Re: [Linux-ia64] reader-writer livelock problem

[William Lee Irwin III <wli@holomorphy.com>]

In message <1036777105.13021.13.camel@ixodes.goop.org> you write:
>> Even without interrupts that would be a bug.  It isn't ever safe to
>> attempt to retake a read lock if you already hold it, because you may
>> deadlock with a pending writer.  Fair multi-reader locks aren't
>> recursive locks.

On Sat, Nov 09, 2002 at 01:48:17PM +1100, Rusty Russell wrote:
> That's the point.  This is explicitly guaranteed with the current
> locks, and you are allowed to recurse on them.  The netfilter code
> explicitly uses this to retake the net brlock, since it gets called
> from different paths.
> Implement "read_lock_yield" or "wrlock_t" but don't break existing
> semantics until 2.7 *please*!

My only interest is doing this specifically for problematic cases,
e.g. the tasklist_lock. Other usages probably shouldn't be altered
during this release cycle unless they too prove problematic, in which
case their usages should be fixed by their maintainers as bugfixes.

Of course, I'm not happy to hear about this explicit recursion.


Bill