diff options
Diffstat (limited to 'net/core')
30 files changed, 1490 insertions, 574 deletions
diff --git a/net/core/Makefile b/net/core/Makefile index 6e6548011f..62be9aef25 100644 --- a/net/core/Makefile +++ b/net/core/Makefile @@ -26,6 +26,7 @@ obj-$(CONFIG_NETPOLL) += netpoll.o obj-$(CONFIG_FIB_RULES) += fib_rules.o obj-$(CONFIG_TRACEPOINTS) += net-traces.o obj-$(CONFIG_NET_DROP_MONITOR) += drop_monitor.o +obj-$(CONFIG_NET_IEEE8021Q_HELPERS) += ieee8021q_helpers.o obj-$(CONFIG_NET_SELFTESTS) += selftests.o obj-$(CONFIG_NETWORK_PHY_TIMESTAMPING) += timestamping.o obj-$(CONFIG_NET_PTP_CLASSIFY) += ptp_classifier.o @@ -41,4 +42,4 @@ obj-$(CONFIG_NET_SOCK_MSG) += skmsg.o obj-$(CONFIG_BPF_SYSCALL) += sock_map.o obj-$(CONFIG_BPF_SYSCALL) += bpf_sk_storage.o obj-$(CONFIG_OF) += of_net.o -obj-$(CONFIG_NET_TEST) += gso_test.o +obj-$(CONFIG_NET_TEST) += net_test.o diff --git a/net/core/bpf_sk_storage.c b/net/core/bpf_sk_storage.c index 6c4d90b24d..bc01b3aa6b 100644 --- a/net/core/bpf_sk_storage.c +++ b/net/core/bpf_sk_storage.c @@ -496,27 +496,22 @@ bpf_sk_storage_diag_alloc(const struct nlattr *nla_stgs) if (!bpf_capable()) return ERR_PTR(-EPERM); - nla_for_each_nested(nla, nla_stgs, rem) { - if (nla_type(nla) == SK_DIAG_BPF_STORAGE_REQ_MAP_FD) { - if (nla_len(nla) != sizeof(u32)) - return ERR_PTR(-EINVAL); - nr_maps++; - } + nla_for_each_nested_type(nla, SK_DIAG_BPF_STORAGE_REQ_MAP_FD, + nla_stgs, rem) { + if (nla_len(nla) != sizeof(u32)) + return ERR_PTR(-EINVAL); + nr_maps++; } diag = kzalloc(struct_size(diag, maps, nr_maps), GFP_KERNEL); if (!diag) return ERR_PTR(-ENOMEM); - nla_for_each_nested(nla, nla_stgs, rem) { - struct bpf_map *map; - int map_fd; - - if (nla_type(nla) != SK_DIAG_BPF_STORAGE_REQ_MAP_FD) - continue; + nla_for_each_nested_type(nla, SK_DIAG_BPF_STORAGE_REQ_MAP_FD, + nla_stgs, rem) { + int map_fd = nla_get_u32(nla); + struct bpf_map *map = bpf_map_get(map_fd); - map_fd = nla_get_u32(nla); - map = bpf_map_get(map_fd); if (IS_ERR(map)) { err = PTR_ERR(map); goto err_free; diff --git a/net/core/datagram.c b/net/core/datagram.c index a8b625abe2..e72dd78471 100644 --- a/net/core/datagram.c +++ b/net/core/datagram.c @@ -324,25 +324,6 @@ void skb_free_datagram(struct sock *sk, struct sk_buff *skb) } EXPORT_SYMBOL(skb_free_datagram); -void __skb_free_datagram_locked(struct sock *sk, struct sk_buff *skb, int len) -{ - bool slow; - - if (!skb_unref(skb)) { - sk_peek_offset_bwd(sk, len); - return; - } - - slow = lock_sock_fast(sk); - sk_peek_offset_bwd(sk, len); - skb_orphan(skb); - unlock_sock_fast(sk, slow); - - /* skb is now orphaned, can be freed outside of locked section */ - __kfree_skb(skb); -} -EXPORT_SYMBOL(__skb_free_datagram_locked); - int __sk_queue_drop_skb(struct sock *sk, struct sk_buff_head *sk_queue, struct sk_buff *skb, unsigned int flags, void (*destructor)(struct sock *sk, @@ -435,15 +416,23 @@ static int __skb_datagram_iter(const struct sk_buff *skb, int offset, end = start + skb_frag_size(frag); if ((copy = end - offset) > 0) { - struct page *page = skb_frag_page(frag); - u8 *vaddr = kmap(page); + u32 p_off, p_len, copied; + struct page *p; + u8 *vaddr; if (copy > len) copy = len; - n = INDIRECT_CALL_1(cb, simple_copy_to_iter, - vaddr + skb_frag_off(frag) + offset - start, - copy, data, to); - kunmap(page); + + n = 0; + skb_frag_foreach_page(frag, + skb_frag_off(frag) + offset - start, + copy, p, p_off, p_len, copied) { + vaddr = kmap_local_page(p); + n += INDIRECT_CALL_1(cb, simple_copy_to_iter, + vaddr + p_off, p_len, data, to); + kunmap_local(vaddr); + } + offset += n; if (n != copy) goto short_copy; diff --git a/net/core/dev.c b/net/core/dev.c index e8fb4ef8a8..2b4819b610 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -77,7 +77,9 @@ #include <linux/hash.h> #include <linux/slab.h> #include <linux/sched.h> +#include <linux/sched/isolation.h> #include <linux/sched/mm.h> +#include <linux/smpboot.h> #include <linux/mutex.h> #include <linux/rwsem.h> #include <linux/string.h> @@ -197,35 +199,60 @@ static inline struct hlist_head *dev_index_hash(struct net *net, int ifindex) return &net->dev_index_head[ifindex & (NETDEV_HASHENTRIES - 1)]; } -static inline void rps_lock_irqsave(struct softnet_data *sd, - unsigned long *flags) +#ifndef CONFIG_PREEMPT_RT + +static DEFINE_STATIC_KEY_FALSE(use_backlog_threads_key); + +static int __init setup_backlog_napi_threads(char *arg) +{ + static_branch_enable(&use_backlog_threads_key); + return 0; +} +early_param("thread_backlog_napi", setup_backlog_napi_threads); + +static bool use_backlog_threads(void) +{ + return static_branch_unlikely(&use_backlog_threads_key); +} + +#else + +static bool use_backlog_threads(void) { - if (IS_ENABLED(CONFIG_RPS)) + return true; +} + +#endif + +static inline void backlog_lock_irq_save(struct softnet_data *sd, + unsigned long *flags) +{ + if (IS_ENABLED(CONFIG_RPS) || use_backlog_threads()) spin_lock_irqsave(&sd->input_pkt_queue.lock, *flags); else if (!IS_ENABLED(CONFIG_PREEMPT_RT)) local_irq_save(*flags); } -static inline void rps_lock_irq_disable(struct softnet_data *sd) +static inline void backlog_lock_irq_disable(struct softnet_data *sd) { - if (IS_ENABLED(CONFIG_RPS)) + if (IS_ENABLED(CONFIG_RPS) || use_backlog_threads()) spin_lock_irq(&sd->input_pkt_queue.lock); else if (!IS_ENABLED(CONFIG_PREEMPT_RT)) local_irq_disable(); } -static inline void rps_unlock_irq_restore(struct softnet_data *sd, - unsigned long *flags) +static inline void backlog_unlock_irq_restore(struct softnet_data *sd, + unsigned long *flags) { - if (IS_ENABLED(CONFIG_RPS)) + if (IS_ENABLED(CONFIG_RPS) || use_backlog_threads()) spin_unlock_irqrestore(&sd->input_pkt_queue.lock, *flags); else if (!IS_ENABLED(CONFIG_PREEMPT_RT)) local_irq_restore(*flags); } -static inline void rps_unlock_irq_enable(struct softnet_data *sd) +static inline void backlog_unlock_irq_enable(struct softnet_data *sd) { - if (IS_ENABLED(CONFIG_RPS)) + if (IS_ENABLED(CONFIG_RPS) || use_backlog_threads()) spin_unlock_irq(&sd->input_pkt_queue.lock); else if (!IS_ENABLED(CONFIG_PREEMPT_RT)) local_irq_enable(); @@ -912,6 +939,18 @@ struct net_device *dev_get_by_napi_id(unsigned int napi_id) } EXPORT_SYMBOL(dev_get_by_napi_id); +static DEFINE_SEQLOCK(netdev_rename_lock); + +void netdev_copy_name(struct net_device *dev, char *name) +{ + unsigned int seq; + + do { + seq = read_seqbegin(&netdev_rename_lock); + strscpy(name, dev->name, IFNAMSIZ); + } while (read_seqretry(&netdev_rename_lock, seq)); +} + /** * netdev_get_name - get a netdevice name, knowing its ifindex. * @net: network namespace @@ -923,7 +962,6 @@ int netdev_get_name(struct net *net, char *name, int ifindex) struct net_device *dev; int ret; - down_read(&devnet_rename_sem); rcu_read_lock(); dev = dev_get_by_index_rcu(net, ifindex); @@ -932,12 +970,11 @@ int netdev_get_name(struct net *net, char *name, int ifindex) goto out; } - strcpy(name, dev->name); + netdev_copy_name(dev, name); ret = 0; out: rcu_read_unlock(); - up_read(&devnet_rename_sem); return ret; } @@ -1189,7 +1226,10 @@ int dev_change_name(struct net_device *dev, const char *newname) memcpy(oldname, dev->name, IFNAMSIZ); + write_seqlock_bh(&netdev_rename_lock); err = dev_get_valid_name(net, dev, newname); + write_sequnlock_bh(&netdev_rename_lock); + if (err < 0) { up_write(&devnet_rename_sem); return err; @@ -1229,7 +1269,9 @@ rollback: if (err >= 0) { err = ret; down_write(&devnet_rename_sem); + write_seqlock_bh(&netdev_rename_lock); memcpy(dev->name, oldname, IFNAMSIZ); + write_sequnlock_bh(&netdev_rename_lock); memcpy(oldname, newname, IFNAMSIZ); WRITE_ONCE(dev->name_assign_type, old_assign_type); old_assign_type = NET_NAME_RENAMED; @@ -2057,6 +2099,11 @@ void net_dec_egress_queue(void) EXPORT_SYMBOL_GPL(net_dec_egress_queue); #endif +#ifdef CONFIG_NET_CLS_ACT +DEFINE_STATIC_KEY_FALSE(tcf_bypass_check_needed_key); +EXPORT_SYMBOL(tcf_bypass_check_needed_key); +#endif + DEFINE_STATIC_KEY_FALSE(netstamp_needed_key); EXPORT_SYMBOL(netstamp_needed_key); #ifdef CONFIG_JUMP_LABEL @@ -3917,6 +3964,11 @@ static int tc_run(struct tcx_entry *entry, struct sk_buff *skb, if (!miniq) return ret; + if (static_branch_unlikely(&tcf_bypass_check_needed_key)) { + if (tcf_block_bypass_sw(miniq->block)) + return ret; + } + tc_skb_cb(skb)->mru = 0; tc_skb_cb(skb)->post_ct = false; tcf_set_drop_reason(skb, *drop_reason); @@ -4410,8 +4462,8 @@ EXPORT_SYMBOL(__dev_direct_xmit); /************************************************************************* * Receiver routines *************************************************************************/ +static DEFINE_PER_CPU(struct task_struct *, backlog_napi); -unsigned int sysctl_skb_defer_max __read_mostly = 64; int weight_p __read_mostly = 64; /* old backlog weight */ int dev_weight_rx_bias __read_mostly = 1; /* bias for backlog weight */ int dev_weight_tx_bias __read_mostly = 1; /* bias for output_queue quota */ @@ -4433,18 +4485,16 @@ static inline void ____napi_schedule(struct softnet_data *sd, */ thread = READ_ONCE(napi->thread); if (thread) { - /* Avoid doing set_bit() if the thread is in - * INTERRUPTIBLE state, cause napi_thread_wait() - * makes sure to proceed with napi polling - * if the thread is explicitly woken from here. - */ - if (READ_ONCE(thread->__state) != TASK_INTERRUPTIBLE) - set_bit(NAPI_STATE_SCHED_THREADED, &napi->state); + if (use_backlog_threads() && thread == raw_cpu_read(backlog_napi)) + goto use_local_napi; + + set_bit(NAPI_STATE_SCHED_THREADED, &napi->state); wake_up_process(thread); return; } } +use_local_napi: list_add_tail(&napi->poll_list, &sd->poll_list); WRITE_ONCE(napi->list_owner, smp_processor_id()); /* If not called from net_rx_action() @@ -4466,12 +4516,13 @@ set_rps_cpu(struct net_device *dev, struct sk_buff *skb, struct rps_dev_flow *rflow, u16 next_cpu) { if (next_cpu < nr_cpu_ids) { + u32 head; #ifdef CONFIG_RFS_ACCEL struct netdev_rx_queue *rxqueue; struct rps_dev_flow_table *flow_table; struct rps_dev_flow *old_rflow; - u32 flow_id; u16 rxq_index; + u32 flow_id; int rc; /* Should we steer this flow to a different hardware queue? */ @@ -4493,16 +4544,16 @@ set_rps_cpu(struct net_device *dev, struct sk_buff *skb, goto out; old_rflow = rflow; rflow = &flow_table->flows[flow_id]; - rflow->filter = rc; - if (old_rflow->filter == rflow->filter) - old_rflow->filter = RPS_NO_FILTER; + WRITE_ONCE(rflow->filter, rc); + if (old_rflow->filter == rc) + WRITE_ONCE(old_rflow->filter, RPS_NO_FILTER); out: #endif - rflow->last_qtail = - per_cpu(softnet_data, next_cpu).input_queue_head; + head = READ_ONCE(per_cpu(softnet_data, next_cpu).input_queue_head); + rps_input_queue_tail_save(&rflow->last_qtail, head); } - rflow->cpu = next_cpu; + WRITE_ONCE(rflow->cpu, next_cpu); return rflow; } @@ -4581,7 +4632,7 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb, */ if (unlikely(tcpu != next_cpu) && (tcpu >= nr_cpu_ids || !cpu_online(tcpu) || - ((int)(per_cpu(softnet_data, tcpu).input_queue_head - + ((int)(READ_ONCE(per_cpu(softnet_data, tcpu).input_queue_head) - rflow->last_qtail)) >= 0)) { tcpu = next_cpu; rflow = set_rps_cpu(dev, skb, rflow, next_cpu); @@ -4635,9 +4686,9 @@ bool rps_may_expire_flow(struct net_device *dev, u16 rxq_index, if (flow_table && flow_id <= flow_table->mask) { rflow = &flow_table->flows[flow_id]; cpu = READ_ONCE(rflow->cpu); - if (rflow->filter == filter_id && cpu < nr_cpu_ids && - ((int)(per_cpu(softnet_data, cpu).input_queue_head - - rflow->last_qtail) < + if (READ_ONCE(rflow->filter) == filter_id && cpu < nr_cpu_ids && + ((int)(READ_ONCE(per_cpu(softnet_data, cpu).input_queue_head) - + READ_ONCE(rflow->last_qtail)) < (int)(10 * flow_table->mask))) expire = false; } @@ -4684,6 +4735,11 @@ static void napi_schedule_rps(struct softnet_data *sd) #ifdef CONFIG_RPS if (sd != mysd) { + if (use_backlog_threads()) { + __napi_schedule_irqoff(&sd->backlog); + return; + } + sd->rps_ipi_next = mysd->rps_ipi_list; mysd->rps_ipi_list = sd; @@ -4698,6 +4754,23 @@ static void napi_schedule_rps(struct softnet_data *sd) __napi_schedule_irqoff(&mysd->backlog); } +void kick_defer_list_purge(struct softnet_data *sd, unsigned int cpu) +{ + unsigned long flags; + + if (use_backlog_threads()) { + backlog_lock_irq_save(sd, &flags); + + if (!__test_and_set_bit(NAPI_STATE_SCHED, &sd->backlog.state)) + __napi_schedule_irqoff(&sd->backlog); + + backlog_unlock_irq_restore(sd, &flags); + + } else if (!cmpxchg(&sd->defer_ipi_scheduled, 0, 1)) { + smp_call_function_single_async(cpu, &sd->defer_csd); + } +} + #ifdef CONFIG_NET_FLOW_LIMIT int netdev_flow_limit_table_len __read_mostly = (1 << 12); #endif @@ -4749,37 +4822,45 @@ static int enqueue_to_backlog(struct sk_buff *skb, int cpu, struct softnet_data *sd; unsigned long flags; unsigned int qlen; + int max_backlog; + u32 tail; - reason = SKB_DROP_REASON_NOT_SPECIFIED; + reason = SKB_DROP_REASON_DEV_READY; + if (!netif_running(skb->dev)) + goto bad_dev; + + reason = SKB_DROP_REASON_CPU_BACKLOG; sd = &per_cpu(softnet_data, cpu); - rps_lock_irqsave(sd, &flags); - if (!netif_running(skb->dev)) - goto drop; + qlen = skb_queue_len_lockless(&sd->input_pkt_queue); + max_backlog = READ_ONCE(net_hotdata.max_backlog); + if (unlikely(qlen > max_backlog)) + goto cpu_backlog_drop; + backlog_lock_irq_save(sd, &flags); qlen = skb_queue_len(&sd->input_pkt_queue); - if (qlen <= READ_ONCE(net_hotdata.max_backlog) && - !skb_flow_limit(skb, qlen)) { - if (qlen) { -enqueue: - __skb_queue_tail(&sd->input_pkt_queue, skb); - input_queue_tail_incr_save(sd, qtail); - rps_unlock_irq_restore(sd, &flags); - return NET_RX_SUCCESS; + if (qlen <= max_backlog && !skb_flow_limit(skb, qlen)) { + if (!qlen) { + /* Schedule NAPI for backlog device. We can use + * non atomic operation as we own the queue lock. + */ + if (!__test_and_set_bit(NAPI_STATE_SCHED, + &sd->backlog.state)) + napi_schedule_rps(sd); } + __skb_queue_tail(&sd->input_pkt_queue, skb); + tail = rps_input_queue_tail_incr(sd); + backlog_unlock_irq_restore(sd, &flags); - /* Schedule NAPI for backlog device - * We can use non atomic operation since we own the queue lock - */ - if (!__test_and_set_bit(NAPI_STATE_SCHED, &sd->backlog.state)) - napi_schedule_rps(sd); - goto enqueue; + /* save the tail outside of the critical section */ + rps_input_queue_tail_save(qtail, tail); + return NET_RX_SUCCESS; } - reason = SKB_DROP_REASON_CPU_BACKLOG; -drop: - sd->dropped++; - rps_unlock_irq_restore(sd, &flags); + backlog_unlock_irq_restore(sd, &flags); +cpu_backlog_drop: + atomic_inc(&sd->dropped); +bad_dev: dev_core_stats_rx_dropped_inc(skb->dev); kfree_skb_reason(skb, reason); return NET_RX_DROP; @@ -5844,21 +5925,21 @@ static void flush_backlog(struct work_struct *work) local_bh_disable(); sd = this_cpu_ptr(&softnet_data); - rps_lock_irq_disable(sd); + backlog_lock_irq_disable(sd); skb_queue_walk_safe(&sd->input_pkt_queue, skb, tmp) { if (skb->dev->reg_state == NETREG_UNREGISTERING) { __skb_unlink(skb, &sd->input_pkt_queue); dev_kfree_skb_irq(skb); - input_queue_head_incr(sd); + rps_input_queue_head_incr(sd); } } - rps_unlock_irq_enable(sd); + backlog_unlock_irq_enable(sd); skb_queue_walk_safe(&sd->process_queue, skb, tmp) { if (skb->dev->reg_state == NETREG_UNREGISTERING) { __skb_unlink(skb, &sd->process_queue); kfree_skb(skb); - input_queue_head_incr(sd); + rps_input_queue_head_incr(sd); } } local_bh_enable(); @@ -5870,14 +5951,14 @@ static bool flush_required(int cpu) struct softnet_data *sd = &per_cpu(softnet_data, cpu); bool do_flush; - rps_lock_irq_disable(sd); + backlog_lock_irq_disable(sd); /* as insertion into process_queue happens with the rps lock held, * process_queue access may race only with dequeue */ do_flush = !skb_queue_empty(&sd->input_pkt_queue) || !skb_queue_empty_lockless(&sd->process_queue); - rps_unlock_irq_enable(sd); + backlog_unlock_irq_enable(sd); return do_flush; #endif @@ -5943,7 +6024,7 @@ static void net_rps_action_and_irq_enable(struct softnet_data *sd) #ifdef CONFIG_RPS struct softnet_data *remsd = sd->rps_ipi_list; - if (remsd) { + if (!use_backlog_threads() && remsd) { sd->rps_ipi_list = NULL; local_irq_enable(); @@ -5958,7 +6039,7 @@ static void net_rps_action_and_irq_enable(struct softnet_data *sd) static bool sd_has_rps_ipi_waiting(struct softnet_data *sd) { #ifdef CONFIG_RPS - return sd->rps_ipi_list != NULL; + return !use_backlog_threads() && sd->rps_ipi_list; #else return false; #endif @@ -5986,13 +6067,14 @@ static int process_backlog(struct napi_struct *napi, int quota) rcu_read_lock(); __netif_receive_skb(skb); rcu_read_unlock(); - input_queue_head_incr(sd); - if (++work >= quota) + if (++work >= quota) { + rps_input_queue_head_add(sd, work); return work; + } } - rps_lock_irq_disable(sd); + backlog_lock_irq_disable(sd); if (skb_queue_empty(&sd->input_pkt_queue)) { /* * Inline a custom version of __napi_complete(). @@ -6002,15 +6084,17 @@ static int process_backlog(struct napi_struct *napi, int quota) * We can use a plain write instead of clear_bit(), * and we dont need an smp_mb() memory barrier. */ - napi->state = 0; + napi->state &= NAPIF_STATE_THREADED; again = false; } else { skb_queue_splice_tail_init(&sd->input_pkt_queue, &sd->process_queue); } - rps_unlock_irq_enable(sd); + backlog_unlock_irq_enable(sd); } + if (work) + rps_input_queue_head_add(sd, work); return work; } @@ -6447,7 +6531,7 @@ int dev_set_threaded(struct net_device *dev, bool threaded) } } - dev->threaded = threaded; + WRITE_ONCE(dev->threaded, threaded); /* Make sure kthread is created before THREADED bit * is set. @@ -6538,7 +6622,7 @@ void netif_napi_add_weight(struct net_device *dev, struct napi_struct *napi, * threaded mode will not be enabled in napi_enable(). */ if (dev->threaded && napi_kthread_create(napi)) - dev->threaded = 0; + dev->threaded = false; netif_napi_set_irq(napi, -1); } EXPORT_SYMBOL(netif_napi_add_weight); @@ -6716,8 +6800,6 @@ static int napi_poll(struct napi_struct *n, struct list_head *repoll) static int napi_thread_wait(struct napi_struct *napi) { - bool woken = false; - set_current_state(TASK_INTERRUPTIBLE); while (!kthread_should_stop()) { @@ -6726,15 +6808,13 @@ static int napi_thread_wait(struct napi_struct *napi) * Testing SCHED bit is not enough because SCHED bit might be * set by some other busy poll thread or by napi_disable(). */ - if (test_bit(NAPI_STATE_SCHED_THREADED, &napi->state) || woken) { + if (test_bit(NAPI_STATE_SCHED_THREADED, &napi->state)) { WARN_ON(!list_empty(&napi->poll_list)); __set_current_state(TASK_RUNNING); return 0; } schedule(); - /* woken being true indicates this thread owns this napi. */ - woken = true; set_current_state(TASK_INTERRUPTIBLE); } __set_current_state(TASK_RUNNING); @@ -6742,43 +6822,48 @@ static int napi_thread_wait(struct napi_struct *napi) return -1; } -static int napi_threaded_poll(void *data) +static void napi_threaded_poll_loop(struct napi_struct *napi) { - struct napi_struct *napi = data; struct softnet_data *sd; - void *have; + unsigned long last_qs = jiffies; - while (!napi_thread_wait(napi)) { - unsigned long last_qs = jiffies; - - for (;;) { - bool repoll = false; + for (;;) { + bool repoll = false; + void *have; - local_bh_disable(); - sd = this_cpu_ptr(&softnet_data); - sd->in_napi_threaded_poll = true; + local_bh_disable(); + sd = this_cpu_ptr(&softnet_data); + sd->in_napi_threaded_poll = true; - have = netpoll_poll_lock(napi); - __napi_poll(napi, &repoll); - netpoll_poll_unlock(have); + have = netpoll_poll_lock(napi); + __napi_poll(napi, &repoll); + netpoll_poll_unlock(have); - sd->in_napi_threaded_poll = false; - barrier(); + sd->in_napi_threaded_poll = false; + barrier(); - if (sd_has_rps_ipi_waiting(sd)) { - local_irq_disable(); - net_rps_action_and_irq_enable(sd); - } - skb_defer_free_flush(sd); - local_bh_enable(); + if (sd_has_rps_ipi_waiting(sd)) { + local_irq_disable(); + net_rps_action_and_irq_enable(sd); + } + skb_defer_free_flush(sd); + local_bh_enable(); - if (!repoll) - break; + if (!repoll) + break; - rcu_softirq_qs_periodic(last_qs); - cond_resched(); - } + rcu_softirq_qs_periodic(last_qs); + cond_resched(); } +} + +static int napi_threaded_poll(void *data) +{ + struct napi_struct *napi = data; + + while (!napi_thread_wait(napi)) + napi_threaded_poll_loop(napi); + return 0; } @@ -8459,27 +8544,29 @@ static void dev_change_rx_flags(struct net_device *dev, int flags) static int __dev_set_promiscuity(struct net_device *dev, int inc, bool notify) { unsigned int old_flags = dev->flags; + unsigned int promiscuity, flags; kuid_t uid; kgid_t gid; ASSERT_RTNL(); - dev->flags |= IFF_PROMISC; - dev->promiscuity += inc; - if (dev->promiscuity == 0) { + promiscuity = dev->promiscuity + inc; + if (promiscuity == 0) { /* * Avoid overflow. * If inc causes overflow, untouch promisc and return error. */ - if (inc < 0) - dev->flags &= ~IFF_PROMISC; - else { - dev->promiscuity -= inc; + if (unlikely(inc > 0)) { netdev_warn(dev, "promiscuity touches roof, set promiscuity failed. promiscuity feature of device might be broken.\n"); return -EOVERFLOW; } + flags = old_flags & ~IFF_PROMISC; + } else { + flags = old_flags | IFF_PROMISC; } - if (dev->flags != old_flags) { + WRITE_ONCE(dev->promiscuity, promiscuity); + if (flags != old_flags) { + WRITE_ONCE(dev->flags, flags); netdev_info(dev, "%s promiscuous mode\n", dev->flags & IFF_PROMISC ? "entered" : "left"); if (audit_enabled) { @@ -8530,25 +8617,27 @@ EXPORT_SYMBOL(dev_set_promiscuity); static int __dev_set_allmulti(struct net_device *dev, int inc, bool notify) { unsigned int old_flags = dev->flags, old_gflags = dev->gflags; + unsigned int allmulti, flags; ASSERT_RTNL(); - dev->flags |= IFF_ALLMULTI; - dev->allmulti += inc; - if (dev->allmulti == 0) { + allmulti = dev->allmulti + inc; + if (allmulti == 0) { /* * Avoid overflow. * If inc causes overflow, untouch allmulti and return error. */ - if (inc < 0) - dev->flags &= ~IFF_ALLMULTI; - else { - dev->allmulti -= inc; + if (unlikely(inc > 0)) { netdev_warn(dev, "allmulti touches roof, set allmulti failed. allmulti feature of device might be broken.\n"); return -EOVERFLOW; } + flags = old_flags & ~IFF_ALLMULTI; + } else { + flags = old_flags | IFF_ALLMULTI; } - if (dev->flags ^ old_flags) { + WRITE_ONCE(dev->allmulti, allmulti); + if (flags != old_flags) { + WRITE_ONCE(dev->flags, flags); netdev_info(dev, "%s allmulticast mode\n", dev->flags & IFF_ALLMULTI ? "entered" : "left"); dev_change_rx_flags(dev, IFF_ALLMULTI); @@ -8874,7 +8963,7 @@ int dev_change_tx_queue_len(struct net_device *dev, unsigned long new_len) return -ERANGE; if (new_len != orig_len) { - dev->tx_queue_len = new_len; + WRITE_ONCE(dev->tx_queue_len, new_len); res = call_netdevice_notifiers(NETDEV_CHANGE_TX_QUEUE_LEN, dev); res = notifier_to_errno(res); if (res) @@ -8888,7 +8977,7 @@ int dev_change_tx_queue_len(struct net_device *dev, unsigned long new_len) err_rollback: netdev_err(dev, "refused to change device tx_queue_len\n"); - dev->tx_queue_len = orig_len; + WRITE_ONCE(dev->tx_queue_len, orig_len); return res; } @@ -9134,7 +9223,7 @@ int dev_change_proto_down(struct net_device *dev, bool proto_down) netif_carrier_off(dev); else netif_carrier_on(dev); - dev->proto_down = proto_down; + WRITE_ONCE(dev->proto_down, proto_down); return 0; } @@ -9148,18 +9237,21 @@ int dev_change_proto_down(struct net_device *dev, bool proto_down) void dev_change_proto_down_reason(struct net_device *dev, unsigned long mask, u32 value) { + u32 proto_down_reason; int b; if (!mask) { - dev->proto_down_reason = value; + proto_down_reason = value; } else { + proto_down_reason = dev->proto_down_reason; for_each_set_bit(b, &mask, 32) { if (value & (1 << b)) - dev->proto_down_reason |= BIT(b); + proto_down_reason |= BIT(b); else - dev->proto_down_reason &= ~BIT(b); + proto_down_reason &= ~BIT(b); } } + WRITE_ONCE(dev->proto_down_reason, proto_down_reason); } struct bpf_xdp_link { @@ -10349,25 +10441,12 @@ err_free_name: } EXPORT_SYMBOL(register_netdevice); -/** - * init_dummy_netdev - init a dummy network device for NAPI - * @dev: device to init - * - * This takes a network device structure and initialize the minimum - * amount of fields so it can be used to schedule NAPI polls without - * registering a full blown interface. This is to be used by drivers - * that need to tie several hardware interfaces to a single NAPI - * poll scheduler due to HW limitations. +/* Initialize the core of a dummy net device. + * This is useful if you are calling this function after alloc_netdev(), + * since it does not memset the net_device fields. */ -void init_dummy_netdev(struct net_device *dev) +static void init_dummy_netdev_core(struct net_device *dev) { - /* Clear everything. Note we don't initialize spinlocks - * are they aren't supposed to be taken by any of the - * NAPI code and this dummy netdev is supposed to be - * only ever used for NAPI polls - */ - memset(dev, 0, sizeof(struct net_device)); - /* make sure we BUG if trying to hit standard * register/unregister code path */ @@ -10388,8 +10467,28 @@ void init_dummy_netdev(struct net_device *dev) * its refcount. */ } -EXPORT_SYMBOL_GPL(init_dummy_netdev); +/** + * init_dummy_netdev - init a dummy network device for NAPI + * @dev: device to init + * + * This takes a network device structure and initializes the minimum + * amount of fields so it can be used to schedule NAPI polls without + * registering a full blown interface. This is to be used by drivers + * that need to tie several hardware interfaces to a single NAPI + * poll scheduler due to HW limitations. + */ +void init_dummy_netdev(struct net_device *dev) +{ + /* Clear everything. Note we don't initialize spinlocks + * as they aren't supposed to be taken by any of the + * NAPI code and this dummy netdev is supposed to be + * only ever used for NAPI polls + */ + memset(dev, 0, sizeof(struct net_device)); + init_dummy_netdev_core(dev); +} +EXPORT_SYMBOL_GPL(init_dummy_netdev); /** * register_netdev - register a network device @@ -10988,7 +11087,8 @@ void free_netdev(struct net_device *dev) dev->xdp_bulkq = NULL; /* Compatibility with error handling in drivers */ - if (dev->reg_state == NETREG_UNINITIALIZED) { + if (dev->reg_state == NETREG_UNINITIALIZED || + dev->reg_state == NETREG_DUMMY) { netdev_freemem(dev); return; } @@ -11002,6 +11102,19 @@ void free_netdev(struct net_device *dev) EXPORT_SYMBOL(free_netdev); /** + * alloc_netdev_dummy - Allocate and initialize a dummy net device. + * @sizeof_priv: size of private data to allocate space for + * + * Return: the allocated net_device on success, NULL otherwise + */ +struct net_device *alloc_netdev_dummy(int sizeof_priv) +{ + return alloc_netdev(sizeof_priv, "dummy#", NET_NAME_UNKNOWN, + init_dummy_netdev_core); +} +EXPORT_SYMBOL_GPL(alloc_netdev_dummy); + +/** * synchronize_net - Synchronize with packet receive processing * * Wait for packets currently being received to be done. @@ -11304,8 +11417,12 @@ int __dev_change_net_namespace(struct net_device *dev, struct net *net, dev_net_set(dev, net); dev->ifindex = new_ifindex; - if (new_name[0]) /* Rename the netdev to prepared name */ + if (new_name[0]) { + /* Rename the netdev to prepared name */ + write_seqlock_bh(&netdev_rename_lock); strscpy(dev->name, new_name, IFNAMSIZ); + write_sequnlock_bh(&netdev_rename_lock); + } /* Fixup kobjects */ dev_set_uevent_suppress(&dev->dev, 1); @@ -11380,7 +11497,7 @@ static int dev_cpu_dead(unsigned int oldcpu) list_del_init(&napi->poll_list); if (napi->poll == process_backlog) - napi->state = 0; + napi->state &= NAPIF_STATE_THREADED; else ____napi_schedule(sd, napi); } @@ -11388,21 +11505,23 @@ static int dev_cpu_dead(unsigned int oldcpu) raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_enable(); + if (!use_backlog_threads()) { #ifdef CONFIG_RPS - remsd = oldsd->rps_ipi_list; - oldsd->rps_ipi_list = NULL; + remsd = oldsd->rps_ipi_list; + oldsd->rps_ipi_list = NULL; #endif - /* send out pending IPI's on offline CPU */ - net_rps_send_ipi(remsd); + /* send out pending IPI's on offline CPU */ + net_rps_send_ipi(remsd); + } /* Process offline CPU's input_pkt_queue */ while ((skb = __skb_dequeue(&oldsd->process_queue))) { netif_rx(skb); - input_queue_head_incr(oldsd); + rps_input_queue_head_incr(oldsd); } while ((skb = skb_dequeue(&oldsd->input_pkt_queue))) { netif_rx(skb); - input_queue_head_incr(oldsd); + rps_input_queue_head_incr(oldsd); } return 0; @@ -11719,7 +11838,7 @@ static int net_page_pool_create(int cpuid) struct page_pool_params page_pool_params = { .pool_size = SYSTEM_PERCPU_PAGE_POOL_SIZE, .flags = PP_FLAG_SYSTEM_POOL, - .nid = NUMA_NO_NODE, + .nid = cpu_to_mem(cpuid), }; struct page_pool *pp_ptr; @@ -11732,6 +11851,38 @@ static int net_page_pool_create(int cpuid) return 0; } +static int backlog_napi_should_run(unsigned int cpu) +{ + struct softnet_data *sd = per_cpu_ptr(&softnet_data, cpu); + struct napi_struct *napi = &sd->backlog; + + return test_bit(NAPI_STATE_SCHED_THREADED, &napi->state); +} + +static void run_backlog_napi(unsigned int cpu) +{ + struct softnet_data *sd = per_cpu_ptr(&softnet_data, cpu); + + napi_threaded_poll_loop(&sd->backlog); +} + +static void backlog_napi_setup(unsigned int cpu) +{ + struct softnet_data *sd = per_cpu_ptr(&softnet_data, cpu); + struct napi_struct *napi = &sd->backlog; + + napi->thread = this_cpu_read(backlog_napi); + set_bit(NAPI_STATE_THREADED, &napi->state); +} + +static struct smp_hotplug_thread backlog_threads = { + .store = &backlog_napi, + .thread_should_run = backlog_napi_should_run, + .thread_fn = run_backlog_napi, + .thread_comm = "backlog_napi/%u", + .setup = backlog_napi_setup, +}; + /* * This is called single threaded during boot, so no need * to take the rtnl semaphore. @@ -11783,10 +11934,13 @@ static int __init net_dev_init(void) init_gro_hash(&sd->backlog); sd->backlog.poll = process_backlog; sd->backlog.weight = weight_p; + INIT_LIST_HEAD(&sd->backlog.poll_list); if (net_page_pool_create(i)) goto out; } + if (use_backlog_threads()) + smpboot_register_percpu_thread(&backlog_threads); dev_boot_phase = 0; @@ -11812,6 +11966,10 @@ static int __init net_dev_init(void) NULL, dev_cpu_dead); WARN_ON(rc < 0); rc = 0; + + /* avoid static key IPIs to isolated CPUs */ + if (housekeeping_enabled(HK_TYPE_MISC)) + net_enable_timestamp(); out: if (rc < 0) { for_each_possible_cpu(i) { diff --git a/net/core/dev.h b/net/core/dev.h index 2bcaf8eee5..b7b518bc2b 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -4,11 +4,9 @@ #include <linux/types.h> #include <linux/rwsem.h> +#include <linux/netdevice.h> struct net; -struct net_device; -struct netdev_bpf; -struct netdev_phys_item_id; struct netlink_ext_ack; struct cpumask; @@ -38,7 +36,6 @@ int dev_addr_init(struct net_device *dev); void dev_addr_check(struct net_device *dev); /* sysctls not referred to from outside net/core/ */ -extern unsigned int sysctl_skb_defer_max; extern int netdev_unregister_timeout_secs; extern int weight_p; extern int dev_weight_rx_bias; @@ -150,4 +147,23 @@ static inline void xdp_do_check_flushed(struct napi_struct *napi) { } #endif struct napi_struct *napi_by_id(unsigned int napi_id); +void kick_defer_list_purge(struct softnet_data *sd, unsigned int cpu); + +#define XMIT_RECURSION_LIMIT 8 +static inline bool dev_xmit_recursion(void) +{ + return unlikely(__this_cpu_read(softnet_data.xmit.recursion) > + XMIT_RECURSION_LIMIT); +} + +static inline void dev_xmit_recursion_inc(void) +{ + __this_cpu_inc(softnet_data.xmit.recursion); +} + +static inline void dev_xmit_recursion_dec(void) +{ + __this_cpu_dec(softnet_data.xmit.recursion); +} + #endif diff --git a/net/core/dev_addr_lists_test.c b/net/core/dev_addr_lists_test.c index 4dbd0dc6ae..8e1dba825e 100644 --- a/net/core/dev_addr_lists_test.c +++ b/net/core/dev_addr_lists_test.c @@ -49,7 +49,6 @@ static int dev_addr_test_init(struct kunit *test) KUNIT_FAIL(test, "Can't register netdev %d", err); } - rtnl_lock(); return 0; } @@ -57,7 +56,6 @@ static void dev_addr_test_exit(struct kunit *test) { struct net_device *netdev = test->priv; - rtnl_unlock(); unregister_netdev(netdev); free_netdev(netdev); } @@ -67,6 +65,7 @@ static void dev_addr_test_basic(struct kunit *test) struct net_device *netdev = test->priv; u8 addr[ETH_ALEN]; + rtnl_lock(); KUNIT_EXPECT_TRUE(test, !!netdev->dev_addr); memset(addr, 2, sizeof(addr)); @@ -76,6 +75,7 @@ static void dev_addr_test_basic(struct kunit *test) memset(addr, 3, sizeof(addr)); dev_addr_set(netdev, addr); KUNIT_EXPECT_MEMEQ(test, netdev->dev_addr, addr, sizeof(addr)); + rtnl_unlock(); } static void dev_addr_test_sync_one(struct kunit *test) @@ -86,6 +86,7 @@ static void dev_addr_test_sync_one(struct kunit *test) datp = netdev_priv(netdev); + rtnl_lock(); memset(addr, 1, sizeof(addr)); eth_hw_addr_set(netdev, addr); @@ -103,6 +104,7 @@ static void dev_addr_test_sync_one(struct kunit *test) * considered synced and we overwrite in place. */ KUNIT_EXPECT_EQ(test, 0, datp->addr_seen); + rtnl_unlock(); } static void dev_addr_test_add_del(struct kunit *test) @@ -114,6 +116,7 @@ static void dev_addr_test_add_del(struct kunit *test) datp = netdev_priv(netdev); + rtnl_lock(); for (i = 1; i < 4; i++) { memset(addr, i, sizeof(addr)); KUNIT_EXPECT_EQ(test, 0, dev_addr_add(netdev, addr, @@ -143,6 +146,7 @@ static void dev_addr_test_add_del(struct kunit *test) __hw_addr_sync_dev(&netdev->dev_addrs, netdev, dev_addr_test_sync, dev_addr_test_unsync); KUNIT_EXPECT_EQ(test, 1, datp->addr_seen); + rtnl_unlock(); } static void dev_addr_test_del_main(struct kunit *test) @@ -150,6 +154,7 @@ static void dev_addr_test_del_main(struct kunit *test) struct net_device *netdev = test->priv; u8 addr[ETH_ALEN]; + rtnl_lock(); memset(addr, 1, sizeof(addr)); eth_hw_addr_set(netdev, addr); @@ -161,6 +166,7 @@ static void dev_addr_test_del_main(struct kunit *test) NETDEV_HW_ADDR_T_LAN)); KUNIT_EXPECT_EQ(test, -ENOENT, dev_addr_del(netdev, addr, NETDEV_HW_ADDR_T_LAN)); + rtnl_unlock(); } static void dev_addr_test_add_set(struct kunit *test) @@ -172,6 +178,7 @@ static void dev_addr_test_add_set(struct kunit *test) datp = netdev_priv(netdev); + rtnl_lock(); /* There is no external API like dev_addr_add_excl(), * so shuffle the tree a little bit and exploit aliasing. */ @@ -191,6 +198,7 @@ static void dev_addr_test_add_set(struct kunit *test) __hw_addr_sync_dev(&netdev->dev_addrs, netdev, dev_addr_test_sync, dev_addr_test_unsync); KUNIT_EXPECT_EQ(test, 0xffff, datp->addr_seen); + rtnl_unlock(); } static void dev_addr_test_add_excl(struct kunit *test) @@ -199,6 +207,7 @@ static void dev_addr_test_add_excl(struct kunit *test) u8 addr[ETH_ALEN]; int i; + rtnl_lock(); for (i = 0; i < 10; i++) { memset(addr, i, sizeof(addr)); KUNIT_EXPECT_EQ(test, 0, dev_uc_add_excl(netdev, addr)); @@ -213,6 +222,7 @@ static void dev_addr_test_add_excl(struct kunit *test) memset(addr, i, sizeof(addr)); KUNIT_EXPECT_EQ(test, -EEXIST, dev_uc_add_excl(netdev, addr)); } + rtnl_unlock(); } static struct kunit_case dev_addr_test_cases[] = { diff --git a/net/core/dst_cache.c b/net/core/dst_cache.c index 0c0bdb058c..70c634b9e7 100644 --- a/net/core/dst_cache.c +++ b/net/core/dst_cache.c @@ -27,6 +27,7 @@ struct dst_cache_pcpu { static void dst_cache_per_cpu_dst_set(struct dst_cache_pcpu *dst_cache, struct dst_entry *dst, u32 cookie) { + DEBUG_NET_WARN_ON_ONCE(!in_softirq()); dst_release(dst_cache->dst); if (dst) dst_hold(dst); @@ -40,6 +41,7 @@ static struct dst_entry *dst_cache_per_cpu_get(struct dst_cache *dst_cache, { struct dst_entry *dst; + DEBUG_NET_WARN_ON_ONCE(!in_softirq()); dst = idst->dst; if (!dst) goto fail; @@ -47,7 +49,8 @@ static struct dst_entry *dst_cache_per_cpu_get(struct dst_cache *dst_cache, /* the cache already hold a dst reference; it can't go away */ dst_hold(dst); - if (unlikely(!time_after(idst->refresh_ts, dst_cache->reset_ts) || + if (unlikely(!time_after(idst->refresh_ts, + READ_ONCE(dst_cache->reset_ts)) || (dst->obsolete && !dst->ops->check(dst, idst->cookie)))) { dst_cache_per_cpu_dst_set(idst, NULL, 0); dst_release(dst); @@ -111,7 +114,7 @@ void dst_cache_set_ip6(struct dst_cache *dst_cache, struct dst_entry *dst, return; idst = this_cpu_ptr(dst_cache->cache); - dst_cache_per_cpu_dst_set(this_cpu_ptr(dst_cache->cache), dst, + dst_cache_per_cpu_dst_set(idst, dst, rt6_get_cookie(dst_rt6_info(dst))); idst->in6_saddr = *saddr; } @@ -170,7 +173,7 @@ void dst_cache_reset_now(struct dst_cache *dst_cache) if (!dst_cache->cache) return; - dst_cache->reset_ts = jiffies; + dst_cache_reset(dst_cache); for_each_possible_cpu(i) { struct dst_cache_pcpu *idst = per_cpu_ptr(dst_cache->cache, i); struct dst_entry *dst = idst->dst; diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c index 3f933ffcef..6ebffbc632 100644 --- a/net/core/fib_rules.c +++ b/net/core/fib_rules.c @@ -1142,10 +1142,10 @@ static int fib_nl_dumprule(struct sk_buff *skb, struct netlink_callback *cb) const struct nlmsghdr *nlh = cb->nlh; struct net *net = sock_net(skb->sk); struct fib_rules_ops *ops; - int idx = 0, family; + int err, idx = 0, family; if (cb->strict_check) { - int err = fib_valid_dumprule_req(nlh, cb->extack); + err = fib_valid_dumprule_req(nlh, cb->extack); if (err < 0) return err; @@ -1158,17 +1158,17 @@ static int fib_nl_dumprule(struct sk_buff *skb, struct netlink_callback *cb) if (ops == NULL) return -EAFNOSUPPORT; - dump_rules(skb, cb, ops); - - return skb->len; + return dump_rules(skb, cb, ops); } + err = 0; rcu_read_lock(); list_for_each_entry_rcu(ops, &net->rules_ops, list) { if (idx < cb->args[0] || !try_module_get(ops->owner)) goto skip; - if (dump_rules(skb, cb, ops) < 0) + err = dump_rules(skb, cb, ops); + if (err < 0) break; cb->args[1] = 0; @@ -1178,7 +1178,7 @@ skip: rcu_read_unlock(); cb->args[0] = idx; - return skb->len; + return err; } static void notify_rule_change(int event, struct fib_rule *rule, @@ -1293,7 +1293,8 @@ static int __init fib_rules_init(void) int err; rtnl_register(PF_UNSPEC, RTM_NEWRULE, fib_nl_newrule, NULL, 0); rtnl_register(PF_UNSPEC, RTM_DELRULE, fib_nl_delrule, NULL, 0); - rtnl_register(PF_UNSPEC, RTM_GETRULE, NULL, fib_nl_dumprule, 0); + rtnl_register(PF_UNSPEC, RTM_GETRULE, NULL, fib_nl_dumprule, + RTNL_FLAG_DUMP_UNLOCKED); err = register_pernet_subsys(&fib_rules_net_ops); if (err < 0) diff --git a/net/core/filter.c b/net/core/filter.c index ce255e0a2f..ab0455c64e 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -87,6 +87,9 @@ #include "dev.h" +/* Keep the struct bpf_fib_lookup small so that it fits into a cacheline */ +static_assert(sizeof(struct bpf_fib_lookup) == 64, "struct bpf_fib_lookup size check"); + static const struct bpf_func_proto * bpf_sk_base_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog); @@ -2276,12 +2279,12 @@ static int __bpf_redirect_neigh_v6(struct sk_buff *skb, struct net_device *dev, err = bpf_out_neigh_v6(net, skb, dev, nh); if (unlikely(net_xmit_eval(err))) - dev->stats.tx_errors++; + DEV_STATS_INC(dev, tx_errors); else ret = NET_XMIT_SUCCESS; goto out_xmit; out_drop: - dev->stats.tx_errors++; + DEV_STATS_INC(dev, tx_errors); kfree_skb(skb); out_xmit: return ret; @@ -2382,12 +2385,12 @@ static int __bpf_redirect_neigh_v4(struct sk_buff *skb, struct net_device *dev, err = bpf_out_neigh_v4(net, skb, dev, nh); if (unlikely(net_xmit_eval(err))) - dev->stats.tx_errors++; + DEV_STATS_INC(dev, tx_errors); else ret = NET_XMIT_SUCCESS; goto out_xmit; out_drop: - dev->stats.tx_errors++; + DEV_STATS_INC(dev, tx_errors); kfree_skb(skb); out_xmit: return ret; @@ -3541,13 +3544,20 @@ static int bpf_skb_net_grow(struct sk_buff *skb, u32 off, u32 len_diff, if (skb_is_gso(skb)) { struct skb_shared_info *shinfo = skb_shinfo(skb); - /* Due to header grow, MSS needs to be downgraded. */ - if (!(flags & BPF_F_ADJ_ROOM_FIXED_GSO)) - skb_decrease_gso_size(shinfo, len_diff); - /* Header must be checked, and gso_segs recomputed. */ shinfo->gso_type |= gso_type; shinfo->gso_segs = 0; + + /* Due to header growth, MSS needs to be downgraded. + * There is a BUG_ON() when segmenting the frag_list with + * head_frag true, so linearize the skb after downgrading + * the MSS. + */ + if (!(flags & BPF_F_ADJ_ROOM_FIXED_GSO)) { + skb_decrease_gso_size(shinfo, len_diff); + if (shinfo->frag_list) + return skb_linearize(skb); + } } return 0; @@ -4688,7 +4698,7 @@ set_compat: to->tunnel_tos = info->key.tos; to->tunnel_ttl = info->key.ttl; if (flags & BPF_F_TUNINFO_FLAGS) - to->tunnel_flags = info->key.tun_flags; + to->tunnel_flags = ip_tunnel_flags_to_be16(info->key.tun_flags); else to->tunnel_ext = 0; @@ -4731,7 +4741,7 @@ BPF_CALL_3(bpf_skb_get_tunnel_opt, struct sk_buff *, skb, u8 *, to, u32, size) int err; if (unlikely(!info || - !(info->key.tun_flags & TUNNEL_OPTIONS_PRESENT))) { + !ip_tunnel_is_options_present(info->key.tun_flags))) { err = -ENOENT; goto err_clear; } @@ -4801,15 +4811,15 @@ BPF_CALL_4(bpf_skb_set_tunnel_key, struct sk_buff *, skb, memset(info, 0, sizeof(*info)); info->mode = IP_TUNNEL_INFO_TX; - info->key.tun_flags = TUNNEL_KEY | TUNNEL_CSUM | TUNNEL_NOCACHE; - if (flags & BPF_F_DONT_FRAGMENT) - info->key.tun_flags |= TUNNEL_DONT_FRAGMENT; - if (flags & BPF_F_ZERO_CSUM_TX) - info->key.tun_flags &= ~TUNNEL_CSUM; - if (flags & BPF_F_SEQ_NUMBER) - info->key.tun_flags |= TUNNEL_SEQ; - if (flags & BPF_F_NO_TUNNEL_KEY) - info->key.tun_flags &= ~TUNNEL_KEY; + __set_bit(IP_TUNNEL_NOCACHE_BIT, info->key.tun_flags); + __assign_bit(IP_TUNNEL_DONT_FRAGMENT_BIT, info->key.tun_flags, + flags & BPF_F_DONT_FRAGMENT); + __assign_bit(IP_TUNNEL_CSUM_BIT, info->key.tun_flags, + !(flags & BPF_F_ZERO_CSUM_TX)); + __assign_bit(IP_TUNNEL_SEQ_BIT, info->key.tun_flags, + flags & BPF_F_SEQ_NUMBER); + __assign_bit(IP_TUNNEL_KEY_BIT, info->key.tun_flags, + !(flags & BPF_F_NO_TUNNEL_KEY)); info->key.tun_id = cpu_to_be64(from->tunnel_id); info->key.tos = from->tunnel_tos; @@ -4847,13 +4857,15 @@ BPF_CALL_3(bpf_skb_set_tunnel_opt, struct sk_buff *, skb, { struct ip_tunnel_info *info = skb_tunnel_info(skb); const struct metadata_dst *md = this_cpu_ptr(md_dst); + IP_TUNNEL_DECLARE_FLAGS(present) = { }; if (unlikely(info != &md->u.tun_info || (size & (sizeof(u32) - 1)))) return -EINVAL; if (unlikely(size > IP_TUNNEL_OPTS_MAX)) return -ENOMEM; - ip_tunnel_info_opts_set(info, from, size, TUNNEL_OPTIONS_PRESENT); + ip_tunnel_set_options_present(present); + ip_tunnel_info_opts_set(info, from, size, present); return 0; } @@ -5910,7 +5922,10 @@ static int bpf_ipv4_fib_lookup(struct net *net, struct bpf_fib_lookup *params, err = fib_table_lookup(tb, &fl4, &res, FIB_LOOKUP_NOREF); } else { - fl4.flowi4_mark = 0; + if (flags & BPF_FIB_LOOKUP_MARK) + fl4.flowi4_mark = params->mark; + else + fl4.flowi4_mark = 0; fl4.flowi4_secid = 0; fl4.flowi4_tun_key.tun_id = 0; fl4.flowi4_uid = sock_net_uid(net, NULL); @@ -6053,7 +6068,10 @@ static int bpf_ipv6_fib_lookup(struct net *net, struct bpf_fib_lookup *params, err = ipv6_stub->fib6_table_lookup(net, tb, oif, &fl6, &res, strict); } else { - fl6.flowi6_mark = 0; + if (flags & BPF_FIB_LOOKUP_MARK) + fl6.flowi6_mark = params->mark; + else + fl6.flowi6_mark = 0; fl6.flowi6_secid = 0; fl6.flowi6_tun_key.tun_id = 0; fl6.flowi6_uid = sock_net_uid(net, NULL); @@ -6131,7 +6149,7 @@ set_fwd_params: #define BPF_FIB_LOOKUP_MASK (BPF_FIB_LOOKUP_DIRECT | BPF_FIB_LOOKUP_OUTPUT | \ BPF_FIB_LOOKUP_SKIP_NEIGH | BPF_FIB_LOOKUP_TBID | \ - BPF_FIB_LOOKUP_SRC) + BPF_FIB_LOOKUP_SRC | BPF_FIB_LOOKUP_MARK) BPF_CALL_4(bpf_xdp_fib_lookup, struct xdp_buff *, ctx, struct bpf_fib_lookup *, params, int, plen, u32, flags) @@ -8368,8 +8386,6 @@ sk_msg_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog) return &bpf_event_output_data_proto; case BPF_FUNC_get_current_uid_gid: return &bpf_get_current_uid_gid_proto; - case BPF_FUNC_get_current_pid_tgid: - return &bpf_get_current_pid_tgid_proto; case BPF_FUNC_sk_storage_get: return &bpf_sk_storage_get_proto; case BPF_FUNC_sk_storage_delete: diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c index 272f092513..7b54f44f53 100644 --- a/net/core/flow_dissector.c +++ b/net/core/flow_dissector.c @@ -455,17 +455,25 @@ skb_flow_dissect_tunnel_info(const struct sk_buff *skb, if (dissector_uses_key(flow_dissector, FLOW_DISSECTOR_KEY_ENC_OPTS)) { struct flow_dissector_key_enc_opts *enc_opt; + IP_TUNNEL_DECLARE_FLAGS(flags) = { }; + u32 val; enc_opt = skb_flow_dissector_target(flow_dissector, FLOW_DISSECTOR_KEY_ENC_OPTS, target_container); - if (info->options_len) { - enc_opt->len = info->options_len; - ip_tunnel_info_opts_get(enc_opt->data, info); - enc_opt->dst_opt_type = info->key.tun_flags & - TUNNEL_OPTIONS_PRESENT; - } + if (!info->options_len) + return; + + enc_opt->len = info->options_len; + ip_tunnel_info_opts_get(enc_opt->data, info); + + ip_tunnel_set_options_present(flags); + ip_tunnel_flags_and(flags, info->key.tun_flags, flags); + + val = find_next_bit(flags, __IP_TUNNEL_FLAG_NUM, + IP_TUNNEL_GENEVE_OPT_BIT); + enc_opt->dst_opt_type = val < __IP_TUNNEL_FLAG_NUM ? val : 0; } } EXPORT_SYMBOL(skb_flow_dissect_tunnel_info); @@ -1093,7 +1101,7 @@ bool __skb_flow_dissect(const struct net *net, } } - WARN_ON_ONCE(!net); + DEBUG_NET_WARN_ON_ONCE(!net); if (net) { enum netns_bpf_attach_type type = NETNS_BPF_FLOW_DISSECTOR; struct bpf_prog_array *run_array; diff --git a/net/core/gro.c b/net/core/gro.c index c7901253a1..b3b43de1a6 100644 --- a/net/core/gro.c +++ b/net/core/gro.c @@ -3,6 +3,7 @@ #include <net/dst_metadata.h> #include <net/busy_poll.h> #include <trace/events/net.h> +#include <linux/skbuff_ref.h> #define MAX_GRO_SKBS 8 @@ -230,6 +231,33 @@ done: return 0; } +int skb_gro_receive_list(struct sk_buff *p, struct sk_buff *skb) +{ + if (unlikely(p->len + skb->len >= 65536)) + return -E2BIG; + + if (NAPI_GRO_CB(p)->last == p) + skb_shinfo(p)->frag_list = skb; + else + NAPI_GRO_CB(p)->last->next = skb; + + skb_pull(skb, skb_gro_offset(skb)); + + NAPI_GRO_CB(p)->last = skb; + NAPI_GRO_CB(p)->count++; + p->data_len += skb->len; + + /* sk ownership - if any - completely transferred to the aggregated packet */ + skb->destructor = NULL; + skb->sk = NULL; + p->truesize += skb->truesize; + p->len += skb->len; + + NAPI_GRO_CB(skb)->same_flow = 1; + + return 0; +} + static void napi_gro_complete(struct napi_struct *napi, struct sk_buff *skb) { @@ -330,8 +358,6 @@ static void gro_list_prepare(const struct list_head *head, list_for_each_entry(p, head, list) { unsigned long diffs; - NAPI_GRO_CB(p)->flush = 0; - if (hash != skb_get_hash_raw(p)) { NAPI_GRO_CB(p)->same_flow = 0; continue; @@ -471,7 +497,6 @@ found_ptype: sizeof(u32))); /* Avoid slow unaligned acc */ *(u32 *)&NAPI_GRO_CB(skb)->zeroed = 0; NAPI_GRO_CB(skb)->flush = skb_has_frag_list(skb); - NAPI_GRO_CB(skb)->is_atomic = 1; NAPI_GRO_CB(skb)->count = 1; if (unlikely(skb_is_gso(skb))) { NAPI_GRO_CB(skb)->count = skb_shinfo(skb)->gso_segs; diff --git a/net/core/hotdata.c b/net/core/hotdata.c index c8a7a451c1..d0aaaaa556 100644 --- a/net/core/hotdata.c +++ b/net/core/hotdata.c @@ -1,9 +1,9 @@ // SPDX-License-Identifier: GPL-2.0-or-later -#include <net/hotdata.h> #include <linux/cache.h> #include <linux/jiffies.h> #include <linux/list.h> - +#include <net/hotdata.h> +#include <net/proto_memory.h> struct net_hotdata net_hotdata __cacheline_aligned = { .offload_base = LIST_HEAD_INIT(net_hotdata.offload_base), @@ -18,5 +18,8 @@ struct net_hotdata net_hotdata __cacheline_aligned = { .max_backlog = 1000, .dev_tx_weight = 64, .dev_rx_weight = 64, + .sysctl_max_skb_frags = MAX_SKB_FRAGS, + .sysctl_skb_defer_max = 64, + .sysctl_mem_pcpu_rsv = SK_MEMORY_PCPU_RESERVE }; EXPORT_SYMBOL(net_hotdata); diff --git a/net/core/ieee8021q_helpers.c b/net/core/ieee8021q_helpers.c new file mode 100644 index 0000000000..759a9b9f3f --- /dev/null +++ b/net/core/ieee8021q_helpers.c @@ -0,0 +1,242 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2024 Pengutronix, Oleksij Rempel <kernel@pengutronix.de> + +#include <linux/array_size.h> +#include <linux/printk.h> +#include <linux/types.h> +#include <net/dscp.h> +#include <net/ieee8021q.h> + +/* The following arrays map Traffic Types (TT) to traffic classes (TC) for + * different number of queues as shown in the example provided by + * IEEE 802.1Q-2022 in Annex I "I.3 Traffic type to traffic class mapping" and + * Table I-1 "Traffic type to traffic class mapping". + */ +static const u8 ieee8021q_8queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, + [IEEE8021Q_TT_BE] = 1, + [IEEE8021Q_TT_EE] = 2, + [IEEE8021Q_TT_CA] = 3, + [IEEE8021Q_TT_VI] = 4, + [IEEE8021Q_TT_VO] = 5, + [IEEE8021Q_TT_IC] = 6, + [IEEE8021Q_TT_NC] = 7, +}; + +static const u8 ieee8021q_7queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, + [IEEE8021Q_TT_BE] = 1, + [IEEE8021Q_TT_EE] = 2, + [IEEE8021Q_TT_CA] = 3, + [IEEE8021Q_TT_VI] = 4, [IEEE8021Q_TT_VO] = 4, + [IEEE8021Q_TT_IC] = 5, + [IEEE8021Q_TT_NC] = 6, +}; + +static const u8 ieee8021q_6queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, + [IEEE8021Q_TT_BE] = 1, + [IEEE8021Q_TT_EE] = 2, [IEEE8021Q_TT_CA] = 2, + [IEEE8021Q_TT_VI] = 3, [IEEE8021Q_TT_VO] = 3, + [IEEE8021Q_TT_IC] = 4, + [IEEE8021Q_TT_NC] = 5, +}; + +static const u8 ieee8021q_5queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, [IEEE8021Q_TT_BE] = 0, + [IEEE8021Q_TT_EE] = 1, [IEEE8021Q_TT_CA] = 1, + [IEEE8021Q_TT_VI] = 2, [IEEE8021Q_TT_VO] = 2, + [IEEE8021Q_TT_IC] = 3, + [IEEE8021Q_TT_NC] = 4, +}; + +static const u8 ieee8021q_4queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, [IEEE8021Q_TT_BE] = 0, + [IEEE8021Q_TT_EE] = 1, [IEEE8021Q_TT_CA] = 1, + [IEEE8021Q_TT_VI] = 2, [IEEE8021Q_TT_VO] = 2, + [IEEE8021Q_TT_IC] = 3, [IEEE8021Q_TT_NC] = 3, +}; + +static const u8 ieee8021q_3queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, [IEEE8021Q_TT_BE] = 0, + [IEEE8021Q_TT_EE] = 0, [IEEE8021Q_TT_CA] = 0, + [IEEE8021Q_TT_VI] = 1, [IEEE8021Q_TT_VO] = 1, + [IEEE8021Q_TT_IC] = 2, [IEEE8021Q_TT_NC] = 2, +}; + +static const u8 ieee8021q_2queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, [IEEE8021Q_TT_BE] = 0, + [IEEE8021Q_TT_EE] = 0, [IEEE8021Q_TT_CA] = 0, + [IEEE8021Q_TT_VI] = 1, [IEEE8021Q_TT_VO] = 1, + [IEEE8021Q_TT_IC] = 1, [IEEE8021Q_TT_NC] = 1, +}; + +static const u8 ieee8021q_1queue_tt_tc_map[] = { + [IEEE8021Q_TT_BK] = 0, [IEEE8021Q_TT_BE] = 0, + [IEEE8021Q_TT_EE] = 0, [IEEE8021Q_TT_CA] = 0, + [IEEE8021Q_TT_VI] = 0, [IEEE8021Q_TT_VO] = 0, + [IEEE8021Q_TT_IC] = 0, [IEEE8021Q_TT_NC] = 0, +}; + +/** + * ieee8021q_tt_to_tc - Map IEEE 802.1Q Traffic Type to Traffic Class + * @tt: IEEE 802.1Q Traffic Type + * @num_queues: Number of queues + * + * This function maps an IEEE 802.1Q Traffic Type to a Traffic Class (TC) based + * on the number of queues configured on the NIC. The mapping is based on the + * example provided by IEEE 802.1Q-2022 in Annex I "I.3 Traffic type to traffic + * class mapping" and Table I-1 "Traffic type to traffic class mapping". + * + * Return: Traffic Class corresponding to the given Traffic Type or negative + * value in case of error. + */ +int ieee8021q_tt_to_tc(enum ieee8021q_traffic_type tt, unsigned int num_queues) +{ + if (tt < 0 || tt >= IEEE8021Q_TT_MAX) { + pr_err("Requested Traffic Type (%d) is out of range (%d)\n", tt, + IEEE8021Q_TT_MAX); + return -EINVAL; + } + + switch (num_queues) { + case 8: + compiletime_assert(ARRAY_SIZE(ieee8021q_8queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_8queue_tt_tc_map != max - 1"); + return ieee8021q_8queue_tt_tc_map[tt]; + case 7: + compiletime_assert(ARRAY_SIZE(ieee8021q_7queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_7queue_tt_tc_map != max - 1"); + + return ieee8021q_7queue_tt_tc_map[tt]; + case 6: + compiletime_assert(ARRAY_SIZE(ieee8021q_6queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_6queue_tt_tc_map != max - 1"); + + return ieee8021q_6queue_tt_tc_map[tt]; + case 5: + compiletime_assert(ARRAY_SIZE(ieee8021q_5queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_5queue_tt_tc_map != max - 1"); + + return ieee8021q_5queue_tt_tc_map[tt]; + case 4: + compiletime_assert(ARRAY_SIZE(ieee8021q_4queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_4queue_tt_tc_map != max - 1"); + + return ieee8021q_4queue_tt_tc_map[tt]; + case 3: + compiletime_assert(ARRAY_SIZE(ieee8021q_3queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_3queue_tt_tc_map != max - 1"); + + return ieee8021q_3queue_tt_tc_map[tt]; + case 2: + compiletime_assert(ARRAY_SIZE(ieee8021q_2queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_2queue_tt_tc_map != max - 1"); + + return ieee8021q_2queue_tt_tc_map[tt]; + case 1: + compiletime_assert(ARRAY_SIZE(ieee8021q_1queue_tt_tc_map) != + IEEE8021Q_TT_MAX - 1, + "ieee8021q_1queue_tt_tc_map != max - 1"); + + return ieee8021q_1queue_tt_tc_map[tt]; + } + + pr_err("Invalid number of queues %d\n", num_queues); + + return -EINVAL; +} +EXPORT_SYMBOL_GPL(ieee8021q_tt_to_tc); + +/** + * ietf_dscp_to_ieee8021q_tt - Map IETF DSCP to IEEE 802.1Q Traffic Type + * @dscp: IETF DSCP value + * + * This function maps an IETF DSCP value to an IEEE 802.1Q Traffic Type (TT). + * Since there is no corresponding mapping between DSCP and IEEE 802.1Q Traffic + * Type, this function is inspired by the RFC8325 documentation which describe + * the mapping between DSCP and 802.11 User Priority (UP) values. + * + * Return: IEEE 802.1Q Traffic Type corresponding to the given DSCP value + */ +int ietf_dscp_to_ieee8021q_tt(u8 dscp) +{ + switch (dscp) { + case DSCP_CS0: + /* Comment from RFC8325: + * [RFC4594], Section 4.8, recommends High-Throughput Data be marked + * AF1x (that is, AF11, AF12, and AF13, according to the rules defined + * in [RFC2475]). + * + * By default (as described in Section 2.3), High-Throughput Data will + * map to UP 1 and, thus, to the Background Access Category (AC_BK), + * which is contrary to the intent expressed in [RFC4594]. + + * Unfortunately, there really is no corresponding fit for the High- + * Throughput Data service class within the constrained 4 Access + * Category [IEEE.802.11-2016] model. If the High-Throughput Data + * service class is assigned to the Best Effort Access Category (AC_BE), + * then it would contend with Low-Latency Data (while [RFC4594] + * recommends a distinction in servicing between these service classes) + * as well as with the default service class; alternatively, if it is + * assigned to the Background Access Category (AC_BK), then it would + * receive a less-then-best-effort service and contend with Low-Priority + * Data (as discussed in Section 4.2.10). + * + * As such, since there is no directly corresponding fit for the High- + * Throughout Data service class within the [IEEE.802.11-2016] model, it + * is generally RECOMMENDED to map High-Throughput Data to UP 0, thereby + * admitting it to the Best Effort Access Category (AC_BE). + * + * Note: The above text is from RFC8325 which is describing the mapping + * between DSCP and 802.11 User Priority (UP) values. The mapping + * between UP and IEEE 802.1Q Traffic Type is not defined in the RFC but + * the 802.11 AC_BK and AC_BE are closely related to the IEEE 802.1Q + * Traffic Types BE and BK. + */ + case DSCP_AF11: + case DSCP_AF12: + case DSCP_AF13: + return IEEE8021Q_TT_BE; + /* Comment from RFC8325: + * RFC3662 and RFC4594 both recommend Low-Priority Data be marked + * with DSCP CS1. The Low-Priority Data service class loosely + * corresponds to the [IEEE.802.11-2016] Background Access Category + */ + case DSCP_CS1: + return IEEE8021Q_TT_BK; + case DSCP_CS2: + case DSCP_AF21: + case DSCP_AF22: + case DSCP_AF23: + return IEEE8021Q_TT_EE; + case DSCP_CS3: + case DSCP_AF31: + case DSCP_AF32: + case DSCP_AF33: + return IEEE8021Q_TT_CA; + case DSCP_CS4: + case DSCP_AF41: + case DSCP_AF42: + case DSCP_AF43: + return IEEE8021Q_TT_VI; + case DSCP_CS5: + case DSCP_EF: + case DSCP_VOICE_ADMIT: + return IEEE8021Q_TT_VO; + case DSCP_CS6: + return IEEE8021Q_TT_IC; + case DSCP_CS7: + return IEEE8021Q_TT_NC; + } + + return SIMPLE_IETF_DSCP_TO_IEEE8021Q_TT(dscp); +} +EXPORT_SYMBOL_GPL(ietf_dscp_to_ieee8021q_tt); diff --git a/net/core/link_watch.c b/net/core/link_watch.c index 8ec35194bf..ab15064114 100644 --- a/net/core/link_watch.c +++ b/net/core/link_watch.c @@ -148,9 +148,9 @@ static void linkwatch_schedule_work(int urgent) * override the existing timer. */ if (test_bit(LW_URGENT, &linkwatch_flags)) - mod_delayed_work(system_wq, &linkwatch_work, 0); + mod_delayed_work(system_unbound_wq, &linkwatch_work, 0); else - schedule_delayed_work(&linkwatch_work, delay); + queue_delayed_work(system_unbound_wq, &linkwatch_work, delay); } diff --git a/net/core/neighbour.c b/net/core/neighbour.c index 552719c3bb..45fd88405b 100644 --- a/net/core/neighbour.c +++ b/net/core/neighbour.c @@ -734,7 +734,9 @@ out_neigh_release: struct neighbour *__neigh_create(struct neigh_table *tbl, const void *pkey, struct net_device *dev, bool want_ref) { - return ___neigh_create(tbl, pkey, dev, 0, false, want_ref); + bool exempt_from_gc = !!(dev->flags & IFF_LOOPBACK); + + return ___neigh_create(tbl, pkey, dev, 0, exempt_from_gc, want_ref); } EXPORT_SYMBOL(__neigh_create); @@ -1769,7 +1771,7 @@ static void neigh_parms_destroy(struct neigh_parms *parms) static struct lock_class_key neigh_table_proxy_queue_class; -static struct neigh_table *neigh_tables[NEIGH_NR_TABLES] __read_mostly; +static struct neigh_table __rcu *neigh_tables[NEIGH_NR_TABLES] __read_mostly; void neigh_table_init(int index, struct neigh_table *tbl) { @@ -1826,13 +1828,19 @@ void neigh_table_init(int index, struct neigh_table *tbl) tbl->last_flush = now; tbl->last_rand = now + tbl->parms.reachable_time * 20; - neigh_tables[index] = tbl; + rcu_assign_pointer(neigh_tables[index], tbl); } EXPORT_SYMBOL(neigh_table_init); +/* + * Only called from ndisc_cleanup(), which means this is dead code + * because we no longer can unload IPv6 module. + */ int neigh_table_clear(int index, struct neigh_table *tbl) { - neigh_tables[index] = NULL; + RCU_INIT_POINTER(neigh_tables[index], NULL); + synchronize_rcu(); + /* It is not clean... Fix it to unload IPv6 module safely */ cancel_delayed_work_sync(&tbl->managed_work); cancel_delayed_work_sync(&tbl->gc_work); @@ -1864,10 +1872,10 @@ static struct neigh_table *neigh_find_table(int family) switch (family) { case AF_INET: - tbl = neigh_tables[NEIGH_ARP_TABLE]; + tbl = rcu_dereference_rtnl(neigh_tables[NEIGH_ARP_TABLE]); break; case AF_INET6: - tbl = neigh_tables[NEIGH_ND_TABLE]; + tbl = rcu_dereference_rtnl(neigh_tables[NEIGH_ND_TABLE]); break; } @@ -2331,7 +2339,7 @@ static int neightbl_set(struct sk_buff *skb, struct nlmsghdr *nlh, ndtmsg = nlmsg_data(nlh); for (tidx = 0; tidx < NEIGH_NR_TABLES; tidx++) { - tbl = neigh_tables[tidx]; + tbl = rcu_dereference_rtnl(neigh_tables[tidx]); if (!tbl) continue; if (ndtmsg->ndtm_family && tbl->family != ndtmsg->ndtm_family) @@ -2519,7 +2527,7 @@ static int neightbl_dump_info(struct sk_buff *skb, struct netlink_callback *cb) for (tidx = 0; tidx < NEIGH_NR_TABLES; tidx++) { struct neigh_parms *p; - tbl = neigh_tables[tidx]; + tbl = rcu_dereference_rtnl(neigh_tables[tidx]); if (!tbl) continue; @@ -2674,7 +2682,7 @@ static bool neigh_master_filtered(struct net_device *dev, int master_idx) if (!master_idx) return false; - master = dev ? netdev_master_upper_dev_get(dev) : NULL; + master = dev ? netdev_master_upper_dev_get_rcu(dev) : NULL; /* 0 is already used to denote NDA_MASTER wasn't passed, therefore need another * invalid value for ifindex to denote "no master". @@ -2707,7 +2715,7 @@ static int neigh_dump_table(struct neigh_table *tbl, struct sk_buff *skb, { struct net *net = sock_net(skb->sk); struct neighbour *n; - int rc, h, s_h = cb->args[1]; + int err = 0, h, s_h = cb->args[1]; int idx, s_idx = idx = cb->args[2]; struct neigh_hash_table *nht; unsigned int flags = NLM_F_MULTI; @@ -2715,7 +2723,6 @@ static int neigh_dump_table(struct neigh_table *tbl, struct sk_buff *skb, if (filter->dev_idx || filter->master_idx) flags |= NLM_F_DUMP_FILTERED; - rcu_read_lock(); nht = rcu_dereference(tbl->nht); for (h = s_h; h < (1 << nht->hash_shift); h++) { @@ -2729,23 +2736,19 @@ static int neigh_dump_table(struct neigh_table *tbl, struct sk_buff *skb, if (neigh_ifindex_filtered(n->dev, filter->dev_idx) || neigh_master_filtered(n->dev, filter->master_idx)) goto next; - if (neigh_fill_info(skb, n, NETLINK_CB(cb->skb).portid, - cb->nlh->nlmsg_seq, - RTM_NEWNEIGH, - flags) < 0) { - rc = -1; + err = neigh_fill_info(skb, n, NETLINK_CB(cb->skb).portid, + cb->nlh->nlmsg_seq, + RTM_NEWNEIGH, flags); + if (err < 0) goto out; - } next: idx++; } } - rc = skb->len; out: - rcu_read_unlock(); cb->args[1] = h; cb->args[2] = idx; - return rc; + return err; } static int pneigh_dump_table(struct neigh_table *tbl, struct sk_buff *skb, @@ -2754,7 +2757,7 @@ static int pneigh_dump_table(struct neigh_table *tbl, struct sk_buff *skb, { struct pneigh_entry *n; struct net *net = sock_net(skb->sk); - int rc, h, s_h = cb->args[3]; + int err = 0, h, s_h = cb->args[3]; int idx, s_idx = idx = cb->args[4]; unsigned int flags = NLM_F_MULTI; @@ -2772,11 +2775,11 @@ static int pneigh_dump_table(struct neigh_table *tbl, struct sk_buff *skb, if (neigh_ifindex_filtered(n->dev, filter->dev_idx) || neigh_master_filtered(n->dev, filter->master_idx)) goto next; - if (pneigh_fill_info(skb, n, NETLINK_CB(cb->skb).portid, - cb->nlh->nlmsg_seq, - RTM_NEWNEIGH, flags, tbl) < 0) { + err = pneigh_fill_info(skb, n, NETLINK_CB(cb->skb).portid, + cb->nlh->nlmsg_seq, + RTM_NEWNEIGH, flags, tbl); + if (err < 0) { read_unlock_bh(&tbl->lock); - rc = -1; goto out; } next: @@ -2785,12 +2788,10 @@ static int pneigh_dump_table(struct neigh_table *tbl, struct sk_buff *skb, } read_unlock_bh(&tbl->lock); - rc = skb->len; out: cb->args[3] = h; cb->args[4] = idx; - return rc; - + return err; } static int neigh_valid_dump_req(const struct nlmsghdr *nlh, @@ -2878,8 +2879,9 @@ static int neigh_dump_info(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; + rcu_read_lock(); for (t = 0; t < NEIGH_NR_TABLES; t++) { - tbl = neigh_tables[t]; + tbl = rcu_dereference(neigh_tables[t]); if (!tbl) continue; @@ -2895,9 +2897,10 @@ static int neigh_dump_info(struct sk_buff *skb, struct netlink_callback *cb) if (err < 0) break; } + rcu_read_unlock(); cb->args[0] = t; - return skb->len; + return err; } static int neigh_valid_get_req(const struct nlmsghdr *nlh, @@ -3143,14 +3146,15 @@ int neigh_xmit(int index, struct net_device *dev, const void *addr, struct sk_buff *skb) { int err = -EAFNOSUPPORT; + if (likely(index < NEIGH_NR_TABLES)) { struct neigh_table *tbl; struct neighbour *neigh; - tbl = neigh_tables[index]; - if (!tbl) - goto out; rcu_read_lock(); + tbl = rcu_dereference(neigh_tables[index]); + if (!tbl) + goto out_unlock; if (index == NEIGH_ARP_TABLE) { u32 key = *((u32 *)addr); @@ -3166,6 +3170,7 @@ int neigh_xmit(int index, struct net_device *dev, goto out_kfree_skb; } err = READ_ONCE(neigh->output)(neigh, skb); +out_unlock: rcu_read_unlock(); } else if (index == NEIGH_LINK_TABLE) { @@ -3728,7 +3733,7 @@ static int neigh_proc_base_reachable_time(struct ctl_table *ctl, int write, static struct neigh_sysctl_table { struct ctl_table_header *sysctl_header; - struct ctl_table neigh_vars[NEIGH_VAR_MAX + 1]; + struct ctl_table neigh_vars[NEIGH_VAR_MAX]; } neigh_sysctl_template __read_mostly = { .neigh_vars = { NEIGH_SYSCTL_ZERO_INTMAX_ENTRY(MCAST_PROBES, "mcast_solicit"), @@ -3779,7 +3784,6 @@ static struct neigh_sysctl_table { .extra2 = SYSCTL_INT_MAX, .proc_handler = proc_dointvec_minmax, }, - {}, }, }; @@ -3807,8 +3811,6 @@ int neigh_sysctl_register(struct net_device *dev, struct neigh_parms *p, if (dev) { dev_name_source = dev->name; /* Terminate the table early */ - memset(&t->neigh_vars[NEIGH_VAR_GC_INTERVAL], 0, - sizeof(t->neigh_vars[NEIGH_VAR_GC_INTERVAL])); neigh_vars_size = NEIGH_VAR_BASE_REACHABLE_TIME_MS + 1; } else { struct neigh_table *tbl = p->tbl; @@ -3889,7 +3891,8 @@ static int __init neigh_init(void) { rtnl_register(PF_UNSPEC, RTM_NEWNEIGH, neigh_add, NULL, 0); rtnl_register(PF_UNSPEC, RTM_DELNEIGH, neigh_delete, NULL, 0); - rtnl_register(PF_UNSPEC, RTM_GETNEIGH, neigh_get, neigh_dump_info, 0); + rtnl_register(PF_UNSPEC, RTM_GETNEIGH, neigh_get, neigh_dump_info, + RTNL_FLAG_DUMP_UNLOCKED); rtnl_register(PF_UNSPEC, RTM_GETNEIGHTBL, NULL, neightbl_dump_info, 0); diff --git a/net/core/net-procfs.c b/net/core/net-procfs.c index a97eceb84e..fa6d396973 100644 --- a/net/core/net-procfs.c +++ b/net/core/net-procfs.c @@ -144,7 +144,8 @@ static int softnet_seq_show(struct seq_file *seq, void *v) seq_printf(seq, "%08x %08x %08x %08x %08x %08x %08x %08x %08x %08x %08x %08x %08x " "%08x %08x\n", - sd->processed, sd->dropped, sd->time_squeeze, 0, + sd->processed, atomic_read(&sd->dropped), + sd->time_squeeze, 0, 0, 0, 0, 0, /* was fastroute */ 0, /* was cpu_collision */ sd->received_rps, flow_limit_count, diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index e3d7a8cfa2..4c27a360c2 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -605,13 +605,13 @@ static ssize_t threaded_show(struct device *dev, struct net_device *netdev = to_net_dev(dev); ssize_t ret = -EINVAL; - if (!rtnl_trylock()) - return restart_syscall(); + rcu_read_lock(); if (dev_isalive(netdev)) - ret = sysfs_emit(buf, fmt_dec, netdev->threaded); + ret = sysfs_emit(buf, fmt_dec, READ_ONCE(netdev->threaded)); + + rcu_read_unlock(); - rtnl_unlock(); return ret; } @@ -1419,7 +1419,7 @@ static ssize_t bql_show_stall_thrs(struct netdev_queue *queue, char *buf) { struct dql *dql = &queue->dql; - return sprintf(buf, "%u\n", jiffies_to_msecs(dql->stall_thrs)); + return sysfs_emit(buf, "%u\n", jiffies_to_msecs(dql->stall_thrs)); } static ssize_t bql_set_stall_thrs(struct netdev_queue *queue, @@ -1451,7 +1451,7 @@ static struct netdev_queue_attribute bql_stall_thrs_attribute __ro_after_init = static ssize_t bql_show_stall_max(struct netdev_queue *queue, char *buf) { - return sprintf(buf, "%u\n", READ_ONCE(queue->dql.stall_max)); + return sysfs_emit(buf, "%u\n", READ_ONCE(queue->dql.stall_max)); } static ssize_t bql_set_stall_max(struct netdev_queue *queue, @@ -1468,7 +1468,7 @@ static ssize_t bql_show_stall_cnt(struct netdev_queue *queue, char *buf) { struct dql *dql = &queue->dql; - return sprintf(buf, "%lu\n", dql->stall_cnt); + return sysfs_emit(buf, "%lu\n", dql->stall_cnt); } static struct netdev_queue_attribute bql_stall_cnt_attribute __ro_after_init = @@ -2046,7 +2046,7 @@ static void net_get_ownership(const struct device *d, kuid_t *uid, kgid_t *gid) net_ns_get_ownership(net, uid, gid); } -static struct class net_class __ro_after_init = { +static const struct class net_class = { .name = "net", .dev_release = netdev_release, .dev_groups = net_class_groups, diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c index b1dc84c4fd..6a823ba906 100644 --- a/net/core/net_namespace.c +++ b/net/core/net_namespace.c @@ -1098,7 +1098,7 @@ static int rtnl_net_dumpid(struct sk_buff *skb, struct netlink_callback *cb) end: if (net_cb.fillargs.add_ref) put_net(net_cb.tgt_net); - return err < 0 ? err : skb->len; + return err; } static void rtnl_net_notifyid(struct net *net, int cmd, int id, u32 portid, @@ -1213,7 +1213,8 @@ void __init net_ns_init(void) rtnl_register(PF_UNSPEC, RTM_NEWNSID, rtnl_net_newid, NULL, RTNL_FLAG_DOIT_UNLOCKED); rtnl_register(PF_UNSPEC, RTM_GETNSID, rtnl_net_getid, rtnl_net_dumpid, - RTNL_FLAG_DOIT_UNLOCKED); + RTNL_FLAG_DOIT_UNLOCKED | + RTNL_FLAG_DUMP_UNLOCKED); } static void free_exit_list(struct pernet_operations *ops, struct list_head *net_exit_list) diff --git a/net/core/gso_test.c b/net/core/net_test.c index 358c44680d..9c3a590865 100644 --- a/net/core/gso_test.c +++ b/net/core/net_test.c @@ -1,6 +1,9 @@ // SPDX-License-Identifier: GPL-2.0-or-later #include <kunit/test.h> + +/* GSO */ + #include <linux/skbuff.h> static const char hdr[] = "abcdefgh"; @@ -258,17 +261,127 @@ free_gso_skb: consume_skb(skb); } -static struct kunit_case gso_test_cases[] = { - KUNIT_CASE_PARAM(gso_test_func, gso_test_gen_params), - {} +/* IP tunnel flags */ + +#include <net/ip_tunnels.h> + +struct ip_tunnel_flags_test { + const char *name; + + const u16 *src_bits; + const u16 *exp_bits; + u8 src_num; + u8 exp_num; + + __be16 exp_val; + bool exp_comp; +}; + +#define IP_TUNNEL_FLAGS_TEST(n, src, comp, eval, exp) { \ + .name = (n), \ + .src_bits = (src), \ + .src_num = ARRAY_SIZE(src), \ + .exp_comp = (comp), \ + .exp_val = (eval), \ + .exp_bits = (exp), \ + .exp_num = ARRAY_SIZE(exp), \ +} + +/* These are __be16-compatible and can be compared as is */ +static const u16 ip_tunnel_flags_1[] = { + IP_TUNNEL_KEY_BIT, + IP_TUNNEL_STRICT_BIT, + IP_TUNNEL_ERSPAN_OPT_BIT, +}; + +/* Due to the previous flags design limitation, setting either + * ``IP_TUNNEL_CSUM_BIT`` (on Big Endian) or ``IP_TUNNEL_DONT_FRAGMENT_BIT`` + * (on Little) also sets VTI/ISATAP bit. In the bitmap implementation, they + * correspond to ``BIT(16)``, which is bigger than ``U16_MAX``, but still is + * backward-compatible. + */ +#ifdef __LITTLE_ENDIAN +#define IP_TUNNEL_CONFLICT_BIT IP_TUNNEL_DONT_FRAGMENT_BIT +#else +#define IP_TUNNEL_CONFLICT_BIT IP_TUNNEL_CSUM_BIT +#endif + +static const u16 ip_tunnel_flags_2_src[] = { + IP_TUNNEL_CONFLICT_BIT, +}; + +static const u16 ip_tunnel_flags_2_exp[] = { + IP_TUNNEL_CONFLICT_BIT, + IP_TUNNEL_SIT_ISATAP_BIT, }; -static struct kunit_suite gso_test_suite = { - .name = "net_core_gso", - .test_cases = gso_test_cases, +/* Bits 17 and higher are not compatible with __be16 flags */ +static const u16 ip_tunnel_flags_3_src[] = { + IP_TUNNEL_VXLAN_OPT_BIT, + 17, + 18, + 20, }; -kunit_test_suite(gso_test_suite); +static const u16 ip_tunnel_flags_3_exp[] = { + IP_TUNNEL_VXLAN_OPT_BIT, +}; + +static const struct ip_tunnel_flags_test ip_tunnel_flags_test[] = { + IP_TUNNEL_FLAGS_TEST("compat", ip_tunnel_flags_1, true, + cpu_to_be16(BIT(IP_TUNNEL_KEY_BIT) | + BIT(IP_TUNNEL_STRICT_BIT) | + BIT(IP_TUNNEL_ERSPAN_OPT_BIT)), + ip_tunnel_flags_1), + IP_TUNNEL_FLAGS_TEST("conflict", ip_tunnel_flags_2_src, true, + VTI_ISVTI, ip_tunnel_flags_2_exp), + IP_TUNNEL_FLAGS_TEST("new", ip_tunnel_flags_3_src, false, + cpu_to_be16(BIT(IP_TUNNEL_VXLAN_OPT_BIT)), + ip_tunnel_flags_3_exp), +}; + +static void +ip_tunnel_flags_test_case_to_desc(const struct ip_tunnel_flags_test *t, + char *desc) +{ + strscpy(desc, t->name, KUNIT_PARAM_DESC_SIZE); +} +KUNIT_ARRAY_PARAM(ip_tunnel_flags_test, ip_tunnel_flags_test, + ip_tunnel_flags_test_case_to_desc); + +static void ip_tunnel_flags_test_run(struct kunit *test) +{ + const struct ip_tunnel_flags_test *t = test->param_value; + IP_TUNNEL_DECLARE_FLAGS(src) = { }; + IP_TUNNEL_DECLARE_FLAGS(exp) = { }; + IP_TUNNEL_DECLARE_FLAGS(out); + + for (u32 j = 0; j < t->src_num; j++) + __set_bit(t->src_bits[j], src); + for (u32 j = 0; j < t->exp_num; j++) + __set_bit(t->exp_bits[j], exp); + + KUNIT_ASSERT_EQ(test, t->exp_comp, + ip_tunnel_flags_is_be16_compat(src)); + KUNIT_ASSERT_EQ(test, (__force u16)t->exp_val, + (__force u16)ip_tunnel_flags_to_be16(src)); + + ip_tunnel_flags_from_be16(out, t->exp_val); + KUNIT_ASSERT_TRUE(test, __ipt_flag_op(bitmap_equal, exp, out)); +} + +static struct kunit_case net_test_cases[] = { + KUNIT_CASE_PARAM(gso_test_func, gso_test_gen_params), + KUNIT_CASE_PARAM(ip_tunnel_flags_test_run, + ip_tunnel_flags_test_gen_params), + { }, +}; + +static struct kunit_suite net_test_suite = { + .name = "net_core", + .test_cases = net_test_cases, +}; +kunit_test_suite(net_test_suite); +MODULE_DESCRIPTION("KUnit tests for networking core"); MODULE_LICENSE("GPL"); -MODULE_DESCRIPTION("KUnit tests for segmentation offload"); diff --git a/net/core/netdev-genl-gen.c b/net/core/netdev-genl-gen.c index 8d8ace9ef8..8350a0afa9 100644 --- a/net/core/netdev-genl-gen.c +++ b/net/core/netdev-genl-gen.c @@ -70,6 +70,7 @@ static const struct nla_policy netdev_napi_get_dump_nl_policy[NETDEV_A_NAPI_IFIN /* NETDEV_CMD_QSTATS_GET - dump */ static const struct nla_policy netdev_qstats_get_nl_policy[NETDEV_A_QSTATS_SCOPE + 1] = { + [NETDEV_A_QSTATS_IFINDEX] = NLA_POLICY_MIN(NLA_U32, 1), [NETDEV_A_QSTATS_SCOPE] = NLA_POLICY_MASK(NLA_UINT, 0x1), }; diff --git a/net/core/netdev-genl.c b/net/core/netdev-genl.c index 8c2d5a0bc2..05f9515d2c 100644 --- a/net/core/netdev-genl.c +++ b/net/core/netdev-genl.c @@ -489,7 +489,17 @@ netdev_nl_stats_write_rx(struct sk_buff *rsp, struct netdev_queue_stats_rx *rx) { if (netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_PACKETS, rx->packets) || netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_BYTES, rx->bytes) || - netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_ALLOC_FAIL, rx->alloc_fail)) + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_ALLOC_FAIL, rx->alloc_fail) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_DROPS, rx->hw_drops) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_DROP_OVERRUNS, rx->hw_drop_overruns) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_CSUM_UNNECESSARY, rx->csum_unnecessary) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_CSUM_NONE, rx->csum_none) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_CSUM_BAD, rx->csum_bad) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_PACKETS, rx->hw_gro_packets) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_BYTES, rx->hw_gro_bytes) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_WIRE_PACKETS, rx->hw_gro_wire_packets) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_WIRE_BYTES, rx->hw_gro_wire_bytes) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_DROP_RATELIMITS, rx->hw_drop_ratelimits)) return -EMSGSIZE; return 0; } @@ -498,7 +508,18 @@ static int netdev_nl_stats_write_tx(struct sk_buff *rsp, struct netdev_queue_stats_tx *tx) { if (netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_PACKETS, tx->packets) || - netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_BYTES, tx->bytes)) + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_BYTES, tx->bytes) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_DROPS, tx->hw_drops) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_DROP_ERRORS, tx->hw_drop_errors) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_CSUM_NONE, tx->csum_none) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_NEEDS_CSUM, tx->needs_csum) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_PACKETS, tx->hw_gso_packets) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_BYTES, tx->hw_gso_bytes) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_WIRE_PACKETS, tx->hw_gso_wire_packets) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_WIRE_BYTES, tx->hw_gso_wire_bytes) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_DROP_RATELIMITS, tx->hw_drop_ratelimits) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_STOP, tx->stop) || + netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_WAKE, tx->wake)) return -EMSGSIZE; return 0; } @@ -639,6 +660,24 @@ nla_put_failure: return -EMSGSIZE; } +static int +netdev_nl_qstats_get_dump_one(struct net_device *netdev, unsigned int scope, + struct sk_buff *skb, const struct genl_info *info, + struct netdev_nl_dump_ctx *ctx) +{ + if (!netdev->stat_ops) + return 0; + + switch (scope) { + case 0: + return netdev_nl_stats_by_netdev(netdev, skb, info); + case NETDEV_QSTATS_SCOPE_QUEUE: + return netdev_nl_stats_by_queue(netdev, skb, info, ctx); + } + + return -EINVAL; /* Should not happen, per netlink policy */ +} + int netdev_nl_qstats_get_dumpit(struct sk_buff *skb, struct netlink_callback *cb) { @@ -646,6 +685,7 @@ int netdev_nl_qstats_get_dumpit(struct sk_buff *skb, const struct genl_info *info = genl_info_dump(cb); struct net *net = sock_net(skb->sk); struct net_device *netdev; + unsigned int ifindex; unsigned int scope; int err = 0; @@ -653,21 +693,28 @@ int netdev_nl_qstats_get_dumpit(struct sk_buff *skb, if (info->attrs[NETDEV_A_QSTATS_SCOPE]) scope = nla_get_uint(info->attrs[NETDEV_A_QSTATS_SCOPE]); - rtnl_lock(); - for_each_netdev_dump(net, netdev, ctx->ifindex) { - if (!netdev->stat_ops) - continue; + ifindex = 0; + if (info->attrs[NETDEV_A_QSTATS_IFINDEX]) + ifindex = nla_get_u32(info->attrs[NETDEV_A_QSTATS_IFINDEX]); - switch (scope) { - case 0: - err = netdev_nl_stats_by_netdev(netdev, skb, info); - break; - case NETDEV_QSTATS_SCOPE_QUEUE: - err = netdev_nl_stats_by_queue(netdev, skb, info, ctx); - break; + rtnl_lock(); + if (ifindex) { + netdev = __dev_get_by_index(net, ifindex); + if (netdev && netdev->stat_ops) { + err = netdev_nl_qstats_get_dump_one(netdev, scope, skb, + info, ctx); + } else { + NL_SET_BAD_ATTR(info->extack, + info->attrs[NETDEV_A_QSTATS_IFINDEX]); + err = netdev ? -EOPNOTSUPP : -ENODEV; + } + } else { + for_each_netdev_dump(net, netdev, ctx->ifindex) { + err = netdev_nl_qstats_get_dump_one(netdev, scope, skb, + info, ctx); + if (err < 0) + break; } - if (err < 0) - break; } rtnl_unlock(); diff --git a/net/core/page_pool.c b/net/core/page_pool.c index dd364d738c..3772eb63dc 100644 --- a/net/core/page_pool.c +++ b/net/core/page_pool.c @@ -5,6 +5,7 @@ * Copyright (C) 2016 Red Hat, Inc. */ +#include <linux/error-injection.h> #include <linux/types.h> #include <linux/kernel.h> #include <linux/slab.h> @@ -123,9 +124,9 @@ int page_pool_ethtool_stats_get_count(void) } EXPORT_SYMBOL(page_pool_ethtool_stats_get_count); -u64 *page_pool_ethtool_stats_get(u64 *data, void *stats) +u64 *page_pool_ethtool_stats_get(u64 *data, const void *stats) { - struct page_pool_stats *pool_stats = stats; + const struct page_pool_stats *pool_stats = stats; *data++ = pool_stats->alloc_stats.fast; *data++ = pool_stats->alloc_stats.slow; @@ -172,19 +173,29 @@ static void page_pool_producer_unlock(struct page_pool *pool, spin_unlock_bh(&pool->ring.producer_lock); } +static void page_pool_struct_check(void) +{ + CACHELINE_ASSERT_GROUP_MEMBER(struct page_pool, frag, frag_users); + CACHELINE_ASSERT_GROUP_MEMBER(struct page_pool, frag, frag_page); + CACHELINE_ASSERT_GROUP_MEMBER(struct page_pool, frag, frag_offset); + CACHELINE_ASSERT_GROUP_SIZE(struct page_pool, frag, 4 * sizeof(long)); +} + static int page_pool_init(struct page_pool *pool, const struct page_pool_params *params, int cpuid) { unsigned int ring_qsize = 1024; /* Default */ + page_pool_struct_check(); + memcpy(&pool->p, ¶ms->fast, sizeof(pool->p)); memcpy(&pool->slow, ¶ms->slow, sizeof(pool->slow)); pool->cpuid = cpuid; /* Validate only known flags were used */ - if (pool->p.flags & ~(PP_FLAG_ALL)) + if (pool->slow.flags & ~PP_FLAG_ALL) return -EINVAL; if (pool->p.pool_size) @@ -198,22 +209,26 @@ static int page_pool_init(struct page_pool *pool, * DMA_BIDIRECTIONAL is for allowing page used for DMA sending, * which is the XDP_TX use-case. */ - if (pool->p.flags & PP_FLAG_DMA_MAP) { + if (pool->slow.flags & PP_FLAG_DMA_MAP) { if ((pool->p.dma_dir != DMA_FROM_DEVICE) && (pool->p.dma_dir != DMA_BIDIRECTIONAL)) return -EINVAL; + + pool->dma_map = true; } - if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV) { + if (pool->slow.flags & PP_FLAG_DMA_SYNC_DEV) { /* In order to request DMA-sync-for-device the page * needs to be mapped */ - if (!(pool->p.flags & PP_FLAG_DMA_MAP)) + if (!(pool->slow.flags & PP_FLAG_DMA_MAP)) return -EINVAL; if (!pool->p.max_len) return -EINVAL; + pool->dma_sync = true; + /* pool->p.offset has to be set according to the address * offset used by the DMA engine to start copying rx data */ @@ -222,7 +237,7 @@ static int page_pool_init(struct page_pool *pool, pool->has_init_callback = !!pool->slow.init_callback; #ifdef CONFIG_PAGE_POOL_STATS - if (!(pool->p.flags & PP_FLAG_SYSTEM_POOL)) { + if (!(pool->slow.flags & PP_FLAG_SYSTEM_POOL)) { pool->recycle_stats = alloc_percpu(struct page_pool_recycle_stats); if (!pool->recycle_stats) return -ENOMEM; @@ -232,12 +247,13 @@ static int page_pool_init(struct page_pool *pool, * (also percpu) page pool instance. */ pool->recycle_stats = &pp_system_recycle_stats; + pool->system = true; } #endif if (ptr_ring_init(&pool->ring, ring_qsize, GFP_KERNEL) < 0) { #ifdef CONFIG_PAGE_POOL_STATS - if (!(pool->p.flags & PP_FLAG_SYSTEM_POOL)) + if (!pool->system) free_percpu(pool->recycle_stats); #endif return -ENOMEM; @@ -248,7 +264,7 @@ static int page_pool_init(struct page_pool *pool, /* Driver calling page_pool_create() also call page_pool_destroy() */ refcount_set(&pool->user_cnt, 1); - if (pool->p.flags & PP_FLAG_DMA_MAP) + if (pool->dma_map) get_device(pool->p.dev); return 0; @@ -258,11 +274,11 @@ static void page_pool_uninit(struct page_pool *pool) { ptr_ring_cleanup(&pool->ring, NULL); - if (pool->p.flags & PP_FLAG_DMA_MAP) + if (pool->dma_map) put_device(pool->p.dev); #ifdef CONFIG_PAGE_POOL_STATS - if (!(pool->p.flags & PP_FLAG_SYSTEM_POOL)) + if (!pool->system) free_percpu(pool->recycle_stats); #endif } @@ -383,16 +399,26 @@ static struct page *__page_pool_get_cached(struct page_pool *pool) return page; } -static void page_pool_dma_sync_for_device(struct page_pool *pool, - struct page *page, - unsigned int dma_sync_size) +static void __page_pool_dma_sync_for_device(const struct page_pool *pool, + const struct page *page, + u32 dma_sync_size) { +#if defined(CONFIG_HAS_DMA) && defined(CONFIG_DMA_NEED_SYNC) dma_addr_t dma_addr = page_pool_get_dma_addr(page); dma_sync_size = min(dma_sync_size, pool->p.max_len); - dma_sync_single_range_for_device(pool->p.dev, dma_addr, - pool->p.offset, dma_sync_size, - pool->p.dma_dir); + __dma_sync_single_for_device(pool->p.dev, dma_addr + pool->p.offset, + dma_sync_size, pool->p.dma_dir); +#endif +} + +static __always_inline void +page_pool_dma_sync_for_device(const struct page_pool *pool, + const struct page *page, + u32 dma_sync_size) +{ + if (pool->dma_sync && dma_dev_need_sync(pool->p.dev)) + __page_pool_dma_sync_for_device(pool, page, dma_sync_size); } static bool page_pool_dma_map(struct page_pool *pool, struct page *page) @@ -414,13 +440,12 @@ static bool page_pool_dma_map(struct page_pool *pool, struct page *page) if (page_pool_set_dma_addr(page, dma)) goto unmap_failed; - if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV) - page_pool_dma_sync_for_device(pool, page, pool->p.max_len); + page_pool_dma_sync_for_device(pool, page, pool->p.max_len); return true; unmap_failed: - WARN_ON_ONCE("unexpected DMA address, please report to netdev@"); + WARN_ONCE(1, "unexpected DMA address, please report to netdev@"); dma_unmap_page_attrs(pool->p.dev, dma, PAGE_SIZE << pool->p.order, pool->p.dma_dir, DMA_ATTR_SKIP_CPU_SYNC | DMA_ATTR_WEAK_ORDERING); @@ -460,8 +485,7 @@ static struct page *__page_pool_alloc_page_order(struct page_pool *pool, if (unlikely(!page)) return NULL; - if ((pool->p.flags & PP_FLAG_DMA_MAP) && - unlikely(!page_pool_dma_map(pool, page))) { + if (pool->dma_map && unlikely(!page_pool_dma_map(pool, page))) { put_page(page); return NULL; } @@ -481,8 +505,8 @@ static struct page *__page_pool_alloc_pages_slow(struct page_pool *pool, gfp_t gfp) { const int bulk = PP_ALLOC_CACHE_REFILL; - unsigned int pp_flags = pool->p.flags; unsigned int pp_order = pool->p.order; + bool dma_map = pool->dma_map; struct page *page; int i, nr_pages; @@ -507,8 +531,7 @@ static struct page *__page_pool_alloc_pages_slow(struct page_pool *pool, */ for (i = 0; i < nr_pages; i++) { page = pool->alloc.cache[i]; - if ((pp_flags & PP_FLAG_DMA_MAP) && - unlikely(!page_pool_dma_map(pool, page))) { + if (dma_map && unlikely(!page_pool_dma_map(pool, page))) { put_page(page); continue; } @@ -550,6 +573,7 @@ struct page *page_pool_alloc_pages(struct page_pool *pool, gfp_t gfp) return page; } EXPORT_SYMBOL(page_pool_alloc_pages); +ALLOW_ERROR_INJECTION(page_pool_alloc_pages, NULL); /* Calculate distance between two u32 values, valid if distance is below 2^(31) * https://en.wikipedia.org/wiki/Serial_number_arithmetic#General_Solution @@ -580,7 +604,7 @@ void __page_pool_release_page_dma(struct page_pool *pool, struct page *page) { dma_addr_t dma; - if (!(pool->p.flags & PP_FLAG_DMA_MAP)) + if (!pool->dma_map) /* Always account for inflight pages, even if we didn't * map them */ @@ -663,7 +687,7 @@ static bool __page_pool_page_can_be_recycled(const struct page *page) } /* If the page refcnt == 1, this will try to recycle the page. - * if PP_FLAG_DMA_SYNC_DEV is set, we'll try to sync the DMA area for + * If pool->dma_sync is set, we'll try to sync the DMA area for * the configured size min(dma_sync_size, pool->max_len). * If the page refcnt != 1, then the page will be returned to memory * subsystem. @@ -686,12 +710,9 @@ __page_pool_put_page(struct page_pool *pool, struct page *page, if (likely(__page_pool_page_can_be_recycled(page))) { /* Read barrier done in page_ref_count / READ_ONCE */ - if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV) - page_pool_dma_sync_for_device(pool, page, - dma_sync_size); + page_pool_dma_sync_for_device(pool, page, dma_sync_size); - if (allow_direct && in_softirq() && - page_pool_recycle_in_cache(page, pool)) + if (allow_direct && page_pool_recycle_in_cache(page, pool)) return NULL; /* Page found as candidate for recycling */ @@ -716,9 +737,35 @@ __page_pool_put_page(struct page_pool *pool, struct page *page, return NULL; } +static bool page_pool_napi_local(const struct page_pool *pool) +{ + const struct napi_struct *napi; + u32 cpuid; + + if (unlikely(!in_softirq())) + return false; + + /* Allow direct recycle if we have reasons to believe that we are + * in the same context as the consumer would run, so there's + * no possible race. + * __page_pool_put_page() makes sure we're not in hardirq context + * and interrupts are enabled prior to accessing the cache. + */ + cpuid = smp_processor_id(); + if (READ_ONCE(pool->cpuid) == cpuid) + return true; + + napi = READ_ONCE(pool->p.napi); + + return napi && READ_ONCE(napi->list_owner) == cpuid; +} + void page_pool_put_unrefed_page(struct page_pool *pool, struct page *page, unsigned int dma_sync_size, bool allow_direct) { + if (!allow_direct) + allow_direct = page_pool_napi_local(pool); + page = __page_pool_put_page(pool, page, dma_sync_size, allow_direct); if (page && !page_pool_recycle_in_ring(pool, page)) { /* Cache full, fallback to free pages */ @@ -747,8 +794,11 @@ void page_pool_put_page_bulk(struct page_pool *pool, void **data, int count) { int i, bulk_len = 0; + bool allow_direct; bool in_softirq; + allow_direct = page_pool_napi_local(pool); + for (i = 0; i < count; i++) { struct page *page = virt_to_head_page(data[i]); @@ -756,13 +806,13 @@ void page_pool_put_page_bulk(struct page_pool *pool, void **data, if (!page_pool_is_last_ref(page)) continue; - page = __page_pool_put_page(pool, page, -1, false); + page = __page_pool_put_page(pool, page, -1, allow_direct); /* Approved for bulk recycling in ptr_ring cache */ if (page) data[bulk_len++] = page; } - if (unlikely(!bulk_len)) + if (!bulk_len) return; /* Bulk producer into ptr_ring page_pool cache */ @@ -799,9 +849,7 @@ static struct page *page_pool_drain_frag(struct page_pool *pool, return NULL; if (__page_pool_page_can_be_recycled(page)) { - if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV) - page_pool_dma_sync_for_device(pool, page, -1); - + page_pool_dma_sync_for_device(pool, page, -1); return page; } @@ -959,7 +1007,7 @@ static void page_pool_release_retry(struct work_struct *wq) } void page_pool_use_xdp_mem(struct page_pool *pool, void (*disconnect)(void *), - struct xdp_mem_info *mem) + const struct xdp_mem_info *mem) { refcount_inc(&pool->user_cnt); pool->disconnect = disconnect; @@ -969,7 +1017,7 @@ void page_pool_use_xdp_mem(struct page_pool *pool, void (*disconnect)(void *), static void page_pool_disable_direct_recycling(struct page_pool *pool) { /* Disable direct recycling based on pool->cpuid. - * Paired with READ_ONCE() in napi_pp_put_page(). + * Paired with READ_ONCE() in page_pool_napi_local(). */ WRITE_ONCE(pool->cpuid, -1); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 74e6f9746f..5e589f0a62 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -1036,8 +1036,8 @@ static size_t rtnl_proto_down_size(const struct net_device *dev) { size_t size = nla_total_size(1); - if (dev->proto_down_reason) - size += nla_total_size(0) + nla_total_size(4); + /* Assume dev->proto_down_reason is not zero. */ + size += nla_total_size(0) + nla_total_size(4); return size; } @@ -1477,13 +1477,15 @@ static int rtnl_fill_link_ifmap(struct sk_buff *skb, static u32 rtnl_xdp_prog_skb(struct net_device *dev) { const struct bpf_prog *generic_xdp_prog; + u32 res = 0; - ASSERT_RTNL(); + rcu_read_lock(); + generic_xdp_prog = rcu_dereference(dev->xdp_prog); + if (generic_xdp_prog) + res = generic_xdp_prog->aux->id; + rcu_read_unlock(); - generic_xdp_prog = rtnl_dereference(dev->xdp_prog); - if (!generic_xdp_prog) - return 0; - return generic_xdp_prog->aux->id; + return res; } static u32 rtnl_xdp_prog_drv(struct net_device *dev) @@ -1603,7 +1605,8 @@ static int put_master_ifindex(struct sk_buff *skb, struct net_device *dev) upper_dev = netdev_master_upper_dev_get_rcu(dev); if (upper_dev) - ret = nla_put_u32(skb, IFLA_MASTER, upper_dev->ifindex); + ret = nla_put_u32(skb, IFLA_MASTER, + READ_ONCE(upper_dev->ifindex)); rcu_read_unlock(); return ret; @@ -1736,10 +1739,10 @@ static int rtnl_fill_proto_down(struct sk_buff *skb, struct nlattr *pr; u32 preason; - if (nla_put_u8(skb, IFLA_PROTO_DOWN, dev->proto_down)) + if (nla_put_u8(skb, IFLA_PROTO_DOWN, READ_ONCE(dev->proto_down))) goto nla_put_failure; - preason = dev->proto_down_reason; + preason = READ_ONCE(dev->proto_down_reason); if (!preason) return 0; @@ -1812,6 +1815,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, u32 event, int *new_nsid, int new_ifindex, int tgt_netnsid, gfp_t gfp) { + char devname[IFNAMSIZ]; struct ifinfomsg *ifm; struct nlmsghdr *nlh; struct Qdisc *qdisc; @@ -1824,41 +1828,51 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, ifm = nlmsg_data(nlh); ifm->ifi_family = AF_UNSPEC; ifm->__ifi_pad = 0; - ifm->ifi_type = dev->type; - ifm->ifi_index = dev->ifindex; + ifm->ifi_type = READ_ONCE(dev->type); + ifm->ifi_index = READ_ONCE(dev->ifindex); ifm->ifi_flags = dev_get_flags(dev); ifm->ifi_change = change; if (tgt_netnsid >= 0 && nla_put_s32(skb, IFLA_TARGET_NETNSID, tgt_netnsid)) goto nla_put_failure; - qdisc = rtnl_dereference(dev->qdisc); - if (nla_put_string(skb, IFLA_IFNAME, dev->name) || - nla_put_u32(skb, IFLA_TXQLEN, dev->tx_queue_len) || + netdev_copy_name(dev, devname); + if (nla_put_string(skb, IFLA_IFNAME, devname)) + goto nla_put_failure; + + if (nla_put_u32(skb, IFLA_TXQLEN, READ_ONCE(dev->tx_queue_len)) || nla_put_u8(skb, IFLA_OPERSTATE, - netif_running(dev) ? dev->operstate : IF_OPER_DOWN) || - nla_put_u8(skb, IFLA_LINKMODE, dev->link_mode) || - nla_put_u32(skb, IFLA_MTU, dev->mtu) || - nla_put_u32(skb, IFLA_MIN_MTU, dev->min_mtu) || - nla_put_u32(skb, IFLA_MAX_MTU, dev->max_mtu) || - nla_put_u32(skb, IFLA_GROUP, dev->group) || - nla_put_u32(skb, IFLA_PROMISCUITY, dev->promiscuity) || - nla_put_u32(skb, IFLA_ALLMULTI, dev->allmulti) || - nla_put_u32(skb, IFLA_NUM_TX_QUEUES, dev->num_tx_queues) || - nla_put_u32(skb, IFLA_GSO_MAX_SEGS, dev->gso_max_segs) || - nla_put_u32(skb, IFLA_GSO_MAX_SIZE, dev->gso_max_size) || - nla_put_u32(skb, IFLA_GRO_MAX_SIZE, dev->gro_max_size) || - nla_put_u32(skb, IFLA_GSO_IPV4_MAX_SIZE, dev->gso_ipv4_max_size) || - nla_put_u32(skb, IFLA_GRO_IPV4_MAX_SIZE, dev->gro_ipv4_max_size) || - nla_put_u32(skb, IFLA_TSO_MAX_SIZE, dev->tso_max_size) || - nla_put_u32(skb, IFLA_TSO_MAX_SEGS, dev->tso_max_segs) || + netif_running(dev) ? READ_ONCE(dev->operstate) : + IF_OPER_DOWN) || + nla_put_u8(skb, IFLA_LINKMODE, READ_ONCE(dev->link_mode)) || + nla_put_u32(skb, IFLA_MTU, READ_ONCE(dev->mtu)) || + nla_put_u32(skb, IFLA_MIN_MTU, READ_ONCE(dev->min_mtu)) || + nla_put_u32(skb, IFLA_MAX_MTU, READ_ONCE(dev->max_mtu)) || + nla_put_u32(skb, IFLA_GROUP, READ_ONCE(dev->group)) || + nla_put_u32(skb, IFLA_PROMISCUITY, READ_ONCE(dev->promiscuity)) || + nla_put_u32(skb, IFLA_ALLMULTI, READ_ONCE(dev->allmulti)) || + nla_put_u32(skb, IFLA_NUM_TX_QUEUES, + READ_ONCE(dev->num_tx_queues)) || + nla_put_u32(skb, IFLA_GSO_MAX_SEGS, + READ_ONCE(dev->gso_max_segs)) || + nla_put_u32(skb, IFLA_GSO_MAX_SIZE, + READ_ONCE(dev->gso_max_size)) || + nla_put_u32(skb, IFLA_GRO_MAX_SIZE, + READ_ONCE(dev->gro_max_size)) || + nla_put_u32(skb, IFLA_GSO_IPV4_MAX_SIZE, + READ_ONCE(dev->gso_ipv4_max_size)) || + nla_put_u32(skb, IFLA_GRO_IPV4_MAX_SIZE, + READ_ONCE(dev->gro_ipv4_max_size)) || + nla_put_u32(skb, IFLA_TSO_MAX_SIZE, + READ_ONCE(dev->tso_max_size)) || + nla_put_u32(skb, IFLA_TSO_MAX_SEGS, + READ_ONCE(dev->tso_max_segs)) || #ifdef CONFIG_RPS - nla_put_u32(skb, IFLA_NUM_RX_QUEUES, dev->num_rx_queues) || + nla_put_u32(skb, IFLA_NUM_RX_QUEUES, + READ_ONCE(dev->num_rx_queues)) || #endif put_master_ifindex(skb, dev) || nla_put_u8(skb, IFLA_CARRIER, netif_carrier_ok(dev)) || - (qdisc && - nla_put_string(skb, IFLA_QDISC, qdisc->ops->id)) || nla_put_ifalias(skb, dev) || nla_put_u32(skb, IFLA_CARRIER_CHANGES, atomic_read(&dev->carrier_up_count) + @@ -1909,9 +1923,6 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, goto nla_put_failure; } - if (rtnl_fill_link_netnsid(skb, dev, src_net, gfp)) - goto nla_put_failure; - if (new_nsid && nla_put_s32(skb, IFLA_NEW_NETNSID, *new_nsid) < 0) goto nla_put_failure; @@ -1924,6 +1935,11 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, goto nla_put_failure; rcu_read_lock(); + if (rtnl_fill_link_netnsid(skb, dev, src_net, GFP_ATOMIC)) + goto nla_put_failure_rcu; + qdisc = rcu_dereference(dev->qdisc); + if (qdisc && nla_put_string(skb, IFLA_QDISC, qdisc->ops->id)) + goto nla_put_failure_rcu; if (rtnl_fill_link_af(skb, dev, ext_filter_mask)) goto nla_put_failure_rcu; if (rtnl_fill_link_ifmap(skb, dev)) @@ -3272,7 +3288,7 @@ static int rtnl_dellink(struct sk_buff *skb, struct nlmsghdr *nlh, if (ifm->ifi_index > 0) dev = __dev_get_by_index(tgt_net, ifm->ifi_index); else if (tb[IFLA_IFNAME] || tb[IFLA_ALT_IFNAME]) - dev = rtnl_dev_get(net, tb); + dev = rtnl_dev_get(tgt_net, tb); else if (tb[IFLA_GROUP]) err = rtnl_group_dellink(tgt_net, nla_get_u32(tb[IFLA_GROUP])); else @@ -5245,15 +5261,14 @@ static int rtnl_bridge_dellink(struct sk_buff *skb, struct nlmsghdr *nlh, br_spec = nlmsg_find_attr(nlh, sizeof(struct ifinfomsg), IFLA_AF_SPEC); if (br_spec) { - nla_for_each_nested(attr, br_spec, rem) { - if (nla_type(attr) == IFLA_BRIDGE_FLAGS) { - if (nla_len(attr) < sizeof(flags)) - return -EINVAL; + nla_for_each_nested_type(attr, IFLA_BRIDGE_FLAGS, br_spec, + rem) { + if (nla_len(attr) < sizeof(flags)) + return -EINVAL; - have_flags = true; - flags = nla_get_u16(attr); - break; - } + have_flags = true; + flags = nla_get_u16(attr); + break; } } @@ -5962,19 +5977,17 @@ static int rtnl_stats_get(struct sk_buff *skb, struct nlmsghdr *nlh, static int rtnl_stats_dump(struct sk_buff *skb, struct netlink_callback *cb) { struct netlink_ext_ack *extack = cb->extack; - int h, s_h, err, s_idx, s_idxattr, s_prividx; struct rtnl_stats_dump_filters filters; struct net *net = sock_net(skb->sk); unsigned int flags = NLM_F_MULTI; struct if_stats_msg *ifsm; - struct hlist_head *head; + struct { + unsigned long ifindex; + int idxattr; + int prividx; + } *ctx = (void *)cb->ctx; struct net_device *dev; - int idx = 0; - - s_h = cb->args[0]; - s_idx = cb->args[1]; - s_idxattr = cb->args[2]; - s_prividx = cb->args[3]; + int err; cb->seq = net->dev_base_seq; @@ -5993,39 +6006,26 @@ static int rtnl_stats_dump(struct sk_buff *skb, struct netlink_callback *cb) if (err) return err; - for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) { - idx = 0; - head = &net->dev_index_head[h]; - hlist_for_each_entry(dev, head, index_hlist) { - if (idx < s_idx) - goto cont; - err = rtnl_fill_statsinfo(skb, dev, RTM_NEWSTATS, - NETLINK_CB(cb->skb).portid, - cb->nlh->nlmsg_seq, 0, - flags, &filters, - &s_idxattr, &s_prividx, - extack); - /* If we ran out of room on the first message, - * we're in trouble - */ - WARN_ON((err == -EMSGSIZE) && (skb->len == 0)); + for_each_netdev_dump(net, dev, ctx->ifindex) { + err = rtnl_fill_statsinfo(skb, dev, RTM_NEWSTATS, + NETLINK_CB(cb->skb).portid, + cb->nlh->nlmsg_seq, 0, + flags, &filters, + &ctx->idxattr, &ctx->prividx, + extack); + /* If we ran out of room on the first message, + * we're in trouble. + */ + WARN_ON((err == -EMSGSIZE) && (skb->len == 0)); - if (err < 0) - goto out; - s_prividx = 0; - s_idxattr = 0; - nl_dump_check_consistent(cb, nlmsg_hdr(skb)); -cont: - idx++; - } + if (err < 0) + break; + ctx->prividx = 0; + ctx->idxattr = 0; + nl_dump_check_consistent(cb, nlmsg_hdr(skb)); } -out: - cb->args[3] = s_prividx; - cb->args[2] = s_idxattr; - cb->args[1] = idx; - cb->args[0] = h; - return skb->len; + return err; } void rtnl_offload_xstats_notify(struct net_device *dev) diff --git a/net/core/scm.c b/net/core/scm.c index 9cd4b0a01c..4f6a14babe 100644 --- a/net/core/scm.c +++ b/net/core/scm.c @@ -89,6 +89,12 @@ static int scm_fp_copy(struct cmsghdr *cmsg, struct scm_fp_list **fplp) fpl->count_unix = 0; fpl->max = SCM_MAX_FD; fpl->user = NULL; +#if IS_ENABLED(CONFIG_UNIX) + fpl->inflight = false; + fpl->dead = false; + fpl->edges = NULL; + INIT_LIST_HEAD(&fpl->vertices); +#endif } fpp = &fpl->fp[fpl->count]; @@ -376,8 +382,14 @@ struct scm_fp_list *scm_fp_dup(struct scm_fp_list *fpl) if (new_fpl) { for (i = 0; i < fpl->count; i++) get_file(fpl->fp[i]); + new_fpl->max = new_fpl->count; new_fpl->user = get_uid(fpl->user); +#if IS_ENABLED(CONFIG_UNIX) + new_fpl->inflight = false; + new_fpl->edges = NULL; + INIT_LIST_HEAD(&new_fpl->vertices); +#endif } return new_fpl; } diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 4096e679f6..466999a751 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -51,6 +51,7 @@ #endif #include <linux/string.h> #include <linux/skbuff.h> +#include <linux/skbuff_ref.h> #include <linux/splice.h> #include <linux/cache.h> #include <linux/rtnetlink.h> @@ -108,9 +109,6 @@ static struct kmem_cache *skbuff_ext_cache __ro_after_init; #define SKB_SMALL_HEAD_HEADROOM \ SKB_WITH_OVERHEAD(SKB_SMALL_HEAD_CACHE_SIZE) -int sysctl_max_skb_frags __read_mostly = MAX_SKB_FRAGS; -EXPORT_SYMBOL(sysctl_max_skb_frags); - /* kcm_write_msgs() relies on casting paged frags to bio_vec to use * iov_iter_bvec(). These static asserts ensure the cast is valid is long as the * netmem is a page. @@ -775,10 +773,9 @@ skb_fail: EXPORT_SYMBOL(__netdev_alloc_skb); /** - * __napi_alloc_skb - allocate skbuff for rx in a specific NAPI instance + * napi_alloc_skb - allocate skbuff for rx in a specific NAPI instance * @napi: napi instance this buffer was allocated for * @len: length to allocate - * @gfp_mask: get_free_pages mask, passed to alloc_skb and alloc_pages * * Allocate a new sk_buff for use in NAPI receive. This buffer will * attempt to allocate the head from a special reserved region used @@ -787,9 +784,9 @@ EXPORT_SYMBOL(__netdev_alloc_skb); * * %NULL is returned if there is no free memory. */ -struct sk_buff *__napi_alloc_skb(struct napi_struct *napi, unsigned int len, - gfp_t gfp_mask) +struct sk_buff *napi_alloc_skb(struct napi_struct *napi, unsigned int len) { + gfp_t gfp_mask = GFP_ATOMIC | __GFP_NOWARN; struct napi_alloc_cache *nc; struct sk_buff *skb; bool pfmemalloc; @@ -860,7 +857,7 @@ skb_success: skb_fail: return skb; } -EXPORT_SYMBOL(__napi_alloc_skb); +EXPORT_SYMBOL(napi_alloc_skb); void skb_add_rx_frag_netmem(struct sk_buff *skb, int i, netmem_ref netmem, int off, int size, unsigned int truesize) @@ -1005,11 +1002,8 @@ int skb_cow_data_for_xdp(struct page_pool *pool, struct sk_buff **pskb, EXPORT_SYMBOL(skb_cow_data_for_xdp); #if IS_ENABLED(CONFIG_PAGE_POOL) -bool napi_pp_put_page(struct page *page, bool napi_safe) +bool napi_pp_put_page(struct page *page) { - bool allow_direct = false; - struct page_pool *pp; - page = compound_head(page); /* page->pp_magic is OR'ed with PP_SIGNATURE after the allocation @@ -1022,39 +1016,18 @@ bool napi_pp_put_page(struct page *page, bool napi_safe) if (unlikely(!is_pp_page(page))) return false; - pp = page->pp; - - /* Allow direct recycle if we have reasons to believe that we are - * in the same context as the consumer would run, so there's - * no possible race. - * __page_pool_put_page() makes sure we're not in hardirq context - * and interrupts are enabled prior to accessing the cache. - */ - if (napi_safe || in_softirq()) { - const struct napi_struct *napi = READ_ONCE(pp->p.napi); - unsigned int cpuid = smp_processor_id(); - - allow_direct = napi && READ_ONCE(napi->list_owner) == cpuid; - allow_direct |= READ_ONCE(pp->cpuid) == cpuid; - } - - /* Driver set this to memory recycling info. Reset it on recycle. - * This will *not* work for NIC using a split-page memory model. - * The page will be returned to the pool here regardless of the - * 'flipped' fragment being in use or not. - */ - page_pool_put_full_page(pp, page, allow_direct); + page_pool_put_full_page(page->pp, page, false); return true; } EXPORT_SYMBOL(napi_pp_put_page); #endif -static bool skb_pp_recycle(struct sk_buff *skb, void *data, bool napi_safe) +static bool skb_pp_recycle(struct sk_buff *skb, void *data) { if (!IS_ENABLED(CONFIG_PAGE_POOL) || !skb->pp_recycle) return false; - return napi_pp_put_page(virt_to_page(data), napi_safe); + return napi_pp_put_page(virt_to_page(data)); } /** @@ -1096,12 +1069,12 @@ static void skb_kfree_head(void *head, unsigned int end_offset) kfree(head); } -static void skb_free_head(struct sk_buff *skb, bool napi_safe) +static void skb_free_head(struct sk_buff *skb) { unsigned char *head = skb->head; if (skb->head_frag) { - if (skb_pp_recycle(skb, head, napi_safe)) + if (skb_pp_recycle(skb, head)) return; skb_free_frag(head); } else { @@ -1109,8 +1082,7 @@ static void skb_free_head(struct sk_buff *skb, bool napi_safe) } } -static void skb_release_data(struct sk_buff *skb, enum skb_drop_reason reason, - bool napi_safe) +static void skb_release_data(struct sk_buff *skb, enum skb_drop_reason reason) { struct skb_shared_info *shinfo = skb_shinfo(skb); int i; @@ -1127,13 +1099,13 @@ static void skb_release_data(struct sk_buff *skb, enum skb_drop_reason reason, } for (i = 0; i < shinfo->nr_frags; i++) - napi_frag_unref(&shinfo->frags[i], skb->pp_recycle, napi_safe); + __skb_frag_unref(&shinfo->frags[i], skb->pp_recycle); free_head: if (shinfo->frag_list) kfree_skb_list_reason(shinfo->frag_list, reason); - skb_free_head(skb, napi_safe); + skb_free_head(skb); exit: /* When we clone an SKB we copy the reycling bit. The pp_recycle * bit is only set on the head though, so in order to avoid races @@ -1194,12 +1166,11 @@ void skb_release_head_state(struct sk_buff *skb) } /* Free everything but the sk_buff shell. */ -static void skb_release_all(struct sk_buff *skb, enum skb_drop_reason reason, - bool napi_safe) +static void skb_release_all(struct sk_buff *skb, enum skb_drop_reason reason) { skb_release_head_state(skb); if (likely(skb->head)) - skb_release_data(skb, reason, napi_safe); + skb_release_data(skb, reason); } /** @@ -1213,7 +1184,7 @@ static void skb_release_all(struct sk_buff *skb, enum skb_drop_reason reason, void __kfree_skb(struct sk_buff *skb) { - skb_release_all(skb, SKB_DROP_REASON_NOT_SPECIFIED, false); + skb_release_all(skb, SKB_DROP_REASON_NOT_SPECIFIED); kfree_skbmem(skb); } EXPORT_SYMBOL(__kfree_skb); @@ -1270,7 +1241,7 @@ static void kfree_skb_add_bulk(struct sk_buff *skb, return; } - skb_release_all(skb, reason, false); + skb_release_all(skb, reason); sa->skb_array[sa->skb_count++] = skb; if (unlikely(sa->skb_count == KFREE_SKB_BULK_SIZE)) { @@ -1331,22 +1302,28 @@ void skb_dump(const char *level, const struct sk_buff *skb, bool full_pkt) has_trans = skb_transport_header_was_set(skb); printk("%sskb len=%u headroom=%u headlen=%u tailroom=%u\n" - "mac=(%d,%d) net=(%d,%d) trans=%d\n" + "mac=(%d,%d) mac_len=%u net=(%d,%d) trans=%d\n" "shinfo(txflags=%u nr_frags=%u gso(size=%hu type=%u segs=%hu))\n" - "csum(0x%x ip_summed=%u complete_sw=%u valid=%u level=%u)\n" - "hash(0x%x sw=%u l4=%u) proto=0x%04x pkttype=%u iif=%d\n", + "csum(0x%x start=%u offset=%u ip_summed=%u complete_sw=%u valid=%u level=%u)\n" + "hash(0x%x sw=%u l4=%u) proto=0x%04x pkttype=%u iif=%d\n" + "priority=0x%x mark=0x%x alloc_cpu=%u vlan_all=0x%x\n" + "encapsulation=%d inner(proto=0x%04x, mac=%u, net=%u, trans=%u)\n", level, skb->len, headroom, skb_headlen(skb), tailroom, has_mac ? skb->mac_header : -1, has_mac ? skb_mac_header_len(skb) : -1, + skb->mac_len, skb->network_header, has_trans ? skb_network_header_len(skb) : -1, has_trans ? skb->transport_header : -1, sh->tx_flags, sh->nr_frags, sh->gso_size, sh->gso_type, sh->gso_segs, - skb->csum, skb->ip_summed, skb->csum_complete_sw, - skb->csum_valid, skb->csum_level, + skb->csum, skb->csum_start, skb->csum_offset, skb->ip_summed, + skb->csum_complete_sw, skb->csum_valid, skb->csum_level, skb->hash, skb->sw_hash, skb->l4_hash, - ntohs(skb->protocol), skb->pkt_type, skb->skb_iif); + ntohs(skb->protocol), skb->pkt_type, skb->skb_iif, + skb->priority, skb->mark, skb->alloc_cpu, skb->vlan_all, + skb->encapsulation, skb->inner_protocol, skb->inner_mac_header, + skb->inner_network_header, skb->inner_transport_header); if (dev) printk("%sdev name=%s feat=%pNF\n", @@ -1444,7 +1421,7 @@ EXPORT_SYMBOL(consume_skb); void __consume_stateless_skb(struct sk_buff *skb) { trace_consume_skb(skb, __builtin_return_address(0)); - skb_release_data(skb, SKB_CONSUMED, false); + skb_release_data(skb, SKB_CONSUMED); kfree_skbmem(skb); } @@ -1471,7 +1448,7 @@ static void napi_skb_cache_put(struct sk_buff *skb) void __napi_kfree_skb(struct sk_buff *skb, enum skb_drop_reason reason) { - skb_release_all(skb, reason, true); + skb_release_all(skb, reason); napi_skb_cache_put(skb); } @@ -1509,7 +1486,7 @@ void napi_consume_skb(struct sk_buff *skb, int budget) return; } - skb_release_all(skb, SKB_CONSUMED, !!budget); + skb_release_all(skb, SKB_CONSUMED); napi_skb_cache_put(skb); } EXPORT_SYMBOL(napi_consume_skb); @@ -1640,7 +1617,7 @@ EXPORT_SYMBOL_GPL(alloc_skb_for_msg); */ struct sk_buff *skb_morph(struct sk_buff *dst, struct sk_buff *src) { - skb_release_all(dst, SKB_CONSUMED, false); + skb_release_all(dst, SKB_CONSUMED); return __skb_clone(dst, src); } EXPORT_SYMBOL_GPL(skb_morph); @@ -1708,7 +1685,7 @@ static struct ubuf_info *msg_zerocopy_alloc(struct sock *sk, size_t size) return NULL; } - uarg->ubuf.callback = msg_zerocopy_callback; + uarg->ubuf.ops = &msg_zerocopy_ubuf_ops; uarg->id = ((u32)atomic_inc_return(&sk->sk_zckey)) - 1; uarg->len = 1; uarg->bytelen = size; @@ -1734,7 +1711,7 @@ struct ubuf_info *msg_zerocopy_realloc(struct sock *sk, size_t size, u32 bytelen, next; /* there might be non MSG_ZEROCOPY users */ - if (uarg->callback != msg_zerocopy_callback) + if (uarg->ops != &msg_zerocopy_ubuf_ops) return NULL; /* realloc only when socket is locked (TCP, UDP cork), @@ -1845,8 +1822,8 @@ release: sock_put(sk); } -void msg_zerocopy_callback(struct sk_buff *skb, struct ubuf_info *uarg, - bool success) +static void msg_zerocopy_complete(struct sk_buff *skb, struct ubuf_info *uarg, + bool success) { struct ubuf_info_msgzc *uarg_zc = uarg_to_msgzc(uarg); @@ -1855,7 +1832,6 @@ void msg_zerocopy_callback(struct sk_buff *skb, struct ubuf_info *uarg, if (refcount_dec_and_test(&uarg->refcnt)) __msg_zerocopy_callback(uarg_zc); } -EXPORT_SYMBOL_GPL(msg_zerocopy_callback); void msg_zerocopy_put_abort(struct ubuf_info *uarg, bool have_uref) { @@ -1865,10 +1841,15 @@ void msg_zerocopy_put_abort(struct ubuf_info *uarg, bool have_uref) uarg_to_msgzc(uarg)->len--; if (have_uref) - msg_zerocopy_callback(NULL, uarg, true); + msg_zerocopy_complete(NULL, uarg, true); } EXPORT_SYMBOL_GPL(msg_zerocopy_put_abort); +const struct ubuf_info_ops msg_zerocopy_ubuf_ops = { + .complete = msg_zerocopy_complete, +}; +EXPORT_SYMBOL_GPL(msg_zerocopy_ubuf_ops); + int skb_zerocopy_iter_stream(struct sock *sk, struct sk_buff *skb, struct msghdr *msg, int len, struct ubuf_info *uarg) @@ -1876,11 +1857,18 @@ int skb_zerocopy_iter_stream(struct sock *sk, struct sk_buff *skb, struct ubuf_info *orig_uarg = skb_zcopy(skb); int err, orig_len = skb->len; - /* An skb can only point to one uarg. This edge case happens when - * TCP appends to an skb, but zerocopy_realloc triggered a new alloc. - */ - if (orig_uarg && uarg != orig_uarg) - return -EEXIST; + if (uarg->ops->link_skb) { + err = uarg->ops->link_skb(skb, uarg); + if (err) + return err; + } else { + /* An skb can only point to one uarg. This edge case happens + * when TCP appends to an skb, but zerocopy_realloc triggered + * a new alloc. + */ + if (orig_uarg && uarg != orig_uarg) + return -EEXIST; + } err = __zerocopy_sg_from_iter(msg, sk, skb, &msg->msg_iter, len); if (err == -EFAULT || (err == -EMSGSIZE && skb->len == orig_len)) { @@ -1894,7 +1882,8 @@ int skb_zerocopy_iter_stream(struct sock *sk, struct sk_buff *skb, return err; } - skb_zcopy_set(skb, uarg, NULL); + if (!uarg->ops->link_skb) + skb_zcopy_set(skb, uarg, NULL); return skb->len - orig_len; } EXPORT_SYMBOL_GPL(skb_zerocopy_iter_stream); @@ -2278,9 +2267,9 @@ int pskb_expand_head(struct sk_buff *skb, int nhead, int ntail, if (skb_has_frag_list(skb)) skb_clone_fraglist(skb); - skb_release_data(skb, SKB_CONSUMED, false); + skb_release_data(skb, SKB_CONSUMED); } else { - skb_free_head(skb, false); + skb_free_head(skb); } off = (data + nhead) - skb->head; @@ -6586,12 +6575,12 @@ static int pskb_carve_inside_header(struct sk_buff *skb, const u32 off, skb_frag_ref(skb, i); if (skb_has_frag_list(skb)) skb_clone_fraglist(skb); - skb_release_data(skb, SKB_CONSUMED, false); + skb_release_data(skb, SKB_CONSUMED); } else { /* we can reuse existing recount- all we did was * relocate values */ - skb_free_head(skb, false); + skb_free_head(skb); } skb->head = data; @@ -6726,7 +6715,7 @@ static int pskb_carve_inside_nonlinear(struct sk_buff *skb, const u32 off, skb_kfree_head(data, size); return -ENOMEM; } - skb_release_data(skb, SKB_CONSUMED, false); + skb_release_data(skb, SKB_CONSUMED); skb->head = data; skb->head_frag = 0; @@ -7006,6 +6995,19 @@ free_now: EXPORT_SYMBOL(__skb_ext_put); #endif /* CONFIG_SKB_EXTENSIONS */ +static void kfree_skb_napi_cache(struct sk_buff *skb) +{ + /* if SKB is a clone, don't handle this case */ + if (skb->fclone != SKB_FCLONE_UNAVAILABLE) { + __kfree_skb(skb); + return; + } + + local_bh_disable(); + __napi_kfree_skb(skb, SKB_CONSUMED); + local_bh_enable(); +} + /** * skb_attempt_defer_free - queue skb for remote freeing * @skb: buffer @@ -7021,10 +7023,10 @@ void skb_attempt_defer_free(struct sk_buff *skb) unsigned int defer_max; bool kick; - if (WARN_ON_ONCE(cpu >= nr_cpu_ids) || - !cpu_online(cpu) || - cpu == raw_smp_processor_id()) { -nodefer: __kfree_skb(skb); + if (cpu == raw_smp_processor_id() || + WARN_ON_ONCE(cpu >= nr_cpu_ids) || + !cpu_online(cpu)) { +nodefer: kfree_skb_napi_cache(skb); return; } @@ -7032,7 +7034,7 @@ nodefer: __kfree_skb(skb); DEBUG_NET_WARN_ON_ONCE(skb->destructor); sd = &per_cpu(softnet_data, cpu); - defer_max = READ_ONCE(sysctl_skb_defer_max); + defer_max = READ_ONCE(net_hotdata.sysctl_skb_defer_max); if (READ_ONCE(sd->defer_count) >= defer_max) goto nodefer; @@ -7050,8 +7052,8 @@ nodefer: __kfree_skb(skb); /* Make sure to trigger NET_RX_SOFTIRQ on the remote CPU * if we are unlucky enough (this seems very unlikely). */ - if (unlikely(kick) && !cmpxchg(&sd->defer_ipi_scheduled, 0, 1)) - smp_call_function_single_async(cpu, &sd->defer_csd); + if (unlikely(kick)) + kick_defer_list_purge(sd, cpu); } static void skb_splice_csum_page(struct sk_buff *skb, struct page *page, @@ -7084,7 +7086,7 @@ static void skb_splice_csum_page(struct sk_buff *skb, struct page *page, ssize_t skb_splice_from_iter(struct sk_buff *skb, struct iov_iter *iter, ssize_t maxsize, gfp_t gfp) { - size_t frag_limit = READ_ONCE(sysctl_max_skb_frags); + size_t frag_limit = READ_ONCE(net_hotdata.sysctl_max_skb_frags); struct page *pages[8], **ppages = pages; ssize_t spliced = 0, ret = 0; unsigned int i; diff --git a/net/core/skmsg.c b/net/core/skmsg.c index fd20aae30b..bbf40b9997 100644 --- a/net/core/skmsg.c +++ b/net/core/skmsg.c @@ -434,7 +434,8 @@ int sk_msg_recvmsg(struct sock *sk, struct sk_psock *psock, struct msghdr *msg, page = sg_page(sge); if (copied + copy > len) copy = len - copied; - copy = copy_page_to_iter(page, sge->offset, copy, iter); + if (copy) + copy = copy_page_to_iter(page, sge->offset, copy, iter); if (!copy) { copied = copied ? copied : -EFAULT; goto out; diff --git a/net/core/sock.c b/net/core/sock.c index 09eccc9c50..100e975073 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -127,6 +127,7 @@ #include <net/net_namespace.h> #include <net/request_sock.h> #include <net/sock.h> +#include <net/proto_memory.h> #include <linux/net_tstamp.h> #include <net/xfrm.h> #include <linux/ipsec.h> @@ -283,7 +284,6 @@ __u32 sysctl_rmem_max __read_mostly = SK_RMEM_MAX; EXPORT_SYMBOL(sysctl_rmem_max); __u32 sysctl_wmem_default __read_mostly = SK_WMEM_MAX; __u32 sysctl_rmem_default __read_mostly = SK_RMEM_MAX; -int sysctl_mem_pcpu_rsv __read_mostly = SK_MEMORY_PCPU_RESERVE; int sysctl_tstamp_allow_data __read_mostly = 1; @@ -2526,13 +2526,12 @@ EXPORT_SYMBOL(skb_set_owner_w); static bool can_skb_orphan_partial(const struct sk_buff *skb) { -#ifdef CONFIG_TLS_DEVICE /* Drivers depend on in-order delivery for crypto offload, * partial orphan breaks out-of-order-OK logic. */ - if (skb->decrypted) + if (skb_is_decrypted(skb)) return false; -#endif + return (skb->destructor == sock_wfree || (IS_ENABLED(CONFIG_INET) && skb->destructor == tcp_wfree)); } @@ -3242,8 +3241,8 @@ int sock_no_socketpair(struct socket *sock1, struct socket *sock2) } EXPORT_SYMBOL(sock_no_socketpair); -int sock_no_accept(struct socket *sock, struct socket *newsock, int flags, - bool kern) +int sock_no_accept(struct socket *sock, struct socket *newsock, + struct proto_accept_arg *arg) { return -EOPNOTSUPP; } @@ -3338,7 +3337,7 @@ static void sock_def_error_report(struct sock *sk) wq = rcu_dereference(sk->sk_wq); if (skwq_has_sleeper(wq)) wake_up_interruptible_poll(&wq->wait, EPOLLERR); - sk_wake_async(sk, SOCK_WAKE_IO, POLL_ERR); + sk_wake_async_rcu(sk, SOCK_WAKE_IO, POLL_ERR); rcu_read_unlock(); } @@ -3353,7 +3352,7 @@ void sock_def_readable(struct sock *sk) if (skwq_has_sleeper(wq)) wake_up_interruptible_sync_poll(&wq->wait, EPOLLIN | EPOLLPRI | EPOLLRDNORM | EPOLLRDBAND); - sk_wake_async(sk, SOCK_WAKE_WAITD, POLL_IN); + sk_wake_async_rcu(sk, SOCK_WAKE_WAITD, POLL_IN); rcu_read_unlock(); } @@ -3373,7 +3372,7 @@ static void sock_def_write_space(struct sock *sk) EPOLLWRNORM | EPOLLWRBAND); /* Should agree with poll, otherwise some programs break */ - sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT); + sk_wake_async_rcu(sk, SOCK_WAKE_SPACE, POLL_OUT); } rcu_read_unlock(); @@ -3398,7 +3397,7 @@ static void sock_def_write_space_wfree(struct sock *sk) EPOLLWRNORM | EPOLLWRBAND); /* Should agree with poll, otherwise some programs break */ - sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT); + sk_wake_async_rcu(sk, SOCK_WAKE_SPACE, POLL_OUT); } } diff --git a/net/core/sock_map.c b/net/core/sock_map.c index 01be07b485..d3dbb92153 100644 --- a/net/core/sock_map.c +++ b/net/core/sock_map.c @@ -24,8 +24,16 @@ struct bpf_stab { #define SOCK_CREATE_FLAG_MASK \ (BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY) +/* This mutex is used to + * - protect race between prog/link attach/detach and link prog update, and + * - protect race between releasing and accessing map in bpf_link. + * A single global mutex lock is used since it is expected contention is low. + */ +static DEFINE_MUTEX(sockmap_mutex); + static int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog, - struct bpf_prog *old, u32 which); + struct bpf_prog *old, struct bpf_link *link, + u32 which); static struct sk_psock_progs *sock_map_progs(struct bpf_map *map); static struct bpf_map *sock_map_alloc(union bpf_attr *attr) @@ -71,7 +79,9 @@ int sock_map_get_from_fd(const union bpf_attr *attr, struct bpf_prog *prog) map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); - ret = sock_map_prog_update(map, prog, NULL, attr->attach_type); + mutex_lock(&sockmap_mutex); + ret = sock_map_prog_update(map, prog, NULL, NULL, attr->attach_type); + mutex_unlock(&sockmap_mutex); fdput(f); return ret; } @@ -103,7 +113,9 @@ int sock_map_prog_detach(const union bpf_attr *attr, enum bpf_prog_type ptype) goto put_prog; } - ret = sock_map_prog_update(map, NULL, prog, attr->attach_type); + mutex_lock(&sockmap_mutex); + ret = sock_map_prog_update(map, NULL, prog, NULL, attr->attach_type); + mutex_unlock(&sockmap_mutex); put_prog: bpf_prog_put(prog); put_map: @@ -411,9 +423,6 @@ static int __sock_map_delete(struct bpf_stab *stab, struct sock *sk_test, struct sock *sk; int err = 0; - if (irqs_disabled()) - return -EOPNOTSUPP; /* locks here are hardirq-unsafe */ - spin_lock_bh(&stab->lock); sk = *psk; if (!sk_test || sk_test == sk) @@ -936,9 +945,6 @@ static long sock_hash_delete_elem(struct bpf_map *map, void *key) struct bpf_shtab_elem *elem; int ret = -ENOENT; - if (irqs_disabled()) - return -EOPNOTSUPP; /* locks here are hardirq-unsafe */ - hash = sock_hash_bucket_hash(key, key_size); bucket = sock_hash_select_bucket(htab, hash); @@ -1460,55 +1466,84 @@ static struct sk_psock_progs *sock_map_progs(struct bpf_map *map) return NULL; } -static int sock_map_prog_lookup(struct bpf_map *map, struct bpf_prog ***pprog, - u32 which) +static int sock_map_prog_link_lookup(struct bpf_map *map, struct bpf_prog ***pprog, + struct bpf_link ***plink, u32 which) { struct sk_psock_progs *progs = sock_map_progs(map); + struct bpf_prog **cur_pprog; + struct bpf_link **cur_plink; if (!progs) return -EOPNOTSUPP; switch (which) { case BPF_SK_MSG_VERDICT: - *pprog = &progs->msg_parser; + cur_pprog = &progs->msg_parser; + cur_plink = &progs->msg_parser_link; break; #if IS_ENABLED(CONFIG_BPF_STREAM_PARSER) case BPF_SK_SKB_STREAM_PARSER: - *pprog = &progs->stream_parser; + cur_pprog = &progs->stream_parser; + cur_plink = &progs->stream_parser_link; break; #endif case BPF_SK_SKB_STREAM_VERDICT: if (progs->skb_verdict) return -EBUSY; - *pprog = &progs->stream_verdict; + cur_pprog = &progs->stream_verdict; + cur_plink = &progs->stream_verdict_link; break; case BPF_SK_SKB_VERDICT: if (progs->stream_verdict) return -EBUSY; - *pprog = &progs->skb_verdict; + cur_pprog = &progs->skb_verdict; + cur_plink = &progs->skb_verdict_link; break; default: return -EOPNOTSUPP; } + *pprog = cur_pprog; + if (plink) + *plink = cur_plink; return 0; } +/* Handle the following four cases: + * prog_attach: prog != NULL, old == NULL, link == NULL + * prog_detach: prog == NULL, old != NULL, link == NULL + * link_attach: prog != NULL, old == NULL, link != NULL + * link_detach: prog == NULL, old != NULL, link != NULL + */ static int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog, - struct bpf_prog *old, u32 which) + struct bpf_prog *old, struct bpf_link *link, + u32 which) { struct bpf_prog **pprog; + struct bpf_link **plink; int ret; - ret = sock_map_prog_lookup(map, &pprog, which); + ret = sock_map_prog_link_lookup(map, &pprog, &plink, which); if (ret) return ret; - if (old) - return psock_replace_prog(pprog, prog, old); + /* for prog_attach/prog_detach/link_attach, return error if a bpf_link + * exists for that prog. + */ + if ((!link || prog) && *plink) + return -EBUSY; - psock_set_prog(pprog, prog); - return 0; + if (old) { + ret = psock_replace_prog(pprog, prog, old); + if (!ret) + *plink = NULL; + } else { + psock_set_prog(pprog, prog); + if (link) + *plink = link; + } + + return ret; } int sock_map_bpf_prog_query(const union bpf_attr *attr, @@ -1533,7 +1568,7 @@ int sock_map_bpf_prog_query(const union bpf_attr *attr, rcu_read_lock(); - ret = sock_map_prog_lookup(map, &pprog, attr->query.attach_type); + ret = sock_map_prog_link_lookup(map, &pprog, NULL, attr->query.attach_type); if (ret) goto end; @@ -1667,6 +1702,196 @@ no_psock: } EXPORT_SYMBOL_GPL(sock_map_close); +struct sockmap_link { + struct bpf_link link; + struct bpf_map *map; + enum bpf_attach_type attach_type; +}; + +static void sock_map_link_release(struct bpf_link *link) +{ + struct sockmap_link *sockmap_link = container_of(link, struct sockmap_link, link); + + mutex_lock(&sockmap_mutex); + if (!sockmap_link->map) + goto out; + + WARN_ON_ONCE(sock_map_prog_update(sockmap_link->map, NULL, link->prog, link, + sockmap_link->attach_type)); + + bpf_map_put_with_uref(sockmap_link->map); + sockmap_link->map = NULL; +out: + mutex_unlock(&sockmap_mutex); +} + +static int sock_map_link_detach(struct bpf_link *link) +{ + sock_map_link_release(link); + return 0; +} + +static void sock_map_link_dealloc(struct bpf_link *link) +{ + kfree(link); +} + +/* Handle the following two cases: + * case 1: link != NULL, prog != NULL, old != NULL + * case 2: link != NULL, prog != NULL, old == NULL + */ +static int sock_map_link_update_prog(struct bpf_link *link, + struct bpf_prog *prog, + struct bpf_prog *old) +{ + const struct sockmap_link *sockmap_link = container_of(link, struct sockmap_link, link); + struct bpf_prog **pprog, *old_link_prog; + struct bpf_link **plink; + int ret = 0; + + mutex_lock(&sockmap_mutex); + + /* If old prog is not NULL, ensure old prog is the same as link->prog. */ + if (old && link->prog != old) { + ret = -EPERM; + goto out; + } + /* Ensure link->prog has the same type/attach_type as the new prog. */ + if (link->prog->type != prog->type || + link->prog->expected_attach_type != prog->expected_attach_type) { + ret = -EINVAL; + goto out; + } + + ret = sock_map_prog_link_lookup(sockmap_link->map, &pprog, &plink, + sockmap_link->attach_type); + if (ret) + goto out; + + /* return error if the stored bpf_link does not match the incoming bpf_link. */ + if (link != *plink) { + ret = -EBUSY; + goto out; + } + + if (old) { + ret = psock_replace_prog(pprog, prog, old); + if (ret) + goto out; + } else { + psock_set_prog(pprog, prog); + } + + bpf_prog_inc(prog); + old_link_prog = xchg(&link->prog, prog); + bpf_prog_put(old_link_prog); + +out: + mutex_unlock(&sockmap_mutex); + return ret; +} + +static u32 sock_map_link_get_map_id(const struct sockmap_link *sockmap_link) +{ + u32 map_id = 0; + + mutex_lock(&sockmap_mutex); + if (sockmap_link->map) + map_id = sockmap_link->map->id; + mutex_unlock(&sockmap_mutex); + return map_id; +} + +static int sock_map_link_fill_info(const struct bpf_link *link, + struct bpf_link_info *info) +{ + const struct sockmap_link *sockmap_link = container_of(link, struct sockmap_link, link); + u32 map_id = sock_map_link_get_map_id(sockmap_link); + + info->sockmap.map_id = map_id; + info->sockmap.attach_type = sockmap_link->attach_type; + return 0; +} + +static void sock_map_link_show_fdinfo(const struct bpf_link *link, + struct seq_file *seq) +{ + const struct sockmap_link *sockmap_link = container_of(link, struct sockmap_link, link); + u32 map_id = sock_map_link_get_map_id(sockmap_link); + + seq_printf(seq, "map_id:\t%u\n", map_id); + seq_printf(seq, "attach_type:\t%u\n", sockmap_link->attach_type); +} + +static const struct bpf_link_ops sock_map_link_ops = { + .release = sock_map_link_release, + .dealloc = sock_map_link_dealloc, + .detach = sock_map_link_detach, + .update_prog = sock_map_link_update_prog, + .fill_link_info = sock_map_link_fill_info, + .show_fdinfo = sock_map_link_show_fdinfo, +}; + +int sock_map_link_create(const union bpf_attr *attr, struct bpf_prog *prog) +{ + struct bpf_link_primer link_primer; + struct sockmap_link *sockmap_link; + enum bpf_attach_type attach_type; + struct bpf_map *map; + int ret; + + if (attr->link_create.flags) + return -EINVAL; + + map = bpf_map_get_with_uref(attr->link_create.target_fd); + if (IS_ERR(map)) + return PTR_ERR(map); + if (map->map_type != BPF_MAP_TYPE_SOCKMAP && map->map_type != BPF_MAP_TYPE_SOCKHASH) { + ret = -EINVAL; + goto out; + } + + sockmap_link = kzalloc(sizeof(*sockmap_link), GFP_USER); + if (!sockmap_link) { + ret = -ENOMEM; + goto out; + } + + attach_type = attr->link_create.attach_type; + bpf_link_init(&sockmap_link->link, BPF_LINK_TYPE_SOCKMAP, &sock_map_link_ops, prog); + sockmap_link->map = map; + sockmap_link->attach_type = attach_type; + + ret = bpf_link_prime(&sockmap_link->link, &link_primer); + if (ret) { + kfree(sockmap_link); + goto out; + } + + mutex_lock(&sockmap_mutex); + ret = sock_map_prog_update(map, prog, NULL, &sockmap_link->link, attach_type); + mutex_unlock(&sockmap_mutex); + if (ret) { + bpf_link_cleanup(&link_primer); + goto out; + } + + /* Increase refcnt for the prog since when old prog is replaced with + * psock_replace_prog() and psock_set_prog() its refcnt will be decreased. + * + * Actually, we do not need to increase refcnt for the prog since bpf_link + * will hold a reference. But in order to have less complexity w.r.t. + * replacing/setting prog, let us increase the refcnt to make things simpler. + */ + bpf_prog_inc(prog); + + return bpf_link_settle(&link_primer); + +out: + bpf_map_put_with_uref(map); + return ret; +} + static int sock_map_iter_attach_target(struct bpf_prog *prog, union bpf_iter_link_info *linfo, struct bpf_iter_aux_info *aux) diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c index 6973dda3ab..c9fb9ad874 100644 --- a/net/core/sysctl_net_core.c +++ b/net/core/sysctl_net_core.c @@ -24,6 +24,7 @@ #include <net/busy_poll.h> #include <net/pkt_sched.h> #include <net/hotdata.h> +#include <net/proto_memory.h> #include <net/rps.h> #include "dev.h" @@ -415,7 +416,7 @@ static struct ctl_table net_core_table[] = { }, { .procname = "mem_pcpu_rsv", - .data = &sysctl_mem_pcpu_rsv, + .data = &net_hotdata.sysctl_mem_pcpu_rsv, .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, @@ -595,7 +596,7 @@ static struct ctl_table net_core_table[] = { }, { .procname = "max_skb_frags", - .data = &sysctl_max_skb_frags, + .data = &net_hotdata.sysctl_max_skb_frags, .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, @@ -654,13 +655,12 @@ static struct ctl_table net_core_table[] = { }, { .procname = "skb_defer_max", - .data = &sysctl_skb_defer_max, + .data = &net_hotdata.sysctl_skb_defer_max, .maxlen = sizeof(unsigned int), .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, }, - { } }; static struct ctl_table netns_core_table[] = { @@ -697,7 +697,6 @@ static struct ctl_table netns_core_table[] = { .extra2 = SYSCTL_ONE, .proc_handler = proc_dou8vec_minmax, }, - { } }; static int __init fb_tunnels_only_for_init_net_sysctl_setup(char *str) @@ -715,20 +714,21 @@ __setup("fb_tunnels=", fb_tunnels_only_for_init_net_sysctl_setup); static __net_init int sysctl_core_net_init(struct net *net) { - struct ctl_table *tbl, *tmp; + size_t table_size = ARRAY_SIZE(netns_core_table); + struct ctl_table *tbl; tbl = netns_core_table; if (!net_eq(net, &init_net)) { + int i; tbl = kmemdup(tbl, sizeof(netns_core_table), GFP_KERNEL); if (tbl == NULL) goto err_dup; - for (tmp = tbl; tmp->procname; tmp++) - tmp->data += (char *)net - (char *)&init_net; + for (i = 0; i < table_size; ++i) + tbl[i].data += (char *)net - (char *)&init_net; } - net->core.sysctl_hdr = register_net_sysctl_sz(net, "net/core", tbl, - ARRAY_SIZE(netns_core_table)); + net->core.sysctl_hdr = register_net_sysctl_sz(net, "net/core", tbl, table_size); if (net->core.sysctl_hdr == NULL) goto err_reg; @@ -743,7 +743,7 @@ err_dup: static __net_exit void sysctl_core_net_exit(struct net *net) { - struct ctl_table *tbl; + const struct ctl_table *tbl; tbl = net->core.sysctl_hdr->ctl_table_arg; unregister_net_sysctl_table(net->core.sysctl_hdr); diff --git a/net/core/xdp.c b/net/core/xdp.c index 41693154e4..bcc5551c64 100644 --- a/net/core/xdp.c +++ b/net/core/xdp.c @@ -127,10 +127,8 @@ void xdp_unreg_mem_model(struct xdp_mem_info *mem) return; if (type == MEM_TYPE_PAGE_POOL) { - rcu_read_lock(); - xa = rhashtable_lookup(mem_id_ht, &id, mem_id_rht_params); + xa = rhashtable_lookup_fast(mem_id_ht, &id, mem_id_rht_params); page_pool_destroy(xa->page_pool); - rcu_read_unlock(); } } EXPORT_SYMBOL_GPL(xdp_unreg_mem_model); @@ -295,10 +293,8 @@ static struct xdp_mem_allocator *__xdp_reg_mem_model(struct xdp_mem_info *mem, mutex_lock(&mem_id_lock); ret = __mem_id_init_hash_table(); mutex_unlock(&mem_id_lock); - if (ret < 0) { - WARN_ON(1); + if (ret < 0) return ERR_PTR(ret); - } } xdp_alloc = kzalloc(sizeof(*xdp_alloc), gfp); |