diff options
| author | Eric Dumazet <eric.dumazet@gmail.com> | 2010-10-29 03:09:24 +0000 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2010-11-08 13:50:08 -0800 |
| commit | fc766e4c4965915ab52a1d1fa3c7a7b3e7bc07f0 (patch) | |
| tree | d45160f52eea37d4e5149d511c3c577bef253801 /net/decnet/dn_fib.c | |
| parent | e4a7b93bd5d84e1e79917d024d17d745d190fc9a (diff) | |
| download | kernel_samsung_tuna-fc766e4c4965915ab52a1d1fa3c7a7b3e7bc07f0.zip kernel_samsung_tuna-fc766e4c4965915ab52a1d1fa3c7a7b3e7bc07f0.tar.gz kernel_samsung_tuna-fc766e4c4965915ab52a1d1fa3c7a7b3e7bc07f0.tar.bz2 | |
decnet: RCU conversion and get rid of dev_base_lock
While tracking dev_base_lock users, I found decnet used it in
dnet_select_source(), but for a wrong purpose:
Writers only hold RTNL, not dev_base_lock, so readers must use RCU if
they cannot use RTNL.
Adds an rcu_head in struct dn_ifaddr and handle proper RCU management.
Adds __rcu annotation in dn_route as well.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/decnet/dn_fib.c')
| -rw-r--r-- | net/decnet/dn_fib.c | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/net/decnet/dn_fib.c b/net/decnet/dn_fib.c index 4ab96c1..0ef0a81 100644 --- a/net/decnet/dn_fib.c +++ b/net/decnet/dn_fib.c @@ -610,10 +610,12 @@ static void dn_fib_del_ifaddr(struct dn_ifaddr *ifa) /* Scan device list */ rcu_read_lock(); for_each_netdev_rcu(&init_net, dev) { - dn_db = dev->dn_ptr; + dn_db = rcu_dereference(dev->dn_ptr); if (dn_db == NULL) continue; - for(ifa2 = dn_db->ifa_list; ifa2; ifa2 = ifa2->ifa_next) { + for (ifa2 = rcu_dereference(dn_db->ifa_list); + ifa2 != NULL; + ifa2 = rcu_dereference(ifa2->ifa_next)) { if (ifa2->ifa_local == ifa->ifa_local) { found_it = 1; break; |