From 680aea08e78c003292415518ad270bc20f9c80b1 Mon Sep 17 00:00:00 2001 From: Amit Cohen Date: Mon, 1 Feb 2021 21:47:52 +0200 Subject: net: ipv4: Emit notification when fib hardware flags are changed After installing a route to the kernel, user space receives an acknowledgment, which means the route was installed in the kernel, but not necessarily in hardware. The asynchronous nature of route installation in hardware can lead to a routing daemon advertising a route before it was actually installed in hardware. This can result in packet loss or mis-routed packets until the route is installed in hardware. It is also possible for a route already installed in hardware to change its action and therefore its flags. For example, a host route that is trapping packets can be "promoted" to perform decapsulation following the installation of an IPinIP/VXLAN tunnel. Emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/RTM_F_TRAP flags are changed. The aim is to provide an indication to user-space (e.g., routing daemons) about the state of the route in hardware. Introduce a sysctl that controls this behavior. Keep the default value at 0 (i.e., do not emit notifications) for several reasons: - Multiple RTM_NEWROUTE notification per-route might confuse existing routing daemons. - Convergence reasons in routing daemons. - The extra notifications will negatively impact the insertion rate. - Not all users are interested in these notifications. Signed-off-by: Amit Cohen Acked-by: Roopa Prabhu Signed-off-by: Ido Schimmel Reviewed-by: David Ahern Signed-off-by: Jakub Kicinski --- include/net/netns/ipv4.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h index 8e4fcac4df72..70a2a085dd1a 100644 --- a/include/net/netns/ipv4.h +++ b/include/net/netns/ipv4.h @@ -188,6 +188,8 @@ struct netns_ipv4 { int sysctl_udp_wmem_min; int sysctl_udp_rmem_min; + int sysctl_fib_notify_on_flag_change; + #ifdef CONFIG_NET_L3_MASTER_DEV int sysctl_udp_l3mdev_accept; #endif -- cgit v1.3-14-g43fede From fbaca8f895a6855b0b442227bc21da2d4cf77a01 Mon Sep 17 00:00:00 2001 From: Amit Cohen Date: Mon, 1 Feb 2021 21:47:53 +0200 Subject: net: Pass 'net' struct as first argument to fib6_info_hw_flags_set() The next patch will emit notification when hardware flags are changed, in case that fib_notify_on_flag_change sysctl is set to 1. To know sysctl values, net struct is needed. This change is consistent with the IPv4 version, which gets 'net' struct as its first argument. Currently, the only callers of this function are mlxsw and netdevsim. Patch the callers to pass net. Signed-off-by: Amit Cohen Signed-off-by: Ido Schimmel Reviewed-by: David Ahern Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 7 ++++--- drivers/net/netdevsim/fib.c | 14 ++++++++------ include/net/ip6_fib.h | 5 +++-- 3 files changed, 15 insertions(+), 11 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c index 0ac7014703aa..5516e141f5bf 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c +++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c @@ -5004,8 +5004,8 @@ mlxsw_sp_fib6_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp, fib6_entry = container_of(fib_entry, struct mlxsw_sp_fib6_entry, common); list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) - fib6_info_hw_flags_set(mlxsw_sp_rt6->rt, should_offload, - !should_offload); + fib6_info_hw_flags_set(mlxsw_sp_net(mlxsw_sp), mlxsw_sp_rt6->rt, + should_offload, !should_offload); } static void @@ -5018,7 +5018,8 @@ mlxsw_sp_fib6_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp, fib6_entry = container_of(fib_entry, struct mlxsw_sp_fib6_entry, common); list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) - fib6_info_hw_flags_set(mlxsw_sp_rt6->rt, false, false); + fib6_info_hw_flags_set(mlxsw_sp_net(mlxsw_sp), mlxsw_sp_rt6->rt, + false, false); } static void diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index d557533d43dd..cb68f0cc6740 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -585,13 +585,15 @@ err_fib6_rt_nh_del: return err; } -static void nsim_fib6_rt_hw_flags_set(const struct nsim_fib6_rt *fib6_rt, +static void nsim_fib6_rt_hw_flags_set(struct nsim_fib_data *data, + const struct nsim_fib6_rt *fib6_rt, bool trap) { + struct net *net = devlink_net(data->devlink); struct nsim_fib6_rt_nh *fib6_rt_nh; list_for_each_entry(fib6_rt_nh, &fib6_rt->nh_list, list) - fib6_info_hw_flags_set(fib6_rt_nh->rt, false, trap); + fib6_info_hw_flags_set(net, fib6_rt_nh->rt, false, trap); } static int nsim_fib6_rt_add(struct nsim_fib_data *data, @@ -607,7 +609,7 @@ static int nsim_fib6_rt_add(struct nsim_fib_data *data, goto err_fib_dismiss; msleep(1); - nsim_fib6_rt_hw_flags_set(fib6_rt, true); + nsim_fib6_rt_hw_flags_set(data, fib6_rt, true); return 0; @@ -641,9 +643,9 @@ static int nsim_fib6_rt_replace(struct nsim_fib_data *data, return err; msleep(1); - nsim_fib6_rt_hw_flags_set(fib6_rt, true); + nsim_fib6_rt_hw_flags_set(data, fib6_rt, true); - nsim_fib6_rt_hw_flags_set(fib6_rt_old, false); + nsim_fib6_rt_hw_flags_set(data, fib6_rt_old, false); nsim_fib6_rt_destroy(fib6_rt_old); return 0; @@ -954,7 +956,7 @@ static void nsim_fib6_rt_free(struct nsim_fib_rt *fib_rt, struct nsim_fib6_rt *fib6_rt; fib6_rt = container_of(fib_rt, struct nsim_fib6_rt, common); - nsim_fib6_rt_hw_flags_set(fib6_rt, false); + nsim_fib6_rt_hw_flags_set(data, fib6_rt, false); nsim_fib_account(&data->ipv6.fib, false); nsim_fib6_rt_destroy(fib6_rt); } diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h index ac5ff3c3afb1..cc189e668adf 100644 --- a/include/net/ip6_fib.h +++ b/include/net/ip6_fib.h @@ -336,8 +336,9 @@ static inline void fib6_info_release(struct fib6_info *f6i) call_rcu(&f6i->rcu, fib6_info_destroy_rcu); } -static inline void fib6_info_hw_flags_set(struct fib6_info *f6i, bool offload, - bool trap) +static inline void +fib6_info_hw_flags_set(struct net *net, struct fib6_info *f6i, bool offload, + bool trap) { f6i->offload = offload; f6i->trap = trap; -- cgit v1.3-14-g43fede From 907eea486888cfe118c19759bbc4b8fca2e004df Mon Sep 17 00:00:00 2001 From: Amit Cohen Date: Mon, 1 Feb 2021 21:47:55 +0200 Subject: net: ipv6: Emit notification when fib hardware flags are changed After installing a route to the kernel, user space receives an acknowledgment, which means the route was installed in the kernel, but not necessarily in hardware. The asynchronous nature of route installation in hardware can lead to a routing daemon advertising a route before it was actually installed in hardware. This can result in packet loss or mis-routed packets until the route is installed in hardware. It is also possible for a route already installed in hardware to change its action and therefore its flags. For example, a host route that is trapping packets can be "promoted" to perform decapsulation following the installation of an IPinIP/VXLAN tunnel. Emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/RTM_F_TRAP flags are changed. The aim is to provide an indication to user-space (e.g., routing daemons) about the state of the route in hardware. Introduce a sysctl that controls this behavior. Keep the default value at 0 (i.e., do not emit notifications) for several reasons: - Multiple RTM_NEWROUTE notification per-route might confuse existing routing daemons. - Convergence reasons in routing daemons. - The extra notifications will negatively impact the insertion rate. - Not all users are interested in these notifications. Move fib6_info_hw_flags_set() to C file because it is no longer a short function. Signed-off-by: Amit Cohen Signed-off-by: Ido Schimmel Reviewed-by: David Ahern Signed-off-by: Jakub Kicinski --- Documentation/networking/ip-sysctl.rst | 20 ++++++++++++++++ include/net/ip6_fib.h | 10 ++------ include/net/netns/ipv6.h | 1 + net/ipv6/af_inet6.c | 1 + net/ipv6/route.c | 44 ++++++++++++++++++++++++++++++++++ net/ipv6/sysctl_net_ipv6.c | 9 +++++++ 6 files changed, 77 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/Documentation/networking/ip-sysctl.rst b/Documentation/networking/ip-sysctl.rst index b0e800e68366..61a358301f12 100644 --- a/Documentation/networking/ip-sysctl.rst +++ b/Documentation/networking/ip-sysctl.rst @@ -1795,6 +1795,26 @@ nexthop_compat_mode - BOOLEAN and extraneous notifications. Default: true (backward compat mode) +fib_notify_on_flag_change - INTEGER + Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/ + RTM_F_TRAP flags are changed. + + After installing a route to the kernel, user space receives an + acknowledgment, which means the route was installed in the kernel, + but not necessarily in hardware. + It is also possible for a route already installed in hardware to change + its action and therefore its flags. For example, a host route that is + trapping packets can be "promoted" to perform decapsulation following + the installation of an IPinIP/VXLAN tunnel. + The notifications will indicate to user-space the state of the route. + + Default: 0 (Do not emit notifications.) + + Possible values: + + - 0 - Do not emit notifications. + - 1 - Emit notifications. + IPv6 Fragmentation: ip6frag_high_thresh - INTEGER diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h index cc189e668adf..1e262b23c68b 100644 --- a/include/net/ip6_fib.h +++ b/include/net/ip6_fib.h @@ -336,14 +336,6 @@ static inline void fib6_info_release(struct fib6_info *f6i) call_rcu(&f6i->rcu, fib6_info_destroy_rcu); } -static inline void -fib6_info_hw_flags_set(struct net *net, struct fib6_info *f6i, bool offload, - bool trap) -{ - f6i->offload = offload; - f6i->trap = trap; -} - enum fib6_walk_state { #ifdef CONFIG_IPV6_SUBTREES FWS_S, @@ -546,6 +538,8 @@ static inline bool fib6_metric_locked(struct fib6_info *f6i, int metric) { return !!(f6i->fib6_metrics->metrics[RTAX_LOCK - 1] & (1 << metric)); } +void fib6_info_hw_flags_set(struct net *net, struct fib6_info *f6i, + bool offload, bool trap); #if IS_BUILTIN(CONFIG_IPV6) && defined(CONFIG_BPF_SYSCALL) struct bpf_iter__ipv6_route { diff --git a/include/net/netns/ipv6.h b/include/net/netns/ipv6.h index 5ec054473d81..21c0debbd39e 100644 --- a/include/net/netns/ipv6.h +++ b/include/net/netns/ipv6.h @@ -51,6 +51,7 @@ struct netns_sysctl_ipv6 { int max_hbh_opts_len; int seg6_flowlabel; bool skip_notify_on_dev_down; + int fib_notify_on_flag_change; }; struct netns_ipv6 { diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index 8e9c3e9ea36e..0e9994e0ecd7 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -954,6 +954,7 @@ static int __net_init inet6_net_init(struct net *net) net->ipv6.sysctl.max_hbh_opts_cnt = IP6_DEFAULT_MAX_HBH_OPTS_CNT; net->ipv6.sysctl.max_dst_opts_len = IP6_DEFAULT_MAX_DST_OPTS_LEN; net->ipv6.sysctl.max_hbh_opts_len = IP6_DEFAULT_MAX_HBH_OPTS_LEN; + net->ipv6.sysctl.fib_notify_on_flag_change = 0; atomic_set(&net->ipv6.fib6_sernum, 1); err = ipv6_init_mibs(net); diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 41d8f801b75f..92b400eb8476 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -6064,6 +6064,50 @@ errout: rtnl_set_sk_err(net, RTNLGRP_IPV6_ROUTE, err); } +void fib6_info_hw_flags_set(struct net *net, struct fib6_info *f6i, + bool offload, bool trap) +{ + struct sk_buff *skb; + int err; + + if (f6i->offload == offload && f6i->trap == trap) + return; + + f6i->offload = offload; + f6i->trap = trap; + + if (!rcu_access_pointer(f6i->fib6_node)) + /* The route was removed from the tree, do not send + * notfication. + */ + return; + + if (!net->ipv6.sysctl.fib_notify_on_flag_change) + return; + + skb = nlmsg_new(rt6_nlmsg_size(f6i), GFP_KERNEL); + if (!skb) { + err = -ENOBUFS; + goto errout; + } + + err = rt6_fill_node(net, skb, f6i, NULL, NULL, NULL, 0, RTM_NEWROUTE, 0, + 0, 0); + if (err < 0) { + /* -EMSGSIZE implies BUG in rt6_nlmsg_size() */ + WARN_ON(err == -EMSGSIZE); + kfree_skb(skb); + goto errout; + } + + rtnl_notify(skb, net, 0, RTNLGRP_IPV6_ROUTE, NULL, GFP_KERNEL); + return; + +errout: + rtnl_set_sk_err(net, RTNLGRP_IPV6_ROUTE, err); +} +EXPORT_SYMBOL(fib6_info_hw_flags_set); + static int ip6_route_dev_notify(struct notifier_block *this, unsigned long event, void *ptr) { diff --git a/net/ipv6/sysctl_net_ipv6.c b/net/ipv6/sysctl_net_ipv6.c index 5b60a4bdd36a..392ef01e3366 100644 --- a/net/ipv6/sysctl_net_ipv6.c +++ b/net/ipv6/sysctl_net_ipv6.c @@ -160,6 +160,15 @@ static struct ctl_table ipv6_table_template[] = { .mode = 0644, .proc_handler = proc_dointvec }, + { + .procname = "fib_notify_on_flag_change", + .data = &init_net.ipv6.sysctl.fib_notify_on_flag_change, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = SYSCTL_ZERO, + .extra2 = SYSCTL_ONE, + }, { } }; -- cgit v1.3-14-g43fede