Re: [RFC,PATCH] use rcu for fasync_lock

[Linus Torvalds]

On Sun, 21 Dec 2003, Manfred Spraul wrote:
> >
> >Just make the caller do the locking.
>
> It's not that simple:

It _is_ that simple.

The choices are:
 - let the caller do the locking
 - make the callee locking be statically determinable

Those are the choices. Your kind of code is not goign to be integrated.

> the function does
>     kmalloc();
>     spin_lock();
>     use_allocation.

This is trivially handled by splitting out the allocation as a separate 
phase. 

Yes, it requires that the caller be changed, but if the choice is between 
insane locking and making a caller change, then the choice is very very 
clear.

> But: as far as I can see, these lines usually run under lock_kernel(). 
> If this is true, then the spin_lock(&fasync_lock) won't cause any 
> scalability regression, and I'll use that lock instead of lock_sock, 
> even for network sockets.

Don't.

Here's a big clue: if you make code worse than it is today, it won't be 
accepted. I don't even see why you'd bother in the first place.

So go back to the drawing board, and just do it _right_. Or don't do it at 
all. There's no point to making the code look and behave worse than it 
does today.

		Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/