IPoIB: Fix deadlock on RTNL between bcast join comp and ipoib_stop()

Taking rtnl_lock in ipoib_mcast_join_complete() causes a deadlock with
ipoib_stop().  We avoid it by scheduling the piece of code that takes
the lock on ipoib_workqueue instead of executing it directly.  This
works because we only flush the ipoib_workqueue with the RTNL not held.

The deadlock happens because ipoib_stop() calls ipoib_ib_dev_down()
which calls ipoib_mcast_dev_flush(), which calls ipoib_mcast_free(),
which calls ipoib_mcast_leave(). The latter calls
ib_sa_free_multicast(), and this waits until the multicast completion
handler finishes.  This handler is ipoib_mcast_join_complete(), which
waits for the rtnl_lock(), which was already taken by ipoib_stop().

This bug was introduced in commit a77a57a1 ("IPoIB: Fix deadlock on
RTNL in ipoib_stop()").

Signed-off-by: Yossi Etigin <yosefe@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
This commit is contained in:
Yossi Etigin
2008-09-16 11:57:45 -07:00
committed by Roland Dreier
parent 1941246dd9
commit e8224e4b80
3 changed files with 24 additions and 10 deletions

View File

@@ -366,6 +366,21 @@ static int ipoib_mcast_sendonly_join(struct ipoib_mcast *mcast)
return ret;
}
void ipoib_mcast_carrier_on_task(struct work_struct *work)
{
struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv,
carrier_on_task);
/*
* Take rtnl_lock to avoid racing with ipoib_stop() and
* turning the carrier back on while a device is being
* removed.
*/
rtnl_lock();
netif_carrier_on(priv->dev);
rtnl_unlock();
}
static int ipoib_mcast_join_complete(int status,
struct ib_sa_multicast *multicast)
{
@@ -392,16 +407,12 @@ static int ipoib_mcast_join_complete(int status,
&priv->mcast_task, 0);
mutex_unlock(&mcast_mutex);
if (mcast == priv->broadcast) {
/*
* Take RTNL lock here to avoid racing with
* ipoib_stop() and turning the carrier back
* on while a device is being removed.
*/
rtnl_lock();
netif_carrier_on(dev);
rtnl_unlock();
}
/*
* Defer carrier on work to ipoib_workqueue to avoid a
* deadlock on rtnl_lock here.
*/
if (mcast == priv->broadcast)
queue_work(ipoib_workqueue, &priv->carrier_on_task);
return 0;
}