diff options
| author | Eric Dumazet <eric.dumazet@gmail.com> | 2010-09-13 16:24:54 +0400 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2010-09-15 01:27:29 +0400 |
| commit | ef885afbf8a37689afc1d9d545e2f3e7a8276c17 (patch) | |
| tree | 8defdc814be3668f906213b5f32729341eec2aa5 | |
| parent | ab12811c89e88f2e66746790b1fe4469ccb7bdd9 (diff) | |
| download | linux-ef885afbf8a37689afc1d9d545e2f3e7a8276c17.tar.xz | |
net: use rcu_barrier() in rollback_registered_many
netdev_wait_allrefs() waits that all references to a device vanishes.
It currently uses a _very_ pessimistic 250 ms delay between each probe.
Some users reported that no more than 4 devices can be dismantled per
second, this is a pretty serious problem for some setups.
Most of the time, a refcount is about to be released by an RCU callback,
that is still in flight because rollback_registered_many() uses a
synchronize_rcu() call instead of rcu_barrier(). Problem is visible if
number of online cpus is one, because synchronize_rcu() is then a no op.
time to remove 50 ipip tunnels on a UP machine :
before patch : real 11.910s
after patch : real 1.250s
Reported-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Reported-by: Octavian Purdila <opurdila@ixiacom.com>
Reported-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
| -rw-r--r-- | net/core/dev.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/net/core/dev.c b/net/core/dev.c index b9b22a3c4c8f..660dd41aaaa6 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4845,7 +4845,7 @@ static void rollback_registered_many(struct list_head *head) dev = list_first_entry(head, struct net_device, unreg_list); call_netdevice_notifiers(NETDEV_UNREGISTER_BATCH, dev); - synchronize_net(); + rcu_barrier(); list_for_each_entry(dev, head, unreg_list) dev_put(dev); |