Re: [PATCH] Introduce down_try() so we can move away from down_trylock()

[Rusty Russell <rusty@rustcorp.com.au>]

On Saturday 02 August 2008 03:26:33 Linus Torvalds wrote:
> On Tue, 29 Jul 2008, Rusty Russell wrote:
> > Introduce down_try()
>
> I hate that name. Everybody else uses "xxx_trylock()", now you introduce a
> short version of that that just has the same return value as everybody
> else except for semaphores that admittedly were odd.

spin_lock => spin_trylock, so down => trydown.  But everyone hated that, too.

I love your suggestion tho.  Oh wait, you didn't make one...

> Also, all actual _users_ of down_trylock() seem to be prime candidates for
> turning into mutexes anyway - with the _possible_ exception of the console
> semaphore which has problems with the mutex debugging code.

And Willy is working on that.  Still.  Frankly, I gave up waiting.

> > Andrew suggested introducing "down_try" as a wrapper now, to make
> > the transition easier.
>
> The transition to WHAT? To crap?
>
> There is no need to introduce yet another temporary thing just to make
> things even _more_ confusing.

And so my patch series replaces all 21 of them.  It's a trivial replace, 
unlike sem -> mutex.

> Guys, some quality control and critical thinking, please.

Good idea.  If we'd done that we wouldn't have the down_trylock() brain 
damage.

Rusty.