From 8d50b53d66a8a6ae41bafbdcabe401467803f33a Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 30 Jul 2008 02:37:46 -0700 Subject: pkt_sched: Fix OOPS on ingress qdisc add. Bug report from Steven Jan Springl: Issuing the following command causes a kernel oops: tc qdisc add dev eth0 handle ffff: ingress The problem mostly stems from all of the special case handling of ingress qdiscs. So, to fix this, do the grafting operation the same way we do for TX qdiscs. Which means that dev_activate() and dev_deactivate() now do the "qdisc_sleeping <--> qdisc" transitions on dev->rx_queue too. Future simplifications are possible now, mainly because it is impossible for dev_queue->{qdisc,qdisc_sleeping} to be NULL. There are NULL checks all over to handle the ingress qdisc special case that used to exist before this commit. Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 8d13a9b9f1d..63d6bcddbf4 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2100,7 +2100,7 @@ static int ing_filter(struct sk_buff *skb) rxq = &dev->rx_queue; q = rxq->qdisc; - if (q) { + if (q != &noop_qdisc) { spin_lock(qdisc_lock(q)); result = qdisc_enqueue_root(skb, q); spin_unlock(qdisc_lock(q)); @@ -2113,7 +2113,7 @@ 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->rx_queue.qdisc) + if (skb->dev->rx_queue.qdisc == &noop_qdisc) goto out; if (*pt_prev) { -- cgit v1.2.3-70-g09d2 From c3f26a269c2421f97f10cf8ed05d5099b573af4d Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 31 Jul 2008 16:58:50 -0700 Subject: netdev: Fix lockdep warnings in multiqueue configurations. When support for multiple TX queues were added, the netif_tx_lock() routines we converted to iterate over all TX queues and grab each queue's spinlock. This causes heartburn for lockdep and it's not a healthy thing to do with lots of TX queues anyways. So modify this to use a top-level lock and a "frozen" state for the individual TX queues. Signed-off-by: David S. Miller --- drivers/net/ifb.c | 12 ++++--- include/linux/netdevice.h | 86 ++++++++++++++++++++++++++++++----------------- net/core/dev.c | 1 + net/core/netpoll.c | 1 + net/core/pktgen.c | 7 ++-- net/sched/sch_generic.c | 6 ++-- net/sched/sch_teql.c | 9 ++--- 7 files changed, 78 insertions(+), 44 deletions(-) (limited to 'net/core/dev.c') diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c index 0960e69b2da..e4fbefc8c82 100644 --- a/drivers/net/ifb.c +++ b/drivers/net/ifb.c @@ -69,18 +69,20 @@ static void ri_tasklet(unsigned long dev) struct net_device *_dev = (struct net_device *)dev; struct ifb_private *dp = netdev_priv(_dev); struct net_device_stats *stats = &_dev->stats; + struct netdev_queue *txq; struct sk_buff *skb; + txq = netdev_get_tx_queue(_dev, 0); dp->st_task_enter++; if ((skb = skb_peek(&dp->tq)) == NULL) { dp->st_txq_refl_try++; - if (netif_tx_trylock(_dev)) { + if (__netif_tx_trylock(txq)) { dp->st_rxq_enter++; while ((skb = skb_dequeue(&dp->rq)) != NULL) { skb_queue_tail(&dp->tq, skb); dp->st_rx2tx_tran++; } - netif_tx_unlock(_dev); + __netif_tx_unlock(txq); } else { /* reschedule */ dp->st_rxq_notenter++; @@ -115,7 +117,7 @@ static void ri_tasklet(unsigned long dev) BUG(); } - if (netif_tx_trylock(_dev)) { + if (__netif_tx_trylock(txq)) { dp->st_rxq_check++; if ((skb = skb_peek(&dp->rq)) == NULL) { dp->tasklet_pending = 0; @@ -123,10 +125,10 @@ static void ri_tasklet(unsigned long dev) netif_wake_queue(_dev); } else { dp->st_rxq_rsch++; - netif_tx_unlock(_dev); + __netif_tx_unlock(txq); goto resched; } - netif_tx_unlock(_dev); + __netif_tx_unlock(txq); } else { resched: dp->tasklet_pending = 1; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index b4d056ceab9..ee583f642a9 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -440,6 +440,7 @@ static inline void napi_synchronize(const struct napi_struct *n) enum netdev_queue_state_t { __QUEUE_STATE_XOFF, + __QUEUE_STATE_FROZEN, }; struct netdev_queue { @@ -636,7 +637,7 @@ struct net_device unsigned int real_num_tx_queues; unsigned long tx_queue_len; /* Max frames per queue allowed */ - + spinlock_t tx_global_lock; /* * One part is mostly used on xmit path (device) */ @@ -1099,6 +1100,11 @@ static inline int netif_queue_stopped(const struct net_device *dev) return netif_tx_queue_stopped(netdev_get_tx_queue(dev, 0)); } +static inline int netif_tx_queue_frozen(const struct netdev_queue *dev_queue) +{ + return test_bit(__QUEUE_STATE_FROZEN, &dev_queue->state); +} + /** * netif_running - test if up * @dev: network device @@ -1475,6 +1481,26 @@ static inline void __netif_tx_lock_bh(struct netdev_queue *txq) txq->xmit_lock_owner = smp_processor_id(); } +static inline int __netif_tx_trylock(struct netdev_queue *txq) +{ + int ok = spin_trylock(&txq->_xmit_lock); + if (likely(ok)) + txq->xmit_lock_owner = smp_processor_id(); + return ok; +} + +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_bh(struct netdev_queue *txq) +{ + txq->xmit_lock_owner = -1; + spin_unlock_bh(&txq->_xmit_lock); +} + /** * netif_tx_lock - grab network device transmit lock * @dev: network device @@ -1484,12 +1510,23 @@ static inline void __netif_tx_lock_bh(struct netdev_queue *txq) */ static inline void netif_tx_lock(struct net_device *dev) { - int cpu = smp_processor_id(); unsigned int i; + int cpu; + spin_lock(&dev->tx_global_lock); + cpu = smp_processor_id(); for (i = 0; i < dev->num_tx_queues; i++) { struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + + /* We are the only thread of execution doing a + * freeze, but we have to grab the _xmit_lock in + * order to synchronize with threads which are in + * the ->hard_start_xmit() handler and already + * checked the frozen bit. + */ __netif_tx_lock(txq, cpu); + set_bit(__QUEUE_STATE_FROZEN, &txq->state); + __netif_tx_unlock(txq); } } @@ -1499,40 +1536,22 @@ static inline void netif_tx_lock_bh(struct net_device *dev) netif_tx_lock(dev); } -static inline int __netif_tx_trylock(struct netdev_queue *txq) -{ - int ok = spin_trylock(&txq->_xmit_lock); - if (likely(ok)) - txq->xmit_lock_owner = smp_processor_id(); - return ok; -} - -static inline int netif_tx_trylock(struct net_device *dev) -{ - return __netif_tx_trylock(netdev_get_tx_queue(dev, 0)); -} - -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_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; for (i = 0; i < dev->num_tx_queues; i++) { struct netdev_queue *txq = netdev_get_tx_queue(dev, i); - __netif_tx_unlock(txq); - } + /* No need to grab the _xmit_lock here. If the + * queue is not stopped for another reason, we + * force a schedule. + */ + clear_bit(__QUEUE_STATE_FROZEN, &txq->state); + if (!test_bit(__QUEUE_STATE_XOFF, &txq->state)) + __netif_schedule(txq->qdisc); + } + spin_unlock(&dev->tx_global_lock); } static inline void netif_tx_unlock_bh(struct net_device *dev) @@ -1556,13 +1575,18 @@ static inline void netif_tx_unlock_bh(struct net_device *dev) static inline void netif_tx_disable(struct net_device *dev) { unsigned int i; + int cpu; - netif_tx_lock_bh(dev); + local_bh_disable(); + cpu = 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); netif_tx_stop_queue(txq); + __netif_tx_unlock(txq); } - netif_tx_unlock_bh(dev); + local_bh_enable(); } static inline void netif_addr_lock(struct net_device *dev) diff --git a/net/core/dev.c b/net/core/dev.c index 63d6bcddbf4..69320a56a08 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4200,6 +4200,7 @@ static void netdev_init_queues(struct net_device *dev) { netdev_init_one_queue(dev, &dev->rx_queue, NULL); netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL); + spin_lock_init(&dev->tx_global_lock); } /** diff --git a/net/core/netpoll.c b/net/core/netpoll.c index c12720895ec..6c7af390be0 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -70,6 +70,7 @@ static void queue_process(struct work_struct *work) local_irq_save(flags); __netif_tx_lock(txq, smp_processor_id()); if (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq) || dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) { skb_queue_head(&npinfo->txq, skb); __netif_tx_unlock(txq); diff --git a/net/core/pktgen.c b/net/core/pktgen.c index c7d484f7e1c..3284605f2ec 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -3305,6 +3305,7 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) txq = netdev_get_tx_queue(odev, queue_map); if (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq) || need_resched()) { idle_start = getCurUs(); @@ -3320,7 +3321,8 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) pkt_dev->idle_acc += getCurUs() - idle_start; - if (netif_tx_queue_stopped(txq)) { + if (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq)) { pkt_dev->next_tx_us = getCurUs(); /* TODO */ pkt_dev->next_tx_ns = 0; goto out; /* Try the next interface */ @@ -3352,7 +3354,8 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) txq = netdev_get_tx_queue(odev, queue_map); __netif_tx_lock_bh(txq); - if (!netif_tx_queue_stopped(txq)) { + if (!netif_tx_queue_stopped(txq) && + !netif_tx_queue_frozen(txq)) { atomic_inc(&(pkt_dev->skb->users)); retry_now: diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 345838a2e36..9c9cd4d9489 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -135,7 +135,8 @@ static inline int qdisc_restart(struct Qdisc *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)) + if (!netif_tx_queue_stopped(txq) && + !netif_tx_queue_frozen(txq)) ret = dev_hard_start_xmit(skb, dev, txq); HARD_TX_UNLOCK(dev, txq); @@ -162,7 +163,8 @@ static inline int qdisc_restart(struct Qdisc *q) break; } - if (ret && netif_tx_queue_stopped(txq)) + if (ret && (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq))) ret = 0; return ret; diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 537223642b6..2c35c678563 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -305,10 +305,11 @@ restart: switch (teql_resolve(skb, skb_res, slave)) { case 0: - if (netif_tx_trylock(slave)) { - if (!__netif_subqueue_stopped(slave, subq) && + if (__netif_tx_trylock(slave_txq)) { + if (!netif_tx_queue_stopped(slave_txq) && + !netif_tx_queue_frozen(slave_txq) && slave->hard_start_xmit(skb, slave) == 0) { - netif_tx_unlock(slave); + __netif_tx_unlock(slave_txq); master->slaves = NEXT_SLAVE(q); netif_wake_queue(dev); master->stats.tx_packets++; @@ -316,7 +317,7 @@ restart: qdisc_pkt_len(skb); return 0; } - netif_tx_unlock(slave); + __netif_tx_unlock(slave_txq); } if (netif_queue_stopped(dev)) busy = 1; -- cgit v1.2.3-70-g09d2 From 5fb662297b8a4bdadd60371c34b760efca948ebc Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sat, 2 Aug 2008 20:02:43 -0700 Subject: pkt_sched: Use qdisc_lock() on already sampled root qdisc. Based upon a bug report by Jeff Kirsher. Don't use qdisc_root_lock() in these cases as the root qdisc could have been changed, and we'd thus lock the wrong object. Tested by Emil S Tantilov who confirms that this seems to fix the problem. Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- net/sched/sch_generic.c | 12 ++++++------ 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 69320a56a08..da7acacf02b 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1796,7 +1796,7 @@ gso: skb->tc_verd = SET_TC_AT(skb->tc_verd,AT_EGRESS); #endif if (q->enqueue) { - spinlock_t *root_lock = qdisc_root_lock(q); + spinlock_t *root_lock = qdisc_lock(q); spin_lock(root_lock); @@ -1995,7 +1995,7 @@ static void net_tx_action(struct softirq_action *h) smp_mb__before_clear_bit(); clear_bit(__QDISC_STATE_SCHED, &q->state); - root_lock = qdisc_root_lock(q); + root_lock = qdisc_lock(q); if (spin_trylock(root_lock)) { qdisc_run(q); spin_unlock(root_lock); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 9c9cd4d9489..7cf83b37459 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -29,7 +29,7 @@ /* Main transmission queue. */ /* Modifications to data participating in scheduling must be protected with - * qdisc_root_lock(qdisc) spinlock. + * qdisc_lock(qdisc) spinlock. * * The idea is the following: * - enqueue, dequeue are serialized via qdisc root lock @@ -126,7 +126,7 @@ static inline int qdisc_restart(struct Qdisc *q) if (unlikely((skb = dequeue_skb(q)) == NULL)) return 0; - root_lock = qdisc_root_lock(q); + root_lock = qdisc_lock(q); /* And release qdisc */ spin_unlock(root_lock); @@ -507,7 +507,7 @@ errout: } EXPORT_SYMBOL(qdisc_create_dflt); -/* Under qdisc_root_lock(qdisc) and BH! */ +/* Under qdisc_lock(qdisc) and BH! */ void qdisc_reset(struct Qdisc *qdisc) { @@ -543,7 +543,7 @@ static void __qdisc_destroy(struct rcu_head *head) kfree((char *) qdisc - qdisc->padded); } -/* Under qdisc_root_lock(qdisc) and BH! */ +/* Under qdisc_lock(qdisc) and BH! */ void qdisc_destroy(struct Qdisc *qdisc) { @@ -659,7 +659,7 @@ static bool some_qdisc_is_running(struct net_device *dev, int lock) dev_queue = netdev_get_tx_queue(dev, i); q = dev_queue->qdisc; - root_lock = qdisc_root_lock(q); + root_lock = qdisc_lock(q); if (lock) spin_lock_bh(root_lock); @@ -735,7 +735,7 @@ static void shutdown_scheduler_queue(struct net_device *dev, struct Qdisc *qdisc_default = _qdisc_default; if (qdisc) { - spinlock_t *root_lock = qdisc_root_lock(qdisc); + spinlock_t *root_lock = qdisc_lock(qdisc); dev_queue->qdisc = qdisc_default; dev_queue->qdisc_sleeping = qdisc_default; -- cgit v1.2.3-70-g09d2 From e5a4a72d4f88f4389e9340d383ca67031d1b8536 Mon Sep 17 00:00:00 2001 From: Lennert Buytenhek Date: Sun, 3 Aug 2008 01:23:10 -0700 Subject: net: use software GSO for SG+CSUM capable netdevices If a netdevice does not support hardware GSO, allowing the stack to use GSO anyway and then splitting the GSO skb into MSS-sized pieces as it is handed to the netdevice for transmitting is likely still a win as far as throughput and/or CPU usage are concerned, since it reduces the number of trips through the output path. This patch enables the use of GSO on any netdevice that supports SG. If a GSO skb is then sent to a netdevice that supports SG but does not support hardware GSO, net/core/dev.c:dev_hard_start_xmit() will take care of doing the necessary GSO segmentation in software. Signed-off-by: Lennert Buytenhek Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index da7acacf02b..cbf80098980 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3988,6 +3988,10 @@ int register_netdevice(struct net_device *dev) } } + /* Enable software GSO if SG is supported. */ + if (dev->features & NETIF_F_SG) + dev->features |= NETIF_F_GSO; + netdev_initialize_kobject(dev); ret = netdev_register_kobject(dev); if (ret) -- cgit v1.2.3-70-g09d2 From 6e583ce5242f32e925dcb198f7123256d0798370 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Sun, 3 Aug 2008 21:29:57 -0700 Subject: net: eliminate refcounting in backlog queue Avoid the overhead of atomic increment/decrement on each received packet. This helps performance of non-NAPI devices (like loopback). Use cleanup function to walk queue on each cpu and clean out any left over packets. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/core/dev.c | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index cbf80098980..fc6c9881eca 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1909,7 +1909,6 @@ int netif_rx(struct sk_buff *skb) if (queue->input_pkt_queue.qlen <= netdev_max_backlog) { if (queue->input_pkt_queue.qlen) { enqueue: - dev_hold(skb->dev); __skb_queue_tail(&queue->input_pkt_queue, skb); local_irq_restore(flags); return NET_RX_SUCCESS; @@ -2270,6 +2269,20 @@ out: return ret; } +/* Network device is going away, flush any packets still pending */ +static void flush_backlog(void *arg) +{ + struct net_device *dev = arg; + struct softnet_data *queue = &__get_cpu_var(softnet_data); + struct sk_buff *skb, *tmp; + + skb_queue_walk_safe(&queue->input_pkt_queue, skb, tmp) + if (skb->dev == dev) { + __skb_unlink(skb, &queue->input_pkt_queue); + kfree_skb(skb); + } +} + static int process_backlog(struct napi_struct *napi, int quota) { int work = 0; @@ -2279,7 +2292,6 @@ static int process_backlog(struct napi_struct *napi, int quota) napi->weight = weight_p; do { struct sk_buff *skb; - struct net_device *dev; local_irq_disable(); skb = __skb_dequeue(&queue->input_pkt_queue); @@ -2288,14 +2300,9 @@ static int process_backlog(struct napi_struct *napi, int quota) local_irq_enable(); break; } - local_irq_enable(); - dev = skb->dev; - netif_receive_skb(skb); - - dev_put(dev); } while (++work < quota && jiffies == start_time); return work; @@ -4169,6 +4176,8 @@ void netdev_run_todo(void) dev->reg_state = NETREG_UNREGISTERED; + on_each_cpu(flush_backlog, dev, 1); + netdev_wait_allrefs(dev); /* paranoia */ -- cgit v1.2.3-70-g09d2 From c27f339af90bb874a7a9c680b17abfd32d4a727b Mon Sep 17 00:00:00 2001 From: Jarek Poplawski Date: Mon, 4 Aug 2008 22:39:11 -0700 Subject: net_sched: Add qdisc __NET_XMIT_BYPASS flag Patrick McHardy noticed that it would be nice to handle NET_XMIT_BYPASS by NET_XMIT_SUCCESS with an internal qdisc flag __NET_XMIT_BYPASS and to remove the mapping from dev_queue_xmit(). David Miller spotted a serious bug in the first version of this patch. Signed-off-by: Jarek Poplawski Signed-off-by: David S. Miller --- include/net/sch_generic.h | 6 +++--- net/core/dev.c | 1 - net/sched/sch_atm.c | 2 +- net/sched/sch_cbq.c | 4 ++-- net/sched/sch_dsmark.c | 2 +- net/sched/sch_hfsc.c | 4 ++-- net/sched/sch_htb.c | 6 +++--- net/sched/sch_netem.c | 2 +- net/sched/sch_prio.c | 6 +++--- net/sched/sch_sfq.c | 6 +++--- 10 files changed, 19 insertions(+), 20 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index f15b045a85e..a7abfda3e44 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -343,14 +343,14 @@ static inline unsigned int qdisc_pkt_len(struct sk_buff *skb) return qdisc_skb_cb(skb)->pkt_len; } -#ifdef CONFIG_NET_CLS_ACT -/* additional qdisc xmit flags */ +/* additional qdisc xmit flags (NET_XMIT_MASK in linux/netdevice.h) */ enum net_xmit_qdisc_t { __NET_XMIT_STOLEN = 0x00010000, + __NET_XMIT_BYPASS = 0x00020000, }; +#ifdef CONFIG_NET_CLS_ACT #define net_xmit_drop_count(e) ((e) & __NET_XMIT_STOLEN ? 0 : 1) - #else #define net_xmit_drop_count(e) (1) #endif diff --git a/net/core/dev.c b/net/core/dev.c index fc6c9881eca..01993ad74e7 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1805,7 +1805,6 @@ gso: spin_unlock(root_lock); - rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; goto out; } diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 27dd773481b..43d37256c15 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -457,7 +457,7 @@ drop: __maybe_unused return 0; } tasklet_schedule(&p->task); - return NET_XMIT_BYPASS; + return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; } /* diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 765ae565900..4e261ce62f4 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -230,7 +230,7 @@ cbq_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) (cl = cbq_class_lookup(q, prio)) != NULL) return cl; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; for (;;) { int result = 0; defmap = head->defaults; @@ -377,7 +377,7 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch) q->rx_class = cl; #endif if (cl == NULL) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index 7170275d9f9..edd1298f85f 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -268,7 +268,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch) drop: kfree_skb(skb); sch->qstats.drops++; - return NET_XMIT_BYPASS; + return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; } static struct sk_buff *dsmark_dequeue(struct Qdisc *sch) diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 5cf9ae71611..c2b8d9cce3d 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1159,7 +1159,7 @@ hfsc_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) if (cl->level == 0) return cl; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; tcf = q->root.filter_list; while (tcf && (result = tc_classify(skb, tcf, &res)) >= 0) { #ifdef CONFIG_NET_CLS_ACT @@ -1578,7 +1578,7 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) cl = hfsc_classify(skb, sch, &err); if (cl == NULL) { - if (err == NET_XMIT_BYPASS) + if (err & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return err; diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 538d79b489a..be35422711a 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -214,7 +214,7 @@ static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, if ((cl = htb_find(skb->priority, sch)) != NULL && cl->level == 0) return cl; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; tcf = q->filter_list; while (tcf && (result = tc_classify(skb, tcf, &res)) >= 0) { #ifdef CONFIG_NET_CLS_ACT @@ -567,7 +567,7 @@ static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch) } #ifdef CONFIG_NET_CLS_ACT } else if (!cl) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; @@ -612,7 +612,7 @@ static int htb_requeue(struct sk_buff *skb, struct Qdisc *sch) } #ifdef CONFIG_NET_CLS_ACT } else if (!cl) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 6cd6f2bc749..fb0294d0b55 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -176,7 +176,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) if (count == 0) { sch->qstats.drops++; kfree_skb(skb); - return NET_XMIT_BYPASS; + return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; } skb_orphan(skb); diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index adb1a52b77d..eac197610ed 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -38,7 +38,7 @@ prio_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) struct tcf_result res; int err; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; if (TC_H_MAJ(skb->priority) != sch->handle) { err = tc_classify(skb, q->filter_list, &res); #ifdef CONFIG_NET_CLS_ACT @@ -74,7 +74,7 @@ prio_enqueue(struct sk_buff *skb, struct Qdisc *sch) #ifdef CONFIG_NET_CLS_ACT if (qdisc == NULL) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; @@ -103,7 +103,7 @@ prio_requeue(struct sk_buff *skb, struct Qdisc* sch) qdisc = prio_classify(skb, sch, &ret); #ifdef CONFIG_NET_CLS_ACT if (qdisc == NULL) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c index 3a456e1b829..6e041d10dbd 100644 --- a/net/sched/sch_sfq.c +++ b/net/sched/sch_sfq.c @@ -171,7 +171,7 @@ static unsigned int sfq_classify(struct sk_buff *skb, struct Qdisc *sch, if (!q->filter_list) return sfq_hash(q, skb) + 1; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; result = tc_classify(skb, q->filter_list, &res); if (result >= 0) { #ifdef CONFIG_NET_CLS_ACT @@ -285,7 +285,7 @@ sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch) hash = sfq_classify(skb, sch, &ret); if (hash == 0) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; @@ -339,7 +339,7 @@ sfq_requeue(struct sk_buff *skb, struct Qdisc *sch) hash = sfq_classify(skb, sch, &ret); if (hash == 0) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; -- cgit v1.2.3-70-g09d2 From cc9bd5cebc0825e0fabc0186ab85806a0891104f Mon Sep 17 00:00:00 2001 From: Joe Eykholt Date: Wed, 2 Jul 2008 18:22:00 -0700 Subject: net/core: Uninline skb_bond(). Otherwise subsequent changes need multiple return values. Signed-off-by: Joe Eykholt Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- net/core/dev.c | 28 ++++++++-------------------- 1 file changed, 8 insertions(+), 20 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 01993ad74e7..4a09833331f 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1939,22 +1939,6 @@ int netif_rx_ni(struct sk_buff *skb) EXPORT_SYMBOL(netif_rx_ni); -static inline struct net_device *skb_bond(struct sk_buff *skb) -{ - struct net_device *dev = skb->dev; - - if (dev->master) { - if (skb_bond_should_drop(skb)) { - kfree_skb(skb); - return NULL; - } - skb->dev = dev->master; - } - - return dev; -} - - static void net_tx_action(struct softirq_action *h) { struct softnet_data *sd = &__get_cpu_var(softnet_data); @@ -2194,10 +2178,14 @@ int netif_receive_skb(struct sk_buff *skb) if (!skb->iif) skb->iif = skb->dev->ifindex; - orig_dev = skb_bond(skb); - - if (!orig_dev) - return NET_RX_DROP; + orig_dev = skb->dev; + if (orig_dev->master) { + if (skb_bond_should_drop(skb)) { + kfree_skb(skb); + return NET_RX_DROP; + } + skb->dev = orig_dev->master; + } __get_cpu_var(netdev_rx_stat).total++; -- cgit v1.2.3-70-g09d2 From 0d7a3681232f545c6a59f77e60f7667673ef0e93 Mon Sep 17 00:00:00 2001 From: Joe Eykholt Date: Wed, 2 Jul 2008 18:22:01 -0700 Subject: net/core: Allow certain receives on inactive slave. Allow a packet_type that specifies the exact device to receive even on an inactive bonding slave devices. This is important for some L2 protocols such as LLDP and FCoE. This can eventually be used for the bonding special cases as well. Signed-off-by: Joe Eykholt Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- net/core/dev.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 4a09833331f..dab97c7cf27 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2165,6 +2165,7 @@ int netif_receive_skb(struct sk_buff *skb) { struct packet_type *ptype, *pt_prev; struct net_device *orig_dev; + struct net_device *null_or_orig; int ret = NET_RX_DROP; __be16 type; @@ -2178,13 +2179,13 @@ int netif_receive_skb(struct sk_buff *skb) if (!skb->iif) skb->iif = skb->dev->ifindex; + null_or_orig = NULL; orig_dev = skb->dev; if (orig_dev->master) { - if (skb_bond_should_drop(skb)) { - kfree_skb(skb); - return NET_RX_DROP; - } - skb->dev = orig_dev->master; + if (skb_bond_should_drop(skb)) + null_or_orig = orig_dev; /* deliver only exact match */ + else + skb->dev = orig_dev->master; } __get_cpu_var(netdev_rx_stat).total++; @@ -2209,7 +2210,7 @@ int netif_receive_skb(struct sk_buff *skb) #endif list_for_each_entry_rcu(ptype, &ptype_all, list) { - if (!ptype->dev || ptype->dev == skb->dev) { + if (ptype->dev == null_or_orig || ptype->dev == skb->dev) { if (pt_prev) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; @@ -2234,7 +2235,7 @@ ncls: list_for_each_entry_rcu(ptype, &ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) { if (ptype->type == type && - (!ptype->dev || ptype->dev == skb->dev)) { + (ptype->dev == null_or_orig || ptype->dev == skb->dev)) { if (pt_prev) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; -- cgit v1.2.3-70-g09d2 From f982307f22db96201e41540295f24e8dcc10c78f Mon Sep 17 00:00:00 2001 From: Joe Eykholt Date: Wed, 2 Jul 2008 18:22:02 -0700 Subject: net/core: Allow receive on active slaves. If a packet_type specifies an active slave to bonding and not just any interface, allow it to receive frames that came in on that interface. Signed-off-by: Joe Eykholt Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- net/core/dev.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index dab97c7cf27..600bb23c4c2 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2210,7 +2210,8 @@ int netif_receive_skb(struct sk_buff *skb) #endif list_for_each_entry_rcu(ptype, &ptype_all, list) { - if (ptype->dev == null_or_orig || ptype->dev == skb->dev) { + if (ptype->dev == null_or_orig || ptype->dev == skb->dev || + ptype->dev == orig_dev) { if (pt_prev) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; @@ -2235,7 +2236,8 @@ ncls: list_for_each_entry_rcu(ptype, &ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) { if (ptype->type == type && - (ptype->dev == null_or_orig || ptype->dev == skb->dev)) { + (ptype->dev == null_or_orig || ptype->dev == skb->dev || + ptype->dev == orig_dev)) { if (pt_prev) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; -- cgit v1.2.3-70-g09d2