diff options
Diffstat (limited to 'net/sched/sch_htb.c')
| -rw-r--r-- | net/sched/sch_htb.c | 671 |
1 files changed, 621 insertions, 50 deletions
diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 8184c87da8be..e5b4bbf3ce3d 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -52,7 +52,7 @@ */ static int htb_hysteresis __read_mostly = 0; /* whether to use mode hysteresis for speedup */ -#define HTB_VER 0x30011 /* major must be matched with number suplied by TC as version */ +#define HTB_VER 0x30011 /* major must be matched with number supplied by TC as version */ #if HTB_VER >> 16 != TC_HTB_PROTOVER #error "Mismatched sch_htb.c and pkt_sch.h" @@ -113,7 +113,8 @@ struct htb_class { /* * Written often fields */ - struct gnet_stats_basic_packed bstats; + struct gnet_stats_basic_sync bstats; + struct gnet_stats_basic_sync bstats_bias; struct tc_htb_xstats xstats; /* our special stats */ /* token bucket parameters */ @@ -124,6 +125,7 @@ struct htb_class { struct htb_class_leaf { int deficit[TC_HTB_MAXDEPTH]; struct Qdisc *q; + struct netdev_queue *offload_queue; } leaf; struct htb_class_inner { struct htb_prio clprio[TC_HTB_NUMPRIO]; @@ -174,6 +176,11 @@ struct htb_sched { int row_mask[TC_HTB_MAXDEPTH]; struct htb_level hlevel[TC_HTB_MAXDEPTH]; + + struct Qdisc **direct_qdiscs; + unsigned int num_direct_qdiscs; + + bool offload; }; /* find class in global hash table using given handle */ @@ -232,14 +239,14 @@ static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, } *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; - while (tcf && (result = tcf_classify(skb, tcf, &res, false)) >= 0) { + while (tcf && (result = tcf_classify(skb, NULL, tcf, &res, false)) >= 0) { #ifdef CONFIG_NET_CLS_ACT switch (result) { case TC_ACT_QUEUED: case TC_ACT_STOLEN: case TC_ACT_TRAP: *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; - /* fall through */ + fallthrough; case TC_ACT_SHOT: return NULL; } @@ -267,6 +274,9 @@ static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, /** * htb_add_to_id_tree - adds class to the round robin list + * @root: the root of the tree + * @cl: the class to add + * @prio: the give prio in class * * Routine adds class to the list (actually tree) sorted by classid. * Make sure that class is not already on such list for given prio. @@ -292,6 +302,9 @@ static void htb_add_to_id_tree(struct rb_root *root, /** * htb_add_to_wait_tree - adds class to the event queue with delay + * @q: the priority event queue + * @cl: the class to add + * @delay: delay in microseconds * * The class is added to priority event queue to indicate that class will * change its mode in cl->pq_key microseconds. Make sure that class is not @@ -325,6 +338,7 @@ static void htb_add_to_wait_tree(struct htb_sched *q, /** * htb_next_rb_node - finds next node in binary tree + * @n: the current node in binary tree * * When we are past last key we return NULL. * Average complexity is 2 steps per call. @@ -336,6 +350,9 @@ static inline void htb_next_rb_node(struct rb_node **n) /** * htb_add_class_to_row - add class to its row + * @q: the priority event queue + * @cl: the class to add + * @mask: the given priorities in class in bitmap * * The class is added to row at priorities marked in mask. * It does nothing if mask == 0. @@ -365,6 +382,9 @@ static void htb_safe_rb_erase(struct rb_node *rb, struct rb_root *root) /** * htb_remove_class_from_row - removes class from its row + * @q: the priority event queue + * @cl: the class to add + * @mask: the given priorities in class in bitmap * * The class is removed from row at priorities marked in mask. * It does nothing if mask == 0. @@ -392,6 +412,8 @@ static inline void htb_remove_class_from_row(struct htb_sched *q, /** * htb_activate_prios - creates active classe's feed chain + * @q: the priority event queue + * @cl: the class to activate * * The class is connected to ancestors and/or appropriate rows * for priorities it is participating on. cl->cmode must be new @@ -427,6 +449,8 @@ static void htb_activate_prios(struct htb_sched *q, struct htb_class *cl) /** * htb_deactivate_prios - remove class from feed chain + * @q: the priority event queue + * @cl: the class to deactivate * * cl->cmode must represent old mode (before deactivation). It does * nothing if cl->prio_activity == 0. Class is removed from all feed @@ -487,6 +511,8 @@ static inline s64 htb_hiwater(const struct htb_class *cl) /** * htb_class_mode - computes and returns current class mode + * @cl: the target class + * @diff: diff time in microseconds * * It computes cl's mode at time cl->t_c+diff and returns it. If mode * is not HTB_CAN_SEND then cl->pq_key is updated to time difference @@ -515,9 +541,12 @@ htb_class_mode(struct htb_class *cl, s64 *diff) /** * htb_change_class_mode - changes classe's mode + * @q: the priority event queue + * @cl: the target class + * @diff: diff time in microseconds * * This should be the only way how to change classe's mode under normal - * cirsumstances. Routine will update feed lists linkage, change mode + * circumstances. Routine will update feed lists linkage, change mode * and add class to the wait event queue if appropriate. New mode should * be different from old one and cl->pq_key has to be valid if changing * to mode other than HTB_CAN_SEND (see htb_add_to_wait_tree). @@ -547,6 +576,8 @@ htb_change_class_mode(struct htb_sched *q, struct htb_class *cl, s64 *diff) /** * htb_activate - inserts leaf cl into appropriate active feeds + * @q: the priority event queue + * @cl: the target class * * Routine learns (new) priority of leaf and activates feed chain * for the prio. It can be called on already active leaf safely. @@ -564,6 +595,8 @@ static inline void htb_activate(struct htb_sched *q, struct htb_class *cl) /** * htb_deactivate - remove leaf cl from active feeds + * @q: the priority event queue + * @cl: the target class * * Make sure that leaf is active. In the other words it can't be called * with non-active leaf. It also removes class from the drop list. @@ -579,7 +612,7 @@ static inline void htb_deactivate(struct htb_sched *q, struct htb_class *cl) static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch, struct sk_buff **to_free) { - int uninitialized_var(ret); + int ret; unsigned int len = qdisc_pkt_len(skb); struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl = htb_classify(skb, sch, &ret); @@ -643,6 +676,10 @@ static inline void htb_accnt_ctokens(struct htb_class *cl, int bytes, s64 diff) /** * htb_charge_class - charges amount "bytes" to leaf and ancestors + * @q: the priority event queue + * @cl: the class to start iterate + * @level: the minimum level to account + * @skb: the socket buffer * * Routine assumes that packet "bytes" long was dequeued from leaf cl * borrowing from "level". It accounts bytes to ceil leaky bucket for @@ -692,6 +729,9 @@ static void htb_charge_class(struct htb_sched *q, struct htb_class *cl, /** * htb_do_events - make mode changes to classes at the level + * @q: the priority event queue + * @level: which wait_pq in 'q->hlevel' + * @start: start jiffies * * Scans event queue for pending events and applies them. Returns time of * next pending event (0 for no event in pq, q->now for too many events). @@ -760,6 +800,8 @@ static struct rb_node *htb_id_find_next_upper(int prio, struct rb_node *n, /** * htb_lookup_leaf - returns next leaf class in DRR order + * @hprio: the current one + * @prio: which prio in class * * Find leaf where current feed pointers points to. */ @@ -957,7 +999,7 @@ static void htb_reset(struct Qdisc *sch) if (cl->level) memset(&cl->inner, 0, sizeof(cl->inner)); else { - if (cl->leaf.q) + if (cl->leaf.q && !q->offload) qdisc_reset(cl->leaf.q); } cl->prio_activity = 0; @@ -966,8 +1008,6 @@ static void htb_reset(struct Qdisc *sch) } qdisc_watchdog_cancel(&q->watchdog); __qdisc_reset_queue(&q->direct_queue); - sch->q.qlen = 0; - sch->qstats.backlog = 0; memset(q->hlevel, 0, sizeof(q->hlevel)); memset(q->row_mask, 0, sizeof(q->row_mask)); } @@ -980,6 +1020,7 @@ static const struct nla_policy htb_policy[TCA_HTB_MAX + 1] = { [TCA_HTB_DIRECT_QLEN] = { .type = NLA_U32 }, [TCA_HTB_RATE64] = { .type = NLA_U64 }, [TCA_HTB_CEIL64] = { .type = NLA_U64 }, + [TCA_HTB_OFFLOAD] = { .type = NLA_FLAG }, }; static void htb_work_func(struct work_struct *work) @@ -992,12 +1033,28 @@ static void htb_work_func(struct work_struct *work) rcu_read_unlock(); } +static void htb_set_lockdep_class_child(struct Qdisc *q) +{ + static struct lock_class_key child_key; + + lockdep_set_class(qdisc_lock(q), &child_key); +} + +static int htb_offload(struct net_device *dev, struct tc_htb_qopt_offload *opt) +{ + return dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_QDISC_HTB, opt); +} + static int htb_init(struct Qdisc *sch, struct nlattr *opt, struct netlink_ext_ack *extack) { + struct net_device *dev = qdisc_dev(sch); + struct tc_htb_qopt_offload offload_opt; struct htb_sched *q = qdisc_priv(sch); struct nlattr *tb[TCA_HTB_MAX + 1]; struct tc_htb_glob *gopt; + unsigned int ntx; + bool offload; int err; qdisc_watchdog_init(&q->watchdog, sch); @@ -1022,12 +1079,31 @@ static int htb_init(struct Qdisc *sch, struct nlattr *opt, if (gopt->version != HTB_VER >> 16) return -EINVAL; + offload = nla_get_flag(tb[TCA_HTB_OFFLOAD]); + + if (offload) { + if (sch->parent != TC_H_ROOT) { + NL_SET_ERR_MSG(extack, "HTB must be the root qdisc to use offload"); + return -EOPNOTSUPP; + } + + if (!tc_can_offload(dev) || !dev->netdev_ops->ndo_setup_tc) { + NL_SET_ERR_MSG(extack, "hw-tc-offload ethtool feature flag must be on"); + return -EOPNOTSUPP; + } + + q->num_direct_qdiscs = dev->real_num_tx_queues; + q->direct_qdiscs = kcalloc(q->num_direct_qdiscs, + sizeof(*q->direct_qdiscs), + GFP_KERNEL); + if (!q->direct_qdiscs) + return -ENOMEM; + } + err = qdisc_class_hash_init(&q->clhash); if (err < 0) return err; - qdisc_skb_head_init(&q->direct_queue); - if (tb[TCA_HTB_DIRECT_QLEN]) q->direct_qlen = nla_get_u32(tb[TCA_HTB_DIRECT_QLEN]); else @@ -1037,15 +1113,105 @@ static int htb_init(struct Qdisc *sch, struct nlattr *opt, q->rate2quantum = 1; q->defcls = gopt->defcls; + if (!offload) + return 0; + + for (ntx = 0; ntx < q->num_direct_qdiscs; ntx++) { + struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, ntx); + struct Qdisc *qdisc; + + qdisc = qdisc_create_dflt(dev_queue, &pfifo_qdisc_ops, + TC_H_MAKE(sch->handle, 0), extack); + if (!qdisc) { + return -ENOMEM; + } + + htb_set_lockdep_class_child(qdisc); + q->direct_qdiscs[ntx] = qdisc; + qdisc->flags |= TCQ_F_ONETXQUEUE | TCQ_F_NOPARENT; + } + + sch->flags |= TCQ_F_MQROOT; + + offload_opt = (struct tc_htb_qopt_offload) { + .command = TC_HTB_CREATE, + .parent_classid = TC_H_MAJ(sch->handle) >> 16, + .classid = TC_H_MIN(q->defcls), + .extack = extack, + }; + err = htb_offload(dev, &offload_opt); + if (err) + return err; + + /* Defer this assignment, so that htb_destroy skips offload-related + * parts (especially calling ndo_setup_tc) on errors. + */ + q->offload = true; + return 0; } +static void htb_attach_offload(struct Qdisc *sch) +{ + struct net_device *dev = qdisc_dev(sch); + struct htb_sched *q = qdisc_priv(sch); + unsigned int ntx; + + for (ntx = 0; ntx < q->num_direct_qdiscs; ntx++) { + struct Qdisc *old, *qdisc = q->direct_qdiscs[ntx]; + + old = dev_graft_qdisc(qdisc->dev_queue, qdisc); + qdisc_put(old); + qdisc_hash_add(qdisc, false); + } + for (ntx = q->num_direct_qdiscs; ntx < dev->num_tx_queues; ntx++) { + struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, ntx); + struct Qdisc *old = dev_graft_qdisc(dev_queue, NULL); + + qdisc_put(old); + } + + kfree(q->direct_qdiscs); + q->direct_qdiscs = NULL; +} + +static void htb_attach_software(struct Qdisc *sch) +{ + struct net_device *dev = qdisc_dev(sch); + unsigned int ntx; + + /* Resemble qdisc_graft behavior. */ + for (ntx = 0; ntx < dev->num_tx_queues; ntx++) { + struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, ntx); + struct Qdisc *old = dev_graft_qdisc(dev_queue, sch); + + qdisc_refcount_inc(sch); + + qdisc_put(old); + } +} + +static void htb_attach(struct Qdisc *sch) +{ + struct htb_sched *q = qdisc_priv(sch); + + if (q->offload) + htb_attach_offload(sch); + else + htb_attach_software(sch); +} + static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) { struct htb_sched *q = qdisc_priv(sch); struct nlattr *nest; struct tc_htb_glob gopt; + if (q->offload) + sch->flags |= TCQ_F_OFFLOADED; + else + sch->flags &= ~TCQ_F_OFFLOADED; + sch->qstats.overlimits = q->overlimits; /* Its safe to not acquire qdisc lock. As we hold RTNL, * no change can happen on the qdisc parameters. @@ -1063,6 +1229,8 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) if (nla_put(skb, TCA_HTB_INIT, sizeof(gopt), &gopt) || nla_put_u32(skb, TCA_HTB_DIRECT_QLEN, q->direct_qlen)) goto nla_put_failure; + if (q->offload && nla_put_flag(skb, TCA_HTB_OFFLOAD)) + goto nla_put_failure; return nla_nest_end(skb, nest); @@ -1075,6 +1243,7 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, struct sk_buff *skb, struct tcmsg *tcm) { struct htb_class *cl = (struct htb_class *)arg; + struct htb_sched *q = qdisc_priv(sch); struct nlattr *nest; struct tc_htb_opt opt; @@ -1101,6 +1270,8 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, opt.level = cl->level; if (nla_put(skb, TCA_HTB_PARMS, sizeof(opt), &opt)) goto nla_put_failure; + if (q->offload && nla_put_flag(skb, TCA_HTB_OFFLOAD)) + goto nla_put_failure; if ((cl->rate.rate_bytes_ps >= (1ULL << 32)) && nla_put_u64_64bit(skb, TCA_HTB_RATE64, cl->rate.rate_bytes_ps, TCA_HTB_PAD)) @@ -1117,10 +1288,41 @@ nla_put_failure: return -1; } +static void htb_offload_aggregate_stats(struct htb_sched *q, + struct htb_class *cl) +{ + u64 bytes = 0, packets = 0; + struct htb_class *c; + unsigned int i; + + gnet_stats_basic_sync_init(&cl->bstats); + + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(c, &q->clhash.hash[i], common.hnode) { + struct htb_class *p = c; + + while (p && p->level < cl->level) + p = p->parent; + + if (p != cl) + continue; + + bytes += u64_stats_read(&c->bstats_bias.bytes); + packets += u64_stats_read(&c->bstats_bias.packets); + if (c->level == 0) { + bytes += u64_stats_read(&c->leaf.q->bstats.bytes); + packets += u64_stats_read(&c->leaf.q->bstats.packets); + } + } + } + _bstats_update(&cl->bstats, bytes, packets); +} + static int htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, struct gnet_dump *d) { struct htb_class *cl = (struct htb_class *)arg; + struct htb_sched *q = qdisc_priv(sch); struct gnet_stats_queue qs = { .drops = cl->drops, .overlimits = cl->overlimits, @@ -1135,8 +1337,21 @@ htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, struct gnet_dump *d) cl->xstats.ctokens = clamp_t(s64, PSCHED_NS2TICKS(cl->ctokens), INT_MIN, INT_MAX); - if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch), - d, NULL, &cl->bstats) < 0 || + if (q->offload) { + if (!cl->level) { + if (cl->leaf.q) + cl->bstats = cl->leaf.q->bstats; + else + gnet_stats_basic_sync_init(&cl->bstats); + _bstats_update(&cl->bstats, + u64_stats_read(&cl->bstats_bias.bytes), + u64_stats_read(&cl->bstats_bias.packets)); + } else { + htb_offload_aggregate_stats(q, cl); + } + } + + if (gnet_stats_copy_basic(d, NULL, &cl->bstats, true) < 0 || gnet_stats_copy_rate_est(d, &cl->rate_est) < 0 || gnet_stats_copy_queue(d, NULL, &qs, qlen) < 0) return -1; @@ -1144,19 +1359,122 @@ htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, struct gnet_dump *d) return gnet_stats_copy_app(d, &cl->xstats, sizeof(cl->xstats)); } +static struct netdev_queue * +htb_select_queue(struct Qdisc *sch, struct tcmsg *tcm) +{ + struct net_device *dev = qdisc_dev(sch); + struct tc_htb_qopt_offload offload_opt; + struct htb_sched *q = qdisc_priv(sch); + int err; + + if (!q->offload) + return sch->dev_queue; + + offload_opt = (struct tc_htb_qopt_offload) { + .command = TC_HTB_LEAF_QUERY_QUEUE, + .classid = TC_H_MIN(tcm->tcm_parent), + }; + err = htb_offload(dev, &offload_opt); + if (err || offload_opt.qid >= dev->num_tx_queues) + return NULL; + return netdev_get_tx_queue(dev, offload_opt.qid); +} + +static struct Qdisc * +htb_graft_helper(struct netdev_queue *dev_queue, struct Qdisc *new_q) +{ + struct net_device *dev = dev_queue->dev; + struct Qdisc *old_q; + + if (dev->flags & IFF_UP) + dev_deactivate(dev); + old_q = dev_graft_qdisc(dev_queue, new_q); + if (new_q) + new_q->flags |= TCQ_F_ONETXQUEUE | TCQ_F_NOPARENT; + if (dev->flags & IFF_UP) + dev_activate(dev); + + return old_q; +} + +static struct netdev_queue *htb_offload_get_queue(struct htb_class *cl) +{ + struct netdev_queue *queue; + + queue = cl->leaf.offload_queue; + if (!(cl->leaf.q->flags & TCQ_F_BUILTIN)) + WARN_ON(cl->leaf.q->dev_queue != queue); + + return queue; +} + +static void htb_offload_move_qdisc(struct Qdisc *sch, struct htb_class *cl_old, + struct htb_class *cl_new, bool destroying) +{ + struct netdev_queue *queue_old, *queue_new; + struct net_device *dev = qdisc_dev(sch); + + queue_old = htb_offload_get_queue(cl_old); + queue_new = htb_offload_get_queue(cl_new); + + if (!destroying) { + struct Qdisc *qdisc; + + if (dev->flags & IFF_UP) + dev_deactivate(dev); + qdisc = dev_graft_qdisc(queue_old, NULL); + WARN_ON(qdisc != cl_old->leaf.q); + } + + if (!(cl_old->leaf.q->flags & TCQ_F_BUILTIN)) + cl_old->leaf.q->dev_queue = queue_new; + cl_old->leaf.offload_queue = queue_new; + + if (!destroying) { + struct Qdisc *qdisc; + + qdisc = dev_graft_qdisc(queue_new, cl_old->leaf.q); + if (dev->flags & IFF_UP) + dev_activate(dev); + WARN_ON(!(qdisc->flags & TCQ_F_BUILTIN)); + } +} + static int htb_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, struct Qdisc **old, struct netlink_ext_ack *extack) { + struct netdev_queue *dev_queue = sch->dev_queue; struct htb_class *cl = (struct htb_class *)arg; + struct htb_sched *q = qdisc_priv(sch); + struct Qdisc *old_q; if (cl->level) return -EINVAL; - if (new == NULL && - (new = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, - cl->common.classid, extack)) == NULL) - return -ENOBUFS; + + if (q->offload) + dev_queue = htb_offload_get_queue(cl); + + if (!new) { + new = qdisc_create_dflt(dev_queue, &pfifo_qdisc_ops, + cl->common.classid, extack); + if (!new) + return -ENOBUFS; + } + + if (q->offload) { + htb_set_lockdep_class_child(new); + /* One ref for cl->leaf.q, the other for dev_queue->qdisc. */ + qdisc_refcount_inc(new); + old_q = htb_graft_helper(dev_queue, new); + } *old = qdisc_replace(sch, new, &cl->leaf.q); + + if (q->offload) { + WARN_ON(old_q != *old); + qdisc_put(old_q); + } + return 0; } @@ -1184,9 +1502,10 @@ static inline int htb_parent_last_child(struct htb_class *cl) return 1; } -static void htb_parent_to_leaf(struct htb_sched *q, struct htb_class *cl, +static void htb_parent_to_leaf(struct Qdisc *sch, struct htb_class *cl, struct Qdisc *new_q) { + struct htb_sched *q = qdisc_priv(sch); struct htb_class *parent = cl->parent; WARN_ON(cl->level || !cl->leaf.q || cl->prio_activity); @@ -1202,6 +1521,79 @@ static void htb_parent_to_leaf(struct htb_sched *q, struct htb_class *cl, parent->ctokens = parent->cbuffer; parent->t_c = ktime_get_ns(); parent->cmode = HTB_CAN_SEND; + if (q->offload) + parent->leaf.offload_queue = cl->leaf.offload_queue; +} + +static void htb_parent_to_leaf_offload(struct Qdisc *sch, + struct netdev_queue *dev_queue, + struct Qdisc *new_q) +{ + struct Qdisc *old_q; + + /* One ref for cl->leaf.q, the other for dev_queue->qdisc. */ + if (new_q) + qdisc_refcount_inc(new_q); + old_q = htb_graft_helper(dev_queue, new_q); + WARN_ON(!(old_q->flags & TCQ_F_BUILTIN)); +} + +static int htb_destroy_class_offload(struct Qdisc *sch, struct htb_class *cl, + bool last_child, bool destroying, + struct netlink_ext_ack *extack) +{ + struct tc_htb_qopt_offload offload_opt; + struct netdev_queue *dev_queue; + struct Qdisc *q = cl->leaf.q; + struct Qdisc *old = NULL; + int err; + + if (cl->level) + return -EINVAL; + + WARN_ON(!q); + dev_queue = htb_offload_get_queue(cl); + old = htb_graft_helper(dev_queue, NULL); + if (destroying) + /* Before HTB is destroyed, the kernel grafts noop_qdisc to + * all queues. + */ + WARN_ON(!(old->flags & TCQ_F_BUILTIN)); + else + WARN_ON(old != q); + + if (cl->parent) { + _bstats_update(&cl->parent->bstats_bias, + u64_stats_read(&q->bstats.bytes), + u64_stats_read(&q->bstats.packets)); + } + + offload_opt = (struct tc_htb_qopt_offload) { + .command = !last_child ? TC_HTB_LEAF_DEL : + destroying ? TC_HTB_LEAF_DEL_LAST_FORCE : + TC_HTB_LEAF_DEL_LAST, + .classid = cl->common.classid, + .extack = extack, + }; + err = htb_offload(qdisc_dev(sch), &offload_opt); + + if (!err || destroying) + qdisc_put(old); + else + htb_graft_helper(dev_queue, old); + + if (last_child) + return err; + + if (!err && offload_opt.classid != TC_H_MIN(cl->common.classid)) { + u32 classid = TC_H_MAJ(sch->handle) | + TC_H_MIN(offload_opt.classid); + struct htb_class *moved_cl = htb_find(classid, sch); + + htb_offload_move_qdisc(sch, moved_cl, cl, destroying); + } + + return err; } static void htb_destroy_class(struct Qdisc *sch, struct htb_class *cl) @@ -1217,8 +1609,11 @@ static void htb_destroy_class(struct Qdisc *sch, struct htb_class *cl) static void htb_destroy(struct Qdisc *sch) { + struct net_device *dev = qdisc_dev(sch); + struct tc_htb_qopt_offload offload_opt; struct htb_sched *q = qdisc_priv(sch); struct hlist_node *next; + bool nonempty, changed; struct htb_class *cl; unsigned int i; @@ -1237,21 +1632,67 @@ static void htb_destroy(struct Qdisc *sch) cl->block = NULL; } } - for (i = 0; i < q->clhash.hashsize; i++) { - hlist_for_each_entry_safe(cl, next, &q->clhash.hash[i], - common.hnode) - htb_destroy_class(sch, cl); - } + + do { + nonempty = false; + changed = false; + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry_safe(cl, next, &q->clhash.hash[i], + common.hnode) { + bool last_child; + + if (!q->offload) { + htb_destroy_class(sch, cl); + continue; + } + + nonempty = true; + + if (cl->level) + continue; + + changed = true; + + last_child = htb_parent_last_child(cl); + htb_destroy_class_offload(sch, cl, last_child, + true, NULL); + qdisc_class_hash_remove(&q->clhash, + &cl->common); + if (cl->parent) + cl->parent->children--; + if (last_child) + htb_parent_to_leaf(sch, cl, NULL); + htb_destroy_class(sch, cl); + } + } + } while (changed); + WARN_ON(nonempty); + qdisc_class_hash_destroy(&q->clhash); __qdisc_reset_queue(&q->direct_queue); + + if (q->offload) { + offload_opt = (struct tc_htb_qopt_offload) { + .command = TC_HTB_DESTROY, + }; + htb_offload(dev, &offload_opt); + } + + if (!q->direct_qdiscs) + return; + for (i = 0; i < q->num_direct_qdiscs && q->direct_qdiscs[i]; i++) + qdisc_put(q->direct_qdiscs[i]); + kfree(q->direct_qdiscs); } -static int htb_delete(struct Qdisc *sch, unsigned long arg) +static int htb_delete(struct Qdisc *sch, unsigned long arg, + struct netlink_ext_ack *extack) { struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl = (struct htb_class *)arg; struct Qdisc *new_q = NULL; int last_child = 0; + int err; /* TODO: why don't allow to delete subtree ? references ? does * tc subsys guarantee us that in htb_destroy it holds no class @@ -1260,11 +1701,30 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) if (cl->children || cl->filter_cnt) return -EBUSY; - if (!cl->level && htb_parent_last_child(cl)) { - new_q = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, + if (!cl->level && htb_parent_last_child(cl)) + last_child = 1; + + if (q->offload) { + err = htb_destroy_class_offload(sch, cl, last_child, false, + extack); + if (err) + return err; + } + + if (last_child) { + struct netdev_queue *dev_queue = sch->dev_queue; + + if (q->offload) + dev_queue = htb_offload_get_queue(cl); + + new_q = qdisc_create_dflt(dev_queue, &pfifo_qdisc_ops, cl->parent->common.classid, NULL); - last_child = 1; + if (q->offload) { + if (new_q) + htb_set_lockdep_class_child(new_q); + htb_parent_to_leaf_offload(sch, dev_queue, new_q); + } } sch_tree_lock(sch); @@ -1285,7 +1745,7 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) &q->hlevel[cl->level].wait_pq); if (last_child) - htb_parent_to_leaf(q, cl, new_q); + htb_parent_to_leaf(sch, cl, new_q); sch_tree_unlock(sch); @@ -1300,9 +1760,11 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, int err = -EINVAL; struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl = (struct htb_class *)*arg, *parent; + struct tc_htb_qopt_offload offload_opt; struct nlattr *opt = tca[TCA_OPTIONS]; struct nlattr *tb[TCA_HTB_MAX + 1]; struct Qdisc *parent_qdisc = NULL; + struct netdev_queue *dev_queue; struct tc_htb_opt *hopt; u64 rate64, ceil64; int warn = 0; @@ -1326,6 +1788,26 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, if (!hopt->rate.rate || !hopt->ceil.rate) goto failure; + if (q->offload) { + /* Options not supported by the offload. */ + if (hopt->rate.overhead || hopt->ceil.overhead) { + NL_SET_ERR_MSG(extack, "HTB offload doesn't support the overhead parameter"); + goto failure; + } + if (hopt->rate.mpu || hopt->ceil.mpu) { + NL_SET_ERR_MSG(extack, "HTB offload doesn't support the mpu parameter"); + goto failure; + } + if (hopt->quantum) { + NL_SET_ERR_MSG(extack, "HTB offload doesn't support the quantum parameter"); + goto failure; + } + if (hopt->prio) { + NL_SET_ERR_MSG(extack, "HTB offload doesn't support the prio parameter"); + goto failure; + } + } + /* Keeping backward compatible with rate_table based iproute2 tc */ if (hopt->rate.linklayer == TC_LINKLAYER_UNAWARE) qdisc_put_rtab(qdisc_get_rtab(&hopt->rate, tb[TCA_HTB_RTAB], @@ -1335,8 +1817,12 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, qdisc_put_rtab(qdisc_get_rtab(&hopt->ceil, tb[TCA_HTB_CTAB], NULL)); + rate64 = tb[TCA_HTB_RATE64] ? nla_get_u64(tb[TCA_HTB_RATE64]) : 0; + ceil64 = tb[TCA_HTB_CEIL64] ? nla_get_u64(tb[TCA_HTB_CEIL64]) : 0; + if (!cl) { /* new class */ - struct Qdisc *new_q; + struct net_device *dev = qdisc_dev(sch); + struct Qdisc *new_q, *old_q; int prio; struct { struct nlattr nla; @@ -1368,6 +1854,9 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, if (!cl) goto failure; + gnet_stats_basic_sync_init(&cl->bstats); + gnet_stats_basic_sync_init(&cl->bstats_bias); + err = tcf_block_get(&cl->block, &cl->filter_list, sch, extack); if (err) { kfree(cl); @@ -1377,13 +1866,10 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, err = gen_new_estimator(&cl->bstats, NULL, &cl->rate_est, NULL, - qdisc_root_sleeping_running(sch), + true, tca[TCA_RATE] ? : &est.nla); - if (err) { - tcf_block_put(cl->block); - kfree(cl); - goto failure; - } + if (err) + goto err_block_put; } cl->children = 0; @@ -1392,12 +1878,77 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, for (prio = 0; prio < TC_HTB_NUMPRIO; prio++) RB_CLEAR_NODE(&cl->node[prio]); + cl->common.classid = classid; + + /* Make sure nothing interrupts us in between of two + * ndo_setup_tc calls. + */ + ASSERT_RTNL(); + /* 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_queue, &pfifo_qdisc_ops, + if (!q->offload) { + dev_queue = sch->dev_queue; + } else if (!(parent && !parent->level)) { + /* Assign a dev_queue to this classid. */ + offload_opt = (struct tc_htb_qopt_offload) { + .command = TC_HTB_LEAF_ALLOC_QUEUE, + .classid = cl->common.classid, + .parent_classid = parent ? + TC_H_MIN(parent->common.classid) : + TC_HTB_CLASSID_ROOT, + .rate = max_t(u64, hopt->rate.rate, rate64), + .ceil = max_t(u64, hopt->ceil.rate, ceil64), + .extack = extack, + }; + err = htb_offload(dev, &offload_opt); + if (err) { + pr_err("htb: TC_HTB_LEAF_ALLOC_QUEUE failed with err = %d\n", + err); + goto err_kill_estimator; + } + dev_queue = netdev_get_tx_queue(dev, offload_opt.qid); + } else { /* First child. */ + dev_queue = htb_offload_get_queue(parent); + old_q = htb_graft_helper(dev_queue, NULL); + WARN_ON(old_q != parent->leaf.q); + offload_opt = (struct tc_htb_qopt_offload) { + .command = TC_HTB_LEAF_TO_INNER, + .classid = cl->common.classid, + .parent_classid = + TC_H_MIN(parent->common.classid), + .rate = max_t(u64, hopt->rate.rate, rate64), + .ceil = max_t(u64, hopt->ceil.rate, ceil64), + .extack = extack, + }; + err = htb_offload(dev, &offload_opt); + if (err) { + pr_err("htb: TC_HTB_LEAF_TO_INNER failed with err = %d\n", + err); + htb_graft_helper(dev_queue, old_q); + goto err_kill_estimator; + } + _bstats_update(&parent->bstats_bias, + u64_stats_read(&old_q->bstats.bytes), + u64_stats_read(&old_q->bstats.packets)); + qdisc_put(old_q); + } + new_q = qdisc_create_dflt(dev_queue, &pfifo_qdisc_ops, classid, NULL); + if (q->offload) { + if (new_q) { + htb_set_lockdep_class_child(new_q); + /* One ref for cl->leaf.q, the other for + * dev_queue->qdisc. + */ + qdisc_refcount_inc(new_q); + } + old_q = htb_graft_helper(dev_queue, new_q); + /* No qdisc_put needed. */ + WARN_ON(!(old_q->flags & TCQ_F_BUILTIN)); + } sch_tree_lock(sch); if (parent && !parent->level) { /* turn parent into inner node */ @@ -1415,10 +1966,12 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, : TC_HTB_MAXDEPTH) - 1; memset(&parent->inner, 0, sizeof(parent->inner)); } + /* leaf (we) needs elementary qdisc */ cl->leaf.q = new_q ? new_q : &noop_qdisc; + if (q->offload) + cl->leaf.offload_queue = dev_queue; - cl->common.classid = classid; cl->parent = parent; /* set class to be in HTB_CAN_SEND state */ @@ -1439,17 +1992,35 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, err = gen_replace_estimator(&cl->bstats, NULL, &cl->rate_est, NULL, - qdisc_root_sleeping_running(sch), + true, tca[TCA_RATE]); if (err) return err; } - sch_tree_lock(sch); - } - rate64 = tb[TCA_HTB_RATE64] ? nla_get_u64(tb[TCA_HTB_RATE64]) : 0; + if (q->offload) { + struct net_device *dev = qdisc_dev(sch); + + offload_opt = (struct tc_htb_qopt_offload) { + .command = TC_HTB_NODE_MODIFY, + .classid = cl->common.classid, + .rate = max_t(u64, hopt->rate.rate, rate64), + .ceil = max_t(u64, hopt->ceil.rate, ceil64), + .extack = extack, + }; + err = htb_offload(dev, &offload_opt); + if (err) + /* Estimator was replaced, and rollback may fail + * as well, so we don't try to recover it, and + * the estimator won't work property with the + * offload anyway, because bstats are updated + * only when the stats are queried. + */ + return err; + } - ceil64 = tb[TCA_HTB_CEIL64] ? nla_get_u64(tb[TCA_HTB_CEIL64]) : 0; + sch_tree_lock(sch); + } psched_ratecfg_precompute(&cl->rate, &hopt->rate, rate64); psched_ratecfg_precompute(&cl->ceil, &hopt->ceil, ceil64); @@ -1492,6 +2063,11 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, *arg = (unsigned long)cl; return 0; +err_kill_estimator: + gen_kill_estimator(&cl->rate_est); +err_block_put: + tcf_block_put(cl->block); + kfree(cl); failure: return err; } @@ -1543,20 +2119,14 @@ static void htb_walk(struct Qdisc *sch, struct qdisc_walker *arg) for (i = 0; i < q->clhash.hashsize; i++) { hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) { - if (arg->count < arg->skip) { - arg->count++; - continue; - } - if (arg->fn(sch, (unsigned long)cl, arg) < 0) { - arg->stop = 1; + if (!tc_qdisc_stats_dump(sch, (unsigned long)cl, arg)) return; - } - arg->count++; } } } static const struct Qdisc_class_ops htb_class_ops = { + .select_queue = htb_select_queue, .graft = htb_graft, .leaf = htb_leaf, .qlen_notify = htb_qlen_notify, @@ -1579,6 +2149,7 @@ static struct Qdisc_ops htb_qdisc_ops __read_mostly = { .dequeue = htb_dequeue, .peek = qdisc_peek_dequeued, .init = htb_init, + .attach = htb_attach, .reset = htb_reset, .destroy = htb_destroy, .dump = htb_dump, |