Re: [rfc] "fair" rw spinlocks

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Sat, Nov 28, 2009 at 09:30:18AM -0800, Linus Torvalds wrote:
> 
> 
> On Mon, 23 Nov 2009, Nick Piggin wrote:
> > 
> > Last time this issue came up that I could see, I don't think
> > there were objections to making rwlocks fair, the main
> > difficulty seemed to be that we allow reentrant read locks
> > (so a write lock waiting must not block arbitrary read lockers).
> 
> We have at least one major rwlock user - tasklist_lock or whatever. And 
> that one definitely depends on being able to do 'rwlock()' in an 
> interrupt, without other rwlock'ers having to disable irq's (even if there 
> might be a new writer coming in on another cpu).

In other words, any fair rwlock must unconditionally grant read access
to any CPU that already read-holds the lock, regardless of the state of
any writer wannabees.  Or is there another requirement I am missing?

> That usage case _might_ be turned into RCU or something similar, in which 
> case I don't think any major rwlock users remain. However, if that's the 
> case, then why should anybody care about fairness any more either?

Indeed, RCU goes a step further, permitting new readers unconditionally,
regardless of writer state.  ;-)

> So as far as I can tell, we have only one real user of rwlocks where 
> livelocks might be relevant, but that one real user absolutely _requires_ 
> the unfair behavior.

But the required unfairness is limited to unconditionally granting
recursive read requests, right?  If I understand correctly, if a given
CPU does not already read-hold the lock, then we can safely make that
CPU wait for a writer that might otherwise be starved.  Again, is there
another requirement that I am missing?

							Thanx, Paul