bonding: Deadlock between bonding_store_bonds and bond_destroy_sysfs.

The sysfs layer has an internal protection, that ensures, that
all the process sitting inside ->sore/->show callback exits
before the appropriate entry is unregistered (the calltraces
are rather big, but I can provide them if required).

On the other hand, bonding takes rtnl_lock in
a) the bonding_store_bonds, i.e. in ->store callback,
b) module exit before calling the sysfs unregister routines.

Thus, the classical AB-BA deadlock may occur. To reproduce run
# while :; do modprobe bonding; rmmod bonding; done
and
# while :; do echo '+bond%d' > /sys/class/net/bonding_masters ; done
in parallel.

The fix is to move the bond_destroy_sysfs out of the rtnl_lock,
but _before_ bond_free_all to make sure no bonding devices exist
after module unload.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Acked-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

authored by Pavel Emelyanov and committed by Jeff Garzik ae68c398 c4ebc66a

+4 -2
+4 -2
drivers/net/bonding/bond_main.c
··· 4992 4992 destroy_workqueue(bond->wq); 4993 4993 } 4994 4994 4995 + bond_destroy_sysfs(); 4996 + 4995 4997 rtnl_lock(); 4996 4998 bond_free_all(); 4997 - bond_destroy_sysfs(); 4998 4999 rtnl_unlock(); 4999 5000 out: 5000 5001 return res; ··· 5007 5006 unregister_netdevice_notifier(&bond_netdev_notifier); 5008 5007 unregister_inetaddr_notifier(&bond_inetaddr_notifier); 5009 5008 5009 + bond_destroy_sysfs(); 5010 + 5010 5011 rtnl_lock(); 5011 5012 bond_free_all(); 5012 - bond_destroy_sysfs(); 5013 5013 rtnl_unlock(); 5014 5014 } 5015 5015