From 6f2667b98ef2ea81aad6761409604480d14c9095 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:34 -0700 Subject: [PATCH 01/16] vxlan: Convert vxlan_exit_batch_rtnl() to ->exit_rtnl(). vxlan_exit_batch_rtnl() iterates the dying netns list and performs the same operations for each. Let's use ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Nikolay Aleksandrov Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-6-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- drivers/net/vxlan/vxlan_core.c | 18 +++++++----------- 1 file changed, 7 insertions(+), 11 deletions(-) diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c index 9ccc3f09f71b..56aee539c235 100644 --- a/drivers/net/vxlan/vxlan_core.c +++ b/drivers/net/vxlan/vxlan_core.c @@ -4966,19 +4966,15 @@ static void __net_exit vxlan_destroy_tunnels(struct vxlan_net *vn, vxlan_dellink(vxlan->dev, dev_to_kill); } -static void __net_exit vxlan_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) +static void __net_exit vxlan_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - struct net *net; - - ASSERT_RTNL(); - list_for_each_entry(net, net_list, exit_list) { - struct vxlan_net *vn = net_generic(net, vxlan_net_id); + struct vxlan_net *vn = net_generic(net, vxlan_net_id); - __unregister_nexthop_notifier(net, &vn->nexthop_notifier_block); + ASSERT_RTNL_NET(net); - vxlan_destroy_tunnels(vn, dev_to_kill); - } + __unregister_nexthop_notifier(net, &vn->nexthop_notifier_block); + vxlan_destroy_tunnels(vn, dev_to_kill); } static void __net_exit vxlan_exit_net(struct net *net) @@ -4992,7 +4988,7 @@ static void __net_exit vxlan_exit_net(struct net *net) static struct pernet_operations vxlan_net_ops = { .init = vxlan_init_net, - .exit_batch_rtnl = vxlan_exit_batch_rtnl, + .exit_rtnl = vxlan_exit_rtnl, .exit = vxlan_exit_net, .id = &vxlan_net_id, .size = sizeof(struct vxlan_net), -- 2.50.1 From a967e01e2ad201f6ddc778ed65a5dae1c68ee8a5 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:35 -0700 Subject: [PATCH 02/16] ipv4: ip_tunnel: Convert ip_tunnel_delete_nets() callers to ->exit_rtnl(). ip_tunnel_delete_nets() iterates the dying netns list and performs the same operations for each. Let's export ip_tunnel_destroy() as ip_tunnel_delete_net() and call it from ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: David Ahern Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-7-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- include/net/ip_tunnels.h | 7 +++---- net/ipv4/ip_gre.c | 27 ++++++++++++--------------- net/ipv4/ip_tunnel.c | 25 +++++++------------------ net/ipv4/ip_vti.c | 9 ++++----- net/ipv4/ipip.c | 9 ++++----- 5 files changed, 30 insertions(+), 47 deletions(-) diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h index a36a335cef9f..0c3d571a04a1 100644 --- a/include/net/ip_tunnels.h +++ b/include/net/ip_tunnels.h @@ -377,10 +377,9 @@ struct net *ip_tunnel_get_link_net(const struct net_device *dev); int ip_tunnel_get_iflink(const struct net_device *dev); int ip_tunnel_init_net(struct net *net, unsigned int ip_tnl_net_id, struct rtnl_link_ops *ops, char *devname); - -void ip_tunnel_delete_nets(struct list_head *list_net, unsigned int id, - struct rtnl_link_ops *ops, - struct list_head *dev_to_kill); +void ip_tunnel_delete_net(struct net *net, unsigned int id, + struct rtnl_link_ops *ops, + struct list_head *dev_to_kill); void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, const struct iphdr *tnl_params, const u8 protocol); diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c index 26d15f907551..f5b9004d6938 100644 --- a/net/ipv4/ip_gre.c +++ b/net/ipv4/ip_gre.c @@ -1066,16 +1066,15 @@ static int __net_init ipgre_init_net(struct net *net) return ip_tunnel_init_net(net, ipgre_net_id, &ipgre_link_ops, NULL); } -static void __net_exit ipgre_exit_batch_rtnl(struct list_head *list_net, - struct list_head *dev_to_kill) +static void __net_exit ipgre_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - ip_tunnel_delete_nets(list_net, ipgre_net_id, &ipgre_link_ops, - dev_to_kill); + ip_tunnel_delete_net(net, ipgre_net_id, &ipgre_link_ops, dev_to_kill); } static struct pernet_operations ipgre_net_ops = { .init = ipgre_init_net, - .exit_batch_rtnl = ipgre_exit_batch_rtnl, + .exit_rtnl = ipgre_exit_rtnl, .id = &ipgre_net_id, .size = sizeof(struct ip_tunnel_net), }; @@ -1752,16 +1751,15 @@ static int __net_init ipgre_tap_init_net(struct net *net) return ip_tunnel_init_net(net, gre_tap_net_id, &ipgre_tap_ops, "gretap0"); } -static void __net_exit ipgre_tap_exit_batch_rtnl(struct list_head *list_net, - struct list_head *dev_to_kill) +static void __net_exit ipgre_tap_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - ip_tunnel_delete_nets(list_net, gre_tap_net_id, &ipgre_tap_ops, - dev_to_kill); + ip_tunnel_delete_net(net, gre_tap_net_id, &ipgre_tap_ops, dev_to_kill); } static struct pernet_operations ipgre_tap_net_ops = { .init = ipgre_tap_init_net, - .exit_batch_rtnl = ipgre_tap_exit_batch_rtnl, + .exit_rtnl = ipgre_tap_exit_rtnl, .id = &gre_tap_net_id, .size = sizeof(struct ip_tunnel_net), }; @@ -1772,16 +1770,15 @@ static int __net_init erspan_init_net(struct net *net) &erspan_link_ops, "erspan0"); } -static void __net_exit erspan_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) +static void __net_exit erspan_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - ip_tunnel_delete_nets(net_list, erspan_net_id, &erspan_link_ops, - dev_to_kill); + ip_tunnel_delete_net(net, erspan_net_id, &erspan_link_ops, dev_to_kill); } static struct pernet_operations erspan_net_ops = { .init = erspan_init_net, - .exit_batch_rtnl = erspan_exit_batch_rtnl, + .exit_rtnl = erspan_exit_rtnl, .id = &erspan_net_id, .size = sizeof(struct ip_tunnel_net), }; diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c index 1024f961ec9a..3913ec89ad20 100644 --- a/net/ipv4/ip_tunnel.c +++ b/net/ipv4/ip_tunnel.c @@ -1174,13 +1174,16 @@ int ip_tunnel_init_net(struct net *net, unsigned int ip_tnl_net_id, } EXPORT_SYMBOL_GPL(ip_tunnel_init_net); -static void ip_tunnel_destroy(struct net *net, struct ip_tunnel_net *itn, - struct list_head *head, - struct rtnl_link_ops *ops) +void ip_tunnel_delete_net(struct net *net, unsigned int id, + struct rtnl_link_ops *ops, + struct list_head *head) { + struct ip_tunnel_net *itn = net_generic(net, id); struct net_device *dev, *aux; int h; + ASSERT_RTNL_NET(net); + for_each_netdev_safe(net, dev, aux) if (dev->rtnl_link_ops == ops) unregister_netdevice_queue(dev, head); @@ -1198,21 +1201,7 @@ static void ip_tunnel_destroy(struct net *net, struct ip_tunnel_net *itn, unregister_netdevice_queue(t->dev, head); } } - -void ip_tunnel_delete_nets(struct list_head *net_list, unsigned int id, - struct rtnl_link_ops *ops, - struct list_head *dev_to_kill) -{ - struct ip_tunnel_net *itn; - struct net *net; - - ASSERT_RTNL(); - list_for_each_entry(net, net_list, exit_list) { - itn = net_generic(net, id); - ip_tunnel_destroy(net, itn, dev_to_kill, ops); - } -} -EXPORT_SYMBOL_GPL(ip_tunnel_delete_nets); +EXPORT_SYMBOL_GPL(ip_tunnel_delete_net); int ip_tunnel_newlink(struct net *net, struct net_device *dev, struct nlattr *tb[], struct ip_tunnel_parm_kern *p, diff --git a/net/ipv4/ip_vti.c b/net/ipv4/ip_vti.c index 159b4473290e..686e4f3d83aa 100644 --- a/net/ipv4/ip_vti.c +++ b/net/ipv4/ip_vti.c @@ -523,16 +523,15 @@ static int __net_init vti_init_net(struct net *net) return 0; } -static void __net_exit vti_exit_batch_rtnl(struct list_head *list_net, - struct list_head *dev_to_kill) +static void __net_exit vti_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - ip_tunnel_delete_nets(list_net, vti_net_id, &vti_link_ops, - dev_to_kill); + ip_tunnel_delete_net(net, vti_net_id, &vti_link_ops, dev_to_kill); } static struct pernet_operations vti_net_ops = { .init = vti_init_net, - .exit_batch_rtnl = vti_exit_batch_rtnl, + .exit_rtnl = vti_exit_rtnl, .id = &vti_net_id, .size = sizeof(struct ip_tunnel_net), }; diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c index bab0bf90c908..3e03af073a1c 100644 --- a/net/ipv4/ipip.c +++ b/net/ipv4/ipip.c @@ -604,16 +604,15 @@ static int __net_init ipip_init_net(struct net *net) return ip_tunnel_init_net(net, ipip_net_id, &ipip_link_ops, "tunl0"); } -static void __net_exit ipip_exit_batch_rtnl(struct list_head *list_net, - struct list_head *dev_to_kill) +static void __net_exit ipip_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - ip_tunnel_delete_nets(list_net, ipip_net_id, &ipip_link_ops, - dev_to_kill); + ip_tunnel_delete_net(net, ipip_net_id, &ipip_link_ops, dev_to_kill); } static struct pernet_operations ipip_net_ops = { .init = ipip_init_net, - .exit_batch_rtnl = ipip_exit_batch_rtnl, + .exit_rtnl = ipip_exit_rtnl, .id = &ipip_net_id, .size = sizeof(struct ip_tunnel_net), }; -- 2.50.1 From f76758f18fb8677464a71cdcdd1920ffe6d59d85 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:36 -0700 Subject: [PATCH 03/16] ipv6: Convert tunnel devices' ->exit_batch_rtnl() to ->exit_rtnl(). The following functions iterates the dying netns list and performs the same operations for each netns. * ip6gre_exit_batch_rtnl() * ip6_tnl_exit_batch_rtnl() * vti6_exit_batch_rtnl() * sit_exit_batch_rtnl() Let's use ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: David Ahern Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-8-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- net/ipv6/ip6_gre.c | 22 ++++++---------------- net/ipv6/ip6_tunnel.c | 24 ++++++++---------------- net/ipv6/ip6_vti.c | 27 +++++++-------------------- net/ipv6/sit.c | 23 ++++++----------------- 4 files changed, 27 insertions(+), 69 deletions(-) diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c index 957ca98fa70f..2dc9dcffe2ca 100644 --- a/net/ipv6/ip6_gre.c +++ b/net/ipv6/ip6_gre.c @@ -1570,7 +1570,7 @@ static struct inet6_protocol ip6gre_protocol __read_mostly = { .flags = INET6_PROTO_FINAL, }; -static void ip6gre_destroy_tunnels(struct net *net, struct list_head *head) +static void __net_exit ip6gre_exit_rtnl_net(struct net *net, struct list_head *head) { struct ip6gre_net *ign = net_generic(net, ip6gre_net_id); struct net_device *dev, *aux; @@ -1587,16 +1587,16 @@ static void ip6gre_destroy_tunnels(struct net *net, struct list_head *head) for (h = 0; h < IP6_GRE_HASH_SIZE; h++) { struct ip6_tnl *t; - t = rtnl_dereference(ign->tunnels[prio][h]); + t = rtnl_net_dereference(net, ign->tunnels[prio][h]); while (t) { /* If dev is in the same netns, it has already * been added to the list by the previous loop. */ if (!net_eq(dev_net(t->dev), net)) - unregister_netdevice_queue(t->dev, - head); - t = rtnl_dereference(t->next); + unregister_netdevice_queue(t->dev, head); + + t = rtnl_net_dereference(net, t->next); } } } @@ -1640,19 +1640,9 @@ err_alloc_dev: return err; } -static void __net_exit ip6gre_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) -{ - struct net *net; - - ASSERT_RTNL(); - list_for_each_entry(net, net_list, exit_list) - ip6gre_destroy_tunnels(net, dev_to_kill); -} - static struct pernet_operations ip6gre_net_ops = { .init = ip6gre_init_net, - .exit_batch_rtnl = ip6gre_exit_batch_rtnl, + .exit_rtnl = ip6gre_exit_rtnl_net, .id = &ip6gre_net_id, .size = sizeof(struct ip6gre_net), }; diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index a04dd1bb4b19..894d3158a6f0 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -2210,7 +2210,7 @@ static struct xfrm6_tunnel mplsip6_handler __read_mostly = { .priority = 1, }; -static void __net_exit ip6_tnl_destroy_tunnels(struct net *net, struct list_head *list) +static void __net_exit ip6_tnl_exit_rtnl_net(struct net *net, struct list_head *list) { struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id); struct net_device *dev, *aux; @@ -2222,25 +2222,27 @@ static void __net_exit ip6_tnl_destroy_tunnels(struct net *net, struct list_head unregister_netdevice_queue(dev, list); for (h = 0; h < IP6_TUNNEL_HASH_SIZE; h++) { - t = rtnl_dereference(ip6n->tnls_r_l[h]); + t = rtnl_net_dereference(net, ip6n->tnls_r_l[h]); while (t) { /* If dev is in the same netns, it has already * been added to the list by the previous loop. */ if (!net_eq(dev_net(t->dev), net)) unregister_netdevice_queue(t->dev, list); - t = rtnl_dereference(t->next); + + t = rtnl_net_dereference(net, t->next); } } - t = rtnl_dereference(ip6n->tnls_wc[0]); + t = rtnl_net_dereference(net, ip6n->tnls_wc[0]); while (t) { /* If dev is in the same netns, it has already * been added to the list by the previous loop. */ if (!net_eq(dev_net(t->dev), net)) unregister_netdevice_queue(t->dev, list); - t = rtnl_dereference(t->next); + + t = rtnl_net_dereference(net, t->next); } } @@ -2287,19 +2289,9 @@ err_alloc_dev: return err; } -static void __net_exit ip6_tnl_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) -{ - struct net *net; - - ASSERT_RTNL(); - list_for_each_entry(net, net_list, exit_list) - ip6_tnl_destroy_tunnels(net, dev_to_kill); -} - static struct pernet_operations ip6_tnl_net_ops = { .init = ip6_tnl_init_net, - .exit_batch_rtnl = ip6_tnl_exit_batch_rtnl, + .exit_rtnl = ip6_tnl_exit_rtnl_net, .id = &ip6_tnl_net_id, .size = sizeof(struct ip6_tnl_net), }; diff --git a/net/ipv6/ip6_vti.c b/net/ipv6/ip6_vti.c index 09ec4b0ad7dc..40464a88bca6 100644 --- a/net/ipv6/ip6_vti.c +++ b/net/ipv6/ip6_vti.c @@ -1112,21 +1112,21 @@ static struct rtnl_link_ops vti6_link_ops __read_mostly = { .get_link_net = ip6_tnl_get_link_net, }; -static void __net_exit vti6_destroy_tunnels(struct vti6_net *ip6n, - struct list_head *list) +static void __net_exit vti6_exit_rtnl_net(struct net *net, struct list_head *list) { - int h; + struct vti6_net *ip6n = net_generic(net, vti6_net_id); struct ip6_tnl *t; + int h; for (h = 0; h < IP6_VTI_HASH_SIZE; h++) { - t = rtnl_dereference(ip6n->tnls_r_l[h]); + t = rtnl_net_dereference(net, ip6n->tnls_r_l[h]); while (t) { unregister_netdevice_queue(t->dev, list); - t = rtnl_dereference(t->next); + t = rtnl_net_dereference(net, t->next); } } - t = rtnl_dereference(ip6n->tnls_wc[0]); + t = rtnl_net_dereference(net, ip6n->tnls_wc[0]); if (t) unregister_netdevice_queue(t->dev, list); } @@ -1170,22 +1170,9 @@ err_alloc_dev: return err; } -static void __net_exit vti6_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) -{ - struct vti6_net *ip6n; - struct net *net; - - ASSERT_RTNL(); - list_for_each_entry(net, net_list, exit_list) { - ip6n = net_generic(net, vti6_net_id); - vti6_destroy_tunnels(ip6n, dev_to_kill); - } -} - static struct pernet_operations vti6_net_ops = { .init = vti6_init_net, - .exit_batch_rtnl = vti6_exit_batch_rtnl, + .exit_rtnl = vti6_exit_rtnl_net, .id = &vti6_net_id, .size = sizeof(struct vti6_net), }; diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c index 9a0f32acb750..a72dbca9e8fc 100644 --- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -1804,8 +1804,7 @@ static struct xfrm_tunnel mplsip_handler __read_mostly = { }; #endif -static void __net_exit sit_destroy_tunnels(struct net *net, - struct list_head *head) +static void __net_exit sit_exit_rtnl_net(struct net *net, struct list_head *head) { struct sit_net *sitn = net_generic(net, sit_net_id); struct net_device *dev, *aux; @@ -1820,15 +1819,15 @@ static void __net_exit sit_destroy_tunnels(struct net *net, for (h = 0; h < (prio ? IP6_SIT_HASH_SIZE : 1); h++) { struct ip_tunnel *t; - t = rtnl_dereference(sitn->tunnels[prio][h]); + t = rtnl_net_dereference(net, sitn->tunnels[prio][h]); while (t) { /* If dev is in the same netns, it has already * been added to the list by the previous loop. */ if (!net_eq(dev_net(t->dev), net)) - unregister_netdevice_queue(t->dev, - head); - t = rtnl_dereference(t->next); + unregister_netdevice_queue(t->dev, head); + + t = rtnl_net_dereference(net, t->next); } } } @@ -1881,19 +1880,9 @@ err_alloc_dev: return err; } -static void __net_exit sit_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) -{ - struct net *net; - - ASSERT_RTNL(); - list_for_each_entry(net, net_list, exit_list) - sit_destroy_tunnels(net, dev_to_kill); -} - static struct pernet_operations sit_net_ops = { .init = sit_init_net, - .exit_batch_rtnl = sit_exit_batch_rtnl, + .exit_rtnl = sit_exit_rtnl_net, .id = &sit_net_id, .size = sizeof(struct sit_net), }; -- 2.50.1 From 9571ab5a98fefa2698f52adfb2823eaac5ca8ce2 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:37 -0700 Subject: [PATCH 04/16] xfrm: Convert xfrmi_exit_batch_rtnl() to ->exit_rtnl(). xfrmi_exit_batch_rtnl() iterates the dying netns list and performs the same operations for each. Let's use ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-9-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- net/xfrm/xfrm_interface_core.c | 34 +++++++++++++++------------------- 1 file changed, 15 insertions(+), 19 deletions(-) diff --git a/net/xfrm/xfrm_interface_core.c b/net/xfrm/xfrm_interface_core.c index 622445f041d3..cb1e12740c87 100644 --- a/net/xfrm/xfrm_interface_core.c +++ b/net/xfrm/xfrm_interface_core.c @@ -952,32 +952,28 @@ static struct rtnl_link_ops xfrmi_link_ops __read_mostly = { .get_link_net = xfrmi_get_link_net, }; -static void __net_exit xfrmi_exit_batch_rtnl(struct list_head *net_exit_list, - struct list_head *dev_to_kill) +static void __net_exit xfrmi_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - struct net *net; + struct xfrmi_net *xfrmn = net_generic(net, xfrmi_net_id); + struct xfrm_if __rcu **xip; + struct xfrm_if *xi; + int i; - ASSERT_RTNL(); - list_for_each_entry(net, net_exit_list, exit_list) { - struct xfrmi_net *xfrmn = net_generic(net, xfrmi_net_id); - struct xfrm_if __rcu **xip; - struct xfrm_if *xi; - int i; - - for (i = 0; i < XFRMI_HASH_SIZE; i++) { - for (xip = &xfrmn->xfrmi[i]; - (xi = rtnl_dereference(*xip)) != NULL; - xip = &xi->next) - unregister_netdevice_queue(xi->dev, dev_to_kill); - } - xi = rtnl_dereference(xfrmn->collect_md_xfrmi); - if (xi) + for (i = 0; i < XFRMI_HASH_SIZE; i++) { + for (xip = &xfrmn->xfrmi[i]; + (xi = rtnl_net_dereference(net, *xip)) != NULL; + xip = &xi->next) unregister_netdevice_queue(xi->dev, dev_to_kill); } + + xi = rtnl_net_dereference(net, xfrmn->collect_md_xfrmi); + if (xi) + unregister_netdevice_queue(xi->dev, dev_to_kill); } static struct pernet_operations xfrmi_net_ops = { - .exit_batch_rtnl = xfrmi_exit_batch_rtnl, + .exit_rtnl = xfrmi_exit_rtnl, .id = &xfrmi_net_id, .size = sizeof(struct xfrmi_net), }; -- 2.50.1 From b7924f50be1525543d261a209790c86240b0df97 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:38 -0700 Subject: [PATCH 05/16] bridge: Convert br_net_exit_batch_rtnl() to ->exit_rtnl(). br_net_exit_batch_rtnl() iterates the dying netns list and performs the same operation for each. Let's use ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Ido Schimmel Acked-by: Nikolay Aleksandrov Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-10-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- net/bridge/br.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) diff --git a/net/bridge/br.c b/net/bridge/br.c index 25dda554ca5b..0adeafe11a36 100644 --- a/net/bridge/br.c +++ b/net/bridge/br.c @@ -368,21 +368,20 @@ void br_opt_toggle(struct net_bridge *br, enum net_bridge_opts opt, bool on) clear_bit(opt, &br->options); } -static void __net_exit br_net_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) +static void __net_exit br_net_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { struct net_device *dev; - struct net *net; - ASSERT_RTNL(); - list_for_each_entry(net, net_list, exit_list) - for_each_netdev(net, dev) - if (netif_is_bridge_master(dev)) - br_dev_delete(dev, dev_to_kill); + ASSERT_RTNL_NET(net); + + for_each_netdev(net, dev) + if (netif_is_bridge_master(dev)) + br_dev_delete(dev, dev_to_kill); } static struct pernet_operations br_net_ops = { - .exit_batch_rtnl = br_net_exit_batch_rtnl, + .exit_rtnl = br_net_exit_rtnl, }; static const struct stp_proto br_stp_proto = { -- 2.50.1 From baf720334c02f07f434feb6b6977ecef20cd58a1 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:39 -0700 Subject: [PATCH 06/16] bonding: Convert bond_net_exit_batch_rtnl() to ->exit_rtnl(). bond_net_exit_batch_rtnl() iterates the dying netns list and performs the same operation for each. Let's use ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Nikolay Aleksandrov Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-11-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- drivers/net/bonding/bond_main.c | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 950d8e4d86f8..bdd36409dd9b 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -6558,7 +6558,7 @@ static int __net_init bond_net_init(struct net *net) /* According to commit 69b0216ac255 ("bonding: fix bonding_masters * race condition in bond unloading") we need to remove sysfs files - * before we remove our devices (done later in bond_net_exit_batch_rtnl()) + * before we remove our devices (done later in bond_net_exit_rtnl()) */ static void __net_exit bond_net_pre_exit(struct net *net) { @@ -6567,25 +6567,20 @@ static void __net_exit bond_net_pre_exit(struct net *net) bond_destroy_sysfs(bn); } -static void __net_exit bond_net_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_kill_list) +static void __net_exit bond_net_exit_rtnl(struct net *net, + struct list_head *dev_kill_list) { - struct bond_net *bn; - struct net *net; + struct bond_net *bn = net_generic(net, bond_net_id); + struct bonding *bond, *tmp_bond; /* Kill off any bonds created after unregistering bond rtnl ops */ - list_for_each_entry(net, net_list, exit_list) { - struct bonding *bond, *tmp_bond; - - bn = net_generic(net, bond_net_id); - list_for_each_entry_safe(bond, tmp_bond, &bn->dev_list, bond_list) - unregister_netdevice_queue(bond->dev, dev_kill_list); - } + list_for_each_entry_safe(bond, tmp_bond, &bn->dev_list, bond_list) + unregister_netdevice_queue(bond->dev, dev_kill_list); } /* According to commit 23fa5c2caae0 ("bonding: destroy proc directory * only after all bonds are gone") bond_destroy_proc_dir() is called - * after bond_net_exit_batch_rtnl() has completed. + * after bond_net_exit_rtnl() has completed. */ static void __net_exit bond_net_exit_batch(struct list_head *net_list) { @@ -6601,7 +6596,7 @@ static void __net_exit bond_net_exit_batch(struct list_head *net_list) static struct pernet_operations bond_net_ops = { .init = bond_net_init, .pre_exit = bond_net_pre_exit, - .exit_batch_rtnl = bond_net_exit_batch_rtnl, + .exit_rtnl = bond_net_exit_rtnl, .exit_batch = bond_net_exit_batch, .id = &bond_net_id, .size = sizeof(struct bond_net), -- 2.50.1 From bc7eaf7a40bba2172af3033fbece13af206c97fe Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:40 -0700 Subject: [PATCH 07/16] gtp: Convert gtp_net_exit_batch_rtnl() to ->exit_rtnl(). gtp_net_exit_batch_rtnl() iterates the dying netns list and performs the same operations for each. Let's use ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-12-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- drivers/net/gtp.c | 18 +++++++----------- 1 file changed, 7 insertions(+), 11 deletions(-) diff --git a/drivers/net/gtp.c b/drivers/net/gtp.c index ef793607890d..d4dec741c7f4 100644 --- a/drivers/net/gtp.c +++ b/drivers/net/gtp.c @@ -2475,23 +2475,19 @@ static int __net_init gtp_net_init(struct net *net) return 0; } -static void __net_exit gtp_net_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) +static void __net_exit gtp_net_exit_rtnl(struct net *net, + struct list_head *dev_to_kill) { - struct net *net; - - list_for_each_entry(net, net_list, exit_list) { - struct gtp_net *gn = net_generic(net, gtp_net_id); - struct gtp_dev *gtp, *gtp_next; + struct gtp_net *gn = net_generic(net, gtp_net_id); + struct gtp_dev *gtp, *gtp_next; - list_for_each_entry_safe(gtp, gtp_next, &gn->gtp_dev_list, list) - gtp_dellink(gtp->dev, dev_to_kill); - } + list_for_each_entry_safe(gtp, gtp_next, &gn->gtp_dev_list, list) + gtp_dellink(gtp->dev, dev_to_kill); } static struct pernet_operations gtp_net_ops = { .init = gtp_net_init, - .exit_batch_rtnl = gtp_net_exit_batch_rtnl, + .exit_rtnl = gtp_net_exit_rtnl, .id = >p_net_id, .size = sizeof(struct gtp_net), }; -- 2.50.1 From fc3dc33f668c12285d82ad7a57b9dc3ca2922f8e Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:41 -0700 Subject: [PATCH 08/16] bareudp: Convert bareudp_exit_batch_rtnl() to ->exit_rtnl(). bareudp_exit_batch_rtnl() iterates the dying netns list and performs the same operation for each. Let's use ->exit_rtnl(). While at it, we replace unregister_netdevice_queue() with bareudp_dellink() for better cleanup. It unlinks the device from net_generic(net, bareudp_net_id)->bareudp_list, but there is no real issue as both the dev and the list are freed later. Signed-off-by: Kuniyuki Iwashima Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-13-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- drivers/net/bareudp.c | 16 ++++------------ 1 file changed, 4 insertions(+), 12 deletions(-) diff --git a/drivers/net/bareudp.c b/drivers/net/bareudp.c index d1473c5f8eef..a9dffdcac805 100644 --- a/drivers/net/bareudp.c +++ b/drivers/net/bareudp.c @@ -777,27 +777,19 @@ static __net_init int bareudp_init_net(struct net *net) return 0; } -static void bareudp_destroy_tunnels(struct net *net, struct list_head *head) +static void __net_exit bareudp_exit_rtnl_net(struct net *net, + struct list_head *dev_kill_list) { struct bareudp_net *bn = net_generic(net, bareudp_net_id); struct bareudp_dev *bareudp, *next; list_for_each_entry_safe(bareudp, next, &bn->bareudp_list, next) - unregister_netdevice_queue(bareudp->dev, head); -} - -static void __net_exit bareudp_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_kill_list) -{ - struct net *net; - - list_for_each_entry(net, net_list, exit_list) - bareudp_destroy_tunnels(net, dev_kill_list); + bareudp_dellink(bareudp->dev, dev_kill_list); } static struct pernet_operations bareudp_net_ops = { .init = bareudp_init_net, - .exit_batch_rtnl = bareudp_exit_batch_rtnl, + .exit_rtnl = bareudp_exit_rtnl_net, .id = &bareudp_net_id, .size = sizeof(struct bareudp_net), }; -- 2.50.1 From 4e53b32d74f052458309a10867e642a8aeafd2f5 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:42 -0700 Subject: [PATCH 09/16] geneve: Convert geneve_exit_batch_rtnl() to ->exit_rtnl(). geneve_exit_batch_rtnl() iterates the dying netns list and performs the same operation for each. Let's use ->exit_rtnl(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-14-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- drivers/net/geneve.c | 16 ++++------------ 1 file changed, 4 insertions(+), 12 deletions(-) diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c index 66e38ce9cd1d..ffc15a432689 100644 --- a/drivers/net/geneve.c +++ b/drivers/net/geneve.c @@ -1946,22 +1946,14 @@ static __net_init int geneve_init_net(struct net *net) return 0; } -static void geneve_destroy_tunnels(struct net *net, struct list_head *head) +static void __net_exit geneve_exit_rtnl_net(struct net *net, + struct list_head *dev_to_kill) { struct geneve_net *gn = net_generic(net, geneve_net_id); struct geneve_dev *geneve, *next; list_for_each_entry_safe(geneve, next, &gn->geneve_list, next) - geneve_dellink(geneve->dev, head); -} - -static void __net_exit geneve_exit_batch_rtnl(struct list_head *net_list, - struct list_head *dev_to_kill) -{ - struct net *net; - - list_for_each_entry(net, net_list, exit_list) - geneve_destroy_tunnels(net, dev_to_kill); + geneve_dellink(geneve->dev, dev_to_kill); } static void __net_exit geneve_exit_net(struct net *net) @@ -1973,7 +1965,7 @@ static void __net_exit geneve_exit_net(struct net *net) static struct pernet_operations geneve_net_ops = { .init = geneve_init_net, - .exit_batch_rtnl = geneve_exit_batch_rtnl, + .exit_rtnl = geneve_exit_rtnl_net, .exit = geneve_exit_net, .id = &geneve_net_id, .size = sizeof(struct geneve_net), -- 2.50.1 From c57a9c503543cd8829eeaaf88362199e0491c0d7 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 11 Apr 2025 13:52:43 -0700 Subject: [PATCH 10/16] net: Remove ->exit_batch_rtnl(). There are no ->exit_batch_rtnl() users remaining. Let's remove the hook. Signed-off-by: Kuniyuki Iwashima Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250411205258.63164-15-kuniyu@amazon.com Signed-off-by: Jakub Kicinski --- include/net/net_namespace.h | 2 -- net/core/net_namespace.c | 8 +------- 2 files changed, 1 insertion(+), 9 deletions(-) diff --git a/include/net/net_namespace.h b/include/net/net_namespace.h index b071e6eed9d5..025a7574b275 100644 --- a/include/net/net_namespace.h +++ b/include/net/net_namespace.h @@ -477,8 +477,6 @@ struct pernet_operations { /* Following method is called with RTNL held. */ void (*exit_rtnl)(struct net *net, struct list_head *dev_kill_list); - void (*exit_batch_rtnl)(struct list_head *net_exit_list, - struct list_head *dev_kill_list); unsigned int * const id; const size_t size; }; diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c index afaa3d1bda8d..0a2b24af4028 100644 --- a/net/core/net_namespace.c +++ b/net/core/net_namespace.c @@ -185,12 +185,6 @@ static void ops_exit_rtnl_list(const struct list_head *ops_list, __rtnl_net_unlock(net); } - ops = saved_ops; - list_for_each_entry_continue_reverse(ops, ops_list, list) { - if (ops->exit_batch_rtnl) - ops->exit_batch_rtnl(net_exit_list, &dev_kill_list); - } - unregister_netdevice_many(&dev_kill_list); rtnl_unlock(); @@ -263,7 +257,7 @@ static void ops_undo_list(const struct list_head *ops_list, static void ops_undo_single(struct pernet_operations *ops, struct list_head *net_exit_list) { - bool hold_rtnl = ops->exit_rtnl || ops->exit_batch_rtnl; + bool hold_rtnl = !!ops->exit_rtnl; LIST_HEAD(ops_list); list_add(&ops->list, &ops_list); -- 2.50.1 From dadc3a6be46987d9417732e5b59476267bba76ed Mon Sep 17 00:00:00 2001 From: "Russell King (Oracle)" Date: Sat, 12 Apr 2025 09:08:29 +0100 Subject: [PATCH 11/16] net: stmmac: dwc-qos: remove tegra_eqos_init() tegra_eqos_init() initialises the 1US TIC counter for the EEE timers. However, the DWGMAC core is reset after this write, which clears this register to its default. However, dwmac4_core_init() configures this register using the same clock, which happens after reset - thus this is the write which ensures that the register is correctly configured. Therefore, tegra_eqos_init() is not required and is removed. This also means eqos->clk_slave can also be removed. Reviewed-by: Andrew Lunn Signed-off-by: Russell King (Oracle) Link: https://patch.msgid.link/E1u3Vuf-000E7g-U4@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski --- .../stmicro/stmmac/dwmac-dwc-qos-eth.c | 24 +------------------ 1 file changed, 1 insertion(+), 23 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c index 5db318327d33..583b5c071cd1 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c @@ -29,7 +29,6 @@ struct tegra_eqos { void __iomem *regs; struct reset_control *rst; - struct clk *clk_slave; struct gpio_desc *reset; }; @@ -199,20 +198,6 @@ static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode) } } -static int tegra_eqos_init(struct platform_device *pdev, void *priv) -{ - struct tegra_eqos *eqos = priv; - unsigned long rate; - u32 value; - - rate = clk_get_rate(eqos->clk_slave); - - value = (rate / 1000000) - 1; - writel(value, eqos->regs + GMAC_1US_TIC_COUNTER); - - return 0; -} - static int tegra_eqos_probe(struct platform_device *pdev, struct plat_stmmacenet_data *plat_dat, struct stmmac_resources *res) @@ -227,7 +212,6 @@ static int tegra_eqos_probe(struct platform_device *pdev, eqos->dev = &pdev->dev; eqos->regs = res->addr; - eqos->clk_slave = plat_dat->stmmac_clk; if (!is_of_node(dev->fwnode)) goto bypass_clk_reset_gpio; @@ -267,17 +251,11 @@ static int tegra_eqos_probe(struct platform_device *pdev, bypass_clk_reset_gpio: plat_dat->fix_mac_speed = tegra_eqos_fix_speed; plat_dat->set_clk_tx_rate = stmmac_set_clk_tx_rate; - plat_dat->init = tegra_eqos_init; plat_dat->bsp_priv = eqos; plat_dat->flags |= STMMAC_FLAG_SPH_DISABLE; - err = tegra_eqos_init(pdev, eqos); - if (err < 0) - goto reset; - return 0; -reset: - reset_control_assert(eqos->rst); + reset_phy: gpiod_set_value(eqos->reset, 1); -- 2.50.1 From 17ec6dbaaed3e2447c8aac93c7161823402aa2d9 Mon Sep 17 00:00:00 2001 From: "Russell King (Oracle)" Date: Sat, 12 Apr 2025 09:08:35 +0100 Subject: [PATCH 12/16] net: stmmac: intel: remove eee_usecs_rate and hardware write Remove the write to GMAC_1US_TIC_COUNTER for two reasons: 1. during initialisation or reinitialisation of the DWMAC core, the core is reset, which sets this register back to its default value. Writing it prior to stmmac_dvr_probe() has no effect. 2. Since commit 8efbdbfa9938 ("net: stmmac: Initialize MAC_ONEUS_TIC_COUNTER register"), GMAC4/5 core code will set this register based on the rate of plat->stmmac_clk. This clock is created by the same code which initialises plat->eee_usecs_rate, which is also created to run at this same rate. Since Marek's commit, this will set this register appropriately using the rate of this clock. Therefore, dwmac-intel.c writing GMAC_1US_TIC_COUNTER serves no useful purpose and can be removed. Reviewed-by: Andrew Lunn Signed-off-by: Russell King (Oracle) Link: https://patch.msgid.link/E1u3Vul-000E7m-1j@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/stmicro/stmmac/dwmac-intel.c | 8 -------- 1 file changed, 8 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-intel.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-intel.c index c8bb9265bbb4..54db5b778304 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-intel.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-intel.c @@ -682,7 +682,6 @@ static int intel_mgbe_common_data(struct pci_dev *pdev, plat->axi->axi_blen[2] = 16; plat->ptp_max_adj = plat->clk_ptp_rate; - plat->eee_usecs_rate = plat->clk_ptp_rate; /* Set system clock */ sprintf(clk_name, "%s-%s", "stmmac", pci_name(pdev)); @@ -1313,13 +1312,6 @@ static int intel_eth_pci_probe(struct pci_dev *pdev, memset(&res, 0, sizeof(res)); res.addr = pcim_iomap_table(pdev)[0]; - if (plat->eee_usecs_rate > 0) { - u32 tx_lpi_usec; - - tx_lpi_usec = (plat->eee_usecs_rate / 1000000) - 1; - writel(tx_lpi_usec, res.addr + GMAC_1US_TIC_COUNTER); - } - ret = stmmac_config_multi_msi(pdev, plat, &res); if (ret) { ret = stmmac_config_single_msi(pdev, plat, &res); -- 2.50.1 From 35031c6256f19773b7e46df80ef4e6d6b782b088 Mon Sep 17 00:00:00 2001 From: "Russell King (Oracle)" Date: Sat, 12 Apr 2025 09:08:40 +0100 Subject: [PATCH 13/16] net: stmmac: intel-plat: remove eee_usecs_rate and hardware write Remove the write to GMAC_1US_TIC_COUNTER for two reasons: 1. during initialisation or reinitialisation of the DWMAC core, the core is reset, which sets this register back to its default value. Writing it prior to stmmac_dvr_probe() has no effect. 2. Since commit 8efbdbfa9938 ("net: stmmac: Initialize MAC_ONEUS_TIC_COUNTER register"), GMAC4/5 core code will set this register based on the rate of plat->stmmac_clk. This clock is fetched by devm_stmmac_probe_config_dt(), and plat->clk_ptp_rate will be set to its rate profided a "ptp_ref" clock is not provided. In any case, Marek's commit will set the effectual value of this register. Therefore, dwmac-intel-plat.c writing GMAC_1US_TIC_COUNTER serves no useful purpose and can be removed. Reviewed-by: Andrew Lunn Signed-off-by: Russell King (Oracle) Link: https://patch.msgid.link/E1u3Vuq-000E7s-5Y@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/stmicro/stmmac/dwmac-intel-plat.c | 9 --------- 1 file changed, 9 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-intel-plat.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-intel-plat.c index 599def7b3a64..4ea7b0a803d7 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-intel-plat.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-intel-plat.c @@ -113,16 +113,7 @@ static int intel_eth_plat_probe(struct platform_device *pdev) plat_dat->clk_tx_i = dwmac->tx_clk; plat_dat->set_clk_tx_rate = stmmac_set_clk_tx_rate; - plat_dat->bsp_priv = dwmac; - plat_dat->eee_usecs_rate = plat_dat->clk_ptp_rate; - - if (plat_dat->eee_usecs_rate > 0) { - u32 tx_lpi_usec; - - tx_lpi_usec = (plat_dat->eee_usecs_rate / 1000000) - 1; - writel(tx_lpi_usec, stmmac_res.addr + GMAC_1US_TIC_COUNTER); - } ret = stmmac_dvr_probe(&pdev->dev, plat_dat, &stmmac_res); if (ret) -- 2.50.1 From 651f88cb046c5e002f7c11de2cebf207787d2346 Mon Sep 17 00:00:00 2001 From: "Russell King (Oracle)" Date: Sat, 12 Apr 2025 09:08:45 +0100 Subject: [PATCH 14/16] net: stmmac: remove eee_usecs_rate plat_dat->eee_users_rate is now unused, so remove this member. Reviewed-by: Andrew Lunn Signed-off-by: Russell King (Oracle) Link: https://patch.msgid.link/E1u3Vuv-000E7y-9k@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski --- include/linux/stmmac.h | 1 - 1 file changed, 1 deletion(-) diff --git a/include/linux/stmmac.h b/include/linux/stmmac.h index c4ec8bb8144e..8aed09d65b4a 100644 --- a/include/linux/stmmac.h +++ b/include/linux/stmmac.h @@ -276,7 +276,6 @@ struct plat_stmmacenet_data { int mac_port_sel_speed; int has_xgmac; u8 vlan_fail_q; - unsigned long eee_usecs_rate; struct pci_dev *pdev; int int_snapshot_num; int msi_mac_vec; -- 2.50.1 From 25af74ed68c4e3de35f5314fe4f8e2853a7e79ac Mon Sep 17 00:00:00 2001 From: "Russell King (Oracle)" Date: Sat, 12 Apr 2025 09:08:50 +0100 Subject: [PATCH 15/16] net: stmmac: remove GMAC_1US_TIC_COUNTER definition GMAC_1US_TIC_COUNTER is now no longer used, so remove the definition. This was duplicated by GMAC4_MAC_ONEUS_TIC_COUNTER further down in the same file. Reviewed-by: Andrew Lunn Signed-off-by: Russell King (Oracle) Link: https://patch.msgid.link/E1u3Vv0-000E87-DQ@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/stmicro/stmmac/dwmac4.h | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h index 42fe29a4e300..5f387ec27c8c 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4.h +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4.h @@ -31,7 +31,6 @@ #define GMAC_RXQ_CTRL3 0x000000ac #define GMAC_INT_STATUS 0x000000b0 #define GMAC_INT_EN 0x000000b4 -#define GMAC_1US_TIC_COUNTER 0x000000dc #define GMAC_PCS_BASE 0x000000e0 #define GMAC_PHYIF_CONTROL_STATUS 0x000000f8 #define GMAC_PMT 0x000000c0 -- 2.50.1 From 4129a75a76a755c151be1ebc0e7d618a4d21e41f Mon Sep 17 00:00:00 2001 From: Oleksij Rempel Date: Thu, 10 Apr 2025 14:42:49 +0200 Subject: [PATCH 16/16] net: dsa: microchip: add ETS scheduler support for KSZ88x3 switches Implement Enhanced Transmission Selection scheduler (ETS) support for KSZ88x3 devices, which support two fixed egress scheduling modes: Strict Priority and Weighted Fair Queuing (WFQ). Since the switch does not allow remapping priorities to queues or adjusting weights, this implementation only supports enabling strict priority mode. If strict mode is not explicitly requested, the switch falls back to its default WFQ mode. This patch introduces KSZ88x3-specific handlers for ETS add and delete operations and uses TXQ Split Control registers to toggle the WFQ enable bit per queue. Corresponding macros are also added for register access. Signed-off-by: Oleksij Rempel Link: https://patch.msgid.link/20250410124249.2728568-1-o.rempel@pengutronix.de Signed-off-by: Jakub Kicinski --- drivers/net/dsa/microchip/ksz_common.c | 97 +++++++++++++++++++++++++- drivers/net/dsa/microchip/ksz_common.h | 19 +++++ 2 files changed, 113 insertions(+), 3 deletions(-) diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c index 89f0796894af..b45052497f8a 100644 --- a/drivers/net/dsa/microchip/ksz_common.c +++ b/drivers/net/dsa/microchip/ksz_common.c @@ -3999,6 +3999,89 @@ static int ksz_ets_band_to_queue(struct tc_ets_qopt_offload_replace_params *p, return p->bands - 1 - band; } +/** + * ksz88x3_tc_ets_add - Configure ETS (Enhanced Transmission Selection) + * for a port on KSZ88x3 switch + * @dev: Pointer to the KSZ switch device structure + * @port: Port number to configure + * @p: Pointer to offload replace parameters describing ETS bands and mapping + * + * The KSZ88x3 supports two scheduling modes: Strict Priority and + * Weighted Fair Queuing (WFQ). Both modes have fixed behavior: + * - No configurable queue-to-priority mapping + * - No weight adjustment in WFQ mode + * + * This function configures the switch to use strict priority mode by + * clearing the WFQ enable bit for all queues associated with ETS bands. + * If strict priority is not explicitly requested, the switch will default + * to WFQ mode. + * + * Return: 0 on success, or a negative error code on failure + */ +static int ksz88x3_tc_ets_add(struct ksz_device *dev, int port, + struct tc_ets_qopt_offload_replace_params *p) +{ + int ret, band; + + /* Only strict priority mode is supported for now. + * WFQ is implicitly enabled when strict mode is disabled. + */ + for (band = 0; band < p->bands; band++) { + int queue = ksz_ets_band_to_queue(p, band); + u8 reg; + + /* Calculate TXQ Split Control register address for this + * port/queue + */ + reg = KSZ8873_TXQ_SPLIT_CTRL_REG(port, queue); + + /* Clear WFQ enable bit to select strict priority scheduling */ + ret = ksz_rmw8(dev, reg, KSZ8873_TXQ_WFQ_ENABLE, 0); + if (ret) + return ret; + } + + return 0; +} + +/** + * ksz88x3_tc_ets_del - Reset ETS (Enhanced Transmission Selection) config + * for a port on KSZ88x3 switch + * @dev: Pointer to the KSZ switch device structure + * @port: Port number to reset + * + * The KSZ88x3 supports only fixed scheduling modes: Strict Priority or + * Weighted Fair Queuing (WFQ), with no reconfiguration of weights or + * queue mapping. This function resets the port’s scheduling mode to + * the default, which is WFQ, by enabling the WFQ bit for all queues. + * + * Return: 0 on success, or a negative error code on failure + */ +static int ksz88x3_tc_ets_del(struct ksz_device *dev, int port) +{ + int ret, queue; + + /* Iterate over all transmit queues for this port */ + for (queue = 0; queue < dev->info->num_tx_queues; queue++) { + u8 reg; + + /* Calculate TXQ Split Control register address for this + * port/queue + */ + reg = KSZ8873_TXQ_SPLIT_CTRL_REG(port, queue); + + /* Set WFQ enable bit to revert back to default scheduling + * mode + */ + ret = ksz_rmw8(dev, reg, KSZ8873_TXQ_WFQ_ENABLE, + KSZ8873_TXQ_WFQ_ENABLE); + if (ret) + return ret; + } + + return 0; +} + static int ksz_queue_set_strict(struct ksz_device *dev, int port, int queue) { int ret; @@ -4080,6 +4163,7 @@ static int ksz_tc_ets_del(struct ksz_device *dev, int port) for (queue = 0; queue < dev->info->num_tx_queues; queue++) { ret = ksz_queue_set_wrr(dev, port, queue, KSZ9477_DEFAULT_WRR_WEIGHT); + if (ret) return ret; } @@ -4132,7 +4216,7 @@ static int ksz_tc_setup_qdisc_ets(struct dsa_switch *ds, int port, struct ksz_device *dev = ds->priv; int ret; - if (is_ksz8(dev)) + if (is_ksz8(dev) && !ksz_is_ksz88x3(dev)) return -EOPNOTSUPP; if (qopt->parent != TC_H_ROOT) { @@ -4146,9 +4230,16 @@ static int ksz_tc_setup_qdisc_ets(struct dsa_switch *ds, int port, if (ret) return ret; - return ksz_tc_ets_add(dev, port, &qopt->replace_params); + if (ksz_is_ksz88x3(dev)) + return ksz88x3_tc_ets_add(dev, port, + &qopt->replace_params); + else + return ksz_tc_ets_add(dev, port, &qopt->replace_params); case TC_ETS_DESTROY: - return ksz_tc_ets_del(dev, port); + if (ksz_is_ksz88x3(dev)) + return ksz88x3_tc_ets_del(dev, port); + else + return ksz_tc_ets_del(dev, port); case TC_ETS_STATS: case TC_ETS_GRAFT: return -EOPNOTSUPP; diff --git a/drivers/net/dsa/microchip/ksz_common.h b/drivers/net/dsa/microchip/ksz_common.h index af17a9c030d4..dd5429ff16ee 100644 --- a/drivers/net/dsa/microchip/ksz_common.h +++ b/drivers/net/dsa/microchip/ksz_common.h @@ -836,6 +836,25 @@ static inline bool is_lan937x_tx_phy(struct ksz_device *dev, int port) #define SW_HI_SPEED_DRIVE_STRENGTH_S 4 #define SW_LO_SPEED_DRIVE_STRENGTH_S 0 +/* TXQ Split Control Register for per-port, per-queue configuration. + * Register 0xAF is TXQ Split for Q3 on Port 1. + * Register offset formula: 0xAF + (port * 4) + (3 - queue) + * where: port = 0..2, queue = 0..3 + */ +#define KSZ8873_TXQ_SPLIT_CTRL_REG(port, queue) \ + (0xAF + ((port) * 4) + (3 - (queue))) + +/* Bit 7 selects between: + * 0 = Strict priority mode (highest-priority queue first) + * 1 = Weighted Fair Queuing (WFQ) mode: + * Queue weights: Q3:Q2:Q1:Q0 = 8:4:2:1 + * If any queues are empty, weight is redistributed. + * + * Note: This is referred to as "Weighted Fair Queuing" (WFQ) in KSZ8863/8873 + * documentation, and as "Weighted Round Robin" (WRR) in KSZ9477 family docs. + */ +#define KSZ8873_TXQ_WFQ_ENABLE BIT(7) + #define KSZ9477_REG_PORT_OUT_RATE_0 0x0420 #define KSZ9477_OUT_RATE_NO_LIMIT 0 -- 2.50.1