Re: [Bonding-devel] [PATCH] bonding: cancel_delayed_work() -> cancel_delayed_work_sync()

[Jarek Poplawski <jarkao2@gmail.com>]

On Thu, Dec 17, 2009 at 10:58:23PM +0100, Jarek Poplawski wrote:
> On Thu, Dec 17, 2009 at 01:40:29PM -0800, Jay Vosburgh wrote:
> > 	A less evil alternative would be to punt and reschedule the work
> > if the rtnl_trylock failed, e.g.,
> > 
> > 	if (bond_miimon_inspect(bond)) {
> > 		read_unlock(&bond->lock);
> > 		if (!rtnl_trylock()) {
> > 			queue_work(...);
> > 			return;
> > 		}
> > 
> > 		read_lock(&bond->lock);
> > 
> > 		bond_miimon_commit(bond);
> > 		[...]
> > 
> > 	I'm not sure what the usual contention level on rtnl is (and,
> > therefore, how often this will punt for the normal case that's not the
> > race we're trying to avoid here).
> 
> Even if there is not much contention there is usually a lot of work
> inside, so this looks most reasonable to me.

On the other hand, there could be considered other alternatives yet,
like separating these works with rtnl_lock, and killing them reliably
from some better place (after dev close).

Jarek P.