* [PATCH net-next] ipv6: invert join/leave anycast rtnl/socket locking order
@ 2015-03-20 13:20 Marcelo Ricardo Leitner
2015-03-20 14:24 ` Sergei Shtylyov
0 siblings, 1 reply; 4+ messages in thread
From: Marcelo Ricardo Leitner @ 2015-03-20 13:20 UTC (permalink / raw)
To: netdev; +Cc: hannes
Commit baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
missed to update two setsockopt options, IPV6_JOIN_ANYCAST and
IPV6_LEAVE_ANYCAST, causing a lock inverstion regarding the updated ones.
As ipv6_sock_ac_join and ipv6_sock_ac_leave are only called from
do_ipv6_setsockopt, we are good to just move the rtnl lock upper.
Fixes: baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
Reported-by: Ying Huang <ying.huang@intel.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
net/ipv6/anycast.c | 9 ++++-----
net/ipv6/ipv6_sockglue.c | 2 ++
2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/net/ipv6/anycast.c b/net/ipv6/anycast.c
index baf2742d1ec47ee1b9cf62d087afb8b7d7195921..a1db95e863ffdd1aa1cb5112da11ccdef9e945fd 100644
--- a/net/ipv6/anycast.c
+++ b/net/ipv6/anycast.c
@@ -60,6 +60,8 @@ int ipv6_sock_ac_join(struct sock *sk, int ifindex, const struct in6_addr *addr)
int ishost = !net->ipv6.devconf_all->forwarding;
int err = 0;
+ ASSERT_RTNL();
+
if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
return -EPERM;
if (ipv6_addr_is_multicast(addr))
@@ -73,7 +75,6 @@ int ipv6_sock_ac_join(struct sock *sk, int ifindex, const struct in6_addr *addr)
pac->acl_next = NULL;
pac->acl_addr = *addr;
- rtnl_lock();
if (ifindex == 0) {
struct rt6_info *rt;
@@ -130,7 +131,6 @@ int ipv6_sock_ac_join(struct sock *sk, int ifindex, const struct in6_addr *addr)
}
error:
- rtnl_unlock();
if (pac)
sock_kfree_s(sk, pac, sizeof(*pac));
return err;
@@ -146,7 +146,8 @@ int ipv6_sock_ac_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
struct ipv6_ac_socklist *pac, *prev_pac;
struct net *net = sock_net(sk);
- rtnl_lock();
+ ASSERT_RTNL();
+
prev_pac = NULL;
for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
if ((ifindex == 0 || pac->acl_ifindex == ifindex) &&
@@ -155,7 +156,6 @@ int ipv6_sock_ac_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
prev_pac = pac;
}
if (!pac) {
- rtnl_unlock();
return -ENOENT;
}
if (prev_pac)
@@ -166,7 +166,6 @@ int ipv6_sock_ac_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
dev = __dev_get_by_index(net, pac->acl_ifindex);
if (dev)
ipv6_dev_ac_dec(dev, &pac->acl_addr);
- rtnl_unlock();
sock_kfree_s(sk, pac, sizeof(*pac));
return 0;
diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c
index cc5883791bac1b32e4bbf9c36ee30f4d078b7dde..9b2cb144423004a02d61373338f3216b6a955043 100644
--- a/net/ipv6/ipv6_sockglue.c
+++ b/net/ipv6/ipv6_sockglue.c
@@ -122,6 +122,8 @@ static bool setsockopt_needs_rtnl(int optname)
switch (optname) {
case IPV6_ADD_MEMBERSHIP:
case IPV6_DROP_MEMBERSHIP:
+ case IPV6_JOIN_ANYCAST:
+ case IPV6_LEAVE_ANYCAST:
case MCAST_JOIN_GROUP:
case MCAST_LEAVE_GROUP:
case MCAST_JOIN_SOURCE_GROUP:
--
2.1.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH net-next] ipv6: invert join/leave anycast rtnl/socket locking order
2015-03-20 13:20 [PATCH net-next] ipv6: invert join/leave anycast rtnl/socket locking order Marcelo Ricardo Leitner
@ 2015-03-20 14:24 ` Sergei Shtylyov
2015-03-20 14:37 ` [PATCH net-next v2] " Marcelo Ricardo Leitner
0 siblings, 1 reply; 4+ messages in thread
From: Sergei Shtylyov @ 2015-03-20 14:24 UTC (permalink / raw)
To: Marcelo Ricardo Leitner, netdev; +Cc: hannes
Hello.
On 3/20/2015 4:20 PM, Marcelo Ricardo Leitner wrote:
> Commit baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
> missed to update two setsockopt options, IPV6_JOIN_ANYCAST and
> IPV6_LEAVE_ANYCAST, causing a lock inverstion regarding the updated ones.
> As ipv6_sock_ac_join and ipv6_sock_ac_leave are only called from
> do_ipv6_setsockopt, we are good to just move the rtnl lock upper.
> Fixes: baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
> Reported-by: Ying Huang <ying.huang@intel.com>
> Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> ---
> net/ipv6/anycast.c | 9 ++++-----
> net/ipv6/ipv6_sockglue.c | 2 ++
> 2 files changed, 6 insertions(+), 5 deletions(-)
> diff --git a/net/ipv6/anycast.c b/net/ipv6/anycast.c
> index baf2742d1ec47ee1b9cf62d087afb8b7d7195921..a1db95e863ffdd1aa1cb5112da11ccdef9e945fd 100644
> --- a/net/ipv6/anycast.c
> +++ b/net/ipv6/anycast.c
[...]
> @@ -155,7 +156,6 @@ int ipv6_sock_ac_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
> prev_pac = pac;
> }
> if (!pac) {
> - rtnl_unlock();
> return -ENOENT;
> }
Get rid of {} please, as they're not needed anymore.
[...]
WBR, Sergei
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH net-next v2] ipv6: invert join/leave anycast rtnl/socket locking order
2015-03-20 14:24 ` Sergei Shtylyov
@ 2015-03-20 14:37 ` Marcelo Ricardo Leitner
2015-03-20 17:33 ` David Miller
0 siblings, 1 reply; 4+ messages in thread
From: Marcelo Ricardo Leitner @ 2015-03-20 14:37 UTC (permalink / raw)
To: netdev; +Cc: hannes, sergei.shtylyov
Commit baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
missed to update two setsockopt options, IPV6_JOIN_ANYCAST and
IPV6_LEAVE_ANYCAST, causing a lock inverstion regarding to the updated ones.
As ipv6_sock_ac_join and ipv6_sock_ac_leave are only called from
do_ipv6_setsockopt, we are good to just move the rtnl lock upper.
Fixes: baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
Reported-by: Ying Huang <ying.huang@intel.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
---
Notes:
v2:
removed {} that aren't needed anymore, as pointed by Sergei
net/ipv6/anycast.c | 12 +++++-------
net/ipv6/ipv6_sockglue.c | 2 ++
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/net/ipv6/anycast.c b/net/ipv6/anycast.c
index baf2742d1ec47ee1b9cf62d087afb8b7d7195921..9e6b0ee563f090d0607f9805d3d2d8c00f787c07 100644
--- a/net/ipv6/anycast.c
+++ b/net/ipv6/anycast.c
@@ -60,6 +60,8 @@ int ipv6_sock_ac_join(struct sock *sk, int ifindex, const struct in6_addr *addr)
int ishost = !net->ipv6.devconf_all->forwarding;
int err = 0;
+ ASSERT_RTNL();
+
if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
return -EPERM;
if (ipv6_addr_is_multicast(addr))
@@ -73,7 +75,6 @@ int ipv6_sock_ac_join(struct sock *sk, int ifindex, const struct in6_addr *addr)
pac->acl_next = NULL;
pac->acl_addr = *addr;
- rtnl_lock();
if (ifindex == 0) {
struct rt6_info *rt;
@@ -130,7 +131,6 @@ int ipv6_sock_ac_join(struct sock *sk, int ifindex, const struct in6_addr *addr)
}
error:
- rtnl_unlock();
if (pac)
sock_kfree_s(sk, pac, sizeof(*pac));
return err;
@@ -146,7 +146,8 @@ int ipv6_sock_ac_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
struct ipv6_ac_socklist *pac, *prev_pac;
struct net *net = sock_net(sk);
- rtnl_lock();
+ ASSERT_RTNL();
+
prev_pac = NULL;
for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
if ((ifindex == 0 || pac->acl_ifindex == ifindex) &&
@@ -154,10 +155,8 @@ int ipv6_sock_ac_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
break;
prev_pac = pac;
}
- if (!pac) {
- rtnl_unlock();
+ if (!pac)
return -ENOENT;
- }
if (prev_pac)
prev_pac->acl_next = pac->acl_next;
else
@@ -166,7 +165,6 @@ int ipv6_sock_ac_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)
dev = __dev_get_by_index(net, pac->acl_ifindex);
if (dev)
ipv6_dev_ac_dec(dev, &pac->acl_addr);
- rtnl_unlock();
sock_kfree_s(sk, pac, sizeof(*pac));
return 0;
diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c
index cc5883791bac1b32e4bbf9c36ee30f4d078b7dde..9b2cb144423004a02d61373338f3216b6a955043 100644
--- a/net/ipv6/ipv6_sockglue.c
+++ b/net/ipv6/ipv6_sockglue.c
@@ -122,6 +122,8 @@ static bool setsockopt_needs_rtnl(int optname)
switch (optname) {
case IPV6_ADD_MEMBERSHIP:
case IPV6_DROP_MEMBERSHIP:
+ case IPV6_JOIN_ANYCAST:
+ case IPV6_LEAVE_ANYCAST:
case MCAST_JOIN_GROUP:
case MCAST_LEAVE_GROUP:
case MCAST_JOIN_SOURCE_GROUP:
--
2.1.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH net-next v2] ipv6: invert join/leave anycast rtnl/socket locking order
2015-03-20 14:37 ` [PATCH net-next v2] " Marcelo Ricardo Leitner
@ 2015-03-20 17:33 ` David Miller
0 siblings, 0 replies; 4+ messages in thread
From: David Miller @ 2015-03-20 17:33 UTC (permalink / raw)
To: marcelo.leitner; +Cc: netdev, hannes, sergei.shtylyov
From: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Date: Fri, 20 Mar 2015 11:37:17 -0300
> Commit baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
> missed to update two setsockopt options, IPV6_JOIN_ANYCAST and
> IPV6_LEAVE_ANYCAST, causing a lock inverstion regarding to the updated ones.
>
> As ipv6_sock_ac_join and ipv6_sock_ac_leave are only called from
> do_ipv6_setsockopt, we are good to just move the rtnl lock upper.
>
> Fixes: baf606d9c9b1 ("ipv4,ipv6: grab rtnl before locking the socket")
> Reported-by: Ying Huang <ying.huang@intel.com>
> Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Applied, thank you.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2015-03-20 17:33 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-03-20 13:20 [PATCH net-next] ipv6: invert join/leave anycast rtnl/socket locking order Marcelo Ricardo Leitner
2015-03-20 14:24 ` Sergei Shtylyov
2015-03-20 14:37 ` [PATCH net-next v2] " Marcelo Ricardo Leitner
2015-03-20 17:33 ` David Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).