Merge branch 'rtnetlink-allow-selected-handlers-to-run-without-rtnl'

Florian Westphal says:

====================
rtnetlink: allow selected handlers to run without rtnl

Changes since v1:
 In patch 6, don't make ipv6 route handlers lockless, they all have
 assumptions on rtnl being held.  Other patches are unchanged.

The RTNL mutex is used to serialize both rtnetlink calls and
dump requests.
Its also used to protect other things such as the list of current
net namespaces.

Unfortunately RTNL mutex is a performance issue, e.g. a cpu adding an
ip address prevents other cpus from seemingly unrelated tasks such as
dumping tc classifiers or doing rtnetlink route lookups.

This patch set adds basic infrastructure to start pushing the rtnl lock
down to those places that need it, or even elide it entirely in some cases.

Subsystems can now indicate that their doit() callback can run without
RTNL mutex, such callbacks can then run in parallel.

This will obviously need a lot of followup work; all current
users need to be audited/changed to benefit from this.
Initial no-rtnl spot is netns new/getid.

We have various 'get' handlers that are also a tempting target,
however, several of these depend on rtnl mutex to prevent information
from changing while objects are being read by rtnl handlers; however,
it doesn't appear impossible to change this.

Dumps are another problem entirely, see
commit 2907c35f ("net: hold rtnl again in dump callbacks"),
this patchset doesn't touch dump requests.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/rtnetlink.h

@@ -7,12 +7,15 @@
 typedef int (*rtnl_doit_func)(struct sk_buff *, struct nlmsghdr *,
 			      struct netlink_ext_ack *);
 typedef int (*rtnl_dumpit_func)(struct sk_buff *, struct netlink_callback *);
-typedef u16 (*rtnl_calcit_func)(struct sk_buff *, struct nlmsghdr *);
+
+enum rtnl_link_flags {
+	RTNL_FLAG_DOIT_UNLOCKED = 1,
+};
 
 int __rtnl_register(int protocol, int msgtype,
-		    rtnl_doit_func, rtnl_dumpit_func, rtnl_calcit_func);
+		    rtnl_doit_func, rtnl_dumpit_func, unsigned int flags);
 void rtnl_register(int protocol, int msgtype,
-		   rtnl_doit_func, rtnl_dumpit_func, rtnl_calcit_func);
+		   rtnl_doit_func, rtnl_dumpit_func, unsigned int flags);
 int rtnl_unregister(int protocol, int msgtype);
 void rtnl_unregister_all(int protocol);
 

net/bridge/br_mdb.c

@@ -713,9 +713,9 @@ static int br_mdb_del(struct sk_buff *skb, struct nlmsghdr *nlh,
 
 void br_mdb_init(void)
 {
-	rtnl_register(PF_BRIDGE, RTM_GETMDB, NULL, br_mdb_dump, NULL);
-	rtnl_register(PF_BRIDGE, RTM_NEWMDB, br_mdb_add, NULL, NULL);
-	rtnl_register(PF_BRIDGE, RTM_DELMDB, br_mdb_del, NULL, NULL);
+	rtnl_register(PF_BRIDGE, RTM_GETMDB, NULL, br_mdb_dump, 0);
+	rtnl_register(PF_BRIDGE, RTM_NEWMDB, br_mdb_add, NULL, 0);
+	rtnl_register(PF_BRIDGE, RTM_DELMDB, br_mdb_del, NULL, 0);
 }
 
 void br_mdb_uninit(void)

net/can/gw.c

@@ -1031,15 +1031,15 @@ static __init int cgw_module_init(void)
 	notifier.notifier_call = cgw_notifier;
 	register_netdevice_notifier(&notifier);
 
-	if (__rtnl_register(PF_CAN, RTM_GETROUTE, NULL, cgw_dump_jobs, NULL)) {
+	if (__rtnl_register(PF_CAN, RTM_GETROUTE, NULL, cgw_dump_jobs, 0)) {
 		unregister_netdevice_notifier(&notifier);
 		kmem_cache_destroy(cgw_cache);
 		return -ENOBUFS;
 	}
 
 	/* Only the first call to __rtnl_register can fail */
-	__rtnl_register(PF_CAN, RTM_NEWROUTE, cgw_create_job, NULL, NULL);
-	__rtnl_register(PF_CAN, RTM_DELROUTE, cgw_remove_job, NULL, NULL);
+	__rtnl_register(PF_CAN, RTM_NEWROUTE, cgw_create_job, NULL, 0);
+	__rtnl_register(PF_CAN, RTM_DELROUTE, cgw_remove_job, NULL, 0);
 
 	return 0;
 }

net/core/fib_rules.c

@@ -1026,9 +1026,9 @@ static struct pernet_operations fib_rules_net_ops = {
 static int __init fib_rules_init(void)
 {
 	int err;
-	rtnl_register(PF_UNSPEC, RTM_NEWRULE, fib_nl_newrule, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_DELRULE, fib_nl_delrule, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_GETRULE, NULL, fib_nl_dumprule, NULL);
+	rtnl_register(PF_UNSPEC, RTM_NEWRULE, fib_nl_newrule, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_DELRULE, fib_nl_delrule, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_GETRULE, NULL, fib_nl_dumprule, 0);
 
 	err = register_pernet_subsys(&fib_rules_net_ops);
 	if (err < 0)

net/core/neighbour.c

@@ -3261,13 +3261,13 @@ EXPORT_SYMBOL(neigh_sysctl_unregister);
 
 static int __init neigh_init(void)
 {
-	rtnl_register(PF_UNSPEC, RTM_NEWNEIGH, neigh_add, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_DELNEIGH, neigh_delete, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_GETNEIGH, NULL, neigh_dump_info, NULL);
+	rtnl_register(PF_UNSPEC, RTM_NEWNEIGH, neigh_add, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_DELNEIGH, neigh_delete, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_GETNEIGH, NULL, neigh_dump_info, 0);
 
 	rtnl_register(PF_UNSPEC, RTM_GETNEIGHTBL, NULL, neightbl_dump_info,
-		      NULL);
-	rtnl_register(PF_UNSPEC, RTM_SETNEIGHTBL, neightbl_set, NULL, NULL);
+		      0);
+	rtnl_register(PF_UNSPEC, RTM_SETNEIGHTBL, neightbl_set, NULL, 0);
 
 	return 0;
 }

net/core/net_namespace.c

@@ -855,9 +855,10 @@ static int __init net_ns_init(void)
 
 	register_pernet_subsys(&net_ns_ops);
 
-	rtnl_register(PF_UNSPEC, RTM_NEWNSID, rtnl_net_newid, NULL, NULL);
+	rtnl_register(PF_UNSPEC, RTM_NEWNSID, rtnl_net_newid, NULL,
+		      RTNL_FLAG_DOIT_UNLOCKED);
 	rtnl_register(PF_UNSPEC, RTM_GETNSID, rtnl_net_getid, rtnl_net_dumpid,
-		      NULL);
+		      RTNL_FLAG_DOIT_UNLOCKED);
 
 	return 0;
 }

net/dcb/dcbnl.c

@@ -1938,8 +1938,8 @@ static int __init dcbnl_init(void)
 {
 	INIT_LIST_HEAD(&dcb_app_list);
 
-	rtnl_register(PF_UNSPEC, RTM_GETDCB, dcb_doit, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_SETDCB, dcb_doit, NULL, NULL);
+	rtnl_register(PF_UNSPEC, RTM_GETDCB, dcb_doit, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_SETDCB, dcb_doit, NULL, 0);
 
 	return 0;
 }

net/decnet/dn_dev.c

@@ -1419,9 +1419,9 @@ void __init dn_dev_init(void)
 
 	dn_dev_devices_on();
 
-	rtnl_register(PF_DECnet, RTM_NEWADDR, dn_nl_newaddr, NULL, NULL);
-	rtnl_register(PF_DECnet, RTM_DELADDR, dn_nl_deladdr, NULL, NULL);
-	rtnl_register(PF_DECnet, RTM_GETADDR, NULL, dn_nl_dump_ifaddr, NULL);
+	rtnl_register(PF_DECnet, RTM_NEWADDR, dn_nl_newaddr, NULL, 0);
+	rtnl_register(PF_DECnet, RTM_DELADDR, dn_nl_deladdr, NULL, 0);
+	rtnl_register(PF_DECnet, RTM_GETADDR, NULL, dn_nl_dump_ifaddr, 0);
 
 	proc_create("decnet_dev", S_IRUGO, init_net.proc_net, &dn_dev_seq_fops);
 

net/decnet/dn_fib.c

@@ -791,8 +791,8 @@ void __init dn_fib_init(void)
 
 	register_dnaddr_notifier(&dn_fib_dnaddr_notifier);
 
-	rtnl_register(PF_DECnet, RTM_NEWROUTE, dn_fib_rtm_newroute, NULL, NULL);
-	rtnl_register(PF_DECnet, RTM_DELROUTE, dn_fib_rtm_delroute, NULL, NULL);
+	rtnl_register(PF_DECnet, RTM_NEWROUTE, dn_fib_rtm_newroute, NULL, 0);
+	rtnl_register(PF_DECnet, RTM_DELROUTE, dn_fib_rtm_delroute, NULL, 0);
 }
 
 

net/decnet/dn_route.c

@@ -1922,10 +1922,10 @@ void __init dn_route_init(void)
 
 #ifdef CONFIG_DECNET_ROUTER
 	rtnl_register(PF_DECnet, RTM_GETROUTE, dn_cache_getroute,
-		      dn_fib_dump, NULL);
+		      dn_fib_dump, 0);
 #else
 	rtnl_register(PF_DECnet, RTM_GETROUTE, dn_cache_getroute,
-		      dn_cache_dump, NULL);
+		      dn_cache_dump, 0);
 #endif
 }
 

net/ipv4/devinet.c

@@ -2491,9 +2491,9 @@ void __init devinet_init(void)
 
 	rtnl_af_register(&inet_af_ops);
 
-	rtnl_register(PF_INET, RTM_NEWADDR, inet_rtm_newaddr, NULL, NULL);
-	rtnl_register(PF_INET, RTM_DELADDR, inet_rtm_deladdr, NULL, NULL);
-	rtnl_register(PF_INET, RTM_GETADDR, NULL, inet_dump_ifaddr, NULL);
+	rtnl_register(PF_INET, RTM_NEWADDR, inet_rtm_newaddr, NULL, 0);
+	rtnl_register(PF_INET, RTM_DELADDR, inet_rtm_deladdr, NULL, 0);
+	rtnl_register(PF_INET, RTM_GETADDR, NULL, inet_dump_ifaddr, 0);
 	rtnl_register(PF_INET, RTM_GETNETCONF, inet_netconf_get_devconf,
-		      inet_netconf_dump_devconf, NULL);
+		      inet_netconf_dump_devconf, 0);
 }

net/ipv4/fib_frontend.c

@@ -1348,7 +1348,7 @@ void __init ip_fib_init(void)
 	register_netdevice_notifier(&fib_netdev_notifier);
 	register_inetaddr_notifier(&fib_inetaddr_notifier);
 
-	rtnl_register(PF_INET, RTM_NEWROUTE, inet_rtm_newroute, NULL, NULL);
-	rtnl_register(PF_INET, RTM_DELROUTE, inet_rtm_delroute, NULL, NULL);
-	rtnl_register(PF_INET, RTM_GETROUTE, NULL, inet_dump_fib, NULL);
+	rtnl_register(PF_INET, RTM_NEWROUTE, inet_rtm_newroute, NULL, 0);
+	rtnl_register(PF_INET, RTM_DELROUTE, inet_rtm_delroute, NULL, 0);
+	rtnl_register(PF_INET, RTM_GETROUTE, NULL, inet_dump_fib, 0);
 }

net/ipv4/ipmr.c

@@ -3114,14 +3114,14 @@ int __init ip_mr_init(void)
 	}
 #endif
 	rtnl_register(RTNL_FAMILY_IPMR, RTM_GETROUTE,
-		      ipmr_rtm_getroute, ipmr_rtm_dumproute, NULL);
+		      ipmr_rtm_getroute, ipmr_rtm_dumproute, 0);
 	rtnl_register(RTNL_FAMILY_IPMR, RTM_NEWROUTE,
-		      ipmr_rtm_route, NULL, NULL);
+		      ipmr_rtm_route, NULL, 0);
 	rtnl_register(RTNL_FAMILY_IPMR, RTM_DELROUTE,
-		      ipmr_rtm_route, NULL, NULL);
+		      ipmr_rtm_route, NULL, 0);
 
 	rtnl_register(RTNL_FAMILY_IPMR, RTM_GETLINK,
-		      NULL, ipmr_rtm_dumplink, NULL);
+		      NULL, ipmr_rtm_dumplink, 0);
 	return 0;
 
 #ifdef CONFIG_IP_PIMSM_V2

net/ipv4/route.c

@@ -3067,7 +3067,7 @@ int __init ip_rt_init(void)
 	xfrm_init();
 	xfrm4_init();
 #endif
-	rtnl_register(PF_INET, RTM_GETROUTE, inet_rtm_getroute, NULL, NULL);
+	rtnl_register(PF_INET, RTM_GETROUTE, inet_rtm_getroute, NULL, 0);
 
 #ifdef CONFIG_SYSCTL
 	register_pernet_subsys(&sysctl_route_ops);

net/ipv6/addrconf.c

@@ -6605,21 +6605,21 @@ int __init addrconf_init(void)
 	rtnl_af_register(&inet6_ops);
 
 	err = __rtnl_register(PF_INET6, RTM_GETLINK, NULL, inet6_dump_ifinfo,
-			      NULL);
+			      0);
 	if (err < 0)
 		goto errout;
 
 	/* Only the first call to __rtnl_register can fail */
-	__rtnl_register(PF_INET6, RTM_NEWADDR, inet6_rtm_newaddr, NULL, NULL);
-	__rtnl_register(PF_INET6, RTM_DELADDR, inet6_rtm_deladdr, NULL, NULL);
+	__rtnl_register(PF_INET6, RTM_NEWADDR, inet6_rtm_newaddr, NULL, 0);
+	__rtnl_register(PF_INET6, RTM_DELADDR, inet6_rtm_deladdr, NULL, 0);
 	__rtnl_register(PF_INET6, RTM_GETADDR, inet6_rtm_getaddr,
-			inet6_dump_ifaddr, NULL);
+			inet6_dump_ifaddr, 0);
 	__rtnl_register(PF_INET6, RTM_GETMULTICAST, NULL,
-			inet6_dump_ifmcaddr, NULL);
+			inet6_dump_ifmcaddr, 0);
 	__rtnl_register(PF_INET6, RTM_GETANYCAST, NULL,
-			inet6_dump_ifacaddr, NULL);
+			inet6_dump_ifacaddr, 0);
 	__rtnl_register(PF_INET6, RTM_GETNETCONF, inet6_netconf_get_devconf,
-			inet6_netconf_dump_devconf, NULL);
+			inet6_netconf_dump_devconf, 0);
 
 	ipv6_addr_label_rtnl_register();
 

net/ipv6/addrlabel.c

@@ -593,10 +593,10 @@ static int ip6addrlbl_get(struct sk_buff *in_skb, struct nlmsghdr *nlh,
 void __init ipv6_addr_label_rtnl_register(void)
 {
 	__rtnl_register(PF_INET6, RTM_NEWADDRLABEL, ip6addrlbl_newdel,
-			NULL, NULL);
+			NULL, 0);
 	__rtnl_register(PF_INET6, RTM_DELADDRLABEL, ip6addrlbl_newdel,
-			NULL, NULL);
+			NULL, 0);
 	__rtnl_register(PF_INET6, RTM_GETADDRLABEL, ip6addrlbl_get,
-			ip6addrlbl_dump, NULL);
+			ip6addrlbl_dump, 0);
 }
 

net/ipv6/ip6_fib.c

@@ -2038,7 +2038,7 @@ int __init fib6_init(void)
 		goto out_kmem_cache_create;
 
 	ret = __rtnl_register(PF_INET6, RTM_GETROUTE, NULL, inet6_dump_fib,
-			      NULL);
+			      0);
 	if (ret)
 		goto out_unregister_subsys;
 

net/ipv6/ip6mr.c

@@ -1427,7 +1427,7 @@ int __init ip6_mr_init(void)
 	}
 #endif
 	rtnl_register(RTNL_FAMILY_IP6MR, RTM_GETROUTE, NULL,
-		      ip6mr_rtm_dumproute, NULL);
+		      ip6mr_rtm_dumproute, 0);
 	return 0;
 #ifdef CONFIG_IPV6_PIMSM_V2
 add_proto_fail:

net/ipv6/route.c

@@ -4105,9 +4105,9 @@ int __init ip6_route_init(void)
 		goto fib6_rules_init;
 
 	ret = -ENOBUFS;
-	if (__rtnl_register(PF_INET6, RTM_NEWROUTE, inet6_rtm_newroute, NULL, NULL) ||
-	    __rtnl_register(PF_INET6, RTM_DELROUTE, inet6_rtm_delroute, NULL, NULL) ||
-	    __rtnl_register(PF_INET6, RTM_GETROUTE, inet6_rtm_getroute, NULL, NULL))
+	if (__rtnl_register(PF_INET6, RTM_NEWROUTE, inet6_rtm_newroute, NULL, 0) ||
+	    __rtnl_register(PF_INET6, RTM_DELROUTE, inet6_rtm_delroute, NULL, 0) ||
+	    __rtnl_register(PF_INET6, RTM_GETROUTE, inet6_rtm_getroute, NULL, 0))
 		goto out_register_late_subsys;
 
 	ret = register_netdevice_notifier(&ip6_route_dev_notifier);

net/mpls/af_mpls.c

@@ -2479,12 +2479,12 @@ static int __init mpls_init(void)
 
 	rtnl_af_register(&mpls_af_ops);
 
-	rtnl_register(PF_MPLS, RTM_NEWROUTE, mpls_rtm_newroute, NULL, NULL);
-	rtnl_register(PF_MPLS, RTM_DELROUTE, mpls_rtm_delroute, NULL, NULL);
+	rtnl_register(PF_MPLS, RTM_NEWROUTE, mpls_rtm_newroute, NULL, 0);
+	rtnl_register(PF_MPLS, RTM_DELROUTE, mpls_rtm_delroute, NULL, 0);
 	rtnl_register(PF_MPLS, RTM_GETROUTE, mpls_getroute, mpls_dump_routes,
-		      NULL);
+		      0);
 	rtnl_register(PF_MPLS, RTM_GETNETCONF, mpls_netconf_get_devconf,
-		      mpls_netconf_dump_devconf, NULL);
+		      mpls_netconf_dump_devconf, 0);
 	err = 0;
 out:
 	return err;

net/phonet/pn_netlink.c

@@ -300,15 +300,15 @@ static int route_dumpit(struct sk_buff *skb, struct netlink_callback *cb)
 int __init phonet_netlink_register(void)
 {
 	int err = __rtnl_register(PF_PHONET, RTM_NEWADDR, addr_doit,
-				  NULL, NULL);
+				  NULL, 0);
 	if (err)
 		return err;
 
 	/* Further __rtnl_register() cannot fail */
-	__rtnl_register(PF_PHONET, RTM_DELADDR, addr_doit, NULL, NULL);
-	__rtnl_register(PF_PHONET, RTM_GETADDR, NULL, getaddr_dumpit, NULL);
-	__rtnl_register(PF_PHONET, RTM_NEWROUTE, route_doit, NULL, NULL);
-	__rtnl_register(PF_PHONET, RTM_DELROUTE, route_doit, NULL, NULL);
-	__rtnl_register(PF_PHONET, RTM_GETROUTE, NULL, route_dumpit, NULL);
+	__rtnl_register(PF_PHONET, RTM_DELADDR, addr_doit, NULL, 0);
+	__rtnl_register(PF_PHONET, RTM_GETADDR, NULL, getaddr_dumpit, 0);
+	__rtnl_register(PF_PHONET, RTM_NEWROUTE, route_doit, NULL, 0);
+	__rtnl_register(PF_PHONET, RTM_DELROUTE, route_doit, NULL, 0);
+	__rtnl_register(PF_PHONET, RTM_GETROUTE, NULL, route_dumpit, 0);
 	return 0;
 }

net/qrtr/qrtr.c

@@ -1081,7 +1081,7 @@ static int __init qrtr_proto_init(void)
 		return rc;
 	}
 
-	rtnl_register(PF_QIPCRTR, RTM_NEWADDR, qrtr_addr_doit, NULL, NULL);
+	rtnl_register(PF_QIPCRTR, RTM_NEWADDR, qrtr_addr_doit, NULL, 0);
 
 	return 0;
 }

net/sched/act_api.c

@@ -1255,10 +1255,10 @@ static int tc_dump_action(struct sk_buff *skb, struct netlink_callback *cb)
 
 static int __init tc_action_init(void)
 {
-	rtnl_register(PF_UNSPEC, RTM_NEWACTION, tc_ctl_action, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_DELACTION, tc_ctl_action, NULL, NULL);
+	rtnl_register(PF_UNSPEC, RTM_NEWACTION, tc_ctl_action, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_DELACTION, tc_ctl_action, NULL, 0);
 	rtnl_register(PF_UNSPEC, RTM_GETACTION, tc_ctl_action, tc_dump_action,
-		      NULL);
+		      0);
 
 	return 0;
 }

net/sched/cls_api.c

@@ -1010,10 +1010,10 @@ EXPORT_SYMBOL(tcf_exts_get_dev);
 
 static int __init tc_filter_init(void)
 {
-	rtnl_register(PF_UNSPEC, RTM_NEWTFILTER, tc_ctl_tfilter, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_DELTFILTER, tc_ctl_tfilter, NULL, NULL);
+	rtnl_register(PF_UNSPEC, RTM_NEWTFILTER, tc_ctl_tfilter, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_DELTFILTER, tc_ctl_tfilter, NULL, 0);
 	rtnl_register(PF_UNSPEC, RTM_GETTFILTER, tc_ctl_tfilter,
-		      tc_dump_tfilter, NULL);
+		      tc_dump_tfilter, 0);
 
 	return 0;
 }

net/sched/sch_api.c

@@ -1952,14 +1952,14 @@ static int __init pktsched_init(void)
 	register_qdisc(&mq_qdisc_ops);
 	register_qdisc(&noqueue_qdisc_ops);
 
-	rtnl_register(PF_UNSPEC, RTM_NEWQDISC, tc_modify_qdisc, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_DELQDISC, tc_get_qdisc, NULL, NULL);
+	rtnl_register(PF_UNSPEC, RTM_NEWQDISC, tc_modify_qdisc, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_DELQDISC, tc_get_qdisc, NULL, 0);
 	rtnl_register(PF_UNSPEC, RTM_GETQDISC, tc_get_qdisc, tc_dump_qdisc,
-		      NULL);
-	rtnl_register(PF_UNSPEC, RTM_NEWTCLASS, tc_ctl_tclass, NULL, NULL);
-	rtnl_register(PF_UNSPEC, RTM_DELTCLASS, tc_ctl_tclass, NULL, NULL);
+		      0);
+	rtnl_register(PF_UNSPEC, RTM_NEWTCLASS, tc_ctl_tclass, NULL, 0);
+	rtnl_register(PF_UNSPEC, RTM_DELTCLASS, tc_ctl_tclass, NULL, 0);
 	rtnl_register(PF_UNSPEC, RTM_GETTCLASS, tc_ctl_tclass, tc_dump_tclass,
-		      NULL);
+		      0);
 
 	return 0;
 }