Re: local_add_return

[Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>]

* Rusty Russell (rusty@rustcorp.com.au) wrote:
> On Wednesday 17 December 2008 10:31:55 Mathieu Desnoyers wrote:
> > I think we have two different use-cases here :
> > 
> > - local_t is useful as-is for things such as a tracer, which need to
> >   modify an element of data atomically wrt local interrupts. The
> >   atomic_long_t, in this case, is the correct fallback.
> > - local_count_t could be used for fast counters.
> 
> Hi Mathieu,
> 
>    Complete agreement.
> 
>    I guess I'm biassed towards local_t = counter version, something else
> = nmi-safe version because that's what it was originally.  Looking through
> the tree, there are only 5 users: module, dmaengine and percpu_counter want
> a counter, and tracing and x86 nmi.c want nmi-safe.  There are several other
> places I know of which want local_t-the-counter.
> 
>    I'll prepare a patch which adds nmi_safe_t, and see how it looks.  There's
> no amazing hurry on this, so I won't race to hit the merge window.
> 

OK,

But can we turn what you call "nmi_safe_t" into "local_atomic_t" then ?
Because we have to specify that this type must only be used as part of
per-cpu data with preemption disabled, and we also specify that it is
atomic.

Plus, nmi_safe_t does not make much sense on architectures without NMIs,
where we sometimes disable interrupts to make the modification "atomic"
wrt all other interrupts that can happen.

Mathieu

> Thanks!
> Rusty.

-- 
Mathieu Desnoyers
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68