From 814abfabef3ceed390c10d06a0cc69a86454b6cf Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Mon, 17 Jul 2017 09:27:07 -0700 Subject: xdp: add bpf_redirect helper function This adds support for a bpf_redirect helper function to the XDP infrastructure. For now this only supports redirecting to the egress path of a port. In order to support drivers handling a xdp_buff natively this patches uses a new ndo operation ndo_xdp_xmit() that takes pushes a xdp_buff to the specified device. If the program specifies either (a) an unknown device or (b) a device that does not support the operation a BPF warning is thrown and the XDP_ABORTED error code is returned. Signed-off-by: John Fastabend Acked-by: Daniel Borkmann Acked-by: Jesper Dangaard Brouer Signed-off-by: David S. Miller --- include/linux/filter.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'include/linux/filter.h') diff --git a/include/linux/filter.h b/include/linux/filter.h index bfef1e5734f8..64cae7a08148 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -711,7 +711,11 @@ bool bpf_helper_changes_pkt_data(void *func); struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off, const struct bpf_insn *patch, u32 len); + +int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp); + void bpf_warn_invalid_xdp_action(u32 act); +void bpf_warn_invalid_xdp_redirect(u32 ifindex); #ifdef CONFIG_BPF_JIT extern int bpf_jit_enable; -- cgit v1.2.3 From 6103aa96ec077c976e851e0b89cc2446cb76573d Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Mon, 17 Jul 2017 09:27:50 -0700 Subject: net: implement XDP_REDIRECT for xdp generic Add support for redirect to xdp generic creating a fall back for devices that do not yet have support and allowing test infrastructure using veth pairs to be built. Signed-off-by: John Fastabend Tested-by: Andy Gospodarek Acked-by: Daniel Borkmann Acked-by: Jesper Dangaard Brouer Signed-off-by: David S. Miller --- include/linux/filter.h | 1 + net/core/dev.c | 22 ++++++++++++++++++++-- net/core/filter.c | 26 ++++++++++++++++++++++++++ 3 files changed, 47 insertions(+), 2 deletions(-) (limited to 'include/linux/filter.h') diff --git a/include/linux/filter.h b/include/linux/filter.h index 64cae7a08148..10df7daf5ec6 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -712,6 +712,7 @@ bool bpf_helper_changes_pkt_data(void *func); struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off, const struct bpf_insn *patch, u32 len); +int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb); int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp); void bpf_warn_invalid_xdp_action(u32 act); diff --git a/net/core/dev.c b/net/core/dev.c index a1ed7b41b3e8..9f3f4083ada5 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3902,6 +3902,7 @@ static u32 netif_receive_generic_xdp(struct sk_buff *skb, __skb_push(skb, -off); switch (act) { + case XDP_REDIRECT: case XDP_TX: __skb_push(skb, mac_len); /* fall through */ @@ -3956,14 +3957,27 @@ static int do_xdp_generic(struct sk_buff *skb) if (xdp_prog) { u32 act = netif_receive_generic_xdp(skb, xdp_prog); + int err; if (act != XDP_PASS) { - if (act == XDP_TX) + switch (act) { + case XDP_REDIRECT: + err = xdp_do_generic_redirect(skb->dev, skb); + if (err) + goto out_redir; + /* fallthru to submit skb */ + case XDP_TX: generic_xdp_tx(skb, xdp_prog); + break; + } return XDP_DROP; } } return XDP_PASS; +out_redir: + trace_xdp_exception(skb->dev, xdp_prog, XDP_REDIRECT); + kfree_skb(skb); + return XDP_DROP; } static int netif_rx_internal(struct sk_buff *skb) @@ -3977,8 +3991,12 @@ static int netif_rx_internal(struct sk_buff *skb) if (static_key_false(&generic_xdp_needed)) { int ret = do_xdp_generic(skb); + /* Consider XDP consuming the packet a success from + * the netdev point of view we do not want to count + * this as an error. + */ if (ret != XDP_PASS) - return NET_RX_DROP; + return NET_RX_SUCCESS; } #ifdef CONFIG_RPS diff --git a/net/core/filter.c b/net/core/filter.c index d606a66d1040..eeb713461c25 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -2437,6 +2437,32 @@ int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp) } EXPORT_SYMBOL_GPL(xdp_do_redirect); +int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb) +{ + struct redirect_info *ri = this_cpu_ptr(&redirect_info); + unsigned int len; + + dev = dev_get_by_index_rcu(dev_net(dev), ri->ifindex); + ri->ifindex = 0; + if (unlikely(!dev)) { + bpf_warn_invalid_xdp_redirect(ri->ifindex); + goto err; + } + + if (unlikely(!(dev->flags & IFF_UP))) + goto err; + + len = dev->mtu + dev->hard_header_len + VLAN_HLEN; + if (skb->len > len) + goto err; + + skb->dev = dev; + return 0; +err: + return -EINVAL; +} +EXPORT_SYMBOL_GPL(xdp_do_generic_redirect); + BPF_CALL_2(bpf_xdp_redirect, u32, ifindex, u64, flags) { struct redirect_info *ri = this_cpu_ptr(&redirect_info); -- cgit v1.2.3 From 5acaee0a8964c9bab7775ab8bedcd1f66a2a1011 Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Mon, 17 Jul 2017 09:28:35 -0700 Subject: xdp: add trace event for xdp redirect This adds a trace event for xdp redirect which may help when debugging XDP programs that use redirect bpf commands. Signed-off-by: John Fastabend Acked-by: Daniel Borkmann Acked-by: Jesper Dangaard Brouer Signed-off-by: David S. Miller --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +- include/linux/filter.h | 4 +++- include/trace/events/xdp.h | 31 ++++++++++++++++++++++++++- net/core/filter.c | 13 +++++++---- 4 files changed, 43 insertions(+), 7 deletions(-) (limited to 'include/linux/filter.h') diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index 3db04736a048..38f7ff97d636 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -2232,7 +2232,7 @@ static struct sk_buff *ixgbe_run_xdp(struct ixgbe_adapter *adapter, result = ixgbe_xmit_xdp_ring(adapter, xdp); break; case XDP_REDIRECT: - err = xdp_do_redirect(adapter->netdev, xdp); + err = xdp_do_redirect(adapter->netdev, xdp, xdp_prog); if (!err) result = IXGBE_XDP_TX; else diff --git a/include/linux/filter.h b/include/linux/filter.h index 10df7daf5ec6..ce8211fa91c7 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -713,7 +713,9 @@ struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off, const struct bpf_insn *patch, u32 len); int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb); -int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp); +int xdp_do_redirect(struct net_device *dev, + struct xdp_buff *xdp, + struct bpf_prog *prog); void bpf_warn_invalid_xdp_action(u32 act); void bpf_warn_invalid_xdp_redirect(u32 ifindex); diff --git a/include/trace/events/xdp.h b/include/trace/events/xdp.h index 1b61357d3f57..7b1eb7b4be41 100644 --- a/include/trace/events/xdp.h +++ b/include/trace/events/xdp.h @@ -12,7 +12,8 @@ FN(ABORTED) \ FN(DROP) \ FN(PASS) \ - FN(TX) + FN(TX) \ + FN(REDIRECT) #define __XDP_ACT_TP_FN(x) \ TRACE_DEFINE_ENUM(XDP_##x); @@ -48,6 +49,34 @@ TRACE_EVENT(xdp_exception, __print_symbolic(__entry->act, __XDP_ACT_SYM_TAB)) ); +TRACE_EVENT(xdp_redirect, + + TP_PROTO(const struct net_device *from, + const struct net_device *to, + const struct bpf_prog *xdp, u32 act), + + TP_ARGS(from, to, xdp, act), + + TP_STRUCT__entry( + __string(name_from, from->name) + __string(name_to, to->name) + __array(u8, prog_tag, 8) + __field(u32, act) + ), + + TP_fast_assign( + BUILD_BUG_ON(sizeof(__entry->prog_tag) != sizeof(xdp->tag)); + memcpy(__entry->prog_tag, xdp->tag, sizeof(xdp->tag)); + __assign_str(name_from, from->name); + __assign_str(name_to, to->name); + __entry->act = act; + ), + + TP_printk("prog=%s from=%s to=%s action=%s", + __print_hex_str(__entry->prog_tag, 8), + __get_str(name_from), __get_str(name_to), + __print_symbolic(__entry->act, __XDP_ACT_SYM_TAB)) +); #endif /* _TRACE_XDP_H */ #include diff --git a/net/core/filter.c b/net/core/filter.c index eeb713461c25..e30d38b27f21 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -55,6 +55,7 @@ #include #include #include +#include /** * sk_filter_trim_cap - run a packet through a socket filter @@ -2422,18 +2423,22 @@ static int __bpf_tx_xdp(struct net_device *dev, struct xdp_buff *xdp) return -EOPNOTSUPP; } -int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp) +int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp, + struct bpf_prog *xdp_prog) { struct redirect_info *ri = this_cpu_ptr(&redirect_info); + struct net_device *fwd; - dev = dev_get_by_index_rcu(dev_net(dev), ri->ifindex); + fwd = dev_get_by_index_rcu(dev_net(dev), ri->ifindex); ri->ifindex = 0; - if (unlikely(!dev)) { + if (unlikely(!fwd)) { bpf_warn_invalid_xdp_redirect(ri->ifindex); return -EINVAL; } - return __bpf_tx_xdp(dev, xdp); + trace_xdp_redirect(dev, fwd, xdp_prog, XDP_REDIRECT); + + return __bpf_tx_xdp(fwd, xdp); } EXPORT_SYMBOL_GPL(xdp_do_redirect); -- cgit v1.2.3 From 11393cc9b9be2a1f61559e6fb9c27bc8fa20b1ff Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Mon, 17 Jul 2017 09:29:40 -0700 Subject: xdp: Add batching support to redirect map For performance reasons we want to avoid updating the tail pointer in the driver tx ring as much as possible. To accomplish this we add batching support to the redirect path in XDP. This adds another ndo op "xdp_flush" that is used to inform the driver that it should bump the tail pointer on the TX ring. Signed-off-by: John Fastabend Signed-off-by: Jesper Dangaard Brouer Acked-by: Daniel Borkmann Signed-off-by: David S. Miller --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 28 ++++++++- include/linux/bpf.h | 2 + include/linux/filter.h | 7 +++ include/linux/netdevice.h | 5 +- kernel/bpf/devmap.c | 84 ++++++++++++++++++++++++++- net/core/filter.c | 55 ++++++++++++++---- 6 files changed, 166 insertions(+), 15 deletions(-) (limited to 'include/linux/filter.h') diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index 38f7ff97d636..0f867dcda65f 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -2415,6 +2415,8 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector, */ wmb(); writel(ring->next_to_use, ring->tail); + + xdp_do_flush_map(); } u64_stats_update_begin(&rx_ring->syncp); @@ -5817,6 +5819,9 @@ void ixgbe_down(struct ixgbe_adapter *adapter) usleep_range(10000, 20000); + /* synchronize_sched() needed for pending XDP buffers to drain */ + if (adapter->xdp_ring[0]) + synchronize_sched(); netif_tx_stop_all_queues(netdev); /* call carrier off first to avoid false dev_watchdog timeouts */ @@ -9850,15 +9855,31 @@ static int ixgbe_xdp_xmit(struct net_device *dev, struct xdp_buff *xdp) if (err != IXGBE_XDP_TX) return -ENOMEM; + return 0; +} + +static void ixgbe_xdp_flush(struct net_device *dev) +{ + struct ixgbe_adapter *adapter = netdev_priv(dev); + struct ixgbe_ring *ring; + + /* Its possible the device went down between xdp xmit and flush so + * we need to ensure device is still up. + */ + if (unlikely(test_bit(__IXGBE_DOWN, &adapter->state))) + return; + + ring = adapter->xdp_prog ? adapter->xdp_ring[smp_processor_id()] : NULL; + if (unlikely(!ring)) + return; + /* Force memory writes to complete before letting h/w know there * are new descriptors to fetch. */ wmb(); - - ring = adapter->xdp_ring[smp_processor_id()]; writel(ring->next_to_use, ring->tail); - return 0; + return; } static const struct net_device_ops ixgbe_netdev_ops = { @@ -9908,6 +9929,7 @@ static const struct net_device_ops ixgbe_netdev_ops = { .ndo_features_check = ixgbe_features_check, .ndo_xdp = ixgbe_xdp, .ndo_xdp_xmit = ixgbe_xdp_xmit, + .ndo_xdp_flush = ixgbe_xdp_flush, }; /** diff --git a/include/linux/bpf.h b/include/linux/bpf.h index d0d3281ac678..6850a760dc94 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -381,5 +381,7 @@ u64 bpf_user_rnd_u32(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5); /* Map specifics */ struct net_device *__dev_map_lookup_elem(struct bpf_map *map, u32 key); +void __dev_map_insert_ctx(struct bpf_map *map, u32 index); +void __dev_map_flush(struct bpf_map *map); #endif /* _LINUX_BPF_H */ diff --git a/include/linux/filter.h b/include/linux/filter.h index ce8211fa91c7..3323ee91c172 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -712,10 +712,17 @@ bool bpf_helper_changes_pkt_data(void *func); struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off, const struct bpf_insn *patch, u32 len); +/* The pair of xdp_do_redirect and xdp_do_flush_map MUST be called in the + * same cpu context. Further for best results no more than a single map + * for the do_redirect/do_flush pair should be used. This limitation is + * because we only track one map and force a flush when the map changes. + * This does not appear to be a real limiation for existing software. + */ int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb); int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp, struct bpf_prog *prog); +void xdp_do_flush_map(void); void bpf_warn_invalid_xdp_action(u32 act); void bpf_warn_invalid_xdp_redirect(u32 ifindex); diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 77f5376005e6..03b104908235 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1142,7 +1142,9 @@ struct xfrmdev_ops { * int (*ndo_xdp_xmit)(struct net_device *dev, struct xdp_buff *xdp); * This function is used to submit a XDP packet for transmit on a * netdevice. - * + * void (*ndo_xdp_flush)(struct net_device *dev); + * This function is used to inform the driver to flush a paticular + * xpd tx queue. Must be called on same CPU as xdp_xmit. */ struct net_device_ops { int (*ndo_init)(struct net_device *dev); @@ -1329,6 +1331,7 @@ struct net_device_ops { struct netdev_xdp *xdp); int (*ndo_xdp_xmit)(struct net_device *dev, struct xdp_buff *xdp); + void (*ndo_xdp_flush)(struct net_device *dev); }; /** diff --git a/kernel/bpf/devmap.c b/kernel/bpf/devmap.c index 36dc13deb2e1..b2ef04a1c86a 100644 --- a/kernel/bpf/devmap.c +++ b/kernel/bpf/devmap.c @@ -53,6 +53,7 @@ struct bpf_dtab_netdev { struct bpf_dtab { struct bpf_map map; struct bpf_dtab_netdev **netdev_map; + unsigned long int __percpu *flush_needed; }; static struct bpf_map *dev_map_alloc(union bpf_attr *attr) @@ -87,6 +88,7 @@ static struct bpf_map *dev_map_alloc(union bpf_attr *attr) /* make sure page count doesn't overflow */ cost = (u64) dtab->map.max_entries * sizeof(struct bpf_dtab_netdev *); + cost += BITS_TO_LONGS(attr->max_entries) * sizeof(unsigned long); if (cost >= U32_MAX - PAGE_SIZE) goto free_dtab; @@ -97,6 +99,14 @@ static struct bpf_map *dev_map_alloc(union bpf_attr *attr) if (err) goto free_dtab; + /* A per cpu bitfield with a bit per possible net device */ + dtab->flush_needed = __alloc_percpu( + BITS_TO_LONGS(attr->max_entries) * + sizeof(unsigned long), + __alignof__(unsigned long)); + if (!dtab->flush_needed) + goto free_dtab; + dtab->netdev_map = bpf_map_area_alloc(dtab->map.max_entries * sizeof(struct bpf_dtab_netdev *)); if (!dtab->netdev_map) @@ -105,6 +115,7 @@ static struct bpf_map *dev_map_alloc(union bpf_attr *attr) return &dtab->map; free_dtab: + free_percpu(dtab->flush_needed); kfree(dtab); return ERR_PTR(err); } @@ -112,7 +123,7 @@ free_dtab: static void dev_map_free(struct bpf_map *map) { struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map); - int i; + int i, cpu; /* At this point bpf_prog->aux->refcnt == 0 and this map->refcnt == 0, * so the programs (can be more than one that used this map) were @@ -123,6 +134,18 @@ static void dev_map_free(struct bpf_map *map) */ synchronize_rcu(); + /* To ensure all pending flush operations have completed wait for flush + * bitmap to indicate all flush_needed bits to be zero on _all_ cpus. + * Because the above synchronize_rcu() ensures the map is disconnected + * from the program we can assume no new bits will be set. + */ + for_each_online_cpu(cpu) { + unsigned long *bitmap = per_cpu_ptr(dtab->flush_needed, cpu); + + while (!bitmap_empty(bitmap, dtab->map.max_entries)) + cpu_relax(); + } + for (i = 0; i < dtab->map.max_entries; i++) { struct bpf_dtab_netdev *dev; @@ -137,6 +160,7 @@ static void dev_map_free(struct bpf_map *map) /* At this point bpf program is detached and all pending operations * _must_ be complete */ + free_percpu(dtab->flush_needed); bpf_map_area_free(dtab->netdev_map); kfree(dtab); } @@ -159,6 +183,14 @@ static int dev_map_get_next_key(struct bpf_map *map, void *key, void *next_key) return 0; } +void __dev_map_insert_ctx(struct bpf_map *map, u32 key) +{ + struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map); + unsigned long *bitmap = this_cpu_ptr(dtab->flush_needed); + + __set_bit(key, bitmap); +} + struct net_device *__dev_map_lookup_elem(struct bpf_map *map, u32 key) { struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map); @@ -171,6 +203,39 @@ struct net_device *__dev_map_lookup_elem(struct bpf_map *map, u32 key) return dev ? dev->dev : NULL; } +/* __dev_map_flush is called from xdp_do_flush_map() which _must_ be signaled + * from the driver before returning from its napi->poll() routine. The poll() + * routine is called either from busy_poll context or net_rx_action signaled + * from NET_RX_SOFTIRQ. Either way the poll routine must complete before the + * net device can be torn down. On devmap tear down we ensure the ctx bitmap + * is zeroed before completing to ensure all flush operations have completed. + */ +void __dev_map_flush(struct bpf_map *map) +{ + struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map); + unsigned long *bitmap = this_cpu_ptr(dtab->flush_needed); + u32 bit; + + for_each_set_bit(bit, bitmap, map->max_entries) { + struct bpf_dtab_netdev *dev = READ_ONCE(dtab->netdev_map[bit]); + struct net_device *netdev; + + /* This is possible if the dev entry is removed by user space + * between xdp redirect and flush op. + */ + if (unlikely(!dev)) + continue; + + netdev = dev->dev; + + __clear_bit(bit, bitmap); + if (unlikely(!netdev || !netdev->netdev_ops->ndo_xdp_flush)) + continue; + + netdev->netdev_ops->ndo_xdp_flush(netdev); + } +} + /* rcu_read_lock (from syscall and BPF contexts) ensures that if a delete and/or * update happens in parallel here a dev_put wont happen until after reading the * ifindex. @@ -188,11 +253,28 @@ static void *dev_map_lookup_elem(struct bpf_map *map, void *key) return dev ? &dev->dev->ifindex : NULL; } +static void dev_map_flush_old(struct bpf_dtab_netdev *old_dev) +{ + if (old_dev->dev->netdev_ops->ndo_xdp_flush) { + struct net_device *fl = old_dev->dev; + unsigned long *bitmap; + int cpu; + + for_each_online_cpu(cpu) { + bitmap = per_cpu_ptr(old_dev->dtab->flush_needed, cpu); + __clear_bit(old_dev->key, bitmap); + + fl->netdev_ops->ndo_xdp_flush(old_dev->dev); + } + } +} + static void __dev_map_entry_free(struct rcu_head *rcu) { struct bpf_dtab_netdev *old_dev; old_dev = container_of(rcu, struct bpf_dtab_netdev, rcu); + dev_map_flush_old(old_dev); dev_put(old_dev->dev); kfree(old_dev); } diff --git a/net/core/filter.c b/net/core/filter.c index e93a558324b5..e23aa6fa1119 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -1780,6 +1780,7 @@ struct redirect_info { u32 ifindex; u32 flags; struct bpf_map *map; + struct bpf_map *map_to_flush; }; static DEFINE_PER_CPU(struct redirect_info, redirect_info); @@ -2438,34 +2439,68 @@ static const struct bpf_func_proto bpf_xdp_adjust_head_proto = { .arg2_type = ARG_ANYTHING, }; -static int __bpf_tx_xdp(struct net_device *dev, struct xdp_buff *xdp) +static int __bpf_tx_xdp(struct net_device *dev, + struct bpf_map *map, + struct xdp_buff *xdp, + u32 index) { - if (dev->netdev_ops->ndo_xdp_xmit) { - dev->netdev_ops->ndo_xdp_xmit(dev, xdp); - return 0; + int err; + + if (!dev->netdev_ops->ndo_xdp_xmit) { + bpf_warn_invalid_xdp_redirect(dev->ifindex); + return -EOPNOTSUPP; } - bpf_warn_invalid_xdp_redirect(dev->ifindex); - return -EOPNOTSUPP; + + err = dev->netdev_ops->ndo_xdp_xmit(dev, xdp); + if (err) + return err; + + if (map) + __dev_map_insert_ctx(map, index); + else + dev->netdev_ops->ndo_xdp_flush(dev); + + return err; } +void xdp_do_flush_map(void) +{ + struct redirect_info *ri = this_cpu_ptr(&redirect_info); + struct bpf_map *map = ri->map_to_flush; + + ri->map = NULL; + ri->map_to_flush = NULL; + + if (map) + __dev_map_flush(map); +} +EXPORT_SYMBOL_GPL(xdp_do_flush_map); + int xdp_do_redirect_map(struct net_device *dev, struct xdp_buff *xdp, struct bpf_prog *xdp_prog) { struct redirect_info *ri = this_cpu_ptr(&redirect_info); struct bpf_map *map = ri->map; + u32 index = ri->ifindex; struct net_device *fwd; int err = -EINVAL; ri->ifindex = 0; ri->map = NULL; - fwd = __dev_map_lookup_elem(map, ri->ifindex); + fwd = __dev_map_lookup_elem(map, index); if (!fwd) goto out; - trace_xdp_redirect(dev, fwd, xdp_prog, XDP_REDIRECT); - err = __bpf_tx_xdp(fwd, xdp); + if (ri->map_to_flush && (ri->map_to_flush != map)) + xdp_do_flush_map(); + + err = __bpf_tx_xdp(fwd, map, xdp, index); + if (likely(!err)) + ri->map_to_flush = map; + out: + trace_xdp_redirect(dev, fwd, xdp_prog, XDP_REDIRECT); return err; } @@ -2488,7 +2523,7 @@ int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp, trace_xdp_redirect(dev, fwd, xdp_prog, XDP_REDIRECT); - return __bpf_tx_xdp(fwd, xdp); + return __bpf_tx_xdp(fwd, NULL, xdp, 0); } EXPORT_SYMBOL_GPL(xdp_do_redirect); -- cgit v1.2.3 From 2ddf71e23cc246e95af72a6deed67b4a50a7b81c Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Mon, 17 Jul 2017 09:30:02 -0700 Subject: net: add notifier hooks for devmap bpf map The BPF map devmap holds a refcnt on the net_device structure when it is in the map. We need to do this to ensure on driver unload we don't lose a dev reference. However, its not very convenient to have to manually unload the map when destroying a net device so add notifier handlers to do the cleanup automatically. But this creates a race between update/destroy BPF syscall and programs and the unregister netdev hook. Unfortunately, the best I could come up with is either to live with requiring manual removal of net devices from the map before removing the net device OR to add a mutex in devmap to ensure the map is not modified while we are removing a device. The fallout also requires that BPF programs no longer update/delete the map from the BPF program side because the mutex may sleep and this can not be done from inside an rcu critical section. This is not a real problem though because I have not come up with any use cases where this is actually useful in practice. If/when we come up with a compelling user for this we may need to revisit this. Signed-off-by: John Fastabend Acked-by: Daniel Borkmann Acked-by: Jesper Dangaard Brouer Signed-off-by: David S. Miller --- include/linux/filter.h | 2 +- kernel/bpf/devmap.c | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++ kernel/bpf/verifier.c | 2 +- 3 files changed, 75 insertions(+), 2 deletions(-) (limited to 'include/linux/filter.h') diff --git a/include/linux/filter.h b/include/linux/filter.h index 3323ee91c172..d19ed3c15e1e 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -716,7 +716,7 @@ struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off, * same cpu context. Further for best results no more than a single map * for the do_redirect/do_flush pair should be used. This limitation is * because we only track one map and force a flush when the map changes. - * This does not appear to be a real limiation for existing software. + * This does not appear to be a real limitation for existing software. */ int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb); int xdp_do_redirect(struct net_device *dev, diff --git a/kernel/bpf/devmap.c b/kernel/bpf/devmap.c index b2ef04a1c86a..899364d097f5 100644 --- a/kernel/bpf/devmap.c +++ b/kernel/bpf/devmap.c @@ -34,6 +34,17 @@ * netdev_map consistent in this case. From the devmap side BPF programs * calling into these operations are the same as multiple user space threads * making system calls. + * + * Finally, any of the above may race with a netdev_unregister notifier. The + * unregister notifier must search for net devices in the map structure that + * contain a reference to the net device and remove them. This is a two step + * process (a) dereference the bpf_dtab_netdev object in netdev_map and (b) + * check to see if the ifindex is the same as the net_device being removed. + * Unfortunately, the xchg() operations do not protect against this. To avoid + * potentially removing incorrect objects the dev_map_list_mutex protects + * conflicting netdev unregister and BPF syscall operations. Updates and + * deletes from a BPF program (done in rcu critical section) are blocked + * because of this mutex. */ #include #include @@ -54,8 +65,12 @@ struct bpf_dtab { struct bpf_map map; struct bpf_dtab_netdev **netdev_map; unsigned long int __percpu *flush_needed; + struct list_head list; }; +static DEFINE_MUTEX(dev_map_list_mutex); +static LIST_HEAD(dev_map_list); + static struct bpf_map *dev_map_alloc(union bpf_attr *attr) { struct bpf_dtab *dtab; @@ -112,6 +127,9 @@ static struct bpf_map *dev_map_alloc(union bpf_attr *attr) if (!dtab->netdev_map) goto free_dtab; + mutex_lock(&dev_map_list_mutex); + list_add_tail(&dtab->list, &dev_map_list); + mutex_unlock(&dev_map_list_mutex); return &dtab->map; free_dtab: @@ -146,6 +164,11 @@ static void dev_map_free(struct bpf_map *map) cpu_relax(); } + /* Although we should no longer have datapath or bpf syscall operations + * at this point we we can still race with netdev notifier, hence the + * lock. + */ + mutex_lock(&dev_map_list_mutex); for (i = 0; i < dtab->map.max_entries; i++) { struct bpf_dtab_netdev *dev; @@ -160,6 +183,8 @@ static void dev_map_free(struct bpf_map *map) /* At this point bpf program is detached and all pending operations * _must_ be complete */ + list_del(&dtab->list); + mutex_unlock(&dev_map_list_mutex); free_percpu(dtab->flush_needed); bpf_map_area_free(dtab->netdev_map); kfree(dtab); @@ -296,9 +321,11 @@ static int dev_map_delete_elem(struct bpf_map *map, void *key) * the driver tear down ensures all soft irqs are complete before * removing the net device in the case of dev_put equals zero. */ + mutex_lock(&dev_map_list_mutex); old_dev = xchg(&dtab->netdev_map[k], NULL); if (old_dev) call_rcu(&old_dev->rcu, __dev_map_entry_free); + mutex_unlock(&dev_map_list_mutex); return 0; } @@ -341,9 +368,11 @@ static int dev_map_update_elem(struct bpf_map *map, void *key, void *value, * Remembering the driver side flush operation will happen before the * net device is removed. */ + mutex_lock(&dev_map_list_mutex); old_dev = xchg(&dtab->netdev_map[i], dev); if (old_dev) call_rcu(&old_dev->rcu, __dev_map_entry_free); + mutex_unlock(&dev_map_list_mutex); return 0; } @@ -356,3 +385,47 @@ const struct bpf_map_ops dev_map_ops = { .map_update_elem = dev_map_update_elem, .map_delete_elem = dev_map_delete_elem, }; + +static int dev_map_notification(struct notifier_block *notifier, + ulong event, void *ptr) +{ + struct net_device *netdev = netdev_notifier_info_to_dev(ptr); + struct bpf_dtab *dtab; + int i; + + switch (event) { + case NETDEV_UNREGISTER: + mutex_lock(&dev_map_list_mutex); + list_for_each_entry(dtab, &dev_map_list, list) { + for (i = 0; i < dtab->map.max_entries; i++) { + struct bpf_dtab_netdev *dev; + + dev = dtab->netdev_map[i]; + if (!dev || + dev->dev->ifindex != netdev->ifindex) + continue; + dev = xchg(&dtab->netdev_map[i], NULL); + if (dev) + call_rcu(&dev->rcu, + __dev_map_entry_free); + } + } + mutex_unlock(&dev_map_list_mutex); + break; + default: + break; + } + return NOTIFY_OK; +} + +static struct notifier_block dev_map_notifier = { + .notifier_call = dev_map_notification, +}; + +static int __init dev_map_init(void) +{ + register_netdevice_notifier(&dev_map_notifier); + return 0; +} + +subsys_initcall(dev_map_init); diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index df05d65f0c87..ebe9b38ff522 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -1281,7 +1281,7 @@ static int check_map_func_compatibility(struct bpf_map *map, int func_id) * for now. */ case BPF_MAP_TYPE_DEVMAP: - if (func_id == BPF_FUNC_map_lookup_elem) + if (func_id != BPF_FUNC_redirect_map) goto error; break; case BPF_MAP_TYPE_ARRAY_OF_MAPS: -- cgit v1.2.3 From 174a79ff9515f400b9a6115643dafd62a635b7e6 Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Tue, 15 Aug 2017 22:32:47 -0700 Subject: bpf: sockmap with sk redirect support Recently we added a new map type called dev map used to forward XDP packets between ports (6093ec2dc313). This patches introduces a similar notion for sockets. A sockmap allows users to add participating sockets to a map. When sockets are added to the map enough context is stored with the map entry to use the entry with a new helper bpf_sk_redirect_map(map, key, flags) This helper (analogous to bpf_redirect_map in XDP) is given the map and an entry in the map. When called from a sockmap program, discussed below, the skb will be sent on the socket using skb_send_sock(). With the above we need a bpf program to call the helper from that will then implement the send logic. The initial site implemented in this series is the recv_sock hook. For this to work we implemented a map attach command to add attributes to a map. In sockmap we add two programs a parse program and a verdict program. The parse program uses strparser to build messages and pass them to the verdict program. The parse programs use the normal strparser semantics. The verdict program is of type SK_SKB. The verdict program returns a verdict SK_DROP, or SK_REDIRECT for now. Additional actions may be added later. When SK_REDIRECT is returned, expected when bpf program uses bpf_sk_redirect_map(), the sockmap logic will consult per cpu variables set by the helper routine and pull the sock entry out of the sock map. This pattern follows the existing redirect logic in cls and xdp programs. This gives the flow, recv_sock -> str_parser (parse_prog) -> verdict_prog -> skb_send_sock \ -> kfree_skb As an example use case a message based load balancer may use specific logic in the verdict program to select the sock to send on. Sample programs are provided in future patches that hopefully illustrate the user interfaces. Also selftests are in follow-on patches. Signed-off-by: John Fastabend Signed-off-by: David S. Miller --- include/linux/bpf.h | 7 +- include/linux/bpf_types.h | 1 + include/linux/filter.h | 2 + include/uapi/linux/bpf.h | 33 +- kernel/bpf/Makefile | 2 +- kernel/bpf/sockmap.c | 792 ++++++++++++++++++++++++++++++++++++++++++++++ kernel/bpf/syscall.c | 51 ++- kernel/bpf/verifier.c | 14 + net/core/filter.c | 43 +++ 9 files changed, 940 insertions(+), 5 deletions(-) create mode 100644 kernel/bpf/sockmap.c (limited to 'include/linux/filter.h') diff --git a/include/linux/bpf.h b/include/linux/bpf.h index d6e1de8ce0fc..a4145e9c74b5 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -16,6 +16,7 @@ #include struct perf_event; +struct bpf_prog; struct bpf_map; /* map is generic key/value storage optionally accesible by eBPF programs */ @@ -37,6 +38,8 @@ struct bpf_map_ops { void (*map_fd_put_ptr)(void *ptr); u32 (*map_gen_lookup)(struct bpf_map *map, struct bpf_insn *insn_buf); u32 (*map_fd_sys_lookup_elem)(void *ptr); + int (*map_attach)(struct bpf_map *map, + struct bpf_prog *p1, struct bpf_prog *p2); }; struct bpf_map { @@ -138,8 +141,6 @@ enum bpf_reg_type { PTR_TO_PACKET_END, /* skb->data + headlen */ }; -struct bpf_prog; - /* The information passed from prog-specific *_is_valid_access * back to the verifier. */ @@ -312,6 +313,7 @@ int bpf_check(struct bpf_prog **fp, union bpf_attr *attr); /* Map specifics */ struct net_device *__dev_map_lookup_elem(struct bpf_map *map, u32 key); +struct sock *__sock_map_lookup_elem(struct bpf_map *map, u32 key); void __dev_map_insert_ctx(struct bpf_map *map, u32 index); void __dev_map_flush(struct bpf_map *map); @@ -391,6 +393,7 @@ extern const struct bpf_func_proto bpf_get_current_comm_proto; extern const struct bpf_func_proto bpf_skb_vlan_push_proto; extern const struct bpf_func_proto bpf_skb_vlan_pop_proto; extern const struct bpf_func_proto bpf_get_stackid_proto; +extern const struct bpf_func_proto bpf_sock_map_update_proto; /* Shared helpers among cBPF and eBPF. */ void bpf_user_rnd_init_once(void); diff --git a/include/linux/bpf_types.h b/include/linux/bpf_types.h index 4b72db30dacf..fa805074d168 100644 --- a/include/linux/bpf_types.h +++ b/include/linux/bpf_types.h @@ -38,4 +38,5 @@ BPF_MAP_TYPE(BPF_MAP_TYPE_ARRAY_OF_MAPS, array_of_maps_map_ops) BPF_MAP_TYPE(BPF_MAP_TYPE_HASH_OF_MAPS, htab_of_maps_map_ops) #ifdef CONFIG_NET BPF_MAP_TYPE(BPF_MAP_TYPE_DEVMAP, dev_map_ops) +BPF_MAP_TYPE(BPF_MAP_TYPE_SOCKMAP, sock_map_ops) #endif diff --git a/include/linux/filter.h b/include/linux/filter.h index d19ed3c15e1e..7015116331af 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -727,6 +727,8 @@ void xdp_do_flush_map(void); void bpf_warn_invalid_xdp_action(u32 act); void bpf_warn_invalid_xdp_redirect(u32 ifindex); +struct sock *do_sk_redirect_map(void); + #ifdef CONFIG_BPF_JIT extern int bpf_jit_enable; extern int bpf_jit_harden; diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h index 2e796e384aeb..7f774769e3f5 100644 --- a/include/uapi/linux/bpf.h +++ b/include/uapi/linux/bpf.h @@ -110,6 +110,7 @@ enum bpf_map_type { BPF_MAP_TYPE_ARRAY_OF_MAPS, BPF_MAP_TYPE_HASH_OF_MAPS, BPF_MAP_TYPE_DEVMAP, + BPF_MAP_TYPE_SOCKMAP, }; enum bpf_prog_type { @@ -135,11 +136,15 @@ enum bpf_attach_type { BPF_CGROUP_INET_EGRESS, BPF_CGROUP_INET_SOCK_CREATE, BPF_CGROUP_SOCK_OPS, + BPF_CGROUP_SMAP_INGRESS, __MAX_BPF_ATTACH_TYPE }; #define MAX_BPF_ATTACH_TYPE __MAX_BPF_ATTACH_TYPE +/* If BPF_SOCKMAP_STRPARSER is used sockmap will use strparser on receive */ +#define BPF_SOCKMAP_STRPARSER (1U << 0) + /* If BPF_F_ALLOW_OVERRIDE flag is used in BPF_PROG_ATTACH command * to the given target_fd cgroup the descendent cgroup will be able to * override effective bpf program that was inherited from this cgroup @@ -211,6 +216,7 @@ union bpf_attr { __u32 attach_bpf_fd; /* eBPF program to attach */ __u32 attach_type; __u32 attach_flags; + __u32 attach_bpf_fd2; }; struct { /* anonymous struct used by BPF_PROG_TEST_RUN command */ @@ -557,6 +563,23 @@ union bpf_attr { * @mode: operation mode (enum bpf_adj_room_mode) * @flags: reserved for future use * Return: 0 on success or negative error code + * + * int bpf_sk_redirect_map(map, key, flags) + * Redirect skb to a sock in map using key as a lookup key for the + * sock in map. + * @map: pointer to sockmap + * @key: key to lookup sock in map + * @flags: reserved for future use + * Return: SK_REDIRECT + * + * int bpf_sock_map_update(skops, map, key, flags, map_flags) + * @skops: pointer to bpf_sock_ops + * @map: pointer to sockmap to update + * @key: key to insert/update sock in map + * @flags: same flags as map update elem + * @map_flags: sock map specific flags + * bit 1: Enable strparser + * other bits: reserved */ #define __BPF_FUNC_MAPPER(FN) \ FN(unspec), \ @@ -610,7 +633,9 @@ union bpf_attr { FN(set_hash), \ FN(setsockopt), \ FN(skb_adjust_room), \ - FN(redirect_map), + FN(redirect_map), \ + FN(sk_redirect_map), \ + FN(sock_map_update), \ /* integer value in 'imm' field of BPF_CALL instruction selects which helper * function eBPF program intends to call @@ -747,6 +772,12 @@ struct xdp_md { __u32 data_end; }; +enum sk_action { + SK_ABORTED = 0, + SK_DROP, + SK_REDIRECT, +}; + #define BPF_TAG_SIZE 8 struct bpf_prog_info { diff --git a/kernel/bpf/Makefile b/kernel/bpf/Makefile index 2f0bcda40e90..aa24287db888 100644 --- a/kernel/bpf/Makefile +++ b/kernel/bpf/Makefile @@ -3,7 +3,7 @@ obj-y := core.o obj-$(CONFIG_BPF_SYSCALL) += syscall.o verifier.o inode.o helpers.o tnum.o obj-$(CONFIG_BPF_SYSCALL) += hashtab.o arraymap.o percpu_freelist.o bpf_lru_list.o lpm_trie.o map_in_map.o ifeq ($(CONFIG_NET),y) -obj-$(CONFIG_BPF_SYSCALL) += devmap.o +obj-$(CONFIG_BPF_SYSCALL) += devmap.o sockmap.o endif ifeq ($(CONFIG_PERF_EVENTS),y) obj-$(CONFIG_BPF_SYSCALL) += stackmap.o diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c new file mode 100644 index 000000000000..792f0addfafa --- /dev/null +++ b/kernel/bpf/sockmap.c @@ -0,0 +1,792 @@ +/* Copyright (c) 2017 Covalent IO, Inc. http://covalent.io + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of version 2 of the GNU General Public + * License as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ + +/* A BPF sock_map is used to store sock objects. This is primarly used + * for doing socket redirect with BPF helper routines. + * + * A sock map may have two BPF programs attached to it, a program used + * to parse packets and a program to provide a verdict and redirect + * decision on the packet. If no BPF parse program is provided it is + * assumed that every skb is a "message" (skb->len). Otherwise the + * parse program is attached to strparser and used to build messages + * that may span multiple skbs. The verdict program will either select + * a socket to send/receive the skb on or provide the drop code indicating + * the skb should be dropped. More actions may be added later as needed. + * The default program will drop packets. + * + * For reference this program is similar to devmap used in XDP context + * reviewing these together may be useful. For an example please review + * ./samples/bpf/sockmap/. + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +struct bpf_stab { + struct bpf_map map; + struct sock **sock_map; + struct bpf_prog *bpf_parse; + struct bpf_prog *bpf_verdict; + refcount_t refcnt; +}; + +enum smap_psock_state { + SMAP_TX_RUNNING, +}; + +struct smap_psock { + struct rcu_head rcu; + + /* datapath variables */ + struct sk_buff_head rxqueue; + bool strp_enabled; + + /* datapath error path cache across tx work invocations */ + int save_rem; + int save_off; + struct sk_buff *save_skb; + + struct strparser strp; + struct bpf_prog *bpf_parse; + struct bpf_prog *bpf_verdict; + struct bpf_stab *stab; + + /* Back reference used when sock callback trigger sockmap operations */ + int key; + struct sock *sock; + unsigned long state; + + struct work_struct tx_work; + struct work_struct gc_work; + + void (*save_data_ready)(struct sock *sk); + void (*save_write_space)(struct sock *sk); + void (*save_state_change)(struct sock *sk); +}; + +static inline struct smap_psock *smap_psock_sk(const struct sock *sk) +{ + return (struct smap_psock *)rcu_dereference_sk_user_data(sk); +} + +static int smap_verdict_func(struct smap_psock *psock, struct sk_buff *skb) +{ + struct bpf_prog *prog = READ_ONCE(psock->bpf_verdict); + int rc; + + if (unlikely(!prog)) + return SK_DROP; + + skb_orphan(skb); + skb->sk = psock->sock; + bpf_compute_data_end(skb); + rc = (*prog->bpf_func)(skb, prog->insnsi); + skb->sk = NULL; + + return rc; +} + +static void smap_do_verdict(struct smap_psock *psock, struct sk_buff *skb) +{ + struct sock *sock; + int rc; + + /* Because we use per cpu values to feed input from sock redirect + * in BPF program to do_sk_redirect_map() call we need to ensure we + * are not preempted. RCU read lock is not sufficient in this case + * with CONFIG_PREEMPT_RCU enabled so we must be explicit here. + */ + preempt_disable(); + rc = smap_verdict_func(psock, skb); + switch (rc) { + case SK_REDIRECT: + sock = do_sk_redirect_map(); + preempt_enable(); + if (likely(sock)) { + struct smap_psock *peer = smap_psock_sk(sock); + + if (likely(peer && + test_bit(SMAP_TX_RUNNING, &peer->state) && + sk_stream_memory_free(peer->sock))) { + peer->sock->sk_wmem_queued += skb->truesize; + sk_mem_charge(peer->sock, skb->truesize); + skb_queue_tail(&peer->rxqueue, skb); + schedule_work(&peer->tx_work); + break; + } + } + /* Fall through and free skb otherwise */ + case SK_DROP: + default: + preempt_enable(); + kfree_skb(skb); + } +} + +static void smap_report_sk_error(struct smap_psock *psock, int err) +{ + struct sock *sk = psock->sock; + + sk->sk_err = err; + sk->sk_error_report(sk); +} + +static void smap_release_sock(struct sock *sock); + +/* Called with lock_sock(sk) held */ +static void smap_state_change(struct sock *sk) +{ + struct smap_psock *psock; + struct sock *osk; + + rcu_read_lock(); + + /* Allowing transitions into an established syn_recv states allows + * for early binding sockets to a smap object before the connection + * is established. + */ + switch (sk->sk_state) { + case TCP_SYN_RECV: + case TCP_ESTABLISHED: + break; + case TCP_CLOSE_WAIT: + case TCP_CLOSING: + case TCP_LAST_ACK: + case TCP_FIN_WAIT1: + case TCP_FIN_WAIT2: + case TCP_LISTEN: + break; + case TCP_CLOSE: + /* Only release if the map entry is in fact the sock in + * question. There is a case where the operator deletes + * the sock from the map, but the TCP sock is closed before + * the psock is detached. Use cmpxchg to verify correct + * sock is removed. + */ + psock = smap_psock_sk(sk); + if (unlikely(!psock)) + break; + osk = cmpxchg(&psock->stab->sock_map[psock->key], sk, NULL); + if (osk == sk) + smap_release_sock(sk); + break; + default: + smap_report_sk_error(psock, EPIPE); + break; + } + rcu_read_unlock(); +} + +static void smap_read_sock_strparser(struct strparser *strp, + struct sk_buff *skb) +{ + struct smap_psock *psock; + + rcu_read_lock(); + psock = container_of(strp, struct smap_psock, strp); + smap_do_verdict(psock, skb); + rcu_read_unlock(); +} + +/* Called with lock held on socket */ +static void smap_data_ready(struct sock *sk) +{ + struct smap_psock *psock; + + write_lock_bh(&sk->sk_callback_lock); + psock = smap_psock_sk(sk); + if (likely(psock)) + strp_data_ready(&psock->strp); + write_unlock_bh(&sk->sk_callback_lock); +} + +static void smap_tx_work(struct work_struct *w) +{ + struct smap_psock *psock; + struct sk_buff *skb; + int rem, off, n; + + psock = container_of(w, struct smap_psock, tx_work); + + /* lock sock to avoid losing sk_socket at some point during loop */ + lock_sock(psock->sock); + if (psock->save_skb) { + skb = psock->save_skb; + rem = psock->save_rem; + off = psock->save_off; + psock->save_skb = NULL; + goto start; + } + + while ((skb = skb_dequeue(&psock->rxqueue))) { + rem = skb->len; + off = 0; +start: + do { + if (likely(psock->sock->sk_socket)) + n = skb_send_sock_locked(psock->sock, + skb, off, rem); + else + n = -EINVAL; + if (n <= 0) { + if (n == -EAGAIN) { + /* Retry when space is available */ + psock->save_skb = skb; + psock->save_rem = rem; + psock->save_off = off; + goto out; + } + /* Hard errors break pipe and stop xmit */ + smap_report_sk_error(psock, n ? -n : EPIPE); + clear_bit(SMAP_TX_RUNNING, &psock->state); + sk_mem_uncharge(psock->sock, skb->truesize); + psock->sock->sk_wmem_queued -= skb->truesize; + kfree_skb(skb); + goto out; + } + rem -= n; + off += n; + } while (rem); + sk_mem_uncharge(psock->sock, skb->truesize); + psock->sock->sk_wmem_queued -= skb->truesize; + kfree_skb(skb); + } +out: + release_sock(psock->sock); +} + +static void smap_write_space(struct sock *sk) +{ + struct smap_psock *psock; + + rcu_read_lock(); + psock = smap_psock_sk(sk); + if (likely(psock && test_bit(SMAP_TX_RUNNING, &psock->state))) + schedule_work(&psock->tx_work); + rcu_read_unlock(); +} + +static void smap_stop_sock(struct smap_psock *psock, struct sock *sk) +{ + write_lock_bh(&sk->sk_callback_lock); + if (!psock->strp_enabled) + goto out; + sk->sk_data_ready = psock->save_data_ready; + sk->sk_write_space = psock->save_write_space; + sk->sk_state_change = psock->save_state_change; + psock->save_data_ready = NULL; + psock->save_write_space = NULL; + psock->save_state_change = NULL; + strp_stop(&psock->strp); + psock->strp_enabled = false; +out: + write_unlock_bh(&sk->sk_callback_lock); +} + +static void smap_destroy_psock(struct rcu_head *rcu) +{ + struct smap_psock *psock = container_of(rcu, + struct smap_psock, rcu); + + /* Now that a grace period has passed there is no longer + * any reference to this sock in the sockmap so we can + * destroy the psock, strparser, and bpf programs. But, + * because we use workqueue sync operations we can not + * do it in rcu context + */ + schedule_work(&psock->gc_work); +} + +static void smap_release_sock(struct sock *sock) +{ + struct smap_psock *psock = smap_psock_sk(sock); + + smap_stop_sock(psock, sock); + clear_bit(SMAP_TX_RUNNING, &psock->state); + rcu_assign_sk_user_data(sock, NULL); + call_rcu_sched(&psock->rcu, smap_destroy_psock); +} + +static int smap_parse_func_strparser(struct strparser *strp, + struct sk_buff *skb) +{ + struct smap_psock *psock; + struct bpf_prog *prog; + int rc; + + rcu_read_lock(); + psock = container_of(strp, struct smap_psock, strp); + prog = READ_ONCE(psock->bpf_parse); + + if (unlikely(!prog)) { + rcu_read_unlock(); + return skb->len; + } + + /* Attach socket for bpf program to use if needed we can do this + * because strparser clones the skb before handing it to a upper + * layer, meaning skb_orphan has been called. We NULL sk on the + * way out to ensure we don't trigger a BUG_ON in skb/sk operations + * later and because we are not charging the memory of this skb to + * any socket yet. + */ + skb->sk = psock->sock; + bpf_compute_data_end(skb); + rc = (*prog->bpf_func)(skb, prog->insnsi); + skb->sk = NULL; + rcu_read_unlock(); + return rc; +} + + +static int smap_read_sock_done(struct strparser *strp, int err) +{ + return err; +} + +static int smap_init_sock(struct smap_psock *psock, + struct sock *sk) +{ + struct strp_callbacks cb; + + memset(&cb, 0, sizeof(cb)); + cb.rcv_msg = smap_read_sock_strparser; + cb.parse_msg = smap_parse_func_strparser; + cb.read_sock_done = smap_read_sock_done; + return strp_init(&psock->strp, sk, &cb); +} + +static void smap_init_progs(struct smap_psock *psock, + struct bpf_stab *stab, + struct bpf_prog *verdict, + struct bpf_prog *parse) +{ + struct bpf_prog *orig_parse, *orig_verdict; + + orig_parse = xchg(&psock->bpf_parse, parse); + orig_verdict = xchg(&psock->bpf_verdict, verdict); + + if (orig_verdict) + bpf_prog_put(orig_verdict); + if (orig_parse) + bpf_prog_put(orig_parse); +} + +static void smap_start_sock(struct smap_psock *psock, struct sock *sk) +{ + if (sk->sk_data_ready == smap_data_ready) + return; + psock->save_data_ready = sk->sk_data_ready; + psock->save_write_space = sk->sk_write_space; + psock->save_state_change = sk->sk_state_change; + sk->sk_data_ready = smap_data_ready; + sk->sk_write_space = smap_write_space; + sk->sk_state_change = smap_state_change; + psock->strp_enabled = true; +} + +static void sock_map_remove_complete(struct bpf_stab *stab) +{ + bpf_map_area_free(stab->sock_map); + kfree(stab); +} + +static void smap_gc_work(struct work_struct *w) +{ + struct smap_psock *psock; + + psock = container_of(w, struct smap_psock, gc_work); + + /* no callback lock needed because we already detached sockmap ops */ + if (psock->strp_enabled) + strp_done(&psock->strp); + + cancel_work_sync(&psock->tx_work); + __skb_queue_purge(&psock->rxqueue); + + /* At this point all strparser and xmit work must be complete */ + if (psock->bpf_parse) + bpf_prog_put(psock->bpf_parse); + if (psock->bpf_verdict) + bpf_prog_put(psock->bpf_verdict); + + if (refcount_dec_and_test(&psock->stab->refcnt)) + sock_map_remove_complete(psock->stab); + + sock_put(psock->sock); + kfree(psock); +} + +static struct smap_psock *smap_init_psock(struct sock *sock, + struct bpf_stab *stab) +{ + struct smap_psock *psock; + + psock = kzalloc(sizeof(struct smap_psock), GFP_ATOMIC | __GFP_NOWARN); + if (!psock) + return ERR_PTR(-ENOMEM); + + psock->sock = sock; + skb_queue_head_init(&psock->rxqueue); + INIT_WORK(&psock->tx_work, smap_tx_work); + INIT_WORK(&psock->gc_work, smap_gc_work); + + rcu_assign_sk_user_data(sock, psock); + sock_hold(sock); + return psock; +} + +static struct bpf_map *sock_map_alloc(union bpf_attr *attr) +{ + struct bpf_stab *stab; + int err = -EINVAL; + u64 cost; + + /* check sanity of attributes */ + if (attr->max_entries == 0 || attr->key_size != 4 || + attr->value_size != 4 || attr->map_flags) + return ERR_PTR(-EINVAL); + + if (attr->value_size > KMALLOC_MAX_SIZE) + return ERR_PTR(-E2BIG); + + stab = kzalloc(sizeof(*stab), GFP_USER); + if (!stab) + return ERR_PTR(-ENOMEM); + + /* mandatory map attributes */ + stab->map.map_type = attr->map_type; + stab->map.key_size = attr->key_size; + stab->map.value_size = attr->value_size; + stab->map.max_entries = attr->max_entries; + stab->map.map_flags = attr->map_flags; + + /* make sure page count doesn't overflow */ + cost = (u64) stab->map.max_entries * sizeof(struct sock *); + if (cost >= U32_MAX - PAGE_SIZE) + goto free_stab; + + stab->map.pages = round_up(cost, PAGE_SIZE) >> PAGE_SHIFT; + + /* if map size is larger than memlock limit, reject it early */ + err = bpf_map_precharge_memlock(stab->map.pages); + if (err) + goto free_stab; + + stab->sock_map = bpf_map_area_alloc(stab->map.max_entries * + sizeof(struct sock *)); + if (!stab->sock_map) + goto free_stab; + + refcount_set(&stab->refcnt, 1); + return &stab->map; +free_stab: + kfree(stab); + return ERR_PTR(err); +} + +static void sock_map_free(struct bpf_map *map) +{ + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + int i; + + synchronize_rcu(); + + /* At this point no update, lookup or delete operations can happen. + * However, be aware we can still get a socket state event updates, + * and data ready callabacks that reference the psock from sk_user_data + * Also psock worker threads are still in-flight. So smap_release_sock + * will only free the psock after cancel_sync on the worker threads + * and a grace period expire to ensure psock is really safe to remove. + */ + rcu_read_lock(); + for (i = 0; i < stab->map.max_entries; i++) { + struct sock *sock; + + sock = xchg(&stab->sock_map[i], NULL); + if (!sock) + continue; + + smap_release_sock(sock); + } + rcu_read_unlock(); + + if (stab->bpf_verdict) + bpf_prog_put(stab->bpf_verdict); + if (stab->bpf_parse) + bpf_prog_put(stab->bpf_parse); + + if (refcount_dec_and_test(&stab->refcnt)) + sock_map_remove_complete(stab); +} + +static int sock_map_get_next_key(struct bpf_map *map, void *key, void *next_key) +{ + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + u32 i = key ? *(u32 *)key : U32_MAX; + u32 *next = (u32 *)next_key; + + if (i >= stab->map.max_entries) { + *next = 0; + return 0; + } + + if (i == stab->map.max_entries - 1) + return -ENOENT; + + *next = i + 1; + return 0; +} + +struct sock *__sock_map_lookup_elem(struct bpf_map *map, u32 key) +{ + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + + if (key >= map->max_entries) + return NULL; + + return READ_ONCE(stab->sock_map[key]); +} + +static int sock_map_delete_elem(struct bpf_map *map, void *key) +{ + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + int k = *(u32 *)key; + struct sock *sock; + + if (k >= map->max_entries) + return -EINVAL; + + sock = xchg(&stab->sock_map[k], NULL); + if (!sock) + return -EINVAL; + + smap_release_sock(sock); + return 0; +} + +/* Locking notes: Concurrent updates, deletes, and lookups are allowed and are + * done inside rcu critical sections. This ensures on updates that the psock + * will not be released via smap_release_sock() until concurrent updates/deletes + * complete. All operations operate on sock_map using cmpxchg and xchg + * operations to ensure we do not get stale references. Any reads into the + * map must be done with READ_ONCE() because of this. + * + * A psock is destroyed via call_rcu and after any worker threads are cancelled + * and syncd so we are certain all references from the update/lookup/delete + * operations as well as references in the data path are no longer in use. + * + * A psock object holds a refcnt on the sockmap it is attached to and this is + * not decremented until after a RCU grace period and garbage collection occurs. + * This ensures the map is not free'd until psocks linked to it are removed. The + * map link is used when the independent sock events trigger map deletion. + * + * Psocks may only participate in one sockmap at a time. Users that try to + * join a single sock to multiple maps will get an error. + * + * Last, but not least, it is possible the socket is closed while running + * an update on an existing psock. This will release the psock, but again + * not until the update has completed due to rcu grace period rules. + */ +static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, + struct bpf_map *map, + void *key, u64 flags, u64 map_flags) +{ + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + struct bpf_prog *verdict, *parse; + struct smap_psock *psock = NULL; + struct sock *old_sock, *sock; + u32 i = *(u32 *)key; + bool update = false; + int err = 0; + + if (unlikely(flags > BPF_EXIST)) + return -EINVAL; + + if (unlikely(i >= stab->map.max_entries)) + return -E2BIG; + + if (unlikely(map_flags > BPF_SOCKMAP_STRPARSER)) + return -EINVAL; + + verdict = parse = NULL; + sock = READ_ONCE(stab->sock_map[i]); + + if (flags == BPF_EXIST || flags == BPF_ANY) { + if (!sock && flags == BPF_EXIST) { + return -ENOENT; + } else if (sock && sock != skops->sk) { + return -EINVAL; + } else if (sock) { + psock = smap_psock_sk(sock); + if (unlikely(!psock)) + return -EBUSY; + update = true; + } + } else if (sock && BPF_NOEXIST) { + return -EEXIST; + } + + /* reserve BPF programs early so can abort easily on failures */ + if (map_flags & BPF_SOCKMAP_STRPARSER) { + verdict = READ_ONCE(stab->bpf_verdict); + parse = READ_ONCE(stab->bpf_parse); + + if (!verdict || !parse) + return -ENOENT; + + /* bpf prog refcnt may be zero if a concurrent attach operation + * removes the program after the above READ_ONCE() but before + * we increment the refcnt. If this is the case abort with an + * error. + */ + verdict = bpf_prog_inc_not_zero(stab->bpf_verdict); + if (IS_ERR(verdict)) + return PTR_ERR(verdict); + + parse = bpf_prog_inc_not_zero(stab->bpf_parse); + if (IS_ERR(parse)) { + bpf_prog_put(verdict); + return PTR_ERR(parse); + } + } + + if (!psock) { + sock = skops->sk; + if (rcu_dereference_sk_user_data(sock)) + return -EEXIST; + psock = smap_init_psock(sock, stab); + if (IS_ERR(psock)) { + if (verdict) + bpf_prog_put(verdict); + if (parse) + bpf_prog_put(parse); + return PTR_ERR(psock); + } + psock->key = i; + psock->stab = stab; + refcount_inc(&stab->refcnt); + set_bit(SMAP_TX_RUNNING, &psock->state); + } + + if (map_flags & BPF_SOCKMAP_STRPARSER) { + write_lock_bh(&sock->sk_callback_lock); + if (psock->strp_enabled) + goto start_done; + err = smap_init_sock(psock, sock); + if (err) + goto out; + smap_init_progs(psock, stab, verdict, parse); + smap_start_sock(psock, sock); +start_done: + write_unlock_bh(&sock->sk_callback_lock); + } else if (update) { + smap_stop_sock(psock, sock); + } + + if (!update) { + old_sock = xchg(&stab->sock_map[i], skops->sk); + if (old_sock) + smap_release_sock(old_sock); + } + + return 0; +out: + write_unlock_bh(&sock->sk_callback_lock); + if (!update) + smap_release_sock(sock); + return err; +} + +static int sock_map_attach_prog(struct bpf_map *map, + struct bpf_prog *parse, + struct bpf_prog *verdict) +{ + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + struct bpf_prog *_parse, *_verdict; + + _parse = xchg(&stab->bpf_parse, parse); + _verdict = xchg(&stab->bpf_verdict, verdict); + + if (_parse) + bpf_prog_put(_parse); + if (_verdict) + bpf_prog_put(_verdict); + + return 0; +} + +static void *sock_map_lookup(struct bpf_map *map, void *key) +{ + return NULL; +} + +static int sock_map_update_elem(struct bpf_map *map, + void *key, void *value, u64 flags) +{ + struct bpf_sock_ops_kern skops; + u32 fd = *(u32 *)value; + struct socket *socket; + int err; + + socket = sockfd_lookup(fd, &err); + if (!socket) + return err; + + skops.sk = socket->sk; + if (!skops.sk) { + fput(socket->file); + return -EINVAL; + } + + err = sock_map_ctx_update_elem(&skops, map, key, + flags, BPF_SOCKMAP_STRPARSER); + fput(socket->file); + return err; +} + +const struct bpf_map_ops sock_map_ops = { + .map_alloc = sock_map_alloc, + .map_free = sock_map_free, + .map_lookup_elem = sock_map_lookup, + .map_get_next_key = sock_map_get_next_key, + .map_update_elem = sock_map_update_elem, + .map_delete_elem = sock_map_delete_elem, + .map_attach = sock_map_attach_prog, +}; + +BPF_CALL_5(bpf_sock_map_update, struct bpf_sock_ops_kern *, bpf_sock, + struct bpf_map *, map, void *, key, u64, flags, u64, map_flags) +{ + WARN_ON_ONCE(!rcu_read_lock_held()); + return sock_map_ctx_update_elem(bpf_sock, map, key, flags, map_flags); +} + +const struct bpf_func_proto bpf_sock_map_update_proto = { + .func = bpf_sock_map_update, + .gpl_only = false, + .pkt_access = true, + .ret_type = RET_INTEGER, + .arg1_type = ARG_PTR_TO_CTX, + .arg2_type = ARG_CONST_MAP_PTR, + .arg3_type = ARG_PTR_TO_MAP_KEY, + .arg4_type = ARG_ANYTHING, + .arg5_type = ARG_ANYTHING, +}; diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index 17e29f596de1..d2f2bdf71ffa 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -1087,7 +1087,50 @@ static int bpf_obj_get(const union bpf_attr *attr) #ifdef CONFIG_CGROUP_BPF -#define BPF_PROG_ATTACH_LAST_FIELD attach_flags +#define BPF_PROG_ATTACH_LAST_FIELD attach_bpf_fd2 + +static int sockmap_get_from_fd(const union bpf_attr *attr, int ptype) +{ + struct bpf_prog *prog1, *prog2; + int ufd = attr->target_fd; + struct bpf_map *map; + struct fd f; + int err; + + f = fdget(ufd); + map = __bpf_map_get(f); + if (IS_ERR(map)) + return PTR_ERR(map); + + if (!map->ops->map_attach) { + fdput(f); + return -EOPNOTSUPP; + } + + prog1 = bpf_prog_get_type(attr->attach_bpf_fd, ptype); + if (IS_ERR(prog1)) { + fdput(f); + return PTR_ERR(prog1); + } + + prog2 = bpf_prog_get_type(attr->attach_bpf_fd2, ptype); + if (IS_ERR(prog2)) { + fdput(f); + bpf_prog_put(prog1); + return PTR_ERR(prog2); + } + + err = map->ops->map_attach(map, prog1, prog2); + if (err) { + fdput(f); + bpf_prog_put(prog1); + bpf_prog_put(prog2); + return PTR_ERR(map); + } + + fdput(f); + return err; +} static int bpf_prog_attach(const union bpf_attr *attr) { @@ -1116,10 +1159,16 @@ static int bpf_prog_attach(const union bpf_attr *attr) case BPF_CGROUP_SOCK_OPS: ptype = BPF_PROG_TYPE_SOCK_OPS; break; + case BPF_CGROUP_SMAP_INGRESS: + ptype = BPF_PROG_TYPE_SK_SKB; + break; default: return -EINVAL; } + if (attr->attach_type == BPF_CGROUP_SMAP_INGRESS) + return sockmap_get_from_fd(attr, ptype); + prog = bpf_prog_get_type(attr->attach_bpf_fd, ptype); if (IS_ERR(prog)) return PTR_ERR(prog); diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index 7dd96d064be1..a71bc0996572 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -1522,6 +1522,12 @@ static int check_map_func_compatibility(struct bpf_map *map, int func_id) case BPF_MAP_TYPE_HASH_OF_MAPS: if (func_id != BPF_FUNC_map_lookup_elem) goto error; + case BPF_MAP_TYPE_SOCKMAP: + if (func_id != BPF_FUNC_sk_redirect_map && + func_id != BPF_FUNC_sock_map_update && + func_id != BPF_FUNC_map_delete_elem) + goto error; + break; default: break; } @@ -1550,6 +1556,14 @@ static int check_map_func_compatibility(struct bpf_map *map, int func_id) if (map->map_type != BPF_MAP_TYPE_DEVMAP) goto error; break; + case BPF_FUNC_sk_redirect_map: + if (map->map_type != BPF_MAP_TYPE_SOCKMAP) + goto error; + break; + case BPF_FUNC_sock_map_update: + if (map->map_type != BPF_MAP_TYPE_SOCKMAP) + goto error; + break; default: break; } diff --git a/net/core/filter.c b/net/core/filter.c index 46321033ae0e..8e136578488c 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -1858,6 +1858,45 @@ static const struct bpf_func_proto bpf_redirect_map_proto = { .arg3_type = ARG_ANYTHING, }; +BPF_CALL_3(bpf_sk_redirect_map, struct bpf_map *, map, u32, key, u64, flags) +{ + struct redirect_info *ri = this_cpu_ptr(&redirect_info); + + if (unlikely(flags)) + return SK_ABORTED; + + ri->ifindex = key; + ri->flags = flags; + ri->map = map; + + return SK_REDIRECT; +} + +struct sock *do_sk_redirect_map(void) +{ + struct redirect_info *ri = this_cpu_ptr(&redirect_info); + struct sock *sk = NULL; + + if (ri->map) { + sk = __sock_map_lookup_elem(ri->map, ri->ifindex); + + ri->ifindex = 0; + ri->map = NULL; + /* we do not clear flags for future lookup */ + } + + return sk; +} + +static const struct bpf_func_proto bpf_sk_redirect_map_proto = { + .func = bpf_sk_redirect_map, + .gpl_only = false, + .ret_type = RET_INTEGER, + .arg1_type = ARG_CONST_MAP_PTR, + .arg2_type = ARG_ANYTHING, + .arg3_type = ARG_ANYTHING, +}; + BPF_CALL_1(bpf_get_cgroup_classid, const struct sk_buff *, skb) { return task_get_classid(skb); @@ -3229,6 +3268,8 @@ static const struct bpf_func_proto * switch (func_id) { case BPF_FUNC_setsockopt: return &bpf_setsockopt_proto; + case BPF_FUNC_sock_map_update: + return &bpf_sock_map_update_proto; default: return bpf_base_func_proto(func_id); } @@ -3243,6 +3284,8 @@ static const struct bpf_func_proto *sk_skb_func_proto(enum bpf_func_id func_id) return &bpf_get_socket_cookie_proto; case BPF_FUNC_get_socket_uid: return &bpf_get_socket_uid_proto; + case BPF_FUNC_sk_redirect_map: + return &bpf_sk_redirect_map_proto; default: return bpf_base_func_proto(func_id); } -- cgit v1.2.3 From 2facaad6000f2322eb40ca379aced31c957f0a41 Mon Sep 17 00:00:00 2001 From: Jesper Dangaard Brouer Date: Thu, 24 Aug 2017 12:33:08 +0200 Subject: xdp: make generic xdp redirect use tracepoint trace_xdp_redirect If the xdp_do_generic_redirect() call fails, it trigger the trace_xdp_exception tracepoint. It seems better to use the same tracepoint trace_xdp_redirect, as the native xdp_do_redirect{,_map} does. Signed-off-by: Jesper Dangaard Brouer Acked-by: Daniel Borkmann Signed-off-by: David S. Miller --- include/linux/filter.h | 3 ++- net/core/dev.c | 4 ++-- net/core/filter.c | 35 +++++++++++++++++++++++------------ 3 files changed, 27 insertions(+), 15 deletions(-) (limited to 'include/linux/filter.h') diff --git a/include/linux/filter.h b/include/linux/filter.h index 7015116331af..d29e58fde364 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -718,7 +718,8 @@ struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off, * because we only track one map and force a flush when the map changes. * This does not appear to be a real limitation for existing software. */ -int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb); +int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb, + struct bpf_prog *prog); int xdp_do_redirect(struct net_device *dev, struct xdp_buff *xdp, struct bpf_prog *prog); diff --git a/net/core/dev.c b/net/core/dev.c index 40b28e417072..270b54754821 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3953,7 +3953,8 @@ int do_xdp_generic(struct bpf_prog *xdp_prog, struct sk_buff *skb) if (act != XDP_PASS) { switch (act) { case XDP_REDIRECT: - err = xdp_do_generic_redirect(skb->dev, skb); + err = xdp_do_generic_redirect(skb->dev, skb, + xdp_prog); if (err) goto out_redir; /* fallthru to submit skb */ @@ -3966,7 +3967,6 @@ int do_xdp_generic(struct bpf_prog *xdp_prog, struct sk_buff *skb) } return XDP_PASS; out_redir: - trace_xdp_exception(skb->dev, xdp_prog, XDP_REDIRECT); kfree_skb(skb); return XDP_DROP; } diff --git a/net/core/filter.c b/net/core/filter.c index a5e5f31b41b9..a04680331033 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -2553,26 +2553,37 @@ out: } EXPORT_SYMBOL_GPL(xdp_do_redirect); -int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb) +int xdp_do_generic_redirect(struct net_device *dev, struct sk_buff *skb, + struct bpf_prog *xdp_prog) { struct redirect_info *ri = this_cpu_ptr(&redirect_info); - unsigned int len; u32 index = ri->ifindex; + struct net_device *fwd; + unsigned int len; + int err = 0; - dev = dev_get_by_index_rcu(dev_net(dev), index); + fwd = dev_get_by_index_rcu(dev_net(dev), index); ri->ifindex = 0; - if (unlikely(!dev)) { - return -EINVAL; + if (unlikely(!fwd)) { + err = -EINVAL; + goto out; } - if (unlikely(!(dev->flags & IFF_UP))) - return -ENETDOWN; - len = dev->mtu + dev->hard_header_len + VLAN_HLEN; - if (skb->len > len) - return -E2BIG; + if (unlikely(!(fwd->flags & IFF_UP))) { + err = -ENETDOWN; + goto out; + } - skb->dev = dev; - return 0; + len = fwd->mtu + fwd->hard_header_len + VLAN_HLEN; + if (skb->len > len) { + err = -EMSGSIZE; + goto out; + } + + skb->dev = fwd; +out: + trace_xdp_redirect(dev, fwd, xdp_prog, XDP_REDIRECT, err); + return err; } EXPORT_SYMBOL_GPL(xdp_do_generic_redirect); -- cgit v1.2.3