On Fri, 2002-11-08 at 09:34, David Howells wrote:
>
> > The normal way of solving this fairness problem is to make pending write
> > locks block read lock attempts, so that the reader count is guaranteed
> > to drop to zero as read locks are released. I haven't looked at the
> > Linux implementation of rwlocks, so I don't know how hard this is to
> > do. Or perhaps there's some other reason for not implementing it this
> > way?
>
> Actually implementing a fair spinlocks and fair rwlocks on the x86 arch are
> very easy (at least, if you have XADD it is). Any arch which has CMPXCHG can
> also do it, just not so easily.
>
> I've attached an implementation of a fair spinlock and an implementation of a
> fair rwlock (which can be compiled and played with in u-space).
>
> David
There are a selection of similar algorithms here:
http://www.cs.rochester.edu/u/scott/synchronization/pseudocode/rw.html#s_f
How does yours compare?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Nov 15 2002 - 22:00:16 EST