + * pipe is full or if we already spliced the requested length. + */ +int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, + unsigned int *len, + struct splice_pipe_desc *spd) +{ + int seg; + + /* + * map the linear part + */ + if (__splice_segment(virt_to_page(skb->data), + (unsigned long) skb->data & (PAGE_SIZE - 1), + skb_headlen(skb), + offset, len, skb, spd)) + return 1; /* * then map the fragments */ -map_frag: for (seg = 0; seg < skb_shinfo(skb)->nr_frags; seg++) { const skb_frag_t *f = &skb_shinfo(skb)->frags[seg]; - plen = f->size; - poff = f->page_offset; - - if (toff) { - if (plen <= toff) { - toff -= plen; - continue; - } - plen -= toff; - poff += toff; - toff = 0; - } - - plen = min(plen, tlen); - if (!plen) - break; - - error = spd_fill_page(spd, f->page, plen, poff, skb); - if (error) - break; - - tlen -= plen; + if (__splice_segment(f->page, f->page_offset, f->size, + offset, len, skb, spd)) + return 1; } -done: - if (spd->nr_pages - nr_pages) { - *offset = 0; - *total_len = tlen; - return 0; - } -err: - /* update the offset to reflect the linear part skip, if any */ - if (!error) - *offset = toff; - return error; + return 0; } /* -- cgit v1.3-8-gc7d7 From 7b1c65faa27f5ade3915e4bbc9186b6e64d2d6ec Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 16 Jul 2008 20:12:30 -0700 Subject: net: make __skb_splice_bits static net/core/skbuff.c:1335:5: warning: symbol '__skb_splice_bits' was not declared. Should it be static? Signed-off-by: Harvey Harrison Signed-off-by: David S. Miller --- net/core/skbuff.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 48a43727d12f..e4115672b6cf 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -1332,7 +1332,7 @@ static inline int __splice_segment(struct page *page, unsigned int poff, * Map linear and fragment data from the skb to spd. It reports failure if the * pipe is full or if we already spliced the requested length. */ -int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, +static int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, unsigned int *len, struct splice_pipe_desc *spd) { -- cgit v1.3-8-gc7d7 From 5c52ba170f8167511bdb65b981f4582100c40675 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:28:10 -0700 Subject: sock: add net to prot->enter_memory_pressure callback The tcp_enter_memory_pressure calls NET_INC_STATS, but doesn't have where to get the net from. I decided to add a sk argument, not the net itself, only to factor all the required sock_net(sk) calls inside the enter_memory_pressure callback itself. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- include/net/sock.h | 4 ++-- include/net/tcp.h | 2 +- net/core/sock.c | 2 +- net/decnet/af_decnet.c | 2 +- net/ipv4/tcp.c | 4 ++-- net/sctp/socket.c | 2 +- 6 files changed, 8 insertions(+), 8 deletions(-) (limited to 'net/core') diff --git a/include/net/sock.h b/include/net/sock.h index 3f4897ab432e..06c5259aff30 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -565,7 +565,7 @@ struct proto { #endif /* Memory pressure */ - void (*enter_memory_pressure)(void); + void (*enter_memory_pressure)(struct sock *sk); atomic_t *memory_allocated; /* Current allocated memory. */ atomic_t *sockets_allocated; /* Current number of sockets. */ /* @@ -1210,7 +1210,7 @@ static inline struct page *sk_stream_alloc_page(struct sock *sk) page = alloc_pages(sk->sk_allocation, 0); if (!page) { - sk->sk_prot->enter_memory_pressure(); + sk->sk_prot->enter_memory_pressure(sk); sk_stream_moderate_sndbuf(sk); } return page; diff --git a/include/net/tcp.h b/include/net/tcp.h index 4d788181654e..c25cb5278b95 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -975,7 +975,7 @@ static inline void tcp_openreq_init(struct request_sock *req, ireq->rmt_port = tcp_hdr(skb)->source; } -extern void tcp_enter_memory_pressure(void); +extern void tcp_enter_memory_pressure(struct sock *sk); static inline int keepalive_intvl_when(const struct tcp_sock *tp) { diff --git a/net/core/sock.c b/net/core/sock.c index 2c0ba52e5303..10a64d57078c 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -1442,7 +1442,7 @@ int __sk_mem_schedule(struct sock *sk, int size, int kind) /* Under pressure. */ if (allocated > prot->sysctl_mem[1]) if (prot->enter_memory_pressure) - prot->enter_memory_pressure(); + prot->enter_memory_pressure(sk); /* Over hard limit. */ if (allocated > prot->sysctl_mem[2]) diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c index 931bdf9cb756..61b7df577ddd 100644 --- a/net/decnet/af_decnet.c +++ b/net/decnet/af_decnet.c @@ -451,7 +451,7 @@ static void dn_destruct(struct sock *sk) static int dn_memory_pressure; -static void dn_enter_memory_pressure(void) +static void dn_enter_memory_pressure(struct sock *sk) { if (!dn_memory_pressure) { dn_memory_pressure = 1; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 525dcf534153..bc8559a6f7e5 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -316,7 +316,7 @@ int tcp_memory_pressure __read_mostly; EXPORT_SYMBOL(tcp_memory_pressure); -void tcp_enter_memory_pressure(void) +void tcp_enter_memory_pressure(struct sock *sk) { if (!tcp_memory_pressure) { NET_INC_STATS(LINUX_MIB_TCPMEMORYPRESSURES); @@ -649,7 +649,7 @@ struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp) } __kfree_skb(skb); } else { - sk->sk_prot->enter_memory_pressure(); + sk->sk_prot->enter_memory_pressure(sk); sk_stream_moderate_sndbuf(sk); } return NULL; diff --git a/net/sctp/socket.c b/net/sctp/socket.c index df5572c39f0c..6aba01b0ce4e 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -116,7 +116,7 @@ static int sctp_memory_pressure; static atomic_t sctp_memory_allocated; static atomic_t sctp_sockets_allocated; -static void sctp_enter_memory_pressure(void) +static void sctp_enter_memory_pressure(struct sock *sk) { sctp_memory_pressure = 1; } -- cgit v1.3-8-gc7d7 From 9a6d276e85aa3d8f308fc5e8de6892daeb60ae5f Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Wed, 16 Jul 2008 20:50:49 -0700 Subject: core: add stat to track unresolved discards in neighbor cache in __neigh_event_send, if we have a neighbour entry which is in NUD_INCOMPLETE state, we enqueue any outbound frames to that neighbour to the neighbours arp_queue, which is default capped to a length of 3 skbs. If that queue exceeds its set length, it will drop an skb on the queue to enqueue the newly arrived skb. This results in a drop for which we have no statistics incremented. This patch adds an unresolved_discards stat to /proc/net/stat/ndisc_cache to track these lost frames. Signed-off-by: Neil Horman Signed-off-by: David S. Miller --- include/net/neighbour.h | 4 +++- net/core/neighbour.c | 8 +++++--- 2 files changed, 8 insertions(+), 4 deletions(-) (limited to 'net/core') diff --git a/include/net/neighbour.h b/include/net/neighbour.h index dc420fecafb9..aa4b708654a4 100644 --- a/include/net/neighbour.h +++ b/include/net/neighbour.h @@ -75,7 +75,7 @@ struct neigh_statistics unsigned long destroys; /* number of destroyed neighs */ unsigned long hash_grows; /* number of hash resizes */ - unsigned long res_failed; /* nomber of failed resolutions */ + unsigned long res_failed; /* number of failed resolutions */ unsigned long lookups; /* number of lookups */ unsigned long hits; /* number of hits (among lookups) */ @@ -85,6 +85,8 @@ struct neigh_statistics unsigned long periodic_gc_runs; /* number of periodic GC runs */ unsigned long forced_gc_runs; /* number of forced GC runs */ + + unsigned long unres_discards; /* number of unresolved drops */ }; #define NEIGH_CACHE_STAT_INC(tbl, field) \ diff --git a/net/core/neighbour.c b/net/core/neighbour.c index 65f01f71b3f3..f62c8af85d38 100644 --- a/net/core/neighbour.c +++ b/net/core/neighbour.c @@ -930,6 +930,7 @@ int __neigh_event_send(struct neighbour *neigh, struct sk_buff *skb) buff = neigh->arp_queue.next; __skb_unlink(buff, &neigh->arp_queue); kfree_skb(buff); + NEIGH_CACHE_STAT_INC(neigh->tbl, unres_discards); } __skb_queue_tail(&neigh->arp_queue, skb); } @@ -2462,12 +2463,12 @@ static int neigh_stat_seq_show(struct seq_file *seq, void *v) struct neigh_statistics *st = v; if (v == SEQ_START_TOKEN) { - seq_printf(seq, "entries allocs destroys hash_grows lookups hits res_failed rcv_probes_mcast rcv_probes_ucast periodic_gc_runs forced_gc_runs\n"); + seq_printf(seq, "entries allocs destroys hash_grows lookups hits res_failed rcv_probes_mcast rcv_probes_ucast periodic_gc_runs forced_gc_runs unresolved_discards\n"); return 0; } seq_printf(seq, "%08x %08lx %08lx %08lx %08lx %08lx %08lx " - "%08lx %08lx %08lx %08lx\n", + "%08lx %08lx %08lx %08lx %08lx\n", atomic_read(&tbl->entries), st->allocs, @@ -2483,7 +2484,8 @@ static int neigh_stat_seq_show(struct seq_file *seq, void *v) st->rcv_probes_ucast, st->periodic_gc_runs, - st->forced_gc_runs + st->forced_gc_runs, + st->unres_discards ); return 0; -- cgit v1.3-8-gc7d7 From e8a0464cc950972824e2e128028ae3db666ec1ed Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 00:34:19 -0700 Subject: netdev: Allocate multiple queues for TX. alloc_netdev_mq() now allocates an array of netdev_queue structures for TX, based upon the queue_count argument. Furthermore, all accesses to the TX queues are now vectored through the netdev_get_tx_queue() and netdev_for_each_tx_queue() interfaces. This makes it easy to grep the tree for all things that want to get to a TX queue of a net device. Problem spots which are not really multiqueue aware yet, and only work with one queue, can easily be spotted by grepping for all netdev_get_tx_queue() calls that pass in a zero index. Signed-off-by: David S. Miller --- drivers/net/bonding/bond_main.c | 6 +- drivers/net/hamradio/bpqether.c | 6 +- drivers/net/ifb.c | 12 ++- drivers/net/macvlan.c | 6 +- drivers/net/wireless/hostap/hostap_hw.c | 6 +- include/linux/netdevice.h | 69 ++++++++----- include/net/sch_generic.h | 37 +++++-- net/8021q/vlan_dev.c | 10 +- net/core/dev.c | 40 +++++-- net/core/rtnetlink.c | 2 +- net/mac80211/main.c | 4 +- net/mac80211/wme.c | 12 +-- net/netrom/af_netrom.c | 6 +- net/rose/af_rose.c | 6 +- net/sched/cls_api.c | 4 +- net/sched/sch_api.c | 32 ++++-- net/sched/sch_generic.c | 178 +++++++++++++++++++++++--------- net/sched/sch_teql.c | 21 ++-- 18 files changed, 320 insertions(+), 137 deletions(-) (limited to 'net/core') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index fd87dbe7999a..9737c06045d6 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -5042,7 +5042,9 @@ static int bond_check_params(struct bond_params *params) static struct lock_class_key bonding_netdev_xmit_lock_key; -static void bond_set_lockdep_class_one(struct netdev_queue *txq) +static void bond_set_lockdep_class_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &bonding_netdev_xmit_lock_key); @@ -5050,7 +5052,7 @@ static void bond_set_lockdep_class_one(struct netdev_queue *txq) static void bond_set_lockdep_class(struct net_device *dev) { - bond_set_lockdep_class_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, bond_set_lockdep_class_one, NULL); } /* Create a new bond based on the specified name and bonding parameters. diff --git a/drivers/net/hamradio/bpqether.c b/drivers/net/hamradio/bpqether.c index fb186b8c3d4d..b6500b2aacf2 100644 --- a/drivers/net/hamradio/bpqether.c +++ b/drivers/net/hamradio/bpqether.c @@ -124,14 +124,16 @@ static LIST_HEAD(bpq_devices); */ static struct lock_class_key bpq_netdev_xmit_lock_key; -static void bpq_set_lockdep_class_one(struct netdev_queue *txq) +static void bpq_set_lockdep_class_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &bpq_netdev_xmit_lock_key); } static void bpq_set_lockdep_class(struct net_device *dev) { - bpq_set_lockdep_class_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, bpq_set_lockdep_class_one, NULL); } /* ------------------------------------------------------------------------ */ diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c index ccbd6554f6eb..897b05e79ed0 100644 --- a/drivers/net/ifb.c +++ b/drivers/net/ifb.c @@ -229,14 +229,20 @@ module_param(numifbs, int, 0); MODULE_PARM_DESC(numifbs, "Number of ifb devices"); /* - * dev_ifb->tx_queue.lock is usually taken after dev->rx_queue.lock, + * dev_ifb's TX queue lock is usually taken after dev->rx_queue.lock, * reversely to e.g. qdisc_lock_tree(). It should be safe until - * ifb doesn't take dev->tx_queue.lock with dev_ifb->rx_queue.lock. + * ifb doesn't take dev's TX queue lock with dev_ifb->rx_queue.lock. * But lockdep should know that ifb has different locks from dev. */ static struct lock_class_key ifb_tx_queue_lock_key; static struct lock_class_key ifb_rx_queue_lock_key; +static void set_tx_lockdep_key(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) +{ + lockdep_set_class(&txq->lock, &ifb_tx_queue_lock_key); +} static int __init ifb_init_one(int index) { @@ -258,7 +264,7 @@ static int __init ifb_init_one(int index) if (err < 0) goto err; - lockdep_set_class(&dev_ifb->tx_queue.lock, &ifb_tx_queue_lock_key); + netdev_for_each_tx_queue(dev_ifb, set_tx_lockdep_key, NULL); lockdep_set_class(&dev_ifb->rx_queue.lock, &ifb_rx_queue_lock_key); return 0; diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c index 980001c2cf96..72745ce588c6 100644 --- a/drivers/net/macvlan.c +++ b/drivers/net/macvlan.c @@ -285,7 +285,9 @@ static struct lock_class_key macvlan_netdev_xmit_lock_key; #define MACVLAN_STATE_MASK \ ((1<<__LINK_STATE_NOCARRIER) | (1<<__LINK_STATE_DORMANT)) -static void macvlan_set_lockdep_class_one(struct netdev_queue *txq) +static void macvlan_set_lockdep_class_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &macvlan_netdev_xmit_lock_key); @@ -293,7 +295,7 @@ static void macvlan_set_lockdep_class_one(struct netdev_queue *txq) static void macvlan_set_lockdep_class(struct net_device *dev) { - macvlan_set_lockdep_class_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, macvlan_set_lockdep_class_one, NULL); } static int macvlan_init(struct net_device *dev) diff --git a/drivers/net/wireless/hostap/hostap_hw.c b/drivers/net/wireless/hostap/hostap_hw.c index c1f4bb005d92..13d5882f1f21 100644 --- a/drivers/net/wireless/hostap/hostap_hw.c +++ b/drivers/net/wireless/hostap/hostap_hw.c @@ -3102,7 +3102,9 @@ static void prism2_clear_set_tim_queue(local_info_t *local) */ static struct lock_class_key hostap_netdev_xmit_lock_key; -static void prism2_set_lockdep_class_one(struct netdev_queue *txq) +static void prism2_set_lockdep_class_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &hostap_netdev_xmit_lock_key); @@ -3110,7 +3112,7 @@ static void prism2_set_lockdep_class_one(struct netdev_queue *txq) static void prism2_set_lockdep_class(struct net_device *dev) { - prism2_set_lockdep_class_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, prism2_set_lockdep_class_one, NULL); } static struct net_device * diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 570cf7affa72..f25d4f5a31b0 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -463,7 +463,7 @@ struct netdev_queue { struct Qdisc *qdisc_sleeping; struct list_head qdisc_list; struct netdev_queue *next_sched; -}; +} ____cacheline_aligned_in_smp; /* * The DEVICE structure. @@ -641,7 +641,9 @@ struct net_device unsigned char broadcast[MAX_ADDR_LEN]; /* hw bcast add */ struct netdev_queue rx_queue; - struct netdev_queue tx_queue ____cacheline_aligned_in_smp; + + struct netdev_queue *_tx ____cacheline_aligned_in_smp; + unsigned int num_tx_queues; unsigned long tx_queue_len; /* Max frames per queue allowed */ /* @@ -764,6 +766,25 @@ struct net_device #define NETDEV_ALIGN 32 #define NETDEV_ALIGN_CONST (NETDEV_ALIGN - 1) +static inline +struct netdev_queue *netdev_get_tx_queue(const struct net_device *dev, + unsigned int index) +{ + return &dev->_tx[index]; +} + +static inline void netdev_for_each_tx_queue(struct net_device *dev, + void (*f)(struct net_device *, + struct netdev_queue *, + void *), + void *arg) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) + f(dev, &dev->_tx[i], arg); +} + /* * Net namespace inlines */ @@ -977,7 +998,7 @@ static inline void netif_schedule_queue(struct netdev_queue *txq) static inline void netif_schedule(struct net_device *dev) { - netif_schedule_queue(&dev->tx_queue); + netif_schedule_queue(netdev_get_tx_queue(dev, 0)); } /** @@ -993,7 +1014,7 @@ static inline void netif_tx_start_queue(struct netdev_queue *dev_queue) static inline void netif_start_queue(struct net_device *dev) { - netif_tx_start_queue(&dev->tx_queue); + netif_tx_start_queue(netdev_get_tx_queue(dev, 0)); } /** @@ -1017,7 +1038,7 @@ static inline void netif_tx_wake_queue(struct netdev_queue *dev_queue) static inline void netif_wake_queue(struct net_device *dev) { - netif_tx_wake_queue(&dev->tx_queue); + netif_tx_wake_queue(netdev_get_tx_queue(dev, 0)); } /** @@ -1034,7 +1055,7 @@ static inline void netif_tx_stop_queue(struct netdev_queue *dev_queue) static inline void netif_stop_queue(struct net_device *dev) { - netif_tx_stop_queue(&dev->tx_queue); + netif_tx_stop_queue(netdev_get_tx_queue(dev, 0)); } /** @@ -1050,7 +1071,7 @@ static inline int netif_tx_queue_stopped(const struct netdev_queue *dev_queue) static inline int netif_queue_stopped(const struct net_device *dev) { - return netif_tx_queue_stopped(&dev->tx_queue); + return netif_tx_queue_stopped(netdev_get_tx_queue(dev, 0)); } /** @@ -1134,7 +1155,7 @@ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index) #endif if (test_and_clear_bit(__QUEUE_STATE_XOFF, &dev->egress_subqueue[queue_index].state)) - __netif_schedule(&dev->tx_queue); + __netif_schedule(netdev_get_tx_queue(dev, 0)); } /** @@ -1430,18 +1451,19 @@ static inline void __netif_tx_lock(struct netdev_queue *txq, int cpu) static inline void netif_tx_lock(struct net_device *dev) { - __netif_tx_lock(&dev->tx_queue, smp_processor_id()); -} + int cpu = smp_processor_id(); + unsigned int i; -static inline void __netif_tx_lock_bh(struct netdev_queue *txq) -{ - spin_lock_bh(&txq->_xmit_lock); - txq->xmit_lock_owner = smp_processor_id(); + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + __netif_tx_lock(txq, cpu); + } } static inline void netif_tx_lock_bh(struct net_device *dev) { - __netif_tx_lock_bh(&dev->tx_queue); + local_bh_disable(); + netif_tx_lock(dev); } static inline int __netif_tx_trylock(struct netdev_queue *txq) @@ -1454,7 +1476,7 @@ static inline int __netif_tx_trylock(struct netdev_queue *txq) static inline int netif_tx_trylock(struct net_device *dev) { - return __netif_tx_trylock(&dev->tx_queue); + return __netif_tx_trylock(netdev_get_tx_queue(dev, 0)); } static inline void __netif_tx_unlock(struct netdev_queue *txq) @@ -1465,18 +1487,19 @@ static inline void __netif_tx_unlock(struct netdev_queue *txq) static inline void netif_tx_unlock(struct net_device *dev) { - __netif_tx_unlock(&dev->tx_queue); -} + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + __netif_tx_unlock(txq); + } -static inline void __netif_tx_unlock_bh(struct netdev_queue *txq) -{ - txq->xmit_lock_owner = -1; - spin_unlock_bh(&txq->_xmit_lock); } static inline void netif_tx_unlock_bh(struct net_device *dev) { - __netif_tx_unlock_bh(&dev->tx_queue); + netif_tx_unlock(dev); + local_bh_enable(); } #define HARD_TX_LOCK(dev, txq, cpu) { \ diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index 5ba66b555578..b47f556c66f8 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -230,32 +230,47 @@ extern void tcf_destroy_chain(struct tcf_proto **fl); /* Reset all TX qdiscs of a device. */ static inline void qdisc_reset_all_tx(struct net_device *dev) { - qdisc_reset(dev->tx_queue.qdisc); + unsigned int i; + for (i = 0; i < dev->num_tx_queues; i++) + qdisc_reset(netdev_get_tx_queue(dev, i)->qdisc); } /* Are all TX queues of the device empty? */ static inline bool qdisc_all_tx_empty(const struct net_device *dev) { - const struct netdev_queue *txq = &dev->tx_queue; - const struct Qdisc *q = txq->qdisc; + unsigned int i; + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + const struct Qdisc *q = txq->qdisc; - return (q->q.qlen == 0); + if (q->q.qlen) + return false; + } + return true; } /* Are any of the TX qdiscs changing? */ static inline bool qdisc_tx_changing(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; - - return (txq->qdisc != txq->qdisc_sleeping); + unsigned int i; + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + if (txq->qdisc != txq->qdisc_sleeping) + return true; + } + return false; } -/* Is the device using the noop qdisc? */ +/* Is the device using the noop qdisc on all queues? */ static inline bool qdisc_tx_is_noop(const struct net_device *dev) { - const struct netdev_queue *txq = &dev->tx_queue; - - return (txq->qdisc == &noop_qdisc); + unsigned int i; + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + if (txq->qdisc != &noop_qdisc) + return false; + } + return true; } static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch, diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 6b985f23fd9f..f42bc2b26b85 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -570,16 +570,18 @@ static void vlan_dev_set_rx_mode(struct net_device *vlan_dev) */ static struct lock_class_key vlan_netdev_xmit_lock_key; -static void vlan_dev_set_lockdep_one(struct netdev_queue *txq, - int subclass) +static void vlan_dev_set_lockdep_one(struct net_device *dev, + struct netdev_queue *txq, + void *_subclass) { lockdep_set_class_and_subclass(&txq->_xmit_lock, - &vlan_netdev_xmit_lock_key, subclass); + &vlan_netdev_xmit_lock_key, + *(int *)_subclass); } static void vlan_dev_set_lockdep_class(struct net_device *dev, int subclass) { - vlan_dev_set_lockdep_one(&dev->tx_queue, subclass); + netdev_for_each_tx_queue(dev, vlan_dev_set_lockdep_one, &subclass); } static const struct header_ops vlan_header_ops = { diff --git a/net/core/dev.c b/net/core/dev.c index 9b49f74a9820..69378f250695 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1666,6 +1666,12 @@ out_kfree_skb: * --BLG */ +static struct netdev_queue *dev_pick_tx(struct net_device *dev, + struct sk_buff *skb) +{ + return netdev_get_tx_queue(dev, 0); +} + int dev_queue_xmit(struct sk_buff *skb) { struct net_device *dev = skb->dev; @@ -1702,7 +1708,7 @@ int dev_queue_xmit(struct sk_buff *skb) } gso: - txq = &dev->tx_queue; + txq = dev_pick_tx(dev, skb); spin_lock_prefetch(&txq->lock); /* Disable soft irqs for various locks below. Also @@ -3788,8 +3794,9 @@ static void rollback_registered(struct net_device *dev) dev_put(dev); } -static void __netdev_init_queue_locks_one(struct netdev_queue *dev_queue, - struct net_device *dev) +static void __netdev_init_queue_locks_one(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_unused) { spin_lock_init(&dev_queue->_xmit_lock); netdev_set_lockdep_class(&dev_queue->_xmit_lock, dev->type); @@ -3798,8 +3805,8 @@ static void __netdev_init_queue_locks_one(struct netdev_queue *dev_queue, static void netdev_init_queue_locks(struct net_device *dev) { - __netdev_init_queue_locks_one(&dev->tx_queue, dev); - __netdev_init_queue_locks_one(&dev->rx_queue, dev); + netdev_for_each_tx_queue(dev, __netdev_init_queue_locks_one, NULL); + __netdev_init_queue_locks_one(dev, &dev->rx_queue, NULL); } /** @@ -4119,7 +4126,8 @@ static struct net_device_stats *internal_stats(struct net_device *dev) } static void netdev_init_one_queue(struct net_device *dev, - struct netdev_queue *queue) + struct netdev_queue *queue, + void *_unused) { spin_lock_init(&queue->lock); queue->dev = dev; @@ -4127,8 +4135,8 @@ static void netdev_init_one_queue(struct net_device *dev, static void netdev_init_queues(struct net_device *dev) { - netdev_init_one_queue(dev, &dev->rx_queue); - netdev_init_one_queue(dev, &dev->tx_queue); + netdev_init_one_queue(dev, &dev->rx_queue, NULL); + netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL); } /** @@ -4145,9 +4153,10 @@ static void netdev_init_queues(struct net_device *dev) struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, void (*setup)(struct net_device *), unsigned int queue_count) { - void *p; + struct netdev_queue *tx; struct net_device *dev; int alloc_size; + void *p; BUG_ON(strlen(name) >= sizeof(dev->name)); @@ -4167,11 +4176,22 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, return NULL; } + tx = kzalloc(sizeof(struct netdev_queue) * queue_count, GFP_KERNEL); + if (!tx) { + printk(KERN_ERR "alloc_netdev: Unable to allocate " + "tx qdiscs.\n"); + kfree(p); + return NULL; + } + dev = (struct net_device *) (((long)p + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST); dev->padded = (char *)dev - (char *)p; dev_net_set(dev, &init_net); + dev->_tx = tx; + dev->num_tx_queues = queue_count; + if (sizeof_priv) { dev->priv = ((char *)dev + ((sizeof(struct net_device) + @@ -4205,6 +4225,8 @@ void free_netdev(struct net_device *dev) { release_net(dev_net(dev)); + kfree(dev->_tx); + /* Compatibility with error handling in drivers */ if (dev->reg_state == NETREG_UNINITIALIZED) { kfree((char *)dev - dev->padded); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 8ef9f1db610e..71edb8b36341 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -636,7 +636,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, if (dev->master) NLA_PUT_U32(skb, IFLA_MASTER, dev->master->ifindex); - txq = &dev->tx_queue; + txq = netdev_get_tx_queue(dev, 0); if (txq->qdisc_sleeping) NLA_PUT_STRING(skb, IFLA_QDISC, txq->qdisc_sleeping->ops->id); diff --git a/net/mac80211/main.c b/net/mac80211/main.c index af0056e7e5b3..b486e634f4fe 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -621,7 +621,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* ensure that TX flow won't interrupt us * until the end of the call to requeue function */ - txq = &local->mdev->tx_queue; + txq = netdev_get_tx_queue(local->mdev, 0); spin_lock_bh(&txq->lock); /* create a new queue for this aggregation */ @@ -862,7 +862,7 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) /* avoid ordering issues: we are the only one that can modify * the content of the qdiscs */ - txq = &local->mdev->tx_queue; + txq = netdev_get_tx_queue(local->mdev, 0); spin_lock_bh(&txq->lock); /* remove the queue for this aggregation */ ieee80211_ht_agg_queue_remove(local, sta, tid, 1); diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 6ae43a3c7726..f014cd38c2d0 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -574,7 +574,7 @@ static struct Qdisc_ops wme_qdisc_ops __read_mostly = void ieee80211_install_qdisc(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); struct Qdisc *qdisc; qdisc = qdisc_create_dflt(dev, txq, @@ -596,7 +596,7 @@ void ieee80211_install_qdisc(struct net_device *dev) int ieee80211_qdisc_installed(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); return txq->qdisc_sleeping->ops == &wme_qdisc_ops; } @@ -617,7 +617,7 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, struct sta_info *sta, u16 tid) { int i; - struct netdev_queue *txq = &local->mdev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); struct ieee80211_sched_data *q = qdisc_priv(txq->qdisc_sleeping); DECLARE_MAC_BUF(mac); @@ -652,14 +652,14 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, } /** - * the caller needs to hold local->mdev->tx_queue.lock + * the caller needs to hold netdev_get_tx_queue(local->mdev, X)->lock */ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, struct sta_info *sta, u16 tid, u8 requeue) { struct ieee80211_hw *hw = &local->hw; - struct netdev_queue *txq = &local->mdev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); struct ieee80211_sched_data *q = qdisc_priv(txq->qdisc_sleeping); int agg_queue = sta->tid_to_tx_q[tid]; @@ -676,7 +676,7 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, void ieee80211_requeue(struct ieee80211_local *local, int queue) { - struct netdev_queue *txq = &local->mdev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); struct Qdisc *root_qd = txq->qdisc_sleeping; struct ieee80211_sched_data *q = qdisc_priv(root_qd); struct Qdisc *qdisc = q->queues[queue]; diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index 819afc449e1e..d41be0d66eb0 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -74,14 +74,16 @@ static const struct proto_ops nr_proto_ops; */ static struct lock_class_key nr_netdev_xmit_lock_key; -static void nr_set_lockdep_one(struct netdev_queue *txq) +static void nr_set_lockdep_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &nr_netdev_xmit_lock_key); } static void nr_set_lockdep_key(struct net_device *dev) { - nr_set_lockdep_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, nr_set_lockdep_one, NULL); } /* diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index 7dbbc0891623..f3a691f34909 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -75,14 +75,16 @@ ax25_address rose_callsign; */ static struct lock_class_key rose_netdev_xmit_lock_key; -static void rose_set_lockdep_one(struct netdev_queue *txq) +static void rose_set_lockdep_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &rose_netdev_xmit_lock_key); } static void rose_set_lockdep_key(struct net_device *dev) { - rose_set_lockdep_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, rose_set_lockdep_one, NULL); } /* diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index b483bbea6118..d0b0a9b14394 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -166,7 +166,7 @@ replay: /* Find qdisc */ if (!parent) { - struct netdev_queue *dev_queue = &dev->tx_queue; + struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, 0); q = dev_queue->qdisc_sleeping; parent = q->handle; } else { @@ -410,7 +410,7 @@ static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb) if ((dev = dev_get_by_index(&init_net, tcm->tcm_ifindex)) == NULL) return skb->len; - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); if (!tcm->tcm_parent) q = dev_queue->qdisc_sleeping; else diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 95873f8dd37c..830ccc544a15 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -183,9 +183,8 @@ EXPORT_SYMBOL(unregister_qdisc); (root qdisc, all its children, children of children etc.) */ -struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) +static struct Qdisc *__qdisc_lookup(struct netdev_queue *dev_queue, u32 handle) { - struct netdev_queue *dev_queue = &dev->tx_queue; struct Qdisc *q; list_for_each_entry(q, &dev_queue->qdisc_list, list) { @@ -195,6 +194,19 @@ struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) return NULL; } +struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + struct Qdisc *q = __qdisc_lookup(txq, handle); + if (q) + return q; + } + return NULL; +} + static struct Qdisc *qdisc_leaf(struct Qdisc *p, u32 classid) { unsigned long cl; @@ -462,7 +474,7 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) } } else { - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); oqdisc = dev_queue->qdisc_sleeping; /* Prune old scheduler */ @@ -742,7 +754,8 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) q = dev->rx_queue.qdisc; } } else { - struct netdev_queue *dev_queue = &dev->tx_queue; + struct netdev_queue *dev_queue; + dev_queue = netdev_get_tx_queue(dev, 0); q = dev_queue->qdisc_sleeping; } if (!q) @@ -817,7 +830,8 @@ replay: q = dev->rx_queue.qdisc; } } else { - struct netdev_queue *dev_queue = &dev->tx_queue; + struct netdev_queue *dev_queue; + dev_queue = netdev_get_tx_queue(dev, 0); q = dev_queue->qdisc_sleeping; } @@ -899,7 +913,7 @@ create_n_graft: tcm->tcm_parent, tcm->tcm_parent, tca, &err); else - q = qdisc_create(dev, &dev->tx_queue, + q = qdisc_create(dev, netdev_get_tx_queue(dev, 0), tcm->tcm_parent, tcm->tcm_handle, tca, &err); if (q == NULL) { @@ -1025,7 +1039,7 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) if (idx > s_idx) s_q_idx = 0; q_idx = 0; - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (q_idx < s_q_idx) { q_idx++; @@ -1098,7 +1112,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) /* Step 1. Determine qdisc handle X:0 */ - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); if (pid != TC_H_ROOT) { u32 qid1 = TC_H_MAJ(pid); @@ -1275,7 +1289,7 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (t < s_t || !q->ops->cl_ops || (tcm->tcm_parent && diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 243de935b182..4e2b865cbba0 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -40,20 +40,30 @@ */ void qdisc_lock_tree(struct net_device *dev) - __acquires(dev->tx_queue.lock) __acquires(dev->rx_queue.lock) { - spin_lock_bh(&dev->tx_queue.lock); + unsigned int i; + + local_bh_disable(); + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + spin_lock(&txq->lock); + } spin_lock(&dev->rx_queue.lock); } EXPORT_SYMBOL(qdisc_lock_tree); void qdisc_unlock_tree(struct net_device *dev) __releases(dev->rx_queue.lock) - __releases(dev->tx_queue.lock) { + unsigned int i; + spin_unlock(&dev->rx_queue.lock); - spin_unlock_bh(&dev->tx_queue.lock); + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + spin_unlock(&txq->lock); + } + local_bh_enable(); } EXPORT_SYMBOL(qdisc_unlock_tree); @@ -212,22 +222,37 @@ void __qdisc_run(struct netdev_queue *txq) static void dev_watchdog(unsigned long arg) { struct net_device *dev = (struct net_device *)arg; - struct netdev_queue *txq = &dev->tx_queue; netif_tx_lock(dev); - if (txq->qdisc != &noop_qdisc) { + if (!qdisc_tx_is_noop(dev)) { if (netif_device_present(dev) && netif_running(dev) && netif_carrier_ok(dev)) { - if (netif_queue_stopped(dev) && - time_after(jiffies, dev->trans_start + dev->watchdog_timeo)) { + int some_queue_stopped = 0; + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq; + + txq = netdev_get_tx_queue(dev, i); + if (netif_tx_queue_stopped(txq)) { + some_queue_stopped = 1; + break; + } + } - printk(KERN_INFO "NETDEV WATCHDOG: %s: transmit timed out\n", + if (some_queue_stopped && + time_after(jiffies, (dev->trans_start + + dev->watchdog_timeo))) { + printk(KERN_INFO "NETDEV WATCHDOG: %s: " + "transmit timed out\n", dev->name); dev->tx_timeout(dev); WARN_ON_ONCE(1); } - if (!mod_timer(&dev->watchdog_timer, round_jiffies(jiffies + dev->watchdog_timeo))) + if (!mod_timer(&dev->watchdog_timer, + round_jiffies(jiffies + + dev->watchdog_timeo))) dev_hold(dev); } } @@ -542,9 +567,55 @@ void qdisc_destroy(struct Qdisc *qdisc) } EXPORT_SYMBOL(qdisc_destroy); +static bool dev_all_qdisc_sleeping_noop(struct net_device *dev) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + + if (txq->qdisc_sleeping != &noop_qdisc) + return false; + } + return true; +} + +static void attach_one_default_qdisc(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_unused) +{ + struct Qdisc *qdisc; + + if (dev->tx_queue_len) { + qdisc = qdisc_create_dflt(dev, dev_queue, + &pfifo_fast_ops, TC_H_ROOT); + if (!qdisc) { + printk(KERN_INFO "%s: activation failed\n", dev->name); + return; + } + list_add_tail(&qdisc->list, &dev_queue->qdisc_list); + } else { + qdisc = &noqueue_qdisc; + } + dev_queue->qdisc_sleeping = qdisc; +} + +static void transition_one_qdisc(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_need_watchdog) +{ + int *need_watchdog_p = _need_watchdog; + + spin_lock_bh(&dev_queue->lock); + rcu_assign_pointer(dev_queue->qdisc, dev_queue->qdisc_sleeping); + if (dev_queue->qdisc != &noqueue_qdisc) + *need_watchdog_p = 1; + spin_unlock_bh(&dev_queue->lock); +} + void dev_activate(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; + int need_watchdog; /* No queueing discipline is attached to device; create default one i.e. pfifo_fast for devices, @@ -552,39 +623,27 @@ void dev_activate(struct net_device *dev) virtual interfaces */ - if (txq->qdisc_sleeping == &noop_qdisc) { - struct Qdisc *qdisc; - if (dev->tx_queue_len) { - qdisc = qdisc_create_dflt(dev, txq, - &pfifo_fast_ops, - TC_H_ROOT); - if (qdisc == NULL) { - printk(KERN_INFO "%s: activation failed\n", dev->name); - return; - } - list_add_tail(&qdisc->list, &txq->qdisc_list); - } else { - qdisc = &noqueue_qdisc; - } - txq->qdisc_sleeping = qdisc; - } + if (dev_all_qdisc_sleeping_noop(dev)) + netdev_for_each_tx_queue(dev, attach_one_default_qdisc, NULL); if (!netif_carrier_ok(dev)) /* Delay activation until next carrier-on event */ return; - spin_lock_bh(&txq->lock); - rcu_assign_pointer(txq->qdisc, txq->qdisc_sleeping); - if (txq->qdisc != &noqueue_qdisc) { + need_watchdog = 0; + netdev_for_each_tx_queue(dev, transition_one_qdisc, &need_watchdog); + + if (need_watchdog) { dev->trans_start = jiffies; dev_watchdog_up(dev); } - spin_unlock_bh(&txq->lock); } -static void dev_deactivate_queue(struct netdev_queue *dev_queue, - struct Qdisc *qdisc_default) +static void dev_deactivate_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_qdisc_default) { + struct Qdisc *qdisc_default = _qdisc_default; struct Qdisc *qdisc; struct sk_buff *skb; @@ -603,12 +662,35 @@ static void dev_deactivate_queue(struct netdev_queue *dev_queue, kfree_skb(skb); } +static bool some_qdisc_is_running(struct net_device *dev, int lock) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *dev_queue; + int val; + + dev_queue = netdev_get_tx_queue(dev, i); + + if (lock) + spin_lock_bh(&dev_queue->lock); + + val = test_bit(__QUEUE_STATE_QDISC_RUNNING, &dev_queue->state); + + if (lock) + spin_unlock_bh(&dev_queue->lock); + + if (val) + return true; + } + return false; +} + void dev_deactivate(struct net_device *dev) { - struct netdev_queue *dev_queue = &dev->tx_queue; - int running; + bool running; - dev_deactivate_queue(dev_queue, &noop_qdisc); + netdev_for_each_tx_queue(dev, dev_deactivate_queue, &noop_qdisc); dev_watchdog_down(dev); @@ -617,17 +699,14 @@ void dev_deactivate(struct net_device *dev) /* Wait for outstanding qdisc_run calls. */ do { - while (test_bit(__QUEUE_STATE_QDISC_RUNNING, &dev_queue->state)) + while (some_qdisc_is_running(dev, 0)) yield(); /* * Double-check inside queue lock to ensure that all effects * of the queue run are visible when we return. */ - spin_lock_bh(&dev_queue->lock); - running = test_bit(__QUEUE_STATE_QDISC_RUNNING, - &dev_queue->state); - spin_unlock_bh(&dev_queue->lock); + running = some_qdisc_is_running(dev, 1); /* * The running flag should never be set at this point because @@ -642,8 +721,10 @@ void dev_deactivate(struct net_device *dev) static void dev_init_scheduler_queue(struct net_device *dev, struct netdev_queue *dev_queue, - struct Qdisc *qdisc) + void *_qdisc) { + struct Qdisc *qdisc = _qdisc; + dev_queue->qdisc = qdisc; dev_queue->qdisc_sleeping = qdisc; INIT_LIST_HEAD(&dev_queue->qdisc_list); @@ -652,18 +733,19 @@ static void dev_init_scheduler_queue(struct net_device *dev, void dev_init_scheduler(struct net_device *dev) { qdisc_lock_tree(dev); - dev_init_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); + netdev_for_each_tx_queue(dev, dev_init_scheduler_queue, &noop_qdisc); dev_init_scheduler_queue(dev, &dev->rx_queue, NULL); qdisc_unlock_tree(dev); setup_timer(&dev->watchdog_timer, dev_watchdog, (unsigned long)dev); } -static void dev_shutdown_scheduler_queue(struct net_device *dev, - struct netdev_queue *dev_queue, - struct Qdisc *qdisc_default) +static void shutdown_scheduler_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_qdisc_default) { struct Qdisc *qdisc = dev_queue->qdisc_sleeping; + struct Qdisc *qdisc_default = _qdisc_default; if (qdisc) { dev_queue->qdisc = qdisc_default; @@ -676,8 +758,8 @@ static void dev_shutdown_scheduler_queue(struct net_device *dev, void dev_shutdown(struct net_device *dev) { qdisc_lock_tree(dev); - dev_shutdown_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); - dev_shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); + netdev_for_each_tx_queue(dev, shutdown_scheduler_queue, &noop_qdisc); + shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); BUG_TRAP(!timer_pending(&dev->watchdog_timer)); qdisc_unlock_tree(dev); } diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 8ac05981be20..44a2c3451f4d 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -111,7 +111,7 @@ teql_dequeue(struct Qdisc* sch) struct sk_buff *skb; skb = __skb_dequeue(&dat->q); - dat_queue = &dat->m->dev->tx_queue; + dat_queue = netdev_get_tx_queue(dat->m->dev, 0); if (skb == NULL) { struct net_device *m = qdisc_dev(dat_queue->qdisc); if (m) { @@ -155,10 +155,13 @@ teql_destroy(struct Qdisc* sch) if (q == master->slaves) { master->slaves = NEXT_SLAVE(q); if (q == master->slaves) { + struct netdev_queue *txq; + + txq = netdev_get_tx_queue(master->dev, 0); master->slaves = NULL; - spin_lock_bh(&master->dev->tx_queue.lock); - qdisc_reset(master->dev->tx_queue.qdisc); - spin_unlock_bh(&master->dev->tx_queue.lock); + spin_lock_bh(&txq->lock); + qdisc_reset(txq->qdisc); + spin_unlock_bh(&txq->lock); } } skb_queue_purge(&dat->q); @@ -218,7 +221,8 @@ static int teql_qdisc_init(struct Qdisc *sch, struct nlattr *opt) static int __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - struct teql_sched_data *q = qdisc_priv(dev->tx_queue.qdisc); + struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, 0); + struct teql_sched_data *q = qdisc_priv(dev_queue->qdisc); struct neighbour *mn = skb->dst->neighbour; struct neighbour *n = q->ncache; @@ -254,7 +258,8 @@ __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device * static inline int teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - if (dev->tx_queue.qdisc == &noop_qdisc) + struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); + if (txq->qdisc == &noop_qdisc) return -ENODEV; if (dev->header_ops == NULL || @@ -285,8 +290,10 @@ restart: do { struct net_device *slave = qdisc_dev(q); + struct netdev_queue *slave_txq; - if (slave->tx_queue.qdisc_sleeping != q) + slave_txq = netdev_get_tx_queue(slave, 0); + if (slave_txq->qdisc_sleeping != q) continue; if (netif_queue_stopped(slave) || __netif_subqueue_stopped(slave, subq) || -- cgit v1.3-8-gc7d7 From fd2ea0a79faad824258af5dcec1927aa24d81c16 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 01:56:23 -0700 Subject: net: Use queue aware tests throughout. This effectively "flips the switch" by making the core networking and multiqueue-aware drivers use the new TX multiqueue structures. Non-multiqueue drivers need no changes. The interfaces they use such as netif_stop_queue() degenerate into an operation on TX queue zero. So everything "just works" for them. Code that really wants to do "X" to all TX queues now invokes a routine that does so, such as netif_tx_wake_all_queues(), netif_tx_stop_all_queues(), etc. pktgen and netpoll required a little bit more surgery than the others. In particular the pktgen changes, whilst functional, could be largely improved. The initial check in pktgen_xmit() will sometimes check the wrong queue, which is mostly harmless. The thing to do is probably to invoke fill_packet() earlier. The bulk of the netpoll changes is to make the code operate solely on the TX queue indicated by by the SKB queue mapping. Setting of the SKB queue mapping is entirely confined inside of net/core/dev.c:dev_pick_tx(). If we end up needing any kind of special semantics (drops, for example) it will be implemented here. Finally, we now have a "real_num_tx_queues" which is where the driver indicates how many TX queues are actually active. With IGB changes from Jeff Kirsher. Signed-off-by: David S. Miller --- drivers/net/cpmac.c | 20 ++++------ drivers/net/igb/igb_main.c | 19 +++------ drivers/net/ixgbe/ixgbe_ethtool.c | 10 +---- drivers/net/ixgbe/ixgbe_main.c | 15 +++---- drivers/net/s2io.c | 48 +++++++++-------------- include/linux/netdevice.h | 82 ++++++++++++++++++++++++++++++++++----- include/net/pkt_sched.h | 4 +- net/core/dev.c | 28 ++++++------- net/core/netpoll.c | 24 +++++++----- net/core/pktgen.c | 69 ++++++++++++++++++++------------ net/sched/sch_generic.c | 5 +-- net/sched/sch_teql.c | 6 +-- 12 files changed, 187 insertions(+), 143 deletions(-) (limited to 'net/core') diff --git a/drivers/net/cpmac.c b/drivers/net/cpmac.c index 7c7b54e4828e..fbd4280c102c 100644 --- a/drivers/net/cpmac.c +++ b/drivers/net/cpmac.c @@ -544,7 +544,7 @@ fatal_error: spin_unlock(&priv->rx_lock); netif_rx_complete(priv->dev, napi); - netif_stop_queue(priv->dev); + netif_tx_stop_all_queues(priv->dev); napi_disable(&priv->napi); atomic_inc(&priv->reset_pending); @@ -750,9 +750,7 @@ static void cpmac_hw_error(struct work_struct *work) barrier(); atomic_dec(&priv->reset_pending); - for (i = 0; i < CPMAC_QUEUES; i++) - netif_wake_subqueue(priv->dev, i); - netif_wake_queue(priv->dev); + netif_tx_wake_all_queues(priv->dev); cpmac_write(priv->regs, CPMAC_MAC_INT_ENABLE, 3); } @@ -781,7 +779,7 @@ static void cpmac_check_status(struct net_device *dev) dev->name, tx_code, tx_channel, macstatus); } - netif_stop_queue(dev); + netif_tx_stop_all_queues(dev); cpmac_hw_stop(dev); if (schedule_work(&priv->reset_work)) atomic_inc(&priv->reset_pending); @@ -842,9 +840,7 @@ static void cpmac_tx_timeout(struct net_device *dev) barrier(); atomic_dec(&priv->reset_pending); - netif_wake_queue(priv->dev); - for (i = 0; i < CPMAC_QUEUES; i++) - netif_wake_subqueue(dev, i); + netif_tx_wake_all_queues(priv->dev); } static int cpmac_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) @@ -935,7 +931,7 @@ static void cpmac_adjust_link(struct net_device *dev) spin_lock(&priv->lock); if (priv->phy->link) { - netif_start_queue(dev); + netif_tx_start_all_queues(dev); if (priv->phy->duplex != priv->oldduplex) { new_state = 1; priv->oldduplex = priv->phy->duplex; @@ -949,10 +945,10 @@ static void cpmac_adjust_link(struct net_device *dev) if (!priv->oldlink) { new_state = 1; priv->oldlink = 1; - netif_schedule(dev); + netif_tx_schedule_all(dev); } } else if (priv->oldlink) { - netif_stop_queue(dev); + netif_tx_stop_all_queues(dev); new_state = 1; priv->oldlink = 0; priv->oldspeed = 0; @@ -1072,7 +1068,7 @@ static int cpmac_stop(struct net_device *dev) struct cpmac_priv *priv = netdev_priv(dev); struct resource *mem; - netif_stop_queue(dev); + netif_tx_stop_all_queues(dev); cancel_work_sync(&priv->reset_work); napi_disable(&priv->napi); diff --git a/drivers/net/igb/igb_main.c b/drivers/net/igb/igb_main.c index 471c194cd54e..81bba6983dde 100644 --- a/drivers/net/igb/igb_main.c +++ b/drivers/net/igb/igb_main.c @@ -533,7 +533,7 @@ msi_only: adapter->flags |= IGB_FLAG_HAS_MSI; /* Notify the stack of the (possibly) reduced Tx Queue count. */ - adapter->netdev->egress_subqueue_count = adapter->num_tx_queues; + adapter->netdev->real_num_tx_queues = adapter->num_tx_queues; return; } @@ -821,9 +821,7 @@ void igb_down(struct igb_adapter *adapter) wr32(E1000_RCTL, rctl & ~E1000_RCTL_EN); /* flush and sleep below */ - netif_stop_queue(netdev); - for (i = 0; i < adapter->num_tx_queues; i++) - netif_stop_subqueue(netdev, i); + netif_tx_stop_all_queues(netdev); /* disable transmits in the hardware */ tctl = rd32(E1000_TCTL); @@ -1266,9 +1264,7 @@ static int __devinit igb_probe(struct pci_dev *pdev, /* tell the stack to leave us alone until igb_open() is called */ netif_carrier_off(netdev); - netif_stop_queue(netdev); - for (i = 0; i < adapter->num_tx_queues; i++) - netif_stop_subqueue(netdev, i); + netif_tx_stop_all_queues(netdev); strcpy(netdev->name, "eth%d"); err = register_netdev(netdev); @@ -2315,7 +2311,6 @@ static void igb_watchdog_task(struct work_struct *work) struct e1000_mac_info *mac = &adapter->hw.mac; u32 link; s32 ret_val; - int i; if ((netif_carrier_ok(netdev)) && (rd32(E1000_STATUS) & E1000_STATUS_LU)) @@ -2371,9 +2366,7 @@ static void igb_watchdog_task(struct work_struct *work) } netif_carrier_on(netdev); - netif_wake_queue(netdev); - for (i = 0; i < adapter->num_tx_queues; i++) - netif_wake_subqueue(netdev, i); + netif_tx_wake_all_queues(netdev); if (!test_bit(__IGB_DOWN, &adapter->state)) mod_timer(&adapter->phy_info_timer, @@ -2385,9 +2378,7 @@ static void igb_watchdog_task(struct work_struct *work) adapter->link_duplex = 0; dev_info(&adapter->pdev->dev, "NIC Link is Down\n"); netif_carrier_off(netdev); - netif_stop_queue(netdev); - for (i = 0; i < adapter->num_tx_queues; i++) - netif_stop_subqueue(netdev, i); + netif_tx_stop_all_queues(netdev); if (!test_bit(__IGB_DOWN, &adapter->state)) mod_timer(&adapter->phy_info_timer, round_jiffies(jiffies + 2 * HZ)); diff --git a/drivers/net/ixgbe/ixgbe_ethtool.c b/drivers/net/ixgbe/ixgbe_ethtool.c index 81b769093d22..3efe5dda10af 100644 --- a/drivers/net/ixgbe/ixgbe_ethtool.c +++ b/drivers/net/ixgbe/ixgbe_ethtool.c @@ -252,16 +252,10 @@ static int ixgbe_set_tso(struct net_device *netdev, u32 data) netdev->features |= NETIF_F_TSO; netdev->features |= NETIF_F_TSO6; } else { - struct ixgbe_adapter *adapter = netdev_priv(netdev); - int i; - netif_stop_queue(netdev); - for (i = 0; i < adapter->num_tx_queues; i++) - netif_stop_subqueue(netdev, i); + netif_tx_stop_all_queues(netdev); netdev->features &= ~NETIF_F_TSO; netdev->features &= ~NETIF_F_TSO6; - for (i = 0; i < adapter->num_tx_queues; i++) - netif_start_subqueue(netdev, i); - netif_start_queue(netdev); + netif_tx_start_all_queues(netdev); } return 0; } diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c index e6df9233f5ef..6af8fb5c4b5f 100644 --- a/drivers/net/ixgbe/ixgbe_main.c +++ b/drivers/net/ixgbe/ixgbe_main.c @@ -2013,7 +2013,7 @@ void ixgbe_down(struct ixgbe_adapter *adapter) del_timer_sync(&adapter->watchdog_timer); netif_carrier_off(netdev); - netif_stop_queue(netdev); + netif_tx_stop_all_queues(netdev); if (!pci_channel_offline(adapter->pdev)) ixgbe_reset(adapter); @@ -2359,7 +2359,7 @@ try_msi: out: /* Notify the stack of the (possibly) reduced Tx Queue count. */ - adapter->netdev->egress_subqueue_count = adapter->num_tx_queues; + adapter->netdev->real_num_tx_queues = adapter->num_tx_queues; return err; } @@ -2896,7 +2896,6 @@ static void ixgbe_watchdog(unsigned long data) struct net_device *netdev = adapter->netdev; bool link_up; u32 link_speed = 0; - int i; adapter->hw.mac.ops.check_link(&adapter->hw, &(link_speed), &link_up); @@ -2917,9 +2916,7 @@ static void ixgbe_watchdog(unsigned long data) (FLOW_TX ? "TX" : "None")))); netif_carrier_on(netdev); - netif_wake_queue(netdev); - for (i = 0; i < adapter->num_tx_queues; i++) - netif_wake_subqueue(netdev, i); + netif_tx_wake_all_queues(netdev); } else { /* Force detection of hung controller */ adapter->detect_tx_hung = true; @@ -2928,7 +2925,7 @@ static void ixgbe_watchdog(unsigned long data) if (netif_carrier_ok(netdev)) { DPRINTK(LINK, INFO, "NIC Link is Down\n"); netif_carrier_off(netdev); - netif_stop_queue(netdev); + netif_tx_stop_all_queues(netdev); } } @@ -3631,9 +3628,7 @@ static int __devinit ixgbe_probe(struct pci_dev *pdev, ixgbe_start_hw(hw); netif_carrier_off(netdev); - netif_stop_queue(netdev); - for (i = 0; i < adapter->num_tx_queues; i++) - netif_stop_subqueue(netdev, i); + netif_tx_stop_all_queues(netdev); ixgbe_napi_add_all(adapter); diff --git a/drivers/net/s2io.c b/drivers/net/s2io.c index 5f0fcb04afff..9dae40ccf048 100644 --- a/drivers/net/s2io.c +++ b/drivers/net/s2io.c @@ -545,63 +545,53 @@ static struct pci_driver s2io_driver = { /* netqueue manipulation helper functions */ static inline void s2io_stop_all_tx_queue(struct s2io_nic *sp) { - int i; - if (sp->config.multiq) { - for (i = 0; i < sp->config.tx_fifo_num; i++) - netif_stop_subqueue(sp->dev, i); - } else { + if (!sp->config.multiq) { + int i; + for (i = 0; i < sp->config.tx_fifo_num; i++) sp->mac_control.fifos[i].queue_state = FIFO_QUEUE_STOP; - netif_stop_queue(sp->dev); } + netif_tx_stop_all_queues(sp->dev); } static inline void s2io_stop_tx_queue(struct s2io_nic *sp, int fifo_no) { - if (sp->config.multiq) - netif_stop_subqueue(sp->dev, fifo_no); - else { + if (!sp->config.multiq) sp->mac_control.fifos[fifo_no].queue_state = FIFO_QUEUE_STOP; - netif_stop_queue(sp->dev); - } + + netif_tx_stop_all_queues(sp->dev); } static inline void s2io_start_all_tx_queue(struct s2io_nic *sp) { - int i; - if (sp->config.multiq) { - for (i = 0; i < sp->config.tx_fifo_num; i++) - netif_start_subqueue(sp->dev, i); - } else { + if (!sp->config.multiq) { + int i; + for (i = 0; i < sp->config.tx_fifo_num; i++) sp->mac_control.fifos[i].queue_state = FIFO_QUEUE_START; - netif_start_queue(sp->dev); } + netif_tx_start_all_queues(sp->dev); } static inline void s2io_start_tx_queue(struct s2io_nic *sp, int fifo_no) { - if (sp->config.multiq) - netif_start_subqueue(sp->dev, fifo_no); - else { + if (!sp->config.multiq) sp->mac_control.fifos[fifo_no].queue_state = FIFO_QUEUE_START; - netif_start_queue(sp->dev); - } + + netif_tx_start_all_queues(sp->dev); } static inline void s2io_wake_all_tx_queue(struct s2io_nic *sp) { - int i; - if (sp->config.multiq) { - for (i = 0; i < sp->config.tx_fifo_num; i++) - netif_wake_subqueue(sp->dev, i); - } else { + if (!sp->config.multiq) { + int i; + for (i = 0; i < sp->config.tx_fifo_num; i++) sp->mac_control.fifos[i].queue_state = FIFO_QUEUE_START; - netif_wake_queue(sp->dev); } + netif_tx_wake_all_queues(sp->dev); } static inline void s2io_wake_tx_queue( @@ -8691,5 +8681,5 @@ static void s2io_io_resume(struct pci_dev *pdev) } netif_device_attach(netdev); - netif_wake_queue(netdev); + netif_tx_wake_all_queues(netdev); } diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index c02227b9dd7b..b5c1e7df64fc 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -642,7 +642,13 @@ struct net_device struct netdev_queue rx_queue; struct netdev_queue *_tx ____cacheline_aligned_in_smp; + + /* Number of TX queues allocated at alloc_netdev_mq() time */ unsigned int num_tx_queues; + + /* Number of TX queues currently active in device */ + unsigned int real_num_tx_queues; + unsigned long tx_queue_len; /* Max frames per queue allowed */ /* @@ -1000,6 +1006,14 @@ static inline void netif_schedule(struct net_device *dev) netif_schedule_queue(netdev_get_tx_queue(dev, 0)); } +static inline void netif_tx_schedule_all(struct net_device *dev) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) + netif_schedule_queue(netdev_get_tx_queue(dev, i)); +} + /** * netif_start_queue - allow transmit * @dev: network device @@ -1016,6 +1030,16 @@ static inline void netif_start_queue(struct net_device *dev) netif_tx_start_queue(netdev_get_tx_queue(dev, 0)); } +static inline void netif_tx_start_all_queues(struct net_device *dev) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + netif_tx_start_queue(txq); + } +} + /** * netif_wake_queue - restart transmit * @dev: network device @@ -1040,6 +1064,16 @@ static inline void netif_wake_queue(struct net_device *dev) netif_tx_wake_queue(netdev_get_tx_queue(dev, 0)); } +static inline void netif_tx_wake_all_queues(struct net_device *dev) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + netif_tx_wake_queue(txq); + } +} + /** * netif_stop_queue - stop transmitted packets * @dev: network device @@ -1057,6 +1091,16 @@ static inline void netif_stop_queue(struct net_device *dev) netif_tx_stop_queue(netdev_get_tx_queue(dev, 0)); } +static inline void netif_tx_stop_all_queues(struct net_device *dev) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + netif_tx_stop_queue(txq); + } +} + /** * netif_queue_stopped - test if transmit queue is flowblocked * @dev: network device @@ -1100,7 +1144,8 @@ static inline int netif_running(const struct net_device *dev) */ static inline void netif_start_subqueue(struct net_device *dev, u16 queue_index) { - clear_bit(__QUEUE_STATE_XOFF, &dev->egress_subqueue[queue_index].state); + struct netdev_queue *txq = netdev_get_tx_queue(dev, queue_index); + clear_bit(__QUEUE_STATE_XOFF, &txq->state); } /** @@ -1112,11 +1157,12 @@ static inline void netif_start_subqueue(struct net_device *dev, u16 queue_index) */ static inline void netif_stop_subqueue(struct net_device *dev, u16 queue_index) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, queue_index); #ifdef CONFIG_NETPOLL_TRAP if (netpoll_trap()) return; #endif - set_bit(__QUEUE_STATE_XOFF, &dev->egress_subqueue[queue_index].state); + set_bit(__QUEUE_STATE_XOFF, &txq->state); } /** @@ -1129,8 +1175,8 @@ static inline void netif_stop_subqueue(struct net_device *dev, u16 queue_index) static inline int __netif_subqueue_stopped(const struct net_device *dev, u16 queue_index) { - return test_bit(__QUEUE_STATE_XOFF, - &dev->egress_subqueue[queue_index].state); + struct netdev_queue *txq = netdev_get_tx_queue(dev, queue_index); + return test_bit(__QUEUE_STATE_XOFF, &txq->state); } static inline int netif_subqueue_stopped(const struct net_device *dev, @@ -1148,13 +1194,13 @@ static inline int netif_subqueue_stopped(const struct net_device *dev, */ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, queue_index); #ifdef CONFIG_NETPOLL_TRAP if (netpoll_trap()) return; #endif - if (test_and_clear_bit(__QUEUE_STATE_XOFF, - &dev->egress_subqueue[queue_index].state)) - __netif_schedule(netdev_get_tx_queue(dev, 0)); + if (test_and_clear_bit(__QUEUE_STATE_XOFF, &txq->state)) + __netif_schedule(txq); } /** @@ -1198,7 +1244,8 @@ extern int dev_set_mtu(struct net_device *, int); extern int dev_set_mac_address(struct net_device *, struct sockaddr *); extern int dev_hard_start_xmit(struct sk_buff *skb, - struct net_device *dev); + struct net_device *dev, + struct netdev_queue *txq); extern int netdev_budget; @@ -1447,6 +1494,12 @@ static inline void __netif_tx_lock(struct netdev_queue *txq, int cpu) txq->xmit_lock_owner = cpu; } +static inline void __netif_tx_lock_bh(struct netdev_queue *txq) +{ + spin_lock_bh(&txq->_xmit_lock); + txq->xmit_lock_owner = smp_processor_id(); +} + static inline void netif_tx_lock(struct net_device *dev) { int cpu = smp_processor_id(); @@ -1483,6 +1536,12 @@ static inline void __netif_tx_unlock(struct netdev_queue *txq) spin_unlock(&txq->_xmit_lock); } +static inline void __netif_tx_unlock_bh(struct netdev_queue *txq) +{ + txq->xmit_lock_owner = -1; + spin_unlock_bh(&txq->_xmit_lock); +} + static inline void netif_tx_unlock(struct net_device *dev) { unsigned int i; @@ -1514,8 +1573,13 @@ static inline void netif_tx_unlock_bh(struct net_device *dev) static inline void netif_tx_disable(struct net_device *dev) { + unsigned int i; + netif_tx_lock_bh(dev); - netif_stop_queue(dev); + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + netif_tx_stop_queue(txq); + } netif_tx_unlock_bh(dev); } diff --git a/include/net/pkt_sched.h b/include/net/pkt_sched.h index d58c1a5eb845..cb9527815606 100644 --- a/include/net/pkt_sched.h +++ b/include/net/pkt_sched.h @@ -88,9 +88,7 @@ extern void __qdisc_run(struct netdev_queue *txq); static inline void qdisc_run(struct netdev_queue *txq) { - struct net_device *dev = txq->dev; - - if (!netif_queue_stopped(dev) && + if (!netif_tx_queue_stopped(txq) && !test_and_set_bit(__QUEUE_STATE_QDISC_RUNNING, &txq->state)) __qdisc_run(txq); } diff --git a/net/core/dev.c b/net/core/dev.c index 69378f250695..f027a1ac4fbb 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1598,7 +1598,8 @@ static int dev_gso_segment(struct sk_buff *skb) return 0; } -int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev) +int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev, + struct netdev_queue *txq) { if (likely(!skb->next)) { if (!list_empty(&ptype_all)) @@ -1627,9 +1628,7 @@ gso: skb->next = nskb; return rc; } - if (unlikely((netif_queue_stopped(dev) || - netif_subqueue_stopped(dev, skb)) && - skb->next)) + if (unlikely(netif_tx_queue_stopped(txq) && skb->next)) return NETDEV_TX_BUSY; } while (skb->next); @@ -1669,7 +1668,10 @@ out_kfree_skb: static struct netdev_queue *dev_pick_tx(struct net_device *dev, struct sk_buff *skb) { - return netdev_get_tx_queue(dev, 0); + u16 queue_index = 0; + + skb_set_queue_mapping(skb, queue_index); + return netdev_get_tx_queue(dev, queue_index); } int dev_queue_xmit(struct sk_buff *skb) @@ -1737,8 +1739,6 @@ gso: spin_lock(&txq->lock); q = txq->qdisc; if (q->enqueue) { - /* reset queue_mapping to zero */ - skb_set_queue_mapping(skb, 0); rc = q->enqueue(skb, q); qdisc_run(txq); spin_unlock(&txq->lock); @@ -1768,10 +1768,9 @@ gso: HARD_TX_LOCK(dev, txq, cpu); - if (!netif_queue_stopped(dev) && - !netif_subqueue_stopped(dev, skb)) { + if (!netif_tx_queue_stopped(txq)) { rc = 0; - if (!dev_hard_start_xmit(skb, dev)) { + if (!dev_hard_start_xmit(skb, dev, txq)) { HARD_TX_UNLOCK(dev, txq); goto out; } @@ -4160,8 +4159,7 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, BUG_ON(strlen(name) >= sizeof(dev->name)); - alloc_size = sizeof(struct net_device) + - sizeof(struct net_device_subqueue) * (queue_count - 1); + alloc_size = sizeof(struct net_device); if (sizeof_priv) { /* ensure 32-byte alignment of private area */ alloc_size = (alloc_size + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST; @@ -4191,16 +4189,14 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, dev->_tx = tx; dev->num_tx_queues = queue_count; + dev->real_num_tx_queues = queue_count; if (sizeof_priv) { dev->priv = ((char *)dev + - ((sizeof(struct net_device) + - (sizeof(struct net_device_subqueue) * - (queue_count - 1)) + NETDEV_ALIGN_CONST) + ((sizeof(struct net_device) + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST)); } - dev->egress_subqueue_count = queue_count; dev->gso_max_size = GSO_MAX_SIZE; netdev_init_queues(dev); diff --git a/net/core/netpoll.c b/net/core/netpoll.c index 8fb134da0346..c12720895ecf 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -58,25 +58,27 @@ static void queue_process(struct work_struct *work) while ((skb = skb_dequeue(&npinfo->txq))) { struct net_device *dev = skb->dev; + struct netdev_queue *txq; if (!netif_device_present(dev) || !netif_running(dev)) { __kfree_skb(skb); continue; } + txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb)); + local_irq_save(flags); - netif_tx_lock(dev); - if ((netif_queue_stopped(dev) || - netif_subqueue_stopped(dev, skb)) || - dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) { + __netif_tx_lock(txq, smp_processor_id()); + if (netif_tx_queue_stopped(txq) || + dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) { skb_queue_head(&npinfo->txq, skb); - netif_tx_unlock(dev); + __netif_tx_unlock(txq); local_irq_restore(flags); schedule_delayed_work(&npinfo->tx_work, HZ/10); return; } - netif_tx_unlock(dev); + __netif_tx_unlock(txq); local_irq_restore(flags); } } @@ -278,17 +280,19 @@ static void netpoll_send_skb(struct netpoll *np, struct sk_buff *skb) /* don't get messages out of order, and no recursion */ if (skb_queue_len(&npinfo->txq) == 0 && !netpoll_owner_active(dev)) { + struct netdev_queue *txq; unsigned long flags; + txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb)); + local_irq_save(flags); /* try until next clock tick */ for (tries = jiffies_to_usecs(1)/USEC_PER_POLL; tries > 0; --tries) { - if (netif_tx_trylock(dev)) { - if (!netif_queue_stopped(dev) && - !netif_subqueue_stopped(dev, skb)) + if (__netif_tx_trylock(txq)) { + if (!netif_tx_queue_stopped(txq)) status = dev->hard_start_xmit(skb, dev); - netif_tx_unlock(dev); + __netif_tx_unlock(txq); if (status == NETDEV_TX_OK) break; diff --git a/net/core/pktgen.c b/net/core/pktgen.c index fdf537707e51..906802db4ed4 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -2123,6 +2123,24 @@ static void get_ipsec_sa(struct pktgen_dev *pkt_dev, int flow) } } #endif +static void set_cur_queue_map(struct pktgen_dev *pkt_dev) +{ + if (pkt_dev->queue_map_min < pkt_dev->queue_map_max) { + __u16 t; + if (pkt_dev->flags & F_QUEUE_MAP_RND) { + t = random32() % + (pkt_dev->queue_map_max - + pkt_dev->queue_map_min + 1) + + pkt_dev->queue_map_min; + } else { + t = pkt_dev->cur_queue_map + 1; + if (t > pkt_dev->queue_map_max) + t = pkt_dev->queue_map_min; + } + pkt_dev->cur_queue_map = t; + } +} + /* Increment/randomize headers according to flags and current values * for IP src/dest, UDP src/dst port, MAC-Addr src/dst */ @@ -2325,19 +2343,7 @@ static void mod_cur_headers(struct pktgen_dev *pkt_dev) pkt_dev->cur_pkt_size = t; } - if (pkt_dev->queue_map_min < pkt_dev->queue_map_max) { - __u16 t; - if (pkt_dev->flags & F_QUEUE_MAP_RND) { - t = random32() % - (pkt_dev->queue_map_max - pkt_dev->queue_map_min + 1) - + pkt_dev->queue_map_min; - } else { - t = pkt_dev->cur_queue_map + 1; - if (t > pkt_dev->queue_map_max) - t = pkt_dev->queue_map_min; - } - pkt_dev->cur_queue_map = t; - } + set_cur_queue_map(pkt_dev); pkt_dev->flows[flow].count++; } @@ -2458,7 +2464,7 @@ static struct sk_buff *fill_packet_ipv4(struct net_device *odev, __be16 *vlan_encapsulated_proto = NULL; /* packet type ID field (or len) for VLAN tag */ __be16 *svlan_tci = NULL; /* Encapsulates priority and SVLAN ID */ __be16 *svlan_encapsulated_proto = NULL; /* packet type ID field (or len) for SVLAN tag */ - + u16 queue_map; if (pkt_dev->nr_labels) protocol = htons(ETH_P_MPLS_UC); @@ -2469,6 +2475,7 @@ static struct sk_buff *fill_packet_ipv4(struct net_device *odev, /* Update any of the values, used when we're incrementing various * fields. */ + queue_map = pkt_dev->cur_queue_map; mod_cur_headers(pkt_dev); datalen = (odev->hard_header_len + 16) & ~0xf; @@ -2507,7 +2514,7 @@ static struct sk_buff *fill_packet_ipv4(struct net_device *odev, skb->network_header = skb->tail; skb->transport_header = skb->network_header + sizeof(struct iphdr); skb_put(skb, sizeof(struct iphdr) + sizeof(struct udphdr)); - skb_set_queue_mapping(skb, pkt_dev->cur_queue_map); + skb_set_queue_mapping(skb, queue_map); iph = ip_hdr(skb); udph = udp_hdr(skb); @@ -2797,6 +2804,7 @@ static struct sk_buff *fill_packet_ipv6(struct net_device *odev, __be16 *vlan_encapsulated_proto = NULL; /* packet type ID field (or len) for VLAN tag */ __be16 *svlan_tci = NULL; /* Encapsulates priority and SVLAN ID */ __be16 *svlan_encapsulated_proto = NULL; /* packet type ID field (or len) for SVLAN tag */ + u16 queue_map; if (pkt_dev->nr_labels) protocol = htons(ETH_P_MPLS_UC); @@ -2807,6 +2815,7 @@ static struct sk_buff *fill_packet_ipv6(struct net_device *odev, /* Update any of the values, used when we're incrementing various * fields. */ + queue_map = pkt_dev->cur_queue_map; mod_cur_headers(pkt_dev); skb = alloc_skb(pkt_dev->cur_pkt_size + 64 + 16 + @@ -2844,7 +2853,7 @@ static struct sk_buff *fill_packet_ipv6(struct net_device *odev, skb->network_header = skb->tail; skb->transport_header = skb->network_header + sizeof(struct ipv6hdr); skb_put(skb, sizeof(struct ipv6hdr) + sizeof(struct udphdr)); - skb_set_queue_mapping(skb, pkt_dev->cur_queue_map); + skb_set_queue_mapping(skb, queue_map); iph = ipv6_hdr(skb); udph = udp_hdr(skb); @@ -3263,7 +3272,9 @@ static void pktgen_rem_thread(struct pktgen_thread *t) static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) { struct net_device *odev = NULL; + struct netdev_queue *txq; __u64 idle_start = 0; + u16 queue_map; int ret; odev = pkt_dev->odev; @@ -3285,9 +3296,15 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) } } - if ((netif_queue_stopped(odev) || - (pkt_dev->skb && - netif_subqueue_stopped(odev, pkt_dev->skb))) || + if (!pkt_dev->skb) { + set_cur_queue_map(pkt_dev); + queue_map = pkt_dev->cur_queue_map; + } else { + queue_map = skb_get_queue_mapping(pkt_dev->skb); + } + + txq = netdev_get_tx_queue(odev, queue_map); + if (netif_tx_queue_stopped(txq) || need_resched()) { idle_start = getCurUs(); @@ -3303,8 +3320,7 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) pkt_dev->idle_acc += getCurUs() - idle_start; - if (netif_queue_stopped(odev) || - netif_subqueue_stopped(odev, pkt_dev->skb)) { + if (netif_tx_queue_stopped(txq)) { pkt_dev->next_tx_us = getCurUs(); /* TODO */ pkt_dev->next_tx_ns = 0; goto out; /* Try the next interface */ @@ -3331,9 +3347,12 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) } } - netif_tx_lock_bh(odev); - if (!netif_queue_stopped(odev) && - !netif_subqueue_stopped(odev, pkt_dev->skb)) { + /* fill_packet() might have changed the queue */ + queue_map = skb_get_queue_mapping(pkt_dev->skb); + txq = netdev_get_tx_queue(odev, queue_map); + + __netif_tx_lock_bh(txq); + if (!netif_tx_queue_stopped(txq)) { atomic_inc(&(pkt_dev->skb->users)); retry_now: @@ -3377,7 +3396,7 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) pkt_dev->next_tx_ns = 0; } - netif_tx_unlock_bh(odev); + __netif_tx_unlock_bh(txq); /* If pkt_dev->count is zero, then run forever */ if ((pkt_dev->count != 0) && (pkt_dev->sofar >= pkt_dev->count)) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 4e2b865cbba0..2f575b9017d1 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -166,7 +166,7 @@ static inline int qdisc_restart(struct netdev_queue *txq) HARD_TX_LOCK(dev, txq, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) - ret = dev_hard_start_xmit(skb, dev); + ret = dev_hard_start_xmit(skb, dev, txq); HARD_TX_UNLOCK(dev, txq); spin_lock(&txq->lock); @@ -198,11 +198,10 @@ static inline int qdisc_restart(struct netdev_queue *txq) void __qdisc_run(struct netdev_queue *txq) { - struct net_device *dev = txq->dev; unsigned long start_time = jiffies; while (qdisc_restart(txq)) { - if (netif_queue_stopped(dev)) + if (netif_tx_queue_stopped(txq)) break; /* diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 44a2c3451f4d..ade3372221c7 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -295,8 +295,7 @@ restart: slave_txq = netdev_get_tx_queue(slave, 0); if (slave_txq->qdisc_sleeping != q) continue; - if (netif_queue_stopped(slave) || - __netif_subqueue_stopped(slave, subq) || + if (__netif_subqueue_stopped(slave, subq) || !netif_running(slave)) { busy = 1; continue; @@ -305,8 +304,7 @@ restart: switch (teql_resolve(skb, skb_res, slave)) { case 0: if (netif_tx_trylock(slave)) { - if (!netif_queue_stopped(slave) && - !__netif_subqueue_stopped(slave, subq) && + if (!__netif_subqueue_stopped(slave, subq) && slave->hard_start_xmit(skb, slave) == 0) { netif_tx_unlock(slave); master->slaves = NEXT_SLAVE(q); -- cgit v1.3-8-gc7d7 From eae792b722fef08dcf3aee88266ee7def9710757 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 03:03:33 -0700 Subject: netdev: Add netdev->select_queue() method. Devices or device layers can set this to control the queue selection performed by dev_pick_tx(). This function runs under RCU protection, which allows overriding functions to have some way of synchronizing with things like dynamic ->real_num_tx_queues adjustments. This makes the spinlock prefetch in dev_queue_xmit() a little bit less effective, but that's the price right now for correctness. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 3 +++ net/core/dev.c | 9 ++++++--- 2 files changed, 9 insertions(+), 3 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index fdac1159253e..9464e6452967 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -724,6 +724,9 @@ struct net_device void (*poll_controller)(struct net_device *dev); #endif + u16 (*select_queue)(struct net_device *dev, + struct sk_buff *skb); + #ifdef CONFIG_NET_NS /* Network namespace this network device is inside */ struct net *nd_net; diff --git a/net/core/dev.c b/net/core/dev.c index f027a1ac4fbb..7ca9564d2f44 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1670,6 +1670,9 @@ static struct netdev_queue *dev_pick_tx(struct net_device *dev, { u16 queue_index = 0; + if (dev->select_queue) + queue_index = dev->select_queue(dev, skb); + skb_set_queue_mapping(skb, queue_index); return netdev_get_tx_queue(dev, queue_index); } @@ -1710,14 +1713,14 @@ int dev_queue_xmit(struct sk_buff *skb) } gso: - txq = dev_pick_tx(dev, skb); - spin_lock_prefetch(&txq->lock); - /* Disable soft irqs for various locks below. Also * stops preemption for RCU. */ rcu_read_lock_bh(); + txq = dev_pick_tx(dev, skb); + spin_lock_prefetch(&txq->lock); + /* Updates of qdisc are serialized by queue->lock. * The struct Qdisc which is pointed to by qdisc is now a * rcu structure - it may be accessed without acquiring -- cgit v1.3-8-gc7d7 From 8f0f2223cc08a5ae9a77f40edfe02e8a9f1abd77 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 03:47:03 -0700 Subject: net: Implement simple sw TX hashing. It just xor hashes over IPv4/IPv6 addresses and ports of transport. The only assumption it makes is that skb_network_header() is set correctly. With bug fixes from Eric Dumazet. Signed-off-by: David S. Miller --- net/core/dev.c | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 52 insertions(+) (limited to 'net/core') diff --git a/net/core/dev.c b/net/core/dev.c index 7ca9564d2f44..467bfb325123 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -121,6 +121,9 @@ #include #include #include +#include +#include +#include #include "net-sysfs.h" @@ -1665,6 +1668,53 @@ out_kfree_skb: * --BLG */ +static u16 simple_tx_hash(struct net_device *dev, struct sk_buff *skb) +{ + u32 *addr, *ports, hash, ihl; + u8 ip_proto; + int alen; + + switch (skb->protocol) { + case __constant_htons(ETH_P_IP): + ip_proto = ip_hdr(skb)->protocol; + addr = &ip_hdr(skb)->saddr; + ihl = ip_hdr(skb)->ihl; + alen = 2; + break; + case __constant_htons(ETH_P_IPV6): + ip_proto = ipv6_hdr(skb)->nexthdr; + addr = &ipv6_hdr(skb)->saddr.s6_addr32[0]; + ihl = (40 >> 2); + alen = 8; + break; + default: + return 0; + } + + ports = (u32 *) (skb_network_header(skb) + (ihl * 4)); + + hash = 0; + while (alen--) + hash ^= *addr++; + + switch (ip_proto) { + case IPPROTO_TCP: + case IPPROTO_UDP: + case IPPROTO_DCCP: + case IPPROTO_ESP: + case IPPROTO_AH: + case IPPROTO_SCTP: + case IPPROTO_UDPLITE: + hash ^= *ports; + break; + + default: + break; + } + + return hash % dev->real_num_tx_queues; +} + static struct netdev_queue *dev_pick_tx(struct net_device *dev, struct sk_buff *skb) { @@ -1672,6 +1722,8 @@ static struct netdev_queue *dev_pick_tx(struct net_device *dev, if (dev->select_queue) queue_index = dev->select_queue(dev, skb); + else if (dev->real_num_tx_queues > 1) + queue_index = simple_tx_hash(dev, skb); skb_set_queue_mapping(skb, queue_index); return netdev_get_tx_queue(dev, queue_index); -- cgit v1.3-8-gc7d7 From 37437bb2e1ae8af470dfcd5b4ff454110894ccaf Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 02:15:04 -0700 Subject: pkt_sched: Schedule qdiscs instead of netdev_queue. When we have shared qdiscs, packets come out of the qdiscs for multiple transmit queues. Therefore it doesn't make any sense to schedule the transmit queue when logically we cannot know ahead of time the TX queue of the SKB that the qdisc->dequeue() will give us. Just for sanity I added a BUG check to make sure we never get into a state where the noop_qdisc is scheduled. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 12 ++++----- include/net/pkt_sched.h | 11 +++----- include/net/sch_generic.h | 2 ++ net/core/dev.c | 68 +++++++++++++++++++---------------------------- net/sched/sch_api.c | 3 +-- net/sched/sch_cbq.c | 2 +- net/sched/sch_generic.c | 30 ++++++++++----------- 7 files changed, 55 insertions(+), 73 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 9240a95793be..1e839fa01434 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -275,7 +275,6 @@ enum netdev_state_t { __LINK_STATE_START, __LINK_STATE_PRESENT, - __LINK_STATE_SCHED, __LINK_STATE_NOCARRIER, __LINK_STATE_LINKWATCH_PENDING, __LINK_STATE_DORMANT, @@ -452,7 +451,6 @@ struct netdev_queue { int xmit_lock_owner; struct Qdisc *qdisc_sleeping; struct list_head qdisc_list; - struct netdev_queue *next_sched; } ____cacheline_aligned_in_smp; /* @@ -969,7 +967,7 @@ static inline int unregister_gifconf(unsigned int family) */ struct softnet_data { - struct netdev_queue *output_queue; + struct Qdisc *output_queue; struct sk_buff_head input_pkt_queue; struct list_head poll_list; struct sk_buff *completion_queue; @@ -984,12 +982,12 @@ DECLARE_PER_CPU(struct softnet_data,softnet_data); #define HAVE_NETIF_QUEUE -extern void __netif_schedule(struct netdev_queue *txq); +extern void __netif_schedule(struct Qdisc *q); static inline void netif_schedule_queue(struct netdev_queue *txq) { if (!test_bit(__QUEUE_STATE_XOFF, &txq->state)) - __netif_schedule(txq); + __netif_schedule(txq->qdisc); } static inline void netif_tx_schedule_all(struct net_device *dev) @@ -1042,7 +1040,7 @@ static inline void netif_tx_wake_queue(struct netdev_queue *dev_queue) } #endif if (test_and_clear_bit(__QUEUE_STATE_XOFF, &dev_queue->state)) - __netif_schedule(dev_queue); + __netif_schedule(dev_queue->qdisc); } static inline void netif_wake_queue(struct net_device *dev) @@ -1186,7 +1184,7 @@ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index) return; #endif if (test_and_clear_bit(__QUEUE_STATE_XOFF, &txq->state)) - __netif_schedule(txq); + __netif_schedule(txq->qdisc); } /** diff --git a/include/net/pkt_sched.h b/include/net/pkt_sched.h index 06a442d85186..e4e30052e4e2 100644 --- a/include/net/pkt_sched.h +++ b/include/net/pkt_sched.h @@ -84,15 +84,12 @@ extern struct qdisc_rate_table *qdisc_get_rtab(struct tc_ratespec *r, struct nlattr *tab); extern void qdisc_put_rtab(struct qdisc_rate_table *tab); -extern void __qdisc_run(struct netdev_queue *txq); +extern void __qdisc_run(struct Qdisc *q); -static inline void qdisc_run(struct netdev_queue *txq) +static inline void qdisc_run(struct Qdisc *q) { - struct Qdisc *q = txq->qdisc; - - if (!netif_tx_queue_stopped(txq) && - !test_and_set_bit(__QDISC_STATE_RUNNING, &q->state)) - __qdisc_run(txq); + if (!test_and_set_bit(__QDISC_STATE_RUNNING, &q->state)) + __qdisc_run(q); } extern int tc_classify_compat(struct sk_buff *skb, struct tcf_proto *tp, diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index 92417825d387..3cc4b5cd8c6a 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -26,6 +26,7 @@ struct qdisc_rate_table enum qdisc_state_t { __QDISC_STATE_RUNNING, + __QDISC_STATE_SCHED, }; struct Qdisc @@ -45,6 +46,7 @@ struct Qdisc struct sk_buff *gso_skb; struct sk_buff_head q; struct netdev_queue *dev_queue; + struct Qdisc *next_sched; struct list_head list; struct gnet_stats_basic bstats; diff --git a/net/core/dev.c b/net/core/dev.c index 467bfb325123..0b909b74f698 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1323,18 +1323,18 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) } -void __netif_schedule(struct netdev_queue *txq) +void __netif_schedule(struct Qdisc *q) { - struct net_device *dev = txq->dev; + BUG_ON(q == &noop_qdisc); - if (!test_and_set_bit(__LINK_STATE_SCHED, &dev->state)) { + if (!test_and_set_bit(__QDISC_STATE_SCHED, &q->state)) { struct softnet_data *sd; unsigned long flags; local_irq_save(flags); sd = &__get_cpu_var(softnet_data); - txq->next_sched = sd->output_queue; - sd->output_queue = txq; + q->next_sched = sd->output_queue; + sd->output_queue = q; raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_restore(flags); } @@ -1771,37 +1771,23 @@ gso: rcu_read_lock_bh(); txq = dev_pick_tx(dev, skb); - spin_lock_prefetch(&txq->lock); - - /* Updates of qdisc are serialized by queue->lock. - * The struct Qdisc which is pointed to by qdisc is now a - * rcu structure - it may be accessed without acquiring - * a lock (but the structure may be stale.) The freeing of the - * qdisc will be deferred until it's known that there are no - * more references to it. - * - * If the qdisc has an enqueue function, we still need to - * hold the queue->lock before calling it, since queue->lock - * also serializes access to the device queue. - */ - q = rcu_dereference(txq->qdisc); + #ifdef CONFIG_NET_CLS_ACT skb->tc_verd = SET_TC_AT(skb->tc_verd,AT_EGRESS); #endif if (q->enqueue) { - /* Grab device queue */ - spin_lock(&txq->lock); - q = txq->qdisc; - if (q->enqueue) { - rc = q->enqueue(skb, q); - qdisc_run(txq); - spin_unlock(&txq->lock); - - rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; - goto out; - } - spin_unlock(&txq->lock); + spinlock_t *root_lock = qdisc_root_lock(q); + + spin_lock(root_lock); + + rc = q->enqueue(skb, q); + qdisc_run(q); + + spin_unlock(root_lock); + + rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; + goto out; } /* The device has no queue. Common case for software devices: @@ -1974,7 +1960,7 @@ static void net_tx_action(struct softirq_action *h) } if (sd->output_queue) { - struct netdev_queue *head; + struct Qdisc *head; local_irq_disable(); head = sd->output_queue; @@ -1982,18 +1968,20 @@ static void net_tx_action(struct softirq_action *h) local_irq_enable(); while (head) { - struct netdev_queue *txq = head; - struct net_device *dev = txq->dev; + struct Qdisc *q = head; + spinlock_t *root_lock; + head = head->next_sched; smp_mb__before_clear_bit(); - clear_bit(__LINK_STATE_SCHED, &dev->state); + clear_bit(__QDISC_STATE_SCHED, &q->state); - if (spin_trylock(&txq->lock)) { - qdisc_run(txq); - spin_unlock(&txq->lock); + root_lock = qdisc_root_lock(q); + if (spin_trylock(root_lock)) { + qdisc_run(q); + spin_unlock(root_lock); } else { - netif_schedule_queue(txq); + __netif_schedule(q); } } } @@ -4459,7 +4447,7 @@ static int dev_cpu_callback(struct notifier_block *nfb, void *ocpu) { struct sk_buff **list_skb; - struct netdev_queue **list_net; + struct Qdisc **list_net; struct sk_buff *skb; unsigned int cpu, oldcpu = (unsigned long)ocpu; struct softnet_data *sd, *oldsd; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 19c244a00839..8e8c5becc348 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -294,11 +294,10 @@ static enum hrtimer_restart qdisc_watchdog(struct hrtimer *timer) { struct qdisc_watchdog *wd = container_of(timer, struct qdisc_watchdog, timer); - struct netdev_queue *txq = wd->qdisc->dev_queue; wd->qdisc->flags &= ~TCQ_F_THROTTLED; smp_wmb(); - netif_schedule_queue(txq); + __netif_schedule(wd->qdisc); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 37ae653db683..a3953bbe2d79 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -650,7 +650,7 @@ static enum hrtimer_restart cbq_undelay(struct hrtimer *timer) } sch->flags &= ~TCQ_F_THROTTLED; - netif_schedule_queue(sch->dev_queue); + __netif_schedule(sch); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 739a8711ab30..dd5c4e70abe4 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -72,16 +72,14 @@ static inline int qdisc_qlen(struct Qdisc *q) return q->q.qlen; } -static inline int dev_requeue_skb(struct sk_buff *skb, - struct netdev_queue *dev_queue, - struct Qdisc *q) +static inline int dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q) { if (unlikely(skb->next)) q->gso_skb = skb; else q->ops->requeue(skb, q); - netif_schedule_queue(dev_queue); + __netif_schedule(q); return 0; } @@ -121,7 +119,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * some time. */ __get_cpu_var(netdev_rx_stat).cpu_collision++; - ret = dev_requeue_skb(skb, dev_queue, q); + ret = dev_requeue_skb(skb, q); } return ret; @@ -146,9 +144,9 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * >0 - queue is not empty. * */ -static inline int qdisc_restart(struct netdev_queue *txq, - struct Qdisc *q) +static inline int qdisc_restart(struct Qdisc *q) { + struct netdev_queue *txq; int ret = NETDEV_TX_BUSY; struct net_device *dev; spinlock_t *root_lock; @@ -163,7 +161,8 @@ static inline int qdisc_restart(struct netdev_queue *txq, /* And release qdisc */ spin_unlock(root_lock); - dev = txq->dev; + dev = qdisc_dev(q); + txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb)); HARD_TX_LOCK(dev, txq, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) @@ -189,29 +188,28 @@ static inline int qdisc_restart(struct netdev_queue *txq, printk(KERN_WARNING "BUG %s code %d qlen %d\n", dev->name, ret, q->q.qlen); - ret = dev_requeue_skb(skb, txq, q); + ret = dev_requeue_skb(skb, q); break; } + if (ret && netif_tx_queue_stopped(txq)) + ret = 0; + return ret; } -void __qdisc_run(struct netdev_queue *txq) +void __qdisc_run(struct Qdisc *q) { unsigned long start_time = jiffies; - struct Qdisc *q = txq->qdisc; - - while (qdisc_restart(txq, q)) { - if (netif_tx_queue_stopped(txq)) - break; + while (qdisc_restart(q)) { /* * Postpone processing if * 1. another process needs the CPU; * 2. we've been doing it for too long. */ if (need_resched() || jiffies != start_time) { - netif_schedule_queue(txq); + __netif_schedule(q); break; } } -- cgit v1.3-8-gc7d7 From ead81cc5fc6d996db6afb20f211241612610a07a Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 00:50:32 -0700 Subject: netdevice: Move qdisc_list back into net_device proper. And give it it's own lock. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 3 ++- net/core/dev.c | 2 ++ net/sched/sch_api.c | 31 +++++++------------------------ net/sched/sch_generic.c | 9 +++++++-- 4 files changed, 18 insertions(+), 27 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 1e839fa01434..3170bcef734b 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -450,7 +450,6 @@ struct netdev_queue { spinlock_t _xmit_lock; int xmit_lock_owner; struct Qdisc *qdisc_sleeping; - struct list_head qdisc_list; } ____cacheline_aligned_in_smp; /* @@ -638,6 +637,8 @@ struct net_device unsigned int real_num_tx_queues; unsigned long tx_queue_len; /* Max frames per queue allowed */ + spinlock_t qdisc_list_lock; + struct list_head qdisc_list; /* * One part is mostly used on xmit path (device) diff --git a/net/core/dev.c b/net/core/dev.c index 0b909b74f698..6741e344ac59 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3886,6 +3886,8 @@ int register_netdevice(struct net_device *dev) net = dev_net(dev); spin_lock_init(&dev->addr_list_lock); + spin_lock_init(&dev->qdisc_list_lock); + INIT_LIST_HEAD(&dev->qdisc_list); netdev_init_queue_locks(dev); dev->iflink = -1; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 8e8c5becc348..6958fe7c9a77 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -183,30 +183,17 @@ EXPORT_SYMBOL(unregister_qdisc); (root qdisc, all its children, children of children etc.) */ -static struct Qdisc *__qdisc_lookup(struct netdev_queue *dev_queue, u32 handle) +struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) { struct Qdisc *q; - list_for_each_entry(q, &dev_queue->qdisc_list, list) { + list_for_each_entry(q, &dev->qdisc_list, list) { if (q->handle == handle) return q; } return NULL; } -struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) -{ - unsigned int i; - - for (i = 0; i < dev->num_tx_queues; i++) { - struct netdev_queue *txq = netdev_get_tx_queue(dev, i); - struct Qdisc *q = __qdisc_lookup(txq, handle); - if (q) - return q; - } - return NULL; -} - static struct Qdisc *qdisc_leaf(struct Qdisc *p, u32 classid) { unsigned long cl; @@ -645,9 +632,9 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, goto err_out3; } } - qdisc_lock_tree(dev); - list_add_tail(&sch->list, &dev_queue->qdisc_list); - qdisc_unlock_tree(dev); + spin_lock_bh(&dev->qdisc_list_lock); + list_add_tail(&sch->list, &dev->qdisc_list); + spin_unlock_bh(&dev->qdisc_list_lock); return sch; } @@ -1032,14 +1019,12 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) read_lock(&dev_base_lock); idx = 0; for_each_netdev(&init_net, dev) { - struct netdev_queue *dev_queue; if (idx < s_idx) goto cont; if (idx > s_idx) s_q_idx = 0; q_idx = 0; - dev_queue = netdev_get_tx_queue(dev, 0); - list_for_each_entry(q, &dev_queue->qdisc_list, list) { + list_for_each_entry(q, &dev->qdisc_list, list) { if (q_idx < s_q_idx) { q_idx++; continue; @@ -1269,7 +1254,6 @@ static int qdisc_class_dump(struct Qdisc *q, unsigned long cl, struct qdisc_walk static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); - struct netdev_queue *dev_queue; int t; int s_t; struct net_device *dev; @@ -1288,8 +1272,7 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - dev_queue = netdev_get_tx_queue(dev, 0); - list_for_each_entry(q, &dev_queue->qdisc_list, list) { + list_for_each_entry(q, &dev->qdisc_list, list) { if (t < s_t || !q->ops->cl_ops || (tcm->tcm_parent && TC_H_MAJ(tcm->tcm_parent) != q->handle)) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index efa418a1b34e..8cdf0b4a6a5a 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -563,11 +563,15 @@ static void __qdisc_destroy(struct rcu_head *head) void qdisc_destroy(struct Qdisc *qdisc) { + struct net_device *dev = qdisc_dev(qdisc); + if (qdisc->flags & TCQ_F_BUILTIN || !atomic_dec_and_test(&qdisc->refcnt)) return; + spin_lock_bh(&dev->qdisc_list_lock); list_del(&qdisc->list); + spin_unlock_bh(&dev->qdisc_list_lock); call_rcu(&qdisc->q_rcu, __qdisc_destroy); } @@ -599,7 +603,9 @@ static void attach_one_default_qdisc(struct net_device *dev, printk(KERN_INFO "%s: activation failed\n", dev->name); return; } - list_add_tail(&qdisc->list, &dev_queue->qdisc_list); + spin_lock_bh(&dev->qdisc_list_lock); + list_add_tail(&qdisc->list, &dev->qdisc_list); + spin_unlock_bh(&dev->qdisc_list_lock); } else { qdisc = &noqueue_qdisc; } @@ -738,7 +744,6 @@ static void dev_init_scheduler_queue(struct net_device *dev, dev_queue->qdisc = qdisc; dev_queue->qdisc_sleeping = qdisc; - INIT_LIST_HEAD(&dev_queue->qdisc_list); } void dev_init_scheduler(struct net_device *dev) -- cgit v1.3-8-gc7d7 From 83874000929ed63aef30b44083a9f713135ff040 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 00:53:03 -0700 Subject: pkt_sched: Kill netdev_queue lock. We can simply use the qdisc->q.lock for all of the qdisc tree synchronization. Signed-off-by: David S. Miller --- drivers/net/ifb.c | 20 -------------------- include/linux/netdevice.h | 1 - include/net/sch_generic.h | 7 ++++++- net/core/dev.c | 9 +++++---- net/mac80211/wme.c | 19 ++++++++++++------- net/sched/sch_generic.c | 32 +++++++++++++++----------------- net/sched/sch_teql.c | 7 +++++-- 7 files changed, 43 insertions(+), 52 deletions(-) (limited to 'net/core') diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c index 897b05e79ed0..0960e69b2da4 100644 --- a/drivers/net/ifb.c +++ b/drivers/net/ifb.c @@ -35,7 +35,6 @@ #include #include #include -#include #define TX_TIMEOUT (2*HZ) @@ -228,22 +227,6 @@ static struct rtnl_link_ops ifb_link_ops __read_mostly = { module_param(numifbs, int, 0); MODULE_PARM_DESC(numifbs, "Number of ifb devices"); -/* - * dev_ifb's TX queue lock is usually taken after dev->rx_queue.lock, - * reversely to e.g. qdisc_lock_tree(). It should be safe until - * ifb doesn't take dev's TX queue lock with dev_ifb->rx_queue.lock. - * But lockdep should know that ifb has different locks from dev. - */ -static struct lock_class_key ifb_tx_queue_lock_key; -static struct lock_class_key ifb_rx_queue_lock_key; - -static void set_tx_lockdep_key(struct net_device *dev, - struct netdev_queue *txq, - void *_unused) -{ - lockdep_set_class(&txq->lock, &ifb_tx_queue_lock_key); -} - static int __init ifb_init_one(int index) { struct net_device *dev_ifb; @@ -264,9 +247,6 @@ static int __init ifb_init_one(int index) if (err < 0) goto err; - netdev_for_each_tx_queue(dev_ifb, set_tx_lockdep_key, NULL); - lockdep_set_class(&dev_ifb->rx_queue.lock, &ifb_rx_queue_lock_key); - return 0; err: diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 3170bcef734b..9c5a68850114 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -443,7 +443,6 @@ enum netdev_queue_state_t }; struct netdev_queue { - spinlock_t lock; struct net_device *dev; struct Qdisc *qdisc; unsigned long state; diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index 1eef8d0c9990..2902a42564f0 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -163,6 +163,11 @@ struct tcf_proto struct tcf_proto_ops *ops; }; +static inline spinlock_t *qdisc_lock(struct Qdisc *qdisc) +{ + return &qdisc->q.lock; +} + static inline struct Qdisc *qdisc_root(struct Qdisc *qdisc) { return qdisc->dev_queue->qdisc; @@ -172,7 +177,7 @@ static inline spinlock_t *qdisc_root_lock(struct Qdisc *qdisc) { struct Qdisc *root = qdisc_root(qdisc); - return &root->dev_queue->lock; + return qdisc_lock(root); } static inline struct net_device *qdisc_dev(struct Qdisc *qdisc) diff --git a/net/core/dev.c b/net/core/dev.c index 6741e344ac59..32a13772c1cb 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2080,10 +2080,12 @@ static int ing_filter(struct sk_buff *skb) rxq = &dev->rx_queue; - spin_lock(&rxq->lock); - if ((q = rxq->qdisc) != NULL) + q = rxq->qdisc; + if (q) { + spin_lock(qdisc_lock(q)); result = q->enqueue(skb, q); - spin_unlock(&rxq->lock); + spin_unlock(qdisc_lock(q)); + } return result; } @@ -4173,7 +4175,6 @@ static void netdev_init_one_queue(struct net_device *dev, struct netdev_queue *queue, void *_unused) { - spin_lock_init(&queue->lock); queue->dev = dev; } diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index b21cfec4b6ce..6e8099e77043 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -237,12 +237,14 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, ieee80211_requeue(local, agg_queue); } else { struct netdev_queue *txq; + spinlock_t *root_lock; txq = netdev_get_tx_queue(local->mdev, agg_queue); + root_lock = qdisc_root_lock(txq->qdisc); - spin_lock_bh(&txq->lock); + spin_lock_bh(root_lock); qdisc_reset(txq->qdisc); - spin_unlock_bh(&txq->lock); + spin_unlock_bh(root_lock); } } @@ -250,6 +252,7 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) { struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, queue); struct sk_buff_head list; + spinlock_t *root_lock; struct Qdisc *qdisc; u32 len; @@ -261,14 +264,15 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) skb_queue_head_init(&list); - spin_lock(&txq->lock); + root_lock = qdisc_root_lock(qdisc); + spin_lock(root_lock); for (len = qdisc->q.qlen; len > 0; len--) { struct sk_buff *skb = qdisc->dequeue(qdisc); if (skb) __skb_queue_tail(&list, skb); } - spin_unlock(&txq->lock); + spin_unlock(root_lock); for (len = list.qlen; len > 0; len--) { struct sk_buff *skb = __skb_dequeue(&list); @@ -280,12 +284,13 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) txq = netdev_get_tx_queue(local->mdev, new_queue); - spin_lock(&txq->lock); qdisc = rcu_dereference(txq->qdisc); - qdisc->enqueue(skb, qdisc); + root_lock = qdisc_root_lock(qdisc); - spin_unlock(&txq->lock); + spin_lock(root_lock); + qdisc->enqueue(skb, qdisc); + spin_unlock(root_lock); } out_unlock: diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 3d53e92ad9c8..8fc580b3e173 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -96,15 +96,15 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, } /* - * NOTE: Called under queue->lock with locally disabled BH. + * NOTE: Called under qdisc_lock(q) with locally disabled BH. * * __QDISC_STATE_RUNNING guarantees only one CPU can process - * this qdisc at a time. queue->lock serializes queue accesses for - * this queue AND txq->qdisc pointer itself. + * this qdisc at a time. qdisc_lock(q) serializes queue accesses for + * this queue. * * netif_tx_lock serializes accesses to device driver. * - * queue->lock and netif_tx_lock are mutually exclusive, + * qdisc_lock(q) and netif_tx_lock are mutually exclusive, * if one is grabbed, another must be free. * * Note, that this procedure can be called by a watchdog timer @@ -317,7 +317,6 @@ struct Qdisc_ops noop_qdisc_ops __read_mostly = { }; static struct netdev_queue noop_netdev_queue = { - .lock = __SPIN_LOCK_UNLOCKED(noop_netdev_queue.lock), .qdisc = &noop_qdisc, }; @@ -327,6 +326,7 @@ struct Qdisc noop_qdisc = { .flags = TCQ_F_BUILTIN, .ops = &noop_qdisc_ops, .list = LIST_HEAD_INIT(noop_qdisc.list), + .q.lock = __SPIN_LOCK_UNLOCKED(noop_qdisc.q.lock), .dev_queue = &noop_netdev_queue, }; EXPORT_SYMBOL(noop_qdisc); @@ -498,7 +498,7 @@ errout: } EXPORT_SYMBOL(qdisc_create_dflt); -/* Under queue->lock and BH! */ +/* Under qdisc_root_lock(qdisc) and BH! */ void qdisc_reset(struct Qdisc *qdisc) { @@ -526,10 +526,12 @@ static void __qdisc_destroy(struct rcu_head *head) module_put(ops->owner); dev_put(qdisc_dev(qdisc)); + kfree_skb(qdisc->gso_skb); + kfree((char *) qdisc - qdisc->padded); } -/* Under queue->lock and BH! */ +/* Under qdisc_root_lock(qdisc) and BH! */ void qdisc_destroy(struct Qdisc *qdisc) { @@ -586,13 +588,12 @@ static void transition_one_qdisc(struct net_device *dev, struct netdev_queue *dev_queue, void *_need_watchdog) { + struct Qdisc *new_qdisc = dev_queue->qdisc_sleeping; int *need_watchdog_p = _need_watchdog; - spin_lock_bh(&dev_queue->lock); - rcu_assign_pointer(dev_queue->qdisc, dev_queue->qdisc_sleeping); - if (dev_queue->qdisc != &noqueue_qdisc) + rcu_assign_pointer(dev_queue->qdisc, new_qdisc); + if (new_qdisc != &noqueue_qdisc) *need_watchdog_p = 1; - spin_unlock_bh(&dev_queue->lock); } void dev_activate(struct net_device *dev) @@ -629,19 +630,16 @@ static void dev_deactivate_queue(struct net_device *dev, struct sk_buff *skb = NULL; struct Qdisc *qdisc; - spin_lock_bh(&dev_queue->lock); - qdisc = dev_queue->qdisc; if (qdisc) { + spin_lock_bh(qdisc_lock(qdisc)); + dev_queue->qdisc = qdisc_default; qdisc_reset(qdisc); - skb = qdisc->gso_skb; - qdisc->gso_skb = NULL; + spin_unlock_bh(qdisc_lock(qdisc)); } - spin_unlock_bh(&dev_queue->lock); - kfree_skb(skb); } diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index ade3372221c7..8b0ff345f9da 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -156,12 +156,15 @@ teql_destroy(struct Qdisc* sch) master->slaves = NEXT_SLAVE(q); if (q == master->slaves) { struct netdev_queue *txq; + spinlock_t *root_lock; txq = netdev_get_tx_queue(master->dev, 0); master->slaves = NULL; - spin_lock_bh(&txq->lock); + + root_lock = qdisc_root_lock(txq->qdisc); + spin_lock_bh(root_lock); qdisc_reset(txq->qdisc); - spin_unlock_bh(&txq->lock); + spin_unlock_bh(root_lock); } } skb_queue_purge(&dat->q); -- cgit v1.3-8-gc7d7 From 3072367300aa8c779e3a14ee8e89de079e90f3ad Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 18 Jul 2008 22:50:15 -0700 Subject: pkt_sched: Manage qdisc list inside of root qdisc. Idea is from Patrick McHardy. Instead of managing the list of qdiscs on the device level, manage it in the root qdisc of a netdev_queue. This solves all kinds of visibility issues during qdisc destruction. The way to iterate over all qdiscs of a netdev_queue is to visit the netdev_queue->qdisc, and then traverse it's list. The only special case is to ignore builting qdiscs at the root when dumping or doing a qdisc_lookup(). That was not needed previously because builtin qdiscs were not added to the device's qdisc_list. Signed-off-by: David S. Miller --- include/linux/netdevice.h | 2 - net/core/dev.c | 2 - net/sched/sch_api.c | 175 ++++++++++++++++++++++++++++++++++------------ net/sched/sch_generic.c | 10 +-- 4 files changed, 133 insertions(+), 56 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 9c5a68850114..812bcd8b4363 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -636,8 +636,6 @@ struct net_device unsigned int real_num_tx_queues; unsigned long tx_queue_len; /* Max frames per queue allowed */ - spinlock_t qdisc_list_lock; - struct list_head qdisc_list; /* * One part is mostly used on xmit path (device) diff --git a/net/core/dev.c b/net/core/dev.c index e54acde839da..065b9817e209 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3888,8 +3888,6 @@ int register_netdevice(struct net_device *dev) net = dev_net(dev); spin_lock_init(&dev->addr_list_lock); - spin_lock_init(&dev->qdisc_list_lock); - INIT_LIST_HEAD(&dev->qdisc_list); netdev_init_queue_locks(dev); dev->iflink = -1; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index b3ef8307204e..fb43731c9860 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -185,11 +185,20 @@ EXPORT_SYMBOL(unregister_qdisc); struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) { - struct Qdisc *q; + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + struct Qdisc *q, *txq_root = txq->qdisc; - list_for_each_entry(q, &dev->qdisc_list, list) { - if (q->handle == handle) - return q; + if (!(txq_root->flags & TCQ_F_BUILTIN) && + txq_root->handle == handle) + return txq_root; + + list_for_each_entry(q, &txq_root->list, list) { + if (q->handle == handle) + return q; + } } return NULL; } @@ -676,9 +685,8 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, goto err_out3; } } - spin_lock_bh(&dev->qdisc_list_lock); - list_add_tail(&sch->list, &dev->qdisc_list); - spin_unlock_bh(&dev->qdisc_list_lock); + if (parent) + list_add_tail(&sch->list, &dev_queue->qdisc->list); return sch; } @@ -1037,13 +1045,57 @@ err_out: return -EINVAL; } +static bool tc_qdisc_dump_ignore(struct Qdisc *q) +{ + return (q->flags & TCQ_F_BUILTIN) ? true : false; +} + +static int tc_dump_qdisc_root(struct Qdisc *root, struct sk_buff *skb, + struct netlink_callback *cb, + int *q_idx_p, int s_q_idx) +{ + int ret = 0, q_idx = *q_idx_p; + struct Qdisc *q; + + if (!root) + return 0; + + q = root; + if (q_idx < s_q_idx) { + q_idx++; + } else { + if (!tc_qdisc_dump_ignore(q) && + tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid, + cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) + goto done; + q_idx++; + } + list_for_each_entry(q, &root->list, list) { + if (q_idx < s_q_idx) { + q_idx++; + continue; + } + if (!tc_qdisc_dump_ignore(q) && + tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid, + cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) + goto done; + q_idx++; + } + +out: + *q_idx_p = q_idx; + return ret; +done: + ret = -1; + goto out; +} + static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); int idx, q_idx; int s_idx, s_q_idx; struct net_device *dev; - struct Qdisc *q; if (net != &init_net) return 0; @@ -1053,21 +1105,22 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) read_lock(&dev_base_lock); idx = 0; for_each_netdev(&init_net, dev) { + struct netdev_queue *dev_queue; + if (idx < s_idx) goto cont; if (idx > s_idx) s_q_idx = 0; q_idx = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { - if (q_idx < s_q_idx) { - q_idx++; - continue; - } - if (tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid, - cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) - goto done; - q_idx++; - } + + dev_queue = netdev_get_tx_queue(dev, 0); + if (tc_dump_qdisc_root(dev_queue->qdisc, skb, cb, &q_idx, s_q_idx) < 0) + goto done; + + dev_queue = &dev->rx_queue; + if (tc_dump_qdisc_root(dev_queue->qdisc, skb, cb, &q_idx, s_q_idx) < 0) + goto done; + cont: idx++; } @@ -1285,15 +1338,62 @@ static int qdisc_class_dump(struct Qdisc *q, unsigned long cl, struct qdisc_walk a->cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTCLASS); } +static int tc_dump_tclass_qdisc(struct Qdisc *q, struct sk_buff *skb, + struct tcmsg *tcm, struct netlink_callback *cb, + int *t_p, int s_t) +{ + struct qdisc_dump_args arg; + + if (tc_qdisc_dump_ignore(q) || + *t_p < s_t || !q->ops->cl_ops || + (tcm->tcm_parent && + TC_H_MAJ(tcm->tcm_parent) != q->handle)) { + (*t_p)++; + return 0; + } + if (*t_p > s_t) + memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0])); + arg.w.fn = qdisc_class_dump; + arg.skb = skb; + arg.cb = cb; + arg.w.stop = 0; + arg.w.skip = cb->args[1]; + arg.w.count = 0; + q->ops->cl_ops->walk(q, &arg.w); + cb->args[1] = arg.w.count; + if (arg.w.stop) + return -1; + (*t_p)++; + return 0; +} + +static int tc_dump_tclass_root(struct Qdisc *root, struct sk_buff *skb, + struct tcmsg *tcm, struct netlink_callback *cb, + int *t_p, int s_t) +{ + struct Qdisc *q; + + if (!root) + return 0; + + if (tc_dump_tclass_qdisc(root, skb, tcm, cb, t_p, s_t) < 0) + return -1; + + list_for_each_entry(q, &root->list, list) { + if (tc_dump_tclass_qdisc(q, skb, tcm, cb, t_p, s_t) < 0) + return -1; + } + + return 0; +} + static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) { + struct tcmsg *tcm = (struct tcmsg*)NLMSG_DATA(cb->nlh); struct net *net = sock_net(skb->sk); - int t; - int s_t; + struct netdev_queue *dev_queue; struct net_device *dev; - struct Qdisc *q; - struct tcmsg *tcm = (struct tcmsg*)NLMSG_DATA(cb->nlh); - struct qdisc_dump_args arg; + int t, s_t; if (net != &init_net) return 0; @@ -1306,28 +1406,15 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { - if (t < s_t || !q->ops->cl_ops || - (tcm->tcm_parent && - TC_H_MAJ(tcm->tcm_parent) != q->handle)) { - t++; - continue; - } - if (t > s_t) - memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0])); - arg.w.fn = qdisc_class_dump; - arg.skb = skb; - arg.cb = cb; - arg.w.stop = 0; - arg.w.skip = cb->args[1]; - arg.w.count = 0; - q->ops->cl_ops->walk(q, &arg.w); - cb->args[1] = arg.w.count; - if (arg.w.stop) - break; - t++; - } + dev_queue = netdev_get_tx_queue(dev, 0); + if (tc_dump_tclass_root(dev_queue->qdisc, skb, tcm, cb, &t, s_t) < 0) + goto done; + + dev_queue = &dev->rx_queue; + if (tc_dump_tclass_root(dev_queue->qdisc, skb, tcm, cb, &t, s_t) < 0) + goto done; +done: cb->args[0] = t; dev_put(dev); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index e244c462e6bd..14cc443d0490 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -480,15 +480,12 @@ static void __qdisc_destroy(struct rcu_head *head) void qdisc_destroy(struct Qdisc *qdisc) { - struct net_device *dev = qdisc_dev(qdisc); - if (qdisc->flags & TCQ_F_BUILTIN || !atomic_dec_and_test(&qdisc->refcnt)) return; - spin_lock_bh(&dev->qdisc_list_lock); - list_del(&qdisc->list); - spin_unlock_bh(&dev->qdisc_list_lock); + if (qdisc->parent) + list_del(&qdisc->list); call_rcu(&qdisc->q_rcu, __qdisc_destroy); } @@ -520,9 +517,6 @@ static void attach_one_default_qdisc(struct net_device *dev, printk(KERN_INFO "%s: activation failed\n", dev->name); return; } - spin_lock_bh(&dev->qdisc_list_lock); - list_add_tail(&qdisc->list, &dev->qdisc_list); - spin_unlock_bh(&dev->qdisc_list_lock); } else { qdisc = &noqueue_qdisc; } -- cgit v1.3-8-gc7d7 From 721499e8931c5732202481ae24f2dfbf9910f129 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 19 Jul 2008 22:34:43 -0700 Subject: netns: Use net_eq() to compare net-namespaces for optimization. Without CONFIG_NET_NS, namespace is always &init_net. Compiler will be able to omit namespace comparisons with this patch. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- net/appletalk/aarp.c | 4 ++-- net/appletalk/ddp.c | 6 +++--- net/atm/clip.c | 2 +- net/atm/mpc.c | 2 +- net/ax25/af_ax25.c | 2 +- net/ax25/ax25_in.c | 2 +- net/bridge/br_notify.c | 2 +- net/bridge/br_stp_bpdu.c | 2 +- net/can/af_can.c | 4 ++-- net/can/bcm.c | 2 +- net/can/raw.c | 2 +- net/core/pktgen.c | 2 +- net/decnet/af_decnet.c | 2 +- net/decnet/dn_route.c | 2 +- net/econet/af_econet.c | 4 ++-- net/ipv4/igmp.c | 26 +++++++++++++------------- net/ipv4/ipconfig.c | 4 ++-- net/ipv4/ipmr.c | 2 +- net/ipv4/netfilter/ip_queue.c | 2 +- net/ipv4/netfilter/ipt_MASQUERADE.c | 2 +- net/ipv6/ip6mr.c | 2 +- net/ipv6/netfilter/ip6_queue.c | 2 +- net/ipv6/proc.c | 2 +- net/ipx/af_ipx.c | 4 ++-- net/irda/irlap_frame.c | 2 +- net/llc/llc_input.c | 2 +- net/netfilter/nf_sockopt.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_unlabeled.c | 2 +- net/netrom/af_netrom.c | 2 +- net/packet/af_packet.c | 2 +- net/rose/af_rose.c | 2 +- net/sctp/protocol.c | 2 +- net/tipc/eth_media.c | 4 ++-- net/wireless/wext.c | 2 +- net/x25/af_x25.c | 2 +- net/x25/x25_dev.c | 2 +- net/xfrm/xfrm_policy.c | 2 +- 38 files changed, 58 insertions(+), 58 deletions(-) (limited to 'net/core') diff --git a/net/appletalk/aarp.c b/net/appletalk/aarp.c index 25aa37ce9430..b25c1e909d14 100644 --- a/net/appletalk/aarp.c +++ b/net/appletalk/aarp.c @@ -333,7 +333,7 @@ static int aarp_device_event(struct notifier_block *this, unsigned long event, struct net_device *dev = ptr; int ct; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_DOWN) { @@ -716,7 +716,7 @@ static int aarp_rcv(struct sk_buff *skb, struct net_device *dev, struct atalk_addr sa, *ma, da; struct atalk_iface *ifa; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto out0; /* We only do Ethernet SNAP AARP. */ diff --git a/net/appletalk/ddp.c b/net/appletalk/ddp.c index 44cd42f7786b..07b5b82c5eab 100644 --- a/net/appletalk/ddp.c +++ b/net/appletalk/ddp.c @@ -648,7 +648,7 @@ static int ddp_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_DOWN) @@ -1405,7 +1405,7 @@ static int atalk_rcv(struct sk_buff *skb, struct net_device *dev, int origlen; __u16 len_hops; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto freeit; /* Don't mangle buffer if shared */ @@ -1493,7 +1493,7 @@ freeit: static int ltalk_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev) { - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto freeit; /* Expand any short form frames */ diff --git a/net/atm/clip.c b/net/atm/clip.c index 6f8223ebf551..5b5b96344ce6 100644 --- a/net/atm/clip.c +++ b/net/atm/clip.c @@ -612,7 +612,7 @@ static int clip_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = arg; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_UNREGISTER) { diff --git a/net/atm/mpc.c b/net/atm/mpc.c index 9db332e7a6c0..4fccaa1e07be 100644 --- a/net/atm/mpc.c +++ b/net/atm/mpc.c @@ -964,7 +964,7 @@ static int mpoa_event_listener(struct notifier_block *mpoa_notifier, unsigned lo dev = (struct net_device *)dev_ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->name == NULL || strncmp(dev->name, "lec", 3)) diff --git a/net/ax25/af_ax25.c b/net/ax25/af_ax25.c index 97eaa23ad9ea..01c83e2a4c19 100644 --- a/net/ax25/af_ax25.c +++ b/net/ax25/af_ax25.c @@ -116,7 +116,7 @@ static int ax25_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Reject non AX.25 devices */ diff --git a/net/ax25/ax25_in.c b/net/ax25/ax25_in.c index 33790a8efbc8..4a5ba978a804 100644 --- a/net/ax25/ax25_in.c +++ b/net/ax25/ax25_in.c @@ -451,7 +451,7 @@ int ax25_kiss_rcv(struct sk_buff *skb, struct net_device *dev, skb->sk = NULL; /* Initially we don't know who it's for */ skb->destructor = NULL; /* Who initializes this, dammit?! */ - if (dev_net(dev) != &init_net) { + if (!net_eq(dev_net(dev), &init_net)) { kfree_skb(skb); return 0; } diff --git a/net/bridge/br_notify.c b/net/bridge/br_notify.c index 88d8ec7b3142..76340bdd052e 100644 --- a/net/bridge/br_notify.c +++ b/net/bridge/br_notify.c @@ -35,7 +35,7 @@ static int br_device_event(struct notifier_block *unused, unsigned long event, v struct net_bridge_port *p = dev->br_port; struct net_bridge *br; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* not a port of a bridge */ diff --git a/net/bridge/br_stp_bpdu.c b/net/bridge/br_stp_bpdu.c index 996476174517..8b200f96f722 100644 --- a/net/bridge/br_stp_bpdu.c +++ b/net/bridge/br_stp_bpdu.c @@ -140,7 +140,7 @@ void br_stp_rcv(const struct stp_proto *proto, struct sk_buff *skb, struct net_bridge *br; const unsigned char *buf; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto err; if (!p) diff --git a/net/can/af_can.c b/net/can/af_can.c index 484bbf6dd032..8035fbf526ae 100644 --- a/net/can/af_can.c +++ b/net/can/af_can.c @@ -615,7 +615,7 @@ static int can_rcv(struct sk_buff *skb, struct net_device *dev, struct can_frame *cf = (struct can_frame *)skb->data; int matches; - if (dev->type != ARPHRD_CAN || dev_net(dev) != &init_net) { + if (dev->type != ARPHRD_CAN || !net_eq(dev_net(dev), &init_net)) { kfree_skb(skb); return 0; } @@ -728,7 +728,7 @@ static int can_notifier(struct notifier_block *nb, unsigned long msg, struct net_device *dev = (struct net_device *)data; struct dev_rcv_lists *d; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type != ARPHRD_CAN) diff --git a/net/can/bcm.c b/net/can/bcm.c index 72c2ce904f83..d0dd382001e2 100644 --- a/net/can/bcm.c +++ b/net/can/bcm.c @@ -1303,7 +1303,7 @@ static int bcm_notifier(struct notifier_block *nb, unsigned long msg, struct bcm_op *op; int notify_enodev = 0; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type != ARPHRD_CAN) diff --git a/net/can/raw.c b/net/can/raw.c index 3e46ee36a1aa..6e0663faaf9f 100644 --- a/net/can/raw.c +++ b/net/can/raw.c @@ -210,7 +210,7 @@ static int raw_notifier(struct notifier_block *nb, struct raw_sock *ro = container_of(nb, struct raw_sock, notifier); struct sock *sk = &ro->sk; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type != ARPHRD_CAN) diff --git a/net/core/pktgen.c b/net/core/pktgen.c index 906802db4ed4..c7d484f7e1c4 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -1875,7 +1875,7 @@ static int pktgen_device_event(struct notifier_block *unused, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* It is OK that we do not hold the group lock right now, diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c index 61b7df577ddd..3c23ab33dbc0 100644 --- a/net/decnet/af_decnet.c +++ b/net/decnet/af_decnet.c @@ -2089,7 +2089,7 @@ static int dn_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; switch(event) { diff --git a/net/decnet/dn_route.c b/net/decnet/dn_route.c index f50e88bf2661..821bd1cdec04 100644 --- a/net/decnet/dn_route.c +++ b/net/decnet/dn_route.c @@ -580,7 +580,7 @@ int dn_route_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type struct dn_dev *dn = (struct dn_dev *)dev->dn_ptr; unsigned char padlen = 0; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto dump_it; if (dn == NULL) diff --git a/net/econet/af_econet.c b/net/econet/af_econet.c index d35127bb84e1..8789d2bb1b06 100644 --- a/net/econet/af_econet.c +++ b/net/econet/af_econet.c @@ -1062,7 +1062,7 @@ static int econet_rcv(struct sk_buff *skb, struct net_device *dev, struct packet struct sock *sk; struct ec_device *edev = dev->ec_ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; if (skb->pkt_type == PACKET_OTHERHOST) @@ -1119,7 +1119,7 @@ static int econet_notifier(struct notifier_block *this, unsigned long msg, void struct net_device *dev = (struct net_device *)data; struct ec_device *edev; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; switch (msg) { diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index 68e84a933e90..6203ece53606 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -1196,7 +1196,7 @@ void ip_mc_inc_group(struct in_device *in_dev, __be32 addr) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; for (im=in_dev->mc_list; im; im=im->next) { @@ -1278,7 +1278,7 @@ void ip_mc_dec_group(struct in_device *in_dev, __be32 addr) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; for (ip=&in_dev->mc_list; (i=*ip)!=NULL; ip=&i->next) { @@ -1308,7 +1308,7 @@ void ip_mc_down(struct in_device *in_dev) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; for (i=in_dev->mc_list; i; i=i->next) @@ -1331,7 +1331,7 @@ void ip_mc_init_dev(struct in_device *in_dev) { ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; in_dev->mc_tomb = NULL; @@ -1357,7 +1357,7 @@ void ip_mc_up(struct in_device *in_dev) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; ip_mc_inc_group(in_dev, IGMP_ALL_HOSTS); @@ -1376,7 +1376,7 @@ void ip_mc_destroy_dev(struct in_device *in_dev) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; /* Deactivate timers */ @@ -1760,7 +1760,7 @@ int ip_mc_join_group(struct sock *sk , struct ip_mreqn *imr) if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -1831,7 +1831,7 @@ int ip_mc_leave_group(struct sock *sk, struct ip_mreqn *imr) u32 ifindex; int ret = -EADDRNOTAVAIL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -1879,7 +1879,7 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2015,7 +2015,7 @@ int ip_mc_msfilter(struct sock *sk, struct ip_msfilter *msf, int ifindex) msf->imsf_fmode != MCAST_EXCLUDE) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2098,7 +2098,7 @@ int ip_mc_msfget(struct sock *sk, struct ip_msfilter *msf, if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2163,7 +2163,7 @@ int ip_mc_gsfget(struct sock *sk, struct group_filter *gsf, if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2250,7 +2250,7 @@ void ip_mc_drop_socket(struct sock *sk) if (inet->mc_list == NULL) return; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return; rtnl_lock(); diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c index b88aa9afa42e..42065fff46c4 100644 --- a/net/ipv4/ipconfig.c +++ b/net/ipv4/ipconfig.c @@ -432,7 +432,7 @@ ic_rarp_recv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt unsigned char *sha, *tha; /* s for "source", t for "target" */ struct ic_device *d; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) @@ -852,7 +852,7 @@ static int __init ic_bootp_recv(struct sk_buff *skb, struct net_device *dev, str struct ic_device *d; int len, ext_len; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; /* Perform verifications before taking the lock. */ diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 033c712c3a5d..c519b8d30eee 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -1124,7 +1124,7 @@ static int ipmr_device_event(struct notifier_block *this, unsigned long event, v struct vif_device *v; int ct; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_UNREGISTER) diff --git a/net/ipv4/netfilter/ip_queue.c b/net/ipv4/netfilter/ip_queue.c index aa33a4a7a715..432ce9d1c11c 100644 --- a/net/ipv4/netfilter/ip_queue.c +++ b/net/ipv4/netfilter/ip_queue.c @@ -477,7 +477,7 @@ ipq_rcv_dev_event(struct notifier_block *this, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Drop any packets associated with the downed device */ diff --git a/net/ipv4/netfilter/ipt_MASQUERADE.c b/net/ipv4/netfilter/ipt_MASQUERADE.c index 84c26dd27d81..0841aefaa503 100644 --- a/net/ipv4/netfilter/ipt_MASQUERADE.c +++ b/net/ipv4/netfilter/ipt_MASQUERADE.c @@ -120,7 +120,7 @@ static int masq_device_event(struct notifier_block *this, { const struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_DOWN) { diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index 0b41aa2675f5..095bc453ff4c 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -935,7 +935,7 @@ static int ip6mr_device_event(struct notifier_block *this, struct mif_device *v; int ct; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_UNREGISTER) diff --git a/net/ipv6/netfilter/ip6_queue.c b/net/ipv6/netfilter/ip6_queue.c index 1b8815f6153d..5859c046cbc4 100644 --- a/net/ipv6/netfilter/ip6_queue.c +++ b/net/ipv6/netfilter/ip6_queue.c @@ -480,7 +480,7 @@ ipq_rcv_dev_event(struct notifier_block *this, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Drop any packets associated with the downed device */ diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c index 70940b3654a1..f82f6074cf85 100644 --- a/net/ipv6/proc.c +++ b/net/ipv6/proc.c @@ -214,7 +214,7 @@ int snmp6_register_dev(struct inet6_dev *idev) if (!idev || !idev->dev) return -EINVAL; - if (dev_net(idev->dev) != &init_net) + if (!net_eq(dev_net(idev->dev), &init_net)) return 0; if (!proc_net_devsnmp6) diff --git a/net/ipx/af_ipx.c b/net/ipx/af_ipx.c index 81ae8735f5e3..b6e70f92e7fb 100644 --- a/net/ipx/af_ipx.c +++ b/net/ipx/af_ipx.c @@ -335,7 +335,7 @@ static int ipxitf_device_event(struct notifier_block *notifier, struct net_device *dev = ptr; struct ipx_interface *i, *tmp; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_DOWN && event != NETDEV_UP) @@ -1636,7 +1636,7 @@ static int ipx_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_ty u16 ipx_pktsize; int rc = 0; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; /* Not ours */ diff --git a/net/irda/irlap_frame.c b/net/irda/irlap_frame.c index 90894534f3cc..f17b65af9c9b 100644 --- a/net/irda/irlap_frame.c +++ b/net/irda/irlap_frame.c @@ -1326,7 +1326,7 @@ int irlap_driver_rcv(struct sk_buff *skb, struct net_device *dev, int command; __u8 control; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto out; /* FIXME: should we get our own field? */ diff --git a/net/llc/llc_input.c b/net/llc/llc_input.c index 1c45f172991e..57ad974e4d94 100644 --- a/net/llc/llc_input.c +++ b/net/llc/llc_input.c @@ -150,7 +150,7 @@ int llc_rcv(struct sk_buff *skb, struct net_device *dev, int (*rcv)(struct sk_buff *, struct net_device *, struct packet_type *, struct net_device *); - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; /* diff --git a/net/netfilter/nf_sockopt.c b/net/netfilter/nf_sockopt.c index 69d699f95f4c..01489681fa96 100644 --- a/net/netfilter/nf_sockopt.c +++ b/net/netfilter/nf_sockopt.c @@ -65,7 +65,7 @@ static struct nf_sockopt_ops *nf_sockopt_find(struct sock *sk, int pf, { struct nf_sockopt_ops *ops; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return ERR_PTR(-ENOPROTOOPT); if (mutex_lock_interruptible(&nf_sockopt_mutex) != 0) diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c index 04e9c965f8ca..8c860112ce05 100644 --- a/net/netfilter/nfnetlink_queue.c +++ b/net/netfilter/nfnetlink_queue.c @@ -555,7 +555,7 @@ nfqnl_rcv_dev_event(struct notifier_block *this, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Drop any packets associated with the downed device */ diff --git a/net/netlabel/netlabel_unlabeled.c b/net/netlabel/netlabel_unlabeled.c index 56f80872924e..921c118ead89 100644 --- a/net/netlabel/netlabel_unlabeled.c +++ b/net/netlabel/netlabel_unlabeled.c @@ -954,7 +954,7 @@ static int netlbl_unlhsh_netdev_handler(struct notifier_block *this, struct net_device *dev = ptr; struct netlbl_unlhsh_iface *iface = NULL; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* XXX - should this be a check for NETDEV_DOWN or _UNREGISTER? */ diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index d41be0d66eb0..fccc250f95f5 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -118,7 +118,7 @@ static int nr_device_event(struct notifier_block *this, unsigned long event, voi { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_DOWN) diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index de73bcb5235a..d56cae112dc8 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -1677,7 +1677,7 @@ static int packet_ioctl(struct socket *sock, unsigned int cmd, case SIOCGIFDSTADDR: case SIOCSIFDSTADDR: case SIOCSIFFLAGS: - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -ENOIOCTLCMD; return inet_dgram_ops.ioctl(sock, cmd, arg); #endif diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index f3a691f34909..dbc963b4f5fb 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -209,7 +209,7 @@ static int rose_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_DOWN) diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index ed9acffe8105..a6e0818bcff5 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -668,7 +668,7 @@ static int sctp_inetaddr_event(struct notifier_block *this, unsigned long ev, struct sctp_sockaddr_entry *temp; int found = 0; - if (dev_net(ifa->ifa_dev->dev) != &init_net) + if (!net_eq(dev_net(ifa->ifa_dev->dev), &init_net)) return NOTIFY_DONE; switch (ev) { diff --git a/net/tipc/eth_media.c b/net/tipc/eth_media.c index bc72fbc4f8b8..fe43ef7dd7e3 100644 --- a/net/tipc/eth_media.c +++ b/net/tipc/eth_media.c @@ -101,7 +101,7 @@ static int recv_msg(struct sk_buff *buf, struct net_device *dev, struct eth_bearer *eb_ptr = (struct eth_bearer *)pt->af_packet_priv; u32 size; - if (dev_net(dev) != &init_net) { + if (!net_eq(dev_net(dev), &init_net)) { kfree_skb(buf); return 0; } @@ -198,7 +198,7 @@ static int recv_notification(struct notifier_block *nb, unsigned long evt, struct eth_bearer *eb_ptr = ð_bearers[0]; struct eth_bearer *stop = ð_bearers[MAX_ETH_BEARERS]; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; while ((eb_ptr->dev != dev)) { diff --git a/net/wireless/wext.c b/net/wireless/wext.c index 273a84359998..df5b3886c36b 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -1299,7 +1299,7 @@ static void rtmsg_iwinfo(struct net_device *dev, char *event, int event_len) struct sk_buff *skb; int err; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return; skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_ATOMIC); diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c index 7b1c6ef04553..9fc5b023d111 100644 --- a/net/x25/af_x25.c +++ b/net/x25/af_x25.c @@ -191,7 +191,7 @@ static int x25_device_event(struct notifier_block *this, unsigned long event, struct net_device *dev = ptr; struct x25_neigh *nb; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type == ARPHRD_X25 diff --git a/net/x25/x25_dev.c b/net/x25/x25_dev.c index 3ff206c0ae94..3e1efe534645 100644 --- a/net/x25/x25_dev.c +++ b/net/x25/x25_dev.c @@ -95,7 +95,7 @@ int x25_lapb_receive_frame(struct sk_buff *skb, struct net_device *dev, struct sk_buff *nskb; struct x25_neigh *nb; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; nskb = skb_copy(skb, GFP_ATOMIC); diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index cae9fd815543..841b32a2e680 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -2360,7 +2360,7 @@ static int xfrm_dev_event(struct notifier_block *this, unsigned long event, void { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; switch (event) { -- cgit v1.3-8-gc7d7 From 230b183921ecbaa5fedc0d35ad6ba7bb64b6e06a Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 19 Jul 2008 22:35:47 -0700 Subject: net: Use standard structures for generic socket address structures. Use sockaddr_storage{} for generic socket address storage and ensures proper alignment. Use sockaddr{} for pointers to omit several casts. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- include/linux/socket.h | 6 ++-- include/net/compat.h | 2 +- net/compat.c | 2 +- net/core/iovec.c | 2 +- net/socket.c | 82 ++++++++++++++++++++++++++++---------------------- 5 files changed, 52 insertions(+), 42 deletions(-) (limited to 'net/core') diff --git a/include/linux/socket.h b/include/linux/socket.h index bd2b30a74e76..950af631e7fb 100644 --- a/include/linux/socket.h +++ b/include/linux/socket.h @@ -306,10 +306,10 @@ extern int csum_partial_copy_fromiovecend(unsigned char *kdata, int offset, unsigned int len, __wsum *csump); -extern int verify_iovec(struct msghdr *m, struct iovec *iov, char *address, int mode); +extern int verify_iovec(struct msghdr *m, struct iovec *iov, struct sockaddr *address, int mode); extern int memcpy_toiovec(struct iovec *v, unsigned char *kdata, int len); -extern int move_addr_to_user(void *kaddr, int klen, void __user *uaddr, int __user *ulen); -extern int move_addr_to_kernel(void __user *uaddr, int ulen, void *kaddr); +extern int move_addr_to_user(struct sockaddr *kaddr, int klen, void __user *uaddr, int __user *ulen); +extern int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr *kaddr); extern int put_cmsg(struct msghdr*, int level, int type, int len, void *data); #endif diff --git a/include/net/compat.h b/include/net/compat.h index 164cb682e220..5bbf8bf9efea 100644 --- a/include/net/compat.h +++ b/include/net/compat.h @@ -32,7 +32,7 @@ extern int compat_sock_get_timestampns(struct sock *, struct timespec __user *); #endif /* defined(CONFIG_COMPAT) */ extern int get_compat_msghdr(struct msghdr *, struct compat_msghdr __user *); -extern int verify_compat_iovec(struct msghdr *, struct iovec *, char *, int); +extern int verify_compat_iovec(struct msghdr *, struct iovec *, struct sockaddr *, int); extern asmlinkage long compat_sys_sendmsg(int,struct compat_msghdr __user *,unsigned); extern asmlinkage long compat_sys_recvmsg(int,struct compat_msghdr __user *,unsigned); extern asmlinkage long compat_sys_getsockopt(int, int, int, char __user *, int __user *); diff --git a/net/compat.c b/net/compat.c index c823f6f290cb..6e1b03b51933 100644 --- a/net/compat.c +++ b/net/compat.c @@ -75,7 +75,7 @@ int get_compat_msghdr(struct msghdr *kmsg, struct compat_msghdr __user *umsg) /* I've named the args so it is easy to tell whose space the pointers are in. */ int verify_compat_iovec(struct msghdr *kern_msg, struct iovec *kern_iov, - char *kern_address, int mode) + struct sockaddr *kern_address, int mode) { int tot_len; diff --git a/net/core/iovec.c b/net/core/iovec.c index 755c37fdaee7..4c9c0121c9da 100644 --- a/net/core/iovec.c +++ b/net/core/iovec.c @@ -36,7 +36,7 @@ * in any case. */ -int verify_iovec(struct msghdr *m, struct iovec *iov, char *address, int mode) +int verify_iovec(struct msghdr *m, struct iovec *iov, struct sockaddr *address, int mode) { int size, err, ct; diff --git a/net/socket.c b/net/socket.c index 81fe82513046..1ba57d888981 100644 --- a/net/socket.c +++ b/net/socket.c @@ -180,9 +180,9 @@ static DEFINE_PER_CPU(int, sockets_in_use) = 0; * invalid addresses -EFAULT is returned. On a success 0 is returned. */ -int move_addr_to_kernel(void __user *uaddr, int ulen, void *kaddr) +int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr *kaddr) { - if (ulen < 0 || ulen > MAX_SOCK_ADDR) + if (ulen < 0 || ulen > sizeof(struct sockaddr_storage)) return -EINVAL; if (ulen == 0) return 0; @@ -208,7 +208,7 @@ int move_addr_to_kernel(void __user *uaddr, int ulen, void *kaddr) * specified. Zero is returned for a success. */ -int move_addr_to_user(void *kaddr, int klen, void __user *uaddr, +int move_addr_to_user(struct sockaddr *kaddr, int klen, void __user *uaddr, int __user *ulen) { int err; @@ -219,7 +219,7 @@ int move_addr_to_user(void *kaddr, int klen, void __user *uaddr, return err; if (len > klen) len = klen; - if (len < 0 || len > MAX_SOCK_ADDR) + if (len < 0 || len > sizeof(struct sockaddr_storage)) return -EINVAL; if (len) { if (audit_sockaddr(klen, kaddr)) @@ -1342,20 +1342,20 @@ out_fd: asmlinkage long sys_bind(int fd, struct sockaddr __user *umyaddr, int addrlen) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); if (sock) { - err = move_addr_to_kernel(umyaddr, addrlen, address); + err = move_addr_to_kernel(umyaddr, addrlen, (struct sockaddr *)&address); if (err >= 0) { err = security_socket_bind(sock, - (struct sockaddr *)address, + (struct sockaddr *)&address, addrlen); if (!err) err = sock->ops->bind(sock, (struct sockaddr *) - address, addrlen); + &address, addrlen); } fput_light(sock->file, fput_needed); } @@ -1407,7 +1407,7 @@ asmlinkage long sys_accept(int fd, struct sockaddr __user *upeer_sockaddr, struct socket *sock, *newsock; struct file *newfile; int err, len, newfd, fput_needed; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; sock = sockfd_lookup_light(fd, &err, &fput_needed); if (!sock) @@ -1446,13 +1446,13 @@ asmlinkage long sys_accept(int fd, struct sockaddr __user *upeer_sockaddr, goto out_fd; if (upeer_sockaddr) { - if (newsock->ops->getname(newsock, (struct sockaddr *)address, + if (newsock->ops->getname(newsock, (struct sockaddr *)&address, &len, 2) < 0) { err = -ECONNABORTED; goto out_fd; } - err = move_addr_to_user(address, len, upeer_sockaddr, - upeer_addrlen); + err = move_addr_to_user((struct sockaddr *)&address, + len, upeer_sockaddr, upeer_addrlen); if (err < 0) goto out_fd; } @@ -1495,22 +1495,22 @@ asmlinkage long sys_connect(int fd, struct sockaddr __user *uservaddr, int addrlen) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); if (!sock) goto out; - err = move_addr_to_kernel(uservaddr, addrlen, address); + err = move_addr_to_kernel(uservaddr, addrlen, (struct sockaddr *)&address); if (err < 0) goto out_put; err = - security_socket_connect(sock, (struct sockaddr *)address, addrlen); + security_socket_connect(sock, (struct sockaddr *)&address, addrlen); if (err) goto out_put; - err = sock->ops->connect(sock, (struct sockaddr *)address, addrlen, + err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen, sock->file->f_flags); out_put: fput_light(sock->file, fput_needed); @@ -1527,7 +1527,7 @@ asmlinkage long sys_getsockname(int fd, struct sockaddr __user *usockaddr, int __user *usockaddr_len) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int len, err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); @@ -1538,10 +1538,10 @@ asmlinkage long sys_getsockname(int fd, struct sockaddr __user *usockaddr, if (err) goto out_put; - err = sock->ops->getname(sock, (struct sockaddr *)address, &len, 0); + err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0); if (err) goto out_put; - err = move_addr_to_user(address, len, usockaddr, usockaddr_len); + err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr, usockaddr_len); out_put: fput_light(sock->file, fput_needed); @@ -1558,7 +1558,7 @@ asmlinkage long sys_getpeername(int fd, struct sockaddr __user *usockaddr, int __user *usockaddr_len) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int len, err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); @@ -1570,10 +1570,10 @@ asmlinkage long sys_getpeername(int fd, struct sockaddr __user *usockaddr, } err = - sock->ops->getname(sock, (struct sockaddr *)address, &len, + sock->ops->getname(sock, (struct sockaddr *)&address, &len, 1); if (!err) - err = move_addr_to_user(address, len, usockaddr, + err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr, usockaddr_len); fput_light(sock->file, fput_needed); } @@ -1591,7 +1591,7 @@ asmlinkage long sys_sendto(int fd, void __user *buff, size_t len, int addr_len) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err; struct msghdr msg; struct iovec iov; @@ -1610,10 +1610,10 @@ asmlinkage long sys_sendto(int fd, void __user *buff, size_t len, msg.msg_controllen = 0; msg.msg_namelen = 0; if (addr) { - err = move_addr_to_kernel(addr, addr_len, address); + err = move_addr_to_kernel(addr, addr_len, (struct sockaddr *)&address); if (err < 0) goto out_put; - msg.msg_name = address; + msg.msg_name = (struct sockaddr *)&address; msg.msg_namelen = addr_len; } if (sock->file->f_flags & O_NONBLOCK) @@ -1649,7 +1649,7 @@ asmlinkage long sys_recvfrom(int fd, void __user *ubuf, size_t size, struct socket *sock; struct iovec iov; struct msghdr msg; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err, err2; int fput_needed; @@ -1663,14 +1663,15 @@ asmlinkage long sys_recvfrom(int fd, void __user *ubuf, size_t size, msg.msg_iov = &iov; iov.iov_len = size; iov.iov_base = ubuf; - msg.msg_name = address; - msg.msg_namelen = MAX_SOCK_ADDR; + msg.msg_name = (struct sockaddr *)&address; + msg.msg_namelen = sizeof(address); if (sock->file->f_flags & O_NONBLOCK) flags |= MSG_DONTWAIT; err = sock_recvmsg(sock, &msg, size, flags); if (err >= 0 && addr != NULL) { - err2 = move_addr_to_user(address, msg.msg_namelen, addr, addr_len); + err2 = move_addr_to_user((struct sockaddr *)&address, + msg.msg_namelen, addr, addr_len); if (err2 < 0) err = err2; } @@ -1790,7 +1791,7 @@ asmlinkage long sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags) struct compat_msghdr __user *msg_compat = (struct compat_msghdr __user *)msg; struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; struct iovec iovstack[UIO_FASTIOV], *iov = iovstack; unsigned char ctl[sizeof(struct cmsghdr) + 20] __attribute__ ((aligned(sizeof(__kernel_size_t)))); @@ -1828,9 +1829,13 @@ asmlinkage long sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags) /* This will also move the address data into kernel space */ if (MSG_CMSG_COMPAT & flags) { - err = verify_compat_iovec(&msg_sys, iov, address, VERIFY_READ); + err = verify_compat_iovec(&msg_sys, iov, + (struct sockaddr *)&address, + VERIFY_READ); } else - err = verify_iovec(&msg_sys, iov, address, VERIFY_READ); + err = verify_iovec(&msg_sys, iov, + (struct sockaddr *)&address, + VERIFY_READ); if (err < 0) goto out_freeiov; total_len = err; @@ -1901,7 +1906,7 @@ asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, int fput_needed; /* kernel mode address */ - char addr[MAX_SOCK_ADDR]; + struct sockaddr_storage addr; /* user mode address pointers */ struct sockaddr __user *uaddr; @@ -1939,9 +1944,13 @@ asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, uaddr = (__force void __user *)msg_sys.msg_name; uaddr_len = COMPAT_NAMELEN(msg); if (MSG_CMSG_COMPAT & flags) { - err = verify_compat_iovec(&msg_sys, iov, addr, VERIFY_WRITE); + err = verify_compat_iovec(&msg_sys, iov, + (struct sockaddr *)&addr, + VERIFY_WRITE); } else - err = verify_iovec(&msg_sys, iov, addr, VERIFY_WRITE); + err = verify_iovec(&msg_sys, iov, + (struct sockaddr *)&addr, + VERIFY_WRITE); if (err < 0) goto out_freeiov; total_len = err; @@ -1957,7 +1966,8 @@ asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, len = err; if (uaddr != NULL) { - err = move_addr_to_user(addr, msg_sys.msg_namelen, uaddr, + err = move_addr_to_user((struct sockaddr *)&addr, + msg_sys.msg_namelen, uaddr, uaddr_len); if (err < 0) goto out_freeiov; -- cgit v1.3-8-gc7d7 From 5f86173bdf15981ca49d0434f638b68f70a35644 Mon Sep 17 00:00:00 2001 From: Jussi Kivilinna Date: Sun, 20 Jul 2008 00:08:04 -0700 Subject: net_sched: Add qdisc_enqueue wrapper Signed-off-by: Jussi Kivilinna Signed-off-by: David S. Miller --- include/net/sch_generic.h | 10 ++++++++++ net/core/dev.c | 4 ++-- net/mac80211/wme.c | 2 +- net/sched/sch_atm.c | 2 +- net/sched/sch_cbq.c | 5 +++-- net/sched/sch_dsmark.c | 2 +- net/sched/sch_hfsc.c | 2 +- net/sched/sch_htb.c | 3 +-- net/sched/sch_netem.c | 20 ++++++++++++-------- net/sched/sch_prio.c | 3 ++- net/sched/sch_red.c | 2 +- net/sched/sch_tbf.c | 3 ++- 12 files changed, 37 insertions(+), 21 deletions(-) (limited to 'net/core') diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index 8a44386b35cf..f396dff335a3 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -306,6 +306,16 @@ static inline bool qdisc_tx_is_noop(const struct net_device *dev) return true; } +static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch) +{ + return sch->enqueue(skb, sch); +} + +static inline int qdisc_enqueue_root(struct sk_buff *skb, struct Qdisc *sch) +{ + return qdisc_enqueue(skb, sch); +} + static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch, struct sk_buff_head *list) { diff --git a/net/core/dev.c b/net/core/dev.c index 065b9817e209..2eed17bcb2dd 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1781,7 +1781,7 @@ gso: spin_lock(root_lock); - rc = q->enqueue(skb, q); + rc = qdisc_enqueue_root(skb, q); qdisc_run(q); spin_unlock(root_lock); @@ -2083,7 +2083,7 @@ static int ing_filter(struct sk_buff *skb) q = rxq->qdisc; if (q) { spin_lock(qdisc_lock(q)); - result = q->enqueue(skb, q); + result = qdisc_enqueue_root(skb, q); spin_unlock(qdisc_lock(q)); } diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 6e8099e77043..07edda0b8a5c 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -289,7 +289,7 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) root_lock = qdisc_root_lock(qdisc); spin_lock(root_lock); - qdisc->enqueue(skb, qdisc); + qdisc_enqueue_root(skb, qdisc); spin_unlock(root_lock); } diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 0de757e3be4a..68ed35e2a763 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -429,7 +429,7 @@ static int atm_tc_enqueue(struct sk_buff *skb, struct Qdisc *sch) #endif } - ret = flow->q->enqueue(skb, flow->q); + ret = qdisc_enqueue(skb, flow->q); if (ret != 0) { drop: __maybe_unused sch->qstats.drops++; diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index a3953bbe2d79..1afe3eece627 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -387,7 +387,8 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch) #ifdef CONFIG_NET_CLS_ACT cl->q->__parent = sch; #endif - if ((ret = cl->q->enqueue(skb, cl->q)) == NET_XMIT_SUCCESS) { + ret = qdisc_enqueue(skb, cl->q); + if (ret == NET_XMIT_SUCCESS) { sch->q.qlen++; sch->bstats.packets++; sch->bstats.bytes+=len; @@ -671,7 +672,7 @@ static int cbq_reshape_fail(struct sk_buff *skb, struct Qdisc *child) q->rx_class = cl; cl->q->__parent = sch; - if (cl->q->enqueue(skb, cl->q) == 0) { + if (qdisc_enqueue(skb, cl->q) == 0) { sch->q.qlen++; sch->bstats.packets++; sch->bstats.bytes+=len; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index 3aafbd17393a..44d347e831cf 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -252,7 +252,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch) } } - err = p->q->enqueue(skb, p->q); + err = qdisc_enqueue(skb, p->q); if (err != NET_XMIT_SUCCESS) { sch->qstats.drops++; return err; diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 5090708ba384..fd61ed6ee1e7 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1586,7 +1586,7 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) } len = skb->len; - err = cl->qdisc->enqueue(skb, cl->qdisc); + err = qdisc_enqueue(skb, cl->qdisc); if (unlikely(err != NET_XMIT_SUCCESS)) { cl->qstats.drops++; sch->qstats.drops++; diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index ee48457eaa4a..72b5a946178f 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -572,8 +572,7 @@ static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch) kfree_skb(skb); return ret; #endif - } else if (cl->un.leaf.q->enqueue(skb, cl->un.leaf.q) != - NET_XMIT_SUCCESS) { + } else if (qdisc_enqueue(skb, cl->un.leaf.q) != NET_XMIT_SUCCESS) { sch->qstats.drops++; cl->qstats.drops++; return NET_XMIT_DROP; diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index c5ea40c9eb21..13c4821e42b8 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -82,6 +82,12 @@ struct netem_skb_cb { psched_time_t time_to_send; }; +static inline struct netem_skb_cb *netem_skb_cb(struct sk_buff *skb) +{ + BUILD_BUG_ON(sizeof(skb->cb) < sizeof(struct netem_skb_cb)); + return (struct netem_skb_cb *)skb->cb; +} + /* init_crandom - initialize correlated random number generator * Use entropy source for initial seed. */ @@ -184,7 +190,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) u32 dupsave = q->duplicate; /* prevent duplicating a dup... */ q->duplicate = 0; - rootq->enqueue(skb2, rootq); + qdisc_enqueue_root(skb2, rootq); q->duplicate = dupsave; } @@ -205,7 +211,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) skb->data[net_random() % skb_headlen(skb)] ^= 1<<(net_random() % 8); } - cb = (struct netem_skb_cb *)skb->cb; + cb = netem_skb_cb(skb); if (q->gap == 0 /* not doing reordering */ || q->counter < q->gap /* inside last reordering gap */ || q->reorder < get_crandom(&q->reorder_cor)) { @@ -218,7 +224,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) now = psched_get_time(); cb->time_to_send = now + delay; ++q->counter; - ret = q->qdisc->enqueue(skb, q->qdisc); + ret = qdisc_enqueue(skb, q->qdisc); } else { /* * Do re-ordering by putting one out of N packets at the front @@ -277,8 +283,7 @@ static struct sk_buff *netem_dequeue(struct Qdisc *sch) skb = q->qdisc->dequeue(q->qdisc); if (skb) { - const struct netem_skb_cb *cb - = (const struct netem_skb_cb *)skb->cb; + const struct netem_skb_cb *cb = netem_skb_cb(skb); psched_time_t now = psched_get_time(); /* if more time remaining? */ @@ -457,7 +462,7 @@ static int tfifo_enqueue(struct sk_buff *nskb, struct Qdisc *sch) { struct fifo_sched_data *q = qdisc_priv(sch); struct sk_buff_head *list = &sch->q; - psched_time_t tnext = ((struct netem_skb_cb *)nskb->cb)->time_to_send; + psched_time_t tnext = netem_skb_cb(nskb)->time_to_send; struct sk_buff *skb; if (likely(skb_queue_len(list) < q->limit)) { @@ -468,8 +473,7 @@ static int tfifo_enqueue(struct sk_buff *nskb, struct Qdisc *sch) } skb_queue_reverse_walk(list, skb) { - const struct netem_skb_cb *cb - = (const struct netem_skb_cb *)skb->cb; + const struct netem_skb_cb *cb = netem_skb_cb(skb); if (tnext >= cb->time_to_send) break; diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index 536ca474dc69..d29c2f87fc0b 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -81,7 +81,8 @@ prio_enqueue(struct sk_buff *skb, struct Qdisc *sch) } #endif - if ((ret = qdisc->enqueue(skb, qdisc)) == NET_XMIT_SUCCESS) { + ret = qdisc_enqueue(skb, qdisc); + if (ret == NET_XMIT_SUCCESS) { sch->bstats.bytes += skb->len; sch->bstats.packets++; sch->q.qlen++; diff --git a/net/sched/sch_red.c b/net/sched/sch_red.c index 77098acf0adc..b48a391bc129 100644 --- a/net/sched/sch_red.c +++ b/net/sched/sch_red.c @@ -92,7 +92,7 @@ static int red_enqueue(struct sk_buff *skb, struct Qdisc* sch) break; } - ret = child->enqueue(skb, child); + ret = qdisc_enqueue(skb, child); if (likely(ret == NET_XMIT_SUCCESS)) { sch->bstats.bytes += skb->len; sch->bstats.packets++; diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c index 444c227fcb6b..7d705b86dae5 100644 --- a/net/sched/sch_tbf.c +++ b/net/sched/sch_tbf.c @@ -133,7 +133,8 @@ static int tbf_enqueue(struct sk_buff *skb, struct Qdisc* sch) return NET_XMIT_DROP; } - if ((ret = q->qdisc->enqueue(skb, q->qdisc)) != 0) { + ret = qdisc_enqueue(skb, q->qdisc); + if (ret != 0) { sch->qstats.drops++; return ret; } -- cgit v1.3-8-gc7d7 From b6b2fed1f4802b8fcc9d7548a8f785225d38f9a3 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Mon, 21 Jul 2008 09:48:06 -0700 Subject: net: Improve simple_tx_hash(). Based upon feedback from Eric Dumazet and Andi Kleen. Cure several deficiencies in simple_tx_hash() by using jhash + reciprocol multiply. 1) Eliminates expensive modulus operation. 2) Makes hash less attackable by using random seed. 3) Eliminates endianness hash distribution issues. Signed-off-by: David S. Miller --- net/core/dev.c | 34 +++++++++++++++++++++------------- 1 file changed, 21 insertions(+), 13 deletions(-) (limited to 'net/core') diff --git a/net/core/dev.c b/net/core/dev.c index 2eed17bcb2dd..7e2d5274333f 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -124,6 +124,8 @@ #include #include #include +#include +#include #include "net-sysfs.h" @@ -1668,34 +1670,37 @@ out_kfree_skb: * --BLG */ +static u32 simple_tx_hashrnd; +static int simple_tx_hashrnd_initialized = 0; + static u16 simple_tx_hash(struct net_device *dev, struct sk_buff *skb) { - u32 *addr, *ports, hash, ihl; + u32 addr1, addr2, ports; + u32 hash, ihl; u8 ip_proto; - int alen; + + if (unlikely(!simple_tx_hashrnd_initialized)) { + get_random_bytes(&simple_tx_hashrnd, 4); + simple_tx_hashrnd_initialized = 1; + } switch (skb->protocol) { case __constant_htons(ETH_P_IP): ip_proto = ip_hdr(skb)->protocol; - addr = &ip_hdr(skb)->saddr; + addr1 = ip_hdr(skb)->saddr; + addr2 = ip_hdr(skb)->daddr; ihl = ip_hdr(skb)->ihl; - alen = 2; break; case __constant_htons(ETH_P_IPV6): ip_proto = ipv6_hdr(skb)->nexthdr; - addr = &ipv6_hdr(skb)->saddr.s6_addr32[0]; + addr1 = ipv6_hdr(skb)->saddr.s6_addr32[3]; + addr2 = ipv6_hdr(skb)->daddr.s6_addr32[3]; ihl = (40 >> 2); - alen = 8; break; default: return 0; } - ports = (u32 *) (skb_network_header(skb) + (ihl * 4)); - - hash = 0; - while (alen--) - hash ^= *addr++; switch (ip_proto) { case IPPROTO_TCP: @@ -1705,14 +1710,17 @@ static u16 simple_tx_hash(struct net_device *dev, struct sk_buff *skb) case IPPROTO_AH: case IPPROTO_SCTP: case IPPROTO_UDPLITE: - hash ^= *ports; + ports = *((u32 *) (skb_network_header(skb) + (ihl * 4))); break; default: + ports = 0; break; } - return hash % dev->real_num_tx_queues; + hash = jhash_3words(addr1, addr2, ports, simple_tx_hashrnd); + + return (u16) (((u64) hash * dev->real_num_tx_queues) >> 32); } static struct netdev_queue *dev_pick_tx(struct net_device *dev, -- cgit v1.3-8-gc7d7 From 867d79fb9a4d5929ad8335c896fcfe11c3b2ef14 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 21 Jul 2008 09:54:18 -0700 Subject: net: In __netif_schedule() use WARN_ON instead of BUG_ON Signed-off-by: David S. Miller --- net/core/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net/core') diff --git a/net/core/dev.c b/net/core/dev.c index 7e2d5274333f..cbc34c0db376 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1327,7 +1327,8 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) void __netif_schedule(struct Qdisc *q) { - BUG_ON(q == &noop_qdisc); + if (WARN_ON_ONCE(q == &noop_qdisc)) + return; if (!test_and_set_bit(__QDISC_STATE_SCHED, &q->state)) { struct softnet_data *sd; -- cgit v1.3-8-gc7d7 From 7943986ca1138ac99597b1aa4dc893012dcfdc08 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Mon, 21 Jul 2008 13:28:44 -0700 Subject: net: use kcalloc in netdev_queue alloc Minor nit, use size_t for allocation size and kcalloc to allocate an array. Probably makes no actual code difference. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/core') diff --git a/net/core/dev.c b/net/core/dev.c index cbc34c0db376..1698b3998981 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4207,7 +4207,7 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, { struct netdev_queue *tx; struct net_device *dev; - int alloc_size; + size_t alloc_size; void *p; BUG_ON(strlen(name) >= sizeof(dev->name)); @@ -4227,7 +4227,7 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, return NULL; } - tx = kzalloc(sizeof(struct netdev_queue) * queue_count, GFP_KERNEL); + tx = kcalloc(queue_count, sizeof(struct netdev_queue), GFP_KERNEL); if (!tx) { printk(KERN_ERR "alloc_netdev: Unable to allocate " "tx qdiscs.\n"); -- cgit v1.3-8-gc7d7 From 6579e57b31d79d31d9b806e41ba48774e73257dc Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Mon, 21 Jul 2008 13:31:48 -0700 Subject: net: Print the module name as part of the watchdog message As suggested by Dave: This patch adds a function to get the driver name from a struct net_device, and consequently uses this in the watchdog timeout handler to print as part of the message. Signed-off-by: Arjan van de Ven Signed-off-by: David S. Miller --- include/linux/netdevice.h | 2 ++ net/core/dev.c | 20 ++++++++++++++++++++ net/sched/sch_generic.c | 6 +++--- 3 files changed, 25 insertions(+), 3 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 812bcd8b4363..f5ea445f89f0 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1645,6 +1645,8 @@ extern void dev_seq_stop(struct seq_file *seq, void *v); extern int netdev_class_create_file(struct class_attribute *class_attr); extern void netdev_class_remove_file(struct class_attribute *class_attr); +extern char *netdev_drivername(struct net_device *dev, char *buffer, int len); + extern void linkwatch_run_queue(void); extern int netdev_compute_features(unsigned long all, unsigned long one); diff --git a/net/core/dev.c b/net/core/dev.c index 1698b3998981..ad5598d2bb37 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4686,6 +4686,26 @@ err_name: return -ENOMEM; } +char *netdev_drivername(struct net_device *dev, char *buffer, int len) +{ + struct device_driver *driver; + struct device *parent; + + if (len <= 0 || !buffer) + return buffer; + buffer[0] = 0; + + parent = dev->dev.parent; + + if (!parent) + return buffer; + + driver = parent->driver; + if (driver && driver->name) + strlcpy(buffer, driver->name, len); + return buffer; +} + static void __net_exit netdev_exit(struct net *net) { kfree(net->dev_name_head); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index cb625b4d6da5..4ac7e3a8c253 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -212,9 +212,9 @@ static void dev_watchdog(unsigned long arg) if (some_queue_stopped && time_after(jiffies, (dev->trans_start + dev->watchdog_timeo))) { - printk(KERN_INFO "NETDEV WATCHDOG: %s: " - "transmit timed out\n", - dev->name); + char drivername[64]; + printk(KERN_INFO "NETDEV WATCHDOG: %s (%s): transmit timed out\n", + dev->name, netdev_drivername(dev, drivername, 64)); dev->tx_timeout(dev); WARN_ON_ONCE(1); } -- cgit v1.3-8-gc7d7 From d29f749e252bcdbfe7a75a58f0ee92da16f127c0 Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Tue, 22 Jul 2008 14:09:06 -0700 Subject: net: Fix build failure with 'make mandocs'. The function header comments have to go with the functions they are documenting, or things go horribly wrong when we try to process them with the docbook tools. Warning(include/linux/netdevice.h:1006): No description found for parameter 'dev_queue' Warning(include/linux/netdevice.h:1033): No description found for parameter 'dev_queue' Warning(include/linux/netdevice.h:1067): No description found for parameter 'dev_queue' Warning(include/linux/netdevice.h:1093): No description found for parameter 'dev_queue' Warning(include/linux/netdevice.h:1474): No description found for parameter 'txq' Error(net/core/dev.c:1674): cannot understand prototype: 'u32 simple_tx_hashrnd; ' Signed-off-by: Dave Jones Acked-by: Randy Dunlap Signed-off-by: David S. Miller --- include/linux/netdevice.h | 58 +++++++++++++++++++++++------------------------ net/core/dev.c | 51 ++++++++++++++++++++--------------------- 2 files changed, 54 insertions(+), 55 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index f5ea445f89f0..b4d056ceab96 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -996,17 +996,17 @@ static inline void netif_tx_schedule_all(struct net_device *dev) netif_schedule_queue(netdev_get_tx_queue(dev, i)); } +static inline void netif_tx_start_queue(struct netdev_queue *dev_queue) +{ + clear_bit(__QUEUE_STATE_XOFF, &dev_queue->state); +} + /** * netif_start_queue - allow transmit * @dev: network device * * Allow upper layers to call the device hard_start_xmit routine. */ -static inline void netif_tx_start_queue(struct netdev_queue *dev_queue) -{ - clear_bit(__QUEUE_STATE_XOFF, &dev_queue->state); -} - static inline void netif_start_queue(struct net_device *dev) { netif_tx_start_queue(netdev_get_tx_queue(dev, 0)); @@ -1022,13 +1022,6 @@ static inline void netif_tx_start_all_queues(struct net_device *dev) } } -/** - * netif_wake_queue - restart transmit - * @dev: network device - * - * Allow upper layers to call the device hard_start_xmit routine. - * Used for flow control when transmit resources are available. - */ static inline void netif_tx_wake_queue(struct netdev_queue *dev_queue) { #ifdef CONFIG_NETPOLL_TRAP @@ -1041,6 +1034,13 @@ static inline void netif_tx_wake_queue(struct netdev_queue *dev_queue) __netif_schedule(dev_queue->qdisc); } +/** + * netif_wake_queue - restart transmit + * @dev: network device + * + * Allow upper layers to call the device hard_start_xmit routine. + * Used for flow control when transmit resources are available. + */ static inline void netif_wake_queue(struct net_device *dev) { netif_tx_wake_queue(netdev_get_tx_queue(dev, 0)); @@ -1056,6 +1056,11 @@ static inline void netif_tx_wake_all_queues(struct net_device *dev) } } +static inline void netif_tx_stop_queue(struct netdev_queue *dev_queue) +{ + set_bit(__QUEUE_STATE_XOFF, &dev_queue->state); +} + /** * netif_stop_queue - stop transmitted packets * @dev: network device @@ -1063,11 +1068,6 @@ static inline void netif_tx_wake_all_queues(struct net_device *dev) * Stop upper layers calling the device hard_start_xmit routine. * Used for flow control when transmit resources are unavailable. */ -static inline void netif_tx_stop_queue(struct netdev_queue *dev_queue) -{ - set_bit(__QUEUE_STATE_XOFF, &dev_queue->state); -} - static inline void netif_stop_queue(struct net_device *dev) { netif_tx_stop_queue(netdev_get_tx_queue(dev, 0)); @@ -1083,17 +1083,17 @@ static inline void netif_tx_stop_all_queues(struct net_device *dev) } } +static inline int netif_tx_queue_stopped(const struct netdev_queue *dev_queue) +{ + return test_bit(__QUEUE_STATE_XOFF, &dev_queue->state); +} + /** * netif_queue_stopped - test if transmit queue is flowblocked * @dev: network device * * Test if transmit queue on device is currently unable to send. */ -static inline int netif_tx_queue_stopped(const struct netdev_queue *dev_queue) -{ - return test_bit(__QUEUE_STATE_XOFF, &dev_queue->state); -} - static inline int netif_queue_stopped(const struct net_device *dev) { return netif_tx_queue_stopped(netdev_get_tx_queue(dev, 0)); @@ -1463,13 +1463,6 @@ static inline void netif_rx_complete(struct net_device *dev, local_irq_restore(flags); } -/** - * netif_tx_lock - grab network device transmit lock - * @dev: network device - * @cpu: cpu number of lock owner - * - * Get network device transmit lock - */ static inline void __netif_tx_lock(struct netdev_queue *txq, int cpu) { spin_lock(&txq->_xmit_lock); @@ -1482,6 +1475,13 @@ static inline void __netif_tx_lock_bh(struct netdev_queue *txq) txq->xmit_lock_owner = smp_processor_id(); } +/** + * netif_tx_lock - grab network device transmit lock + * @dev: network device + * @cpu: cpu number of lock owner + * + * Get network device transmit lock + */ static inline void netif_tx_lock(struct net_device *dev) { int cpu = smp_processor_id(); diff --git a/net/core/dev.c b/net/core/dev.c index ad5598d2bb37..65eea83613ef 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1645,32 +1645,6 @@ out_kfree_skb: return 0; } -/** - * dev_queue_xmit - transmit a buffer - * @skb: buffer to transmit - * - * Queue a buffer for transmission to a network device. The caller must - * have set the device and priority and built the buffer before calling - * this function. The function can be called from an interrupt. - * - * A negative errno code is returned on a failure. A success does not - * guarantee the frame will be transmitted as it may be dropped due - * to congestion or traffic shaping. - * - * ----------------------------------------------------------------------------------- - * I notice this method can also return errors from the queue disciplines, - * including NET_XMIT_DROP, which is a positive value. So, errors can also - * be positive. - * - * Regardless of the return value, the skb is consumed, so it is currently - * difficult to retry a send to this method. (You can bump the ref count - * before sending to hold a reference for retry if you are careful.) - * - * When calling this method, interrupts MUST be enabled. This is because - * the BH enable code must have IRQs enabled so that it will not deadlock. - * --BLG - */ - static u32 simple_tx_hashrnd; static int simple_tx_hashrnd_initialized = 0; @@ -1738,6 +1712,31 @@ static struct netdev_queue *dev_pick_tx(struct net_device *dev, return netdev_get_tx_queue(dev, queue_index); } +/** + * dev_queue_xmit - transmit a buffer + * @skb: buffer to transmit + * + * Queue a buffer for transmission to a network device. The caller must + * have set the device and priority and built the buffer before calling + * this function. The function can be called from an interrupt. + * + * A negative errno code is returned on a failure. A success does not + * guarantee the frame will be transmitted as it may be dropped due + * to congestion or traffic shaping. + * + * ----------------------------------------------------------------------------------- + * I notice this method can also return errors from the queue disciplines, + * including NET_XMIT_DROP, which is a positive value. So, errors can also + * be positive. + * + * Regardless of the return value, the skb is consumed, so it is currently + * difficult to retry a send to this method. (You can bump the ref count + * before sending to hold a reference for retry if you are careful.) + * + * When calling this method, interrupts MUST be enabled. This is because + * the BH enable code must have IRQs enabled so that it will not deadlock. + * --BLG + */ int dev_queue_xmit(struct sk_buff *skb) { struct net_device *dev = skb->dev; -- cgit v1.3-8-gc7d7 From cf508b1211dbe576778ff445ea1b4b0bcfa5c4ea Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 22 Jul 2008 14:16:42 -0700 Subject: netdev: Handle ->addr_list_lock just like ->_xmit_lock for lockdep. The new address list lock needs to handle the same device layering issues that the _xmit_lock one does. This integrates work done by Patrick McHardy. Signed-off-by: David S. Miller --- drivers/net/bonding/bond_main.c | 3 +++ drivers/net/hamradio/bpqether.c | 2 ++ drivers/net/macvlan.c | 3 +++ drivers/net/wireless/hostap/hostap_hw.c | 3 +++ net/8021q/vlan_dev.c | 4 ++++ net/core/dev.c | 27 +++++++++++++++++++++------ net/netrom/af_netrom.c | 2 ++ net/rose/af_rose.c | 2 ++ 8 files changed, 40 insertions(+), 6 deletions(-) (limited to 'net/core') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 9737c06045d6..a641eeaa2a2f 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -5041,6 +5041,7 @@ static int bond_check_params(struct bond_params *params) } static struct lock_class_key bonding_netdev_xmit_lock_key; +static struct lock_class_key bonding_netdev_addr_lock_key; static void bond_set_lockdep_class_one(struct net_device *dev, struct netdev_queue *txq, @@ -5052,6 +5053,8 @@ static void bond_set_lockdep_class_one(struct net_device *dev, static void bond_set_lockdep_class(struct net_device *dev) { + lockdep_set_class(&dev->addr_list_lock, + &bonding_netdev_addr_lock_key); netdev_for_each_tx_queue(dev, bond_set_lockdep_class_one, NULL); } diff --git a/drivers/net/hamradio/bpqether.c b/drivers/net/hamradio/bpqether.c index b6500b2aacf2..58f4b1d7bf1f 100644 --- a/drivers/net/hamradio/bpqether.c +++ b/drivers/net/hamradio/bpqether.c @@ -123,6 +123,7 @@ static LIST_HEAD(bpq_devices); * off into a separate class since they always nest. */ static struct lock_class_key bpq_netdev_xmit_lock_key; +static struct lock_class_key bpq_netdev_addr_lock_key; static void bpq_set_lockdep_class_one(struct net_device *dev, struct netdev_queue *txq, @@ -133,6 +134,7 @@ static void bpq_set_lockdep_class_one(struct net_device *dev, static void bpq_set_lockdep_class(struct net_device *dev) { + lockdep_set_class(&dev->addr_list_lock, &bpq_netdev_addr_lock_key); netdev_for_each_tx_queue(dev, bpq_set_lockdep_class_one, NULL); } diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c index efbc15567dd3..42394505bb50 100644 --- a/drivers/net/macvlan.c +++ b/drivers/net/macvlan.c @@ -276,6 +276,7 @@ static int macvlan_change_mtu(struct net_device *dev, int new_mtu) * separate class since they always nest. */ static struct lock_class_key macvlan_netdev_xmit_lock_key; +static struct lock_class_key macvlan_netdev_addr_lock_key; #define MACVLAN_FEATURES \ (NETIF_F_SG | NETIF_F_ALL_CSUM | NETIF_F_HIGHDMA | NETIF_F_FRAGLIST | \ @@ -295,6 +296,8 @@ static void macvlan_set_lockdep_class_one(struct net_device *dev, static void macvlan_set_lockdep_class(struct net_device *dev) { + lockdep_set_class(&dev->addr_list_lock, + &macvlan_netdev_addr_lock_key); netdev_for_each_tx_queue(dev, macvlan_set_lockdep_class_one, NULL); } diff --git a/drivers/net/wireless/hostap/hostap_hw.c b/drivers/net/wireless/hostap/hostap_hw.c index 13d5882f1f21..3153fe9d7ce0 100644 --- a/drivers/net/wireless/hostap/hostap_hw.c +++ b/drivers/net/wireless/hostap/hostap_hw.c @@ -3101,6 +3101,7 @@ static void prism2_clear_set_tim_queue(local_info_t *local) * This is a natural nesting, which needs a split lock type. */ static struct lock_class_key hostap_netdev_xmit_lock_key; +static struct lock_class_key hostap_netdev_addr_lock_key; static void prism2_set_lockdep_class_one(struct net_device *dev, struct netdev_queue *txq, @@ -3112,6 +3113,8 @@ static void prism2_set_lockdep_class_one(struct net_device *dev, static void prism2_set_lockdep_class(struct net_device *dev) { + lockdep_set_class(&dev->addr_list_lock, + &hostap_netdev_addr_lock_key); netdev_for_each_tx_queue(dev, prism2_set_lockdep_class_one, NULL); } diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index f42bc2b26b85..4bf014e51f8c 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -569,6 +569,7 @@ static void vlan_dev_set_rx_mode(struct net_device *vlan_dev) * separate class since they always nest. */ static struct lock_class_key vlan_netdev_xmit_lock_key; +static struct lock_class_key vlan_netdev_addr_lock_key; static void vlan_dev_set_lockdep_one(struct net_device *dev, struct netdev_queue *txq, @@ -581,6 +582,9 @@ static void vlan_dev_set_lockdep_one(struct net_device *dev, static void vlan_dev_set_lockdep_class(struct net_device *dev, int subclass) { + lockdep_set_class_and_subclass(&dev->addr_list_lock, + &vlan_netdev_addr_lock_key, + subclass); netdev_for_each_tx_queue(dev, vlan_dev_set_lockdep_one, &subclass); } diff --git a/net/core/dev.c b/net/core/dev.c index 65eea83613ef..6bf217da9d8f 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -261,7 +261,7 @@ static RAW_NOTIFIER_HEAD(netdev_chain); DEFINE_PER_CPU(struct softnet_data, softnet_data); -#ifdef CONFIG_DEBUG_LOCK_ALLOC +#ifdef CONFIG_LOCKDEP /* * register_netdevice() inits txq->_xmit_lock and sets lockdep class * according to dev->type @@ -301,6 +301,7 @@ static const char *netdev_lock_name[] = "_xmit_NONE"}; static struct lock_class_key netdev_xmit_lock_key[ARRAY_SIZE(netdev_lock_type)]; +static struct lock_class_key netdev_addr_lock_key[ARRAY_SIZE(netdev_lock_type)]; static inline unsigned short netdev_lock_pos(unsigned short dev_type) { @@ -313,8 +314,8 @@ static inline unsigned short netdev_lock_pos(unsigned short dev_type) return ARRAY_SIZE(netdev_lock_type) - 1; } -static inline void netdev_set_lockdep_class(spinlock_t *lock, - unsigned short dev_type) +static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock, + unsigned short dev_type) { int i; @@ -322,9 +323,22 @@ static inline void netdev_set_lockdep_class(spinlock_t *lock, lockdep_set_class_and_name(lock, &netdev_xmit_lock_key[i], netdev_lock_name[i]); } + +static inline void netdev_set_addr_lockdep_class(struct net_device *dev) +{ + int i; + + i = netdev_lock_pos(dev->type); + lockdep_set_class_and_name(&dev->addr_list_lock, + &netdev_addr_lock_key[i], + netdev_lock_name[i]); +} #else -static inline void netdev_set_lockdep_class(spinlock_t *lock, - unsigned short dev_type) +static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock, + unsigned short dev_type) +{ +} +static inline void netdev_set_addr_lockdep_class(struct net_device *dev) { } #endif @@ -3851,7 +3865,7 @@ static void __netdev_init_queue_locks_one(struct net_device *dev, void *_unused) { spin_lock_init(&dev_queue->_xmit_lock); - netdev_set_lockdep_class(&dev_queue->_xmit_lock, dev->type); + netdev_set_xmit_lockdep_class(&dev_queue->_xmit_lock, dev->type); dev_queue->xmit_lock_owner = -1; } @@ -3896,6 +3910,7 @@ int register_netdevice(struct net_device *dev) net = dev_net(dev); spin_lock_init(&dev->addr_list_lock); + netdev_set_addr_lockdep_class(dev); netdev_init_queue_locks(dev); dev->iflink = -1; diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index fccc250f95f5..532e4faa29f7 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -73,6 +73,7 @@ static const struct proto_ops nr_proto_ops; * separate class since they always nest. */ static struct lock_class_key nr_netdev_xmit_lock_key; +static struct lock_class_key nr_netdev_addr_lock_key; static void nr_set_lockdep_one(struct net_device *dev, struct netdev_queue *txq, @@ -83,6 +84,7 @@ static void nr_set_lockdep_one(struct net_device *dev, static void nr_set_lockdep_key(struct net_device *dev) { + lockdep_set_class(&dev->addr_list_lock, &nr_netdev_addr_lock_key); netdev_for_each_tx_queue(dev, nr_set_lockdep_one, NULL); } diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index dbc963b4f5fb..a7f1ce11bc22 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -74,6 +74,7 @@ ax25_address rose_callsign; * separate class since they always nest. */ static struct lock_class_key rose_netdev_xmit_lock_key; +static struct lock_class_key rose_netdev_addr_lock_key; static void rose_set_lockdep_one(struct net_device *dev, struct netdev_queue *txq, @@ -84,6 +85,7 @@ static void rose_set_lockdep_one(struct net_device *dev, static void rose_set_lockdep_key(struct net_device *dev) { + lockdep_set_class(&dev->addr_list_lock, &rose_netdev_addr_lock_key); netdev_for_each_tx_queue(dev, rose_set_lockdep_one, NULL); } -- cgit v1.3-8-gc7d7 From 16a37acaaf4aaa631ba3f83710ed6cdb1a597520 Mon Sep 17 00:00:00 2001 From: Maciej Sosnowski Date: Tue, 22 Jul 2008 17:30:57 -0700 Subject: I/OAT: tcp_dma_copybreak default value dependent on I/OAT version I/OAT DMA performance tuning showed different optimal values of tcp_dma_copybreak for different I/OAT versions (4096 for 1.2 and 2048 for 2.0). This patch lets ioatdma driver set tcp_dma_copybreak value according to these results. [dan.j.williams@intel.com: remove some ifdefs] Signed-off-by: Maciej Sosnowski Signed-off-by: Dan Williams --- drivers/dma/ioat_dma.c | 2 ++ drivers/dma/ioatdma.h | 15 +++++++++++++++ net/core/user_dma.c | 1 + 3 files changed, 18 insertions(+) (limited to 'net/core') diff --git a/drivers/dma/ioat_dma.c b/drivers/dma/ioat_dma.c index da572968a7db..ece5a0e3a335 100644 --- a/drivers/dma/ioat_dma.c +++ b/drivers/dma/ioat_dma.c @@ -1581,6 +1581,8 @@ struct ioatdma_device *ioat_dma_probe(struct pci_dev *pdev, if (err) goto err_self_test; + ioat_set_tcp_copy_break(device); + dma_async_device_register(&device->common); INIT_DELAYED_WORK(&device->work, ioat_dma_chan_watchdog); diff --git a/drivers/dma/ioatdma.h b/drivers/dma/ioatdma.h index c6ec933f9895..685adb62aa5a 100644 --- a/drivers/dma/ioatdma.h +++ b/drivers/dma/ioatdma.h @@ -27,6 +27,7 @@ #include #include #include +#include #define IOAT_DMA_VERSION "2.18" @@ -129,6 +130,20 @@ struct ioat_desc_sw { struct dma_async_tx_descriptor async_tx; }; +static inline void ioat_set_tcp_copy_break(struct ioatdma_device *dev) +{ + #ifdef CONFIG_NET_DMA + switch (dev->version) { + case IOAT_VER_1_2: + sysctl_tcp_dma_copybreak = 4096; + break; + case IOAT_VER_2_0: + sysctl_tcp_dma_copybreak = 2048; + break; + } + #endif +} + #if defined(CONFIG_INTEL_IOATDMA) || defined(CONFIG_INTEL_IOATDMA_MODULE) struct ioatdma_device *ioat_dma_probe(struct pci_dev *pdev, void __iomem *iobase); diff --git a/net/core/user_dma.c b/net/core/user_dma.c index 0ad1cd57bc39..de760504f6fe 100644 --- a/net/core/user_dma.c +++ b/net/core/user_dma.c @@ -34,6 +34,7 @@ #define NET_DMA_DEFAULT_COPYBREAK 4096 int sysctl_tcp_dma_copybreak = NET_DMA_DEFAULT_COPYBREAK; +EXPORT_SYMBOL(sysctl_tcp_dma_copybreak); /** * dma_skb_copy_datagram_iovec - Copy a datagram to an iovec. -- cgit v1.3-8-gc7d7