github-mirror/linux
1
0
Fork 0
mirror of https://github.com/torvalds/linux.git synced 2026-07-27 17:47:41 +02:00
Code Issues Actions Packages Projects Releases Wiki Activity Graph

mpls: Hold dev refcnt for mpls_nh.

Browse Source 
MPLS uses RTNL

  1) to guarantee the lifetime of struct mpls_nh.nh_dev
  2) to protect net->mpls.platform_label

, but neither actually requires RTNL.

If we do not call dev_put() in find_outdev() and call it
just before freeing struct mpls_route, we can drop RTNL for 1).

Let's hold the refcnt of mpls_nh.nh_dev and track it with
netdevice_tracker.

Two notable changes:

If mpls_nh_build_multi() fails to set up a neighbour, we need
to call netdev_put() for successfully created neighbours in
mpls_rt_free_rcu(), so the number of neighbours (rt->rt_nhn)
is now updated in each iteration.

When a dev is unregistered, mpls_ifdown() clones mpls_route
and replaces it with the clone, so the clone requires extra
netdev_hold().

Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Link: https://patch.msgid.link/20251029173344.2934622-3-kuniyu@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This commit is contained in:
Kuniyuki Iwashima 2025-10-29 17:32:54 +00:00 committed by Jakub Kicinski
parent 2214ca1ff6
commit f0914b8436
2 changed files with 45 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

63
net/mpls/af_mpls.c
View File

@ -530,10 +530,23 @@ static struct mpls_route *mpls_rt_alloc(u8 num_nh, u8 max_alen, u8 max_labels)
return rt; return rt;
} }
static void mpls_rt_free_rcu(struct rcu_head *head)
{
struct mpls_route *rt;
rt = container_of(head, struct mpls_route, rt_rcu);
change_nexthops(rt) {
netdev_put(nh->nh_dev, &nh->nh_dev_tracker);
} endfor_nexthops(rt);
kfree(rt);
}
static void mpls_rt_free(struct mpls_route *rt) static void mpls_rt_free(struct mpls_route *rt)
{ {
if (rt) if (rt)
kfree_rcu(rt, rt_rcu); call_rcu(&rt->rt_rcu, mpls_rt_free_rcu);
} }
static void mpls_notify_route(struct net *net, unsigned index, static void mpls_notify_route(struct net *net, unsigned index,
@ -587,6 +600,7 @@ static unsigned find_free_label(struct net *net)
#if IS_ENABLED(CONFIG_INET) #if IS_ENABLED(CONFIG_INET)
static struct net_device *inet_fib_lookup_dev(struct net *net, static struct net_device *inet_fib_lookup_dev(struct net *net,
struct mpls_nh *nh,
const void *addr) const void *addr)
{ {
struct net_device *dev; struct net_device *dev;
@ -599,14 +613,14 @@ static struct net_device *inet_fib_lookup_dev(struct net *net,
return ERR_CAST(rt); return ERR_CAST(rt);
dev = rt->dst.dev; dev = rt->dst.dev;
dev_hold(dev); netdev_hold(dev, &nh->nh_dev_tracker, GFP_KERNEL);
ip_rt_put(rt); ip_rt_put(rt);
return dev; return dev;
} }
#else #else
static struct net_device *inet_fib_lookup_dev(struct net *net, static struct net_device *inet_fib_lookup_dev(struct net *net,
struct mpls_nh *nh,
const void *addr) const void *addr)
{ {
return ERR_PTR(-EAFNOSUPPORT); return ERR_PTR(-EAFNOSUPPORT);
@ -615,6 +629,7 @@ static struct net_device *inet_fib_lookup_dev(struct net *net,
#if IS_ENABLED(CONFIG_IPV6) #if IS_ENABLED(CONFIG_IPV6)
static struct net_device *inet6_fib_lookup_dev(struct net *net, static struct net_device *inet6_fib_lookup_dev(struct net *net,
struct mpls_nh *nh,
const void *addr) const void *addr)
{ {
struct net_device *dev; struct net_device *dev;
@ -631,13 +646,14 @@ static struct net_device *inet6_fib_lookup_dev(struct net *net,
return ERR_CAST(dst); return ERR_CAST(dst);
dev = dst->dev; dev = dst->dev;
dev_hold(dev); netdev_hold(dev, &nh->nh_dev_tracker, GFP_KERNEL);
dst_release(dst); dst_release(dst);
return dev; return dev;
} }
#else #else
static struct net_device *inet6_fib_lookup_dev(struct net *net, static struct net_device *inet6_fib_lookup_dev(struct net *net,
struct mpls_nh *nh,
const void *addr) const void *addr)
{ {
return ERR_PTR(-EAFNOSUPPORT); return ERR_PTR(-EAFNOSUPPORT);
@ -653,16 +669,17 @@ static struct net_device *find_outdev(struct net *net,
if (!oif) { if (!oif) {
switch (nh->nh_via_table) { switch (nh->nh_via_table) {
case NEIGH_ARP_TABLE: case NEIGH_ARP_TABLE:
dev = inet_fib_lookup_dev(net, mpls_nh_via(rt, nh)); dev = inet_fib_lookup_dev(net, nh, mpls_nh_via(rt, nh));
break; break;
case NEIGH_ND_TABLE: case NEIGH_ND_TABLE:
dev = inet6_fib_lookup_dev(net, mpls_nh_via(rt, nh)); dev = inet6_fib_lookup_dev(net, nh, mpls_nh_via(rt, nh));
break; break;
case NEIGH_LINK_TABLE: case NEIGH_LINK_TABLE:
break; break;
} }
} else { } else {
dev = dev_get_by_index(net, oif); dev = netdev_get_by_index(net, oif,
&nh->nh_dev_tracker, GFP_KERNEL);
} }
if (!dev) if (!dev)
@ -671,8 +688,7 @@ static struct net_device *find_outdev(struct net *net,
if (IS_ERR(dev)) if (IS_ERR(dev))
return dev; return dev;
/* The caller is holding rtnl anyways, so release the dev reference */ nh->nh_dev = dev;
dev_put(dev);
return dev; return dev;
} }
@ -686,20 +702,17 @@ static int mpls_nh_assign_dev(struct net *net, struct mpls_route *rt,
dev = find_outdev(net, rt, nh, oif); dev = find_outdev(net, rt, nh, oif);
if (IS_ERR(dev)) { if (IS_ERR(dev)) {
err = PTR_ERR(dev); err = PTR_ERR(dev);
dev = NULL;
goto errout; goto errout;
} }
/* Ensure this is a supported device */ /* Ensure this is a supported device */
err = -EINVAL; err = -EINVAL;
if (!mpls_dev_get(dev)) if (!mpls_dev_get(dev))
goto errout; goto errout_put;
if ((nh->nh_via_table == NEIGH_LINK_TABLE) && if ((nh->nh_via_table == NEIGH_LINK_TABLE) &&
(dev->addr_len != nh->nh_via_alen)) (dev->addr_len != nh->nh_via_alen))
goto errout; goto errout_put;
nh->nh_dev = dev;
if (!(dev->flags & IFF_UP)) { if (!(dev->flags & IFF_UP)) {
nh->nh_flags |= RTNH_F_DEAD; nh->nh_flags |= RTNH_F_DEAD;
@ -713,6 +726,9 @@ static int mpls_nh_assign_dev(struct net *net, struct mpls_route *rt,
return 0; return 0;
errout_put:
netdev_put(nh->nh_dev, &nh->nh_dev_tracker);
nh->nh_dev = NULL;
errout: errout:
return err; return err;
} }
@ -890,7 +906,8 @@ static int mpls_nh_build_multi(struct mpls_route_config *cfg,
struct nlattr *nla_via, *nla_newdst; struct nlattr *nla_via, *nla_newdst;
int remaining = cfg->rc_mp_len; int remaining = cfg->rc_mp_len;
int err = 0; int err = 0;
u8 nhs = 0;
rt->rt_nhn = 0;
change_nexthops(rt) { change_nexthops(rt) {
int attrlen; int attrlen;
@ -926,11 +943,9 @@ static int mpls_nh_build_multi(struct mpls_route_config *cfg,
rt->rt_nhn_alive--; rt->rt_nhn_alive--;
rtnh = rtnh_next(rtnh, &remaining); rtnh = rtnh_next(rtnh, &remaining);
nhs++; rt->rt_nhn++;
} endfor_nexthops(rt); } endfor_nexthops(rt);
rt->rt_nhn = nhs;
return 0; return 0;
errout: errout:
@ -1523,8 +1538,12 @@ static int mpls_ifdown(struct net_device *dev, int event)
change_nexthops(rt) { change_nexthops(rt) {
unsigned int nh_flags = nh->nh_flags; unsigned int nh_flags = nh->nh_flags;
if (nh->nh_dev != dev) if (nh->nh_dev != dev) {
if (nh_del)
netdev_hold(nh->nh_dev, &nh->nh_dev_tracker,
GFP_KERNEL);
goto next; goto next;
}
switch (event) { switch (event) {
case NETDEV_DOWN: case NETDEV_DOWN:
@ -2518,10 +2537,13 @@ static int resize_platform_label_table(struct net *net, size_t limit)
/* In case the predefined labels need to be populated */ /* In case the predefined labels need to be populated */
if (limit > MPLS_LABEL_IPV4NULL) { if (limit > MPLS_LABEL_IPV4NULL) {
struct net_device *lo = net->loopback_dev; struct net_device *lo = net->loopback_dev;
rt0 = mpls_rt_alloc(1, lo->addr_len, 0); rt0 = mpls_rt_alloc(1, lo->addr_len, 0);
if (IS_ERR(rt0)) if (IS_ERR(rt0))
goto nort0; goto nort0;
rt0->rt_nh->nh_dev = lo; rt0->rt_nh->nh_dev = lo;
netdev_hold(lo, &rt0->rt_nh->nh_dev_tracker, GFP_KERNEL);
rt0->rt_protocol = RTPROT_KERNEL; rt0->rt_protocol = RTPROT_KERNEL;
rt0->rt_payload_type = MPT_IPV4; rt0->rt_payload_type = MPT_IPV4;
rt0->rt_ttl_propagate = MPLS_TTL_PROP_DEFAULT; rt0->rt_ttl_propagate = MPLS_TTL_PROP_DEFAULT;
@ -2532,10 +2554,13 @@ static int resize_platform_label_table(struct net *net, size_t limit)
} }
if (limit > MPLS_LABEL_IPV6NULL) { if (limit > MPLS_LABEL_IPV6NULL) {
struct net_device *lo = net->loopback_dev; struct net_device *lo = net->loopback_dev;
rt2 = mpls_rt_alloc(1, lo->addr_len, 0); rt2 = mpls_rt_alloc(1, lo->addr_len, 0);
if (IS_ERR(rt2)) if (IS_ERR(rt2))
goto nort2; goto nort2;
rt2->rt_nh->nh_dev = lo; rt2->rt_nh->nh_dev = lo;
netdev_hold(lo, &rt2->rt_nh->nh_dev_tracker, GFP_KERNEL);
rt2->rt_protocol = RTPROT_KERNEL; rt2->rt_protocol = RTPROT_KERNEL;
rt2->rt_payload_type = MPT_IPV6; rt2->rt_payload_type = MPT_IPV6;
rt2->rt_ttl_propagate = MPLS_TTL_PROP_DEFAULT; rt2->rt_ttl_propagate = MPLS_TTL_PROP_DEFAULT;

1
net/mpls/internal.h
View File

@ -88,6 +88,7 @@ enum mpls_payload_type {
struct mpls_nh { /* next hop label forwarding entry */ struct mpls_nh { /* next hop label forwarding entry */
struct net_device *nh_dev; struct net_device *nh_dev;
netdevice_tracker nh_dev_tracker;
/* nh_flags is accessed under RCU in the packet path; it is /* nh_flags is accessed under RCU in the packet path; it is
* modified handling netdev events with rtnl lock held * modified handling netdev events with rtnl lock held