From b824c3e16c1904bf80df489e293d1e3cbf98896d Mon Sep 17 00:00:00 2001 From: Sebastian Andrzej Siewior Date: Mon, 2 Mar 2026 17:26:31 +0100 Subject: net: Provide a PREEMPT_RT specific check for netdev_queue::_xmit_lock After acquiring netdev_queue::_xmit_lock the number of the CPU owning the lock is recorded in netdev_queue::xmit_lock_owner. This works as long as the BH context is not preemptible. On PREEMPT_RT the softirq context is preemptible and without the softirq-lock it is possible to have multiple user in __dev_queue_xmit() submitting a skb on the same CPU. This is fine in general but this means also that the current CPU is recorded as netdev_queue::xmit_lock_owner. This in turn leads to the recursion alert and the skb is dropped. Instead checking the for CPU number, that owns the lock, PREEMPT_RT can check if the lockowner matches the current task. Add netif_tx_owned() which returns true if the current context owns the lock by comparing the provided CPU number with the recorded number. This resembles the current check by negating the condition (the current check returns true if the lock is not owned). On PREEMPT_RT use rt_mutex_owner() to return the lock owner and compare the current task against it. Use the new helper in __dev_queue_xmit() and netif_local_xmit_active() which provides a similar check. Update comments regarding pairing READ_ONCE(). Reported-by: Bert Karwatzki Closes: https://lore.kernel.org/all/20260216134333.412332-1-spasswolf@web.de Fixes: 3253cb49cbad4 ("softirq: Allow to drop the softirq-BKL lock on PREEMPT_RT") Signed-off-by: Sebastian Andrzej Siewior Reported-by: Bert Karwatzki Signed-off-by: Sebastian Andrzej Siewior Link: https://patch.msgid.link/20260302162631.uGUyIqDT@linutronix.de Signed-off-by: Paolo Abeni --- include/linux/netdevice.h | 27 ++++++++++++++++++++++----- 1 file changed, 22 insertions(+), 5 deletions(-) (limited to 'include/linux/netdevice.h') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index d4e6e00bb90a..67e25f6d15a4 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -4711,7 +4711,7 @@ static inline u32 netif_msg_init(int debug_value, int default_msg_enable_bits) static inline void __netif_tx_lock(struct netdev_queue *txq, int cpu) { spin_lock(&txq->_xmit_lock); - /* Pairs with READ_ONCE() in __dev_queue_xmit() */ + /* Pairs with READ_ONCE() in netif_tx_owned() */ WRITE_ONCE(txq->xmit_lock_owner, cpu); } @@ -4729,7 +4729,7 @@ static inline void __netif_tx_release(struct netdev_queue *txq) static inline void __netif_tx_lock_bh(struct netdev_queue *txq) { spin_lock_bh(&txq->_xmit_lock); - /* Pairs with READ_ONCE() in __dev_queue_xmit() */ + /* Pairs with READ_ONCE() in netif_tx_owned() */ WRITE_ONCE(txq->xmit_lock_owner, smp_processor_id()); } @@ -4738,7 +4738,7 @@ static inline bool __netif_tx_trylock(struct netdev_queue *txq) bool ok = spin_trylock(&txq->_xmit_lock); if (likely(ok)) { - /* Pairs with READ_ONCE() in __dev_queue_xmit() */ + /* Pairs with READ_ONCE() in netif_tx_owned() */ WRITE_ONCE(txq->xmit_lock_owner, smp_processor_id()); } return ok; @@ -4746,14 +4746,14 @@ static inline bool __netif_tx_trylock(struct netdev_queue *txq) static inline void __netif_tx_unlock(struct netdev_queue *txq) { - /* Pairs with READ_ONCE() in __dev_queue_xmit() */ + /* Pairs with READ_ONCE() in netif_tx_owned() */ WRITE_ONCE(txq->xmit_lock_owner, -1); spin_unlock(&txq->_xmit_lock); } static inline void __netif_tx_unlock_bh(struct netdev_queue *txq) { - /* Pairs with READ_ONCE() in __dev_queue_xmit() */ + /* Pairs with READ_ONCE() in netif_tx_owned() */ WRITE_ONCE(txq->xmit_lock_owner, -1); spin_unlock_bh(&txq->_xmit_lock); } @@ -4846,6 +4846,23 @@ static inline void netif_tx_disable(struct net_device *dev) local_bh_enable(); } +#ifndef CONFIG_PREEMPT_RT +static inline bool netif_tx_owned(struct netdev_queue *txq, unsigned int cpu) +{ + /* Other cpus might concurrently change txq->xmit_lock_owner + * to -1 or to their cpu id, but not to our id. + */ + return READ_ONCE(txq->xmit_lock_owner) == cpu; +} + +#else +static inline bool netif_tx_owned(struct netdev_queue *txq, unsigned int cpu) +{ + return rt_mutex_owner(&txq->_xmit_lock.lock) == current; +} + +#endif + static inline void netif_addr_lock(struct net_device *dev) { unsigned char nest_level = 0; -- cgit v1.2.3 From 6f1a9140ecda3baba3d945b9a6155af4268aafc4 Mon Sep 17 00:00:00 2001 From: Weiming Shi Date: Sat, 7 Mar 2026 00:01:34 +0800 Subject: net: add xmit recursion limit to tunnel xmit functions Tunnel xmit functions (iptunnel_xmit, ip6tunnel_xmit) lack their own recursion limit. When a bond device in broadcast mode has GRE tap interfaces as slaves, and those GRE tunnels route back through the bond, multicast/broadcast traffic triggers infinite recursion between bond_xmit_broadcast() and ip_tunnel_xmit()/ip6_tnl_xmit(), causing kernel stack overflow. The existing XMIT_RECURSION_LIMIT (8) in the no-qdisc path is not sufficient because tunnel recursion involves route lookups and full IP output, consuming much more stack per level. Use a lower limit of 4 (IP_TUNNEL_RECURSION_LIMIT) to prevent overflow. Add recursion detection using dev_xmit_recursion helpers directly in iptunnel_xmit() and ip6tunnel_xmit() to cover all IPv4/IPv6 tunnel paths including UDP encapsulated tunnels (VXLAN, Geneve, etc.). Move dev_xmit_recursion helpers from net/core/dev.h to public header include/linux/netdevice.h so they can be used by tunnel code. BUG: KASAN: stack-out-of-bounds in blake2s.constprop.0+0xe7/0x160 Write of size 32 at addr ffff88810033fed0 by task kworker/0:1/11 Workqueue: mld mld_ifc_work Call Trace: __build_flow_key.constprop.0 (net/ipv4/route.c:515) ip_rt_update_pmtu (net/ipv4/route.c:1073) iptunnel_xmit (net/ipv4/ip_tunnel_core.c:84) ip_tunnel_xmit (net/ipv4/ip_tunnel.c:847) gre_tap_xmit (net/ipv4/ip_gre.c:779) dev_hard_start_xmit (net/core/dev.c:3887) sch_direct_xmit (net/sched/sch_generic.c:347) __dev_queue_xmit (net/core/dev.c:4802) bond_dev_queue_xmit (drivers/net/bonding/bond_main.c:312) bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5279) bond_start_xmit (drivers/net/bonding/bond_main.c:5530) dev_hard_start_xmit (net/core/dev.c:3887) __dev_queue_xmit (net/core/dev.c:4841) ip_finish_output2 (net/ipv4/ip_output.c:237) ip_output (net/ipv4/ip_output.c:438) iptunnel_xmit (net/ipv4/ip_tunnel_core.c:86) gre_tap_xmit (net/ipv4/ip_gre.c:779) dev_hard_start_xmit (net/core/dev.c:3887) sch_direct_xmit (net/sched/sch_generic.c:347) __dev_queue_xmit (net/core/dev.c:4802) bond_dev_queue_xmit (drivers/net/bonding/bond_main.c:312) bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5279) bond_start_xmit (drivers/net/bonding/bond_main.c:5530) dev_hard_start_xmit (net/core/dev.c:3887) __dev_queue_xmit (net/core/dev.c:4841) ip_finish_output2 (net/ipv4/ip_output.c:237) ip_output (net/ipv4/ip_output.c:438) iptunnel_xmit (net/ipv4/ip_tunnel_core.c:86) ip_tunnel_xmit (net/ipv4/ip_tunnel.c:847) gre_tap_xmit (net/ipv4/ip_gre.c:779) dev_hard_start_xmit (net/core/dev.c:3887) sch_direct_xmit (net/sched/sch_generic.c:347) __dev_queue_xmit (net/core/dev.c:4802) bond_dev_queue_xmit (drivers/net/bonding/bond_main.c:312) bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5279) bond_start_xmit (drivers/net/bonding/bond_main.c:5530) dev_hard_start_xmit (net/core/dev.c:3887) __dev_queue_xmit (net/core/dev.c:4841) mld_sendpack mld_ifc_work process_one_work worker_thread Fixes: 745e20f1b626 ("net: add a recursion limit in xmit path") Reported-by: Xiang Mei Signed-off-by: Weiming Shi Link: https://patch.msgid.link/20260306160133.3852900-2-bestswngs@gmail.com Signed-off-by: Paolo Abeni --- include/linux/netdevice.h | 32 ++++++++++++++++++++++++++++++++ include/net/ip6_tunnel.h | 12 ++++++++++++ include/net/ip_tunnels.h | 7 +++++++ net/core/dev.h | 35 ----------------------------------- net/ipv4/ip_tunnel_core.c | 13 +++++++++++++ 5 files changed, 64 insertions(+), 35 deletions(-) (limited to 'include/linux/netdevice.h') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 67e25f6d15a4..ae269a2e7f4d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -3576,17 +3576,49 @@ struct page_pool_bh { }; DECLARE_PER_CPU(struct page_pool_bh, system_page_pool); +#define XMIT_RECURSION_LIMIT 8 + #ifndef CONFIG_PREEMPT_RT static inline int dev_recursion_level(void) { return this_cpu_read(softnet_data.xmit.recursion); } + +static inline bool dev_xmit_recursion(void) +{ + return unlikely(__this_cpu_read(softnet_data.xmit.recursion) > + XMIT_RECURSION_LIMIT); +} + +static inline void dev_xmit_recursion_inc(void) +{ + __this_cpu_inc(softnet_data.xmit.recursion); +} + +static inline void dev_xmit_recursion_dec(void) +{ + __this_cpu_dec(softnet_data.xmit.recursion); +} #else static inline int dev_recursion_level(void) { return current->net_xmit.recursion; } +static inline bool dev_xmit_recursion(void) +{ + return unlikely(current->net_xmit.recursion > XMIT_RECURSION_LIMIT); +} + +static inline void dev_xmit_recursion_inc(void) +{ + current->net_xmit.recursion++; +} + +static inline void dev_xmit_recursion_dec(void) +{ + current->net_xmit.recursion--; +} #endif void __netif_schedule(struct Qdisc *q); diff --git a/include/net/ip6_tunnel.h b/include/net/ip6_tunnel.h index 120db2865811..1253cbb4b0a4 100644 --- a/include/net/ip6_tunnel.h +++ b/include/net/ip6_tunnel.h @@ -156,6 +156,16 @@ static inline void ip6tunnel_xmit(struct sock *sk, struct sk_buff *skb, { int pkt_len, err; + if (dev_recursion_level() > IP_TUNNEL_RECURSION_LIMIT) { + net_crit_ratelimited("Dead loop on virtual device %s, fix it urgently!\n", + dev->name); + DEV_STATS_INC(dev, tx_errors); + kfree_skb(skb); + return; + } + + dev_xmit_recursion_inc(); + memset(skb->cb, 0, sizeof(struct inet6_skb_parm)); IP6CB(skb)->flags = ip6cb_flags; pkt_len = skb->len - skb_inner_network_offset(skb); @@ -166,6 +176,8 @@ static inline void ip6tunnel_xmit(struct sock *sk, struct sk_buff *skb, pkt_len = -1; iptunnel_xmit_stats(dev, pkt_len); } + + dev_xmit_recursion_dec(); } #endif #endif diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h index 4021e6a73e32..80662f812080 100644 --- a/include/net/ip_tunnels.h +++ b/include/net/ip_tunnels.h @@ -27,6 +27,13 @@ #include #endif +/* Recursion limit for tunnel xmit to detect routing loops. + * Unlike XMIT_RECURSION_LIMIT (8) used in the no-qdisc path, tunnel + * recursion involves route lookups and full IP output, consuming much + * more stack per level, so a lower limit is needed. + */ +#define IP_TUNNEL_RECURSION_LIMIT 4 + /* Keep error state on tunnel for 30 sec */ #define IPTUNNEL_ERR_TIMEO (30*HZ) diff --git a/net/core/dev.h b/net/core/dev.h index 98793a738f43..781619e76b3e 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -366,41 +366,6 @@ static inline void napi_assert_will_not_race(const struct napi_struct *napi) void kick_defer_list_purge(unsigned int cpu); -#define XMIT_RECURSION_LIMIT 8 - -#ifndef CONFIG_PREEMPT_RT -static inline bool dev_xmit_recursion(void) -{ - return unlikely(__this_cpu_read(softnet_data.xmit.recursion) > - XMIT_RECURSION_LIMIT); -} - -static inline void dev_xmit_recursion_inc(void) -{ - __this_cpu_inc(softnet_data.xmit.recursion); -} - -static inline void dev_xmit_recursion_dec(void) -{ - __this_cpu_dec(softnet_data.xmit.recursion); -} -#else -static inline bool dev_xmit_recursion(void) -{ - return unlikely(current->net_xmit.recursion > XMIT_RECURSION_LIMIT); -} - -static inline void dev_xmit_recursion_inc(void) -{ - current->net_xmit.recursion++; -} - -static inline void dev_xmit_recursion_dec(void) -{ - current->net_xmit.recursion--; -} -#endif - int dev_set_hwtstamp_phylib(struct net_device *dev, struct kernel_hwtstamp_config *cfg, struct netlink_ext_ack *extack); diff --git a/net/ipv4/ip_tunnel_core.c b/net/ipv4/ip_tunnel_core.c index 2e61ac137128..b1b6bf949f65 100644 --- a/net/ipv4/ip_tunnel_core.c +++ b/net/ipv4/ip_tunnel_core.c @@ -58,6 +58,17 @@ void iptunnel_xmit(struct sock *sk, struct rtable *rt, struct sk_buff *skb, struct iphdr *iph; int err; + if (dev_recursion_level() > IP_TUNNEL_RECURSION_LIMIT) { + net_crit_ratelimited("Dead loop on virtual device %s, fix it urgently!\n", + dev->name); + DEV_STATS_INC(dev, tx_errors); + ip_rt_put(rt); + kfree_skb(skb); + return; + } + + dev_xmit_recursion_inc(); + skb_scrub_packet(skb, xnet); skb_clear_hash_if_not_l4(skb); @@ -88,6 +99,8 @@ void iptunnel_xmit(struct sock *sk, struct rtable *rt, struct sk_buff *skb, pkt_len = 0; iptunnel_xmit_stats(dev, pkt_len); } + + dev_xmit_recursion_dec(); } EXPORT_SYMBOL_GPL(iptunnel_xmit); -- cgit v1.2.3