diff options
Diffstat (limited to 'net/core/dev.c')
| -rw-r--r-- | net/core/dev.c | 4404 |
1 files changed, 3119 insertions, 1285 deletions
diff --git a/net/core/dev.c b/net/core/dev.c index b76fb37b381e..9094c0fb8c68 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -69,7 +69,7 @@ */ #include <linux/uaccess.h> -#include <linux/bitops.h> +#include <linux/bitmap.h> #include <linux/capability.h> #include <linux/cpu.h> #include <linux/types.h> @@ -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> @@ -90,6 +92,7 @@ #include <linux/netdevice.h> #include <linux/etherdevice.h> #include <linux/ethtool.h> +#include <linux/ethtool_netlink.h> #include <linux/skbuff.h> #include <linux/kthread.h> #include <linux/bpf.h> @@ -103,10 +106,12 @@ #include <net/dst.h> #include <net/dst_metadata.h> #include <net/gro.h> +#include <net/netdev_queues.h> #include <net/pkt_sched.h> #include <net/pkt_cls.h> #include <net/checksum.h> #include <net/xfrm.h> +#include <net/tcx.h> #include <linux/highmem.h> #include <linux/init.h> #include <linux/module.h> @@ -132,6 +137,7 @@ #include <trace/events/net.h> #include <trace/events/skb.h> #include <trace/events/qdisc.h> +#include <trace/events/xdp.h> #include <linux/inetdevice.h> #include <linux/cpu_rmap.h> #include <linux/static_key.h> @@ -150,44 +156,25 @@ #include <linux/pm_runtime.h> #include <linux/prandom.h> #include <linux/once_lite.h> +#include <net/netdev_lock.h> +#include <net/netdev_rx_queue.h> +#include <net/page_pool/types.h> +#include <net/page_pool/helpers.h> +#include <net/page_pool/memory_provider.h> +#include <net/rps.h> +#include <linux/phy_link_topology.h> #include "dev.h" +#include "devmem.h" #include "net-sysfs.h" - static DEFINE_SPINLOCK(ptype_lock); struct list_head ptype_base[PTYPE_HASH_SIZE] __read_mostly; -struct list_head ptype_all __read_mostly; /* Taps */ static int netif_rx_internal(struct sk_buff *skb); -static int call_netdevice_notifiers_info(unsigned long val, - struct netdev_notifier_info *info); static int call_netdevice_notifiers_extack(unsigned long val, struct net_device *dev, struct netlink_ext_ack *extack); -static struct napi_struct *napi_by_id(unsigned int napi_id); - -/* - * The @dev_base_head list is protected by @dev_base_lock and the rtnl - * semaphore. - * - * Pure readers hold dev_base_lock for reading, or rcu_read_lock() - * - * Writers must hold the rtnl semaphore while they loop through the - * dev_base_head list, and hold dev_base_lock for writing when they do the - * actual updates. This allows pure readers to access the list even - * while a writer is preparing to update it. - * - * To put it another way, dev_base_lock is held for writing only to - * protect against pure readers; the rtnl semaphore provides the - * protection against other writers. - * - * See, for example usages, register_netdevice() and - * unregister_netdevice(), which must be called with the rtnl - * semaphore held. - */ -DEFINE_RWLOCK(dev_base_lock); -EXPORT_SYMBOL(dev_base_lock); static DEFINE_MUTEX(ifalias_mutex); @@ -197,12 +184,11 @@ static DEFINE_SPINLOCK(napi_hash_lock); static unsigned int napi_gen_id = NR_CPUS; static DEFINE_READ_MOSTLY_HASHTABLE(napi_hash, 8); -static DECLARE_RWSEM(devnet_rename_sem); - static inline void dev_base_seq_inc(struct net *net) { - while (++net->dev_base_seq == 0) - ; + unsigned int val = net->dev_base_seq + 1; + + WRITE_ONCE(net->dev_base_seq, val ?: 1); } static inline struct hlist_head *dev_name_hash(struct net *net, const char *name) @@ -217,37 +203,62 @@ 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) { - if (IS_ENABLED(CONFIG_RPS)) + 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) +{ + 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)) + else 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)) + else 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)) + else 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)) + else local_irq_enable(); } @@ -337,19 +348,27 @@ int netdev_name_node_alt_create(struct net_device *dev, const char *name) return -ENOMEM; netdev_name_node_add(net, name_node); /* The node that holds dev->name acts as a head of per-device list. */ - list_add_tail(&name_node->list, &dev->name_node->list); + list_add_tail_rcu(&name_node->list, &dev->name_node->list); return 0; } -static void __netdev_name_node_alt_destroy(struct netdev_name_node *name_node) +static void netdev_name_node_alt_free(struct rcu_head *head) { - list_del(&name_node->list); - netdev_name_node_del(name_node); + struct netdev_name_node *name_node = + container_of(head, struct netdev_name_node, rcu); + kfree(name_node->name); netdev_name_node_free(name_node); } +static void __netdev_name_node_alt_destroy(struct netdev_name_node *name_node) +{ + netdev_name_node_del(name_node); + list_del(&name_node->list); + call_rcu(&name_node->rcu, netdev_name_node_alt_free); +} + int netdev_name_node_alt_destroy(struct net_device *dev, const char *name) { struct netdev_name_node *name_node; @@ -365,7 +384,6 @@ int netdev_name_node_alt_destroy(struct net_device *dev, const char *name) return -EINVAL; __netdev_name_node_alt_destroy(name_node); - return 0; } @@ -373,23 +391,30 @@ static void netdev_name_node_alt_flush(struct net_device *dev) { struct netdev_name_node *name_node, *tmp; - list_for_each_entry_safe(name_node, tmp, &dev->name_node->list, list) - __netdev_name_node_alt_destroy(name_node); + list_for_each_entry_safe(name_node, tmp, &dev->name_node->list, list) { + list_del(&name_node->list); + netdev_name_node_alt_free(&name_node->rcu); + } } /* Device list insertion */ static void list_netdevice(struct net_device *dev) { + struct netdev_name_node *name_node; struct net *net = dev_net(dev); ASSERT_RTNL(); - write_lock(&dev_base_lock); list_add_tail_rcu(&dev->dev_list, &net->dev_base_head); netdev_name_node_add(net, dev->name_node); hlist_add_head_rcu(&dev->index_hlist, dev_index_hash(net, dev->ifindex)); - write_unlock(&dev_base_lock); + + netdev_for_each_altname(dev, name_node) + netdev_name_node_add(net, name_node); + + /* We reserved the ifindex, this can't fail */ + WARN_ON(xa_store(&net->dev_by_index, dev->ifindex, dev, GFP_KERNEL)); dev_base_seq_inc(net); } @@ -397,18 +422,22 @@ static void list_netdevice(struct net_device *dev) /* Device list removal * caller must respect a RCU grace period before freeing/reusing dev */ -static void unlist_netdevice(struct net_device *dev, bool lock) +static void unlist_netdevice(struct net_device *dev) { + struct netdev_name_node *name_node; + struct net *net = dev_net(dev); + ASSERT_RTNL(); + xa_erase(&net->dev_by_index, dev->ifindex); + + netdev_for_each_altname(dev, name_node) + netdev_name_node_del(name_node); + /* Unlink dev from the device chain */ - if (lock) - write_lock(&dev_base_lock); list_del_rcu(&dev->dev_list); netdev_name_node_del(dev->name_node); hlist_del_rcu(&dev->index_hlist); - if (lock) - write_unlock(&dev_base_lock); dev_base_seq_inc(dev_net(dev)); } @@ -424,9 +453,19 @@ static RAW_NOTIFIER_HEAD(netdev_chain); * queue in the local softnet handler. */ -DEFINE_PER_CPU_ALIGNED(struct softnet_data, softnet_data); +DEFINE_PER_CPU_ALIGNED(struct softnet_data, softnet_data) = { + .process_queue_bh_lock = INIT_LOCAL_LOCK(process_queue_bh_lock), +}; EXPORT_PER_CPU_SYMBOL(softnet_data); +/* Page_pool has a lockless array/stack to alloc/recycle pages. + * PP consumers must pay attention to run APIs in the appropriate context + * (e.g. NAPI context). + */ +DEFINE_PER_CPU(struct page_pool_bh, system_page_pool) = { + .bh_lock = INIT_LOCAL_LOCK(bh_lock), +}; + #ifdef CONFIG_LOCKDEP /* * register_netdevice() inits txq->_xmit_lock and sets lockdep class @@ -535,10 +574,18 @@ static inline void netdev_set_addr_lockdep_class(struct net_device *dev) static inline struct list_head *ptype_head(const struct packet_type *pt) { - if (pt->type == htons(ETH_P_ALL)) - return pt->dev ? &pt->dev->ptype_all : &ptype_all; - else - return pt->dev ? &pt->dev->ptype_specific : + if (pt->type == htons(ETH_P_ALL)) { + if (!pt->af_packet_net && !pt->dev) + return NULL; + + return pt->dev ? &pt->dev->ptype_all : + &pt->af_packet_net->ptype_all; + } + + if (pt->dev) + return &pt->dev->ptype_specific; + + return pt->af_packet_net ? &pt->af_packet_net->ptype_specific : &ptype_base[ntohs(pt->type) & PTYPE_HASH_MASK]; } @@ -559,6 +606,9 @@ void dev_add_pack(struct packet_type *pt) { struct list_head *head = ptype_head(pt); + if (WARN_ON_ONCE(!head)) + return; + spin_lock(&ptype_lock); list_add_rcu(&pt->list, head); spin_unlock(&ptype_lock); @@ -583,6 +633,9 @@ void __dev_remove_pack(struct packet_type *pt) struct list_head *head = ptype_head(pt); struct packet_type *pt1; + if (!head) + return; + spin_lock(&ptype_lock); list_for_each_entry(pt1, head, list) { @@ -638,7 +691,7 @@ int dev_get_iflink(const struct net_device *dev) if (dev->netdev_ops && dev->netdev_ops->ndo_get_iflink) return dev->netdev_ops->ndo_get_iflink(dev); - return dev->ifindex; + return READ_ONCE(dev->ifindex); } EXPORT_SYMBOL(dev_get_iflink); @@ -718,14 +771,88 @@ int dev_fill_forward_path(const struct net_device *dev, const u8 *daddr, } EXPORT_SYMBOL_GPL(dev_fill_forward_path); +/* must be called under rcu_read_lock(), as we dont take a reference */ +static struct napi_struct *napi_by_id(unsigned int napi_id) +{ + unsigned int hash = napi_id % HASH_SIZE(napi_hash); + struct napi_struct *napi; + + hlist_for_each_entry_rcu(napi, &napi_hash[hash], napi_hash_node) + if (napi->napi_id == napi_id) + return napi; + + return NULL; +} + +/* must be called under rcu_read_lock(), as we dont take a reference */ +static struct napi_struct * +netdev_napi_by_id(struct net *net, unsigned int napi_id) +{ + struct napi_struct *napi; + + napi = napi_by_id(napi_id); + if (!napi) + return NULL; + + if (WARN_ON_ONCE(!napi->dev)) + return NULL; + if (!net_eq(net, dev_net(napi->dev))) + return NULL; + + return napi; +} + +/** + * netdev_napi_by_id_lock() - find a device by NAPI ID and lock it + * @net: the applicable net namespace + * @napi_id: ID of a NAPI of a target device + * + * Find a NAPI instance with @napi_id. Lock its device. + * The device must be in %NETREG_REGISTERED state for lookup to succeed. + * netdev_unlock() must be called to release it. + * + * Return: pointer to NAPI, its device with lock held, NULL if not found. + */ +struct napi_struct * +netdev_napi_by_id_lock(struct net *net, unsigned int napi_id) +{ + struct napi_struct *napi; + struct net_device *dev; + + rcu_read_lock(); + napi = netdev_napi_by_id(net, napi_id); + if (!napi || READ_ONCE(napi->dev->reg_state) != NETREG_REGISTERED) { + rcu_read_unlock(); + return NULL; + } + + dev = napi->dev; + dev_hold(dev); + rcu_read_unlock(); + + dev = __netdev_put_lock(dev, net); + if (!dev) + return NULL; + + rcu_read_lock(); + napi = netdev_napi_by_id(net, napi_id); + if (napi && napi->dev != dev) + napi = NULL; + rcu_read_unlock(); + + if (!napi) + netdev_unlock(dev); + return napi; +} + /** * __dev_get_by_name - find a device by its name * @net: the applicable net namespace * @name: name to find * - * Find an interface by name. Must be called under RTNL semaphore - * or @dev_base_lock. If the name is found a pointer to the device - * is returned. If the name is not found then %NULL is returned. The + * Find an interface by name. Must be called under RTNL semaphore. + * If the name is found a pointer to the device is returned. + * If the name is not found then %NULL is returned. The * reference counters are not incremented so the caller must be * careful with locks. */ @@ -760,29 +887,43 @@ struct net_device *dev_get_by_name_rcu(struct net *net, const char *name) } EXPORT_SYMBOL(dev_get_by_name_rcu); +/* Deprecated for new users, call netdev_get_by_name() instead */ +struct net_device *dev_get_by_name(struct net *net, const char *name) +{ + struct net_device *dev; + + rcu_read_lock(); + dev = dev_get_by_name_rcu(net, name); + dev_hold(dev); + rcu_read_unlock(); + return dev; +} +EXPORT_SYMBOL(dev_get_by_name); + /** - * dev_get_by_name - find a device by its name + * netdev_get_by_name() - find a device by its name * @net: the applicable net namespace * @name: name to find + * @tracker: tracking object for the acquired reference + * @gfp: allocation flags for the tracker * * Find an interface by name. This can be called from any * context and does its own locking. The returned handle has - * the usage count incremented and the caller must use dev_put() to + * the usage count incremented and the caller must use netdev_put() to * release it when it is no longer needed. %NULL is returned if no * matching device is found. */ - -struct net_device *dev_get_by_name(struct net *net, const char *name) +struct net_device *netdev_get_by_name(struct net *net, const char *name, + netdevice_tracker *tracker, gfp_t gfp) { struct net_device *dev; - rcu_read_lock(); - dev = dev_get_by_name_rcu(net, name); - dev_hold(dev); - rcu_read_unlock(); + dev = dev_get_by_name(net, name); + if (dev) + netdev_tracker_alloc(dev, tracker, gfp); return dev; } -EXPORT_SYMBOL(dev_get_by_name); +EXPORT_SYMBOL(netdev_get_by_name); /** * __dev_get_by_index - find a device by its ifindex @@ -792,8 +933,7 @@ EXPORT_SYMBOL(dev_get_by_name); * Search for an interface by index. Returns %NULL if the device * is not found or a pointer to the device. The device has not * had its reference counter increased so the caller must be careful - * about locking. The caller must hold either the RTNL semaphore - * or @dev_base_lock. + * about locking. The caller must hold the RTNL semaphore. */ struct net_device *__dev_get_by_index(struct net *net, int ifindex) @@ -833,29 +973,42 @@ struct net_device *dev_get_by_index_rcu(struct net *net, int ifindex) } EXPORT_SYMBOL(dev_get_by_index_rcu); +/* Deprecated for new users, call netdev_get_by_index() instead */ +struct net_device *dev_get_by_index(struct net *net, int ifindex) +{ + struct net_device *dev; + + rcu_read_lock(); + dev = dev_get_by_index_rcu(net, ifindex); + dev_hold(dev); + rcu_read_unlock(); + return dev; +} +EXPORT_SYMBOL(dev_get_by_index); /** - * dev_get_by_index - find a device by its ifindex + * netdev_get_by_index() - find a device by its ifindex * @net: the applicable net namespace * @ifindex: index of device + * @tracker: tracking object for the acquired reference + * @gfp: allocation flags for the tracker * * Search for an interface by index. Returns NULL if the device * is not found or a pointer to the device. The device returned has * had a reference added and the pointer is safe until the user calls - * dev_put to indicate they have finished with it. + * netdev_put() to indicate they have finished with it. */ - -struct net_device *dev_get_by_index(struct net *net, int ifindex) +struct net_device *netdev_get_by_index(struct net *net, int ifindex, + netdevice_tracker *tracker, gfp_t gfp) { struct net_device *dev; - rcu_read_lock(); - dev = dev_get_by_index_rcu(net, ifindex); - dev_hold(dev); - rcu_read_unlock(); + dev = dev_get_by_index(net, ifindex); + if (dev) + netdev_tracker_alloc(dev, tracker, gfp); return dev; } -EXPORT_SYMBOL(dev_get_by_index); +EXPORT_SYMBOL(netdev_get_by_index); /** * dev_get_by_napi_id - find a device by napi_id @@ -866,21 +1019,151 @@ EXPORT_SYMBOL(dev_get_by_index); * its reference counter increased so the caller must be careful * about locking. The caller must hold RCU lock. */ - struct net_device *dev_get_by_napi_id(unsigned int napi_id) { struct napi_struct *napi; WARN_ON_ONCE(!rcu_read_lock_held()); - if (napi_id < MIN_NAPI_ID) + if (!napi_id_valid(napi_id)) return NULL; napi = napi_by_id(napi_id); return napi ? napi->dev : NULL; } -EXPORT_SYMBOL(dev_get_by_napi_id); + +/* Release the held reference on the net_device, and if the net_device + * is still registered try to lock the instance lock. If device is being + * unregistered NULL will be returned (but the reference has been released, + * either way!) + * + * This helper is intended for locking net_device after it has been looked up + * using a lockless lookup helper. Lock prevents the instance from going away. + */ +struct net_device *__netdev_put_lock(struct net_device *dev, struct net *net) +{ + netdev_lock(dev); + if (dev->reg_state > NETREG_REGISTERED || + dev->moving_ns || !net_eq(dev_net(dev), net)) { + netdev_unlock(dev); + dev_put(dev); + return NULL; + } + dev_put(dev); + return dev; +} + +static struct net_device * +__netdev_put_lock_ops_compat(struct net_device *dev, struct net *net) +{ + netdev_lock_ops_compat(dev); + if (dev->reg_state > NETREG_REGISTERED || + dev->moving_ns || !net_eq(dev_net(dev), net)) { + netdev_unlock_ops_compat(dev); + dev_put(dev); + return NULL; + } + dev_put(dev); + return dev; +} + +/** + * netdev_get_by_index_lock() - find a device by its ifindex + * @net: the applicable net namespace + * @ifindex: index of device + * + * Search for an interface by index. If a valid device + * with @ifindex is found it will be returned with netdev->lock held. + * netdev_unlock() must be called to release it. + * + * Return: pointer to a device with lock held, NULL if not found. + */ +struct net_device *netdev_get_by_index_lock(struct net *net, int ifindex) +{ + struct net_device *dev; + + dev = dev_get_by_index(net, ifindex); + if (!dev) + return NULL; + + return __netdev_put_lock(dev, net); +} + +struct net_device * +netdev_get_by_index_lock_ops_compat(struct net *net, int ifindex) +{ + struct net_device *dev; + + dev = dev_get_by_index(net, ifindex); + if (!dev) + return NULL; + + return __netdev_put_lock_ops_compat(dev, net); +} + +struct net_device * +netdev_xa_find_lock(struct net *net, struct net_device *dev, + unsigned long *index) +{ + if (dev) + netdev_unlock(dev); + + do { + rcu_read_lock(); + dev = xa_find(&net->dev_by_index, index, ULONG_MAX, XA_PRESENT); + if (!dev) { + rcu_read_unlock(); + return NULL; + } + dev_hold(dev); + rcu_read_unlock(); + + dev = __netdev_put_lock(dev, net); + if (dev) + return dev; + + (*index)++; + } while (true); +} + +struct net_device * +netdev_xa_find_lock_ops_compat(struct net *net, struct net_device *dev, + unsigned long *index) +{ + if (dev) + netdev_unlock_ops_compat(dev); + + do { + rcu_read_lock(); + dev = xa_find(&net->dev_by_index, index, ULONG_MAX, XA_PRESENT); + if (!dev) { + rcu_read_unlock(); + return NULL; + } + dev_hold(dev); + rcu_read_unlock(); + + dev = __netdev_put_lock_ops_compat(dev, net); + if (dev) + return dev; + + (*index)++; + } while (true); +} + +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)); +} +EXPORT_IPV6_MOD_GPL(netdev_copy_name); /** * netdev_get_name - get a netdevice name, knowing its ifindex. @@ -893,7 +1176,6 @@ 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); @@ -902,15 +1184,20 @@ 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; } +static bool dev_addr_cmp(struct net_device *dev, unsigned short type, + const char *ha) +{ + return dev->type == type && !memcmp(dev->dev_addr, ha, dev->addr_len); +} + /** * dev_getbyhwaddr_rcu - find a device by its hardware address * @net: the applicable net namespace @@ -919,7 +1206,7 @@ out: * * Search for an interface by MAC address. Returns NULL if the device * is not found or a pointer to the device. - * The caller must hold RCU or RTNL. + * The caller must hold RCU. * The returned device has not had its ref count increased * and the caller must therefore be careful about locking * @@ -931,14 +1218,39 @@ struct net_device *dev_getbyhwaddr_rcu(struct net *net, unsigned short type, struct net_device *dev; for_each_netdev_rcu(net, dev) - if (dev->type == type && - !memcmp(dev->dev_addr, ha, dev->addr_len)) + if (dev_addr_cmp(dev, type, ha)) return dev; return NULL; } EXPORT_SYMBOL(dev_getbyhwaddr_rcu); +/** + * dev_getbyhwaddr() - find a device by its hardware address + * @net: the applicable net namespace + * @type: media type of device + * @ha: hardware address + * + * Similar to dev_getbyhwaddr_rcu(), but the owner needs to hold + * rtnl_lock. + * + * Context: rtnl_lock() must be held. + * Return: pointer to the net_device, or NULL if not found + */ +struct net_device *dev_getbyhwaddr(struct net *net, unsigned short type, + const char *ha) +{ + struct net_device *dev; + + ASSERT_RTNL(); + for_each_netdev(net, dev) + if (dev_addr_cmp(dev, type, ha)) + return dev; + + return NULL; +} +EXPORT_SYMBOL(dev_getbyhwaddr); + struct net_device *dev_getfirstbyhwtype(struct net *net, unsigned short type) { struct net_device *dev, *ret = NULL; @@ -956,33 +1268,32 @@ struct net_device *dev_getfirstbyhwtype(struct net *net, unsigned short type) EXPORT_SYMBOL(dev_getfirstbyhwtype); /** - * __dev_get_by_flags - find any device with given flags - * @net: the applicable net namespace - * @if_flags: IFF_* values - * @mask: bitmask of bits in if_flags to check + * netdev_get_by_flags_rcu - find any device with given flags + * @net: the applicable net namespace + * @tracker: tracking object for the acquired reference + * @if_flags: IFF_* values + * @mask: bitmask of bits in if_flags to check + * + * Search for any interface with the given flags. * - * Search for any interface with the given flags. Returns NULL if a device - * is not found or a pointer to the device. Must be called inside - * rtnl_lock(), and result refcount is unchanged. + * Context: rcu_read_lock() must be held. + * Returns: NULL if a device is not found or a pointer to the device. */ - -struct net_device *__dev_get_by_flags(struct net *net, unsigned short if_flags, - unsigned short mask) +struct net_device *netdev_get_by_flags_rcu(struct net *net, netdevice_tracker *tracker, + unsigned short if_flags, unsigned short mask) { - struct net_device *dev, *ret; - - ASSERT_RTNL(); + struct net_device *dev; - ret = NULL; - for_each_netdev(net, dev) { - if (((dev->flags ^ if_flags) & mask) == 0) { - ret = dev; - break; + for_each_netdev_rcu(net, dev) { + if (((READ_ONCE(dev->flags) ^ if_flags) & mask) == 0) { + netdev_hold(dev, tracker, GFP_ATOMIC); + return dev; } } - return ret; + + return NULL; } -EXPORT_SYMBOL(__dev_get_by_flags); +EXPORT_IPV6_MOD(netdev_get_by_flags_rcu); /** * dev_valid_name - check if name is okay for network device @@ -1014,7 +1325,7 @@ EXPORT_SYMBOL(dev_valid_name); * __dev_alloc_name - allocate a name for a device * @net: network namespace to allocate the device name in * @name: name format string - * @buf: scratch buffer and result name string + * @res: result name string * * Passed a format string - eg "lt%d" it will try and find a suitable * id. It scans list of devices to build up a free map, then chooses @@ -1025,83 +1336,79 @@ EXPORT_SYMBOL(dev_valid_name); * Returns the number of the unit assigned or a negative errno code. */ -static int __dev_alloc_name(struct net *net, const char *name, char *buf) +static int __dev_alloc_name(struct net *net, const char *name, char *res) { int i = 0; const char *p; const int max_netdevices = 8*PAGE_SIZE; unsigned long *inuse; struct net_device *d; + char buf[IFNAMSIZ]; - if (!dev_valid_name(name)) - return -EINVAL; - + /* Verify the string as this thing may have come from the user. + * There must be one "%d" and no other "%" characters. + */ p = strchr(name, '%'); - if (p) { - /* - * Verify the string as this thing may have come from - * the user. There must be either one "%d" and no other "%" - * characters. - */ - if (p[1] != 'd' || strchr(p + 2, '%')) - return -EINVAL; + if (!p || p[1] != 'd' || strchr(p + 2, '%')) + return -EINVAL; - /* Use one page as a bit array of possible slots */ - inuse = (unsigned long *) get_zeroed_page(GFP_ATOMIC); - if (!inuse) - return -ENOMEM; + /* Use one page as a bit array of possible slots */ + inuse = bitmap_zalloc(max_netdevices, GFP_ATOMIC); + if (!inuse) + return -ENOMEM; - for_each_netdev(net, d) { - struct netdev_name_node *name_node; - list_for_each_entry(name_node, &d->name_node->list, list) { - if (!sscanf(name_node->name, name, &i)) - continue; - if (i < 0 || i >= max_netdevices) - continue; + for_each_netdev(net, d) { + struct netdev_name_node *name_node; - /* avoid cases where sscanf is not exact inverse of printf */ - snprintf(buf, IFNAMSIZ, name, i); - if (!strncmp(buf, name_node->name, IFNAMSIZ)) - __set_bit(i, inuse); - } - if (!sscanf(d->name, name, &i)) + netdev_for_each_altname(d, name_node) { + if (!sscanf(name_node->name, name, &i)) continue; if (i < 0 || i >= max_netdevices) continue; - /* avoid cases where sscanf is not exact inverse of printf */ + /* avoid cases where sscanf is not exact inverse of printf */ snprintf(buf, IFNAMSIZ, name, i); - if (!strncmp(buf, d->name, IFNAMSIZ)) + if (!strncmp(buf, name_node->name, IFNAMSIZ)) __set_bit(i, inuse); } + if (!sscanf(d->name, name, &i)) + continue; + if (i < 0 || i >= max_netdevices) + continue; - i = find_first_zero_bit(inuse, max_netdevices); - free_page((unsigned long) inuse); + /* avoid cases where sscanf is not exact inverse of printf */ + snprintf(buf, IFNAMSIZ, name, i); + if (!strncmp(buf, d->name, IFNAMSIZ)) + __set_bit(i, inuse); } - snprintf(buf, IFNAMSIZ, name, i); - if (!netdev_name_in_use(net, buf)) - return i; + i = find_first_zero_bit(inuse, max_netdevices); + bitmap_free(inuse); + if (i == max_netdevices) + return -ENFILE; - /* It is possible to run out of possible slots - * when the name is long and there isn't enough space left - * for the digits, or if all bits are used. - */ - return -ENFILE; + /* 'res' and 'name' could overlap, use 'buf' as an intermediate buffer */ + strscpy(buf, name, IFNAMSIZ); + snprintf(res, IFNAMSIZ, buf, i); + return i; } -static int dev_alloc_name_ns(struct net *net, - struct net_device *dev, - const char *name) +/* Returns negative errno or allocated unit id (see __dev_alloc_name()) */ +static int dev_prep_valid_name(struct net *net, struct net_device *dev, + const char *want_name, char *out_name, + int dup_errno) { - char buf[IFNAMSIZ]; - int ret; + if (!dev_valid_name(want_name)) + return -EINVAL; - BUG_ON(!net); - ret = __dev_alloc_name(net, name, buf); - if (ret >= 0) - strscpy(dev->name, buf, IFNAMSIZ); - return ret; + if (strchr(want_name, '%')) + return __dev_alloc_name(net, want_name, out_name); + + if (netdev_name_in_use(net, want_name)) + return -dup_errno; + if (out_name != want_name) + strscpy(out_name, want_name, IFNAMSIZ); + return 0; } /** @@ -1120,93 +1427,65 @@ static int dev_alloc_name_ns(struct net *net, int dev_alloc_name(struct net_device *dev, const char *name) { - return dev_alloc_name_ns(dev_net(dev), dev, name); + return dev_prep_valid_name(dev_net(dev), dev, name, dev->name, ENFILE); } EXPORT_SYMBOL(dev_alloc_name); static int dev_get_valid_name(struct net *net, struct net_device *dev, const char *name) { - BUG_ON(!net); - - if (!dev_valid_name(name)) - return -EINVAL; - - if (strchr(name, '%')) - return dev_alloc_name_ns(net, dev, name); - else if (netdev_name_in_use(net, name)) - return -EEXIST; - else if (dev->name != name) - strscpy(dev->name, name, IFNAMSIZ); + int ret; - return 0; + ret = dev_prep_valid_name(net, dev, name, dev->name, EEXIST); + return ret < 0 ? ret : 0; } -/** - * dev_change_name - change name of a device - * @dev: device - * @newname: name (or format string) must be at least IFNAMSIZ - * - * Change name of a device, can pass format strings "eth%d". - * for wildcarding. - */ -int dev_change_name(struct net_device *dev, const char *newname) +int netif_change_name(struct net_device *dev, const char *newname) { + struct net *net = dev_net(dev); unsigned char old_assign_type; char oldname[IFNAMSIZ]; int err = 0; int ret; - struct net *net; - - ASSERT_RTNL(); - BUG_ON(!dev_net(dev)); - net = dev_net(dev); + ASSERT_RTNL_NET(net); - down_write(&devnet_rename_sem); - - if (strncmp(newname, dev->name, IFNAMSIZ) == 0) { - up_write(&devnet_rename_sem); + if (!strncmp(newname, dev->name, IFNAMSIZ)) return 0; - } memcpy(oldname, dev->name, IFNAMSIZ); + write_seqlock_bh(&netdev_rename_lock); err = dev_get_valid_name(net, dev, newname); - if (err < 0) { - up_write(&devnet_rename_sem); + write_sequnlock_bh(&netdev_rename_lock); + + if (err < 0) return err; - } if (oldname[0] && !strchr(oldname, '%')) netdev_info(dev, "renamed from %s%s\n", oldname, dev->flags & IFF_UP ? " (while UP)" : ""); old_assign_type = dev->name_assign_type; - dev->name_assign_type = NET_NAME_RENAMED; + WRITE_ONCE(dev->name_assign_type, NET_NAME_RENAMED); rollback: ret = device_rename(&dev->dev, dev->name); if (ret) { + write_seqlock_bh(&netdev_rename_lock); memcpy(dev->name, oldname, IFNAMSIZ); - dev->name_assign_type = old_assign_type; - up_write(&devnet_rename_sem); + write_sequnlock_bh(&netdev_rename_lock); + WRITE_ONCE(dev->name_assign_type, old_assign_type); return ret; } - up_write(&devnet_rename_sem); - netdev_adjacent_rename_links(dev, oldname); - write_lock(&dev_base_lock); netdev_name_node_del(dev->name_node); - write_unlock(&dev_base_lock); - synchronize_rcu(); + synchronize_net(); - write_lock(&dev_base_lock); netdev_name_node_add(net, dev->name_node); - write_unlock(&dev_base_lock); ret = call_netdevice_notifiers(NETDEV_CHANGENAME, dev); ret = notifier_to_errno(ret); @@ -1215,10 +1494,11 @@ rollback: /* err >= 0 after dev_alloc_name() or stores the first errno */ 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); - dev->name_assign_type = old_assign_type; + WRITE_ONCE(dev->name_assign_type, old_assign_type); old_assign_type = NET_NAME_RENAMED; goto rollback; } else { @@ -1230,15 +1510,7 @@ rollback: return err; } -/** - * dev_set_alias - change ifalias of a device - * @dev: device - * @alias: name up to IFALIASZ - * @len: limit of bytes to copy from info - * - * Set ifalias for a device, - */ -int dev_set_alias(struct net_device *dev, const char *alias, size_t len) +int netif_set_alias(struct net_device *dev, const char *alias, size_t len) { struct dev_ifalias *new_alias = NULL; @@ -1264,7 +1536,6 @@ int dev_set_alias(struct net_device *dev, const char *alias, size_t len) return len; } -EXPORT_SYMBOL(dev_set_alias); /** * dev_get_alias - get ifalias of a device @@ -1301,16 +1572,10 @@ void netdev_features_change(struct net_device *dev) } EXPORT_SYMBOL(netdev_features_change); -/** - * netdev_state_change - device changes state - * @dev: device to cause notification - * - * Called to indicate a device has changed state. This function calls - * the notifier chains for netdev_chain and sends a NEWLINK message - * to the routing socket. - */ -void netdev_state_change(struct net_device *dev) +void netif_state_change(struct net_device *dev) { + netdev_ops_assert_locked_or_invisible(dev); + if (dev->flags & IFF_UP) { struct netdev_notifier_change_info change_info = { .info.dev = dev, @@ -1321,7 +1586,6 @@ void netdev_state_change(struct net_device *dev) rtmsg_ifinfo(RTM_NEWLINK, dev, 0, GFP_KERNEL, 0, NULL); } } -EXPORT_SYMBOL(netdev_state_change); /** * __netdev_notify_peers - notify network peers about existence of @dev, @@ -1410,6 +1674,8 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack) set_bit(__LINK_STATE_START, &dev->state); + netdev_ops_assert_locked(dev); + if (ops->ndo_validate_addr) ret = ops->ndo_validate_addr(dev); @@ -1421,7 +1687,7 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack) if (ret) clear_bit(__LINK_STATE_START, &dev->state); else { - dev->flags |= IFF_UP; + netif_set_up(dev, true); dev_set_rx_mode(dev); dev_activate(dev); add_device_randomness(dev->dev_addr, dev->addr_len); @@ -1430,20 +1696,7 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack) return ret; } -/** - * dev_open - prepare an interface for use. - * @dev: device to open - * @extack: netlink extended ack - * - * Takes a device from down to up state. The device's private open - * function is invoked and then the multicast lists are loaded. Finally - * the device is moved into the up state and a %NETDEV_UP message is - * sent to the netdev notifier chain. - * - * Calling this function on an active interface is a nop. On a failure - * a negative errno code is returned. - */ -int dev_open(struct net_device *dev, struct netlink_ext_ack *extack) +int netif_open(struct net_device *dev, struct netlink_ext_ack *extack) { int ret; @@ -1459,7 +1712,6 @@ int dev_open(struct net_device *dev, struct netlink_ext_ack *extack) return ret; } -EXPORT_SYMBOL(dev_open); static void __dev_close_many(struct list_head *head) { @@ -1497,10 +1749,13 @@ static void __dev_close_many(struct list_head *head) * We allow it to be called even after a DETACH hot-plug * event. */ + + netdev_ops_assert_locked(dev); + if (ops->ndo_stop) ops->ndo_stop(dev); - dev->flags &= ~IFF_UP; + netif_set_up(dev, false); netpoll_poll_enable(dev); } } @@ -1514,7 +1769,7 @@ static void __dev_close(struct net_device *dev) list_del(&single); } -void dev_close_many(struct list_head *head, bool unlink) +void netif_close_many(struct list_head *head, bool unlink) { struct net_device *dev, *tmp; @@ -1532,39 +1787,21 @@ void dev_close_many(struct list_head *head, bool unlink) list_del_init(&dev->close_list); } } -EXPORT_SYMBOL(dev_close_many); +EXPORT_SYMBOL_NS_GPL(netif_close_many, "NETDEV_INTERNAL"); -/** - * dev_close - shutdown an interface. - * @dev: device to shutdown - * - * This function moves an active device into down state. A - * %NETDEV_GOING_DOWN is sent to the netdev notifier chain. The device - * is then deactivated and finally a %NETDEV_DOWN is sent to the notifier - * chain. - */ -void dev_close(struct net_device *dev) +void netif_close(struct net_device *dev) { if (dev->flags & IFF_UP) { LIST_HEAD(single); list_add(&dev->close_list, &single); - dev_close_many(&single, true); + netif_close_many(&single, true); list_del(&single); } } -EXPORT_SYMBOL(dev_close); - +EXPORT_SYMBOL(netif_close); -/** - * dev_disable_lro - disable Large Receive Offload on a device - * @dev: device - * - * Disable Large Receive Offload (LRO) on a net device. Must be - * called under RTNL. This is needed if received packets may be - * forwarded to another interface. - */ -void dev_disable_lro(struct net_device *dev) +void netif_disable_lro(struct net_device *dev) { struct net_device *lower_dev; struct list_head *iter; @@ -1575,10 +1812,13 @@ void dev_disable_lro(struct net_device *dev) if (unlikely(dev->features & NETIF_F_LRO)) netdev_WARN(dev, "failed to disable LRO!\n"); - netdev_for_each_lower_dev(dev, lower_dev, iter) - dev_disable_lro(lower_dev); + netdev_for_each_lower_dev(dev, lower_dev, iter) { + netdev_lock_ops(lower_dev); + netif_disable_lro(lower_dev); + netdev_unlock_ops(lower_dev); + } } -EXPORT_SYMBOL(dev_disable_lro); +EXPORT_IPV6_MOD(netif_disable_lro); /** * dev_disable_gro_hw - disable HW Generic Receive Offload on a device @@ -1614,6 +1854,7 @@ const char *netdev_cmd_to_name(enum netdev_cmd cmd) N(SVLAN_FILTER_PUSH_INFO) N(SVLAN_FILTER_DROP_INFO) N(PRE_CHANGEADDR) N(OFFLOAD_XSTATS_ENABLE) N(OFFLOAD_XSTATS_DISABLE) N(OFFLOAD_XSTATS_REPORT_USED) N(OFFLOAD_XSTATS_REPORT_DELTA) + N(XDP_FEAT_CHANGE) } #undef N return "UNKNOWN_NETDEV_EVENT"; @@ -1665,7 +1906,9 @@ static int call_netdevice_register_net_notifiers(struct notifier_block *nb, int err; for_each_netdev(net, dev) { + netdev_lock_ops(dev); err = call_netdevice_register_notifiers(nb, dev); + netdev_unlock_ops(dev); if (err) goto rollback; } @@ -1709,14 +1952,19 @@ int register_netdevice_notifier(struct notifier_block *nb) /* Close race with setup_net() and cleanup_net() */ down_write(&pernet_ops_rwsem); + + /* When RTNL is removed, we need protection for netdev_chain. */ rtnl_lock(); + err = raw_notifier_chain_register(&netdev_chain, nb); if (err) goto unlock; if (dev_boot_phase) goto unlock; for_each_net(net) { + __rtnl_net_lock(net); err = call_netdevice_register_net_notifiers(nb, net); + __rtnl_net_unlock(net); if (err) goto rollback; } @@ -1727,8 +1975,11 @@ unlock: return err; rollback: - for_each_net_continue_reverse(net) + for_each_net_continue_reverse(net) { + __rtnl_net_lock(net); call_netdevice_unregister_net_notifiers(nb, net); + __rtnl_net_unlock(net); + } raw_notifier_chain_unregister(&netdev_chain, nb); goto unlock; @@ -1761,8 +2012,11 @@ int unregister_netdevice_notifier(struct notifier_block *nb) if (err) goto unlock; - for_each_net(net) + for_each_net(net) { + __rtnl_net_lock(net); call_netdevice_unregister_net_notifiers(nb, net); + __rtnl_net_unlock(net); + } unlock: rtnl_unlock(); @@ -1826,9 +2080,10 @@ int register_netdevice_notifier_net(struct net *net, struct notifier_block *nb) { int err; - rtnl_lock(); + rtnl_net_lock(net); err = __register_netdevice_notifier_net(net, nb, false); - rtnl_unlock(); + rtnl_net_unlock(net); + return err; } EXPORT_SYMBOL(register_netdevice_notifier_net); @@ -1840,7 +2095,7 @@ EXPORT_SYMBOL(register_netdevice_notifier_net); * @nb: notifier * * Unregister a notifier previously registered by - * register_netdevice_notifier(). The notifier is unlinked into the + * register_netdevice_notifier_net(). The notifier is unlinked from the * kernel structures and may then be reused. A negative errno code * is returned on a failure. * @@ -1854,9 +2109,10 @@ int unregister_netdevice_notifier_net(struct net *net, { int err; - rtnl_lock(); + rtnl_net_lock(net); err = __unregister_netdevice_notifier_net(net, nb); - rtnl_unlock(); + rtnl_net_unlock(net); + return err; } EXPORT_SYMBOL(unregister_netdevice_notifier_net); @@ -1869,12 +2125,40 @@ static void __move_netdevice_notifier_net(struct net *src_net, __register_netdevice_notifier_net(dst_net, nb, true); } -void move_netdevice_notifier_net(struct net *src_net, struct net *dst_net, - struct notifier_block *nb) +static void rtnl_net_dev_lock(struct net_device *dev) { - rtnl_lock(); - __move_netdevice_notifier_net(src_net, dst_net, nb); - rtnl_unlock(); + bool again; + + do { + struct net *net; + + again = false; + + /* netns might be being dismantled. */ + rcu_read_lock(); + net = dev_net_rcu(dev); + net_passive_inc(net); + rcu_read_unlock(); + + rtnl_net_lock(net); + +#ifdef CONFIG_NET_NS + /* dev might have been moved to another netns. */ + if (!net_eq(net, rcu_access_pointer(dev->nd_net.net))) { + rtnl_net_unlock(net); + net_passive_dec(net); + again = true; + } +#endif + } while (again); +} + +static void rtnl_net_dev_unlock(struct net_device *dev) +{ + struct net *net = dev_net(dev); + + rtnl_net_unlock(net); + net_passive_dec(net); } int register_netdevice_notifier_dev_net(struct net_device *dev, @@ -1883,13 +2167,14 @@ int register_netdevice_notifier_dev_net(struct net_device *dev, { int err; - rtnl_lock(); + rtnl_net_dev_lock(dev); err = __register_netdevice_notifier_net(dev_net(dev), nb, false); if (!err) { nn->nb = nb; list_add(&nn->list, &dev->net_notifier_list); } - rtnl_unlock(); + rtnl_net_dev_unlock(dev); + return err; } EXPORT_SYMBOL(register_netdevice_notifier_dev_net); @@ -1900,10 +2185,11 @@ int unregister_netdevice_notifier_dev_net(struct net_device *dev, { int err; - rtnl_lock(); + rtnl_net_dev_lock(dev); list_del(&nn->list); err = __unregister_netdevice_notifier_net(dev_net(dev), nb); - rtnl_unlock(); + rtnl_net_dev_unlock(dev); + return err; } EXPORT_SYMBOL(unregister_netdevice_notifier_dev_net); @@ -1926,8 +2212,8 @@ static void move_netdevice_notifiers_dev_net(struct net_device *dev, * are as for raw_notifier_call_chain(). */ -static int call_netdevice_notifiers_info(unsigned long val, - struct netdev_notifier_info *info) +int call_netdevice_notifiers_info(unsigned long val, + struct netdev_notifier_info *info) { struct net *net = dev_net(info->dev); int ret; @@ -2051,6 +2337,11 @@ void net_dec_egress_queue(void) EXPORT_SYMBOL_GPL(net_dec_egress_queue); #endif +#ifdef CONFIG_NET_CLS_ACT +DEFINE_STATIC_KEY_FALSE(tcf_sw_enabled_key); +EXPORT_SYMBOL(tcf_sw_enabled_key); +#endif + DEFINE_STATIC_KEY_FALSE(netstamp_needed_key); EXPORT_SYMBOL(netstamp_needed_key); #ifdef CONFIG_JUMP_LABEL @@ -2107,7 +2398,7 @@ EXPORT_SYMBOL(net_disable_timestamp); static inline void net_timestamp_set(struct sk_buff *skb) { skb->tstamp = 0; - skb->mono_delivery_time = 0; + skb->tstamp_type = SKB_CLOCK_REALTIME; if (static_branch_unlikely(&netstamp_needed_key)) skb->tstamp = ktime_get_real(); } @@ -2172,9 +2463,9 @@ int dev_forward_skb_nomtu(struct net_device *dev, struct sk_buff *skb) return __dev_forward_skb2(dev, skb, false) ?: netif_rx_internal(skb); } -static inline int deliver_skb(struct sk_buff *skb, - struct packet_type *pt_prev, - struct net_device *orig_dev) +static int deliver_skb(struct sk_buff *skb, + struct packet_type *pt_prev, + struct net_device *orig_dev) { if (unlikely(skb_orphan_frags_rx(skb, GFP_ATOMIC))) return -ENOMEM; @@ -2193,7 +2484,7 @@ static inline void deliver_ptype_list_skb(struct sk_buff *skb, list_for_each_entry_rcu(ptype, ptype_list, list) { if (ptype->type != type) continue; - if (pt_prev) + if (unlikely(pt_prev)) deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; } @@ -2214,15 +2505,21 @@ static inline bool skb_loop_sk(struct packet_type *ptype, struct sk_buff *skb) } /** - * dev_nit_active - return true if any network interface taps are in use + * dev_nit_active_rcu - return true if any network interface taps are in use + * + * The caller must hold the RCU lock * * @dev: network device to check for the presence of taps */ -bool dev_nit_active(struct net_device *dev) +bool dev_nit_active_rcu(const struct net_device *dev) { - return !list_empty(&ptype_all) || !list_empty(&dev->ptype_all); + /* Callers may hold either RCU or RCU BH lock */ + WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_bh_held()); + + return !list_empty(&dev_net(dev)->ptype_all) || + !list_empty(&dev->ptype_all); } -EXPORT_SYMBOL_GPL(dev_nit_active); +EXPORT_SYMBOL_GPL(dev_nit_active_rcu); /* * Support routine. Sends outgoing frames to any network @@ -2231,15 +2528,15 @@ EXPORT_SYMBOL_GPL(dev_nit_active); void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) { - struct packet_type *ptype; + struct packet_type *ptype, *pt_prev = NULL; + struct list_head *ptype_list; struct sk_buff *skb2 = NULL; - struct packet_type *pt_prev = NULL; - struct list_head *ptype_list = &ptype_all; rcu_read_lock(); + ptype_list = &dev_net_rcu(dev)->ptype_all; again: list_for_each_entry_rcu(ptype, ptype_list, list) { - if (ptype->ignore_outgoing) + if (READ_ONCE(ptype->ignore_outgoing)) continue; /* Never send packets back to the socket @@ -2248,7 +2545,7 @@ again: if (skb_loop_sk(ptype, skb)) continue; - if (pt_prev) { + if (unlikely(pt_prev)) { deliver_skb(skb2, pt_prev, skb->dev); pt_prev = ptype; continue; @@ -2280,7 +2577,7 @@ again: pt_prev = ptype; } - if (ptype_list == &ptype_all) { + if (ptype_list != &dev->ptype_all) { ptype_list = &dev->ptype_all; goto again; } @@ -2366,8 +2663,7 @@ static bool remove_xps_queue(struct xps_dev_maps *dev_maps, struct xps_map *map = NULL; int pos; - if (dev_maps) - map = xmap_dereference(dev_maps->attr_map[tci]); + map = xmap_dereference(dev_maps->attr_map[tci]); if (!map) return false; @@ -2542,6 +2838,8 @@ int __netif_set_xps_queue(struct net_device *dev, const unsigned long *mask, struct xps_map *map, *new_map; unsigned int nr_ids; + WARN_ON_ONCE(index >= dev->num_tx_queues); + if (dev->num_tc) { /* Do not allow XPS on subordinate device directly */ num_tc = dev->num_tc; @@ -2881,7 +3179,7 @@ int netif_set_real_num_tx_queues(struct net_device *dev, unsigned int txq) if (dev->reg_state == NETREG_REGISTERED || dev->reg_state == NETREG_UNREGISTERING) { - ASSERT_RTNL(); + netdev_ops_assert_locked(dev); rc = netdev_queue_update_kobjects(dev, dev->real_num_tx_queues, txq); @@ -2891,6 +3189,8 @@ int netif_set_real_num_tx_queues(struct net_device *dev, unsigned int txq) if (dev->num_tc) netif_setup_tc(dev, txq); + net_shaper_set_real_num_tx_queues(dev, txq); + dev_qdisc_change_real_num_tx(dev, txq); dev->real_num_tx_queues = txq; @@ -2910,7 +3210,6 @@ int netif_set_real_num_tx_queues(struct net_device *dev, unsigned int txq) } EXPORT_SYMBOL(netif_set_real_num_tx_queues); -#ifdef CONFIG_SYSFS /** * netif_set_real_num_rx_queues - set actual number of RX queues used * @dev: Network device @@ -2929,7 +3228,7 @@ int netif_set_real_num_rx_queues(struct net_device *dev, unsigned int rxq) return -EINVAL; if (dev->reg_state == NETREG_REGISTERED) { - ASSERT_RTNL(); + netdev_ops_assert_locked(dev); rc = net_rx_queue_update_kobjects(dev, dev->real_num_rx_queues, rxq); @@ -2941,7 +3240,6 @@ int netif_set_real_num_rx_queues(struct net_device *dev, unsigned int rxq) return 0; } EXPORT_SYMBOL(netif_set_real_num_rx_queues); -#endif /** * netif_set_real_num_queues - set actual number of RX and TX queues used @@ -3001,6 +3299,8 @@ void netif_set_tso_max_size(struct net_device *dev, unsigned int size) dev->tso_max_size = min(GSO_MAX_SIZE, size); if (size < READ_ONCE(dev->gso_max_size)) netif_set_gso_max_size(dev, size); + if (size < READ_ONCE(dev->gso_ipv4_max_size)) + netif_set_gso_ipv4_max_size(dev, size); } EXPORT_SYMBOL(netif_set_tso_max_size); @@ -3074,13 +3374,20 @@ static void __netif_reschedule(struct Qdisc *q) void __netif_schedule(struct Qdisc *q) { + /* If q->defer_list is not empty, at least one thread is + * in __dev_xmit_skb() before llist_del_all(&q->defer_list). + * This thread will attempt to run the queue. + */ + if (!llist_empty(&q->defer_list)) + return; + if (!test_and_set_bit(__QDISC_STATE_SCHED, &q->state)) __netif_reschedule(q); } EXPORT_SYMBOL(__netif_schedule); struct dev_kfree_skb_cb { - enum skb_free_reason reason; + enum skb_drop_reason reason; }; static struct dev_kfree_skb_cb *get_kfree_skb_cb(const struct sk_buff *skb) @@ -3113,7 +3420,7 @@ void netif_tx_wake_queue(struct netdev_queue *dev_queue) } EXPORT_SYMBOL(netif_tx_wake_queue); -void __dev_kfree_skb_irq(struct sk_buff *skb, enum skb_free_reason reason) +void dev_kfree_skb_irq_reason(struct sk_buff *skb, enum skb_drop_reason reason) { unsigned long flags; @@ -3133,16 +3440,16 @@ void __dev_kfree_skb_irq(struct sk_buff *skb, enum skb_free_reason reason) raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_restore(flags); } -EXPORT_SYMBOL(__dev_kfree_skb_irq); +EXPORT_SYMBOL(dev_kfree_skb_irq_reason); -void __dev_kfree_skb_any(struct sk_buff *skb, enum skb_free_reason reason) +void dev_kfree_skb_any_reason(struct sk_buff *skb, enum skb_drop_reason reason) { if (in_hardirq() || irqs_disabled()) - __dev_kfree_skb_irq(skb, reason); + dev_kfree_skb_irq_reason(skb, reason); else - dev_kfree_skb(skb); + kfree_skb_reason(skb, reason); } -EXPORT_SYMBOL(__dev_kfree_skb_any); +EXPORT_SYMBOL(dev_kfree_skb_any_reason); /** @@ -3171,7 +3478,7 @@ void netif_device_attach(struct net_device *dev) if (!test_and_set_bit(__LINK_STATE_PRESENT, &dev->state) && netif_running(dev)) { netif_tx_wake_all_queues(dev); - __netdev_watchdog_up(dev); + netdev_watchdog_up(dev); } } EXPORT_SYMBOL(netif_device_attach); @@ -3202,6 +3509,7 @@ static u16 skb_tx_hash(const struct net_device *dev, } if (skb_rx_queue_recorded(skb)) { + DEBUG_NET_WARN_ON_ONCE(qcount == 0); hash = skb_get_rx_queue(skb); if (hash >= qoffset) hash -= qoffset; @@ -3213,7 +3521,7 @@ static u16 skb_tx_hash(const struct net_device *dev, return (u16) reciprocal_scale(skb_get_hash(skb), qcount) + qoffset; } -static void skb_warn_bad_offload(const struct sk_buff *skb) +void skb_warn_bad_offload(const struct sk_buff *skb) { static const netdev_features_t null_features; struct net_device *dev = skb->dev; @@ -3251,6 +3559,10 @@ int skb_checksum_help(struct sk_buff *skb) return -EINVAL; } + if (!skb_frags_readable(skb)) { + return -EFAULT; + } + /* Before computing a checksum, we should make sure no frag could * be modified by an external entity : checksum could be wrong. */ @@ -3262,15 +3574,19 @@ int skb_checksum_help(struct sk_buff *skb) offset = skb_checksum_start_offset(skb); ret = -EINVAL; - if (WARN_ON_ONCE(offset >= skb_headlen(skb))) { + if (unlikely(offset >= skb_headlen(skb))) { DO_ONCE_LITE(skb_dump, KERN_ERR, skb, false); + WARN_ONCE(true, "offset (%d) >= skb_headlen() (%u)\n", + offset, skb_headlen(skb)); goto out; } csum = skb_checksum(skb, offset, skb->len - offset, 0); offset += skb->csum_offset; - if (WARN_ON_ONCE(offset + sizeof(__sum16) > skb_headlen(skb))) { + if (unlikely(offset + sizeof(__sum16) > skb_headlen(skb))) { DO_ONCE_LITE(skb_dump, KERN_ERR, skb, false); + WARN_ONCE(true, "offset+2 (%zu) > skb_headlen() (%u)\n", + offset + sizeof(__sum16), skb_headlen(skb)); goto out; } ret = skb_ensure_writable(skb, offset + sizeof(__sum16)); @@ -3285,9 +3601,10 @@ out: } EXPORT_SYMBOL(skb_checksum_help); +#ifdef CONFIG_NET_CRC32C int skb_crc32c_csum_help(struct sk_buff *skb) { - __le32 crc32c_csum; + u32 crc; int ret = 0, offset, start; if (skb->ip_summed != CHECKSUM_PARTIAL) @@ -3315,15 +3632,14 @@ int skb_crc32c_csum_help(struct sk_buff *skb) if (ret) goto out; - crc32c_csum = cpu_to_le32(~__skb_checksum(skb, start, - skb->len - start, ~(__u32)0, - crc32c_csum_stub)); - *(__le32 *)(skb->data + offset) = crc32c_csum; - skb->ip_summed = CHECKSUM_NONE; - skb->csum_not_inet = 0; + crc = ~skb_crc32c(skb, start, skb->len - start, ~0); + *(__le32 *)(skb->data + offset) = cpu_to_le32(crc); + skb_reset_csum_not_inet(skb); out: return ret; } +EXPORT_SYMBOL(skb_crc32c_csum_help); +#endif /* CONFIG_NET_CRC32C */ __be16 skb_network_protocol(struct sk_buff *skb, int *depth) { @@ -3340,77 +3656,9 @@ __be16 skb_network_protocol(struct sk_buff *skb, int *depth) type = eth->h_proto; } - return __vlan_get_protocol(skb, type, depth); + return vlan_get_protocol_and_depth(skb, type, depth); } -/* openvswitch calls this on rx path, so we need a different check. - */ -static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path) -{ - if (tx_path) - return skb->ip_summed != CHECKSUM_PARTIAL && - skb->ip_summed != CHECKSUM_UNNECESSARY; - - return skb->ip_summed == CHECKSUM_NONE; -} - -/** - * __skb_gso_segment - Perform segmentation on skb. - * @skb: buffer to segment - * @features: features for the output path (see dev->features) - * @tx_path: whether it is called in TX path - * - * This function segments the given skb and returns a list of segments. - * - * It may return NULL if the skb requires no segmentation. This is - * only possible when GSO is used for verifying header integrity. - * - * Segmentation preserves SKB_GSO_CB_OFFSET bytes of previous skb cb. - */ -struct sk_buff *__skb_gso_segment(struct sk_buff *skb, - netdev_features_t features, bool tx_path) -{ - struct sk_buff *segs; - - if (unlikely(skb_needs_check(skb, tx_path))) { - int err; - - /* We're going to init ->check field in TCP or UDP header */ - err = skb_cow_head(skb, 0); - if (err < 0) - return ERR_PTR(err); - } - - /* Only report GSO partial support if it will enable us to - * support segmentation on this frame without needing additional - * work. - */ - if (features & NETIF_F_GSO_PARTIAL) { - netdev_features_t partial_features = NETIF_F_GSO_ROBUST; - struct net_device *dev = skb->dev; - - partial_features |= dev->features & dev->gso_partial_features; - if (!skb_gso_ok(skb, features | partial_features)) - features &= ~NETIF_F_GSO_PARTIAL; - } - - BUILD_BUG_ON(SKB_GSO_CB_OFFSET + - sizeof(*SKB_GSO_CB(skb)) > sizeof(skb->cb)); - - SKB_GSO_CB(skb)->mac_offset = skb_headroom(skb); - SKB_GSO_CB(skb)->encap_level = 0; - - skb_reset_mac_header(skb); - skb_reset_mac_len(skb); - - segs = skb_mac_gso_segment(skb, features); - - if (segs != skb && unlikely(skb_needs_check(skb, tx_path) && !IS_ERR(segs))) - skb_warn_bad_offload(skb); - - return segs; -} -EXPORT_SYMBOL(__skb_gso_segment); /* Take action when hardware reception checksum errors are detected. */ #ifdef CONFIG_BUG @@ -3437,8 +3685,9 @@ static int illegal_highdma(struct net_device *dev, struct sk_buff *skb) if (!(dev->features & NETIF_F_HIGHDMA)) { for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) { skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; + struct page *page = skb_frag_page(frag); - if (PageHighMem(skb_frag_page(frag))) + if (page && PageHighMem(page)) return 1; } } @@ -3510,6 +3759,9 @@ static netdev_features_t gso_features_check(const struct sk_buff *skb, if (gso_segs > READ_ONCE(dev->gso_max_segs)) return features & ~NETIF_F_GSO_MASK; + if (unlikely(skb->len >= netif_get_gso_max_size(dev, skb))) + return features & ~NETIF_F_GSO_MASK; + if (!skb_shinfo(skb)->gso_type) { skb_warn_bad_offload(skb); return features & ~NETIF_F_GSO_MASK; @@ -3524,8 +3776,14 @@ static netdev_features_t gso_features_check(const struct sk_buff *skb, if (!(skb_shinfo(skb)->gso_type & SKB_GSO_PARTIAL)) features &= ~dev->gso_partial_features; - /* Make sure to clear the IPv4 ID mangling feature if the - * IPv4 header has the potential to be fragmented. + /* Make sure to clear the IPv4 ID mangling feature if the IPv4 header + * has the potential to be fragmented so that TSO does not generate + * segments with the same ID. For encapsulated packets, the ID mangling + * feature is guaranteed not to use the same ID for the outer IPv4 + * headers of the generated segments if the headers have the potential + * to be fragmented, so there is no need to clear the IPv4 ID mangling + * feature (see the section about NETIF_F_TSO_MANGLEID in + * segmentation-offloads.rst). */ if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4) { struct iphdr *iph = skb->encapsulation ? @@ -3535,6 +3793,18 @@ static netdev_features_t gso_features_check(const struct sk_buff *skb, features &= ~NETIF_F_TSO_MANGLEID; } + /* NETIF_F_IPV6_CSUM does not support IPv6 extension headers, + * so neither does TSO that depends on it. + */ + if (features & NETIF_F_IPV6_CSUM && + (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6 || + (skb_shinfo(skb)->gso_type & SKB_GSO_UDP_L4 && + vlan_get_protocol(skb) == htons(ETH_P_IPV6))) && + skb_transport_header_was_set(skb) && + skb_network_header_len(skb) != sizeof(struct ipv6hdr) && + !ipv6_has_hopopt_jumbo(skb)) + features &= ~(NETIF_F_IPV6_CSUM | NETIF_F_TSO6 | NETIF_F_GSO_UDP_L4); + return features; } @@ -3575,7 +3845,7 @@ static int xmit_one(struct sk_buff *skb, struct net_device *dev, unsigned int len; int rc; - if (dev_nit_active(dev)) + if (dev_nit_active_rcu(dev)) dev_queue_xmit_nit(skb, dev); len = skb->len; @@ -3634,6 +3904,11 @@ int skb_csum_hwoffload_help(struct sk_buff *skb, return 0; if (features & (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM)) { + if (vlan_get_protocol(skb) == htons(ETH_P_IPV6) && + skb_network_header_len(skb) != sizeof(struct ipv6hdr) && + !ipv6_has_hopopt_jumbo(skb)) + goto sw_checksum; + switch (skb->csum_offset) { case offsetof(struct tcphdr, check): case offsetof(struct udphdr, check): @@ -3641,14 +3916,80 @@ int skb_csum_hwoffload_help(struct sk_buff *skb, } } +sw_checksum: return skb_checksum_help(skb); } EXPORT_SYMBOL(skb_csum_hwoffload_help); +/* Checks if this SKB belongs to an HW offloaded socket + * and whether any SW fallbacks are required based on dev. + * Check decrypted mark in case skb_orphan() cleared socket. + */ +static struct sk_buff *sk_validate_xmit_skb(struct sk_buff *skb, + struct net_device *dev) +{ +#ifdef CONFIG_SOCK_VALIDATE_XMIT + struct sk_buff *(*sk_validate)(struct sock *sk, struct net_device *dev, + struct sk_buff *skb); + struct sock *sk = skb->sk; + + sk_validate = NULL; + if (sk) { + if (sk_fullsock(sk)) + sk_validate = sk->sk_validate_xmit_skb; + else if (sk_is_inet(sk) && sk->sk_state == TCP_TIME_WAIT) + sk_validate = inet_twsk(sk)->tw_validate_xmit_skb; + } + + if (sk_validate) { + skb = sk_validate(sk, dev, skb); + } else if (unlikely(skb_is_decrypted(skb))) { + pr_warn_ratelimited("unencrypted skb with no associated socket - dropping\n"); + kfree_skb(skb); + skb = NULL; + } +#endif + + return skb; +} + +static struct sk_buff *validate_xmit_unreadable_skb(struct sk_buff *skb, + struct net_device *dev) +{ + struct skb_shared_info *shinfo; + struct net_iov *niov; + + if (likely(skb_frags_readable(skb))) + goto out; + + if (!dev->netmem_tx) + goto out_free; + + shinfo = skb_shinfo(skb); + + if (shinfo->nr_frags > 0) { + niov = netmem_to_net_iov(skb_frag_netmem(&shinfo->frags[0])); + if (net_is_devmem_iov(niov) && + net_devmem_iov_binding(niov)->dev != dev) + goto out_free; + } + +out: + return skb; + +out_free: + kfree_skb(skb); + return NULL; +} + static struct sk_buff *validate_xmit_skb(struct sk_buff *skb, struct net_device *dev, bool *again) { netdev_features_t features; + skb = validate_xmit_unreadable_skb(skb, dev); + if (unlikely(!skb)) + goto out_null; + features = netif_skb_features(skb); skb = validate_xmit_vlan(skb, features); if (unlikely(!skb)) @@ -3708,7 +4049,7 @@ struct sk_buff *validate_xmit_skb_list(struct sk_buff *skb, struct net_device *d next = skb->next; skb_mark_not_on_list(skb); - /* in case skb wont be segmented, point to itself */ + /* in case skb won't be segmented, point to itself */ skb->prev = skb; skb = validate_xmit_skb(skb, dev, again); @@ -3728,43 +4069,58 @@ struct sk_buff *validate_xmit_skb_list(struct sk_buff *skb, struct net_device *d } EXPORT_SYMBOL_GPL(validate_xmit_skb_list); -static void qdisc_pkt_len_init(struct sk_buff *skb) +static void qdisc_pkt_len_segs_init(struct sk_buff *skb) { - const struct skb_shared_info *shinfo = skb_shinfo(skb); + struct skb_shared_info *shinfo = skb_shinfo(skb); + u16 gso_segs; qdisc_skb_cb(skb)->pkt_len = skb->len; + if (!shinfo->gso_size) { + qdisc_skb_cb(skb)->pkt_segs = 1; + return; + } + + qdisc_skb_cb(skb)->pkt_segs = gso_segs = shinfo->gso_segs; /* To get more precise estimation of bytes sent on wire, * we add to pkt_len the headers size of all segments */ - if (shinfo->gso_size && skb_transport_header_was_set(skb)) { + if (skb_transport_header_was_set(skb)) { unsigned int hdr_len; - u16 gso_segs = shinfo->gso_segs; /* mac layer + network layer */ - hdr_len = skb_transport_header(skb) - skb_mac_header(skb); + if (!skb->encapsulation) + hdr_len = skb_transport_offset(skb); + else + hdr_len = skb_inner_transport_offset(skb); /* + transport layer */ if (likely(shinfo->gso_type & (SKB_GSO_TCPV4 | SKB_GSO_TCPV6))) { const struct tcphdr *th; struct tcphdr _tcphdr; - th = skb_header_pointer(skb, skb_transport_offset(skb), + th = skb_header_pointer(skb, hdr_len, sizeof(_tcphdr), &_tcphdr); if (likely(th)) hdr_len += __tcp_hdrlen(th); - } else { + } else if (shinfo->gso_type & SKB_GSO_UDP_L4) { struct udphdr _udphdr; - if (skb_header_pointer(skb, skb_transport_offset(skb), + if (skb_header_pointer(skb, hdr_len, sizeof(_udphdr), &_udphdr)) hdr_len += sizeof(struct udphdr); } - if (shinfo->gso_type & SKB_GSO_DODGY) - gso_segs = DIV_ROUND_UP(skb->len - hdr_len, - shinfo->gso_size); + if (unlikely(shinfo->gso_type & SKB_GSO_DODGY)) { + int payload = skb->len - hdr_len; + /* Malicious packet. */ + if (payload <= 0) + return; + gso_segs = DIV_ROUND_UP(payload, shinfo->gso_size); + shinfo->gso_segs = gso_segs; + qdisc_skb_cb(skb)->pkt_segs = gso_segs; + } qdisc_skb_cb(skb)->pkt_len += (gso_segs - 1) * hdr_len; } } @@ -3785,13 +4141,16 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q, struct net_device *dev, struct netdev_queue *txq) { + struct sk_buff *next, *to_free = NULL, *to_free2 = NULL; spinlock_t *root_lock = qdisc_lock(q); - struct sk_buff *to_free = NULL; - bool contended; + struct llist_node *ll_list, *first_n; + unsigned long defer_count = 0; int rc; qdisc_calculate_pkt_len(skb, q); + tcf_set_drop_reason(skb, SKB_DROP_REASON_QDISC_DROP); + if (q->flags & TCQ_F_NOLOCK) { if (q->flags & TCQ_F_CAN_BYPASS && nolock_qdisc_is_empty(q) && qdisc_run_begin(q)) { @@ -3801,9 +4160,9 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q, if (unlikely(!nolock_qdisc_is_empty(q))) { rc = dev_qdisc_enqueue(skb, q, &to_free, txq); __qdisc_run(q); - qdisc_run_end(q); + to_free2 = qdisc_run_end(q); - goto no_lock_out; + goto free_skbs; } qdisc_bstats_cpu_update(q, skb); @@ -3811,74 +4170,93 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q, !nolock_qdisc_is_empty(q)) __qdisc_run(q); - qdisc_run_end(q); - return NET_XMIT_SUCCESS; + to_free2 = qdisc_run_end(q); + rc = NET_XMIT_SUCCESS; + goto free_skbs; } rc = dev_qdisc_enqueue(skb, q, &to_free, txq); - qdisc_run(q); - -no_lock_out: - if (unlikely(to_free)) - kfree_skb_list_reason(to_free, - SKB_DROP_REASON_QDISC_DROP); - return rc; + to_free2 = qdisc_run(q); + goto free_skbs; } - /* - * Heuristic to force contended enqueues to serialize on a - * separate lock before trying to get qdisc main lock. - * This permits qdisc->running owner to get the lock more - * often and dequeue packets faster. - * On PREEMPT_RT it is possible to preempt the qdisc owner during xmit - * and then other tasks will only enqueue packets. The packets will be - * sent after the qdisc owner is scheduled again. To prevent this - * scenario the task always serialize on the lock. + /* Open code llist_add(&skb->ll_node, &q->defer_list) + queue limit. + * In the try_cmpxchg() loop, we want to increment q->defer_count + * at most once to limit the number of skbs in defer_list. + * We perform the defer_count increment only if the list is not empty, + * because some arches have slow atomic_long_inc_return(). + */ + first_n = READ_ONCE(q->defer_list.first); + do { + if (first_n && !defer_count) { + defer_count = atomic_long_inc_return(&q->defer_count); + if (unlikely(defer_count > READ_ONCE(q->limit))) { + kfree_skb_reason(skb, SKB_DROP_REASON_QDISC_DROP); + return NET_XMIT_DROP; + } + } + skb->ll_node.next = first_n; + } while (!try_cmpxchg(&q->defer_list.first, &first_n, &skb->ll_node)); + + /* If defer_list was not empty, we know the cpu which queued + * the first skb will process the whole list for us. */ - contended = qdisc_is_running(q) || IS_ENABLED(CONFIG_PREEMPT_RT); - if (unlikely(contended)) - spin_lock(&q->busylock); + if (first_n) + return NET_XMIT_SUCCESS; spin_lock(root_lock); + + ll_list = llist_del_all(&q->defer_list); + /* There is a small race because we clear defer_count not atomically + * with the prior llist_del_all(). This means defer_list could grow + * over q->limit. + */ + atomic_long_set(&q->defer_count, 0); + + ll_list = llist_reverse_order(ll_list); + if (unlikely(test_bit(__QDISC_STATE_DEACTIVATED, &q->state))) { - __qdisc_drop(skb, &to_free); + llist_for_each_entry_safe(skb, next, ll_list, ll_node) + __qdisc_drop(skb, &to_free); rc = NET_XMIT_DROP; - } else if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) && - qdisc_run_begin(q)) { + goto unlock; + } + if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) && + !llist_next(ll_list) && qdisc_run_begin(q)) { /* * This is a work-conserving queue; there are no old skbs * waiting to be sent out; and the qdisc is not running - * xmit the skb directly. */ + DEBUG_NET_WARN_ON_ONCE(skb != llist_entry(ll_list, + struct sk_buff, + ll_node)); qdisc_bstats_update(q, skb); - - if (sch_direct_xmit(skb, q, dev, txq, root_lock, true)) { - if (unlikely(contended)) { - spin_unlock(&q->busylock); - contended = false; - } + if (sch_direct_xmit(skb, q, dev, txq, root_lock, true)) __qdisc_run(q); - } - - qdisc_run_end(q); + to_free2 = qdisc_run_end(q); rc = NET_XMIT_SUCCESS; } else { - rc = dev_qdisc_enqueue(skb, q, &to_free, txq); - if (qdisc_run_begin(q)) { - if (unlikely(contended)) { - spin_unlock(&q->busylock); - contended = false; - } - __qdisc_run(q); - qdisc_run_end(q); + int count = 0; + + llist_for_each_entry_safe(skb, next, ll_list, ll_node) { + prefetch(next); + prefetch(&next->priority); + skb_mark_not_on_list(skb); + rc = dev_qdisc_enqueue(skb, q, &to_free, txq); + count++; } + to_free2 = qdisc_run(q); + if (count != 1) + rc = NET_XMIT_SUCCESS; } +unlock: spin_unlock(root_lock); - if (unlikely(to_free)) - kfree_skb_list_reason(to_free, SKB_DROP_REASON_QDISC_DROP); - if (unlikely(contended)) - spin_unlock(&q->busylock); + +free_skbs: + tcf_kfree_skb_list(to_free); + tcf_kfree_skb_list(to_free2); return rc; } @@ -3928,69 +4306,242 @@ int dev_loopback_xmit(struct net *net, struct sock *sk, struct sk_buff *skb) EXPORT_SYMBOL(dev_loopback_xmit); #ifdef CONFIG_NET_EGRESS -static struct sk_buff * -sch_handle_egress(struct sk_buff *skb, int *ret, struct net_device *dev) +static struct netdev_queue * +netdev_tx_queue_mapping(struct net_device *dev, struct sk_buff *skb) +{ + int qm = skb_get_queue_mapping(skb); + + return netdev_get_tx_queue(dev, netdev_cap_txqueue(dev, qm)); +} + +#ifndef CONFIG_PREEMPT_RT +static bool netdev_xmit_txqueue_skipped(void) { + return __this_cpu_read(softnet_data.xmit.skip_txqueue); +} + +void netdev_xmit_skip_txqueue(bool skip) +{ + __this_cpu_write(softnet_data.xmit.skip_txqueue, skip); +} +EXPORT_SYMBOL_GPL(netdev_xmit_skip_txqueue); + +#else +static bool netdev_xmit_txqueue_skipped(void) +{ + return current->net_xmit.skip_txqueue; +} + +void netdev_xmit_skip_txqueue(bool skip) +{ + current->net_xmit.skip_txqueue = skip; +} +EXPORT_SYMBOL_GPL(netdev_xmit_skip_txqueue); +#endif +#endif /* CONFIG_NET_EGRESS */ + +#ifdef CONFIG_NET_XGRESS +static int tc_run(struct tcx_entry *entry, struct sk_buff *skb, + enum skb_drop_reason *drop_reason) +{ + int ret = TC_ACT_UNSPEC; #ifdef CONFIG_NET_CLS_ACT - struct mini_Qdisc *miniq = rcu_dereference_bh(dev->miniq_egress); - struct tcf_result cl_res; + struct mini_Qdisc *miniq = rcu_dereference_bh(entry->miniq); + struct tcf_result res; if (!miniq) - return skb; + return ret; + + /* Global bypass */ + if (!static_branch_likely(&tcf_sw_enabled_key)) + return ret; + + /* Block-wise bypass */ + if (tcf_block_bypass_sw(miniq->block)) + return ret; - /* qdisc_skb_cb(skb)->pkt_len was already set by the caller. */ tc_skb_cb(skb)->mru = 0; - tc_skb_cb(skb)->post_ct = false; - mini_qdisc_bstats_cpu_update(miniq, skb); + qdisc_skb_cb(skb)->post_ct = false; + tcf_set_drop_reason(skb, *drop_reason); - switch (tcf_classify(skb, miniq->block, miniq->filter_list, &cl_res, false)) { + mini_qdisc_bstats_cpu_update(miniq, skb); + ret = tcf_classify(skb, miniq->block, miniq->filter_list, &res, false); + /* Only tcf related quirks below. */ + switch (ret) { + case TC_ACT_SHOT: + *drop_reason = tcf_get_drop_reason(skb); + mini_qdisc_qstats_cpu_drop(miniq); + break; case TC_ACT_OK: case TC_ACT_RECLASSIFY: - skb->tc_index = TC_H_MIN(cl_res.classid); + skb->tc_index = TC_H_MIN(res.classid); break; + } +#endif /* CONFIG_NET_CLS_ACT */ + return ret; +} + +static DEFINE_STATIC_KEY_FALSE(tcx_needed_key); + +void tcx_inc(void) +{ + static_branch_inc(&tcx_needed_key); +} + +void tcx_dec(void) +{ + static_branch_dec(&tcx_needed_key); +} + +static __always_inline enum tcx_action_base +tcx_run(const struct bpf_mprog_entry *entry, struct sk_buff *skb, + const bool needs_mac) +{ + const struct bpf_mprog_fp *fp; + const struct bpf_prog *prog; + int ret = TCX_NEXT; + + if (needs_mac) + __skb_push(skb, skb->mac_len); + bpf_mprog_foreach_prog(entry, fp, prog) { + bpf_compute_data_pointers(skb); + ret = bpf_prog_run(prog, skb); + if (ret != TCX_NEXT) + break; + } + if (needs_mac) + __skb_pull(skb, skb->mac_len); + return tcx_action_code(skb, ret); +} + +static __always_inline struct sk_buff * +sch_handle_ingress(struct sk_buff *skb, struct packet_type **pt_prev, int *ret, + struct net_device *orig_dev, bool *another) +{ + struct bpf_mprog_entry *entry = rcu_dereference_bh(skb->dev->tcx_ingress); + enum skb_drop_reason drop_reason = SKB_DROP_REASON_TC_INGRESS; + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; + int sch_ret; + + if (!entry) + return skb; + + bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx); + if (unlikely(*pt_prev)) { + *ret = deliver_skb(skb, *pt_prev, orig_dev); + *pt_prev = NULL; + } + + qdisc_pkt_len_segs_init(skb); + tcx_set_ingress(skb, true); + + if (static_branch_unlikely(&tcx_needed_key)) { + sch_ret = tcx_run(entry, skb, true); + if (sch_ret != TC_ACT_UNSPEC) + goto ingress_verdict; + } + sch_ret = tc_run(tcx_entry(entry), skb, &drop_reason); +ingress_verdict: + switch (sch_ret) { + case TC_ACT_REDIRECT: + /* skb_mac_header check was done by BPF, so we can safely + * push the L2 header back before redirecting to another + * netdev. + */ + __skb_push(skb, skb->mac_len); + if (skb_do_redirect(skb) == -EAGAIN) { + __skb_pull(skb, skb->mac_len); + *another = true; + break; + } + *ret = NET_RX_SUCCESS; + bpf_net_ctx_clear(bpf_net_ctx); + return NULL; case TC_ACT_SHOT: - mini_qdisc_qstats_cpu_drop(miniq); - *ret = NET_XMIT_DROP; - kfree_skb_reason(skb, SKB_DROP_REASON_TC_EGRESS); + kfree_skb_reason(skb, drop_reason); + *ret = NET_RX_DROP; + bpf_net_ctx_clear(bpf_net_ctx); return NULL; + /* used by tc_run */ case TC_ACT_STOLEN: case TC_ACT_QUEUED: case TC_ACT_TRAP: - *ret = NET_XMIT_SUCCESS; consume_skb(skb); + fallthrough; + case TC_ACT_CONSUMED: + *ret = NET_RX_SUCCESS; + bpf_net_ctx_clear(bpf_net_ctx); return NULL; + } + bpf_net_ctx_clear(bpf_net_ctx); + + return skb; +} + +static __always_inline struct sk_buff * +sch_handle_egress(struct sk_buff *skb, int *ret, struct net_device *dev) +{ + struct bpf_mprog_entry *entry = rcu_dereference_bh(dev->tcx_egress); + enum skb_drop_reason drop_reason = SKB_DROP_REASON_TC_EGRESS; + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; + int sch_ret; + + if (!entry) + return skb; + + bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx); + + /* qdisc_skb_cb(skb)->pkt_len & tcx_set_ingress() was + * already set by the caller. + */ + if (static_branch_unlikely(&tcx_needed_key)) { + sch_ret = tcx_run(entry, skb, false); + if (sch_ret != TC_ACT_UNSPEC) + goto egress_verdict; + } + sch_ret = tc_run(tcx_entry(entry), skb, &drop_reason); +egress_verdict: + switch (sch_ret) { case TC_ACT_REDIRECT: /* No need to push/pop skb's mac_header here on egress! */ skb_do_redirect(skb); *ret = NET_XMIT_SUCCESS; + bpf_net_ctx_clear(bpf_net_ctx); + return NULL; + case TC_ACT_SHOT: + kfree_skb_reason(skb, drop_reason); + *ret = NET_XMIT_DROP; + bpf_net_ctx_clear(bpf_net_ctx); + return NULL; + /* used by tc_run */ + case TC_ACT_STOLEN: + case TC_ACT_QUEUED: + case TC_ACT_TRAP: + consume_skb(skb); + fallthrough; + case TC_ACT_CONSUMED: + *ret = NET_XMIT_SUCCESS; + bpf_net_ctx_clear(bpf_net_ctx); return NULL; - default: - break; } -#endif /* CONFIG_NET_CLS_ACT */ + bpf_net_ctx_clear(bpf_net_ctx); return skb; } - -static struct netdev_queue * -netdev_tx_queue_mapping(struct net_device *dev, struct sk_buff *skb) -{ - int qm = skb_get_queue_mapping(skb); - - return netdev_get_tx_queue(dev, netdev_cap_txqueue(dev, qm)); -} - -static bool netdev_xmit_txqueue_skipped(void) +#else +static __always_inline struct sk_buff * +sch_handle_ingress(struct sk_buff *skb, struct packet_type **pt_prev, int *ret, + struct net_device *orig_dev, bool *another) { - return __this_cpu_read(softnet_data.xmit.skip_txqueue); + return skb; } -void netdev_xmit_skip_txqueue(bool skip) +static __always_inline struct sk_buff * +sch_handle_egress(struct sk_buff *skb, int *ret, struct net_device *dev) { - __this_cpu_write(softnet_data.xmit.skip_txqueue, skip); + return skb; } -EXPORT_SYMBOL_GPL(netdev_xmit_skip_txqueue); -#endif /* CONFIG_NET_EGRESS */ +#endif /* CONFIG_NET_XGRESS */ #ifdef CONFIG_XPS static int __get_xps_queue_idx(struct net_device *dev, struct sk_buff *skb, @@ -4069,12 +4620,31 @@ u16 dev_pick_tx_zero(struct net_device *dev, struct sk_buff *skb, } EXPORT_SYMBOL(dev_pick_tx_zero); -u16 dev_pick_tx_cpu_id(struct net_device *dev, struct sk_buff *skb, - struct net_device *sb_dev) +int sk_tx_queue_get(const struct sock *sk) { - return (u16)raw_smp_processor_id() % dev->real_num_tx_queues; + int resel, val; + + if (!sk) + return -1; + /* Paired with WRITE_ONCE() in sk_tx_queue_clear() + * and sk_tx_queue_set(). + */ + val = READ_ONCE(sk->sk_tx_queue_mapping); + + if (val == NO_QUEUE_MAPPING) + return -1; + + if (!sk_fullsock(sk)) + return val; + + resel = READ_ONCE(sock_net(sk)->core.sysctl_txq_reselection); + if (resel && time_is_before_jiffies( + READ_ONCE(sk->sk_tx_queue_mapping_jiffies) + resel)) + return -1; + + return val; } -EXPORT_SYMBOL(dev_pick_tx_cpu_id); +EXPORT_SYMBOL(sk_tx_queue_get); u16 netdev_pick_tx(struct net_device *dev, struct sk_buff *skb, struct net_device *sb_dev) @@ -4091,8 +4661,7 @@ u16 netdev_pick_tx(struct net_device *dev, struct sk_buff *skb, if (new_index < 0) new_index = skb_tx_hash(dev, sb_dev, skb); - if (queue_index != new_index && sk && - sk_fullsock(sk) && + if (sk && sk_fullsock(sk) && rcu_access_pointer(sk->sk_dst_cache)) sk_tx_queue_set(sk, new_index); @@ -4163,7 +4732,8 @@ int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev) skb_reset_mac_header(skb); skb_assert_len(skb); - if (unlikely(skb_shinfo(skb)->tx_flags & SKBTX_SCHED_TSTAMP)) + if (unlikely(skb_shinfo(skb)->tx_flags & + (SKBTX_SCHED_TSTAMP | SKBTX_BPF))) __skb_tstamp_tx(skb, NULL, NULL, skb->sk, SCM_TSTAMP_SCHED); /* Disable soft irqs for various locks below. Also @@ -4173,10 +4743,8 @@ int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev) skb_update_prio(skb); - qdisc_pkt_len_init(skb); -#ifdef CONFIG_NET_CLS_ACT - skb->tc_at_ingress = 0; -#endif + qdisc_pkt_len_segs_init(skb); + tcx_set_ingress(skb, false); #ifdef CONFIG_NET_EGRESS if (static_branch_unlikely(&egress_needed_key)) { if (nf_hook_egress_active()) { @@ -4318,20 +4886,11 @@ EXPORT_SYMBOL(__dev_direct_xmit); /************************************************************************* * Receiver routines *************************************************************************/ +static DEFINE_PER_CPU(struct task_struct *, backlog_napi); -int netdev_max_backlog __read_mostly = 1000; -EXPORT_SYMBOL(netdev_max_backlog); - -int netdev_tstamp_prequeue __read_mostly = 1; -unsigned int sysctl_skb_defer_max __read_mostly = 64; -int netdev_budget __read_mostly = 300; -/* Must be at least 2 jiffes to guarantee 1 jiffy timeout */ -unsigned int __read_mostly netdev_budget_usecs = 2 * USEC_PER_SEC / HZ; 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 */ -int dev_rx_weight __read_mostly = 64; -int dev_tx_weight __read_mostly = 64; /* Called with irq disabled */ static inline void ____napi_schedule(struct softnet_data *sd, @@ -4343,52 +4902,88 @@ static inline void ____napi_schedule(struct softnet_data *sd, if (test_bit(NAPI_STATE_THREADED, &napi->state)) { /* Paired with smp_mb__before_atomic() in - * napi_enable()/dev_set_threaded(). + * napi_enable()/netif_set_threaded(). * Use READ_ONCE() to guarantee a complete * read on napi->thread. Only call * wake_up_process() when it's not NULL. */ 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: + DEBUG_NET_WARN_ON_ONCE(!list_empty(&napi->poll_list)); list_add_tail(&napi->poll_list, &sd->poll_list); - __raise_softirq_irqoff(NET_RX_SOFTIRQ); + WRITE_ONCE(napi->list_owner, smp_processor_id()); + /* If not called from net_rx_action() + * we have to raise NET_RX_SOFTIRQ. + */ + if (!sd->in_net_rx_action) + raise_softirq_irqoff(NET_RX_SOFTIRQ); } #ifdef CONFIG_RPS -/* One global table that all flow-based protocols share. */ -struct rps_sock_flow_table __rcu *rps_sock_flow_table __read_mostly; -EXPORT_SYMBOL(rps_sock_flow_table); -u32 rps_cpu_mask __read_mostly; -EXPORT_SYMBOL(rps_cpu_mask); - struct static_key_false rps_needed __read_mostly; EXPORT_SYMBOL(rps_needed); struct static_key_false rfs_needed __read_mostly; EXPORT_SYMBOL(rfs_needed); +static u32 rfs_slot(u32 hash, const struct rps_dev_flow_table *flow_table) +{ + return hash_32(hash, flow_table->log); +} + +#ifdef CONFIG_RFS_ACCEL +/** + * rps_flow_is_active - check whether the flow is recently active. + * @rflow: Specific flow to check activity. + * @flow_table: per-queue flowtable that @rflow belongs to. + * @cpu: CPU saved in @rflow. + * + * If the CPU has processed many packets since the flow's last activity + * (beyond 10 times the table size), the flow is considered stale. + * + * Return: true if flow was recently active. + */ +static bool rps_flow_is_active(struct rps_dev_flow *rflow, + struct rps_dev_flow_table *flow_table, + unsigned int cpu) +{ + unsigned int flow_last_active; + unsigned int sd_input_head; + + if (cpu >= nr_cpu_ids) + return false; + + sd_input_head = READ_ONCE(per_cpu(softnet_data, cpu).input_queue_head); + flow_last_active = READ_ONCE(rflow->last_qtail); + + return (int)(sd_input_head - flow_last_active) < + (int)(10 << flow_table->log); +} +#endif + static struct rps_dev_flow * set_rps_cpu(struct net_device *dev, struct sk_buff *skb, - struct rps_dev_flow *rflow, u16 next_cpu) + struct rps_dev_flow *rflow, u16 next_cpu, u32 hash, + u32 flow_id) { 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; + struct rps_dev_flow *tmp_rflow; + unsigned int tmp_cpu; u16 rxq_index; int rc; @@ -4404,23 +4999,38 @@ set_rps_cpu(struct net_device *dev, struct sk_buff *skb, flow_table = rcu_dereference(rxqueue->rps_flow_table); if (!flow_table) goto out; - flow_id = skb_get_hash(skb) & flow_table->mask; + + tmp_rflow = &flow_table->flows[flow_id]; + tmp_cpu = READ_ONCE(tmp_rflow->cpu); + + if (READ_ONCE(tmp_rflow->filter) != RPS_NO_FILTER) { + if (rps_flow_is_active(tmp_rflow, flow_table, + tmp_cpu)) { + if (hash != READ_ONCE(tmp_rflow->hash) || + next_cpu == tmp_cpu) + goto out; + } + } + rc = dev->netdev_ops->ndo_rx_flow_steer(dev, skb, rxq_index, flow_id); if (rc < 0) 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; + rflow = tmp_rflow; + WRITE_ONCE(rflow->filter, rc); + WRITE_ONCE(rflow->hash, hash); + + 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; } @@ -4437,6 +5047,7 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb, struct rps_dev_flow_table *flow_table; struct rps_map *map; int cpu = -1; + u32 flow_id; u32 tcpu; u32 hash; @@ -4465,23 +5076,26 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb, if (!hash) goto done; - sock_flow_table = rcu_dereference(rps_sock_flow_table); + sock_flow_table = rcu_dereference(net_hotdata.rps_sock_flow_table); if (flow_table && sock_flow_table) { struct rps_dev_flow *rflow; u32 next_cpu; u32 ident; - /* First check into global flow table if there is a match */ - ident = sock_flow_table->ents[hash & sock_flow_table->mask]; - if ((ident ^ hash) & ~rps_cpu_mask) + /* First check into global flow table if there is a match. + * This READ_ONCE() pairs with WRITE_ONCE() from rps_record_sock_flow(). + */ + ident = READ_ONCE(sock_flow_table->ents[hash & sock_flow_table->mask]); + if ((ident ^ hash) & ~net_hotdata.rps_cpu_mask) goto try_rps; - next_cpu = ident & rps_cpu_mask; + next_cpu = ident & net_hotdata.rps_cpu_mask; /* OK, now we know there is a match, * we can look at the local (per receive queue) flow table */ - rflow = &flow_table->flows[hash & flow_table->mask]; + flow_id = rfs_slot(hash, flow_table); + rflow = &flow_table->flows[flow_id]; tcpu = rflow->cpu; /* @@ -4497,10 +5111,11 @@ 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); + rflow = set_rps_cpu(dev, skb, rflow, next_cpu, hash, + flow_id); } if (tcpu < nr_cpu_ids && cpu_online(tcpu)) { @@ -4544,17 +5159,16 @@ bool rps_may_expire_flow(struct net_device *dev, u16 rxq_index, struct rps_dev_flow_table *flow_table; struct rps_dev_flow *rflow; bool expire = true; - unsigned int cpu; rcu_read_lock(); flow_table = rcu_dereference(rxqueue->rps_flow_table); - if (flow_table && flow_id <= flow_table->mask) { + if (flow_table && flow_id < (1UL << flow_table->log)) { + unsigned int cpu; + 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) < - (int)(10 * flow_table->mask))) + if (READ_ONCE(rflow->filter) == filter_id && + rps_flow_is_active(rflow, flow_table, cpu)) expire = false; } rcu_read_unlock(); @@ -4570,7 +5184,8 @@ static void rps_trigger_softirq(void *data) struct softnet_data *sd = data; ____napi_schedule(sd, &sd->backlog); - sd->received_rps++; + /* Pairs with READ_ONCE() in softnet_seq_show() */ + WRITE_ONCE(sd->received_rps, sd->received_rps + 1); } #endif /* CONFIG_RPS */ @@ -4585,39 +5200,71 @@ static void trigger_rx_softirq(void *data) } /* - * Check if this softnet_data structure is another cpu one - * If yes, queue it to our IPI list and return 1 - * If no, return 0 + * After we queued a packet into sd->input_pkt_queue, + * we need to make sure this queue is serviced soon. + * + * - If this is another cpu queue, link it to our rps_ipi_list, + * and make sure we will process rps_ipi_list from net_rx_action(). + * + * - If this is our own queue, NAPI schedule our backlog. + * Note that this also raises NET_RX_SOFTIRQ. */ -static int napi_schedule_rps(struct softnet_data *sd) +static void napi_schedule_rps(struct softnet_data *sd) { struct softnet_data *mysd = this_cpu_ptr(&softnet_data); #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; - __raise_softirq_irqoff(NET_RX_SOFTIRQ); - return 1; + /* If not called from net_rx_action() or napi_threaded_poll() + * we have to raise NET_RX_SOFTIRQ. + */ + if (!mysd->in_net_rx_action && !mysd->in_napi_threaded_poll) + __raise_softirq_irqoff(NET_RX_SOFTIRQ); + return; } #endif /* CONFIG_RPS */ __napi_schedule_irqoff(&mysd->backlog); - return 0; +} + +void kick_defer_list_purge(unsigned int cpu) +{ + struct softnet_data *sd = &per_cpu(softnet_data, 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 -static bool skb_flow_limit(struct sk_buff *skb, unsigned int qlen) +static bool skb_flow_limit(struct sk_buff *skb, unsigned int qlen, + int max_backlog) { #ifdef CONFIG_NET_FLOW_LIMIT - struct sd_flow_limit *fl; - struct softnet_data *sd; unsigned int old_flow, new_flow; + const struct softnet_data *sd; + struct sd_flow_limit *fl; - if (qlen < (READ_ONCE(netdev_max_backlog) >> 1)) + if (likely(qlen < (max_backlog >> 1))) return false; sd = this_cpu_ptr(&softnet_data); @@ -4625,7 +5272,7 @@ static bool skb_flow_limit(struct sk_buff *skb, unsigned int qlen) rcu_read_lock(); fl = rcu_dereference(sd->flow_limit); if (fl) { - new_flow = skb_get_hash(skb) & (fl->num_buckets - 1); + new_flow = hash_32(skb_get_hash(skb), fl->log_buckets); old_flow = fl->history[fl->history_head]; fl->history[fl->history_head] = new_flow; @@ -4636,7 +5283,8 @@ static bool skb_flow_limit(struct sk_buff *skb, unsigned int qlen) fl->buckets[old_flow]--; if (++fl->buckets[new_flow] > (FLOW_LIMIT_HISTORY >> 1)) { - fl->count++; + /* Pairs with READ_ONCE() in softnet_seq_show() */ + WRITE_ONCE(fl->count, fl->count + 1); rcu_read_unlock(); return true; } @@ -4657,36 +5305,46 @@ 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_DEV_READY; + if (unlikely(!netif_running(skb->dev))) + goto bad_dev; - reason = SKB_DROP_REASON_NOT_SPECIFIED; 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) || + skb_flow_limit(skb, 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(netdev_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 (likely(qlen <= max_backlog)) { + 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: + reason = SKB_DROP_REASON_CPU_BACKLOG; + numa_drop_add(&sd->drop_counters, 1); +bad_dev: dev_core_stats_rx_dropped_inc(skb->dev); kfree_skb_reason(skb, reason); return NET_RX_DROP; @@ -4716,7 +5374,7 @@ static struct netdev_rx_queue *netif_get_rxqueue(struct sk_buff *skb) } u32 bpf_prog_run_generic_xdp(struct sk_buff *skb, struct xdp_buff *xdp, - struct bpf_prog *xdp_prog) + const struct bpf_prog *xdp_prog) { void *orig_data, *orig_data_end, *hard_start; struct netdev_rx_queue *rxqueue; @@ -4741,6 +5399,12 @@ u32 bpf_prog_run_generic_xdp(struct sk_buff *skb, struct xdp_buff *xdp, xdp_init_buff(xdp, frame_sz, &rxqueue->xdp_rxq); xdp_prepare_buff(xdp, hard_start, skb_headroom(skb) - mac_len, skb_headlen(skb) + mac_len, true); + if (skb_is_nonlinear(skb)) { + skb_shinfo(skb)->xdp_frags_size = skb->data_len; + xdp_buff_set_frags_flag(xdp); + } else { + xdp_buff_clear_frags_flag(xdp); + } orig_data_end = xdp->data_end; orig_data = xdp->data; @@ -4770,6 +5434,14 @@ u32 bpf_prog_run_generic_xdp(struct sk_buff *skb, struct xdp_buff *xdp, skb->len += off; /* positive on grow, negative on shrink */ } + /* XDP frag metadata (e.g. nr_frags) are updated in eBPF helpers + * (e.g. bpf_xdp_adjust_tail), we need to update data_len here. + */ + if (xdp_buff_has_frags(xdp)) + skb->data_len = skb_shinfo(skb)->xdp_frags_size; + else + skb->data_len = 0; + /* check if XDP changed eth hdr such SKB needs update */ eth = (struct ethhdr *)xdp->data; if ((orig_eth_type != eth->h_proto) || @@ -4803,11 +5475,38 @@ u32 bpf_prog_run_generic_xdp(struct sk_buff *skb, struct xdp_buff *xdp, return act; } -static u32 netif_receive_generic_xdp(struct sk_buff *skb, +static int +netif_skb_check_for_xdp(struct sk_buff **pskb, const struct bpf_prog *prog) +{ + struct sk_buff *skb = *pskb; + int err, hroom, troom; + + local_lock_nested_bh(&system_page_pool.bh_lock); + err = skb_cow_data_for_xdp(this_cpu_read(system_page_pool.pool), pskb, prog); + local_unlock_nested_bh(&system_page_pool.bh_lock); + if (!err) + return 0; + + /* In case we have to go down the path and also linearize, + * then lets do the pskb_expand_head() work just once here. + */ + hroom = XDP_PACKET_HEADROOM - skb_headroom(skb); + troom = skb->tail + skb->data_len - skb->end; + err = pskb_expand_head(skb, + hroom > 0 ? ALIGN(hroom, NET_SKB_PAD) : 0, + troom > 0 ? troom + 128 : 0, GFP_ATOMIC); + if (err) + return err; + + return skb_linearize(skb); +} + +static u32 netif_receive_generic_xdp(struct sk_buff **pskb, struct xdp_buff *xdp, - struct bpf_prog *xdp_prog) + const struct bpf_prog *xdp_prog) { - u32 act = XDP_DROP; + struct sk_buff *skb = *pskb; + u32 mac_len, act = XDP_DROP; /* Reinjected packets coming from act_mirred or similar should * not get XDP generic processing. @@ -4815,41 +5514,36 @@ static u32 netif_receive_generic_xdp(struct sk_buff *skb, if (skb_is_redirected(skb)) return XDP_PASS; - /* XDP packets must be linear and must have sufficient headroom - * of XDP_PACKET_HEADROOM bytes. This is the guarantee that also - * native XDP provides, thus we need to do it here as well. + /* XDP packets must have sufficient headroom of XDP_PACKET_HEADROOM + * bytes. This is the guarantee that also native XDP provides, + * thus we need to do it here as well. */ + mac_len = skb->data - skb_mac_header(skb); + __skb_push(skb, mac_len); + if (skb_cloned(skb) || skb_is_nonlinear(skb) || skb_headroom(skb) < XDP_PACKET_HEADROOM) { - int hroom = XDP_PACKET_HEADROOM - skb_headroom(skb); - int troom = skb->tail + skb->data_len - skb->end; - - /* In case we have to go down the path and also linearize, - * then lets do the pskb_expand_head() work just once here. - */ - if (pskb_expand_head(skb, - hroom > 0 ? ALIGN(hroom, NET_SKB_PAD) : 0, - troom > 0 ? troom + 128 : 0, GFP_ATOMIC)) - goto do_drop; - if (skb_linearize(skb)) + if (netif_skb_check_for_xdp(pskb, xdp_prog)) goto do_drop; } - act = bpf_prog_run_generic_xdp(skb, xdp, xdp_prog); + __skb_pull(*pskb, mac_len); + + act = bpf_prog_run_generic_xdp(*pskb, xdp, xdp_prog); switch (act) { case XDP_REDIRECT: case XDP_TX: case XDP_PASS: break; default: - bpf_warn_invalid_xdp_action(skb->dev, xdp_prog, act); + bpf_warn_invalid_xdp_action((*pskb)->dev, xdp_prog, act); fallthrough; case XDP_ABORTED: - trace_xdp_exception(skb->dev, xdp_prog, act); + trace_xdp_exception((*pskb)->dev, xdp_prog, act); fallthrough; case XDP_DROP: do_drop: - kfree_skb(skb); + kfree_skb(*pskb); break; } @@ -4862,7 +5556,7 @@ static u32 netif_receive_generic_xdp(struct sk_buff *skb, * and DDOS attacks will be more effective. In-driver-XDP use dedicated TX * queues, so they do not have this starvation issue. */ -void generic_xdp_tx(struct sk_buff *skb, struct bpf_prog *xdp_prog) +void generic_xdp_tx(struct sk_buff *skb, const struct bpf_prog *xdp_prog) { struct net_device *dev = skb->dev; struct netdev_queue *txq; @@ -4887,32 +5581,38 @@ void generic_xdp_tx(struct sk_buff *skb, struct bpf_prog *xdp_prog) static DEFINE_STATIC_KEY_FALSE(generic_xdp_needed_key); -int do_xdp_generic(struct bpf_prog *xdp_prog, struct sk_buff *skb) +int do_xdp_generic(const struct bpf_prog *xdp_prog, struct sk_buff **pskb) { + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; + if (xdp_prog) { struct xdp_buff xdp; u32 act; int err; - act = netif_receive_generic_xdp(skb, &xdp, xdp_prog); + bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx); + act = netif_receive_generic_xdp(pskb, &xdp, xdp_prog); if (act != XDP_PASS) { switch (act) { case XDP_REDIRECT: - err = xdp_do_generic_redirect(skb->dev, skb, + err = xdp_do_generic_redirect((*pskb)->dev, *pskb, &xdp, xdp_prog); if (err) goto out_redir; break; case XDP_TX: - generic_xdp_tx(skb, xdp_prog); + generic_xdp_tx(*pskb, xdp_prog); break; } + bpf_net_ctx_clear(bpf_net_ctx); return XDP_DROP; } + bpf_net_ctx_clear(bpf_net_ctx); } return XDP_PASS; out_redir: - kfree_skb_reason(skb, SKB_DROP_REASON_XDP); + bpf_net_ctx_clear(bpf_net_ctx); + kfree_skb_reason(*pskb, SKB_DROP_REASON_XDP); return XDP_DROP; } EXPORT_SYMBOL_GPL(do_xdp_generic); @@ -4921,7 +5621,7 @@ static int netif_rx_internal(struct sk_buff *skb) { int ret; - net_timestamp_check(READ_ONCE(netdev_tstamp_prequeue), skb); + net_timestamp_check(READ_ONCE(net_hotdata.tstamp_prequeue), skb); trace_netif_rx(skb); @@ -5005,7 +5705,7 @@ int netif_rx(struct sk_buff *skb) } EXPORT_SYMBOL(netif_rx); -static __latent_entropy void net_tx_action(struct softirq_action *h) +static __latent_entropy void net_tx_action(void) { struct softnet_data *sd = this_cpu_ptr(&softnet_data); @@ -5023,16 +5723,17 @@ static __latent_entropy void net_tx_action(struct softirq_action *h) clist = clist->next; WARN_ON(refcount_read(&skb->users)); - if (likely(get_kfree_skb_cb(skb)->reason == SKB_REASON_CONSUMED)) - trace_consume_skb(skb); + if (likely(get_kfree_skb_cb(skb)->reason == SKB_CONSUMED)) + trace_consume_skb(skb, net_tx_action); else trace_kfree_skb(skb, net_tx_action, - SKB_DROP_REASON_NOT_SPECIFIED); + get_kfree_skb_cb(skb)->reason, NULL); if (skb->fclone != SKB_FCLONE_UNAVAILABLE) __kfree_skb(skb); else - __kfree_skb_defer(skb); + __napi_kfree_skb(skb, + get_kfree_skb_cb(skb)->reason); } } @@ -5048,8 +5749,9 @@ static __latent_entropy void net_tx_action(struct softirq_action *h) rcu_read_lock(); while (head) { - struct Qdisc *q = head; spinlock_t *root_lock = NULL; + struct sk_buff *to_free; + struct Qdisc *q = head; head = head->next_sched; @@ -5076,9 +5778,10 @@ static __latent_entropy void net_tx_action(struct softirq_action *h) } clear_bit(__QDISC_STATE_SCHED, &q->state); - qdisc_run(q); + to_free = qdisc_run(q); if (root_lock) spin_unlock(root_lock); + tcf_kfree_skb_list(to_free); } rcu_read_unlock(); @@ -5094,72 +5797,6 @@ int (*br_fdb_test_addr_hook)(struct net_device *dev, EXPORT_SYMBOL_GPL(br_fdb_test_addr_hook); #endif -static inline struct sk_buff * -sch_handle_ingress(struct sk_buff *skb, struct packet_type **pt_prev, int *ret, - struct net_device *orig_dev, bool *another) -{ -#ifdef CONFIG_NET_CLS_ACT - struct mini_Qdisc *miniq = rcu_dereference_bh(skb->dev->miniq_ingress); - struct tcf_result cl_res; - - /* If there's at least one ingress present somewhere (so - * we get here via enabled static key), remaining devices - * that are not configured with an ingress qdisc will bail - * out here. - */ - if (!miniq) - return skb; - - if (*pt_prev) { - *ret = deliver_skb(skb, *pt_prev, orig_dev); - *pt_prev = NULL; - } - - qdisc_skb_cb(skb)->pkt_len = skb->len; - tc_skb_cb(skb)->mru = 0; - tc_skb_cb(skb)->post_ct = false; - skb->tc_at_ingress = 1; - mini_qdisc_bstats_cpu_update(miniq, skb); - - switch (tcf_classify(skb, miniq->block, miniq->filter_list, &cl_res, false)) { - case TC_ACT_OK: - case TC_ACT_RECLASSIFY: - skb->tc_index = TC_H_MIN(cl_res.classid); - break; - case TC_ACT_SHOT: - mini_qdisc_qstats_cpu_drop(miniq); - kfree_skb_reason(skb, SKB_DROP_REASON_TC_INGRESS); - *ret = NET_RX_DROP; - return NULL; - case TC_ACT_STOLEN: - case TC_ACT_QUEUED: - case TC_ACT_TRAP: - consume_skb(skb); - *ret = NET_RX_SUCCESS; - return NULL; - case TC_ACT_REDIRECT: - /* skb_mac_header check was done by cls/act_bpf, so - * we can safely push the L2 header back before - * redirecting to another netdev - */ - __skb_push(skb, skb->mac_len); - if (skb_do_redirect(skb) == -EAGAIN) { - __skb_pull(skb, skb->mac_len); - *another = true; - break; - } - *ret = NET_RX_SUCCESS; - return NULL; - case TC_ACT_CONSUMED: - *ret = NET_RX_SUCCESS; - return NULL; - default: - break; - } -#endif /* CONFIG_NET_CLS_ACT */ - return skb; -} - /** * netdev_is_rx_handler_busy - check if receive handler is registered * @dev: device to check @@ -5254,7 +5891,7 @@ static inline int nf_ingress(struct sk_buff *skb, struct packet_type **pt_prev, if (nf_hook_ingress_active(skb)) { int ingress_retval; - if (*pt_prev) { + if (unlikely(*pt_prev)) { *ret = deliver_skb(skb, *pt_prev, orig_dev); *pt_prev = NULL; } @@ -5270,6 +5907,7 @@ static inline int nf_ingress(struct sk_buff *skb, struct packet_type **pt_prev, static int __netif_receive_skb_core(struct sk_buff **pskb, bool pfmemalloc, struct packet_type **ppt_prev) { + enum skb_drop_reason drop_reason = SKB_DROP_REASON_UNHANDLED_PROTO; struct packet_type *ptype, *pt_prev; rx_handler_func_t *rx_handler; struct sk_buff *skb = *pskb; @@ -5278,15 +5916,21 @@ static int __netif_receive_skb_core(struct sk_buff **pskb, bool pfmemalloc, int ret = NET_RX_DROP; __be16 type; - net_timestamp_check(!READ_ONCE(netdev_tstamp_prequeue), skb); + net_timestamp_check(!READ_ONCE(net_hotdata.tstamp_prequeue), skb); trace_netif_receive_skb(skb); orig_dev = skb->dev; skb_reset_network_header(skb); +#if !defined(CONFIG_DEBUG_NET) + /* We plan to no longer reset the transport header here. + * Give some time to fuzzers and dev build to catch bugs + * in network stacks. + */ if (!skb_transport_header_was_set(skb)) skb_reset_transport_header(skb); +#endif skb_reset_mac_len(skb); pt_prev = NULL; @@ -5300,7 +5944,8 @@ another_round: int ret2; migrate_disable(); - ret2 = do_xdp_generic(rcu_dereference(skb->dev->xdp_prog), skb); + ret2 = do_xdp_generic(rcu_dereference(skb->dev->xdp_prog), + &skb); migrate_enable(); if (ret2 != XDP_PASS) { @@ -5321,14 +5966,15 @@ another_round: if (pfmemalloc) goto skip_taps; - list_for_each_entry_rcu(ptype, &ptype_all, list) { - if (pt_prev) + list_for_each_entry_rcu(ptype, &dev_net_rcu(skb->dev)->ptype_all, + list) { + if (unlikely(pt_prev)) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; } list_for_each_entry_rcu(ptype, &skb->dev->ptype_all, list) { - if (pt_prev) + if (unlikely(pt_prev)) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; } @@ -5353,11 +5999,13 @@ skip_taps: #endif skb_reset_redirect(skb); skip_classify: - if (pfmemalloc && !skb_pfmemalloc_protocol(skb)) + if (pfmemalloc && !skb_pfmemalloc_protocol(skb)) { + drop_reason = SKB_DROP_REASON_PFMEMALLOC; goto drop; + } if (skb_vlan_tag_present(skb)) { - if (pt_prev) { + if (unlikely(pt_prev)) { ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = NULL; } @@ -5369,7 +6017,7 @@ skip_classify: rx_handler = rcu_dereference(skb->dev->rx_handler); if (rx_handler) { - if (pt_prev) { + if (unlikely(pt_prev)) { ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = NULL; } @@ -5433,6 +6081,14 @@ check_vlan_id: deliver_ptype_list_skb(skb, &pt_prev, orig_dev, type, &ptype_base[ntohs(type) & PTYPE_HASH_MASK]); + + /* orig_dev and skb->dev could belong to different netns; + * Even in such case we need to traverse only the list + * coming from skb->dev, as the ptype owner (packet socket) + * will use dev_net(skb->dev) to do namespace filtering. + */ + deliver_ptype_list_skb(skb, &pt_prev, orig_dev, type, + &dev_net_rcu(skb->dev)->ptype_specific); } deliver_ptype_list_skb(skb, &pt_prev, orig_dev, type, @@ -5444,8 +6100,6 @@ check_vlan_id: } if (pt_prev) { - if (unlikely(skb_orphan_frags_rx(skb, GFP_ATOMIC))) - goto drop; *ppt_prev = pt_prev; } else { drop: @@ -5453,7 +6107,8 @@ drop: dev_core_stats_rx_dropped_inc(skb->dev); else dev_core_stats_rx_nohandler_inc(skb->dev); - kfree_skb_reason(skb, SKB_DROP_REASON_UNHANDLED_PROTO); + + kfree_skb_reason(skb, drop_reason); /* Jamal, now you will not able to escape explaining * me how you were going to use this. :-) */ @@ -5546,10 +6201,9 @@ static void __netif_receive_skb_list_core(struct list_head *head, bool pfmemallo struct packet_type *pt_curr = NULL; /* Current (common) orig_dev of sublist */ struct net_device *od_curr = NULL; - struct list_head sublist; struct sk_buff *skb, *next; + LIST_HEAD(sublist); - INIT_LIST_HEAD(&sublist); list_for_each_entry_safe(skb, next, head, list) { struct net_device *orig_dev = skb->dev; struct packet_type *pt_prev = NULL; @@ -5644,7 +6298,7 @@ static int generic_xdp_install(struct net_device *dev, struct netdev_bpf *xdp) static_branch_dec(&generic_xdp_needed_key); } else if (new && !old) { static_branch_inc(&generic_xdp_needed_key); - dev_disable_lro(dev); + netif_disable_lro(dev); dev_disable_gro_hw(dev); } break; @@ -5661,7 +6315,7 @@ static int netif_receive_skb_internal(struct sk_buff *skb) { int ret; - net_timestamp_check(READ_ONCE(netdev_tstamp_prequeue), skb); + net_timestamp_check(READ_ONCE(net_hotdata.tstamp_prequeue), skb); if (skb_defer_rx_timestamp(skb)) return NET_RX_SUCCESS; @@ -5687,11 +6341,11 @@ static int netif_receive_skb_internal(struct sk_buff *skb) void netif_receive_skb_list_internal(struct list_head *head) { struct sk_buff *skb, *next; - struct list_head sublist; + LIST_HEAD(sublist); - INIT_LIST_HEAD(&sublist); list_for_each_entry_safe(skb, next, head, list) { - net_timestamp_check(READ_ONCE(netdev_tstamp_prequeue), skb); + net_timestamp_check(READ_ONCE(net_hotdata.tstamp_prequeue), + skb); skb_list_del_init(skb); if (!skb_defer_rx_timestamp(skb)) list_add_tail(&skb->list, &sublist); @@ -5770,35 +6424,39 @@ void netif_receive_skb_list(struct list_head *head) } EXPORT_SYMBOL(netif_receive_skb_list); -static DEFINE_PER_CPU(struct work_struct, flush_works); - /* Network device is going away, flush any packets still pending */ static void flush_backlog(struct work_struct *work) { struct sk_buff *skb, *tmp; + struct sk_buff_head list; struct softnet_data *sd; + __skb_queue_head_init(&list); 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) { + if (READ_ONCE(skb->dev->reg_state) == NETREG_UNREGISTERING) { __skb_unlink(skb, &sd->input_pkt_queue); - dev_kfree_skb_irq(skb); - input_queue_head_incr(sd); + __skb_queue_tail(&list, skb); + rps_input_queue_head_incr(sd); } } - rps_unlock_irq_enable(sd); + backlog_unlock_irq_enable(sd); + local_lock_nested_bh(&softnet_data.process_queue_bh_lock); skb_queue_walk_safe(&sd->process_queue, skb, tmp) { - if (skb->dev->reg_state == NETREG_UNREGISTERING) { + if (READ_ONCE(skb->dev->reg_state) == NETREG_UNREGISTERING) { __skb_unlink(skb, &sd->process_queue); - kfree_skb(skb); - input_queue_head_incr(sd); + __skb_queue_tail(&list, skb); + rps_input_queue_head_incr(sd); } } + local_unlock_nested_bh(&softnet_data.process_queue_bh_lock); local_bh_enable(); + + __skb_queue_purge_reason(&list, SKB_DROP_REASON_DEV_READY); } static bool flush_required(int cpu) @@ -5807,14 +6465,14 @@ 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 @@ -5826,36 +6484,54 @@ static bool flush_required(int cpu) return true; } +struct flush_backlogs { + cpumask_t flush_cpus; + struct work_struct w[]; +}; + +static struct flush_backlogs *flush_backlogs_alloc(void) +{ + return kmalloc(struct_size_t(struct flush_backlogs, w, nr_cpu_ids), + GFP_KERNEL); +} + +static struct flush_backlogs *flush_backlogs_fallback; +static DEFINE_MUTEX(flush_backlogs_mutex); + static void flush_all_backlogs(void) { - static cpumask_t flush_cpus; + struct flush_backlogs *ptr = flush_backlogs_alloc(); unsigned int cpu; - /* since we are under rtnl lock protection we can use static data - * for the cpumask and avoid allocating on stack the possibly - * large mask - */ - ASSERT_RTNL(); + if (!ptr) { + mutex_lock(&flush_backlogs_mutex); + ptr = flush_backlogs_fallback; + } + cpumask_clear(&ptr->flush_cpus); cpus_read_lock(); - cpumask_clear(&flush_cpus); for_each_online_cpu(cpu) { if (flush_required(cpu)) { - queue_work_on(cpu, system_highpri_wq, - per_cpu_ptr(&flush_works, cpu)); - cpumask_set_cpu(cpu, &flush_cpus); + INIT_WORK(&ptr->w[cpu], flush_backlog); + queue_work_on(cpu, system_highpri_wq, &ptr->w[cpu]); + __cpumask_set_cpu(cpu, &ptr->flush_cpus); } } /* we can have in flight packet[s] on the cpus we are not flushing, * synchronize_net() in unregister_netdevice_many() will take care of - * them + * them. */ - for_each_cpu(cpu, &flush_cpus) - flush_work(per_cpu_ptr(&flush_works, cpu)); + for_each_cpu(cpu, &ptr->flush_cpus) + flush_work(&ptr->w[cpu]); cpus_read_unlock(); + + if (ptr != flush_backlogs_fallback) + kfree(ptr); + else + mutex_unlock(&flush_backlogs_mutex); } static void net_rps_send_ipi(struct softnet_data *remsd) @@ -5880,7 +6556,7 @@ 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(); @@ -5895,7 +6571,7 @@ 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 @@ -5915,21 +6591,26 @@ static int process_backlog(struct napi_struct *napi, int quota) net_rps_action_and_irq_enable(sd); } - napi->weight = READ_ONCE(dev_rx_weight); + napi->weight = READ_ONCE(net_hotdata.dev_rx_weight); while (again) { struct sk_buff *skb; + local_lock_nested_bh(&softnet_data.process_queue_bh_lock); while ((skb = __skb_dequeue(&sd->process_queue))) { + local_unlock_nested_bh(&softnet_data.process_queue_bh_lock); 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; + } + local_lock_nested_bh(&softnet_data.process_queue_bh_lock); } + local_unlock_nested_bh(&softnet_data.process_queue_bh_lock); - 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(). @@ -5939,15 +6620,19 @@ 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 { + local_lock_nested_bh(&softnet_data.process_queue_bh_lock); skb_queue_splice_tail_init(&sd->input_pkt_queue, &sd->process_queue); + local_unlock_nested_bh(&softnet_data.process_queue_bh_lock); } - rps_unlock_irq_enable(sd); + backlog_unlock_irq_enable(sd); } + if (work) + rps_input_queue_head_add(sd, work); return work; } @@ -6035,25 +6720,23 @@ bool napi_complete_done(struct napi_struct *n, int work_done) return false; if (work_done) { - if (n->gro_bitmask) - timeout = READ_ONCE(n->dev->gro_flush_timeout); - n->defer_hard_irqs_count = READ_ONCE(n->dev->napi_defer_hard_irqs); + if (n->gro.bitmask) + timeout = napi_get_gro_flush_timeout(n); + n->defer_hard_irqs_count = napi_get_defer_hard_irqs(n); } if (n->defer_hard_irqs_count > 0) { n->defer_hard_irqs_count--; - timeout = READ_ONCE(n->dev->gro_flush_timeout); + timeout = napi_get_gro_flush_timeout(n); if (timeout) ret = false; } - if (n->gro_bitmask) { - /* When the NAPI instance uses a timeout and keeps postponing - * it, we need to bound somehow the time packets are kept in - * the GRO layer - */ - napi_gro_flush(n, !!timeout); - } - gro_normal_list(n); + /* + * When the NAPI instance uses a timeout and keeps postponing + * it, we need to bound somehow the time packets are kept in + * the GRO layer. + */ + gro_flush_normal(&n->gro, !!timeout); if (unlikely(!list_empty(&n->poll_list))) { /* If n->poll_list is not empty, we need to mask irqs */ @@ -6061,6 +6744,7 @@ bool napi_complete_done(struct napi_struct *n, int work_done) list_del_init(&n->poll_list); local_irq_restore(flags); } + WRITE_ONCE(n->list_owner, -1); val = READ_ONCE(n->state); do { @@ -6090,17 +6774,26 @@ bool napi_complete_done(struct napi_struct *n, int work_done) } EXPORT_SYMBOL(napi_complete_done); -/* must be called under rcu_read_lock(), as we dont take a reference */ -static struct napi_struct *napi_by_id(unsigned int napi_id) +static void skb_defer_free_flush(void) { - unsigned int hash = napi_id % HASH_SIZE(napi_hash); - struct napi_struct *napi; + struct llist_node *free_list; + struct sk_buff *skb, *next; + struct skb_defer_node *sdn; + int node; - hlist_for_each_entry_rcu(napi, &napi_hash[hash], napi_hash_node) - if (napi->napi_id == napi_id) - return napi; + for_each_node(node) { + sdn = this_cpu_ptr(net_hotdata.skb_defer_nodes) + node; - return NULL; + if (llist_empty(&sdn->defer_list)) + continue; + atomic_long_set(&sdn->defer_count, 0); + free_list = llist_del_all(&sdn->defer_list); + + llist_for_each_entry_safe(skb, next, free_list, ll_node) { + prefetch(next); + napi_consume_skb(skb, 1); + } + } } #if defined(CONFIG_NET_RX_BUSY_POLL) @@ -6108,25 +6801,26 @@ static struct napi_struct *napi_by_id(unsigned int napi_id) static void __busy_poll_stop(struct napi_struct *napi, bool skip_schedule) { if (!skip_schedule) { - gro_normal_list(napi); + gro_normal_list(&napi->gro); __napi_schedule(napi); return; } - if (napi->gro_bitmask) { - /* flush too old packets - * If HZ < 1000, flush all packets. - */ - napi_gro_flush(napi, HZ >= 1000); - } + /* Flush too old packets. If HZ < 1000, flush all packets */ + gro_flush_normal(&napi->gro, HZ >= 1000); - gro_normal_list(napi); clear_bit(NAPI_STATE_SCHED, &napi->state); } -static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, bool prefer_busy_poll, - u16 budget) +enum { + NAPI_F_PREFER_BUSY_POLL = 1, + NAPI_F_END_ON_RESCHED = 2, +}; + +static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, + unsigned flags, u16 budget) { + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; bool skip_schedule = false; unsigned long timeout; int rc; @@ -6144,10 +6838,11 @@ static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, bool clear_bit(NAPI_STATE_IN_BUSY_POLL, &napi->state); local_bh_disable(); + bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx); - if (prefer_busy_poll) { - napi->defer_hard_irqs_count = READ_ONCE(napi->dev->napi_defer_hard_irqs); - timeout = READ_ONCE(napi->dev->gro_flush_timeout); + if (flags & NAPI_F_PREFER_BUSY_POLL) { + napi->defer_hard_irqs_count = napi_get_defer_hard_irqs(napi); + timeout = napi_get_gro_flush_timeout(napi); if (napi->defer_hard_irqs_count && timeout) { hrtimer_start(&napi->timer, ns_to_ktime(timeout), HRTIMER_MODE_REL_PINNED); skip_schedule = true; @@ -6166,32 +6861,36 @@ static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, bool netpoll_poll_unlock(have_poll_lock); if (rc == budget) __busy_poll_stop(napi, skip_schedule); + bpf_net_ctx_clear(bpf_net_ctx); local_bh_enable(); } -void napi_busy_loop(unsigned int napi_id, - bool (*loop_end)(void *, unsigned long), - void *loop_end_arg, bool prefer_busy_poll, u16 budget) +static void __napi_busy_loop(unsigned int napi_id, + bool (*loop_end)(void *, unsigned long), + void *loop_end_arg, unsigned flags, u16 budget) { unsigned long start_time = loop_end ? busy_loop_current_time() : 0; int (*napi_poll)(struct napi_struct *napi, int budget); + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; void *have_poll_lock = NULL; struct napi_struct *napi; + WARN_ON_ONCE(!rcu_read_lock_held()); + restart: napi_poll = NULL; - rcu_read_lock(); - napi = napi_by_id(napi_id); if (!napi) - goto out; + return; - preempt_disable(); + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) + preempt_disable(); for (;;) { int work = 0; local_bh_disable(); + bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx); if (!napi_poll) { unsigned long val = READ_ONCE(napi->state); @@ -6200,14 +6899,14 @@ restart: */ if (val & (NAPIF_STATE_DISABLE | NAPIF_STATE_SCHED | NAPIF_STATE_IN_BUSY_POLL)) { - if (prefer_busy_poll) + if (flags & NAPI_F_PREFER_BUSY_POLL) set_bit(NAPI_STATE_PREFER_BUSY_POLL, &napi->state); goto count; } if (cmpxchg(&napi->state, val, val | NAPIF_STATE_IN_BUSY_POLL | NAPIF_STATE_SCHED) != val) { - if (prefer_busy_poll) + if (flags & NAPI_F_PREFER_BUSY_POLL) set_bit(NAPI_STATE_PREFER_BUSY_POLL, &napi->state); goto count; } @@ -6216,22 +6915,28 @@ restart: } work = napi_poll(napi, budget); trace_napi_poll(napi, work, budget); - gro_normal_list(napi); + gro_normal_list(&napi->gro); count: if (work > 0) __NET_ADD_STATS(dev_net(napi->dev), LINUX_MIB_BUSYPOLLRXPACKETS, work); + skb_defer_free_flush(); + bpf_net_ctx_clear(bpf_net_ctx); local_bh_enable(); if (!loop_end || loop_end(loop_end_arg, start_time)) break; if (unlikely(need_resched())) { + if (flags & NAPI_F_END_ON_RESCHED) + break; if (napi_poll) - busy_poll_stop(napi, have_poll_lock, prefer_busy_poll, budget); - preempt_enable(); + busy_poll_stop(napi, have_poll_lock, flags, budget); + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) + preempt_enable(); rcu_read_unlock(); cond_resched(); + rcu_read_lock(); if (loop_end(loop_end_arg, start_time)) return; goto restart; @@ -6239,33 +6944,113 @@ count: cpu_relax(); } if (napi_poll) - busy_poll_stop(napi, have_poll_lock, prefer_busy_poll, budget); - preempt_enable(); -out: + busy_poll_stop(napi, have_poll_lock, flags, budget); + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) + preempt_enable(); +} + +void napi_busy_loop_rcu(unsigned int napi_id, + bool (*loop_end)(void *, unsigned long), + void *loop_end_arg, bool prefer_busy_poll, u16 budget) +{ + unsigned flags = NAPI_F_END_ON_RESCHED; + + if (prefer_busy_poll) + flags |= NAPI_F_PREFER_BUSY_POLL; + + __napi_busy_loop(napi_id, loop_end, loop_end_arg, flags, budget); +} + +void napi_busy_loop(unsigned int napi_id, + bool (*loop_end)(void *, unsigned long), + void *loop_end_arg, bool prefer_busy_poll, u16 budget) +{ + unsigned flags = prefer_busy_poll ? NAPI_F_PREFER_BUSY_POLL : 0; + + rcu_read_lock(); + __napi_busy_loop(napi_id, loop_end, loop_end_arg, flags, budget); rcu_read_unlock(); } EXPORT_SYMBOL(napi_busy_loop); +void napi_suspend_irqs(unsigned int napi_id) +{ + struct napi_struct *napi; + + rcu_read_lock(); + napi = napi_by_id(napi_id); + if (napi) { + unsigned long timeout = napi_get_irq_suspend_timeout(napi); + + if (timeout) + hrtimer_start(&napi->timer, ns_to_ktime(timeout), + HRTIMER_MODE_REL_PINNED); + } + rcu_read_unlock(); +} + +void napi_resume_irqs(unsigned int napi_id) +{ + struct napi_struct *napi; + + rcu_read_lock(); + napi = napi_by_id(napi_id); + if (napi) { + /* If irq_suspend_timeout is set to 0 between the call to + * napi_suspend_irqs and now, the original value still + * determines the safety timeout as intended and napi_watchdog + * will resume irq processing. + */ + if (napi_get_irq_suspend_timeout(napi)) { + local_bh_disable(); + napi_schedule(napi); + local_bh_enable(); + } + } + rcu_read_unlock(); +} + #endif /* CONFIG_NET_RX_BUSY_POLL */ +static void __napi_hash_add_with_id(struct napi_struct *napi, + unsigned int napi_id) +{ + napi->gro.cached_napi_id = napi_id; + + WRITE_ONCE(napi->napi_id, napi_id); + hlist_add_head_rcu(&napi->napi_hash_node, + &napi_hash[napi->napi_id % HASH_SIZE(napi_hash)]); +} + +static void napi_hash_add_with_id(struct napi_struct *napi, + unsigned int napi_id) +{ + unsigned long flags; + + spin_lock_irqsave(&napi_hash_lock, flags); + WARN_ON_ONCE(napi_by_id(napi_id)); + __napi_hash_add_with_id(napi, napi_id); + spin_unlock_irqrestore(&napi_hash_lock, flags); +} + static void napi_hash_add(struct napi_struct *napi) { + unsigned long flags; + if (test_bit(NAPI_STATE_NO_BUSY_POLL, &napi->state)) return; - spin_lock(&napi_hash_lock); + spin_lock_irqsave(&napi_hash_lock, flags); /* 0..NR_CPUS range is reserved for sender_cpu use */ do { - if (unlikely(++napi_gen_id < MIN_NAPI_ID)) + if (unlikely(!napi_id_valid(++napi_gen_id))) napi_gen_id = MIN_NAPI_ID; } while (napi_by_id(napi_gen_id)); - napi->napi_id = napi_gen_id; - hlist_add_head_rcu(&napi->napi_hash_node, - &napi_hash[napi->napi_id % HASH_SIZE(napi_hash)]); + __napi_hash_add_with_id(napi, napi_gen_id); - spin_unlock(&napi_hash_lock); + spin_unlock_irqrestore(&napi_hash_lock, flags); } /* Warning : caller is responsible to make sure rcu grace period @@ -6273,11 +7058,13 @@ static void napi_hash_add(struct napi_struct *napi) */ static void napi_hash_del(struct napi_struct *napi) { - spin_lock(&napi_hash_lock); + unsigned long flags; + + spin_lock_irqsave(&napi_hash_lock, flags); hlist_del_init_rcu(&napi->napi_hash_node); - spin_unlock(&napi_hash_lock); + spin_unlock_irqrestore(&napi_hash_lock, flags); } static enum hrtimer_restart napi_watchdog(struct hrtimer *timer) @@ -6298,75 +7085,423 @@ static enum hrtimer_restart napi_watchdog(struct hrtimer *timer) return HRTIMER_NORESTART; } -static void init_gro_hash(struct napi_struct *napi) +static void napi_stop_kthread(struct napi_struct *napi) { - int i; + unsigned long val, new; + + /* Wait until the napi STATE_THREADED is unset. */ + while (true) { + val = READ_ONCE(napi->state); - for (i = 0; i < GRO_HASH_BUCKETS; i++) { - INIT_LIST_HEAD(&napi->gro_hash[i].list); - napi->gro_hash[i].count = 0; + /* If napi kthread own this napi or the napi is idle, + * STATE_THREADED can be unset here. + */ + if ((val & NAPIF_STATE_SCHED_THREADED) || + !(val & NAPIF_STATE_SCHED)) { + new = val & (~(NAPIF_STATE_THREADED | + NAPIF_STATE_THREADED_BUSY_POLL)); + } else { + msleep(20); + continue; + } + + if (try_cmpxchg(&napi->state, &val, new)) + break; + } + + /* Once STATE_THREADED is unset, wait for SCHED_THREADED to be unset by + * the kthread. + */ + while (true) { + if (!test_bit(NAPI_STATE_SCHED_THREADED, &napi->state)) + break; + + msleep(20); } - napi->gro_bitmask = 0; + + kthread_stop(napi->thread); + napi->thread = NULL; } -int dev_set_threaded(struct net_device *dev, bool threaded) +static void napi_set_threaded_state(struct napi_struct *napi, + enum netdev_napi_threaded threaded_mode) +{ + bool threaded = threaded_mode != NETDEV_NAPI_THREADED_DISABLED; + bool busy_poll = threaded_mode == NETDEV_NAPI_THREADED_BUSY_POLL; + + assign_bit(NAPI_STATE_THREADED, &napi->state, threaded); + assign_bit(NAPI_STATE_THREADED_BUSY_POLL, &napi->state, busy_poll); +} + +int napi_set_threaded(struct napi_struct *napi, + enum netdev_napi_threaded threaded) +{ + if (threaded) { + if (!napi->thread) { + int err = napi_kthread_create(napi); + + if (err) + return err; + } + } + + if (napi->config) + napi->config->threaded = threaded; + + /* Setting/unsetting threaded mode on a napi might not immediately + * take effect, if the current napi instance is actively being + * polled. In this case, the switch between threaded mode and + * softirq mode will happen in the next round of napi_schedule(). + * This should not cause hiccups/stalls to the live traffic. + */ + if (!threaded && napi->thread) { + napi_stop_kthread(napi); + } else { + /* Make sure kthread is created before THREADED bit is set. */ + smp_mb__before_atomic(); + napi_set_threaded_state(napi, threaded); + } + + return 0; +} + +int netif_set_threaded(struct net_device *dev, + enum netdev_napi_threaded threaded) { struct napi_struct *napi; - int err = 0; + int i, err = 0; - if (dev->threaded == threaded) - return 0; + netdev_assert_locked_or_invisible(dev); if (threaded) { list_for_each_entry(napi, &dev->napi_list, dev_list) { if (!napi->thread) { err = napi_kthread_create(napi); if (err) { - threaded = false; + threaded = NETDEV_NAPI_THREADED_DISABLED; break; } } } } - dev->threaded = threaded; + WRITE_ONCE(dev->threaded, threaded); - /* Make sure kthread is created before THREADED bit - * is set. - */ - smp_mb__before_atomic(); + /* The error should not occur as the kthreads are already created. */ + list_for_each_entry(napi, &dev->napi_list, dev_list) + WARN_ON_ONCE(napi_set_threaded(napi, threaded)); - /* Setting/unsetting threaded mode on a napi might not immediately - * take effect, if the current napi instance is actively being - * polled. In this case, the switch between threaded mode and - * softirq mode will happen in the next round of napi_schedule(). - * This should not cause hiccups/stalls to the live traffic. + /* Override the config for all NAPIs even if currently not listed */ + for (i = 0; i < dev->num_napi_configs; i++) + dev->napi_config[i].threaded = threaded; + + return err; +} + +/** + * netif_threaded_enable() - enable threaded NAPIs + * @dev: net_device instance + * + * Enable threaded mode for the NAPI instances of the device. This may be useful + * for devices where multiple NAPI instances get scheduled by a single + * interrupt. Threaded NAPI allows moving the NAPI processing to cores other + * than the core where IRQ is mapped. + * + * This function should be called before @dev is registered. + */ +void netif_threaded_enable(struct net_device *dev) +{ + WARN_ON_ONCE(netif_set_threaded(dev, NETDEV_NAPI_THREADED_ENABLED)); +} +EXPORT_SYMBOL(netif_threaded_enable); + +/** + * netif_queue_set_napi - Associate queue with the napi + * @dev: device to which NAPI and queue belong + * @queue_index: Index of queue + * @type: queue type as RX or TX + * @napi: NAPI context, pass NULL to clear previously set NAPI + * + * Set queue with its corresponding napi context. This should be done after + * registering the NAPI handler for the queue-vector and the queues have been + * mapped to the corresponding interrupt vector. + */ +void netif_queue_set_napi(struct net_device *dev, unsigned int queue_index, + enum netdev_queue_type type, struct napi_struct *napi) +{ + struct netdev_rx_queue *rxq; + struct netdev_queue *txq; + + if (WARN_ON_ONCE(napi && !napi->dev)) + return; + netdev_ops_assert_locked_or_invisible(dev); + + switch (type) { + case NETDEV_QUEUE_TYPE_RX: + rxq = __netif_get_rx_queue(dev, queue_index); + rxq->napi = napi; + return; + case NETDEV_QUEUE_TYPE_TX: + txq = netdev_get_tx_queue(dev, queue_index); + txq->napi = napi; + return; + default: + return; + } +} +EXPORT_SYMBOL(netif_queue_set_napi); + +static void +netif_napi_irq_notify(struct irq_affinity_notify *notify, + const cpumask_t *mask) +{ + struct napi_struct *napi = + container_of(notify, struct napi_struct, notify); +#ifdef CONFIG_RFS_ACCEL + struct cpu_rmap *rmap = napi->dev->rx_cpu_rmap; + int err; +#endif + + if (napi->config && napi->dev->irq_affinity_auto) + cpumask_copy(&napi->config->affinity_mask, mask); + +#ifdef CONFIG_RFS_ACCEL + if (napi->dev->rx_cpu_rmap_auto) { + err = cpu_rmap_update(rmap, napi->napi_rmap_idx, mask); + if (err) + netdev_warn(napi->dev, "RMAP update failed (%d)\n", + err); + } +#endif +} + +#ifdef CONFIG_RFS_ACCEL +static void netif_napi_affinity_release(struct kref *ref) +{ + struct napi_struct *napi = + container_of(ref, struct napi_struct, notify.kref); + struct cpu_rmap *rmap = napi->dev->rx_cpu_rmap; + + netdev_assert_locked(napi->dev); + WARN_ON(test_and_clear_bit(NAPI_STATE_HAS_NOTIFIER, + &napi->state)); + + if (!napi->dev->rx_cpu_rmap_auto) + return; + rmap->obj[napi->napi_rmap_idx] = NULL; + napi->napi_rmap_idx = -1; + cpu_rmap_put(rmap); +} + +int netif_enable_cpu_rmap(struct net_device *dev, unsigned int num_irqs) +{ + if (dev->rx_cpu_rmap_auto) + return 0; + + dev->rx_cpu_rmap = alloc_irq_cpu_rmap(num_irqs); + if (!dev->rx_cpu_rmap) + return -ENOMEM; + + dev->rx_cpu_rmap_auto = true; + return 0; +} +EXPORT_SYMBOL(netif_enable_cpu_rmap); + +static void netif_del_cpu_rmap(struct net_device *dev) +{ + struct cpu_rmap *rmap = dev->rx_cpu_rmap; + + if (!dev->rx_cpu_rmap_auto) + return; + + /* Free the rmap */ + cpu_rmap_put(rmap); + dev->rx_cpu_rmap = NULL; + dev->rx_cpu_rmap_auto = false; +} + +#else +static void netif_napi_affinity_release(struct kref *ref) +{ +} + +int netif_enable_cpu_rmap(struct net_device *dev, unsigned int num_irqs) +{ + return 0; +} +EXPORT_SYMBOL(netif_enable_cpu_rmap); + +static void netif_del_cpu_rmap(struct net_device *dev) +{ +} +#endif + +void netif_set_affinity_auto(struct net_device *dev) +{ + unsigned int i, maxqs, numa; + + maxqs = max(dev->num_tx_queues, dev->num_rx_queues); + numa = dev_to_node(&dev->dev); + + for (i = 0; i < maxqs; i++) + cpumask_set_cpu(cpumask_local_spread(i, numa), + &dev->napi_config[i].affinity_mask); + + dev->irq_affinity_auto = true; +} +EXPORT_SYMBOL(netif_set_affinity_auto); + +void netif_napi_set_irq_locked(struct napi_struct *napi, int irq) +{ + int rc; + + netdev_assert_locked_or_invisible(napi->dev); + + if (napi->irq == irq) + return; + + /* Remove existing resources */ + if (test_and_clear_bit(NAPI_STATE_HAS_NOTIFIER, &napi->state)) + irq_set_affinity_notifier(napi->irq, NULL); + + napi->irq = irq; + if (irq < 0 || + (!napi->dev->rx_cpu_rmap_auto && !napi->dev->irq_affinity_auto)) + return; + + /* Abort for buggy drivers */ + if (napi->dev->irq_affinity_auto && WARN_ON_ONCE(!napi->config)) + return; + +#ifdef CONFIG_RFS_ACCEL + if (napi->dev->rx_cpu_rmap_auto) { + rc = cpu_rmap_add(napi->dev->rx_cpu_rmap, napi); + if (rc < 0) + return; + + cpu_rmap_get(napi->dev->rx_cpu_rmap); + napi->napi_rmap_idx = rc; + } +#endif + + /* Use core IRQ notifier */ + napi->notify.notify = netif_napi_irq_notify; + napi->notify.release = netif_napi_affinity_release; + rc = irq_set_affinity_notifier(irq, &napi->notify); + if (rc) { + netdev_warn(napi->dev, "Unable to set IRQ notifier (%d)\n", + rc); + goto put_rmap; + } + + set_bit(NAPI_STATE_HAS_NOTIFIER, &napi->state); + return; + +put_rmap: +#ifdef CONFIG_RFS_ACCEL + if (napi->dev->rx_cpu_rmap_auto) { + napi->dev->rx_cpu_rmap->obj[napi->napi_rmap_idx] = NULL; + cpu_rmap_put(napi->dev->rx_cpu_rmap); + napi->napi_rmap_idx = -1; + } +#endif + napi->notify.notify = NULL; + napi->notify.release = NULL; +} +EXPORT_SYMBOL(netif_napi_set_irq_locked); + +static void napi_restore_config(struct napi_struct *n) +{ + n->defer_hard_irqs = n->config->defer_hard_irqs; + n->gro_flush_timeout = n->config->gro_flush_timeout; + n->irq_suspend_timeout = n->config->irq_suspend_timeout; + + if (n->dev->irq_affinity_auto && + test_bit(NAPI_STATE_HAS_NOTIFIER, &n->state)) + irq_set_affinity(n->irq, &n->config->affinity_mask); + + /* a NAPI ID might be stored in the config, if so use it. if not, use + * napi_hash_add to generate one for us. */ - list_for_each_entry(napi, &dev->napi_list, dev_list) { - if (threaded) - set_bit(NAPI_STATE_THREADED, &napi->state); + if (n->config->napi_id) { + napi_hash_add_with_id(n, n->config->napi_id); + } else { + napi_hash_add(n); + n->config->napi_id = n->napi_id; + } + + WARN_ON_ONCE(napi_set_threaded(n, n->config->threaded)); +} + +static void napi_save_config(struct napi_struct *n) +{ + n->config->defer_hard_irqs = n->defer_hard_irqs; + n->config->gro_flush_timeout = n->gro_flush_timeout; + n->config->irq_suspend_timeout = n->irq_suspend_timeout; + napi_hash_del(n); +} + +/* Netlink wants the NAPI list to be sorted by ID, if adding a NAPI which will + * inherit an existing ID try to insert it at the right position. + */ +static void +netif_napi_dev_list_add(struct net_device *dev, struct napi_struct *napi) +{ + unsigned int new_id, pos_id; + struct list_head *higher; + struct napi_struct *pos; + + new_id = UINT_MAX; + if (napi->config && napi->config->napi_id) + new_id = napi->config->napi_id; + + higher = &dev->napi_list; + list_for_each_entry(pos, &dev->napi_list, dev_list) { + if (napi_id_valid(pos->napi_id)) + pos_id = pos->napi_id; + else if (pos->config) + pos_id = pos->config->napi_id; else - clear_bit(NAPI_STATE_THREADED, &napi->state); + pos_id = UINT_MAX; + + if (pos_id <= new_id) + break; + higher = &pos->dev_list; } + list_add_rcu(&napi->dev_list, higher); /* adds after higher */ +} - return err; +/* Double check that napi_get_frags() allocates skbs with + * skb->head being backed by slab, not a page fragment. + * This is to make sure bug fixed in 3226b158e67c + * ("net: avoid 32 x truesize under-estimation for tiny skbs") + * does not accidentally come back. + */ +static void napi_get_frags_check(struct napi_struct *napi) +{ + struct sk_buff *skb; + + local_bh_disable(); + skb = napi_get_frags(napi); + WARN_ON_ONCE(skb && skb->head_frag); + napi_free_frags(napi); + local_bh_enable(); } -EXPORT_SYMBOL(dev_set_threaded); -void netif_napi_add_weight(struct net_device *dev, struct napi_struct *napi, - int (*poll)(struct napi_struct *, int), int weight) +void netif_napi_add_weight_locked(struct net_device *dev, + struct napi_struct *napi, + int (*poll)(struct napi_struct *, int), + int weight) { + netdev_assert_locked(dev); if (WARN_ON(test_and_set_bit(NAPI_STATE_LISTED, &napi->state))) return; INIT_LIST_HEAD(&napi->poll_list); INIT_HLIST_NODE(&napi->napi_hash_node); - hrtimer_init(&napi->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED); - napi->timer.function = napi_watchdog; - init_gro_hash(napi); + hrtimer_setup(&napi->timer, napi_watchdog, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED); + gro_init(&napi->gro); napi->skb = NULL; - INIT_LIST_HEAD(&napi->rx_list); - napi->rx_count = 0; napi->poll = poll; if (weight > NAPI_POLL_WEIGHT) netdev_err_once(dev, "%s() called with weight %d\n", __func__, @@ -6376,25 +7511,36 @@ void netif_napi_add_weight(struct net_device *dev, struct napi_struct *napi, #ifdef CONFIG_NETPOLL napi->poll_owner = -1; #endif + napi->list_owner = -1; set_bit(NAPI_STATE_SCHED, &napi->state); set_bit(NAPI_STATE_NPSVC, &napi->state); - list_add_rcu(&napi->dev_list, &dev->napi_list); - napi_hash_add(napi); + netif_napi_dev_list_add(dev, napi); + + /* default settings from sysfs are applied to all NAPIs. any per-NAPI + * configuration will be loaded in napi_enable + */ + napi_set_defer_hard_irqs(napi, READ_ONCE(dev->napi_defer_hard_irqs)); + napi_set_gro_flush_timeout(napi, READ_ONCE(dev->gro_flush_timeout)); + napi_get_frags_check(napi); /* Create kthread for this napi if dev->threaded is set. * Clear dev->threaded if kthread creation failed so that * threaded mode will not be enabled in napi_enable(). */ - if (dev->threaded && napi_kthread_create(napi)) - dev->threaded = 0; + if (napi_get_threaded_config(dev, napi)) + if (napi_kthread_create(napi)) + dev->threaded = NETDEV_NAPI_THREADED_DISABLED; + netif_napi_set_irq_locked(napi, -1); } -EXPORT_SYMBOL(netif_napi_add_weight); +EXPORT_SYMBOL(netif_napi_add_weight_locked); -void napi_disable(struct napi_struct *n) +void napi_disable_locked(struct napi_struct *n) { unsigned long val, new; might_sleep(); + netdev_assert_locked(n->dev); + set_bit(NAPI_STATE_DISABLE, &n->state); val = READ_ONCE(n->state); @@ -6405,26 +7551,47 @@ void napi_disable(struct napi_struct *n) } new = val | NAPIF_STATE_SCHED | NAPIF_STATE_NPSVC; - new &= ~(NAPIF_STATE_THREADED | NAPIF_STATE_PREFER_BUSY_POLL); + new &= ~(NAPIF_STATE_THREADED | + NAPIF_STATE_THREADED_BUSY_POLL | + NAPIF_STATE_PREFER_BUSY_POLL); } while (!try_cmpxchg(&n->state, &val, new)); hrtimer_cancel(&n->timer); + if (n->config) + napi_save_config(n); + else + napi_hash_del(n); + clear_bit(NAPI_STATE_DISABLE, &n->state); } -EXPORT_SYMBOL(napi_disable); +EXPORT_SYMBOL(napi_disable_locked); /** - * napi_enable - enable NAPI scheduling - * @n: NAPI context + * napi_disable() - prevent NAPI from scheduling + * @n: NAPI context * - * Resume NAPI from being scheduled on this context. - * Must be paired with napi_disable. + * Stop NAPI from being scheduled on this context. + * Waits till any outstanding processing completes. + * Takes netdev_lock() for associated net_device. */ -void napi_enable(struct napi_struct *n) +void napi_disable(struct napi_struct *n) +{ + netdev_lock(n->dev); + napi_disable_locked(n); + netdev_unlock(n->dev); +} +EXPORT_SYMBOL(napi_disable); + +void napi_enable_locked(struct napi_struct *n) { unsigned long new, val = READ_ONCE(n->state); + if (n->config) + napi_restore_config(n); + else + napi_hash_add(n); + do { BUG_ON(!test_bit(NAPI_STATE_SCHED, &val)); @@ -6433,40 +7600,54 @@ void napi_enable(struct napi_struct *n) new |= NAPIF_STATE_THREADED; } while (!try_cmpxchg(&n->state, &val, new)); } -EXPORT_SYMBOL(napi_enable); +EXPORT_SYMBOL(napi_enable_locked); -static void flush_gro_hash(struct napi_struct *napi) +/** + * napi_enable() - enable NAPI scheduling + * @n: NAPI context + * + * Enable scheduling of a NAPI instance. + * Must be paired with napi_disable(). + * Takes netdev_lock() for associated net_device. + */ +void napi_enable(struct napi_struct *n) { - int i; - - for (i = 0; i < GRO_HASH_BUCKETS; i++) { - struct sk_buff *skb, *n; - - list_for_each_entry_safe(skb, n, &napi->gro_hash[i].list, list) - kfree_skb(skb); - napi->gro_hash[i].count = 0; - } + netdev_lock(n->dev); + napi_enable_locked(n); + netdev_unlock(n->dev); } +EXPORT_SYMBOL(napi_enable); /* Must be called in process context */ -void __netif_napi_del(struct napi_struct *napi) +void __netif_napi_del_locked(struct napi_struct *napi) { + netdev_assert_locked(napi->dev); + if (!test_and_clear_bit(NAPI_STATE_LISTED, &napi->state)) return; - napi_hash_del(napi); + /* Make sure NAPI is disabled (or was never enabled). */ + WARN_ON(!test_bit(NAPI_STATE_SCHED, &napi->state)); + + if (test_and_clear_bit(NAPI_STATE_HAS_NOTIFIER, &napi->state)) + irq_set_affinity_notifier(napi->irq, NULL); + + if (napi->config) { + napi->index = -1; + napi->config = NULL; + } + list_del_rcu(&napi->dev_list); napi_free_frags(napi); - flush_gro_hash(napi); - napi->gro_bitmask = 0; + gro_cleanup(&napi->gro); if (napi->thread) { kthread_stop(napi->thread); napi->thread = NULL; } } -EXPORT_SYMBOL(__netif_napi_del); +EXPORT_SYMBOL(__netif_napi_del_locked); static int __napi_poll(struct napi_struct *n, bool *repoll) { @@ -6481,9 +7662,11 @@ static int __napi_poll(struct napi_struct *n, bool *repoll) * accidentally calling ->poll() when NAPI is not scheduled. */ work = 0; - if (test_bit(NAPI_STATE_SCHED, &n->state)) { + if (napi_is_scheduled(n)) { work = n->poll(n, weight); trace_napi_poll(n, work, weight); + + xdp_do_check_flushed(n); } if (unlikely(work > weight)) @@ -6516,14 +7699,8 @@ static int __napi_poll(struct napi_struct *n, bool *repoll) return work; } - if (n->gro_bitmask) { - /* flush too old packets - * If HZ < 1000, flush all packets. - */ - napi_gro_flush(n, HZ >= 1000); - } - - gro_normal_list(n); + /* Flush too old packets. If HZ < 1000, flush all packets */ + gro_flush_normal(&n->gro, HZ >= 1000); /* Some drivers may have called napi_schedule * prior to exhausting their budget. @@ -6551,9 +7728,14 @@ static int napi_poll(struct napi_struct *n, struct list_head *repoll) work = __napi_poll(n, &do_repoll); - if (do_repoll) + if (do_repoll) { +#if defined(CONFIG_DEBUG_NET) + if (unlikely(!napi_is_scheduled(n))) + pr_crit("repoll requested for device %s %ps but napi is not scheduled.\n", + n->dev->name, n->poll); +#endif list_add_tail(&n->poll_list, repoll); - + } netpoll_poll_unlock(have); return work; @@ -6561,8 +7743,6 @@ 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()) { @@ -6571,15 +7751,13 @@ 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); @@ -6587,63 +7765,93 @@ 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, bool busy_poll) { - struct napi_struct *napi = data; - void *have; + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; + struct softnet_data *sd; + unsigned long last_qs = jiffies; - while (!napi_thread_wait(napi)) { - for (;;) { - bool repoll = false; + for (;;) { + bool repoll = false; + void *have; - local_bh_disable(); + local_bh_disable(); + bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx); - have = netpoll_poll_lock(napi); - __napi_poll(napi, &repoll); - netpoll_poll_unlock(have); + sd = this_cpu_ptr(&softnet_data); + sd->in_napi_threaded_poll = true; - local_bh_enable(); + have = netpoll_poll_lock(napi); + __napi_poll(napi, &repoll); + netpoll_poll_unlock(have); - if (!repoll) - break; + 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(); + bpf_net_ctx_clear(bpf_net_ctx); + + /* When busy poll is enabled, the old packets are not flushed in + * napi_complete_done. So flush them here. + */ + if (busy_poll) + gro_flush_normal(&napi->gro, HZ >= 1000); + local_bh_enable(); + /* Call cond_resched here to avoid watchdog warnings. */ + if (repoll || busy_poll) { + rcu_softirq_qs_periodic(last_qs); cond_resched(); } + + if (!repoll) + break; } - return 0; } -static void skb_defer_free_flush(struct softnet_data *sd) +static int napi_threaded_poll(void *data) { - struct sk_buff *skb, *next; - unsigned long flags; + struct napi_struct *napi = data; + bool want_busy_poll; + bool in_busy_poll; + unsigned long val; - /* Paired with WRITE_ONCE() in skb_attempt_defer_free() */ - if (!READ_ONCE(sd->defer_list)) - return; + while (!napi_thread_wait(napi)) { + val = READ_ONCE(napi->state); - spin_lock_irqsave(&sd->defer_lock, flags); - skb = sd->defer_list; - sd->defer_list = NULL; - sd->defer_count = 0; - spin_unlock_irqrestore(&sd->defer_lock, flags); + want_busy_poll = val & NAPIF_STATE_THREADED_BUSY_POLL; + in_busy_poll = val & NAPIF_STATE_IN_BUSY_POLL; - while (skb != NULL) { - next = skb->next; - napi_consume_skb(skb, 1); - skb = next; + if (unlikely(val & NAPIF_STATE_DISABLE)) + want_busy_poll = false; + + if (want_busy_poll != in_busy_poll) + assign_bit(NAPI_STATE_IN_BUSY_POLL, &napi->state, + want_busy_poll); + + napi_threaded_poll_loop(napi, want_busy_poll); } + + return 0; } -static __latent_entropy void net_rx_action(struct softirq_action *h) +static __latent_entropy void net_rx_action(void) { struct softnet_data *sd = this_cpu_ptr(&softnet_data); unsigned long time_limit = jiffies + - usecs_to_jiffies(READ_ONCE(netdev_budget_usecs)); - int budget = READ_ONCE(netdev_budget); + usecs_to_jiffies(READ_ONCE(net_hotdata.netdev_budget_usecs)); + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; + int budget = READ_ONCE(net_hotdata.netdev_budget); LIST_HEAD(list); LIST_HEAD(repoll); + bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx); +start: + sd->in_net_rx_action = true; local_irq_disable(); list_splice_init(&sd->poll_list, &list); local_irq_enable(); @@ -6651,11 +7859,21 @@ static __latent_entropy void net_rx_action(struct softirq_action *h) for (;;) { struct napi_struct *n; - skb_defer_free_flush(sd); + skb_defer_free_flush(); if (list_empty(&list)) { - if (!sd_has_rps_ipi_waiting(sd) && list_empty(&repoll)) - goto end; + if (list_empty(&repoll)) { + sd->in_net_rx_action = false; + barrier(); + /* We need to check if ____napi_schedule() + * had refilled poll_list while + * sd->in_net_rx_action was true. + */ + if (!list_empty(&sd->poll_list)) + goto start; + if (!sd_has_rps_ipi_waiting(sd)) + goto end; + } break; } @@ -6668,7 +7886,8 @@ static __latent_entropy void net_rx_action(struct softirq_action *h) */ if (unlikely(budget <= 0 || time_after_eq(jiffies, time_limit))) { - sd->time_squeeze++; + /* Pairs with READ_ONCE() in softnet_seq_show() */ + WRITE_ONCE(sd->time_squeeze, sd->time_squeeze + 1); break; } } @@ -6680,9 +7899,12 @@ static __latent_entropy void net_rx_action(struct softirq_action *h) list_splice(&list, &sd->poll_list); if (!list_empty(&sd->poll_list)) __raise_softirq_irqoff(NET_RX_SOFTIRQ); + else + sd->in_net_rx_action = false; net_rps_action_and_irq_enable(sd); -end:; +end: + bpf_net_ctx_clear(bpf_net_ctx); } struct netdev_adjacent { @@ -8298,30 +9520,31 @@ 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) { - pr_info("device %s %s promiscuous mode\n", - dev->name, - dev->flags & IFF_PROMISC ? "entered" : "left"); + 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) { current_uid_gid(&uid, &gid); audit_log(audit_context(), GFP_ATOMIC, @@ -8337,23 +9560,20 @@ static int __dev_set_promiscuity(struct net_device *dev, int inc, bool notify) dev_change_rx_flags(dev, IFF_PROMISC); } - if (notify) + if (notify) { + /* The ops lock is only required to ensure consistent locking + * for `NETDEV_CHANGE` notifiers. This function is sometimes + * called without the lock, even for devices that are ops + * locked, such as in `dev_uc_sync_multiple` when using + * bonding or teaming. + */ + netdev_ops_assert_locked(dev); __dev_notify_flags(dev, old_flags, IFF_PROMISC, 0, NULL); + } return 0; } -/** - * dev_set_promiscuity - update promiscuity count on a device - * @dev: device - * @inc: modifier - * - * Add or remove promiscuity from a device. While the count in the device - * remains above zero the interface remains promiscuous. Once it hits zero - * the device reverts back to normal filtering operation. A negative inc - * value is used to drop promiscuity on the device. - * Return 0 if successful or a negative errno code on error. - */ -int dev_set_promiscuity(struct net_device *dev, int inc) +int netif_set_promiscuity(struct net_device *dev, int inc) { unsigned int old_flags = dev->flags; int err; @@ -8365,30 +9585,33 @@ int dev_set_promiscuity(struct net_device *dev, int inc) dev_set_rx_mode(dev); return err; } -EXPORT_SYMBOL(dev_set_promiscuity); -static int __dev_set_allmulti(struct net_device *dev, int inc, bool notify) +int netif_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); dev_set_rx_mode(dev); if (notify) @@ -8398,25 +9621,6 @@ static int __dev_set_allmulti(struct net_device *dev, int inc, bool notify) return 0; } -/** - * dev_set_allmulti - update allmulti count on a device - * @dev: device - * @inc: modifier - * - * Add or remove reception of all multicast frames to a device. While the - * count in the device remains above zero the interface remains listening - * to all interfaces. Once it hits zero the device reverts back to normal - * filtering operation. A negative @inc value is used to drop the counter - * when releasing a resource needing all multicasts. - * Return 0 if successful or a negative errno code on error. - */ - -int dev_set_allmulti(struct net_device *dev, int inc) -{ - return __dev_set_allmulti(dev, inc, true); -} -EXPORT_SYMBOL(dev_set_allmulti); - /* * Upload unicast and multicast address lists to device and * configure RX filtering. When the device doesn't support unicast @@ -8459,21 +9663,21 @@ void dev_set_rx_mode(struct net_device *dev) } /** - * dev_get_flags - get flags reported to userspace - * @dev: device + * netif_get_flags() - get flags reported to userspace + * @dev: device * - * Get the combination of flag bits exported through APIs to userspace. + * Get the combination of flag bits exported through APIs to userspace. */ -unsigned int dev_get_flags(const struct net_device *dev) +unsigned int netif_get_flags(const struct net_device *dev) { unsigned int flags; - flags = (dev->flags & ~(IFF_PROMISC | + flags = (READ_ONCE(dev->flags) & ~(IFF_PROMISC | IFF_ALLMULTI | IFF_RUNNING | IFF_LOWER_UP | IFF_DORMANT)) | - (dev->gflags & (IFF_PROMISC | + (READ_ONCE(dev->gflags) & (IFF_PROMISC | IFF_ALLMULTI)); if (netif_running(dev)) { @@ -8487,7 +9691,7 @@ unsigned int dev_get_flags(const struct net_device *dev) return flags; } -EXPORT_SYMBOL(dev_get_flags); +EXPORT_SYMBOL(netif_get_flags); int __dev_change_flags(struct net_device *dev, unsigned int flags, struct netlink_ext_ack *extack) @@ -8532,7 +9736,7 @@ int __dev_change_flags(struct net_device *dev, unsigned int flags, if ((flags ^ dev->gflags) & IFF_PROMISC) { int inc = (flags & IFF_PROMISC) ? 1 : -1; - unsigned int old_flags = dev->flags; + old_flags = dev->flags; dev->gflags ^= IFF_PROMISC; @@ -8549,7 +9753,7 @@ int __dev_change_flags(struct net_device *dev, unsigned int flags, int inc = (flags & IFF_ALLMULTI) ? 1 : -1; dev->gflags ^= IFF_ALLMULTI; - __dev_set_allmulti(dev, inc, false); + netif_set_allmulti(dev, inc, false); } return ret; @@ -8584,17 +9788,8 @@ void __dev_notify_flags(struct net_device *dev, unsigned int old_flags, } } -/** - * dev_change_flags - change device settings - * @dev: device - * @flags: device state flags - * @extack: netlink extended ack - * - * Change settings on device based state flags. The flags are - * in the userspace exported format. - */ -int dev_change_flags(struct net_device *dev, unsigned int flags, - struct netlink_ext_ack *extack) +int netif_change_flags(struct net_device *dev, unsigned int flags, + struct netlink_ext_ack *extack) { int ret; unsigned int changes, old_flags = dev->flags, old_gflags = dev->gflags; @@ -8607,9 +9802,8 @@ int dev_change_flags(struct net_device *dev, unsigned int flags, __dev_notify_flags(dev, old_flags, changes, 0, NULL); return ret; } -EXPORT_SYMBOL(dev_change_flags); -int __dev_set_mtu(struct net_device *dev, int new_mtu) +int __netif_set_mtu(struct net_device *dev, int new_mtu) { const struct net_device_ops *ops = dev->netdev_ops; @@ -8620,7 +9814,7 @@ int __dev_set_mtu(struct net_device *dev, int new_mtu) WRITE_ONCE(dev->mtu, new_mtu); return 0; } -EXPORT_SYMBOL(__dev_set_mtu); +EXPORT_SYMBOL_NS_GPL(__netif_set_mtu, "NETDEV_INTERNAL"); int dev_validate_mtu(struct net_device *dev, int new_mtu, struct netlink_ext_ack *extack) @@ -8639,18 +9833,22 @@ int dev_validate_mtu(struct net_device *dev, int new_mtu, } /** - * dev_set_mtu_ext - Change maximum transfer unit - * @dev: device - * @new_mtu: new transfer unit - * @extack: netlink extended ack + * netif_set_mtu_ext() - Change maximum transfer unit + * @dev: device + * @new_mtu: new transfer unit + * @extack: netlink extended ack + * + * Change the maximum transfer size of the network device. * - * Change the maximum transfer size of the network device. + * Return: 0 on success, -errno on failure. */ -int dev_set_mtu_ext(struct net_device *dev, int new_mtu, - struct netlink_ext_ack *extack) +int netif_set_mtu_ext(struct net_device *dev, int new_mtu, + struct netlink_ext_ack *extack) { int err, orig_mtu; + netdev_ops_assert_locked(dev); + if (new_mtu == dev->mtu) return 0; @@ -8667,7 +9865,7 @@ int dev_set_mtu_ext(struct net_device *dev, int new_mtu, return err; orig_mtu = dev->mtu; - err = __dev_set_mtu(dev, new_mtu); + err = __netif_set_mtu(dev, new_mtu); if (!err) { err = call_netdevice_notifiers_mtu(NETDEV_CHANGEMTU, dev, @@ -8677,7 +9875,7 @@ int dev_set_mtu_ext(struct net_device *dev, int new_mtu, /* setting mtu back and notifying everyone again, * so that they have a chance to revert changes. */ - __dev_set_mtu(dev, orig_mtu); + __netif_set_mtu(dev, orig_mtu); call_netdevice_notifiers_mtu(NETDEV_CHANGEMTU, dev, new_mtu); } @@ -8685,25 +9883,20 @@ int dev_set_mtu_ext(struct net_device *dev, int new_mtu, return err; } -int dev_set_mtu(struct net_device *dev, int new_mtu) +int netif_set_mtu(struct net_device *dev, int new_mtu) { struct netlink_ext_ack extack; int err; memset(&extack, 0, sizeof(extack)); - err = dev_set_mtu_ext(dev, new_mtu, &extack); + err = netif_set_mtu_ext(dev, new_mtu, &extack); if (err && extack._msg) net_err_ratelimited("%s: %s\n", dev->name, extack._msg); return err; } -EXPORT_SYMBOL(dev_set_mtu); +EXPORT_SYMBOL(netif_set_mtu); -/** - * dev_change_tx_queue_len - Change TX queue length of a netdevice - * @dev: device - * @new_len: new tx queue length - */ -int dev_change_tx_queue_len(struct net_device *dev, unsigned long new_len) +int netif_change_tx_queue_len(struct net_device *dev, unsigned long new_len) { unsigned int orig_len = dev->tx_queue_len; int res; @@ -8712,7 +9905,7 @@ 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) @@ -8726,28 +9919,25 @@ 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; } -/** - * dev_set_group - Change group this device belongs to - * @dev: device - * @new_group: group this device should belong to - */ -void dev_set_group(struct net_device *dev, int new_group) +void netif_set_group(struct net_device *dev, int new_group) { dev->group = new_group; } /** - * dev_pre_changeaddr_notify - Call NETDEV_PRE_CHANGEADDR. - * @dev: device - * @addr: new address - * @extack: netlink extended ack + * netif_pre_changeaddr_notify() - Call NETDEV_PRE_CHANGEADDR. + * @dev: device + * @addr: new address + * @extack: netlink extended ack + * + * Return: 0 on success, -errno on failure. */ -int dev_pre_changeaddr_notify(struct net_device *dev, const char *addr, - struct netlink_ext_ack *extack) +int netif_pre_changeaddr_notify(struct net_device *dev, const char *addr, + struct netlink_ext_ack *extack) { struct netdev_notifier_pre_changeaddr_info info = { .info.dev = dev, @@ -8759,58 +9949,40 @@ int dev_pre_changeaddr_notify(struct net_device *dev, const char *addr, rc = call_netdevice_notifiers_info(NETDEV_PRE_CHANGEADDR, &info.info); return notifier_to_errno(rc); } -EXPORT_SYMBOL(dev_pre_changeaddr_notify); +EXPORT_SYMBOL_NS_GPL(netif_pre_changeaddr_notify, "NETDEV_INTERNAL"); -/** - * dev_set_mac_address - Change Media Access Control Address - * @dev: device - * @sa: new address - * @extack: netlink extended ack - * - * Change the hardware (MAC) address of the device - */ -int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack) +int netif_set_mac_address(struct net_device *dev, struct sockaddr_storage *ss, + struct netlink_ext_ack *extack) { const struct net_device_ops *ops = dev->netdev_ops; int err; if (!ops->ndo_set_mac_address) return -EOPNOTSUPP; - if (sa->sa_family != dev->type) + if (ss->ss_family != dev->type) return -EINVAL; if (!netif_device_present(dev)) return -ENODEV; - err = dev_pre_changeaddr_notify(dev, sa->sa_data, extack); - if (err) - return err; - err = ops->ndo_set_mac_address(dev, sa); + err = netif_pre_changeaddr_notify(dev, ss->__data, extack); if (err) return err; + if (memcmp(dev->dev_addr, ss->__data, dev->addr_len)) { + err = ops->ndo_set_mac_address(dev, ss); + if (err) + return err; + } dev->addr_assign_type = NET_ADDR_SET; call_netdevice_notifiers(NETDEV_CHANGEADDR, dev); add_device_randomness(dev->dev_addr, dev->addr_len); return 0; } -EXPORT_SYMBOL(dev_set_mac_address); -static DECLARE_RWSEM(dev_addr_sem); +DECLARE_RWSEM(dev_addr_sem); -int dev_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack) +/* "sa" is a true struct sockaddr with limited "sa_data" member. */ +int netif_get_mac_address(struct sockaddr *sa, struct net *net, char *dev_name) { - int ret; - - down_write(&dev_addr_sem); - ret = dev_set_mac_address(dev, sa, extack); - up_write(&dev_addr_sem); - return ret; -} -EXPORT_SYMBOL(dev_set_mac_address_user); - -int dev_get_mac_address(struct sockaddr *sa, struct net *net, char *dev_name) -{ - size_t size = sizeof(sa->sa_data_min); + size_t size = sizeof(sa->sa_data); struct net_device *dev; int ret = 0; @@ -8834,16 +10006,9 @@ unlock: up_read(&dev_addr_sem); return ret; } -EXPORT_SYMBOL(dev_get_mac_address); +EXPORT_SYMBOL_NS_GPL(netif_get_mac_address, "NETDEV_INTERNAL"); -/** - * dev_change_carrier - Change device carrier - * @dev: device - * @new_carrier: new value - * - * Change device carrier - */ -int dev_change_carrier(struct net_device *dev, bool new_carrier) +int netif_change_carrier(struct net_device *dev, bool new_carrier) { const struct net_device_ops *ops = dev->netdev_ops; @@ -8894,16 +10059,17 @@ int dev_get_phys_port_name(struct net_device *dev, } /** - * dev_get_port_parent_id - Get the device's port parent identifier - * @dev: network device - * @ppid: pointer to a storage for the port's parent identifier - * @recurse: allow/disallow recursion to lower devices + * netif_get_port_parent_id() - Get the device's port parent identifier + * @dev: network device + * @ppid: pointer to a storage for the port's parent identifier + * @recurse: allow/disallow recursion to lower devices + * + * Get the devices's port parent identifier. * - * Get the devices's port parent identifier + * Return: 0 on success, -errno on failure. */ -int dev_get_port_parent_id(struct net_device *dev, - struct netdev_phys_item_id *ppid, - bool recurse) +int netif_get_port_parent_id(struct net_device *dev, + struct netdev_phys_item_id *ppid, bool recurse) { const struct net_device_ops *ops = dev->netdev_ops; struct netdev_phys_item_id first = { }; @@ -8922,7 +10088,7 @@ int dev_get_port_parent_id(struct net_device *dev, return err; netdev_for_each_lower_dev(dev, lower_dev, iter) { - err = dev_get_port_parent_id(lower_dev, ppid, true); + err = netif_get_port_parent_id(lower_dev, ppid, true); if (err) break; if (!first.id_len) @@ -8933,7 +10099,7 @@ int dev_get_port_parent_id(struct net_device *dev, return err; } -EXPORT_SYMBOL(dev_get_port_parent_id); +EXPORT_SYMBOL(netif_get_port_parent_id); /** * netdev_port_same_parent_id - Indicate if two network devices have @@ -8946,23 +10112,17 @@ bool netdev_port_same_parent_id(struct net_device *a, struct net_device *b) struct netdev_phys_item_id a_id = { }; struct netdev_phys_item_id b_id = { }; - if (dev_get_port_parent_id(a, &a_id, true) || - dev_get_port_parent_id(b, &b_id, true)) + if (netif_get_port_parent_id(a, &a_id, true) || + netif_get_port_parent_id(b, &b_id, true)) return false; return netdev_phys_item_id_same(&a_id, &b_id); } EXPORT_SYMBOL(netdev_port_same_parent_id); -/** - * dev_change_proto_down - set carrier according to proto_down. - * - * @dev: device - * @proto_down: new value - */ -int dev_change_proto_down(struct net_device *dev, bool proto_down) +int netif_change_proto_down(struct net_device *dev, bool proto_down) { - if (!(dev->priv_flags & IFF_CHANGE_PROTO_DOWN)) + if (!dev->change_proto_down) return -EOPNOTSUPP; if (!netif_device_present(dev)) return -ENODEV; @@ -8970,32 +10130,35 @@ 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; } /** - * dev_change_proto_down_reason - proto down reason + * netdev_change_proto_down_reason_locked - proto down reason * * @dev: device * @mask: proto down mask * @value: proto down value */ -void dev_change_proto_down_reason(struct net_device *dev, unsigned long mask, - u32 value) +void netdev_change_proto_down_reason_locked(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 { @@ -9056,6 +10219,40 @@ u8 dev_xdp_prog_count(struct net_device *dev) } EXPORT_SYMBOL_GPL(dev_xdp_prog_count); +u8 dev_xdp_sb_prog_count(struct net_device *dev) +{ + u8 count = 0; + int i; + + for (i = 0; i < __MAX_XDP_MODE; i++) + if (dev->xdp_state[i].prog && + !dev->xdp_state[i].prog->aux->xdp_has_frags) + count++; + return count; +} + +int netif_xdp_propagate(struct net_device *dev, struct netdev_bpf *bpf) +{ + if (!dev->netdev_ops->ndo_bpf) + return -EOPNOTSUPP; + + if (dev->cfg->hds_config == ETHTOOL_TCP_DATA_SPLIT_ENABLED && + bpf->command == XDP_SETUP_PROG && + bpf->prog && !bpf->prog->aux->xdp_has_frags) { + NL_SET_ERR_MSG(bpf->extack, + "unable to propagate XDP to device using tcp-data-split"); + return -EBUSY; + } + + if (dev_get_min_mp_channel_count(dev)) { + NL_SET_ERR_MSG(bpf->extack, "unable to propagate XDP to device using memory provider"); + return -EBUSY; + } + + return dev->netdev_ops->ndo_bpf(dev, bpf); +} +EXPORT_SYMBOL_GPL(netif_xdp_propagate); + u32 dev_xdp_prog_id(struct net_device *dev, enum bpf_xdp_mode mode) { struct bpf_prog *prog = dev_xdp_prog(dev, mode); @@ -9084,6 +10281,19 @@ static int dev_xdp_install(struct net_device *dev, enum bpf_xdp_mode mode, struct netdev_bpf xdp; int err; + netdev_ops_assert_locked(dev); + + if (dev->cfg->hds_config == ETHTOOL_TCP_DATA_SPLIT_ENABLED && + prog && !prog->aux->xdp_has_frags) { + NL_SET_ERR_MSG(extack, "unable to install XDP to device using tcp-data-split"); + return -EBUSY; + } + + if (dev_get_min_mp_channel_count(dev)) { + NL_SET_ERR_MSG(extack, "unable to install XDP to device using memory provider"); + return -EBUSY; + } + memset(&xdp, 0, sizeof(xdp)); xdp.command = mode == XDP_MODE_HW ? XDP_SETUP_PROG_HW : XDP_SETUP_PROG; xdp.extack = extack; @@ -9224,8 +10434,16 @@ static int dev_xdp_attach(struct net_device *dev, struct netlink_ext_ack *extack NL_SET_ERR_MSG(extack, "Native and generic XDP can't be active at the same time"); return -EEXIST; } - if (!offload && bpf_prog_is_dev_bound(new_prog->aux)) { - NL_SET_ERR_MSG(extack, "Using device-bound program without HW_MODE flag is not supported"); + if (!offload && bpf_prog_is_offloaded(new_prog->aux)) { + NL_SET_ERR_MSG(extack, "Using offloaded program without HW_MODE flag is not supported"); + return -EINVAL; + } + if (bpf_prog_is_dev_bound(new_prog->aux) && !bpf_offload_dev_match(new_prog, dev)) { + NL_SET_ERR_MSG(extack, "Program bound to different device"); + return -EINVAL; + } + if (bpf_prog_is_dev_bound(new_prog->aux) && mode == XDP_MODE_SKB) { + NL_SET_ERR_MSG(extack, "Can't attach device-bound programs in generic mode"); return -EINVAL; } if (new_prog->expected_attach_type == BPF_XDP_DEVMAP) { @@ -9297,7 +10515,9 @@ static void bpf_xdp_link_release(struct bpf_link *link) * already NULL, in which case link was already auto-detached */ if (xdp_link->dev) { + netdev_lock_ops(xdp_link->dev); WARN_ON(dev_xdp_detach_link(xdp_link->dev, NULL, xdp_link)); + netdev_unlock_ops(xdp_link->dev); xdp_link->dev = NULL; } @@ -9379,10 +10599,12 @@ static int bpf_xdp_link_update(struct bpf_link *link, struct bpf_prog *new_prog, goto out_unlock; } + netdev_lock_ops(xdp_link->dev); mode = dev_xdp_mode(xdp_link->dev, xdp_link->flags); bpf_op = dev_xdp_bpf_op(xdp_link->dev, mode); err = dev_xdp_install(xdp_link->dev, mode, bpf_op, NULL, xdp_link->flags, new_prog); + netdev_unlock_ops(xdp_link->dev); if (err) goto out_unlock; @@ -9407,6 +10629,7 @@ int bpf_xdp_link_attach(const union bpf_attr *attr, struct bpf_prog *prog) { struct net *net = current->nsproxy->net_ns; struct bpf_link_primer link_primer; + struct netlink_ext_ack extack = {}; struct bpf_xdp_link *link; struct net_device *dev; int err, fd; @@ -9424,7 +10647,8 @@ int bpf_xdp_link_attach(const union bpf_attr *attr, struct bpf_prog *prog) goto unlock; } - bpf_link_init(&link->link, BPF_LINK_TYPE_XDP, &bpf_xdp_link_lops, prog); + bpf_link_init(&link->link, BPF_LINK_TYPE_XDP, &bpf_xdp_link_lops, prog, + attr->link_create.attach_type); link->dev = dev; link->flags = attr->link_create.flags; @@ -9434,12 +10658,15 @@ int bpf_xdp_link_attach(const union bpf_attr *attr, struct bpf_prog *prog) goto unlock; } - err = dev_xdp_attach_link(dev, NULL, link); + netdev_lock_ops(dev); + err = dev_xdp_attach_link(dev, &extack, link); + netdev_unlock_ops(dev); rtnl_unlock(); if (err) { link->dev = NULL; bpf_link_cleanup(&link_primer); + trace_bpf_xdp_link_attach_failed(extack._msg); goto out_put_dev; } @@ -9502,34 +10729,74 @@ err_out: return err; } +u32 dev_get_min_mp_channel_count(const struct net_device *dev) +{ + int i; + + netdev_ops_assert_locked(dev); + + for (i = dev->real_num_rx_queues - 1; i >= 0; i--) + if (dev->_rx[i].mp_params.mp_priv) + /* The channel count is the idx plus 1. */ + return i + 1; + + return 0; +} + /** - * dev_new_index - allocate an ifindex - * @net: the applicable net namespace + * dev_index_reserve() - allocate an ifindex in a namespace + * @net: the applicable net namespace + * @ifindex: requested ifindex, pass %0 to get one allocated + * + * Allocate a ifindex for a new device. Caller must either use the ifindex + * to store the device (via list_netdevice()) or call dev_index_release() + * to give the index up. * - * Returns a suitable unique value for a new device interface - * number. The caller must hold the rtnl semaphore or the - * dev_base_lock to be sure it remains unique. + * Return: a suitable unique value for a new device interface number or -errno. */ -static int dev_new_index(struct net *net) +static int dev_index_reserve(struct net *net, u32 ifindex) { - int ifindex = net->ifindex; + int err; - for (;;) { - if (++ifindex <= 0) - ifindex = 1; - if (!__dev_get_by_index(net, ifindex)) - return net->ifindex = ifindex; + if (ifindex > INT_MAX) { + DEBUG_NET_WARN_ON_ONCE(1); + return -EINVAL; } + + if (!ifindex) + err = xa_alloc_cyclic(&net->dev_by_index, &ifindex, NULL, + xa_limit_31b, &net->ifindex, GFP_KERNEL); + else + err = xa_insert(&net->dev_by_index, ifindex, NULL, GFP_KERNEL); + if (err < 0) + return err; + + return ifindex; +} + +static void dev_index_release(struct net *net, int ifindex) +{ + /* Expect only unused indexes, unlist_netdevice() removes the used */ + WARN_ON(xa_erase(&net->dev_by_index, ifindex)); +} + +static bool from_cleanup_net(void) +{ +#ifdef CONFIG_NET_NS + return current == READ_ONCE(cleanup_net_task); +#else + return false; +#endif } /* Delayed registration/unregisteration */ LIST_HEAD(net_todo_list); DECLARE_WAIT_QUEUE_HEAD(netdev_unregistering_wq); +atomic_t dev_unreg_count = ATOMIC_INIT(0); static void net_set_todo(struct net_device *dev) { list_add_tail(&dev->todo_list, &net_todo_list); - atomic_inc(&dev_net(dev)->dev_unreg_count); } static netdev_features_t netdev_sync_upper_features(struct net_device *lower, @@ -9564,6 +10831,7 @@ static void netdev_sync_lower_features(struct net_device *upper, if (!(features & feature) && (lower->features & feature)) { netdev_dbg(upper, "Disabling feature %pNF on lower dev %s.\n", &feature, lower->name); + netdev_lock_ops(lower); lower->wanted_features &= ~feature; __netdev_update_features(lower); @@ -9572,10 +10840,20 @@ static void netdev_sync_lower_features(struct net_device *upper, &feature, lower->name); else netdev_features_change(lower); + netdev_unlock_ops(lower); } } } +static bool netdev_has_ip_or_hw_csum(netdev_features_t features) +{ + netdev_features_t ip_csum_mask = NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM; + bool ip_csum = (features & ip_csum_mask) == ip_csum_mask; + bool hw_csum = features & NETIF_F_HW_CSUM; + + return ip_csum || hw_csum; +} + static netdev_features_t netdev_fix_features(struct net_device *dev, netdev_features_t features) { @@ -9657,15 +10935,9 @@ static netdev_features_t netdev_fix_features(struct net_device *dev, features &= ~NETIF_F_LRO; } - if (features & NETIF_F_HW_TLS_TX) { - bool ip_csum = (features & (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM)) == - (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM); - bool hw_csum = features & NETIF_F_HW_CSUM; - - if (!ip_csum && !hw_csum) { - netdev_dbg(dev, "Dropping TLS TX HW offload feature since no CSUM feature.\n"); - features &= ~NETIF_F_HW_TLS_TX; - } + if ((features & NETIF_F_HW_TLS_TX) && !netdev_has_ip_or_hw_csum(features)) { + netdev_dbg(dev, "Dropping TLS TX HW offload feature since no CSUM feature.\n"); + features &= ~NETIF_F_HW_TLS_TX; } if ((features & NETIF_F_HW_TLS_RX) && !(features & NETIF_F_RXCSUM)) { @@ -9673,6 +10945,11 @@ static netdev_features_t netdev_fix_features(struct net_device *dev, features &= ~NETIF_F_HW_TLS_RX; } + if ((features & NETIF_F_GSO_UDP_L4) && !netdev_has_ip_or_hw_csum(features)) { + netdev_dbg(dev, "Dropping USO feature since no CSUM feature.\n"); + features &= ~NETIF_F_GSO_UDP_L4; + } + return features; } @@ -9684,6 +10961,7 @@ int __netdev_update_features(struct net_device *dev) int err = -1; ASSERT_RTNL(); + netdev_ops_assert_locked(dev); features = netdev_get_wanted_features(dev); @@ -9736,12 +11014,14 @@ sync_lower: * *before* calling udp_tunnel_get_rx_info, * but *after* calling udp_tunnel_drop_rx_info. */ + udp_tunnel_nic_lock(dev); if (features & NETIF_F_RX_UDP_TUNNEL_PORT) { dev->features = features; udp_tunnel_get_rx_info(dev); } else { udp_tunnel_drop_rx_info(dev); } + udp_tunnel_nic_unlock(dev); } if (diff & NETIF_F_HW_VLAN_CTAG_FILTER) { @@ -9929,6 +11209,65 @@ void netif_tx_stop_all_queues(struct net_device *dev) } EXPORT_SYMBOL(netif_tx_stop_all_queues); +static int netdev_do_alloc_pcpu_stats(struct net_device *dev) +{ + void __percpu *v; + + /* Drivers implementing ndo_get_peer_dev must support tstat + * accounting, so that skb_do_redirect() can bump the dev's + * RX stats upon network namespace switch. + */ + if (dev->netdev_ops->ndo_get_peer_dev && + dev->pcpu_stat_type != NETDEV_PCPU_STAT_TSTATS) + return -EOPNOTSUPP; + + switch (dev->pcpu_stat_type) { + case NETDEV_PCPU_STAT_NONE: + return 0; + case NETDEV_PCPU_STAT_LSTATS: + v = dev->lstats = netdev_alloc_pcpu_stats(struct pcpu_lstats); + break; + case NETDEV_PCPU_STAT_TSTATS: + v = dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats); + break; + case NETDEV_PCPU_STAT_DSTATS: + v = dev->dstats = netdev_alloc_pcpu_stats(struct pcpu_dstats); + break; + default: + return -EINVAL; + } + + return v ? 0 : -ENOMEM; +} + +static void netdev_do_free_pcpu_stats(struct net_device *dev) +{ + switch (dev->pcpu_stat_type) { + case NETDEV_PCPU_STAT_NONE: + return; + case NETDEV_PCPU_STAT_LSTATS: + free_percpu(dev->lstats); + break; + case NETDEV_PCPU_STAT_TSTATS: + free_percpu(dev->tstats); + break; + case NETDEV_PCPU_STAT_DSTATS: + free_percpu(dev->dstats); + break; + } +} + +static void netdev_free_phy_link_topology(struct net_device *dev) +{ + struct phy_link_topology *topo = dev->link_topo; + + if (IS_ENABLED(CONFIG_PHYLIB) && topo) { + xa_destroy(&topo->phys); + kfree(topo); + dev->link_topo = NULL; + } +} + /** * register_netdevice() - register a network device * @dev: device to register @@ -9958,6 +11297,10 @@ int register_netdevice(struct net_device *dev) if (ret) return ret; + /* rss ctx ID 0 is reserved for the default context, start from 1 */ + xa_init_flags(&dev->ethtool->rss_ctx, XA_FLAGS_ALLOC1); + mutex_init(&dev->ethtool->rss_lock); + spin_lock_init(&dev->addr_list_lock); netdev_set_addr_lockdep_class(dev); @@ -9989,12 +11332,15 @@ int register_netdevice(struct net_device *dev) goto err_uninit; } - ret = -EBUSY; - if (!dev->ifindex) - dev->ifindex = dev_new_index(net); - else if (__dev_get_by_index(net, dev->ifindex)) + ret = netdev_do_alloc_pcpu_stats(dev); + if (ret) goto err_uninit; + ret = dev_index_reserve(net, dev->ifindex); + if (ret < 0) + goto err_free_pcpu; + dev->ifindex = ret; + /* Transfer changeable features to wanted_features and enable * software offloads (GSO and GRO). */ @@ -10040,16 +11386,20 @@ int register_netdevice(struct net_device *dev) ret = call_netdevice_notifiers(NETDEV_POST_INIT, dev); ret = notifier_to_errno(ret); if (ret) - goto err_uninit; + goto err_ifindex_release; ret = netdev_register_kobject(dev); - write_lock(&dev_base_lock); - dev->reg_state = ret ? NETREG_UNREGISTERED : NETREG_REGISTERED; - write_unlock(&dev_base_lock); + + netdev_lock(dev); + WRITE_ONCE(dev->reg_state, ret ? NETREG_UNREGISTERED : NETREG_REGISTERED); + netdev_unlock(dev); + if (ret) goto err_uninit_notify; + netdev_lock_ops(dev); __netdev_update_features(dev); + netdev_unlock_ops(dev); /* * Default initial state at registry is that the @@ -10075,7 +11425,9 @@ int register_netdevice(struct net_device *dev) memcpy(dev->perm_addr, dev->dev_addr, dev->addr_len); /* Notify protocols, that a new device appeared. */ + netdev_lock_ops(dev); ret = call_netdevice_notifiers(NETDEV_REGISTER, dev); + netdev_unlock_ops(dev); ret = notifier_to_errno(ret); if (ret) { /* Expect explicit free_netdev() on failure */ @@ -10087,8 +11439,7 @@ int register_netdevice(struct net_device *dev) * Prevent userspace races by waiting until the network * device is fully setup before sending notifications. */ - if (!dev->rtnl_link_ops || - dev->rtnl_link_state == RTNL_LINK_INITIALIZED) + if (!(dev->rtnl_link_ops && dev->rtnl_link_initializing)) rtmsg_ifinfo(RTM_NEWLINK, dev, ~0U, GFP_KERNEL, 0, NULL); out: @@ -10096,6 +11447,10 @@ out: err_uninit_notify: call_netdevice_notifiers(NETDEV_PRE_UNINIT, dev); +err_ifindex_release: + dev_index_release(net, dev->ifindex); +err_free_pcpu: + netdev_do_free_pcpu_stats(dev); err_uninit: if (dev->netdev_ops->ndo_uninit) dev->netdev_ops->ndo_uninit(dev); @@ -10107,49 +11462,26 @@ 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. + * The setup steps dummy netdevs need which normal netdevs get by going + * through register_netdevice(). */ -int init_dummy_netdev(struct net_device *dev) +static void init_dummy_netdev(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 */ dev->reg_state = NETREG_DUMMY; - /* NAPI wants this */ - INIT_LIST_HEAD(&dev->napi_list); - /* a dummy interface is started by default */ set_bit(__LINK_STATE_PRESENT, &dev->state); set_bit(__LINK_STATE_START, &dev->state); - /* napi_busy_loop stats accounting wants this */ - dev_net_set(dev, &init_net); - /* Note : We dont allocate pcpu_refcnt for dummy devices, * because users of this 'device' dont need to change * its refcount. */ - - return 0; } -EXPORT_SYMBOL_GPL(init_dummy_netdev); - /** * register_netdev - register a network device @@ -10166,12 +11498,16 @@ EXPORT_SYMBOL_GPL(init_dummy_netdev); */ int register_netdev(struct net_device *dev) { + struct net *net = dev_net(dev); int err; - if (rtnl_lock_killable()) + if (rtnl_net_lock_killable(net)) return -EINTR; + err = register_netdevice(dev); - rtnl_unlock(); + + rtnl_net_unlock(net); + return err; } EXPORT_SYMBOL(register_netdev); @@ -10248,8 +11584,9 @@ static struct net_device *netdev_wait_allrefs_any(struct list_head *list) rebroadcast_time = jiffies; } + rcu_barrier(); + if (!wait) { - rcu_barrier(); wait = WAIT_REFS_MIN_MSECS; } else { msleep(wait); @@ -10301,15 +11638,15 @@ void netdev_run_todo(void) { struct net_device *dev, *tmp; struct list_head list; + int cnt; #ifdef CONFIG_LOCKDEP struct list_head unlink_list; list_replace_init(&net_unlink_list, &unlink_list); while (!list_empty(&unlink_list)) { - struct net_device *dev = list_first_entry(&unlink_list, - struct net_device, - unlink_list); + dev = list_first_entry(&unlink_list, struct net_device, + unlink_list); list_del_init(&dev->unlink_list); dev->nested_level = dev->lower_level - 1; } @@ -10331,12 +11668,13 @@ void netdev_run_todo(void) continue; } - write_lock(&dev_base_lock); - dev->reg_state = NETREG_UNREGISTERED; - write_unlock(&dev_base_lock); - linkwatch_forget_dev(dev); + netdev_lock(dev); + WRITE_ONCE(dev->reg_state, NETREG_UNREGISTERED); + netdev_unlock(dev); + linkwatch_sync_dev(dev); } + cnt = 0; while (!list_empty(&list)) { dev = netdev_wait_allrefs_any(&list); list_del(&dev->todo_list); @@ -10348,17 +11686,67 @@ void netdev_run_todo(void) WARN_ON(rcu_access_pointer(dev->ip_ptr)); WARN_ON(rcu_access_pointer(dev->ip6_ptr)); + netdev_do_free_pcpu_stats(dev); if (dev->priv_destructor) dev->priv_destructor(dev); if (dev->needs_free_netdev) free_netdev(dev); - if (atomic_dec_and_test(&dev_net(dev)->dev_unreg_count)) - wake_up(&netdev_unregistering_wq); + cnt++; /* Free network device */ kobject_put(&dev->dev.kobj); } + if (cnt && atomic_sub_and_test(cnt, &dev_unreg_count)) + wake_up(&netdev_unregistering_wq); +} + +/* Collate per-cpu network dstats statistics + * + * Read per-cpu network statistics from dev->dstats and populate the related + * fields in @s. + */ +static void dev_fetch_dstats(struct rtnl_link_stats64 *s, + const struct pcpu_dstats __percpu *dstats) +{ + int cpu; + + for_each_possible_cpu(cpu) { + u64 rx_packets, rx_bytes, rx_drops; + u64 tx_packets, tx_bytes, tx_drops; + const struct pcpu_dstats *stats; + unsigned int start; + + stats = per_cpu_ptr(dstats, cpu); + do { + start = u64_stats_fetch_begin(&stats->syncp); + rx_packets = u64_stats_read(&stats->rx_packets); + rx_bytes = u64_stats_read(&stats->rx_bytes); + rx_drops = u64_stats_read(&stats->rx_drops); + tx_packets = u64_stats_read(&stats->tx_packets); + tx_bytes = u64_stats_read(&stats->tx_bytes); + tx_drops = u64_stats_read(&stats->tx_drops); + } while (u64_stats_fetch_retry(&stats->syncp, start)); + + s->rx_packets += rx_packets; + s->rx_bytes += rx_bytes; + s->rx_dropped += rx_drops; + s->tx_packets += tx_packets; + s->tx_bytes += tx_bytes; + s->tx_dropped += tx_drops; + } +} + +/* ndo_get_stats64 implementation for dtstats-based accounting. + * + * Populate @s from dev->stats and dev->dstats. This is used internally by the + * core for NETDEV_PCPU_STAT_DSTAT-type stats collection. + */ +static void dev_get_dstats64(const struct net_device *dev, + struct rtnl_link_stats64 *s) +{ + netdev_stats_to_stats64(s, &dev->stats); + dev_fetch_dstats(s, dev->dstats); } /* Convert net_device_stats to rtnl_link_stats64. rtnl_link_stats64 has @@ -10375,14 +11763,15 @@ void netdev_stats_to_stats64(struct rtnl_link_stats64 *stats64, BUILD_BUG_ON(n > sizeof(*stats64) / sizeof(u64)); for (i = 0; i < n; i++) - dst[i] = atomic_long_read(&src[i]); + dst[i] = (unsigned long)atomic_long_read(&src[i]); /* zero out counters that only exist in rtnl_link_stats64 */ memset((char *)stats64 + n * sizeof(u64), 0, sizeof(*stats64) - n * sizeof(u64)); } EXPORT_SYMBOL(netdev_stats_to_stats64); -struct net_device_core_stats __percpu *netdev_core_stats_alloc(struct net_device *dev) +static __cold struct net_device_core_stats __percpu *netdev_core_stats_alloc( + struct net_device *dev) { struct net_device_core_stats __percpu *p; @@ -10395,7 +11784,23 @@ struct net_device_core_stats __percpu *netdev_core_stats_alloc(struct net_device /* This READ_ONCE() pairs with the cmpxchg() above */ return READ_ONCE(dev->core_stats); } -EXPORT_SYMBOL(netdev_core_stats_alloc); + +noinline void netdev_core_stats_inc(struct net_device *dev, u32 offset) +{ + /* This READ_ONCE() pairs with the write in netdev_core_stats_alloc() */ + struct net_device_core_stats __percpu *p = READ_ONCE(dev->core_stats); + unsigned long __percpu *field; + + if (unlikely(!p)) { + p = netdev_core_stats_alloc(dev); + if (!p) + return; + } + + field = (unsigned long __percpu *)((void __percpu *)p + offset); + this_cpu_inc(*field); +} +EXPORT_SYMBOL_GPL(netdev_core_stats_inc); /** * dev_get_stats - get network device statistics @@ -10413,11 +11818,29 @@ struct rtnl_link_stats64 *dev_get_stats(struct net_device *dev, const struct net_device_ops *ops = dev->netdev_ops; const struct net_device_core_stats __percpu *p; + /* + * IPv{4,6} and udp tunnels share common stat helpers and use + * different stat type (NETDEV_PCPU_STAT_TSTATS vs + * NETDEV_PCPU_STAT_DSTATS). Ensure the accounting is consistent. + */ + BUILD_BUG_ON(offsetof(struct pcpu_sw_netstats, rx_bytes) != + offsetof(struct pcpu_dstats, rx_bytes)); + BUILD_BUG_ON(offsetof(struct pcpu_sw_netstats, rx_packets) != + offsetof(struct pcpu_dstats, rx_packets)); + BUILD_BUG_ON(offsetof(struct pcpu_sw_netstats, tx_bytes) != + offsetof(struct pcpu_dstats, tx_bytes)); + BUILD_BUG_ON(offsetof(struct pcpu_sw_netstats, tx_packets) != + offsetof(struct pcpu_dstats, tx_packets)); + if (ops->ndo_get_stats64) { memset(storage, 0, sizeof(*storage)); ops->ndo_get_stats64(dev, storage); } else if (ops->ndo_get_stats) { netdev_stats_to_stats64(storage, ops->ndo_get_stats(dev)); + } else if (dev->pcpu_stat_type == NETDEV_PCPU_STAT_TSTATS) { + dev_get_tstats64(dev, storage); + } else if (dev->pcpu_stat_type == NETDEV_PCPU_STAT_DSTATS) { + dev_get_dstats64(dev, storage); } else { netdev_stats_to_stats64(storage, &dev->stats); } @@ -10501,7 +11924,7 @@ struct netdev_queue *dev_ingress_queue_create(struct net_device *dev) return NULL; netdev_init_one_queue(dev, queue, NULL); RCU_INIT_POINTER(queue->qdisc, &noop_qdisc); - queue->qdisc_sleeping = &noop_qdisc; + RCU_INIT_POINTER(queue->qdisc_sleeping, &noop_qdisc); rcu_assign_pointer(dev->ingress_queue, queue); #endif return queue; @@ -10528,18 +11951,13 @@ void netdev_sw_irq_coalesce_default_on(struct net_device *dev) { WARN_ON(dev->reg_state == NETREG_REGISTERED); - dev->gro_flush_timeout = 20000; - dev->napi_defer_hard_irqs = 1; + if (!IS_ENABLED(CONFIG_PREEMPT_RT)) { + netdev_set_gro_flush_timeout(dev, 20000); + netdev_set_defer_hard_irqs(dev, 1); + } } EXPORT_SYMBOL_GPL(netdev_sw_irq_coalesce_default_on); -void netdev_freemem(struct net_device *dev) -{ - char *addr = (char *)dev - dev->padded; - - kvfree(addr); -} - /** * alloc_netdev_mqs - allocate network device * @sizeof_priv: size of private data to allocate space for @@ -10559,8 +11977,8 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name, unsigned int txqs, unsigned int rxqs) { struct net_device *dev; - unsigned int alloc_size; - struct net_device *p; + size_t napi_config_sz; + unsigned int maxqs; BUG_ON(strlen(name) >= sizeof(dev->name)); @@ -10574,23 +11992,16 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name, return NULL; } - alloc_size = sizeof(struct net_device); - if (sizeof_priv) { - /* ensure 32-byte alignment of private area */ - alloc_size = ALIGN(alloc_size, NETDEV_ALIGN); - alloc_size += sizeof_priv; - } - /* ensure 32-byte alignment of whole construct */ - alloc_size += NETDEV_ALIGN - 1; + maxqs = max(txqs, rxqs); - p = kvzalloc(alloc_size, GFP_KERNEL_ACCOUNT | __GFP_RETRY_MAYFAIL); - if (!p) + dev = kvzalloc(struct_size(dev, priv, sizeof_priv), + GFP_KERNEL_ACCOUNT | __GFP_RETRY_MAYFAIL); + if (!dev) return NULL; - dev = PTR_ALIGN(p, NETDEV_ALIGN); - dev->padded = (char *)dev - (char *)p; + dev->priv_len = sizeof_priv; - ref_tracker_dir_init(&dev->refcnt_tracker, 128); + ref_tracker_dir_init(&dev->refcnt_tracker, 128, "netdev"); #ifdef CONFIG_PCPU_DEV_REFCNT dev->pcpu_refcnt = alloc_percpu(int); if (!dev->pcpu_refcnt) @@ -10609,8 +12020,11 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name, dev_net_set(dev, &init_net); dev->gso_max_size = GSO_LEGACY_MAX_SIZE; + dev->xdp_zc_max_segs = 1; dev->gso_max_segs = GSO_MAX_SEGS; dev->gro_max_size = GRO_LEGACY_MAX_SIZE; + dev->gso_ipv4_max_size = GSO_LEGACY_MAX_SIZE; + dev->gro_ipv4_max_size = GRO_LEGACY_MAX_SIZE; dev->tso_max_size = TSO_LEGACY_MAX_SIZE; dev->tso_max_segs = TSO_MAX_SEGS; dev->upper_level = 1; @@ -10632,6 +12046,9 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name, #ifdef CONFIG_NET_SCHED hash_init(dev->qdisc_hash); #endif + + mutex_init(&dev->lock); + dev->priv_flags = IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM; setup(dev); @@ -10649,8 +12066,22 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name, dev->real_num_rx_queues = rxqs; if (netif_alloc_rx_queues(dev)) goto free_all; + dev->ethtool = kzalloc(sizeof(*dev->ethtool), GFP_KERNEL_ACCOUNT); + if (!dev->ethtool) + goto free_all; - strcpy(dev->name, name); + dev->cfg = kzalloc(sizeof(*dev->cfg), GFP_KERNEL_ACCOUNT); + if (!dev->cfg) + goto free_all; + dev->cfg_pending = dev->cfg; + + dev->num_napi_configs = maxqs; + napi_config_sz = array_size(maxqs, sizeof(*dev->napi_config)); + dev->napi_config = kvzalloc(napi_config_sz, GFP_KERNEL_ACCOUNT); + if (!dev->napi_config) + goto free_all; + + strscpy(dev->name, name); dev->name_assign_type = name_assign_type; dev->group = INIT_NETDEV_GROUP; if (!dev->ethtool_ops) @@ -10669,11 +12100,27 @@ free_pcpu: free_percpu(dev->pcpu_refcnt); free_dev: #endif - netdev_freemem(dev); + kvfree(dev); return NULL; } EXPORT_SYMBOL(alloc_netdev_mqs); +static void netdev_napi_exit(struct net_device *dev) +{ + if (!list_empty(&dev->napi_list)) { + struct napi_struct *p, *n; + + netdev_lock(dev); + list_for_each_entry_safe(p, n, &dev->napi_list, dev_list) + __netif_napi_del_locked(p); + netdev_unlock(dev); + + synchronize_net(); + } + + kvfree(dev->napi_config); +} + /** * free_netdev - free network device * @dev: device @@ -10685,8 +12132,6 @@ EXPORT_SYMBOL(alloc_netdev_mqs); */ void free_netdev(struct net_device *dev) { - struct napi_struct *p, *n; - might_sleep(); /* When called immediately after register_netdevice() failed the unwind @@ -10699,6 +12144,9 @@ void free_netdev(struct net_device *dev) return; } + WARN_ON(dev->cfg != dev->cfg_pending); + kfree(dev->cfg); + kfree(dev->ethtool); netif_free_tx_queues(dev); netif_free_rx_queues(dev); @@ -10707,8 +12155,9 @@ void free_netdev(struct net_device *dev) /* Flush device addresses */ dev_addr_flush(dev); - list_for_each_entry_safe(p, n, &dev->napi_list, dev_list) - netif_napi_del(p); + netdev_napi_exit(dev); + + netif_del_cpu_rmap(dev); ref_tracker_dir_exit(&dev->refcnt_tracker); #ifdef CONFIG_PCPU_DEV_REFCNT @@ -10720,14 +12169,19 @@ void free_netdev(struct net_device *dev) free_percpu(dev->xdp_bulkq); dev->xdp_bulkq = NULL; + netdev_free_phy_link_topology(dev); + + mutex_destroy(&dev->lock); + /* Compatibility with error handling in drivers */ - if (dev->reg_state == NETREG_UNINITIALIZED) { - netdev_freemem(dev); + if (dev->reg_state == NETREG_UNINITIALIZED || + dev->reg_state == NETREG_DUMMY) { + kvfree(dev); return; } BUG_ON(dev->reg_state != NETREG_UNREGISTERED); - dev->reg_state = NETREG_RELEASED; + WRITE_ONCE(dev->reg_state, NETREG_RELEASED); /* will free via device release */ put_device(&dev->dev); @@ -10735,6 +12189,19 @@ 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); +} +EXPORT_SYMBOL_GPL(alloc_netdev_dummy); + +/** * synchronize_net - Synchronize with packet receive processing * * Wait for packets currently being received to be done. @@ -10743,13 +12210,28 @@ EXPORT_SYMBOL(free_netdev); void synchronize_net(void) { might_sleep(); - if (rtnl_is_locked()) + if (from_cleanup_net() || rtnl_is_locked()) synchronize_rcu_expedited(); else synchronize_rcu(); } EXPORT_SYMBOL(synchronize_net); +static void netdev_rss_contexts_free(struct net_device *dev) +{ + struct ethtool_rxfh_context *ctx; + unsigned long context; + + mutex_lock(&dev->ethtool->rss_lock); + xa_for_each(&dev->ethtool->rss_ctx, context, ctx) { + xa_erase(&dev->ethtool->rss_ctx, context); + dev->ethtool_ops->remove_rxfh_context(dev, ctx, context, NULL); + kfree(ctx); + } + xa_destroy(&dev->ethtool->rss_ctx); + mutex_unlock(&dev->ethtool->rss_lock); +} + /** * unregister_netdevice_queue - remove device from the kernel * @dev: device @@ -10778,11 +12260,54 @@ void unregister_netdevice_queue(struct net_device *dev, struct list_head *head) } EXPORT_SYMBOL(unregister_netdevice_queue); +static void dev_memory_provider_uninstall(struct net_device *dev) +{ + unsigned int i; + + for (i = 0; i < dev->real_num_rx_queues; i++) { + struct netdev_rx_queue *rxq = &dev->_rx[i]; + struct pp_memory_provider_params *p = &rxq->mp_params; + + if (p->mp_ops && p->mp_ops->uninstall) + p->mp_ops->uninstall(rxq->mp_params.mp_priv, rxq); + } +} + +/* devices must be UP and netdev_lock()'d */ +static void netif_close_many_and_unlock(struct list_head *close_head) +{ + struct net_device *dev, *tmp; + + netif_close_many(close_head, false); + + /* ... now unlock them */ + list_for_each_entry_safe(dev, tmp, close_head, close_list) { + netdev_unlock(dev); + list_del_init(&dev->close_list); + } +} + +static void netif_close_many_and_unlock_cond(struct list_head *close_head) +{ +#ifdef CONFIG_LOCKDEP + /* We can only track up to MAX_LOCK_DEPTH locks per task. + * + * Reserve half the available slots for additional locks possibly + * taken by notifiers and (soft)irqs. + */ + unsigned int limit = MAX_LOCK_DEPTH / 2; + + if (lockdep_depth(current) > limit) + netif_close_many_and_unlock(close_head); +#endif +} + void unregister_netdevice_many_notify(struct list_head *head, u32 portid, const struct nlmsghdr *nlh) { struct net_device *dev, *tmp; LIST_HEAD(close_head); + int cnt = 0; BUG_ON(dev_boot_phase); ASSERT_RTNL(); @@ -10807,17 +12332,30 @@ void unregister_netdevice_many_notify(struct list_head *head, BUG_ON(dev->reg_state != NETREG_REGISTERED); } - /* If device is running, close it first. */ - list_for_each_entry(dev, head, unreg_list) - list_add_tail(&dev->close_list, &close_head); - dev_close_many(&close_head, true); + /* If device is running, close it first. Start with ops locked... */ + list_for_each_entry(dev, head, unreg_list) { + if (!(dev->flags & IFF_UP)) + continue; + if (netdev_need_ops_lock(dev)) { + list_add_tail(&dev->close_list, &close_head); + netdev_lock(dev); + } + netif_close_many_and_unlock_cond(&close_head); + } + netif_close_many_and_unlock(&close_head); + /* ... now go over the rest. */ + list_for_each_entry(dev, head, unreg_list) { + if (!netdev_need_ops_lock(dev)) + list_add_tail(&dev->close_list, &close_head); + } + netif_close_many(&close_head, true); list_for_each_entry(dev, head, unreg_list) { /* And unlink it from device chain. */ - write_lock(&dev_base_lock); - unlist_netdevice(dev, false); - dev->reg_state = NETREG_UNREGISTERING; - write_unlock(&dev_base_lock); + unlist_netdevice(dev); + netdev_lock(dev); + WRITE_ONCE(dev->reg_state, NETREG_UNREGISTERING); + netdev_unlock(dev); } flush_all_backlogs(); @@ -10827,9 +12365,13 @@ void unregister_netdevice_many_notify(struct list_head *head, struct sk_buff *skb = NULL; /* Shutdown queueing discipline. */ + netdev_lock_ops(dev); dev_shutdown(dev); - + dev_tcx_uninstall(dev); dev_xdp_uninstall(dev); + dev_memory_provider_uninstall(dev); + netdev_unlock_ops(dev); + bpf_dev_bound_netdev_unregister(dev); netdev_offload_xstats_disable_all(dev); @@ -10838,11 +12380,10 @@ void unregister_netdevice_many_notify(struct list_head *head, */ call_netdevice_notifiers(NETDEV_UNREGISTER, dev); - if (!dev->rtnl_link_ops || - dev->rtnl_link_state == RTNL_LINK_INITIALIZED) + if (!(dev->rtnl_link_ops && dev->rtnl_link_initializing)) skb = rtmsg_ifinfo_build_skb(RTM_DELLINK, dev, ~0U, 0, GFP_KERNEL, NULL, 0, - portid, nlmsg_seq(nlh)); + portid, nlh); /* * Flush the unicast and multicast chains @@ -10853,11 +12394,17 @@ void unregister_netdevice_many_notify(struct list_head *head, netdev_name_node_alt_flush(dev); netdev_name_node_free(dev->name_node); + netdev_rss_contexts_free(dev); + call_netdevice_notifiers(NETDEV_PRE_UNINIT, dev); if (dev->netdev_ops->ndo_uninit) dev->netdev_ops->ndo_uninit(dev); + mutex_destroy(&dev->ethtool->rss_lock); + + net_shaper_flush_netdev(dev); + if (skb) rtmsg_ifinfo_send(skb, dev, GFP_KERNEL, portid, nlh); @@ -10878,7 +12425,9 @@ void unregister_netdevice_many_notify(struct list_head *head, list_for_each_entry(dev, head, unreg_list) { netdev_put(dev, &dev->dev_registered_tracker); net_set_todo(dev); + cnt++; } + atomic_add(cnt, &dev_unreg_count); list_del(head); } @@ -10888,7 +12437,7 @@ void unregister_netdevice_many_notify(struct list_head *head, * @head: list of devices * * Note: As most callers use a stack allocated list_head, - * we force a list_del() to make sure stack wont be corrupted later. + * we force a list_del() to make sure stack won't be corrupted later. */ void unregister_netdevice_many(struct list_head *head) { @@ -10909,44 +12458,35 @@ EXPORT_SYMBOL(unregister_netdevice_many); */ void unregister_netdev(struct net_device *dev) { - rtnl_lock(); + rtnl_net_dev_lock(dev); unregister_netdevice(dev); - rtnl_unlock(); + rtnl_net_dev_unlock(dev); } EXPORT_SYMBOL(unregister_netdev); -/** - * __dev_change_net_namespace - move device to different nethost namespace - * @dev: device - * @net: network namespace - * @pat: If not NULL name pattern to try if the current device name - * is already taken in the destination network namespace. - * @new_ifindex: If not zero, specifies device index in the target - * namespace. - * - * This function shuts down a device interface and moves it - * to a new network namespace. On success 0 is returned, on - * a failure a netagive errno code is returned. - * - * Callers must hold the rtnl semaphore. - */ - int __dev_change_net_namespace(struct net_device *dev, struct net *net, - const char *pat, int new_ifindex) + const char *pat, int new_ifindex, + struct netlink_ext_ack *extack) { + struct netdev_name_node *name_node; struct net *net_old = dev_net(dev); + char new_name[IFNAMSIZ] = {}; int err, new_nsid; ASSERT_RTNL(); /* Don't allow namespace local devices to be moved. */ err = -EINVAL; - if (dev->features & NETIF_F_NETNS_LOCAL) + if (dev->netns_immutable) { + NL_SET_ERR_MSG(extack, "The interface netns is immutable"); goto out; + } - /* Ensure the device has been registrered */ - if (dev->reg_state != NETREG_REGISTERED) + /* Ensure the device has been registered */ + if (dev->reg_state != NETREG_REGISTERED) { + NL_SET_ERR_MSG(extack, "The interface isn't registered"); goto out; + } /* Get out if there is nothing todo */ err = 0; @@ -10959,32 +12499,73 @@ int __dev_change_net_namespace(struct net_device *dev, struct net *net, err = -EEXIST; if (netdev_name_in_use(net, dev->name)) { /* We get here if we can't use the current device name */ - if (!pat) + if (!pat) { + NL_SET_ERR_MSG(extack, + "An interface with the same name exists in the target netns"); goto out; - err = dev_get_valid_name(net, dev, pat); - if (err < 0) + } + err = dev_prep_valid_name(net, dev, pat, new_name, EEXIST); + if (err < 0) { + NL_SET_ERR_MSG_FMT(extack, + "Unable to use '%s' for the new interface name in the target netns", + pat); + goto out; + } + } + /* Check that none of the altnames conflicts. */ + err = -EEXIST; + netdev_for_each_altname(dev, name_node) { + if (netdev_name_in_use(net, name_node->name)) { + NL_SET_ERR_MSG_FMT(extack, + "An interface with the altname %s exists in the target netns", + name_node->name); goto out; + } } /* Check that new_ifindex isn't used yet. */ - err = -EBUSY; - if (new_ifindex && __dev_get_by_index(net, new_ifindex)) - goto out; + if (new_ifindex) { + err = dev_index_reserve(net, new_ifindex); + if (err < 0) { + NL_SET_ERR_MSG_FMT(extack, + "The ifindex %d is not available in the target netns", + new_ifindex); + goto out; + } + } else { + /* If there is an ifindex conflict assign a new one */ + err = dev_index_reserve(net, dev->ifindex); + if (err == -EBUSY) + err = dev_index_reserve(net, 0); + if (err < 0) { + NL_SET_ERR_MSG(extack, + "Unable to allocate a new ifindex in the target netns"); + goto out; + } + new_ifindex = err; + } /* * And now a mini version of register_netdevice unregister_netdevice. */ + netdev_lock_ops(dev); /* If device is running close it first. */ - dev_close(dev); - + netif_close(dev); /* And unlink it from device chain */ - unlist_netdevice(dev, true); + unlist_netdevice(dev); + + if (!netdev_need_ops_lock(dev)) + netdev_lock(dev); + dev->moving_ns = true; + netdev_unlock(dev); synchronize_net(); /* Shutdown queueing discipline. */ + netdev_lock_ops(dev); dev_shutdown(dev); + netdev_unlock_ops(dev); /* Notify protocols, that we are about to destroy * this device. They should clean all the things. @@ -10997,13 +12578,6 @@ int __dev_change_net_namespace(struct net_device *dev, struct net *net, rcu_barrier(); new_nsid = peernet2id_alloc(dev_net(dev), net, GFP_KERNEL); - /* If there is an ifindex conflict assign a new one */ - if (!new_ifindex) { - if (__dev_get_by_index(net, dev->ifindex)) - new_ifindex = dev_new_index(net); - else - new_ifindex = dev->ifindex; - } rtmsg_ifinfo_newnet(RTM_DELLINK, dev, ~0U, GFP_KERNEL, &new_nsid, new_ifindex); @@ -11022,28 +12596,44 @@ int __dev_change_net_namespace(struct net_device *dev, struct net *net, move_netdevice_notifiers_dev_net(dev, net); /* Actually switch the network namespace */ + netdev_lock(dev); dev_net_set(dev, net); + netdev_unlock(dev); dev->ifindex = new_ifindex; - /* Send a netdev-add uevent to the new namespace */ - kobject_uevent(&dev->dev.kobj, KOBJ_ADD); - netdev_adjacent_add_links(dev); + 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); err = device_rename(&dev->dev, dev->name); + dev_set_uevent_suppress(&dev->dev, 0); WARN_ON(err); + /* Send a netdev-add uevent to the new namespace */ + kobject_uevent(&dev->dev.kobj, KOBJ_ADD); + netdev_adjacent_add_links(dev); + /* Adapt owner in case owning user namespace of target network * namespace is different from the original one. */ err = netdev_change_owner(dev, net_old, net); WARN_ON(err); + netdev_lock(dev); + dev->moving_ns = false; + if (!netdev_need_ops_lock(dev)) + netdev_unlock(dev); + /* Add the device back in the hashes */ list_netdevice(dev); - /* Notify protocols, that a new device appeared. */ call_netdevice_notifiers(NETDEV_REGISTER, dev); + netdev_unlock_ops(dev); /* * Prevent userspace races by waiting until the network @@ -11056,7 +12646,6 @@ int __dev_change_net_namespace(struct net_device *dev, struct net *net, out: return err; } -EXPORT_SYMBOL_GPL(__dev_change_net_namespace); static int dev_cpu_dead(unsigned int oldcpu) { @@ -11096,7 +12685,7 @@ 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); } @@ -11104,21 +12693,23 @@ 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; @@ -11152,6 +12743,94 @@ netdev_features_t netdev_increment_features(netdev_features_t all, } EXPORT_SYMBOL(netdev_increment_features); +/** + * netdev_compute_master_upper_features - compute feature from lowers + * @dev: the upper device + * @update_header: whether to update upper device's header_len/headroom/tailroom + * + * Recompute the upper device's feature based on all lower devices. + */ +void netdev_compute_master_upper_features(struct net_device *dev, bool update_header) +{ + unsigned int dst_release_flag = IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM; + netdev_features_t gso_partial_features = MASTER_UPPER_DEV_GSO_PARTIAL_FEATURES; + netdev_features_t xfrm_features = MASTER_UPPER_DEV_XFRM_FEATURES; + netdev_features_t mpls_features = MASTER_UPPER_DEV_MPLS_FEATURES; + netdev_features_t vlan_features = MASTER_UPPER_DEV_VLAN_FEATURES; + netdev_features_t enc_features = MASTER_UPPER_DEV_ENC_FEATURES; + unsigned short max_header_len = ETH_HLEN; + unsigned int tso_max_size = TSO_MAX_SIZE; + unsigned short max_headroom = 0; + unsigned short max_tailroom = 0; + u16 tso_max_segs = TSO_MAX_SEGS; + struct net_device *lower_dev; + struct list_head *iter; + + mpls_features = netdev_base_features(mpls_features); + vlan_features = netdev_base_features(vlan_features); + enc_features = netdev_base_features(enc_features); + + netdev_for_each_lower_dev(dev, lower_dev, iter) { + gso_partial_features = netdev_increment_features(gso_partial_features, + lower_dev->gso_partial_features, + MASTER_UPPER_DEV_GSO_PARTIAL_FEATURES); + + vlan_features = netdev_increment_features(vlan_features, + lower_dev->vlan_features, + MASTER_UPPER_DEV_VLAN_FEATURES); + + enc_features = netdev_increment_features(enc_features, + lower_dev->hw_enc_features, + MASTER_UPPER_DEV_ENC_FEATURES); + + if (IS_ENABLED(CONFIG_XFRM_OFFLOAD)) + xfrm_features = netdev_increment_features(xfrm_features, + lower_dev->hw_enc_features, + MASTER_UPPER_DEV_XFRM_FEATURES); + + mpls_features = netdev_increment_features(mpls_features, + lower_dev->mpls_features, + MASTER_UPPER_DEV_MPLS_FEATURES); + + dst_release_flag &= lower_dev->priv_flags; + + if (update_header) { + max_header_len = max(max_header_len, lower_dev->hard_header_len); + max_headroom = max(max_headroom, lower_dev->needed_headroom); + max_tailroom = max(max_tailroom, lower_dev->needed_tailroom); + } + + tso_max_size = min(tso_max_size, lower_dev->tso_max_size); + tso_max_segs = min(tso_max_segs, lower_dev->tso_max_segs); + } + + dev->gso_partial_features = gso_partial_features; + dev->vlan_features = vlan_features; + dev->hw_enc_features = enc_features | NETIF_F_GSO_ENCAP_ALL | + NETIF_F_HW_VLAN_CTAG_TX | + NETIF_F_HW_VLAN_STAG_TX; + if (IS_ENABLED(CONFIG_XFRM_OFFLOAD)) + dev->hw_enc_features |= xfrm_features; + dev->mpls_features = mpls_features; + + dev->priv_flags &= ~IFF_XMIT_DST_RELEASE; + if ((dev->priv_flags & IFF_XMIT_DST_RELEASE_PERM) && + dst_release_flag == (IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM)) + dev->priv_flags |= IFF_XMIT_DST_RELEASE; + + if (update_header) { + dev->hard_header_len = max_header_len; + dev->needed_headroom = max_headroom; + dev->needed_tailroom = max_tailroom; + } + + netif_set_tso_max_segs(dev, tso_max_segs); + netif_set_tso_max_size(dev, tso_max_size); + + netdev_change_features(dev); +} +EXPORT_SYMBOL(netdev_compute_master_upper_features); + static struct hlist_head * __net_init netdev_create_hash(void) { int i; @@ -11169,7 +12848,7 @@ static struct hlist_head * __net_init netdev_create_hash(void) static int __net_init netdev_init(struct net *net) { BUILD_BUG_ON(GRO_HASH_BUCKETS > - 8 * sizeof_field(struct napi_struct, gro_bitmask)); + BITS_PER_BYTE * sizeof_field(struct gro_node, bitmask)); INIT_LIST_HEAD(&net->dev_base_head); @@ -11181,6 +12860,8 @@ static int __net_init netdev_init(struct net *net) if (net->dev_index_head == NULL) goto err_idx; + xa_init_flags(&net->dev_by_index, XA_FLAGS_ALLOC1); + RAW_INIT_NOTIFIER_HEAD(&net->netdev_chain); return 0; @@ -11278,6 +12959,7 @@ static void __net_exit netdev_exit(struct net *net) { kfree(net->dev_name_head); kfree(net->dev_index_head); + xa_destroy(&net->dev_by_index); if (net != &init_net) WARN_ON_ONCE(!list_empty(&net->dev_base_head)); } @@ -11289,6 +12971,7 @@ static struct pernet_operations __net_initdata netdev_net_ops = { static void __net_exit default_device_exit_net(struct net *net) { + struct netdev_name_node *name_node, *tmp; struct net_device *dev, *aux; /* * Push all migratable network devices back to the @@ -11300,7 +12983,7 @@ static void __net_exit default_device_exit_net(struct net *net) char fb_name[IFNAMSIZ]; /* Ignore unmoveable devices (i.e. loopback) */ - if (dev->features & NETIF_F_NETNS_LOCAL) + if (dev->netns_immutable) continue; /* Leave virtual devices for the generic cleanup */ @@ -11311,6 +12994,11 @@ static void __net_exit default_device_exit_net(struct net *net) snprintf(fb_name, IFNAMSIZ, "dev%d", dev->ifindex); if (netdev_name_in_use(&init_net, fb_name)) snprintf(fb_name, IFNAMSIZ, "dev%%d"); + + netdev_for_each_altname_safe(dev, name_node, tmp) + if (netdev_name_in_use(&init_net, name_node->name)) + __netdev_name_node_alt_destroy(name_node); + err = dev_change_net_namespace(dev, &init_net, fb_name); if (err) { pr_emerg("%s: failed to move %s to init_net: %d\n", @@ -11353,6 +13041,61 @@ static struct pernet_operations __net_initdata default_device_ops = { .exit_batch = default_device_exit_batch, }; +static void __init net_dev_struct_check(void) +{ + /* TX read-mostly hotpath */ + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, priv_flags_fast); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, netdev_ops); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, header_ops); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, _tx); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, real_num_tx_queues); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, gso_max_size); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, gso_ipv4_max_size); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, gso_max_segs); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, gso_partial_features); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, num_tc); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, mtu); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, needed_headroom); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, tc_to_txq); +#ifdef CONFIG_XPS + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, xps_maps); +#endif +#ifdef CONFIG_NETFILTER_EGRESS + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, nf_hooks_egress); +#endif +#ifdef CONFIG_NET_XGRESS + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_tx, tcx_egress); +#endif + CACHELINE_ASSERT_GROUP_SIZE(struct net_device, net_device_read_tx, 160); + + /* TXRX read-mostly hotpath */ + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_txrx, lstats); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_txrx, state); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_txrx, flags); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_txrx, hard_header_len); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_txrx, features); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_txrx, ip6_ptr); + CACHELINE_ASSERT_GROUP_SIZE(struct net_device, net_device_read_txrx, 46); + + /* RX read-mostly hotpath */ + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, ptype_specific); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, ifindex); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, real_num_rx_queues); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, _rx); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, gro_max_size); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, gro_ipv4_max_size); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, rx_handler); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, rx_handler_data); + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, nd_net); +#ifdef CONFIG_NETPOLL + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, npinfo); +#endif +#ifdef CONFIG_NET_XGRESS + CACHELINE_ASSERT_GROUP_MEMBER(struct net_device, net_device_read_rx, tcx_ingress); +#endif + CACHELINE_ASSERT_GROUP_SIZE(struct net_device, net_device_read_rx, 92); +} + /* * Initialize the DEV module. At boot time this walks the device list and * unhooks any devices that fail to initialise (normally hardware not @@ -11360,6 +13103,67 @@ static struct pernet_operations __net_initdata default_device_ops = { * */ +/* We allocate 256 pages for each CPU if PAGE_SHIFT is 12 */ +#define SYSTEM_PERCPU_PAGE_POOL_SIZE ((1 << 20) / PAGE_SIZE) + +static int net_page_pool_create(int cpuid) +{ +#if IS_ENABLED(CONFIG_PAGE_POOL) + struct page_pool_params page_pool_params = { + .pool_size = SYSTEM_PERCPU_PAGE_POOL_SIZE, + .flags = PP_FLAG_SYSTEM_POOL, + .nid = cpu_to_mem(cpuid), + }; + struct page_pool *pp_ptr; + int err; + + pp_ptr = page_pool_create_percpu(&page_pool_params, cpuid); + if (IS_ERR(pp_ptr)) + return -ENOMEM; + + err = xdp_reg_page_pool(pp_ptr); + if (err) { + page_pool_destroy(pp_ptr); + return err; + } + + per_cpu(system_page_pool.pool, cpuid) = pp_ptr; +#endif + 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, false); +} + +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. @@ -11370,13 +13174,14 @@ static int __init net_dev_init(void) BUG_ON(!dev_boot_phase); + net_dev_struct_check(); + if (dev_proc_init()) goto out; if (netdev_kobject_init()) goto out; - INIT_LIST_HEAD(&ptype_all); for (i = 0; i < PTYPE_HASH_SIZE; i++) INIT_LIST_HEAD(&ptype_base[i]); @@ -11387,12 +13192,13 @@ static int __init net_dev_init(void) * Initialise the packet receive queues. */ + flush_backlogs_fallback = flush_backlogs_alloc(); + if (!flush_backlogs_fallback) + goto out; + for_each_possible_cpu(i) { - struct work_struct *flush = per_cpu_ptr(&flush_works, i); struct softnet_data *sd = &per_cpu(softnet_data, i); - INIT_WORK(flush, flush_backlog); - skb_queue_head_init(&sd->input_pkt_queue); skb_queue_head_init(&sd->process_queue); #ifdef CONFIG_XFRM_OFFLOAD @@ -11405,12 +13211,22 @@ static int __init net_dev_init(void) sd->cpu = i; #endif INIT_CSD(&sd->defer_csd, trigger_rx_softirq, sd); - spin_lock_init(&sd->defer_lock); - init_gro_hash(&sd->backlog); + gro_init(&sd->backlog.gro); 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; } + net_hotdata.skb_defer_nodes = + __alloc_percpu(sizeof(struct skb_defer_node) * nr_node_ids, + __alignof__(struct skb_defer_node)); + if (!net_hotdata.skb_defer_nodes) + goto out; + if (use_backlog_threads()) + smpboot_register_percpu_thread(&backlog_threads); dev_boot_phase = 0; @@ -11436,7 +13252,25 @@ 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) { + struct page_pool *pp_ptr; + + pp_ptr = per_cpu(system_page_pool.pool, i); + if (!pp_ptr) + continue; + + xdp_unreg_page_pool(pp_ptr); + page_pool_destroy(pp_ptr); + per_cpu(system_page_pool.pool, i) = NULL; + } + } + return rc; } |