diff options
| -rw-r--r-- | drivers/net/ifb.c | 3 | ||||
| -rw-r--r-- | include/linux/netdevice.h | 4 | ||||
| -rw-r--r-- | include/linux/netfilter_netdev.h | 86 | ||||
| -rw-r--r-- | include/linux/skbuff.h | 4 | ||||
| -rw-r--r-- | include/uapi/linux/netfilter.h | 1 | ||||
| -rw-r--r-- | net/core/dev.c | 15 | ||||
| -rw-r--r-- | net/netfilter/Kconfig | 11 | ||||
| -rw-r--r-- | net/netfilter/core.c | 34 | ||||
| -rw-r--r-- | net/netfilter/nfnetlink_hook.c | 16 | ||||
| -rw-r--r-- | net/netfilter/nft_chain_filter.c | 4 |
10 files changed, 168 insertions, 10 deletions
diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c index e9258a9f3702..2c319dd27f29 100644 --- a/drivers/net/ifb.c +++ b/drivers/net/ifb.c @@ -31,6 +31,7 @@ #include <linux/init.h> #include <linux/interrupt.h> #include <linux/moduleparam.h> +#include <linux/netfilter_netdev.h> #include <net/pkt_sched.h> #include <net/net_namespace.h> @@ -75,8 +76,10 @@ static void ifb_ri_tasklet(struct tasklet_struct *t) } while ((skb = __skb_dequeue(&txp->tq)) != NULL) { + /* Skip tc and netfilter to prevent redirection loop. */ skb->redirected = 0; skb->tc_skip_classify = 1; + nf_skip_egress(skb, true); u64_stats_update_begin(&txp->tsync); txp->tx_packets++; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index d79163208dfd..e9a48068f306 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1861,6 +1861,7 @@ enum netdev_ml_priv_type { * @xps_maps: XXX: need comments on this one * @miniq_egress: clsact qdisc specific data for * egress processing + * @nf_hooks_egress: netfilter hooks executed for egress packets * @qdisc_hash: qdisc hash table * @watchdog_timeo: Represents the timeout that is used by * the watchdog (see dev_watchdog()) @@ -2161,6 +2162,9 @@ struct net_device { #ifdef CONFIG_NET_CLS_ACT struct mini_Qdisc __rcu *miniq_egress; #endif +#ifdef CONFIG_NETFILTER_EGRESS + struct nf_hook_entries __rcu *nf_hooks_egress; +#endif #ifdef CONFIG_NET_SCHED DECLARE_HASHTABLE (qdisc_hash, 4); diff --git a/include/linux/netfilter_netdev.h b/include/linux/netfilter_netdev.h index 5812b0fb0278..b71b57a83bb4 100644 --- a/include/linux/netfilter_netdev.h +++ b/include/linux/netfilter_netdev.h @@ -50,11 +50,97 @@ static inline int nf_hook_ingress(struct sk_buff *skb) } #endif /* CONFIG_NETFILTER_INGRESS */ +#ifdef CONFIG_NETFILTER_EGRESS +static inline bool nf_hook_egress_active(void) +{ +#ifdef CONFIG_JUMP_LABEL + if (!static_key_false(&nf_hooks_needed[NFPROTO_NETDEV][NF_NETDEV_EGRESS])) + return false; +#endif + return true; +} + +/** + * nf_hook_egress - classify packets before transmission + * @skb: packet to be classified + * @rc: result code which shall be returned by __dev_queue_xmit() on failure + * @dev: netdev whose egress hooks shall be applied to @skb + * + * Returns @skb on success or %NULL if the packet was consumed or filtered. + * Caller must hold rcu_read_lock. + * + * On ingress, packets are classified first by tc, then by netfilter. + * On egress, the order is reversed for symmetry. Conceptually, tc and + * netfilter can be thought of as layers, with netfilter layered above tc: + * When tc redirects a packet to another interface, netfilter is not applied + * because the packet is on the tc layer. + * + * The nf_skip_egress flag controls whether netfilter is applied on egress. + * It is updated by __netif_receive_skb_core() and __dev_queue_xmit() when the + * packet passes through tc and netfilter. Because __dev_queue_xmit() may be + * called recursively by tunnel drivers such as vxlan, the flag is reverted to + * false after sch_handle_egress(). This ensures that netfilter is applied + * both on the overlay and underlying network. + */ +static inline struct sk_buff *nf_hook_egress(struct sk_buff *skb, int *rc, + struct net_device *dev) +{ + struct nf_hook_entries *e; + struct nf_hook_state state; + int ret; + +#ifdef CONFIG_NETFILTER_SKIP_EGRESS + if (skb->nf_skip_egress) + return skb; +#endif + + e = rcu_dereference(dev->nf_hooks_egress); + if (!e) + return skb; + + nf_hook_state_init(&state, NF_NETDEV_EGRESS, + NFPROTO_NETDEV, dev, NULL, NULL, + dev_net(dev), NULL); + ret = nf_hook_slow(skb, &state, e, 0); + + if (ret == 1) { + return skb; + } else if (ret < 0) { + *rc = NET_XMIT_DROP; + return NULL; + } else { /* ret == 0 */ + *rc = NET_XMIT_SUCCESS; + return NULL; + } +} +#else /* CONFIG_NETFILTER_EGRESS */ +static inline bool nf_hook_egress_active(void) +{ + return false; +} + +static inline struct sk_buff *nf_hook_egress(struct sk_buff *skb, int *rc, + struct net_device *dev) +{ + return skb; +} +#endif /* CONFIG_NETFILTER_EGRESS */ + +static inline void nf_skip_egress(struct sk_buff *skb, bool skip) +{ +#ifdef CONFIG_NETFILTER_SKIP_EGRESS + skb->nf_skip_egress = skip; +#endif +} + static inline void nf_hook_netdev_init(struct net_device *dev) { #ifdef CONFIG_NETFILTER_INGRESS RCU_INIT_POINTER(dev->nf_hooks_ingress, NULL); #endif +#ifdef CONFIG_NETFILTER_EGRESS + RCU_INIT_POINTER(dev->nf_hooks_egress, NULL); +#endif } #endif /* _NETFILTER_NETDEV_H_ */ diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 841e2f0f5240..cb96f1e6460c 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -652,6 +652,7 @@ typedef unsigned char *sk_buff_data_t; * @tc_at_ingress: used within tc_classify to distinguish in/egress * @redirected: packet was redirected by packet classifier * @from_ingress: packet was redirected from the ingress path + * @nf_skip_egress: packet shall skip nf egress - see netfilter_netdev.h * @peeked: this packet has been seen already, so stats have been * done for it, don't do them again * @nf_trace: netfilter packet trace flag @@ -868,6 +869,9 @@ struct sk_buff { #ifdef CONFIG_NET_REDIRECT __u8 from_ingress:1; #endif +#ifdef CONFIG_NETFILTER_SKIP_EGRESS + __u8 nf_skip_egress:1; +#endif #ifdef CONFIG_TLS_DEVICE __u8 decrypted:1; #endif diff --git a/include/uapi/linux/netfilter.h b/include/uapi/linux/netfilter.h index ef9a44286e23..53411ccc69db 100644 --- a/include/uapi/linux/netfilter.h +++ b/include/uapi/linux/netfilter.h @@ -51,6 +51,7 @@ enum nf_inet_hooks { enum nf_dev_hooks { NF_NETDEV_INGRESS, + NF_NETDEV_EGRESS, NF_NETDEV_NUMHOOKS }; diff --git a/net/core/dev.c b/net/core/dev.c index e4c683029c61..09d74798b440 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3920,6 +3920,7 @@ EXPORT_SYMBOL(dev_loopback_xmit); static struct sk_buff * sch_handle_egress(struct sk_buff *skb, int *ret, struct net_device *dev) { +#ifdef CONFIG_NET_CLS_ACT struct mini_Qdisc *miniq = rcu_dereference_bh(dev->miniq_egress); struct tcf_result cl_res; @@ -3955,6 +3956,7 @@ sch_handle_egress(struct sk_buff *skb, int *ret, struct net_device *dev) default: break; } +#endif /* CONFIG_NET_CLS_ACT */ return skb; } @@ -4148,13 +4150,20 @@ static int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev) qdisc_pkt_len_init(skb); #ifdef CONFIG_NET_CLS_ACT skb->tc_at_ingress = 0; -# ifdef CONFIG_NET_EGRESS +#endif +#ifdef CONFIG_NET_EGRESS if (static_branch_unlikely(&egress_needed_key)) { + if (nf_hook_egress_active()) { + skb = nf_hook_egress(skb, &rc, dev); + if (!skb) + goto out; + } + nf_skip_egress(skb, true); skb = sch_handle_egress(skb, &rc, dev); if (!skb) goto out; + nf_skip_egress(skb, false); } -# endif #endif /* If device/qdisc don't need skb->dst, release it right now while * its hot in this cpu cache. @@ -5296,6 +5305,7 @@ skip_taps: if (static_branch_unlikely(&ingress_needed_key)) { bool another = false; + nf_skip_egress(skb, true); skb = sch_handle_ingress(skb, &pt_prev, &ret, orig_dev, &another); if (another) @@ -5303,6 +5313,7 @@ skip_taps: if (!skb) goto out; + nf_skip_egress(skb, false); if (nf_ingress(skb, &pt_prev, &ret, orig_dev) < 0) goto out; } diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig index 54395266339d..49c9fae9c62c 100644 --- a/net/netfilter/Kconfig +++ b/net/netfilter/Kconfig @@ -10,6 +10,17 @@ config NETFILTER_INGRESS This allows you to classify packets from ingress using the Netfilter infrastructure. +config NETFILTER_EGRESS + bool "Netfilter egress support" + default y + select NET_EGRESS + help + This allows you to classify packets before transmission using the + Netfilter infrastructure. + +config NETFILTER_SKIP_EGRESS + def_bool NETFILTER_EGRESS && (NET_CLS_ACT || IFB) + config NETFILTER_NETLINK tristate diff --git a/net/netfilter/core.c b/net/netfilter/core.c index 63d032191e62..3a32a813fcde 100644 --- a/net/netfilter/core.c +++ b/net/netfilter/core.c @@ -317,6 +317,12 @@ nf_hook_entry_head(struct net *net, int pf, unsigned int hooknum, return &dev->nf_hooks_ingress; } #endif +#ifdef CONFIG_NETFILTER_EGRESS + if (hooknum == NF_NETDEV_EGRESS) { + if (dev && dev_net(dev) == net) + return &dev->nf_hooks_egress; + } +#endif WARN_ON_ONCE(1); return NULL; } @@ -344,6 +350,11 @@ static inline bool nf_ingress_hook(const struct nf_hook_ops *reg, int pf) return false; } +static inline bool nf_egress_hook(const struct nf_hook_ops *reg, int pf) +{ + return pf == NFPROTO_NETDEV && reg->hooknum == NF_NETDEV_EGRESS; +} + static void nf_static_key_inc(const struct nf_hook_ops *reg, int pf) { #ifdef CONFIG_JUMP_LABEL @@ -383,9 +394,18 @@ static int __nf_register_net_hook(struct net *net, int pf, switch (pf) { case NFPROTO_NETDEV: - err = nf_ingress_check(net, reg, NF_NETDEV_INGRESS); - if (err < 0) - return err; +#ifndef CONFIG_NETFILTER_INGRESS + if (reg->hooknum == NF_NETDEV_INGRESS) + return -EOPNOTSUPP; +#endif +#ifndef CONFIG_NETFILTER_EGRESS + if (reg->hooknum == NF_NETDEV_EGRESS) + return -EOPNOTSUPP; +#endif + if ((reg->hooknum != NF_NETDEV_INGRESS && + reg->hooknum != NF_NETDEV_EGRESS) || + !reg->dev || dev_net(reg->dev) != net) + return -EINVAL; break; case NFPROTO_INET: if (reg->hooknum != NF_INET_INGRESS) @@ -418,6 +438,10 @@ static int __nf_register_net_hook(struct net *net, int pf, if (nf_ingress_hook(reg, pf)) net_inc_ingress_queue(); #endif +#ifdef CONFIG_NETFILTER_EGRESS + if (nf_egress_hook(reg, pf)) + net_inc_egress_queue(); +#endif nf_static_key_inc(reg, pf); BUG_ON(p == new_hooks); @@ -475,6 +499,10 @@ static void __nf_unregister_net_hook(struct net *net, int pf, if (nf_ingress_hook(reg, pf)) net_dec_ingress_queue(); #endif +#ifdef CONFIG_NETFILTER_EGRESS + if (nf_egress_hook(reg, pf)) + net_dec_egress_queue(); +#endif nf_static_key_dec(reg, pf); } else { WARN_ONCE(1, "hook not found, pf %d num %d", pf, reg->hooknum); diff --git a/net/netfilter/nfnetlink_hook.c b/net/netfilter/nfnetlink_hook.c index f554e2ea32ee..d5c719c9e36c 100644 --- a/net/netfilter/nfnetlink_hook.c +++ b/net/netfilter/nfnetlink_hook.c @@ -185,7 +185,7 @@ static const struct nf_hook_entries * nfnl_hook_entries_head(u8 pf, unsigned int hook, struct net *net, const char *dev) { const struct nf_hook_entries *hook_head = NULL; -#ifdef CONFIG_NETFILTER_INGRESS +#if defined(CONFIG_NETFILTER_INGRESS) || defined(CONFIG_NETFILTER_EGRESS) struct net_device *netdev; #endif @@ -221,9 +221,9 @@ nfnl_hook_entries_head(u8 pf, unsigned int hook, struct net *net, const char *de hook_head = rcu_dereference(net->nf.hooks_decnet[hook]); break; #endif -#ifdef CONFIG_NETFILTER_INGRESS +#if defined(CONFIG_NETFILTER_INGRESS) || defined(CONFIG_NETFILTER_EGRESS) case NFPROTO_NETDEV: - if (hook != NF_NETDEV_INGRESS) + if (hook >= NF_NETDEV_NUMHOOKS) return ERR_PTR(-EOPNOTSUPP); if (!dev) @@ -233,7 +233,15 @@ nfnl_hook_entries_head(u8 pf, unsigned int hook, struct net *net, const char *de if (!netdev) return ERR_PTR(-ENODEV); - return rcu_dereference(netdev->nf_hooks_ingress); +#ifdef CONFIG_NETFILTER_INGRESS + if (hook == NF_NETDEV_INGRESS) + return rcu_dereference(netdev->nf_hooks_ingress); +#endif +#ifdef CONFIG_NETFILTER_EGRESS + if (hook == NF_NETDEV_EGRESS) + return rcu_dereference(netdev->nf_hooks_egress); +#endif + fallthrough; #endif default: return ERR_PTR(-EPROTONOSUPPORT); diff --git a/net/netfilter/nft_chain_filter.c b/net/netfilter/nft_chain_filter.c index 5b02408a920b..680fe557686e 100644 --- a/net/netfilter/nft_chain_filter.c +++ b/net/netfilter/nft_chain_filter.c @@ -310,9 +310,11 @@ static const struct nft_chain_type nft_chain_filter_netdev = { .name = "filter", .type = NFT_CHAIN_T_DEFAULT, .family = NFPROTO_NETDEV, - .hook_mask = (1 << NF_NETDEV_INGRESS), + .hook_mask = (1 << NF_NETDEV_INGRESS) | + (1 << NF_NETDEV_EGRESS), .hooks = { [NF_NETDEV_INGRESS] = nft_do_chain_netdev, + [NF_NETDEV_EGRESS] = nft_do_chain_netdev, }, }; |