RE: [Algorithms] RE: simple threading control

[Jon W. <hp...@mi...>]

> One thing that you are missing is tracking which thread
> already has the lock, so that further locks on that same
> thread always succeed.  Otherwise, you'll deadlock against
> your own thread.

There's a not-insignificant contingent of system programmers who believe
that recursive locks promote laziness, and that a design is better off if it
doesn't need them. You could argue that if you are surprised by deadlocking
on yourself, you don't understand your own design; you should design to be
locking correct, and you shouldn't be doing enough with the lock held to
have to worry about recursive locking.

I'm not a convert myself, but I attend service now and then, because they're
an interesting bunch to hang out with :-)

Cheers,

			/ h+


PS: "real" spin-locks, including clearing interrupts and not page-faulting
while holding them, of course cannot be recursive on thread Id in any
system.