X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=net%2Fsched%2Fsch_generic.c;h=bc116bd6937c01ffba6adcb142161f38b05b7e8c;hb=97bf2856c6014879bd04983a3e9dfcdac1e7fe85;hp=82117f9ba7d1907dba3dcb800f95dc5236ad49c6;hpb=9bf4aaab3e101692164d49b7ca357651eb691cb6;p=linux-2.6.git diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 82117f9ba..bc116bd69 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -13,8 +13,7 @@ #include #include -#include -#include +#include #include #include #include @@ -46,35 +45,34 @@ The idea is the following: - enqueue, dequeue are serialized via top level device spinlock dev->queue_lock. - - tree walking is protected by read_lock_bh(qdisc_tree_lock) + - tree walking is protected by read_lock(qdisc_tree_lock) and this lock is used only in process context. - - updates to tree are made under rtnl semaphore or - from softirq context (__qdisc_destroy rcu-callback) - hence this lock needs local bh disabling. + - updates to tree are made only under rtnl semaphore, + hence this lock may be made without local bh disabling. qdisc_tree_lock must be grabbed BEFORE dev->queue_lock! */ -rwlock_t qdisc_tree_lock = RW_LOCK_UNLOCKED; +DEFINE_RWLOCK(qdisc_tree_lock); void qdisc_lock_tree(struct net_device *dev) { - write_lock_bh(&qdisc_tree_lock); + write_lock(&qdisc_tree_lock); spin_lock_bh(&dev->queue_lock); } void qdisc_unlock_tree(struct net_device *dev) { spin_unlock_bh(&dev->queue_lock); - write_unlock_bh(&qdisc_tree_lock); + write_unlock(&qdisc_tree_lock); } /* dev->queue_lock serializes queue accesses for this device AND dev->qdisc pointer itself. - dev->xmit_lock serializes accesses to device driver. + netif_tx_lock serializes accesses to device driver. - dev->queue_lock and dev->xmit_lock are mutually exclusive, + dev->queue_lock and netif_tx_lock are mutually exclusive, if one is grabbed, another must be free. */ @@ -90,53 +88,75 @@ void qdisc_unlock_tree(struct net_device *dev) NOTE: Called under dev->queue_lock with locally disabled BH. */ -int qdisc_restart(struct net_device *dev) +static inline int qdisc_restart(struct net_device *dev) { struct Qdisc *q = dev->qdisc; struct sk_buff *skb; /* Dequeue packet */ - if ((skb = q->dequeue(q)) != NULL) { - if (spin_trylock(&dev->xmit_lock)) { - /* Remember that the driver is grabbed by us. */ - dev->xmit_lock_owner = smp_processor_id(); - + if (((skb = dev->gso_skb)) || ((skb = q->dequeue(q)))) { + unsigned nolock = (dev->features & NETIF_F_LLTX); + + dev->gso_skb = NULL; + + /* + * When the driver has LLTX set it does its own locking + * in start_xmit. No need to add additional overhead by + * locking again. These checks are worth it because + * even uncongested locks can be quite expensive. + * The driver can do trylock like here too, in case + * of lock congestion it should return -1 and the packet + * will be requeued. + */ + if (!nolock) { + if (!netif_tx_trylock(dev)) { + collision: + /* So, someone grabbed the driver. */ + + /* It may be transient configuration error, + when hard_start_xmit() recurses. We detect + it by checking xmit owner and drop the + packet when deadloop is detected. + */ + if (dev->xmit_lock_owner == smp_processor_id()) { + kfree_skb(skb); + if (net_ratelimit()) + printk(KERN_DEBUG "Dead loop on netdevice %s, fix it urgently!\n", dev->name); + return -1; + } + __get_cpu_var(netdev_rx_stat).cpu_collision++; + goto requeue; + } + } + + { /* And release queue */ spin_unlock(&dev->queue_lock); if (!netif_queue_stopped(dev)) { - if (netdev_nit) - dev_queue_xmit_nit(skb, dev); - - if (dev->hard_start_xmit(skb, dev) == 0) { - dev->xmit_lock_owner = -1; - spin_unlock(&dev->xmit_lock); + int ret; + ret = dev_hard_start_xmit(skb, dev); + if (ret == NETDEV_TX_OK) { + if (!nolock) { + netif_tx_unlock(dev); + } spin_lock(&dev->queue_lock); return -1; } + if (ret == NETDEV_TX_LOCKED && nolock) { + spin_lock(&dev->queue_lock); + goto collision; + } } + /* NETDEV_TX_BUSY - we need to requeue */ /* Release the driver */ - dev->xmit_lock_owner = -1; - spin_unlock(&dev->xmit_lock); + if (!nolock) { + netif_tx_unlock(dev); + } spin_lock(&dev->queue_lock); q = dev->qdisc; - } else { - /* So, someone grabbed the driver. */ - - /* It may be transient configuration error, - when hard_start_xmit() recurses. We detect - it by checking xmit owner and drop the - packet when deadloop is detected. - */ - if (dev->xmit_lock_owner == smp_processor_id()) { - kfree_skb(skb); - if (net_ratelimit()) - printk(KERN_DEBUG "Dead loop on netdevice %s, fix it urgently!\n", dev->name); - return -1; - } - __get_cpu_var(netdev_rx_stat).cpu_collision++; } /* Device kicked us out :( @@ -149,32 +169,51 @@ int qdisc_restart(struct net_device *dev) 3. device is buggy (ppp) */ - q->ops->requeue(skb, q); +requeue: + if (skb->next) + dev->gso_skb = skb; + else + q->ops->requeue(skb, q); netif_schedule(dev); return 1; } + BUG_ON((int) q->q.qlen < 0); return q->q.qlen; } +void __qdisc_run(struct net_device *dev) +{ + if (unlikely(dev->qdisc == &noop_qdisc)) + goto out; + + while (qdisc_restart(dev) < 0 && !netif_queue_stopped(dev)) + /* NOTHING */; + +out: + clear_bit(__LINK_STATE_QDISC_RUNNING, &dev->state); +} + static void dev_watchdog(unsigned long arg) { struct net_device *dev = (struct net_device *)arg; - spin_lock(&dev->xmit_lock); + netif_tx_lock(dev); if (dev->qdisc != &noop_qdisc) { if (netif_device_present(dev) && netif_running(dev) && netif_carrier_ok(dev)) { if (netif_queue_stopped(dev) && - (jiffies - dev->trans_start) > dev->watchdog_timeo) { - printk(KERN_INFO "NETDEV WATCHDOG: %s: transmit timed out\n", dev->name); + 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); } if (!mod_timer(&dev->watchdog_timer, jiffies + dev->watchdog_timeo)) dev_hold(dev); } } - spin_unlock(&dev->xmit_lock); + netif_tx_unlock(dev); dev_put(dev); } @@ -198,17 +237,29 @@ void __netdev_watchdog_up(struct net_device *dev) static void dev_watchdog_up(struct net_device *dev) { - spin_lock_bh(&dev->xmit_lock); __netdev_watchdog_up(dev); - spin_unlock_bh(&dev->xmit_lock); } static void dev_watchdog_down(struct net_device *dev) { - spin_lock_bh(&dev->xmit_lock); + netif_tx_lock_bh(dev); if (del_timer(&dev->watchdog_timer)) - __dev_put(dev); - spin_unlock_bh(&dev->xmit_lock); + dev_put(dev); + netif_tx_unlock_bh(dev); +} + +void netif_carrier_on(struct net_device *dev) +{ + if (test_and_clear_bit(__LINK_STATE_NOCARRIER, &dev->state)) + linkwatch_fire_event(dev); + if (netif_running(dev)) + __netdev_watchdog_up(dev); +} + +void netif_carrier_off(struct net_device *dev) +{ + if (!test_and_set_bit(__LINK_STATE_NOCARRIER, &dev->state)) + linkwatch_fire_event(dev); } /* "NOOP" scheduler: the best scheduler, recommended for all interfaces @@ -216,31 +267,27 @@ static void dev_watchdog_down(struct net_device *dev) cheaper. */ -static int -noop_enqueue(struct sk_buff *skb, struct Qdisc * qdisc) +static int noop_enqueue(struct sk_buff *skb, struct Qdisc * qdisc) { kfree_skb(skb); return NET_XMIT_CN; } -static struct sk_buff * -noop_dequeue(struct Qdisc * qdisc) +static struct sk_buff *noop_dequeue(struct Qdisc * qdisc) { return NULL; } -static int -noop_requeue(struct sk_buff *skb, struct Qdisc* qdisc) +static int noop_requeue(struct sk_buff *skb, struct Qdisc* qdisc) { if (net_ratelimit()) - printk(KERN_DEBUG "%s deferred output. It is buggy.\n", skb->dev->name); + printk(KERN_DEBUG "%s deferred output. It is buggy.\n", + skb->dev->name); kfree_skb(skb); return NET_XMIT_CN; } struct Qdisc_ops noop_qdisc_ops = { - .next = NULL, - .cl_ops = NULL, .id = "noop", .priv_size = 0, .enqueue = noop_enqueue, @@ -254,11 +301,10 @@ struct Qdisc noop_qdisc = { .dequeue = noop_dequeue, .flags = TCQ_F_BUILTIN, .ops = &noop_qdisc_ops, + .list = LIST_HEAD_INIT(noop_qdisc.list), }; -struct Qdisc_ops noqueue_qdisc_ops = { - .next = NULL, - .cl_ops = NULL, +static struct Qdisc_ops noqueue_qdisc_ops = { .id = "noqueue", .priv_size = 0, .enqueue = noop_enqueue, @@ -267,11 +313,12 @@ struct Qdisc_ops noqueue_qdisc_ops = { .owner = THIS_MODULE, }; -struct Qdisc noqueue_qdisc = { +static struct Qdisc noqueue_qdisc = { .enqueue = NULL, .dequeue = noop_dequeue, .flags = TCQ_F_BUILTIN, .ops = &noqueue_qdisc_ops, + .list = LIST_HEAD_INIT(noqueue_qdisc.list), }; @@ -282,96 +329,86 @@ static const u8 prio2band[TC_PRIO_MAX+1] = generic prio+fifo combination. */ -static int -pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc* qdisc) +#define PFIFO_FAST_BANDS 3 + +static inline struct sk_buff_head *prio2list(struct sk_buff *skb, + struct Qdisc *qdisc) { struct sk_buff_head *list = qdisc_priv(qdisc); + return list + prio2band[skb->priority & TC_PRIO_MAX]; +} - list += prio2band[skb->priority&TC_PRIO_MAX]; +static int pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc* qdisc) +{ + struct sk_buff_head *list = prio2list(skb, qdisc); - if (list->qlen < qdisc->dev->tx_queue_len) { - __skb_queue_tail(list, skb); + if (skb_queue_len(list) < qdisc->dev->tx_queue_len) { qdisc->q.qlen++; - qdisc->stats.bytes += skb->len; - qdisc->stats.packets++; - return 0; + return __qdisc_enqueue_tail(skb, qdisc, list); } - qdisc->stats.drops++; - kfree_skb(skb); - return NET_XMIT_DROP; + + return qdisc_drop(skb, qdisc); } -static struct sk_buff * -pfifo_fast_dequeue(struct Qdisc* qdisc) +static struct sk_buff *pfifo_fast_dequeue(struct Qdisc* qdisc) { int prio; struct sk_buff_head *list = qdisc_priv(qdisc); - struct sk_buff *skb; - for (prio = 0; prio < 3; prio++, list++) { - skb = __skb_dequeue(list); - if (skb) { + for (prio = 0; prio < PFIFO_FAST_BANDS; prio++) { + if (!skb_queue_empty(list + prio)) { qdisc->q.qlen--; - return skb; + return __qdisc_dequeue_head(qdisc, list + prio); } } + return NULL; } -static int -pfifo_fast_requeue(struct sk_buff *skb, struct Qdisc* qdisc) +static int pfifo_fast_requeue(struct sk_buff *skb, struct Qdisc* qdisc) { - struct sk_buff_head *list = qdisc_priv(qdisc); - - list += prio2band[skb->priority&TC_PRIO_MAX]; - - __skb_queue_head(list, skb); qdisc->q.qlen++; - return 0; + return __qdisc_requeue(skb, qdisc, prio2list(skb, qdisc)); } -static void -pfifo_fast_reset(struct Qdisc* qdisc) +static void pfifo_fast_reset(struct Qdisc* qdisc) { int prio; struct sk_buff_head *list = qdisc_priv(qdisc); - for (prio=0; prio < 3; prio++) - skb_queue_purge(list+prio); + for (prio = 0; prio < PFIFO_FAST_BANDS; prio++) + __qdisc_reset_queue(qdisc, list + prio); + + qdisc->qstats.backlog = 0; qdisc->q.qlen = 0; } static int pfifo_fast_dump(struct Qdisc *qdisc, struct sk_buff *skb) { - unsigned char *b = skb->tail; - struct tc_prio_qopt opt; + struct tc_prio_qopt opt = { .bands = PFIFO_FAST_BANDS }; - opt.bands = 3; memcpy(&opt.priomap, prio2band, TC_PRIO_MAX+1); RTA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt); return skb->len; rtattr_failure: - skb_trim(skb, b - skb->data); return -1; } static int pfifo_fast_init(struct Qdisc *qdisc, struct rtattr *opt) { - int i; + int prio; struct sk_buff_head *list = qdisc_priv(qdisc); - for (i=0; i<3; i++) - skb_queue_head_init(list+i); + for (prio = 0; prio < PFIFO_FAST_BANDS; prio++) + skb_queue_head_init(list + prio); return 0; } static struct Qdisc_ops pfifo_fast_ops = { - .next = NULL, - .cl_ops = NULL, .id = "pfifo_fast", - .priv_size = 3 * sizeof(struct sk_buff_head), + .priv_size = PFIFO_FAST_BANDS * sizeof(struct sk_buff_head), .enqueue = pfifo_fast_enqueue, .dequeue = pfifo_fast_dequeue, .requeue = pfifo_fast_requeue, @@ -381,24 +418,22 @@ static struct Qdisc_ops pfifo_fast_ops = { .owner = THIS_MODULE, }; -struct Qdisc * qdisc_create_dflt(struct net_device *dev, struct Qdisc_ops *ops) +struct Qdisc *qdisc_alloc(struct net_device *dev, struct Qdisc_ops *ops) { void *p; struct Qdisc *sch; - int size; + unsigned int size; + int err = -ENOBUFS; /* ensure that the Qdisc and the private data are 32-byte aligned */ - size = ((sizeof(*sch) + QDISC_ALIGN_CONST) & ~QDISC_ALIGN_CONST); - size += ops->priv_size + QDISC_ALIGN_CONST; + size = QDISC_ALIGN(sizeof(*sch)); + size += ops->priv_size + (QDISC_ALIGNTO - 1); - p = kmalloc(size, GFP_KERNEL); + p = kzalloc(size, GFP_KERNEL); if (!p) - return NULL; - memset(p, 0, size); - - sch = (struct Qdisc *)(((unsigned long)p + QDISC_ALIGN_CONST) - & ~QDISC_ALIGN_CONST); - sch->padded = (char *)sch - (char *)p; + goto errout; + sch = (struct Qdisc *) QDISC_ALIGN((unsigned long) p); + sch->padded = (char *) sch - (char *) p; INIT_LIST_HEAD(&sch->list); skb_queue_head_init(&sch->q); @@ -409,13 +444,27 @@ struct Qdisc * qdisc_create_dflt(struct net_device *dev, struct Qdisc_ops *ops) dev_hold(dev); sch->stats_lock = &dev->queue_lock; atomic_set(&sch->refcnt, 1); - /* enqueue is accessed locklessly - make sure it's visible - * before we set a netdevice's qdisc pointer to sch */ - smp_wmb(); + + return sch; +errout: + return ERR_PTR(-err); +} + +struct Qdisc * qdisc_create_dflt(struct net_device *dev, struct Qdisc_ops *ops, + unsigned int parentid) +{ + struct Qdisc *sch; + + sch = qdisc_alloc(dev, ops); + if (IS_ERR(sch)) + goto errout; + sch->parent = parentid; + if (!ops->init || ops->init(sch, NULL) == 0) return sch; - kfree(p); + qdisc_destroy(sch); +errout: return NULL; } @@ -435,31 +484,30 @@ void qdisc_reset(struct Qdisc *qdisc) static void __qdisc_destroy(struct rcu_head *head) { struct Qdisc *qdisc = container_of(head, struct Qdisc, q_rcu); + kfree((char *) qdisc - qdisc->padded); +} + +/* Under dev->queue_lock and BH! */ + +void qdisc_destroy(struct Qdisc *qdisc) +{ struct Qdisc_ops *ops = qdisc->ops; + if (qdisc->flags & TCQ_F_BUILTIN || + !atomic_dec_and_test(&qdisc->refcnt)) + return; + + list_del(&qdisc->list); #ifdef CONFIG_NET_ESTIMATOR - qdisc_kill_estimator(&qdisc->stats); + gen_kill_estimator(&qdisc->bstats, &qdisc->rate_est); #endif - write_lock(&qdisc_tree_lock); if (ops->reset) ops->reset(qdisc); if (ops->destroy) ops->destroy(qdisc); - write_unlock(&qdisc_tree_lock); - module_put(ops->owner); + module_put(ops->owner); dev_put(qdisc->dev); - if (!(qdisc->flags&TCQ_F_BUILTIN)) - kfree((char *) qdisc - qdisc->padded); -} - -/* Under dev->queue_lock and BH! */ - -void qdisc_destroy(struct Qdisc *qdisc) -{ - if (!atomic_dec_and_test(&qdisc->refcnt)) - return; - list_del(&qdisc->list); call_rcu(&qdisc->q_rcu, __qdisc_destroy); } @@ -474,24 +522,30 @@ 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, &pfifo_fast_ops, + TC_H_ROOT); if (qdisc == NULL) { printk(KERN_INFO "%s: activation failed\n", dev->name); return; } - write_lock_bh(&qdisc_tree_lock); + write_lock(&qdisc_tree_lock); list_add_tail(&qdisc->list, &dev->qdisc_list); - write_unlock_bh(&qdisc_tree_lock); + write_unlock(&qdisc_tree_lock); } else { qdisc = &noqueue_qdisc; } - write_lock_bh(&qdisc_tree_lock); + write_lock(&qdisc_tree_lock); dev->qdisc_sleeping = qdisc; - write_unlock_bh(&qdisc_tree_lock); + write_unlock(&qdisc_tree_lock); } + if (!netif_carrier_ok(dev)) + /* Delay activation until next carrier-on event */ + return; + spin_lock_bh(&dev->queue_lock); - if ((dev->qdisc = dev->qdisc_sleeping) != &noqueue_qdisc) { + rcu_assign_pointer(dev->qdisc, dev->qdisc_sleeping); + if (dev->qdisc != &noqueue_qdisc) { dev->trans_start = jiffies; dev_watchdog_up(dev); } @@ -512,10 +566,17 @@ void dev_deactivate(struct net_device *dev) dev_watchdog_down(dev); - while (test_bit(__LINK_STATE_SCHED, &dev->state)) + /* Wait for outstanding dev_queue_xmit calls. */ + synchronize_rcu(); + + /* Wait for outstanding qdisc_run calls. */ + while (test_bit(__LINK_STATE_QDISC_RUNNING, &dev->state)) yield(); - spin_unlock_wait(&dev->xmit_lock); + if (dev->gso_skb) { + kfree_skb(dev->gso_skb); + dev->gso_skb = NULL; + } } void dev_init_scheduler(struct net_device *dev) @@ -548,12 +609,11 @@ void dev_shutdown(struct net_device *dev) qdisc_unlock_tree(dev); } -EXPORT_SYMBOL(__netdev_watchdog_up); +EXPORT_SYMBOL(netif_carrier_on); +EXPORT_SYMBOL(netif_carrier_off); EXPORT_SYMBOL(noop_qdisc); -EXPORT_SYMBOL(noop_qdisc_ops); EXPORT_SYMBOL(qdisc_create_dflt); EXPORT_SYMBOL(qdisc_destroy); EXPORT_SYMBOL(qdisc_reset); -EXPORT_SYMBOL(qdisc_restart); EXPORT_SYMBOL(qdisc_lock_tree); EXPORT_SYMBOL(qdisc_unlock_tree);