ipv6: Split ip6_route_info_create().

We will get rid of RTNL from RTM_NEWROUTE and SIOCADDRT and rely
on RCU to guarantee dev and nexthop lifetime.

Then, we want to allocate as much as possible before entering
the RCU section.

The RCU section will start in the middle of ip6_route_info_create(),
and this is problematic for ip6_route_multipath_add() that calls
ip6_route_info_create() multiple times.

Let's split ip6_route_info_create() into two parts; one for memory
allocation and another for nexthop setup.

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Link: https://patch.msgid.link/20250418000443.43734-7-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

1 files changed, 62 insertions, 33 deletions

diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index f66f90f8f153..bba35cb2dc25 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -3729,15 +3729,13 @@ void fib6_nh_release_dsts(struct fib6_nh *fib6_nh)
 }
 
 static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
-					      gfp_t gfp_flags,
-					      struct netlink_ext_ack *extack)
+					       gfp_t gfp_flags,
+					       struct netlink_ext_ack *extack)
 {
 	struct net *net = cfg->fc_nlinfo.nl_net;
-	struct fib6_info *rt = NULL;
 	struct fib6_table *table;
-	struct fib6_nh *fib6_nh;
-	int err = -ENOBUFS;
-	int addr_type;
+	struct fib6_info *rt;
+	int err;
 
 	if (cfg->fc_nlinfo.nlh &&
 	    !(cfg->fc_nlinfo.nlh->nlmsg_flags & NLM_F_CREATE)) {
@@ -3749,22 +3747,22 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 	} else {
 		table = fib6_new_table(net, cfg->fc_table);
 	}
+	if (!table) {
+		err = -ENOBUFS;
+		goto err;
+	}
 
-	if (!table)
-		goto out;
-
-	err = -ENOMEM;
 	rt = fib6_info_alloc(gfp_flags, !cfg->fc_nh_id);
-	if (!rt)
-		goto out;
+	if (!rt) {
+		err = -ENOMEM;
+		goto err;
+	}
 
 	rt->fib6_metrics = ip_fib_metrics_init(cfg->fc_mx, cfg->fc_mx_len,
 					       extack);
 	if (IS_ERR(rt->fib6_metrics)) {
 		err = PTR_ERR(rt->fib6_metrics);
-		/* Do not leave garbage there. */
-		rt->fib6_metrics = (struct dst_metrics *)&dst_default_metrics;
-		goto out_free;
+		goto free;
 	}
 
 	if (cfg->fc_flags & RTF_ADDRCONF)
@@ -3772,12 +3770,12 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 
 	if (cfg->fc_flags & RTF_EXPIRES)
 		fib6_set_expires(rt, jiffies +
-				clock_t_to_jiffies(cfg->fc_expires));
+				 clock_t_to_jiffies(cfg->fc_expires));
 
 	if (cfg->fc_protocol == RTPROT_UNSPEC)
 		cfg->fc_protocol = RTPROT_BOOT;
-	rt->fib6_protocol = cfg->fc_protocol;
 
+	rt->fib6_protocol = cfg->fc_protocol;
 	rt->fib6_table = table;
 	rt->fib6_metric = cfg->fc_metric;
 	rt->fib6_type = cfg->fc_type ? : RTN_UNICAST;
@@ -3790,6 +3788,20 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 	ipv6_addr_prefix(&rt->fib6_src.addr, &cfg->fc_src, cfg->fc_src_len);
 	rt->fib6_src.plen = cfg->fc_src_len;
 #endif
+	return rt;
+free:
+	kfree(rt);
+err:
+	return ERR_PTR(err);
+}
+
+static int ip6_route_info_create_nh(struct fib6_info *rt,
+				    struct fib6_config *cfg,
+				    struct netlink_ext_ack *extack)
+{
+	struct net *net = cfg->fc_nlinfo.nl_net;
+	struct fib6_nh *fib6_nh;
+	int err;
 
 	if (cfg->fc_nh_id) {
 		struct nexthop *nh;
@@ -3814,9 +3826,11 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 		rt->nh = nh;
 		fib6_nh = nexthop_fib6_nh(rt->nh);
 	} else {
-		err = fib6_nh_init(net, rt->fib6_nh, cfg, gfp_flags, extack);
+		int addr_type;
+
+		err = fib6_nh_init(net, rt->fib6_nh, cfg, GFP_ATOMIC, extack);
 		if (err)
-			goto out;
+			goto out_release;
 
 		fib6_nh = rt->fib6_nh;
 
@@ -3835,21 +3849,20 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 		if (!ipv6_chk_addr(net, &cfg->fc_prefsrc, dev, 0)) {
 			NL_SET_ERR_MSG(extack, "Invalid source address");
 			err = -EINVAL;
-			goto out;
+			goto out_release;
 		}
 		rt->fib6_prefsrc.addr = cfg->fc_prefsrc;
 		rt->fib6_prefsrc.plen = 128;
-	} else
-		rt->fib6_prefsrc.plen = 0;
+	}
 
-	return rt;
-out:
+	return 0;
+out_release:
 	fib6_info_release(rt);
-	return ERR_PTR(err);
+	return err;
 out_free:
 	ip_fib_metrics_put(rt->fib6_metrics);
 	kfree(rt);
-	return ERR_PTR(err);
+	return err;
 }
 
 int ip6_route_add(struct fib6_config *cfg, gfp_t gfp_flags,
@@ -3862,6 +3875,10 @@ int ip6_route_add(struct fib6_config *cfg, gfp_t gfp_flags,
 	if (IS_ERR(rt))
 		return PTR_ERR(rt);
 
+	err = ip6_route_info_create_nh(rt, cfg, extack);
+	if (err)
+		return err;
+
 	err = __ip6_ins_rt(rt, &cfg->fc_nlinfo, extack);
 	fib6_info_release(rt);
 
@@ -4585,6 +4602,7 @@ struct fib6_info *addrconf_f6i_alloc(struct net *net,
 		.fc_ignore_dev_down = true,
 	};
 	struct fib6_info *f6i;
+	int err;
 
 	if (anycast) {
 		cfg.fc_type = RTN_ANYCAST;
@@ -4595,14 +4613,19 @@ struct fib6_info *addrconf_f6i_alloc(struct net *net,
 	}
 
 	f6i = ip6_route_info_create(&cfg, gfp_flags, extack);
-	if (!IS_ERR(f6i)) {
-		f6i->dst_nocount = true;
+	if (IS_ERR(f6i))
+		return f6i;
 
-		if (!anycast &&
-		    (READ_ONCE(net->ipv6.devconf_all->disable_policy) ||
-		     READ_ONCE(idev->cnf.disable_policy)))
-			f6i->dst_nopolicy = true;
-	}
+	err = ip6_route_info_create_nh(f6i, &cfg, extack);
+	if (err)
+		return ERR_PTR(err);
+
+	f6i->dst_nocount = true;
+
+	if (!anycast &&
+	    (READ_ONCE(net->ipv6.devconf_all->disable_policy) ||
+	     READ_ONCE(idev->cnf.disable_policy)))
+		f6i->dst_nopolicy = true;
 
 	return f6i;
 }
@@ -5400,6 +5423,12 @@ static int ip6_route_multipath_add(struct fib6_config *cfg,
 			goto cleanup;
 		}
 
+		err = ip6_route_info_create_nh(rt, &r_cfg, extack);
+		if (err) {
+			rt = NULL;
+			goto cleanup;
+		}
+
 		rt->fib6_nh->fib_nh_weight = rtnh->rtnh_hops + 1;
 
 		err = ip6_route_info_append(info->nl_net, &rt6_nh_list,